Mapping the Sequences That Mediate Interaction of the Equine Herpesvirus 1 Immediate-Early Protein and Human TFIIB

Abstract

The sole immediate-early (IE) gene of equine herpesvirus 1 encodes a 1,487-amino-acid (aa) regulatory phosphoprotein that independently activates expression of early viral genes. Coimmunoprecipitation assays demonstrated that the IE protein physically interacts with the general transcription factor TFIIB. Using a variety of protein-binding assays that employed a panel of IE truncation and deletion mutants expressed as in vitro-synthesized or glutathione S-transferase fusion proteins, we mapped a TFIIB-binding domain to aa 407 to 757 of the IE protein. IE mutants carrying internal deletions of aa 426 to 578 and 621 to 757 were partially defective for TFIIB binding, indicating that aa 407 to 757 may harbor more than one TFIIB-binding domain. The interaction between the IE protein and TFIIB is of physiological importance, as evidenced by transient-cotransfection assays. Partial deletion of the TFIIB-binding domain within the IE protein inhibited its ability to activate expression of the viral thymidine kinase gene, a representative early promoter, and of the IR5 gene, a representative late promoter, by greater than 20 and 50%, respectively. These results indicate that the interaction of the IE protein with TFIIB is necessary for its full transactivation function and that the IE-TFIIB interaction may be part of the mechanism by which the IE protein activates transcription.

Among the eight equid herpesviruses identified to date (52), equine herpesvirus 1 (EHV-1) is one of the most pathogenic herpesviruses of horses, causing spontaneous abortions in pregnant mares, as well as respiratory tract infections and neurological disorders (1, 12, 45). The virus is a member of the subfamily Alphaherpesvirinae and serves as a model for the investigation of alphaherpesvirus gene regulation during both productive and persistent infections. The 77 EHV-1 genes are temporally and coordinately expressed at immediate-early (IE), early, and late (γ1 and γ2) times of the lytic infection cycle (8, 18), analogous to that of herpes simplex virus type 1 (HSV-1) (11, 33). In contrast to HSV-1, EHV-1 carries only one IE gene (also termed IR1 gene) that is expressed without prior viral protein synthesis due to the EHV-1 α-trans-inducing factor (ETIF), a homolog of the HSV-1 VP16 protein (14, 41, 47). The EHV-1 IE gene (i) is located within each inverted-repeat region and encodes a polypeptide of 1,487 amino acids (aa) with a predicted molecular mass of approximately 155 kDa (19, 21, 27), (ii) has a product with a high degree of homology with HSV-1 ICP4 and the varicella-zoster virus ORF62 gene products (21), and (iii) is transcribed as a 6.0-kb spliced mRNA (19, 27, 51) that gives rise to both structurally and antigenically related protein species ranging from 125 to 200 kDa (7, 8, 51). In transient-cotransfection assays, the IE protein is a bifunctional regulatory protein capable of (i) negatively autoregulating its own promoter (55), (ii) independently activating EHV-1 early and heterologous viral promoters (55, 56), (iii) cooperating synergistically with two early auxiliary regulatory proteins (EICP22 and EICP27) to activate EHV-1 early and γ1 late promoters (32, 44, 55, 57, 64), and (iv) acting antagonistically with a third early major regulatory protein, EICP0, to selectively repress expression of certain promoters from all classes of EHV-1 promoters, including γ2 late promoters (3, 35).

Sequence alignment of the EHV-1 IE protein and other homologs in the subfamily Alphaherpesvirinae defined five colinear regions that harbor specific functional domains. Region 1 contains an acidic transactivation domain (TAD; aa 3 to 89) (58) and a serine-rich tract (SRT; aa 181 to 220). Regions 2 and 3 harbor a helix-loop-helix motif that mediates a sequence-specific DNA-binding activity (aa 422 to 597) (38), while the nuclear localization signal (aa 963 to 970) lies within region 3 (56). Region 5 contains a transcriptional-enhancement domain that is required for the full transactivation activity of the IE protein (5, 56). Most of these functional domains are essential for EHV-1 lytic growth, since mutant viruses with deletions within the SRT, nuclear localization signal, or DNA-binding domain are capable of growth only on IE protein-producing cells (15; K. A. Buczynski and D. J. O'Callaghan, unpublished data). In addition, viruses expressing an IE protein with a truncated carboxyl terminus are impaired for growth in cell culture and express reduced levels of viral early and late genes (Buczynski and O'Callaghan, unpublished). A novel feature of the IE gene is that an early gene, IR2, maps within the IE gene and is expressed as a 4.4-kb mRNA that is 3′ coterminal with the 6.0-kb IE mRNA. The IR2 gene encodes an N-terminally truncated form (aa 323 to 1487) of the IE protein lacking two functional domains, TAD and SRT (29). The IR2 protein is not capable of transactivating any viral genes tested to date, indicating that the TAD and SRT are essential for the transactivating functions of the IE protein (5, 36, 58). Analysis of the DNA-binding activity of the IE protein revealed that it recognizes the consensus sequence 5′-ATCGT-3′ (38). The IE promoter contains the consensus target sequence near its transcription start site. It is speculated that the IE and IR2 proteins, which harbor the DNA-binding domain, repress transcription of the IE promoter by binding to this sequence and blocking access to the transcription start site (28, 38).

The mechanism by which the IE protein transactivates target viral genes has not yet been fully elucidated. However, transcription of viral genes in a productive infection is mediated by the interaction between viral activator proteins and various components of the cellular transcriptional machinery (2, 6, 9, 16, 20, 30, 34, 43, 49, 54, 60, 61, 63). These interactions are important, at least in part, for facilitating transcription by increasing assembly of a preinitiation complex (PIC) (9, 10, 20, 30, 40, 42, 54). The PIC contains multiple components of the cellular transcriptional machinery, including RNA polymerase II (Pol II) and general transcription factors (GTFs), and can be formed on Pol II promoters in a sequential order from the individual assembly of the GTFs, TFIIA, TFIIB, TFIID, TFIIE, TFIIF, TFIIH, TFIIJ, and Pol II (4, 26). Among the GTFs, TFIID and TFIIB have been studied extensively and function in an early stage of PIC assembly by acting as a scaffold for the assembly of the remaining transcriptional machinery. The recognition of promoters is most frequently mediated by TFIID through the binding of the TATA binding protein (TBP) subunit to TATA box elements and/or recognition of non-TATA box cis elements by TBP-associated factors (TAFs) (4, 25, 26). TFIIB plays a pivotal role in PIC formation, providing a bridge between promoter-bound TBP and Pol II-TFIIF. This observation is supported by the interaction between TFIIB and GTFs, including TBP, TFIIF, and Pol II (24). The complex of TBP-TFIIB-Pol II represents the minimal requirement for PIC formation but is not sufficient for activation of gene expression. A recent study of PIC formation and transcriptional activation demonstrated that PIC assembly occurs by at least two stages and that the TATA box and TFIIB can also affect transcription subsequent to PIC assembly (48). Thus, processes other than factor recruitment are potentially influenced by transactivators. In many cases, TFIID and TFIIB have been implicated as direct targets for viral transactivators. HSV-1 VP16 facilitates PIC formation by enhancing TFIIB binding to the complex, as well as TFIIA and TAF_II40, which in turn increase recruitment of Pol II and stabilize its association with TFIID (10, 17, 39, 40, 42, 60). HSV-1 ICP4 is part of a tripartite complex involving TFIIB and TFIID (54) and facilitates PIC formation by enhancing the binding of TFIID to the TATA box element (20). Further, the interaction of the carboxyl-terminal region of HSV-1 ICP4 with TFIID is mediated through an interaction with TAF_II250, which appears to be critical for the ICP4 transactivation function (6). In addition, the large T antigen of simian virus 40 (34), the EIA protein of adenovirus (16), the E2 protein of papillomaviruses (2, 49, 63), and the EBNA2 and ZEBRA proteins of Epstein-Barr virus (9, 43) are other noteworthy examples of viral transactivators that directly interact with TFIID and/or TFIIB. These interactions may also stimulate mRNA elongation by the Pol II-associated complex (62).

In this report, we present our findings that the EHV-1 IE protein can interact and functionally cooperate with one of the GTFs, TFIIB, to stimulate transcription synergistically. Using a variety of protein-binding assays, we have assessed the potential interaction of the IE protein with TFIIB. In addition, we have mapped regions of the IE protein responsible for the IE-TFIIB interaction. Finally, we have used transient-cotransfection assays to assess the in vivo importance of this interaction. Our results suggest that the IE protein residues 407 to 757 bind to TFIIB and that this interaction may contribute to transcriptional activation of EHV-1 genes.

MATERIALS AND METHODS

Cell culture, virus infection, and preparation of nuclear extracts from infected cells.

Cultures of murine fibroblast L-M cells were grown as monolayers in Eagle's minimum essential medium (EMEM) supplemented with penicillin (100 μg/ml), streptomycin (100 μg/ml), nonessential amino acids, and 5% fetal bovine serum (FBS) (8, 55). Nuclear extracts of infected cells were prepared as described previously, with some modifications (46). L-M cells (2.3 × 10⁷) were infected with the wild-type EHV-1 Kentucky A (KyA) strain at a multiplicity of infection of 15 to 20 PFU per cell. At 6 h postinfection, the cells were scraped into phosphate-buffered saline containing 0.1 mM (each) Nα-p-tosyl-l-lysine chloromethyl ketone (TLCK) and N-tosyl-l-phenylalanine chloromethyl ketone (TPCK), pelleted, and resuspended in 4 volumes of buffer A (10 mM HEPES [pH 7.9], 1.5 mM MgCl₂, 10 mM KCl, 0.5% NP-40, 0.5 mM dithiothreitol, 0.1 mM TLCK, and 0.1 mM TPCK). After incubation for 10 min on ice, the nuclei were pelleted at 14,000 rpm for 5 min in a microcentrifuge. The supernatant was discarded, and proteins were eluted from the nuclei by incubation for 30 min on ice in 2 volumes of buffer B (10 mM HEPES [pH 7.9] 25% glycerol, 0.42 M NaCl, 1.5 mM MgCl₂, 10 mM KCl, 0.2 mM EDTA, 0.5 mM phenylmethylsulfonyl fluoride, 0.5 mM dithiothreitol, 0.1 mM TLCK, and 0.1 mM TPCK). The nuclear debris was pelleted by centrifugation at 14,000 rpm for 15 min in an Eppendorf microcentrifuge, and the supernatants containing the nuclear IE proteins were stored at −70°C.

Plasmid construction.

All recombinant DNA methods were performed according to standard protocols (53). Two expression plasmids (pN254 and pM270) carrying the entire human TFIIB gene were kindly provided by D. Reinberg and M. Hampsey (23). The pGST-IE (1–1487) plasmid producing a full-length glutathione S-transferase (GST)-IE fusion protein, and the pGEM44 plasmid expressing the IR2 protein were previously generated as described elsewhere (29, 38). The generation of the effector constructs pSVEICP27 (pSVUL3) and pcDR4 (pEICP22) and the reporter constructs pTK-CAT and pIR5-CAT has been described previously (31, 55, 64).

(i) Human TFIIB expression plasmids.

The full-length TFIIB gene from pM270 was amplified by PCR, using the 5′ primer hIIB#F1 and the 3′ primer hIIB#R1, which contained SmaI and BamHI restriction sites, respectively, and was cloned in-frame into the same sites of pGBKT7 yeast two-hybrid vector (Clontech) to obtain pGBKhIIB. pGBKhIIB encodes a chimeric protein containing an amino-terminal GAL4 DNA-binding domain and a carboxyl-terminal TFIIB protein and is being used for other investigations. For in vitro protein-binding experiments, the SmaI-SalI fragment containing the TFIIB insert from the pGBKhIIB was subcloned into the pGEX-KG vector (22), and the resultant plasmid, pGSTKG-hIIB (1–316), was used to produce the GST-hIIB fusion protein. Plasmid pG3hIIB (1–316) was generated by inserting the TFIIB-containing SmaI-BamHI fragment of pGSTKG-hIIB (1–316) into the HincII-BamHI sites of pGEM-3Z (Promega) and was employed to generate in vitro-transcribed and -translated TFIIB. To obtain the mammalian TFIIB expression plasmid, pTriExhIIB (1–316), the TFIIB gene from pM270 was amplified by PCR with the primers hIIB#F2 and hIIB#R2, which contained NcoI and HindIII sites, respectively, and cloned into the same sites of the pTriEx-1 vector (Novagen). This vector contains HSV tag and His tag sequences at the distal end of the multiple cloning site to enable the construction of carboxyl-terminally tagged fusion proteins, which will facilitate detection and purification of the fusion protein in multiple expression systems.

(ii) IE mutant plasmids for in vitro transcription and translation (IVTT).

Plasmid pG3IE (1–1487) was constructed by cloning an NheI-DraI fragment containing the entire IE gene (21) into the XbaI and SmaI sites of the pGEM-3Z vector (Promega). The generation of the carboxyl-terminal truncation mutant plasmids pG3IE (1–1254), pG3IE (1–904), pG3IE (1–620), pG3IE (1–424), and pG3IE (1–289) was performed by a modification of the method described elsewhere (56). Briefly, pG3IE (1–1487) was digested with restriction enzymes (NruI, EcoNI, RsrII, BamHI, or StuI), and the linear plasmids were purified with the GeneClean II kit (ISC Bioexpress) following agarose gel electrophoresis. The 5′ overhangs were filled in with Klenow enzyme, and the resulting blunt-ended DNA was ligated to the double-stranded 14-bp SpeI linker oligonucleotide, which contains a nonsense codon in all three reading frames (New England Biolabs). Positive clones were cleaved with SpeI, gel purified, and religated to yield a single inserted copy of the linker. The PCR primers IE#F1 and IE#R1 each contain an NcoI site and were used to amplify a 350-bp fragment (corresponding to aa 422 to 539) to generate pG3IEΔ2–421 by inserting the NcoI-digested PCR product into the 5.9-kbp NcoI fragment of pG3IE (1–1487). In addition, pG3IEΔ2–539 was generated by self-ligation of the 5.9-kbp NcoI fragment of pG3IE (1–1487). pG3IE (1–1487) was digested with NcoI and EcoNI to release a 4.79-kbp fragment and was utilized to construct the additional amino-terminal-deletion mutants pG3IEΔ2–654, pG3IEΔ2–719, and pG3IEΔ2–826. The fragments (750, 550, or 230 bp; corresponding to aa 655 to 904, 720 to 904, or 827 to 904) were created by PCR amplification with a set of NcoI-containing 5′ primers (IE#F2, IE#F3, and IE#F4) and the 3′ primer IE#R2 containing an EcoNI site, were digested with the same enzymes, and were ligated to the 4.79-kbp NcoI-EcoNI fragment of pG3IE (1–1487). pG3IEΔ2–757 was constructed by a two-step cloning procedure. First, an 8-bp NcoI linker (New England Biolabs) was inserted into the blunt-ended RsrII site of the 7.09-kbp fragment derived from RsrII-digested pG3IE (1–1487) to obtain the pG3IENΔ621–757 intermediate. Second, pG3IENΔ621–757 was cut with NcoI to release a 5.23-kbp fragment that was self-ligated to generate pG3IEΔ2–757. The internal mutant pG3IEΔ290–757 was constructed by digesting pG3IE (1–1487) with RsrII and StuI, filling in the 5′ overhangs with Klenow enzyme, and self-ligation of the plasmid DNA. To generate plasmid pG3IEΔ407–757, the 1.22-kbp fragment (corresponding to aa 1 to 406) was amplified by PCR using primers IE#F5 and IE#R3, which created NcoI and RsrII sites on the 5′ and 3′ termini of the fragment, respectively. The PCR product was cleaved with NcoI and RsrII, and the resulting fragment was cloned into the 5.23-kbp NcoI-RsrII fragment of pG3IE (1–1487). Plasmid pG3IEΔ426–578 was constructed by cloning the PCR-amplified 0.98-kbp fragment (corresponding to aa 579 to 904) with primers IE#F6 and IE#R2, containing BamHI and EcoNI sites, respectively, into the 6.06-kbp BamHI-EcoNI fragment of pG3IE (1–1487). pG3IEΔ621–757 was constructed by inserting an 8-bp ClaI linker (New England Biolabs) in frame into the blunt-ended RsrII site of pG3IE (1–1487).

(iii) GST-IE fusion plasmids.

To generate plasmids pGST-IE (1–960), pGST-IE (1–289), and pGST-IE (1–180), pGST-IE (1–1487) was digested with restriction enzymes (SalI, StuI, or BspEI), the 5′ overhangs were filled in with Klenow enzyme, and it was ligated to the SpeI linker (amber codons) as described above. To generate pGST-IE (1–424), the 1.61-kbp NcoI fragment of pGST-IE (1–1487) was swapped for the corresponding domain of pG3IE (1–424). pGST-IE (179–424) was constructed by cloning the PCR-amplified fragment (using primers IE#F7 and IE#R4, containing NcoI and XhoI sites, respectively) into the 5.41-kbp NcoI-XhoI fragment of pGST-IE (1–1487). Plasmids pGST-IE (407–539), pGST-IE (407–757), and pGST-IE (539–910) were constructed by PCR amplification of the insert fragments using primers IE#F8 and IE#R5, IE#F8 and IE#R6, and IE#F9 and IE#R7, each pair of which contained NcoI and EcoRI sites, respectively, and ligating the fragments to the same sites of pGEX-2TN (38). pGST-IE (758–1487) was constructed by digestion of pGST-IE (1–1487) with BamHI and RsrII, filling in the 5′ overhangs with Klenow enzyme, and self-ligation of the resulting plasmid. pGST-IR2 (323–1487) was constructed by cloning a StuI-EcoRI fragment from pGEM44 (29) into the SmaI-EcoRI sites of pGEX-KG (22).

(iv) Mammalian IE expression plasmids.

pG3IE (1–1487) was digested with NcoI and BamHI to release the 5′ 1.27-kbp portion of the IE gene, which was ligated to an NcoI-BamHI-digested pTriEx-1 (Novagen) to generate the pTriExIEΔBK intermediate. The pG3IE (1–1487) plasmid was then cut with BamHI and KpnI to release the 3′ 3.49-kbp portion of the IE gene, which was ligated to the BamHI-KpnI-digested pTriExIEΔBK to generate pTriExIENF (1–1487). To create a fusion between the full-length IE sequence and the carboxyl-terminally tagged sequences (HSV tag and His tag), the carboxyl-terminal NruI-HindIII portion of pTriExIENF (1–1487) was swapped for the 700-bp fragment (corresponding to aa 1255 to 1487) amplified by PCR using primers IE#F10 and IE#R8, which contained NruI and HindIII sites on the 5′ and 3′ termini of the fragment, respectively. The resultant plasmid, pTriExIE (1–1487), encodes a tagged protein in the mammalian L-M cells and was used in this study. pTriExIEΔ407–757, pTriExIEΔ426–578, and pTriExIEΔ621–757 were constructed by domain swapping of the 2.71-kbp NcoI-EcoNI or the 1.44-kbp BamHI-EcoNI fragment of pTriExIE (1–1487) for the corresponding domain (1.66, 2.25, or 1.3 kb) of pG3IEΔ407–757, pTriExIEΔ426–578, or pTriExIEΔ621–757, respectively.

Purification of GST fusion proteins.

Expression and purification of GST fusion proteins were carried out by a modification of the purification procedures described elsewhere (13, 37, 38). The pGEX expression vectors encoding TFIIB or each IE derivative were transformed into the Escherichia coli BL21(DE3) pLysE strain. The transformed bacteria were grown overnight at 37°C in 2× YT (yeast extract and tryptone) medium supplemented with 2% glucose and the appropriate antibiotics (100 μg of ampicillin/ml and 34 μg of chloramphenicol/ml). The cultures were diluted 1:100 in 250 ml of fresh prewarmed 2× YT medium containing the appropriate antibiotics and grown for 1.5 to 2 h at 37°C. Fusion protein synthesis was then induced by incubating the cells with 0.5 mM isopropyl-β-d-thiogalactoside (IPTG) for 2 to 3 h at 37°C. The cells were lysed, and the proteins were purified with the BugBuster GST-Bind purification kit (Novagen) according to the manufacturer's instructions with slight modifications. The cells were lysed in 12 ml of protein extraction reagent containing 25 U of Benzonase nuclease/ml for 20 min with shaking at room temperature (RT). Insoluble debris was removed by centrifugation at 12,000 rpm for 20 min in the Beckman JA-20 rotor, and the GST proteins were purified from the soluble extract by batch binding the supernatant with GST-Bind resin (1 ml of resin per 5 to 8 mg of protein) at RT for 40 min. The beads were then washed twice with 5 ml of GST-Bind and wash buffer. The bound proteins were eluted from the resin in 1.5 to 2 ml of GST elution buffer. The eluates were then loaded into Centricon columns (Amicon) as directed by the manufacturer to both desalt and concentrate the purified proteins by ultrafiltration. Protein purity was determined by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE), and concentrations were estimated by densitometric analysis with the Gel Doc 1000/2000 gel documentation system (Bio-Rad) by comparing protein intensity to known amounts of bovine serum albumin (Pierce). Aliquots of proteins were stored at −70°C.

In vitro transcription and translation.

All of the in vitro expression plasmids used in this study were transcribed with SP6 RNA polymerase and translated in rabbit reticulocyte lysate (Promega) in the presence of [³⁵S]methionine (40 μCi/ml; specific activity, 1,175 Ci/mmol; New England Nuclear Corp.) as recommended by the manufacturer. Experiments involving the in vitro-synthesized proteins were performed in parallel in the presence or absence (competition assays) of ³⁵S-labeled proteins. Radioactive products were analyzed by SDS-PAGE followed by autoradiography and either stored at −70°C or used immediately for in vitro protein-binding assays.

Protein-binding and competition assays.

Protein-protein interactions of the EHV-1 IE proteins (in vitro synthesized, bacterially expressed, or nuclear extracted) with TFIIB were carried out by a modification of previously described procedures (13). Aliquots of proteins were thawed on ice, and 2 μg of the appropriate GST fusion protein was incubated with the ³⁵S-labeled proteins (see Fig. 3 to 5, lanes INPUT) in a final volume of 600 μl of NETN buffer (100 mM NaCl, 1 mM EDTA, 20 mM Tris-HCl [pH 8.0], 0.5% NP-40). After incubation for 90 min at RT with gentle rocking, 30 μl of a 50% slurry of glutathione-Sepharose beads (Pharmacia) was added and the proteins were incubated an additional hour at RT. The beads were then centrifuged and washed five times with 600 μl of NETN buffer. The bound proteins were eluted by boiling them for 5 min in 20 μl of 2× SDS sample buffer (120 mM Tris-HCl [pH 6.8], 4% SDS, 20% glycerol, 0.001% bromphenol blue, 2% 2-mercaptoethanol), and analyzed by SDS-PAGE. The gels were dried, and the bands were quantitated by PhosphorImager analysis (Molecular Dynamics). The competition assays were performed essentially as before except that unlabeled in vitro-synthesized proteins were employed to compete for the binding of the ³⁵S-labeled IE protein. To assess interactions with native IE proteins, 20 μl of nuclear extracts from mock-infected or EHV-1-infected cells were allowed to bind with 2 μg of GST or GST-hIIB protein in a final volume of 600 μl of NETN buffer as described above. The bound proteins were resolved by SDS-PAGE, and the gel was subjected to Western blot analysis.

FIG. 3.

Mapping the TFIIB-binding domain of the IE protein. (A) Schematic representation of the functional regions of the wild type, truncation mutants, and deletion mutants of the IE protein and the quantitative results of their interactions with TFIIB as presented in panel B. The IE protein regions (R1 to R5), along with the corresponding functional domains that contribute to its various activities, are indicated by the solid rectangles. The IE truncation and deletion mutants used in the protein-binding assays (panel B) are schematically shown with respect to their positions within the full-length IE protein. The amount of each protein binding to TFIIB is given as the percentage of binding of each input protein, which was set at 100%, as determined by PhosphorImager analysis. (B) Results from protein-binding assays in which various forms of [³⁵S]methionine-labeled IE proteins were incubated with 2 μg of either GST or GST-hIIB. After extensive washing, the bound proteins were eluted, resolved by SDS-PAGE, and visualized by autoradiography. The masses (in kilodaltons) of ¹⁴C-methylated protein markers (Pharmacia) are indicated on the left of each gel. The first lane for each IVTT protein shows the amount of input for each protein. +, present; −, absent.

FIG. 5.

IE amino acids 407 to 757 contain the putative TFIIB-binding domain(s). (A) Various constructs expressing selected domains of the IE protein as GST fusion proteins are schematically represented by solid rectangles, with their amino acid sequences indicated by the scale above the proteins. (B) Each purified GST-IE fusion protein containing the wild-type IE (lane 2), IR2 proteins (lane 12), or other IE derivatives (lanes 3 to 11) used in the protein-binding experiments (panel C) were separated by SDS-PAGE. The gel was stained with Coomassie blue to assess whether each GST fusion protein migrated at the predicted mass and to confirm the stability of the GST fusion proteins. Molecular mass markers are indicated on the left (lane 1). (C) Results of protein-binding assays. The efficiencies with which the different GST-IE fusion proteins bound to in vitro-synthesized radiolabeled human TFIIB ([³⁵S]hIIB) are indicated under the lanes. The additional bands detected in the [³⁵S]hIIB precipitates most likely represent degradation of the in vitro-synthesized product or minor protein species originating from internal translation start sites.

DNA transfection and coimmunoprecipitation.

For coimmunoprecipitation, the pTriEx constructs were transfected as described previously (32, 55) and expressed in L-M cells under the control of the chicken β-actin promoter. Briefly, L-M cells were plated at a density of 4 × 10⁶ cells per 60-mm-diameter dish and grown overnight in EMEM supplemented with 5% FBS. The cells were washed with serum-free EMEM, and liposome-mediated DNA transfection was performed with 22 μl of Lipofectin reagent (Gibco BRL) containing the appropriate pTriEx constructs (10 μg each of the pTriExIE derivative and/or pTriExhIIB DNA). After a 5-h incubation, the transfectant was removed, fresh EMEM containing 5% FBS was added, and the cells were incubated for an additional 48 to 50 h. The cells were then washed extensively with phosphate-buffered saline, and protein extracts were prepared by lysis with RIPA buffer (150 mM NaCl, 50 mM Tris-HCl [pH 8.0], 0.1% SDS, 1% NP-40, 0.5% sodium deoxycholate, 0.5 mM phenylmethylsulfonyl fluoride, 0.1 mM TLCK, 0.1 mM TPCK). To reduce nonspecific binding of the extracts to protein A-Sepharose, normal rabbit serum plus protein A-Sepharose (Sigma) was added to the extracts and the mixtures were incubated for 1 h at 4°C. The coimmunoprecipitations were performed with the precleared extracts as described previously (50). The extracts from the transfected L-M cells were incubated overnight at 4°C with anti-TFIIB antibody (Santa Cruz Biotechnology) in a total volume of 300 μl of RIPA buffer. The protein-antibody complexes were precipitated with 30 μl of a 50% mixture of protein A-Sepharose beads for 2 h at 4°C. After being washed to eliminate nonspecific precipitates, the bound proteins were analyzed by SDS-PAGE and subsequent Western blot (immunoblot) analysis. To prepare total extracts containing native IE proteins, L-M cells (2.3 × 10⁷) were infected with EHV-1 KyA at a multiplicity of infection of 1.0 PFU per cell as described above. At 48 to 50 h postinfection, the cells were harvested and lysed in RIPA buffer, and the IE protein in the lysates was used as a molecular mass marker for the IE proteins.

Western blot analyses and antibodies.

Proteins were separated in mass–8% SDS-PAGE gels and electrotransferred to a nitrocellulose membrane (Bio-Rad) for 1 h. After transfer, the membrane was blocked for 1 h at RT in TBST buffer (10 mM Tris-HCl [pH 8.0], 150 mM NaCl, 0.05% Tween 20) containing 5% nonfat powdered milk. The membrane was then incubated with anti-IE peptide-specific antiserum at a dilution of 1:1,000 or anti-TFIIB antibody (Santa Cruz Biotechnology) at a dilution of 1:1,500 in TBST for 30 min at RT. After three 10-min washes with TBST, the membrane was incubated with alkaline phosphatase-conjugated goat anti-rabbit antibody (Sigma), diluted in TBST at a dilution of 1:5,000, for 30 min at RT followed with three TBST washes to remove unbound antibody. Immunocomplexes were visualized by incubation in AP buffer (100 mM Tris-HCl [pH 9.5], 100 mM NaCl, 5 mM MgCl₂) containing the AP substrates and 0.165 mg of 5-bromo-4-chloro-3-indolylphosphate p-toluidine salt/ml and 0.3 mg of nitroblue tetrazolium chloride/ml (Gibco BRL). The anti-IE peptide-specific antiserum was raised against a peptide derived from aa 925 to 943 of the IE protein and has been demonstrated in previous studies to be highly reactive to the IE protein (15, 58). The protein expressions produced by the pTriEx constructs were confirmed prior to use in the coimmunoprecipitations by Western blot analyses with anti-HSV tag monoclonal antibody (Novagen) and goat anti-mouse antibody conjugated to alkaline phosphatase (Novagen) at a dilution of 1:5,000 each.

CAT assays.

L-M cells (3 × 10⁶) were plated in tissue culture dishes (60-mm diameter) at approximately 75% confluence and transfected with plasmid DNAs by the Lipofectin-mediated DNA transfection method (Gibco BRL) (32, 55). The cells were transfected with either 1.4 pmol of the EHV-1 early thymidine kinase (TK) promoter-chloramphenicol acetyltransferase (CAT) reporter (pTK-CAT) or with 2.0 pmol of the late IR5 promoter-CAT reporter (pIR5-CAT). The effector constructs, pTriExIE (1–1487), pTriExIE (Δ407–757), pTriExIE (Δ621–757), pSVEICP27, and pcDR4 (EICP22), were transfected in amounts of 0.3 pmol. In all transfections, the total amount of DNA was adjusted to 10 μg by the addition of pUC19 DNA. The cells were harvested at 48 to 50 h posttransfection, and CAT assays were performed as described previously (32, 55). Each CAT assay was independently repeated at least three times, and within individual experiments, each sample was assayed in triplicate. Data were analyzed for statistical significance by the Student t test.

Oligonucleotide primers.

The sequences of oligonucleotides (restriction sites are underlined) are as follows: hIIB#F1, TCCCCCGGGGATGGCGTCTACCAGCCGTTT; hIIB#F2, CATGCCATGGCGTCTACCAGCCGTTTGGAT; hIIB#R1, CGGGATCCTTATAGCTGTCCTAGTTTGTCC; hIIB#R2, CCCAAG CTTTAGCTGTGGTAGTTTGTCCAC; IE#F1, CATGCC ATGGGGTCGGATCCACCACCGATG; IE#F2, CATGCCATGGCCTTCAGACTCAGGGAGCTC; IE#F3, CATGCCATGGAACTCAGGGACCTGGTCGAG; IE#F4, CATGCCATG GAGGGG AGCCTTCAGACCCTG; IE#F5, CACGCCATG GCCAGCCAGCGCAGCGACTTC; IE#F6, CGGGATCCAGACATGGCCTACCCGAGAGAC; IE#F7, CATGCCATGGTCCGGAGCATCTCCATCTCA; IE#F8, CATGCCATGGCCTACCAGCGCGAGCCGCTTCTC; IE#F9, CTGCCCCACATCGGGGACGCCATGGCGGCC; IE#F10, GGACTGTTCGCGAGGGCCGAGGCCGCGTTC; IE#F11, CATGCCATGGCTTCTCCGCCGGGCCGGAGC; IE#R1, GGCCGCCATGGCGTCGGGGATGTGGGGCAG; IE#R2, GGAGCCGGGCATCGCCTCCTGGCTGAGGTG; IE#R3, CGCCACGGTCCGGACGCCCGGGGCGCACGA; IE#R4, CCGCTCGAGTATCGCACCCTCCCCATGGGT; IE#R5, CCGGAATTCCATGGCGTCCCCGATGTGGGGCAG; IE#R6, CCGGAATTCCCGGGCCACCACCAGAACGGCGCG; IE#R7, CCGGAATTCGGGCATCGCCTCCTGGCTGAGGTG; IE#R8, CCCAAGCTTGGAGAGGTACGGATTGCACCA; IE#R9, TCGCACCCTCCCCATCGGTGGTGGATC CGA.

RESULTS

Native EHV-1 IE protein interacts with TFIIB.

Initially, we sought to determine whether the native EHV-1 IE protein interacts with TFIIB in vitro. Human TFIIB was purified from bacteria as a GST fusion protein (GST-hIIB) with GST-Bind resin as described in Materials and Methods and was incubated with nuclear extracts derived from either mock-infected or KyA-infected cells. The level of nonspecific binding by GST-hIIB was determined with GST alone as a negative control. The bound proteins were then analyzed by immunoblotting them with the anti-IE peptide-specific antibody. The results (Fig. 1) showed that nuclear extracts derived from mock-infected cells bound to neither the control GST alone nor GST-hIIB (lanes 4 and 6), whereas nuclear extracts derived from KyA-infected cells bound to GST-hIIB (lane 7) but not to GST alone (lane 5). Several species of antigenically cross-reactive IE proteins characterized in previous studies (7, 8, 51) were detected in the virus-infected nuclear extract precipitate by the anti-IE antibody (lane 3). The two largest IE protein species migrated on the SDS-PAGE gel with apparent molecular masses of 175 to 200 kDa as expected, and significant amounts of these IE species bound to GST-hIIB (approximately 70% bound compared to the input proteins). These results suggested that the EHV-1 IE proteins interact with TFIIB.

FIG. 1.

The native EHV-1 IE protein binds directly to TFIIB. Nuclear extracts derived from mock-infected or EHV-1 KyA-infected cells were prepared at 6 h postinfection and preincubated with either GST (lanes 4 and 5) or GST-hIIB (lanes 6 and 7) prior to precipitation with glutathione-Sepharose beads. Lanes 2 and 3 are the input proteins used in the binding reactions. The eluted proteins were analyzed by SDS-PAGE and immunoblotting with an anti-IE peptide-specific antibody. The arrowheads point to IE protein species bound to GST-hIIB. The molecular masses of the protein markers (lane 1) are given.

IE protein coimmunoprecipitates with TFIIB.

To show further that the IE protein interacts with TFIIB, we performed coimmunoprecipitation assays. Cells were transfected singly or cotransfected with pTriExIE (1–1487) and/or pTriExhIIB (1–316) expressing either the wild-type EHV-1 IE protein or TFIIB, respectively, as carboxyl-terminally HSV-tagged proteins. The presence of the proteins in the resulting cell extracts was first confirmed by immunoblotting them with anti-HSV tag monoclonal antibody prior to use in the coimmunoprecipitation assays (Fig. 2A). No HSV-tagged protein was detected in the extract of cells transfected with the parent vector pTriEx-1 (lane 2), whereas the TFIIB and/or the IE protein was detected as an HSV-tagged protein in the extracts of cells transfected either with the IE expression vector alone (lane 3) or cotransfected with the IE and TFIIB expression vectors (lane 4). Additionally, no cross-reactive recognition of the IE protein with the anti-TFIIB antibody (or, conversely, TFIIB with the anti-IE antibody) occurred with the cell extracts used in this study (data not shown). The cell extracts were immunoprecipitated with anti-TFIIB antibody (Santa Cruz Biotechnology), and the precipitated proteins were then analyzed by immunoblotting them with the anti-IE peptide-specific antibody (Fig. 2B). The HSV-tagged IE protein was effectively coimmunoprecipitated (approximately 20%) with the overexpressed HSV-tagged TFIIB (Fig. 2B, compare lane 4 to lane 3, which contains approximately 20% of the input protein). When the IE protein was singly expressed, a much weaker band was observed (lane 5), which is probably due to the interaction of the IE protein with endogenous TFIIB; this is similar to observations with other viral transactivators (30, 61). However, no HSV-tagged protein was immunoprecipitated with anti-TFIIB antibody in the extract of cells transfected with the parent vector (lane 6) as a negative control. Furthermore, preimmune sera were unable to immunoprecipitate the HSV-tagged proteins (lanes 7 to 9), demonstrating the specificity of our assays. The results of the coimmunoprecipitation experiments support the notion that the IE protein interacts with TFIIB and even suggest that the physical interaction occurs in vivo.

FIG. 2.

Coimmunoprecipitation of the IE protein and TFIIB. (A) Cells were transfected with only the parent vector pTriEx-1 (lane 2) or with pTriExIE (1–1487) containing the entire IE gene (lane 3) or were cotransfected with pTriExIE and pTriExhIIB containing the entire human TFIIB (hTFIIB) gene (lane 4). The resulting proteins, indicated by arrowheads, were efficiently expressed as HSV-tagged proteins as determined by immunoblotting with an anti-HSV tag monoclonal antibody (Novagen). Lane 1 contains the molecular mass markers. (B) Cell extracts of the IE protein and hTFIIB shown as approximately 20% of the input proteins (lane 3), as well as fractions that were immunoprecipitated (IP) with either anti-hTFIIB antibody (lanes 4 to 6) or preimmune serum (lanes 7 to 9), were analyzed by immunoblotting them with an anti-IE peptide-specific antibody (top). The presence of hTFIIB (approximately 20% of the input proteins) used in each coimmunoprecipitation reaction was confirmed by immunoblotting them with an anti-HSV tag antibody (bottom, solid arrowhead). Lane 1 contains the protein mass markers. Total extract from KyA-infected cells was also used as a molecular mass marker for the IE proteins (lane 2). Proteins nonspecifically precipitated with both anti-hTFIIB and preimmune sera are indicated by the open arrowhead.

Mapping the domain of the IE protein required for the IE protein-TFIIB interaction.

We next attempted to map the domain(s) of the EHV-1 IE protein responsible for binding TFIIB. We first generated a series of in-frame truncation and deletion mutants of the IE protein, which included amino-terminal, carboxyl-terminal, and internal regions, and expressed the mutants as radiolabeled, in vitro-synthesized proteins under the control of the SP6 promoter. The resulting radiolabeled proteins were then analyzed for the ability to bind full-length GST-hIIB tethered to glutathione-Sepharose beads. The autoradiographic results shown in Fig. 3B were quantitated, and the percentages of the radiolabeled IE derivatives bound by GST-hIIB were compared to the radiolabeled protein used in binding reactions (the first lane of each IVTT protein shows the amount of input), which was set at 100%. The results, shown as percent bound by TFIIB, are summarized to the right of the corresponding schematic diagram (Fig. 3A).

The results shown in Fig. 3B show that the three IE truncation mutants IE (1–1254), IE (1–904), and IE (1–620) were comparable to the wild-type protein IE (1–1487) in binding to GST-hIIB. Whereas removal of an additional 196 residues from the carboxyl terminus resulted in a truncated product, IE (1–424), that was severely impaired for binding GST-hIIB (∼8%), further truncation at aa 290 led to the complete loss of the ability of IE (1–289) to bind GST-hIIB; these results are comparable to the background level of nonspecific binding by GST alone when used as a negative control. These results suggest that the region between approximately aa 400 and 620 is necessary and sufficient for binding to TFIIB. Deletion mutants of the IE TAD (IEΔ2–88 and IEΔ2–178) bound to GST-hIIB as efficiently as did the wild-type IE protein (data not shown). The IR2 protein, a naturally truncated form of the IE protein lacking the first 322 residues, including the TAD within aa 3 to 89 (29, 58), also effectively bound to GST-hIIB. These data support the above-mentioned results and indicate that the amino-terminal portion of the IE protein is not responsible for binding TFIIB. Proteins with deletions that removed the amino-terminal 538 residues (IEΔ2–421 and IEΔ2–539) were as active in binding GST-hIIB as was the wild-type IE protein. However, removal of an additional 116 or 181 residues from the amino-terminal region of the IE protein (IEΔ2–654 and IEΔ2–719) significantly reduced binding of GST-hIIB (∼31 and ∼37%, respectively). Further removal of 38 or 107 residues from IEΔ2–719 produced the proteins IEΔ2–757 and IEΔ2–826, respectively, which did not bind to GST-hIIB. These results suggest that the region of the IE protein between approximately residues 540 and 757 contains sequences important for binding to TFIIB.

To confirm and finely map the putative TFIIB interaction domain, we generated a further series of internal deletion mutants based on the results from the initial TFIIB-binding assays and tested their binding activity to GST-hIIB as described above (Fig. 3B). Our results indicate that internal deletions of the amino-terminal portion of the IE protein (IEΔ181–249, IEΔ181–423, and IEΔ256–368) did not significantly inhibit binding to GST-hIIB (data not shown). Additional removal of the amino-terminal boundary (IEΔ426–578) as well as the carboxyl-terminal boundary (IEΔ621–757) of the potential TFIIB-binding domain reduced binding efficiency to ∼67 and ∼73%, respectively. However, removal of approximately the central region of the IE protein (IEΔ290–757 [Fig. 3B] and IEΔ181–904 [data not shown]) essentially abolished its binding to GST-hIIB. Furthermore, removal of an additional 351 residues from the central region of the IE protein produced an inactive protein (IEΔ407–757) which bound to GST-hIIB at least 14-fold less efficiently (∼7%). These results, taken together with those from the experiments that employed the amino-terminal and carboxyl-terminal deletion proteins, identify an ∼351-aa region between residues 407 and 757 that likely contains an important interface for binding TFIIB.

IE protein amino acids 407 to 757 specifically interact with TFIIB.

Considering the fact that deletion of an internal domain (aa 407 to 757) of the IE protein definitely reduced its binding activity to TFIIB, we next tested whether this domain of the IE protein specifically binds to TFIIB in vitro. Either GST or GST-hIIB was incubated with equivalent amounts of ³⁵S-labeled IE (407–757) produced from in vitro transcription and translation of pG3IE (407–757) or luciferase, and the binding proteins were detected by autoradiography and quantitated by PhosphorImager analysis. In vitro-synthesized luciferase was used as a negative control to ensure the specificity of the observed GST-hIIB interactions. The autoradiogram shown in Fig. 4A indicates that significant amounts of the radiolabeled IE (407–757) were bound to GST-hIIB (approximately 86% compared to the input protein). The binding activity of IE (407–757) is apparently the same as that of the wild-type IE protein (Fig. 3B). No binding was observed between luciferase and GST-hIIB. Negative control reactions showed that GST itself was unable to bind either radiolabeled IE (407–757) or luciferase. Our previous findings (13, 37, 38), as well as additional control reactions (Fig. 1), revealed that GST did not react with the IE proteins. The failure of GST to interact with the IE proteins, including the IE mutants, was a reproducible observation, as shown in Fig. 3B.

FIG. 4.

Mutant form of the IE protein harboring amino acids 407 to 757 specifically binds to TFIIB. (A) Protein-binding experiments assessing binding of the in vitro-synthesized IE aa 407 to 757 and the luciferase control by either GST or GST-hIIB were performed as described in the legend to Fig. 3, and the quantitative results of the autoradiogram are indicated below each lane. (B) Competition assays were carried out as outlined previously (13). GST-hIIB was preincubated with equivalent amounts (lane INPUT) of ³⁵S-labeled IE protein in the presence of increasing amounts (i.e., 1, 2, or 4 times that of the ³⁵S-IE) of in vitro-synthesized unlabeled competitor proteins, including wild-type IE (1–1487), IR2 (323–1487), and IE (407–757), which contains the IE-TFIIB interaction domain. As negative controls, IEΔ407–757, containing a deletion of the TFIIB-binding domain, or luciferase was included in some reactions. The percentage of the amount of ³⁵S-IE bound by GST-hIIB in the presence of each competitor compared to the amount of ³⁵S-IE bound by GST-hIIB in the absence of the above competitors, which was set at 100% (lane No competitor), is indicated below each lane.

Furthermore, Fig. 4B shows the results of competition assays in which the levels of binding of the ³⁵S-labeled wild-type IE protein to GST-hIIB in the presence of increasing amounts (i.e., approximately 1, 2, or 4 times that of the radiolabeled IE proteins) of the unlabeled wild-type IE (1–1487), IR2 (323–1487), or IE (407–757) were compared. Competition assays were also performed in parallel in the presence of increasing amounts (1 or 4 times that of the radiolabeled IE proteins) of unlabeled IEΔ407–757, which contains a deletion of the TFIIB-binding domain, and luciferase as negative controls. The precipitated ³⁵S-IE proteins were compared to the input lane, which did not contain any of the competitors described above and which was set at 100%. Quantitation of these results by PhosphorImager analysis is shown in Fig. 4B (the percentage of ³⁵S-IE proteins precipitated is given below each lane). Addition of increasing amounts of unlabeled IE, IR2, or IE (407–757) protein decreased the amounts of precipitated ³⁵S-IE proteins (from an average of 47 to 8%). These results indicate that the unlabeled IE (407–757), IE, and IR2 proteins competed with the radiolabeled IE proteins for binding to GST-hIIB. The unlabeled deletion mutant IEΔ407–757 and luciferase only minimally decreased the amount of ³⁵S-IE that was precipitated. These results demonstrate that the aa 407-to-757 region of the IE protein specifically interacts with TFIIB.

IE protein amino acids 407 to 757 harbor a TFIIB-binding domain.

As mentioned above, the IEΔ407–757 mutant was almost completely defective in TFIIB-binding activity (Fig. 3B). In addition, the observations that TFIIB specifically binds with equal efficiency to IE (407–757) and wild-type IE proteins (Fig. 4A) and, to a lesser extent, the IEΔ426–578 or IEΔ621–757 mutant (Fig. 3B), suggested that IE aa 407 to 757 may contain more than one TFIIB-binding domain. This possibility was tested by constructing a panel of IE truncation and deletion mutants expressed as GST fusion proteins (Fig. 5A). We determined by SDS-PAGE and Coomassie blue staining that the GST-IE derivatives were of the predicted sizes (Fig. 5B). The abilities of these proteins to interact with radiolabeled, in vitro-synthesized human TFIIB (³⁵S-labeled hIIB) was tested in protein-binding assays. The bound proteins shown in Fig. 5C were compared to the input [³⁵S]hIIB and were quantitated by PhosphorImager analysis. The binding efficiency data are indicated below the lanes of the corresponding proteins.

As a negative control, GST itself did not bind [³⁵S]hIIB. The IE (1–960) protein that was truncated at the carboxyl-terminal aa 961 was as effective in binding to [³⁵S]hIIB as was the wild-type IE protein (1–1487). As expected, the four truncation and deletion mutants GST-IE (1–424), GST-IE (1–289), GST-IE (1–180), and GST-IE (179–424) were dramatically deficient in binding to [³⁵S]hIIB. Additional GST-IE proteins encompassing aa 1 to 88 of the IE TAD (58) and aa 214 to 424 were also unable to bind [³⁵S]hIIB (data not shown). These results are consistent with those of the protein-binding assays, in which the radiolabeled truncation mutants IE (1–289) and IE (1–424), spanning the amino-terminal region, were not capable of binding GST-hIIB (Fig. 3B). Interestingly, GST-IE (407–757), which contains the entire IE-TFIIB interaction domain identified above, was equal to the wild-type IE protein in binding [³⁵S]hIIB. In contrast, the GST-IE derivatives GST-IE (407–539) and GST-IE (539–910), which contained partial upstream or downstream sequences of the TFIIB-binding domain, respectively, also bound to [³⁵S]hIIB, but to a lesser extent (approximately 70% compared to the GST-IE derivatives containing the entire TFIIB-binding domain). In data not shown, various GST-IE derivatives partially encompassing the TFIIB-binding domain of the IE protein (aa 422 to 597, 539 to 904, 539 to 757, 539 to 951, and 579 to 904) bound [³⁵S]hIIB with efficiencies similar to those observed for GST-IE (407–539) and GST-IE (539–910). GST-IR2 (323–1487) was as active in binding [³⁵S]hIIB as the wild-type IE protein, whereas binding was greatly reduced, to background levels, with GST-IE (758–1487) (Fig. 5C) and GST-IE (898–1487) (data not shown).

Interactions between TFIIB and IE mutant proteins lacking the TFIIB-binding domain were further analyzed in a set of coimmunoprecipitation experiments, as shown in Fig. 2. Cells were either transiently transfected with only pTriExIE (1–1487), (Δ407–757), (Δ426–578), or (Δ621–757), which express the IE mutants as HSV-tagged proteins, or cotransfected with the IE mutants and pTriExhIIB, which expresses an HSV-tagged TFIIB. Cell extracts were subjected to immunoprecipitation with an anti-TFIIB antibody (Santa Cruz Biotechnology), and the precipitated immunocomplexes were analyzed by immunoblotting them with an anti-IE peptide-specific antibody (Fig. 6, top). The presence of HSV-tagged TFIIB in each reaction was detected by using an anti-HSV tag monoclonal antibody (Fig. 6, bottom; as indicated, some lanes contained approximately 20% of the input proteins). As shown in Fig. 2B and C and Fig. 6 (lanes 3 and 4), the wild-type IE protein interacted with the exogenous TFIIB, as well as the much more weakly expressed endogenous TFIIB, in the transiently transfected cells. The IE mutants IEΔ426–578 and IEΔ621–757, which lack either part of the TFIIB-binding domain within the aa 407-to-757 region, were also efficiently coimmunoprecipitated with HSV-tagged TFIIB (compare lanes 9 and 12 to the input lanes 8 and 11). However, IEΔ407–757, which lacks the entire TFIIB-binding domain, was not coimmunoprecipitated with either HSV-tagged or endogenous TFIIB (lanes 5 [input] to 7). The absence of the immunocomplexes is similar to the results for the negative controls, in which cells were cotransfected with only pTriExhIIB and/or the parent plasmid, pTriEx-1 (lanes 14 and 15). Only small amounts of both IE mutants IEΔ426–578 and IEΔ621–757 were immunoprecipitated with endogenous TFIIB (lanes 10 and 13) compared to the full-length IE protein (lane 4). However, no specific protein was detected when preimmune sera were included in some of the reactions that were performed in parallel, which indicates the specificity of the immunoprecipitations (data not shown). Taken together, these results suggest that the aa 407-to-757 region of the IE protein contains one or more TFIIB-binding domains.

FIG. 6.

Coimmunoprecipitation experiments employing TFIIB and IE mutant proteins lacking the TFIIB-binding domain(s). Cells were cotransfected with vectors expressing HSV-tagged human TFIIB (hTFIIB) (lanes 3, 6, 9, 12, and 14) plus either the wild-type IE protein (lane 3), IEΔ407–757 (lane 6), IEΔ426–578 (lane 9), IEΔ621–757 (lane 12), or their cloning vector as a negative control (lane 14). Cells were transfected in parallel with only vectors expressing either the wild-type IE (lane 4), the IE mutants described above (lanes 7, 10, and 13), or the parent vector (lane 15). Coimmunoprecipitations were performed with anti-hTFIIB antibody, and the resulting proteins were analyzed by immunoblotting them with an anti-IE antibody. The amounts of immunoprecipitated proteins were compared to those of the corresponding input proteins indicated by asterisks (lanes 2, 5, 8, and 11). The presence of the overexpressed HSV-tagged hTFIIB in cotransfected cell extracts, indicated by the solid arrowhead, was determined by immunoblotting with an anti-HSV tag monoclonal antibody (bottom; ∗, lanes receiving approximately 20% of each input protein). Lane 1 contains the protein mass markers. Nonspecifically precipitated proteins are indicated by the open arrowhead. +, present; −, absent.

The TFIIB-binding domain of the IE protein is required for full transactivation of the EHV-1 early and late promoters.

To address the physiological significance of the interaction between the IE protein and TFIIB, transient-cotransfection experiments were performed. The pTriExIE constructs [IE (1–1487), IEΔ407–757, IEΔ426–578, and IEΔ621–757] were first individually cotransfected with the pTK-CAT reporter, in which the EHV-1 early TK promoter regulates expression of the CAT reporter gene, into L-M cells with or without the EHV-1 EICP27 and EICP22 effector plasmids (Fig. 7A). As reported previously (32, 55), the wild-type IE protein independently induced reporter gene expression by 23-fold compared to the reporter gene expression in the absence of an effector plasmid. Also, the transactivation ability of the IE protein increased in combination with the auxiliary regulatory proteins (EICP27 and EICP22) by 66-fold compared to the basal levels of the early TK promoter-CAT reporter alone. The IE protein's TFIIB-binding domain overlaps its DNA-binding domain within aa 422 to 597, which recognizes the consensus sequence motif (ATCGT) (38). As expected, IEΔ407–757, which has a deletion of the entire TFIIB-binding domain as well as the DNA-binding domain, was defective in activating expression of the pTK-CAT reporter. In addition and as expected, IEΔ426–578 failed to exhibit any trans-activation activity, as it has a deletion within the DNA-binding domain (data not shown). Interestingly, the IEΔ621–757 mutant, which lacks part of the TFIIB interaction domain but contains the DNA-binding domain, demonstrated reduced levels of activation compared to the wild-type IE protein alone or to the IE protein in combination with the auxiliary proteins (approximately 21 and 44% acetylation, respectively). Similar patterns of reporter gene expression were observed with the EHV-1 representative late IR5 promoter (pIR5-CAT) (Fig. 7B). Consistent with our previous findings (3, 32), the wild-type IE protein activated expression of the IR5 promoter approximately 3.2-fold alone and 12.2-fold in combination with the auxiliary proteins over basal levels (pIR5-CAT alone). Expression of pIR5-CAT was not observed in the presence of the IEΔ407–757 mutant, which lacks the TFIIB-binding domain. In contrast, IEΔ621–757, containing a partial deletion of the TFIIB-binding domain, activated the IR5-CAT reporter approximately 46% compared to the wild-type IE protein alone or in combination. These results suggest that the TFIIB-binding activity of the IE protein is essential for the IE protein to fully activate gene expression.

FIG. 7.

The IE-TFIIB interaction domain(s) is physiologically significant for the transactivating ability of the IE protein. Transient-cotransfection assays assessing transactivation of the representative EHV-1 early TK and late IR5 promoters linked to the CAT reporter gene by effector constructs expressing the IE deletion mutants targeting the TFIIB-binding domain were performed as described in Materials and Methods. All transfections were performed in triplicate. The error bars show standard deviations. (A) Transactivation of the viral pTK-CAT (early) reporter construct (1.4 pmol). (B) Transactivation of the viral pIR5-CAT (late) reporter construct (2.0 pmol).

DISCUSSION

A common theme of viral regulatory proteins in activating gene expression is interaction with general transcription factors in a manner that enhances recruitment of Pol II. Examples include the bovine papillomavirus E2 protein interactions with TFIIB and TBP (49, 63); HSV-1 VP16 associations with TFIIB, TFIIA, and TAF_II40 (10, 17, 39, 40, 42, 60); HSV-1 ICP4 recruitment of TFIIB, TBP, and TAF_II250 (6, 20, 54); and Epstein-Barr virus EBNA2 binding to TFIIB, TAF_II40, and RPA70 (43). In this report, we documented the specific interaction between the EHV-1 IE protein and TFIIB in in vitro protein-binding assays as well as in vivo coimmunoprecipitation studies. Interestingly, GST-hIIB interacted with multiple species of the IE protein produced during EHV-1 infections (7). Experiments with a panel of IE deletion and truncation mutants employed in in vitro protein-binding assays indicated that aa 407 to 757 of the IE protein specifically interacted with TFIIB. This observation was confirmed in competition assays in which a peptide encompassing aa 407 to 757 specifically prevented precipitation of ³⁵S-labeled full-length IE protein by TFIIB. All GST-IE derivatives containing the aa 407-to-757 domain tested appeared to bind TFIIB more strongly than GST-IE derivatives that partially encompassed this TFIIB-binding domain, suggesting the presence of one or more TFIIB-binding domains within the aa 407-to-757 region. To determine whether the IE protein contains one or more TFIIB-binding domain(s), additional protein-binding assays were performed which revealed that although IE (407–539) and IE (539–910) bound TFIIB with equal efficiency, these deletion mutants were less effective in binding TFIIB than the full-length IE protein. Coimmunoprecipitation experiments confirmed these results and further indicated that deletion of aa 426 to 578 caused a pronounced loss of TFIIB-binding activity. These results suggest that the DNA-binding domain (aa 422 to 597) may harbor the major site for interaction with TFIIB, since IEΔ621–757 appeared to interact more efficiently with TFIIB than did IEΔ426–578. A similar observation was reported for the bovine papillomavirus type 1 E2 transactivator protein (49, 63). Overall, our results defined a region between residues 407 and 757 required for maximal interaction of the IE protein and TFIIB.

Despite the fact that part of the TFIIB-binding site(s) maps within the DNA-binding domain of the IE protein (38), our recent results show that there is a direct interaction between purified IE protein and TFIIB in in vitro reactions (H. K. Jang, R. A. Albrecht, S. K. Kim, and D. J. O'Callaghan, unpublished data). It is possible that the interaction of the IE protein with DNA enhances its association with the TFIIB transcription factor. Our recent results also reveal that both the EHV-1 IE protein and the early EICP0 regulatory protein interact with the TATA-binding protein, as well as TFIIB. Since these two EHV-1 regulatory proteins have an antagonistic relationship (35), it is possible that competition for these transcription factors plays a role in this relationship.

The in vitro protein-binding assays do not address the physiological relevance of the interaction between this viral regulatory protein and TFIIB. As is the case in functional assays of transcription factors within a cellular context, it is difficult to assess the contributions of specific interactions between viral regulatory proteins and transcription factors in initiating transcription. We have attempted to gain some insight into the importance of the IE-TFIIB interaction by in vivo transient-cotransfection assays. An essential characteristic of transcriptional activators is their ability to cooperate with viral and/or cellular components to stimulate transcription synergistically (9, 10, 59). Transient-cotransfection assays to elucidate whether the TFIIB-binding domain is required for transactivation of viral promoters revealed that deletion of either the entire TFIIB-binding domain (aa 407 to 757) or a part of the domain (aa 621 to 757), completely inhibited the ability of the IE protein to activate expression of EHV-1 promoter-CAT reporters, supporting the importance of this interaction. The auxiliary regulatory proteins, EICP22 and EICP27, failed to overcome the inability of the IE mutant proteins to activate expression of the test promoters.

The observation that the TFIIB-binding domain overlaps the IE protein's DNA-binding domain was surprising, considering that some viral regulatory proteins bind to TFIIB via their acidic activation domains (42). Sequence analysis of this TFIIB-binding domain revealed that the region spanning aa 407 to 757 is neutral with respect to charge and contains an abundance of hydrophobic residues. Within the identified IE protein's TFIIB-binding domain is a helix-loop-helix motif, a homeodomain DNA recognition helix, that mediates the DNA-binding activity of the IE protein (37, 38). Even though the TFIIB-binding domain overlaps the DNA-binding domain, partial deletion of the identified TFIIB interaction sequence inhibited the ability of the IE protein to interact with TFIIB and to activate gene expression to normal levels.

Based on the data presented and our finding that the IE protein forms dimers (13), we present the following model to explain how the IE protein possibly activates transcription of viral genes via recruitment of TFIIB. Initially, the IE protein homodimerizes via the adjacent alpha helices present within the DNA-binding helix-loop-helix domain (aa 422 to 597 [38]). This self-interaction results in a conformational change of the loop structure such that the loop becomes accessible to bind to the major groove at the consensus target sequence, ATCGT. Concomitantly, dimerization of the IE protein may increase the exposure of key hydrophobic residues within the TFIIB-binding domain (aa 407 to 757), resulting in the increased efficiency with which the dimeric IE proteins bind to TFIIB. Computer analysis of TFIIB with the program ProtScale (available at http://www.expasy.ch/cgi-bin/protscale.pl?1) identifies regions of TFIIB that are rich in hydrophobic residues (i.e., aa 1 to 26, 55 to 64, 171 to 180, and 277 to 282) and that could interact with the exposed hydrophobic residues within the IE protein's TFIIB-binding domain.

The findings in this report add TFIIB to a growing list of proteins that interact with the sole multifunctional IE protein of EHV-1. Our recent studies have revealed that the IE protein interacts with itself, the EICP22 protein, and the cellular protein EAP and possibly with the EICP27 protein (unpublished observation), as well as proteins that posttranscriptionally modify the IE phosphoprotein and allow its transport to the nucleus. A major goal of our future work will concern efforts to identify other viral proteins and cellular factors that interact with the IE protein and hopefully to gain some understanding of how these interactions influence specific functions of this interesting and essential viral protein.