A Testis-specific Transcription Factor IIA (TFIIAτ) Stimulates TATA-binding Protein-DNA Binding and Transcription Activation*

The general transcription factor IIA (TFIIA) stimulates RNA polymerase II-specific transcription by stabilizing the association of the TATA-binding protein (TBP) with promoter DNA, inhibiting repressors of TBP, and facilitating activator-dependent conformational changes in the preinitiation complex. TFIIA is encoded by two genes (αβ and γ) that are highly conserved between human and yeast. Here, we report the molecular cloning of a novel human gene that shares significant sequence similarity to the evolutionarily conserved amino- and carboxyl-terminal domains of TFIIAαβ. The TFIIA-related protein (TFIIAτ) was cloned from a testis-specific cDNA library, and its mRNA is expressed predominantly in testis tissue as determined by expressed sequence tag data base analysis and Northern blotting analysis. The TFIIA complex reconstituted with the testis-specific subunit, TFIIA (τ+γ), formed the TFIIA-TBP-TATA DNA (T-A) and TFIIA-TFIIB-TBP-TATA DNA (TAB) complexes indistinguishably from TFIIA (αβ+γ). TFIIA (τ+γ) supported basal and activated transcription for most activators in reactions reconstituted with TFIIA-depleted nuclear extracts. However, TFIIA (τ+γ) was reduced relative to TFIIA (αβ+γ) for stimulating transcription with at least one activator, suggesting that these two forms of TFIIA have activator specificity. These results suggest that TFIIAτ may be important for testis-specific transcription regulation.

The general transcription factor IIA (TFIIA) stimulates RNA polymerase II-specific transcription by stabilizing the association of the TATA-binding protein (TBP) with promoter DNA, inhibiting repressors of TBP, and facilitating activator-dependent conformational changes in the preinitiation complex. TFIIA is encoded by two genes (␣␤ and ␥) that are highly conserved between human and yeast. Here, we report the molecular cloning of a novel human gene that shares significant sequence similarity to the evolutionarily conserved amino-and carboxyl-terminal domains of TFIIA␣␤. The TFIIA-related protein (TFIIA) was cloned from a testis-specific cDNA library, and its mRNA is expressed predominantly in testis tissue as determined by expressed sequence tag data base analysis and Northern blotting analysis. The TFIIA complex reconstituted with the testis-specific subunit, TFIIA (؉␥), formed the TFIIA-TBP-TATA DNA (T-A) and TFIIA-TFIIB-TBP-TATA DNA (TAB) complexes indistinguishably from TFIIA (␣␤؉␥). TFIIA (؉␥) supported basal and activated transcription for most activators in reactions reconstituted with TFIIAdepleted nuclear extracts. However, TFIIA (؉␥) was reduced relative to TFIIA (␣␤؉␥) for stimulating transcription with at least one activator, suggesting that these two forms of TFIIA have activator specificity. These results suggest that TFIIA may be important for testis-specific transcription regulation.
The general transcription factors were initially identified as a set of accessory proteins required for accurate RNA polymerase II transcription initiation from the majority of viral and cellular promoters examined (reviewed in Refs. [1][2][3]. The general factors can assemble into a stable preinitiation complex nucleated by general transcription factor IID (TFIID) 1 binding the TATA box found at the Ϫ30 position of many viral and cellular promoters (4,5). The general transcription factors TFIIA and TFIIB bind directly to the TATA-binding protein (TBP) of the TFIID multiprotein complex and stabilize its as-sociation with promoter DNA (6). TFIIB provides a docking site for polymerase II and additional general factors that associate tightly with polymerase II. TFIIA binds directly to TBP and at least one TBP-associated factor (TAF) in the TFIID complex (7). TFIIA has a regulatory function in preinitiation complex formation that can be modulated by a variety of positive and negative acting transcriptional modifiers. TFIIA binding to TBP can prevent the inhibitory effects of transcriptional repressors, Dr1/DRAP1, HMG1, DSP1, MOT1, and TAF II 250 (8 -14). TFIIA can also directly mediate a stimulatory effect of promoter-specific transcriptional activators on the binding of TFIID to the promoter (15)(16)(17). TFIIA induces a conformational reorganization of TAFs in the TFIID complex bound to DNA, which can be further stimulated by a subset of transcriptional activators and correlates with an increase in transcription activity (18 -20). Inhibition of this conformational change may also be an important target for transcriptional repressors, pRb and RBP(CBF) (21,22).
TFIIA can be isolated from metazoan cells as a heterotrimeric complex encoded by two evolutionarily conserved genes (7,(23)(24)(25)(26)(27)(28). TFIIA␣␤ encodes the two largest subunits of the trimeric form of human TFIIA (referred to as ␣ and ␤), which are derived by a sequence-specific post-translational proteolytic cleavage of the ␣␤ polypeptide (7,24,27). The small subunit (referred to as ␥) is encoded by the TFIIA ␥ gene. In the budding yeast Saccharomyces cerevisiae, TFIIA can be isolated as a heterodimer, and the large subunit (TOA1) is not proteolytically cleaved (29,30). The amino-terminal (␣) and carboxyterminal (␤)domains of TFIIA␣␤ share significant sequence similarity with the amino-and carboxyl-terminal domains of yeast TOA1. The spacer regions between these two domains share little homology, and genetic experiments in yeast indicate that this region is not essential for viability (31). The conserved regions of TFIIA ␣, ␤, and ␥ are essential for viability in yeast (31).
Despite an absolute requirement for yeast viability, TFIIA may not be essential for the transcription of all promoters in yeast or metazoan genomes (29,32). Early biochemical experiments revealed that TFIIA was not required for basal transcription reactions reconstituted with TBP and highly purified or recombinant general transcription factors (33). However, these highly purified systems do not respond robustly to transcriptional modulation by promoter-specific activators and repressors. Transcription activation reactions reconstituted with TFIID (TBPϩTAF II s) have shown a strong dependence on TFIIA (23,26). However, transcription reactions reconstituted with yeast general transcription factors and the mediator-SRB complex does not have any apparent requirement for TFIIA (34,35). Similarly, there was not a strong dependence on TFIIA in transcription reactions reconstituted with holo-RNA polymerase II that lacked TAF II s (36). Thus, TFIIA may have a specific function in TAF II -dependent transcription activation. Consistent with this model is the observation that TFIIA can overcome the inhibitory activity of the amino-terminal domain of TAF II 250 and function as a high copy suppressor of a conditionally lethal mutation in the yeast TAF II 250 homologue, yTAF II 145 (13,14).
The variable requirement for TFIIA in transcription activation reactions suggests that TFIIA functions at a subset of promoters in vivo and is likely to have activator and promoter specificity. Recent experiments with TFIIA mutants compromised for their ability to bind to TBP reveal promoter and activator-specific defects in vivo, consistent with the prediction that TFIIA may be rate-limiting at a subset of promoters (32). Although the precise determinants conferring this specificity are not clear, it appears that at least a subset of activators can bind directly to TFIIA subunits. In metazoan systems, the viral activators Zta and HTLV-I Tax and the carboxyl-terminal activation domain of VP16 can all bind directly to TFIIA and stimulate the formation of a stable TFIIA-TFIID-promoter complex (15,16,26,37). These data support a role for TFIIA as a coactivator for TAF II -specific transcription regulation.
Variant and tissue-specific general transcription factors and coactivators may be required for complex combinatorial gene regulation. The Drosophila TBP-related factor (TRF1) is expressed primarily in nervous system tissue and can substitute for TBP with a corresponding novel set of associated factors (38,39). Similarly, a B cell-specific TAF II 135 has been identified in human tissue, but the function of this subunit is unclear (40). We have searched an expressed sequence tag data base and found that a gene related to TFIIA ␣␤ exists in human tissue. In this work we describe the isolation of the full-length gene encoding this TFIIA␣␤-related protein and demonstrate that it can functionally substitute for TFIIA␣␤ in T-A formation and stimulation of basal transcription in vitro. Interestingly, we found that this TFIIA␣␤-related gene is only detected in mRNA derived from testis. This testis-specific TFIIA (referred to as TFIIA) supported transcription activation for some but not all activators tested.

MATERIALS AND METHODS
Isolation of a cDNA encoding TFIIA-The Human Genome Sciences data base of over 2 million expressed sequence tags (ESTs) generated from over 800 human normal and disease tissue-specific cDNA libraries was screened for altered forms of TFIIA subunits (␣␤ and ␥) using the BLAST algorithm (41)(42)(43). Several ESTs with significant sequence similarity to the conserved amino-and carboxyl-terminal domains of human TFIIA␣␤ were found in testis cDNA libraries. To obtain the complete open reading frame, a random primed testis cDNA library (CLONTECH) was screened using a fragment of the partial cDNA clone. Filters were hybridized with the probe in PIPES, pH 8.0, 1% SDS, 50% formamide, and 1 g/l of denatured salmon sperm DNA at 42°C overnight and then washed at a final stringency of 0.2ϫ SSC, 0.1% SDS at 50°C. Multiple independent cDNA clones were identified, and the longest clone was isolated. Sequence analysis revealed an open reading frame of 478 amino acids with a predicted molecular mass of 52.4 kDa and an isoelectric point of 4.78.
E. coli DH5␣ cells transformed with pGEX-TBP, pGEX-TRP, pGEX-IIA␣␤, pGEX-IIA, and pGEX-2T were grown in LB at 37°C until an absorbance at 600 nm of 0.4 was reached. Isopropyl-␤-D-thiogalactopyranoside was added to 1 mM, and cultures were induced at 28°C for 4 h. Cells were harvested and washed once in 1ϫ phosphate-buffered saline and then sonicated in 20 mM Tris, pH 8.0, 5.0 mM EDTA, 1 mM DTT, 2 g/ml aprotinin, 2 g/ml leupeptin, and 1 mM PMSF. After sonication, 0.1% Nonidet 40 and 100 mM NaCl were added and incubated on ice for 15 min. The clarified lysates were mixed with glutathione-Sepharose (Amersham Pharmacia Biotech) for 2 h at 4°C with gentle shaking. The beads were then washed two times in 20 mM HEPES, pH 7.9, 12% glycerol, 200 mM KCl, 0.1% Nonidet P-40, 1 mM DTT, and 1 mM PMSF, and the fusion proteins were eluted with 10 mM reduced glutathione in 50 mM Tris-Cl, pH 8.0, 1 mM DTT at room temperature for 10 min. Eluted material was dialyzed directly into D100 buffer and stored in small aliquots at Ϫ70°C. Purification of GAL4-VP16, GAL4-CTF, GAL4-AH, and Zta has been described previously (48). Human TBP, TFIIA, and TFIIB have been isolated from E. coli BL21 cells as described (48).
GST Interaction, Electrophoretic Mobility Shift, and in Vitro Transcription Assays-pBS-TFIIA ␣␤, -␥, and -were used as templates for T7 or T3 polymerase directed coupled in vitro transcription-translation reactions using rabbit reticulocyte lysate (Promega) supplemented with [ 35 S]methionine. GST interaction assays were described previously (26). Electrophoresis mobility shift assays were described previously (26). In vitro transcription reactions with nickel-nitrilotriacetic acid agarose depletion of TFIIA has been described previously (26).
Chromosome Mapping-A panel of 24 monochromosomal somatic cell hybrids were obtained from Quantum Biotechnology, and the G3 panel of 83 radiation hybrids was obtained from Research Genetics. The following oligonucleotides which span a 150-base pair region of the TFIIA coding region were used for polymerase chain reaction analysis on 100 ng of template: DNA, GTGACACTACAGACTGTATCTG (5Ј primer), and CATGGATTCAATTGTGGAACAC (3Ј primer). 35 cycles of polymerase chain reaction amplification (94°C for 30 s, and 58°C for 45 s, and 72°C for 1 min) were performed on 100 ng of each hybrid in a 50-l reaction. Analysis of the radiation hybrid data was performed using the server at the Stanford Human Genome Center.
Antibodies and Western Blotting-A rabbit antiserum was raised against a synthetic peptide corresponding to nonconserved amino acids 275-294 of TFIIA (HDESLSTSPHGALHQHVTDI) according to the manufacturer's protocol (Research Genetics, Inc.). A 1:1000 dilution of polyclonal serum (second bleed) was used to perform Western blot analysis of the indicated protein extracts. A 1:3000 dilution of goat anti-rabbit (Bio-Rad) secondary antibody was used to detect the TFIIA antigen by chemiluminescence (ECL). Human testis protein extract was purchased from CLONTECH.

RESULTS
Isolation of a Human cDNA Encoding a TFIIA-related Protein-To determine whether a human gene encoding a protein with sequence similarity to the hTFIIA␣␤ subunit exists, a data base of expressed sequence tags generated from multiple tissues was searched by the BLAST algorithm (41). Several ESTs that showed significant homology to the conserved amino-and carboxyl-terminal domains of hTFIIA␣␤ were identified in testis-specific mRNA and used to probe a cDNA library derived from testis tissue. A full-length cDNA was isolated from testis-specific library encoding a 478-amino acid protein with significant similarity to hTFIIA␣␤ amino-and carboxylterminal domains (Fig. 1). The amino-terminal domain was 57% identical (32 of 56 amino acids), whereas the carboxylterminal domains were 76% identical (54 of the carboxyl-terminal 71 amino acids). A stretch of homology was also detected in the nonconserved spacer region near the presumptive ␣␤ cleavage site (amino acids 272-281 in TFIIA␣␤). No significant sequence similarity was detected in the nonconserved domain between amino acids 56 and 271 of TFIIA␣␤ or amino acids 53 and 347 in TFIIA. Although no in-frame stop codon resides upstream of the predicted open reading frame, it is likely that the full-length open reading has been isolated based on the following: the length of the TFIIA transcript identified by Northern analysis, which is consistent with the length of the cDNA identified; the presence of a reasonable Kozak sequence around the putative start codon; and the length of the identified open reading frame, which is consistent with that of TFIIA␣␤.
To determine the chromosomal location of the TFIIA gene, a panel of monochromosomal somatic cell hybrids retaining individual chromosomes was screened by polymerase chain reaction using TFIIA-specific primers. A polymerase chain reaction product was detected in human chromosome 2, whereas no amplification was observed in any other sample (data not shown). To sublocalize TFIIA on chromosome 2, a panel of 83 radiation hybrids were used. In addition to the human genomic DNA, amplicons were observed in hybrids 1, 2, 44, 50, 51, 61, 66, and 68 (data not shown). Analysis of this data using the Stanford Human Genome Center RH server revealed linkage to the SHGC-34039 marker on chromosome 2. Superposition of this map with the cytogenetic map of human chromosome 2 allowed the assignment of TFIIA to chromosomal band 2p15-16.
Testis-specific Expression of TFIIA-The cDNA encoding TFIIA␣␤-related protein was radiolabeled by random priming and used to probe mRNA derived from various human tissues by Northern blot analysis (Fig. 2). A specific hybridization with an RNA species of ϳ1.9 kilobases was detected in testis tissue but not in any other tissue examined, including spleen, thymus, ovary, prostate, lung, and brain (Fig. 2, upper panels). The Northern blot was stripped and hybridized with human TFIIA ␣␤-specific probe. A prominent RNA species migrating at ϳ7 kilobases was detected in spleen, thymus, prostate, testis, ovary, and small intestine. We did not detect significant expression of TFIIA␣␤ in colon or peripheral blood lymphocytes (Fig. 2, right panels). The blot was then stripped and hybridized with a ␤-actin-specific probe that reacted strongly with a ϳ2.0-kilobase species in all tissues tested. Note that although TFIIA and TFIIA␣␤ were expressed highest in testis, ␤-actin expression was slightly decreased in this tissue. These Northern blotting results are consistent with the detection of all TFIIA ESTs in testis-derived tissue (data not shown), which further suggests that TFIIA is a testis-specific form of the large TFIIA subunit.
To determine whether TFIIA protein was expressed in testis tissue, polyclonal rabbit antiserum was raised against a 20-amino acid peptide derived from the nonconserved domain of TFIIA. Serum derived from the peptide-immunized rabbits was reactive to recombinant TFIIA, whereas preimmune serum demonstrated no reactivity toward recombinant TFIIA (Fig. 3A). The TFIIA-specific antiserum was tested for its ability to recognize a specific polypeptide in testis tissue by Western blotting analysis of protein extract derived from human testis or from HeLa cell lines (Fig. 3B). A ϳ52-kDa polypeptide was detected by the peptide-immunized rabbit serum in testis tissue but not in HeLa extract (Fig. 3B). The preimmune serum did not detect any proteins of similar mobility in testis or HeLa extracts. We next tested whether TFIIA protein was expressed in a testis-derived teratocarcinoma cell line, NTERA-2 (Fig. 3C). We found that NTERA-2 expressed a ϳ52-kDa protein that was reactive to the TFIIA-specific antiserum, whereas extracts derived from HeLa (cervical carcinoma) and HepG2 (hepatic) cell lines had much lower or undetectable levels of expression. Again, the preimmune serum did not recognize any polypeptides of similar mobility in the NTERA-2 cell line extracts. These results demonstrate that TFIIA is expressed as a 52-kDa protein and that its expression is significantly enriched in testis-derived tissue.
TFIIA Stimulates TBP-DNA Binding-TFIIA was expressed and purified from E. coli and then renatured with the TFIIA-␥ subunit. TFIIA (ϩ␥) was compared with TFIIA (␣␤ϩ␥) for the ability to form the T-A complex with human TBP and the adenovirus E4 TATA box oligonucleotide probe (Fig.  4A). Addition of 10 -20 ng of recombinant human TBP does not form a very stable complex with the E4 TATA under these conditions of electrophoretic mobility shift assay (Fig. 4A, lanes  2 and 3). However, addition of TFIIA (␣␤ϩ␥) to TBP results in a stable TATA bound complex with slower mobility (Fig. 4A,  lane 4). Addition of TFIIB to the T-A complex further alters the mobility, indicating that TFIIB can efficiently bind to the T-A complex (Fig. 4A, lane 5). We next tested whether TFIIA (ϩ␥) could form a T-A complex similar to TFIIA (␣␤ϩ␥). Indeed, TFIIA (ϩ␥) formed a stable T-A complex at similar concentration ranges but slower mobility than TFIIA (␣␤ϩ␥) (Fig. 4A,  lane 6). Addition of TFIIB altered the mobility of the T-A complex formed with TFIIA (ϩ␥) (Fig. 4A, lane 7), further indicating that TFIIA can support T-A-B complex formation.
TFIIA Can Bind dTAF II 110 in Vitro-TFIIA␣␤ has been shown to interact with several polypeptides in vitro, including TFIIA␥, TBP, a TBP-related protein (TRP), and dTAF II 110 (7,26). We next determined whether TFIIA could similarly interact with these polypeptides using GST pull-down assays. TFIIA␣␤ and TFIIA were in vitro translated in the presence of [ 35 S]methionine and compared for their ability to bind to GST, GST-TBP, GST-TRP, and GST-IIA␥ (Fig. 4B). As reported previously, TFIIA␣␤ bound to GST-TBP, GST-TRP, and GST-IIA␥ (top panel). Under identical binding conditions, we found that TFIIA bound to GST-IIA␥ but did not demonstrate any specific binding for GST-TBP or GST-TRP (Fig. 4B, middle panel). Luciferase control did not bind significantly to GST, GST-TBP, GST-TRP, or GST-IIA␥ (bottom panel). Thus, although TFIIA (ϩ␥) could form a stable T-A complex with TBP in electrophoretic mobility shift assay (Fig. 4A), it appears that some TBP-specific contacts made by TFIIA␣␤ may be absent in TFIIA.
TFIIA was compared with TFIIA␣␤ for its ability to bind to dTAF II 110 (Fig. 4C). In vitro translated [ 35 S]methionine-labeled dTAF II 110 was tested for its ability to bind GST, GST-IIA␣␤, or GST-IIA. As shown previously, dTAF II 110 bound GST-IIA␣␤ but did not bind significantly to GST alone (Fig. 4C,  upper panel). Interestingly, dTAF II 110 also bound GST-IIA more abundantly than it bound to GST-␣␤ (Fig. 4C, lane 4). Luciferase control protein did not bind significantly to GST, GST-IIA␣␤, or GST-IIA. These results indicate that TFIIA FIG. 3. TFIIA protein is expressed in testis tissue. A, GST-TFIIA protein (rTFIIA) was probed by Western blotting with rabbit antiserum immunized against TFIIA-specific peptide (I) or preimmune (PI) serum. B, human testis tissue or HeLa cell line-derived extract was probed with TFIIA immune or preimmune rabbit serum by Western blotting. C, extracts derived from NTERA-2, HeLa, or HepG2 cell lines were probed with TFIIA-specific immune serum or preimmune rabbit serum by Western blotting. shares with TFIIA␣␤ the ability to bind dTAF II 110 and suggests that TFIIA will be capable of interacting with TFIID subunits to stimulate transcription.
TFIIA Can Stimulate Basal and Activated Transcription-HeLa nuclear extracts were depleted of TFIIA activity by incubation with nickel-nitrilotriacetic acid beads, which bind specifically to the stretch of seven histidines present in the nonconserved region of TFIIA␣␤. These extracts have been used previously to demonstrate the activity of recombinant TFIIA (␣␤ϩ␥) in transcription reactions with various activators (26). TFIIA-depleted extracts were shown to be defective for basal transcription with a promoter (TATA/Inr) containing the strong core promoter elements derived from the adenovirus major late promoter TATA box and the terminal deoxynucleotidyl transferase promoter initiator (Fig. 5A). Depleted extracts (Fig. 5A, lane 1) in the absence of exogenous TFIIA were significantly reduced for transcription compared with undepleted extracts (Fig. 5A, lane 4). Addition of recombinant TFIIA reconstituted with TFIIA (␣␤ϩ␥) stimulated transcription from the basal promoter in depleted extracts (Fig. 5A, lane 2). Similarly, addition of TFIIA (ϩ␥) to TFIIA-depleted extracts stimulated transcription from the TATA/Inr promoter (Fig. 5A, lane  3), indicating that TFIIA functions in basal transcription.
We next compared the ability of TFIIA (␣␤ϩ␥) and TFIIA (ϩ␥) to stimulate transcription with the Epstein-Barr virus lytic activator, Zta. Zta can form a stable complex with TFIIA (␣␤ϩ␥) and TFIID and may interact with TFIIA in a manner different than the activators described in Fig. 5B. In TFIIAdepleted extracts, Zta did not stimulate transcription in the absence of exogenous TFIIA (Fig. 5C, lane 1). Addition of TFIIA (␣␤ϩ␥) strongly stimulated Zta activated transcription (Fig.  5C, lane 2). Interestingly, addition of TFIIA(ϩ␥) was significantly reduced (ϳ3-fold) for Zta-activated transcription (Fig.  5C, lane 3). TFIIA (␣␤ϩ␥) and TFIIA (ϩ␥) were carefully normalized for T-A complex formation and for transcription stimulation by GAL4-AH (data not shown). At no concentration of TFIIA (ϩ␥) was Zta transcriptional activation restored to levels detected with TFIIA (␣␤ϩ␥). These results indicate that TFIIA is defective for Zta-mediated transcription and suggest that TFIIA may have specificity for a subset of transcriptional activators.

DISCUSSION
Tissue-specific gene expression is thought to occur largely through the combinatorial effects of tissue-and promoter-specific activators and repressors (49). The general transcription factors are thought to be essentially the same in all tissue types and may respond in a limited number of mechanisms to the synergistic signals generated by enhancers and promoters (1,2). Variation of the composition of general transcription factors has been observed in vitro for several promoters and has been implicated by some genetic experiments in yeast S. cerevisiae (50 -54). The identification of tissue-specific general transcription factors and coactivators suggests that tissue-specific gene expression may also be controlled by variation in the components of the preinitiation complex. The most provocative example is the identification of a neural tissue-specific form of TBP that has been isolated from Drosophila (38, 39). However, a  -6), or GAL4-AH (lanes 7-9) was assayed in TFIIA-depleted extracts. TFIIA (␣␤ϩ␥) was added to the reactions shown in lanes 2, 5, and 8. TFIIA (ϩ␥) was added to the reactions shown in lanes 3, 6, and 9. The primer extension reflecting the correctly initiated transcript for the G 5 E4TCAT template is indicated. C, transcription activated by Zta was assayed in TFIIA-depleted nuclear extracts. TFIIA (␣␤ϩ␥) was included in the reaction shown in lane 2. TFIIA (ϩ␥) was included in the reaction shown in lane 3. The primer extension product for the correctly initiated transcript from the Z 7 E4TCAT template is indicated. similar neural-specific form of TBP has not yet been identified in human tissue.
In this work we describe the first example of a human tissuespecific general transcription factor. TFIIA was discovered by analysis of a data base of expressed sequence tags derived from various human tissues. TFIIA-specific ESTs were found in testis tissue and Northern blot analysis detected TFIIA mRNA only in testis. Antibodies raised against a TFIIA-specific peptide reacted to a 52-kDa protein found in testis-derived tissue and cell line extract. TFIIA protein complexed with TFIIA␥ to form a functional TFIIA heterodimer that could stimulate TBP-TATA binding and support TFIIB assembly in electrophoretic mobility shift assays. TFIIA (ϩ␥) stimulated transcription in vitro from the GAL4-AH, VP16, and CTF transcriptional activators, with similar activity to that of TFIIA (␣␤ϩ␥). However, for at least one transcriptional activator, Zta, we found that TFIIA (ϩ␥) was reduced for activity relative to TFIIA (␣␤ϩ␥). Thus, TFIIA may confer activator specificity on the general transcriptional machinery by providing novel interacting surface for activation domain targets. Thus, TFIIA appears to be a tissue-specific subunit of a general transcription factor that may have overlapping, as well as unique features relative to the ubiquitously expressed TFIIA␣␤.
Several genes have been identified that are closely related to general transcription factor polypeptides. A TFIIB-related polypeptide (BRF), identified originally as a high copy suppressor of a TFIIA mutation, has been found to be an essential component of the RNA polymerase III general transcription factor TFIIIB (55). TRF1 has been identified from a Drosophila genetic screen for neurological phenotypes (38). TRF1 has subsequently been found to function homologously to TBP but is expressed predominantly in neural tissue (39). TRF1 can substitute for TBP in TATA binding and transcription reactions reconstituted with purified remaining general transcription factors. Like TBP, TRF1 associates with a novel set of associated factors, which are likely to confer promoter and activator specificity to TRF1 (39). A mouse TBP-like protein has been cloned (56). Unlike Drosophila TRF1, TBP-like protein is expressed in most tissue examined and does not bind TATA-DNA. We have also identified a human TRP that is highly related to TBP-like protein. 2 TRP does not bind to TATA-DNA but does interact with general transcription factors TFIIA and TFIIB. Unlike, TBP and TRF, TRP inhibits transcription in vitro and in vivo, presumably by sequestering essential general transcription factors, like TFIIA and TFIIB. 3 The isolation of a testis-specific form of TFIIA raised the question of whether TFIIA was the counterpart of an altered form of TBP, perhaps TRP. However, TFIIA does not bind TRP in solution (Fig. 4), and we have been unable to demonstrate any association or stimulation of TRP-DNA binding activity by TFIIA (data not shown).
The designation of TFIIA as a general transcription factor may be a misnomer. TFIIA is not required for reconstitution of basal transcription with highly purified essential general factors (33). TFIIA may better be characterized as a global but not universal transcriptional coactivator and derepressor. Several tissue-specific coactivators have been previously characterized. These include, among others, a B cell-specific form of TAF II 135 (40) and OCA-B, a B cell-specific coactivator for the octamer-binding protein 1 (57). More recently, a testis-specific coactivator (ACT) has been identified for the testis-specific isoform of the cyclic AMP response element modulator (CREM) (58,59). Interestingly, both the promoterspecific factor CREM and the CREM-specific coactivator ACT have expression restricted to testis tissue. Testis-specific gene expression may require specialized coactivators and global regulators, like ACT and TFIIA, to function in the protamine-rich environment unique to post-meiotic testis tissue. Future work can address the role of TFIIA in germ line-specific transcription regulation.