The Germ Cell-specific Transcription Factor ALF
2002; Elsevier BV; Volume: 277; Issue: 37 Linguagem: Inglês
10.1074/jbc.m204808200
ISSN1083-351X
AutoresAshok B. Upadhyaya, Mohammed Abdul Sattar Khan, Tung‐Chung Mou, Matt Junker, Donald M. Gray, Jeff DeJong,
Tópico(s)RNA Interference and Gene Delivery
ResumoThe assembly and stability of the RNA polymerase II transcription preinitiation complex on a eukaryotic core promoter involves the effects of TFIIA on the interaction between TATA-binding protein (TBP) and DNA. To extend our understanding of these interactions, we characterized properties of ALF, a germ cell-specific TFIIA-like factor. ALF was able to stabilize the binding of TBP to DNA, but it could not stabilize TBP mutants A184E, N189E, E191R, and R205E nor could it facilitate binding of the TBP-like factor TRF2/TLF to a consensus TATA element. However, phosphorylation of ALF with casein kinase II resulted in the partial restoration of complex formation using mutant TBPs. Studies of ALF-TBP complexes formed on the Adenovirus Major Late (AdML) promoter revealed protection of the TATA box and upstream sequences from −38 to −20 (top strand) and −40 to −22 (bottom strand). The half-life and apparentKD of this complex was determined to be 650 min and 4.8 ± 2.7 nm, respectively. The presence of ALF or TFIIA did not significantly alter the ability of TBP to bind TATA elements from several testis-specific genes. Finally, analysis of the distinct, nonhomologous internal regions of ALF and TFIIAα/β using circular dichroism spectroscopy provided the first evidence to suggest that these domains are unordered, a result consistent with other genetic and biochemical properties. Overall, the results show that while the sequence and regulation of the ALF gene are distinct from its somatic cell counterpart TFIIAα/β, the TFIIAγ-dependent interactions of these factors with TBP are nearly indistinguishable in vitro. Thus, a role for ALF in the assembly and stabilization of initiation complexes in germ cells is likely to be similar or identical to the role of TFIIA in somatic cells. The assembly and stability of the RNA polymerase II transcription preinitiation complex on a eukaryotic core promoter involves the effects of TFIIA on the interaction between TATA-binding protein (TBP) and DNA. To extend our understanding of these interactions, we characterized properties of ALF, a germ cell-specific TFIIA-like factor. ALF was able to stabilize the binding of TBP to DNA, but it could not stabilize TBP mutants A184E, N189E, E191R, and R205E nor could it facilitate binding of the TBP-like factor TRF2/TLF to a consensus TATA element. However, phosphorylation of ALF with casein kinase II resulted in the partial restoration of complex formation using mutant TBPs. Studies of ALF-TBP complexes formed on the Adenovirus Major Late (AdML) promoter revealed protection of the TATA box and upstream sequences from −38 to −20 (top strand) and −40 to −22 (bottom strand). The half-life and apparentKD of this complex was determined to be 650 min and 4.8 ± 2.7 nm, respectively. The presence of ALF or TFIIA did not significantly alter the ability of TBP to bind TATA elements from several testis-specific genes. Finally, analysis of the distinct, nonhomologous internal regions of ALF and TFIIAα/β using circular dichroism spectroscopy provided the first evidence to suggest that these domains are unordered, a result consistent with other genetic and biochemical properties. Overall, the results show that while the sequence and regulation of the ALF gene are distinct from its somatic cell counterpart TFIIAα/β, the TFIIAγ-dependent interactions of these factors with TBP are nearly indistinguishable in vitro. Thus, a role for ALF in the assembly and stabilization of initiation complexes in germ cells is likely to be similar or identical to the role of TFIIA in somatic cells. general transcription factor(s) TATA-binding protein adenovirus major late preinitiation complex circular dichroism Transcription of eukaryotic genes in vitro requires RNA polymerase II and a set of general transcription factors (GTFs)1 (TFIIB, -D, -E, -F, and -H) (1Roeder R.G. Trends Biochem. Sci. 1996; 21: 327-335Abstract Full Text PDF PubMed Scopus (718) Google Scholar, 2Orphanides G. Lagrange T. Reinberg D. Genes Dev. 1996; 10: 2657-2683Crossref PubMed Scopus (856) Google Scholar). An additional factor, TFIIA, interacts with the TATA-binding protein (TBP) to stabilize binding to core promoter DNA so that a transcriptionally active preinitiation complex (PIC) can be formed. TFIIA consists of large (e.g. yeast TOA1,Drosophila TFIIA-L, human TFIIAα/β) and small (e.g. yeast TOA2, Drosophila TFIIA-S, human TFIIAγ) subunits, which form a two-domain, boot-shaped structure (3Geiger J.H. Hahn S. Lee S. Sigler P.B. Science. 1996; 272: 830-836Crossref PubMed Scopus (241) Google Scholar,4Tan S. Hunziker Y. Sargent D.F. Richmond T.J. Nature. 1996; 381: 127-134Crossref PubMed Scopus (261) Google Scholar). The C-terminal domains of TOA1 and TOA2 form a six-stranded β-barrel that lies parallel to promoter DNA. Residues at the end of the barrel contact the first direct repeat of the saddle-shaped TBP protein. The N-terminal domains of TOA1 and TOA2 form an α-helical bundle that extends at a right angle away from DNA. Mutations in this domain affect viability in yeast, although its function is not known (5Kang J.J. Auble D.T. Ranish J.A. Hahn S. Mol. Cell. Biol. 1995; 15: 1234-1243Crossref PubMed Google Scholar). The N terminus (also referred to as region I) and C terminus (region IV) of the large subunit are separated by a nonconserved spacer (region II) whose structure is unknown and which is not required for activity (5Kang J.J. Auble D.T. Ranish J.A. Hahn S. Mol. Cell. Biol. 1995; 15: 1234-1243Crossref PubMed Google Scholar). However, in higher eukaryotes the large subunit is post-translationally cleaved into two separate subunits (α and β) at a site that is probably close to the junction between region II and an adjacent acidic domain (region III) (6DeJong J. Roeder R.G. Genes Dev. 1993; 7: 2220-2234Crossref PubMed Scopus (94) Google Scholar, 7DeJong J. Bernstein R. Roeder R.G. Proc. Natl. Acad. Sci. 1995; 92: 3313-3317Crossref PubMed Scopus (65) Google Scholar, 8Ma D. Watanabe H. Mermelstein F. Admon A. Oguri K. Sun X. Wada T. Imai T. Shiroya T. Reinberg D. Handa H. Genes Dev. 1993; 7: 2246-2257Crossref PubMed Scopus (77) Google Scholar, 9Yokomori K. Admon A. Goodrich J.A. Chen J.-L. Tjian R. Genes Dev. 1993; 7: 2235-2245Crossref PubMed Scopus (86) Google Scholar). The TFIIAα/β and TFIIAγ genes are ubiquitously expressed in somatic tissues and are transcriptionally up-regulated in male germ cells (10Han S.Y. Zhou L. Upadhyaya A. Lee S.H. Parker K.L. DeJong J. Biol. Reprod. 2001; 64: 507-517Crossref PubMed Scopus (48) Google Scholar). A factor related to TFIIAα/β, called ALF (TFIIAτ), is expressed only in male germ cells (10Han S.Y. Zhou L. Upadhyaya A. Lee S.H. Parker K.L. DeJong J. Biol. Reprod. 2001; 64: 507-517Crossref PubMed Scopus (48) Google Scholar, 11Upadhyaya A. Lee S.H. DeJong J. J. Biol. Chem. 1999; 274: 18040-18048Abstract Full Text Full Text PDF PubMed Scopus (73) Google Scholar, 12Ozer J. Moore P.A. Lieberman P.M. J. Biol. Chem. 2000; 275: 122-128Abstract Full Text Full Text PDF PubMed Scopus (57) Google Scholar). Although larger than TFIIAα/β (478 versus 376 residues) due to a longer internal region, ALF is able to interact with TFIIAγ and can restore activity to TFIIA-depleted HeLa cell nuclear extracts (11Upadhyaya A. Lee S.H. DeJong J. J. Biol. Chem. 1999; 274: 18040-18048Abstract Full Text Full Text PDF PubMed Scopus (73) Google Scholar, 12Ozer J. Moore P.A. Lieberman P.M. J. Biol. Chem. 2000; 275: 122-128Abstract Full Text Full Text PDF PubMed Scopus (57) Google Scholar). The discovery of ALF and other germ cell-specific factors such as TRF2/TLF (13Maldonado E. J. Biol. Chem. 1999; 274: 12963-12966Abstract Full Text Full Text PDF PubMed Scopus (31) Google Scholar, 14Ohbayashi T. Kishimoto T. Makino Y. Shimada M. Nakadai T. Aoki T. Kawata T. Niwa S. Tamura T. Biochem. Biophys. Res. Commun. 1999; 255: 137-142Crossref PubMed Scopus (31) Google Scholar, 15Rabenstein M.D. Zhou S. Lis J.T. Tjian R. Proc. Natl. Acad. Sci. 1999; 96: 4791-4796Crossref PubMed Scopus (146) Google Scholar, 16Teichmann M. Wang Z. Martinez E. Tjernberg A. Zhang D. Vollmer F. Chait B.T. Roeder R.G. Proc. Natl. Acad. Sci. 1999; 96: 13720-13725Crossref PubMed Scopus (101) Google Scholar, 17Moore P.A. Ozer J. Salunek M. Jan G. Zerby D. Campbell S. Lieberman P.M. Mol. Cell. Biol. 1999; 19: 7610-7620Crossref PubMed Scopus (83) Google Scholar), TAFII105 (18Freiman R.N. Albright S.R. Zheng S. Sha W.C. Hammer R.E. Tjian R. Science. 2001; 293: 2084-2087Crossref PubMed Scopus (168) Google Scholar), and cannonball(19Hiller M.A. Lin T.-Y. Wood C. Fuller M.T. Genes Dev. 2001; 15: 1021-1030Crossref PubMed Scopus (166) Google Scholar) has added a new layer of complexity to the mechanisms of eukaryotic gene regulation. Gene regulation in male germ cells is characterized by several unusual features (20Hecht N.B. BioEssays. 1998; 20: 555-561Crossref PubMed Scopus (391) Google Scholar, 21Eddy E.M. O'Brien D.A. Curr. Topics Dev. Biol. 1998; 37: 141-200Crossref PubMed Scopus (162) Google Scholar, 22Kleene K.C. Mech. Dev. 2001; 106: 3-23Crossref PubMed Scopus (183) Google Scholar). For instance, the core promoter regions of many germ cell-specific genes are GC-rich, and sequences of 100 bp or less are sufficient for germ cell-specific expression and somatic cell-silencing in transgenic mice. In addition, transcription can initiate at sites that are not normally used in somatic cells, resulting in multiple transcripts with unique 5′-untranslated regions. For example, the mousetbp gene initiates from at least six different sites within ∼4 kb, only one of which is used in somatic cells (23Schmidt E.E. Ohbayashi T. Makino Y. Tamura T Schibler U. J. Biol. Chem. 1997; 272: 5326-5334Abstract Full Text Full Text PDF PubMed Scopus (48) Google Scholar), while theACE gene initiates within the twelfth intron (24Howard T. Shai S.Y. Langford K.G. Martin B.M. Bernstein K.E. Mol. Cell. Biol. 1990; 10: 4294-4302Crossref PubMed Scopus (187) Google Scholar). These phenomena may be related to the fact that germ cells display elevated expression of GTFs and GTF-like factors such as ALF and TRF2/TLF (10Han S.Y. Zhou L. Upadhyaya A. Lee S.H. Parker K.L. DeJong J. Biol. Reprod. 2001; 64: 507-517Crossref PubMed Scopus (48) Google Scholar,25Schmidt E.E. Schibler U. Development. 1995; 121: 2373-2383Crossref PubMed Google Scholar, 26Persengiev S.P. Robert S. Kilpatrick D.L. Mol. Endocrin. 1996; 10: 742-747PubMed Google Scholar) that are involved in PIC assembly. Interestingly, TRF2/TLF is unable to recognize canonical TATA elements, and its specificity for core promoter DNA is uncertain (14Ohbayashi T. Kishimoto T. Makino Y. Shimada M. Nakadai T. Aoki T. Kawata T. Niwa S. Tamura T. Biochem. Biophys. Res. Commun. 1999; 255: 137-142Crossref PubMed Scopus (31) Google Scholar, 15Rabenstein M.D. Zhou S. Lis J.T. Tjian R. Proc. Natl. Acad. Sci. 1999; 96: 4791-4796Crossref PubMed Scopus (146) Google Scholar, 16Teichmann M. Wang Z. Martinez E. Tjernberg A. Zhang D. Vollmer F. Chait B.T. Roeder R.G. Proc. Natl. Acad. Sci. 1999; 96: 13720-13725Crossref PubMed Scopus (101) Google Scholar, 17Moore P.A. Ozer J. Salunek M. Jan G. Zerby D. Campbell S. Lieberman P.M. Mol. Cell. Biol. 1999; 19: 7610-7620Crossref PubMed Scopus (83) Google Scholar). In addition, new patterns of gene expression in male germ cells may be related to changes in chromatin structure that occur during meiosis and spermatogenesis as these may influence accessibility of DNA to the transcription factor machinery (27Lambert W.C. Lambert M.W. Schwartz L.M. Azar M.M. Advanced Cell Biology. Van Nostrand Reinhold Co., New York1981: 1077-1104Google Scholar). Here we have used both qualitative and quantitative analyses to describe the interactions between the germ cell-specific factor ALF with TBP and promoter DNA because this interplay is central to the function of the corresponding somatic cell factor, TFIIA. Among the questions considered are whether ALF interacts with residues in the first repeat of TBP and extends upstream protection of promoter DNA, whether phosphorylation of ALF affects its activity, and whether the interactions among ALF, TBP, and DNA are of comparable affinity and stability as those among TFIIA, TBP, and DNA. In addition, we have used circular dichroism (CD) spectroscopy to predict that the nonhomologous internal region of ALF and TFIIAα/β are mostly unordered. Overall, the results demonstrate that ALF interacts with the TFIIAγ subunit to form a heterodimeric complex that is biochemically and structurally similar to its somatic cell counterpart. These results and the implications for germ cell-specific gene expression are discussed. Recombinant proteins were typically expressed from the pRSET vector (Invitrogen) in Escherichia coli strain BL21(DE3)pLysS (Novagen) and purified over nickel-nitrilotriacetic acid-agarose (Qiagen). ALF region II was amplified with ALFrII-1 (5′-GCTCATATGGCACATCACCATCACCATCACCTTCAGTTGCCGCACAGCT-3′) and ALFrII-2 (5′-TGCGGATCCCTAAAGCTGAATATCAGTCACG-3′) and cloned into pRSET. This construct begins with the N-terminal extension MAHHHHHH- followed by ALF residues 68–296. TFIIAα/β region II was amplified with TFIIArII-1 (5′-GGCCATATGGAGCAGCAGCTTCTACTG-3′) and TFIIArII-2 (5′-CGCGGATCCTTAATCAACTTGTAAGACCAATGG-3′), and encodes residues 63 to 274. Human TFIIB was amplified with TFIIB-1 (5′-CGCCATATGCACCATCACCATCACCATGCGTCTACCAGCCGTTTGG-3′) and TFIIB-2 (5′-CGCGGATCCTTATAGCTGTGGTAGTTTG-3′) and cloned into the pRSET vector. This construct begins with the N-terminal extension MAHHHHHHV-. Human TRF2/TLF was amplified with TLF-1 (5′-CGCCATATGCACCATCACCATCACCATGATGCAGACAGTGATGT-3′) and TLF-2 (5′-CGCGGATCCTTATAAAATTTCTTTCC-3′) and purified as described (13Maldonado E. J. Biol. Chem. 1999; 274: 12963-12966Abstract Full Text Full Text PDF PubMed Scopus (31) Google Scholar). Relaxed specificity TBP mutant constructs were obtained from Dr. Arnold Berk (28Bryant G.O. Martel L.S. Burley S.K. Berk A.J. Genes Dev. 1996; 10: 2491-2504Crossref PubMed Scopus (99) Google Scholar). Mobility shift assays were performed using a [γ-32P]ATP kinase-labeled TATA-containing oligonucleotide that spanned −40 to −16 of the Adenovirus Major Late promoter (5′-AAGGGGGGCTATAAAAGGGGGTGGG-3′), or with core promoter sequences from protamine 1 (5′-CCTGGCATCTATAACAGGCCGCAGA-3′), protamine 2 (5′-GTCCCCCTTTATATACAAGCTCCCG-3′), a testis-specific TBP promoter (5′-TTATCTGTCTATATGTGCACCACAT-3′), transition protein 2 (5′-AGCCCCAACTATATAACCAGGTGGG-3′), and an AdML mutant (5′-AAGGGGGGCTAGAGAAGGGGTGGG-3′). Binding was performed in 10 mm HEPES (pH 7.9), 2% (w/v) PEG-8000, 5 mmdithiothreitol, 0.2 mm EDTA, 5 mm ammonium sulfate, 8% glycerol (made as a 5× stock). Reactions also included 5 mm MgCl2, 100 mm KCl, 1 mg/ml bovine serum albumin, and 10 fmol probe in a reaction volume of 25 μl. Typical protein amounts were 15 ng of TBP, 33 ng of A814E, 12 ng of N189E, 7 ng of E191R, 10 ng of R205E, 250 ng of TFIIB, 180 ng ALF, 80 ng TFIIAγ, and 6 ng TFIIAα/β. Reactions were usually incubated at room temperature for 30 min and were separated on 4% acrylamide 0.5× TBE gels that contained 5% glycerol. TFIIAα/β and ALF were phosphorylated in bandshift buffer with 134 nm [γ-32P]ATP and 2000 units/ml of casein kinase II (CKII, New England Biolabs) for 1 h at 30 °C (29Solow S. Salunek M. Ryan R. Lieberman P.M. J. Biol. Chem. 2001; 276: 15886-15892Abstract Full Text Full Text PDF PubMed Scopus (19) Google Scholar), separated by SDS-PAGE, and visualized by autoradiography. For bandshift reactions, ALF and TFIIAα/β were phosphorylated with 200 μm cold ATP prior to addition to binding reactions. A 158-base pair PCR fragment containing the AdML promoter (−101 to +40) was amplified with primers AdML-1 (5′-CGGAATTCTCTTCGGCATCAAGGAAGGTGATTGGTTTATAGG-3′) and AdML-2 (5′-CGCGGATCCCCAACAGCTGGCCCTCGCAGAC-3′) and subcloned into pBlueScript II SK (Stratagene). Coding or noncoding strands were labeled at the 3′-end with Klenow (New England Biolabs) and [32P]dATP (EcoRI end) or [32P]dCTP (BamHI end), followed by a cold chase with all four dNTPs. Binding reactions (50 μl) were incubated at room temperature for 30 min. CaCl2 and MgCl2were then added to a final concentration of 2.5 mmCa2+ and 7.5 mm MgCl2, followed by 0.5 units (5 μl) of diluted (1:10) DNaseI (Promega) for 2 min. Reactions were stopped with 70 μl of stop solution (5.32m ammonium acetate, 168 μg/ml tRNA), and 200 μl of phenol:chloroform:isoamyl alcohol and resolved on 8% sequencing gels. Mobility shift assays were performed with varying concentrations of cold competitor AdML TATA oligonucleotide (0, 10, 50, 100, and 500 nm and 1, 5 μm) in reactions where TBP was limiting. The amount of complex and free probe was quantitated by PhosphorImager (Molecular Dynamics) and, with the concentration of competitor oligonucleotide, were used to solve the equation KD = [S]tot/(X/Y + X −1 − Y), where [S]tot is the total concentration of competitor TATA, X is the ratio of shifted complex to free probe as determined from the 0-nm lane, andY is the ratio of shifted complex to free probe as determined in lanes with oligonucleotides present (30Carey M. Smale S.T. Transcriptional Regulation in Eukaryotes: Concepts, Strategies, and Techniques. Cold Spring Harbor Press, Cold Spring Harbor, NY2000: 263-265Google Scholar). Lanes with 5 μm oligonucleotide were excluded from the analysis. ALF and TFIIA master bandshift cocktails containing recombinant factors and reaction buffer were prepared. One aliquot was removed and combined with cold poly[d(A-T)] (final concentration of 21.3 μm) followed by the [γ-32P]ATP-labeled AdML TATA probe. This reaction served as a control to show competition by dAdT. The remainder of the master mixture received labeled probe and was incubated at 30 °C for 1 h. One third of this mixture was removed; one aliquot was loaded immediately (t = 0), one aliquot was frozen at −80 °C for 6 h, and one continued to incubate at 30 °C for 6 h. The latter two reactions were loaded at the end of the experiment to correct for loss of activity over time. The two thirds of the master mixture that remained received cold dAdT competitor, and aliquots were taken at t = 0, 0.5, 1, 2, 3, 4, 5, and 6 h for loading. Experiments were performed in triplicate, and band intensities were measured by PhosphorImager analysis. Dissociation curves were generated with SigmaPlot and extrapolated to 50% remaining complex (t½). Polypeptides were dialyzed overnight in 10 mm sodium phosphate and 150 mmsodium fluoride (pH 7.6). CD spectra were measured with a Jasco Model J715 spectropolarimeter, using a 0.05-cm path length cylindrical cell. Calibration of the spectropolarimeter was done as described (31Gray D.M. Hung S.-H. Johnson K.H. Methods Enzymol. 1995; 246: 19-34Crossref PubMed Scopus (246) Google Scholar). Protein concentrations were 0.75 mg/ml (1.42 × 10−5m) for full-length ALF, 0.16 mg/ml (0.70 × 10−5m) for TFIIAα/β region II, and 0.11 mg/ml (0.43 × 10−5m) for ALF region II. CD spectra were taken at 25 °C with a 2-nm spectral bandwidth, run scan speed of 50 nm/min, and a response time of 1 s. Each spectrum was an average of 12 accumulations, and data were collected at 0.1-nm intervals. CD data were smoothed by the Savitzky-Golay method using the program provided by Jasco, and the data were plotted at 1-nm intervals as εL − εR in units ofm−1 cm−1 per mole of residue. Protein CD spectra over the range of 250–190 nm were analyzed for percentages of secondary structure using CDPRO software CONTINLL (32Provencher S.W. Glockner J. Biochemistry. 1981; 20: 33-37Crossref PubMed Scopus (1937) Google Scholar), SELCON3 (33Sreerama N. Venyaminov S.Y. Woody R.W. Protein Sci. 1999; 8: 370-380Crossref PubMed Scopus (655) Google Scholar), and CDSSTR (34Johnson W.C. Proteins. 1999; 35: 307-312Crossref PubMed Scopus (633) Google Scholar) available at the website lamar.colostate.edu/∼sreeram. The analyses were performed using a 22-protein reference set that included references for the poly(Pro)II structure. Fluorescence emission spectra were measured with a SLM8000C (SLM Instruments) spectrofluorimeter at 25 °C between 300–400 nm using an excitation wavelength of 280 nm and a 4-nm bandpass width. Native proteins were in 10 mm sodium phosphate and 150 mm sodium fluoride (pH 7.6), and denatured proteins were in 8 m urea (pH 8.0). We wanted to evaluate whether ALF interacts with TFIIAγ or whether it might interact with some germ cell-specific TFIIAγ-like factor. To address this issue, we performed BLAST searches (35Altschul S.F. Gish W. Miller W. Myers E.W. Lipman D.J. J. Mol. Biol. 1990; 215: 403-410Crossref PubMed Scopus (73592) Google Scholar) of the human genome with a TFIIAγ query. The results showed that the TFIIAγ gene is located on chromosome 15q22 and is composed of five exons that span ∼19 kb (GenBankTM accession number AC092755; Fig.1) (10Han S.Y. Zhou L. Upadhyaya A. Lee S.H. Parker K.L. DeJong J. Biol. Reprod. 2001; 64: 507-517Crossref PubMed Scopus (48) Google Scholar). Sequences homologous to TFIIAγ were also present on chromosomes 1 (accession number AL451070), 8 (accession number AF252825), and 9 (accession number AL358934) (Fig.1). These sequences did not contain introns, suggesting they are processed pseudogenes, and we denote them as ΨTFIIAγ1, ΨTFIIAγ2, and ΨTFIIAγ3, respectively. Two of these, ΨTFIIAγ1 and ΨTFIIAγ2, are interrupted by Alu repetitive elements, while ΨTFIIAγ1 and ΨTFIIAγ3 do not contain an initiating methionine codon. All three sequences display nonconservative amino acid changes and contain frameshifts that would appear to prevent production of an intact, functional protein. A search of human expressed sequence tags with the ΨTFIIAγ2 or ΨTFIIAγ3 sequences gave no matches, suggesting that these sequences are not expressed. In contrast, the 5′-end of ΨTFIIAγ1 was present in cDNA libraries from three normal tissues (testis, fetal liver/spleen, and head/neck), three tumors (neuroendocrine lung carcinoid, amelanotic melanoma, and lung carcinoid), and one fibrosarcoma cell line (HT1080). However, none of the sequences identified in these libraries contained the 3′-end of ΨTFIIAγ1. Overall, the available data suggest that ΨTFIIAγ1, 2, and 3 do not produce functional proteins and that ALF and TFIIAα/β both use a small subunit encoded by the TFIIAγ gene. We have used this subunit to form active ALF and TFIIA complexes for biochemical studies, and unless otherwise specified, we extend the convention whereby complexed TFIIAα/β/γ is referred to as "TFIIA", and ALF/γ is referred to as "ALF". Furthermore, we have used TBP for our experiments because, unlike TRF2/TLF, TBP binds core promoter DNA efficiently and because there is a body of literature on the biochemical and physical properties of the TFIIA-TBP complex to which comparisons can be made. We generated homology-based structural models that describe the ALF complex using Swiss PDBViewer (36Guex N. Peitsch M.C. Electrophoresis. 1997; 18: 2714-2723Crossref PubMed Scopus (9795) Google Scholar). Structures are based on an alignment and energy-minimization of the conserved regions from ALF, TFIIAα/β, and TFIIAγ using the structure of yeast TFIIA as the template (3,4) (Fig. 2A). The models do not include the internal nonconserved region or the acidic region as these are divergent and were absent from the yeast TFIIA structure. As depicted in Fig. 2, B and C, the models contain characteristic β-barrel and α-helical bundle domains. Both models predict a domain of positive charge along the surface of the β-barrel that faces the negatively charged phosphodiester backbone (Fig. 2B). The figure also depicts residues in TBP whose mutation affects the TFIIA-TBP interaction (Fig. 2D). The models suggest that the ALF-dependent complexes are held together by similar electrostatic interactions as TFIIA-dependent complexes and may therefore share functional properties. ALF can form stable ternary complexes with TBP and the AdML TATA element (Fig.3A, lane 3) under conditions (5 mm Mg2+) where TBP alone cannot bind (lane 2). The ALF-TBP-DNA complex, although formed with a subunit that is 102 amino acids longer than the one present in TFIIA, migrates slightly faster than the TFIIA-TBP-DNA complex (lane 4). Maximal complex formation for both factors was observed in 0.1m KCl but was diminished by ∼90% in 0.4 mKCl (data not shown), consistent with chromatographic studies showing dissociation of TFIIA from a TBP-affinity column at ∼0.3m KCl (6DeJong J. Roeder R.G. Genes Dev. 1993; 7: 2220-2234Crossref PubMed Scopus (94) Google Scholar). The TBP-related factor TRF2/TLF does not interact with a canonical TATA element sequence, even in the presence of TFIIA (14Ohbayashi T. Kishimoto T. Makino Y. Shimada M. Nakadai T. Aoki T. Kawata T. Niwa S. Tamura T. Biochem. Biophys. Res. Commun. 1999; 255: 137-142Crossref PubMed Scopus (31) Google Scholar, 15Rabenstein M.D. Zhou S. Lis J.T. Tjian R. Proc. Natl. Acad. Sci. 1999; 96: 4791-4796Crossref PubMed Scopus (146) Google Scholar, 16Teichmann M. Wang Z. Martinez E. Tjernberg A. Zhang D. Vollmer F. Chait B.T. Roeder R.G. Proc. Natl. Acad. Sci. 1999; 96: 13720-13725Crossref PubMed Scopus (101) Google Scholar, 17Moore P.A. Ozer J. Salunek M. Jan G. Zerby D. Campbell S. Lieberman P.M. Mol. Cell. Biol. 1999; 19: 7610-7620Crossref PubMed Scopus (83) Google Scholar). However, we speculated that since ALF and TRF2/TLF are selectively transcribed in male germ cells, ALF might facilitate the binding of TRF2/TLF to a consensus TATA element. Addition of intact, recombinant TRF2 to bandshift reactions that contained TFIIA, TBP, and labeled promoter DNA caused a reduction in the TFIIA-TBP-DNA complex (data not shown), consistent with the idea that TRF2 was active and that it could prevent the association of TFIIA with TBP (17Moore P.A. Ozer J. Salunek M. Jan G. Zerby D. Campbell S. Lieberman P.M. Mol. Cell. Biol. 1999; 19: 7610-7620Crossref PubMed Scopus (83) Google Scholar). However, the presence of ALF did not cause TRF2/TLF to interact with the AdML TATA element (Fig.3A, lanes 6 and 7), and the addition of TFIIB (lanes 9 and 10) had no further effect. Thus, ALF-TRF2/TLF complexes formed in vivo may recognize variant TATA-like sequences or sequester ALF and other GTFs into transcriptionally nonproductive complexes (16Teichmann M. Wang Z. Martinez E. Tjernberg A. Zhang D. Vollmer F. Chait B.T. Roeder R.G. Proc. Natl. Acad. Sci. 1999; 96: 13720-13725Crossref PubMed Scopus (101) Google Scholar, 17Moore P.A. Ozer J. Salunek M. Jan G. Zerby D. Campbell S. Lieberman P.M. Mol. Cell. Biol. 1999; 19: 7610-7620Crossref PubMed Scopus (83) Google Scholar). We also examined the possibility that ALF and TFIIA could compete with one another for TBP binding. Taking advantage of their distinct migration, we set up bandshift reactions that contained fixed concentrations of ALF and increasing amounts of TFIIA (Fig.3B, lanes 2–7) or vice versa(lanes 8–12). The data show an that increase in TFIIA complex formation coincided with a loss of the ALF complex (lanes 2–7) and that an increase in ALF complex formation coincided with a loss of the TFIIA complex (lanes 8–12). The results demonstrate an interplay between ALF and TFIIA for TBP which could potentially regulate patterns of gene expression in vivo and which might also occur with TRF2/TLF. We next examined whether mutations in the first repeat of TBP that compromise interactions with TFIIA were also defective for interactions with ALF. To test this possibility, TBP and four derivatives (A184E, N189E, E191R, and R205E) were normalized for activity by forming TFIIB-dependent complexes (Fig.4A, lanes 1–5). In experiments with ALF, complexes did not form with the A184E, N189E, and E191R mutants (lanes 9, 11, and 13), while weak complexes were seen with the R205E mutant (lane 15). TFIIA was also unable to stabilize these mutant TBPs (lanes 8, 10, 12, and 14), and together with the competition assay shown in Fig. 3B, we conclude that ALF and TFIIA interact at the same surface of TBP. Phosphorylation of TFIIAα/β by casein kinase II overcomes the destabilizing effect of a Y65A mutation in TFIIAγ on the TFIIA-TBP interaction (29Solow S. Salunek M. Ryan R. Lieberman P.M. J. Biol. Chem. 2001; 276: 15886-15892Abstract Full Text Full Text PDF PubMed Scopus (19) Google Scholar). Because the residues involved (Ser-280, Ser-281, Ser-316, and Ser-321) are conserved in ALF (Ser-356, Ser-357, Ser-418, and Ser-423), we tested whether its activity might also be affected by this modification. We first showed that ALF and TFIIAα/β polypeptides incubated with casein kinase II and [γ-32P]ATP (Fig. 4C, lanes 1 and2) could be labeled (Fig. 4B, lanes 1and 2). Bandshift results showed that phosphorylation of ALF with cold ATP did not affect complex formation when TBP was used (Fig.4C, lanes 1 and 2), consistent with the previous report on TFIIAα/β (29Solow S. Salunek M. Ryan R. Lieberman P.M. J. Biol. Chem. 2001; 276: 15886-15892Abstract Full Text Full Text PDF PubMed Scopus (19) Google Scholar). However, P-ALF (and P-TFIIAα/β, data not shown) could partially restore interactions with each of the TBP mutants (Fig. 4C, lanes 3–10). The results suggest that ALF, like TFIIA, is potentially subject to a kinase-mediated post-translational modification. ALF contains the largest internal nonconserved region identified so far (∼287 residues), exceeding region II in TFIIAα/β by ∼80 residues. We wished to determine whether the greater size or structure of this region would affect the disposition of the ALF-dependent complex on promoter DNA, as judged by DNase I footprint analysis. In the presence of ALF, increasing amounts of TBP resulted in DNase I protection from approximately −40 to −22 of the noncoding strand of the AdML promoter (Fig. 5A,lanes 7–11) and −38 to −20 of the coding strand (Fig.5B, lanes 2–6), with DNase I hypersensitive sites appearing a
Referência(s)