HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

J. Biol. Chem., Vol. 277, Issue 37, 34208-34216, September 13, 2002

This Article Abstract Full Text (PDF) All Versions of this Article: 277/37/34208    most recent M204808200v1 Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Upadhyaya, A. B. Articles by DeJong, J. Search for Related Content PubMed PubMed Citation Articles by Upadhyaya, A. B. Articles by DeJong, J. Social Bookmarking                      What's this?

The Germ Cell-specific Transcription Factor ALF

STRUCTURAL PROPERTIES AND STABILIZATION OF THE TATA-BINDING PROTEIN (TBP)-DNA COMPLEX^*

Ashok B. Upadhyaya, Mohammed Khan, Tung-Chung Mou, Matt Junker, Donald M. Gray, and Jeff DeJong^§

From the Department of Molecular and Cell Biology, University of Texas at Dallas, Richardson, Texas 75080

Received for publication, May 16, 2002, and in revised form, July 9, 2002

	ABSTRACT

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION CONCLUSION REFERENCES

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 apparent K_D 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 TFIIA/ 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 TFIIA/, the TFIIA-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.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION CONCLUSION REFERENCES

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) (1, 2). 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 TFIIA/) and small (e.g. yeast TOA2, Drosophila TFIIA-S, human TFIIA) subunits, which form a two-domain, boot-shaped structure (3, 4). The C-terminal domains of TOA1 and TOA2 form a six-stranded -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 -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 (5). 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 (5). However, in higher eukaryotes the large subunit is post-translationally cleaved into two separate subunits ( and ) at a site that is probably close to the junction between region II and an adjacent acidic domain (region III) (6-9).

The TFIIA/ and TFIIA genes are ubiquitously expressed in somatic tissues and are transcriptionally up-regulated in male germ cells (10). A factor related to TFIIA/, called ALF (TFIIA), is expressed only in male germ cells (10-12). Although larger than TFIIA/ (478 versus 376 residues) due to a longer internal region, ALF is able to interact with TFIIA and can restore activity to TFIIA-depleted HeLa cell nuclear extracts (11, 12). The discovery of ALF and other germ cell-specific factors such as TRF2/TLF (13-17), TAF_II105 (18), and cannonball (19) 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 (20-22). 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 mouse tbp gene initiates from at least six different sites within ~4 kb, only one of which is used in somatic cells (23), while the ACE gene initiates within the twelfth intron (24). 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 (10, 25, 26) 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 (14-17). 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 (27).

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 TFIIA/ are mostly unordered. Overall, the results demonstrate that ALF interacts with the TFIIA 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.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION CONCLUSION REFERENCES

Protein Expression-- 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. TFIIA/ 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 (13). Relaxed specificity TBP mutant constructs were obtained from Dr. Arnold Berk (28).

Bandshift Reactions-- Mobility shift assays were performed using a [-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 mM dithiothreitol, 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 TFIIA, and 6 ng TFIIA/. Reactions were usually incubated at room temperature for 30 min and were separated on 4% acrylamide 0.5× TBE gels that contained 5% glycerol.

Kinase Reactions-- TFIIA/ and ALF were phosphorylated in bandshift buffer with 134 nM [-³²P]ATP and 2000 units/ml of casein kinase II (CKII, New England Biolabs) for 1 h at 30 °C (29), separated by SDS-PAGE, and visualized by autoradiography. For bandshift reactions, ALF and TFIIA/ were phosphorylated with 200 µM cold ATP prior to addition to binding reactions.

DNase I Footprinting-- 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 MgCl₂ were then added to a final concentration of 2.5 mM Ca²⁺ 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 (5.32 M ammonium acetate, 168 µg/ml tRNA), and 200 µl of phenol:chloroform:isoamyl alcohol and resolved on 8% sequencing gels.

K_D Determinations-- 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 K_D = [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, and Y is the ratio of shifted complex to free probe as determined in lanes with oligonucleotides present (30). Lanes with 5 µM oligonucleotide were excluded from the analysis.

Off Rate 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 [-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 (t_1/2).

Circular Dichroism Spectroscopy-- Polypeptides were dialyzed overnight in 10 mM sodium phosphate and 150 mM sodium 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 (31). Protein concentrations were 0.75 mg/ml (1.42 × 10⁵ M) for full-length ALF, 0.16 mg/ml (0.70 × 10⁵ M) for TFIIA/ region II, and 0.11 mg/ml (0.43 × 10⁵ M) 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 _L  R in units of M¹ cm¹ 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 (32), SELCON3 (33), and CDSSTR (34) 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 Spectroscopy-- 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).

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION CONCLUSION REFERENCES

Identification of TFIIA-like Genes-- We wanted to evaluate whether ALF interacts with TFIIA or whether it might interact with some germ cell-specific TFIIA-like factor. To address this issue, we performed BLAST searches (35) of the human genome with a TFIIA query. The results showed that the TFIIA gene is located on chromosome 15q22 and is composed of five exons that span ~19 kb (GenBank^TM accession number AC092755; Fig. 1) (10). Sequences homologous to TFIIA 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 TFIIA1, TFIIA2, and TFIIA3, respectively. Two of these, TFIIA1 and TFIIA2, are interrupted by Alu repetitive elements, while TFIIA1 and TFIIA3 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.

View larger version (20K): [in this window] [in a new window]   Fig. 1.   Structural organization of the human TFIIA gene and three pseudogenes. A, the human TFIIA gene on chromosome 15q22 contains five exons (I-V) and spans ~20 kb. B, three apparent TFIIA pseudogenes were identified on chromosome 1 (denoted TFIIA1), chromosome 8 (TFIIA2), and chromosome 9 (TFIIA3). Sequences homologous to the TFIIA transcript are shown in dark gray, those similar to the reading frame are shown in white, and repetitive elements (ALU) are shown in light gray. The line beneath TFIIA1 indicates sequences found as expressed sequence tag cDNAs.

A search of human expressed sequence tags with the TFIIA2 or TFIIA3 sequences gave no matches, suggesting that these sequences are not expressed. In contrast, the 5'-end of TFIIA1 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 TFIIA1. Overall, the available data suggest that TFIIA1, 2, and 3 do not produce functional proteins and that ALF and TFIIA/ both use a small subunit encoded by the TFIIA 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 TFIIA// is referred to as "TFIIA", and ALF/ 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.

Computer Modeling of Heterodimeric ALF Complexes-- We generated homology-based structural models that describe the ALF complex using Swiss PDBViewer (36). Structures are based on an alignment and energy-minimization of the conserved regions from ALF, TFIIA/, and TFIIA 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 -barrel and -helical bundle domains. Both models predict a domain of positive charge along the surface of the -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.

View larger version (37K): [in this window] [in a new window]   Fig. 2.   Structural modeling of ALF. A, structure of the yeast TFIIA-TBP-DNA complex showing the TFIIA-TBP interaction and the barrel and bundle domains of yeast TFIIA (PDB ID 1YTF). The N and C termini of the large TOA1 subunit are shown in yellow, and the small subunit is shown in blue. B, ribbon diagram of the modeled ALF structure. C, ribbon diagram of the modeled ALF structure showing the end of the barrel predicted to interact with TBP. Beneath each model are surface charge illustrations of the same view. Acidic regions are red, nonpolar regions are white, and basic regions are blue. D, space-filling model of human TBP (red) and TATA DNA (green). Residues defective for interactions with TFIIA are labeled and shown in black.

Interactions of ALF with TBP and TRF2/TLF-- 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 0.1 M KCl but was diminished by ~90% in 0.4 M KCl (data not shown), consistent with chromatographic studies showing dissociation of TFIIA from a TBP-affinity column at ~0.3 M KCl (6).

View larger version (41K): [in this window] [in a new window]   Fig. 3.   Effect of ALF on TBP- and TRF2/TLF-DNA interactions. A, mobility shift assays show that ALF does not stabilize binding of recombinant TRF2/TLF to the AdML TATA element. Proteins present in each reaction are shown at the top of the figure. B, increasing amounts of TFIIA (lanes 3-7) or ALF (lanes 9-12) compete with one another for interactions with TBP. The lower shifted species is ALF-TBP-DNA; the upper shifted species is TFIIA-TBP-DNA.

The TBP-related factor TRF2/TLF does not interact with a canonical TATA element sequence, even in the presence of TFIIA (14-17). 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 (17). 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 (16, 17).

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.

Interaction of ALF and P-ALF with TBP Mutants-- 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.

View larger version (71K): [in this window] [in a new window]   Fig. 4.   Interaction of ALF and P-ALF with TBP mutants. A, TBP and four mutants (A184E, N189E, E191R, and R205E) were incubated with TFIIB to normalize activity in each preparation (lanes 1-5). Each TBP was then tested for its ability to support interactions with either ALF (lanes 7, 9, 11, 13, and 15) or TFIIA (lanes 6, 8, 10, 12, and 14). B, autoradiograph of recombinant ALF or TFIIA/ proteins phosphorylated in vitro with casein kinase II and [-32P]ATP and separated by SDS-PAGE. C, ALF proteins were incubated with or without casein kinase II and cold ATP and tested for their ability to form complexes with wild-type and mutant TBPs. All experiments used the AdML promoter oligonucleotide.

Phosphorylation of TFIIA/ by casein kinase II overcomes the destabilizing effect of a Y65A mutation in TFIIA on the TFIIA-TBP interaction (29). 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 TFIIA/ polypeptides incubated with casein kinase II and [-³²P]ATP (Fig. 4C, lanes 1 and 2) could be labeled (Fig. 4B, lanes 1 and 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 TFIIA/ (29). However, P-ALF (and P-TFIIA/, 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.

Protection of Promoter DNA by ALF-TBP Complexes-- ALF contains the largest internal nonconserved region identified so far (~287 residues), exceeding region II in TFIIA/ 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 at positions 40 to 42 (coding) and 42 to 44 and 48 to 50 (noncoding). This pattern matched that using TFIIA (Fig. 5A, lanes 13-17 and Fig. 5B, lanes 8-12) showing that despite its significantly larger size, ALF retains a TFIIA-like disposition at the promoter.

View larger version (103K): [in this window] [in a new window]   Fig. 5.   DNase I footprint analysis of ALF-TBP and TFIIA-TBP complexes on the AdML promoter. A, footprinting experiments with the noncoding strand labeled. Reactions contained fixed amounts of ALF (lanes 6-11) or TFIIA (lanes 12-17) and increasing amounts of TBP (indicated by the black ramp). B, footprinting experiments with the coding strand labeled. Reactions contained fixed amounts of ALF (lanes 1-6) or TFIIA (lanes 7-12) and increasing amounts of TBP.

Affinity of ALF-TBP Complexes-- TFIIB and TFIIA increase the affinity of the TBP-DNA interaction (37), thereby pushing the equilibrium toward completion of PIC formation and transcriptional initiation. To determine the apparent dissociation constant (K_D) for ALF-TBP on the AdML promoter, mobility shift assays were performed with increasing concentrations of unlabeled competitor (10, 50, 100, and 500 nM; 1 and 5 µM) (Fig. 6). Assays were performed under conditions that do not allow TBP to interact with DNA by itself. The ratio of bound complex to free probe was determined by PhosphorImager analysis and used to calculate an apparent K_D as described (30). Calculations using the 10 nM to 1 µM lanes revealed K_D values of 4.8 ± 2.7 nM for ALF-TBP and 5.5 ± 2.6 nM for TFIIA-TBP, indicating that ALF- and TFIIA-stabilized TBP complexes have comparable affinities for the AdML TATA element.

View larger version (83K): [in this window] [in a new window]   Fig. 6.   Determination of the apparent dissociation constant (KD). TBP-dependent mobility shift reactions were performed with either ALF or TFIIA in the presence of varying concentrations of unlabeled AdML TATA oligonucleotide competitor. The amounts of complex and free probe were determined by PhosphorImager analysis and analyzed as described under "Experimental Procedures."

We also wanted to determine whether ALF and TFIIA had differential effects on the formation of TBP-dependent complexes with other TATA elements (38-40). To address this point, TATA elements were synthesized from several testis-specific genes, including protamine 1, protamine 2, a testis-specific TBP promoter, and transition protein 2, and from the adenovirus major late promoter. Each oligonucleotide differed in its TATA and flanking sequences (Fig. 7A), and each was differentially recognized by TBP (data not shown). Assays that contained ALF or TFIIA gave results similar to those seen with TBP and did not depend on which of these two factors was present (Fig. 7B). In addition, the ability to compete for complexes formed on the AdML promoter was specific for each oligonucleotide and was not altered in the presence of ALF or TFIIA (Fig. 7C). The results show that TBP-dependent complex formation, at least as these promoters, was not subject to differential recognition and stabilization by ALF or TFIIA.

View larger version (31K): [in this window] [in a new window]   Fig. 7.   Recognition of variant TATA elements by ALF and TFIIA-containing TBP complexes. A, oligonucleotides used in this experiment were from the AdML promoter (AdML) and derived mutant (AdMLmut), protamine 1 (PRM1), protamine 2 (PRM2), transition nuclear protein 2 (TNP2), and a testis-specific mouse TBP gene promoter (TBP). B, oligonucleotides were kinase-labeled and incubated with TBP and either ALF or TFIIA. The amount of complex formation was measured by PhosphorImager analysis and is indicated in the bar graph. C, ALF-TBP and TFIIA-TBP complexes were formed on the AdML promoter and competed with cold oligonucleotide competitors. The percent competition is indicated in the bar graph.

Stability of ALF-TBP Complexes-- The stability of the TFIIA-TBP-DNA interaction is also an important aspect of gene regulation because the TFIIA-TFIID complex probably remains bound to provide a stable platform for reinitiation (41-43). In fact, TFIID persists at some promoters throughout the cell cycle (44). We therefore determined the t_1/2 of the ALF-containing TBP complex by challenge with an excess of a poly[d(A-T)] DNA competitor. Samples were removed from a master bandshift reaction at various time points (t = 0, 10 s, 30 s, 1 h, 2 h, 3 h, 4 h, 5 h, 6 h) and loaded onto a running polyacrylamide gel (Fig. 7A, lanes 1 and 4-11). Because the experiment occurred over several hours, two aliquots taken at zero time (before addition of the competitor) were either frozen at 80 °C or incubated at 30 °C and loaded with the final time point (Fig. 8A, lanes 2 and 3). This allowed normalization for any loss of activity that occurred during the course of the experiment. Data from triplicate experiments were used to calculate a t_1/2 of 650 and 770 min for the ALF-TBP and TFIIA-TBP complexes, respectively (Fig. 8B). The stability of ALF-TBP complexes (or ALF-TRF2/TLF-complexes) may allow them to persist on promoter DNA during meiosis.

View larger version (25K): [in this window] [in a new window]   Fig. 8.   Off-rate analysis of ALF-TBP-DNA complexes. A, complexes were formed on a AdML promoter oligonucleotide and challenged with an excess of unlabeled poly[d(A-T)]. Aliquots were taken at t = 0 and at 10 min, 30 min, 1 h, 2 h, 3 h, 4 h, 5 h, and 6 h and loaded onto the gel (lanes 1, 4-11). One reaction was performed with poly[d(A-T)] added first (lane 12). Two additional reactions were formed at t = 0 and incubated at 80 °C or 30 °C for 6 h and loaded with the final time point (lanes 2 and 3). B, graphical presentation of off-rate data using ALF (top) or TFIIA (bottom).

Circular Dichroism-based Prediction of ALF Secondary Structure-- The nonconserved internal region of ALF is larger and less biased in its amino acid composition than the internal region of TFIIA/. The corresponding region of TOA1 was not present in polypeptides used to crystallize yeast TFIIA. To obtain evidence as to what structure this region might adopt, we expressed and purified full-length ALF, internal ALF (residues 68-296), and internal TFIIA/ (residues 63-274) for CD spectroscopy (Fig. 9A). The CD spectrum of full-length ALF protein had negative CD bands near 208 and 222 nm, characteristic of -helices (Fig. 9B). The spectrum, with a crossover at about 205 nm, was comparable with that reported for an artificial bundle domain from TOA1 and TOA2 (45). Analysis of the spectrum suggested that 41 ± 0.5% of the protein was unordered. In the entire protein, this could include as many as 196 residues.

View larger version (25K): [in this window] [in a new window]   Fig. 9.   Analysis of full-length and internal ALF polypeptides by circular dichroism and fluorescence emission spectroscopy. A, polypeptides used for the analysis include a full-length ALF protein (ALF), a region of ALF from the nonconserved internal region II, and a region of TFIIA/ from region II. B, circular dichroism spectra of ALF full-length, ALF region II, and TFIIA/ region II. C, fluorescence emission spectra of the full-length ALF protein in native or denatured conditions show a shift of ~10 nm. D, Fluorescence emission spectra of ALF region II in native or denatured conditions show a shift of ~2 nm.

The CD spectra of the internal regions from ALF and TFIIA/ were much different from the full-length ALF protein. Both showed a broad minima at 200 nm typical of unordered polypeptides (Fig. 9B). Secondary structure analysis confirmed that the structures of these proteins were low in helix (9% for ALF and 13% for TFIIA/) and -sheet (15% for both ALF and TFIIA/) content, but there were relatively high amounts (55-60%) of unordered residues (Table I). Although it is possible that the internal regions adopt a different configuration after association with TFIIA and TBP, our analysis of these proteins in isolation suggest they are responsible for most of the unordered secondary structure content predicted in the full-length protein.

View this table: [in this window] [in a new window]   Table I Percentages of secondary structures for full-length ALF and internal region II segments of ALF and TFIIA/ The analyses were of spectra over the range of 250-190 nm using a 22-protein reference set that included references for the poly(Pro)II structure.

Two tests were performed to verify folding of the proteins used for CD. First, we confirmed that the full-length ALF protein, together with TFIIA, could form TBP-dependent complexes (data not shown). Second, we measured the fluorescence emission signals from tryptophan and tyrosine residues in aqueous and 8 M urea-containing buffers. The full-length ALF protein showed a red-shift of ~10 nm (Fig. 9C), suggestive of a transition from an ordered, or folded, state in which residues were solvent-protected to a disordered, or denatured, state in which internal residues were solvent-exposed. In contrast, the internal ALF polypeptide showed a shift of less than 2 nm (Fig. 9D), suggesting that residues in this polypeptide were in a partially solvent-exposed configuration even prior to denaturation. Overall, the data suggest that region II of ALF and TFIIA both form an irregular secondary structure, possibly part of a flexible "`shoelace" (4, 46).

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION CONCLUSION REFERENCES

In this paper we describe experiments aimed at understanding how the properties of the germ cell-specific factor ALF compare with its somatic counterpart TFIIA. In particular, we have examined interactions with TBP and the ability to stabilize binding of TBP to DNA.

Properties of ALF and TFIIA-- The studies here suggest that TFIIA probably serves as a common subunit for the formation of ALF and TFIIA heterodimers. This conclusion is based on a search of what is assumed to be an essentially complete human genome sequence data base, the apparent inability of three TFIIA-like sequences identified in that search to produce functional polypeptides, and the detection of a single TFIIA transcript in multiple tissues including testis (11). This is a potentially important result as it suggests the amount of ALF formed might involve a competition with TFIIA/ for the small subunit. The existence of a single TFIIA also places a limit on possible biochemical distinctions of the tissue-specific ALF complex, since important contacts with TBP are made by the small subunit (e.g. residues 66-73 of TOA2) (3, 4).

Nevertheless, ALF and TFIIA/ also possess characteristics that may contribute to these interactions and to other functional properties. For instance, the N- and C-terminal domains are not identical (82 and 88% similar, respectively) and at least two residues in the C terminus interact with TBP (3, 4). The acidic stretch located just before the first -strand of the large subunit may interact with helix 2 of TBP atop the saddle (4, 5, 38, 46-48). This region is less acidic in ALF than it is in TFIIA/. Interactions also occur between positively charged residues within the large subunit and the phosphodiester backbone of promoter DNA (3, 4). Finally, the large subunit is a target for phosphorylation, a modification that can partially restore TBP interactions using mutant TBP (Fig. 4, D and E) or mutant TFIIA proteins (29).

Our studies show that ALF exhibits properties that are consistent with those reported here and elsewhere on TFIIA. First, ALF-TFIIA and TFIIA/-TFIIA heterodimers can be modeled into three-dimensional structures that are similar to the yeast TFIIA crystal structure (Fig. 2). Second, ALF- and TFIIA-dependent complexes produce nearly identical patterns of DNase I protection and enhanced cleavage sites (Fig. 5). The results closely match those showing that TFIIA extends the area protected by TBP to include the sequence upstream of the TATA box (49, 50). Third, ALF-TBP and TFIIA-TBP complexes were shown to have half-lives (t_1/2) of 650 and 770 min, respectively (Fig. 8). These values are lower but comparable with those observed with recombinant or native TFIIA-containing complexes formed on a 13-base pair AdML TATA oligonucleotide (51). Third, the apparent K_D for the ALF-TBP-DNA complex was 4.8 ± 2.7 nM, and for the TFIIA-TBP-DNA complex it was 5.5 ± 2.6 nM (Fig. 6). Both values are similar to the previously reported K_D of 6 ± 2 nM for the TFIIA-TBP-DNA complex determined by DNase I footprinting (37). These similarities might be due in part to the fact that ALF contains a sequence of eight amino acids (HRSKNKWK) identical to those in TFIIA/ which interacts with the DNA backbone (3, 4). The one distinction observed in our bandshift assays is that ALF-TBP-DNA complexes migrate more rapidly than TFIIA-TBP-DNA complexes (Fig. 3B), suggesting differences in surface charge or conformation reverse the mobility expected based on size.

The results suggest that the ability to interact with TBP and stabilize the TBP-DNA interaction is a conserved and perhaps interchangeable function of ALF and TFIIA and raise the possibility that ALF substitutes for TFIIA in male germ cells. A precedent for such a notion comes from the observation that the autosomal PGK-2 gene is up-regulated in male germ cells to compensate for meiotic silencing of the X-linked PGK-1 gene (52). However, this model may not apply in the case of ALF and TFIIA, as both genes are autosomal (chromosomes 2 and 14, respectively) and both are up-regulated in male germ cells (10). Thus, if ALF does replace TFIIA/ in germ cells, it might involve translational control rather than gene inactivation. Another issue of interest is whether all possible combinations of factors, TFIIA-TBP, TFIIA-TRF2, ALF-TBP, and ALF-TRF2, actually occur in germ cells in vivo and in what proportions.

Studies of other somatic- and germ cell-specific isoforms also bear on the relationship between ALF and TFIIA/. For instance, in Drosophila the developmentally regulated 85E and germ cell-specific 84B tubulin isoforms are 98% similar, yet ectopic overexpression of 85E to levels of 50% or more in post-mitotic germ cells causes defects in axoneme assembly while being normal for formation of cytoplasmic microtubules and meiotic spindles (53). Likewise, the testis-specific factor TRF2/TLF and its somatic cell counterpart TBP both interact with TFIIA and TFIIB, but unlike TBP, TRF2/TLF is unable to bind a consensus TATA element (14-17). These studies demonstrate that somatic and germ cell isoforms, even those that are nearly identical, are not interchangeable. Thus, while our studies show ALF and TFIIA to be similar for interactions with TBP, distinctions could be revealed in genetic or biochemical assays that depend on interactions with TRF2/TLF, formation of a complete PIC, or activator function. In fact, TFIIA is involved in events that occur after PIC assembly even at TATA-less promoters whose initial recognition does not involve TFIIA-TBP complex formation (54-57). Moreover, one report suggests that ALF is 3-fold less effective than TFIIA in supporting activation in vitro by the Epstein-Barr virus Zta protein (12).

The extent of the similarities between ALF and TFIIA suggest that gene expression in somatic and germ-line tissues share a similar requirement for stabilization of the PIC and that differences in core promoter recognition and PIC stability are less likely to derive from ability of these factors to directly modulate the specificity of the core promoter factors. Instead, other mechanisms might account for the new patterns of core promoter recognition seen in germ cells. One possibility is that germ cell-specific TFIID complexes composed of ALF, TRF2/TLF, and others play a role promoter selection and PIC stability similar to the role that TFIIA, TBP, and TAFs are proposed to play in somatic cells (although the ability of TRF2/TLF to direct sequence-specific binding is still uncertain). Another possibility is that promoter selection in germ cells involves increased accessibility of the transcription machinery to genomic DNA. In this scenario, PIC assembly still involves ALF- or TFIIA-dependent stabilization of TBP or TRF2/TLF on the core promoter, but new assembly sites are available because of changes in chromatin packaging that occur in meiotic cells.

Nonconserved Regions in ALF and TFIIA/ Are Unordered-- A striking characteristic of ALF is that its internal domain is longer and different in amino acid composition compared with that of TFIIA/. This region was not analyzed in crystallographic studies of the yeast factor, and its structural make-up has not been previously addressed. The CD data (Fig. 9 and Table I) provide the first evidence to suggest that the internal regions of ALF and TFIIA/ are unordered, at least in the absence of other proteins with which they might interact. In contrast, the CD spectrum of the full-length protein predicts -helix, -sheet, and unordered structures that are consistent with an unordered internal domain and with the -helical and -sheet structures predicted from the yeast TFIIA crystal structure and the homology-modeled ALF heterodimer (Fig. 2).

The internal domains of TFIIA large subunits display a number of curious properties, some of which are consistent with the notion that this region is unordered. First, the length varies in human ALF (~287 residues), human TFIIA/ (~208 residues), Drosophila TFIIA-L (~195 residues), Arabidopsis TFIIA-L (~189 residues), Caenorhabditis elegans TFIIA-L (~182 residues), and yeast TOA1 (~112 residues) and is absent altogether in the large subunit encoded by the microsporidian Encephalitozoon cuniculi (58). Second, homology in this domain only exists in orthologous subunits from closely related species such as mouse and human (10). Third, the amino acid sequence is sometimes biased in its composition, as in human TFIIA/ (25% Gln, 12% Pro), Arabidopsis TFIIA-L (15% Pro), and yeast TOA1 (25% Asn). Fourth, the large subunit is in some species cleaved to form separate  and  subunits. In fact, recombinant  (amino acids 1-274) and  (amino acids 275-376) subunits of human TFIIA/ separated at QVDGT migrate near the position of the purified subunits (7). Interestingly, this site is conserved in human ALF (QVDGSGDTSS), human TFIIA/ (QVDGTGDTSS), zebrafish TFIIA/ (QVDGAGDTSS), and Drosophila TFIIA-L (QLDGALDSSD) and is less conserved but still present in C. elegans (QLDGGGGGMS), Arabidopsis (QVDGPMPDPY) and Schizosaccharomyces pombe (QIDGTIEDNE). Finally, genetic studies of TOA1 have shown that, except for a complete deletion, mutations in this region have little effect on the ability to support growth. Thus, any role for this domain, including that of a spacer, is predicted here to involve an unordered and perhaps flexible structure.

	CONCLUSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION CONCLUSION REFERENCES

In this report we present a series of in vitro protein-DNA interaction assays that show that ALF and TFIIA are nearly indistinguishable in their ability to interact with TBP. Overall, the results imply that ALF probably plays a role in the assembly and stabilization of TBP- or TRF2/TLF-dependent complexes in germ cells that is similar or identical to the role played by TFIIA in somatic cells.

	ACKNOWLEDGEMENTS

We thank Dr. Arnold Berk for TBP constructs, Dr. Wensheng Xie for assistance with bandshift reactions, and Asligul Yalcin for a TFIIA/ expression vector.

	FOOTNOTES

^* This work was supported by Grants AT-503 (to D. M. G.) and AT-1343 (to J. D.) from the Robert A. Welch Foundation and by a grant from the American Cancer Society (to J. D.).The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

Both authors contributed equally to this work.

^§ To whom correspondence should be addressed: Dept. of Molecular and Cell Biology, University of Texas at Dallas, 2601 N. Floyd Rd., Richardson, TX 75080. Tel.: 972-883-6882; Fax: 972-883-2409; E-mail: dejong@utdallas.edu.

Published, JBC Papers in Press, July 9, 2002, DOI 10.1074/jbc.M204808200

	ABBREVIATIONS

The abbreviations used are: GTF(s), general transcription factor(s); TBP, TATA-binding protein; AdML, adenovirus major late; PIC, preinitiation complex; CD, circular dichroism.

	REFERENCES

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION CONCLUSION REFERENCES