INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

The promoters of protein-encoding genes in metazoans have a bipartite structure, consisting of (i) a core promoter that encompasses the transcription start site and is directly responsible for the formation of the RNA polymerase II-containing preinitiation complex and (ii) binding sites for gene-specific regulators (34). Two common core promoter elements in higher eukaryotes are the TATA box, located 25 to 30 bp upstream of the transcription start site, and the initiator (Inr), a pyrimidine-rich element surrounding the start site. These elements are functionally analogous, since each is capable of directing accurate transcription initiation. In higher eukaryotes, core promoters have been shown to be quite heterogeneous; some contain both of these elements, while others contain either a TATA box or an Inr. In addition, for some genes the core promoter elements are undefined, as they lack either of these elements (reviewed in reference 28).

Despite significant progress in understanding the biochemical mechanisms of transcription in higher eukaryotes, the properties governing gene expression in deep-branching eukaryotic lineages remain poorly understood. The best studied of these organisms are the parasitic protists, for which recent studies have begun to define motifs critical for promoter function (8, 20, 22, 25, 29, 30, 33). The most striking result from these studies is the apparent lack of typical eukaryotic promoter elements, such as the TATA box or Inr. One group of parasites, the kinetoplastids, appears to lack sequence-specific promoters entirely. Instead, their protein-encoding genes are transcribed in large polycistronic units, and the process of trans-splicing regulates the formation of discrete mRNAs (reviewed in references 4 and 15). This lack of typical eukaryotic promoter elements may reflect the great divergence that has occurred since these organisms branched from the main line of eukaryotic descent.

Recently, we have examined the promoters of protein-encoding genes in the parasitic protist Trichomonas vaginalis (16, 17). This organism is a common human pathogen belonging to the phylum Parabasalia, one of the deepest-branching eukaryotic lineages (13). Trichomonads are characterized by a number of primitive cellular features, including the absence of two hallmark eukaryotic organelles, peroxisomes and mitochondria (5). Our studies have shown that promoters of protein-encoding genes in T. vaginalis, such as those of higher eukaryotes, consist of gene-specific regulatory elements and a conserved core promoter containing the transcription start site (16, 17). Core promoters in T. vaginalis appear to consist solely of a highly conserved Inr element whose consensus sequence, T C A + 1 Py (T/A), matches that of the metazoan Inr, Py Py A + 1 N (T/A) Py Py (17, 23, 27). This Inr has been found in all examined T. vaginalis genes and is both a structural and a functional homologue of the metazoan Inr. The T. vaginalis Inr is responsible for transcription start site selection, and its sequence requirements for activity are the same as those for metazoan Inr function (9, 17). The Inr, therefore, represents the first cognate core promoter element found in both deep-branching and metazoan eukaryotes, suggesting that it evolved early during eukaryotic evolution.

Since T. vaginalis appears to rely exclusively on the Inr to direct transcription initiation, it is likely that the trichomonad transcription machinery is highly optimized for Inr function. For this reason, we believe T. vaginalis is an excellent system in which to explore the mechanisms and evolution of Inr-mediated transcription in eukaryotes. A fundamental question in understanding Inr function in metazoans is the identity of the protein(s) that recognizes the Inr during transcription initiation. While several proteins, such as YY1, USF, TFII-I, and RNA polymerase II itself, have been shown to bind to sites that coincide with the Inrs of specific genes, TFIID has emerged as the prime candidate for recognizing the basal Inr during transcription initiation (reviewed in references 2, 3, and 26). In metazoans, the TFIID subunits TAF_II150 and TAF_II250 are required for Inr function and are in close contact with the core promoter DNA (11, 12, 21, 31, 32). In addition, a TAF_II150-TAF_II250 heterodimer has been shown to specifically recognize the Inr in a binding site selection assay, providing direct evidence of Inr binding by one or both of these TAFs (3). The exact mechanism of Inr binding, including the identity of the Inr-binding domain(s), remains to be defined.

To identify the protein(s) involved in Inr recognition in trichomonads, we have used DNA affinity chromatography to isolate an Inr-binding protein from T. vaginalis. This 39-kDa protein, IBP39, appears to be a novel DNA-binding protein, since database searches have failed to reveal homologues in other organisms. Using DNase I footprinting assays to assess the DNA-binding specificity of IBP39, we demonstrate that the sequence requirements for binding correspond to the conserved T. vaginalis Inr motif. Our results describe a previously unidentified, sequence-specific Inr-binding protein that is likely to be involved in promoter recognition by the T. vaginalis transcription machinery.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Cell culture. T. vaginalis strain C1 (ATCC 3001) was grown in Diamond's medium supplemented with 10% (vol/vol) horse serum and iron as described previously (23).

Plasmid construction. The SCS-CAT, Inr1, Inr2, Inr3, and Inr13 plasmids have been described previously (17). The FP plasmids (see Fig. 9) were generated by using the QuikChange site-directed mutagenesis method (Stratagene). Each mutagenesis reaction contained the SCS-CAT plasmid as a template DNA and two complementary oligonucleotides, each containing the desired mutation surrounded by 15 bp of flanking sequence on both the 5' and the 3' sides. Constructs were verified by sequencing by using the Sequenase kit (USB).

Preparation of the DNA affinity column. Multimers of the wild-type SCS Inr (Fig. 2A) were generated by a PCR-based method (7) by using the complementary primers wtInrA (5'-CTTGTTCACTTCACATTAATGCCCTTGTTCACTTCACATTAATGCC-3') and wtInrB (5'-GGCATTAATGTGAAGTGAACAAGGGCATTAATGTGAAGTGAACAAG-3'). PCRs contained 50 ng of each primer, 1× PFU buffer [20 mM Tris-HCl (pH 8.8), 10 mM (NH₄)₂SO₄, 2 mM MgSO₄, 100 µg of bovine serum albumin/ml, 0.1% Triton X-100], a 10 µM concentration of each deoxynucleoside triphosphate, and Pfu polymerase. Cycling conditions were 95°C for 2 min, followed by 30 cycles of 95°C for 1 min, 65°C for 2 min, and 37°C for 1 min. PCR products were purified by phenol-chloroform extraction, followed by precipitation with ammonium acetate and ethanol. Each PCR yielded ~6 µg of multimers that were between ~200 and ~5 kb in length. Multimer DNA (~200 µg) was then bound to CNBr-activated Sepharose 4B (Pharmacia) according to the manufacturer's instructions.

Purification of IBP39. Nuclear extracts from an 80-liter culture of T. vaginalis were prepared as described previously (17). The extracts were then fractionated by (NH₄)₂SO₄ precipitation; (NH₄)₂SO₄ was added to 60% saturation (0.36 g/ml), stirred for 45 min at 4°C, and then spun at 10,000 rpm in a JA-14 rotor (Beckman) for 45 min at 4°C. To the supernatant, additional (NH₄)₂SO₄ was added to 80% saturation (0.13 g/ml) and precipitated as described above. The protein pellets were each resuspended in 40 ml of HGED0.1 (20 mM HEPES-KOH [pH 7.9], 100 mM KCl, 1 mM EDTA, 1 mM dithiothreitol [DTT], 10 µg of leupeptin/ml, 50 µg of TLCK [N-p-tosyl-L-lysine chloromethyl ketone]/ml, 1 mM phenylmethylsulfonyl fluoride, and 20% glycerol) and dialyzed against 4 liters of HGED0.1 for 8 h with three buffer changes. The 80% fraction contained 500 to 600 mg of total protein.

Mobility shift assay. Probes corresponding to 15 to +15 of the SCS (Fig. 1) or ferredoxin Inrs (see Fig. 7) were prepared by annealing complementary oligonucleotides and end labeling them with [-³²P]ATP and T4 polynucleotide kinase. The labeled probe was then purified on an 8% polyacrylamide gel. Unlabeled complementary primers containing the Inr3 and Inr13 mutations (Fig. 1) were annealed and used in competition assays. Each binding reaction contained the protein fraction, 5,000 cpm of labeled probe, 500 ng of poly(dG-dC), 20 mM HEPES-KOH (pH 7.9), 100 mM KCl, 1 mM EDTA, 1 mM DTT, 0.01% NP-40, and 10% glycerol. Reaction mixtures were incubated for 20 min at room temperature and run on a 6% 0.5× TBE polyacrylamide gel at 100 V for 2 h. For the supershift experiments, the binding reaction mixtures were preincubated with either preimmune or anti-IBP39 immunoglobulin G (IgG) for 30 min on ice prior to addition of the probe.

View larger version (34K): [in this window] [in a new window]   FIG. 1.   Initiator-specific binding activity in T. vaginalis nuclear extracts. A 37-bp probe, corresponding to positions 15 to +15 of the SCS Inr region, was used in mobility shift assays with 20 µg of T. vaginalis crude nuclear extract. The SCS Inr is a tandem Inr (underlined) that is known to contain two transcription initiation sites (indicated by the arrows) (17). Competition assays were performed with a 100× molar ratio of unlabeled wild-type and mutant SCS Inr probes. Lane 1, no extract; lane 2, with extract; lane 3, competition with wild-type SCS Inr; lane 4, competition with Inr13; lane 5, competition with Inr3. DNA sequences of the three probes are shown with the mutations in the Inr13 and Inr3 probes in boldface.

DNase I footprinting assay. The 184-bp probes were generated by PCR with the 7 primer (5'-ACTTACGCTTCAATTAAGGG-3'), which was end labeled with [-³²P]ATP and T4 polynucleotide kinase; the CAT-PE primer (5'-TATATCAACGGTGGTATATCCAGTG-3'); and either the SCS-CAT, Inr1, Inr2, Inr3, or Inr13 plasmids (17) or the FP1-16 plasmids (this study) as a template. Labeled probes were purified on a 6% polyacrylamide gel. DNase I footprinting reactions were then performed as described by Marshak et al. (18). Briefly, each 50-µl binding reaction mixture contained the protein fraction, 10,000 cpm of labeled probe, 10 mM HEPES-KOH (pH 7.9), 50 mM KCl, 1 mM EDTA, 1 mM DTT, 10% glycerol, and 0.2% polyvinyl alcohol. Reaction mixtures were incubated at room temperature for 20 min. Samples were then digested with either a 1:2,500 (for no-protein controls) or a 1:1,000 dilution of 2.5 mg of DNase I/ml, extracted with phenol-chloroform, ethanol precipitated, and separated on an 8% sequencing gel.

Denaturation-renaturation analysis. Denaturation-renaturation analyses were performed as described by Baeuerle and Baltimore (1). Proteins were separated by SDS-PAGE, and slices were excised from the gel. Proteins from the gel slices were eluted, denatured, and renatured individually. Then, 10 µl of the renatured protein was used in each mobility shift reaction; 25 µl was used in each DNase I footprinting reaction.

Isolation of IBP39 genomic and cDNA clones. To generate a fragment of the IBP39 gene, degenerate primers based on peptides 1 and 3 (Fig. 4B) were used in PCR with T. vaginalis genomic DNA. Use of the forward primer 5'-AACGTYGCYYTYGTYATGGG-3' (corresponding to amino acids NVALVMG) and the reverse primer 5'-CTTRTCGTGYTGRAGYTGYTC-3' (corresponding to amino acids EQLQHDK) resulted in a 90-bp product after PCR with 100 ng of T. vaginalis genomic DNA. PCR products were cloned into the pCR2.1-TOPO vector by using the TOPO-TA kit (Invitrogen) and sequenced with the Sequenase kit (USB). The PCR product was then labeled by random-primed synthesis with [-³²P]dATP and Klenow fragment and used to screen T. vaginalis cDNA and EcoRI genomic DNA libraries constructed in ZAPII (Stratagene). Plaques of positive clones were isolated, and pBluescript plasmids containing the genomic and cDNA clones were excised according to manufacturer's instructions. Clones with the largest genomic (1.6-kb EcoRI fragment) and cDNA (1,067-bp EcoRI-XhoI fragment) inserts were completely sequenced by the UCLA Sequencing/Genotyping Core Facility.

Expression of recombinant IBP39. To express recombinant IBP39, the coding region of the IBP39 gene was amplified from the cDNA clone by PCR by using the forward primer 5'-CACACCATGGATTCCAATGACCTTGAAGCAAGTTTTACATCTCGTC-3', which contains an NcoI site, and the reverse primer 5'-CCCAGATCTCATTGGAGCGAAAGTAGG-3', which contains a BglII site. The PCR products were cloned into the pCR2.1-TOPO vector (Invitrogen), and then the 1-kb NcoI-BglII fragment was subcloned into the pQE60 expression vector (Qiagen). IBP39 was then expressed as a C-terminal six-histidine fusion protein in Escherichia coli M15(pREP4) cells and purified by nickel column chromatography according to the manufacturer's instructions (Qiagen). To express the recombinant 14.5-kDa polypeptide, the coding region of IBP39 starting with amino acid 1 and ending with amino acid 126 was amplified by PCR by using the same forward primer as the full-length sequence and 5'-CCCAGATCTCATCGGGGAATCGTTTTG-3' as the reverse primer. Expression and purification of the 14.5-kDa polypeptide were performed as described above.

IBP39 antisera and immunoblots. Rabbit antiserum to IBP39 was prepared by Animal Pharm Services, Inc. (Healdsburg, Calif.), by using the recombinant IBP39-6 histidine fusion protein described above. For use in mobility shift assays, IgG from preimmune and anti-IBP39 antisera was purified by using protein A-Sepharose (Pharmacia) as recommended by the manufacturer. Nuclear extracts were prepared in the presence of protease inhibitors (Complete Mini Protease Inhibitor Cocktail [Roche], supplemented with 10 µg of leupeptin and 50 µg of TLCK/ml). Extracts were then separated by SDS-12% PAGE, followed by immunoblotting with anti-IBP39 antisera (used at a 1:4,000 dilution) and detection with protein A-horseradish peroxidase (1:3,500 dilution) and enhanced chemiluminescence reagents (Amersham).

Nucleotide sequence accession number. The sequence of IBP39 has been deposited with GenBank under the accession number AF409099.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

T. vaginalis nuclear extracts contain an Inr-specific binding activity. As a first step toward the identification of T. vaginalis proteins that recognize the Inr, we identified an Inr-specific binding activity in T. vaginalis nuclear extracts (Fig. 1) (17). In an electrophoretic mobility shift assay with a 37-bp DNA probe containing the tandem Inr sequences (TCACTTCACA) from the -succinyl coenzyme A synthetase (SCS) gene (14), a strong binding activity was detected (Fig. 1, lane 2). This protein-DNA complex was abolished when an unlabeled competitor DNA containing a wild-type SCS Inr was included but not when competitor DNAs lacking an Inr element (Inr13) or containing a specific mutation at the +1 positions (Inr3) was included (Fig. 1, lanes 3 to 5). Both of these Inr mutations have been shown previously to abolish Inr activity in vivo (17), indicating that this binding activity is specific for a functional Inr. No DNA-binding activity was observed by using either the coding or the noncoding strand of the 37-bp probe, demonstrating that the protein recognizing this motif binds only double-stranded DNA (data not shown). These data indicate that the binding activity observed in this assay is likely to play a role in Inr recognition in T. vaginalis.

Purification of the Inr-binding protein by DNA affinity chromatography. We have purified the Inr-binding protein by using the strategy outlined in Fig. 2A. Briefly, T. vaginalis crude nuclear extracts were fractionated by ammonium sulfate precipitation, and the fraction containing the Inr-binding activity (80% ammonium sulfate) was subjected to heparin-Sepharose chromatography. The activity, which eluted in 0.3 M KCl, was then applied at a low salt concentration (0.1 M KCl) to a DNA affinity column containing multimers of a 23-bp sequence encompassing the SCS Inr element. The SCS Inr was used because it contains two Inr motifs in tandem (17), thus increasing the number of protein-binding sites on the column. After elution with higher KCl concentrations, fractions containing the binding activity were pooled, diluted to 0.1 M KCl, and applied to a second DNA affinity column. As shown in Fig. 2B, the Inr binding activity was detected by DNase I footprinting assays in fractions eluted from the second DNA affinity column with 0.3 to 0.5 M KCl. Strong protection of the DNA sequence surrounding the SCS Inr was observed. This protection was centered over each of the two Inr motifs on the probe, and the binding was specific for a functional Inr element, since it was not seen with probes containing either a mutation of the +1 position (Inr3) or lacking an Inr sequence (Inr13) (Fig. 2D).

View larger version (78K): [in this window] [in a new window]   FIG. 2.   Purification of the initiator-specific binding activity. (A) Purification scheme for Inr-specific binding activity from T. vaginalis nuclear extracts. Column elutions were performed by using the indicated concentrations of KCl. The fractions that contained binding activity, as assessed by DNase I footprinting assays, are shown in boldface. The sequence of the SCS Inr attached as a multimer to the DNA affinity columns is shown with the Inr motif underlined. (B) DNase I footprinting assay with fractions from the final SCS Inr column. Footprinting reactions were performed with either the 0.2 to 0.7 M KCl SCS Inr column fractions (lanes 1 to 6, respectively) or no protein (lane 7) and the wild-type SCS Inr probe. No binding activity was detected in the flowthrough fraction (not shown). The DNA sequence of the Inr region of the probe in shown in Fig. 1. The location of the Inr element is shown with the two transcription start sites indicated by arrows. (C) Protein profile of fractions from the final SCS Inr column. Protein from each fraction was precipitated with 25% TCA and separated on a 17.5% Tris-glycine SDS-PAGE gel, followed by silver staining. Lane 1, flowthrough fraction; lanes 2 to 9, 0.2 to 0.9 M KCl eluates from the SCS Inr column. Positions of the molecular mass markers are indicated to the left. The 14.5-kDa protein band enriched in the 0.3 to 0.5 M KCl fractions is indicated by the arrow. (D) DNase I footprinting assay with the 0.5 M KCl eluate from the final SCS Inr column. Footprinting reactions were performed with either no protein (lanes 1, 4, 5, 8, 9, and 12) or 5 µl (lanes 2, 6, and 10) or 10 µl (lanes 3, 7, and 11) of the 0.5 M KCl Inr column eluate and the wild-type SCS Inr (lanes 1 to 4), Inr3 (lanes 5 to 8), or Inr13 (lanes 9 to 12) probes. The DNA sequence of the Inr region of the probes is shown in Fig. 1. The location of the Inr element is shown at the left with the two transcription start sites indicated by arrows.

View larger version (56K): [in this window] [in a new window]   FIG. 3.   The 14.5-kDa protein band contains the Inr-binding activity. (A) Protein from the 0.4 M KCl final Inr column fraction was precipitated with 25% TCA and separated in duplicate lanes on a 17.5% Tris-glycine SDS-PAGE gel. Gel slices, ca. 4 mm in length, were excised from one lane for the denaturation-renaturation analysis, while the remainder of the gel was silver stained. The approximate position of each gel slice is indicated to the right of the silver-stained gel, and the molecular mass markers are shown on the left. (B) Gel shift assay with the 15/+15 SCS Inr probe (Fig. 1) and either the 0.4 M KCl fraction (lane 2) or an equal volume of the eluate recovered from the individual gel slices following the denaturation-renaturation protocol (lanes 3 to 9). The Inr-specific binding activity is indicated by the arrow. (C) DNase I footprinting assay performed with either no protein (lanes 1, 6), the 0.4 M KCl fraction (lane 2), or an equal volume of the eluate recovered from gel slices 4, 5, and 6 (lanes 3 to 5, respectively) and the wild-type SCS Inr probe. The location of the Inr element is shown with the two transcription start sites indicated by the arrows.

Cloning and characterization of the Inr-binding protein. To identify the protein responsible for the Inr-binding activity, the 0.4 M KCl eluates from two purification runs were separated by Tris-Tricine SDS-PAGE to obtain increased resolution in the 14.5-kDa region (Fig. 4A) (24). The 14.5-kDa band was excised from the gel and subjected to proteolysis with trypsin. Five of the resulting peptides were then sequenced (Fig. 4B). To isolate the gene encoding the protein, degenerate primers based on the five peptide sequences were used in PCR with T. vaginalis genomic DNA. PCR with primers based on peptides 1 and 3 resulted in a 90-bp fragment that encoded peptides 1, 3, and 4. This PCR product was then used to screen a T. vaginalis genomic library and a clone containing a 1,024-bp open reading frame, encoding a 341-amino-acid, 39.3-kDa protein was obtained (Fig. 4C). This protein, which we have named IBP39 (for initiator binding protein, 39 kDa), has all five peptides obtained by microsequencing (underlined in Fig. 4C) clustered near the N terminus.

View larger version (26K): [in this window] [in a new window]   FIG. 4.   Amino acid sequence of IBP39. (A) Proteins from the 0.4 M KCl Inr column eluate were precipitated with 25% TCA and separated on a 16.5% Tris-Tricine SDS-PAGE gel, followed by silver staining. The arrow indicates the 14.5-kDa band excised for peptide sequence analysis. The positions of the molecular mass standards are shown on the left. (B) Peptide sequences obtained from tryptic fragments of the 14.5-kDa band. (C) Predicted amino acid sequence of IBP39. Amino acid residues matching those of the tryptic peptides derived from the purified IBP39 are underlined. The predicted proteolytic cleavage site at methionine 126 that resulted in the 14.5-kDa N-terminal peptide is marked by an arrow. Four regions enriched in either basic or acidic residues are overlined and indicated by BR1 and BR2 for the basic regions and AR1 and AR2 for the acidic regions.

View larger version (41K): [in this window] [in a new window]   FIG. 5.   IBP39 exists as a 39-kDa polypeptide in vivo. (A) Anti-IBP39 antiserum recognizes full-length recombinant IBP39 and ~25- and 14.5-kDa cleavage products from E. coli. Recombinant IBP39 was expressed as a C-terminal six-histidine fusion protein in E. coli and purified by nickel column chromatography in the absence of protease inhibitors. Nickel column eluates were then separated by SDS-PAGE, blotted, and reacted with the anti-IBP39 antiserum. Lane 1, nickel column eluate from cells expressing the full-length IBP39 six-histidine fusion protein; this protein and a degradation product of ~25 kDa (marked by arrow) were purified. Lane 2, eluate from cells expressing the recombinant 14.5-kDa N-terminal peptide, corresponding to amino acids 1 to 126 of IBP39, with a C-terminal six-histidine tag. E. coli that expressed the 14.5-kDa peptide received the same treatment as those expressing full-length IBP39. (B) Western blot analysis of T. vaginalis nuclear extracts prepared in the presence of protease inhibitors. Lane 1, 25 µg of T. vaginalis freshly prepared nuclear extracts; lane 2, 25 µg of the same extract subjected to one freeze-thaw cycle at 80°C.

IBP39 encodes a novel DNA-binding protein. Extensive database searches using IBP39 failed to reveal significant homology to any known proteins in other organisms (for example, BLAST searches did not find any matches with E values of <0.6). Likewise, searches of protein domain databases revealed no previously identified protein domains. Thus, IBP39 appears to be a novel DNA-binding protein, based on its primary sequence and its lack of known DNA-binding domains. This protein is rich in leucine (8.5%) and lysine (9.1%) residues and has an overall basic charge, with a predicted pI of 8.4 (Fig. 4C). Secondary structure predictions using the GOR4 algorithm, indicated a mostly -helical (42.5%) and random coil (44%) conformation for IBP39 (6). Two regions of IBP39 are predicted to be mostly -helical, from amino acids 1 to 100 and 160 to 250, separated by a 60-amino-acid random coil region. This suggests that the structure of IBP39 may consist of two compact domains separated by an unstructured linker region. It is noteworthy that protease cleavage within this proposed linker region could result in an ~14.5-kDa N-terminal domain, the form of IBP39 purified from T. vaginalis extracts (see arrow, Fig. 4A). IBP39 contains four regions enriched in charged amino acids (Fig. 4C). There are two basic regions near the N terminus, a 17-residue region from amino acids 23 to 39, and a cluster of four lysine residues from amino acids 66 to 70. There are also two acidic regions, a cluster of three glutamic acid residues at amino acids 52 to 54, and a 28-residue region from amino acids 302 to 329. Since the two basic regions are located within the N-terminal 14.5 kDa of IBP39, which likely contains the DNA-binding domain (discussed in detail below), they may be involved in DNA binding. The functional significance of the two acidic regions is unclear at present.

Recombinant IBP39 has Inr-specific binding activity. To confirm that IBP39 has Inr-specific binding activity, the Inr-binding activity of recombinant IBP39 was tested in a DNase I footprinting assay (Fig. 6A). Recombinant IBP39 bound to the SCS Inr element, resulting in a protection pattern similar to that seen with the 0.5 M Inr affinity column fraction (Fig. 6A, compare lanes 2 and 5). As with the purified protein, the binding of recombinant IBP39 was specific for a functional Inr; no binding was seen to a probe containing a mutation at the +1 position (Fig. 6A, lanes 8 to 11). This result shows that recombinant IBP39 has the same binding properties as the purified protein and is therefore most likely responsible for the Inr-binding activity in T. vaginalis extracts.

View larger version (63K): [in this window] [in a new window]   FIG. 6.   IBP39 specifically binds the initiator. (A) DNase I footprinting assays were performed with either the wild-type SCS Inr (lanes 1 to 6) or the Inr3 (lanes 7 to 12) probes and either no protein (lanes 1, 6, 7, and 12), 5 µl of the 0.5 M KCl final Inr column fraction (lanes 2 and 8), 1 ng of recombinant IBP39 (lanes 3, 9), 5 ng of recombinant IBP39 (lanes 4 and 10), or 10 ng of recombinant IBP39 (lanes 5, 11). The location of the Inr element is shown with the two transcription start sites indicated by arrows. (B) Recombinant IBP39 (lane 1) and a 14.5-kDa N-terminal peptide corresponding to amino acids 1 to 125 of IBP39 (lane 2) were expressed in E. coli and separated on an SDS-PAGE gel. (C) Mobility shift assay using the 15/+15 SCS Inr probe and the 39- and 25-kDa proteins recovered from the gel after the denaturation-renaturation protocol. Lane 1, no protein; lanes 2 and 3, either 1 or 5 µl, respectively, of the 39-kDa protein; lanes 4 and 5, either 1 or 5 µl, respectively, of the 25-kDa protein. (D) Mobility shift assay using the 15 or +15 SCS Inr probe and 5 ng (lane 1) or 10 ng (lane 2) of the 14.5-kDa recombinant protein. The Inr-specific binding activities are indicated by the arrows.

View larger version (67K): [in this window] [in a new window]   FIG. 7.   IBP39 recognizes a single functional Inr motif. Mobility shift assays were performed with 10 ng of recombinant IBP39 and wild-type and mutant ferredoxin Inr probes. The sequences of the probes are listed below. (A) Recombinant IBP39 binds the ferredoxin single Inr element. Lanes 1 to 3, wild-type ferredoxin Inr; lanes 4 to 6, +1 A-to-G mutation; lanes 7 to 9, +1 A-to-T mutation; lanes 10 to 12, 5 A-to-G mutation. Lanes 1, 4, 7, and 10 are DNA probe-only controls. (B) Mutation of non-Inr sequences just upstream of the ferredoxin Inr does not affect IBP39 binding. Lanes 1 to 3, 3 and 4 TC-to-CT mutation; lanes 4 to 6, 3 and 4 TC-to-AG mutation; lanes 7 to 9, wild-type ferredoxin Inr. Lanes 1, 4, and 7 are DNA probe-only controls.

View larger version (58K): [in this window] [in a new window]   FIG. 8.   IBP39 is responsible for the initiator-binding activity in T. vaginalis extracts. Mobility shift assays were performed with the 15/+15 SCS (A) or the 15/+15 ferredoxin initiator (B) probes. Lane 1, no protein; lanes 2 to 6, 10 ng of recombinant IBP39; lanes 7 to 11, 12.5 µg of T. vaginalis nuclear extracts. Binding reaction mixtures were preincubated without probe with either no IgG (lanes 2 and 7); 1.5 µg (lanes 3 and 8) or 3.0 µg (lanes 4 and 9) of preimmune IgG; or 1.5 µg (lanes 5 and 10) or 3.0 µg (lanes 6 and 11) of anti-IBP39 IgG. The Inr-specific binding activities are indicated by the arrows.

IBP39 recognizes the Inr motif. Previously, we have shown that the binding of IBP39 depends on the presence of an Inr motif and is blocked by mutations at the +1, 1, and 2 positions (Fig. 1 and 2D) (17). In order to define precisely the nucleotides recognized by IBP39, a series of individual point mutations were generated in the region containing the SCS Inr. At each position, a mutation was introduced to alter both the type of base (pyrimidine versus purine) and the base pair (A-T versus G-C). This panel of mutations was then tested in DNase I footprinting assays with recombinant IBP39 to determine the effect of each mutation on DNA binding (Fig. 9). As expected, the mutations that disrupted IBP39 binding were located within both SCS Inr motifs. In vivo analysis of T. vaginalis Inr function has shown that the most important nucleotides are a Py at 2, a C at 1, an A at +1, and a T or A at +3 (17). Likewise, mutation of these nucleotides disrupted recombinant IBP39 binding. Mutations at 2 (FP3 and -7), 1 (FP4 and -8) and +1 (Inr1 and Inr2) abolished binding by recombinant IBP39 (Fig. 9). Interestingly, mutations at the +3 position had less of an effect on IBP39 binding. While mutation of the 5' +3 T to a G (FP6) greatly reduced IBP39 binding, mutation of the 3' +3 A to a C (FP10) had little effect. This difference correlates with the different effects of these mutations on Inr activity in vivo: a G at the +3 position resulted in a reduction of Inr activity to 15% of wild-type, whereas a C at this position only reduced activity to 55% (17). The in vitro binding of IBP39 thus parallels in vivo Inr activity. Other positions in and around the Inr were also found to contribute to IBP39 binding. Mutation of the +2 position was found to abolish IBP39 binding in vitro (Fig. 9, FP5 and -9). This was somewhat surprising, since mutations at this position in vivo have a modest effect on Inr activity. For example, a +2 G mutation reduces Inr activity to 60% of that of the wild type (17). However, there does appear to be a strong preference for a pyrimidine at this position, since 28 of 33 Inr elements examined contained a T or C at this position (17). Pyrimidine nucleotides downstream of the Inr also appear to contribute to IBP39 binding. Mutation of the two T bases just downstream of the Inr motifs slightly reduced the binding of IBP39, as indicated by the weaker protection seen when 10 ng of protein was used (Fig. 9, FP11 and -12). This likely reflects a preference for pyrimidines surrounding the Inr, since 23 of 33 Inr elements examined have pyrimidines at both of these positions (17). Taken together, these results demonstrate that the sequence requirements for recombinant IBP39 binding are reflected in the nucleotide conservation of the Inr motif.

View larger version (93K): [in this window] [in a new window]   FIG. 9.   Recombinant IBP39 recognizes the initiator motif. DNase I footprinting assays were performed with recombinant IBP39 at either 10 or 25 ng and probes containing either the wild-type SCS Inr or individual point mutations in the Inr region. The wild-type SCS Inr region is shown. At each indicated position, a point mutation was introduced that altered both the type of base (pyrimidine versus purine) and the base pair (A-T versus G-C). The two Inr motifs on the probe are underlined, with the +1 positions indicated by arrows. Nucleotides which, when mutated, disrupted IBP39 binding are shown in boldface.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

This study describes the isolation and cloning of a novel Inr-binding protein, IBP39, from the protist T. vaginalis. IBP39 was isolated by DNA affinity chromatography as a 14.5-kDa polypeptide, derived from the N-terminal region of a 39-kDa protein. IBP39 was shown to have Inr-specific binding activity in both mobility shift and DNase I footprinting assays and is responsible for the binding activity in crude T. vaginalis nuclear extracts. Using a DNase I footprinting assay, we have also shown that the sequence requirements for IBP39 binding correspond to the conserved nucleotides of the T. vaginalis Inr element. Nucleotides that are critical for Inr function in vivo are also necessary for IBP39 binding. These results are consistent with a role for IBP39 in Inr recognition by the T. vaginalis transcription machinery.

Somewhat surprisingly, database searches with the IBP39 amino acid sequence have failed to identify homologues in other organisms. This suggests that IBP39 may represent a novel DNA-binding protein, one not yet identified in other eukaryotes. Since IBP39 is one of the first sequence-specific DNA-binding proteins isolated from a deep-branching eukaryote, it is not surprising that no closely related homologues have been reported from other protists. Furthermore, the great divergence between T. vaginalis and higher eukaryotes may obscure similarity at the protein sequence level to homologous sequence-specific DNA-binding proteins identified in higher eukaryotes. It is also possible that IBP39 homologues are not yet included in databases of eukaryotes whose genomic sequences are incomplete. Alternatively, a true functional homologue may not exist in other eukaryotes. We have also been unable to identify known DNA-binding domains within the IBP39 sequence. Again, this may reflect the great distance between T. vaginalis and higher eukaryotes, such that similarities at the sequence level are unclear, or IBP39 may contain a unique DNA-binding structure, one not yet identified in other eukaryotes.

Given the structural and functional similarities between their Inrs, trichomonads and metazoans may share a common mechanism for Inr recognition. In metazoans, TFIID is the main general transcription factor involved in core promoter recognition (reviewed in reference 26). The TFIID subunits TAF_II150 and TAF_II250 have been shown to be necessary for Inr-mediated transcription (12, 32), and a recent study has demonstrated direct Inr binding by a TAF_II150-TAF_II250 heterodimer (3). These results strongly suggest that these two TAFs are responsible for Inr recognition by TFIID. As noted above, similarity between IBP39 and either TAF_II150 or TAF_II250 is not apparent at the sequence level. However, since the DNA-binding domains of these two TAFs have yet to be identified, IBP39 may have limited homology to the Inr recognition domains of TAF_II150 and/or TAF_II250 that a computer search might miss. Alternatively, T. vaginalis may contain other, currently unidentified, Inr-binding proteins that may be homologues of various TFIID subunits. Another possibility is that IBP39 may represent a homologue of an as-yet-unidentified metazoan Inr-binding protein. Two studies have shown that TFIID is not sufficient to direct Inr-mediated transcription in a highly purified in vitro transcription system (12, 19). These studies identified several cofactors which were found to be required for Inr-mediated transcription: the CIF fraction, which contains the human homologue of TAF_II150, among other proteins (12), and the TIC cofactors, which do not contain TAF_II150 or the previously identified Inr-binding proteins YY1, USF, and TFII-I (19). Therefore, IBP39 may be homologous to a component of the CIF or TIC fraction which has yet to be characterized. Since all T. vaginalis core promoters appear to consist solely of an Inr, the trichomonad transcription machinery presumably does not have to distinguish between multiple core promoter elements, as does the metazoan transcription apparatus. Therefore, it is possible is that trichomonads have evolved a unique mechanism for Inr recognition, which is distinct from that of higher eukaryotes. Alternatively, IBP39 may be found to function like a TAF without sharing sequence homology with other eukaryotic TAFs.

The conservation of both structure and function of the Inr between trichomonads and metazoans strongly suggests that this core promoter element arose early during eukaryotic evolution. Interestingly, a sequence similar to the Inr is required for transcription of a gene in the protist Toxoplasma gondii, although it was not determined whether this element actually functions as an Inr (20). In addition, conserved sequences surround the transcription start sites of genes in the protists Entamoeba histolytica (22) and Giardia lamblia (30, 33). These elements do not share any sequence similarity with either each other or the trichomonad Inr. However, they may play a similar role during transcription initiation. This is the case for E. histolytica, since its Inr-like element has been shown to direct transcription start sites in vivo (25). Therefore, it is possible that these protists share a common mechanism for recognition and selection of transcription start sites with trichomonads. The identification of IBP39 represents the first step in characterizing the transcription machinery of T. vaginalis. Further work on the function of IBP39 and the identification of additional components of the trichomonad transcription apparatus should lead to a greater understanding of the evolution of gene regulation and the mechanisms of Inr function in eukaryotes.