Transcription initiation from a poly(dA) tract - NCBI - NIH

Download PDF
0 downloads 0 Views 416KB Size Report
Comment
INTRODUCTION. Transcription initiation is among the most highly regulated steps .... that the poly(dA) tract must be continuous for binding to occur. Consistent ...
 1998 Oxford University Press

Nucleic Acids Research, 1998, Vol. 26, No. 4

1019–1025

Transcription initiation from a poly(dA) tract Daniel H. Shain*, Mauricio X. Zuber§ and Toomas Neuman1,+ Department of Biochemistry and Molecular Biology and 1Department of Anatomy and Neurobiology, Colorado State University, Fort Collins, CO 80523, USA Received October 9, 1997; Revised and Accepted December 11, 1997

ABSTRACT The mammalian ME1 gene encodes a non-tissuespecific, helix–loop–helix transcription factor that is enriched in morphogenetically active regions during development. Regulation of mouse ME1 gene expression is controlled by a novel initiator (ME1 Inr) that promotes transcription from the center of a 13 bp poly(dA) tract. We show here that the ME1 Inr autonomously directs initiation from the poly(dA) tract both in vitro and in vivo. This transcription was dependent upon two protein complexes; MBPα, which associated directly with the poly(dA) tract, and MBPβ, which introduced an ∼60 bend immediately downstream of the poly(dA) tract. The MBPα and MBPβ binding sites were strikingly conserved in homologous DNA from several mammalian species and the frog Xenopus laevis. These results suggest that the ME1 Inr constitutes a robust nucleation site that promotes transcription initiation in the absence of conventional promoter elements. INTRODUCTION Transcription initiation is among the most highly regulated steps in the process of gene expression. Many eukaryotic promoters utilize a TATA box to recruit and position components of the Pol II transcriptional machinery at their transcription start site. Those promoters lacking a TATA box appear to recruit a similar complement of proteins by utilizing sequences proximal to their initiation site (1–3). To date, many TATA-less promoters contain sequences related to the terminal deoxynucleotidyltransferase initiator (TdT Inr), a 17 bp element (5′-GCCCTCATTCTGGAGAC-3′) that is sufficient for initiating accurate basal transcription (4–6). One exception to this paradigm is the TATA-less ME1 promoter, which regulates expression of the mammalian ME1 gene (7), a helix–loop–helix transcription factor associated with morphogenetically active regions during development (8,9). Transcription initiation from the ME1 gene occurs at the center of a 13 bp poly(dA) tract which is flanked upstream by the 16 bp palindrome 5′-GCTGAGGCGCCTCAGC-3′ and downstream by the 9 bp inverted repeat 5′-GTCCGCCTG-3′ (7). Collectively

these elements comprise the ME1 Inr. Our aim in this study was to determine the role of the poly(dA) tract and its flanking sequences in ME1 Inr-dependent transcription initiation. We show here that the ME1 Inr autonomously promotes transcription in vitro and in vivo and this transcription is dependent upon two protein complexes, MBPα and MBPβ. These proteins interact directly with the ME1 Inr and are likely to play pivotal roles in regulating ME1 gene expression. MATERIALS AND METHODS Electrophoretic mobility shift assays (EMSA) Extraction buffer contained 20 mM HEPES, pH 7.8, 450 mM NaCl, 0.2 mM EDTA, 0.5 mM DTT, 25% glycerol and a mixture of protease inhibitors: PMSF (0.5 mM), leupeptin (0.5 µg/ml), pepstatin (0.7 µg/ml), aprotinin (1 µg/ml) and bestatin (40 µg/ml). All steps were performed at 4C. After sonication extracts were cleared by microcentrifugation for 5 min. Binding conditions for EMSA were 10 mM HEPES, pH 7.8, 1 mM spermidine, 5 mM MgCl2, 50 mM KCl, 0.5 mM DTT, 9% glycerol, 0.8 mg poly(dI·dC), 100 000 c.p.m. labeled oligonucleotide (cold competitor DNA was used at a 200-fold molar excess) and 10 µg nuclear extract. Following incubation for 15 min at 37C reactions were applied to a polyacrylamide gel, electrophoresed to separate DNA–protein complexes and analyzed by autoradiography. Methylation interference analysis End-labeled DNA fragments (coding, XhoI–SalI; non-coding, XhoI–XbaI; see Fig. 3A) were partially methylated at guanidine residues using dimethylsulfate as described (10). Following EMSA, free and complexed DNA fragments were excised and purified by electroelution, followed by ethanol precipitation. DNAs were then suspended in 1 M piperidine, incubated for 30 min at 90C and electrophoresed through a 10% polyacrylamide gel containing 8 M urea. Interference patterns caused by methylation were analyzed by autoradiography. Gel permutation assay The ME1/pBend2 vector containing the ME1 Inr insert was digested with the appropriate restriction endonucleases (shown in Fig. 3). DNA fragments were then filled in with [α-32P]dCTP as

*To whom correspondence should be addressed at present address: Department of Molecular and Cell Biology, 385 LSA, University of California, Berkeley, CA 94720-3200, USA. Tel: +1 510 642 2697; Fax: +1 510 643 6791; Email: [email protected] +Present

address: Department of Surgery, Cedars-Sinai Medical Center, Davis Building, Room 4018, 8700 Beverly Boulevard, Los Angeles, CA 90048, USA

§Deceased

1020 Nucleic Acids Research, 1998, Vol. 26, No. 4

Figure 1. EMSA analysis of the ME1 Inr and truncated oligonucleotides in PCC7, HT-4 and NIH 3T3 cells. (A) A synthetic oligonucleotide comprising the ME1 Inr and flanking restriction endonuclease sites is shown. The 16 bp palindrome upstream from the poly(dA) tract and the downstream 9 bp inverted repeat are indicated. The arrow designates the transcription start site (in all figures) (7). Positions of the ME1 Inr, 5′ and 3′ oligonucleotides are shown below. (B) EMSA analysis of the ME1 Inr using PCC7, HT-4 and NIH 3T3 nuclear extracts. The ME1 Inr oligonucleotide was [α-32P]dCTP labeled and was present in each lane. Cold competitor DNAs in each panel were: lane 1, none; lane 2, random 26mer; lane 3, ME1 Inr; lane 4, 3′ oligo; lane 5, 5′ oligo; lane 6, 5′ + 3′ oligos. Competitor DNA was used at a 200-fold molar excess (in all figures). The predominant protein complexes that bound the ME1 Inr, designated MBPα and MBPβ, are indicated.

described (10), electrophoresed on a 5% polyacrylamide gel and visualized by autoradiography. Relative mobilities were measured from a point originating at the loading well.

Figure 2. Methylation interference pattern of ME1 Inr protein binding complexes. Alternating free and bound lanes are shown in the coding and non-coding strands of the ME1 Inr, respectively. Bound lanes were incubated with PCC7 nuclear extracts. Arrowheads indicate regions of interference in bound lanes; their corresponding positions in the ME1 Inr are shown below.

Cloning genomic DNA Genomic clones were obtained by screening human (Stratagene) and X.laevis (Stratagene) genomic libraries, respectively, at low stringency with a DNA fragment containing the ME1 Inr (–45 to +121 bp in the ME1 promoter; 7). Conditions of hybridization were 25% formamide, 5× SSPE, 5% SDS at 42C; washes were in 0.5× SSC at 55C. RESULTS

Transcription assays

Identification of protein binding domains in the ME1 Inr

HeLa Scribe (Promega) nuclear extracts were used for all in vitro transcription assays according to the manufacturer’s specifications. Total RNA purification and primer extension analyses were performed as described (7).

To determine the protein binding sites in the ME1 Inr, electrophoretic mobility shift assays (EMSA) were performed using the ME1 Inr oligonucleotides shown in Figure 1. Nuclear extracts were prepared from three cell types: PCC7 (mouse teratocarcinoma), HT-4 (rat immortalized hippocampal) and NIH 3T3 (fibroblast). Levels of endogenous ME1 mRNA are ∼10-fold greater in PCC7 and HT-4 cells when compared with NIH 3T3 cells (7). Two ME1 Inr binding complexes, MBPα (ME1 binding protein α) and MBPβ, became the focus of our study due to their specific interactions with the ME1 Inr. MBPα was observed in PCC7 and HT-4 cells and was competed for by the ME1 Inr oligonucleotide (Fig. 1). The 5′ and 3′ oligos, either independently or in combination, were not effective competitors of MBPα, suggesting

Cell culture PCC7, HT-4 and NIH 3T3 cells were maintained in Dulbecco’s modified Eagle’s medium (Gibco) supplemented with 10% fetal calf serum, 2 mM L-glutamine, 100 U/ml penicillin G and 100 mg/ml streptomycin. Transient transfections were performed by calcium phosphate co-precipitation (11); transfection efficiencies were determined by β-galactosidase assay as described (12).

1021 Nucleic Acids Acids Research, Research,1994, 1998,Vol. Vol.22, 26,No. No.14 Nucleic

1021

Figure 3. Protein-induced bending of the ME1 Inr. (A) Vector map of the ME1 Inr oligonucleotide cloned into the XbaI and SalI sites of pBend2 (ME1/pBend2). Restriction endonuclease sites used for the bending analysis are indicated; numbering initiates at the EcoRI site. (B) Static conformation of the ME1 Inr in the absence of nuclear extract. The left panel shows restriction fragments (indicated by arrows) from (A) electrophoresed on a native 5% polyacrylamide gel (the upper band is vector DNA). The EcoRI–XbaI fragment was included as a negative control since it lacks the ME1 Inr sequence. The right panel plots the relative mobility (EcoRI–XbaI fragment designated 1.0) against the distance from the EcoRI site. The linearity of this plot indicates the lack of intrinsic curvature in the ME1 Inr sequence. (C) Bending analysis of the ME1 Inr in the presence of PCC7, HT-4 and NIH 3T3 nuclear extracts. Conditions were as in (B). The absence of protein binding in the EcoRI–XbaI lane indicates that protein complexes did not recognize vector DNA. MBPα and MBPβ protein complexes, which were not resolved by this analysis, bent the ME1 Inr ∼60 according to the equation µM/µE = cosα/2, where µM is the distance of bound probe, µE is the distance of unbound probe and α is the angle of bending (15). The bend mapped to the center of the inverted repeat 5′-GTCCGCCTG-3′ [shown in (A) as a double underlined sequence].

that the poly(dA) tract must be continuous for binding to occur. Consistent with this observation, an oligo mimicking the ME1 poly(dA) tract was sufficient to compete for MBPα binding (see below). MBPβ was detected in all three cell types and was competed for by the ME1 Inr and 3′ oligonucleotides (Fig. 1). Methylation interference analysis supported the notion that MBPβ bound the ME1 Inr within the downstream inverted repeat 5′-GTCCGCCTG-3′ (Fig. 2).

binding site; (ii) the observed band was competed for by the 3′ ME1 Inr oligonucleotide (Fig. 4); (iii) bending was observed in NIH 3T3 cells which contain only MBPβ (and not MBPα). However, the MBPα complex is likely to contain components of MBPβ, since these bands co-migrated in our bending assays (these complexes were resolved at lower gel concentrations).

Bending of the ME1 Inr induced by MBPβ

To examine the putative role of the MBPα and MBPβ protein complexes in transcriptional initiation, in vitro transcription assays were performed using HeLa Scribe nuclear extracts (Promega) in the presence of competing ME1 Inr oligonucleotides (Fig. 5). The counterparts of MBPα and MBPβ in HeLa cells were identified on the basis of which oligonucleotides competed with their binding (Fig. 5A; compare with Fig. 1). An oligonucleotide mimicking the 13 bp poly(dA) tract in the ME1 Inr competed with MBPα in HeLa Scribe (and also in PCC7 and HT-4 extracts; not shown), suggesting that MBPα is recruited by the poly(dA) tract. This is consistent with the lack of MBPα competition when using 5′ and 3′ oligos, which lack a complete poly(dA) tract (see Fig. 1). The TATA-containing oligo failed to compete for ME1 Inr-specific protein complexes in

The presence of a poly(dA) tract within the ME1 Inr prompted us to examine its putative role in DNA bending, since related sequences can produce topologically complex shapes in DNA regulatory domains (13). To test the intrinsic bending potential of the ME1 Inr, it was cloned into the pBend2 vector (14) and subjected to gel permeation analysis (15). By itself the ME1 Inr contained no detectable intrinsic curvature (Fig. 3B). However, upon incubation with PCC7, HT-4 or NIH 3T3 nuclear extracts an ∼60 bend resulted which mapped to the center of the inverted repeat sequence (Fig. 3C). Bending was likely due to MBPβ because: (i) the bend center was coincident with the MBPβ

The ME1 Inr autonomously promotes transcription initiation

1022 Nucleic Acids Research, 1998, Vol. 26, No. 4

Figure 4. EMSA competition analysis of ME1 Inr bending complexes. The NheI and SmaI restriction endonuclease fragments from ME1/pBend2 were incubated with PCC7 nuclear extracts and competed with ME1 Inr-specific oligonucleotides. Cold competitor DNAs in each panel were: lane 1, none; lane 2, ME1 Inr; lane 3, 3′ oligo; lane 4, 5′ oligo.

HeLa Scribe, implying that components of the Pol II transcription machinery (i.e. TFIID) do not interact directly with the ME1 Inr. Transcription assays in vitro with HeLa Scribe demonstrated that a –240 ME1–CAT promoter construct (7) could direct accurate transcription from the center of the poly(dA) tract (Fig. 5B). The ME1 Inr, 3′, poly(dA) and TATA-containing oligonucleotides were effective competitors of this transcription, while the 5′ and non-specific oligonucleotides were not. Thus MBPα and MBPβ [competed by the poly(dA) and 3′ oligonucleotides respectively] were required for full transcriptional activity of the ME1 Inr. It is also noteworthy that the TATA-containing oligonucleotide abrogated ME1 Inr-dependent transcription. Thus while components of the general transcriptional machinery do not interact directly with the ME1 Inr, they appear to be required in steps subsequent to MBPα and MBPβ binding. In a complementary set of experiments it was determined that the ME1 Inr in pBend2 (ME1 Inr/pBend2) could independently direct accurate transcription from the poly(dA) tract (Fig. 5C). To exclude the possibility that upstream sequences in pBend2 contributed to this transcription, the ME1 Inr was linearized at its 5′-end with the restriction endonuclease XhoI. In concordance with –240 ME1–CAT (Fig. 5B), transcription was reduced by the ME1 Inr, 3′, poly(dA) and TATA-containing oligonucleotides but not the 5′ or non-specific oligos. Inclusion of a sense RNA strand in primer extension analysis indicated that reverse transcriptase could read through the poly(dA) tract and, therefore, the possibility of artifactual stalling within this region can be negated. Taken together, the results suggest that the ME1 Inr appears to encode sufficient information for recruiting and positioning the appropriate transcriptional machinery for accurate initiation in vitro. To determine whether the ME1 Inr could function autonomously in vivo, we next transfected the ME1 Inr/pBend2 construct into PCC7, HT-4, HeLa and NIH 3T3 cell lines. In all cells the ME1 Inr directed transcription initiation from within the poly(dA) tract (Fig. 5D). Relative levels of ME1 Inr-dependent transcription were determined by co-transfecting the β-galactosidase (β-gal) gene [driven by the Rous sarcoma virus (RSV) promoter]. β-Gal assays on cell culture plates revealed that approximately the same number of cells/plate were transfected in each cell type (PCC7, 17 ± 7; HT-4, 25 ± 10; HeLa, 42 ± 11; NIH 3T3, 29 ± 11; numbers represent cells transfected per 100 cells). β-Gal assays of cell

Figure 5. The ME1 Inr directs accurate transcription initiation in vitro and in vivo. (A) EMSA analysis of the ME1 Inr in HeLa Scribe nuclear extracts (Promega). MBPα and MBPβ counterparts were identified on the basis of their respective competitor DNAs (compare with PCC7, HT-4 and NIH 3T3 extracts, Fig. 1). An oligonucleotide mimicking the ME1 poly(dA) tract (5′-GGAAAAAAAAAAAAATG-3′) competed with the MBPα complex (lane 6), while a TATA-containing oligonucleotide (5′-GTCAGAGTCCCGGCGTATAAATAGAGGTGACTA-3′) did not affect the banding pattern (lane 7). Competitor DNAs in each lane were: lane 1, none; lane 2, random 26mer; lane 3, ME1 Inr oligo; lane 4, 5′ oligo; lane 5, 3′ oligo; lane 6, poly(dA) oligo; lane 7, TATA-containing oligo. (B) HeLa Scribe in vitro transcription from a –240 ME1–CAT construct (21). Ladder shows –240 ME1–CAT sequenced with the CAT-specific oligonucleotide (5′-GCAACTGACTGAAATGCCTC-3′), which was used for subsequent primer extension analysis. Competitor DNAs in each lane are as in (A). (C) HeLa Scribe in vitro transcription from the ME1 Inr in pBend2 (ME1 Inr/pBend2), linearized at the 5′-end with the restriction endonuclease XbaI. The ladder is from ME1 Inr/pBend2 sequenced with the SP6 primer, which was used in subsequent primer extension analysis. Lanes 1–7 are as described in (A). Lane 8 contained sense RNA transcribed from the T7 promoter, which was located upstream of the ME1 Inr. (D) Transcription from the ME1 Inr in vivo. The ME1 Inr/pBend plasmid (5 µg) was transfected into PCC7 (lane 2), HT-4 (lane 4), HeLa (lane 6) and NIH 3T3 cells (lane 8) respectively and analyzed for transcriptional activity as described in (C). Mock transfection followed by primer extension analysis was performed in each cell type (lane 1, PCC7; lane 3, HT-4; lane 5, HeLa; lane 7, NIH 3T3). Lane 9 contained a sense RNA template transcribed from the T7 promoter, as described in (C).

extracts suggested that the transfection efficiency of NIH 3T3 cells (25 ± 2) was ∼1.5-fold greater than PCC7 (13.5 ± 1.5) or HT-4 (15.3 ± 0.4). Normalized values obtained by phosphorimager quantitation of primer extension products (Fig. 5D) suggested that the ME1 Inr/pBend2 plasmid was about an order of

1023 Nucleic Acids Acids Research, Research,1994, 1998,Vol. Vol.22, 26,No. No.14 Nucleic

Figure 6. EMSA analysis of ME1 Inr binding complexes in adult mouse tissues. Nuclear extracts were prepared from the tissues indicated and incubated with [α-32P]dCTP-labeled ME1 Inr oligonucleotide. The MBPβ complex was present in all tissues, while MBPα was detected at low levels in liver, lung and spleen.

1023

Figure 7. Ion dependence of ME1 Inr binding complexes. PCC7 nuclear extracts were incubated with labeled ME1 Inr oligonucleotide in the presence of varying concentrations of Na+, Mg2+ and Ca2+. Ion concentrations in each lane were: lane 1, 40 mM; lane 2, 55 mM; lane 3, 70 mM; lane 4, 150 mM; lane 5, 0 mM; lane 6, 0.1 mM; lane 7, 0.5 mM; lane 8, 2 mM; lane 9, 0 mM; lane 10, 0.1 mM; lane 11, 0.5 mM; lane 12, 2 mM. Positions of the MBPα and MBPβ complexes are indicated.

magnitude more active in PCC7 and HT-4 when compared with NIH 3T3 cells. Properties of ME1 Inr binding complexes To determine the distribution of the MBPα and MBPβ complexes in the adult mouse, nuclear extracts were prepared from various tissues and subjected to EMSA analysis (Fig. 6). The MBPβ complex was present in all tissues examined, albeit at low levels in the brain stem and heart. Liver, lung and spleen were the only tissues that showed detectable levels of MBPα. Several faster migrating species were observed in liver, lung, kidney and spleen. The sensitivity of ME1 Inr binding complexes to monovalent and divalent cations is shown in Figure 7. Binding of the MBPβ complex was enhanced dramatically with increasing concentrations of Mg2+ or Ca2+, while modest increases were observed with MBPα and faster migrating species. MBPβ binding was significantly reduced with increasing concentrations of Na+ while other complexes were relatively insensitive (below 70 mM). The kinetics of MBPβ binding were determined by titrating increasing amounts of ME1 Inr oligonucleotide into the EMSA assay (Fig. 8). Maximum binding reached a plateau at ∼800 pM, which translated into a Kd of 0.89 nM (Fig. 8B). Similar kinetics were observed with MBPα (not shown), suggesting that the MBPα and MBPβ complexes bind tightly to their target DNAs. Conservation of the ME1 Inr during vertebrate evolution To determine whether the MBPα and MBPβ binding sites have been conserved in other vertebrates, we cloned homologous ME1 Inr DNA from human and the clawed frog X.laevis (Fig. 9). Fortuitously, two related clones were obtained from X.laevis, which likely represent duplicated alleles from Xenopus tetraploidization ∼30 million years ago (16). All the sequences encoded a poly(dA) tract, the shortest of which contained seven consecutive adenosine

Figure 8. Kinetics of MBPβ binding. (A) PCC7 nuclear extracts were incubated with increasing concentrations of labeled ME1 Inr oligonucleotide. DNA bound by MBPβ was quantitated by a PhosphorImager and plotted against [total DNA]. (B) A Scatchard plot indicated the Kd of MBPβ binding to be ∼0.89 nM.

1024 Nucleic Acids Research, 1998, Vol. 26, No. 4

Figure 9. Homologous ME1 Inr DNA from distant vertebrates. The mouse ME1 Inr and proximal nucleotide sequences are shown in their entirety, while differences in counterpart DNA from human (HEB) and X.laevis (XE1.1 and XE1.2) sequences are indicated. Dashes represent gaps in the sequence and dots indicate sequence identity.

residues (XE1.2). Although the length of the poly(dA) tract varied between species, preliminary EMSA and RNase protection analyses suggest that a relatively short stretch of dA residues within the ME1 Inr (i.e. 8 bp, HEB) is sufficient for MBPα and MBPβ binding and accurate transcription initiation from the poly(dA) tract (not shown). The region downstream of the poly(dA) tract, corresponding to the MBPβ binding site (5′-TGCCGCCGT-3′), was strikingly conserved in all species. DISCUSSION This study demonstrates the autonomy of the ME1 Inr as a transcriptional activator and the putative roles of ME1 Inr binding proteins MBPα and MBPβ. A model depicting one possible mechanism of transcription initiation by the ME1 Inr is shown in Figure 10. MBPα is recruited by the 13 bp poly(dA) tract (–6 to +7) and is likely to associate directly with MBPβ. MBPβ independently, or in combination with MBPα, introduces an ∼60 bend within the inverted repeat (+9 to +17). Upon binding its full complement of proteins the ME1 Inr provides a robust substrate for the Pol II machinery, catalyzing transcription initiation consistently from the center of the 13 bp poly(dA) tract in vitro and in vivo. Since a TATA-containing oligonucleotide does not bind the ME1 Inr directly (Fig. 5A) but competes with ME1 Inr transcriptional activity in vitro (Fig. 5B and C), it appears as though Pol II is recruited to the ME1 Inr by components of TFIID (which recognizes the TATA consensus sequence) and not by the ME1 Inr or its associated proteins (e.g. MBPα and MBPβ). In this perspective, ME1 Inr-dependent transcription appears related to the mechanism of TdT-dependent transcription, which also requires TFIID for activity (17,18). This is in contrast to the YY1 intiator element, which directs accurate and efficient transcriptional initiation by Pol II in the absence of TBP (6). Interestingly, the levels of ME1 Inr-dependent transcription were significantly reduced in NIH 3T3 cells (Fig. 5D), consistent with the relatively low levels of endogenous ME1 transcripts observed in these cells (7). Since NIH 3T3 cells lack detectable levels of MBPα [which binds the poly(dA) tract], it appears that MBPα may play a role in regulating cell type-specific transcription in other cells through its interaction with the poly(dA) tract. Whether the presence or absence of MBPα can account for the dramatically different levels of ME1 transcription observed between PCC7, HT-4 and NIH 3T3 cells remains to be determined. It is worth noting, however, that ME1 gene expression is down-regulated in the adult mouse (9) and this is consistent with the relatively low levels of MBPα observed in the adult tissues examined here (see Fig. 6). Finally, it remains to be

Figure 10. Model of ME1 Inr transcription initiation. MBPα and MBPβ protein complexes bind the poly(dA) tract and the inverted repeat respectively, introducing an ∼60 bend downstream of the transcription start site. (It should be emphasized that the bend direction is only hypothesized, since circular permutation analysis does not provide directional information.) Components of the basal transcriptional machinery are then recruited to this complex, resulting in accurate transcription from the center of the ME1 poly(dA) tract. The 16 bp palindromic sequence (5′-GCTGAGGCGCCTCAGC-3′) is shown in black, the 13 bp poly(dA) tract in white and the 9 bp inverted repeat (5′-GTCCGCCTG-3′) in gray.

established whether MBPα represents a mammalian homolog of Datin, an oligo(dA)·oligo(dT) binding protein that acts as a transcriptional repressor in yeast (19,20). Several features of the ME1 Inr were conserved in diverse vertebrate species ranging from human to frog, strongly suggesting that an ME1 Inr-like element functioned in an ancestral vertebrate >350 million years ago (the presumptive divergence between amphibians and the lineage leading to mammals; 21). Particularly striking was the region downstream of the poly(dA) tract (and including the MBPβ binding site), which maintained ∼80% sequence identity between human and frog. This degree of conservation exceeds that observed in the HLH domain of these genes (XE1 and HEB respectively), a region that is highly conserved among all vertebrate taxa that have been examined (22). The 5′ palindromic sequence of the ME1 Inr was highly conserved in human but not in the X.laevis clones, suggesting that this site was acquired more recently during vertebrate evolution. However, since the palindromic sequence does not recruit a detectable protein complex by our analysis nor does it interfere with ME1 Inr-dependent transcription in vitro, the role of the 5′ palindrome, if any, remains unclear. Perhaps the most intriguing feature of the ME1 Inr is the presence of a 13 bp poly(dA) tract, which serves as a template for transcription initiation in vitro and in vivo. Although adenine tracts have been associated with numerous aspects of transcription, including mRNA polyadenylation [i.e. poly(A) tail addition; 23], nucleosome positioning (24,25) and DNA bending (26), we believe

1025 Nucleic Acids Acids Research, Research,1994, 1998,Vol. Vol.22, 26,No. No.14 Nucleic this is the first example in which a poly(dA) tract plays a direct role in transcription initiation. The function of the poly(dA) tract and the MBPβ-induced bend in this process could be several fold. DNA bending has been associated with multiple steps of transcription initiation, ranging from juxtapositioning of transcription factors in the proximal promoter (27–30) to providing torsional energy for driving formation of an open complex (31,32). In some cases DNA bending alone has been shown to enhance transcription by 100- to 200-fold (33,34). The proximity of the ME1 poly(dA) tract (–6 to +7) to the bend center (+13) suggests that these two motifs may interact directly; for example, torsional energy from the bending site may be sufficient to melt the relatively weak base pairing of the poly(dA) tract, thus allowing Pol II entry. Consistent with this notion, Boettcher et al. (35) have identified multiple poly(dA) tracts in the ε-globin promoter that could provide entry sites for Pol II. In addition, numerous intergenic poly(dA) tracts have been identified in yeast that influence transcription of adjacent genes (19,20). Thus poly(dA)-dependent transcription may not be restricted to the ME1 promoter; rather, it may represent an alternative mechanism for recruiting components of the Pol II transcriptional apparatus in the absence of a TATA box, a TdT-like sequence or other conventional promoter elements. ACKNOWLEDGEMENTS We thank Jakyoung Yoo for helpful comments and suggestions. This work was supported by March of Dimes Basil O’Conner Starter Grant Fy94-0835, NSF grant BNS-91558411 and NIH grant NS33804 to M.X.Z. REFERENCES 1 Wiley,S.R., Kraus,R.J. and Mertz,J.E. (1992) Proc. Natl. Acad. Sci. USA, 89, 5814–5818. 2 Zenzie-Gregory,B., O’Shea-Greenfield,A. and Smale,S.T. (1992) J. Biol. Chem., 267, 2823–2830. 3 Zhou,Q., Lieberman,P.M., Boyer,T.G. and Berk,A.J. (1992) Genes Dev., 6, 1964–1973. 4 Martinez,E., Chiang,C.M., Ge,H. and Roeder,R.G. (1994) EMBO J., 13, 3115–3126. 5 Smale,S.T. and Baltimore,D. (1989) Cell, 57, 103–113.

1025

6 Usheva,A. and Shenk,T. (1994) Cell, 76, 1115–1121. 7 Shain,D.H., Neuman,T. and Zuber,M.X. (1995) Nucleic Acids Res., 23, 1696–1703. 8 Klein,E.S., Simmons,D.M., Swanson,L.W. and Rosenfeld,M.G. (1993) Genes Dev., 7, 55–71. 9 Neuman,T., Keen,A., Knapik,E., Shain,D., Ross,M., Nornes,H.O. and Zuber,M.X. (1993) Eur. J. Neurosci., 5, 311–318. 10 Ausubel,F.M., Brent,R., Kingston,R.E., Moore,D.D., Seidman,J.G., Smith,J.A. and Struhl,K. (1994) Current Protocols in Molecular Biology. John Wiley & Sons, New York, NY. 11 Graham,F.L. and van der Eb,A.J. (1973) Virology, 52, 456–467. 12 Sambrook,J., Fritsch,E.F. and Maniatis,T. (1989) Molecular Cloning: A Laboratory Manual, 2nd Edn. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY. 13 Nelson,H.C.M., Finch,J.T., Luisi,B.F. and Klug,A. (1987) Nature, 330, 221–226. 14 Kim,J., Zwieb,C., Wu,C. and Adhya,S. (1989) Gene, 85, 15–23. 15 Wu,H.M. and Crothers,D.M. (1984) Nature, 308, 509–513. 16 Bisbee,C.A., Baker,M.A., Wilson,A.C., Hadji-azimi,I. and Fischberg,M. (1977) Science, 195, 785–787. 17 Smale,S.T., Schmidt,M.C., Berk,A.J. and Baltimore,D. (1990) Proc. Natl. Acad. Sci. USA, 87, 4509–4513. 18 Pugh,B.F. and Tjian,R. (1990) Cell, 61, 1187–1197. 19 Winter,E. and Varshavsky,A. (1989) EMBO J., 8, 1867–1878. 20 Iyer,V. and Struhl,K. (1995) EMBO J., 14, 2570–2579. 21 Strickberger,M.W. (1990) Evolution. Jones and Bartlett, Boston, MA. 22 Shain,D.H., Neuman,T. and Zuber,M.X. (1997) Genetics, 146, 345–353. 23 Proudfoot,N.J. and Brownlee,G.G. (1976) Nature, 263, 211–214. 24 Kunkel,G.R. and Martinson,H.G. (1981) Nucleic Acids Res., 9, 6869–6888. 25 Prunell,A. (1982) EMBO J., 1, 173–179. 26 Zinkel,S.S. and Crothers,D.M. (1987) Nature, 328, 178–181. 27 Becker,J.C., Nikroo,A., Brabletz,T. and Reisfeld,R.A. (1995) Proc. Natl. Acad. Sci. USA, 92, 9727–9731. 28 Huber,H.E., Goodhart,P.J. and Huang,P.S. (1994) J. Biol. Chem., 269, 6999–7005. 29 Nardulli,A.M., Grobner,C. and Cotter,D. (1995) Mol. Endocrinol., 9, 1064–1075. 30 Sharrocks,A.D. and Shore,P. (1995) Nucleic Acids Res., 23, 2442–2449. 31 Horikoshi,M., Bertuccioli,C., Takada,R., Wang,J., Yamamoto,T. and Roeder,R.G. (1992) Proc. Natl. Acad. Sci. USA, 89, 1060–1064. 32 Perez-Martin,J., Rojo,F. and de Lorenzo,V. (1994) Microbiol. Rev., 58, 268–290. 33 Kahn,J.D. and Crothers,D.M. (1993) Cold Spring Harbor Symp. Quant. Biol., LVIII, 115–122. 34 Parvin,J.D., McCormick,R.J., Sharp,P.A. and Fisher,D.E. (1995) Nature, 373, 724–727. 35 Boettcher,B. (1990) J. Theor. Biol., 146, 333–339.