Transcription initiation from a poly(dA) tractDaniel 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-tissue-specific, 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; MBP[alpha], which associated directly with the poly(dA) tract, and MBP[beta], which introduced an ~60° bend immediately downstream of the poly(dA) tract. The MBP[alpha] and MBP[beta] 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.
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, MBP[alpha] and MBP[beta]. These proteins interact directly with the ME1 Inr and are likely to play pivotal roles in regulating ME1 gene expression.
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 4°C. 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 37°C reactions were applied to a polyacrylamide gel, electrophoresed to separate DNA-protein complexes and analyzed by autoradiography.
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 90°C and electrophoresed through a 10% polyacrylamide gel containing 8 M urea. Interference patterns caused by methylation were analyzed by autoradiography.
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 [[alpha]-32P]dCTP as described (10), electrophoresed on a 5% polyacrylamide gel and visualized by autoradiography. Relative mobilities were measured from a point originating at the loading well.
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).
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 [beta]-galactosidase assay as described (12).
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 42°C; washes were in 0.5× SSC at 55°C.
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, MBP[alpha] (ME1 binding protein [alpha]) and MBP[beta], became the focus of our study due to their specific interactions with the ME1 Inr. MBP[alpha] 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 MBP[alpha], suggesting 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 MBP[alpha] binding (see below). MBP[beta] 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 MBP[beta] bound the ME1 Inr within the downstream inverted repeat 5'-GTCCGCCTG-3' (Fig. 2).
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 MBP[beta] because: (i) the bend center was coincident with the MBP[beta] 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 MBP[beta] (and not MBP[alpha]). However, the MBP[alpha] complex is likely to contain components of MBP[beta], since these bands co-migrated in our bending assays (these complexes were resolved at lower gel concentrations).
To examine the putative role of the MBP[alpha] and MBP[beta] 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 MBP[alpha] and MBP[beta] 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 MBP[alpha] in HeLa Scribe (and also in PCC7 and HT-4 extracts; not shown), suggesting that MBP[alpha] is recruited by the poly(dA) tract. This is consistent with the lack of MBP[alpha] 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 HeLa Scribe, implying that components of the Pol II transcription machinery (i.e. TFIID) do not interact directly with the ME1 Inr.
To determine the distribution of the MBP[alpha] and MBP[beta] complexes in the adult mouse, nuclear extracts were prepared from various tissues and subjected to EMSA analysis (Fig. 6). The MBP[beta] 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 MBP[alpha]. Several faster migrating species were observed in liver, lung, kidney and spleen.
To determine whether the MBP[alpha] and MBP[beta] 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 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 MBP[alpha] and MBP[beta] binding and accurate transcription initiation from the poly(dA) tract (not shown). The region downstream of the poly(dA) tract, corresponding to the MBP[beta] binding site (5'-TGCCGCCGT-3'), was strikingly conserved in all species.
This study demonstrates the autonomy of the ME1 Inr as a transcriptional activator and the putative roles of ME1 Inr binding proteins MBP[alpha] and MBP[beta]. A model depicting one possible mechanism of transcription initiation by the ME1 Inr is shown in Figure 10. MBP[alpha] is recruited by the 13 bp poly(dA) tract (-6 to +7) and is likely to associate directly with MBP[beta]. MBP[beta] independently, or in combination with MBP[alpha], 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. MBP[alpha] and MBP[beta]). 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 MBP[alpha] [which binds the poly(dA) tract], it appears that MBP[alpha] 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 MBP[alpha] 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 MBP[alpha] observed in the adult tissues examined here (see Fig. 6). Finally, it remains to be established whether MBP[alpha] 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 MBP[beta] 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.
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.
8Klein,E.S., Simmons,D.M., Swanson,L.W. and Rosenfeld,M.G. (1993) Genes Dev., 7,55-71.MEDLINE Abstract
9Neuman,T., Keen,A., Knapik,E., Shain,D., Ross,M., Nornes,H.O. and Zuber,M.X. (1993) Eur. J. Neurosci., 5, 311-318.MEDLINE Abstract
10Ausubel,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.
11Graham,F.L. and van der Eb,A.J. (1973) Virology, 52, 456-467.MEDLINE Abstract
12Sambrook,J., Fritsch,E.F. and Maniatis,T. (1989) Molecular Cloning: A Laboratory Manual, 2nd Edn. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY.
13Nelson,H.C.M., Finch,J.T., Luisi,B.F. and Klug,A. (1987) Nature, 330, 221-226.
14Kim,J., Zwieb,C., Wu,C. and Adhya,S. (1989) Gene, 85, 15-23.MEDLINE Abstract
15Wu,H.M. and Crothers,D.M. (1984) Nature, 308, 509-513.MEDLINE Abstract
16Bisbee,C.A., Baker,M.A., Wilson,A.C., Hadji-azimi,I. and Fischberg,M. (1977) Science, 195,785-787.MEDLINE Abstract
17Smale,S.T., Schmidt,M.C., Berk,A.J. and Baltimore,D. (1990) Proc. Natl. Acad. Sci. USA, 87, 4509-4513.MEDLINE Abstract
A. BRUNETTI, G. MANFIOLETTI, E. CHIEFARI, I. D. GOLDFINE, and D. FOTI Transcriptional regulation of human insulin receptor gene by the high-mobility group protein HMGI(Y)
FASEB J,
February 1, 2001;
15(2):
492 - 500.
[Abstract][Full Text][PDF]
CiteULike 
Connotea 
Del.icio.us What's this?
