Biochimica et Biophysica Acta 1399 (1998) 117^125
A sequence based computational identi¢cation of a Drosophila developmentally regulated TATA-less RNA polymerase II promoter and its experimental validation Marie-Jose¨e Santoni a , Ounissa A|«t-Ahmed
a;
*, Monique Marilley
b
a
b
LGPD/IBDM, CNRS/Universite¨ de la Me¨diterrane¨e, Parc Scienti¢que et Technologique de Luminy, CNRS case 907, 13288 Marseilles Cedex 9, France RGFCP, UPRES 2059, Universite¨ de la Me¨diterrane¨e, Faculte¨ de Me¨decine 27, Boulevard Jean Moulin, 13385 Marseilles Cedex 5, France Received 26 February 1998; revised 13 May 1998; accepted 27 May 1998
Abstract Many RNA polymerase II promoters lack the characteristic TATA box sequence located 325/330 nucleotides upstream from the transcription start. In Drosophila, half of the promoters identified so far are TATA-deficient. The yemanuclein-K gene whose promoter activity is restricted to oogenesis, falls in this class. A number of upstream and downstream promoter elements have been identified for some TATA-less promoters. The yem-K promoter contains none of the consensus elements identified so far. Our work was based on the assumption that the physical parameters of the DNA could be used to predict the location of the yem-K promoter. A sequence based computational analysis allowed us to determine the characteristic changes of DNA curvature and helix stability in the presumptive regulatory region. Our experimental data were in good agreement with the computational analysis. We have started to investigate the general value of this approach by analyzing other promoters. ß 1998 Elsevier Science B.V. All rights reserved. Keywords: TATA-less promoter; yem-K ; Bent DNA ; Duplex stability ; (Drosophila)
1. Introduction Although eukaryotic promoters di¡er widely, one can distinguish two main classes, the TATA box containing and the TATA-less promoters. An important proportion of cellular genes (50% in Drosophila) fall in this class. Interestingly, a large number of Drosophila regulatory factors encoding genes (which
* Corresponding author. Institut de Ge¨ne¨tique Humaine, CNRS UPR 1142, 141, Rue de la Cardonille, 34396 Montpellier Cedex 5, France. E-mail:
[email protected]
may be themselves submitted to developmental regulation) are TATA-de¢cient [3]. It is noteworthy that the same TFIID transcription factor is involved in their basal activity although the mechanisms of the transcription complex may di¡er. It is clearly established both by biochemical analysis and X-ray analysis of crystal structure of the TBP/ TATA-box complex, that the TATA binding protein (TBP) interacts directly with the TATA box, resulting in a dramatic bending of the DNA structure [23,24]. The mechanisms of the initiation complex formation at the TATA-less promoters are less well documented. While the TATA box is found around 325/330 from the RNA start, the location of the
0167-4781 / 98 / $19.00 ß 1998 Elsevier Science B.V. All rights reserved. PII: S 0 1 6 7 - 4 7 8 1 ( 9 8 ) 0 0 0 9 3 - 1
BBAEXP 93125 7-8-98
118
M.-J. Santoni et al. / Biochimica et Biophysica Acta 1399 (1998) 117^125
control region for the basal transcriptional machinery varies widely in the so-called TATA-less promoters. The presence of promoter consensus sequences such as the initiator element has been reported for a number of genes. They map at the close vicinity of the transcription start [37,36]. In some cases, the control regions may lie quite far upstream from the transcription site as shown for the Drosophila vermilion gene [17], or downstream within the transcribed region [8]. The only systematic search for consensus sequences within TATA-less promoters has been reported in [3]. The variety in location and the sequence diversity of the promoter regions led us to the hypothesis that structural elements rather than the sequence itself may play a critical role. According to this view the TBP factor could act by exploiting the deformability of the TATA sequence to initiate the unwinding of the DNA helix, which is a prerequisite in transcription [25]. Indeed there is some evidence that the T-A step in a double helix is easily deformable [25,26]. The above observation and others led to the conclusion that bent forms of DNA are important features in the control of a number of essential biological functions such as replication, recombination and, of course, transcription [20]. Recently, it has been shown that RNA polymerase I promoters share common structural features even in the absence of sequence homology [29]. Given the necessity of a number of common general transcription factors for the activity of the three RNA polymerases, we suggest that the promoters they act on could share some of these structural features. TATA-less promoters cannot be identi¢ed even by sophisticated computer programs such as PromoterScan [35]. Thus, it is a real challenge to try to identify them on the ground of their nucleotide sequence. In this paper, we show that in the absence of any conspicuous sequence motif, it was possible to use a sequence based computational method to predict and identify the yem-K gene promoter. This computer based identi¢cation was con¢rmed by an experimental approach. We show that the two parameters analyzed, DNA curvature and helix stability, reveal su¤cient information to identify the promoter sequence in a candidate region. More promoters have been analyzed in order to test the generalization of our model.
2. Material and methods 2.1. DNA curvature calculation The algorithm for calculating DNA bending from nucleotide sequence was published by Eckdahl and Anderson [14]. Three-dimensional coordinates of the helical axis are calculated along the sequence as previously described [33], using parameters of the wedge model for bent DNA from Bolshoy et al. [6] and twist angles from Kabsch et al. [22]. The magnitude of bending is expressed as the ENDS ratio, de¢ned as the ratio of the contour length of a segment of the helical axis to the shortest distance between its ends. The ENDS ratio was computed at a window width of 200 nucleotides and at a window step of 1 nucleotide. This value was chosen to allow comparison with the results of [40,29]. 2.2. DNA duplex stability determination The thermodynamic libraries, that characterize all ten Watson-Crick nearest-neighbor interactions in DNA [7], provide an empirical basis for predicting the stability (vG) of any DNA duplex region by inspection of its primary sequence. The local energy required for the strand separation of a 200 bp segment of DNA was calculated. Each value takes into account the contribution of the surrounding nucleotides and is plotted instead of plotting the individual values. All calculations were carried out with a 1 bp step movement and parameters corresponding to 1 M NaCl, 25³C and pH 7. Calculations were made with the PACS DNA program already exploited in previous studies [11,30, 29,33]. 2.3. Constructs used in the cotransfection assay The deletion constructs were made with the pCaspeR AUG-L-galactosidase (L-Gal) vector used in Drosophila transformation experiments [38]. To generate a vA construct, we used a pBluescript SK yem-K clone obtained by exonuclease III nested deletions [2]. An EcoRI-KpnI fragment containing the ¢rst 1378 nucleotides (nt) was subcloned into the corresponding sites of the pCasper AUG-L-Gal vector.
BBAEXP 93125 7-8-98
M.-J. Santoni et al. / Biochimica et Biophysica Acta 1399 (1998) 117^125
119
349), SnaBI (nt 473) and BssHII (nt 633). These deleted clones were then blunted and ligated; they are called, respectively, pGEM 3-SacII, pGem 3-BglII, pGem 3-SnaBI, pGem 3-BssHII. The various SphIKpnI deletion fragments were then cut out, puri¢ed and subcloned into the vA (SphI-KpnI cut) vector to generate the constructs used in transfection assays. The numbering of the sequence is as reported in [2]. The transfection e¤ciency was assessed using an hsp-70 CAT construct. The hsp-70 CAT construct was made by cloning the SalI-BamHI fragment containing hsp-70 promoter from pCasper-hs into the pBLCAT-3 [34,28]. 2.4. Drosophila cell culture, cell transfection and L-galactosidase activity measurements
Fig. 1. Curvature map and DNA duplex stability variation in the restriction fragment which contains the yem-K promoter. Upper part: curvature map of the promoter region. The nucleotide sequence is analyzed by computer modeling to reveal sequence-directed curvature. The 3D helical path of the molecule was calculated using the model of Trifonov [6]; the mean curvature map is shown. ENDS ratios were computed at a 200 bp window size and a 1 bp step. Lower part: variation of duplex stability along the yem-K promoter region. Local vG variations and G+C content maps (thick solid line, vG curves; thin solid line, G+C content). The vG is calculated as the sum of nearestneighbor interaction values for a 200 bp window sliding along the sequence as indicated in Section 2. Restriction map and positions of nucleotides relative to the start point of transcription are shown.
All the other deletion constructs were vA derivatives. First the SphI-KpnI fragment containing the entire yem-K 5P sequences including the 5P UTR, was recloned into the HincII-KpnI cut pGEM 3 vector, after ¢lling in the SphI protruding end. To generate vA20, vA15, vA10 and vA5 constructs, the pGEM-yem-K intermediate construct was double cut in the pGEM PstI site and in a downstream yem-K site, respectively SacII (nt 205), BglII (nt
Drosophila S2 cells were grown at 25³C in modi¢ed Schneider's medium supplemented with 10% fetal calf serum, 50 U/ml of penicillin and 50 Wg/ml streptomycin. Cell transfections were performed using the calcium phosphate precipitation technique [19]. 3 Wg of reporter DNA (vA, vA20, vA15, vA10, vA5) were used in the transfection assay. 0.7 Wg of hsp CAT vector was also incorporated in each assay.
Fig. 2. Deletion constructs and promoter activity. On the left part are ¢gured the remaining 5P yem-K sequences (bold lines) fused to the lacZ reporter gene; the complete structure of the fused construct £anked with SV40 3P sequences is shown for vA. The arrow indicates the transcription start site. Restriction sites: E, EcoRI; Sp, SphI; Sa, SacII; Bg, BglII; Sn, SnaBI; Bs, BssHII; Kp, KpnI. On the right side are shown the relative value of the promoter activity for each deletion construct with the standard deviation, and the number (n) of samples tested independently.
BBAEXP 93125 7-8-98
120
M.-J. Santoni et al. / Biochimica et Biophysica Acta 1399 (1998) 117^125
Fig. 3. Deletion mutations and their structural e¡ects as detected by computational analysis. Local bending and DNA double helix stability variations following progressive nucleotide sequence deletions. The bent structure, positioned about 300 bp upstream from the transcription start is progressively altered and then eliminated in vA15, vA10 and vA5 deletion mutants. Changes in the helical stability are also observed but they are not so obvious since they may be more or less substituted by the low GC% content of the neighboring vector. Contribution of the vector to the double strand stability is indicated by a solid gray line.
36 h after transfection, the cells were heat shocked for 15 min and let to recover for 1.5 h at 25³C. They were then lysed according to the protocol provided with the CAT ELISA kit (Boehringer). The protein
concentration in the cell extract was determined by the Bradford kit from Bio-Rad. The L-galactosidase activity measurement was carried out in 30 Wl of the cell lysate (a 20 Wg protein equivalent) diluted in 140
BBAEXP 93125 7-8-98
M.-J. Santoni et al. / Biochimica et Biophysica Acta 1399 (1998) 117^125
121
Table 1 Size of the bent element mapping immediately upstream (1) and downstream (2) from the transcription start Promoter
Organism
Reference
ENDS ratio (1)
ENDS ratio (2)
1. ADP-ribosylation factor 3 2. N-Cam 3. N-Cam 4. Ras2 5. Stellate 6. TPA 7. Thymidylate synthetase 8. Utrophin 9. Utrophin 10. White
Human Human Mouse Drosophila Drosophila Rat Mouse Human Mouse Drosophila
[39] [5] [21] [10] [27] [16] [12] [13] [13] [31]
1.13 1.15 1.08 1.27 1.18 1.14 1.06 1.09 1.03 1.22
ND 1.12 1.07 ND 1.19 ND 1.08 1.09 1.07 1.15
The size corresponds to the maximum ENDS ratio value calculated with a window size of 200 bp and a 1 bp step. ND, not determined because the sequence is not long enough to allow analysis.
Wl of Z bu¡er (60 mM Na2 HPO4, 40 mM NaH2 PO4, 10 mM KCl, 1 mM MgCl2 , 50 mM L-mercaptoethanol), and 30 Wl of a 4 mg/ml solution of the L-galactosidase ONPG chromogenic substrate. The reaction mix was incubated at 37³C; as a result of the enzymatic activity, a yellow coloration was obtained. At this step, 50 Wl of a 2 M Na2 CO3 solution were added and colorimetric measurement was carried out at 420 nm. Seven to ten independent assays were performed for each construct. The values were expressed as a percentage of the activity obtained for the vA reference construct considered arbitrarily as 100%. The mean value and standard deviation were determined for each set of data. 3. Results 3.1. Computational analysis of the yem-K presumptive promoter region The transcription start maps at nucleotide 717 as determined by primer extension and RT-PCR experiments (data not shown). Using transgenic £ies, we have shown that the promoter is located within the ¢rst 716 nucleotides [9]. The promoter containing region is TATA-less, moreover none of the motifs reported so far to be conserved through a number of TATA-less promoters could be identi¢ed, either upstream or downstream to the transcription start site [1,2]. Because DNA bending characteristics as well as
thermodynamic properties might be important parameters for the promoter function we thought that we could investigate these parameters with calculation programs on the ground of the primary sequence. 3.2. DNA bending characteristics A curvature map is shown in the upper part of Fig. 1. The analyzed region is characterized by the presence of two bent structures of nearly equivalent amplitude at positions 350^500 and 750^900. The maximum ENDS ratio value is about 1.2, indicating that the curvature is not a very strong one. As a comparison, the ARS1 curvature which is known to contain a strongly bent DNA was calculated in the same conditions to be 1.5 [40]. A second peak of curvature is found immediately downstream from the transcription start as determined experimentally (nt 717). The signi¢cance of this peak is still unclear. 3.3. Thermodynamic characteristics General information on the duplex stability may be obtained by inspecting the G+C content variations along the nucleotide sequence (Fig. 1), but a better way to describe the promoter duplex stability is to analyze the thermodynamic features of the sequence. Thermodynamic libraries have been established that characterize all ten Watson-Crick nearest-neighbor interactions in DNA [7]. These thermodynamic data provide an empirical basis for predicting the stability (vG) of any DNA duplex region. The var-
BBAEXP 93125 7-8-98
122
M.-J. Santoni et al. / Biochimica et Biophysica Acta 1399 (1998) 117^125
iations of the thermodynamic characteristics along a nucleotide sequence are used here to probe a DNA sequence of about 1200 nucleotides which contains the presumptive TATA-less promoter (Fig. 1). A region of decreased stability is clearly seen between positions 300 and 700. This region is localized upstream from the transcription start. We also ¢nd that the stability of the DNA duplex is highest in the neighborhood of the transcription start, which results in a `thermodynamic barrier'. Interestingly, this feature is shared with the ribosomal gene promoters [29]. These values registered near the transcription start were found to be fairly superior to the average value, calculated for the whole gene sequence (data not shown). 3.4. Experimental determination of the yem-K basal promoter In order to verify the presence of the basal promoter in the DNA region de¢ned above, we have realized deletion constructs whose structure is shown in Fig. 2. The yem-K 5P sequences were fused to a lacZ reporter gene, the SV40 3P polyadenylation sequences being fused downstream from the lacZ gene. The various constructs were used in transient transfection assays of Drosophila S2 cells. The transcription level was assessed indirectly with an assay for Lgalactosidase activity. The activity of vA construct was arbitrarily considered as the 100% activity. vA contains the whole yem-K promoter region and the 5P UTR. We were con¢dent that any variation in Lgalactosidase activity could be accounted for by differences at the transcription level as all the other deletion constructs were vA derivatives with an intact 5P UTR. The ¢rst signi¢cant decrease in enzymatic activity could be readily observed with vA15 in which the DNA is deleted up to the BglII site (nt 349). The enzymatic activity was hardly above background levels with vA10 in which the deleted region extends up to the SnaBI site (nt 473). These data indicate that the basal promoter is located within a fragment which spans a region from nt 349 to nt 473. This result is in good agreement with the computational analysis which pointed out this region as important in promoter activity.
3.5. Alteration of structural parameters within the deletion constructs Fig. 3 shows the computational analysis (ENDS ratio and vG) of the deletion constructs described above. We veri¢ed ¢rst that the vA construct from which the deletion constructs are derived, exhibits the same features as the chromosomal sequence (DNA bending and stability pro¢les). We were con¢dent that the vector did not a¡ect the structural properties of the region analyzed in the present study. The curvature map remains unchanged for the ¢rst deletion, vA20, while the amplitude of the bent structure, upstream from the transcription start, is decreased in vA15. This bent structure is deleted in the following vA10 and vA5 constructs. There is a good correlation between the decrease in promoter activity as shown in Fig. 2, and the changes in curvature map. Changes in DNA stability are also observed. Interestingly, the vA20 construct which exhibits higher transcription activity than the vA reference construct displays a decrease in vG values in its upstream region. Moreover, the adjacent sequences of the vector were found to have a low stability. Both events may contribute to DNA unwinding facilitation and consequently result in the observed increase in transcription activity. Nevertheless, the vector contribution being equal for all the deletion constructs, if we considered only the vG parameter we would expect a higher promoter activity for vA15 and vA10 constructs than observed. This indicates the importance of considering more than one parameter in such an analysis. 3.6. ENDS ratio determination on other promoters In order to test whether the computational method applied to the analysis of yem-K promoter could have a more general value, we have applied the same analysis to other TATA-less promoter sequences available in data bases. Ten such sequences could be considered as two criteria had to be met: a size amenable to analysis and availability of information concerning transcription initiation (Table 1).
BBAEXP 93125 7-8-98
M.-J. Santoni et al. / Biochimica et Biophysica Acta 1399 (1998) 117^125
The curvature map was calculated for all these sequences in the same conditions as for yem-K. Interestingly, all these promoters revealed a similar organization. They show, in every point, very similar curvature and vG maps as the ones calculated for the yem-K promoter (data not shown). Like the yem-K promoter, they all exhibited ENDS ratios whose highest values were rather low, and strikingly the second structure downstream from the transcription start was also found. It is to be noticed that this downstream structure element must not be assimilated to the previously described downstream promoter element (DPE) [8]; although we think that the DPE could correspond to a structural element, it remained undetected in the present analysis. It follows from this ¢rst comparative study that, although corresponding to di¡erent genes and issuing from various organisms, TATA-less promoters share some common features. This is typically a favorable situation for identifying these promoters on the basis of sequence-induced structural elements. Such analyses will be useful to determine the critical structural parameters for promoter function regardless of the presence of characteristic sequence motifs. 4. Discussion 4.1. Analogy between yem-K and other eukaryotic gene promoters Recent results showed that TATA-less promoters may not be deeply di¡erent from the TATA box containing ones. Aso et al. [4] suggest that RNA polymerase II could assemble with several TATA box and TATA-less promoters by a common pathway. Moreover, according to Wiley et al. [41] the two types of promoters could simply di¡er in TFIID af¢nity for their 330 region. This led us to the hypothesis that the computational analysis previously carried out for a TATA box containing promoter [11] and for various ribosomal promoters [29] could also be applied for the analysis of an RNA polymerase II TATA-less promoter. In the present work, we performed a sequence based computational analysis of two structural parameters in the developmentally regulated yem-K
123
gene promoter. In parallel, we have carried out an experimental analysis to attest the pertinence of our computer predictions. Strong analogies were found between the thermodynamic characteristics of yem-K promoter and rRNA promoters. The region of lowest DNA stability was found by deletion mutations to be necessary for promoter activity. When compared to the ribosomal promoter of Pisum sativum which has an equivalent GC content, a similar pattern of vG variation and a similar vG amplitude were observed. Interestingly, in both cases the transcription start maps at the limit of a vG `barrier' [29]. It appears then that the thermodynamic characteristics may follow the same basic rule of variation in very di¡erent promoters. This pro¢le of vG variation along the candidate region made the localization of the promoter sequence unambiguous. 4.2. Bent DNA and promoter function Characteristic ENDS ratio variations have been found both in yem-K promoter and in the other TATA-less promoters we have analyzed. The highest values averaging 1.2 were found for yem-K and the other Drosophila TATA-less promoters analyzed. This might be correlated with the highest A+T content of their DNA. All these bending values are surprisingly low compared to those found for most of the ribosomal promoters. Nevertheless, we think that the bent structure observed in TATA-less promoters is equivalent to the one found in ribosomal promoters. As previously discussed by Marilley and Pasero [29], very di¡erent functions may be carried on by this structure (e.g. protein docking, nucleosome positioning). The function of bent DNA is only documented in prokaryotic promoters. For example in L-lactamase [32] the promoter activity was seen to be related to the gross geometry of the bent structure. In the case of yemK promoter, the inactivation of the promoter is seen to coincide with the deletion of the bent structure. An attractive hypothesis would be the existence of a correlation between the low ENDS ratio value and the fact that the promoter is submitted to a developmental regulation. In the case of a developmentally regulated promoter such as yem-K, the relation of the bent structure and nucleosome positioning is
BBAEXP 93125 7-8-98
124
M.-J. Santoni et al. / Biochimica et Biophysica Acta 1399 (1998) 117^125
interesting to discuss since nucleosome placement is critical for gene regulation. Flexible DNA is more likely to form nucleosomes than other structures such as Z-DNA [18]. Eukaryotic promoters appear to fall into two architectural categories, either preset or remodeled. Induction of transcription from this second type of promoter occurs only after rearrangement of the chromatin environment in order to make a gene control region accessible to transcription components. In the case of yem-K promoter, a poorly bent DNA may be compatible with a continuous nucleosome organization but could also provide, in response to an activation signal, a nucleation site where chromatin rearrangement may occur, allowing an architectural change necessary for the activity of the promoter [42]. 4.3. Predictive value of such a computational analysis An important issue remains to detect TATA-less promoters on the ground of their nucleotide sequence. The two structural components we have studied here revealed su¤cient information to identify a TATA-less promoter within a presumptive region. However, more information is necessary to detect them unambiguously over the background noise of a complete genomic nucleotide sequence. We are aware that such a detection will imply the identi¢cation of a greater number of parameters. The use of structurally mutated promoters might prove valuable to identify which are the important features. Notably, at this step, computational prediction might reveal very useful information. Moreover, some of these parameters may possibly be better understood in analyzing the structural requirements for transcription factor recognition. A growing number of them appear to be based on groove structural characteristics ([23,24]; Marilley et al., unpublished results). Promoter detection remains one of the challenging issues in the genome projects. This is even more critical for TATA-less promoters as their identi¢cation requires alternative methods to the ones already available [15]. We have tried to show in this report that a computational identi¢cation of promoter regions in candidate regions is now possible. Of course it has to be validated by experimental approaches as described in this work.
We think that our model can be used in the search for other structural parameters which may play a key role in promoter function. These studies are only at their beginning. Acknowledgements This work has been partly supported by research grants from the Ministe©re Charge¨ de la Recherche (ACC-SV 4) to O.A.-A. and the Association pour la Recherche sur le Cancer (ARC) to M.M.
References [1] O. A|«t-Ahmed, M. Thomas-Cavallin, R. Rosset, Isolation and characterization of a region of the Drosophila genome which contains a cluster of di¡erentially expressed maternal genes (yema gene region), Dev. Biol. 122 (1987) 153^162. [2] O. A|«t-Ahmed, B. Bellon, M. Capri, C. Joblet, M. ThomasDelaage, The yemanuclein-alpha: a new Drosophila DNA binding protein speci¢c for the oocyte nucleus, Mech. Dev. 37 (1992) 69^80. [3] I.R. Arkhipova, Promoter elements in Drosophila melanogaster revealed by sequence analysis, Genetics 139 (1995) 1359^1369. [4] T. Aso, J.W. Conaway, R.C. Conaway, Role of core promoter structure in assembly of the RNA polymerase II preinitiation complex. A common pathway for formation of preinitiation intermediates at many TATA and TATA-less promoters, J. Biol. Chem. 269 (1994) 26575^26583. [5] C.H. Barton, D.A. Mann, F.S. Walsh, Characterization of the human N-CAM promoter, Biochem. J. 268 (1990) 161^ 168. [6] A. Bolshoy, P. McNamara, R.E. Harrington, E.N. Trifonov, Curved DNA without A-A: experimental estimation of all 16 DNA wedge angles, Proc. Natl. Acad. Sci. USA 88 (1991) 2312^2316. [7] K.J. Breslauer, R. Frank, H. Blocker, L.A. Marky, Predicting DNA duplex stability from the base sequence, Proc. Natl. Acad. Sci. USA 83 (1986) 3746^3750. [8] T.W. Burke, J.T. Kadonaga, Drosophila TFIID binds to a conserved downstream basal promoter element that is present in many TATA-box-de¢cient promoters, Genes Dev. 10 (1996) 711^724. [9] M. Capri, M.-J. Santoni, M. Thomas-Delaage, O. A|«tAhmed, Implication of a 5P coding sequence in targeting maternal mRNA to the Drosophila oocyte, Mech. Dev. 68 (1997) 91^100. [10] N. Cohen, A. Salzberg, Z. Lev, A bidirectional promoter is regulating the Drosophila ras2 gene, Oncogene 3 (1988) 137^ 142.
BBAEXP 93125 7-8-98
M.-J. Santoni et al. / Biochimica et Biophysica Acta 1399 (1998) 117^125 [11] M.-L. Delabre, P. Pasero, M. Marilley, P. Bougis, Promoter structure and intron-exon organization of a scorpion alphatoxin gene, Biochemistry 34 (1995) 6729^6736. [12] T. Deng L, D.W. Li, C.-H. Jenh, L.F. Johnson, Structure of the gene for mouse thymidylate synthase. Locations of introns and multiple transcriptional start sites, J. Biol. Chem. 261 (1986) 16000^16005. [13] C.L. Dennis, J.M. Tinsley, A.E. Deconinck, K.E. Davies, Molecular and functional analysis of the utrophin promoter, Nucleic Acids Res. 24 (1996) 1646^1652. [14] T.T. Eckdahl, J.N. Anderson, Computer modelling of DNA structures involved in chromosome maintenance, Nucleic Acids Res. 15 (1987) 8531^8545. [15] J.W. Fickett, Finding genes by computer: the state of the art, Trends Genet. 12 (1996) 316^320. [16] P. Feng, M. Ohlson, T. Ny, The structure of the TATA-less rat tissue-type plasminogen activator gene. Species-speci¢c sequence divergences in the promoter predict di¡erences in regulation of gene expression, J. Biol. Chem. 265 (1990) 2022^2027. [17] Y.W. Fridell, L.L. Searles, In vivo transcriptional analysis of the TATA-less promoter of the Drosophila melanogaster vermilion gene, Mol. Cell. Biol. 12 (1992) 4571^4577. [18] M.M. Garner, G. Felsenfeld, E¡ect of Z-DNA on nucleosome placement, J. Mol. Biol. 196 (1987) 581^590. [19] C.M. Gorman, L.F. Mo¡at, B.H. Howard, Recombinant genomes which express chloramphenicol acetyltransferase in mammalian cells, Mol. Cell. Biol. 2 (1982) 1044^1051. [20] P. Hagerman, Sequence-directed curvature of DNA, Annu. Rev. Biochem. 59 (1990) 755^781. [21] M.-R. Hirsch, L. Gaugler, H. Deagostini-Bazin, L. BallyCuif, C. Goridis, Identi¢cation of positive and negative regulatory elements governing cell-type-speci¢c expression of the neural cell adhesion molecule gene, Mol. Cell. Biol. 10 (1990) 1959^1968. [22] W. Kabsch, C. Sander, E.N. Trifonov, The ten helical twist angles of B-DNA, Nucleic Acids Res. 10 (1982) 1097^1104. [23] Y. Kim, J.H. Geiger, S. Hahn, P.B. Sigler, Crystal structure of a yeast TBP/TATA-box complex, Nature 365 (1993) 512^ 520. [24] J.L. Kim, D.B. Nikolov, S.K. Burley, Co-crystal structure of TBP recognizing the minor groove of a TATA element, Nature 365 (1993) 520^527. [25] A. Klug, Transcription. Opening the gateway, Nature 365 (1993) 486^487. [26] A. Klug, A. Jack, M.A. Viswamitra, O. Kennard, Z. Shakked, T.A. Steitz, A hypothesis on a speci¢c sequencedependent conformation of DNA and its relation to the binding of the lac-repressor protein, J. Mol. Biol. 131 (1979) 669^680. [27] K.J. Livak, Detailed structure of the Drosophila melanogas-
[28] [29]
[30]
[31]
[32]
[33]
[34]
[35]
[36]
[37] [38]
[39]
[40]
[41]
[42]
125
ter stellate genes and their transcripts, Genetics 124 (1990) 303^316. B. Luckow, G. Schutz, A new method for constructing linker scanning mutants, Nucleic Acids Res. 15 (1987) 417^429. M. Marilley, P. Pasero, Common DNA structural features exhibited by eukaryotic ribosomal gene promoters, Nucleic Acids Res. 24 (1996) 2204^2211. M. Marilley, P. Pasero, A. Humbert, S. Granjeaud, M. Dayez, R. Pierrisnard, B. Jordan, Scanning tunneling microscopy study of a DNA fragment of known size and sequence, Microsc. Microanal. Microstruct. 5 (1994) 47^56. K. O'Hare, C. Murphy, R. Levis, G.M. Rubin, DNA sequence of the white locus of Drosophila melanogaster, J. Mol. Biol. 180 (1984) 437^455. T. Ohyama, M. Nagumo, Y. Hirota, S. Sakuma, Alteration of the curved helical structure located in the upstream region of the beta-lactamase promoter of plasmid pUC19 and its e¡ect on transcription, Nucleic Acids Res. 20 (1992) 1617^ 1622. P. Pasero, N. Sjakste, C. Blettry, C. Got, M. Marilley, Longrange organization and sequence-directed curvature of Xenopus laevis satellite 1 DNA, Nucleic Acids Res. 21 (1993) 4703^4710. V. Pirrotta, in: R.L. Rodriguez, D.T. Denhardt (Eds.), Vectors: a Survey of Molecular Cloning Vectors and their Use, Butterworths, Boston, MA, 1988, pp. 437^456. D.S. Prestridge, Predicting Pol II promoter sequences using transcription factor binding sites, J. Mol. Biol. 249 (1995) 923^932. B.F. Pugh, R. Tjian, Transcription from a TATA-less promoter requires a multisubunit TFIID complex, Genes Dev. 5 (1991) 1935^1945. S.T. Smale, D. Baltimore, The `initiator' as a transcription control element, Cell 5 (1989) 103^113. C.S. Thummel, A.M. Boulet, H.D. Lipshitz, Vectors for Drosophila P-element-mediated transformation and tissue culture transfection, Gene 74 (1988) 445^456. S.-C. Tsai, R.S. Haun, M. Tsuchiya, J. Moss, M. Vaughan, Isolation and characterization of the human gene for ADPribosylation factor 3, a 20-kDa guanine nucleotide-binding protein activator of cholera toxin, J. Biol. Chem. 266 (1991) 23053^23059. J.D. VanWye, E.C. Bronson, J.N. Anderson, Species-speci¢c patterns of DNA bending and sequence, Nucleic Acids Res. 19 (1991) 5253^5261. S.R. Wiley, R.J. Kraus, E. Mertz, Functional binding of the `TATA' box binding component of transcription factor TFIID to the 330 region of TATA-less promoters, Proc. Natl. Acad. Sci. USA 89 (1992) 5814^5818. A.P. Wol¡e, Architectural transcription factors, Science 264 (1994) 1100^1101.
BBAEXP 93125 7-8-98
