Nucleic Acids Research

Plasmids and cell cultures

pGLTH plasmid contains -425/+25 bp 5[prime]-flanking region of the bovine TH gene (34) cloned as described previously (24) into the pGL2_Basic promoterless plasmid (Promega) in front of the luciferase reporter gene. pGLTH derivatives with truncated or mutated DSE1 (Fig. 1A) were obtained using routine cloning procedures. In order to construct the [Delta]DSE1 promoter mutant the pGLTH plasmid was digested with Acc65I and HindIII and the resulting 495 bp fragment containing -425/+25 bp of TH promoter was digested with HphI which cuts TH promoter at -301 bp. The 3[prime]-protruding ends were blunted with T4 polymerase and DNA was digested with MscI (at -365 bp), AatII (at -39 bp) and MluI (between the Acc65I site in polylinker and the 5[prime]-end of the TH promoter). The MluI/MscI and HphI/AatII (262 bp) fragments were ligated to the 5.6 kb pGLTH fragment remaining after the removal of AatII and MluII. To construct the [Delta]L mutant, we isolated 495 bp Acc65I/HindIII fragment of pGLTH containing the TH promoter (-425/+25 bp) and divided it into portions. One portion was further digested with MscI and an 81 bp Acc65I/MscI fragment containing -425/-365 bp of the TH promoter was isolated. The second portion was digested with BstNI (at -330 bp), and its 5[prime]-protruding ends were filled with Klenow enzyme. The DNA was further digested with AatII and the 291 bp BstNI/AatII fragment was ligated along with the 96 bp Acc65I/MscI fragment to the large 5.6 kb Acc65I/AatII fragment of pGLTH. The [Delta]R mutant was constructed in a similar way as the [Delta]L, except that the large Acc65I/AatII fragment of pGLTH was ligated to Acc65I/BstNI and HphI/AatII fragments from the TH promoter. The construction of the reversed orientation (RevL) mutant involved an isolation of 35 bp MscI/BstNI fragment of TH promoter and treatment with Klenow to blunt the BstNI 5[prime]-protruding end. This fragment was combined in the ligation mixture with a 290 bp BstNI/AatII fragment of the TH promoter in which the BstNI ends were blunted prior to AatII digestion and larger 5.7 kb MscI/AatII fragment was derived from pGLTH. The recombinants in which the orientation of -365/-330 bp TH promoter sequence was reversed, were identified by DNA sequencing. All DNA constructs were sequenced with a Sequenase 2.0 kit (US Biochem Corp., Cleveland, OH) to confirm their identities. In all experiments plasmid preparations were used in which >90% of DNA was in supercoiled form as determined by agarose gel electrophoresis and confirmed by atomic force microscopy as described in (35), (Y.Lyubchenko, H.Peng and M.K.Stachowiak, unpublished observations).

Culture conditions for TH expressing bovine adrenal medullary chromaffin (BAMC) cells and human IMR322 neuroblastoma cells have been described previously (24,36). The TE 671 cell line, originally designated as medulloblastoma (37) was cultured as described in (38). pGL2_Basic, pGLTH or its promoter mutants were transfected into cells by calcium phosphate-DNA co-precipitation, and the luciferase activity was determined 24 h later as previously described (24). To estimate TH promoter function accurately, differences in efficiency of transfection among individual wells and plasmids were normalized by measuring the content of transfected DNA (dot blot hybridization) in lysates used for luciferase assay (24,36). Luciferase activity was expressed as the number of photons (c.p.m.) per microgram of protein × picograms DNA. At least two different preparations of each plasmid were tested in all experiments.

S1 nuclease- and OSO₄-mapping of unpaired bases

S1 nuclease-sensitive sites were mapped as previously described (24). Briefly, 1 µg of supercoiled plasmid DNA was incubated in 50 µl of S1 buffer (40 mM sodium acetate, pH 4.5, 100 mM NaC1 and 1 mM ZnSO₄) with 0.05-1.0 U of S1 nuclease for 5 min at room temperature. S1 nuclease-nicked plasmid DNA was digested with Acc65I, which linearized plasmid directly upstream from the TH promoter, and end-labeled with [[gamma]-³²P]ATP and T4 polynucleotide kinase (for mapping top stand) or [[alpha]-³²P]dGTP and Klenow polymerase (bottom strand). The end-labeled DNA was digested with ApaI, and -425/-269 bp TH promoter fragment containing wild-type or mutated DSE (-356/-303 bp) was gel-purified. Single-stranded fragments generated by S1 nuclease were resolved on 6% denaturing polyacrylamide gels.

Figure 1. DSE1 supports a basal TH promoter activity in BAMC cells. (A)Schematic representation of TH promoter mutants. Numbers indicate position relative to the transcription start site (short arrow). Sequences of the wild-type and mutated DSE1 are shown. Long arrows indicate DSE1 arms. Boxes pGL2 and LUC indicate multiple cloning region of pGL2Basic and the reporter luciferase gene, respectively. In the RevL mutant the orientation of the left DSE1 arm (both strands) was reversed. (B)Promoter activities of individual mutants. Observations from four independent experiments, each in triplicate, were combined. Bars represent means " SEM. Luciferase activity was normalized to the intracellular content of transfected DNA and expressed as c.p.m./µg protein × picogram DNA (n = 12).

OsO₄-modifications were performed according to (39,40,41).One microgram of supercoiled DNA was incubated for 20 min at room temperature in 50 mM Tris-HCl buffer, pH 7.5, containing 100 mM NaCl, 2% pyridine and OsO₄ (1, 5 or 10 mM).DNA was ethanol-precipitated, digested with Acc65I and labeled with T4 polynucleotide kinase (top strand) or Klenow enzyme (bottom strand). The label was removed from one end by digesting DNA with ApaI. DSE1-containing fragments were treated with 1 M piperidine at 90°C for 30 min and analyzed on a sequencing gel.

Analysis of cruciform extrusion in supercoiled plasmids with cruciform-specific 2D3 antibody and electrophoretic mobility shift assay (EMSA) (24)

One hundred nanograms of a supercoiled plasmid DNA was incubated for 1 h at 4°C with or without 2D3 (42,43) in 10 mM HEPES, pH 7.8, 100 mM KC1, 1 mM EDTA and 5 mM MgC1₂. After addition of 2 µl of 10× restriction buffer, plasmid was digested for 10 min at room temperature with 5 U of each Acc65I and ApaI. Subsequently, 1 U of Klenow fragment and 10 µCi of [[alpha]-³²P]GTP were added to the reaction mixture, and incubation continued for 10 min. ³²P-labeled fragment I (162 bp containing -425/-269 bp region of TH promoter and 5 bp from pGL2_Basic polylinker) and fragment II (5809 bp of pGLTH with the remaining -268/+25 bp fragment of TH promoter) were resolved on 4% non-denaturing polyacrylamide gels for a 10 V/cm. Gels were dried and exposed overnight. Retardation of TH promoter fragments was not observed with control monoclonal antibodies (24).

In vitro DNase I footprinting of TH promoter DNA in supercoiled and linearized plasmids

Nuclear extracts were prepared from BAMC or TE 671 cells according to (44) as previously described (22), and the footprinting was performed according to (17,43,45,46). Linearized or supercoiled pGLTH plasmid (200 ng) containing the wild-type or RevL mutant TH promoter sequence was incubated with 5-60 µg of nuclear BAMC cell or TE 671 cell proteins in 50 µl reactions containing binding buffer (125 mM HEPES, pH 7.8, 250 mM KCl, 0.25 mM EDTA, 2.5 mM DTT, 25% glycerol), 5 µg BSA, 5 µg of sonicated salmon sperm DNA and 50 mM PMSF. After 20 min of binding at room temperature, DNase I was added (10 ng/µl) and samples were incubated for an additional 2 min. DNA digestion was stopped by adding 1 vol of DNase I stop buffer (10 mM HEPES, pH 7.6, 20 mM EDTA, 1% SDS, 5 µg/ml yeast tRNA). The DNA was immediately extracted with phenol:chloroform:isoamyl alcohol (24:24:1). After extraction, the DNA was precipitated with 0.1 vol of 3 M sodium acetate, pH 5.2 and 2.5 vol of ethanol, washed with 75% ethanol and resuspended in 49.5 µl of PCR mixture containing 5 µl of 10× PCR reaction buffer (Boehringer Mannheim), 4 µl of 10 mM dNTP, and ³²P-labeled primer (5[prime]-GCTCCAGCAGGCGTGTTCTGCAG-3[prime]) complementary to -250/-228 nt on the coding strand of the TH gene. The primer was labeled with [³²P]ATP (6000 Ci/mmol) using T4 polynucleotide kinase, and 2-6 × 10⁶ primer c.p.m. were used per PCR reaction. After a hot start (95°C, 5 min), 2.5 U of Taq DNA Polymerase (Boehringer Mannheim) were added followed by 40 PCR cycles (95°C for 30 s, 55°C for 30 s and 72°C for 1 min). After the final extension (72°C, 10 min), the samples were treated with 50 µl of DNase I stop buffer, extracted with phenol:chloroform, and precipitated by ethanol as described above. After precipitation, the samples were resuspended with 5 µl sequencing-loading buffer, denatured for 5 min at 100°C and electrophoresed on 6% sequencing gels. Sequencing markers have been synthesized from pHISi plasmid template (CH laboratories, Palo Alto, CA) using commercial sequencing primer and were resolved on the same gel. Gels were dried and exposed to Kodak XAR-2 film. Equivalent regions of different lanes were selected and scanned using the Beckman spectrophotometer DU-70. Intensities of protected regions in each lane were normalized to non-binding sequence (within 304/-299 nt) to correct for uneven loading.

Figure 2. Mapping DSE1 secondary structure with S1 nuclease and OsO4. (A) S1 nuclease-sensitive sites in TH promoter of pGLTH (sc, supercoiled and lin, linearized). (B)Compare lack of S1 nuclease sensitive sites in DSE1 of pRevL (lane 1) and nuclease-sensitive sites in pGLTH (lane 2) and (C) OsO4-reactive sites in pGLTH. (D)Cruciform structure of DSE1 deduced from S1 nuclease and OsO4 mapping. `v' indicates mismatched bases interrupting dyad symmetry; `*' indicates S1 nuclease-sensitive and OsO4-reactive bases. Numbers indicate nucleotides relative to the transcription initiation site.

DSE1 acts as a positive regulatory element in TH promoter in BAMC cells

Deletion analysis of the bovine TH gene promoter (-425/+25 bp relative to the transcription start site) revealed that the region from -425 to -300 bp is essential for the maximal expression of the reporter gene in transfected BAMC cells (24) and in human catecholaminergic cell line IMR322 (E.Kim and M.K.Stachowiak, unpublished observations). The search for the potential regulatory sequence revealed a 46 bp element (-352/-307 bp)with palindromic symmetry, which we designated as DSE1 (Fig. 1A). The deletion of -365/-301 bp, which includes DSE1 ([Delta]DSE1 promoter mutant), markedly reduced expression of luciferase in BAMC cells relative to pGLTH plasmid that contains -425/+25 bp TH promoter fragment (Fig. 1B).

To further determine the mechanism of DSE1 action, we constructed deletion mutants in which regions containing the right arm ([Delta]R) or the left arm ([Delta]L) of DSE1 were deleted. In the RevL mutant the orientation of the left DSE1 arm was reversed (Fig. 1A). The pGLTH containing the -425/+25 bp of the bovine TH gene or its mutants was transfected into BAMC cells. Deletion of the right DSE1 arm markedly reduced specific luciferase activity (Fig. 1B). A similar reduction was observed when the left arm and sequences directly upstream were deleted (mutant [Delta]L). Both arms are therefore required for the trans-activating function. Even though the sequences of both DSE1 arms were preserved, the RevL mutation also reduced TH promoter activity. Thus, symmetry is essential for DSE1 to support promoter function. As in BAMC cells, [Delta]R or [Delta]L mutations reduced promoter activity in transfected human IMR322 cells by 92 (n = 6) and 62% (n = 19), respectively, relative to pGLTH (not shown).

DSE1 extrudes cruciform structure in supercoiled TH promoter

A characteristic feature of sequences possessing dyad symmetry is their ability to adopt cruciform-like structures in supercoiled DNA (10,47). A hallmark of non-B-DNA structures is the presence of unpaired bases that can be detected by single-strand specific S1 nuclease (48). To determine whether structural transitions may occur within DSE1, we subjected supercoiled pGLTH to S1 nuclease analysis. The reaction products were end-labeled with ³²P and resolved on sequencing gels. We detected several S1-sensitive bases within DSE1 (Fig. 2A). These bases were not detected when pGLTH was linearized before digestion with S1 nuclease, indicating that the S1 nuclease-sensitivity requires DNA supercoiling(Fig. 2A). Furthermore, in supercoiled RevL pGLTH in which the mutation of DSE1 prevented intrastrand base-pairing, we detected no S1-sensitive sites in DSE1 (Fig. 2B).

In the cruciform that could be formed by the wild-type DSE1 some bases are mismatched and remain unpaired (Fig. 2D). The positions of the most prominent S1-sensitive sites match the locations of the mismatched bases dictated by the predicted structure (compare Fig. 2A and D). The imperfect cruciform structure derived from the S1 analysis contains five unpaired thymidine residues (four on the top and one on the bottom strand). Therefore, to confirm this structure we treated supercoiled pGLTH with osmium tetroxide, which forms stable adducts with unpaired pyrimidines, especially with thymidine (39). The cleavage of the OsO₄-modified DNA with piperidine confirmed that all five thymidine residues in DSE1 were OsO₄-reactive and hence unpaired (Fig. 2C). Additional evidence for the cruciform extrusion was provided by the interaction of DSE1 with the cruciform-specific monoclonal antibody 2D3, isolated and characterized in (42,43). In a modified EMSA (24,43), we incubated supercoiled pGLTH with 2D3 and digested it with Acc65I and ApaI. 2D3 reduced the electrophoretic mobility of the -425/-260 bp TH promoter fragment containing DSE1 retaining almost all of the DNA in a shifted band (Fig. 3, compare lanes 1 and 2 with lane 3). This indicates that DSE1 exists in a cruciform conformation in the majority of plasmid molecules. 2D3 also shifted the larger fragment of the pGLTH, which contains additional cruciforms (24). In pGLTH with a deleted DSE1, the smaller retarded complex did not form, indicating that it is generated by 2D3 binding to the DSE1 cruciform. In contrast, the larger pGLTH fragment remained supershifted, providing an internal control for 2D3 binding (Fig. 3). Thus, the results of S1-nuclease, OsO₄ and 2D3-EMSA assays showed that DSE1 adopts an imperfect cruciform conformation under torsional stress and exists as a cruciform in supercoiled DNA.

Figure 3. Binding of cruciform-specific antibody 2D3 to DSE1 in a supercoiled pGLTH. Arrow indicates a retarded complex of 2D3 and a -425/-269 bp TH promoter fragment (FP 1) from pGLTH. The -425/-269 bp fragment with deleted DSE1 (FP2) from [Delta]DSE1 pGLTH mutant was not retarded by 2D3. The slight difference in the intensity of the FP2 band may reflect small differences in efficiency of plasmid digestion and radioactive end-labeling in the presence of 2D3. In both reactions the larger fragment containing pGLTH, which extrudes additional cruciform(s) (24), was retarded by 2D3, providing an internal control for 2D3 binding.

Binding of nuclear proteins to DSE1 is affected by its structural transformation

Our results suggest that the disruption of intrastrand base-pairing preventing cruciform formation is responsible for DSE1 inactivation in DSE1 mutants. One mechanism by which structural transitions could affect promoter activity is by altering the interaction of transcriptional factors with the DNA. To determine whether DSE1 binds nuclear proteins and whether cruciform extrusion may affect protein binding, we performed DNase I footprinting on linearized pGLTH in which DSE1 is in linear duplex form (Fig. 4, left) and in supercoiled pGLTH in which DSE1 extrudes the cruciform (Fig. 4, middle). When linearized pGLTH plasmid was incubated with nuclear extracts from BAMC cells and digested with DNase I, the subsequent primer extension reaction revealed a short protected region at approximately -325 to -310 bp (Fig. 4, compare lanes 1 and 3, and 2 and 4). This finding indicated a direct protein binding to the right arm of DSE1 and was consistent with promoter inactivation by the [Delta]R mutation. However, the region encompassing the left DSE1 arm and sequences directly upstream showed no protein binding (Fig. 4, left) even though its deletion reduced promoter activity (Fig. 1B). Incubation with nuclear proteins induced DNase I hypersensitivity at -330 nt in the middle of DSE1 and at 307-309 nt in the right arm. DNase I hypersensitive bases were also detected in the 5[prime]-end of the left DSE1 arm (Fig. 4, left).

Figure 4. Binding of nuclear BAMC cell proteins to DSE1 in relaxed and supercoiled TH promoter, DNase I analysis. Footprints of top strand (see also Fig. 1A) are shown. `+' indicates 60 µg of nuclear proteins. Sequence markers were dideoxy chain termination reactions (not shown). Lanes 1, 3, 5 and 7 contain 50 000 c.p.m. of [<sup>32</sup>P]DNA. Lanes 2, 4, 6, 8 and 9-12 contain 200 000 c.p.m. of [<sup>32</sup>P]DNA. Protected region is designated by an oval. The arrows indicate the two arms of DSE1. (Left) Linear pGLTH plasmid was used as a target DNA. Nuclear proteins protected predominantly a short sequence (from -325 to -310 nt) in the left arm of DSE1. (Middle) Supercoiled pGLTH. Nuclear proteins protected the right arm of DSE1. (Right) Linear or supercoiled pRevL plasmid was used as target DNA. No protein binding to the RevL DSE1 mutant was detected. Protection of DSE1 in pGLTH was quantified by scanning selected sequences in the autoradiogram (see Materials and Methods). In relaxed DNA the ratios of signal intensity (nuclear extracts `+'/nuclear extracts `-') in region -342/-335 nt was 1.0 and in region -327/-321 nt was 0.05. In supercoiled pGLTH these ratios were 0.2 and 0.9, respectively.

The same BAMC cell extracts protected DSE1 with extruded cruciform in supercoiled pGLTH more efficiently than its linear-duplex isomer in relaxed plasmid (Fig. 4 compare lanes 1-4 and lanes 5-8). Also, the protected region changed from the short sequence in the right arm in linear-duplex DSE1 to include the entire left arm and the center of DSE1, and it extended to the right DSE1 arm in supercoiled DNA. This pattern of binding was consistent with the results of transfection experiments (Fig. 1B), which showed that both DSE1 arms support promoter activity.

RevL mutation abolished protein binding to both DSE1 in linear and in supercoiled plasmids demonstrating that proteins bind in a sequence-specific maner (Fig. 4, lanes 9-12).

Table 1. Effect of DSE1 and CRE mutations on TH promoter activity in TE671 cells

Plasmid	Luciferase activity (c.p.m./µg protein × picogram DNA)	% of pGLTH activity
pGLTH	431.21 + 15.36	100
pRevL	883.13 + 20.58	205
p[Delta]CRE	173.24 + 16.78	[numsp]40

Numbers represent mean " SEM luciferase activities of six transfected dishes from a representative experiment. Activities of TH promoter mutants were also normalized to the activity of pGLTH (100%). The RevL promoter mutant has a reversed left DSE1 arm (Fig. 1A); [Delta]CRE indicates the mutation of the CRE element in TH promoter (24). The effect of the [Delta]CRE mutation is shown for comparison.

DSE1 controls TH promoter activity in a cell-specific manner

Unlike catecholaminergic BAMC cells, TE 671 cells do not express detectable 2 kb TH mRNA levels (H.Peng and M.K.Stachowiak, unpublished observations). In TE 671 cells, the specific luciferase activity expressed from transfected pGLTH (431 c.p.m./µg protein × picograms DNA; Table 1) was 3-fold lower than in BAMC cells (1291 c.p.m./µg protein × picograms DNA; Fig. 1B). Also, unlike in BAMC cells, in which the RevL mutation inactivated the TH promoter, the RevL mutation increased TH promoter activity in TE 671 cells. In contrast, mutation of a CRE-like sequence that reduced promoter activity in BAMC cells by 60% (24) had a similar reducing effect on TH promoter activity in the TE 671 cells (Table 1). Thus, DSE1 acts as a transcriptional activator in BAMC and IMR322 cells and as a repressor in TE 671 cells.

Figure 5. Binding of nuclear TE 671 cell proteins to DSE1 in relaxed and in supercoiled DNA, DNase I analysis. Footprints of top strand (see also Fig. 1A) are shown. Linear (lanes 1 and 2) or supercoiled pGLTH plasmid (lanes 3 and 4) was used as a target in protein binding reactions (See Materials and Methods). Only supercoiled pRevL was used in footprinting assay (lanes 5 and 6). Nuclear extracts `+' indicates 50 µg of nuclear proteins. Sequence markers were dideoxy chain termination reactions (not shown). The arrows indicate the two arms of DSE1. The small oval indicates a partially protected region on relaxed DNA. The large oval indicates the region with the strongest protection on supercoiled DNA. DNA protection was quantified by scanning selected sequences in the autoradiogram (see Materials and Methods). In linear pGLTH, the ratios of signal intensity (nuclear extract `+'/nuclear extract `-') were 1.4 in -344/-340 nt region and 0.18 in -338/-332 nt region. In supercoiled pGLTH these ratios were and 0.1 and 0.16, respectively. No protein binding to the RevL DSE1 mutant was detected.

Nuclear extracts from TE 671 cells weakly protected a short sequence (from -328 to -335 bp) in the left DSE1 arm of linear pGLTH and induced DNase I hypersensitivity in the center of DSE1 at -330 nt and in the left and right arms of DSE1 (Fig. 5, compare lanes 1 and 2). Protein binding to DSE1 in supercoiled DNA (Fig. 5, lane 4) was more efficient and extended further than in linear DSE1, encompassing the entire left arm, bases upstream from DSE1 and the middle of DSE1. Although the RevL mutation did not alter the protein-binding sequence or its flanking nucleotides, it prevented protein binding to DSE1 and DNase I hypersensitivity (Fig. 5, lanes 5 and 6).

DISCUSSION

We have previously identified an upstream -425/-300 bp region in the bovine TH gene promoter that supports basal promoter activity in BAMC cells (24). In the present report we showed that a 46 bp DSE1 located at -353/-307 bp of the bovine TH promoter is essential for this function. It acts as transcriptional activator in TH-expressing BAMC and IMR322 cells and as a repressor in non-expressing TE671. The presence and location of the reverse palindrome are conserved in human and rat TH promoters (E.L.Kim, S.Maltchenko, M.K.Stachowiak, unpublished observations), further indicating the importance of DSE1 in the regulation of TH-promoter activity. Also, the deletion of DSE1 had the same reducing effect on bovine TH promoter activity in human IMR322 cells as in BAMC cells, indicating its consistent function in different species.

Deletion of the left or right DSE1 arm, and the RevL mutation reduced promoter activity in a similar manner demonstrating that both arms and dyad symmetry are essential for DSE1 function. DNA palindromes are often targeted by dimeric trans-acting factors (49). However, DNase I footprinting in relaxed plasmids showed that nuclear proteins bind only to one arm of the palindrome and could not explain why deletion or mutation of the opposite arm affected protein binding and promoter activity. Protein binding to linearized DNA may not reflect protein-DNA interactions that occur in vivo in supercoiled DNA of circular plasmids, viruses or topologically constrained chromosomal loops. Indeed, we found that nuclear proteins bound to DSE1 more efficiently in supercoiled TH promoter and that the protection included the arm that did not bind protein in relaxed DNA.

How may DNA supercoiling affect protein binding by DSE1? Dyad symmetry sequences endow DNA with the capacity to generate cruciforms through intrastrand base-pairing (47). Whether a cruciform is extruded depends on the length of the palindrome, the number of unpaired bases in the stem-loop structures and the degree of negative supercoiling (10). Cruciform structures arising from palindromes occur at physiological superhelical densities (50). S1 nuclease and OsO₄-analyses indicate extrusion of an imperfect cruciform by DSE1 in supercoiled but not in relaxed TH promoter DNA, confirmed by the binding of the cruciform-specific 2D3 antibody. Thus, the degree of native DNA supercoiling generated in the TH gene promoter in vivo in bacterial cells is sufficient to enforce extrusion of cruciform by DSE1.Near complete shift of DSE1 DNA by the 2D3 antibody observed in modified gel shift assay (Fig. 3) indicated that in most of plasmid molecules, DSE1 exists in a cruciform conformation.In general, the abilityof inverted repeats to extrude cruciforms increases with their length. Consistent with that relationship, we found that the shorter inverted repeat, DSE3 (-217/-187 bp), was only partially reactive with 2D3 in supercoiled pGLTH (24). This suggests that DSE1 extrudes cruciform more efficiently than DSE3 in supercoiled DNA (24).Since the movement of RNA polymerase II along the DNA template adds negative supercoils in the 5[prime] direction from the enzyme (46), the extrusion of cruciform by DSE1 in eukaryotic cells is even more likely.

One mechanism through which conformational transitions in DNA may affect promoter activity is by changing its interaction with trans-acting factors. In the bovine TH gene promoter, a palindromic sequence at -217/-187 bp (DSE3) extrudes a cruciform that overlaps with the AP1-binding site and acts as the AP1 repressor (24). A different mechanism may operate in DSE1. In its cruciform state, DSE1 was targeted by nuclear proteins more efficiently than the linear-duplex isomer. Also, in DSE1 cruciform, nuclear proteins protected longer sequences than in the linear duplex isomer. Formation of a new target site for nuclear proteins by DSE1 cruciform may explain why the deletion of the non-binding arm inactivates DSE1, and why the RevL mutation disrupts protein binding and promoter activity (i.e., the disruption of intrastrand base-pairing and cruciform formation may be responsible for the inactivation of DSE1).

An induction of DNase I hypersensitivity in DSE1 in relaxed DNA by nuclear proteins suggested that promoter-binding proteins also may alter the DNA conformation. The hypersensitive sites were absent in the RevL template indicating that the protein-induced structural changes may involve intrastrand base-pairing.

DNA cruciforms interact with specific cruciform-binding proteins (9,52,53). These proteins differ from the single-stranded DNA-binding proteins required for DNA replication, repair and recombination (54). Single-stranded DNA-binding proteins are present in high concentration in vivo and bind non-specifically to single-stranded DNA. Upon binding they destabilize secondary structure present on single-stranded DNA (47). However, the DNase I-hypersensitive sites observed in DSE1 in supercoiled plasmid in the absence of nuclear proteins persisted when nuclear BAMC cell or TE 671 extracts were added. This finding suggests that the binding of proteins to DSE1 does not erase the pre-existing cruciform structure. Furthermore, nuclear proteins protected extended regions of DSE1 cruciform besides the unpaired bases in the hairpin loop. Also, proteins from BAMC and TE 671 cells showed a distinct pattern of binding to DSE1, and our results indicate that they have specific trans-activator and repressor functions, respectively. The marked increase in protein binding after the transformation of linear DSE1 into cruciform may provide an effective switch that converts an inactive DSE1 into a protein-binding element that controls cell-specific expression of the TH gene.

In the pro-enkephalin gene, a new CRE site is created within the cruciform extruded in the promoter (17,18,55). Because of the presence of unpaired bases, this new site binds CREB with a higher affinity than the CRE in linear duplex DNA. A similar mechanism may operate in the TH promoter DSE1, which also extrudes an imperfect cruciform with stem bulges created by unpaired bases. Alternatively, the DSE1 cruciform could bind nuclear proteins that specifically recognize the cruciform structure. In recent years, a few proteins that bind specifically to non-B-DNA have been identified (2,41,56-59). Factors that target the DSE1 cruciform could belong to this new class of nuclear proteins. Their identification will foster our understanding of the function of the DNA structure in the regulation of gene expression and the mechanisms that determine TH promoter regulation specific for this gene.

ACKNOWLEDGEMENTS

We are very grateful to Dr M.K.Zannis-Hadjopoulos and Dr G.B.Price for the generous gift of 2D3 anti-cruciform antibody. We thank Ewa K.Stachowiak M.S. for the preparation of adrenal medullary cultures, Mr S.Freist and Mrs Terri Gonzales for preparing photographs, and Dr Shelley Kick for editing the manuscript. This work was supported by National Institutes of Health Grant (HL49376) and National Science Foundation (IBN-9411226) (to M.K.S.). F.M.E. was supported by studentship from National Science Foundation Research Experiences for Undergraduates (supplement to IBN-9411226).

REFERENCES

1. Dayn, A., Malkhosyan, S., Duzhy, D., Lyamichev, V., Panchenko, Y. and Mirkin, S. (1991) J. Bacteriol. 173, 2658-2664. MEDLINE Abstract

2. Glucksmann-Kuis, M.A., Xing, D., Markiewicz, P. and Rothman-Denes, L.B. (1996) Cell 84, 147-154. MEDLINE Abstract

3. Van Holde, K. and Zlatanova, J. (1994) Bioessays 16, 59-68. MEDLINE Abstract

4. Hanke, J.H., Hambor, J.E. and Kavathas, P. (1995) J. Mol. Biol. 246, 63-73. MEDLINE Abstract

5. Wittig, B., Dorbic, T. and Rich, A. (1989) J. Cell. Biol. 108, 755-764. MEDLINE Abstract

6. Peck, L.J. and Wang, J.C. (1983) Proc. Natl. Acad. Sci. USA 80, 6206-6210. MEDLINE Abstract

7. Gellert, M., O'Dea, M.H. and Mizuuchi, K. (1983) Proc. Natl. Acad. Sci. USA 80, 5545-5549. MEDLINE Abstract

8. Wells, R.D. (1988) J. Biol. Chem. 263, 1085-1098.

9. Iyer, V. and Struhl, K. (1995) EMBO J. 14, 2570-2579. MEDLINE Abstract

10. Mizuuchi, K., Mizuuchi, M. and Gellert, M. (1982) J. Mol. Biol. 156, 229-243. MEDLINE Abstract

11. Mizuuchi, K., Mizuuchi, M. and Gellert, M. (1982) J. Mol. Biol. 156, 229-243. MEDLINE Abstract

12. Panayotatos, N. and Wells, R.D. (1981) Nature 300, 611-615.

13. Klug, A. (1982) Cold Spring Harbor Symp. Quant. Biol. 47, 1215-1223.

14. Bell, D., Sabloff, M., Zannis-Hadjopoulos, M. and Price, G.B. (1991) Biochim. Biophys. Acta 1089, 299-308. MEDLINE Abstract

15. Ward, G.K., Shihab-e-Deen, A., Zannis-Handjopoulos, M. and Price, G.B. (1991) Exp. Cell Res. 195, 92-98.

16. Delic, J., Onclerq, R. and Moisan-Coppey, M. (1991) Biochem. Biophys. Res. Commun. 181, 818-826. MEDLINE Abstract

17. Spiro, C., Bazett-Jones, D.-P., Wu, X. and McMurray, C.T. (1995) J. Biol. Chem. 270, 27702-27710. MEDLINE Abstract

18. Spiro, C., Richards, J.P., Chandrasekaran, S., Brennan, R.G. and McMurray, C.T. (1993) Proc. Natl. Acad. Sci. USA 90, 4606-4610. MEDLINE Abstract

19. Wang, Z.-Y., Masaharu, N., Qiu, Q.-Q., Takimoto, Y. and Deuel, T.F. (1994) Nucleic Acid Res. 22, 457-464. MEDLINE Abstract

20. Wittig, B., Wolfl, S., Dorbic, T., Vahrson, W. and Rich, A. (1992) EMBO J. 11, 4653-4663. MEDLINE Abstract

21. Kumer, S.C. and Vrana, K.E. (1996) J. Neurochem. 67, 443-462. MEDLINE Abstract

22. Stachowiak, M.K. and Goc, A. (1992) In Holle G. and Wood, J.D. (eds), Advances In the Innervation of the Gastrointestinal Tract. Elsevier Science Publishers B.V., Amsterdam. pp. 251-263.

23. Goc, A. and Stachowiak, M.K. (1994) J. Neurochem. 62, 834-843. MEDLINE Abstract

24. Kim, E., Esparza, F.M. and Stachowiak, M.K. (1996) J. Neurochem. 67, 26-36. MEDLINE Abstract

25. Stachowiak, M.K., Jiang, H.K., Poisner A.M., Tuominen, R.K. and Hong, J.S. (1990) J. Biol. Chem. 265, 4694-4702. MEDLINE Abstract

26. Fung, B.P., Yoon S.O. and Chikaraishi, D.M. (1992) J. Neurochem. 58, 2044-2052. MEDLINE Abstract

27. Kim K.-S., Tinti, C., Song, B., Cubells, J.F. and Joh, T.H. (1994) J. Neurochem. 63, 834-842. MEDLINE Abstract

28. Lazaroff, M., Patankar, S., Yoon, S.O. and Chikaraishi, D.M. (1995) J. Biol. Chem. 270, 21579-21589. MEDLINE Abstract

29. Nankova, B., Hiremagalur, B., Menezes, A., Zeman, R. and Sabban, E. (1996) Mol. Brain Res. 35, 164-172. MEDLINE Abstract

30. Stachowiak, M.K., Goc, A., Hong, J.-S., Poisner, A., Jiang, H.K. and Stachowiak, E.K. (1994) Mol. Brain Res. 22, 309-319. MEDLINE Abstract

31. Banerjee S.A., Hoppe P., Brilliant M. and Chikaraishi, D.M. (1992) J. Neurosci. 12, 4460-4467. MEDLINE Abstract

32. Min, N., Joh, T.H., Kim, K.S., Peng, C. and Son, J.H. (1994) Mol. Brain Res. 27, 281-289. MEDLINE Abstract

33. Yoon, S.O. and Chikaraishi, D.M. (1992) Neuron 9, 55-67. MEDLINE Abstract

34. D'Mello, S.R., Turzai, L.M., Gioio, A.E. and Kaplan, B.B. (1989) J. Neurosci. Res. 23, 31-40. MEDLINE Abstract

35. Lyubchenko, Y.L. and Shlyakhtenko, L.S. (1997) Proc. Natl. Acad. Sci. USA 94, 496-501. MEDLINE Abstract

36. Stachowiak, M.K., Moffett, J., Joy, A., Puchacz, E., Florkiewicz, R. and Stachowiak, E.K. (1994) J. Cell. Biol. 127, 203-223. MEDLINE Abstract

37. McAllister, D.M., Isaacs H., Rogney, R., Peer, M., Au, W., Soukup, S.W. and Gardner, M.B. (1977) Int. J. Cancer 20, 206-213. MEDLINE Abstract

38. Hemmick L.M., Ross, M.E. and Evinger M.J. (1995) Neurosci. Lett. 201, 77-80. MEDLINE Abstract

39. Hanvey, J.C., Klysik, J. and Wells, R.D. (1988) J. Biol. Chem. 263, 7386-7396. MEDLINE Abstract

40. Palecek E. (1992) Methods Enzymol. 212, 139-154. MEDLINE Abstract

41. Klysik, J. and Shimizu, M. (1993) FEBS lett. 333, 261-267.

42. Frappier, L., Price, G.B., Martin, R.G. and Zannis-Handjopoulos, M. (1989) J. Biol. Chem. 264, 334-341. MEDLINE Abstract

43. Steinmetzer, K., Zannis-Hadjopoulos, M K. and Price, G.B. (1995) J. Mol. Biol. 254, 29-37. MEDLINE Abstract

44. Dignam, J.D., Lebovitz, R.M. and Roeder, R.G. (1983) Nucleic Acids Res. 11, 1475-1484. MEDLINE Abstract

45. Tugores, A. and Brenner D.A. (1994) Biotechniques 17, 410-412. MEDLINE Abstract

46. Tugores,A., Magness, S.T. and Brenner, D.A. (1994) J. Biol. Chem. 269, 30789-30797. MEDLINE Abstract

47. Murchie A.I. and Lilley, D.M. (1992) Methods Enzymol. 211, 158-180. MEDLINE Abstract

48. Frank-Kamenetskii, M.D. (1992) Methods Enzymol. 211, 180-191. MEDLINE Abstract

49. Boulikas, T. (1994) Crit. Rev. Gene Exp. 4, 177-321.

50. Vologodskii, A. (1992) In Boca Raton (ed.), Topology of circular DNA. CRC Press Florida.

51. Wu, H.-Y., Shyy, S.H., Wang, J.C. and Liu, L.F. (1988) Cell 53, 433-440. MEDLINE Abstract

52. Pearson, C.E., Ruitz, M.T., Price, G.B. and Zannis-Hadjopoulos, M. (1994) Biochemistry 33, 14185-14196. MEDLINE Abstract

53. Pearson, C.E., Zannis-Hadjopoulos, M., Price, G.B. and Zorbas, H. (1995) EMBO J. 14, 1571-1580. MEDLINE Abstract

54. Chase, J.W. and Williams, K.R. (1986) Annu. Rev. Biochem. 55, 130-136.

55. McMurray, C.T., Wilson, W.D. and Douglass, J.O. (1991) Proc. Natl. Acad. Sci. USA 88, 666-670. MEDLINE Abstract

56. Bianchi, M.E., Beltrame, M. and Paonesa, G. (1988) Science 243, 1056-1059.

57. Solaro, P., Greger, B. and Kemper, B. (1995) Eur. J. Biochem. 230, 926-933. MEDLINE Abstract

58. Waga, S., Mizuno, S. and Yoshida, M. (1990) J. Biol. Chem. 265, 19424-19429. MEDLINE Abstract

59. Weisman-Shromer, P. and Fry, M. (1993) J. Biol. Chem. 268, 3306-3312.

---

*To whom correspondence should be addressed. Tel: +1 602 406 3730; Fax: +1 602 406 4172; Email: mstacho@mha.chw.edu

---

This page is run by Oxford University Press, Great Clarendon Street, Oxford OX2 6DP, as part of the OUP Journals
Comments and feedback: www-admin{at}oup.co.uk
Last modification: 24 Mar 1998
Copyright© Oxford University Press, 1998.

CiteULike    Connotea    Del.icio.us    What's this?

This article has been cited by other articles:

				H. Kurahashi, H. Inagaki, K. Yamada, T. Ohye, M. Taniguchi, B. S. Emanuel, and T. Toda Cruciform DNA Structure Underlies the Etiology for Palindrome-mediated Human Chromosomal Translocations J. Biol. Chem., August 20, 2004; 279(34): 35377 - 35383. [Abstract] [Full Text] [PDF]

				L. A. Cunningham, A. G. Cote, C. Cam-Ozdemir, and S. M. Lewis Rapid, Stabilizing Palindrome Rearrangements in Somatic Cells by the Center-Break Mechanism Mol. Cell. Biol., December 1, 2003; 23(23): 8740 - 8750. [Abstract] [Full Text] [PDF]

				H. Peng, J. Moffett, J. Myers, X. Fang, E. K. Stachowiak, P. Maher, E. Kratz, J. Hines, S. J. Fluharty, E. Mizukoshi, et al. Novel Nuclear Signaling Pathway Mediates Activation of Fibroblast Growth Factor-2 Gene by Type 1 and Type 2 Angiotensin II Receptors Mol. Biol. Cell, February 1, 2001; 12(2): 449 - 462. [Abstract] [Full Text]

				J. L. Wiemels and M. Greaves Structure and Possible Mechanisms of TEL-AML1 Gene Fusions in Childhood Acute Lymphoblastic Leukemia Cancer Res., August 1, 1999; 59(16): 4075 - 4082. [Abstract] [Full Text] [PDF]

				L.B. ROTHMAN-DENES, X. DAI, E. DAVYDOVA, R. CARTER, and K. KAZMIERCZAK Transcriptional Regulation by DNA Structural Transitions and Single-stranded DNA-binding Proteins Cold Spring Harb Symp Quant Biol, January 1, 1998; 63(0): 63 - 74. [Abstract] [PDF]

This Article Abstract Print PDF (404K) Alert me when this article is cited Alert me if a correction is posted Services Email this article to a friend Similar articles in this journal Similar articles in ISI Web of Science Similar articles in PubMed Alert me to new issues of the journal Add to My Personal Archive Download to citation manager Search for citing articles in: ISI Web of Science (17) Request Permissions Commercial Re-use Guidelines for Open Access NAR Content Google Scholar Articles by Kim, E. L. Articles by Stachowiak, M. K. Search for Related Content PubMed PubMed Citation Articles by Kim, E. L. Articles by Stachowiak, M. K. Social Bookmarking          What's this?