C fi G base mutations in the CArG box of c-fos serum response element alter its bending flexibility Consequences for core-SRF recognition Josef Stepanek1,2,*, Michel Vincent3, , Pierre-Yves Turpin1, Denise Paulin2, Serge Fermandjian4,5, Bernard Alpert2 and Christian Zentz1 ´ ´ Laboratoire de Biophysique Moleculaire Cellulaire & Tissulaire, Universite Pierre et Marie Curie, Evry, France ´ ´ ´ Laboratoire de Biologie Moleculaire de la Differenciation, Universite Denis Diderot, Paris, France ´ LURE, Universite Paris-Sud, Orsay, France ´ ´ Departement de Biologie et Pharmacologie Structurales, Ecole Normale Superieure de Cachan, France Institut Gustave Roussy, Villejuif, France Keywords CArG box; c-fos; DNA bending; DNA dynamics; serum response element Correspondence C Zentz, Laboratoire de Biophysique ´ Moleculaire Cellulaire & Tissulaire, ´ Universite Pierre et Marie Curie, CNRS UMR 7033, GENOPOLE Campus 1, rue ` Henri Desbrueres, 91030 Evry Cedex, France Fax: +33 69 87 43 60 Tel: +33 69 87 43 52 E-mail: zentz@ccr.jussieu.fr *Permanent address Charles University, Faculty of Mathematics and Physics, Prague, Czech Republic  Present address ´ IBBMC, Universite Paris-Sud, Orsay France (Received 22 December 2006, revised 20 February 2007, accepted March 2007) doi:10.1111/j.1742-4658.2007.05768.x By binding to the CArG box sequence, the serum response factor (SRF) activates several muscle-specific genes, as well as genes that respond to mitogens The core domain of the SRF (core-SRF) binds as a dimer to the CArG box C)5C)4A)3T)2A)1T+1T+2A+3G+4G+5 of the c-fos serum response element (SREfos) However, previous studies using 20-mer DNAs have shown that the binding stoichiometry of core-SRF is significantly altered by mutations C)5 fi G (SREGfos) and C)5C)4 fi GG (SREGGfos) ` of the CArG box [A Huet, A Parlakian, M-C Arnaud, J-M Glandieres, P Valat, S Fermandjian, D Paulin, B Alpert & C Zentz (2005) FEBS J 272, 3105–3119] To understand these effects, we carried out a comparative analysis of the three 20-mer DNAs SREfos, SREGfos and SREGGfos in aqueous solution Their CD spectra were of the B-DNA type with small differences generated by variations in the mutual arrangement of the base pairs Analysis by singular value decomposition of a set of Raman spectra recorded as a function of temperature, revealed a premelting transition associated with a conformational shift in the DNA double helices from a bent to a linear form Time-resolved fluorescence anisotropy shows that the fluorescein reporter linked to the oligonucleotide 5¢-ends experiences twisting motions of the double helices related to the interconversion between bent and linear conformers The three SREs present various bent populations submitted, however, to particular internal dynamics, decisive for the mutual adjustment of binding partners and therefore specific complex formation Specific binding of the serum response factor (SRF) to the serum response element (SRE) requires a consensus sequence CC(A ⁄ T)6GG, the CArG box [1–7] The transcriptional activity of a number of CArG-dependent genes is associated with SRF-binding activity [8–14] The c-fos gene contains a single high-affinity CArG box, whereas many muscle-specific genes contain two or more CArG boxes However, these carry substitutions with G or C nucleotides within their (A ⁄ T) domain, thus lowering the affinity [15–19] Abbreviations core-SRF, core domain of the serum response factor; SRE, serum response element; SRF, serum response factor; SVD, singular value decomposition FEBS Journal 274 (2007) 2333–2348 ª 2007 The Authors Journal compilation ª 2007 FEBS 2333 Implication of CArG sequence in SRE flexibility J Stepanek et al Strong-affinity CArG boxes are likely to bind SRF constitutively and genes appear to be regulated primarily during the post-SRF binding step, owing to interactions with accessory proteins [20] Weaker affinity CArG boxes may offer additional control through a mechanism that influences SRF binding, i.e by mutual combined interactions of CArG boxes and accessory proteins [21,22] The core domain of the SRF (core-SRF) binds to the CArG box as a homodimer [7,23,24] The specific core-SRF–SREfos complex is characterized by the particular properties of the minor groove in the (A ⁄ T) domain and its flanking G:C base pairs The SRF causes the SRE to bend $ 70° The role of this bending in specific recognition has been emphasized [23,25,26] The efficiency and specificity of SRF-dependent transcription may vary due to changes in the CArG box sequence [22] To understand the origin of these effects this study focuses on the three 20-mer oligonucleotides: SREfos, SREGfos and SREGGfos The SREfos sequence, 5¢-d(GGATGTC)5C)4A)3T)2A)1 T+1T+2A+3G+4G+5ACAT)-3¢, embodies the native CArG box of the c-fos enhancer (CArG box numbered) [2,27], whereas SREGfos carries the single C)5 fi G mutation and SREGGfos the double C)5C)4 fi GG mutation within their CArG box A previous report has shown that the parent SREfos bound a core-SRF homodimer, whereas the single mutant SREGfos and the double mutant SREGGfos bound one and four monomers (on average), respectively [7] This highlights the role of the base sequence at the border of the A ⁄ T track in the specific complex assembly and functional organization How mutations affect binding of the core-SRF and generate a lack of defined stoichiometry is an open question Thus, we carried out a comparative analysis of the three oligonucleotides using Raman, CD and fluorescence spectroscopies in order to detect their mutual structural, electrostatic and dynamical differences CD and Raman are sensitive to small structural changes [28,29] In addition, Raman scattering is a powerful means of clearing up the various sensitivities of the nucleic acid chains to temperature [30,31] Fluorescence studies require a fluorophore reporter, such as fluorescein, chemically fixed to the oligonucleotides The fluorescein fluorescence signal arises from the overlapping emissions of the mono- and dianionic protolytic states [32], which are sensitive to the electric charge distribution on the DNA Chain DNA dynamics have been extensively studied [33–37] DNA is intrinsically flexible, but this flexibility varies from one DNA to another [38] To date, little is known about the relationships between the ability of DNA to bend and its effects on protein binding Previous studies have 2334 shown that association of core-SRF with SREfos reduces the flexibility of each partner, suggesting a strong role for dynamics in the adjustment of protein–DNA contacts and thereby the specificity of the complex formation [7] Time-resolved fluorescence anisotropy decays of the modified and native fluoresceinated SREs allow us to assess differences in dynamics among the three oligonucleotides The results highlight the strong relationships between the base sequence, DNA bending, interactions with water molecules and the internal dynamics in the specific attachment of core-SRF to SREfos Results Electric charge distribution in SRE containing oligonucleotides at 10 °C The electric charge distribution along the phosphate backbone plays a crucial role in the recognition of DNA by proteins [39] Certain changes within this distribution can affect the fluorescence emission of fluorescein linked to the oligonucleotide [32] In solution at pH 8.5, fluorescein exists as an equilibrium of mono- and dianionic forms Upon excitation at 490 nm, the unlinked fluorescein fluoresces with a maximum at 516 nm The emission spectrum shifts to 520 nm when fluorescein is conjugated to SREfos (Fig 1), the electrostatic potential of DNA generating a new equilibrium between the mono- and dianionic populations of the fluorescein [7] By contrast, the mutations performed in the native SREfos sequence not affect the fluorescein emission profile indicating that the fluorescent reporter experiences almost the same environment in SREfos, SREGfos and SREGGfos The electric charges cannot, therefore, be considered responsible for the differences in stoichiometry observed previously between the complexes of SREfos, SREGfos and SREGGfos formed with the core-SRF Interactions between neighboring bases of the SRE oligonucleotides at 10 °C The CD spectra of oligonucleotides are influenced by both the base composition of the nearest neighbor and the mutual arrangement of bases [28] The CD spectra of SREfos, SREGfos and SREGGfos, recorded at 10 °C, are of the B-DNA family with a positive band centered at 272 nm and a negative band close to 250 nm (Fig 2) Slight differences can be assigned to small changes in local interactions introduced by the mutations FEBS Journal 274 (2007) 2333–2348 ª 2007 The Authors Journal compilation ª 2007 FEBS Implication of CArG sequence in SRE flexibility Fluorescence intensity (a.u.) J Stepanek et al 500 520 540 560 580 Wavelength (nm) Fig Conjugation effect of SRE oligonucleotides on fluorescein fluorescence emission Fluorescence emission spectrum of fluorescein (—) Fluorescence emission spectrum of fluorescein conjugated to SREfos (e) Both spectra are normalized The fluorescence emission spectra of fluorescein conjugated to the three oligonucleotides are identical Spectra obtained at 10 °C with an excitation wavelength at 490 nm Ellipticity (m.deg.) 200 240 280 -2 Wavelength (nm) Fig CD spectra of SREfos (e), SREGfos (h) and SREGGfos (n) Temperature 10 °C Ellipticity is expressed in millidegrees Optical path length 0.1 cm Oligonucleotide concentration: 10)6 M Basic character of the temperature effect on SRE oligonucleotides: singular value decomposition analysis of Raman spectra The Raman spectra of the three duplexes are sensitive to temperature variations between 10 and 65 °C To find out the basic character of these changes, each set of spectra was statistically treated by means of singular value decomposition (SVD) [40] SVD outputs were similar for the three SREs Those for SREfos are given in Fig A factor dimension of 3, means that all Raman spectra obtained between 10 and 65 °C can be expressed from three spectral components (Fig 3) The first component, S1, is an invariable spectral residuum with an almost constant V1 contribution in each Raman spectrum The other two components, S2 and S3, account for two types of change induced by temperature Their contributions, V2 and V3, reveal two kinds of temperature processes separated by a boundary between 30 and 40 °C: V2 and V3 exhibit an inflexion and a minimum While the second dimension (V2, S2) shows spectral features that are common for both transitions, the third dimension (V3, S3) reflects the differences between them Above 40 °C, spectral changes are related to the melting of the duplexes V2 and V3 show a parallel increase with temperature and the changes induced by temperature are given by the summation of the spectral components S2 and S3 The most significant changes include (Fig 3): a decrease in the intensity of the Raman bands of deoxyribose phosphate backbone typical of B-type structures [789, 838, 891 (893), 1092 cm)1] [29,30,41] and of some bands characteristic for 2¢-endo ⁄ anti conformation of deoxynucleotides [671 (dT), 681 (dG), 1255 (dA, dC), 1338 cm)1 (dA)] [41–44] By contrast, there is an increase in the Raman bands at 729 (dA), 1238 (dT), 1303 (dA), 1488 (dA, dG), and 1667 cm)1 (dT) in response to base destacking in the oligonucleotides [29,30,41–43,45–49] Between 10 and 25 °C, within the premelting domain, spectral changes are reflected in a gradual increase in the contribution of V2 and a simultaneous decrease in the contribution of V3 V3 is normalized and its amplitude looks very similar for all three duplexes in a temperature region where the premelting is dominant By contrast, V2 is mainly normalized according to melting and weak variations between the three duplexes can be seen during premelting In this study, we are interested in the premelting transitions because they reveal subtle variations without dissociation of the DNA strands Changes occurring in SRE oligonucleotides between 25 and 10 °C B-DNA conformation of duplexes at 25 °C At 25 °C, the three duplexes display very similar Raman spectra Several peaks can be assigned to known characteristic vibrational bands (Figs 4–6, upper) Bands from the deoxyribose-phophate backbone (790, 838, 1093 and 1421 cm)1) at a position diagnostic of the B-type conformation [41,42] are identified together with bands from deoxyoligonucleotides [681 (dG), 750 (dT), 1255 (dC) and 1339 cm)1 (dA)] related to the C2¢-endo ⁄ anti conformation [41,42,44] The resemblances between the spectra FEBS Journal 274 (2007) 2333–2348 ª 2007 The Authors Journal compilation ª 2007 FEBS 2335 Implication of CArG sequence in SRE flexibility J Stepanek et al Fig Results of the factor analysis applied to the set of temperature Raman spectra of SREfos Raman spectrum Yi at each temperature is decomposed into M independent subspectra Sj Upper: (left) Singular values Wj evaluating statistical weight of individual spectral components Sj, (right) residual errors for various numbers of considered spectral components M Both panels show that the true factor dimension, i.e the minimum number of spectral components sufficient to approximate all Raman spectra, is Middle: Relevant spectral components Sj, j ¼ 1, 2, Lower: Coefficients Vij, j ¼ 1, 2, 3, indicating the relative contribution of each spectral component Sj into the spectrum Yi Spectral components (S1, S2, S3) and coefficients (V1, V2, V3) are normalized so that the sum of their squares over spectral points or temperature, respectively, is equal to Dashed lines indicate marker bands of the duplex melting, observable as coincidently oriented peaks in the both S2 and S3 spectral components 2336 FEBS Journal 274 (2007) 2333–2348 ª 2007 The Authors Journal compilation ª 2007 FEBS J Stepanek et al Fig Raman spectrum of SREfos at 25 °C and the effect of a decrease in temperature to 10 °C Upper: Spectrum at 25 °C Lower: Temperature effect on the Raman spectrum: spectrum at 10, 15 or 20 °C minus spectrum at 25 °C The intensity scale is the same in Figs 4–6 Fig As Fig 4, but for SREGfos indicate that the three oligonucleotides have very similar B-DNA conformations Spectral changes in the three duplexes between 25 and 10 °C The effects of decreasing the temperature from 25 to 10 °C are illustrated by the difference Raman spectra Implication of CArG sequence in SRE flexibility Fig As Fig 4, but for SREGGfos (Figs 4–6, lower) For the same oligonucleotide, the shape of the difference spectra between two temperatures is conserved When we compare the shape of the spectra from one oligonucleotide with the two others, high levels of similarity are also apparent Essentially, the band intensities and a few band positions vary slightly Spectral conservation allows us to make a common analysis of the temperature effect on the three duplexes, in agreement with the similarity of the results provided by their respective SVD analysis Changes of intensity and position of the Raman bands are given in Table Effect of the temperature decrease on base stacking and backbone geometry The 790 ⁄ 784 cm)1 doublet undergoes both an upshift of its 784 cm)1 component and an increase in the intensity of its 790 cm)1 component, thus expressing changes in the geometry of the phosphodiester group, and ⁄ or in the conformation of deoxycytidine or deoxythymidine, these becoming closer to the 2¢-endo ⁄ anti geometry [41,42,47] The differential profile around 1339 cm)1 shows that the corresponding adenine band is upshifted to its position of 2¢-endo ⁄ anti conformation [41] For cytosine, the shift in the 1255 cm)1 band to 1265 cm)1 very probably indicates a change in deoxynucleoside sugar pucker from the C3¢-endo ⁄ anti family to the C2¢endo ⁄ anti family [41–43] The shift in the 838 cm)1 band toward higher wavenumbers, though moderate, is generally interpreted as a sign of minor groove narrowing [29,41,45] Wavenumber upshift can be also seen for the sugar vibration at FEBS Journal 274 (2007) 2333–2348 ª 2007 The Authors Journal compilation ª 2007 FEBS 2337 Implication of CArG sequence in SRE flexibility J Stepanek et al Table Temperature-induced change in the Raman spectra of SREfos, SREGfos and SREGGfos and difference in Raman spectra between SREfos and SREGfos and between SREfos and SREGGfos Peak position at 25 °Ca,b Effect of temperature decrease from 25 to 10 °Ca,c 671 681 729 750 m m s m fl › › › 784 m, sh Þ 790 vs › 838 m Þ › › › 1057 w, as 1093 vs › Ü 1144 w 1178 w, as › 1149 1213 w, br, sh 1240 w, sh 1255 s › › 1265–1269 1294 m,sh 1303 s Difference spectrum SREfos-SREGGfos ›Þ 893 w 924 w, br, as 973 w, br 1013 m, br Difference spectrum SREfos-SREGfos › 662 691 731 754 10 °C fl c 25 °C fl fl 755 › 780 10 °C 25 °C fl fl fl 755 › 780 › 780 bk O-P-O str + dr, dT [41] bk O-P-O str [41] dr C2¢H2 rock [29,30] dr ring str [45] T C6H op-def [45], bk [42] G NH def [44], T CH3 rock [45], dr at 1003 [43] bk C-O str [29,30] PO2– sym str [41] › 928 Ü › 926 › 885 › 928 Ü › 928 › 1006 › 1006 › 1006 › 1005 › › Ü › 1218 fl 1186 › 1218 › 1257 › 1298 › 1299 › 1299 fl 1321 › 1258 fl › 1270 fl 1322 fl 1361 fl 1364 › 1379 fl 1361 fl 1363 fl 1396–1403 1510 m 1532 w, sh 1577 vs 2338 › › ›Þ dT [42,45] dT [42,45]; dG [29,43], dC [51] dT, dA [41,42]; dG [43] dT [42], dC [43] dC, dA, dT [42], also dG in [51] dC [42,51] dA, dT [42,45] dG [43,44] dA, dG [41,42] dG [44] T CH3 def [45]; dA, dG [42] B-DNA [41], sensitive to electrostatic environment [46] dT 2¢-endo ⁄ anti at 1208 [41] T hypochromic [47] dC 2¢-endo ⁄ anti at 1255, shift to 1265 for 3¢-endo ⁄ syn [41] against 2¢-endo ⁄ anti at 1268 [43]; signature of adenine non Watson–Crick bonding [45] dA hypochromic [30] dG 2¢-endo ⁄ syn [44]) dA 2¢-endo ⁄ anti at 1339, 3¢-endo ⁄ anti at 1335 [41]; dG 2¢-endo ⁄ anti at 1336 [44] dG 2¢-endo ⁄ anti [44] intensity increase in hydrophobic environment of T methyl [29,47] 1402 ị ị dT 2Â-endo anti at 665 [41] dG 2¢-endo ⁄ anti at 684 [41,44] hypochromic [45,49] dT 2¢-endo ⁄ anti at 748, 3¢-endo ⁄ -anti at 745 [41] dC 2¢-endo ⁄ anti at 782, 3¢-endo ⁄ anti at 780 [41] B-DNA g––g– of a ⁄ f torsion dT 2¢-endo ⁄ anti [41,47] B-DNA, exact position sensitive to minor-groove dimension [41,45] B DNA, sensitive to premelting [30] sensitive to B-B¢ transition [30] dC [41,42] Þ s w, br w vs, as dT, dA [41,42] dG [41,42] A breath [41,42] dT [41] › fl 1321 1421 1444 1462 1488 Significanced › › › 1299 1375 vs Assignmentd,e › 780 fl 1339 vs, br c fl fl fl 1580 fl 1583 fl fl 1574 dr C5¢H2 def, dA [42] dr C5¢H2 def [42] dr C2¢H2 def [42] G imi ring [46], dA, dT [42] dA, dC [42] dC, dG [42] dG, dA [42] B-DNA [41] hypochromic [41,47]; N7 bonding to guanine causes intensity decrease [49] and frequency downshift [47] upshift with A N7 bonding [45] G, A hypochromic [41,45] FEBS Journal 274 (2007) 2333–2348 ª 2007 The Authors Journal compilation ª 2007 FEBS J Stepanek et al Implication of CArG sequence in SRE flexibility Table Continued Peak position at 25 °Ca,b Effect of temperature decrease from 25 to 10 °Ca,c Difference spectrum SREfos-SREGfos 10 °C c 25 °C Difference spectrum SREfos-SREGGfos 10 °C c 25 °C 1602 w, sh 1652 m, sh Assignmentd,e › 1662 › 1695–1720 › 1699 T C2O str [42,48] › 1730 1668 s, br dC [42], G N1H def [41] T (C4O ⁄ C5C6) str [42], dC [43] T (C4O ⁄ C5C6) str [42,48] › 1711 › 1736 dG: CO str [44] › 1662 Ü Significanced shift to 1662 in case of extra H-bonding at C ¼ O [41,45] at 1689, shift to 1681 in case of extra H-bonding at C2 ¼ O [41,45] variable position 1686–1722 [44] Common characteristics for the three DNA duplexes b Peak positions are in wavenumber units (cm)1) Numbers in bold correspond to well-resolved bands; precision of the peak position ± cm)1 Numbers in standard type correspond to shoulders, asymmetrical or partly overlapped bands, and also to peaks in difference spectra; precision of the peak position ± cm)1 Added are basic characteristics of Raman band intensities: w ¼ weak, m ¼ medium, s ¼ strong, vs ¼ very strong, sh ¼ shoulder, br ¼ broad, as ¼ asymmetric c Symbols: › intensity increase, fl intensity decrease, Þ upshift of vibrational frequency, Ü downshift of vibrational frequency If the intensity increase or decrease in the difference spectrum is not pronounced exactly at the frequency corresponding to the basic Raman band position (first column), the position of the peak or nick in the difference spectrum is indicated d Abbreviations: A, C, G, T ¼ adenine, cytosine, guanine, thymine; dA, dC, dG, dT ¼ deoxynucleotide containing given nucleobase; bk ¼ backbone; dr ¼ deoxyribose e In case of overlapping Raman bands of several vibrational modes, the dominating mode is underlined Abbreviations for vibrational modes: str ¼ stretching, def ¼ deformation, breath ¼ breathing, rock ¼ rocking, op ¼ out-of-plane, sym ¼ symmetric a 893 cm)1 [29] The intensity increase for the 731 (729), 754 (750), 1306 (1303), 1379 (1375), 1490 (1488), 1584 (1577), 1662 (1668) and 1695–1730 cm)1 (dT) bands results from partial base unstacking affecting mainly adenine and thymine, and to a lesser degree also guanine [30,41,42,44–49] Globally, Raman bands related to the sugar–phosphate backbone conformation and to base-stacking reflect conformational changes taking place in various regions of the DNA duplexes The changes induced in our spectra by the decrease in temperature from 25 to 10 °C are similar to those resulting from the formation of a sharp bend in the DNA octamer duplex (HMG box) due to binding of the human SRY–HMG protein The decrease in temperature results in striking similarities between both Raman signatures (Figs 4–6, lower) [47,50] We may therefore conclude that SREfos and its two mutants exhibit, at 10 °C, a large population of bent conformers The bend is not limited to the central (A ⁄ T) sequence of the CArG box, but includes the bordering G ⁄ C base pairs, because guanine and cytosine signals (1488, 1578 cm)1 and 780, 1257, 1299 cm)1, respectively) are also affected [41,42,46,51] The structural adjustment resulting in a bent population at 10 °C underlies a more favored linear B form at higher temperatures The increase in intensity of the 926 (924), 1444 and 1462 cm)1 vibrational bands of deoxyribose and of the 790 and 1056 cm)1 bands of backbone reflects the disappearance of this linear population [29,30,41,42,45] The increase in both well-resolved bands at 1444 and 1462 cm)1 correlates with a broad band around 1400 cm)1 at 25 °C of about the same integral intensity, indicating a larger population of linear conformers The upshift and increase in intensity of the peak at 838 cm)1 suggest that the backbone conformation is altered to the detriment of a more canonical B form [41,45] Effect of a decrease in temperature on hydrogen-bond interactions and hydration From 25 to 10 °C, numerous base vibrations exhibit spectral shifts indicating changes in the hydrogen bond array However, these not concern regular Watson– Crick hydrogen bonds The upshift of the adenine bands at 1510 cm)1 (sensitive to binding at N7) and 1577 cm)1, like that of the guanine band at 1488 cm)1 (also sensitive to interaction at N7), are signs of hydrogen-bond formation [42,45–47,49] The downshift of the 1668 cm)1 band to 1662 cm)1 is connected with a change in hydrogen-bond interaction at the O4 of thymine [41,42,45,48] These changes can be assigned to a redistribution of water molecules or hydrated ions on the above-mentioned base This is in accordance with the weak wavenumber downshift of the PO2– FEBS Journal 274 (2007) 2333–2348 ª 2007 The Authors Journal compilation ª 2007 FEBS 2339 Implication of CArG sequence in SRE flexibility J Stepanek et al symmetric stretching vibration (1092 cm)1) expected to be sensitive to solvent charge interactions in the environment of phosphate groups [46] Effect of mutations Even though the temperature difference spectra look similar between one oligonucleotide and the other two (Figs 4–6, lower), their mutual differences reveal some disparities These are visible at 10 and 25 °C in the spectra shown in Figs and 8, respectively, and in Table At a given temperature, spectra of SREfos and SREGfos bearing one mutation are very similar, but they differ much more significantly from the spectrum of SREGGfos bearing two mutations Effect on G:C base pairs Fig As Fig 7, but at 25 °C As expected, the mutations entail visible, local conformational changes between the native C)5C)4 and single mutated C)5G)4 steps (SREGfos), and the double mutated G)5G)4 step (SREGGfos) The main effect of the mutations concerns the region of the two G:C base pairs, whose orientation is reversed There are signs of increased intensity for several guanosine signals (troughs at 679, 1321, 1361, 1488 and 1578 cm)1) [41–44,46,48], including the markers of deoxyguanosine 2¢-endo ⁄ anti conformation (679 and 1361 cm)1) and also the 1321 cm)1 band considered to be a 2¢-endo ⁄ syn conformation marker [44]; the increased intensity of several of these bands reflects increased unstacking of the guanine residue By contrast, several positive peaks in the difference spectra (780, 1257 and 1299 cm)1) are attributable to a decreased cytidine intensity [41,42,51] They indicate that, in the case of cytidine, the mutation causes better stacking and also reduces the probability of the 3¢-endo ⁄ anti conformation (the 780 cm)1 band) [41] In the spectral differences at 10 and 25 °C the mutational effects are conserved for the guanosine bands, whereas they are substantially weaker at increased temperature for the cytidine bands Effect on hydrogen-bond interactions, hydration and stability of the various SREs Fig Difference in Raman spectra at 10 °C between SREfos and SREGfos and between SREfos and SREGGfos The intensity scale is the same as in Figs 4–6 2340 At 10 °C (Fig 7), the negative band at 755 cm)1 attributed to the deoxythymidine 2¢-endo ⁄ anti conformation appears somewhat more pronounced in the mutated versions [41] The two deoxyribose vibration bands (positive peaks at 885 and 928 cm)1) become less intensive in both mutant spectra [29,30,45] For the double mutant SREGGfos, the simultaneous upshift of the 1668 cm)1 band suggests a weakening of the extra hydrogen bonding of the thymine carbonyl with the surrounding water molecules [41,42,45,48] Because no bands appear around 1093 cm)1 the electrostatic environment of the three duplexes cannot be distinguished [46] At 25 °C (Fig 8), the difference in the Raman spectra between the oligonucleotides increases The different intensities of the bands at 1402 cm)1 and at 790, 838, 927, 1056 cm)1 of the deoxyribose and the backbone [29,30,41,45] reflect the relative disappearance of FEBS Journal 274 (2007) 2333–2348 ª 2007 The Authors Journal compilation ª 2007 FEBS J Stepanek et al Implication of CArG sequence in SRE flexibility the bend population for the benefit of the linear one Concurrently, the bands from extra hydrogen bonding at thymine (1662, 1699 cm)1) [41,42,45,48] amplified by the increase in temperature, varies with the band at 1093 cm)1 (positive peak at 1086 cm)1, trough at 1098 cm)1), most probably due to modified interactions between the (A ⁄ T) domain and solvent molecules These changes concern mainly the SREGGfos and to a lesser extent the SREGfos Thus, an increase in temperature decreases the thermal stability of the bent form in the order: SREfos < SREGfos < SREGGfos The bent structure of SREfos is the most stable and preserved of the three duplexes, whereas the double mutation brings about a higher instability of that structure Internal dynamics of SRE helices The dynamics of the three SRE oligonucleotides were assessed using time-resolved fluorescence anisotropy decays with the fluorescein group fixed at the 5¢-end as a fluorescence reporter During the lifetime of its excited state (4 ns), the fluorescein group is involved in several motions: rotation as a whole, together with the internal motions of the oligonucleotide; and the proper rotations of the fluorophore around its link with the oligonucleotide Correlation times for the multiexponential anisotropy decays with their relative proportions are shown in Table The shortest correlation time (i.e F ¼ 0.4 ns) carries the strongest weight in the composite decay This correlation time is linked to the time of fluorescein rotation around its link with the oligonucleotide The correlation time for rotation of the SRE molecule as a whole, estimated to be 10 ns from hydrodynamic measurements [7,52], was hard to Table Relation between parameters of the fluorescence anisotropy decays of fluorescein labeling the various SRE oligonucleotides and the number of bound core-SRF monomers at 10 °C Fa ns (± 0.1 ns) SREGfos SREfos SREGGfos a bb % (± 2%) 0.4 3.9 0.4 3.2 0.4 1.8 86 14 88 12 82 18 m ¼ Uc 10 Hz detect in our experiments In any case, the fast depolarization process due to fluorescein motions prevents monitoring of the entire oligonucleotide rotation Because the fluorescent reporter experiences the same environment for the three oligonucleotides, we conclude that the longest correlation time reflects the internal dynamics of helix strands that drive fluorescein with them The longest correlation time for SREfos, i.e F ¼ 3.2 ns, slows to F ¼ 3.9 ns in SREGfos, whereas the double mutation shortens it to F ¼ 1.8 ns in SREGGfos The inverse of the correlation time (1 ⁄ F) represents the twisting oscillation frequency (m) of the double helix The oscillation frequency increases in the order (Table 2): SREGfos < SREfos < SREGGfos Table also gives the statistical weight (b) for the longest correlation times which increases in the order: SREfos < SREGfos < SREGGfos For each oligonucleotide, this weight decreases when the temperature increases from 10 to 30 °C (not shown), indicating a lower population that depolarizes at higher temperature Because the population of the bent form decreases at higher temperature, we must assume that the linear form does not give a detectable depolarization signal Thus, fluorescence anisotropy decay mainly detects the helix twisting of the bent form offering enough thermal amplitude motions In addition, b-value and thermal instability of the bent form detected using differences in Raman spectra between the oligonucleotides increase in the same order Discussion The C fi G mutations at the )5 and )4 positions of the CArG box alter the binding stoichiometry in a dramatic manner [7] Here we show that, at 10 °C, such mutations not affect electric charge repartition along the oligonucleotides and preserve the same B-DNA conformation Essentially, the interactions at the mutated positions are modified together with the arrangement of water molecules and the internal dynamics Nd 260 ± 10 310 ± 10 560 ± 20 %4 F, correlation time The longest correlation time characterizes the internal motion of the DNA duplex b b, weight of the exponential component c m, oscillation frequency d N, number of core-SRF monomer bound to DNA fragment [7] Premelting effect on the equilibrium of the bent linear form The premelting transition has been studied in detail by Raman spectroscopy for alternating [poly(dA–dT)]2 and homogenous poly(dA):poly(dT) sequences [30,45] The similarity to the effects of temperature on our Raman spectra emphasizes its influence on the six central (A ⁄ T) base pairs of the CArG boxes Detailed analysis of Raman spectra has confirmed that the FEBS Journal 274 (2007) 2333–2348 ª 2007 The Authors Journal compilation ª 2007 FEBS 2341 Implication of CArG sequence in SRE flexibility J Stepanek et al premelting transition conserves the basic local conformation features of B-DNA (A ⁄ T)-rich sequences have been found to be highly polymorphic and depend strongly on the temperature [53,54] Indeed, the change in the array of the hydrogen bonds at thymine of SREGGfos is probably a sign of perturbation in the hydration scheme along the minor groove of the (A ⁄ T) domain The G–C base pair is characterized by a large dipole and both inversions change the local electric charge repartition at )5 and )4 positions of the CArG box, and as a consequence the interactions with water molecules of the (A ⁄ T) domain [55,56] Premelting transitions are ascribed to the disruption of water molecules specifically bound to DNA [31,45,57] The presence of a ‘low-temperature form’, referred to as B¢-type DNA, is correlated with tight binding between water molecules and bases, especially in the narrow minor groove of the (A ⁄ T) domains [53,54,58] At low temperatures, between and 10 °C, free SREfos appears more bent using Raman spectroscopy than was found using electrophoresis [23,59] Relevant to the vibrational timescale (10)14 s), Raman spectroscopy allows the signals of the bent and linear conformers to be differentiated whatever their conversion time, whereas electrophoretic techniques average the signals of both conformers [59] Thus, it is more a transient bent population than a stable one that is observed in solution From one oligonucleotide to the other two, the temperature difference Raman spectra (Figs 4–6), like the difference spectra at 10 and 25 °C (Figs 7,8), exhibit a high degree of spectral pattern conservation with uniform low-intensity variations The SREfos and its two mutants oscillate between a bent and a linear form keeping the same average conformations Thus, an increase in temperature displaces the equilibrium, increasing the amplitude of motion around the regular states within the frame of the same average geometries These results suggest that the conversion process arises from global thermal fluctuations of the oligonucleotides and the mutations mainly influence the probability of their occurrence [60] Bending magnitude of SREfos In order to evaluate the bend angle induced by the decrease in temperature from 25 to 10 °C, the Raman spectral changes for SREfos were compared with those resulting from the formation of a sharp bend in a DNA octamer duplex (HMG box) upon interaction with the SRY(HMG) protein [47] The CArG and HMG boxes have very similar proportions of A:T vs G:C base pairs (6:4 in our case and 5:3 in HMG box), and approximately the same size region is expected to 2342 be subject to a sharp bend Moreover, the SREs used in this study (20-mers) contain 2.5 times more nucleotides than the HMG box (octamer used for comparison) The spectral changes occurring in SREfos between 10 and 25 °C correspond to approximately half of that caused by the SRY–HMG protein in the HMG box Otherwise, the temperature-induced structural changes in the Raman spectra during premelting are mainly characterized, in SVD analysis, by variation in the V2 contribution of the spectral component S2 Actually, the temperature profile of the V2 contribution is in accordance with the reduction in the bent population in the oligonucleotide Thereby, we can deduce that 10 °C corresponds closely to the temperature transition between the bent form and the linear form, since their populations are roughly equivalent The agreement between our results and those reported by Benevides et al [47] for the 70° sharp bend induced in the HMG box seems very interesting Indeed, the bend determined by Raman for the free SREfos in solution is roughly similar to that formed in SREfos in the crystal of its complex with the core-SRF [24] This study does not provide information on the local repartition of the angles involved in the SREfos helix bending Relative effect of bending strain There are several indications of a redistribution of the strains exerted on the oligonucleotide by the bend: partial unstacking of some adenine, thymine and guanine bases and a more distinct presence of 2¢-endo conformations of furanose rings at 10 °C against a higher percentage of 3¢-endo ⁄ anti at 25 °C Because the bend is present at low temperature, its stabilization must be favored from the point of view of enthalpy, but unfavored from the point of view of entropy A 25 °C, the higher entropy of the linear form is likely due to its higher flexibility, the higher mobility of the hydration shell, or both In the curved conformation, the strain exerted on the secondary structure of the double helix increases its torsional stiffness [61] Dynamic effects of mutations on SRE helices G fi C base mutations at positions )4 and )5 of the CArG box induce only slight local structural differences but important interactional changes between the bases The extensive empirical study of El Hassan and Calladine [56] showed that the CA step adopts a wide continuous range of conformations However, the persistence of the backbone conformation restricts FEBS Journal 274 (2007) 2333–2348 ª 2007 The Authors Journal compilation ª 2007 FEBS J Stepanek et al the space devoted to the motions of this step Within the CArG box of SREfos and SREGfos, the C)4A)3 step retains a large part of its original flexibility Conversely, the GA step adopts a restricted range of conformational space Thus, for the CArG box of SREGGfos, the rigid G)4A)3 step inhibits (by mechanical locking) the local freedom brought about by the C)4A)3 step in SREfos We should note that the C)5C)4 step in SREfos, the G)5C)4 step in SREGfos and the G)5G)4 step in SREGGfos belong to the class of loose steps Their conformational bistable space is not limited by mechanical locking, but by strong electrostatic interactions [56] The stiffness of the whole double helix depends on each local elasticity modulus and the strains involved between neighboring bases appear to be the main factor [38,62] Other local strains, such as hydration level and oligonucleotide bending, also act on the elastic properties of the entire SRE helix Premelting transition reveals that the mutations affect the hydration of the SRE minor groove Thus, this alteration contributes to the change in overall elasticity In this way, the global elastic modulus of each oligonucleotide is linked (by Hook’s law) to the elastic force on the central (A ⁄ T) domain of the CArG box and on the base sequence flanking it [63] Thus, the internal dynamics of each oligonucleotide are directly correlated with its global stiffness and the behavior of the torsional oscillations reveals these changes in elasticity Only the fluorescein reporters swept by oligonucleotide molecules with sufficient stiffness experience part of the SRE internal movements If the conversion between a bent and a linear form occurs at a shorter or longer timescale than the fluorescence excited state of fluorescein (4 ns), the frequency m ¼ ⁄ F is linked to the stiffness of the bent form and reflects the speed of helix twisting But if some of the conversion process is performed during fluorescence emission, the anisotropy decreases and the frequency m ¼ ⁄ F mirrors the helix stiffness mediated by the concomitant bent–linear conversion In that case, the weight b of the longest correlation time is related to the population of the bent form, which varies with the speed of conversion between the bent and linear forms Thus, the b population value of the double mutant SREGGfos reveals a faster conversion rate between the bent and linear form, whereas SREfos and SREGfos probably have their conversion rates close to each other Indeed, thermal fluctuations in SREfos and SREGfos make the most of the large conformational range allowed by the C)4A)3 step, whereas, owing to the restricted conformational space allowed by the G)4A)3 step, SREGGfos can only take the effects of the thermal fluctuations by a higher frequency of helix Implication of CArG sequence in SRE flexibility twisting [56] The double mutation brings a higher rigidity and instability of the bent form, correlated to the higher aptitude of the G)4A)3 step (SREGGfos) compared with the C)4A)3 step (SREfos, SREGfos) to fluctuate between the BI and the BII phosphodiester states [64–66] Indeed, a dynamic bend resulting from a ‘BI ⁄ BII’ equilibrium has been observed previously [67] The fluctuations between these phosphodiester states probably interfere with the conversion between linear and bent conformers on a nanosecond timescale [68] BI conformers have a straight helix axis, whereas BII conformers display a global dynamic curvature [67] Upon an increase in temperature the ‘BI ⁄ BII’ equilibrium is displaced from the BII to the BI conformers [69] Basis of SREfos recognition A previous working model of the core-SRF ⁄ DNA interaction suggested that core-SRF forms a stable dimer in solution under physiological conditions [70] But these studies were carried out in the presence of DNA, which induces a conformational change in coreSRF and leads to a particular monomer structure [7,23] The dimerization constant of core-SRF, alone in solution, remains unknown In vivo, binding of coreSRF to DNA is never in equilibrium but rather is a kinetic process Yet, nothing could exclude that two independent monomers bind nonspecifically to DNA and move randomly to associate on the CArG box target in a specific dimer Some features pre-exist in the free DNA and are required for preferential interactions with proteins Analysis of the crystal structure of specific complexes reveals that core-SRF modulates the inducible conformational properties of SRE [23,24] The origin of the adequate conformational deformation of the CArG boxes lies in the polymorphism of the (A ⁄ T) domain Yet within the known CArG box–core-SRF complexes, the CArG box remains bent whatever the mutations performed in the (A ⁄ T) sequence [23,24,26] There is a direct correlation between the degree of DNA bending and the ability of core-SRF to recognize a CArG consensus [25,26] Crystallographic data show that core-SRF residues K154 and K165 of the a1 helix and residues T191 and H193 of the b loop stabilize the DNA into a bent conformation Residue K154 plays a major role in specificity determination [23,25] Thus, DNA bending appears to be a major determinant of SRE–core-SRF binding specificity SREGfos displays stabilization of the stoichiometry brought about by binding of a monomer, whereas SREGGfos recruits an average of four monomers, which FEBS Journal 274 (2007) 2333–2348 ª 2007 The Authors Journal compilation ª 2007 FEBS 2343 Implication of CArG sequence in SRE flexibility J Stepanek et al pile up on this oligonucleotide The bases at positions )4 ⁄ )5 and +4 ⁄ +5 of the CArG box are highly conserved [19] In the case of SREGfos, the G–C base pair inversion at )5 position precludes formation of the hydrogen bond between the K163 residue of the core-SRF proximal subunit and the G)5 base For SREGGfos, in the same manner, inversion of the G–C base pairs at positions )5 and )4, prevents formation of the specific double-hydrogen bond between residue K163 and both G)5 and G)4 bases, and moreover, between residue T140 and C)4 base [23,24] Actually, mutations of this highly conserved K163 residue prevent complex formation [6] This specific double hydrogen bond is essential for stabilization of the a1 helix of core-SRF Thus, both single and double mutations affect binding of the proximal core-SRF subunit to the first half-site of the CArG box As a matter of fact, the structure of the (A ⁄ T) domain of the three oligonucleotides is similar It is likely that both mutations not induce any structural perturbation which could prevent, or favor, fixation of a coreSRF monomer to the second half-site Thus, structural constraints alone cannot explain why a core-SRF monomer could establish or not, some links with the second half-site The main difference between SREfos and SREGfos, on the one hand, and SREGGfos, on the other hand, consists in the distinct strains acting on the helix ply of these oligonucleotides Conversion between the bent and linear forms depends on local SRE constraints A decrease in the strains should freeze some conversion and consequently alter the topology and the population of the bent form A greater strain in the SRE-ply enhances its rigidity, which in turn reduces the amplitude and increases the frequency of the twist If the frequency of helix twisting is too high many core-SRF gather without specific association to DNA (i.e SREGGfos) [7] We should note that SREGGfos possesses a hydration state different from the other two oligonucleotides Thus, more than structural constraints, dynamics and hydration play a key role in the failing of a core-SRF monomer to establish specific links with the second half-site of SREGGfos The changes brought about by C fi G base mutations are a consequence of a complex balance between structural and dynamical effects Previous results indicated that the dynamics of core-SRF and SREfos are crucial during complex assembly [7] The specific recognition is in need of a particular DNA dynamic status of the bent form allowing the core-SRF to lock this transient conformer into a specific stable bent form [23,24] Thus, the dynamics of the bent form determine when the core-SRF switch from nonspecific to 2344 specific interactions with SREfos, even if this complex results from the interplay of interactions between both partners Conclusion CArG box sequences play a key role in hydration and dynamics within SRE double helices A premelting transition of SREs reveals a dynamic equilibrium between a bent and a linear form involving polymorphism of the (A ⁄ T) domain of the CArG box These pre-existing features of free SREfos contribute to the specific recognition with core-SRF The polymorphism of the (A ⁄ T) domain and the dynamics of the bent form one determinant for specific complex formation They appear to be as important as the conservation of the DNA base sequence Therefore, the basic question is no longer what prerequisite site on the DNA determines the specific complex formation, but rather what dynamical scenario leads to stabilization of the core-SRF on the consensus CArG sequence From this point of view, the various interactions connecting the bases of the CArG box play the key role in the physiological activity of DNA Experimental procedures Oligonucleotides The oligonucleotide SREfos 5¢-d(GGATGTCCATATTA GGACAT)-3¢ reproduces the sequence of the SRF recognition element of the c-fos enhancer [2,27] The mutants SREGfos 5¢-d(GGATGTgCATATTAGGACAT)-3¢ and SREGGfos 5¢-d(GGATGTggATATTAGGACAT)-3¢ have one (C fi G) and two (CC fi GG) mutations, respectively, at the end of the CArG consensus sequence underlined above (mutations are indicated by lower case letters) HPLC-purified single strands were purchased from Invitrogen (Cergy Pontoise, France) These sequences and their complementary strands were annealed by two heating cycles followed by slow cooling to room temperature Concentrations of double-stranded DNA were determined from single-strand DNA concentrations, estimated by absorbance measurements at 260 nm and using extinction coefficients (in mm)1 cm)1) of 14.7 (dA), (dC), 11.8 (dG), and 8.7 (dT) For fluorescence experiments, the single-stranded oligonucleotide sequences presented above were labeled with fluorescein (Invitrogen) at their 5¢-ends After association of the complementary strand, the remaining single strands and excess free fluorescein were removed by column chromatography on Sephadex G25 (Pharmacia, Saclay, France) Absorption spectra were recorded on a Varian Cary3E spectrophotometer equipped with a thermostatically FEBS Journal 274 (2007) 2333–2348 ª 2007 The Authors Journal compilation ª 2007 FEBS J Stepanek et al controlled sample holder The cell path length was cm All experiments were done in a mm Tris buffer, pH 8.5, 0.1 m NaCl, mm EDTA and mm dithiothreitol Raman spectra The 488.0 nm line of an Ar+ laser ($ 200 mW at the sample) was used for excitation Samples were placed in a temperature-stabilized microcell of 12 lL volume Raman scattered light was collected in a standard 90° geometry and recorded in a Jobin–Yvon T64000 CCD Raman spectrometer The effective spectral resolution was $ cm)1 Raman spectra of 0.4 mm SREfos, SREGfos and SREGGfos duplex solutions were measured in the 600–1800 cm)1 spectral region between 10 and 65 °C Total exposure time per spectrum ranged from 4000 to 6000 s, depending on the residual fluorescence level Before measurement of each spectrum, the sample temperature was kept constant for 10 To correct for possible drifts in the wavenumber scale, a neon glow-lamp spectrum was recorded after every analyzed sample and the Raman shift values were corrected by using an automatic recalibration procedure Subtle changes in Raman spectra were visualized by calculating difference spectra When spectra of particular duplexes were subtracted, the band of PO2– symmetric stretching vibration commonly used as intensity standard in Raman spectroscopy of DNA [30] was used to determine the right scaling factor No scaling factor was used for subtraction of Raman spectra obtained for the same sample at various temperatures; the correctness of difference spectra was nevertheless evidenced by the zero integral intensity in the region around 1092 cm)1 Raman spectra were corrected by using a unified semiautomatic procedure (subtraction of solvent spectrum, scattered light from the microcell glass walls and background represented by a sixth degree polynomial) Sets of temperature-dependent Raman spectra of each DNA duplex were treated by factor analysis –SVD algorithm [40] Fluorescence measurements To minimize possible inner filter effect, all fluorescence measurements were carried out in fluorescein-labeled oligonucleotide solutions having at the excitation wavelength an optical density < 0.05 on a cm path length Steady-state fluorescence emission spectra were recorded on a SLM Aminco-Bowman series spectrofluorometer between 500 and 600 nm with an excitation at 490 nm Both excitation and emission bandwidths were nm Fluorescence spectra were corrected for the buffer background Time-resolved fluorescence anisotropy decays were obtained by the time correlated single-photon counting method in using synchrotron radiation (superACO, LURE) as a source of exciting pulsed polarized light, with a repeat frequency of 8.33 MHz Excitation (with a nm bandwidth Implication of CArG sequence in SRE flexibility through a Jobin-Yvon double monochromator) and emission (10 nm bandwidth) were set at 490 and 520 nm, respectively Vertical and horizontal polarized light emissions were collected alternately on the experimental set-up (installed on the SB1 window) Automatic sampling cycles included a 30 s accumulation time for the instrument response function (measured with a glycogen scattering solution) and a 90 s acquisition time for each polarized fluorescent component This was repeated until a total of (2–4) · 106 counts was obtained on each intensity decay component Fluorescence anisotropy decay curves A(t) were analyzed by the maximum entropy method  [71], in using the distribuP tion function Atị ẳ i bi exp Àt ; the bi parameters being Ui the contribution weights related to the corresponding correlation times Fi CD measurements UV CD spectra were recorded on a Jasco model J-810 spectropolarimeter equipped with a thermoelectrically controlled cell holder A 0.1 cm quartz cell was used Each spectrum, monitored between 200 and 320 nm by steps of nm, represents an average of · scans Results are presented in ellipticity (millidegrees) as a function of wavelength (nm) Acknowledgements This investigation was supported by the Association Francaise contre les Myopathies (grant N°9693) and ¸ ´ the Fondation pour la Recherche Medicale (ACE 20030 92 61 87) J Stepanek thanks the University Denis Diderot for their invitation We are greatly indebted to A Huet for his help in the samples preparation and to the members of the synchrotron at LURE and the linear Accelerator Laboratory (Orsay, France) for running Super ACO References Minty A & Kedes L (1986) Upstream regions of the human cardiac actin gene that modulate its transcription in muscle cells: presence of an evolutionary conserved repeated motif Mol Cell Biol 6, 2125–2136 Treisman R (1986) Identification of a protein-binding site that mediates the transcriptional response of the c-fos gene to serum factors Cell 46, 567–574 Prywes R & Roeder RG (1986) Inducible binding of a factor to the c-fos enhancer Cell 47, 777–784 Gilman MZ, Wilson RN & Weinberg RA (1986) Multiple protein-binding sites in the 5¢-flanking region regulate c-fos expression Mol Cell Biol 6, 4305–4316 FEBS Journal 274 (2007) 2333–2348 ª 2007 The Authors Journal compilation ª 2007 FEBS 2345 Implication of CArG sequence in SRE flexibility J Stepanek et al Nurrish SJ & Treisman R (1995) DNA binding specificity determinants in MADS-box transcription factors Mol Cell Biol 15, 4076–4085 Shore P & Sharrocks AD (1995) The MADS-box family of transcription factors Eur J Biochem 229, 1–13 ` Huet A, Parlakian A, Arnaud M-C, Glandieres J-M, Valat P, Fermandjian S, Paulin D, Alpert B & Zentz C (2005) Mechanism of binding of serum response factor to serum response element FEBS J 272, 3105– 3119 Lee T, Shi Y & Schwartz RJ (1992) Displacement of BrdUrd-induced YY1 by serum response factor activates skeletal a-actin transcription in embryonic myoblasts Proc Natl Acad Sci USA 89, 9814–9818 Croissant JD, Kim J-H, Eichele G, Goering L, Lough J, Prywes R & Schwartz RJ (1996) Avian serum response factor expression restricted primarily to muscle cell lineages is required for a-actin gene transcription Dev Biol 177, 250–264 10 Kim JH, Johansen F-E, Robertson N, Catino JJ, Prywes R & Kumar CC (1994) Suppression of ras transformation by serum response factor J Biol Chem 269, 13740–13743 11 Hill CS, Wynne J & Treisman R (1994) Serum-regulated transcription by serum response factor (SRF): a novel role for the DNA binding domain EMBO J 13, 5421– 5432 12 Parlakian A, Tuil D, Hamard G, Tavernier G, Hentzen D, Concordet J-P, Paulin D, Li Z & Daegelen D (2004) Targeted inactivation of serum response factor in the developing heart results in myocardial defects and embryonic lethality Mol Cell Biol 24, 5281–5289 13 Xing Zhang S, Garcia-Gras E, Wycuff DR, Marriot SJ, Kadeer N, Yu W, Olson EN, Garry DJ, Parmacek MS & Schwartz RJ (2005) Identification of direct serumresponse factor gene target during Me2SO-induced P19 J Biol Chem 280, 19115–19126 14 Mericskay M, Parlakian A, Porteu A, Dandre F, Bonnet J, Paulin D & Li Z (2000) An overlapping CArG ⁄ octamer element is required for regulation of desmin gene transcription in arterial smooth muscle cells Dev Biol 226, 192–208 15 Leung S & Miyamoto NG (1989) Point mutational analysis of the human c-fos serum response factor binding site Nucleic Acids Res 17, 1177–1195 16 Gustafson TA & Kedes L (1989) Identification of multiple proteins that interact with functional regions of the human cardiac alpha-actin promoter Mol Cell Biol 9, 3269–3283 17 Chen CY, Croissant J, Majesky M, Topouzis S, McQuinn T, Frankovsky MJ & Schwartz RJ (1996) Activation of the cardiac alpha-actin promoter depends upon serum response factor; Tinman homologue, Nkx2.5, and intact serum response element Dev Gen 19, 119–130 2346 18 Madsen CS, Hershey JC, Hautmann MB, White SL & Owens GK (1997) Expression of the smooth muscle myosin heavy chain gene is regulated by a negative-acting GC-rich element located between two positive-acting serum response factor-binding element J Biol Chem 272, 6332–6340 19 Miano JM (2003) Serum response factor: toggling between disparate programs of gene expression J Mol Cell Cardiol 35, 577–593 20 Treisman R (1994) Ternary complex factors: growth factor regulated transcriptional activators Curr Opin Genet Dev 4, 96–101 21 Sartorelli V, Kurabayashi M & Kedes L (1993) Musclespecific gene expression A comparison of cardiac and skeletal muscle transcription strategies Circ Res 72, 925–931 22 Hautmann MB, Madsen CS, Mack CP & Owens GK (1998) Substitution of the degenerate smooth muscle (SM) a-actin CC(A ⁄ T)6GG elements with c-fos serum response elements results in increased basal expression but relaxed SM cell specificity and reduced angiotensin II inducibility J Biol Chem 273, 8398–8406 23 Pellegrini L, Tan S & Richmond TJ (1995) Structure of serum response factor core bound to DNA Nature 376, 490–498 24 Mo Y, Ho W, Johnston K & Marmorstein R (2001) Crystal structure of a ternary SAP-1 ⁄ SRF ⁄ c-fos SRE DNA complex J Mol Biol 314, 495–506 25 West AG, Shore P & Sharrocks AD (1997) DNA binding by MADS-box transcription factors: a molecular mechanism for differential DNA bending Mol Cell Biol 17, 2876–2887 26 West AG & Sharrocks AD (1999) MADS-box transcription factors adopt alternative mechanisms for bending DNA J Mol Biol 286, 1311–1323 27 Deschamps J, Meijlink F & Verma IM (1985) Identification of a transcriptional enhancer element upstream from the proto-oncogene fos Science 230, 1174–1177 28 Rodger A & Norden B (1997) Circular dichroism of biomolecules In Circular Dichroism and Linear Dichroism (Compton RG, Davies SG & Evans J, eds), pp 15–32 Oxford University Press, Oxford 29 Deng H, Bloomfield VA, Benevides JM & Thomas GJ Jr (1999) Dependence of the Raman signature of genomic B-DNA on nucleotide base sequence Biopolymers 50, 656–666 30 Movileanu L, Benevides JM & Thomas GJ Jr (1999) Temperature dependence of the Raman spectrum of DNA Part I – Raman signature of premelting and melting transitions of poly(dA–dT) poly(dA–dT) J Raman Spectrosc 30, 637–649 31 Movileanu L, Benevides JM & Thomas GJ Jr (2002) Determination of base and backbone contributions to the thermodynamics of premelting and melting transitions in DNA Nucleic Acids Res 30, 3767–3777 FEBS Journal 274 (2007) 2333–2348 ª 2007 The Authors Journal compilation ª 2007 FEBS J Stepanek et al 32 Sjoback R, Nygren J & Kubista M (1995) Absorption and fluorescence properties of fluorescein Spectrochim Acta A 51, L7–L21 33 Ashikawa I, Kinosita K Jr, Ikegami A, Nishimura Y, Tsuboi M, Watanabe K, Iso K & Nakano T (1983) Internal motions of deoxyribonucleic acid in chromatin Nanosecond fluorescence studies of intercalated ethidium Biochemistry 22, 6018–6026 34 Nordlund TM, Andersson S, Nilsson L, Rigler R, Graslund A & McLaughlin LW (1989) Structure and dynamics of a fluorescent DNA oligomer containing the EcoRI recognition sequence: fluorescence, molecular dynamics, and NMR studies Biochemistry 28, 9095–9103 35 Strobel OK, Keyes RS & Bobst AM (1990) Base dynamics of local Z-DNA conformations as detected by electron paramagnetic resonance with spin-labeled deoxcycytidine analogues Biochemistry 29, 8522–8528 36 Hustedt EJ, Spaltenstein A, Kirchner JJ, Hopkins PB & Robinson BH (1993) Motions of short DNA duplexes: an analysis of DNA dynamics using an EPR-active probe Biochemistry 32, 1774–1787 37 Ramreddy T, Sen S, Rao BJ & Krishnamoorthy G (2003) DNA dynamics in rec-a filaments: ATP hydrolysis-related flexibility in DNA Biochemistry 42, 12085–12094 38 Hogan ME & Austin RH (1987) Importance of DNA stiffness in protein–DNA binding specificity Nature 329, 263–266 39 Von Hippel PH & Berg OG (1989) Facilitated target location in biological systems J Biol Chem 264, 675– 678 40 Malinowski ER (2002) Factor Analysis in Chemistry, 3rd edn Wiley-Interscience, New York 41 Benevides JM, Overman SA & Thomas GJ Jr (2005) Raman, polarized Raman and ultraviolet resonance Raman spectroscopy of nucleic acids and their complexes J Raman Spectrosc 36, 279–299 42 Benevides JM & Thomas GJ Jr (2005) Local conformational changes induced in B-DNA by ethidium intercalation Biochemistry 44, 2993–2999 43 Hernandez B, Baumruck V, Gouyette C & Ghomi M (2005) Thermal stability, structural features, and B-to-Z transition in DNA tetraloop hair pins as determined by optical spectroscopy in d(CG)3T4(CG)3 and d(CG)3A4(CG)3 oligonucleotides Biopolymers 78, 21–34 44 Krafft C, Benevides JM & Thomas GJ Jr (2002) Secondary structure polymorphism in oxytricha nova telomeric DNA Nucleic Acids Res 30, 3981–3991 45 Movileanu L, Benevides JM & Thomas GJ Jr (2002) Temperature dependence of the Raman spectrum of DNA II Raman signature of premelting and melting transitions of poly(dA) poly(dT) and comparison with poly(dA–dT) poly(dA–dT) Biopolymers 63, 181–194 Implication of CArG sequence in SRE flexibility 46 Deng H, Bloomfield VA, Benevides JM & Thomas GJ Jr (2000) Structural basis of polyamine-DNA recognition: spermidine and spermine interactions with genomic B-DNAs of different GC content probed by Raman spectroscopy Nucleic Acids Res 28, 3379–3385 47 Benevides JM, Chan G, Lu X-J, Olson WK, Weiss MA & Thomas GJ Jr (2000) Protein-directed DNA structure I Raman spectroscopy of a high-mobility-group box with application to human sex reversal Biochemistry 39, 537–547 48 Benevides JM, Wang AH-J, Van der Marel GA, Van Boom JH & Thomas GJ Jr (1989) Effect of the G–T mismatch on backbone and sugar conformation of Z-DNA and B-DNA: analysis by Raman spectroscopy of crystal and solutions structures of d(CGCGTG) and d(CGCGCG) Biochemistry 28, 304–310 49 Puppels GJ, Otto C & Greve J (1994) Raman microscopic study of low-pH induced changes in DNA structure of polytene chromosomes Biochemistry 33, 3386–3395 50 Serban D, Benevides JM & Thomas GJ Jr (2003) HU protein employs similar mechanisms of minor-groove recognition in binding to different B-DNA sites: demonstration by Raman spectroscopy Biochemistry 42, 7390–7399 51 Ruiz-Chica AJ, Medina MA, Sanchez-Jiminez F & Ramirez FJ (2004) Characterization by Raman spectroscopy of conformational changes on guanine–cytosine and adenine–thymine oligonucleotides induced by aminooxy analogues of spermidine J Raman Spectrosc 35, 93–100 52 Duhamel J, Kanyo J, Dinter-Gottlieb G & Lu P (1996) Fluorescence emission of ethidium bromide intercalated in defined DNA duplexes: evaluation of hydrodynamics components Biochemistry 35, 16687–16697 53 Shatzky-Schwartz M, Arbuckle ND, Eisenstein M, Rabinovich D, Bareket-Samish A, Haran TE, Luisi BF & Shakked Z (1997) X-Ray and solution studies of DNA oligomers and implications for the structural basis of A-tract-dependent curvature J Mol Biol 267, 595–623 54 Yuan H, Quintana J & Dickerson RE (1992) Alternative structures for alternating poly(dA–dT) tracts: the structure of the B-DNA decamer C-G-A-T-A-T-A-T-C-G Biochemistry 31, 8009–8021 55 Hunter CA (1993) Sequence-dependent DNA structure The role of base stacking interactions J Mol Biol 230, 1025–1054 56 El Hassan ME & Calladine CR (1997) Conformational characteristics of DNA: empirical classifications and a hypothesis for the conformational behaviour of dinucleotide steps Phil Trans R Soc Lond A 355, 43–100 57 Chen YZ & Prohofsky EW (1992) The role of a minor groove spine of hydration in stabilizing poly(dA) poly(dT) against fluctuational interbase H-bond FEBS Journal 274 (2007) 2333–2348 ª 2007 The Authors Journal compilation ª 2007 FEBS 2347 Implication of CArG sequence in SRE flexibility 58 59 60 61 62 63 64 65 J Stepanek et al disruption in the premelting temperature regime Nucleic Acids Res 20, 415–419 Arai S, Chatake T, Ohhara T, Kurihara K, Tanaka I, Suzuki N, Fujimoto Z, Mizuno H & Niimura N (2005) Complicated water orientation in the minor groove of the B-DNA decamer d(CCATTAATGG)2 observed by neutron diffraction measurements Nucleic Acids Res 33, 3017–3024 Gustafson TA, Taylor A & Kedes L (1989) DNA bending is induced by a transcription factor that interacts with the human c-fos and a-actin promoters Proc Natl Acad Sci USA 86, 2162–2166 Bell AF, Hecht L & Barron LD (1998) Evidence for global mobility in the premelting of a polynucleotide from temperature-dependent Raman optical activity Biospectroscopy 4, 107–111 Heath PJ, Clendenning JB, Fujimoto BS & Schurr JM (1996) Effect of bending strain on the torsion elastic constant of DNA J Mol Biol 260, 718–730 Coleman BD, Olson WK & Swigon D (2003) Theory of sequence-dependent DNA elasticity J Chem Phys 118, 7127–7140 Gavish B (1977) The role of geometry and elastic strains in dynamic states of proteins Biophys Struct Mech 4, 37–52 ´ Grzeskowiak K, Yanagi K, Prive GG & Dickerson RE (1991) The structure of B-helical C-G-A-T-C-G-A-T-C-G and comparison with C-C-A-C-G-T-T-G-G The effect of base pair reversals J Biol Chem 266, 8861–8883 Chou S-H, Cheng J-W, Fedoroff OY, Chuprina VP & Reid BR (1992) Adjacent G:A mismatch base pairs 2348 66 67 68 69 70 71 contain BII phosphodiesters in solution J Am Chem Soc 114, 3114–3115 Chou S-H, Cheng J-W & Reid BR (1992) Adjacent G:A mismatches stabilized by cross-strand base-stacking and BII phosphate groups J Mol Biol 228, 138–155 ´ Tisne C, Hantz E, Hartmann B & Delepierre M (1998) Solution structure of a non-palindromic 16 base-pair DNA related to the HIV jB site: evidence for BI–BII equilibrium inducing a global dynamic curvature of the duplex J Mol Biol 279, 127–142 Winger RH, Liedl KR, Rudisser S, Pichler A, Hallbrucker A & Mayer E (1998) B-DNA BI–BII conformer substate dynamics is coupled with water migration J Phys Chem 102, 8934–8940 Gorenstein DG (1992) 31P NMR of DNA In Methods in Enzymology: DNA Structures, Part A, Synthesis and Physical Analysis of DNA (Lilley DMJ & Dahlberg JE, eds), Vol 211, pp 254–286 Academic Press, London Norman C, Runswick M, Pollock R & Treisman R (1988) Isolation and properties of cDNA clones encoding SRF, a transcription factor that binds to the c-fos serum response element Cell 55, 989–1003 Gentin M, Vincent M, Brochon JC, Livesey AK, Cittanova N & Gallay J (1990) Time-resolved fluorescence of the single tryptophan residue in rat alpha-fetoprotein and rat serum albumin: analysis by the maximum entropy method Biochemistry 29, 10405–10412 FEBS Journal 274 (2007) 2333–2348 ª 2007 The Authors Journal compilation ª 2007 FEBS ... 5¢-d(GGATGTCCATATTA GGACAT)-3¢ reproduces the sequence of the SRF recognition element of the c- fos enhancer [2,27] The mutants SREGfos 5¢-d(GGATGTgCATATTAGGACAT)-3¢ and SREGGfos 5¢-d(GGATGTggATATTAGGACAT)-3¢... a restricted range of conformational space Thus, for the CArG box of SREGGfos, the rigid G) 4A)3 step inhibits (by mechanical locking) the local freedom brought about by the C) 4A)3 step in SREfos... interactions of CArG boxes and accessory proteins [21,22] The core domain of the SRF (core-SRF) binds to the CArG box as a homodimer [7,23,24] The speci? ?c core-SRF? ??SREfos complex is characterized