Indirect readout of DNA sequence at the primary-kink site in the CAP-DNA complex: DNA binding specificity based on energetics of DNA kinking.

The catabolite activator protein (CAP) makes no direct contact with the consensus base-pair T:A at position 6 of the DNA half-site 5'-A(1)A(2)A(3)T(4)G(5)T(6)G(7)A(8)T(9)C(10)T(11)-3' but, nevertheless, exhibits strong specificity for T:A at position 6. Binding of CAP results in formation of a sharp DNA kink, with a roll angle of approximately 40 degrees and a twist angle of approximately 20 degrees, between positions 6 and 7 of the DNA half-site. The consensus base-pair T:A at position 6 and the consensus base-pair G:C at position 7 form a T:A/G:C step, which is known to be associated with DNA flexibility. It has been proposed that specificity for T:A at position 6 is a consequence of formation of the DNA kink between positions 6 and 7, and of effects of the T:A(6)/G:C(7) step on the geometry of DNA kinking, or the energetics of DNA kinking. In this work, we determine crystallographic structures of CAP-DNA complexes having the consensus base-pair T:A at position 6 or the non-consensus base-pair C:G at position 6. We show that complexes containing T:A or C:G at position 6 exhibit similar overall DNA bend angles and local geometries of DNA kinking. We infer that indirect readout in this system does not involve differences in the geometry of DNA kinking but, rather, solely differences in the energetics of DNA kinking. We further infer that the main determinant of DNA conformation in this system is protein-DNA interaction, and not DNA sequence.     

Indirect readout of DNA sequence at the primary-kink site in the CAP-DNA complex: alteration of DNA binding specificity through alteration of DNA kinking.

The catabolite activator protein (CAP) sharply bends DNA in the CAP-DNA complex, introducing a DNA kink, with a roll angle of approximately 40 degrees and a twist angle of approximately 20 degrees, between positions 6 and 7 of the DNA half-site, 5'-A(1)A(2)A(3)T(4)G(5)T(6)G(7)A(8)T(9)C(10)T(11)-3' ("primary kink"). CAP recognizes the base-pair immediately 5' to the primary-kink site, T:A(6), through an "indirect-readout" mechanism involving sequence effects on the energetics of primary-kink formation. CAP recognizes the base-pair immediately 3' to the primary-kink site, G:C(7), through a "direct-readout" mechanism involving formation of a hydrogen bond between Glu181 of CAP and G:C(7). Here, we report that substitution of the carboxylate side-chain of Glu181 of CAP by the one-methylene-group-shorter carboxylate side-chain of Asp changes DNA binding specificity at position 6 of the DNA half site, changing specificity for T:A(6) to specificity for C:G(6), and we report a crystallographic analysis defining the structural basis of the change in specificity. The Glu181-->Asp substitution eliminates the primary kink and thus eliminates indirect-readout-based specificity for T:A(6). The Glu181-->Asp substitution does not eliminate hydrogen-bond formation with G:C(7), and thus does not eliminate direct-readout-based specificity for G:C(7).     

DNA bending: the prevalence of kinkiness and the virtues of normality.

DNA bending in 86 complexes with sequence-specific proteins has been examined using normal vector plots, matrices of normal vector angles between all base pairs in the helix, and one-digit roll/slide/twist tables. FREEHELIX, a new program especially designed to analyze severely bent and kinked duplexes, generates the foregoing quantities plus local roll, tilt, twist, slide, shift and rise parameters that are completely free of any assumptions about an overall helix axis. In nearly every case, bending results from positive roll at pyrimidine-purine base pair steps: C-A (= T-G), T-A, or less frequently C-G, in a direction that compresses the major groove. Normal vector plots reveal three well-defined types of bending among the 86 examples: (i) localized kinks produced by positive roll at one or two discrete base pairs steps, (ii) three-dimensional writhe resulting from positive roll at a series of adjacent base pairs steps, or (iii) continuous curvature produced by alternations of positive and negative roll every 5 bp, with side-to-side zig-zag roll at intermediate position. In no case is tilt a significant component of the bending process. In sequences with two localized kinks, such as CAP and IHF, the dihedral angle formed by the three helix segments is a linear function of the number of base pair steps between kinks: dihedral angle = 36 degrees x kink separation. Twenty-eight of the 86 examples can be described as major bends, and significant elements in the recognition of a given base sequence by protein. But even the minor bends play a role in fine-tuning protein/DNA interactions. Sequence-dependent helix deformability is an important component of protein/DNA recognition, alongside the more generally recognized patterns of hydrogen bonding. The combination of FREEHELIX, normal vector plots, full vector angle matrices, and one-digit roll/slide/twist tables affords a rapid and convenient method for assessing bending in DNA.     

DNA sequence determinants for binding of the Escherichia coli catabolite gene activator protein.

The consensus DNA site for binding of the Escherichia coli catabolite gene activator protein (CAP) is 22 base pairs in length and is 2-fold symmetric: 5'-AAATGTGATCTAGATCACATTT-3'. Positions 4 to 8 of each half of the consensus DNA half-site are the most strongly conserved. In this report, we analyze the effects of substitution of DNA base pairs at positions 4 to 8, the effects of substitution of thymine by uracil and by 5-methylcytosine at positions 4, 6, and 8, and the effect of dam methylation of the 5'-GATC-3' sequence at positions 7 to 10. All DNA sites having substitutions of DNA base pairs at positions 4 to 8 exhibit lower affinities for CAP than does the consensus DNA site, consistent with the proposal that the consensus DNA site is the ideal DNA site for CAP. Specificity for T:A at position 4 appears to be determined solely by the thymine 5-methyl group. Specificity for T:A at position 6 and specificity for A:T at position 8 appear to be determined in part, but not solely, by the thymine 5-methyl group. dam methylation has little effect on CAP.DNA complex formation. The thermodynamically defined consensus DNA site spans 28 base pairs. All, or nearly all, DNA determinants required for maximal affinity for CAP and for maximal thermodynamically defined CAP.DNA ion pair formation are contained within a 28-base pair DNA fragment that has the 22-base pair consensus DNA site at its center. The quantitative data in this report provide base-line thermodynamic data required for detailed investigations of amino acid-base pair and amino acid-phosphate contacts in this protein-DNA complex.     

Intrinsic bending and deformability at the T-A step of CCTTTAAAGG: a comparative analysis of T-A and A-T steps within A-tracts

Introduction of a T-A or pyrimidine-purine step into a straight and rigid A-tract can cause a positive roll deformation that kinks the DNA helix at that step. In CCTTTAAAGG, the central T-A step has an 8.6 degrees bend toward the major groove. We report the structural analysis of CCTTTAAAGG and a comparison with 25 other representative crystal structures from the NDB containing at least four consecutive A or T bases. On average, more local bending occurs at the disruptive T-A step (8.21 degrees ) than at an A-T step (5.71 degrees ). In addition, A-tracts containing an A-T step are more bent than are pure A-tracts, and hence A-A and A-T steps are not equivalent. All T-A steps examined exhibit positive roll, bending towards the major groove, while A-T steps display negative roll and bend slightly towards the minor groove. This illustrates how inherent negative and positive roll are, respectively, at A-T and T-A steps within A-tracts. T-A steps are more deformable, showing larger and more variable deformations of minor groove width, rise, cup, twist, and buckle. Standard deviations of twist, rise, and cup for T-A steps are 6.66 degrees, 0.55 A, and 15.90 degrees, versus 2.28 degrees, 0.21 A, and 2.99 degrees for A-T steps. Packing constraints determine which local values of these helical parameters an individual T-A step will adopt. For instance, with CCTTTAAAGG and three isomorphous structures, CGATTAATCG, CGATATATCG, and CGATCGATCG, crystal packing forces lead to a series of correlated changes: widened minor groove, large slide, low twist, and large rise. The difference in helical parameters between A-T steps lying within A-tracts, versus A-T steps within alternating AT sequences, demonstrates the importance of neighboring steps on the conformation of a given dinucleotide step.     

DNA-sequence recognition by CAP: role of the adenine N6 atom of base pair 6 of the DNA site.

Two similar, but not identical, models have been proposed for the amino acid-base pair contacts in the CAP-DNA complex ('Model I,' Weber, I. and Steitz, T., Proc. Natl. Acad. Sci. USA, 81, 3973-3977, 1984; 'Model II,' Ebright, et al., Proc. Natl. Acad. Sci. USA, 81, 7274-7278, 1984). One difference between the two models involves Glu181 of CAP. Model I predicts that Glu181 of CAP makes two specificity determining contacts: one H-bond with the cytosine N4 atom of G:C at base pair 7 of the DNA half site, and one H-bond with the adenine N6 atom of T:A at base pair 6 of the DNA half site. In contrast, Model II predicts that Glu181 makes only one specificity determining contact: one H-bond with the cytosine N4 atom of G:C at base pair 7 of the DNA half site. In the present work, we show that replacement of T:A at base pair 6 of the DNA half site by T:N6-methyl-adenine has no, or almost no, effect on the binding of CAP. We conclude, contrary to Model I, that Glu181 of CAP makes no contact with the adenine N6 atom of base pair 6 of the DNA half site.     

The crystal structure of d(GTACGTAC) at 2.25 A resolution: are the A-DNA's always unwound approximately 10 degrees at the C-G steps?

The structure of the self-complementary octamer d(GTACGTAC) has been analyzed by a single crystal X-ray diffraction method at 2.25 A resolution. The crystallographic R factor was 0.184 for all 1233 reflections at this resolution. In spite of the alternating purine-pyrimidine sequence, d(GTACGTAC) adopts the A-form conformation rather than the left-handed Z-form. The average helix twist and the mean rise per base pair are 32.1 degrees and 3.18 A, respectively. The d(GTACGTAC) helix is characterized by a wide open major groove and small base-pair tilt (9.7 degrees). The partial unwinding of the helix is observed only at the central pyrimidine-purine C-G step, but not at the other pyrimidine-purine T-A steps. Based on this study and six other X-ray studies, we propose a hypothesis that the A-DNA's are always unwound approximately 10 degrees at the C-G steps. Significant differences in base-pair stacking modes are seen between the purine-pyrimidine step and the pyrimidine-purine step. All deoxyribose rings adopt the C3'-endo conformation. All backbone torsion angles fall into the range expected for the A-DNA form, except for the nucleotide G5, whose alpha and gamma torsion angles adopt the trans, trans conformation instead of the common gauche-, gauche+ conformation.     

Solution structures of a DNA dodecamer duplex with and without a cisplatin 1,2-d(GG) intrastrand cross-link: comparison with the same DNA duplex containing an oxaliplatin 1,2-d(GG) intrastrand cross-link

Proteins that discriminate between cisplatin-DNA adducts and oxaliplatin-DNA adducts are thought to be responsible for the differences in tumor range, toxicity, and mutagenicity of these two important chemotherapeutic agents. However, the structural basis for differential protein recognition of these adducts has not been determined and could be important for the design of more effective platinum anticancer agents. We have determined high-resolution NMR structures for cisplatin-GG and undamaged DNA dodecamers in the AGGC sequence context and have compared these structures with the oxaliplatin-GG structure in the same sequence context determined previously in our laboratory. This structural study allows the first direct comparison of cisplatin-GG DNA and oxaliplatin-GG DNA solution structures referenced to undamaged DNA in the same sequence context. Non-hydrogen atom rmsds of 0.81 and 1.21 were determined for the 15 lowest-energy structures for cisplatin-GG DNA and undamaged DNA, respectively, indicating good structural convergence. The theoretical NOESY spectra obtained by back-calculation from the final average structures showed excellent agreement with the experimental data, indicating that the final structures are consistent with the NMR data. Several significant conformational differences were observed between the cisplatin-GG adduct and the oxaliplatin-GG adduct, including buckle at the 5' G6.C19 base pair, opening at the 3' G7.C18 base pair, twist at the A5G6.T20C19 base pair step, slide, twist, and roll at the G6G7.C19C18 base pair step, slide at the G7C8.C18G17 base pair step, G6G7 dihedral angle, and overall bend angle. We hypothesize that these conformational differences may be related to the ability of various DNA repair proteins, DNA binding proteins, and DNA polymerases to discriminate between cisplatin-GG and oxaliplatin-GG adducts.     

The solution conformation of a carbocyclic analog of the Dickerson-Drew dodecamer: comparison with its own X-ray structure and that of the NMR structure of the native counterpart

The NMR conformation of a carbocyclic analog of the Dickerson-Drew dodecamer [d(CGC-GAAT*T*CGCG)]2 containing 6'-alpha-Me carbocyclic thymidines (T*) has been determined and compared with that of its X-ray structure. The solution structure of the 6'-alpha-Me carbocyclic thymidine modified duplex has also been compared with the solution structure of the corresponding unmodified Dickerson-Drew duplex solved by us under the same experimental conditions. The NMR structures have been based on 24 experimental distance and torsion constraints per residue for [d(CGCGAAT*T*CGCG)]2 (1) and on 21 constraints per residue for the natural counterpart. In general, both final NMR structures are more close to the B-type DNA. The cyclopentane moieties of the carbocyclic thymidine residues adopt C1'-exo B-DNA type puckers (the phase angles P = 136-139 degrees and the puckering amplitudes psi = 36-37 degrees) that are close to their previously published crystal C1'-exo or C2'-endo puckers. The main differences between the two NMR structures are for beta(T*8) and epsilon, xi(T*7) backbone torsions (27-50 degrees ), for basepair twist for the 7-8 and 8-9 basepair steps (5-6 degrees), tilt for the 8-9 step (7 degrees), roll for the 7-8 step (7 degrees), shift for the 7-8 step (0.9A) and slide for the 9-10 step (0.6A). The relatively small deviations of helical structure parameters lead to structural isomorphism of these duplexes in aqueous solutions (atomic RMSD = 1.0A). The difference of the minor groove widths (less than 0.7A) in the core part of the modified duplex in comparison with the native one is much smaller than the difference between the X-ray structures of these duplexes. A detailed comparison of NMR and X-ray structure parameters showed significant monotonic differences (0.9-2.5A) for all basepair slides in both duplexes. Deviations between NMR and X-ray structure parameters for the modified duplex were also found for basepair tilt of the 4-5 step (13 degrees), rolls for the 8-9 and 10-11 steps (16 and 19 degrees), twist of the 3-4 step (8 degrees) and shift of the 9-10 step (0.9A).     

Crystal structure of an RNA duplex r(G GCGC CC)2 with non-adjacent G*U base pairs

The crystal structure of a self-complementary RNA duplex r(GGGCGCUCC)2with non-adjacent G*U and U*G wobble pairs separated by four Watson-Crick base pairs has been determined to 2.5 A resolution. Crystals belong to the space group R3; a = 33.09 A,alpha = 87.30 degrees with a pseudodyad related duplex in the asymmetric unit. The structure was refined to a final Rworkof 17.5% and Rfreeof 24.0%. The duplexes stack head-to-tail forming infinite columns with virtually no twist at the junction steps. The 3'-terminal cytosine nucleosides are disordered and there are no electron densities, but the 3' penultimate phosphates are observed. As expected, the wobble pairs are displaced with guanine towards the minor groove and uracil towards the major groove. The largest twist angles (37.70 and 40.57 degrees ) are at steps G1*C17/G2*U16 and U7*G11/C8*G10, while the smallest twist angles (28.24 and 27.27 degrees ) are at G2*U16/G3*C15 and C6*G12/U7*G11 and conform to the pseudo-dyad symmetry of the duplex. The molecule has two unequal kinks (17 and 11 degrees ) at the wobble sites and a third kink at the central G5 site which may be attributed to trans alpha (O5'-P), trans gamma (C4'-C5') backbone conformations. The 2'-hydroxyl groups in the minor groove form inter-column hydrogen bonding, either directly or through water molecules.     

Crystal structure of an RNA octamer duplex r(CCCIUGGG)2 incorporating tandem I.U wobbles.

The crystal structure of the RNA octamer duplex r(CCCIUGGG)2has been elucidated at 2.5 A resolution. The crystals belong to the space group P21and have unit cell constants a = 33.44 A, b = 43.41 A, c = 49.39 A and beta = 104.7 degrees with three independent duplexes (duplexes 1-3) in the asymmetric unit. The structure was solved by the molecular replacement method and refined to an Rwork/Rfree of 0.185/0.243 using 3765 reflections between 8.0 and 2.5 A. This is the first report of an RNA crystal structure incorporating I.U wobbles and three molecules in the asymmetric unit. Duplex 1 displays a kink of 24 degrees between the mismatch sites, while duplexes 2 and 3 have two kinks each of 19 degrees and 27 degrees, and 24 degrees and 29 degrees, respectively, on either side of the tandem mismatches. At the I.U/U.I mismatch steps, duplex 1 has a twist angle of 33.9 degrees, close to the average for all base pair steps, but duplexes 2 and 3 are underwound, with twist angles of 24.4 degrees and 26.5 degrees, respectively. The tandem I.U wobbles show intrastrand purine-pyrimidine stacking but exhibit interstrand purine-purine stacking with the flanking C.G pairs. The three independent duplexes are stacked non-coaxially in a head-to-tail fashion to form infinite pseudo-continuous helical columns which form intercolumn hydrogen bonding interactions through the 2'-hydroxyl groups where the minor grooves come together.     

Further characterization of intracytoplasmic A particle-associated DNA.

A DNA species with buoyant densities greater than mouse cellular DNA was found associated with intracytoplasmic A particles (CAP) isolated from mouse mammary tumor virus-infected mouse mammary tumors and mouse Leydig cell tumors which produce CAP but no complete mouse mammary tumor virus virions. This DNA species was absent in identically prepared tissue fractions from tumors which did not contain CAP. Treatment of CAP-associated DNA with pancreatic RNase A did not alter the buoyant density although a reduction in apparent molecular weight (broadening of the DNA band at equilibrium) was observed upon analytical equilibrium sedimentation in CsCl. The molecular weight of untreated CAP-associated DNA was estimated to range from 0.8 x 10(6) to 3.1 x 10(6). Base composition analysis showed CAP-DNA to possess an approximate guanine plus cytosine content of 38%. Ninety percent of CAP-associated DNA eluted as single-stranded molecules upon hydroxyapatite column chromatography, a characteristic that accounts in part for its higher buoyant density in neutral CsCl compared to native double-stranded mouse DNA. In two preparations, CAP-DNA had a sedimentation coefficient of 7 to 8S.     

Kinetics and mechanism in the reaction of gene regulatory proteins with DNA

We have measured the kinetic properties of the Escherichia coli cAMP receptor protein (CAP) and lac repressor interacting with lac promoter restriction fragments. Under our reaction conditions (10 mM-Tris X HCl (pH 8.0 at 21 degrees C), 1 mM-EDTA, 10 microM-cAMP, 50 micrograms bovine serum albumin/ml, 5% glycerol), the association of CAP is at least a two-step process, with an initial, unstable complex formed with rate constant kappa a = 5(+/- 2.5) X 10(7) M-1 s-1. Subsequent formation of a stable complex occurs with an apparent bimolecular rate constant kappa a = 6.7 X 10(6) M-1 s-1. At low total DNA concentration, the dissociation rate constant for the specific CAP-DNA complex is 1.2 X 10(-4) s-1. The ratio of formation and dissociation rate constants yields an estimate of the equilibrium constant, Keq = 5 X 10(10) M-1, in good agreement with static results. We observed that the dissociation rate constant of both CAP-DNA and repressor-DNA complexes is increased by adding non-specific "catalytic" DNA to the reaction mixture. CAP dissociation by the concentration-dependent pathway is second-order in added non-specific DNA, consistent with either the simultaneous or the sequential participation of two DNA molecules in the reaction mechanism. The results imply a role for distal DNA in assembly-disassembly of specific CAP-DNA complexes, and are consistent with a model in which the subunits in the CAP dimer separate in the assembly-disassembly process. The dissociation of lac repressor-operator complexes was found to be DNA concentration-dependent as well, although in contrast to CAP, the reaction is first-order in catalytic DNA. Added excess operator-rich DNA gave more rapid dissociation than equivalent concentrations of non-specific DNA, indicating that the sequence content of the competing DNA influences the rate of repressor dissociation. The simplest interpretation of these observations is that lac repressor can be transferred directly from one DNA molecule to another. A comparison of the translocation rates calculated for direct transfer with those predicted by the one-dimensional sliding model indicates that direct transfer may play a role in the binding site search of lac repressor.     

Interaction of unfused tricyclic aromatic cations with DNA: a new class of intercalators

Unfused tricyclic aromatic ring systems 1-6 with one or two cationic side chains have been synthesized and their interactions with DNA and synthetic polymers probed with a variety of techniques. Molecular mechanics calculations indicate that the torsional angle between ring planes in the minimum energy conformation of the tricyclic molecules can range from 0 degree to as high as 50 degrees depending on the type of rings and substituents. Viscometric titrations with linear and supercoiled DNA, linear dichroism, and NMR studies indicated that all compounds with torsional angles of approximately 20 degrees or less bind to DNA by intercalation. The more highly twisted intercalators caused significant perturbation of DNA structure. Unfused intercalators with twist angles of approximately 20 degrees have reduced binding constants, suggesting that they could not form an optimum interaction with the DNA base pairs. Unfused intercalators with twist less than 20 degrees formed strong complexes with DNA. The structures of these unfused intercalators are more analogous to typical groove-binding molecules, and an analysis of their interaction with DNA provides a better understanding of the subtle differences between intercalation and groove-binding modes for aromatic cations. The results indicate that intercalation and groove-binding modes should be viewed as two potential wells on a continuous energy surface. The results also suggest design strategies for intercalators that can optimally complement DNA base pair propeller twist or that can induce bends in DNA at the intercalation site.     

Solution structure of [d(GCGTATACGC)]2.

The solution structure of the alternating pyrimidine-purine DNA duplex [d(GCGTATACGC)]2 has been determined using two-dimensional nuclear magnetic resonance techniques and distance geometry methods. Backbone distance constraints derived from experimental nuclear Overhauser enhancement and J-coupling torsion angle constraints were required to adequately define the conformation of the inter-residue backbone linkages and to avoid underwinding of the duplex. The distance geometry structures were further refined by back-calculation of the two-dimensional nuclear Overhauser enhancement spectra to correct spin-diffusion distance errors. Fifteen final structures for [d(GCGTATACGC)]2 were generated from the refined experimental distance bounds. These structures all exhibit fully wound B-form geometry with small penalty values (< 1.5 A) against the distance bounds and small pair-wise root-mean-square deviation values (typically 0.6 A to 1.5 A). The final structures exhibit positive base-pair inclination with respect to the helix axis, a marked alternation in rise and twist, and are shorter and wider than classical fiber B-form DNA. The purines were found to adopt a sugar pucker close to the C-2'-endo conformation while pyrimidine sugars exhibited significantly lower pseudorotation phase angles in the C-1'-exo to C-2'-endo range. The minor groove cross-strand steric clashes at pyrimidine-purine steps that would exist in pure B-DNA are attenuated by an increased rise at these steps (and an increased roll angle at TpA steps). Concomitantly the backbone torsion angles of the pyrimidine moieties have larger gamma values, larger epsilon values, and smaller zeta values than the purines. The structures generated by distance geometry methods were also compared with those obtained from restrained molecular dynamics with empirical force-field potentials. The results indicate that the nuclear magnetic resonance/distance geometry approach alone is capable of elucidating most of the salient structural features of double-stranded helical nucleic acids in solution without resorting to empirical energy potentials and without using any structural assumptions from crystallographic data.     

Characterization of the base stacking interactions in DNA by means of Lennard-Jones empirical potentials

Three empirical potentials of the Lennard-Jones type taken from literature were used to calculate van der Waals contributions to the base-pair couples stacking energies in B-DNA and A-DNA type double helical conformations. The information obtained can be summarized as follows: (1) Purine-pyrimidine and purine-purine (pyrimidine-pyrimidine in the complementary strand) sequences preferred right-handed helical arrangement, whereas pyrimidine-purine sequences favoured left-handed (C-G) or unwound (T-A) stacking geometry; in the latter case this only held for B- but not A-DNA (the C-G sequence was not studied in A-DNA owing to difficulties (see below) with the G amino group in B-DNA); (2) Positive propeller twist of base-pairs was stable in both B- and A-DNA; the thymine methyl group promoted the propeller and this effect was strongest in the A-T step; (3) Tilt of base pairs occurred around zero in B-DNA and between 15-20 degrees C in A-DNA, in agreement with the experimental observations; (4) Vertical separation of base pairs was optimal within 0.33-0.34 nm for B-DNA and around 0.29 nm for A-DNA using the 9-6 potential. The 12-6 potential gave similar results with B-DNA as the 9-6 potential if, however, base pairs were separated by 0.35-0.36 nm; (5) The calculated effect of the guanine amino group was substantially stronger than expected on the basis of data derived from X-ray diffraction studies of oligonucleotide single crystals; (6) In comparison with the 9-6 potential, the 12-6 potential provided more strict energy minima. In summary, the empirical potentials reproduce, at least semiquantitatively, many but not all DNA properties; this should be taken into account whenever the potentials are used for prediction purposes.