Energetic basis for selective recognition of T·G mismatched base pairs
http://www.100md.com
《核酸研究医学期刊》
1 Department of Chemistry, Georgia State University, Atlanta, GA 30303, USA, 2 Department of Chemistry, Furman University, Greenville, SC 29613, USA and 3 Department of Oncology, Royal Free and University College Medical School, University College London, London W1W 7BS, UK
*To whom correspondence should be addressed. Tel: +1 404 651 3903; Fax: +1 404 651 1416; Email: chewdw@panther.gsu.edu
Correspondence may also be addressed to Moses Lee. Tel: +1 864 294 3368; Fax: +1 864 294 3559; Email: moses.lee@furman.edu
ABSTRACT
To complement available structure and binding results and to develop a detailed understanding of the basis for selective molecular recognition of T·G mismatches in DNA by imidazole containing polyamides, a full thermodynamic profile for formation of the T·G–polyamide complex has been determined. The amide-linked heterocycles f-ImImIm and f-PyImIm (where f is formamido group, Im is imidazole and Py is pyrrole) were studied by using biosensor-surface plasmon resonance (SPR) and isothermal titration calorimetry (ITC) with a T·G mismatch containing DNA hairpin duplex and a similar DNA with only Watson–Crick base pairs. Large negative binding enthalpies for all of the polyamide–DNA complexes indicate that the interactions are enthalpically driven. SPR results show slower complex formation and stronger binding of f-ImImIm to the T·G than to the match site. The thermodynamic analysis indicates that the enhanced binding to the T·G site is the result of better entropic contributions. Negative heat capacity changes for the complex are correlated with calculated solvent accessible surface area changes and indicate hydrophobic contributions to complex formation. DNase I footprinting analysis in a long DNA sequence provided supporting evidence that f-ImImIm binds selectively to T·G mismatch sites.
INTRODUCTION
Polyamide DNA binding agents are of interest due to their sequence-specific binding characteristics (1–3), their ability to recognize mismatched base pairs, and their potential for use in modulation of gene expression (4). Through the combination of pyrrole (Py), imidazole (Im), and other heterocycles in a side-by-side pairing, polyamides capable of specifically recognizing Watson–Crick DNA sequences have been synthesized (1–3). In addition, structural, binding affinity and kinetics studies have shown the potential use of the side-by-side pairing of Im heterocycles for recognition of DNA sequences containing non-Watson–Crick base pairs (5,6). The amide-linked triimidazole, f-ImImIm where f is formamido (Fig. 1A), can discriminate T·G from Watson–Crick C·G base pairs as well as from other mismatches (5,6).
Figure 1. (A) Chemical structures of the polyamides: distamycin A, f-ImImIm and f-PyImIm. (B) Sequences of the DNA hairpins used in this study: CCGG and A3T3 are Watson–Crick sequences while CTGGsm contains one T·G mismatch. (C) Schematic models of the complexes formed between f-ImImIm and the DNA sequences containing a double T·G mismatch (left) and CCGG (right) as inferred from NMR data.
Possible uses of the Im/Im pair include rapid detection of single nucleotide polymorphisms (5,6). Rapid and accurate detection of T·G mismatch base pairs in DNA would be of significant benefit since this mismatch is responsible for most of the common mutations leading to formation of tumors in humans. In human bladder carcinoma, for example, a G·C to A·T transition at the 3'G of the GG doublet in codon 12 of the Ha-ras and Ki-ras proto-oncogene converts them to oncogenes (7–9). A T·G mismatch is often introduced by spontaneous deamination of 5-methylcytosine and can arise from errors in replication (10). Structural analysis of DNAs containing T·G mismatches, indicates that the mismatches adopt a wobble conformation and structural perturbations are mainly in the vicinity of the mismatch (11,12). NMR studies show that f-ImImIm forms an anti-parallel dimer in the DNA minor groove with a T·G mismatch site (5,6,13,14). In the mismatch complex the guanine-NH2 group of the wobble T·G base pair forms two specific hydrogen bonds to the side-by-side Im/Im pair (5). Biosensor-surface plasmon resonance (SPR) analysis has shown that f-ImImIm is capable of strongly and specifically binding to a DNA sequence with only a single T·G mismatch (6,13). This novel T·G mismatch recognition motif has recently been corroborated by results from the Dervan laboratory, in which a hairpin polyamide containing a stacked Im/Im pair was found to bind at a T·G mismatch site (14).
There is a wealth of information on structure, binding affinity and specificity for the interactions of polyamides with DNA (1–3), however, the specific energetic factors that govern the binding events have rarely been characterized. Important calorimetric results have been reported for a very limited number of polyamides in complex with their target DNA sequences (15–17). Thermodynamic studies can provide valuable complementary information to the structural perspective and can improve our understanding of the molecular forces that govern the affinity and specificity of binding. Specifically, they can give additional insights into the energetic basis for the discrimination of the T·G mismatch from matched sites by f-ImImIm. We report here biosensor-SPR and calorimetric studies as a function of temperature as well as footprinting studies in a large sequence context for the interaction of f-ImImIm and f-PyImIm (Fig. 1A) with mismatched and matched sequence DNAs. By conducting the studies as a function of temperature, we were able to determine the heat capacity changes on binding that are critical for a complete thermodynamic understanding of biomolecular interactions (18–20).
MATERIALS AND METHODS
Reagents and biochemicals
Compounds and buffers. Distamycin A was purchased from Sigma Chemical Co. and used without further purification. N--1-methyl-4-{1-methyl-4-imidazole-2- carboxamido}imidazole-2-carboxamide (f-ImImIm) and N--1-methyl-4-{1-methyl-4-imidazole-2-carboxamido}imidazole-2-carboxamide (f-PyImIm) were prepared as described previously (6,21). Solutions of the compounds for calorimetric studies were prepared in MES20 buffer, 0.01 M MES adjusted to pH 6.25, 0.001 M EDTA and 0.2 M NaCl. Concentrations of the compounds were determined spectrophotometrically by absorbance at 304 nm for distamycin and 303 nm for f-ImImIm and f-PyImIm. The extinction coefficients used were 34 000 M–1 cm–1 for distamycin (17) and 30 700 M–1 cm–1 for f-ImImIm and f-PyImIm at 25°C.
DNA sequences. DNAs were obtained as anionic exchange, HPLC purified products (Midland Certified Reagent Co). Biosensor-SPR binding studies were conducted with 5'-biotinated DNAs, and calorimetric studies were performed with non-biotinated DNAs. Solutions for the ITC experiments were prepared in degassed MES20 buffer. To ensure that the DNA was in the hairpin form, the solutions were heated to 90°C and cooled rapidly in ice (22). The concentration of the DNA solutions was determined spectrophotometrically at 260 nm and 85°C using extinction coefficients per nucleotide of 9264, 9059 and 9400 M–1 cm–1 for CCGG, CTGGsm and A3T3, respectively. The extinction coefficients were calculated on a per strand basis by the nearest-neighbor method and divided by the number of nucleotides per strand (23–25). The extinction coefficients extrapolated to 25°C were 7824, 8107 and 7978 M–1 cm–1 for CCGG, CTGGsm and A3T3, respectively.
Methods
DNase I footprinting of TG mismatched oligonucleotides. Five micrograms of an 80 base DNA 5'-AGG TGA GCA GGT CCA TAC TGG TTT GCA CCT CGA GGT TAC CGG TAT CTG CTC CAG CTC AAC TGG TAA CCT GCA CCT GGT CG-3' (MWG-Biotech AG), were 5'-end-labeled with ATP using standard protocols. The labeled oligonucleotide was heated at 95°C for 3 min and annealed to 5 μg of a complementary strand at 65°C for 1 min then cooled slowly to 4°C. The annealed product, containing a core match site (5'-CCGG-3') and a single T·G mismatch site (5'-CTGG-3') that are underlined above, was separated on a 3% agarose gel, excised and purified with a Zymoclean gel/DNA recovery kit (Anachem) using standard procedures. The polyamide f-ImImIm was incubated with 0.5 μg of the double-stranded DNA in 10 mM Tris pH 7.0, 1 mM EDTA, 50 mM KCl, 1 mM MgCl2, 0.5 mM DTT and 20 mM HEPES, at room temperature for 30 min, in a total volume of 50 μl. DNA cleavage was initiated by the addition of the polyamide treated sample to 2 μl (1 U) DNase I diluted in cold 10 mM Tris pH 7.0 from a stock solution (1 U/μl, Promega) and 1 μl of a solution of 250 mM MgCl2, 250 mM CaCl2. The reactions were performed at room temperature and stopped after 3 min by the addition of 100 μl of a stop mix containing 200 mM NaCl, 30 mM EDTA pH 8 and 1% SDS. The cleavage products were phenol/chloroform extracted and ethanol precipitated in the presence of 1 μl of glycogen (20 mg/ml; Roche Diagnostics), washed once in 80% ethanol and lyophilized. The samples were resuspended in formamide loading dye, denatured for 5 min at 90°C, cooled on ice and electrophoresed at 2000 V for 2 h on a 10% denaturing acrylamide gel (Sequagel, National Diagnostics). The gels were dried under vacuum at 80°C and exposed to film for 24 h (X-OMAT, Kodak).
Isothermal titration calorimetry and biosensor-SPR. Calorimetric titrations were performed with a VP-ITC (Microcal, Inc., Northampton, MA) or a CSC 4200 ITC (Calorimetry Science Corp., Spanish Fork, UT). Software provided with the calorimeters is used for control and data collection. Electrical calibration of the instruments was performed by applying heat pulses using built-in modules. Chemical calibrations were performed by using: HCl/THAM, RNase A/2'CMP, and BaCl2/18-Crown-6 (Aldrich Chem. Co., Milwaukee, WI). The enthalpy values obtained in the calibrations differed by <2% from literature values (26–28).
ITC experiments were conducted by injecting 6–15 μl (depending on the compounds concentrations and heat generated) of the polyamides every 300 s for a total of 16–34 injections into DNA solution The concentrations were 0.0585, 0.2228 and 0.200 mM for f-ImImIm, f-PyImIm and distamycin, respectively. DNA hairpin concentrations were from 0.0035 to 0.010 mM. Similar experiments were performed to determine the heats of dilution of the polyamides with the DNA solution substituted by buffer. An experimental upper limit of 35°C was selected based on the Tm of the DNA, which was determined to be over 40°C. The heat produced for each injection of compound into DNA or buffer was obtained by integration of the area under each peak of the titration plots with respect to time. The heats of reaction were obtained by subtraction of the integrated heats of dilution of the compounds from the heats corresponding to the injection of compound into DNA. The heat of titration of the ligand into buffer varies slightly over the ratio range. Because of the variation, point-to-point rather than averaged subtraction was applied to ITC titration data. Data corresponding to the first injection were discarded. The binding enthalpy (H°) for each titration was obtained by fitting the results of heat per mole as a function of total molar ratio (polyamide/DNA).
SPR experiments were conducted using a BIACORE 2000 with an SA chip. The DNA hairpins were purchased with 5'-biotinated. Immobilizations, experimental details and data processing procedures have been previously described (6,21).
Calculation of solvent accessible surface area. Changes in solvent accessible surface area (SASA) upon complex formation were calculated using the GRASP program (29) with a probe radius of 1.4. Computations were performed for all atom NMR structures for f-ImImIm bound to the 5'-GAACTGGTTC-3' duplex (CTGGdm) containing a core having two adjacent T·G mismatches (5). Carbon, carbon-bound hydrogen and phosphorus were defined as nonpolar atoms, while all other atoms were defined as polar. The change in SASA (SASA) on complex formation was calculated as the difference between the SASA of the complex and the sum of the SASAs of the free DNA duplex and the free compound in the bound conformation (SASA = SASAcomplex – ). Total contributions of the polar and nonpolar atoms to the surface areas were also obtained to calculate the predicted heat capacity changes (Cp°). The Cp° values were calculated using the equation derived from protein folding (30) and recently modified for small molecule–DNA complexes (31):
Cp° = (0.382·Anp – 0.121·Ap) cal/mol K
RESULTS
Selective recognition of T·G mismatches by f-ImImIm
Although binding of f-ImImIm to single and double mismatch sites has been demonstrated in short DNA oligomers, selective binding has not been confirmed in the context of a long duplex sequence. For direct comparison of the binding of f-ImImIm to match and mismatch sites, DNase I footprinting was used with an 80 base DNA containing Watson–Crick CCGG and single T·G mismatch (5'-CTGG-3') sites (Fig. 2). Footprints were observed at both the T·G and match sites but footprints at the mismatch site appear at concentrations of 10 μM while binding at the match site required more than 5-fold higher concentrations. No binding was observed at other sites in the DNA duplex under the footprinting conditions. At the highest concentration of 50 μM, the footprints are larger than expected for the polyamide size. This has been observed with other polyamides and probably arises from nonspecific interactions at the high concentrations.
Figure 2. DNase I footprinting results for the f-ImImIm complex with match (top) and single T·G mismatch (bottom) DNA sequences. Concentrations of f-ImImIm increase from 1 to 50 μM and the f-ImImIm gel lanes are compared to a G/A sequencing ladder and to uncleaved DNA (–). The protected regions are denoted by gray and black rectangles, respectively. Density plots (far right) illustrate a decrease in band intensity upon increasing concentrations of f-ImImIm with protection of DNA cleavage at the single T·G mismatch indicated with a bold arrow.
Binding constants (K), cooperativity and stoichiometry
Biosensor-SPR sensorgrams for the interaction of f-ImImIm with DNA hairpin duplexes containing CCGG and CTGGsm binding sites (Fig. 1) are in Figure 3. The maximum instrument response (RUmax) obtained in the steady-state region corresponds to approximately twice the predicted response for binding of one molecule of the compound and indicates a 2:1 stoichiometry (2 mol of compound/mol of DNA hairpin) in agreement with previous studies (5,6). As can be seen in the figure, the kinetics for binding of the polyamide differ markedly for the different DNA sequences: f-ImImIm dissociates much more slowly from the mismatch complex (Fig. 3). Binding constants can be determined from fitting the steady state region of the sensorgrams, as previously described (6,21), or from fitting kinetics in the association and dissociation regions. As expected, all results were best fit with a model that has a 2:1 polyamide to DNA stoichiometry. For binding of f-ImImIm to both matched and mismatched sequences, K1 is significantly less than K2 indicating positive cooperativity in binding of the two molecules of f-ImImIm. Because of correlation between K1 and K2, the error in fitting individual values of K1 and K2 is larger than that of the average (K = 1/2), for dimer complex formation. For comparison of the binding of polyamides to DNA, the average K values determined from the steady-state region are converted to binding free energies and reported in Table 1. Reporting the square root values places all results on a per-bound molecule basis and allows direct comparison between binding of monomers and dimers as well as with literature results. The rate constants obtained from fitting the kinetics curves allow calculation of K and with the polyamides these values agree with those from the steady-state analysis. Biosensor-SPR methods were also used to determine K as a function of temperature and the constants changed very little with temperature (Supplementary Material, Table S1). In summary, the presence of a T·G base pair in a DNA sequence has a pronounced influence on the binding of f-ImImIm and yields a much higher affinity than for a similar perfectly matched sequence.
Figure 3. Typical SPR sensorgrams for the interaction of f-ImImIm with the CCGG hairpin (top) and CTGGsm hairpin (bottom) at 25°C. The maximum responses, with the same amount of DNA immobilized on each surface of the biosensor, are the same in both cases at high concentration indicating the same stoichiometry. Remarkable differences in binding kinetics are observed: fast on/off kinetics for f-ImImIm–CCGG and slower kinetics for the f-ImImIm–CTGGsm complex. The compound concentrations are from 0.75 to 26 μM with CCGG and 0.01 to 7 μM with CTGGsm.
Table 1. Thermodynamic parameters for the interaction of polyamides with matched and mismatched DNA oligomers
Calorimetric analysis of binding. Large exothermic enthalpies are observed in ITC titrations of f-ImImIm into CCGG and CTGGsm (examples in Fig. 4). The titration heats for the interaction of f-ImImIm with the CCGG hairpin are smaller than for CTGGsm and are spread over a broad ratio range as expected for weaker binding (6). The titration heat was converted to heat per mole as a function of total molar ratio, and point-to-point subtraction of dilution heat was applied. The data were fitted with a 2:1 model to give molar binding enthalpies (H°). Due to the large correlation among the H° values for binding of the two f-ImImIm molecules in a cooperative complex, the average binding enthalpy, rather than the individual enthalpies (H1° and H2°), is reported in Table 1. Titrations of f-PyImIm into the CCGG hairpin (Fig. 1) provide an excellent model for comparison of strong matched site binding with the mismatched complex of f-ImImIm. The enthalpic contributions are highly favorable for binding of both polyamides to their DNA recognition sequences. The H° values along with G° values from SPR allow calculation of S° for binding (Table 1). The S° values are negative, an unfavorable term in the G° of binding. Although the average enthalpic contribution for the formation of f-ImImIm/CTGGsm and f-PyImIm/CCGG complexes is comparable, the entropic contribution is less favorable for f-PyImIm binding to CCGG at 298 K. The enthalpy is strongly temperature dependent and decreases (becomes more exothermic) as the temperature is increased (Fig. 5, top). Similar results were obtained for the interaction of f-PyImIm with its cognate CCGG sequence (Fig. 5, bottom). The experimental heat capacity changes for formation of the complexes were obtained from linear fitting of H° as a function of temperature (Table 1) and all Cp° values are negative.
Figure 4. Baseline corrected ITC results, before heat of dilution subtraction, for titration of f-ImImIm into the CCGG hairpin in MES 20 buffer 25°C (top). Every peak represents the heat released from a 10 μl injection of 0.0585 mM f-ImImIm into 3.5 μM of DNA hairpin. The inset is the corrected molar heat/addition, after subtraction heat of dilution, with the best-fit curve for a 2:1 model. Similar titrations with the CTGGsm hairpin are shown (bottom). The relative small heat of dilution of f-ImImIm can be seen in the last two to three injections when all of DNA sites are saturated. It is interesting that distinct differences in the heat profiles are observed but that the averaged total binding heats are comparable (Table 1).
Figure 5. The dependence of the thermodynamic parameters on temperature for binding of f-ImImIm to CTGGsm (top) and f-PyImIm to CCGG (bottom). The entropy TS° (diamonds) was calculated from the experimental G° (squares) obtained from SPR, and H° (filled circles) obtained from ITC. The slopes of H° as a function of temperature provide the Cp° values for complex formation (Table 1).
The interactions of polyamides vary significantly with DNA sequence as shown in previous studies of distamycin binding to AT-rich sequences (2,6,17,21). Unlike the imidazole containing polyamides, formation of the distamycin 2:1 complex with an A3T3 sequence is governed by negative cooperativity with strong affinity in the first interaction (K1 = 3.1 x 108 M–1, G1° = –11.6 kcal/mol) followed by weaker binding of a second molecule (K2 = 8.0 x 105 M–1, G2° = –8.1 kcal/mol) (6). ITC titrations of distamycin were conducted as a model system for comparison of positive and negative cooperativity in dimer interactions (Supplementary Material, Fig. S1). Binding of the two molecules of distamycin to DNA is clearly distinguished in the plot, which has a quite different shape than with the Im polyamide titrations. Binding of the first distamycin has a lower enthalpy (H1° = –12.2 kcal/mol) than that of the second (H2° = –23.9 kcal/mol). The entropic contributions of –0.6 and –15.9 kcal/mol for distamycin binding to the two sites indicate that the entropic contribution is near zero for binding of the first molecule but is very unfavorable for the second. These results indicate that, as with the imidazole polyamides, distamycin binding is enthalpy driven.
Calculated heat capacity changes. Changes in SASA on complex formation have been correlated with Cp° values for protein–protein, protein–DNA (32,33) and more recently for drug–DNA interactions (34,35). The Cp° values for the f-ImImIm–CTGGdm complex were calculated based on buried surface areas determined by using the reported NMR structure (5) with previously described methods (31,36). Although the structure for the complex with the CTGGsm sequence is not available, the values obtained with the CTGGdm sequence could be extrapolated to the complex formed with the single mismatch. The calculated heat capacity changes using two different sets of parameters are negative and in general agreement with the experimental values (Supplementary Material, Table S2).
DISCUSSION
Previous biosensor binding studies have shown that f-ImImIm and other polyamides bind to short oligonucleotides (matched or mismatched sequences) with significantly different affinities (6,21). Fox and coworkers have also recently shown that some, but not all, chemical and enzymatic cleavage agents can recognize T·G (and other) mismatched base pairs in a long DNA sequence (37). The first question that we wished to answer in this study was whether f-ImImIm could selectively recognize a single T·G mismatch site in competition with excess Watson–Crick sites. Footprints from DNase I cleavage of an 80 base DNA sequence with a single T·G mismatch (Fig. 2) show strong protection from cleavage by f-ImImIm at the T·G site. These results are in agreement with the oligonucleotide studies by SPR that indicate stronger binding of f-ImImIm to the mismatched CTGGsm over the matched CCGG site.
A full thermodynamic data set is essential to complement the structural information available for T·G recognition by f-ImImIm and to provide a complete understanding the energetic basis for the T·G recognition. The values of H° and Cp° for binding reactions are most reliably obtained by calorimetry rather than indirectly from van’t Hoff determinations (38–41). Calorimetry, on the other hand, may provide a less accurate K for strong binding reactions due to the relatively high concentrations that may have to be used if reaction heats are low. We have found biosensor-SPR methods to be particularly valuable in studies of DNA oligomer complexes with small molecules such as the polyamides (6,21) since a wide range of concentrations can be used and both kinetic and equilibrium results are obtained. The SPR studies are particularly useful when the K values are high, as with the f-ImImIm/T·G complex.
Biosensor-SPR methods clearly show that the polyamides form cooperative 2:1 complexes with their DNA recognition sites. The G° values for binding are essentially constant with temperature suggesting strong enthalpy–entropy compensation in this temperature range (Fig. 5). Such compensation is observed for many biomolecular processes and suggests a significant hydrophobic component to the binding energetics (31,42–44). The SPR results quantitatively show that f-ImImIm binds much more strongly to the mismatched CTGGsm sequence than to the equivalent DNA with Watson–Crick base pairs. The G° value for binding to T·G sites is 2 kcal/mol more favorable than for binding to equivalent CG sites. The improved binding of f-ImImIm to the T·G mismatch is correlated with much slower dissociation kinetics relative to the matched CCGG complex (6,21). The slow dissociation kinetics for the mismatch complex could arise from at least two factors: the breakage of strong H-bonds (from the amino protons of G to f-ImImIm) and the rearrangement/re-uptake of water molecules in the DNA minor groove as described below.
Calorimetric studies are also consistent with formation of cooperative 2:1 complexes of f-ImImIm with both matched CG and mismatched T·G DNA sequences and with stronger affinity for the mismatch. The difference in the heat profiles for f-ImImIm binding to CTGGsm and CCGG (Fig. 4) arises both from differences in individual enthalpies (H1° and H2°) and binding constants (K1 and K2) in the 2:1 complexes. Somewhat surprisingly given all of these differences, the calorimetric titrations suggest that the averaged total heats for dimer formation are comparable for formation of f-ImImIm–CTGGsm, f-ImImIm–CCGG, and f-PyImIm–CCGG complexes at 298 K. This is particularly surprising since comparison of the H° and G° values for binding (Table 1) indicates that the enthalpic contributions play the primary role in driving the interactions. Given that the calorimetric H° values for binding are more negative than the total free energies of binding, there must be unfavorable compensating entropic contributions in each binding process. Since the H° values for these interactions are similar at 298 K, it is the difference in the unfavorable entropic contribution that dictates the different binding affinities. The higher binding affinity for the f-ImImIm–CTGGsm complex means the S° for formation of this complex must be less unfavorable than that of the complex with CCGG.
It is important at this point to combine results from the structural and thermodynamic studies to suggest possible explanations the similar H° but quite different G° and S° for binding of f-ImImIm to match and mismatch sites. NMR and CD results suggest that significant conformational changes do not occur in the DNA binding sites on formation of these complexes (5,6). These observations imply that the entropic effects do not come from overall structural changes but from factors such as the loss of translational and rotational entropy of the polyamides and the opposing gain of entropy due to release of bound water molecules on complex formation (although the net change in bound water is unknown). Since the polyamides would be expected to be similarly hydrated and to release similar amounts of hydration water on complex formation, any differences in water released, among the complexes, would have to be due to differences in DNA hydration. Structural studies indicate that the T·G base pair adopts a wobble conformation with an unpaired free amino group of G that is strongly hydrated as part of the minor groove hydration matrix (12,45). On complex formation with f-ImImIm some of these tightly bound and highly ordered water molecules would be released from the T·G site to provide a favorable entropic contribution to complex formation. From this model the entropic contributions are predicted to be less unfavorable for formation of the f-ImImIm–CTGGsm complex than for the CCGG complex and this results in the enhanced affinity for the T·G complex. The difference in G° between the f-ImImIm–CTGGsm and –CCGG complexes (2 kcal/mol and mainly from entropy) is about the same order of magnitude as the entropy gain by releasing a single-bound water molecule from its site (46). Binding of a netropsin analog to a mixed sequence DNA suggested that water is actually taken up upon complex formation (47), and it is difficult to see how this could enhance binding entropy. Initial molecular dynamics studies in our laboratory have suggested that minor groove hydration is quite varied among different AT DNAs, and it is possible that water release or uptake is quite dependent on different DNA sequences and ligands.
The negative Cp° values for complex formation can provide useful insights about interactions and hydration (30–36). There are significant differences in the heat capacity changes for the polyamides and DNA sequences used in these experiments. The Cp° for the f-ImImIm–CTGGsm complex, for example, is more negative than that for f-PyImIm–CCGG (Table 1). These Cp° values indicate important differences in the formation of complexes with the CCGG DNA versus the T·G mismatch CTGGsm sequence. Structured water, such as the water of hydrophobic hydration, can have a large heat capacity and release of such water, for example on transfer of nonpolar groups into the interior of proteins or to the minor groove of DNA, contributes a negative term to the binding Cp°. Since all of the polyamide complexes examined in this study have negative Cp°, the binding enthalpy becomes more favorable (more negative) while the binding entropy is less favorable (also more negative) as the temperature increases. Such effects are larger in systems where bound and ordered water molecules are more strongly associated with the DNA groove as with CTGGsm. Calculation of buried surface areas from the NMR structures of f-ImImIm–CTGGdm and f-ImImIm–CCGG indicate that the observed Cp° are correlated with water release from buried nonpolar and polar molecular surface.
Binding of polyamides to GC containing sequences has quite different characteristics than binding to AT-rich sequences. Binding of many imidazole-polyamides to sequences with GC or T·G mismatched base pairs displays positive cooperativity, while distamycin binds to A3T3 with negative cooperativity (6). This difference may be explained in terms of minor groove width. The groove of AT is unusually narrow (0.3–0.4 nm) (48), but it is wider in mixed DNA sequences (0.5–0.6 nm) (49) as well as at T·G base pairs (12). Since the thickness of one polyamide ring is 0.34 nm (50), the first distamycin molecule can fit snugly and make favorable van der Waals contacts with the walls of the minor groove in A3T3. For binding of the second molecule, however, the minor groove must widen sufficiently to accommodate the two stacked distamycin molecules in a 2:1 complex. On the other hand, the groove width of GC-rich or T·G sequences is wider and can bind two polyamides in a positively cooperative manner even at low ratios of compound to DNA. Interestingly, the experimental K and H° for distamycin binding are clearly resolved (17,51 and this work) in ITC while those for polyamide binding with positive cooperativity are overlapped and difficult to separate. In summary, the results presented here provide a thermodynamic rationale for the specificity in molecular recognition of T·G mismatch base pairs by Im-polyamides.
SUPPLEMENTARY MATERIAL
ACKNOWLEDGEMENTS
This work was supported by NIH Grant GM61587, the Georgia Research Alliance (W.D.W.), and by the NSF-REU and Research Corporation (M.L.).
REFERENCES
Bailly,C. and Chaires,J.B. (1998) Sequence-specific DNA minor groove binders. Design and synthesis of netropsin and distamycin analogues. Bioconjug. Chem., 9, 513–538.
Wemmer,D.E. (2000) Designed sequence-specific minor groove ligands. Annu. Rev. Biophys Biomol. Struct., 29, 439–461.
Wemmer,D.E. (2001) Ligands recognizing the minor groove of DNA: Development and applications. Biopolymers, 52, 197–211.
Dervan,P.B. (2001) Molecular recognition of DNA by small molecules. Bioorg. Med. Chem., 9, 2215–2235.
Yang,X.L., Hubbard,R.B., Lee,M., Tao,Z.F., Sugiyama,H. and Wang,A.H. (1999) Imidazole-imidazole pair as a minor groove recognition motif for T:G mismatched base pairs. Nucleic Acids Res., 27, 4183–4190.
Lacy,E.R., Cox,K.K., Wilson,W.D. and Lee,M. (2002) Recognition of T*G mismatched base pairs in DNA by stacked imidazole-containing polyamides: surface plasmon resonance and circular dichroism studies. Nucleic Acids Res., 30, 1834–1841.
Land,H., Parada,L.F. and Weinberg,R.A. (1983) Cellular oncogenes and multistep carcinogenesis. Science, 222, 771–778.
Almoguera,C., Shibata,D., Forrester,K., Martin,J., Arnheim,N. and Perucho,M. (1988) Most human carcinomas of the exocrine pancreas contain mutant c-K-ras genes. Cell, 53, 549–554.
Watanabe,H., Ha,A., Hu,Y.X., Ohtsubo,K., Yamaguchi,Y., Motoo,Y., Okai,T., Toya,D., Tanaka,N. and Sawabu,N. (1999) K-ras mutations in duodenal aspirate without secretin stimulation for screening of pancreatic and biliary tract carcinoma. Cancer, 86, 1441–1448.
Lall,L. and Davidson,R.L. (1998) Sequence-directed base mispairing in human oncogenes. Mol. Cell. Biol., 18, 4659–4669.
Hare,D., Shapiro,L. and Patel,D.J. (1986) Wobble dG X dT pairing in right-handed DNA: solution conformation of the d(C-G-T-G-A-A-T-T-C-G-C-G) duplex deduced from distance geometry analysis of nuclear Overhauser effect spectra. Biochemistry, 25, 7445–7456.
Hunter,W.N., Brown,T., Kneale,G., Anand,N.N., Rabinovich,D. and Kennard,O. (1987) The structure of guanosine-thymidine mismatches in B-DNA at 2.5-? resolution. J. Biol. Chem., 262, 9962–9970.
Lacy,E.R., Madsen,E.M., Lee,M. and Wilson,W.D. (2003) Polyamide dimer stacking in the DNA minor groove and recognition of T.G mismatched base pairs in DNA. In Demeunynck,M., Bailly,C. and Wilson,W.D. (eds), DNA and RNA Binders: From Small Molecules to Drugs. WILEY-VCH, Weinheim, Vol. 2, pp. 384–413.
Rucker,V.C., Foister,S., Melander,C. and Dervan,P.B. (2003) Sequence specific fluorescence detection of double strand DNA. J. Am. Chem. Soc., 125, 1195–1202.
Pilch,D.S., Poklar,N., Gelfand,C.A., Law,S.M., Breslauer,K.J., Baird,E.E. and Dervan,P.B. (1996) Binding of a hairpin polyamide in the minor groove of DNA: sequence-specific enthalpic discrimination. Proc. Natl Acad. Sci. USA, 93, 8306–8311.
Pilch,D.S., Poklar,N., Baird,E.E., Dervan,P.B. and Breslauer,K.J. (1999) The thermodynamics of polyamide-DNA recognition: hairpin polyamide binding in the minor groove of duplex DNA. Biochemistry, 38, 2143–2151.
Rentzeperis,D., Marky,L.A., Dwyer,T.J., Geierstanger,B.H., Pelton,J.G. and Wemmer,D.E. (1995) Interaction of minor groove ligands to an AAATT/AATTT site: correlation of thermodynamic characterization and solution structure. Biochemistry, 34, 2937–2945.
Cooper,A., Johnson,C.M., Lakey,J.H. and Nollmann,M. (2001) Heat does not come in different colours: entropy-enthalpy compensation, free energy windows, quantum confinement, pressure perturbation calorimetry, solvation and the multiple causes of heat capacity effects in biomolecular interactions. Biophys. Chem., 93, 215–230.
Morton,C.J. and Ladbury,J.E. (1996) Water-mediated protein-DNA interactions: the relationship of thermodynamics to structural detail. Protein Sci., 5, 2115–2118.
Privalov,G.P. and Privalov,P.L. (2000) Problems and prospects in microcalorimetry of biological macromolecules. Methods Enzymol., 323, 31–62.
Lacy,E.R., Le,N.M., Price,C.A., Lee,M. and Wilson,W.D. (2002) Influence of a terminal formamido group on the sequence recognition of DNA by polyamides. J. Am. Chem. Soc., 124, 2153–2163.
Xodo,L.E., Manzini,G., Quadrifoglio,F., van der Marel,G.A. and van Boom,J.H. (1988) Oligodeoxynucleotide folding in solution: loop size and stability of B- hairpins. Biochemistry, 27, 6321–6326.
Cantor,C.R. and Tinoco,I.,Jr (1965) Absorption and optical rotatory dispersion of seven trinucleoside diphosphates. J. Mol. Biol., 13, 65–77.
Cantor,C.R., Warshaw,M.M. and Shapiro,H. (1970) Oligonucleotide interactions. 3. Circular dichroism studies of the conformation of deoxyoligonucleotides. Biopolymers, 9, 1059–1077.
Fasman,G.D. (1975) Handbook of Biochemistry and Molecular Biology, Nucleic Acids, 3rd Edn. CRC Press, Cleveland, OH.
Wiseman,T., Williston,S., Brandts,J.F. and Lin,L.N. (1989) Rapid measurement of binding constants and heats of binding using a new titration calorimeter. Anal. Biochem., 179, 131–137.
Horn,J.R., Russell,D., Lewis,E.A. and Murphy,K.P. (2001) Van’t Hoff and calorimetric enthalpies from isothermal titration calorimetry: are there significant discrepancies? Biochemistry, 40, 1774–1778.
Briggner,L.E. and Wadso,I. (1991) Test and calibration processes for microcalorimeters, with special reference to heat conduction instruments used with aqueous systems. J. Biochem. Biophys. Methods, 22, 101–118.
Nicholls,A., Sharp,K.A. and Honig,B. (1991) Protein folding and association: insights from the interfacial and thermodynamic properties of hydrocarbons. Proteins, 11, 281–296.
Spolar,R.S., Livingstone,J.R. and Record,M.T.,Jr (1992) Use of liquid hydrocarbon and amide transfer data to estimate contributions to thermodynamic functions of protein folding from the removal of nonpolar and polar surface from water. Biochemistry, 31, 3947–3955.
Ren,J., Jenkins,T.C. and Chaires,J.B. (2000) Energetics of DNA intercalation reactions. Biochemistry, 39, 8439–8447.
Makhatadze,G.I. and Privalov,P.L. (1990) Heat capacity of proteins. I. Partial molar heat capacity of individual amino acid residues in aqueous solution: hydration effect. J. Mol. Biol., 213, 375–384.
Privalov,P.L. and Makhatadze,G.I. (1990) Heat capacity of proteins. II. Partial molar heat capacity of the unfolded polypeptide chain of proteins: protein unfolding effects. J. Mol. Biol., 213, 385–391.
Mazur,S., Tanious,F.A., Ding,D., Kumar,A., Boykin,D.W., Simpson,I.J., Neidle,S. and Wilson,W.D. (2000) A thermodynamic and structural analysis of DNA minor-groove complex formation. J. Mol. Biol., 300, 321–337.
Haq,I., Ladbury,J.E., Chowdhry,B.Z., Jenkins,T.C. and Chaires,J.B. (1997) Specific binding of hoechst 33258 to the d(CGCAAATTTGCG)2 duplex: calorimetric and spectroscopic studies. J. Mol. Biol., 271, 244–257.
Spolar,R.S. and Record,M.T.,Jr (1994) Coupling of local folding to site-specific binding of proteins to DNA. Science, 263, 777–784.
Brown,J., Brown,T. and Fox,K.R. (2003) Cleavage of fragments containing DNA mismatches by enzymic and chemical probes. Biochem. J., 371, 697–708.
Janin,J. (1995) Elusive affinities. Proteins, 21, 30–39.
Ladbury,J.E. (1995) Counting the calories to stay in the groove. Structure, 3, 635–639.
Naghibi,H., Tamura,A. and Sturtevant,J.M. (1995) Significant discrepancies between van’t Hoff and calorimetric enthalpies. Proc. Natl Acad. Sci. USA, 92, 5597–5599.
Chaires,J.B. (1997) Possible origin of differences between van’t Hoff and calorimetric enthalpy estimates. Biophys. Chem., 64, 15–23.
Chaires,J.B. (1997) Energetics of Drug-DNA Interactions. Biopolymers, 44, 201–215.
Gallagher,K. and Sharp,K. (1998) Electrostatic contributions to heat capacity changes of DNA-ligand binding. Biophys. J., 75, 769–776.
Matulis,D., Rouzina,I. and Bloomfield,V.A. (2002) Thermodynamics of cationic lipid binding to DNA and DNA condensation: roles of electrostatics and hydrophobicity. J. Am. Chem. Soc., 124, 7331–7342.
Yang,X.L., Kaenzig,C., Lee,M. and Wang,A.H. (1999) Binding of AR-1–144, a tri-imidazole DNA minor groove binder, to CCGG sequence analyzed by NMR spectroscopy. Eur. J. Biochem., 263, 646–655.
Dunitz,J.D. (1994) The entropic cost of bound water in crystals and biomolecules. Science, 264, 670.
Sidorova,N. and Rau,D.C. (1995) The osmotic sensitivity of netropsin analogue binding to DNA. Biopolymers, 35, 377–384.
Fratini,A.V., Kopka,M.L., Drew,H.R. and Dickerson,R.E. (1982) Reversible bending and helix geometry in a B-DNA dodecamer: CGCGAATTBrCGCG. J. Biol. Chem., 257, 14686–14707.
Yoon,C., Prive,G.G., Goodsell,D.S. and Dickerson,R.E. (1988) Structure of an alternating-B DNA helix and its relationship to A-tract DNA. Proc. Natl Acad. Sci. USA, 85, 6332–6336.
Pelton,J.G. and Wemmer,D.E. (1990) Binding modes of distamycin A with d(CGCAAATTTGCG)2 determined by two-dimensional NMR. J. Am. Chem. Soc., 112, 1393–1399.
Wilson,W.D., Tanious,F.A., Barton,H.J., Wydra,R.L., Jones,R.L., Boykin,D.W. and Strekowski,L. (1990) The interaction of unfused polyaromatic heterocycles with DNA: intercalation, groove-binding and bleomycin amplification. Anticancer Drug Des., 5, 31–42.(Eilyn R. Lacy1, Binh Nguyen1, Minh Le1,2)
*To whom correspondence should be addressed. Tel: +1 404 651 3903; Fax: +1 404 651 1416; Email: chewdw@panther.gsu.edu
Correspondence may also be addressed to Moses Lee. Tel: +1 864 294 3368; Fax: +1 864 294 3559; Email: moses.lee@furman.edu
ABSTRACT
To complement available structure and binding results and to develop a detailed understanding of the basis for selective molecular recognition of T·G mismatches in DNA by imidazole containing polyamides, a full thermodynamic profile for formation of the T·G–polyamide complex has been determined. The amide-linked heterocycles f-ImImIm and f-PyImIm (where f is formamido group, Im is imidazole and Py is pyrrole) were studied by using biosensor-surface plasmon resonance (SPR) and isothermal titration calorimetry (ITC) with a T·G mismatch containing DNA hairpin duplex and a similar DNA with only Watson–Crick base pairs. Large negative binding enthalpies for all of the polyamide–DNA complexes indicate that the interactions are enthalpically driven. SPR results show slower complex formation and stronger binding of f-ImImIm to the T·G than to the match site. The thermodynamic analysis indicates that the enhanced binding to the T·G site is the result of better entropic contributions. Negative heat capacity changes for the complex are correlated with calculated solvent accessible surface area changes and indicate hydrophobic contributions to complex formation. DNase I footprinting analysis in a long DNA sequence provided supporting evidence that f-ImImIm binds selectively to T·G mismatch sites.
INTRODUCTION
Polyamide DNA binding agents are of interest due to their sequence-specific binding characteristics (1–3), their ability to recognize mismatched base pairs, and their potential for use in modulation of gene expression (4). Through the combination of pyrrole (Py), imidazole (Im), and other heterocycles in a side-by-side pairing, polyamides capable of specifically recognizing Watson–Crick DNA sequences have been synthesized (1–3). In addition, structural, binding affinity and kinetics studies have shown the potential use of the side-by-side pairing of Im heterocycles for recognition of DNA sequences containing non-Watson–Crick base pairs (5,6). The amide-linked triimidazole, f-ImImIm where f is formamido (Fig. 1A), can discriminate T·G from Watson–Crick C·G base pairs as well as from other mismatches (5,6).
Figure 1. (A) Chemical structures of the polyamides: distamycin A, f-ImImIm and f-PyImIm. (B) Sequences of the DNA hairpins used in this study: CCGG and A3T3 are Watson–Crick sequences while CTGGsm contains one T·G mismatch. (C) Schematic models of the complexes formed between f-ImImIm and the DNA sequences containing a double T·G mismatch (left) and CCGG (right) as inferred from NMR data.
Possible uses of the Im/Im pair include rapid detection of single nucleotide polymorphisms (5,6). Rapid and accurate detection of T·G mismatch base pairs in DNA would be of significant benefit since this mismatch is responsible for most of the common mutations leading to formation of tumors in humans. In human bladder carcinoma, for example, a G·C to A·T transition at the 3'G of the GG doublet in codon 12 of the Ha-ras and Ki-ras proto-oncogene converts them to oncogenes (7–9). A T·G mismatch is often introduced by spontaneous deamination of 5-methylcytosine and can arise from errors in replication (10). Structural analysis of DNAs containing T·G mismatches, indicates that the mismatches adopt a wobble conformation and structural perturbations are mainly in the vicinity of the mismatch (11,12). NMR studies show that f-ImImIm forms an anti-parallel dimer in the DNA minor groove with a T·G mismatch site (5,6,13,14). In the mismatch complex the guanine-NH2 group of the wobble T·G base pair forms two specific hydrogen bonds to the side-by-side Im/Im pair (5). Biosensor-surface plasmon resonance (SPR) analysis has shown that f-ImImIm is capable of strongly and specifically binding to a DNA sequence with only a single T·G mismatch (6,13). This novel T·G mismatch recognition motif has recently been corroborated by results from the Dervan laboratory, in which a hairpin polyamide containing a stacked Im/Im pair was found to bind at a T·G mismatch site (14).
There is a wealth of information on structure, binding affinity and specificity for the interactions of polyamides with DNA (1–3), however, the specific energetic factors that govern the binding events have rarely been characterized. Important calorimetric results have been reported for a very limited number of polyamides in complex with their target DNA sequences (15–17). Thermodynamic studies can provide valuable complementary information to the structural perspective and can improve our understanding of the molecular forces that govern the affinity and specificity of binding. Specifically, they can give additional insights into the energetic basis for the discrimination of the T·G mismatch from matched sites by f-ImImIm. We report here biosensor-SPR and calorimetric studies as a function of temperature as well as footprinting studies in a large sequence context for the interaction of f-ImImIm and f-PyImIm (Fig. 1A) with mismatched and matched sequence DNAs. By conducting the studies as a function of temperature, we were able to determine the heat capacity changes on binding that are critical for a complete thermodynamic understanding of biomolecular interactions (18–20).
MATERIALS AND METHODS
Reagents and biochemicals
Compounds and buffers. Distamycin A was purchased from Sigma Chemical Co. and used without further purification. N--1-methyl-4-{1-methyl-4-imidazole-2- carboxamido}imidazole-2-carboxamide (f-ImImIm) and N--1-methyl-4-{1-methyl-4-imidazole-2-carboxamido}imidazole-2-carboxamide (f-PyImIm) were prepared as described previously (6,21). Solutions of the compounds for calorimetric studies were prepared in MES20 buffer, 0.01 M MES adjusted to pH 6.25, 0.001 M EDTA and 0.2 M NaCl. Concentrations of the compounds were determined spectrophotometrically by absorbance at 304 nm for distamycin and 303 nm for f-ImImIm and f-PyImIm. The extinction coefficients used were 34 000 M–1 cm–1 for distamycin (17) and 30 700 M–1 cm–1 for f-ImImIm and f-PyImIm at 25°C.
DNA sequences. DNAs were obtained as anionic exchange, HPLC purified products (Midland Certified Reagent Co). Biosensor-SPR binding studies were conducted with 5'-biotinated DNAs, and calorimetric studies were performed with non-biotinated DNAs. Solutions for the ITC experiments were prepared in degassed MES20 buffer. To ensure that the DNA was in the hairpin form, the solutions were heated to 90°C and cooled rapidly in ice (22). The concentration of the DNA solutions was determined spectrophotometrically at 260 nm and 85°C using extinction coefficients per nucleotide of 9264, 9059 and 9400 M–1 cm–1 for CCGG, CTGGsm and A3T3, respectively. The extinction coefficients were calculated on a per strand basis by the nearest-neighbor method and divided by the number of nucleotides per strand (23–25). The extinction coefficients extrapolated to 25°C were 7824, 8107 and 7978 M–1 cm–1 for CCGG, CTGGsm and A3T3, respectively.
Methods
DNase I footprinting of TG mismatched oligonucleotides. Five micrograms of an 80 base DNA 5'-AGG TGA GCA GGT CCA TAC TGG TTT GCA CCT CGA GGT TAC CGG TAT CTG CTC CAG CTC AAC TGG TAA CCT GCA CCT GGT CG-3' (MWG-Biotech AG), were 5'-end-labeled with ATP using standard protocols. The labeled oligonucleotide was heated at 95°C for 3 min and annealed to 5 μg of a complementary strand at 65°C for 1 min then cooled slowly to 4°C. The annealed product, containing a core match site (5'-CCGG-3') and a single T·G mismatch site (5'-CTGG-3') that are underlined above, was separated on a 3% agarose gel, excised and purified with a Zymoclean gel/DNA recovery kit (Anachem) using standard procedures. The polyamide f-ImImIm was incubated with 0.5 μg of the double-stranded DNA in 10 mM Tris pH 7.0, 1 mM EDTA, 50 mM KCl, 1 mM MgCl2, 0.5 mM DTT and 20 mM HEPES, at room temperature for 30 min, in a total volume of 50 μl. DNA cleavage was initiated by the addition of the polyamide treated sample to 2 μl (1 U) DNase I diluted in cold 10 mM Tris pH 7.0 from a stock solution (1 U/μl, Promega) and 1 μl of a solution of 250 mM MgCl2, 250 mM CaCl2. The reactions were performed at room temperature and stopped after 3 min by the addition of 100 μl of a stop mix containing 200 mM NaCl, 30 mM EDTA pH 8 and 1% SDS. The cleavage products were phenol/chloroform extracted and ethanol precipitated in the presence of 1 μl of glycogen (20 mg/ml; Roche Diagnostics), washed once in 80% ethanol and lyophilized. The samples were resuspended in formamide loading dye, denatured for 5 min at 90°C, cooled on ice and electrophoresed at 2000 V for 2 h on a 10% denaturing acrylamide gel (Sequagel, National Diagnostics). The gels were dried under vacuum at 80°C and exposed to film for 24 h (X-OMAT, Kodak).
Isothermal titration calorimetry and biosensor-SPR. Calorimetric titrations were performed with a VP-ITC (Microcal, Inc., Northampton, MA) or a CSC 4200 ITC (Calorimetry Science Corp., Spanish Fork, UT). Software provided with the calorimeters is used for control and data collection. Electrical calibration of the instruments was performed by applying heat pulses using built-in modules. Chemical calibrations were performed by using: HCl/THAM, RNase A/2'CMP, and BaCl2/18-Crown-6 (Aldrich Chem. Co., Milwaukee, WI). The enthalpy values obtained in the calibrations differed by <2% from literature values (26–28).
ITC experiments were conducted by injecting 6–15 μl (depending on the compounds concentrations and heat generated) of the polyamides every 300 s for a total of 16–34 injections into DNA solution The concentrations were 0.0585, 0.2228 and 0.200 mM for f-ImImIm, f-PyImIm and distamycin, respectively. DNA hairpin concentrations were from 0.0035 to 0.010 mM. Similar experiments were performed to determine the heats of dilution of the polyamides with the DNA solution substituted by buffer. An experimental upper limit of 35°C was selected based on the Tm of the DNA, which was determined to be over 40°C. The heat produced for each injection of compound into DNA or buffer was obtained by integration of the area under each peak of the titration plots with respect to time. The heats of reaction were obtained by subtraction of the integrated heats of dilution of the compounds from the heats corresponding to the injection of compound into DNA. The heat of titration of the ligand into buffer varies slightly over the ratio range. Because of the variation, point-to-point rather than averaged subtraction was applied to ITC titration data. Data corresponding to the first injection were discarded. The binding enthalpy (H°) for each titration was obtained by fitting the results of heat per mole as a function of total molar ratio (polyamide/DNA).
SPR experiments were conducted using a BIACORE 2000 with an SA chip. The DNA hairpins were purchased with 5'-biotinated. Immobilizations, experimental details and data processing procedures have been previously described (6,21).
Calculation of solvent accessible surface area. Changes in solvent accessible surface area (SASA) upon complex formation were calculated using the GRASP program (29) with a probe radius of 1.4. Computations were performed for all atom NMR structures for f-ImImIm bound to the 5'-GAACTGGTTC-3' duplex (CTGGdm) containing a core having two adjacent T·G mismatches (5). Carbon, carbon-bound hydrogen and phosphorus were defined as nonpolar atoms, while all other atoms were defined as polar. The change in SASA (SASA) on complex formation was calculated as the difference between the SASA of the complex and the sum of the SASAs of the free DNA duplex and the free compound in the bound conformation (SASA = SASAcomplex – ). Total contributions of the polar and nonpolar atoms to the surface areas were also obtained to calculate the predicted heat capacity changes (Cp°). The Cp° values were calculated using the equation derived from protein folding (30) and recently modified for small molecule–DNA complexes (31):
Cp° = (0.382·Anp – 0.121·Ap) cal/mol K
RESULTS
Selective recognition of T·G mismatches by f-ImImIm
Although binding of f-ImImIm to single and double mismatch sites has been demonstrated in short DNA oligomers, selective binding has not been confirmed in the context of a long duplex sequence. For direct comparison of the binding of f-ImImIm to match and mismatch sites, DNase I footprinting was used with an 80 base DNA containing Watson–Crick CCGG and single T·G mismatch (5'-CTGG-3') sites (Fig. 2). Footprints were observed at both the T·G and match sites but footprints at the mismatch site appear at concentrations of 10 μM while binding at the match site required more than 5-fold higher concentrations. No binding was observed at other sites in the DNA duplex under the footprinting conditions. At the highest concentration of 50 μM, the footprints are larger than expected for the polyamide size. This has been observed with other polyamides and probably arises from nonspecific interactions at the high concentrations.
Figure 2. DNase I footprinting results for the f-ImImIm complex with match (top) and single T·G mismatch (bottom) DNA sequences. Concentrations of f-ImImIm increase from 1 to 50 μM and the f-ImImIm gel lanes are compared to a G/A sequencing ladder and to uncleaved DNA (–). The protected regions are denoted by gray and black rectangles, respectively. Density plots (far right) illustrate a decrease in band intensity upon increasing concentrations of f-ImImIm with protection of DNA cleavage at the single T·G mismatch indicated with a bold arrow.
Binding constants (K), cooperativity and stoichiometry
Biosensor-SPR sensorgrams for the interaction of f-ImImIm with DNA hairpin duplexes containing CCGG and CTGGsm binding sites (Fig. 1) are in Figure 3. The maximum instrument response (RUmax) obtained in the steady-state region corresponds to approximately twice the predicted response for binding of one molecule of the compound and indicates a 2:1 stoichiometry (2 mol of compound/mol of DNA hairpin) in agreement with previous studies (5,6). As can be seen in the figure, the kinetics for binding of the polyamide differ markedly for the different DNA sequences: f-ImImIm dissociates much more slowly from the mismatch complex (Fig. 3). Binding constants can be determined from fitting the steady state region of the sensorgrams, as previously described (6,21), or from fitting kinetics in the association and dissociation regions. As expected, all results were best fit with a model that has a 2:1 polyamide to DNA stoichiometry. For binding of f-ImImIm to both matched and mismatched sequences, K1 is significantly less than K2 indicating positive cooperativity in binding of the two molecules of f-ImImIm. Because of correlation between K1 and K2, the error in fitting individual values of K1 and K2 is larger than that of the average (K = 1/2), for dimer complex formation. For comparison of the binding of polyamides to DNA, the average K values determined from the steady-state region are converted to binding free energies and reported in Table 1. Reporting the square root values places all results on a per-bound molecule basis and allows direct comparison between binding of monomers and dimers as well as with literature results. The rate constants obtained from fitting the kinetics curves allow calculation of K and with the polyamides these values agree with those from the steady-state analysis. Biosensor-SPR methods were also used to determine K as a function of temperature and the constants changed very little with temperature (Supplementary Material, Table S1). In summary, the presence of a T·G base pair in a DNA sequence has a pronounced influence on the binding of f-ImImIm and yields a much higher affinity than for a similar perfectly matched sequence.
Figure 3. Typical SPR sensorgrams for the interaction of f-ImImIm with the CCGG hairpin (top) and CTGGsm hairpin (bottom) at 25°C. The maximum responses, with the same amount of DNA immobilized on each surface of the biosensor, are the same in both cases at high concentration indicating the same stoichiometry. Remarkable differences in binding kinetics are observed: fast on/off kinetics for f-ImImIm–CCGG and slower kinetics for the f-ImImIm–CTGGsm complex. The compound concentrations are from 0.75 to 26 μM with CCGG and 0.01 to 7 μM with CTGGsm.
Table 1. Thermodynamic parameters for the interaction of polyamides with matched and mismatched DNA oligomers
Calorimetric analysis of binding. Large exothermic enthalpies are observed in ITC titrations of f-ImImIm into CCGG and CTGGsm (examples in Fig. 4). The titration heats for the interaction of f-ImImIm with the CCGG hairpin are smaller than for CTGGsm and are spread over a broad ratio range as expected for weaker binding (6). The titration heat was converted to heat per mole as a function of total molar ratio, and point-to-point subtraction of dilution heat was applied. The data were fitted with a 2:1 model to give molar binding enthalpies (H°). Due to the large correlation among the H° values for binding of the two f-ImImIm molecules in a cooperative complex, the average binding enthalpy, rather than the individual enthalpies (H1° and H2°), is reported in Table 1. Titrations of f-PyImIm into the CCGG hairpin (Fig. 1) provide an excellent model for comparison of strong matched site binding with the mismatched complex of f-ImImIm. The enthalpic contributions are highly favorable for binding of both polyamides to their DNA recognition sequences. The H° values along with G° values from SPR allow calculation of S° for binding (Table 1). The S° values are negative, an unfavorable term in the G° of binding. Although the average enthalpic contribution for the formation of f-ImImIm/CTGGsm and f-PyImIm/CCGG complexes is comparable, the entropic contribution is less favorable for f-PyImIm binding to CCGG at 298 K. The enthalpy is strongly temperature dependent and decreases (becomes more exothermic) as the temperature is increased (Fig. 5, top). Similar results were obtained for the interaction of f-PyImIm with its cognate CCGG sequence (Fig. 5, bottom). The experimental heat capacity changes for formation of the complexes were obtained from linear fitting of H° as a function of temperature (Table 1) and all Cp° values are negative.
Figure 4. Baseline corrected ITC results, before heat of dilution subtraction, for titration of f-ImImIm into the CCGG hairpin in MES 20 buffer 25°C (top). Every peak represents the heat released from a 10 μl injection of 0.0585 mM f-ImImIm into 3.5 μM of DNA hairpin. The inset is the corrected molar heat/addition, after subtraction heat of dilution, with the best-fit curve for a 2:1 model. Similar titrations with the CTGGsm hairpin are shown (bottom). The relative small heat of dilution of f-ImImIm can be seen in the last two to three injections when all of DNA sites are saturated. It is interesting that distinct differences in the heat profiles are observed but that the averaged total binding heats are comparable (Table 1).
Figure 5. The dependence of the thermodynamic parameters on temperature for binding of f-ImImIm to CTGGsm (top) and f-PyImIm to CCGG (bottom). The entropy TS° (diamonds) was calculated from the experimental G° (squares) obtained from SPR, and H° (filled circles) obtained from ITC. The slopes of H° as a function of temperature provide the Cp° values for complex formation (Table 1).
The interactions of polyamides vary significantly with DNA sequence as shown in previous studies of distamycin binding to AT-rich sequences (2,6,17,21). Unlike the imidazole containing polyamides, formation of the distamycin 2:1 complex with an A3T3 sequence is governed by negative cooperativity with strong affinity in the first interaction (K1 = 3.1 x 108 M–1, G1° = –11.6 kcal/mol) followed by weaker binding of a second molecule (K2 = 8.0 x 105 M–1, G2° = –8.1 kcal/mol) (6). ITC titrations of distamycin were conducted as a model system for comparison of positive and negative cooperativity in dimer interactions (Supplementary Material, Fig. S1). Binding of the two molecules of distamycin to DNA is clearly distinguished in the plot, which has a quite different shape than with the Im polyamide titrations. Binding of the first distamycin has a lower enthalpy (H1° = –12.2 kcal/mol) than that of the second (H2° = –23.9 kcal/mol). The entropic contributions of –0.6 and –15.9 kcal/mol for distamycin binding to the two sites indicate that the entropic contribution is near zero for binding of the first molecule but is very unfavorable for the second. These results indicate that, as with the imidazole polyamides, distamycin binding is enthalpy driven.
Calculated heat capacity changes. Changes in SASA on complex formation have been correlated with Cp° values for protein–protein, protein–DNA (32,33) and more recently for drug–DNA interactions (34,35). The Cp° values for the f-ImImIm–CTGGdm complex were calculated based on buried surface areas determined by using the reported NMR structure (5) with previously described methods (31,36). Although the structure for the complex with the CTGGsm sequence is not available, the values obtained with the CTGGdm sequence could be extrapolated to the complex formed with the single mismatch. The calculated heat capacity changes using two different sets of parameters are negative and in general agreement with the experimental values (Supplementary Material, Table S2).
DISCUSSION
Previous biosensor binding studies have shown that f-ImImIm and other polyamides bind to short oligonucleotides (matched or mismatched sequences) with significantly different affinities (6,21). Fox and coworkers have also recently shown that some, but not all, chemical and enzymatic cleavage agents can recognize T·G (and other) mismatched base pairs in a long DNA sequence (37). The first question that we wished to answer in this study was whether f-ImImIm could selectively recognize a single T·G mismatch site in competition with excess Watson–Crick sites. Footprints from DNase I cleavage of an 80 base DNA sequence with a single T·G mismatch (Fig. 2) show strong protection from cleavage by f-ImImIm at the T·G site. These results are in agreement with the oligonucleotide studies by SPR that indicate stronger binding of f-ImImIm to the mismatched CTGGsm over the matched CCGG site.
A full thermodynamic data set is essential to complement the structural information available for T·G recognition by f-ImImIm and to provide a complete understanding the energetic basis for the T·G recognition. The values of H° and Cp° for binding reactions are most reliably obtained by calorimetry rather than indirectly from van’t Hoff determinations (38–41). Calorimetry, on the other hand, may provide a less accurate K for strong binding reactions due to the relatively high concentrations that may have to be used if reaction heats are low. We have found biosensor-SPR methods to be particularly valuable in studies of DNA oligomer complexes with small molecules such as the polyamides (6,21) since a wide range of concentrations can be used and both kinetic and equilibrium results are obtained. The SPR studies are particularly useful when the K values are high, as with the f-ImImIm/T·G complex.
Biosensor-SPR methods clearly show that the polyamides form cooperative 2:1 complexes with their DNA recognition sites. The G° values for binding are essentially constant with temperature suggesting strong enthalpy–entropy compensation in this temperature range (Fig. 5). Such compensation is observed for many biomolecular processes and suggests a significant hydrophobic component to the binding energetics (31,42–44). The SPR results quantitatively show that f-ImImIm binds much more strongly to the mismatched CTGGsm sequence than to the equivalent DNA with Watson–Crick base pairs. The G° value for binding to T·G sites is 2 kcal/mol more favorable than for binding to equivalent CG sites. The improved binding of f-ImImIm to the T·G mismatch is correlated with much slower dissociation kinetics relative to the matched CCGG complex (6,21). The slow dissociation kinetics for the mismatch complex could arise from at least two factors: the breakage of strong H-bonds (from the amino protons of G to f-ImImIm) and the rearrangement/re-uptake of water molecules in the DNA minor groove as described below.
Calorimetric studies are also consistent with formation of cooperative 2:1 complexes of f-ImImIm with both matched CG and mismatched T·G DNA sequences and with stronger affinity for the mismatch. The difference in the heat profiles for f-ImImIm binding to CTGGsm and CCGG (Fig. 4) arises both from differences in individual enthalpies (H1° and H2°) and binding constants (K1 and K2) in the 2:1 complexes. Somewhat surprisingly given all of these differences, the calorimetric titrations suggest that the averaged total heats for dimer formation are comparable for formation of f-ImImIm–CTGGsm, f-ImImIm–CCGG, and f-PyImIm–CCGG complexes at 298 K. This is particularly surprising since comparison of the H° and G° values for binding (Table 1) indicates that the enthalpic contributions play the primary role in driving the interactions. Given that the calorimetric H° values for binding are more negative than the total free energies of binding, there must be unfavorable compensating entropic contributions in each binding process. Since the H° values for these interactions are similar at 298 K, it is the difference in the unfavorable entropic contribution that dictates the different binding affinities. The higher binding affinity for the f-ImImIm–CTGGsm complex means the S° for formation of this complex must be less unfavorable than that of the complex with CCGG.
It is important at this point to combine results from the structural and thermodynamic studies to suggest possible explanations the similar H° but quite different G° and S° for binding of f-ImImIm to match and mismatch sites. NMR and CD results suggest that significant conformational changes do not occur in the DNA binding sites on formation of these complexes (5,6). These observations imply that the entropic effects do not come from overall structural changes but from factors such as the loss of translational and rotational entropy of the polyamides and the opposing gain of entropy due to release of bound water molecules on complex formation (although the net change in bound water is unknown). Since the polyamides would be expected to be similarly hydrated and to release similar amounts of hydration water on complex formation, any differences in water released, among the complexes, would have to be due to differences in DNA hydration. Structural studies indicate that the T·G base pair adopts a wobble conformation with an unpaired free amino group of G that is strongly hydrated as part of the minor groove hydration matrix (12,45). On complex formation with f-ImImIm some of these tightly bound and highly ordered water molecules would be released from the T·G site to provide a favorable entropic contribution to complex formation. From this model the entropic contributions are predicted to be less unfavorable for formation of the f-ImImIm–CTGGsm complex than for the CCGG complex and this results in the enhanced affinity for the T·G complex. The difference in G° between the f-ImImIm–CTGGsm and –CCGG complexes (2 kcal/mol and mainly from entropy) is about the same order of magnitude as the entropy gain by releasing a single-bound water molecule from its site (46). Binding of a netropsin analog to a mixed sequence DNA suggested that water is actually taken up upon complex formation (47), and it is difficult to see how this could enhance binding entropy. Initial molecular dynamics studies in our laboratory have suggested that minor groove hydration is quite varied among different AT DNAs, and it is possible that water release or uptake is quite dependent on different DNA sequences and ligands.
The negative Cp° values for complex formation can provide useful insights about interactions and hydration (30–36). There are significant differences in the heat capacity changes for the polyamides and DNA sequences used in these experiments. The Cp° for the f-ImImIm–CTGGsm complex, for example, is more negative than that for f-PyImIm–CCGG (Table 1). These Cp° values indicate important differences in the formation of complexes with the CCGG DNA versus the T·G mismatch CTGGsm sequence. Structured water, such as the water of hydrophobic hydration, can have a large heat capacity and release of such water, for example on transfer of nonpolar groups into the interior of proteins or to the minor groove of DNA, contributes a negative term to the binding Cp°. Since all of the polyamide complexes examined in this study have negative Cp°, the binding enthalpy becomes more favorable (more negative) while the binding entropy is less favorable (also more negative) as the temperature increases. Such effects are larger in systems where bound and ordered water molecules are more strongly associated with the DNA groove as with CTGGsm. Calculation of buried surface areas from the NMR structures of f-ImImIm–CTGGdm and f-ImImIm–CCGG indicate that the observed Cp° are correlated with water release from buried nonpolar and polar molecular surface.
Binding of polyamides to GC containing sequences has quite different characteristics than binding to AT-rich sequences. Binding of many imidazole-polyamides to sequences with GC or T·G mismatched base pairs displays positive cooperativity, while distamycin binds to A3T3 with negative cooperativity (6). This difference may be explained in terms of minor groove width. The groove of AT is unusually narrow (0.3–0.4 nm) (48), but it is wider in mixed DNA sequences (0.5–0.6 nm) (49) as well as at T·G base pairs (12). Since the thickness of one polyamide ring is 0.34 nm (50), the first distamycin molecule can fit snugly and make favorable van der Waals contacts with the walls of the minor groove in A3T3. For binding of the second molecule, however, the minor groove must widen sufficiently to accommodate the two stacked distamycin molecules in a 2:1 complex. On the other hand, the groove width of GC-rich or T·G sequences is wider and can bind two polyamides in a positively cooperative manner even at low ratios of compound to DNA. Interestingly, the experimental K and H° for distamycin binding are clearly resolved (17,51 and this work) in ITC while those for polyamide binding with positive cooperativity are overlapped and difficult to separate. In summary, the results presented here provide a thermodynamic rationale for the specificity in molecular recognition of T·G mismatch base pairs by Im-polyamides.
SUPPLEMENTARY MATERIAL
ACKNOWLEDGEMENTS
This work was supported by NIH Grant GM61587, the Georgia Research Alliance (W.D.W.), and by the NSF-REU and Research Corporation (M.L.).
REFERENCES
Bailly,C. and Chaires,J.B. (1998) Sequence-specific DNA minor groove binders. Design and synthesis of netropsin and distamycin analogues. Bioconjug. Chem., 9, 513–538.
Wemmer,D.E. (2000) Designed sequence-specific minor groove ligands. Annu. Rev. Biophys Biomol. Struct., 29, 439–461.
Wemmer,D.E. (2001) Ligands recognizing the minor groove of DNA: Development and applications. Biopolymers, 52, 197–211.
Dervan,P.B. (2001) Molecular recognition of DNA by small molecules. Bioorg. Med. Chem., 9, 2215–2235.
Yang,X.L., Hubbard,R.B., Lee,M., Tao,Z.F., Sugiyama,H. and Wang,A.H. (1999) Imidazole-imidazole pair as a minor groove recognition motif for T:G mismatched base pairs. Nucleic Acids Res., 27, 4183–4190.
Lacy,E.R., Cox,K.K., Wilson,W.D. and Lee,M. (2002) Recognition of T*G mismatched base pairs in DNA by stacked imidazole-containing polyamides: surface plasmon resonance and circular dichroism studies. Nucleic Acids Res., 30, 1834–1841.
Land,H., Parada,L.F. and Weinberg,R.A. (1983) Cellular oncogenes and multistep carcinogenesis. Science, 222, 771–778.
Almoguera,C., Shibata,D., Forrester,K., Martin,J., Arnheim,N. and Perucho,M. (1988) Most human carcinomas of the exocrine pancreas contain mutant c-K-ras genes. Cell, 53, 549–554.
Watanabe,H., Ha,A., Hu,Y.X., Ohtsubo,K., Yamaguchi,Y., Motoo,Y., Okai,T., Toya,D., Tanaka,N. and Sawabu,N. (1999) K-ras mutations in duodenal aspirate without secretin stimulation for screening of pancreatic and biliary tract carcinoma. Cancer, 86, 1441–1448.
Lall,L. and Davidson,R.L. (1998) Sequence-directed base mispairing in human oncogenes. Mol. Cell. Biol., 18, 4659–4669.
Hare,D., Shapiro,L. and Patel,D.J. (1986) Wobble dG X dT pairing in right-handed DNA: solution conformation of the d(C-G-T-G-A-A-T-T-C-G-C-G) duplex deduced from distance geometry analysis of nuclear Overhauser effect spectra. Biochemistry, 25, 7445–7456.
Hunter,W.N., Brown,T., Kneale,G., Anand,N.N., Rabinovich,D. and Kennard,O. (1987) The structure of guanosine-thymidine mismatches in B-DNA at 2.5-? resolution. J. Biol. Chem., 262, 9962–9970.
Lacy,E.R., Madsen,E.M., Lee,M. and Wilson,W.D. (2003) Polyamide dimer stacking in the DNA minor groove and recognition of T.G mismatched base pairs in DNA. In Demeunynck,M., Bailly,C. and Wilson,W.D. (eds), DNA and RNA Binders: From Small Molecules to Drugs. WILEY-VCH, Weinheim, Vol. 2, pp. 384–413.
Rucker,V.C., Foister,S., Melander,C. and Dervan,P.B. (2003) Sequence specific fluorescence detection of double strand DNA. J. Am. Chem. Soc., 125, 1195–1202.
Pilch,D.S., Poklar,N., Gelfand,C.A., Law,S.M., Breslauer,K.J., Baird,E.E. and Dervan,P.B. (1996) Binding of a hairpin polyamide in the minor groove of DNA: sequence-specific enthalpic discrimination. Proc. Natl Acad. Sci. USA, 93, 8306–8311.
Pilch,D.S., Poklar,N., Baird,E.E., Dervan,P.B. and Breslauer,K.J. (1999) The thermodynamics of polyamide-DNA recognition: hairpin polyamide binding in the minor groove of duplex DNA. Biochemistry, 38, 2143–2151.
Rentzeperis,D., Marky,L.A., Dwyer,T.J., Geierstanger,B.H., Pelton,J.G. and Wemmer,D.E. (1995) Interaction of minor groove ligands to an AAATT/AATTT site: correlation of thermodynamic characterization and solution structure. Biochemistry, 34, 2937–2945.
Cooper,A., Johnson,C.M., Lakey,J.H. and Nollmann,M. (2001) Heat does not come in different colours: entropy-enthalpy compensation, free energy windows, quantum confinement, pressure perturbation calorimetry, solvation and the multiple causes of heat capacity effects in biomolecular interactions. Biophys. Chem., 93, 215–230.
Morton,C.J. and Ladbury,J.E. (1996) Water-mediated protein-DNA interactions: the relationship of thermodynamics to structural detail. Protein Sci., 5, 2115–2118.
Privalov,G.P. and Privalov,P.L. (2000) Problems and prospects in microcalorimetry of biological macromolecules. Methods Enzymol., 323, 31–62.
Lacy,E.R., Le,N.M., Price,C.A., Lee,M. and Wilson,W.D. (2002) Influence of a terminal formamido group on the sequence recognition of DNA by polyamides. J. Am. Chem. Soc., 124, 2153–2163.
Xodo,L.E., Manzini,G., Quadrifoglio,F., van der Marel,G.A. and van Boom,J.H. (1988) Oligodeoxynucleotide folding in solution: loop size and stability of B- hairpins. Biochemistry, 27, 6321–6326.
Cantor,C.R. and Tinoco,I.,Jr (1965) Absorption and optical rotatory dispersion of seven trinucleoside diphosphates. J. Mol. Biol., 13, 65–77.
Cantor,C.R., Warshaw,M.M. and Shapiro,H. (1970) Oligonucleotide interactions. 3. Circular dichroism studies of the conformation of deoxyoligonucleotides. Biopolymers, 9, 1059–1077.
Fasman,G.D. (1975) Handbook of Biochemistry and Molecular Biology, Nucleic Acids, 3rd Edn. CRC Press, Cleveland, OH.
Wiseman,T., Williston,S., Brandts,J.F. and Lin,L.N. (1989) Rapid measurement of binding constants and heats of binding using a new titration calorimeter. Anal. Biochem., 179, 131–137.
Horn,J.R., Russell,D., Lewis,E.A. and Murphy,K.P. (2001) Van’t Hoff and calorimetric enthalpies from isothermal titration calorimetry: are there significant discrepancies? Biochemistry, 40, 1774–1778.
Briggner,L.E. and Wadso,I. (1991) Test and calibration processes for microcalorimeters, with special reference to heat conduction instruments used with aqueous systems. J. Biochem. Biophys. Methods, 22, 101–118.
Nicholls,A., Sharp,K.A. and Honig,B. (1991) Protein folding and association: insights from the interfacial and thermodynamic properties of hydrocarbons. Proteins, 11, 281–296.
Spolar,R.S., Livingstone,J.R. and Record,M.T.,Jr (1992) Use of liquid hydrocarbon and amide transfer data to estimate contributions to thermodynamic functions of protein folding from the removal of nonpolar and polar surface from water. Biochemistry, 31, 3947–3955.
Ren,J., Jenkins,T.C. and Chaires,J.B. (2000) Energetics of DNA intercalation reactions. Biochemistry, 39, 8439–8447.
Makhatadze,G.I. and Privalov,P.L. (1990) Heat capacity of proteins. I. Partial molar heat capacity of individual amino acid residues in aqueous solution: hydration effect. J. Mol. Biol., 213, 375–384.
Privalov,P.L. and Makhatadze,G.I. (1990) Heat capacity of proteins. II. Partial molar heat capacity of the unfolded polypeptide chain of proteins: protein unfolding effects. J. Mol. Biol., 213, 385–391.
Mazur,S., Tanious,F.A., Ding,D., Kumar,A., Boykin,D.W., Simpson,I.J., Neidle,S. and Wilson,W.D. (2000) A thermodynamic and structural analysis of DNA minor-groove complex formation. J. Mol. Biol., 300, 321–337.
Haq,I., Ladbury,J.E., Chowdhry,B.Z., Jenkins,T.C. and Chaires,J.B. (1997) Specific binding of hoechst 33258 to the d(CGCAAATTTGCG)2 duplex: calorimetric and spectroscopic studies. J. Mol. Biol., 271, 244–257.
Spolar,R.S. and Record,M.T.,Jr (1994) Coupling of local folding to site-specific binding of proteins to DNA. Science, 263, 777–784.
Brown,J., Brown,T. and Fox,K.R. (2003) Cleavage of fragments containing DNA mismatches by enzymic and chemical probes. Biochem. J., 371, 697–708.
Janin,J. (1995) Elusive affinities. Proteins, 21, 30–39.
Ladbury,J.E. (1995) Counting the calories to stay in the groove. Structure, 3, 635–639.
Naghibi,H., Tamura,A. and Sturtevant,J.M. (1995) Significant discrepancies between van’t Hoff and calorimetric enthalpies. Proc. Natl Acad. Sci. USA, 92, 5597–5599.
Chaires,J.B. (1997) Possible origin of differences between van’t Hoff and calorimetric enthalpy estimates. Biophys. Chem., 64, 15–23.
Chaires,J.B. (1997) Energetics of Drug-DNA Interactions. Biopolymers, 44, 201–215.
Gallagher,K. and Sharp,K. (1998) Electrostatic contributions to heat capacity changes of DNA-ligand binding. Biophys. J., 75, 769–776.
Matulis,D., Rouzina,I. and Bloomfield,V.A. (2002) Thermodynamics of cationic lipid binding to DNA and DNA condensation: roles of electrostatics and hydrophobicity. J. Am. Chem. Soc., 124, 7331–7342.
Yang,X.L., Kaenzig,C., Lee,M. and Wang,A.H. (1999) Binding of AR-1–144, a tri-imidazole DNA minor groove binder, to CCGG sequence analyzed by NMR spectroscopy. Eur. J. Biochem., 263, 646–655.
Dunitz,J.D. (1994) The entropic cost of bound water in crystals and biomolecules. Science, 264, 670.
Sidorova,N. and Rau,D.C. (1995) The osmotic sensitivity of netropsin analogue binding to DNA. Biopolymers, 35, 377–384.
Fratini,A.V., Kopka,M.L., Drew,H.R. and Dickerson,R.E. (1982) Reversible bending and helix geometry in a B-DNA dodecamer: CGCGAATTBrCGCG. J. Biol. Chem., 257, 14686–14707.
Yoon,C., Prive,G.G., Goodsell,D.S. and Dickerson,R.E. (1988) Structure of an alternating-B DNA helix and its relationship to A-tract DNA. Proc. Natl Acad. Sci. USA, 85, 6332–6336.
Pelton,J.G. and Wemmer,D.E. (1990) Binding modes of distamycin A with d(CGCAAATTTGCG)2 determined by two-dimensional NMR. J. Am. Chem. Soc., 112, 1393–1399.
Wilson,W.D., Tanious,F.A., Barton,H.J., Wydra,R.L., Jones,R.L., Boykin,D.W. and Strekowski,L. (1990) The interaction of unfused polyaromatic heterocycles with DNA: intercalation, groove-binding and bleomycin amplification. Anticancer Drug Des., 5, 31–42.(Eilyn R. Lacy1, Binh Nguyen1, Minh Le1,2)