Thermodynamics of DNA binding of MM17, a ‘single chain dimer’ of transcription factor MASH-1

Abstract

MASH-1, a member of the basic helix–loop–helix (bHLH) family of transcriptional regulators, is a central factor for the regulation of the differentiation of committed neuronal precursor cells of the peripheral nervous system. We have previously produced MM17, a single chain version of this dimeric protein, by linking the C-terminal end of the first subunit to the N-terminal residue of the second subunit through a flexible peptide linker. We have now determined by isothermal titration calorimetry the thermodynamic parameters characterising the DNA binding reactions of MM17. The DNA binding specificity was relatively low and comparable to that observed for wild-type MASH bHLH. At 32°C and pH 7, the concentration of MM17 at which 50% DNA binding occurred was determined as 22.8 and 152 nM for binding to MCK-S and the heterologous SP-1, respectively. Similarly to MASH bHLH the free energy of the association was only slightly temperature dependent, while both the entropy and the enthalpy change were strong functions of temperature. The free energy of DNA binding was independent of the pH for the pH range between 6 and 8. Dissection of the entropy change of the association reaction suggested that the two basic domains and the linker region between the subunits underwent a folding transition from a mainly unfolded to a predominantly ordered conformation. Therefore, like wild-type MASH bHLH, the DNA binding reaction of MM17 follows an induced fit mechanism.

INTRODUCTION

The precise recognition of a defined DNA sequence by a DNA-binding protein requires an optimal shape complementarity between the interacting species. Even though in some specific complexes the tightly packed interfaces between the proteins and the DNA result from the docking of well-ordered, pre-existing surfaces, in an increasing number of cases the conformation of the proteins or the DNA is found to change markedly in the complex (1,2).

An example of a protein which undergoes a conformational change on DNA binding is provided by the basic helix–loop–helix (bHLH) protein MASH-1, a transcription factor involved in the regulation of cellular differentiation (3). The expression of MASH-1 induces the differentiation of committed neuronal precursor cells of the peripheral nervous system.

Two X-ray structures of DNA complexes of bHLH proteins have revealed the structural basis for the DNA recognition of this class of transcription factors (4–6). The helix–loop–helix (HLH) domain mediates dimerisation of MASH-1, which binds as a dimer with high affinity and relatively low specificity to DNA sequences containing an E-box (CANNTG) (Fig. 1B) (7,8). The interactions with DNA are mediated through residues of the basic region, which in structural terms forms the N-terminal extension of helix 1. CD spectroscopic experiments with the bHLH proteins MASH-1 and MyoD have shown that in the absence of DNA, bHLH proteins can form stable dimers which are in a concentration-dependent equilibrium with the monomer (8,9). Dimerisation is accompanied by a folding transition from a largely unfolded monomer to a predominantly α-helical dimer. The stability of the dimer is due to a hydrophobic core formed mainly between residues of helices 1 and 2. Isothermal titration calorimetry and CD spectroscopy have indicated that the basic region remains unfolded in the dimer, but adopts an α-helical conformation on DNA binding (10). Folding and dimerisation of the bHLH domains can be induced through the addition of DNA even at concentrations where the bHLH domain alone is mainly unfolded.

Figure 1.

(A) Diagram of the MM17 protein indicating the connectivities between individual bHLH subunits of MASH-1. The amino acid sequences of the bHLH domain and of the linker are indicated (11). (B) Representation of the DNA complex of MM17 based on the X-ray structure of the DNA complex of E47. The first and the second bHLH subunits of MM17 are in blue and green, respectively. The linker connecting the C-terminus of the first bHLH subunit to the N-terminus of the second is coloured yellow. The figure was constructed from the coordinates for the crystal structure of the DNA complex of E47 (4) using the Sybyl (Tripos) package.

We have reported the construction of MM17, a single chain dimer of the MASH bHLH, in which the C-terminus of the bHLH subunit is attached to the N-terminus of the second subunit through a peptide linker of 17 amino acids (11). Characterisation by electrophoretic mobility shift assay and CD spectroscopy revealed that MM17 and wild-type MASH bHLH displayed similar DNA recognition properties. However, in MM17 the hydrophobic core between α-helices 1 and 2 of the two HLH domains was formed even in the absence of DNA. ‘Dimerisation’ was independent of the concentration because of the monomeric nature of MM17. To complement the structural and biochemical studies we have now measured the thermodynamic parameters of the DNA binding reaction of MM17 by isothermal titration calorimetry (ITC). The thermodynamics (and kinetics) of DNA recognition by proteins provide the fundamental criteria that determine which interactions take place and as such are of critical importance in understanding the control of gene expression.

The DNA binding properties of MM17 were found to be similar to those measured for wild-type MASH bHLH. Dissection of the entropy terms of the DNA binding reaction of MM17 showed that, similar to wild-type MASH bHLH, the basic domains of the single chain dimer undergo a folding transition from an unstructured conformation to a mainly α-helical form in the DNA complex. In addition, the thermodynamic analysis revealed that the linker region, which connects the C-terminal end of helix 1 of one subunit to the basic region of the second subunit and which is relatively flexible in the absence of DNA, adopts a well-defined conformation in the DNA complex.

MATERIALS AND METHODS

Expression and purification of MM17

MM17 was purified essentially as described (11). The purified protein was homogeneous as judged by SDS–PAGE and cation exchange chromatography on a Resource-S (Pharmacia) HPLC column. MALDI-TOF mass spectrometry showed a molecular mass of 16 581 for MM17, which corresponded well with the calculated mass of 16 580 for the single chain dimer without the N-terminal methionine. Protein concentrations were determined by measuring the UV absorption at 215 and 220 nm (12).

Oligonucleotides

Oligonucleotides were purchased from Integrated DNA Technologies. Complementary strands were annealed by heat denaturation followed by slow cooling to room temperature.

The following oligonucleotides were used:

MCK-S: 5′-CAGGCAGCAGGTGTTGG-3′

3′-GTCCGTCGTCCACAACC-5′

SP-1: 5′-GATGCGGTCCCGCCCTCAGC-3′

3′-CTACGCCAGGGCGGGAGTCG-5′

Buffers

The following buffers were used. Low salt MES buffer: 10 mM MES–KOH, pH 6.0, 5 mM KCl. Low salt PIPES buffer: 10 mM PIPES–KOH, pH 7.0, 5 mM KCl. Low salt Tris buffer: 10 mM Tris–HCl, pH 8.0, 5 mM KCl. High salt PIPES buffer: 10 mM PIPES–KOH, pH 7.0, 25 mM KCl.

Isothermal titration calorimetry

ITC measurements were performed on a MCS isothermal titration calorimeter (MicroCal, Northampton, MA) at the School of Biological Sciences of the University of Birmingham. The calorimeter has been described in detail by Wiseman et al. (13). The calorimeter was calibrated with electrically generated heat pulses as recommended by the manufacturer. The temperature of the system was kept at least 5°C below the temperature of the experiment with the help of a circulating water bath in order to improve baseline stability. Temperature equilibration was allowed to proceed for ~12 h. Samples of MM17 and DNA were dialysed three times against 400 ml and once against 800 ml of reaction buffer to minimise artefacts as a consequence of minor differences in buffer composition. All solutions were degassed for 10 min by evacuation. The reaction cell contained 1.33 ml of 1 µM double-stranded oligonucleotide. The injection syringe contained 50 µM MM17 and was rotated at 350 r.p.m. during equilibration and experiment. Injections were started automatically by the computer software after reaching the desired temperature and baseline stability. Titrations consisted of one injection of 1 µl followed by 20 injections of 4 µl and 7.9 s duration. The length of the intervals between injections was 240 s. Data analysis was performed with the software provided with the instrument (MicroCal Origin v.2.9). The total heat of binding (ΔH) and association constant (K_A) for the binding reaction were obtained by non-linear least squares fitting of the data to a 1:1 binding model using the Marquardt algorithm.

RESULTS AND DISCUSSION

ITC measurements of the DNA binding reaction of MM17

The thermodynamic parameters of the association reaction between MM17 and the MCK-S and SP-1 oligonucleotides were determined by ITC measurements of the evolved heat as a function of temperature in the range from 12 to 32°C (Tables 1 and 2 and Fig. 2). For all titrations, the best fit of the experimental data to the binding isotherm was obtained for one MM17 molecule binding to one double-stranded DNA oligonucleotide, as would be expected because the two bHLH subunits are covalently linked in MM17. Control titrations of MM17 into buffer which did not contain DNA gave rise to small peaks of equal and negligible size.

Table 1. Dependence of the thermodynamic parameters of the binding of MM17 to E-box-containing DNA on temperature and pH measured by ITC^a.

T (K)	KA (M–1)	ΔG (kcal/mol)b	ΔH (kcal/mol)	TΔS (kcal/mol)c
MCK-S in low salt PIPES buffer pH 7.0d
285	1.11 (± 0.1) E + 07	–9.18 ± 0.05	–35.71 ± 1.87	–26.53 ± 1.93
290	1.77 (± 0.65) E + 07	–9.58 ± 0.24	–42.78 ± 3.94	–33.19 ± 4.17
295	6.04 (± 8.65) E + 07	–10.41 ± 0.73	–36.89 ± 2.95	–26.48 ± 3.31
300	2.40 (± 1.16) E + 07	–10.07 ± 0.35	–44.06 ± 0.83	–33.99 ± 1.04
305	4.38 (± 1.87) E + 07	–10.61 ± 0.29	–49.42 ± 1.91	–38.81 ± 1.87
MCK-S in high salt PIPES buffer pH 7.0e
285	4.07 (± 1.66) E + 06	–8.58 ± 0.26	–32.56 ± 3.64	–23.98 ± 3.65

^aStandard deviations for multiple measurements are given.

^bReaction free energy for the association: ΔG = –RT lnK_A.

^cCalculated from TΔS = ΔH – ΔG.

^dLow salt PIPES buffer: 10 mM PIPES–KOH, pH 7.0, 5 mM KCl.

^eHigh salt buffer: 10 mM PIPES–KOH, pH 7.0, 25 mM KCl.

Table 2. Dependence of the thermodynamic parameters of binding of MM17 to the heterologous SP-1 oligonucleotide on temperature and pH measured by ITC^a.

T (K)	KA (M–1)	ΔG (kcal/mol)b	ΔH (kcal/mol)	TΔS (kcal/mol)c
SP-1 in low salt PIPES buffer pH 7.0d
285	5.04 (± 2.75) E + 06	–8.68 ± 0.32	–28.63 ± 1.13	–19.95 ± 0.88
290	8.98 (± 3.17) E + 06	–9.19 ± 0.22	–32.53 ± 2.72	–23.33 ± 2.90
295	1.70 (± 0.86) E + 07	–9.69 ± 0.33	–31.39 ± 3.87	–21.69 ± 4.20
300	1.33 (± 0.73) E + 07	–9.84 ± 0.32	–37.75 ± 5.88	–27.91 ± 6.20
305	6.58 (± 2.97) E + 06	–9.46 ± 0.34	–39.79 ± 6.38	–30.33 ± 6.05
SP-1 in high salt PIPES buffer pH 7.0e
285	3.20 (± 1.31) E + 06	–8.45 ± 0.22	–32.24 ± 6.42	–23.79 ± 6.64

^aStandard deviations for multiple measurements are given.

^bReaction free energy for the association: ΔG = –RT lnK_A.

^cCalculated from TΔS = ΔH – ΔG.

^dLow salt PIPES buffer: 10 mM PIPES–KOH, pH 7.0, 5 mM KCl.

^eHigh salt buffer: 10 mM PIPES–KOH, pH 7.0, 25 mM KCl.

Figure 2.

Example of a typical ITC experiment. MCK-S DNA was titrated with increasing amounts of MM17 in low salt PIPES buffer at pH 7 and 27°C (closed squares). (Top) Raw data for a titration. (Bottom) Titration curve. Abscissa, moles of MM17 added per mole DNA; [MCK-S], 1 µM. Control titration of MM17 into low salt PIPES buffer at pH 7 (closed triangles).

In low salt buffer at pH 7 the association constants for MM17 binding to MCK-S and SP-1 were (1.11 ± 0.11) × 10⁷ M^–1 and (5.04 ± 2.75) × 10⁶ M^–1 at 12°C and (4.38 ± 1.87) × 10⁷ M^–1 and (6.58 ± 2.97) × 10⁶ M^–1 at 32°C, respectively. Similar results were obtained with oligonucleotides containing other heterologous DNA sequences (data not shown). The stability of SP-1 and MCK-S did not interfere with the accuracy of the ITC experiments because the hyperchromicity had been shown previously to be <1% for temperatures below 35°C (10). The free energy required to transfer MM17 from an E-box to a heterologous DNA sequence was 0.5 kcal/mol at 12°C (Tables 1 and 2), which is in good agreement with the value determined previously in EMSA experiments (11). The specificity of DNA binding of MM17 increased with increasing temperature (Fig. 3) and reached a ΔΔG of 1.15 kcal/mol at 32°C (Tables 1 and 2). These specificities are similar to those measured by ITC for wild-type MASH bHLH (10).

Figure 3.

Thermodynamic profiles of the DNA binding reactions of MM17 in low salt PIPES buffer at pH 7 as determined by ITC. Plots of ΔG, ΔH and TΔS as a function of temperature for titrations of MM17 with MCK-S (red) and SP-1 (blue). The heat capacity change (Δc_p) was calculated by linear regression from the values of ΔH = ΔH(T).

Similar results were obtained in ITC experiments in buffers containing higher salt concentrations (Tables 1 and 2). However, due to the low solubility of MM17 in the presence of even modest concentrations of KCl no measurements above salt concentrations of 25 mM were possible.

In order to measure the effect of pH on the stability of the DNA complexes of MM17, ITC measurements were performed for the pH range between 6 and 8. These measurements showed that within experimental error the complex stability and hence the DNA binding specificity were independent of pH, at least for the narrow pH range examined (Fig. 4). Therefore, no important protonation or deprotonation step occurs during the association reactions of MM17 and DNA. This finding was not surprising, since the protonation state did not change in the binding reaction of wild-type MASH bHLH (10).

Figure 4.

Effects of pH on the free energy of association between MM17 and MCK-S (red) and SP-1 (blue).

Thermodynamic parameters of the binding reaction

The binding reactions between MM17 and MCK-S and SP-1 were strongly exothermic (Tables 1 and 2). The reaction was generally more exothermic for binding to E-box-containing DNA. The reaction enthalpy (ΔH) and the reaction entropy (TΔS), both of which are strong functions of the reaction temperature, compensate to make the reaction free energy almost independent of the temperature. This behaviour mirrors that observed for wild-type MASH bHLH binding to DNA and is a general feature of biomolecular association reactions in aqueous solvents (10,14). T_S, the temperature at which the entropy change of the association reaction of MM17 and MCK-S changes sign, was 232 K. The enthalpy change for MCK-S binding of MM17 was 0 for the temperature T_H = 222 K. As a consequence, the association reaction is strongly enthalpy driven over the whole physiological temperature range.

Due to the strong temperature dependence of the reaction enthalpy a large negative change of the heat capacity was measured. Linear regression of the data in Figure 3 gave a Δc_p of –574 ± 239 cal mol^–1 K^–1 for the E-box-containing oligonucleotide and –550 ± 113 cal mol^–1 K^–1 for the heterologous SP-1 oligonucleotide. These values are comparable to those measured for the DNA binding reactions of MASH bHLH and contain not just the contributions from the docking of pre-ordered structures but also from any conformational changes that occur on DNA binding of MM17 (vide infra) (10).

Interpretation of the observed heat capacity change

At T_S, the temperature at which the overall reaction enthalpy equals 0, the following equations holds:

ΔS(T_S) = 0 = ΔS_HE + ΔS_RT + ΔS_PE + ΔS_other

The large negative heat capacity changes typically observed for the formation of specific protein–DNA complexes are caused to a large extent by changes in the exposure of non-polar surface area (10,15–22). A more detailed analysis revealed, however, that changes in the exposure of non-polar surface area also contribute to Δc_p (19,23). An empirical relation between the heat capacity change and the entropy change of the binding reaction due to the hydrophobic effect (ΔS_HE) could be derived (15,19,24). Calculation of ΔS_HE from the measured values of Δc_p gave ΔS_HE of 394 ± 164 cal mol^–1 K^–1 and 335 ± 69 cal mol^–1 K^–1 for MCK-S and SP-1 binding, respectively.

Experimental data and a statistical mechanical analysis of rigid body associations revealed that the unfavourable term ΔS_RT, which results from the reduction in rotational and translational degrees of freedom of MM17 and the DNA on binding, is relatively insensitive to the reacting species and to T_S. Because one protein molecule reacts with one double-stranded oligonucleotide to form the complex a value of –50 cal mol^–1 K^–1 was used for ΔS_RT (25).

When a protein binds to DNA, the Coulombic interactions with positively charged amino acid side chains lead to the release of bound counterions. The resulting free energy change, which becomes more favourable as the bulk salt concentration is reduced, is primarily entropic (26). Due to the low solubility of MM17 in buffers containing even modest salt concentrations no reliable value for the contribution of this temperature-independent polyelectrolyte effect (ΔS_PE) to the observed entropy change could be obtained directly (26–28). However, CD spectroscopy has shown that the DNA complexes of MM17 and MASH bHLH are very similar, therefore a similar ΔS_PE can be assumed. ΔS_PE had been measured as 50 cal mol^–1 K^–1 for the DNA binding reaction of the MASH bHLH in EMSA experiments (10,26,27,29).

CD spectroscopy had indicated that the conformation of the DNA did not change significantly during the DNA binding reactions of MM17 and MASH bHLH (8,10,11,29). All available X-ray structures of DNA complexes of bHLH proteins indicate that the DNA is essentially in its B-form in the complex (4,5). Therefore, no contribution to the overall entropy change of the association reaction is expected from a conformational change of the DNA. Therefore, the remaining entropy change ΔS_other = –(ΔS_HE + ΔS_PE + ΔS_RT) = –394 cal mol^–1 K^–1 can be attributed to local protein folding transitions coupled to DNA binding. Under the experimentally (19 and references therein) and computationally (30–33) well justified assumption that for every residue which undergoes a folding transition on average 5.6 e.u. are lost, 70 residues which are unordered in free MM17 adopt a folded form on DNA binding. CD spectroscopy had revealed that helix 1 and helix 2 of both HLH domains in MM17 adopted an α-helical structure even in the absence of DNA (11). However, the basic regions of MM17 underwent a conformational change on DNA binding from a largely unstructured to a mainly α-helical structure. Similarly, analysis of the thermodynamic data obtained for the DNA binding reaction of the MASH bHLH indicated that 27 amino acids per subunit underwent a conformational change on DNA binding, indicating that the whole basic region and the N-terminal two amino acids of helix 1 change conformation. In agreement with these data, nuclear magnetic resonance studies of the bHLH protein E47 showed that while the HLH domain was stably folded, the basic region was in an unstructured conformation even at the high concentrations used for the NMR experiments (34). However, X-ray crystallography of E47 revealed that the basic region was fully α-helicial in the DNA complex (4).

The ITC data obtained for the association reaction of MM17 with DNA indicate that in addition to the 54 amino acids of the two basic regions, another 16 amino acids lose conformational degrees of freedom in the complex. According to the crystal structures of the DNA complexes of MyoD (5) and E47 (4) the distance between the N-terminal end of one bHLH subunit and the C-terminal end of the other is ~50 Å. This distance can only just be spanned by a linker of length 18 (Fig. 1B). Therefore, several of the residues at the N-terminal end of the basic region will not be in an α-helical conformation but will be part of the linker region. Analysis of the thermodynamic data suggests that due to its short length the linker region shows little flexibility and adopts a well-defined structure in the DNA complex. This hypothesis is currently being tested in NMR experiments with the free protein and the DNA complex.

In spite of a linker, which in the DNA complex shows little conformational flexibility, the single chain dimer MM17 and wild-type MASH bHLH display similar DNA binding properties. MM17 is therefore an ideal vehicle to be displayed on the surface of filamentous phage particles. Random mutagenesis of MM17 combined with in vitro selection should lead to the identification of MASH bHLH mutants with novel DNA binding specificities.