Introduction

High-resolution 3D structures of proteins have been determined using X-ray crystallography, NMR, and electron microscopy. These techniques provide useful information on protein folding and recognition mechanisms. However, because most of the available structures are static or averaged, information on structural dynamics is needed to understand protein motions in solution. Proteins actually fluctuate in solution and express their functions as the result of conformational changes upon ligand binding. The degree of protein fluctuation varies extensively depending on their function. For example, inhibitor proteins are relatively rigid because they need to oppose hydrolyzation by proteases, whereas some DNA-binding proteins are intrinsically disordered probably because of the advantage to find the specific DNA sequence [1, 2]. Limited knowledge of protein structural dynamics caused difficulties in understanding protein structure–activity relationship and inhibited applications such as protein engineering and structure-based drug design. Thermodynamic analysis of proteins in solution could compensate the lack of information arising from static pictures of proteins. Thermodynamic studies unveiling the differences between a protein and its site-specific mutant could show the effect of the mutated site on the protein dynamics and intramolecular interactions such as hydrogen bonds and electrostatic interactions.

The transcriptional factor c-Myb consists of three functional domains involved in DNA-binding, transcriptional activation, and negative regulation [3]. The DNA-binding domain consists of three repeats, namely R1, R2, and R3, with R2R3 being the minimum unit of specific-DNA binding. We have analyzed the structural and physical properties of c-Myb R2R3 and have shown that its structural dynamics correlate well with its function [4–7]. At physiological temperature, at which c-Myb R2R3 binds to a specific DNA sequence in vivo, our NMR analysis showed that fluctuations were exceptionally large [8]. In the absence of DNA (i.e., DNA-free state), R2R3 exists as “semi-intrinsically disordered state.” The 3D structure of R2R3 in the absence or presence of DNA has been revealed by both NMR and crystallography [9–11]. In the crystal structure of R2R3 in the DNA-free state (PDB code 1GV2), which would be one of the stable structures, Asp-178 is spatially next to His-135, both of which form a surface-exposed salt-bridge with each other (Fig. 1). Upon binding to DNA, the relative positions of these two residues change, resulting in salt-bridge disruption. Based on the available structural information, the putative salt-bridge contributes to the “fragile” 3D structure of R2R3. In this study, we prepared a mutant of R2R3 (i.e., D178N), which lacks the salt-bridge, and analyzed its folding thermodynamics compared to those of wild-type R2R3. The results showed that the salt-bridge actually forms in solution and largely contributes to the 3D structure of R2R3 in the DNA-free state, but it barely contributes to that in the DNA-bound state.

Fig. 1
figure 1

3D structure of c-Myb R2R3 determined by X-ray crystallography (PDB code 1GV2) [11]. The side-chains of His-137 and Asp-178 are shown in sticks

Full size image

Experimental

Sample preparations

Prior to the current mutational analyses, the Cys130 in R2, which is the only cysteine residue in the c-Myb R2R3 and is located at a position equivalent to an isoleucine in R3, was replaced with Ile, to facilitate protein purification and characterization. It has been shown that the affinity and the specificity of mutant C130I are similar to those of the wild-type [12, 13]. Therefore, the C130I protein was used as the standard R2R3 and is denoted as R2R3*. The mutant, C130I/D178N is simply denoted as D178N.

R2R3* and D178N were overexpressed in E. coli BL21 (DE3) and purified. The expression and purification methods were described previously [5, 14, 15] and slightly modified. Briefly, the freshly precultivated cells were inoculated into LB medium containing 100 μg mL−1 ampicillin and were grown at 37 °C. After the culture reached OD600 = 0.8, isopropyl-1-thio-β-d-galactopyranoside was added to a final concentration of 0.5 mM and incubated at 37 °C for 4 h. The harvested cells were lysed by the sonication, and the cell debris was removed by the centrifugation. The supernatant was purified by using ammonium sulfate precipitation, and phosphocellulose (P11, Whatman) and gel filtration (Sephacryl S-100, GE healthcare) columns. The purified proteins were concentrated, and the buffer was exchanged to phosphate-buffered saline (PBS, pH 7.4) or NMR sample buffer using Amicon Ultra-4 (Merck). The purity of each protein was determined to be about 95% by SDS/PAGE analysis. The protein concentrations were determined from UV absorption at 280 nm and were calculated by using the molar absorption coefficient of 3.7 × 104 M−1 cm−1 [13], determined by the amino acid composition analysis.

The 22-mer oligonucleotide CACCCTAACTGACACACATTCT, containing the Myb-binding site in the simian virus 40 enhancer sequence (MBS-I) was chemically synthesized and purified as described previously [13].

CD measurements

Far-UV (195-250 nm) and Near-UV (250-350 nm) CD spectra were recorded on a Jasco J-825 spectropolarimeter at 20 °C equipped with Peltier-type temperature control system. The spectra of R2R3* and D178 N were obtained in PBS (pH 7.4), using quartz cell with 0.1 cm path-length. The protein concentrations were 0.2 and 2.3 mg mL−1 for far- and near-UV, respectively. For the analysis of stability in the presence of DNA, MBS-I, the DNA concentration was equimolar to the protein, 16 μM. CD spectra were obtained using scanning speed of 20 nm min−1, a time response of 1 s, a bandwidth of 1 nm, and an average over 4 or 16 scans, for the far- and near-UV range, respectively.

The melting curves were recorded in temperature mode at 222 nm, from 20 to 80 °C with a heating rate of 1.0 °C min−1. The protein concentrations were 0.02 mg mL−1, and the quartz cell with 1 cm path-length was used. The analysis of the transition curves obtained by temperature-scanning CD measurements was performed on the basis of two-state transition model, as described previously [7, 14, 16].

The molar ellipticity data of temperature-scanning CD (Y) were fitted with the weighted-mean combination of two linear functions for the folded and the unfolded states, that is,

$$y_{\text{n}} = A_{\text{n}} (T{-}T_{\text{m}} ) \, + B_{\text{n}}$$
(1)

and

$$y_{\text{d}} = A_{\text{d}} (T{-}T_{\text{m}} ) \, + B_{\text{d}} ,$$
(2)

from Eqs. (1) and (2), as shown below:

$$Y = y_{\text{n}} (1{-}f_{\text{D}} ) \, + y_{\text{d}} f_{\text{D}}$$
(3)

Here, f D represents the molar fraction of the unfolded state and be expressed with the function of ΔG 01, the free energy of the transition:

$$f_{\text{D}} = \exp \, ({-}\Delta G_{01} /RT)/\{ 1 + \exp \, ({-}\Delta G_{01} /RT)\}$$
(4)

ΔG 01 is also the function of the temperature involving three parameters, which are heat capacity change (ΔC p), transition temperature (T m), and van’t Hoff enthalpy change (ΔH vH) at T m:

$$\Delta {\text{G}}_{01} = \Delta C_{\text{p}} T\ln \left( {T/T_{\text{m}} } \right) + \{ \Delta C_{\text{p}} {-}(\Delta H_{\text{vH}} /T_{\text{m}} )\} \;(T{-}T_{\text{m}} )$$
(5)

In this study, we calculated the equation on the assumption that ΔC p is 0 for either R2R3* or D178N in the DNA-free and DNA-bound states. Substituting Eqs. (4) and (5) into (3), Y results in the function containing six parameters, that is A n, A d, B n, B d, T m, and ΔH vH, to be determined. The curve fitting was carried out by the nonlinear least-square method on Origin 5.0 software.

DSC measurements

DSC experiments were carried out on a Nano-DSC calorimeter (TA instruments), as described previously [14]. The data were collected by heating the solution from 5 to 90 °C at a rate of 1 °C min−1 and the sample concentrations were 2.0 mg mL−1 in PBS (pH 7.4). The outer buffer solution recovered from final exchanged experiment was used in the reference cell for each case. The data were analyzed using CpCalc software supplied by the manufacture, in order to subtract the baseline from heat capacity data and convert to molar heat capacity. The fitting analysis to obtain the denaturation temperature (T d) where the fraction folded is equal to that unfolded, heat capacity change (ΔC p), and calorimetric enthalpy change (ΔH cal) at T d were performed by the least-squares methods on Origin 5.0 software. The ΔC p was calculated from the difference in the C p values between unfolded and folded states at T d. ΔH cal was calculated by integrating the area in each heat capacity curve. The ΔH vH was calculated by the next equation for assuming the two-state transition.

$$\Delta H_{\text{vH}} \left( {T_{\text{d}} } \right) = \, 4RT_{\text{d}}^{2} \{ C_{\text{p}} \left( {T_{\text{d}} } \right)/\Delta H_{\text{cal}} \left( {T_{\text{d}} } \right)\}$$
(6)

where R is the gas constant. Using the ΔC p values with the ΔH cal and T d values, the thermodynamic parameters of unfolding as a function of temperature, ΔG (T), ΔH (T), and ΔS (T), could be calculated from the following equation, on the assumption that the ΔC p is constant within the temperature range analyzed, as described previously [7, 14, 15].

$$\Delta G\left( T \right) = \Delta H_{\text{cal}} \left( {T_{\text{d}} } \right)\;\{ 1{-}\left( {T/T_{\text{d}} } \right)\} {-}\Delta C_{\text{p}} T\ln \, \left( {T/T_{\text{d}} } \right){-}\Delta C_{\text{p}} (T_{\text{d}} {-}T)$$
(7)
$$\Delta H\left( T \right) = \Delta H_{\text{cal}} \left( {T_{\text{d}} } \right) + \Delta C_{\text{p}} (T{-}T_{\text{d}} )$$
(8)
$$\Delta S\left( T \right) = \Delta S\left( {T_{\text{d}} } \right) + \Delta C_{\text{p}} \ln \;\left( {T/T_{\text{d}} } \right)$$
(9)

Results

Secondary and tertiary structure analysis

R2R3* and D178N were overexpressed in E. coli with similar yields. To investigate the effect of the mutation on the protein structure, far- and near-UV CD spectra were measured. The far-UV CD spectrum of D178N shows an α-helical-rich structure, similar to that of R2R3*, whereas the near-UV CD spectrum is slightly different especially at 290-310 nm (Fig. 2a, b). These results indicated that the D178N mutation affects the tertiary structure with little changes in the secondary structure.

Fig. 2
figure 2

Secondary and tertiary structure analysis. a Far-UV CD spectra of R2R3* (black) and D178N (gray) in the absence (thick) or presence (thin) of cognate-DNA. b Near-UV CD spectra of R2R3* (black) and D178N (gray) in the DNA-free states. The measurements were performed at 20 °C in PBS (pH 7.4)

Full size image

Thermal stability analysis by CD

Thermal transition curves were monitored by measuring the intensity changes in the CD spectrum at 222 nm between 20 and 80 °C. The apparent fractions in the unfolded states were plotted as a function of temperature (Fig. 3a). The transition temperature (T m) and van’t Hoff enthalpy change (ΔH vH) were determined by fitting the obtained curves with a two-state transition model (Table 1). The T m value of D178N is 41.1 ± 0.3 °C, which is 7.4 ± 0.4 °C lower than that of R2R3*. The thermal stability of D178N considerably decreased, indicating that Asp-178 is important for the stability of the c-Myb R2R3 structure. The ΔH vH value of D178N is also smaller than that of R2R3*. The analysis of the thermal stabilities of R2R3* and D178N in the presence of the cognate-DNA MBS-I showed that they increased upon DNA binding. Interestingly, the T m value of the D178N-DNA complex is similar to that of R2R3*-DNA complex (Table 1), whereas the ΔH vH value of D178N is higher than that of R2R3*.

Fig. 3
figure 3

Thermal denaturation analysis. a The transition curves of R2R3* (black) and D178N (gray) in the absence (thick) or presence (thin) of cognate-DNA. The apparent fraction of folded protein by monitoring the CD values at 222 nm is shown as a function of the temperature. b The heat capacity curves of R2R3* (black) and D178N (gray)

Full size image
Table 1 Thermodynamic parameters for thermal denaturation of R2R3* and D178N analyzed by CD
Full size table

Thermal stability analysis by DSC

To determine the thermodynamic parameters of the thermal unfolding in further detail, the thermal stability of D178N was also studied using DSC. Figure 3b shows the heat capacity curves of the protein analyzed at heating rate of 1.0 °C min−1. The DSC measurement showed that the thermal unfolding process was quantitatively reversible, similar to those of other R2R3 mutants [7, 14]. The T d, ΔH cal, and ΔC p values were determined using the two-state transition model, as described previously [14]. The data were well fitted on the assumption that the ratio ΔH cal/ΔH vH is 1 (Table 2). The T d value of D178N is 44.3 °C, which is lower than that of R2R3* and similar to the T m value determined using CD. The ΔH cal and ΔC p values of D178N are 172 kJ mol−1 and 7.79 kJ mol−1 K−1, respectively, both of which are smaller than those of R2R3*.

Table 2 Thermodynamic parameters for thermal denaturation of R2R3* and D178N analyzed by DSC
Full size table

Discussion

The salt-bridge between His-137 in R2 and Asp-178 in R3 is important for the conformational stability of c-Myb R2R3. The disruption of the salt-bridge changes the cooperativity between R2 and R3, similar to when the linker connecting R2 and R3 is mutated [7]. Salt-bridge formation was expected by observing the crystal structure of the protein (PDB code 1GV2), and its presence was confirmed by the solution thermodynamic analysis of this study.

To investigate the thermodynamic origin of structural difference, the thermodynamic parameters of D178N at 50.3 °C, the denaturation temperature of R2R3*, were calculated using the ΔC p value, as described previously [14]. Within a narrow range of temperatures around the T d value, the errors on the calculated ΔG and ΔH values at the reference temperature are small, though the ΔC p values obtained have an error. The thermodynamic parameters are summarized in Table 3. The results clearly indicate that the thermal stability of D178N decreases due to unfavorable ΔH, which is partially compensated by favorable ΔS. The lack of a salt-bridge between His-137 in R2 and Asp-178 in R3 in the D178N mutant causes the unfavorable and favorable ΔH and ΔS, respectively. Assuming that the D178N mutation has little effects on the enthalpy and entropy levels in the unfolded state, the decreased ΔH of D178N is due to the increased enthalpy level in the folded state, relative to that of R2R3*. It should be noted that the small ΔH value of D178N (48 kJ mol−1) compared to that of R2R3* is much larger than that could derive from a salt-bridge, suggesting that the disruption of the salt-bridge destabilizes the overall R2R3 structure, similar to what happened in the case of the linker mutation in R2R3 [14]. In addition, the disruption of the salt-bridge increased the dynamics of the folded state, resulting in favorable ΔS value. Previous studies have shown that R2 is very dynamic, but it is stabilized by connecting to R3 [8]. In contrast to R2R3*, the results of well-fitting to the DSC data of D178N on the assumption that the ratio ΔH cal/ΔH vH is 1 indicate that the disruption of the salt-bridge weakens the cooperative interaction between R2 and R3 [14]. Although the ΔH vH determined by the fitting procedure for the change in secondary structure might be different from the accurate ΔH (Fig. 3a), it still suggests the existence of cooperative interaction. Taken together, the salt-bridge between His-137 in R2 and Asp-178 in R3 contributes to the fine-tuned R2R3 architecture.

Table 3 Thermodynamic parameters for denaturation of R2R3* and D178N at 50.3 °C
Full size table

The thermal stability of D178N was significantly increased by the presence of cognate-DNA. As described in Table S1, the differences in the T m values were determined. The ΔT m (bound, D178N) is approximately twice larger than ΔT m (bound, WT), indicating that the conformational stability of D178N is remarkably affected by DNA binding. Likewise, ΔT m (bound, mut) is significantly smaller than ΔT m (R2R3, mut), and the sum of ΔT m (bound, WT) and ΔT m (R2R3, mut) is nearly equal to ΔT m (bound, D178N). These results suggest that the salt-bridge between His-137 in R2 and Asp-178 in R3 only occurs in the DNA-free state (Fig. 4), which is confirmed by the NMR and crystal structures of R2R3 bound to DNA [9, 11]. Although Asp-178 is not directly involved in DNA recognition, its positive charge would contribute to the overall electrostatic interaction with DNA. The unique dynamic properties of R2R3, which result in fine-tuned folding architecture to which the salt-bridge largely contributes, are required for its specific DNA-binding function [8]. R2R3 largely fluctuates like a “semi-intrinsically disordered” protein to find the specific binding site on the target DNA. When R2R3 binds to the specific-site, it is stabilized by inter-molecular interactions with DNA, resulting in the disruption of the salt-bridge between His-137 and Asp-178.

Fig. 4
figure 4

Schematic view of the role of salt-bridge between His-137 in R2 and Asp-178 in R3 and its effects on folding thermodynamics

Full size image

The DNA-binding domains of Myb-family or Myb-like proteins are classified into two groups: One is composed of single repeats (group 1); the other, two repeats that bind DNA cooperatively (group 2). Group 1 includes many telomere-binding proteins such as hTRF1, hTFR2, TRFL, TBP, and YlTay1 [17–22]. These proteins have no aspartic acid residues corresponding to Asp-178 in c-Myb R3. Instead, group 2 includes a variety of DNA-binding proteins containing helix-turn-helix motifs such as A- and B-Myb, v-Myb, tvMyb1-3, and VvMyb series proteins, which have highly conserved aspartic acid residues corresponding to Asp-178 in c-Myb R3 [11, 23–27]. The solution NMR structure of tvMyb1 showed that several distinctive salt-bridges formed between the two repeats contributed to the stabilization of the relative orientation of the repeats themselves, resulting in high stability toward thermal denaturation [24]. In this study, we showed that Asp-178 largely contributes to the stability of R2R3 due to salt-bridge formation, which seems to be the only interaction between the two repeats based on the crystal structure [11].

Conclusions

The minimum DNA-binding unit of c-Myb, R2R3, largely fluctuates in solution. The crystal structure of R2R3 has shown that Asp-178 in R3 is spatially next to His-135 in R3, both of which form a surface-exposed salt-bridge with each other. The present mutational and thermodynamic analyses showed that the thermal stability of D178N mutant in the absence of DNA considerably decreased, indicating that Asp-178 is involved in the salt-bridge formation with His-135. The inter-repeat salt-bridge maintains the “semi-intrinsically disordered state” of R2R3, allowing specific DNA-binding. The large fluctuations of R2R3 are crucial for in vivo specific-site recognition of c-Myb.