Thermodynamic Analyses of Nucleotide Binding to an Isolated Monomeric β Subunit and the α₃β₃γ Subcomplex of F₁-ATPase

Abstract

Rotation of the γ subunit of the F₁-ATPase plays an essential role in energy transduction by F₁-ATPase. Hydrolysis of an ATP molecule induces a 120° step rotation that consists of an 80° substep and 40° substep. ATP binding together with ADP release causes the first 80° step rotation. Thus, nucleotide binding is very important for rotation and energy transduction by F₁-ATPase. In this study, we introduced a βY341W mutation as an optical probe for nucleotide binding to catalytic sites, and a βE190Q mutation that suppresses the hydrolysis of nucleoside triphosphate (NTP). Using a mutant monomeric βY341W subunit and a mutant α₃β₃γ subcomplex containing the βY341W mutation with or without an additional βE190Q mutation, we examined the binding of various NTPs (i.e., ATP, GTP, and ITP) and nucleoside diphosphates (NDPs, i.e., ADP, GDP, and IDP). The affinity (1/K_d) of the nucleotides for the isolated β subunit and third catalytic site in the subcomplex was in the order ATP/ADP > GTP/GDP > ITP/IDP. We performed van’t Hoff analyses to obtain the thermodynamic parameters of nucleotide binding. For the isolated β subunit, NDPs and NTPs with the same base moiety exhibited similar ΔH⁰ and ΔG⁰ values at 25°C. The binding of nucleotides with different bases to the isolated β subunit resulted in different entropy changes. Interestingly, NDP binding to the α₃β(Y341W)₃γ subcomplex had similar K_d and ΔG⁰ values as binding to the isolated β(Y341W) subunit, but the contributions of the enthalpy term and the entropy term were very different. We discuss these results in terms of the change in the tightness of the subunit packing, which reduces the excluded volume between subunits and increases water entropy.

Introduction

F₁-ATPase is a cytoplasmic portion of the F_oF₁-ATP synthase, which synthesizes ATP from ADP and inorganic phosphate (P_i) by using the energy of the proton flow driven by a transmembrane electrochemical proton gradient (ΔμH⁺) (1,2). By analyzing the oxygen exchange reactions, Boyer (3-5) established the binding change mechanism, whereby the proton translocation driven by ΔμH⁺ in the F_o portion is energetically coupled with the substrate binding and product release steps in F₁-ATPase. He further proposed a rotary mechanism in which single-copy subunits rotate within the α₃β₃ cylinder (3–5) during catalysis. The rotary mechanism was strongly supported by the crystal structure of F₁-ATPase (6) and was finally envisioned by means of single-molecule experiments using the α₃β₃γ subcomplex (7). Single-molecule experiments further revealed that hydrolysis of one ATP molecule corresponded to a 120° step rotation (8) consisting of 80° and 40° substeps (9–11). ATP binding (11,12) together with ADP release (13) triggered the first 80° substep without cleavage of the chemical bond, which was in harmony with the binding change mechanism. The crystal structure indicated that the catalytic β subunit undergoes a large bending motion upon binding of nucleotides, and an NMR study revealed that a similar structural change occurs in an isolated β subunit (14,15) and a subcomplex in solution (16). These data indicate that nucleotide binding and release are critical for the rotation and hence the energy transduction by F₁-ATPase or F_oF₁-ATP synthase.

Many experimental techniques have been used to investigate nucleotide binding to both F₁-ATPase and ATP synthase. One of the most powerful of these methods involves the use of Trp mutants, whose responses to fluorescence indicate nucleotide binding or nucleotide-binding-induced conformational change (17–21). In this study, Y341 in the β subunit that interacts directly with the base of bound nucleotide was replaced with tryptophan (17–20). The indole ring of the introduced tryptophan stacks the base moiety of the nucleotide bound to the catalytic site, and its fluorescence quenching can serve as an optical probe for nucleotide binding to catalytic sites. In addition, to suppress the hydrolysis of the bound nucleotide during the experiments, we introduced a βE190Q mutation when necessary. The residue βE190 is known to play an indispensable role in catalysis (22), and this mutation suppressed the hydrolytic activity of the α₃β₃γ subcomplex almost completely. Then, using a mutant monomeric β subunit (βY341W) and a mutant α₃β₃γ subcomplex containing the βY341W mutation with or without the additional βE190Q mutation, we examined the binding of various nucleoside triphosphates (NTPs, i.e., ATP, GTP, and ITP) and nucleoside diphosphates (NDPs, i.e., ADP, GDP, and IDP). For nucleotide binding to the subcomplex, we focused on the third binding site because during the steady-state turnover, two catalytic sites are always occupied and NTP binding to the third site causes an 80° substep.

We obtained thermodynamic parameters by conducting van’t Hoff analyses on the temperature dependency of the dissociation constant (K_d). We observed that the Gibbs free-energy changes induced by NDP binding to the β(Y341W) monomer and the α₃β(Y341W)₃γ subcomplex were similar, but binding to the subcomplex resulted in less heat and more entropy production than binding to the β monomer. ADP binding to the β(Y341W/E190Q) monomer and α₃β(Y341W/E190Q)₃γ subcomplex showed a similar tendency. However, NTP binding to the β(Y341W/E190Q) monomer and α₃β(Y341W/E190Q)₃γ subcomplex showed the opposite tendency: NTP binding to the α₃β(Y341W/E190Q)₃γ subcomplex resulted in more heat and less entropy production. We discuss the results in relation to the effect of complex formation and the recently proposed water entropy effect (23,24).

Materials and Methods

Strains, plasmids, and preparation of subcomplexes

Escherichia coli strain JM109 was used for plasmid amplification. JM103Δ(uncB-uncD) was used for overexpression of the β subunit and the α₃β₃γ subcomplex of F₁-ATPase from thermophilic bacterium PS3 (TF₁). The plasmids used were puc β, which carried the gene for the β subunit for both mutagenesis and expression (21), and pTABG1 and pKABG1 (25,26), which carried the genes for the α, β, and γ subunits of TF₁ for mutagenesis and gene expression, respectively. We introduced the βY341W mutation into puc β by ligating the MluI-SmaI fragment of pTABG1, which contained the βY341W mutation for nucleotide-binding measurements in the corresponding site of puc β (18). An α-W463F mutation was also introduced to reduce the background fluorescence during the nucleotide-binding measurements. We introduced the βE190Q mutation, which greatly reduces ATPase activity (27,28), to analyze NTP binding to the α₃β₃γ subcomplex with or without this mutation. For future studies of single-molecule rotation experiments, the mutations α-C193S, γ-S109C, and γ-I212C for specific biotinylation of the γ subunit (9) and β-His₁₀ at the amino terminus (7) for fixation on Ni-NTA-coated glass were introduced.

The mutant β subunit and subcomplex containing the above mutations were overexpressed and isolated as described previously (25,26). Before conducting fluorescence-measurement experiments, we fractionated the proteins on a Superdex 200 gel filtration column (GE Healthcare, Little Chalfont, United Kingdom) using 50 mM 3-morpholinopropane-sulfonic acid-KOH (pH 7.0) buffer containing 50 mM KCl, and 2 mM MgCl₂ (MKM buffer) to remove aggregated and/or disassembled subunits.

Fluorescence measurements

Nucleotide binding was monitored by the decrease in fluorescence of the genetically introduced Trp (βY341W). Fluorescence measurements were performed with an FP-6500 spectrofluorometer (JASCO, Tokyo, Japan). Typically, the excitation wavelength was 300 nm and the emission wavelength was 345 nm, with a slit width of 3 and 10 nm, respectively, for time-course measurements. In the case of GTP/GDP binding, the excitation wavelength was set to 312 nm to avoid the absorbance of GTP/GDP at 300 nm. The concentration of the subcomplex was typically 50 nM, and for the GTP/GDP-binding studies, it was increased to 300 nM. The concentration of the isolated β subunit was double that of the subcomplex (100 nM or 600 nM for GTP/GDP). For time-course measurements, aliquots of concentrated nucleotide solutions containing equimolar MgCl₂ were injected into solutions of the β subunit or the subcomplex in MKM buffer while stirring. By circulating temperature-controlled antifreezing liquid, we performed the measurements between 2°C and 50°C. Dry air was introduced to prevent condensation at low temperatures.

Data analyses

Fluorescence titration of the isolated β subunit with various nucleotides

Fig. 1 A shows the titration of MgATP binding to the isolated β(Y341W) subunit through successive addition of small aliquots of concentrated MgATP solution. After correcting for the small effects of dilution, the fluorescence decrease (ΔF) was plotted against nucleotide concentration (Fig. 1 B) and the dissociation constant for MgATP was obtained by curve fitting using Eq. 1:

$$\Delta \text{F}=\Delta {\text{F}}_{\text{monomer}}\times \frac{\left[\text{MgATP}\right]}{{\text{Kd}}_{\text{monomer}}+\left[\text{MgATP}\right]}$$ (1)

Here, ΔF, ΔF_monomer, and K_dmonomer represent the fluorescence decrease at each MgATP concentration, the total fluorescence decrease at saturating MgATP, and the dissociation constant for MgATP, respectively. We carried out similar experiments using a combination of the βY341W mutant, βY341W/E190Q double mutant, and ATP, ADP, GTP, GDP, ITP, and IDP.

Figure 1.

Fluorescence titration of MgATP binding to the isolated β(Y341W) subunit. (A) Quenching of the fluorescence of the reporter tryptophan by successive addition of concentrated MgATP solutions. The data point at zero concentration was determined with air. The concentration of the subunit was 100 nM and measurement was carried out at 25°C. (B) Titration curve constructed after correcting for the small effect of dilution. The curve is a best fit based on Eq. 1.

Fluorescence titration of the α₃β₃γ subcomplex with various kinds of nucleotides

Fig. 2 A shows the fluorescence change of the α₃β₃γ subcomplex upon addition of MgADP. Sequential addition of MgADP decreased the fluorescence to a constant level. Above 200 μM MgADP, the fluorescence spectrum virtually vanished and we assumed that the catalytic sites were nearly saturated with MgADP, although some baseline fluorescence remained. From the fluorescence change, a titration curve was constructed, as shown in Fig. 2 B. As can be seen, the fluorescence decreased almost linearly until 2 mol of MgADP was added to 1 mol of the subcomplex, indicating the very high affinity of the first and second catalytic sites for MgADP. The affinity of the third catalytic site was several orders of magnitude smaller than that of the first and second sites, and therefore a plateau appeared between 0.2 and 1 μM of MgADP (inset in Fig. 2 B). Because we focused on the nucleotide affinity of the third catalytic site, we analyzed the data above the plateau using Eq. 2:

$$\Delta \text{F}=\Delta {\text{F}}_{1\text{and}2}+\Delta {\text{F}}_3\times \frac{\left[\text{MgADP}\right]}{{\text{Kd}}_3+\left[\text{MgADP}\right]}$$ (2)

Here ΔF represents the fluorescence decrease at a given [MgADP], ΔF_{1 and 2} represents the sum of the maximum fluorescence decrease of the first and second catalytic sites, and ΔF₃ and K_d3 represent the maximum fluorescence decrease and dissociation constant, respectively, of the third catalytic site.

Figure 2.

(A) Time course of fluorescence change upon addition of MgADP to the α₃β(Y341W)₃γ subcomplex. At ∼50 s, aliquots of MgADP solutions were injected to give the final concentrations indicated. The data point at zero concentration was determined with air. The concentration of the subcomplex was 50 nM and measurement was carried out at 25°C. (B) Titration curve of Trp fluorescence. The fluorescence change increased linearly until the molar ratio of the MgADP/subcomplex was 2 (inset). We fitted the data above 200 nM MgADP with Eq. 2 to obtain the K_d of the nucleotide to the third catalytic site (solid line).

We carried out similar experiments using a combination of the β(Y341W) mutant and ADP, GDP, and IDP, and a combination of the β(Y341W/E190Q) double mutant and ATP, ADP, GTP, and ITP. In the combination of the β(Y341W/E190Q) double mutant and ADP, the affinity of the second catalytic site was somewhat lower than that of the β(Y341W); Eq. 3 was used instead of Eq. 2:

$$\Delta \text{F}=\Delta {\text{F}}_1+\Delta {\text{F}}_2\times \frac{\left[\text{MgADP}\right]}{{\text{Kd}}_2+\left[\text{MgADP}\right]}+\Delta {\text{F}}_3\times \frac{\left[\text{MgADP}\right]}{{\text{Kd}}_3+\left[\text{MgADP}\right]}$$ (3)

In the combination of the β(Y341W/E190Q) double mutant, GDP, and IDP, the affinities of the first and second catalytic sites were not high enough to allow us to resolve the nucleotide binding to the third catalytic site, and we could not obtain quantitative data for the K_d3 of GDP and IDP. See the Supporting Material for details.

van’t Hoff analysis of nucleotide binding

To obtain the thermodynamic parameters, we measured nucleotide binding at various temperatures. Fig. 3 A shows the titration curves for MgATP binding to the isolated β(Y341W) subunit at 4°C, 25°C, and 50°C. The K_d values increased with increasing temperatures, indicating that the binding of MgATP to the isolated β(Y341W) subunit is an exothermic reaction. In Fig. 3, B and C, van’t Hoff plots for the isolated β(Y341W) and α₃β(Y341W/E190Q)₃γ subcomplex for ATP and ADP binding, respectively, are shown. It is noteworthy that the slopes of the plots of ATP and ADP binding to the isolated β(Y341W) subunit are similar, whereas the slope of ATP binding is steeper than that of ADP binding to the α₃β(Y341W/E190Q)₃γ subcomplex.

Figure 3.

(A) Titration curves for MgATP binding to the isolated β(Y341W) subunit at various temperatures. (B) van’t Hoff plot of MgATP (square) and MgADP (circle) binding to the isolated β(Y341W) subunit. (C) van’t Hoff plot of MgATP (square) and MgADP (circle) binding to the α₃β(Y341W/E190Q)₃γ subcomplex. See Supporting Material for van’t Hoff plots of all combinations examined in this study.

We further measured binding of other nucleotides to the isolated β subunit and α₃β₃γ subcomplex that contained the βY341W mutation with or without the βE190Q mutation, and constructed van’t Hoff plots. The plots were analyzed by the linear least-squares method. The van’t Hoff plots for all nucleotide and proteins are shown in the Supporting Material.

The standard errors for the enthalpy change (ΔH⁰), entropy change (ΔS⁰), and Gibbs free-energy change (ΔG⁰) at 25°C were deduced according to Squires (29).

Results and Discussion

Summary of the obtained parameters

The thermodynamic parameters obtained by van’t Hoff analyses are summarized in Tables 1–4. The errors for ΔG⁰ at 25°C are relatively small because the temperature is near the center of the temperature range examined, and the best straight line always passes over the center of gravity of the plotted points (29).

Table 1.

Analyses of nucleotide binding to the β(Y341W) monomer

	ΔH0 (kJ·mol−1)	ΔS0 (J·mol−1·K−1)	ΔG0 at 25°C (kJ·mol−1)	Kd at 25°C (μM)
ATP	−36 ± 1	−24 ± 4	−28.7 ± 0.1	7.1 ± 0.5
ADP	−38 ± 1	−32 ± 4	−28.6 ± 0.1	7.5 ± 0.4
GTP	−38 ± 1	−51 ± 4	−23.0 ± 0.1	74 ± 4
GDP	−36 ± 1	−41 ± 4	−23.4 ± 0.1	65 ± 4
ITP	−39 ± 1	−61 ± 4	−20.6 ± 0.1	200 ± 10
IDP	−33 ± 2	−41 ± 5	−20.9 ± 0.1	180 ± 20

Table 2.

Analyses of nucleotide binding to the β(Y341W/E190Q) monomer

	ΔH0 (kJ·mol−1)	ΔS0 (J·mol−1·K−1)	ΔG0 at 25°C (kJ·mol−1)	Kd at 25°C (μM)
ATP	−19.3 ± 0.5	22 ± 2	−25.9 ± 0	22.7 ± 0.3
ADP	−22 ± 1	20 ± 4	−27.9 ± 0.1	10.1 ± 0.6
GTP	−23 ± 3	−10 ± 10	−20.4 ± 0.2	230 ± 20
GDP	−27 ± 1	−16 ± 4	−22.2 ± 0.1	107 ± 6
ITP	−22.3 ± 0.9	−14 ± 3	−18.3 ± 0	530 ± 20
IDP	−23 ± 1	−10 ± 3	−19.7 ± 0	290 ± 10

Table 3.

Analyses of nucleotide binding to the α₃β(Y341W)₃γ subcomplex

	ΔH0 (kJ·mol−1)	ΔS0 (J·mol−1·K−1)	ΔG0 at 25°C (kJ·mol−1)	Kd at 25°C (μM)
ADP	−18 ± 1	33 ± 5	−28.2 ± 0.1	8.9 ± 0.6
GDP	−23 ± 2	5 ± 6	−23.6 ± 0.1	59 ± 5
IDP	−22 ± 1	0.1 ± 4	−21.7 ± 0.1	128 ± 9

Table 4.

Analyses of nucleotide binding to the α₃β(Y341W/ E190Q)₃γ subcomplex

	ΔH0 (kJ·mol−1)	ΔS0 (J·mol−1·K−1)	ΔG0 at 25°C (kJ·mol−1)	Kd at 25°C (μM)
ATP	−28 ± 3	−4 ± 9	−27.1 ± 0.1	14 ± 2
ADP	−21 ± 3	24 ± 9	−27.9 ± 0.1	10 ± 1
GTP	−28 ± 3	−19 ± 9	−22.3 ± 0.1	100 ± 10
ITP	−28 ± 4	−27 ± 10	−19.9 ± 0.2	270 ± 50

In previous studies, the thermodynamic parameters for MgATP and MgADP binding to the isolated β subunit from thermophilic bacterium PS3 were determined by isothermal titration calorimetry (30,31). The results show remarkable similarities to those presented here, even though different methods were used. The slight difference between our parameters and those obtained in the previous studies may be due to differences in the experimental conditions employed, such as the pH or ionic strength. In this study, we extended the experimental conditions to the α₃β₃γ subcomplex and nucleotides other than ATP and ADP to gain new insights into the mechanism of F₁-ATPase.

Comparison of NTP and NDP binding to the monomeric β subunit, and the effect of the βE190Q mutation

The K_d and ΔG⁰ values for nucleotide binding to the monomeric β(Y341W) subunit were essentially the same for all NTPs and NDPs with the same base (Table 1; Fig. 4). The affinities were in the order ATP/ADP > GTP/GDP > ITP/IDP. The difference between the nucleotides with different bases was detectable by the changes in the amount of negative entropy. This could be because the base moiety of the nucleotide is accommodated in the hydrophobic pocket composed of W341, F414, and F420. At first glance, the monomeric β(Y341W) subunit does not strictly discriminate between NTP and NDP with the same base. It is notable that the entropy decrease in ADP, GTP, and ITP binding is slightly larger than that in ATP, GDP and IDP binding. However, these differences in entropy are relatively small and are compensated for by the subtle differences in enthalpy change to yield similar values of Gibbs free-energy change for the binding of ATP and ADP, GTP and GDP, and ITP and IDP, respectively. Introduction of the βE190Q mutation increases the K_d values for the NTPs by roughly threefold, but increases the K_d for the NDPs by less than twofold compared with the β(Y341W) monomer (Table 2). Accordingly, the Gibbs free-energy change in nucleotide binding increases by ∼3 kJ/mol for NTPs and 1 kJ/mol for NDPs. This increase is due to the smaller negative enthalpy change (less heat production) for the β(Y341W/E190Q) monomer than for the β(Y341W) monomer (Fig. 4). The entropy changes for nucleotide binding to the β(Y341W/E190Q) monomer were larger than those for the β(Y341W) monomer, but at 25°C, the effect of enthalpy changes exceeded the effect of the entropy change on ΔG⁰.

Figure 4.

Comparison of the ΔH⁰ and ΔS⁰ values for nucleotide binding to the β(Y341W) subunit (left) and the β(Y341W/E190Q) monomer (right). Data are taken from rows 1–6 in Table 1 and rows 1–6 in Table 2.

The data suggest that because of the absence of βE190, the polar or hydrogen-bonding interactions between nucleotide and protein seem to be impeded in the β(Y341W/E190Q) monomer, which results in a smaller negative enthalpy change and conformational entropy decrease than in the β(Y341W) monomer. The entropy increase observed in ATP and ADP binding to the β(Y341W/E190Q) monomer may also reflect the release of water molecules upon hydrogen-bond formation. This basal effect seems to be masked by other entropy-decreasing factors in the case of GTP/GDP and ITP/IDP.

Comparison of NTP and NDP binding to the α₃β₃γ subcomplex

A comparison of NTP and NDP binding to the α₃β₃γ subcomplex was possible only in the combination of α₃β(Y341W/E190Q)₃γ and ATP and ADP (Table 4, first and second rows; Fig. 5). The K_d and ΔG⁰ values at 25°C for ATP and ADP were similar, but the contributions of enthalpy and entropy were very different, suggesting that the γ phosphate of ATP still could have a polar interaction with the mutated Gln residue, which causes more heat production than ADP binding, whereas the stabilization effect causes the entropy to decrease.

Figure 5.

Comparison of ΔH⁰ and ΔS⁰ values for MgATP and MgADP binding to the α₃β(Y341W/E190Q)₃γ subcomplex. Data are taken from rows 1 and 2 in Table 4.

Effect of complex formation on nucleotide binding

A comparison of rows 2, 4, and 6 in Table 1 and rows 1–3 in Table 3 shows only a small change in the K_d of NDP binding and the corresponding Gibbs free-energy change between the monomeric β(Y341W) and α₃β(Y341W)₃γ subcomplex (Fig. 6). However, when the enthalpy and entropy changes are plotted, clear differences appear. Apparently, the negative enthalpy change or heat production of ADP, GDP, and IDP binding is smaller in binding to the α₃β(Y341W)₃γ subcomplex than in binding to the β(Y341W) monomer. ADP, GDP, and IDP binding to the β(Y341W) monomer causes an entropy decrease, whereas ADP and GDP binding to α₃β(Y341W)₃γ subcomplex causes an entropy increase, and IDP binding causes a slightly positive or almost negligible entropy increase. The entropy increase resulting from binding ADP to the subcomplex is also seen for the α₃β(Y341W/E190Q)₃γ subcomplex in Fig. 5. At first glance, this result is puzzling because one would expect more hydrogen-bond formation connecting adjacent α, β subunits and bound nucleotide, and hence a larger loss of the flexibility of the protein structure for the α₃β(Y341W)₃γ subcomplex than for the β(Y341W) monomer upon nucleotide binding. As hydrogen-bond formation accompanies a positive bond energy of roughly 10–30 kJ mol⁻¹ (exothermic) (32,33), it would be more reasonable to expect more heat production and a larger loss of the protein conformational entropy for the α₃β(Y341W)₃γ subcomplex; however, this was not the case. This result can be explained only through some endothermic and entropy-producing processes that may arise in the subcomplex upon nucleotide binding, but not in the isolated subunit. One possible candidate for such a process is the nucleotide-binding-induced change in the tightness of the packing at the subunit interfaces. Recently, Yoshidome et al. (23,24) proposed that tight packing of subunits reduces the excluded volume between them and leads to a large translational water entropy gain in F₁-ATPase. It is expected that the water release is accompanied by heat absorption. We therefore suspect that a similar change occurs upon a third nucleotide binding to the α₃β(Y341W)₃γ subcomplex.

Figure 6.

Comparison of ΔH⁰ and ΔS⁰ values for NDP binding to the β(Y341W) monomer and the α₃β(Y341W) ₃γ subcomplex. Data are taken from rows 2, 4, and 6 in Table 1 and rows 1–3 in Table 3.

When we compared nucleotide binding to the monomeric β(Y341W/E190Q) and α₃β(Y341W/E190Q)₃γ subcomplex, we found that ADP binding to the subcomplex caused less of a decrease in enthalpy and more of an increase in entropy than binding to the monomeric subunit. However, binding of other NTPs (ATP, GTP, and ITP) to the subcomplex caused a greater decrease in both enthalpy and entropy than did binding to the monomeric subunit (Fig. 7). In the crystal structure discussed by Yoshidome et al. (23,24), three β subunits are in the nucleotide-free form, ATP-bound form, and ADP-bound form, respectively. They found that the entropy increase by the water entropy effect occurred at the interface of the ADP-bound β subunit and adjacent α and γ subunits, but not at the interface of the ATP-bound β subunit and the adjacent subunits. Although it is difficult to compare our experimental conditions with the crystal structure, our experimental results indicating that NDP binding to the subcomplex causes an entropy increase, whereas NTP binding causes an entropy decrease compared with the monomeric subunit, are reminiscent of the asymmetric packing of the crystal structure and water entropy effect proposed by Yoshidome et al. (23,24).

Figure 7.

Comparison of ΔH⁰ and ΔS⁰ values for nucleotide binding to the β(Y341W/E190Q) monomer and the α₃β(Y341W/E190Q) ₃γ subcomplex. Data are taken from rows 1–3 and 5 in Table 2 and rows 1–4 in Table 4.

In summary, we carried out nucleotide-binding assays with the monomeric β subunit and α₃β₃γ subcomplex of F₁-ATPase with a βY341W mutation and with or without the βE190Q mutation using NTPs (ATP, GTP, and ITP) and NDPs (ADP, GDP, and IDP). In many cases, we observed compensation between enthalpy and entropy changes, which seems to occur by the formation of polar or hydrogen bonds and simultaneous stabilization of the protein conformation. In addition, upon NDP binding to the α₃β(Y341W)₃γ subcomplex, we observed a prominent increase in entropy and enthalpy compared with binding to the β(Y341W) monomer, which suggests a possible subunit-packing-induced water entropy effect (23,24) and accompanying heat absorption.