The rotary mechanism of ATP synthase requires a strong binding within stator subunits. In this work we studied the binding affinity of the b-subunit to F(1)-ATPase of Escherichia coli. The dimerization of the truncated b-subunit without amino acids 1-33, b(34-156)T62C, was investigated by analytical ultracentrifugation, resulting in a dissociation constant of 1.8 microM. The binding of b-subunit monomeric and dimeric forms to the isolated F(1) part was investigated by fluorescence correlation spectroscopy and steady-state fluorescence. The mutants b(34-156)T62C and EF(1)-gammaT106C were labeled with several fluorophores. Fluorescence correlation spectroscopy was used to measure translational diffusion times of the labeled b-subunit, labeled F(1), and a mixture of the labeled b-subunit with unlabeled F(1). Data analysis revealed a dissociation constant of 0.2 nM of the F(1)b(2) complex, yielding a Gibbs free energy of binding of DeltaG(o)= -55 kJ mol(-1). In steady-state fluorescence resonance energy transfer (FRET) measurements it was found that binding of the b-subunit to EF(1)-gammaT106C-Alexa488 resulted in a fluorescence decrease of one-third of the initial FRET donor fluorescence intensity. The decrease of fluorescence was measured as a function of b-concentration, and data were described by a model including equilibria for dimerization of the b-subunit and binding of b and b(2) to F(1). For a quantitative description of fluorescence decrease we used two different models: the binding of the first and the second b-subunit causes the same fluorescence decrease (model 1) or only the binding of the first b-subunit causes fluorescence decrease (model 2). Data evaluation revealed a dissociation constant for the F(1)b(2) complex of 0.6 nM (model 1) or 14 nM (model 2), giving DeltaG(o)= -52 kJ mol(-1) and DeltaG(o)= -45 kJ mol(-1), respectively. The maximal DeltaG observed for ATP synthesis in cells is approximately DeltaG= 55 kJ mol(-1). Therefore, the binding energy of the b-subunit seems to be too low for models in which the free energy for ATP synthesis is accumulated in the elastic strain between rotor and stator subunits and then transduced to the catalytic site in one single step. Models in which energy transduction takes place in at least two steps are favored.