Controversy exists over whether the chaperonin GroEL forms a GroEL–(GroES)2 complex (football-shaped complex) during its reaction cycle. We have revealed previously the existence of the football-shaped complex in the chaperonin reaction cycle using a FRET (fluorescence resonance energy transfer) assay [Sameshima, Ueno, Iizuka, Ishii, Terada, Okabe and Funatsu (2008) J. Biol. Chem. 283, 23765–23773]. Although denatured proteins alter the ATPase activity of GroEL and the dynamics of the GroEL–GroES interaction, the effect of denatured proteins on the formation of the football-shaped complex has not been characterized. In the present study, a FRET assay was used to demonstrate that denatured proteins facilitate the formation of the football-shaped complex. The presence of denatured proteins was also found to increase the rate of association of GroES to the trans-ring of GroEL. Furthermore, denatured proteins decrease the inhibitory influence of ADP on ATP-induced association of GroES to the trans-ring of GroEL. From these findings we conclude that denatured proteins facilitate the dissociation of ADP from the trans-ring of GroEL and the concomitant association of ATP and the second GroES.
Chaperonins belong to a ubiquitous class of molecular chaperones that promote protein folding in the cell and are found in bacteria, chloroplasts, mitochondria, archaea and the eukaryotic cytosol [1–4]. The GroEL–GroES chaperonin system of Escherichia coli is one of the best characterized chaperonin systems. GroEL is a homo-oligomer of 14 identical 57-kDa subunits arranged into two heptameric rings. Each ring of GroEL has a large central open cavity and the two rings are stacked back-to-back. GroES contains seven identical 10-kDa subunits which assemble as a heptamer ring. A large conformational change induced by ATP binding to GroEL promotes formation of the GroEL–GroES complex. Binding of the GroES lid seals the GroEL cavity and releases a denatured protein into the confined space of the GroEL–GroES chamber. Protein folding proceeds in this isolated space for several seconds until the GroEL–GroES complex dismantles and the substrate protein, folded or unfolded, is released into the solution.
It has been widely accepted that GroES binds to each ring of GroEL alternatively (a two-stroke model) and promotes protein folding by GroEL [3,5]. In other words, an asymmetric GroEL–GroES complex (termed a bullet-shaped complex) is considered to be the only form of the GroEL–GroES complex, and a symmetric GroEL–(GroES)2 complex (termed a football-shaped complex) is believed not to be stably formed during the reaction cycle [6–9]. However, some studies using electron microscopy [10–13] or chemical cross-linking [14,15] have reported that two GroES molecules simultaneously bind to one GroEL molecule in the presence of ATP. Furthermore, we have observed previously that bullet- and football-shaped complexes co-exist during the reaction cycle of GroEL using a FRET (fluorescence resonance energy transfer) assay between GroEL and GroES . The question that arises from these studies is why controversy exists over whether GroEL forms the football-shaped complex during its reaction cycle. One possible answer to this question is the accumulation of ADP in solution during the biochemical measurements. We have shown previously that ADP prevents the association of ATP to the ring of GroEL, which does not have a GroES (trans-ring), and strongly inhibits the formation of the football-shaped complex . However, some studies have not observed a football-shaped complex in the absence of ADP [6,8]. Consequently, there is most likely an alternative explanation that rationalizes the observation of the football-shaped complex.
Denatured proteins are known to alter the ATPase activity of GroEL [17–19] and the dynamics of the GroEL–GroES interaction [9,20,21]. Hence, we assumed that the formation of the football-shaped complex could also be affected by denatured proteins. To test this hypothesis, we have studied the effect of denatured proteins on the formation of the football-shaped complex. We found that the football-shaped complex disappears in the absence of denatured proteins, whereas the formation of the football-shaped complex was facilitated by the presence of denatured proteins. In this report, we also discuss the mechanism by which denatured proteins promote the formation of the football-shaped complex.
Reagents and proteins
ATP, ADP, NADH, phosphoenolpyruvate and pyruvate kinase from rabbit muscle were purchased from Roche Diagnostics. BeSO4, oxaloacetic acid, urea, MDH (malate dehydrogenase) from porcine hearts, α-lactalbumin from bovine milk (Type III, calcium-depleted), hexokinase from Saccharomyces cerevisiae (Type F-300, lyophilized powder) and pyruvate kinase/lactic dehydrogenase enzymes from rabbit muscle were obtained from Sigma–Aldrich. NaF and DTT (dithiothreitol) were purchased from Wako Pure Chemical Industries. ADP was treated with hexokinase and glucose to remove contaminating ATP as described previously .
Preparation of denatured proteins
MDH was used after buffer exchange on a PD-10 column (GE Healthcare) equilibrated with HKM buffer (25 mM Hepes/KOH, pH 7.4, containing 100 mM KCl and 5 mM MgCl2). The concentration of MDH was determined by the absorption at 280 nm using the molar extinction co-efficient of 7360 M−1·cm−1 (for the monomer) calculated from the MDH amino acid sequence . dMDH (denatured MDH) was prepared by adding MDH for more than 30 min to the HKM buffer containing 8 M (final concentration 6.4 M) urea at room temperature (23°C). Although up to 133 mM urea remained in a solution upon dilution, when refolding the dMDH in the presence of GroEL and GroES system, the effects of remaining urea on the ATPase activity of GroEL and FRET efficiency between GroEL and GroES were negligible (Supplementary Figure S1 available at http://www.BiochemJ.org/bj/427/bj4270247add.htm). α-Lactalbumin was dissolved in the HKM buffer and centrifuged (20400 g for 10 min at 4°C) to remove aggregated protein. The concentration of α-lactalbumin was determined by the absorption at 280 nm using the molar extinction coefficient of 28500 M−1·cm−1 . rLA (reduced α-lactalbumin) was prepared by reduction of the disulfide bonds of 50 μM α-lactalbumin for more than 1 h by the addition of 10 mM DTT in the HKM buffer at 23°C. rLA was used within 150 min after the addition of DTT to avoid protein aggregation.
Preparation and labelling of GroEL and GroES
Variants of GroEL and GroES with exposed cysteine residues were prepared to allow labelling with fluorescent dyes: the E315C variant of GroEL, GroEL-E315C [9,25] has intact endogenous cysteine residues (, but see below); the GroES mutant GroES-98C has an addition of a single cysteine residue to the C-terminus of each GroES subunit [9,16,25]; and the single-ring variant of GroEL-E315C, GroEL-E315C-SR1, replaces four residues which represent major contacts between its two rings (R452G, E461A, S463A and V464A) . The mutant proteins, and wtGroEL (wild-type GroEL) and wtGroES (wild-type GroES), were expressed in E. coli cells and purified as described previously . Purified proteins were stored at 4°C in 65% saturated ammonium sulfate until use. GroEL and GroES concentrations were expressed as the oligomer (GroEL as a tetradecamer; GroEL-E315C-SR1 and GroES as a heptamers). The concentrations were determined by the absorption at 280 nm using the following molar extinction coefficients: GroEL tetradecamer, 130480 M−1·cm−1; GroEL-E315C-SR1 heptamer, 65240 M−1·cm−1; and GroES heptamer, 8960 M−1·cm−1 .
GroEL or GroES stored in 65% saturated ammonium sulfate were centrifuged (204000 g for 10 min at 4°C) and the pellet was dissolved in HKM buffer containing 10–30 mM DTT. DTT was removed with a NAP™-5 column (GE Healthcare) equilibrated with HKM buffer and the protein concentration was determined by a spectrophotometer (V-570; Jasco). Subsequently, GroEL-E315C was labelled with Cy3 (indocarbocyanine)–maleimide (GE Healthcare) or TMR (tetramethylrhodamine)–maleimide (Invitrogen). Although GroEL-E315C has three endogenous cysteine residues, we confirmed that the Cys-315 was labelled preferentially compared with the other cysteine residues . GroES-98C was labelled with Cy5 (indodicarbocyanine)–maleimide (GE Healthcare). Cy3–maleimide or TMR–maleimide (0.8-fold) was mixed with GroEL-E315C and 8-fold of Cy5–maleimide was mixed with GroES-98C. The reaction was allowed to proceed for 90 min at room temperature in the dark. The labelled proteins were separated from unreacted dyes using a NAP™-5 column or a PD-10 column equilibrated with HKM buffer. Fluorescent dye concentrations were determined using the spectrophotometer and the following extinction coefficients: Cy3, 150000 M−1·cm−1 at 552 nm; TMR, 95000 M−1·cm−1 at 555 nm; and Cy5, 250000 M−1·cm−1 at 650 nm. The concentration of Cy3–GroEL-E315C or TMR–GroEL-E315C was determined by correcting for the 280 nm absorbance of the conjugated dye. The concentration of Cy5–GroES-98C was determined using the Lowry method (DC protein assay; BioRad Laboratories) . A standard curve for GroES concentrations was constructed using known amounts of wtGroES. The molar ratio of Cy3 or TMR to the GroEL-E315C (tetradecamer) was 0.5–1.0, and that of Cy5 to the GroES-98C (heptamer) was 2.5–4.0 throughout this study. Fluorescent labelling does not affect the ATPase or folding activity of GroEL .
FRET experiments in equilibrium conditions
HKM buffer containing GroEL-E315C (Cy3- or TMR-labelled), GroES (Cy5–GroES-98C or wtGroES) and 5 mM DTT was pre-incubated at room temperature for over 10 min. 1 mM BeSO4 and 10 mM NaF were mixed with the solution if necessary. Denatured protein was then added to the solution for 1 min followed by the addition of ATP or ADP. The solution was then incubated in a spectrofluorometer (FP-6500; Jasco) for 3 min at 23°C. Fluorescence spectra were measured at 23°C using a microcuvette (3×3×37 mm; Jasco). The excitation wavelength was set to 520 nm (for Cy3) or 525 nm (for TMR). The FRET efficiency was determined by the following equation:
where Fda is the fluorescence intensity of donor-labelled GroEL-E315C in the presence of acceptor-labelled GroES-98C and Fd is the fluorescence intensity of donor-labelled GroEL-E315C in the presence of unlabelled wtGroES. All experiments were repeated three times. The molar ratio of fluorescent dye to GroEL (tetradecamer) or GroES (heptamer) was held constant in the same experiment because the molar ratio of dye to protein influences the FRET efficiency. The detailed experimental procedures and ratios of dye per protein in each experiment are described in the appropriate Figure legends.
ATPase activity of GroEL
The release of ADP from GroEL was measured spectrophotometrically with an ATP-regeneration system as described previously . Briefly, the assay mixture contained 0.2 mM NADH, 1 mM ATP, 5 mM phosphoenolpyruvate, 5 mM DTT and 2 units/ml pyruvate kinase/lactic dehydrogenase enzymes in HKM buffer. The decreases in the absorbance at 340 nm, due to oxidation of NADH, were monitored continuously with a spectrophotometer. wtGroEL (50 nM) and wtGroES (500 nM) and different concentrations of denatured proteins (dMDH or rLA) were mixed into the stirred assay mixture during the measurement of the absorbance at 340 nm. All assays were carried out at 23°C and repeated three times.
Measurement of MDH activity
A solution containing 50 nM wtGroEL, 500 nM wtGroES, an ATP regeneration system (5 mM phosphoenolpyruvate and 10 μg/ml pyruvate kinase) and 5 mM DTT in HKM buffer was pre-incubated at 23°C. Subsequently, 50 μM dMDH was diluted 100-fold into the solution and at 1 min after the addition of dMDH, 1 mM ATP was added to the solution and the refolding reaction by GroEL was initiated. Refolding of MDH was carried out at 23°C. Aliquots (6 μl) of the solution were taken every 5 min and mixed with 1.2 ml of the MDH assay solution (HKM buffer containing 0.2 mM NADH, 0.5 mM oxaloacetic acid and 10 mM DTT) and stirred. MDH activity was measured by continuously monitoring the decrease in the absorbance of 340 nm due to oxidation of NADH at 25°C. As a control, activity of 2.5 nM native MDH (monomer concentration) was monitored and taken as 100% activity.
Association rate of GroES to each ring of GroEL
ADP–BeFx bound to the TMR-labelled bullet-shaped complex, TMR–bullet, was preformed by the addition of different concentrations of ADP to the HKM buffer containing 500 nM TMR–GroEL-E315C, 1 μM GroES (Cy5–GroES-98C or wtGroES), 5 mM DTT, 1 mM BeSO4 and 10 mM NaF and incubated at 23°C for over 1 h. TMR-bullet (10 nM) or TMR-labelled (10 nM) GroEL-E315C-SR1 (TMR–GroEL-E315C-SR1), 200 nM GroES (Cy5–GroES-98C or wtGroES), 5 mM DTT, 1 mM BeSO4 and 10 mM NaF and different concentrations of dMDH in the HKM buffer were pre-incubated at 23°C for 10 min in a spectrofluorometer. Subsequently, 1 mM ATP was injected into the stirred solution using a syringe and the change in donor fluorescence (excitation at 530 nm; emission at 570 or 580 nm) was monitored at 23°C. Fluorescence intensity was measured with the spectrofluorometer using a 10×10×45 mm cuvette. The same experiments were carried out three times and the FRET efficiency values were averaged.
Football-shaped complexes gradually disappear as the denatured protein is refolded by GroEL
In a previous study we revealed that the football-shaped complex is formed during the GroEL–GroES reaction cycle using FRET between GroEL and GroES . GroEL-E315C and GroES-98C were utilized for site-specific fluorescent labelling. Figure 1(A) shows the change in FRET efficiency between GroEL and GroES as a function of time in the presence of the dMDH. We used TMR–GroEL-E315C and Cy5–GroES-98C or wtGroES for the FRET experiments. Control experiments were conducted in the presence of beryllium fluoride (BeFx), which acts as an inorganic phosphate analogue . Football- and bullet-shaped complexes are known to form in the presence of ATP plus BeFx and ADP plus BeFx respectively . The FRET efficiencies of football- and bullet-shaped complexes were nearly constant throughout the measurement (Figure 1A). Subsequently, the change in FRET efficiency was monitored in the presence of ATP and an ATP regeneration system. Initially, the observed FRET efficiency was a value between the football- and bullet-shaped complexes. Interestingly, this FRET efficiency decreased gradually in a time-dependent manner and reached a value similar to that observed for the bullet-shaped complex (Figure 1A). We examined the time course of MDH refolding in the presence of the GroEL–GroES system and found that MDH activity recovered at a similar rate to the decrease in FRET efficiency (Figure 1B). The rate of spontaneous folding of MDH was much lower than that in the presence of GroEL–GroES system (Supplementary Figure S2 available at http://www.BiochemJ.org/bj/427/bj4270247add.htm). Therefore we assumed that the decrease in the levels of the football-shaped complex was caused by the decrease in the amount of dMDH. A time course of FRET change was then monitored in the presence of rLA, which is not folded by GroEL (Figure 1C) . In contrast with dMDH, FRET efficiency in the presence of ATP and its regeneration system was nearly constant throughout the measurement (Figure 1C). We concluded from these results that the amount of the football-shaped complex present decreased as the concentration of the denatured protein decreased.
Stability of football-shaped complexes during the GroEL–GroES interaction cycle
Bullet shaped-complex is preferentially formed in the absence of denatured proteins
We then conducted FRET titration experiments in the absence of the denatured proteins. Cy3–GroEL-E315C was titrated with increasing concentrations of GroES (Cy5–GroES-98C or wtGroES) and the FRET efficiency was calculated. In the presence of ATP plus BeFx, the FRET efficiency was saturated when a 2-fold amount of GroES was mixed with GroEL (Figure 2). In contrast, the FRET efficiency was saturated when an equal amount of GroES was mixed with GroEL in the presence of ADP plus BeFx (Figure 2). These results indicated that the presence of BeFx leads to the efficient formation of the bullet- and football-shaped complexes irrespective of the presence of denatured proteins. In the presence of ATP without BeFx, the FRET efficiency was saturated at a [GroES]/[GroEL] ratio of 1 in the absence of denatured protein, which was similar to the case of ADP plus BeFx. The detection of the football-shaped complex was essentially lost in the absence of the denatured protein using FCS (fluorescence correlation spectroscopy) analysis (results not shown).
Binding stoichiometry of the GroEL–GroES complex in the absence of denatured protein
The amount of the football-shaped complex increases in accordance with a rise in ATPase activity of GroEL
To examine how the level of the football-shaped complex depends on the level of denatured protein, FRET between GroEL and GroES were measured at different concentrations of dMDH. FRET efficiency increased with increasing amounts of dMDH (Figure 3A). Subsequently, we measured ATPase activity of GroEL in the presence of different concentrations of dMDH. The addition of increasing amounts of dMDH resulted in an increase in ATPase activity of GroEL (Figure 3A) and this result is consistent with previous studies [17–19]. The same experiment was also carried out using rLA as another denatured protein (Figure 3B). Experiments using this protein yielded almost identical results. Higher concentrations of rLA were required to reach a saturated amount of the football-shaped complex than dMDH; this is because GroEL has lower affinity for rLA than dMDH [32,33]. These results showed that the amount of the football-shaped complex increased in accordance with the ATPase activation of GroEL by the denatured proteins.
FRET efficiency of the GroEL–GroES complex and ATPase activity of GroEL at different concentrations of denatured protein
Denatured protein accelerates the association of the second GroES to the trans-ring of GroEL
To clarify the mechanism by which denatured protein facilitates the formation of a football-shaped complex, time courses of GroES binding to each ring of GroEL were monitored by FRET under different concentrations of dMDH. These experiments were carried out in the presence of BeFx to avoid dissociation of GroES from GroEL. At first, association of the first GroES to one of the GroEL rings was monitored at different concentrations of dMDH. In this experiment, we employed the single-ring variant GroEL-E315C (GroEL-E315C-SR1) to prevent the binding of the second GroES. Association of GroES to TMR–GroEL-E315C-SR1 showed little difference with increasing concentrations of dMDH (Figure 4A). Subsequently, the association of the second GroES with the trans-ring of the TMR–bullet complex was monitored. The TMR–bullet was preformed in the presence of 1 mM ADP and BeFx and this solution was diluted by 50-fold upon measurements, i.e. association of the second GroES was monitored in the presence of 20 μM ADP. In the presence of BeFx, the second GroES could associate with the trans-ring of the TMR–bullet by addition of ATP despite the absence of denatured protein (Figure 4B). The association rate of GroES to the trans-ring of the TMR–bullet was significantly accelerated as the dMDH concentration increased (Figure 4B). Therefore the denatured protein did not affect the binding of the first GroES to GroEL, whereas it accelerated the association of the second GroES to the trans-ring of GroEL.
Association rate of GroES to the cis- and the trans-ring of GroEL at different concentrations of dMDH
The association rate of the second GroES to the trans-ring of GroEL was affected by ADP in the absence of denatured protein
We have previously reported that ADP prevents the association of the second GroES to the trans-ring of GroEL . To examine whether ADP affects the association of GroES with the trans-ring of GroEL, association of GroES to the trans-ring of the TMR–bullet was monitored at different concentrations of ADP in the presence of BeFx and in the absence of the denatured protein. The association of GroES with GroEL was accelerated as the ADP concentration decreased (Figure 5), resulting in similar responses to increasing concentrations of dMDH (Figure 4B). In the presence of less than 2 μM ADP, GroES associated with the trans-ring of the TMR–bullet as soon as ATP was added to the solution (Figure 5). From these results, we conclude it is likely that ADP strongly prevents the association of the second GroES to the trans-ring of GroEL in the absence of denatured protein and that the presence of the denatured protein decreases the inhibitory effect of ADP.
Association rate of GroES to the trans-ring of GroEL at different concentrations of ADP in the absence of the denatured protein
There is an ongoing debate as to whether the football-shaped complex exists during the GroEL–GroES interaction cycle [6–8,10,11,13–15]. Currently, only a few studies focus on the influence of the denatured protein on the formation of the football-shaped complex . In the present study, we have revealed that the presence of denatured protein accelerates the binding of the second GroES with the trans-ring of GroEL, leading to the formation of the football-shaped complex. This is the first report that clearly shows the effect of denatured protein on the formation of the football-shaped complex. Thereby the present work provides a clue to resolve the argument regarding the presence of the football-shaped complex in the chaperonin reaction cycle.
Our proposed schematic model of the GroEL–GroES reaction cycle is shown in Figure 6. The model consists of two cycles: a ‘bullet cycle’ and a ‘football cycle’. In the presence of a low concentration of denatured protein, GroEL mainly goes through the bullet cycle due to the inhibitory effect of ADP in the trans-ring. On the other hand, there is a switch to the football cycle in the presence of a high concentration of denatured protein as denatured protein weakens the inhibitory effect of ADP and facilitates formation of the football-shaped complex.
Schematic model of the GroEL–GroES reaction cycles
The mechanism by which the association of the second GroES is accelerated by denatured proteins
Why is the association of the second GroES accelerated by the presence of denatured protein? One possible reason is that denatured protein facilitates the dissociation of ADP from the trans-ring of GroEL. Some reports have suggested that ADP remains in the GroEL ring even after GroES has been detached from the same ring [19,34]. We and other groups have shown that ADP prevents the association of ATP  and the second GroES to the trans-ring of GroEL [16,36]. From these reports, it appears that GroES cannot associate with the trans-ring of GroEL until ADP dissociates from the trans-ring of GroEL. In contrast, Grason et al. [19,21] suggested that the association of denatured protein with the trans-ring promotes dissociation of ADP from the ring. We have also found that denatured protein accelerates the association of the second GroES to the trans-ring of GroEL in the presence of 20 μM ADP, where the association of the second GroES to the trans-ring of GroEL is significantly decreased in the absence of denatured protein (Figure 5). From these findings, we propose a model for the GroEL–GroES interaction mechanism that includes the football-shaped complex. The football-shaped complex is formed when both rings of GroEL are occupied with ATP [16,37]. ATP hydrolysis in one of the rings of GroEL brings about dissociation of GroES because the loss of the γ-phosphate decreases the affinity between GroEL and GroES . Subsequently, denatured protein dissociates ADP from the trans-ring of GroEL and this enables ATP and the second GroES to associate with the trans-ring of GroEL (Figure 6). We showed previously that the football- and bullet-shaped complexes co-exist during the reaction cycle of GroEL . Given that dissociation of ADP is a rate-limiting step for the association of ATP and the second GroES to the trans-ring of GroEL, not all GroEL molecules can form the football-shaped complex.
However, there still remains another possible reason as to why the association of the second GroES is accelerated by denatured proteins. As reported previously [17–19], the ATPase activity of GroEL is enhanced by increasing amounts of denatured protein. This suggests that both rings of GroEL hydrolyse ATP simultaneously. In other words, ATP is able to associate with both rings of GroEL in the presence of denatured protein. Although affinity between ATP and the trans-ring of GroEL is weakened by negative co-operativity between the two rings of GroEL [3,5,39], the amount of the denatured protein present may diminish this negative co-operativity and accelerate the association rate of ATP to the trans-ring of GroEL.
Physiological significance of the formation of the football-shaped complex
Azem et al.  reported that the protein-folding activity of GroEL correlates with the formation of the football-shaped complex. Beissinger et al.  showed the participation of the football-shaped complex in the folding reaction of mutant maltose binding protein, which cannot be folded without the assistance of the GroEL–GroES system at high temperatures . Furthermore, both cavities in the football-shaped complex have been shown to be active in assisting protein folding [31,37]. As indicated in these previous reports [11,14,31,37], the football-shaped complex is considered to be the advantageous form in protein folding. On the basis of these findings, it is expected that the football-shaped complex is active when the amount of denatured protein increases and GroEL prevents denatured proteins from accumulating in living cells (e.g. under stressful conditions). Importantly, we have also found that the ATPase activity of GroEL is higher when the levels of the football-shaped complex increase. GroEL does not have to exhibit maximum ATPase activity when there is a small amount of denatured protein. In other words, GroEL does not form the football-shaped complex when the amount of denatured protein is low in order to prevent futile ATP consumption. Various ATPase proteins are known to regulate their ATPase activity in an environment where they do not have to exert maximum activity: kinesin-1, an intracellular cargo transporter along microtubules, is inhibited from consuming ATP unproductively when not bound with cargo  and the ATPase of myosin V, which has a role in vesicle trafficking along actin filaments in vivo, is activated in the presence of actin filaments . Hence, we propose that the ratio of football- and bullet-shaped complexes is dependent on the concentration of the denatured proteins.
denatured malate dehydrogenase
dithiothreitol, FCS, fluorescence correlation spectroscopy
fluorescence resonance energy transfer
single ring GroEL-E315C mutant
Tomoya Sameshima, Ryo Iizuka and Taro Ueno designed the experiments. The experiments were performed and analysed by Tomoya Sameshima. The manuscript was written by Tomoya Sameshima, Ryo Iizuka and Takashi Funatsu.
The authors thank Dr Naofumi Terada (Bioengineering Laboratory, RIKEN, Saitama, Japan) and Kohki Okabe (Laboratory of Bio-analytical Chemistry, Graduate School of Pharmaceutical Sciences, University of Tokyo, Tokyo, Japan) for their help with the FCS analysis.
This work was partly supported by a Grant-in-Aid for Scientific Research [grant number B 2/370065] and a Grant-in-Aid for Scientific Research under Priority Areas [grant number 20059009] from the Ministry of Education, Science, Sports and Culture of Japan to T.F. T.S. is the recipient of a Research Fellowship from the Japan Society for the Promotion of Science for Young Scientists.