Introduction

Photoinduced electron-transfer (ET) reactions between hemoproteins and small molecules, to transport the electron initiated by the light energy, have received considerable attention in the field of bioinorganic chemistry [1–6]. To date, much research on bimolecular photoinduced ET reactions using metal-substituted hemoproteins, especially for the zinc hemoprotein, has been carried out, because the photoexcited triplet state can act as a strong reductant having a lifetime of several milliseconds [7–11]. A series of organic and inorganic quenchers, such as methylviologen [12, 13], quinones [14], and inorganic complexes [15–17] have been utilized for studying such ET systems. Several reaction mechanisms, including conformational gating [18], have been suggested for zinc cytochrome c [19], zinc hemoglobin [20, 21], and zinc-substituted myoglobin (ZnMb) [5, 6, 12, 13, 15]. As the charge, the conformation, the steric bulk, and other factors of the external quenchers may affect the ET reactions with the zinc hemoprotein, chemical modifications of such conventional redox-active molecules and their suitable application to the hemoprotein surface may be required to conduct selective photoinduced ET reactions [22]. Since an obvious property of the protein surface is chirality, stereoselective bimolecular photoinduced ET with a hemoprotein is now becoming an important research subject. After the first demonstration of the stereoselectivity of the outer-sphere ET reaction of metal complexes [23], the following two main questions concerning the stereoselective ET reaction by optically active compounds should be considered. (1) How does the chiral substitution into small molecules affect the stereoselectivity of ET? and (2) How is the chiral environment around the reactive site at the surface of hemoprotein recognized? Even partially successful ET systems should help to answer these questions, for example, the stereoselective ET reaction of chiral metal complexes with spinach plastocyanin [24, 25], horse cytochrome c [26, 27], and plant ferredoxin [28]. Nevertheless, no experiments on the stereoselectivity regarding the photoinduced ET reactions between a metalloprotein and a chiral organic agent have been conducted so far, except for ours on chiral viologens [29–32]. One of the reasons is the lack of the systematic synthesis of such chiral molecules [33].

We have recently reported the synthesis and photophysical properties of a variety of quinolinium ions, such as 1-[(S)-(1-phenylethyl)carbamoylmethyl]quinolinium hexafluorophosphate [(S)-(PQ)PF6] or 1-[(R)-(1-phenylethyl)carbamoylmethyl]quinolinium hexafluorophosphate [(R)-(PQ)PF6] and 1-[(S)-(1-phenylethyl)carbamoylmethyl]-6-methoxyquinolinium hexafluorophosphate [(S)-(PMQ)PF6] or 1-[(R)-(1-phenylethyl)carbamoylmethyl]-6-methoxyquinolinium hexafluorophosphate [(R)−(PMQ)PF6] (Structure 1) [34]. They certainly acted as electron acceptors for the excited triplet state of ZnMb, 3(ZnMb)*, and demonstrated the bimolecular stereoselective quenching and thermal back ET reactions. Their stereoselectivities evaluated from the ET rate constants, however, are relatively low and this result makes it more difficult to discuss the detailed mechanism of how stereoselectivity was achieved in the bimolecular ET reaction. Therefore, a new design and synthetic strategy of optically active quinolinium derivatives is needed to construct the more apparent stereoselective ET systems.

Structure 1
structure 1

Structures of the optically active electron acceptors presented in this work

Full size image

Cinchonine and cinchonidine, each containing a quinoline moiety within a molecule, are diastereomers of the cinchona alkaloids available in nature. Since each possesses five chiral centers, they have been widely utilized for the enantioselective organic syntheses over the past few decades. For example, their N-alkylated derivatives of N-benzylcinchoninium chloride and O-allyl-N-(9-anthracenylmethyl)-cinchonidinium bromide have been developed and they act as effective catalysts for the asymmetric alkylation reactions during phase-transfer conditions [35–38]. Therefore, we expect that new sets of optically active quinolinium ions can be prepared by the N-substitution of naturally occurring cinchona alkaloids and that their characteristic structures will provide enhanced stereoselectivities in the bimolecular photoinduced ET reactions with zinc hemoprotein for the first time. In this regard, we designed and synthesized the N,N′-dimethylcinchoninium ([MCN]2+) and N,N′-dimethylcinchonidinium ([MCD]2+) salts of which two nitrogen atoms are N-methylated. The redox potentials of these compounds indicate that the cinchona alkaloids undergo effective one-electron redox reactions with photogenerated 3(ZnMb)*. The temperature-dependent kinetics studies performed by flash photolysis are presented and the mechanisms for the stereoselective photoinduced ET reactions are discussed on the basis of crystal structures and activation parameters.

Materials and methods

Materials

Metmyoglobin from horse heart muscle (Sigma) was purified as previously described [39]. The reconstitution of apomyoglobin with protoporphyrinIXato(2-)zinc(II) (Sigma) was carried out at 4 °C in the dark using a previously published method [18]. The concentrations of ZnMb were spectrophotometrically determined (ɛ 428=1.53×105 M−1 cm−1) [40]. The ZnMb solution, whose absorption ratio A 428/A 280 is greater than 9.5, was used for kinetics measurements. [MCN]2+ and [MCD]2+ salts were prepared from cinchonine and cinchonidine (Wako). Methyl iodide was purchased from Tokyo Kasei Kogyo and was used without further purification. The ion-exchange resin, Dowex1-X8 (Cl− form), was obtained from Dow Chemical Co. All other reagents and solvents were of reagent grade. All of the aqueous solutions were prepared from redistilled water. The ionic strength (I) of the solution was adjusted with NaCl.

Preparation of N,N′-dimethylcinchoninium diiodide

N,N′-Dimethylcinchoninium diiodide ([MCN]I2) was prepared from the optically active (11S)-(+)-cinchonine and methyl iodide. Cinchonine (0.79 g, 2.7 mmol) was dissolved in 100 ml of N,N-dimethylformamide at 80 °C. Methyl iodide (1.8 g, 13 mmol) was added to the solution and the mixture was reacted for 48 h at 80 °C. After removal of the solvent, the crude residue was dissolved in 10 ml of methanol and then reprecipitated by adding 100 ml of diethyl ether. The yellow solid was collected by filtration and washed with cold ether. Recrystallization from acetonitrile gave yellow single crystals suitable for X-ray structure analysis. Yield 1.3 g (83%). IR (KBr, ν, cm−1): 1,520 (C=C), 1,450 (C=C). Electrospray ionization (ESI)–mass spectrometry (MS) (CH3OH, m/z): 162.14 ([M−2I−]2+). 1H NMR (270 MHz, CD3OD, 27 °C, δ, ppm): 1.24 (m, 2H, methylene), 2.00 (m, 3H, methylene and –OH), 2.35 (t, 1H, J=11.9 Hz, methine), 2.84 (m, 1H, methine), 3.49 (s, 3H, methyl), 3.60 (d, 2H, J=10.8 Hz, –CH=CH 2), 3.84 (d, 2H, J=8.6 Hz, methylene), 4.44 (m, 1H, methine), 4.71 (s, 3H, methyl), 5.31 [d, 1H, J=8.6 Hz, C(11)–H], 6.05 (m, 1H, −CH=CH2), 8.18 (t, 1H, J=7.8 Hz, Ar–H), 8.33 (t, 1H, J=7.8 Hz, Ar–H), 8.41 (d, 1H, J=6.2 Hz, Ar–H), 8.59 (m, 2H, Ar–H), 9.41 (d, 1H, J=6.2 Hz, Ar–H). UV [CH3CN, λ, nm (ɛ/104 M−1 cm−1)]: 240 (5.15), 320 (0.81). [α]D (MeOH, c=0.99, 25 °C): +123°. Anal. Calcd for C21H28I2N2O: C, 43.62; H, 4.88; N, 4.84%. Found: C, 43.14; H, 4.79; N, 4.62%.

Preparation of N,N′-dimethylcinchonidinium diiodide

N,N′-Dimethylcinchonidinium diiodide ([MCD]I2) was prepared according to methods similar to that of [MCN]I2. Yield 1.3 g (83%). IR (KBr, ν, cm−1): 1,520 (C=C), 1,450 (C=C). ESI-MS (CH3OH, m/z): 162.14 ([M–2I−]2+). 1H NMR (270 MHz, CD3OD, 27 °C, δ, ppm): 1.46 (m, 2H, methylene), 2.10 (m, 4H, methylene, methine, and –OH), 2.86 (m, 1H, methine), 3.49 (s, 3H, methyl), 3.75 (m, 3H, –CH=CH 2 and methylene), 4.36(m, 1H, methine), 4.67 (s, 3H, methyl), 5.13 [m, 1H, C(11)–H], 5.67 (m, 1H, –CH=CH2), 8.15 (t, 1H, J=7.6 Hz, Ar–H), 8.30 (t, 1H, J=7.6 Hz, Ar–H), 8.39 (d, 1H, J=5.9 Hz, Ar–H), 8.53 (d, 2H, J=8.9 Hz, Ar-H), 9.38 (d, 1H, J=5.9 Hz, Ar–H). UV [CH3CN, λ, nm (ɛ/104 M−1 cm−1)]: 240 (5.15), 320 (0.81). [α]D (MeOH, c=0.99, 25 °C): −58°. Anal. Calcd for C21H28I2N2O: C, 43.62; H, 4.88; N, 4.84%. Found: C, 43.25; H, 4.81; N, 4.71%.

To conduct the kinetics measurements in an aqueous solution, the counter anions of [MCN]I2 and [MCD]I2 were converted to the Cl− form. The [MCN]I2 or [MCD]I2 was dissolved in 10 ml of MeOH and column chromatography on Dowex1-X8 (Cl− form, 2.5 cm×20 cm, MeOH) afforded [MCN]Cl2 or [MCD]Cl2, respectively.

Kinetics measurements

The procedure for the kinetics studies was the same as that previously reported [29]. The sample solution was gently purged with Ar gas and then carefully degassed by freeze–pump–thaw cycles. The ratio of A 428/A 280 was checked for each solution. Single flash-photolysis experiments were done in the deaerated solutions containing ZnMb (3.0×10−6 M) and quenchers (0−2.0×10−4 M) at various temperatures (10–30°C), pH 7.0 (0.010 M phosphate buffer), and I=0.020 M using a Photal RA-412 pulse flash apparatus with a 30-μs pulse-width Xe lamp (λ>450 nm; a Toshiba Y-47 glass filter). The spectral absorption changes during the reaction were monitored at 460 nm for the decay of 3(ZnMb)*, and at 680 nm for the formation and decay of the radical cation ZnMb·+.

Other measurements

The IR, UV–vis, optical rotation, and ESI-MS spectra were measured with PerkinElmer 1740 FT-IR, Shimadzu UV-2550, JASCO DIP-140, and Applied Biosystems Mariner instruments, respectively. The 1H NMR spectra were recorded using a JEOL JNM-GX270 FT-NMR spectrometer. Cyclic voltammetry was done in N2-saturated 0.05 M [Bu4N]ClO4 acetonitrile and 0.05 M KCl aqueous solutions using an ALS electrochemical analyzer model 610B instrument. A three-electrode system (BAS) was used with a Pt auxiliary electrode and a Pt working electrode against an Ag/AgClO4 [0.05 M (Bu4N)ClO4 acetonitrile] or an Ag/AgCl (3.33 M NaCl in water) reference electrode. The potentials were calibrated using 1,1′-dimethyl-4,4′-bipyridinium perchlorate [E 0=−0.45 V vs the saturated calomel electrode (SCE)], in acetonitrile. The pHs of the solutions were measured using a Hitachi-Horiba F-14RS pH meter.

Results and discussion

The syntheses of [MCN]I2 and [MCD]I2 were conducted using cinchonine, cinchonidine, and methyl iodide as materials. After recrystallization from acetonitrile, the single crystals of [MCN]I2 and [MCD]I2, very suitable for X-ray crystallography, were obtained in good yields. These were satisfactorily identified by elemental analysis and IR, 1H NMR, UV, ESI-MS, and optical rotation spectroscopies (see “Materials and methods” section). The crystal structures of [MCN]I2 and [MCD]I2 consist of a discrete C21H28N2O2+ cation and two iodide anions. In advance, their crystal data and the data statistics have been deposited as structure reports [41, 42]. The notation refers to the numbering assignments of [MCN]I2 and [MCD]I2 as shown in Structure 1.

In order to evaluate the redox properties of [MCN]I2 and [MCD]I2, cyclic voltammograms were recorded in a 0.05 M [Bu4N]ClO4 acetonitrile solution at 25 °C under an N2 atmosphere with a Pt electrode as the working electrode. It is well known that 1-methyl-quinolinium perchlorate has an one-electron redox potential at −0.90 V vs SCE in acetonitrile [43, 44]. We have reported the optically active [PQ]PF6 and [PMQ]PF6 and showed similar irreversible one-electron redox waves at −0.85 V vs SCE. These agents acted as electron acceptors for the photoinduced ET reactions with the excited 3(ZnMb)* [34]. For the optically active [MCN]2+ and [MCD]2+, the irreversible one-electron redox waves both appeared at −0.79 V ([MCN]2+/[MCN]·+ and [MCD]2+/[MCD]·+; see Fig. S1 in the supplementary material). The slightly higher redox potential compared with that of [PMQ]+ and [PQ]+ is probably due to the net charge of the [MCN]2+ or [MCD]2+ ions that increases the electron affinity. We noted that similar electrochemical behavior, such as the irreversible one-electron redox wave at −0.79 V ([MCN]2+/[MCN]·+ and [MCD]2+/[MCD]·+), was observed using the Cl− forms of [MCN]2+ and [MCD]2+ in an aqueous solution with reference to the normal hydrogen electrode. On the basis of these measurements, the new optically active cinchoninium and cinchonidinium ions, [MCN]2+ and [MCD]2+, can be expected to perform as good electron acceptors for the excited triplet state of ZnMb similar to the [PMQ]+ and [PQ]+ as listed in Structure 1 [34].

Photoinduced ET reaction of ZnMb

The photoexcited 3(ZnMb)* is generally known to be quenched by several electron acceptors, whereas the excited singlet state is not quenched [6]. Such a quenching process has been discussed for the intermolecular ET mechanism. Figure 1a displays the first-order decay of the transient absorption of 3(ZnMb)* monitored at 460 nm in the presence of [MCN]2+ at 20 °C, pH 7.0 (0.010 M phosphate buffer), and I=0.020 M. The rate constant for the quenching reaction of 3(ZnMb)*, k obsd, linearly increased with increasing concentration of [MCN]2+ or [MCD]2+ at various temperatures (Fig. 2), indicating that no appreciable complex was formed between 3(ZnMb)* and [MCN]2+ or [MCD]2+. We also monitored the kinetics trace at 680 nm for the formation and decay of the radical cation of ZnMb·+ [29, 45]. The transient absorption kinetics of the thermal back ET in Fig. 1b obeyed a second-order rate law, suggesting that an equimolar amount of ZnMb·+ with an [MCN]·+ or [MCD]·+ radical cation formed.

Fig. 1
figure 1

Absorbance changes after irradiation of zinc-substituted myoglobin (ZnMb) with a Xe flash lamp in the presence of N,N′-dimethylcinchoninium ion ([MCN] 2+) (1.5×10−4 M) at 20 °C, pH 7.0 (0.010 M phosphate buffer), and I=0.020 M. a Decay of 3(ZnMb)* at 460 nm. The dotted line is fitted to the first-order decay kinetics. b Decay of ZnMb·+ at 680 nm. The dotted line is fitted to Eq. 1

Full size image
Fig. 2
figure 2

k obsd versus initial quencher concentration for the quenching of 3(ZnMb)* by [MCN]2+ (open symbols) and N,N′-dimethylcinchonidinium ion ([MCD] 2+) (closed symbols) at pH 7.0 (0.010 M phosphate buffer) and I=0.020 M: 10 °C (up triangles); 15 °C (squares); 20 °C (diamonds); 25 °C (circles); 30 °C (down triangles)

Full size image

The photoinduced ET reaction of ZnMb with [MCN]2+ or [MCD]2+ is represented in Scheme 1. The quenching ET process in Scheme 1 contains the elementary steps, including the pre-equilibrium of binding of [MCN]2+ (or [MCD]2+) with the excited triplet state of 3(ZnMb)*, where the quenching and the ET rate constants are represented as k q and k et, respectively. When the binding between 3(ZnMb)* and [MCN]2+ is weak, the quenching rate constant, k q, corresponds to k q k 1/k −1 [46]. The quenching rate constant (k q) was thus obtained from the slope of the plots of k obsd versus the concentration of [MCN]2+ or [MCD]2+. As the thermal ET reaction to the ground state was much slower than the quenching reaction (Fig. 1b) and the quantum yield of the quenching ET reaction would be nearly unity (see next section), the second-order rate constant of the thermal ET reaction (k b) was evaluated during the latter portion of the decay of ZnMb·+ at 680 nm after the quenching of 3(ZnMb)* was completed (Eq. 1):

$$A_{t} =\frac {A_0 + k_{\text{b}} [A]_{\text{0}} A_\infty t}{1 + k_{\text{b}} [A]_0 t}.$$
(1)

Here A 0, A t , and A ∞ are the absorbances at time 0, t, and infinity, respectively, and [A]0 is the initial concentration of ZnMb·+. Three unknown parameters, A 0, k[A]0, and A ∞ were simultaneously estimated. The value of k b was then determined using the value of [A]0 calculated from the concentration of 3(ZnMb)* [Δɛ 428=ɛ(ground)−ɛ(triplet)=1.00×105 M−1 cm−1] [7].

Scheme 1
scheme 1

Photoinduced electron-transfer reaction between zinc-substituted myoglobin (ZnMb) and N,N′-dimethylcinchoninium ion ([MCN] 2+)

Full size image

Stereoselective ET quenching reactions at 25 °C

The ET quenching reactions of 3(ZnMb)* by [MCN]2+ and [MCD]2+ are slightly exothermic, ΔG 0=0.01 eV, based on the redox potentials. Under the present experimental conditions, in which the [MCN]2+ and [MCD]2+ ions are used in large excess over 3(ZnMb)*, with an initial concentration of 3(ZnMb)* of 1.8×10−6  M, the ET quenching was almost completed: mote than 98% for [MCN]2+ or [MCD]2+ and more than 98% for [PQ]2+. The rate constants (k q) for the ET quenching reactions of [MCN]2+ and [MCD]2+ at 25 °C were determined as k q(MCN)=(1.9±0.1)×106  M−1 s−1 and k q(MCD)=(3.0±0.2)×106  M−1 s−1 (Table 1). Previous results for the quenching reaction of 3(ZnMb)* by the optically active quinolinium ions, [PQ]+ and [PMQ]+ (Structure 1), are also summarized in Table 1 and the ratios of k q(S)/k q(R) appeared in the range from 1.3 to 1.4 at 25 °C. During the quenching reaction of [MCN]2+ and [MCD]2+, the ratio of the rate constants was found to be k q(MCD)/k q(MCN)=1.6 at 25 °C.

Table 1 Rate constants determined for the electron-transfer (ET) reactions between zinc-substituted myoglobin (ZnMb) and optically active N,N′-dimethylcinchoninium ([MCN] 2+), N,N′-dimethylcinchonidinium ([MCD] 2+), 1-[(1-phenylethyl)carbamoylmethyl]quinolinium ([PQ] +), and 1-[(1-phenylethyl)carbamoylmethyl]-6-methoxyquinolinium ([PMQ] +) ions at pH 7.0 (0.010 M phosphate buffer) and I=0.020 M
Full size table

In the case of [PQ]+ and [PMQ]+, we have proposed that the quenching reactions were controlled by an ET step, not by the conformational gating of ZnMb, according to the driving-force dependence of the ET rate constants. This may partly arise because the driving force of the reaction is slightly positive, 0.05 and 0.08 eV for [PQ]+ and [PMQ]+, respectively, and the ET rate is not fast enough to be controlled by the conformational change in myoglobin. In order to evaluate the present quenching mechanism between 3(ZnMb)* and [MCN]2+ (or [MCD]2+), we calculated the rate constants using the following Marcus theory (Eq. 2) [47, 48]:

$$k_{12} = (k_{11} k_{22} f_{12} K_{12})^\frac{1}{2},$$
(2)

where k 12 is the rate constant for the cross reaction, k 11 and k 22 are those for the self-exchange reactions of the donor and acceptor, respectively, K 12 is the equilibrium constant for the cross reaction, and f 12 is given by

$$\hbox{ln}\,f_{12} = \frac{(\hbox{ln}\,K_{12})^2}{4\hbox{ln} (k_{11} k_{22}/10^{22})}.$$
(3)

Equations 2 and 3 are also represented by

$$\hbox{ln}\,k_{12} = 25.3 - \frac{(\lambda _{12} + \Delta G^0)^2}{4\lambda _{12} RT},$$
(4)

where λ 12 is the reorganization energy for the reaction, and equals the average of those of the donor and acceptor, (λ 11+λ 12)/2. By using the data k 11=2.6×105 M−1 s−1 for ZnMb·+/3(ZnMb)* (λ 11=1.32 eV) [49], k 22=1.0×108 M−1 s−1 for both [MCN]2+/[MCN]·+ couples (λ 22=0.71 eV) [34], and K 12=1.5 for [MCN]2+ (the redox potential of ZnMb·+/ZnMb is 0.98 V and ΔG 0=−0.01 eV) [50], the calculated rate constants, k 12, are 6.2×106 M−1 s−1 (f 12=0.99) for [MCN]2+. The calculated rate constant is close to the observed ones (Table 1, Fig. 3, open triangles). On the basis of these results we can assume that the present quenching reactions of 3(ZnMb)* by [MCN]2+ and [MCD]2+ are controlled by the ET step, as can be similarly expected for the [PQ]+ and [PMQ]+ ions.

Fig. 3
figure 3

Driving-force dependence of the rate constants for the quenching and thermal back electron-transfer (ET) reactions of ZnMb with quinolinium ions: 1-[(S)-(1-phenylethyl)carbamoylmethyl]-6-methoxyquinolinium (open circles); 1-[(R)-(1-phenylethyl)carbamoylmethyl]-6-methoxyquinolinium (closed circles); 1-[(S)-(1-phenylethyl)carbamoylmethyl]quinolinium(open squares); 1-[(R)-(1-phenylethyl)carbamoylmethyl]quinolinium (closed squares); [MCN]2+ (open triangles); [MCD]2+ (closed triangles). The solid curve is calculated by Eq. 4 with λ 12=1.02 eV

Full size image

Figure 4 illustrates the plausible structure of the ZnMb/[MCN]2+ complex. The [MCN]2+ ion was displayed by RasMol software according to the crystallographic data [41]. A comparison of the actual size of [MCN]2+ with that of the surface of ZnMb indicates that the [MCN]2+ ion (or [PQ]+ and [PMQ]+ ions) can sufficiently cover the heme pocket of ZnMb. A cationic [MCN]2+, including a bulky quinuclidinium moiety, may not diffuse into the heme pocket but loosely associates with the amino acid residues at the surface of ZnMb. We have recently reported the stereoselective thermal ET reactions between metmyoglobin and optically active viologen radicals [31]. In such a system, metmyoglobin moderately interacted with the viologen-radical cation to form a complex, followed by an intramolecular ET reaction. The calculated intramolecular ET rate constant and the association constant indicate that there is stereoselectivity in both the chiral association and intramolecular ET processes. Although both of these constants could not be clarified for the present [MCN]2+ (or [MCD]2+) system under our experimental conditions, they may also be important for explaining the difference in the ET rate caused by the chirality. However, we note that the quenching rate constants (k q) for the [MCN]2+ (or [MCD]2+) system were higher than those for the previously designed [PQ]+ and [PMQ]+ and the ET stereoselectivities for [MCD]2+ were definitely enhanced as listed in Table 1. Five chiral centers at the C11, C12, C14, C19, and N2 positions are present both in the [MCN]2+ (11S,12R) and in the [MCD]2+ (11R,12S) ions, whereas only one chiral center is introduced for the [PQ]+ and [PMQ]+ ions. These chiral centers can be appropriately recognized by the chiral surface of ZnMb, and the large conformational difference between [MCN]2+ and [MCD]2+ would realize the enhanced stereoselective quenching reaction.

Fig. 4
figure 4

The plausible structure of a ZnMb/[MCN]2+ complex by RasMol (version 2.6). The molecular conformation of [MCN]2+ is displayed according to its crystallographic data

Full size image

Stereoselectivity in thermal ET to the ground state

The thermal back ET reaction between ZnMb·+ and the cinchoninium and cinchonidinium radical cations, [MCN]·+ and [MCD]·+, also has a stereoselectivity for [MCD]·+. The second-order rate constants (k b) for the thermal back ET at 25 °C are shown in Table 1 and were determined as k b(MCN)=(0.79±0.04)×108  M−1 s−1 and k b(MCD)=(1.0±0.1)×108  M−1 s−1. In this study, we similarly obtained the calculated rate constant for the thermal back ET reaction, k 12=4.1×108 M−1 s−1 (f 12=6.8×10−27) for [MCN]·+ using the data k 11=2.6×105 M−1 s−1 for ZnMb·+/ZnMb (λ 11=1.32 eV), and K 12=9.5×1029 for [MCN]·+ (ΔG 0=−1.77 eV). This is also close to the observed rate constant (see Table 1, Fig. 3), indicating that the thermal back ET reaction is similarly controlled by the ET step. The back ET selectivity for [MCN]·+ and [MCD]·+ at 25 °C was found to be k b(MCD)/k b(MCN)=1.3. However, the highest stereoselectivity of 1.7 was observed at low temperature (10 °C). Since the one-electron reduced radicals, [MCN]·+ and [MCD]·+, generated from quenching of the photoexcited 3(ZnMb)* are less positive, we suggest that the charge effect on the quinolinium ion moiety is as important as the conformational effect to explain such a temperature-dependent selectivity in the thermal back ET. In the following section, a further discussion based on the activation parameters will be described to clarify the mechanisms of both ET reactions.

Activation parameters

The second-order rate constants of k et and k b at various temperatures (10–30 °C) provide the activation parameters for the quenching and thermal back ET reactions (see Tables S1, S2 in the supplementary material). Figure 5 displays the Eyring’s plots produced by Eq. 5:

$$\hbox{ln}\,\frac{k}{T} = \frac{- \Delta H^{\ne}}{T} + \frac{\Delta S^{\ne}}{R} + \hbox{ln}\, {k_{\text{B}}}{h}.$$
(5)

Here, T, R, k B, and h are the absolute temperature, the gas constant, the Boltzmann constant, and Planck’s constant, respectively. We estimated the activation enthalpy, ΔH ≠, from the slope and the activation entropy, ΔS ≠, from the intercept of the linear plots of lnk/T versus T −1 that were fitted well to the experimental points. For the quenching reaction, the activation parameters ΔH ≠=21±1 kJ mol−1 and ΔS ≠=−54±4 J mol−1 K−1 for [MCN]2+ and ΔH ≠=20±1 kJ mol−1 and ΔS ≠= −54±4 J mol−1 K−1 for [MCD]2+ were obtained (Table 2). In the case of the thermal back ET reaction, ΔH ≠=17±1 kJ mol−1 and ΔS ≠=−37±3 J mol−1 K−1 for [MCN]·+ and ΔH ≠=6.3±0.5 kJ mol−1 and ΔS ≠=−70±5 J mol−1 K−1 for [MCD]·+ were also determined from Fig. 5.

Fig. 5
figure 5

Eyring’s plots for the quenching and thermal back ET reaction at pH 7.0 (0.010 M phosphate buffer) and I=0.020 M. a Quenching reaction of 3(ZnMb)* by [MCN]2+ (open symbols) and [MCD]2+ (closed symbols). b Thermal ET reaction of ZnMb·+ by [MCN]·+ (open symbols) and [MCD]·+ (closed symbols)

Full size image

To evaluate these activation parameters, the energy differences of ΔΔH ≠(MCD-MCN) and TΔΔS ≠(MCD-MCN) at 25 °C for the quenching and thermal back ET reactions were calculated and are listed in Table 2. The slightly negative ΔΔH ≠(MCD-MCN)=−1.1±2 and TΔΔS ≠ (MCD-MCN)=0 kJ mol−1 appeared in the quenching reaction, indicating that the stereoselective quenching reaction was predominated by the enthalpy term. On the other hand, ΔΔH ≠ (MCD-MCN)=11±2 kJ mol−1 and TΔΔS ≠(MCD-MCN)=−10±2 kJ mol−1 (ΔΔS ≠=−34±8 J mol−1 K−1) were obtained at 25 °C in the thermal back ET reaction. These parameters clearly supported the temperature dependence of the stereoselective back ET reactions, while ΔΔS ≠(MCD-MCN)=0 J mol−1 K−1 for the quenching reactions maintained the selectivity constant from 10 to 30 °C. Although these ΔΔH ≠ and ΔΔS ≠ values are very different from those of the quenching ones, appreciate noncovalent interactions between [MCN]·+ (or [MCD]·+) and the amino acid residues on the ZnMb·+ surface would be present during the thermal back ET reaction. As typical examples of the stereoselective ET reactions, we can use the ET reactions of cytochrome c(III) and metmyoglobin with the chiral viologen cation radicals, possessing a chiral center of the (1-phenyl)carbamoylmethyl or (1-cyclohexyl)carbamoylmethyl group, where the enthalpic parameters are predominant [31, 51]. Here, we would consider the energy differences for the quenching and thermal back ET reactions of ZnMb with [MCN]2+ or [MCD]2+ as follows: the lower ΔH ≠ value for [MCD]2+ and [MCD]·+ allows the quenching and the thermal back ET reactions to be slightly faster than for [MCN]2+ and [MCN]·+, respectively (enthalpy dominating) [32], and the highest stereoselectivity of 1.7 for [MCD]·+ found at low temperature (10 °C) was due to the ΔΔS ≠ value obtained in the thermal back ET reaction. These differences between the [MCN]2+ and [MCD]2+ ions may mainly arise from their special conformations. We conclude that a chiral center at the C11 position and a bulky quinuclidine moiety in the designed [MCN]2+ and [MCD]2+ ions are the significant structural factors resulting in remarkable stereoselective photoinduced ET reactions with ZnMb, as compared with those of the [PQ]+ and [PMQ]+ ions, where each chiral center is far from the quinolinium moiety.

Table 2 Activation parameters for the quenching reaction of 3(ZnMb)* by [MCN]2+ and [MCD]2+ and the thermal ET reaction between ZnMb+ and [MCN]·+ and [MCD]·+ at pH 7.0 (0.010 M phosphate buffer) and I=0.020 M
Full size table

Conclusions

The present studies demonstrate the driving-force-dependent photoinduced ET reactions between the optically active [MCN]2+ and [MCD]2+ and ZnMb with apparent stereoselectivities. Kinetics studies obviously showed the increase in the rate constant and the stereoselectivity for the [MCD]2+ ion both in the quenching and in the thermal back ET reactions, compared with those of the [PQ]+ and [PMQ]+ systems. These results clearly imply that the conformational difference between the [MCN]2+ and [MCD]2+ ions seems to be very important to enhance the stereoselectivity. We believe that further synthetic applications at the N1 and N2 positions of the several cinchona alkaloids will produce diverse sets of chiral organic electron carriers. Investigations of the fundamental photoinduced ET reactions with hemoproteins may provide valuable information for elucidating the complicated mechanisms of biological stereoselective ET.