Protein photoconductors and photodiodes.

Y. Tokita,S. Yamada,W. Luo,Y. Goto,N. Bouley-Ford,H. Nakajima,Y. Watanabe

DOI: https://doi.org/10.1002/anie.201103341

2011-12-02

Angewandte Chemie

Abstract:Long-range electron transfer (ET) through proteins plays important roles in living systems. ET proteins can be regarded as wide-band-gap molecular semiconductors with “dopants” (redox or photoactive species) that provide an electronor hole-localizing site, though the protein scaffold has insulating character. Intramolecular ET in proteins has been investigated in detail by many groups to elucidate the factors that control the rates of these nonadiabatic reactions. 11–13] Dutton and co-workers proposed a square barrier ET rate model, while Gray and colleagues showed that ET depends on the structure of the medium between an electron donor and acceptor. We have investigated photoinduced ET in two proteins, zinc-substituted cytochrome b562 (Zn-cyt b562) and zinc-substituted cytochrome c (Zn-cyt c), the semiconductor properties of which depend on the charge distribution on their molecular surfaces: Zn-cyt b562 has n-type semiconductor character while Zn-cyt c is p-type, although the active center of the two systems is virtually identical. This finding may open up the world of protein-based electronics, because the semiconductor character of proteins could be controlled by variations in surface charge. Zn-cyt b562 on the H2N-SAM/Au electrode (SAM = selfassembled monolayer) was prepared according to the literature (see the Supporting Information). The orientation of Zn-cyt b562 on the H2N-SAM/Au surface (Figure 1) is optimal for favorable heme-region (acidic patch) electrostatic interactions with the SAM positive charge. Figure 2a shows cathodic photocurrents in response to switching on and off of 420 nm illumination for bias voltages of + 300, 0, and 300 mV (vs. Ag jAgCl). The action spectrum (Figure 2b) is similar to the absorption spectrum of Zn-cyt b562 (Figure 2b, inset), which suggests that the photocurrent originates in the protein. Incidentally, there is no photocurrent response in the case that protein changes Zncytb562 into Fe-cytb562. Notably, the Q-band (550 nm) photocurrent response is more prominent in comparison with the absorption spectrum, that is, the Soret/Q ratio in absorbance (I420/I550) of the action spectrum is 3.7 whereas that of the UV/ Vis spectrum is 11.1, thus implying that the photocurrent is derived from migration of the photoexcited electron. These findings are consistent with this mechanism: 1) the photocurrent was observed only in the cathodic direction under the range of bias potential employed here (Figure 2c), and 2) addition of methyl viologen (E0 = 0.62 V vs. Ag j AgCl) to the solution as a redox reagent enhanced the cathodic photocurrent, while addition of potassium ferri/ ferrocyanide (E0 =+ 0.17 V vs. Ag jAgCl) did not show such an effect. 18] This is quite in contrast to the Zn-cyt c/HOOC-SAM/Au system, where the photocurrent action spectrum can be superimposed on the UV/Vis spectrum and the addition of the ferri/ferrocyanide effectively increases the photocurrent that derives from the hole in the occupied molecular orbitals (MOs), or the valence band. Thus, although Zn-cyt b562 and Zn-cyt c are derivatives from similar ET proteins, the systems have a different electronic nature: a photodiode (n-type) and a photoconductor (p-type) character, respectively (Figure 2c). To elucidate the electronic structure and the factors that control the semiconductor properties of the two Zn-cytochrome proteins, we carried out all-electron DFT calculations for Zn-cyt b562. We estimated the photoinduced ET rate constants for Zn-cyt c [Eq. (1)]: