Chinese Journal of Chemical Physics  2018, Vol. 31 Issue (4): 523-528

#### The article information

Junjun Tan, Chuanzhao Li, Jiahui Zhang, Shuji Ye

Real-Time Observation of Protein Transport across Membranes by Femtosecond Sum Frequency Generation Vibrational Spectroscopy

Chinese Journal of Chemical Physics, 2018, 31(4): 523-528

http://dx.doi.org/10.1063/1674-0068/31/cjcp1805128

### Article history

Received on: May 31, 2018
Accepted on: July 3, 2018
Real-Time Observation of Protein Transport across Membranes by Femtosecond Sum Frequency Generation Vibrational Spectroscopy
Junjun Tana,b, Chuanzhao Lia,b, Jiahui Zhanga,b, Shuji Yea,b
Dated: Received on May 31, 2018; Accepted on July 3, 2018
a. Hefei National Laboratory for Physical Sciences at the Microscale, and Department of Chemical Physics, University of Science and Technology of China, Hefei 230026, China;
b. Synergetic Innovation Center of Quantum Information & Quantum Physics, University of Science and Technology of China, Hefei 230026, China
*Author to whom correspondence should be addressed. Shuji Ye, E-mail: shujiye@ustc.edu.cn
Abstract: Characterization of conformation kinetics of proteins at the interfaces is crucial for understanding the biomolecular functions and the mechanisms of interfacial biological action. But it requires to capture the dynamic structures of proteins at the interfaces with sufficient structural and temporal resolutions. Here, we demonstrate that a femtosecond sum frequency generation vibrational spectroscopy (SFG-VS) system developed by our group provides a powerful tool for monitoring the real-time peptide transport across the membranes with time resolution of less than one second. By probing the real-time SFG signals in the amide Ⅰ and amide A bands as WALP23 interacts with DMPG lipid bilayer, it is found that WALP23 is initially absorbed at the gel-phase DMPG bilayer with a random coil structure. The absorption of WALP23 on the surface leads to the surface charge reversal and thus changes the orientation of membrane-bound water. As the DMPG bilayer changes from gel phase into fluid phase, WALP23 inserts into the fluid-phase bilayer with its N-terminal end moving across the membrane, which causes the membrane dehydration and the transition of WALP23 conformation from random coil to mixed helix/loop structure and then to pure α-helical structure. The established system is ready to be employed in characterizing other interfacial fast processes, which will be certainly helpful for providing a clear physical picture of the interfacial phenomena.
Key words: Femtosecond sum frequency generation    Peptide transport across membrane    Real time    Kinetic conformation
Ⅰ. INTRODUCTION

Characterization of conformation kinetics of proteins and peptides at the surfaces and interfaces is crucial for understanding the properties and functions of such biological molecules, as well as the mechanisms of many biological processes at membrane surfaces such as signal transduction, cell adhesion, and antimicrobial selectivity [1]. Particularly, detailed comprehension of the kinetic processes of the protein transport across membranes is important to get insights into the mechanisms of cell transport, membrane protein folding, and protein action [2, 3]. Sum frequency generation vibrational spectroscopy (SFG-VS) has been demonstrated to be a powerful technique for investigating the interfacial structures and interactions at the molecular level [4-26]. In previous studies, frequency-scanning SFG system has been effectively employed to investigate the static structure of proteins and peptides at different interfaces [4, 5, 7, 8, 25]. Despite many progresses, frequency-scanning SFG system takes more than ten minutes to acquire one spectrum. Therefore, it fails to probe the intermediates during the protein transport across membranes due to its low temporal resolution. To access a fast dynamic process, it requires to capture the dynamic structures of proteins at the membrane surface with sufficient structural and temporal resolutions. To solve this problem, we have recently succeeded in developing a highly sensitive femtosecond time-resolved SFG-VS with simultaneous measurement of multiple polarization combinations [26, 27].

This SFG system is capable of acquiring one spectrum with the recording time of less than one second. Such fast recording time can effectively prevent the structural changes during the spectral acquisition, which makes it become a powerful tool to extract the specific structural details about intermediates with both high structural and fast temporal resolution. In this study, we applied this system and systematically investigated the conformation kinetics of transmembrane peptide WALP23 during its transport across the lipid bilayer. WALP peptide is chosen because it provides a 'benchmark' model for complex membrane proteins. It forms continuous $\alpha$-helices in the lipid bilayer and has been widely used for investigating the insertion of peptide into lipid bilayer by simulations and experiments [28-30]. The real-time spectra of C=O (amide Ⅰ) and N-H (amide A) stretch of the peptide bond were used to monitor the conformational changes of WALP23 in lipid bilayer. Time-dependent SFG changes provide a clear physical picture for the dynamic structures and insert direction of WALP23.

Ⅱ. EXPERIMENTAL SECTION A. Materials and sample preparations

1, 2-Ditetradecanoyl-sn-glycero-3-phospho-(1$^{\prime}$-rac-gl- ycerol) (sodium salt) (DMPG) lipid was purchased from Avanti Polar Lipids (Alabaster, AL). WALP23 (sequence: GWW(LA)$_{8}$LWWA) with a purity of $\geq$98% was purchased from Shanghai Apeptide Co., Ltd. Deionized water with resistivity of 18.2 M$\Omega$$\cdotcm was produced by a Milli-Q reference system (Millipore, Bedford, MA). WALP23 was dissolved in methanol (purchased from Sinopharm Chemical Reagent Co., Ltd.) with a concentration of 2 mg/mL. The DMPG solution was prepared in the mixed solvents of chloroform and methanol (with a volume ratio of 2:1, purchased from Sinopharm Chemical Reagent Co., Ltd.) with the concentration of 2.0 mg/mL. The lipid and WALP23 solutions were kept at -20 ℃. Right-angle CaF_{2} prisms were purchased from Chengdu Ya Si Optoelectronics Co., Ltd. (Chengdu, China). All of the chemicals were used as received. A water bath system was used to control the sample temperature (FIG. 1). The prism-cleaning and lipid monolayer/bilayer preparations were performed using a standard procedure given previously [31-34].  FIG. 1 A schematic of the water bath system and the near-total-internal-reflection geometry. B. Femtosecond sum frequency generation (FS-SFG) system SFG is a second-order optical laser technique that permits to determine the molecular species (or chemical groups) and interactions at surfaces and interfaces [4-26]. Recently, FS-SFG has been used to probe proteins and other bioactive molecules dynamics at the interfaces. For example, Yan et al. studied the dynamics of hIAPP misfolding in the DPPG monolayer and LK_{7}$$\beta$ self-assembling at the air/water interfaces [1, 35]. However, it still takes more than 1 min to acquire one SFG spectrum for previously reported FS-SFG system. In our system, we employed several key technical improvements, including adoption of a near-total-internal-reflection geometry (FIG. 1) and employment of two Glan-Laser polarizers (with an extinction ratio (T$_{\textrm{p}}$/T$_{\textrm{s}}$) of $>$10$^{5}$:1) to separate the polarization components. With these technical improvements [26, 27], we can acquire one SFG spectrum in less than one second and achieve the fastest speed of collecting SFG spectra reported so far [27]. The details of the system have been introduced in our recent publications [26, 27]. The energy profiles of the IR pulses were used to normalize the SFG spectra by measuring the SFG signals from the gold surface coated at the prism. All SFG experiments were carried out at room temperature (24 ℃). IR beam was protected by a home-built chamber purged with dry gas (dry gas generator, Peak Scientific) to avoid the IR energy loss due to water vapor absorptions.

Ⅲ. RESULTS AND DISCUSSION A. The conformational changes during the WALP23 transport across the lipid bilayer

We first investigated conformational changes during the WALP23 transport across the DMPG lipid bilayer by probing the amide Ⅰ band. The amide Ⅰ band of proteins arises mainly from the C=O stretching vibration with minor contributions from the out-of-phase C-N stretching vibration. Its frequency is sensitive to the conformation of the peptide backbone [36, 37]. FIG. 2 shows the real-time ppp spectra of DMPG bilayer interface in the amide Ⅰ band region before and after injecting 5 μL WALP23 (with concentration of mg/mL) in the subphase (2 mL) of DMPG lipid bilayer at 15 ℃. At this temperature, the DMPG/DMPG lipid bilayer is in the gel phase. At t$<$0 s, the spectra show a broad peak, which arises from the membrane-bound water. Recent studies have suggested that the membrane-bound water at a charged surface is extraordinarily broad and extends to 1500 cm$^{-1}$ [31, 38]. After injecting WALP23 at t=0 s, the signal from the interfacial water starts to decrease at 15 s and reaches its minimum at t=20 s. After that, two peaks at $\sim$1635 and 1735 cm$^{-1}$ were observed. The intensity of these two peaks increases from t=30 s to 150 s and then reaches a plateau at t$\geq$150 s. The $\sim$1635 cm$^{-1}$ peak is indicative of random coil structure of WALP23, while the 1735 cm$^{-1}$ peak originates from the lipid carbonyl groups of DMPG [25, 36, 37]. Earlier studies have proved that WALP23 molecules lie down on the gel-phase bilayer surface, instead of inserting into the lipid bilayer, when they interact with a gel-phase lipid bilayer at room temperature [26, 30, 39]. Therefore, WALP23 adopts a random coil structure at the surface of lipid bilayer.

 FIG. 2 (a) The real-time ppp spectra of DMPG bilayer interface in the amide Ⅰ band region before and after injecting 5 μL WALP23 in the subphase of DMPG lipid bilayer at 15 ℃, (b) the ppp spectra of WALP23 in amide Ⅰ region at different time interval.

After the interaction reached equilibrium, we heated the sample to the temperature of 35 ℃ using the water bath (FIG. 1). Because the phase transition temperature of DMPG is 24 ℃, the heating can cause the transition from gel-phase state into fluid-phase state of DMPG bilayer, thus promoting the insertion of WALP23 into fluid-phase DMPG bilayer [30]. FIG. 3 shows the real-time ppp SFG spectra in mide Ⅰ region when the temperature of the subphase of lipid bilayer was changed from 15 ℃ to 35 ℃. Following the heating, the intensity of the $\sim$1635 cm$^{-1}$ peak decreases and completely disappears at t$>$650 s (FIG. 3(b)). A new peak is observed at $\sim$1660 cm$^{-1}$ at t$>$200 s, which is assigned to $\alpha$-helical or loop structure [25, 26]. The $\sim$1660 cm$^{-1}$ peak gradually increases and reaches a plateau at t$>$1700 s. The appearance of the $\sim$1660 cm$^{-1}$ peak and weakening of the $\sim$1635 cm$^{-1}$ peak indicates that WALP23 undergoes a transition from random coil structure into loop and $\alpha$-helical structure as it inserts into DMPG bilayer.

 FIG. 3 The real-time ppp spectra of WALP23 at DMPG bilayer interface with the subphase temperature changes from 15 ℃ to 35 ℃: (a) 0-2000 s and (b) 200-600 s, (c) the ppp spectra of WALP23 in amide Ⅰ region at different time interval.
B. The change in membrane-bound water and amide A band during the WALP23 transport across the lipid bilayer

Following similar procedures used in the study of amide Ⅰ band, we investigated the spectra in the frequency range from 3100 cm$^{-1}$ to 3500 cm$^{-1}$. The signals in this region are contributed by the membrane-bound water or N-H stretching of amide A band. FIG. 4 shows the real-time ppp spectra of DMPG bilayer interface in the frequency ranging from 3100 cm$^{-1}$ to 3500 cm$^{-1}$ before and after injecting 5 μL WALP23 in the subphase of DMPG lipid bilayer at 15 ℃. At t$<$0 s, the spectra are dominated by a strong and broad peak. The broad peak arises from membrane-bound water. As WALP23 is injected into the subphase of DMPG bilayer, the intensity of the broad peak decreases and reaches its minimum at t=50 s, and then increases again. The intensity finally becomes stable at t$>$90 s. The SFG spectra of H$_{2}$O molecules at the lipid/water interface have been studied previously [40-47]. In our system, the SFG signal of the interfacial water is dominated by the water molecules at the interface between the outer bilayer leaflet and the bulk water, rather than the confined water between the lipid bilayer and CaF$_{2}$ prism surface [48-50]. Earlier phase-sensitive SFG study indicated that membrane-bound water molecules are oriented preferentially by the electrostatic potential imposed by the phospholipids [40, 41, 43-45, 47]. The O-H stretch of water near negative DMPG lipid adopts a hydrogen-up conformation [51]. After adding WALP23 into the subphase of DMPG bilayer, WALP23 is absorbed on the membrane surface due to coulombic interaction. The absorption of positively charged WALP23 on the membrane surface makes the surface charge become more and more neutral, and finally positive. Such surface charge reversal behavior leads to SFG intensity change observed in FIG. 4(a) [48]. After the absorption reached equilibrium, the interfacial water molecules at the positive surface adopt hydrogen-down conformation.

 FIG. 4 (a) The real-time ppp spectra of DMPG bilayer interface in the amide A band region before and after injecting 5 μL WALP23 in the subphase of DMPG lipid bilayer at 15 ℃, (b) the ppp spectra of WALP23 in amide A region at different time interval.

To get insights into the change in interfacial O-H and N-H stretching during WALP23 transport across the lipid bilayer, we heated the sample to 35 ℃. FIG. 5 shows the real-time ppp spectra of WALP23-bound DMPG bilayer after heating. As WALP23 inserts into lipid bilayer, the intensity of the broad peak from the membrane-bound water decreases due to the dehydration interaction. Besides this, a transient peak centered at $\sim$3300 cm$^{-1}$ starts to appear at t$>$200 s and disappears at 1500 s. The $\sim$3300 cm$^{-1}$ peak originates from N-H group of the peptide band and has an opposite phase relative to the interfacial water (with a hydrogen-down conformation), resulting in a "pit" shown in FIG. 5(a). It means that the N-H group in the peptide band adopts a hydrogen-up conformation, illustrating that WALP23 inserts with its N-terminal end moving across the membrane, which is consistent with previous molecular dynamics simulation [28]. It is worth noticing that at t$<$600 s, SFG intensity at the left and the right sides of the $\sim$3300 cm$^{-1}$ peak is almost the same. However, at t$>$600 s, the intensity at the right side starts to be higher than that at the left side because one new peak at $\sim$3328 cm$^{-1}$ is generated. The peak at $\sim$3328 cm$^{-1}$ becomes more and more intense and finally dominates the spectrum. It is known that the frequency of the N-H stretching modes in peptides or proteins is very sensitive to and correlated with the structural and dynamic properties of hydrogen bonds [52-54]. Here, the peak at $\sim$3328 cm$^{-1}$ is assigned to the N-H mode of $\alpha$-helical structure in the core of lipid bilayer while the peak at $\sim$3300 cm$^{-1}$ is contributed by the N-H group of the $\alpha$-helical and loop structure that is exposed to water. The frequency shift of the N-H group indicates that WALP23 inserts into the hydrophobic part of lipid bilayer and forms $\alpha$-helical structure. This conclusion is approved by the hydrogen-deuterium exchange (HDX) of the amide proton. Previous studies have shown that the residues flanked at the lipid/water interfaces can undergo amide proton HDX rapidly following the sample exposure to deuterium, but the part in the core of lipid bilayer does not exchange in 3-4 days [26, 55, 56]. After the spectra in FIG. 5(a) became stable, we replaced the subphase solution of DMPG bilayer by D$_{2}$O and found that the N-H intensity does not have a significant change [26]. With this information, schematic structure evolution of WALP23 in the DMPG lipid bilayer can be described by FIG. 6. WALP23 binds to membrane surface with a primarily random coil structure. The insertion of WALP23 into lipid bilayer promotes the formation of loop structure initially and finally leads to the formation of $\alpha$-helical structure.

 FIG. 5 (a) The real-time ppp spectra of WALP23 at DMPG bilayer interface with the subphase temperature changes from 15 ℃ to 35 ℃, (b) the ppp spectra of WALP23 in amide A region at different time interval.
 FIG. 6 Schematic structural evolution of WALP23 at DMPG bilayer interface.
Ⅳ. CONCLUSION

In this study, we have demonstrated the power of femtosecond sum frequency generation vibrational spectroscopy in monitoring the peptide transport across the membranes. Our femtosecond SFG system is capable of acquiring the spectrum in the amide Ⅰ and amide A bands with recording time of less than one second. Such fast recording time allows us to probe the specific structural details about intermediates in the interaction between WALP23 and DMPG lipid bilayer. It is found that WALP23 is initially absorbed at the gel-phase DMPG bilayer with a random coil structure. The absorption of WALP23 on the surface leads to the surface charge reversal and thus changes the orientation of membrane-bound water. As the DMPG bilayer changes from gel phase into fluid phase, WALP23 inserts into the fluid-phase bilayer with its N-terminal end moving across the membrane, which causes the membrane dehydration and promotes the formation of loop structure initially and finally the $\alpha$-helical structure.

Ⅴ. ACKNOWLEDGMENTS

This work was supported by the National Natural Science Foundation of China (No.21473177, No.21633007), the National Key Research and Development Program of China (No.2017YFA0303500 and No.2018YFA0208700), Fundamental Research Funds for the Central Universities (No.WK2340000064), Anhui Initiative in Quantum Information Technologies (No.AHY090000).

Reference
 [1] L. Fu, G. Ma, and E. C. Yan, J. Am. Chem. Soc. 132 , 5405 (2010). DOI:10.1021/ja909546b [2] G. van Meer, D. R. Voelker, and G. W. Feigenson, Nat. Rev. Mol. Cell Biol. 9 , 112 (2008). DOI:10.1038/nrm2330 [3] A. J. M. Driessen, and N. Nouwen, Annu. Rev. Biochem. 77 , 643 (2008). DOI:10.1146/annurev.biochem.77.061606.160747 [4] B. Ding, J. Jasensky, Y. Li, and Z. Chen, Acc. Chem. Res. 49 , 1149 (2016). DOI:10.1021/acs.accounts.6b00091 [5] B. Ding, A. Panahi, J. J. Ho, J. E. Laaser, C. L. Brooks, M. T. Zanni, and Z. Chen, J. Am. Chem. Soc. 137 , 10190 (2015). DOI:10.1021/jacs.5b04024 [6] M. Y. Xiao, S. Joglekar, X. X. Zhang, J. Jasensky, J. L. Ma, Q. Y. Cui, L. J. Guo, and Z. Chen, J. Am. Chem. Soc. 139 , 3378 (2017). DOI:10.1021/jacs.6b10651 [7] X. Wang, P. Yang, F. Mondiot, Y. Li, D. S. Miller, Z. Chen, and N. L. Abbott, Chem. Comm. 51 , 16844 (2015). DOI:10.1039/C5CC06996C [8] S. Badieyan, Q. Wang, X. Zou, Y. Li, M. Herron, N. L. Abbott, Z. Chen, and E. N. G. Marsh, J. Am. Chem. Soc. 139 , 2872 (2017). DOI:10.1021/jacs.6b12174 [9] L. Schm, ü ser, S. Roeters, H. Lutz, S. Woutersen, M. Bonn, and T. Weidner, J. Phys. Chem. Lett. 8 , 3101 (2017). DOI:10.1021/acs.jpclett.7b01059 [10] S. J. Roeters, C. N. van Dijk, A. Torres-Knoop, E. H. G. Backus, R. K. Campen, M. Bonn, and S. Woutersen, J. Phys. Chem. A 117 , 6311 (2013). DOI:10.1021/jp401159r [11] D. K. Schach, W. Rock, J. Franz, M. Bonn, S. H. Parekh, and T. Weidner, J. Am. Chem. Soc. 137 , 12199 (2015). DOI:10.1021/jacs.5b06720 [12] Y. Liu, J. J. Tan, J. H. Zhang, C. Z. Li, Y. Luo, and S. J. Ye, Chem. Comm. 54 , 5903 (2018). DOI:10.1039/C8CC01533C [13] H. Lu, M. A. Hood, S. Mauri, J. E. Baio, M. Bonn, R. Munoz Espi, and T. Weidner, Chem. Comm. 51 , 15902 (2015). DOI:10.1039/C5CC05830A [14] L. Fu, J. Liu, and E. C. Y. Yan, J. Am. Chem. Soc. 133 , 8094 (2011). DOI:10.1021/ja201575e [15] L. Fu, Z. Wang, B. T. Psciuk, D. Xiao, V. S. Batista, and E. C. Y. Yan, J. Phys. Chem. Lett. 6 , 1310 (2015). DOI:10.1021/acs.jpclett.5b00326 [16] E. C. Y. Yan, L. Fu, Z. Wang, and W. Liu, Chem. Rev. 114 , 8471 (2014). DOI:10.1021/cr4006044 [17] Z. Wang, M. D. Morales-Acosta, S. Li, W. Liu, T. Kanai, Y. Liu, Y. N. Chen, F. J. Walker, C. H. Ahn, R. M. Leblanc, and E. C. Y. Yan, Chem. Comm. 52 , 2956 (2016). DOI:10.1039/C5CC05743D [18] H. I. Okur, J. Kherb, and P. S. Cremer, J. Am. Chem. Soc. 135 , 5062 (2013). DOI:10.1021/ja3119256 [19] E. Bilkova, R. Pleskot, S. Rissanen, S. Sun, A. Czogalla, L. Cwiklik, I. Vattulainen, P. S. Cremer, and P. Jungwirth, and Ü. Coskun, J. Am. Chem. Soc. 139 , 4019 (2017). DOI:10.1021/jacs.6b11760 [20] A. M. Jubb, W. Hua, and H. C. Allen, Acc. Chem. Res. 45 , 110 (2012). DOI:10.1021/ar200152v [21] W. Hua, D. Verreault, and H. C. Allen, J. Am. Chem. Soc. 137 , 13920 (2015). DOI:10.1021/jacs.5b08636 [22] W. Xiong, J. E. Laaser, R. D. Mehlenbacher, and M. T. Zanni, Proc. Natl. Acad. Sci. USA 108 , 20902 (2011). DOI:10.1073/pnas.1115055108 [23] K. Meister, S. Strazdaite, A. L. DeVries, S. Lotze, L. L. C. Olijve, I. K. Voets, and H. J. Bakker, Proc. Natl. Acad. Sci. USA 111 , 17732 (2014). DOI:10.1073/pnas.1414188111 [24] S. Strazdaite, K. Meister, and H. J. Bakker, J. Am. Chem. Soc. 139 , 3716 (2017). DOI:10.1021/jacs.6b12079 [25] S. J. Ye, H. C. Li, W. L. Yang, and Y. Luo, J. Am. Chem. Soc. 136 , 1206 (2014). DOI:10.1021/ja411081t [26] J. J. Tan, B. X. Zhang, Y. Luo, and S. J. Ye, Angew. Chem. Int. Ed. 56 , 12977 (2017). DOI:10.1002/anie.201706996 [27] J. J. Tan, Y. Luo, and S. J. Ye, Chin. J. Chem. Phys. 30 , 671 (2017). DOI:10.1063/1674-0068/30/cjcp1706114 [28] W. Im, and C. L. Brooks, Proc. Natl. Acad. Sci. USA 102 , 6771 (2005). DOI:10.1073/pnas.0408135102 [29] M. B. Ulmschneider, J. P. Doux, J. A. Killian, J. C. Smith, and J. P. Ulmschneider, J. Am. Chem. Soc. 132 , 3452 (2010). DOI:10.1021/ja909347x [30] L. V. Schäfer, D. H. de Jong, A. Holt, A. J. Rzepiela, A. H. de Vries, B. Poolman, J. A. Killian, and S. J. Marrink, Proc. Natl. Acad. Sci. USA 108 , 1343 (2011). DOI:10.1073/pnas.1009362108 [31] J. H. Zhang, W. L. Yang, J. J. Tan, and S. J. Ye, Phys. Chem. Chem. Phys. 20 , 5657 (2018). DOI:10.1039/C7CP07389E [32] X. Hu, J. J. Tan, and S. J. Ye, J. Phys. Chem. C 121 , 15181 (2017). DOI:10.1021/acs.jpcc.7b03092 [33] T. Yu, G. N. Zhou, X. Hu, and S. J. Ye, Langmuir 32 , 11681 (2016). DOI:10.1021/acs.langmuir.6b02560 [34] J. H. Huang, K. Z. Tian, S. J. Ye, and Y. Luo, J. Phys. Chem. C 120 , 15322 (2016). DOI:10.1021/acs.jpcc.6b05677 [35] L. Fu, D. Xiao, Z. G. Wang, V. S. Batista, and E. C. Y. Yan, J. Am. Chem. Soc. 135 , 3592 (2013). DOI:10.1021/ja3119527 [36] A. Barth, and C. Zscherp, Quar. Rev. Biophys. 35 , 369 (2002). DOI:10.1017/S0033583502003815 [37] L. K. Tamm, and S. A. Tatulian, Quar. Rev. Biophys. 30 , 365 (1997). DOI:10.1017/S0033583597003375 [38] S. Wolf, E. Freier, Q. Cui, and K. Gerwert, J. Chem. Phys. 141 , 22D524 (2014). DOI:10.1063/1.4902237 [39] B. X. Zhang, J. J. Tan, C. Z. Li, J. H. Zhang, and S. J. Ye, Langmuir 34 , 7554 (2018). DOI:10.1021/acs.langmuir.8b00946 [40] W. Hua, D. Verreault, and H. C. Allen, ChemPhysChem 16 , 3910 (2015). DOI:10.1002/cphc.201500720 [41] X. Chen, W. Hua, Z. Huang, and H. C. Allen, J. Am. Chem. Soc. 132 , 11336 (2010). DOI:10.1021/ja1048237 [42] T. Ohto, E. H. G. Backus, C. S. Hsieh, M. Sulpizi, M. Bonn, and Y. Nagata, J. Phys. Chem. Lett. 6 , 4499 (2015). DOI:10.1021/acs.jpclett.5b02141 [43] J. A. Mondal, S. Nihonyanagi, S. Yamaguchi, and T. Tahara, J. Am. Chem. Soc. 134 , 7842 (2012). DOI:10.1021/ja300658h [44] J. A. Mondal, S. Nihonyanagi, S. Yamaguchi, and T. Tahara, J. Am. Chem. Soc. 132 , 10656 (2010). DOI:10.1021/ja104327t [45] A. Kundu, S. Yamaguchi, and T. Tahara, J. Phys. Chem. Lett. 5 , 762 (2014). DOI:10.1021/jz500107e [46] Y. Nagata, and S. Mukamel, J. Am. Chem. Soc. 132 , 6434 (2010). DOI:10.1021/ja100508n [47] T. Ishiyama, D. Terada, and A. Morita, J. Phys. Chem. Lett. 7 , 216 (2016). DOI:10.1021/acs.jpclett.5b02567 [48] K. A. Becraft, F. G. Moore, and G. L. Richmond, J. Phys. Chem. B 107 , 3675 (2003). DOI:10.1021/jp022488r [49] S. L. Ma, H. C. Li, K. Z. Tian, S. J. Ye, and Y. Luo, J. Phys. Chem. Lett. 5 , 419 (2014). DOI:10.1021/jz402537w [50] W. L. Yang, K. Z. Tian, and S. J. Ye, Chin. J. Chem. Phys. 28 , 518 (2015). DOI:10.1063/1674-0068/28/cjcp1504070 [51] W. Sung, S. Seok, D. Kim, C. S. Tian, and Y. R. Shen, Langmuir 26 , 18266 (2010). DOI:10.1021/la103129z [52] J. Park, J. H. Ha, and R. M. Hochstrasser, J. Chem. Phys. 121 , 7281 (2004). DOI:10.1063/1.1792612 [53] M. M. Coleman, K. H. Lee, D. J. Skrovanek, and P. C. Painter, Macromolecules 19 , 2149 (1986). DOI:10.1021/ma00162a008 [54] E. J. Cocinero, P. Carcabal, T. D. Vaden, J. P. Simons, and B. G. Davis, Nature 469 , 76 (2011). DOI:10.1038/nature09693 [55] E. B. Dunkelberger, A. M. Woys, and M. T. Zanni, J. Phys. Chem. B 117 , 15297 (2013). DOI:10.1021/jp402942s [56] J. A. Demmers, J. Haverkamp, A. J. Heck, R. E. Koeppe, and J. A. Killian, Proc. Natl. Acad. Sci. USA 97 , 3189 (2000). DOI:10.1073/pnas.97.7.3189

a. 中国科学技术大学合肥微尺度物质科学国家研究中心, 化学物理系, 合肥 230026;
b. 中国科学技术大学量子信息与量子科技前沿协同创新中心, 能源材料化学协同创新中心, 合肥 230026