Local pH at Nonionic and Zwitterionic Lipid/Water Interfaces Revealed by Heterodyne-Detected Electronic Sum-Frequency Generation: A Unified View to Predict Interfacial pH of Biomembranes

Abstract

For biomembranes, which are composed of neutral as well as charged lipids, the local pH at lipid/water interfaces is extremely important in their structural formation and functional activity. In our previous study of the charged lipid/water interfaces, we found that the local pH at the interface is governed by the positive or negative sign of the charge of the lipid: i.e., the local pH is dictated by the repulsive or attractive electrostatic interaction between the charged lipid headgroup and the proton. Because of the lack of net charge in the headgroup of the neutral lipid, the factor determining the local pH at neutral lipid/water interfaces is less straightforward, and therefore it is more challenging to predict the local pH. Here we apply heterodyne-detected electronic sum frequency generation (HD-ESFG) spectroscopy to nonionic and zwitterionic lipids to investigate the local pH at the neutral lipid/water interfaces. The obtained results indicate that the local pH at the nonionic lipid/water interface is higher than in bulk water by 0.8 whereas the local pH at the zwitterionic lipid/water interface is lower by 0.6, although the latter is subject to significant uncertainty. The present HD-ESFG study on neutral lipids, combined with the previous study on charged lipids, presents a unified view to consider the local pH at biomembranes based on the balance between the electrostatic interaction and the hydrophobicity provided by the lipid.

Figure 1

(a) Acid–base equilibrium of the surface-active pH indicator (4-heptadecyl-7-hydroxycoumarin, HHC). (b) Chemical structures of the nonionic lipid (1,2-dipalmitoyl-sn-glycerol, DPG) and the zwitterionic lipid (1,2-dipalmitoyl-sn-glycerol-3-phosphocholine, DPPC).

Figure 2

(a) Imaginary and (b) real parts of the electronic χ⁽²⁾ spectra of the pH indicator HHC at the nonionic lipid DPG/water interface. The black, pink, green, blue, and red lines represent spectra obtained at bulk pH 12.7, 10.9, 9.5, 9.1, and 6.2, respectively. The dashed lines stand for global fits. (See the text for details.) The interface density of the lipids was approximately 2 molecules nm^–2, which was 10 times higher than that of HHC.

Figure 3

(a) Imaginary and (b) real parts of the electronic χ⁽²⁾ spectra of the pH indicator HHC at the zwitterionic lipid DPPC/water interface. The black, pink, green, blue, and red lines represent spectra obtained at bulk pH 12.6, 11.5, 11.3, 10.8, and 8.4, respectively. The dashed lines stand for global fits. (See the text for details.) The interface density of the lipids was approximately 2 molecules nm^–2, which was 10 times higher than that of HHC.

Figure 4

pH dependence of the dissociation degree of the pH indicator (a) at the nonionic DPG/water interface and (b) at the zwitterionic DPPC/water interface. The solid circles represent the dissociation degree determined from the analyses of the χ⁽²⁾ spectra, and the solid lines represent the fits using eq 2.

Figure 5

(a) Relation between pK_a of the pH indicator (HHC) and ε of the surrounding medium. The black circles and the solid line are taken from the upper dashed curve in Figure 3 of ref (49). The black circles represent data points obtained from dioxane–water mixtures of different ratios. The black circle at ε = 78.5 corresponds to neat water where pK_a is 7.75. The ε values of the nonionic lipid DPG/water and zwitterionic lipid DPPC/water interfaces give red and green points on the solid line, respectively, which allows us to estimate pK_a of the pH indicator at the DPG/water and DPPC/water interfaces as 10.1 and 10.7, respectively. (b) Relation between the peak wavelength of the Im χ⁽²⁾ spectrum of A^– and ε of the surrounding medium. The open circles with error bars and the solid line are taken from Figure 4 of ref (46). Dotted lines placed at ±2 nm from the solid line fully cover the error bars of the individual points, representing an estimate of the experimental error. The open circles represent the data points obtained from different bulk solvents. The peak wavelengths of A^– at the DPG/water interface (372 nm) and DPPC/water interface (376 nm) give red and green points on the solid line, respectively, which allows us to estimate the local ε of the DPG/water and DPPC/water interfaces as 25 and 20, respectively.

Figure 6

ε dependence of δ_H⁺ of the proton in water/dioxane mixtures. The black circles represent experimental data points obtained from a study by Fernández and Fromherz. The solid line stands for a phenomenological exponential fit. The ε values at the DPG/water and DPPC/water interfaces are shown with red and green points on the solid line, respectively, which allows us to estimate δ_H⁺/(RT ln 10) of the proton at the DPG/water and DPPC/water interfaces as 0.9 and 1.2, respectively.