Localized proton microcircuits at the biological membrane–water interface

Results and Discussion

The effect of the membrane composition on the protonation dynamics of the membrane-anchored single fluorescein probe was studied by using FCS. In the FCS measurements, the fluorescence fluctuations generated by protonation/deprotonation of the individual pH sensitive fluorescein molecules were monitored as the liposomes diffused through the small detection volume (1.5 fl). Thus, the recorded fluctuations in fluorescence directly reflected the proton exchange of the fluorophore with its immediate environment at equilibrium conditions (Fig. 1). Moreover, because of the low fluorophore concentrations used, buffer effects from the fluorophore molecules themselves could be neglected. Thus, this approach has great advantages as compared with conventional relaxation experiments where the response time of the fluorescent pH probe to the proton pulse depends not only on the immediate environment of the fluorophores but also on all other interacting proton-accepting groups of the system.

For freely diffusing fluorescein-labeled liposomes, the recorded FCS curves showed independent fluctuations in three different time regimes associated with (i) the diffusion rate constant (k_D), i.e., the rate by which the liposome diffuses through the detection volume (1/k_D = τ_D ≅ 1.2 ms); (ii) the relaxation rate (k_trip) for the singlet–triplet state transitions within the fluorescein molecules (1/k_trip = τ_trip ≈ 1.5 μs); and (iii) the protonation relaxation rate (k_prot). Given the small size of the liposomes (radius = 15 nm) in relation to the detection volume (radius ≅ 300 nm) and their mobility, the fluorescein-labeled lipid molecules enter and leave the detection volume by diffusion of the whole liposomes and not by diffusion of the fluorescein-labeled lipid molecules within the membranes of the liposomes themselves. Therefore, the changes in fluorescence attributable to diffusion of the lipid molecules can be neglected.

To investigate the proton-collecting antenna concept, we prepared liposomes composed of the lipid 1,2-dioleoylsn-glycero-3-[phospho-rac-(1-glycerol)] (DOPG), which has a head group that is negatively charged and protonatable with a pK_a of ≈2. Fig. 2A shows FCS data obtained from measurements on fluorescein-labeled DOPG liposomes at four selected pH values. The amplitude of the protonation component reflects the extent of protonation and increases with the bulk proton concentration, [H⁺], whereas the time constant τ_prot = 1/k_prot reflects the proton exchange relaxation time of the fluorescein molecules. As shown in Fig. 2A Inset, k_prot is linearly dependent on [H⁺], which is consistent with proton transfer between the individual lipid head groups of the membrane surface being much faster than proton transfer from these lipid head groups to the bulk solution (compare proton-collecting antenna):

where k₊₁ is the apparent proton diffusion on-rate constant and k₋₁ is the off-rate constant. From the slope and the intercept of the fitted line with the ordinate, we obtained k₊₁ = 9.4 × 10¹² s⁻¹·M⁻¹ and k₋₁ = 3.3 × 10⁴ s⁻¹, respectively. The pK_a of the dye is obtained from −log(k₋₁/k₊₁), which gives a value of 8.5 for the anchored fluorescein molecule. This pK_a value is equal to that determined from a static pH titration with a spectrofluorometer (data not shown). The corresponding parameters obtained for water-solubilized lipid fluorescein [N-(fluorescein-5-thiocarbamoyl)-1,2-dihexadecanoyl-sn-glycero-3-phosphoethanol-amine (fluorescein-DHPE)] were k₊₁ = 6.5 × 10¹⁰ s⁻¹·M⁻¹, k₋₁ = 2.5 × 10⁴ s⁻¹, yielding a pK_a of 6.4 (data not shown). These values are the same as those obtained previously for free fluorescein in water (28). Thus, the membrane surface itself, composed of negatively charged DOPG head groups, acts to increase k₊₁ for the anchored fluorescein by a factor of 100 while keeping k₋₁ practically unaltered. A relatively large increase in k₊₁ also has been noted in the case of membrane transporters (for review, see ref. 23). Furthermore, these results are qualitatively consistent with those obtained previously in time-resolved kinetic measurements (29).

Next, we investigated whether the increase in k_prot of the membrane-anchored fluorescein molecule could be a consequence of the net negative charge of the membrane surface, which would attract protons and result in a higher proton concentration in the aqueous phase near the surface. We therefore repeated the same experiments with a liposome surface composed of 1,2-dioleoyl-sn-glycero-3-phosphocholine (DOPC) head groups, which are zwitterionic and protonatable (pK_a ≈ 2, i.e., equal to that of DOPG head groups). Also in this case, k₋₁ was found to be unaltered, whereas k₊₁ was increased by a factor of 100 (data not shown), i.e., the behavior was the same as that observed for liposomes composed of DOPG. Thus, the results show that the increased apparent proton diffusion on-rate constant (k₊₁) is a result of an increase in the effective proton collision cross-section of the anchored fluorescein molecule and that proton transfer is rapid between the membrane surface protonatable groups irrespective of the surface charge. In other words, the results show that the surface protonatable groups (negatively charged or neutral) collectively act as a proton-collecting antenna for the anchored fluorescein molecule. Moreover, these observations also indicate that the prerequisite for a localized proton circuit is fulfilled, namely, that the proton transfer along the liposome surface is faster than proton exchange between the surface and the bulk aqueous phase.

To investigate the role of buffering protonatable groups (compare His residues at the surface of a proton transporter) in a proton-collecting antenna, we introduced 1,2-dioleoyl-sn-glycero-3-phosphate (DOPA) lipids into the DOPG liposomes. The DOPA lipid head groups have a pK_a of 7.4 (determined from titration), which is close to the pH of the bulk solution in our experiments as well as to the pK_a of the fluorescein dye molecules (≈8.5). Consequently, the liposomes containing DOPA harbor surface-localized protons that are in rapid equilibrium with the anchored fluorescein molecule. To model this scenario, Eq. 2 has to be extended to account for the equilibrium between the fluorescein molecule and the DOPA head groups in the liposomes. The dependence of the protonation rate constant, k_prot, on the concentration of DOPA (here expressed as the number of DOPA lipid head groups per liposome) was fitted to Eq. 3 (28).

The first two terms of Eq. 3 are the same as those in Eq. 2, and the last term describes the proton exchange rate between the DOPA lipid head groups and the membrane-anchored fluorescein molecule. K_a is the acidity constant of DOPA, which was determined to be 10^−7.4, and k_s+ and k_s− are the protonation on- and off-rate constants of fluorescein for its proton exchange with DOPA along the surface of the DOPG liposomes. As shown in Fig. 2B, the slope of k_prot as a function of [H⁺] increases with [DOPA] (see Fig. 2B Inset), and a fit of the data to Eq. 3 gave k_s+ ≅ 2 × 10⁴ s⁻¹ (liposome) and k_s− ≅ 3 × 10² s⁻¹ (liposome) [note that the units of [DOPA] are (#DOPA)·(liposome⁻¹); see Eq. 3]. This observation clearly demonstrates that the proton-uptake rate to a proton acceptor (e.g., fluorescein) at a membrane surface can be accelerated by surface-bound protonatable groups. Because, as shown above, proton transfer along a liposome surface composed of low-pK_a lipid head groups is fast, these high-pK_a DOPA groups can be rapidly reprotonated from the bulk aqueous phase, analogous with the role of the His residues in the proton-collecting antennae typically found in proton transporters (see above). This role also has been discussed previously for the high-pK_a head group of cardiolipin (30).

The diffusion coefficient for proton transfer along the membrane surface (D_surface) can be estimated from the dependence of k_prot on [DOPA] by using the relation (in a pH range where the DOPG head groups are unprotonated, pH ≫ 2):

where ν is the number of proton “hops” between the DOPA head groups and the anchored fluorescein molecule per unit time and l is the average length of a proton hop. We calculate D_surface as follows. The number of proton hops per unit time is estimated from the rate constant for proton exchange along the surface between fluorescein and the DOPA head groups, i.e., ν = (k_s+ + k_s−)·[DOPA] ≅ 2 × 10⁴·[DOPA] s⁻¹. The surface area of the liposomes used in our study is on average ≈3 × 10⁻¹¹ cm². Assuming that each DOPA and fluorescein head group occupies an area of π(l/2)², we obtain l² ≅ 4 × 10⁻¹¹·[DOPA]⁻¹ cm², which gives (Eq. 4) D_surface ≅ 2 × 10⁻⁷ cm²·s⁻¹.

It should be pointed out that the diffusion coefficient of lipid molecules within a pure lipid bilayer typically is smaller (31). Therefore, the exchange of protons can be expected to occur via collisions of fluorescein with DOPA only to a minor extent. Instead, migration of the proton along the surface of the DOPG head groups appears to be the dominating mechanism. The advantage of using the FCS approach for determining D_surface is that the value can be obtained under equilibrium conditions. Because the rate of direct protonation of a surface group from the bulk solution decreases with increasing pH, the significance of direct proton transfer from a proton source to a proton sink along the surface increases with increased pH.

In conclusion, we have introduced an FCS-based approach to monitor local, unperturbed proton exchange at membrane surfaces. We foresee that this approach can be applied further to a broad range of investigations of membrane-associated ion exchange, also involving ions other than hydrogen. Our results show that the whole DOPG–liposome surface acts to enhance the protonation cross-section of the surface-anchored fluorescein, increasing its protonation rate by a factor of 100 (Fig. 3A). This result therefore demonstrates that proton-collecting antennae are likely to be important for the function of proton transporters. Furthermore, we show that the DOPA lipid head groups can store membrane-localized protons that are in rapid equilibrium with the fluorescein proton acceptor (Fig. 3B). A compatible scenario is found in proton transporters where His residues act as surface-localized buffers near an entry point of a proton-conducting pathway. After delivering a proton to the pathway, the His residue can be rapidly reprotonated from the bulk solution via a surrounding proton-collecting antenna composed of Asp and Glu residues (Fig. 3C). In addition, from the dependence of k_prot on [DOPA], we calculated the diffusion coefficient of proton transfer along the membrane surface (D_surface) to be 2 × 10⁻⁷ cm²·s⁻¹. Even though D_surface < D_water = 9.3 × 10⁻⁵ cm²·s⁻¹ (diffusion coefficient of proton transfer in water), our results show that if the density of surface-bound protonatable groups is such that it corresponds to a higher proton concentration at the surface than in the bulk solution, then the proton exchange rate of a surface group (fluorescein) with the surface can exceed that to the bulk solution.

Materials and Methods