Halide-selective, proton-coupled anion transport by phenylthiosemicarbazones

Phenylthiosemicarbazones (PTSCs) are proton-coupled anion transporters with pH-switchable behaviour known to be regulated by an imine protonation equilibrium. Previously, chloride/nitrate exchange by PTSCs was found to be inactive at pH 7.2 due to locking of the thiourea anion binding site by an intramolecular hydrogen bond, and switched ON upon imine protonation at pH 4.5. The rate-determining process of the pH switch, however, was not examined. We here develop a new series of PTSCs and demonstrate their conformational behaviour by X-ray crystallographic analysis and pH-switchable anion transport properties by liposomal assays. We report the surprising finding that the protonated PTSCs are extremely selective for halides over oxyanions in membrane transport. Owing to the high chloride over nitrate selectivity, the pH-dependent chloride/nitrate exchange of PTSCs originates from the rate-limiting nitrate transport process being inhibited at neutral pH.


Characterisation of (E)-2-(4-butylbenzylidene)-N-(perfluorophenyl)hydrazine-1carbothioamide 9
Afforded white solid (85% yield), mp 187-189 °C  The internal and external solutions vary from experiment to experiment. The buffers for various pHs were prepared with a constant buffer concentration (5.0 mM, citrate buffer for pH 4.5 and phosphate buffer for pH 7.2) and total ionic strength (500 mM) for both intra-and extravesicular solutions. Typically, for each measurement, the lipid stock solution was diluted with the external buffered solution to a standard volume (5.0 mL) with a lipid concentration of 1.0 mM. A DMSO solution (10 μL) of the test compound (carrier) was added to the liposome solution to start the experiment and the chloride efflux was monitored using a chloride-selective electrode. At 5 minutes, the detergent solution (50 μL) was added to lyse the vesicles, and the total chloride reading was taken at 7 min. The initial value was set as 0% chloride efflux and the final chloride reading (at 7 minutes) was set as 100% chloride efflux, to normalize the collected data from each measurement. S16

S2.2 General Procedure for cationophore-coupled assay
Chloride concentrations during transport experiments were monitored using a chloride-selective electrode. POPC LUVs (mean diameter 200 nm) with internal KCl and external K2SO4 were prepared as follows. A lipid film of POPC was formed from a chloroform solution under reduced pressure and dried under vacuum for at least 8 hours. The lipid film was hydrated by vortexing with the intravesicular KCl (300 mM) solution. The lipid suspension was then subjected to 9 freezethaw cycles, where the suspension was alternatingly allowed to freeze in a liquid nitrogen bath, followed by thawing in a water bath. The lipid suspension was allowed to age for 30 minutes at room temperature and was extruded 25 times through a 200 nm polycarbonate membrane using an extruder set. The resulting large unilamellar vesicles (LUVs) with a mean diameter of 200 nm were dialyzed against the extravesicular K2SO4 (100 mM) for a minimum of 2 hours to remove unencapsulated KCl salts. The lipid solution obtained after dialysis was diluted to a standard volume (10 mL) with the extravesicular buffered solution to obtain the lipid stock solution with a known lipid concentration.
A DMSO solution (10 μL) of the anionophore was added to the liposome solution to start the experiment and the chloride efflux was monitored using a chloride-sensitive electrode. When a cationophore (valinomycin or monensin) was used, a DMSO solution of the cationophore (0.5 mM, 10 μL) was added to the liposome solution prior to the addition of the test anionophore. At 5 min, the detergent solution (50 μL) was added to lyse the vesicles, and the total chloride reading was taken at 7 min. The initial value was set at 0% chloride efflux and the final chloride reading (at 7 minutes) was set as 100% chloride efflux, to normalise the collected data from each measurement.
The KCl assay involves two complementary cationophores, monensin and valinomycin, which can determine whether the transport process is electrogenic chloride transport or electroneutral H + /Clcotransport. An electrogenic processes result in a net alteration of charge across the membrane due to ion transport, whereas electroneutral process has a charge balance that can be achieved with transport of a species with the same charge or cotransport of a species with opposite charge.
Monensin functions as a M + /H + antiporter in the lipid bilayer as the deprotonated monensin is unable to diffuse through the membrane unless it is binding a cation, thus it can facilitate electroneutral transport. Valinomycin can only facilitate electrogenic transport, the anionphore transports Clcoupled with K + transport mediated by S17 valinomycin. Chloride efflux is measured by ISE from POPC LUVs loaded with KCl (300 mM) and suspended in K2SO4 (100 mM), adjusted to pH 4.5 with KOH in citrate (5.0 mM) or pH 7.2 in phosphate (5.0 mM). Either monensin (0.1 mol%) or valinomycin (0.1 mol%) are added in combination with an anionophore and chloride efflux is driven by the large chloride concentration gradient.

S3.2 Anion gradient assay
Following vesicle preparation according to the general procedure, the NaCl-containing vesicles were diluted using NaX (100 mM) external solutions to obtain NaCl in /NaX out vesicles suspended in 2.5 mL samples containing 0.1 mM of POPC. The samples were stirred at 298 K.
No base pulse was added so no pH gradient was initially present. Transporter 1 (added in 5 μL of DMSO) was added at time 0 to induce pHin changes. The pHin was monitored by the S41 ratiometric fluorescence response of HPTS I460 / I403 (λex = 460 nm, λem = 510 nm divided by λex = 403 nm, λem = 510 nm). The I460 / I403 values were converted to pHin using previously reported calibration.

S3.3 pH gradient dissipation assay
Following vesicle preparation according to the general procedure, the NaX-containing vesicles were diluted using NaX (100 mM) external solutions to obtain NaX in /NaX out vesicles suspended where Rt is the fluorescence ratio at time t, R0 is the fluorescence ratio at time 0, and Rf is the fluorescence ratio after the addition of detergent.
We have previously shown that the I460 / I403 value, instead of the pHin, is proportional to the amount of H + efflux. [2] Therefore, here for the purpose of indicating the progress of membrane transport, the I460 / I403 values were not converted to pHin values.

S4 Osmotic response (light scattering) assay
To investigate the rates of H + /X − symport at pH 4.5, we conducted an osmotic response assay because we were unable to find a membrane-impermeable fluorescence pH indicator suitable for monitoring intra-vesicular pH around 4.5 in the same way as we conducted the HPTS assay.
Larger LUVs (mean diameter 400 nm) were used compared with the ISE and HPTS assays.
A chloroform solution of POPC was evaporated in a round-bottom flask and the lipid film formed was dried under vacuum for at least 6 h. Then, the lipid film was hydrated by vortexing with an internal solution of NaCl or NaNO3 (300 mM). The lipid suspension was subjected to nine freeze/thaw cycles and then extruded 21 times through a 400 nm polycarbonate membrane.
The NaCl or NaNO3 containing vesicles were diluted using sodium gluconate (300 mM) external solutions buffered at pH 4.5 (5 mM citrate) or pH 7.2 (5 mM phosphate) to obtain 2.5 mL samples containing 0.2 mM of POPC. The samples were stirred at 298 K.

S42
The NaCl or NaNO3 containing vesicles were diluted using sodium gluconate (300 mM) external solutions buffered at pH 4.5 (5 mM citrate) or pH 7.2 (5 mM phosphate) to obtain 2.5 mL samples containing 0.2 mM of POPC. The samples were stirred at 298 K.
The light scattering of the vesicles was monitored using a fluorometer (λex = 600 nm, λem = 600 nm). Efflux of intra-vesicular osmolyte (NaCl or NaNO3) would lead to osmotic shrinkage of vesicles and an increase of light scattering intensity. [3] The samples were initially stirred for ~5 min until the baseline stabilized. A DMSO solution of transporter 1 (1 mM, 10 μL) was then added, followed by a DMSO solution of monensin (0.5 mM, 5 μL). Control experiments with only monensin added were also performed. Monensin facilitates Na + /H + symport, which couples to HCl or HNO3 transport facilitated by an anion transporter or via simple diffusion to give overall NaCl or NaNO3 efflux. The transport of the intra-vesicular salt was monitored for 15 min. At the end of the experiment, the vesicles were treated with a DMSO solution of N,N′−bis [3,5−bis(trifluoromethyl)phenyl]−thiourea (1 mM, 10 μL) for 5 -10 min to complete the salt transport and normalize the light scattering intensity to 100% transport. These data support the hypothesis that the protonated form of 1 is selective for NO3 − over Cl − . The HCl transport induced by 1 is slower than that observed in the ISE assay (Fig. S56 red) because of the larger vesicle sizes used in the osmotic response assay. The use of a higher monensin concentration in this assay and higher transport activity of monensin for Na + than for K + , however, has led to faster efflux of chloride salts (Fig. S65 blue) than in the ISE assay ( Fig. S39 red open symbols) in the absence of an anion transporter.
When the assay was performed at pH 7.2, no salt efflux was observed under all conditions tested (Fig. S66), consistent with the negligible concentration of the free acid of the anions as required for the HX simple diffusion mechanism and the lack of transporter protonation as required for 1-facilitated HX transport at neutral pH.   S44 S5 X-ray crystallography