Transmembrane Transport of Bicarbonate Unravelled

Abstract Invited for the cover of this issue are Dr. Stephen Butler, Dr. Hennie Valkenier and co‐workers at Université Libre de Bruxelles, Loughborough University, Masaryk University, and the University of Bristol. The image depicts the transport of bicarbonate anions versus the spontaneous diffusion of CO2 across the lipid bilayer of a liposome. Read the full text of the article at 10.1002/chem.202100491.


Chloride/Bicarbonate exchange in the lucigenin assay
The transporters that were chosen for the studies in the new EuL1 assay had previously shown to be active Cl − /HCO3 − exchangers in the lucigenin assay. In this assay the influx of Cl − is monitored, making it an indirect method to study the transport of HCO3 − . 1,2 The performance of transporters 1-4 in this lucigenin assay is shown in Figure S1 for comparison. The synthesis of compounds 1, 1 2, 2 and 3 3 has been reported and experimental details for the lucigenin assay can be found in reference 1.  a.
2 Transport experiments with the Eu.L1-based assay

Procedures for the transport experiments with the Eu.L1 assay
General procedure for the preparation of liposomes for transport measurements Lipid solutions of 1-Palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine (POPC) and cholesterol were prepared using chloroform that had been deacidified by passage through a column containing basic alumina. All aqueous solutions were prepared using deionised water that had been passed through a Millipore filtration system. Liposomes were prepared in aqueous salt solutions of 225 mM (cation concentration) and buffered to pH 7 with 5 mM HEPES (unless indicated otherwise). The synthesis and characterisation of the probe [Eu.L 1 ] + have been reported previously. 4 POPC and cholesterol solutions (10-20 mM) in deacidified chloroform were combined with a solution of anionophore in organic solvent (0.1 mM 1 in acetonitrile, 2 and 3 as solutions in methanol) in a 5 mL round bottom flask. The volumes of the aliquots were calculated from the concentrations of the lipid solutions to obtain a POPC to cholesterol ratio of 7:3 (for instance by combining 7 µmol POPC and 3 µmol cholesterol) and the desired anionophore to lipid ratio, as specified in the experiments. The solvents were evaporated under a flow of nitrogen and the resulting lipid film was dried under high vacuum for at least 1 h. The lipid film was then hydrated with 500 µL of an aqueous solution of the probe [Eu.L 1 ] + (50 µM, prepared from a 4.8 mM stock solution in methanol) in a solution of the desired buffer. The resulting mixture was sonicated for 30 s and stirred for 1 h to give heterogeneous vesicles. Multilamellar vesicles were disrupted by 10 freeze-thaw cycles. The mixture was diluted to 1 mL (by adding 0.5 mL of buffer solution) and extruded 29 times through a polycarbonate membrane with 200 nm pores in a mini-extruder (Avestin LiposoFast-Basic). The external probe was removed by passing the liposomes through a pre-packed size exclusion column (containing 8.3 mL Sephadex G-25 medium), eluted with buffer solution. i The collected large unilamellar vesicles were further diluted with buffer solution to obtain total lipid concentration of 0.4 mM and used for transport measurements the same day.
General procedure for bicarbonate transport measurements with Eu.L1 3.00 mL of this liposome solution was placed in a quartz cuvette with a small stir bar and the temperature was allowed to stabilize at 25 ˚C for 3-5 min. inside the sample compartment of a Fluoromax-4 spectrometer, that was equipped with a 420 nm cut-off filter between the sample and the detector. Emission spectra were recorded from 550-725 nm (3 nm slits, 1 nm interval, 0.3 s integration time) with excitation at 330 nm (10 nm slits). During transport measurements the emission intensity at 615 nm (3 nm slits) was monitored over time (15 minutes, 0.2 s interval) with excitation at 330 nm (10 nm slits). If cation transporters (monensin, valinomycin, gramicidin, or CCCP) or prodigiosin were used, then these were added to the liposomes as 5 µL of solution in methanol (0.24 mM to create a 1:1000 cation transport to lipid ratio, unless stated otherwise) 5 minutes before the addition of the bicarbonate. For transport measurements 38 µL of MHCO3 (0.8 M, in 225 mM salt solution) was added to the liposomes 30 s after the start of the emission recording, to create a HCO3 − concentration gradient of 10 mM. The emission intensity was measured for another 10 minutes, followed by lysing of the i The SEC columns could be reused, but required rinsing with large amounts of 0.1 M Na2SO4 solution (~300 mL) to remove any remaining traces of the probe liposomes by addition of 50 µL of Triton X-100 (5% w /w in water). Three runs were performed per experiment (unless indicated otherwise). The averaged data were normalised from 0 (before the addition of MHCO3) to 1 (after lysis).

Fitting of the data
The averaged and normalised curves were fitted in Origin 2019b after removing the first 30 seconds of data (so that the MHCO3 is added at t = 0 s), after which data after 500 s were removed (leaving the data measured at t = 30-530 s).
The normalised transport curves were fitted to a single exponential function to obtain half-lives: The initial values used for the fit parameters were y = 0.8, a = 0.8, b = 0.01 and not boundaries were set.
The half-life t½ was calculated from fit parameter b using: The normalised transport curves were fitted to a double exponential function to obtain initial rates of transport: The initial values used for the fit parameters were y = 1, a = 0.8, b = 0.05, c = 0.2, d = 0.01. All parameters were set to all be ≥ 0 and y, a, and c ≤ 1.
The initial rate I was calculated from fitted parameters a, b, c, and d using: = + Note that in the case of very low rates of transport, I cannot be determined accurately, as the fitting will be dominated by the very small initial response, which results in an overestimation of I. In the extremely high rates of transport, the equilibrium level upon transport can be briefly surpassed, which complicates the fitting of the curve. Therefore, only the first 60 s of the transport process were fitted with a double exponential function if I > 0.06 s -1 (which was the case for 1:2500 anionophore to lipid ratios in NaCl).

DLS analysis of the large unilamellar vesicles
The size of the liposomes was verified by Dynamic Light Scattering (DLS) measurements on a Malvern Zetasizer Ultra at 25 ˚C, using disposable cuvettes. Standard parameters were used for the liposomes (Refractive index 1.45) and the parameters for the aqueous NaCl buffer used were obtained from reference 5 (viscosity of 0.91 mPa·s and refractive index 1.33).
The average intensity DLS data from 6 different batches of LUVs are plotted in Figure S2 and Zaverage values fall in the range of 177 to 190 nm, with an average of 183 nm for the hydrodynamic diameter of the LUVs.

Optimisation of the EuL1 assay
As the emission response of the probe [Eu.L 1 ] + is based on a binding process of HCO3 − , we have verified what concentration of HCO3 − could be added to have a clear response, while avoiding approaching saturation of the probe with the anionic guest. For this we have prepared LUVs with 50 µM probe [Eu.L 1 ] + and compound 1 preincorporated at 1:2500 transporter to lipid ratio, as described in Section 2.1. Instead of adding NaHCO3 as a single pulse, we have added 8 times 19 µL of 0.8 M NaHCO3, corresponding to an increase of the NaHCO3 concentration by 5 mM per addition. After each addition, we waited 3 minutes for transport to occur, which allowed to estimate the final emission levels reached upon addition of different concentrations of HCO3 − in the transport assay ( Figure S3).
While addition of 25 mM NaHCO3 resulted in close to saturation of the probe, upon addition of 10 mM NaHCO3 the probe gives a strong response, but is still far from saturation, ensuring a good sensitivity for the different concentrations of HCO3 − up to this final concentration. For this reason, we have worked with a pulse of 10 mM MHCO3 in all further experiments. See Section 2.6 for further binding studies with the probe [Eu.L 1 ] + and HCO3 − . Figure S3. Transport of HCO3 − by anionophore 1 (preincorporated at 1:2500 transporter to lipid ratio) into LUVs as monitored by the EuL1 assay in 225 mM NaCl with 5 mM HEPES at pH 7, upon 8 additions of 5 mM NaHCO3.
Further optimisation of the assay involved the choice of buffer concentration and pH, which is discussed in Section 2.5.

Increase of bicarbonate concentration and not only of pH
To demonstrate that, although the probe [Eu.L 1 ] + can show a small increase of emission intensity upon increasing pH in absence of HCO3 − , 4 the response observed in our Eu.L1 assay is not due to the increase of the pH, an experiment was performed in which 2 mM NaOH was added instead of 10 mM NaHCO3 to LUVs with monensin. While the increase in pH upon the addition of NaOH was slightly higher (7.0 to 7.7) than upon addition of 10 mM NaHCO3 (7.0 to 7.5), the increase in emission intensity was much smaller ( Figure S4). Figure S4. Transport as monitored by the Eu.L1 assay in 225 mM NaCl with 5 mM HEPES at pH 7, upon addition of 10 mM NaHCO3 (resulting in an exterior pH of 7.5; red curve) or 2 mM NaOH (resulting in an exterior pH of 7.7; dark cyan) after 30 seconds. Monensin was added to the LUVs 5 minutes before addition of the basic solution at a transporter to lipid ratio of 1:1000. The LUVs were lysed 10 minutes after the addition of NaHCO3 or NaOH.

Reproducibility and stability over time
The EuL1 assays shows no significant variation in fluorescence levels, neither before nor after addition of NaHCO3, over the course of several hours, indicating that the probe [Eu.L 1 ] + is stable inside the liposomes and does not leach out. Figure S5. Three blank transport runs as monitored by the Eu.L1 assay in 225 mM NaCl with 5 mM HEPES at pH 7, upon addition of 10 mM NaHCO3. The LUVs were lysed 10 minutes after the addition of NaHCO3.

Compatibility of the assay with different lipids
To demonstrate that the EuL1 assay is also compatible with other lipids than POPC and cholesterol, we prepared DPPC liposomes with 1 incorporated (1:25k ratio) and tested the transport at 25°C and 45°C, which is respectively below and above the phase transition temperature of DPPC (41°C). At 25°C the DPPC lipids are in the gel phase and no transport by ionophores is expected, while the fluid phase is retrieved above 41°C, leading to a restoration of the anion transport by 1. These results are in agreement with those obtained for Cl − /HCO3 − transport in the lucigenin assay. 1 Figure S6. Transport as monitored by the Eu.L1 assay with 1 preincorporated in DPPC vesicles at 1:25k transporter to lipid ratio, below (cyan) and above (dark blue) the phase transition temperature of DPPC. The liposomes were suspended in 225 mM NaCl with 5 mM HEPES at pH 7 and 10 mM NaHCO3 was added after 30 seconds and Triton X-100 was added 10 minutes after the addition of NaHCO3.

25°C
Transmembrane Transport of Bicarbonate Unravelled  While the rate of transport by monensin at 1:10k is clearly inferior to the rates at higher concentration, the curves for 1:1000 and 1:100 can hardly be distinguished, indicating that at a monensin concentration of 1:1000 the Na + /H + antiport by monensin is not rate-limiting for the overall transport process.

Transport at different concentrations of buffer and at different pH
Considering that the role of monensin in the proposed transport process is to prevent the formation of a pH gradient upon CO2 diffusion, we may expect that in the absence of any transporters the concentration of the buffer and the pH at which the experiments are performed determine how much CO2 can diffuse into the LUVs before the pH gradient stops this process. For this reason, we varied the starting pH and HEPES concentration.  When controls for transport experiments were performed by adding 10 mM NaHCO3 to LUVs in 20 mM HEPES at a starting pH of 7.6, the external pH increased to 7.7 and a much larger apparent influx of HCO3 − was observed ( Figure S10) than when working with 5 mM HEPES at pH 7.0 (where addition of 10 mM NaHCO3 increases the pH to 7.4). HEPES has a pKa of 7.5 and the optimal buffer range is reported to be 6.8-8.2, 6 thus the higher concentration of HEPES can compensate the effect of the acidification by CO2 influx. When 20 mM HEPES was used but starting at pH 6.5, the buffer could not compensate for the acidification and hardly any CO2 diffusion was observed ( Figure S11). Figure S12. Transport by thiourea 3 (a) and prodigiosin (4, ii b) as monitored by the Eu.L1 assay in 225 mM NaCl with 5 mM HEPES at pH 7, upon addition of 10 mM NaHCO3 after 30 seconds. Monensin was added to the LUVs at 1:1000 transporter to lipid ratio, 5 minutes before addition of the NaHCO3. The LUVs were lysed after 10 minutes. Figure S13. Transport as monitored by the Eu.L1 assay in 225 mM NaCl with 5 mM HEPES at pH 7, upon addition of 10 mM NaHCO3 after 30 seconds. The LUVs were lysed 10 minutes after the addition of NaHCO3. Anionophores 1 (a.) or 2 (b.) were preincorporated at 1:25k anionophore:lipid ratio. Different concentrations of monensin were added to the LUVs 5 minutes before addition of the NaHCO3, using solutions of 2.4 mM and 0.24 mM in methanol respectively to obtain 1:100 and 1:1000 ratios of monensin:lipid ratios.

Bicarbonate transport by anionophores 1 and 2, with different concentrations of monensin
ii Data for prodigiosin 4 at 1:2500 are not reported, due to the significant fluorescence from this ionophore observed when added at such a high concentration.   We have also observed that HCO3 − /Cl − exchange by bambusuril 1 is inhibited by addition of NaNO3 ( Figure S17), similarly as addition of NaNO3 was found to inhibit Cl − /HCO3 − exchange by 1 in the lucigenin assay. 1 Figure S17. Transport by 1 (preincorporated at 1:50k anionophore:lipid ratio) as monitored by the Eu.L1 assay in 225 mM NaCl with 5 mM HEPES at pH 7, upon addition of 10 mM NaHCO3 after 30 seconds (blue). Addition of 10 mM NaNO3 after 90 s brought the HCO3 − /Cl − exchange to a complete halt (red), while addition of 1 mM NaNO3 reduced the rate of transport. The LUVs were lysed 10 minutes after the addition of MHCO3.

Bicarbonate uniport by anionophores in KGluconate solutions
The EuL1 assay was adapted to study HCO3 − uniport by replacing the NaCl or NaNO3 by 100 mM gluconic acid potassium salt (KGluc) and by addition of the cationophore valinomycin. A bicarbonate gradient was created by addition of 10 mM KHCO3. Gluconate anions are too hydrophilic to participate in an antiport mechanism and instead transport of K + by valinomycin can balance the uniport of HCO3 − by anionophores. 7 While the main results obtained from this HCO3 − uniport assay are discussed in the main text, here we will comment on the mechanisms and effects of the different cationophores in more detail. No response is observed in the presence of anionophores 1-4 or protonophore CCCP alone, as no symport or antiport mechanisms are possible. When monensin is added ( Figure S18e, red curve) or the combination of valinomycin and CCCP ( Figure S18d, green curve), then an increase in the concentration of HCO3 − is observed, which can be attributed to H + /K + exchange in combination with CO2 diffusion (mechanism B). When combined with valinomycin, anionophores can give rise to a response in the EuL1 assay by two potential mechanism: they can either perform true HCO3 − uniport (as indicated in the cartoon in Figure S18), or they could act a H + (or OH − ) transporter in a way identical to CCCP. This latter option would lead to a transport process of which the rate is limited by CO2 diffusion, as for transport by monensin and CCCP+valinomycin. Transport exceeding this rate can thus be attributed to HCO3 − uniport, as is the case for 1 (which is, moreover, incapable of transporting H + or OH − ) and for 2 at higher concentrations (1:2500), see also Table S2 on pS20. Further insight can be obtained by combining the anionophores with monensin. While combining 1 and monensin gives only slightly faster transport than the apparent HCO3 − transport by monensin alone, combining 2 or 3 with monensin leads to significantly faster transport. This can be explained by a mechanism in which 2 and 3 do transport the HCO3 − anion into the liposomes, but return deprotonated (or with a OH − anion bound), leading to HCO3 − /H + symport, which would give acidification of the interior of the liposomes if the pH gradient was not dissipated by monensin.
Especially thiourea 3 appears to prefer this mechanism over HCO3 − uniport in presence of valinomycin. This is analogous to the work by Gale and co-workers on Cl − transport in which they showed that certain anionophores preferred to couple to monensin (acting as Cl − /H + symporters), while other anionophores could couple to valinomycin and perform Cl − uniport. 8 iii Valinomycin was used at a 1:1000 transporter to lipid ratio in most experiments, as higher valinomycin concentrations show a response even in absence of anionophores. While a lower valinomycin concentration (1:10k) gives lower rates of global HCO3 − transport by compounds 2 and 3, the opposite was found for bambusuril 1, which gave better rates when using 1:10k than 1:1000 valinomycin. This is in line with results obtained for 1 in various conditions and assays, showing a detrimental effect of the addition of increasing amounts of valinomycin on the rates of transport by 1. For this reason, we report the HCO3 − uniport data for 1 with a valinomycin concentration of 1:10k. Based on the discussion of the additivity of mechanisms A and B, we could have expected that initial rates of transport by 1 (at different concentrations) and by monensin alone (1:1000 ratio) could be added to predict the initial rates of transport by the combination of 1 and monensin. However, in Table S1 we see that the values found are higher than those predicted by this hypothesis, especially at higher concentrations of 1 (≥ 1:25k ratio). When comparing initial rates of transport as measured in the EuL1 assay quantitively, we should consider the effect of the different pH profiles during the transport measurements (Fig. 6d-f), as the [Eu.L 1 ] + probe is not only sensitive to concentrations of HCO3 − , but also to pH. 4 While the effects of pH on the emission intensity were reported to be small at pH ≤ 8, the pH has a more noticeable effect on the increase of luminescence observed upon addition of HCO3 − . The titrations in Figure S19 and Figure S20 (see next Section) show that this increase of emission intensity is larger at pH 7.4 than at pH 7.0, leading to different apparent affinities of EuL1 for HCO3 − . In the presence of monensin, the pH rapidly equilibrates to ~7.4 (Fig. 6e, green curve), thus resulting in a higher sensitivity of the probe to HCO3 − , which could result in a higher apparent initial rate compared to transport measured under conditions without pH equilibration (i.e., with only 1 present in the LUVs).

Quantification of rates of transport
Furthermore, the normalisation of the data could result in an error on the values of the initial rates, but not the half-lives. Therefore, half-lives are more reliable to compare transport data of 1 in presence and absence of monensin, as these values indicate how fast equilibrium is reached, independent of absolute emission values. The comparison of half-lives of transport by 1 with and without monensin clearly shows that equilibrium is reached much faster in the presence of monensin (Table S1 and Figure S15), confirming the additivity of mechanisms A and B.
For the HCO3 − uniport studies in KGluc solution, we determined the half-lives in presence of valinomycin (typical HCO3 − uniport experiment), but also in presence of monensin and both monensin and valinomycin. See page S19 for a discussion of these data.

Titration at constant pH (Sample 3)
A solution of [Eu.L 1 ] + probe (50 µM) in the adequate buffer (225 mM NaCl, 100 mM HEPES, pH 7.4) was degassed by bubbling argon for 1 h, to remove any atmospheric CO2 that could be dissolved due to the high concentration of buffer. To 1.5 mL of this solution aliquots of 0.8 M NaHCO3 (3.8 -20 µL; in 225 mM NaCl) were added, and after each addition the sample was homogenized and the emission spectrum recorded. The pH of the sample was maintained at 7.4-7.5 (after addition of 30 mM and 40 mM NaHCO3, 10 µL of 0.5 M HCl were added to maintain the pH into the desired range).

Data treatment
The values of fluorescence intensity measured at 615 nm during the titrations were corrected by the factor of dilution resulting from the additions of NaHCO3 and HCl solutions (assuming that under the conditions of the experiment the fluorescence intensity has a linear dependence on the concentration of [Eu.L 1 ] + ). The results are presented in Figure S19a and show a significant dependence of the increase in fluorescence intensity on the pH of the sample. This is illustrated in Figure S19b, where the emission spectra of samples 1 and 2 show only small differences in absence of NaHCO3, while upon addition of 4 mM NaHCO3 (data points in the grey box in Figure S19a) a significant difference in intensity is observed. The corrected titration data have also been fitted to a 1:1 binding model 9 in Origin 9.55, showing a significant variation in apparent affinity constants of [Eu.L 1 ] + for HCO3 − ( Figure S20).
The differences observed here should be taken into consideration when quantitatively comparing data from transporters that give fast pH equilibration (a situation mimicked in sample 3) to those where the pH variation is only caused by the transport of HCO3 − (as by bambusuril 1, mimicked with sample 1).

Procedures for the transport experiments with HPTS
The liposomes used in these experiments were prepared as described for the Eu.L1 assay, but using a solution of 8-hydroxypyrene-1,3,6-trisulfonic acid trisodium (HPTS, 0.1 mM) in the desired buffer (225 mM NaCl, 5 mM HEPES, pH 7.0) to hydrate the lipid films.
3.00 mL of this liposome solution was placed in a quartz cuvette with a small stir bar and the temperature was allowed to stabilize at 25˚C for 3-5 min. inside the sample compartment of a Fluoromax-4 spectrometer. During transport measurements the fluorescence intensity at 511 nm (2 nm slits) was monitored over time (15 minutes, 2 s interval) for two different excitation wavelengths, 403 and 455 nm (2 nm slits). In experiments with monensin or prodigiosin, these were added to the liposomes as 5 µL of solution in methanol 5 minutes before the addition of the bicarbonate. For transport measurements 38 µL of NaHCO3 (0.8 M, in 225 mM salt solution) was added to the liposomes 30 s after the start of the fluorescence recording, to create a HCO3 − concentration gradient of 10 mM and a pH gradient of 7.0 (inside) to 7.5 (outside). The fluorescence intensity was measured for another 10 minutes, followed by lysing of the liposomes by addition of 50 µL of Triton X-100 (5% w /w in water). The transport curves were obtained from the ratio of the emission intensities from the two different excitation wavelengths (F455/F403) and the average of three runs was calculated. The averaged data were normalised from 0 (before the addition of NaHCO3) to 1 (after lysis). Figure S21. Transport by thiourea 3 (a) and prodigiosin (4, b) as monitored by HPTS in 225 mM NaCl with 5 mM HEPES at pH 7, upon addition of 10 mM NaHCO3 after 30 seconds. Monensin was added to the LUVs with 3 at 1:1000 transporter to lipid ratio, 5 minutes before addition of the NaHCO3. The LUVs were lysed after 10 minutes.

Transport experiments with 13 C NMR spectroscopy
Transport experiments on isotopically labelled H 13 CO3 − were performed as described in reference 10.
The following buffer solutions were prepared: -450 mM NaCl and 20 mM HEPES at pH 7.3 in D2O/H2O (1:9) -150 mM Na2SO4 and 20 mM HEPES at pH 7.3 in D2O/H2O (1:9). Lipid films of EYPC were prepared from chloroform, and dried for at least 1 h under high vacuum. The EYPC lipid film (69 mg) was then hydrated with 1 mL NaCl buffer solution, containing 450 mM NaCl and 20 mM HEPES at pH 7.3 in D2O/H2O (9:1), resulting in a lipid concentration of 90 mM. The mixture of lipid and buffer was vortexed for 5 minutes, and the resulting vesicles were frozen and thawed 5 times and extruded 41 times through a polycarbonate membrane with 5 µm pores in a mini-extruder (Avestin LiposoFast-Basic). 13 C NMR spectra were recorded on a Jeol JNM-ECZ600R/S3 Spectrometer with a 5 mm Royal Probe at 25˚C, using a 30˚ pulse, an acquisition time of 0.87 s, relaxation delay of 0.2 s and 160 scans per spectrum.
For the transport experiments, 230 µL of vesicles were combined with 340 µL Na2SO4 buffer in an NMR tube and 38 µL of 0.8 M NaH 13 CO3 in D2O/H2O (1:9) was added to obtain a NaH 13 CO3 concentration of 50 mM. The 13 C signal of HCO3 − was clearly observed at 160 ppm ( Figure S22, red spectra). Then 8.8 µL of 35 mM MnCl2 in D2O/H2O (1:9) was added, resulting in an overall Mn 2+ concentration of 0.5 mM, which resulted in a broadening of the HCO3 − signal ( Figure S22, yellow spectrum). Subsequent addition of 8.65 µL of 24 mM monensin in methanol (1:1000 monensin to lipid ratio) resulted in a re-appearance of a sharp HCO3 − signal ( Figure S22a, green spectra), while this did not happen when methanol without monensin was added ( Figure S22b). This small but sharp HCO3 − signal indicates that HCO3 − is present inside the vesicles, where the signal is not affected by the paramagnetic Mn 2+ . Upon lysis of the vesicles by addition of 30 µL Triton X-100, this sharp interior HCO3 − signal disappeared again ( Figure S22, purple spectra). Figure S22. 13 C NMR spectra of vesicles with NaCl interior and NaH 13 CO3 and exterior (red), followed by addition of paramagnetic MnCl2 (yellow), upon addition of monensin in methanol (a) or methanol only (b) after 10 minutes and 35 minutes (light and dark green), and upon lysis of the vesicles with Triton X-100 (purple). The appearance of a sharp HCO3 − signal after the addition of monensin indicates that HCO3 − has been transferred to the interior of the vesicles.