Identification of a probable pore forming domain in the multimeric vacuolar anion channel AtALMT9

that is poorly permeable compared to malate. Currents were evoked in response to 3 s voltage pulses ranging from +60 mV to -120 mV in -20 mV steps followed by a tail pulse at +60 mV. The holding potential was set +60 mV. Error bars represent Each data point corresponds to 4-11 patches. probability of mM mM malate presence of mM CA mM malate mM malate cyt +10 mM CA cyt The relative open probability was estimated from the initial current amplitude of the tail currents (derived from a mono-exponential fit of the current decay) which followed an activating pulse of various potentials. We were unable to reach the full activation of AtALMT9 channels as it is likely to occur at voltages more negatives than -160 mV, a value at which the vacuolar membrane becomes unstable. Therefore the data was normalized to the I max value that each data the R the T the does not (E) Representative I-V curves obtained with a voltage (from +60 to -160 in 1.5 potential +60 mV) measured in excised cytosolic-side out patches from vacuoles expressing AtALMT9 K87E in control conditions (ctrl) and in presence of 10 mM CA cyt . Dose-response of CA cyt concentration-dependent ratio of AtALMT9 WT and AtALMT9 K87E at -160 mV. To estimate the dissociation constant K dCA the data points were fitted with a Langmuir isotherm (equation 1). The resulting K dCA values were 5.1 ± 0.3 mM and 16.2 ± 2.3 mM for AtALMT9 WT and AtALMT9 K87E , respectively. Error bars sd.

upper traces) and AtALMT9 K193E (C; upper traces) displayed time dependent malate currents in symmetric ionic conditions (circles; 100 mM malate vac / 100 mM malate cyt ). When the cytosolic solution was replaced with 100 mM CA cyt (squares; 100 mM malate vac / 100 mM CA cyt ) the inward currents decreased and displayed only weak time dependent currents (A, C; lower traces). The current ratios I CA /I Malate (AtALMT9 WT ) = 5 ± 5% and I CA /I Malate (AtALMT9 K193E ) = 7 ± 3% show that CA is poorly permeable compared to malate. Currents were evoked in response to 3 s voltage pulses ranging from +60 mV to -120 mV in -20 mV steps followed by a tail pulse at +60 mV. The holding potential was set +60 mV. Error bars represent sd. Each data point corresponds to 4-11 patches. (B) Voltage dependency of the relative open probability of AtALMT9 WT in control conditions (ctrl; 100 mM malate vac / 100 mM malate cyt ; circles) and in presence of 10 mM CA cyt (100 mM malate vac / 100 mM malate cyt +10 mM CA cyt ; diamonds). The relative open probability was estimated from the initial current amplitude of the tail currents (derived from a mono-exponential fit of the current decay) which followed an activating pulse of various potentials. We were unable to reach the full activation of AtALMT9 channels as it is likely to occur at voltages more negatives than -160 mV, a value at which the vacuolar membrane becomes unstable. Therefore the data (n=8-10) was normalized to the I max value that was obtained by fitting each data set with the Boltzmann equation.
The solid lines represent the best fits of the mean relative open probability in control and in presence of 10 mM CA cyt with a Boltzmann equation in the following form: ) in which P O rel is the relative open probability, z the gating charge, F the Faraday constant, R the universal gas constant, T the absolute temperature and V h the voltage of half activation. The fit shows that the presence of 10 mM CA cyt does not change significantly the voltage dependent gating of the channel since V h = -81 ± 1 mV and z = 0.6 ± 2 under control conditions and V h = -76 ± 3 mV and z = 0.7 ± 1 with 10 mM CA cyt . (D) "Kick-out experiment" performed on the mutant AtALMT9 K193E . Grey traces were obtained under 100 mM malate vac / 100 mM malate cyt conditions and black traces in the presence of 10 mM CA cyt (100 mM malate vac / 100 mM malate cyt +10 mM CA cyt ). The currents evoked in response to a 2 s voltage pulse at -140 mV. Subsequently, the membrane potential was transiently stepped for 3 ms to +60 mV and then restored to -140 mV for 1 s which was followed by a tail pulse at +60 mV. The holding potential was set to +60 mV. Error bars display sd. Figure S2. Multiple alignment of the ALMT protein family of Arabidopsis thaliana.

Supplemental
The alignment was conducted with the Jalview software (Waterhouse et al., 2009). Asterisks and red boxes indicate the residues that were targeted by site-directed mutagenesis in the present study.

Supplemental Figure S3. Intracellular localization of the different mutant channels of AtALMT9-GFP.
Fluorescence images of vacuoles extracted from N. benthamiana protoplasts expressing the different AtALMT9-GFP mutants. None of the introduced mutations altered the tonoplastic localization of AtALMT9. The pictures were obtained with an epifluorescence microscope (Nikon Eclipse TS100) and acquired with a digital camera (Nikon DS-Fi1).

Supplemental Figure S4. AtALMT9 point mutants display different channel conductivity and sensitivity to citrate inhibition
Representative traces of current recordings from vacuoles overexpressing AtALMT9 R200N and AtALMT9 R200E (A), AtALMT9 R215N and AtALMT9 R215E (C) in symmetric malate conditions (100 mM malate vac / 100 mM malate cyt ). Currents were evoked in response to 3 s voltage pulses ranging from +60 mV to -120 mV in -20 mV steps followed by a tail pulse at +60 mV. The holding potential was +60 mV. "Kickout experiments" were performed with the mutants AtALMT9 R200N (B) and AtALMT9 R215N (D). Grey traces were obtained in 100 mM malate vac / 100 mM malate cyt conditions and black traces in the presence of 10 mM CA cyt (100 mM malate vac / 100 mM malate cyt + 10 mM CA cyt ). The currents evoked in response to a 2 s voltage pulse at -140 mV. Subsequently the membrane potential was transiently stepped for 3 ms to +60 mV and then restored to -140 mV for 1s which was followed by a tail pulse at +60 mV. The holding potential was set to +60 mV. (E) Representative I-V curves obtained with a voltage ramp (from +60 mV to -160 mV in 1.5 s; holding potential +60 mV) measured in excised cytosolic-side out patches from vacuoles expressing AtALMT9 K87E in control conditions (ctrl) and in presence of 10 mM CA cyt . (F) Dose-response of CA cyt concentration-dependent ratio of AtALMT9 WT and AtALMT9 K87E at -160 mV. To estimate the dissociation constant K d CA the data points were fitted with a Langmuir isotherm (equation 1). The resulting K d CA values were 5.1 ± 0.3 mM and 16.2 ± 2.3 mM for AtALMT9 WT and AtALMT9 K87E , respectively. Error bars represent sd.
Supplemental Figure S5. The double mutant AtALMT9 K93E/E130K is inhibited by intracellular citrate comparable to AtALMT9 WT .
(A) Representative I-V curves obtained with a voltage ramp (from +60 mV to -160 mV in 1.5 s; holding potential +60 mV) measured in excised cytosolic-side out patches from vacuoles expressing the double mutant AtALMT9 K93E/E130K in control conditions (ctrl) and in presence of 10 mM CA cyt . (B) Ratio between currents in presence of 10 mM CA cyt and in control conditions at different membrane potentials. Depicted are AtALMT9 WT (circles; n = 4) and AtALMT9 K93E/E130K (squares; n = 3). Error bars represent sd