Patch clamp experiments on mesophyll vacuoles were performed in the whole-vacuole and excised patch configuration essentially as described by Schulz-Lessdorf and Hedrich (1995) and Ivashikina and Hedrich (2005). Patch pipettes were prepared from Kimax-51 glass capillaries (Kimble products, Vineland, NY, USA) and were covered from the inside with Sigmacoat (Sigma-Aldrich Chemie GmbH, Steinheim, Germany). Close to the tip the outside of the patch pipette was coated with silicone (Sylgard 184 silicone elastomer kit; Dow Corning GmbH, Wiesbaden, Germany). Pipette resistance was about 5 to 6 MΩ for single channel recordings, whereas the pipette resistance for whole vacuole experiments was about 3 MΩ in solutions with 100 mm cytosolic potassium and 2 MΩ in solutions with 150 mm cytosolic potassium. Macroscopic and single channel recordings were performed with an EPC-7 patch clamp amplifier (HEKA, Lambrecht, Germany) at a data acquisition rate of 500 μs and 50 μs, respectively. The macroscopic currents were low-pass filtered at 5 kHz and the single channel currents at 1 kHz. Data were digitized by an ITC-16 interface (Instrutech Corp., Elmont, NY, USA), stored on a Maxdata computer and analysed using different software programs such as Pulse and PulseFit (HEKA Elektronik, Lambrecht, Germany), IGORPro (Wave Metrics Inc., Lake Oswego, OR, USA) and TAC V3.0 (Bruxton Coporation, Seattle, WA, USA). The current recordings were performed according to the convention for electrical measurements on endomembranes (Bertl et al., 1992). The vacuolar membrane was clamped to voltages (V) as indicated in the respective figure legends. To allow comparison of macroscopic current magnitudes among different vacuoles, the current densities (Iss/Cm) were determined upon dividing the macroscopic ionic current through the whole-vacuolar membrane capacitance Cm of the individual vacuole as shown in the figures. The relative voltage-dependent open probability, shown as G(V) curves in Figure 4, was determined from tail current experiments in the whole vacuolar mode. The derived conductance values (G) were plotted against the voltages (V), fitted with a Boltzmann distribution and using a fixed number of apparent gating charges (z = 1.6) and normalized with respect to the maximal conductance (G/Gmax). The half-maximal activation voltage V1/2 was derived from the Boltzmann fit and represents the voltage at which 50% of the maximal conductance is reached. Single channel amplitudes (i) were determined from single channel recordings by using the software program TAC V3.0, and the single channel activity expressed as the single-channel open probability (Po) was estimated as described by Bertl and Slayman (1990).