Molecular biology and cell culture
Reliable expression of Drosophila ion channels is notoriously difficult to achieve in commonly used mammalian cell expression systems. We found that Drosophila S2 cells permit robust expression of dTRPA1, allowing culture below activation temperatures of dTRPA1 and promoting functional expression as confirmed by high sensitivity of channels to heat and chemical activators (Supplementary Fig. S1A–C).
All Drosophila-based channels and chimeras were based upon the highly temperature-dependent canonical TRPA1 isoform (Viswanath et al. 2003). Constructs were cloned into the Drosophila expression vector, pMT/V5-His, containing the inducible metallothionein gene promoter (Invitrogen, Darmstadt, Germany). Drosophila S2 cells were acquired from Invitrogen and maintained according to the manufacturer's protocol at 25°C. Cells were transiently transfected with channel and a yellow fluorescent protein reporter plasmid using FugeneHD (Roche, Mannheim, Germany) at a ratio of 3:1. Twenty-four hours later, expression was initiated using 500 μm CuSO4 and cells were used for experiments after a further 16–24 h of incubation. All experiments were repeated with at least three different transfections. For control cells, yellow fluorescent protein was transfected with empty vector.
Chimeric channels were constructed using the In-Fusion-based PCR cloning system (Clontech, Mountain View, CA, USA). PCR fragments of channel with vector backbone and chimeric insert were generated with 15 bp regions of homology at the 3′ and 5′ ends. Vector and insert were combined following the manufacturer's protocol (Clontech). Point mutagenesis was performed using the QuikChange II XL kit according to the manufacturer's protocol (Agilent, Santa Clara, USA). All mutations were verified by sequencing.
For V5 epitope staining, cells were fixed with paraformaldehyde, permeabilized with triton and sequentially incubated with mouse monoclonal anti-V5 antibody (Sigma, Hamburg, Germany) and Alexa Fluor 555 goat anti-mouse IgG (Invitrogen). Cells were mounted on glass slides and imaged with a fluorescent microscope (Leica, Wetzlar, Germany).
Ratiometric calcium imaging was performed using FURA-2 AM dye (Invitrogen) and analysed using Tillvision software (Till Photonics, Munich, Germany). Transfected cells were plated on glass coverslips, loaded with 3 μm Fura-2 AM in calcium imaging buffer (120 mm NaCl, 5 mm KCl, 2 mm CaCl2, 4 mm MgCl2, 10 mm Tes, 40 mm sucrose, 2.5 mm probenecid, pH 7.2 adjusted with NaOH) and placed in a recording chamber that allowed direct application of buffer at controlled temperature (ESF Electronic, Goettingen, Germany). Fluorescence was measured at excitation wavelengths alternating between 340 nm and 380 nm. Following subtraction of background fluorescence, the ratio of fluorescence at 340 nm and 380 nm was calculated. All graphs are averaged ratios of 50–200 individual cells.
Gigaseals were formed with pipettes that had a resistance between 4 and 6 MΩ in standard pipette solution. For whole cell recordings, bath solution contained 135 mm NaCl, 5.4 mm CsCl, 1.8 mm CaCl2, 1 mm MgCl2, 10 mm glucose and 5 mm Hepes, pH 7.2. Pipette solution contained 20 mm CsCl, 120 mm caesium aspartate, 0.01 mm CaCl2, 2 mm MgCl2, 1 mm EGTA, 1 mm Na2ATP and 5 mm HEPES, pH 7.2. The calculated free Ca2+ concentration was 100 nm.
Whole cell membrane currents were monitored using an EPC-10 patch-clamp amplifier and Patchmaster software (HEKA Elektronic, Lambrecht, Germany). All voltages were corrected for a liquid junction potential of 15 mV (S2 cells) using the Henderson equation (Barry & Diamond, 1970). Capacitance and access resistance were continuously monitored and between 10 and 50% of the series resistance was compensated. Data were sampled at 10–20 kHz and filtered at 2–5 kHz.
Slow temperature ramps were applied using a Peltier device (ESF Electronic, Germany) that changed the temperature of the superfusing buffer at 1°C s−1. The temperature was monitored with a thermocouple placed within the flow of buffer and close to the cells. To measure heat activation of channels, cells were held at −30 mV for 20 ms, stepped to −100 mV for 10 ms and ramped to +100 mV (0.5 mV ms−1) every 5 s for 400 s during which time the temperature was raised and then cooled back to baseline. Current amplitude was extracted at −100 mV and +100 mV for each ramp over the temperature range. We observed that repeated heating in the presence of Ca2+ led to desensitization. We therefore selected cells for analysis based upon the following rules. First, only cells that recovered fully to baseline after heating and cooling were chosen. Secondly, to avoid the effect of desensitization, we analysed only the response to the first heat stimulus and recorded no more than three cells per plate. Thirdly, channels were judged to be functionally expressed based upon their response to voltage with a threshold of 100 pA at 100 mV and the presence of a clearly outwardly rectifying current.
Relative permeability ratios were determined by replacing Na+ with test ions in the bath solution while applying voltage ramps at 32°C. For monovalent cations, bath solution contained 140 mm NaCl or 140 mm KCl or 140 mm CsCl, with 10 mm glucose and 10 mm HEPES. pH was adjusted to 7.2 with NaOH, KOH or CsOH, respectively, and junction potentials were compensated as follows: 16 mV for Na+, 10 mV for K+, 10 mV for Cs+. For divalent cations, bath solutions consisted of 90 mm NaCl, 10 mm glucose, 10 mm HEPES, 30 mm CaCl2 (or MgCl2) and pH was adjusted to 7.2 with NaOH. Junction potentials were compensated at 18 mV. Voltage ramps from −100 mV to 80 mV were applied for 2 s to measure the reversal potential. For monovalent ions, permeability ratios to sodium ion: Na+ (PX/PNa) were calculated according to
where VX–VNa is the reversal potential change, F the Faraday constant, R the universal gas constant and T the absolute temperature. For divalent cations, ratios were calculated according to
To measure whole cell voltage-dependent activation of channels, families of voltage-activated currents were acquired at defined temperatures. These consisted of 200 ms voltage steps from −120 mV to 200 mV in 20 mV increments followed by a 50 ms step at −120 mV. Data were extracted only from cells that underwent testing across the full temperature range. Conductance was calculated from the steady-state current and plotted against voltage. Data were fitted with single Boltzmann functions
Single channel recordings were made in the inside-out or cell-attached configuration using gigaseals larger than 10 GΩ and were acquired with an EPC-10 patch-clamp amplifier and Patchmaster software (HEKA Elektronic, Germany). Inside-out recordings were used to measure unitary conductance and plot single channel I–V curves. Bath solution contained 140 mm caesium aspartate, 10 mm EGTA, 10 mm HEPES and 0.4 mm CaCl2, (pH 7.2). Pipette solution contained 140 mm caesium aspartate, 10 mm EGTA, 1 mm MgCl2 and 5 mm HEPES, (pH 7.2). Cell-attached configuration was used to record extended traces for kinetic analysis. The bath solution contained 120 mm KCl, 5 mm HEPES, 1 mm MgCl2 and 2 mm CaCl2 (pH 7.2) and the pipette solution contained 100 mm KCl, 10 mm HEPES, 1 mm MgCl2 and 10 mm EGTA (pH 7.2). Single channel data were sampled at 20 kHz and filtered at 3 kHz.
Only long recordings showing activity of single channels were selected for kinetic analysis. Traces were visually inspected to exclude those with low signal-to-noise, artifacts and run-down or desensitization of currents. Baseline drift was corrected manually. Data were analysed using the QuB software package (http://www.qub.buffalo.edu) and idealized with the segmental k-means algorithm (Qin, 2004). Following manual inspection of idealization, dwell-time histograms were constructed and fitted with sums of exponentials using a maximum likelihood routine. A dead time of 100 μs corresponding to the 10–90% rise time of the filtered record was imposed upon each trace.
To evaluate kinetic models for each channel we used the maximum interval likelihood (MIL) method (Qin et al. 1997) to fit different kinetic schemes of increasing complexity. Both the model search function of QuB and manual generation of models were employed to ensure that we tested all possible non-cyclic arrangements that corresponded to the number of exponentials calculated from dwell-time histograms. In total this amounted to several hundred schemes for each channel, which were ranked according to their log likelihoods and by visual inspection of fitted data. Additionally, for the H990A-R1004N mutant channel, data were idealized using both an open-subconductance-closed model and an open-closed model (where the subconductance state was combined with the open state). Kinetic schemes were evaluated for each idealization, and in all cases, the open-subconductance-closed model fitted the data more adequately. For each channel, the five highest ranking models with their rates are displayed in Supplementary Table S1.
To simulate temperature dependent single channel responses for dTRPA1 and H990A-R1004N we first generated global models for each channel. Data obtained at 24°C and 32°C for dTRPA1 or 34°C and 40°C for H990A-R1004N were simultaneously fitted with Scheme 1 for each channel using the MIL method. All rate constants were allowed to vary and temperature dependence was calculated using the equation
where K(T) is the rate constant at temperature T, k0 is the pre-exponential rate constant, T is the absolute temperature, and k1 is the exponential rate constant. Single channel currents were simulated by applying these parameters to the simulator function of QuB software at a sampling rate of 20 kHz filtered at 3 kHz. Temperature ramps (and 1/T ramps) were generated from 15°C to 45°C at 0.2°C s−1.