Reply to Benndorff and DiFrancesco: Reliable human HCN4 single-channel recordings using the cell-attached configuration in expression systems

Abstract

We thank Benndorff and DiFrancesco for their interest in our work and stimulating letter (1) in which they comment on the differences between our unitary funny current (If) recordings (2), and others reported in sinoatrial cells, pyramidal neurons, oocytes (XO) or HEK-293 cells (1). Based on our previous experience (3, 4), we anticipated the difficulties in single-channel HCN4 recordings suggested by the controversy among several laboratories (5–8). Therefore, we were extremely rigorous when developing these experiments. We selected CHO instead of HEK-293 cells to avoid the interference of endogenous K channels and further confirmed that nontransfected CHO cells do not generate any interfering current. After each experiment, we perfused the selective If inhibitor ivabradine, which abolished the current in WT and p.V240M HCN4-transfected cells. Furthermore, recordings with more than one channel were discarded (2). Unfortunately, we do not know the exact reasons behind the questioned differences, but they are possibly due to great disparities in the experimental conditions used, including types of cells and channels expressed, patch-clamp configurations, and composition of the solutions (e.g., we added Na+ to external and internal solutions). DiFrancesco’s impressive pioneering recordings were carried out in rabbit sinoatrial cells using whole-cell (which dilutes any critical endogenous mediator because of the intracellular dialysis) combined with cell-attached (5), or inside-out (9) patch-clamp configurations. Besides a low If-channel expression (5), in native cardiomyocytes was suggested the presence of heteromeric HCN channels with dissimilar properties than homomeric channels. Conversely, we recorded single-channel currents generated by homomeric human HCN4 (NM_005477.3) channels using the cell-attached configuration to keep the intracellular media intact preventing run-down. Regarding the disparities with the results obtained in other cell types, apart from biophysical differences between HCN4 and HCN2 channel properties, it is known that the processing and function of HCN4 channels are species-dependent (8). So, in CHO but not in other cells, HCN4 channel activation is constitutively shifted to depolarized potentials and no longer affected by exogenous cAMP due to the expression of an endogenous, isoform-specific modulator of HCN4 (10). Seemingly, If-channels could exhibit long, low-conductance and short, high-conductance openings (8). Our data were filtered at 1 kHz, which, together with the capacitive artifacts, prevented us from reliably detecting small currents at the beginning of the pulses. Moreover, latency upon the first opening is voltage-dependent and short at −90 mV (the membrane potential for our recordings) for the If recorded in rat hippocampal neurons (which exhibits a 9.7 pS conductance) (11). We prioritized the analysis of robust steady-state currents over the search for slow time-dependent activation. Additionally, although long openings were detected, they were not frequent enough to appear in dwell-time histograms. In conclusion, we did not intentionally ignore seminal previous works (1), but due to space limitations, we mainly discussed the coincidence in results obtained by two unrelated independent laboratories using quite similar experimental procedures (2, 6, 7). Therefore, without wishing to polemicize, we reinforce the validity of our single-channel experiments that are supported and give support to the rest of clinical, electrophysiological, and biochemical results.