The hyperpolarization-activated cyclic nucleotide-gated (HCN) current I(f)/I(HCN) is generally thought to be carried by Na(+) and K(+) under physiological conditions. Recently, Ca(2+) influx through HCN channels has indirectly been postulated. However, direct functional evidence of Ca(2+) permeation through I(f)/I(HCN) is still lacking. To possibly provide direct evidence of Ca(2+) influx through I(HCN)/I(f), we performed inside-out and cell-attached single-channel recordings of heterologously expressed HCN channels and native rat and human I(f), since Ca(2+)-mediated I(f)/I(HCN) currents may not readily be recorded using the whole-cell technique. Original current traces demonstrated HCN2 Ca(2+) inward currents upon hyperpolarization with a single-channel amplitude of -0.87+/-0.06 pA, a low open probability of 3.02+/-0.48% (at -110 mV, n=6, Ca(2+) 2 mmol/L), and a Ca(2+) conductance of 8.9+/-1.2 pS. I(HCN2-Ca2+) was significantly activated by the addition of cAMP with an increase in the open probability and suppressed by the specific I(f) inhibitor ivabradine, clearly confirming that Ca(2+) influx indeed was conducted by HCN2 channels. Changing [Na(+)] (10 vs. 100 mmol/L) in the presence or absence of 2 mmol/L Ca(2+) caused a simple shift of the reversal potential along the voltage axis without significantly affecting Na(+)/Ca(2+) conductance, whereas the K(+) conductance of HCN2 increased significantly in the absence of external Ca(2+) with increasing K(+) concentrations. The mixed K(+)-Ca(2+) conductance, however, was unaffected by the external K(+) concentration. Notably, we could also record hyperpolarization-activated Ca(2+) permeation of single native I(f) channels in neonatal rat ventriculocytes and human atrial myocytes in the presence of blockers for all known cardiac calcium conduction pores (Ca(2+) conductance of human I(f), 9.19+/-0.34 pS; amplitude, -0.81+/-0.01 pA; open probability, 1.05+/-0.61% at -90 mV). We directly show Ca(2+) permeability of native rat and, more importantly, human I(f) at physiological extracellular Ca(2+) concentrations at the physiological resting membrane potential. This might have particular implications in diseased states with increased I(f) density and HCN expression.