We have used patch- and two electrode voltage-clamp recording, site directed mutagenesis and molecular model building to ask whether the tandem pore potassium channel TASK-3 is gated at the ‘helix bundle crossing’. Although tandem pore channels are described as constitutively open, the P(open) of wild type TASK-3 channels, heterologously expressed in CHO cells, was only 0.023 ± 0.0004 (mean ± sem, n=6) at -80mV (140mM-K). P(open) increased with depolarisation, but with a low gating charge (z′=0.16). We have used this voltage dependence to examine effects of mutations on gating. Replacement with threonine of an alanine residue, A237, close to the cytoplasmic end of membrane helix M4, raised P(open) to 0.205 ± 0.031 (n=3) in the same conditions. The voltage dependence of P(open) is shifted to more negative potentials without change in gating charge. The shift gives an energy difference of ~1.2kcal/mol, consistent with the mutant channel being held open by a H-bond (or bonds). Contrary to our previous report, we can find no change in the response of A237T to acidification (measured in channels expressed in Xenopus oocytes) (Ashmole et al., 2005). Gating in response to acidification occurs at the selectivity filter (Yuill et al., 2007). Modelling TASK-1 (which has strong identity with TASK-3) using KvAP (Jiang et al, 2003) as template suggests a H-bond between Thr in position 237 and Asn133, at the cytoplasmic end of M2. However, N133A also has a raised P(open), as does the dual mutant N133A/A237T. Neither mutant form of the channel has an altered response to acidification. Gating at the bundle crossing of helices M2 and M4 is expected to occur around hinge glycines. Such residues are conserved through the tandem pore family, in TASK-3 at positions 117 (M2) and 231 (M4). In G117A and G231A; P(open) is reduced to 0.0021 ± 0.0003 (n=3) and 0.0039 ± 0.0013 (n=4) (respectively). The voltage dependence is shifted positive, again without change in gating charge. In spite of a low P(open), macroscopic currents can be recorded from G117A and G231A (oocyte expression). The pH-sensitivity of these currents is again unaltered from wild type. Voltage dependence is weak and its physiological importance is uncertain. But this voltage dependence appears to involve gating by movements of M2 and M4 similar to those found in other potassium channels. It is likely that modulation of TASK-3 by hormones, neurotransmitters and inhalational anaesthetics (Kim, 2005) occurs by gating at this part of the channel.