The HERG potassium channel is a voltage-gated potassium (Kv) channel, which plays a key role in the repolarisation of the cardiac action potential. Like all Kv channels, it is a tetramer consisting of a central ion conducting pore domain (S5-P-S6) surrounded by four voltage-sensing domains (S1-S4). However, unlike most Kv channels, HERG conducts the majority of its current upon repolarisation of the membrane. This is attributed to its unique gating properties: during depolarisation, the channel’s inactivation gates close before its activation gates begin to open, thereby preventing current flow. During hyperpolarisation, inactivation gates open very rapidly, but the activation gates close so slowly that significant amount of current flows before the activation gates close. The molecular mechanism(s) underlying these unique biophysical properties are far from clear. Recent studies on the Shaker Kv channel indicate that the voltage sensing S4 segment interacts with S5 of the pore domain and that these interactions may couple voltage sensor motion to the opening and closing of channel gates. Although similar coupling mechanisms can be envisaged for the HERG channel, there is a paucity of information on the membrane-aqueous boundaries of the transmembrane segments to allow investigation of such mechanisms. In this study, we have used substituted cysteine accessibility method to investigate the outer membrane borders of the HERG channel. HERG channels containing engineered cysteine mutations were expressed in Xenopus oocytes and the effect of extracellular application of the membrane-impermeable cysteine modification reagent p-chloromercurybenzenesulfonic acid (pCMBS) on currents was measured using two-electrode voltage clamp. Application of 100 μM pCMBS to 524C (S4 mutant) abolished the channel activity fully, but only when applied during depolarisation (n=3). This indicates that C at 524 moves out of the membrane electric field and becomes exposed to the extracellular solvent during depolarisation. Currents through mutants with cysteines at deeper positions, 527 to 529 of the S4, were also inhibited significantly (55.7±8.9 %, 22.5±4.5 %, 67.8±12.4 % and 47.8±8.5 %, for 526C, 527C, 528C and 529C respectively [mean ± s.e.m.] n=4-7; p<0.05, Student's paired t-test). pCMBS, however, has no significant effect on 530C (n=4). These results suggest that the extent of S4 exposure to the extracellular phase is likely to be similar to that of the Shaker channel, and that the first and second positively charged residues in HERG S4 (K525 and R528) align with the first and second positively charged residues in the Shaker S4 (R362 and R365) and not the second and third positively charged residues (R365 and R368) as has been suggested in previous sequence alignments.