The ATP-sensitive potassium channel is an inwardly rectifying potassium channel thought to be involved in neuroprotection and glucose sensing. It is a complex consisting of two different proteins: a Kir 6.x pore-forming subunit and a regulatory sulphonylurea receptor (SUR) subunit. We have investigated the distribution of these subunits in the rat brainstem using a non-fluorescent/fluorescent double-labelling in situ hybridisation protocol.Sprague Dawley rats (150 g) were anaesthetised (Sagatal; 60 mg kg^-1 ip) and transcardially perfused with 4% paraformaldehyde. The brain was removed, cryoprotected, and 30 µm coronal brainstem sections cut using a cryostat. Slices were incubated with Biotin- or Digoxigenin-labelled antisense riboprobes for Kir6.2 or SUR1. Biotin labelled probes were visualised using a Fluorescein-avidin conjugate after amplification with an anti-avidin-biotin conjugate (Vector Labs). Digoxigeninlabelled probes were visualised with anti-Digoxigenin-AP conjugate (Roche) and BCIP/NBT (Sigma). Slices were mounted in Vectashield containing DAPI (Vector Labs).Using this method, Kir6.2 and SUR1 mRNA was detected at varying intensities across the lower brainstem and appeared to be co-localised. The most intensely labelled cells were found in the hypoglossal nucleus where approx. 10% of DAPI labelled cells had mRNA for SUR 1. Within the vagal complex we found that approx. 15% of dorsal vagal nucleus (DVN) cells, 30% of nucleus tractus solitarius (NTS) cells and 30% of area postrema (AP) cells expressed SUR1 mRNA. Comparable expression was observed for Kir6.2. Other areas with a relatively strong signal included the cuneate fasciculus and the reticular nucleus. Elsewhere, staining was at a low level.Observation in the DVN and hypoglossal nuclei at a cellular level revealed that SUR1/Kir6.2 mRNA positive cells were mostly large (>20 µ m), had generally larger nuclei than non-positive cells, and Kir6.2 and SUR1 mRNA were expressed in the same cells.We show that mRNA for SUR1 and Kir6.2 is co-localised and distributed widely in the rat brainstem. The percentage of cells expressing Kir6.2 and SUR1 was lower than indicated by our electrophysiological studies (Kruse Hansen et al., 2003). However, the relative abundance between DVN, NTS and AP remained the same. The discrepancy is most likely due to the pre-selection of large neuronal cell bodies in the electrophysiological study, whereas DAPI labels all cell nuclei irrespective of size and cell type.