According to the SNARE hypothesis, every membrane compartment that utilizes this fusion machinery should contain a specific target-SNARE (t-SNARE), which allows docking and fusion only of transport vesicles possessing a matching v-SNARE. Plasma membrane t-SNAREs appear to consist of two subunits: one is a member of the syntaxin family and the other a member of the SNAP-25 family. Recently, the distribution of some syntaxin isoforms has been studied in MDCK, pancreatic acinar and gastric parietal cells, which revealed their differential distribution at the apical and basolateral plasma membrane domains. Syntaxin 4 is restricted to the basolateral domain in both MDCK and acinar cells. Syntaxin 2 was found at both domains in MDCK cells but appeared to be only apical in acinar cells. Syntaxin 3 was studied in all three cell types and was found at the apical domain in MDCK (with some additional lysosomal localization) and possibly also in acinar cells. Interestingly, syntaxin 3 could also be detected on the large secretory granules that ultimately fuse with the (small) apical plasma membrane of acinar cells. Moreover, in parietal cells, at least some syntaxin 3 was localized to the H+/K+-ATPase-containing tubulovesicles that fuse with the apical membrane after gastric stimulation. It is not clear whether this intracellular pool of syntaxin 3 arises from insufficient retention during membrane retrieval from the apical surface or whether it has a specific function. It is striking that, in all cases, syntaxin 3 may be involved in an apically directed pathway.