False interaction of syntaxin 1A with a Ca(2+)-activated K(+) channel revealed by co-immunoprecipitation and pull-down assays: implications for identification of protein-protein interactions.

The techniques of co-immunoprecipitation and immunocytochemical co-labelling are classically used to identify protein-protein interactions. We have used an antibody to the rat small conductance calcium-activated potassium channel subtype 1 (rSK1) to immunoprecipitate proteins from rat brain. A 35 kDa protein was recognized by two monoclonal antibodies to syntaxin 1 and a polyclonal antibody to syntaxin 1A, but not by antibodies to syntaxins 2, 3 or 4. These data suggested that syntaxin 1A is specifically associated with rSK1 in rat brain. A GST construct of the carboxyl terminus of rSK1 was able to pull-down syntaxin 1A from rat brain. Immunocytochemistry showed somatic labelling for both rSK1 and syntaxin 1A in acutely dissociated hippocampal CA1 neurons, confirming that these proteins could interact in vivo. However, control immunoprecipitations showed that antibodies to eight potassium channels could also immunoprecipitate syntaxin, even though some of these channels would not be expected to reside in the same subcellular compartment. Mock immunoprecipitations and pull-down assays showed that syntaxin 1 could directly interact with sepharose and agarose resins. Hence immunoprecipitation and pull-down assays do not provide evidence that syntaxin is specifically associating with a protein, placing doubt on a number of reported interactions with syntaxin 1A.