Interestingly, the three- and seven-residue insertion mutants not only were unable to rescue the desynchronization of release in syntaxin-1 deficient neurons (measured as the SD of rise times and the coefficient of variation of this SD; Maximov and Südhof, 2005), but also strongly aggravated desynchronization of release (Figure 1E). Moreover, these insertion mutations blocked the ability of syntaxin-1A to rescue release evoked by GABA assay hypertonic sucrose, which monitors the readily releasable pool (RRP) of synaptic vesicles (Rosenmund and Stevens, 1996; Figure 1G). The finding that the three-residue insertion blocks release evoked by an action potential supports the notion that the precise

coupling of SNARE-complex assembly to the TMRs drives fusion-pore opening via formation of a continuous α helix (Stein et al., 2009). However, the fact that spontaneous release is not impaired by the same insertion—as previously observed for synaptobrevin-2 (Deák et al., 2006), and reconfirmed in new experiments for

the present study (Figure S2)—suggests alternative explanations. Clearly the three-residue insertion does not block fusion per se, and the coupling of the SNARE motif to the TMR thus is not essential for fusion as such, but only for the rapid synchronous Ca2+-triggering of fusion. We therefore asked whether the function of syntaxin-1 in fusion actually requires a TMR. In considering this question, we noted that the syntaxin-1 homologs syntaxin-11 and syntaxin-19 contain a palmitoyl-lipid anchor instead of a TMR, suggesting that a SNARE TMR may not be universally Navitoclax datasheet involved in fusion. We replaced the TMR of

syntaxin-1A with the lipid anchor of syntaxin-19 without or with a seven-residue linker in case the precise distance of the SNARE motif from the membrane was important (Figure 2A, referred to as Syntaxin-1AΔTMR and as Syntaxin-1AΔTMR+7i, respectively). We then examined the function of lipid-anchored syntaxin-1A in membrane fusion during synaptic vesicle exocytosis. Strikingly, we found that lipid-anchored syntaxin-1A rescued the loss of spontaneous release at excitatory and inhibitory synapses in syntaxin-1-deficient the neurons (Figures 2B, 2C, S3A, and S3B), as well as the impairment in evoked release in these neurons (Figures 2D–2G and S3C). Syntaxin-1AΔTMR partly reversed the decreased speed of release and fully rescued the desynchronization of release, whereas Syntaxin-1AΔTMR+7i completely rescued both (Figure 2E). Moreover, lipid-anchored syntaxin-1A without or with the seven-residue insertion was fully capable of maintaining sustained release evoked by a 10 Hz stimulus train (Figure 2F), and supported release induced by hypertonic sucrose as a measure of the RRP (Figure 2G). Thus, syntaxin-1A does not need a TMR for promoting synaptic membrane fusion.