Figure 7.

These models serve in part as a basis for the model proposed in Figure 7 and provide a rationalization for the liposomes clustering activities observed for C₁C₂BMUN and C₁C₂BMUNC₂C. However, it is important to note that there are multiple potential explanations for these activities. The findings that PS is a major determinant of vesicle clustering by C₁C₂BMUN (Figure 2) without requiring Ca²⁺ (Figure 2—figure supplement 1), but C₁C₂BMUNC₂C requires Ca²⁺ to cluster V-liposomes (Figure 6A,B), suggest that there are multiple membrane binding sites in these large protein fragments that can cooperate in cis to interact with a single membrane or in trans to bind to two membranes. Indeed, the MUN, C₁ and C₂B domains contain several positive patches, the C₁ domain binds to DAG, and the C₂B domain binds to PIP₂ weakly in the absence of Ca²⁺ and more strongly in the presence of Ca²⁺ (Shen et al., 2005; Shin et al., 2010; Yang et al., 2015). The C₂C domain is likely to have at least one lipid-binding site with moderate affinity that explains the stronger overall liposome clustering activity of C₁C₂BMUNC₂C compared to C₁C₂BMUN (Figure 6—figure supplement 1) (see discussion). Moreover, the sequence spanning residues 1517–1531 at the C-terminus of C₁C₂BMUN does not form part of the MUN domain structure (Yang et al. 2015) and contains a highly hydrophobic sequence that could bind to membranes, but this sequence may become structured due to the presence of the C₂C domain in C₁C₂BMUNC₂C, which could render it unable to bind membranes. We speculate that this hydrophobic sequence together with positive patches in the C₁-C₂B region underlie the liposome clustering activity of Ca²⁺-free C₁C₂BMUN (A), while in C₁C₂BMUNC₂C the C₂C domain provides a PS-binding site that cooperates with the C₁-C₂B region to favor binding in cis to the same membrane (B). Ca²⁺ binding to the C₂B domain may favor membrane binding of C₁C₂BMUNC₂C in a different orientation that facilitates interaction of the C₂C domain in trans with another membrane, which would explain why Ca²⁺-bound C₁C₂BMUNC₂C can bridge V-liposomes; this orientation could also be favored by binding of the C₂B domain to PIP₂ and of the C₁ domain to DAG in T-liposomes (C), leading to the overall notion that the C₁-C₂B region binds to the plasma membrane and the C₂C domain to synaptic vesicle membrane (Figure 7). Extensive studies will be required to test this and other plausible models compatible with the liposome clustering data.

DOI: http://dx.doi.org/10.7554/eLife.13696.022