Feline Lindhout

VAP–SCRN1 interaction regulates dynamic endoplasmic reticulum remodeling and presynaptic function 119 4 F153A, or GFP-SCRN1-F402A. Normalized intensity plot represents indicated line (dotted).Scale bars: 10 μm (full size) and 5 μm (zoom). D. Hippocampal neurons (DIV16) co-expressing HA-VAPB with GFP-SCRN1-Y40A, GFP-SCRN1- F144A, GFP-SCRN1-F153A, or GFP-SCRN1-F402A. Scale bars: 10 μm (full size) and 5 μm (zoom). E. Quantifications of SCRN1 recruitment to VAPB-positive structures in COS7 cells (%). Left graph: co-expression of GFP-SCRN1 with HA-VAPB or HA-VAPB-K87D/M89D (N = 2–3, n = 44–46). Right graph: co-expression of HA-VAPB with GFP or GFP-SCRN1-FL, GFP-SCRN1-N, GFP-SCRN1-C, GFP-SCRN1-Y40A, GFP-SCRN1-F144A, GFP-SCRN1-F153A, or GFP-SCRN1-F402A (N = 2–3, n = 41–64). F. Pull-down assay of HEK293T cells co-expressing Myc-VAPB with GFP, GFP-SCRN1-WT, GFP- SCRN1-Y40A, GFP-SCRN1-F144A, GFP-SCRN1-F153A, or GFP-SCRN1-F402A. G. Sequence alignment of the C-terminal human SCRN1, SCRN2, and SCRN3. Amino acid position 3 (asterisk) in FFAT-like motif (orange) of SCRN1 is not shared in SCRN2 and SCRN3. Data information: ***P < 0.001, by chi-square test with post hoc analysis including Bonferroni correction. Source data are available online for this Figure. in SCRN1 knockdown neurons, as well as in neurons expressing GFP-SCRN1-F402A, ER morphology was severely disrupted (Fig 4B). Notably, in these neurons we observed dense ER patches surrounded by less dense or absent ER structures. On the other hand, neurons overexpressing wild-type SCRN1 also showed dense ER patches; however, these did seem properly connected to the rest of the ER structure. Thus, structural disruptions in ER morphology were consistently observed when the VAP-SCRN1 interactions were abrogated (Fig 4A). As such, expression of SCRN1-F402A showed the same phenotype on ER morphology as SCRN1 depletion, suggesting that SCRN1-F402A may act as a dominant-negative. As SCRN1-F402A is cytoplasmic, it could recruit and capture endogenous SCRN1 to the cytoplasmic pool, thereby preventing it from binding to VAP and execute its function at the ER. Indeed, oligomerization is a common feature of the Ntn aminohydrolases superfamily, and pull-down assays showed that both SCRN1 and SCRN1- F402A were associated with other SCRN1 proteins (Fig S4E,F). These results confirm that SCRN1 undergoes oligomerization and that SCRN1-F402A could act as a dominant- negative. Together, these data indicate that both SCRN1 and VAP are required for proper ER morphology, which is mediated by VAP-SCRN1 interactions at the ER membrane. SCRN1 and VAP are engaged in maintaining ER continuity and dynamics The ER structure undergoes constant remodeling while remaining continuous for proper functioning. Thus, next we sought to determine the effect of VAP and SCRN1 on both ER continuity and dynamics. We used live-cell imaging to map ER dynamics in cells expressing luminal ER marker TagRFP-ER. We observed fast remodeling of ER structures in both neurons and COS7 cells, ranging from ER tubule growth events and structural ER “wiggling” events (Fig 4C,D, S4G). The dense ER patches observed when expressing wild-