Feline Lindhout

VAP–SCRN1 interaction regulates dynamic endoplasmic reticulum remodeling and presynaptic function 117 4 like)-containing proteins for the VAP-binding pockets (Fig 2E). Next, we assessed whether the observed recruitment to VAP at ER structures is shared within the SCRN family. Contrarily, we observed no change in GFP-SCRN2 or GFP-SCRN3 localization when co-expressed with HA-VAPB in COS7 cells (Fig 2F). Together, these data indicate that SCRN1, and not SCRN2 and SCRN3, is recruited to VAP at the ER membrane. SCRN1 interacts with VAP through a single FFAT-like motif Next, we sought to determine the specific domains responsible for the interaction between VAP and SCRN1. We found that the C-terminal coiled-coil region of SCRN1 and the N-terminal major sperm protein (MSP) domain of VAPB are the minimal binding domains required for this interaction, as shown by co-expression experiments in COS7 cells and pull-down analysis of HEK293T lysates (Fig S2D–I, 3A,E, S3B). The MSP domain of VAP contains a FFAT(-like) motif binding site, and FFAT(-like) motifs are found in the majority of the VAP-interacting proteins (Loewen et al, 2003; Murphy & Levine, 2016). Indeed, we found that the FFAT binding motif in VAP is responsible for the interaction with SCRN1. The VAP mutant VAP-K87D/M89D, in which FFAT binding is disrupted, was no longer able to recruit GFP-SCRN1 (Fig 3A,B,E) (Kaiser et al, 2005). Next, we searched for FFAT(-like) motifs in SCRN1 using a previously reported algorithm and identified four potential FFAT- like motifs (Fig S3A) (Murphy & Levine, 2016). We generated SCRN1 constructs with single- point mutations for each single FFAT-like motif (Fig 3A). VAP association was disrupted when mutating the most C-terminal FFAT-like motif in SCRN1 (GFP-SCRN1-F402A), but not the other motifs, as shown by pull-down assays and co-expression experiments (Fig 3C–F, S3B–D). Sequence alignment of the SCRN family members revealed that this newly identified FFAT-like motif in SCRN1 is not shared with the other two SCRN family members, which is consistent with our observation that exogenous VAP is unable to recruit GFP- SCRN2 and GFP-SCRN3 (Fig 2F, 3G). Together, these data show that the MSP domains of VAPA and VAPB interact with a single FFAT-like motif in the C-terminal region of SCRN1. SCRN1 and VAP are required for proper ER morphology Previously, it was shown that VAP interactions with FFAT-containing proteins are engaged in maintaining ER morphology (Kaiser et al, 2005). Next, we aimed to investigate whether its interaction with SCRN1 at the ER membrane could be important for this function. Similar to previously reported results, we found that COS7 cells expressing VAP-K87D/M89D, VAP mutant lacking proper FFAT binding, showed a robust phenotype that was characterized by non-reticular VAP structures throughout the cytosol (Fig 3B, 4A, S4A) (Kaiser et al, 2005). Likewise, similar aberrant VAP-positive structures were found in COS7 cells expressing wild-type HA-VAP and GFP-SCRN1-F402A or GFP-SCRN1-N, both lacking a functional FFAT-like motif (Fig 3C, S3C, 4A, S4A,B). These observed phenotypes were also detected with ER luminal marker TagRFP-ER, indicating that the non-reticular VAP signals represent affected underlying ER structures (Fig S4C,D). Similarly, both in VAP and