TRAPP Transport Protein Particle complexes share a

TRAPP (Transport Protein Particle) complexes share a common core architecture and function in FTI 277 HCl and autophagy (Kim et al., 2016). In yeast, the core complex has GEF activity for the Rab1 ortholog Ypt1. The larger TRAPPII and III complexes include regulatory subunits implicated in membrane recruitment. TRAPPIII activates Ypt1 at the Golgi but lacks GEF activity for the redundant Rab11 orthologs Ypt31/32, which are (counterintuitively) activated by TRAPPII at the trans-Golgi network (TGN). TRAPPII also has GEF activity for unprenylated Ypt1 in vitro, although the in vivo relevance of this activity is unclear. In the current issue of Developmental Cell, Thomas et al. (2018) combine in vitro reconstitution experiments with structure-function and complementation analyses to unravel the role of the C-terminal HVDs of Ypt1 and Ypt31/32 in substrate selection by TRAPPII and III. For in vitro GEF assays, prenylated Rabs were purified and reconstituted on liposomes with a Golgi-mimetic lipid composition. To examine GEF-Rab interactions in vivo, an anchor away strategy, dubbed “GRab-IT” for GEF-Rab interaction test, was developed by replacing the cysteine prenylation motifs of the Rabs with the transmembrane domain of a tail-anchored outer mitochondrial membrane protein. Using different fluorescent protein tags, the interaction of TRAPP complexes with wild-type Rabs, or their nucleotide-free (also known as GDP-locked, dominant-negative) mutants predicted to stably associate with cognate GEFs, was detected as co-localization on mitochondria. The functional consequences of Rab mutations, including effects on growth and spatiotemporal dynamics, were assessed by complementation experiments. Swapping the C-terminal HVDs of Ypt1 and Ypt31/32 reversed the interaction specificity for TRAPPII but had little effect on the interaction specificity for TRAPPIII. To determine whether the sequence of the C-terminal HVD of Ypt31/32 is an important determinant, alanine substitutions were introduced in stretches of 3–4 residues (see Figure 1). Length requirements were assessed by GRabIT using truncations and/or insertion of Gly-Ser spacers. Briefly summarizing, Ypt1 or Ypt32 truncations impaired/abrogated interactions with TRAPPIII or TRAPPII, respectively. Ypt32 replacements involving residues GPTISL, which were identified as important by alanine substitution, disrupted TRAPPII interactions. Ypt1 insertions just before the C-terminal cysteines enhanced the interaction as a function of the number of residues inserted. The results support the conclusion that both sequence and length matter. The length requirement is likely related to differences in 3D negative stain-electron microscopy reconstructions of the TRAPPII and TRAPPIII complexes (Tan et al., 2013, Yip et al., 2010). As noted by the authors, the distance from the membrane to the exchange site may differ by ∼20 Å for the two complexes docked in plausible membrane orientations. If fully extended, however, the Rab HVDs would span much larger distances of 100–150 Å. A potential resolution to this conundrum is suggested by the RabGDI complex with prenylated-Ypt1 in which the N- and C-terminal HVDs adopt a compact conformation stabilized by intramolecular interactions with the GTPase domain (Rak et al., 2003). Assuming this compact conformation is maintained after release from RabGDI, the remaining residues would span a much shorter distance of ∼55 Å if fully extended. The sequence dependence suggests a physical interaction with one or more subunits of TRAPPII. Interestingly, the relevant residues (GPTISL) also interact with RabGDI in the complex with prenylated-Ypt1. The specificity of the TRAPP core for Ypt1 in GEF assays is consistent with the in vivo specificity of TRAPPIII for Ypt1 in a context where the length of the C-terminal HVD is sufficient. To explain how the additional regulatory subunits of TRAPPII override the specificity of the core subunits, the authors propose a “steric gating” mechanism whereby Ypt1 is excluded because its C-terminal HVD is too short to reach the exchange site. The C-terminal HVD of Ypt31/32 is long enough and also enhances the interaction with TRAPPII.