The fifth adaptor protein complex

Abstract

Adaptor protein (AP) complexes sort cargo into vesicles for transport from one membrane compartment of the cell to another. Four distinct AP complexes have been identified, which are present in most eukaryotes. We report the existence of a fifth AP complex, AP-5. Tagged AP-5 localises to a late endosomal compartment in HeLa cells. AP-5 does not associate with clathrin and is insensitive to brefeldin A. Knocking down AP-5 subunits interferes with the trafficking of the cation-independent mannose 6-phosphate receptor and causes the cell to form swollen endosomal structures with emanating tubules. AP-5 subunits can be found in all five eukaryotic supergroups, but they have been co-ordinately lost in many organisms. Concatenated phylogenetic analysis provides robust resolution, for the first time, into the evolutionary order of emergence of the adaptor subunit families, showing AP-3 as the basal complex, followed by AP-5, AP-4, and AP-1 and AP-2. Thus, AP-5 is an evolutionarily ancient complex, which is involved in endosomal sorting, and which has links with hereditary spastic paraplegia.

Figure 1. Overview of AP complexes.

(a) Structure of an AP complex, showing the positions of the four subunits, and indicating some of the domains on the μ and β subunits. (b) List of subunits in the four AP complexes and F-COPI. (c) Diagram of trafficking pathways and machinery. We have called the AP-1- and AP-3-positive endosomes “tubular endosomes” because this is how they appear by electron microscopy ,; functional names such as recycling or early endosomes are more contentious.

Figure 2. C14orf108 interacts with DKFZp761E198.

(a) Secondary structure predictions for the four AP μ subunits, the COPI subunit δ-COP, C14orf108, stonin2, and FCHO2. α helices are shown in red and β strands in blue. The domains in μ1 that bind to the β subunit and to cargo proteins are indicated; similar domains are found not only in the other μ-adaptins and δ-COP, but also in C14orf108. In contrast, stonin2 and FCHO2 share only the cargo-binding (μ-homology) domain (MHD). (b) Yeast two-hybrid interactions between C14orf108 and DKFZp761E198. Two clones encoding residues 16–229 of DKFZp761E198, isolated from a human placental cDNA library screen using C14orf108 as bait, were tested for specificity. “C+” rows contain a positive control; “C−” rows contain various negative controls using the empty bait vector (pB66ø), the empty prey vector (pP6ø), or both. Growth in the absence of histidine (SM –trp –leu –his) or in the presence of 3-aminotriazole (3-AT) indicates that the two gene products interact. (c) Secondary structure predictions for the four AP β subunits, the COPI subunit β-COP, and DKFZp761E198. The α-helical solenoid and appendage domains in β1 are indicated. Similar domains are found in the other family members, and also in DKFZp761E198. The diagram also shows that although DKFZp761E198 lacks a long unstructured linker separating the solenoid and appendage domains, its appendage domain contains both the β-sandwich subdomain and the α/β platform domain, unlike β4 which has only the platform subdomain (or γ, which has only the sandwich subdomain—see Figure 8a).

Figure 3. Characterisation of C14orf108 by Western blotting and immunofluorescence.

(a) A rabbit antiserum raised against a C14orf108-derived fusion protein recognises a band of the appropriate size on Western blots of HeLa cell homogenates, which is reduced in intensity when the cells are treated with SMARTpool siRNAs targeting the protein, and also when cells are treated with SMARTpool siRNAs targeting DKFZp761E198. For this and all subsequent siRNA experiments, control cells were treated with RISC-free control siRNA. (b) C14orf108 partitions between membranes and cytosol when cell homogenates are centrifuged at high speed, and does not appear to be associated with clathrin-coated vesicles (CCVs). (c) In cells transfected with GFP-tagged C14orf108, most of the construct is cytosolic. (d) If the cells are treated with Oligo-10, an siRNA that depletes endogenous C14orf108, the tagged construct has a punctate distribution in a limited number of cells. (e) Double labelling for tagged C14orf108 and LAMP1 in cells treated with Oligo-9. By moderating protein levels, treatment with Oligo-9 highlights the membrane association of the construct. There is substantial overlap between tagged C14orf108 and LAMP1. (f and g) Cells treated with Oligo-9 were double labelled for tagged C14orf108 and AP-3, either without (f) or with (g) a 5-min incubation in 20 µM brefeldin A (BFA). Unlike AP-3, C14orf108 is insensitive to BFA. The need to use anti-GFP, instead of relying on GFP fluorescence, indicates that AP-5 is not very abundant. Scale bars: 20 µm.

Figure 4. Phenotype of siRNA-treated cells.

(a–c) Double labelling for the CIMPR and the retromer subunit Vps26 in control cells (a), cells depleted of C14orf108 (b), and cells depleted of DKFZp761E198 (c). Both knockdowns cause a change in the localisation of the CIMPR and Vps26: the individual puncta are larger and brighter, and there are fewer of them per cell (see Figure 8c for quantification). Scale bar: 20 µm. (d and e) Electron micrographs of the Golgi (G) region of control (d) and C14orf108-depleted (e) cells. Knocking down C14orf108 causes the cells to accumulate swollen multivesicular bodies (MVBs), which have tubules emanating from them (black arrows) and flat clathrin bilayer coats (white arrows). Scale bar: 500 nm. (f) Immunogold labelling of C14orf108-depleted cells. The MVBs are positive for the CIMPR, indirectly labelled with 15 nm (left) or 10 nm (right) gold particles. The arrows indicate tubules and vesicles emanating from the MVBs. Scale bar: 200 nm.

Figure 5. Distribution and proposed evolution of the AP-5 complex in eukaryotes.

(a) Coulsen plot with the coloured sectors denoting the presence of μ5, β5, ζ, σ5, SPG11, and SPG15 homologues in 12 of the 29 genomes sampled, spanning the extent of eukaryotic diversity. The light pink shaded sections in Metazoa indicate that an orthologue was not identified in N. vectensis or M. brevicollis but was identified in the nr database from the taxon listed parenthetically. (b) Deduced evolutionary history of the AP-5 complex as present in the Last Eukaryotic Common Ancestor but lost multiple times independently.

Figure 6. Order of evolutionary emergence for the adaptin protein families.

Phylogenetic analysis of a concatenated dataset of β-adaptin and μ-adaptin homologues from diverse eukaryotes, rooted with the β and δ COP homologues. The different complexes are colour coded in the same way as in Figures 1c and 10. AP-3 is demonstrated here to be the earliest diverging adaptor complex, highlighted by the shaded box showing its exclusion from the other AP families. The best Bayesian topology is shown with support values listed in the order of Posterior probability values and ML bootstrap support values for PhyML and RAxML. Numerical values are given for the backbone nodes and for the monophyly of each protein family (illustrated by the shaded boxes). Other values are replaced with symbols as inset.

Figure 7. Schematic drawing illustrating the order of duplications giving rise to the adaptor complexes.

The green circle denotes the hypothetical origin of a primordial hybrid organelle linking the secretory system and the endocytic system, while the purple circle denotes the specialisation into a primordial TGN compartment.

Figure 8. Candidates for other subunits of the AP-5 complex.

(a) KIAA0415/SPG48 and C20orf29, which have been shown to coimmunoprecipitate with μ5 and β5 , have similar predicted secondary structures to the γ/α/δ/ε large subunits and to the small subunits, respectively, although KIAA0415/SPG48 lacks an appendage domain and C20orf29 has longer loops between the folded domains. (b) Knockdown of either KIAA0415/SPG48 or C20orf29 phenocopies μ5 and β5 knockdowns, causing both the CIMPR and Vps26 to localise to larger puncta. Scale bar: 20 µm. (c) Quantification of the changes in CIMPR immunofluorescence observed in cells treated with siRNAs targeting each of the four putative subunits. The changes were quantified using an automated microscope and normalised to the control. In every case, the spots are larger and brighter, and there are fewer of them per cell.

Figure 9. Further evidence for a heterotetrameric complex.

(a) Localisation of GFP-tagged C20orf29. Transiently transfected cells were fixed, extracted with saponin, and double labelled for GFP and LAMP1. There is substantial overlap between the two proteins. (b) Quantification of overlap. The level of overlap between either C14orf108-GFP or C20orf29-GFP, and either LAMP1 and the CIMPR, was quantified using Volocity software, from which Pearson's correlation coefficient was determined. The Pearson's coefficients were 0.850±0.023 for C14orf108-GFP and LAMP1; 0.627±0.029 for C14orf108-GFP and the CIMPR; 0.789±0.003 for C20orf29-GFP and LAMP1; and 0.565±.032 for C20orf29-GFP and the CIMPR. (c) Immunoprecipitation of AP-5 subunits from cells expressing GFP-tagged C20orf29. Control HeLa cells and cells expressing C20orf29-GFP were lysed and immunoprecipitated with anti-GFP, then probed with antibodies against various proteins. In the cells expressing the construct, the anti-GFP antibody not only brings down the construct itself, but also C14orf108 and DKFZp761E198. The anti-DKFZp761E198 antibody produces a high background of non-specific bands that are unaffected by siRNA knockdown, but it labels a band of the expected size (arrow) after the complex is enriched by immunoprecipitation. The blot was also probed with an antibody against the AP-1 γ subunit as a control to ensure that the immunoprecipitation was specific. (d) Proposed organisation of the AP-5 complex, conforming to the established nomenclature for AP subunits: we are calling the other large subunit (KIAA0415/SPG48) ζ, the next letter in the Greek alphabet after ε; and we are calling the small subunit (C20orf29) σ5.

Figure 10. Updated diagram of trafficking machinery and pathways.

AP-5 is unusual in that it is the only coat identified so far that associates with late endosomes.