Nature Structural Biology
9, 527 - 531 (2002)
Published online: 3 June 2002; | doi:10.1038/nsb808
Structural basis for the accessory protein recruitment by the  -adaptin ear domainTerukazu Nogi1, Yoko Shiba2, 3, Masato Kawasaki1, Tomoo Shiba1, 4, Naohiro Matsugaki1, Noriyuki Igarashi1, Mamoru Suzuki1, Ryuichi Kato1, Hiroyuki Takatsu2, Kazuhisa Nakayama2, 3
& Soichi Wakatsuki11 Photon Factory, Institute of Materials Structure Science, High Energy Accelerator Research Organization (KEK), Tsukuba, Ibaraki 305-0801, Japan. 2 Institute of Biological Sciences and Gene Research Center, University of Tsukuba, Tsukuba, Ibaraki 305-8572, Japan. 3 Present address: Faculty of Pharmaceutical Sciences, Kanazawa University, Kanazawa, Ishikawa 920-0934, Japan. 4 Foundation for Advancement of International Science (FAIS) Tsukuba, Ibaraki 305-0062, Japan.
Correspondence should be addressed to Soichi Wakatsuki soichi.wakatsuki@kek.jpThe adaptor proteins AP-1 and GGA regulate membrane traffic between the trans-Golgi network (TGN) and endosomes/lysosomes through ARF-regulated membrane association, recognition of sorting signals, and recruitment of clathrin and accessory proteins. The  1-adaptin subunits of AP-1 and GGA possess homologous ear domains involved in the recruitment of accessory proteins, -synergin and Rabaptin-5. The crystal structure of the human 1-adaptin ear domain consists solely of an immunoglobulin-like fold, unlike the  -adaptin ear domain. Structure-based mutational analyses reveal a binding site for the accessory proteins that is composed of conserved basic residues, indicating that the recruitment mechanism in 1-adaptin and GGA is distinct from that in -adaptin.
Clathrin-mediated membrane traffic is responsible for protein transport from the plasma membrane and trans-Golgi network (TGN) to the endosomal/lysosomal system. Membrane recruitment of clathrin is mediated by heterotetrameric adaptor protein (AP) complexes, which interact with cargo proteins1,
2. AP-1 is one of the AP complexes composed of  (1)-,  1-,  1- and  1-adaptins. The C-terminal protruding part of 1-adaptin, often referred to as an ear domain, is involved in the recruitment of accessory proteins, such as -synergin, Rabaptin-5 and cyclin G-associated kinase, which modulate the functions of AP-1 in the membrane trafficking events3,
4,
5. Similar to endocytic accessory proteins for  -adaptin of the AP-2 complex, -synergin possesses an Eps15 (a phosphorylation substrate of the epidermal growth factor receptor kinase) homology domain and interacts with Asn-Pro-Phe motifs of secretory carrier membrane protein 1 (SCAMP1), which was suggested to serve as an assembly site for the clathrin coat on the donor membrane6. In contrast, Rabaptin-5 is involved in the membrane fusion of transport vesicles at the target membranes.
Until recently, the transport of mannose 6-phoshate-modified lysosomal hydrolases from the TGN to endosomes/lysosomes have been believed to be mediated by AP-1; however, a recent study has suggested that AP-1 is responsible for retrograde endosome-to-TGN transport7. A novel family of adaptor proteins possessing a ear-like domain, termed GGAs (Golgi-localizing, -adaptin ear domain homology, ADP-ribosylation factor (ARF)-binding proteins), have been implicated in the anterograde transport from the TGN4,
8,
9,
10,
11,
12. GGAs are monomeric proteins having four functional domains (from the N- to C-terminus): Vps27p/Hrs/STAM (VHS), GGA and Tom1 (GAT), hinge and -adaptin ear (GAE). The C-terminal GAE domain seems to share a common function with the 1 ear domain — that is, the recruitment of accessory proteins. The GAE domain interacts with -synergin and Rabaptin-5 (refs 4,5,11), suggesting the general importance of the ear-like domain in the membrane traffic between the TGN and endosomes/lysosomes. The N-terminal VHS domain, however, recognizes the acidic dileucine sequence of cargo receptors, such as mannose 6-phosphate receptors and sortilin13,
14,
15,
16. The signal recognition mechanism of the VHS domain has already been demonstrated by X-ray crystallography and biochemical analyses17,
18.
Here we report the crystal structure of the human 1-adaptin ear domain at 1.8 Å resolution, which is the first three-dimensional structure determined in the AP-1 subunits. Based on the resulting structure, we performed mutational analyses to identify the binding site for accessory proteins. The results reveal a novel mechanism of accessory protein recruitment through the 1 ear and GAE domains, distinct from the proposed model for the -adaptin ear domain in endocytosis.
Structure description
The crystal structure of the human 1 ear domain was determined by a combination of SIRAS and MAD phasing (Fig. 1). The resulting structure shows that the 1 ear domain forms an immunoglobulin-like -sandwich fold composed of eight -strands with two short -helices (Fig. 2a). Each strand is composed of residues 707−712 in 1, 715−723 in 2, 730−739 in 3, 746−753 in 4, 759−762 in 5, 778−785 in 6, 795−802 in 7 and 805−812 in 8. Two -sheets of the sandwich structure consist of the strands 1, 2, 3, 5 and 6 and the strands 4, 7 and 8. The helix 1 is composed of residues 772−774 and located between the strands 5 and 6, and 2 is residues 818−820 after 8. The topology of the entire 1 ear domain is similar to those of the N-terminal subdomains in the - and 2-adaptin ear domains of the AP-2 complex19,
20,
21.
 | | |
| | |
A preliminary sequence comparison suggests that the 1 ear domain is only half as long as the and 2 ear domains and that it might have an immunoglobulin-like fold. However, the 1 ear domain has very low sequence identity ( 10%) and homology (20%) to the N-terminal immunoglobulin-like subdomains of the and 2 ear domains. Biochemical analyses of the and 2 ear domains have indicated that their C-terminal platform subdomains, rather than the N-terminal subdomains, are critical for the interaction with the accessory proteins19,
20,
21. Our results confirm that the 1 ear domain indeed folds into an immunoglobulin-like structure but lacks the C-terminal platform subdomain, which leaves the question of how the 1 ear domain recruits the accessory proteins.
Conservation of the 1-ear and GGA GAE domains
A structure-based sequence alignment of the 1 ear domain and the GAE domain reveals conservation of the -sandwich structure (Fig. 3). Several deletions and insertions are present between the 1 ear and GAE domains, but all of them are attributed to the loop regions between -strands. Most of the conserved hydrophobic residues are buried in the core of the molecule, suggesting that they contribute to the structural integrity of the -sandwich structure. According to the sequence alignment, residues 703−822 of the 1 ear domain, which form a rigid -sandwich fold in the resulting structure, show a sequence identity of 34% and homology of 47% to the corresponding regions of the GAE domains of the human GGAs. Moreover, the GAE domains do not possess additional C-terminal residues downstream of the putative rigid -sandwich structure. Thus, the GAE domain is likely to have a similar immunoglobulin-like domain and lack the C-terminal platform regions.
| | Figure 3. Conserved basic surface. | | |  | a, Sequence alignment of ear and GAE from GGA. The numbering for the human 1 ear domain follows the sequence data (GenBank/EBI/DDBJ entry AB015317). In this sequence, the total number of amino acid residues is 822. Human and mouse genomes have 2-adaptin in addition to 1-adaptin. The sequences indicated as plant, yeast and fungus correspond to those of Arabidopsis thaliana, Saccharomyces cerevisiae and Ustilago maydis, respectively. The diagram above the sequences depicts the secondary structures of the human 1 ear domain. The residues conserved among at least 9 out of 12 sequences are colored in red for identity and pink for similarity. In addition, the residues in the vicinity of the highly conserved basic cluster, as shown in (b), are highlighted in blue. Triangles on the sequences indicate the residues whose point mutations were shown to abolish the interaction with -synergin and Rabaptin-5 in yeast two-hybrid screen. The red triangles are for the residues buried in the core of the 1 ear molecule and blue for the exposed residues. In addition, the green triangles indicate the additional point mutants that were designed for GST pull-down assays by the structural data. b, Electrostatic surface potential of the human 1-adaptin ear domain. The molecular surface prepared by GRASP30 is shown in the same orientation as in Fig. 1b. A large basic surface consists of the conserved residues around the C-terminus of strand 4 and the N-terminus of strand 7. The GST pull-down assays in Fig. 4 have indicated that the basic surface serves as the binding sites for -synergin and Rabaptin-5.
 Full Figure and legend (144K) |
| |
Further analysis of the sequence alignment and the X-ray structure reveals a striking feature in the conservation of the charged amino acid residues. The basic residues at the N-terminus of the strand 7 (Arg 793, Arg 795 and Lys 797 in the human 1 ear domain) are highly conserved and form a cluster on the molecular surface. In addition, a sequence at the C-terminus of the strand 4, Ala 753-Val-Pro-Lys 756, is almost completely conserved in all known -adaptin and GGA sequences, and the side chain of Lys 756 is located near the basic cluster. These conserved residues constitute a large basic surface around the C-terminus of strand 4 and the N-terminus of strand 7 (Fig. 2). In contrast, the N-terminal immunoglobulin-like subdomains of the and 2 ear domains do not have conserved basic residues in the corresponding surface area, suggesting that the basic cluster is a unique feature of the ear-like domain.
Binding site for accessory proteins
To identify the residues that interact with accessory proteins, we carried out a reverse two-hybrid screening of 1 ear mutants. By this screening, many mutations in the 1 ear domain were found to abolish the ability to bind Rabaptin-5 (Fig. 3). Most of those identified were introduced into the hydrophobic or uncharged residues embedded in the core of the 1 ear molecule, namely the residues that would not participate in the molecular interaction. These mutations might disrupt the -sandwich structure, which would abolish the binding of accessory proteins altogether.
The screening also highlighted the importance of charged residues in binding to Rabaptin-5, including residues Arg 793, Lys 797 and Glu 812. Arg 793 and Lys 797 are the constituents of the conserved basic cluster, and Glu 812 is located near the cluster. Therefore, we introduced single mutations to all of the basic residues in the cluster (K756Q, R793Q, R795Q and K797Q) and performed a GST pull-down assay, assuming that the conserved basic surface is involved in the interaction with accessory proteins. We also subjected Ala 753 (A753Q) and Glu 812 (E812K) mutants to the pull-down assay; the former is the first residue of the highly conserved sequence, Ala 753-Val-Pro-Lys 756, and is surrounded by the three basic residues in the cluster, Lys 756, Arg 795 and Lys 797.
The 1 ear domain could no longer bind to -synergin if any of the basic residues (Lys 756, Arg 793, Arg 795 and Lys 797) were mutated (Fig. 4), indicating the importance of the basic surface in binding to the accessory protein. Furthermore, the E812K mutation abolishes the binding to -synergin. Glu 812 interacts with the proximal basic residues, Arg 793 and Arg 795, via salt bridges in the present crystal structure, suggesting that the E812K mutation perturbs the spatial arrangement of the basic residues in the binding surface. The A753Q mutation moderately affects binding. With the small side chain, Ala 753 might increase the accessibility of the binding surface to the accessory protein. Therefore, the bulky side chain of the Gln residue might perturb the interaction, at least in the case of -synergin. However, previous binding analyses have shown that the middle portion of human -synergin (residues 518−786) directly interacts with the 1 ear domain3. -Synergin possesses five acidic sequences, Asp-Asp-Phe-X-Asp/Gln-Phe (X represents any amino acid), three of which are located in the 1 ear-binding region. Provided that the acidic sequence serves as a 1 ear-binding motif, the basic surface of the 1 ear domain would be suitable for the signal recognition through electrostatic interactions.
| | |
For Rabaptin-5, similar results were obtained, except for the A753Q and R795Q mutations. The R795Q mutation had a marginal effect on the binding to Rabaptin-5 even though Arg 795 is located at the center of the basic cluster. Our recent two-hybrid analysis5 has revealed that the 1 ear domain interacts with the C-terminal coiled-coil region of Rabaptin-5. However, no acidic sequence similar to the Asp-Asp-Phe-X-Asp/Gln-Phe of -synergin is identified in this region. Although the binding mode of Rabaptin-5 seems different from that of -synergin, the above results strongly indicate that the basic cluster is critical for interaction in both cases.
In contrast, previous structural analyses of the -adaptin ear domain indicated that the N-terminal immunoglobulin-like subdomain serves only as the spacer between the hinge region and the C-terminal platform subdomain19,
20. Furthermore, the ear domain possesses a shallow hydrophobic pocket between the -sheet and the sheet-crossing -helix on the top of the C-terminal subdomain, which recognizes the Asp-Pro-Phe/Trp motifs of endocytic accessory proteins19. Taken together, these findings indicate that 1-adaptin and GGA interact with the accessory proteins in a manner different from that observed in -adaptin of the AP-2 complex. Ultimate confirmation of the recognition mechanism of the accessory proteins awaits the crystal structure of the complex between the 1 ear domain and -synergin or Rabaptin-5. Nevertheless, the combination of the 1 ear structure and the structure-based mutational analyses strongly suggest a novel mechanism in which 1-adaptin and GGA recruit the accessory proteins through the basic cluster of the common immunoglobulin-like ear domain.
Material and methods
Protein expression and purification.
The DNA fragment for residues 677−822 of human 1-adaptin was cloned into the pGEX4T-2 plasmid (Amersham Pharmacia Biotech). The native and selenomethionine (SeMet)-substituted proteins were expressed in Escherichia coli BL21 and DL41 cells, respectively. Addition of preceding residues (677−702) to the 1 ear domain (residues 703−822) improved the solubility of the protein. The GST fusion protein was purified through affinity chromatography using a glutathione-Sepharose 4B column (Amersham Pharmacia Biotech) and cleaved by thrombin. The cleaved 1 ear domain was further purified by Superdex 75 size-exclusion column in 100 mM NaCl and 20 mM Tris-HCl, pH 8.0.
Crystallization and data collection.
Crystals of the native and SeMet-substituted proteins were obtained in a buffer solution containing 10 mg ml-1 protein, 20% (w/v) PEG 4000, 200 mM MgCl2 and 100 mM HEPES-Na, pH 7.5, with equilibration at 277 K for 1 week. All data sets were collected under cryogenic conditions with crystals soaked in the cryoprotectant buffer containing 20% (w/v) glycerol and cooled at 100 K in a nitrogen gas stream. Gold derivatives were prepared by soaking crystals in the cryoprotectant buffer with 10 mM KAu(CN)2. Native and gold-derivative data sets were collected at BL-44XU of SPring-8, and MAD data sets of the SeMet-substituted crystal were collected at BL-6A of Photon Factory. The diffraction data were integrated and scaled using MOSFLM22 and SCALA23.
Structure determination and refinement.
For SIRAS phasing with the gold derivative data, initial phases were determined using MLPHARE23 and improved with a density modification procedure using DM23. For MAD phasing, initial phases were determined using SOLVE24 and improved using RESOLVE25. The initial molecular model was built automatically using ARP/wARP26, which could assign the main chain of 109 out of 147 residues in the 1 ear construct.
Further model fitting to the electron density maps was carried out manually using O27 and followed by structure refinement through REFMAC23. The two 1 ear domain monomers, termed A and B, are present in the asymmetric unit of the crystal and composed of residues 703−822 and 700−822, respectively. No model could be built for the N-terminal hinge region, residues 677−699, due to disorder in the electron density, which might account for the relatively high R-factors; the final values for working and test sets are 22.6 and 24.8% at 1.8 Å resolution, respectively. Stereochemical quality of the final model was assessed by PROCHECK28, where 88.9% of the amino acids were located in the most favored regions and none in the disallowed regions. The crystallographic statistics are summarized in Table 1.
Yeast two-hybrid assay.
Reverse two-hybrid screening was performed according to the reported procedure5,
11. A cDNA fragment of the 1 ear domain (residues 703−822) was subjected to error-prone PCR and subcloned into the pGBT9 bait vector. Y190 reporter cells harboring the pGAD10 prey vector for Rabaptin-5 (ref. 5) were transformed with the bait vector for the mutagenized 1 ear domain and subjected to a filter assay for -galactosidase activity. The pGBT9 plasmids recovered from colonies that did not develop blue color were sequenced.
GST pull-down assay.
N-terminally HA-tagged -synergin was expressed in hEK-293 cells11. The supernatant of cell lysate was incubated for 2 h at 4 °C with 20 g of recombinant GST, GST−1 ear domain5 or its mutant pre-bound to glutathione-Sepharose 4B beads. The beads were then washed three times with homogenization buffer containing 0.5% Triton X-100. Proteins associated with the beads were subjected to immunoblot analysis using anti-Rabaptin-5 (clone 20, Transduction Laboratories) or anti-HA (3F10, Roche Diagnostics)5. The band densities were estimated using a LAS-1000 bioimaging analyzer (Fuji Photo Film).
Coordinates.
The coordinates have been deposited in the Protein Data Bank (accession code 1IU1).
Received 22 February 2002; Accepted 1 May 2002; Published online: 3 June 2002.
REFERENCES
- Kirchhausen, T. Annu. Rev. Cell Dev. Biol. 15, 705−732 (1999). | Article | PubMed | ISI | ChemPort |
- Boehm, M. & Bonifacino, J.S. Mol. Biol. Cell 12, 2907−2920 (2001). | PubMed | ISI | ChemPort |
- Page, L.J., Sowerby, P.J., Lui, W.W. & Robinson, M.S. J. Cell Biol. 146, 993−1004 (1999). | Article | PubMed | ISI | ChemPort |
- Hirst, J. et al. J. Cell. Biol. 149, 67−80 (2000). | Article | PubMed | ISI | ChemPort |
- Shiba, Y., Takatsu, H., Shin, H.W. & Nakayama, K. J. Biochem. 131, 327−336 (2002). | PubMed | ISI | ChemPort |
- Fernandez-Chacon, R., Alvarez de Toledo, G., Hammer, R.E. & Sudhof, T.C. J. Biol. Chem. 274, 32551−32554 (1999). | Article | PubMed | ISI | ChemPort |
- Meyer, C. et al. EMBO J. 19, 2193−2203 (2000). | Article | PubMed | ChemPort |
- Boman, A.L., Zhang, C., Zhu, X. & Kahn, R.A. Mol. Biol. Cell 11, 1241−1255 (2000). | PubMed | ISI | ChemPort |
- Dell'Angelica, E.C. et al. J. Cell Biol. 149, 81−94 (2000). | Article | PubMed | ChemPort |
- Poussu, A., Lohi, O. & Lehto, V.P. J. Biol. Chem. 275, 7176−7183 (2000). | Article | PubMed | ISI | ChemPort |
- Takatsu, H., Yoshino, K. & Nakayama, K. Biochem. Biophys. Res. Commun. 271, 719−725 (2000). | Article | PubMed | ISI | ChemPort |
- Costaguta, G., Stefan, C.J., Bensen, E.S., Emr, S.D. & Payne, G.S. Mol. Biol. Cell 12, 1885−1896 (2001). | PubMed | ISI | ChemPort |
- Nielsen, M.S. et al. EMBO J. 20, 2180−2190 (2001). | Article | PubMed | ChemPort |
- Puertollano, R., Aguilar, R.C., Gorshkova, I., Crouch, R.J. & Bonifacino, J.S. Science 292, 1712−1716 (2001). | Article | PubMed | ISI | ChemPort |
- Takatsu, H., Katoh, Y., Shiba, Y. & Nakayama, K. J. Biol. Chem. 276, 28541−28545 (2001). | Article | PubMed | ISI | ChemPort |
- Zhu, Y., Doray, B., Poussu, A., Lehto, V.P. & Kornfeld, S. Science 292, 1716−1718 (2001). | Article | PubMed | ISI | ChemPort |
- Misra, S., Puertollano, R., Kato, Y., Bonifacino, J.S. & Hurley, J.H. Nature 415, 933−937 (2002). | Article | PubMed | ISI | ChemPort |
- Shiba, T. et al. Nature 415, 937−941 (2002). | Article | PubMed | ISI | ChemPort |
- Owen, D.J. et al. Cell 97, 805−815 (1999). | Article | PubMed | ISI | ChemPort |
- Traub, L.M., Downs, M.A., Westrich, J.L. & Fremont, D.H. Proc. Natl. Acad. Sci. USA 96, 8907−8912 (1999). | Article | PubMed | ChemPort |
- Owen, D.J., Vallis, Y., Pearse, B.M., McMahon, H.T. & Evans, P.R. EMBO J. 19, 4216−4227 (2000). | Article | PubMed | ChemPort |
- Leslie, A.G.W. Crystallographic Computing (Oxford University Press, Oxford; 1990).
- Collaborative Computational Project, Number 4. Acta Crystallogr. D 50, 760−763 (1994). | Article | ISI |
- Terwilliger, T.C. & Berendzen, J. Acta Crystallogr. D 55, 849−861 (1999). | Article | PubMed | ISI |
- Terwilliger, T.C. Acta Crystallogr. D 56, 965−972 (2000). | Article | PubMed | ISI |
- Perrakis, A., Morris, R. & Lamzin, V.S. Nature Struct. Biol. 6, 458−463 (1999). | Article | PubMed | ISI | ChemPort |
- Jones, T.A., Zou, J.Y., Cowan, S.W. & Kjeldgaard, M. Acta Crystallogr. A 47, 110−119 (1991). | Article | PubMed | ISI |
- Laskowski, R.A., MacArthur, M.W., Moss, D.S. & Thornton, J.M. J. Appl. Crystallogr. 26, 283−291 (1993). | Article | ISI | ChemPort |
- Kraulis, P.J. J. Appl. Crystallogr. 24, 946−950 (1991). | Article | ISI |
- Nicholls, A., Sharp, K.A. & Honig, B. Proteins 11, 281−296 (1991). | PubMed | ISI | ChemPort |
Acknowledgments
This work was supported in part by Grants-in-Aid for Scientific Research from the Ministry of Education, Culture, Sports, Science and Technology of Japan, from the Japan Society for Promotion of Science (fellowship to H.T.) and from the University of Tsukuba Research Projects.
Competing interests statement:
The authors declare that they have no competing financial interests. |
2-adaptin ear domains.
