GTC
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE ADVANCED SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

Genes to Cells (2005) 10, 639-654. doi:10.1111/j.1365-2443.2005.00865.x
© 2005 Blackwell Publishing or its licensors
This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Kawasaki, M.
Right arrow Articles by Wakatsuki, S.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Kawasaki, M.
Right arrow Articles by Wakatsuki, S.

Molecular mechanism of ubiquitin recognition by GGA3 GAT domain

Masato Kawasaki1, Tomoo Shiba1, Yoko Shiba2, Yoshiki Yamaguchi3, Naohiro Matsugaki1, Noriyuki Igarashi1, Mamoru Suzuki1,{dagger}, Ryuichi Kato1, Koichi Kato3, Kazuhisa Nakayama2 and Soichi Wakatsuki1,*

1 Structural Biology Research Center, Photon Factory, Institute of Materials Structure Science, High Energy Accelerator Research Organization (KEK), Tsukuba, Ibaraki 305-0801, Japan
2 Department of Physiological Chemistry, Graduate School of Pharmaceutical Sciences, Kyoto University, Yoshida-shimoadachi, Sakyo-ku, Kyoto 606-8501, Japan
3 Department of Structural Biology and Biomolecular Engineering, Graduate School of Pharmaceutical Sciences, Nagoya City University, 3-1 Tanabe-dori, Mizuho-ku, Nagoya 467-8603, Japan


   Abstract
Top
 Abstract
Introduction
 Results
 Discussion
 Experimental procedures
 References
 
GGA (Golgi-localizing, {gamma}-adaptin ear domain homology, ARF-binding) proteins, which constitute a family of clathrin coat adaptor proteins, have recently been shown to be involved in the ubiquitin-dependent sorting of receptors, through the interaction between the C-terminal three-helix-bundle of the GAT (GGA and Tom1) domain (C-GAT) and ubiquitin. We report here the crystal structure of human GGA3 C-GAT in complex with ubiquitin. A hydrophobic patch on C-GAT helices {alpha}1 and {alpha}2 forms a binding site for the hydrophobic Ile44 surface of ubiquitin. Two distinct orientations of ubiquitin Arg42 determine the shape and the charge distribution of ubiquitin Ile44 surface, leading to two different binding modes. Biochemical and NMR data strongly suggest another hydrophobic binding site on C-GAT helices {alpha}2 and {alpha}3, opposite to the first binding site, also binds ubiquitin although weakly. The double-sided ubiquitin binding provides the GAT domain with higher efficiency in recognizing ubiquitinated receptors for lysosomal receptor degradation.


   Introduction
Top
 Abstract
 Introduction
Results
 Discussion
 Experimental procedures
 References
 
Monoubiquitination functions as a signal for protein degradation in lysosomes (Hicke 2001; Katzmann et al. 2001; Di Fiore et al. 2003; Raiborg et al. 2003; Schnell & Hicke 2003). In yeast, ubiquitinated transmembrane proteins such as the general amino acid permease Gap1p are sorted into luminal vesicles of the multivesicular body (MVB) along both endocytic and biosynthetic pathways, and ultimately delivered to the vacuole. In mammalian cells, ubiquitin-dependent transport through the MVB to lysosome has been demonstrated only along the endocytic pathway, as in the case of epidermal growth factor receptor. Ubiquitin is recognized by a series of cytosolic proteins that contain different ubiquitin-binding modules, such as the UBA (ubiquitin-associated), CUE (coupling of ubiquitin to endoplasmic reticulum degradation), UIM (ubiquitin interacting motif), UEV (ubiquitin E2 variant), and NZF (Npl4 zinc finger) domains (Schnell & Hicke 2003 and references therein).

The GAT (GGA and Tom1) domain of GGA (Golgi-localizing, {gamma}-adaptin ear domain homology, ARF-binding) protein is a novel member of ubiquitin-interacting modules. GGAs are a family of monomeric adaptor proteins that regulate clathrin-mediated trafficking of cargo proteins from the trans-Golgi network (TGN) to endosomes (Boman 2001; Hinners & Tooze 2003; Nakayama & Wakatsuki 2003; Bonifacino 2004). GGA is composed of four functional regions: an N-terminal VHS (Vps27/Hrs/Stam) domain that interacts with an acidic cluster dileucine motif found in the cytoplasmic tail of TGN sorting receptors; a GAT domain that binds to the GTP-bound form of ARF; a hinge region that contains clathrin-binding motifs; and a C-terminal GAE ({gamma}-adaptin ear homology) domain that interacts with accessory proteins. Recently the GAT domain of GGAs was found to bind ubiquitin (Puertollano & Bonifacino 2004; Scott et al. 2004; Shiba et al. 2004). In yeast, GGAs are necessary for ubiquitin-dependent sorting of Gap1p from the TGN to endosomes (Scott et al. 2004). In mammalian cells, GGAs are demonstrated to be involved in the endocytic degradation of ubiquitinated epidermal growth factor receptor (Puertollano & Bonifacino 2004). These findings uncovered a novel role for GGAs in ubiquitin-dependent sorting of cargo proteins both in biosynthetic and endocytic pathways, in addition to the previously established function of GGAs in sorting lysosomal cargo receptors from the TGN to endosomes.

The GAT domain of GGA consists of two subdomains (Collins et al. 2003; Shiba et al. 2003; Suer et al. 2003; Zhu et al. 2003). The N-terminal subdomain (N-GAT) is a helix-loop-helix structure that is responsible for ARF binding. The C-terminal subdomain (C-GAT) is a three-helix bundle that has been shown to be responsible for ubiquitin binding (Scott et al. 2004; Shiba et al. 2004). Here we present the crystal structure of the complex between GGA3 C-GAT and ubiquitin which show two modes of primarily hydrophobic interactions between the two molecules. Although the previously proposed ubiquitin binding site on helix {alpha}3 (Shiba et al. 2004) is involved in dimerization of C-GAT in the crystal, we confirm that ubiquitin can also bind to this site in solution based on biochemical and NMR analyses.


   Results
Top
 Abstract
 Introduction
 Results
Discussion
 Experimental procedures
 References
 
Structure determination of GGA3 C-GAT in complex with ubiquitin

We chose to cocrystallize the human GGA3 C-GAT subdomain with ubiquitin for the following two reasons: ubiquitin binds most strongly to GGA3 among three human GGAs (GGA1–3) (Puertollano & Bonifacino 2004; Scott et al. 2004; Shiba et al. 2004) and the GGA3 C-GAT subdomain binds ubiquitinated proteins more efficiently than the full GAT domain does (Shiba et al. 2004). A 1 : 1 mixture of GGA3 C-GAT (residues 209–304) and ubiquitin crystallized under various conditions including salt and polyethylene glycol as precipitants, but the crystals did not diffract well. Therefore we extended C-GAT at either the N- or C-terminus in search for better crystals. A construct with a C-terminal extension (residues 209–319) improved significantly the diffraction quality of the crystal, whereas a construct with an N-terminal extension (residues 192–304) did not improve it. The complex structure was determined by the multiwavelength anomalous dispersion (MAD) and refined to 2.6 Å resolution (Table 1).


View this table:
[in this window]
[in a new window]
  		Table 1 Crystallographic data and refinement statistics
 
The asymmetric unit of the P212121 crystal contains four copies of C-GAT/ubiquitin heterodimers (four pairs AE, BF, CG and DH; letters A to D designate the four C-GAT molecules and letters E to H designate the four ubiquitins; Fig. 1). The GGA3 C-GAT domain consists of a three-helix bundle ({alpha}1, {alpha}2 and {alpha}3). C-GAT molecules dimerize using helices {alpha}2 and {alpha}3 in the crystal lattice (two C-GAT pairs AB and CD; Fig. 1). Ubiquitin is bound on a hydrophobic patch of helices {alpha}1 and {alpha}2 of C-GAT (termed Site 1 hereafter). Site 1 includes Leu227 and Met231 of helix {alpha}1 and Leu247 of helix {alpha}2 of GGA3 C-GAT (see below). It interacts with the hydrophobic patch of the ubiquitin Ile44 surface, which is composed of residues of ß1, ß3, ß4 and ß5 strands of ubiquitin.



View larger version (49K):
[in this window]
[in a new window]
  		Figure 1  Four GGA3 C-GAT/ubiquitin complexes in the asymmetric unit. The asymmetric unit contains four C-GAT/ubiquitin pairs AE, BF, CG and DH (colored by molecule); letters A to D designate the four C-GAT molecules and E to H designate the four ubiquitins. GGA3 C-GAT domain consists of a three-helix bundle ({alpha}1, {alpha}2 and {alpha}3). Ubiquitin is bound on helices {alpha}1 and {alpha}2 of the C-GAT. C-GAT molecules dimerize through helices {alpha}2 and {alpha}3 in the crystal lattice (two C-GAT pairs AB and CD).

 
C-GAT helices {alpha}1, {alpha}2 and {alpha}3 span residues Leu218-Leu233, Asp242-Ser267 and Leu276-Ile296, respectively (Fig. 2A). Although the C-terminal extension (residues 305–319) of C-GAT improved the crystal quality, the extended residues are disordered and invisible in the electron density map. The N-terminal regions before residue 218 of the four C-GAT molecules are affected by the crystal packing. First, the N-terminal six residues of the C-GAT molecules B and D are disordered and invisible. Second, the N-termini of the C-GAT molecules A and C largely bend (Fig. 3A,B) because of the collision with the neighboring C-GAT molecules D and B, respectively, in the crystal lattice. Thus the disorder in the N-termini is likely to be an artifact of crystallization of the C-GAT subdomain that is truncated in the middle of long helix {alpha}1. Normally, helix {alpha}1 should extend to the N-GAT subdomain in the full GAT structures (Collins et al. 2003; Shiba et al. 2003; Suer et al. 2003; Zhu et al. 2003). Otherwise, the four C-GAT molecules (A, B, C and D) in the asymmetric unit are superimposed well on each other with root mean square deviation (rmsd) ranging from 0.61 to 0.98 Å for all C{alpha} atoms of Leu218-Ile296 residues (Fig. 3A). Although the structure of GGA3 GAT alone has not been solved, the three {alpha}-helices of GGA3 C-GAT in the complexes are superimposed well on the corresponding three helices of free GGA1 GAT (Shiba et al. 2003) (Fig. 3A; < 0.97 Å rmsd for the C{alpha} atoms of three helices excluding the loop regions shown in Fig. 2A), suggesting that significant conformational changes do not occur upon the complex formation between GGA3 C-GAT and ubiquitin. The four ubiquitin molecules in the asymmetric unit (E, F, G and H) are also superimposed well with average C{alpha} rmsd of 0.42 Å for residues Met1-Arg72. The structures of ubiquitin in the complex are superimposed well on the structure of the free ubiquitin (Vijay-Kumar et al. 1987) (PDB entry 1UBQ [PDB] ) with C{alpha} rmsd less than 0.57 Å for residues Met1-Arg72. Thus, the complex formation with C-GAT does not cause significant structural changes on the part of ubiquitin either.



View larger version (36K):
[in this window]
[in a new window]
  		Figure 2  Sequence alignment and secondary structures of C-GAT domains and ubiquitin. (A) Secondary structure and sequence alignment of C-GAT. Boxes above the sequences show {alpha}-helix regions of the GGA3 C-GAT domain. Residues conserved in GGA proteins are shown in blue. Residues conserved between Tom1 and Tom1L1 are shown in green. Residues conserved across the GGA and Tom1 subfamilies are shown in red. Red asterisks on top of the sequences indicate GGA3 residues which directly interact with ubiquitin. GGA3 sequence numbers are shown above the sequence. (B) Secondary structure and sequence of ubiquitin. Boxes and arrows above the sequence show {alpha}-helix and ß-strand regions, respectively. Red asterisks indicate residues which directly interact with GGA3 C-GAT. Ubiquitin sequence numbers are shown above the sequence.

 


View larger version (34K):
[in this window]
[in a new window]
  		Figure 3  Superposition of the four molecules in the asymmetric unit. (A) Superposition of four GGA3 C-GAT molecules in the asymmetric unit (blue) and free GGA1 GAT (orange). (B) Stereo view of a superposition of four GGA3 C-GAT/ubiquitin complexes in the asymmetric unit. Four ubiquitin molecules are superimposed. Ubiquitin molecules are colored from green at the N-terminal to red at the C-terminal. GGA3 C-GAT molecules are colored from blue at the N-terminal to green at the C-terminal. Black solid lines represent the axes of helices {alpha}2 and {alpha}3 of the most distant pairs.

 
In spite of the structural integrity of each protein, overall structures of the four GGA3 C-GAT/ubiquitin complexes in the asymmetric unit cannot be superimposed well because their relative orientations of C-GAT and ubiquitin differ significantly. When ubiquitin molecules of the four complexes are used to superimpose the four complexes, the C-GAT molecules have two different orientations, suggesting that the four complexes can be divided into two, AE and CG as one group and BF and DH as the other (Fig. 3B). The relative tilt angles of helices {alpha}1, {alpha}2 and {alpha}3 of C-GAT between the two groups of the C-GAT and ubiquitin complexes are about 4°, 13° and 8°, respectively.

Two modes of interaction between GGA3 C-GAT Site 1 and ubiquitin

Detailed inspection of the two groups of the four complexes reveals two distinct modes of interactions, the ‘Arg42-upright’ and ‘Arg42-inclined’ binding modes (Fig. 4A–D). In the Arg42-upright mode (complexes BF and DH), the guanidium group of Arg42 of ubiquitin forms salt bridges with the carboxyl groups of Glu246 and Glu250 of helix {alpha}2 of C-GAT (Fig. 4A,D). On the other hand, in the Arg42-inclined mode (complexes AE and CG), these three residues somewhat counterintuitively point away from each other (Fig. 4A,C), which enlarges the hydrophobic pocket of ubiquitin which accommodates Leu247 of helix {alpha}2 of C-GAT. This conformational change plays a pivotal role in the tighter binding of ubiquitin by pushing the side chain of Leu247 of C-GAT more deeply into the widened pocket III (see below). As a consequence, the axis of C-GAT helix {alpha}2 is much closer to the ubiquitin Ile44 surface in the Arg42-inclined mode than in the Arg42-upright mode shown in Fig. 4A. The buried surface areas of the GAT-ubiquitin pairs corroborate these differences; the Arg42-inclined mode has larger buried surface areas of 1317 Å2 (AE) and 1351 Å2 (CG) than the Arg42-upright mode with 1254 Å2 (BF) and 1299 Å2 (DH).



View larger version (47K):
[in this window]
[in a new window]
  		Figure 4  GGA3 C–GAT/ubiquitin interface. (A) Side view of the GGA3 C–GAT/ubiquitin interface of the superimposed two complexes (pairs AE and BF). Ribbon diagrams of ubiquitin molecules are colored yellow (molecule E) and pink (molecule F). Ribbon diagrams of C-GAT molecules are colored blue (molecule A) and green (molecule B). The side chains involved in the interaction are shown. (B) Stereo diagram of the refined model of GGA3 C-GAT/ubiquitin complex, molecules B (green) and F (pink), with the 2Fo-Fc electron density map contoured at 1.0 {sigma}, viewed in the same orientation as (A). (C) Stereo view of the binding interface between C-GAT (molecule A, blue) and ubiquitin (molecule E, yellow), rotated 45° about the horizontal axis relative to (A). (D) Stereo view of the binding interface between C-GAT (molecule B, green) and ubiquitin (molecule F, pink), rotated 45° about the horizontal axis relative to (A).

 
Apart from the above differences, the binding interface is almost the same between the two modes. In the center of the contact surface, the side chains of three hydrophobic residues (Leu227 and Met231 of helix {alpha}1 and Leu247 of helix {alpha}2) of C-GAT surround the Ile44 side chain of ubiquitin (Fig. 4C,D). Leu227 of helix {alpha}1 of C-GAT is accommodated in a deep hydrophobic pocket formed by the side chains of ubiquitin Leu8, Ile44, His68 and Val70 residues. Met231 of helix {alpha}1 of C-GAT lies in a shallow groove on ubiquitin surface formed by Ile44, Gly47, Lys48 and Gln49 of ubiquitin. Leu247 of helix {alpha}2 of C-GAT fits in a pocket formed by the side chains of Arg42, Ile44, Gln49 and Val70 of ubiquitin. Outside of the Ile44 hydrophobic patch, Glu220 and Asn223 of C-GAT helix {alpha}1 pack against Leu8 and Thr9 of the ß-turn between ß1 and ß2 strands of ubiquitin. Glu230 of helix {alpha}1 and His234 located at the end of helix {alpha}1 of C-GAT pack against Ala46 and Gly47 of the ß-turn between ß3 and ß4 strands of ubiquitin. Gln49 of ubiquitin makes two hydrogen bonds with helix {alpha}2 of C-GAT. One hydrogen bond is between the side chain amide nitrogen of Gln49 and the main chain carbonyl oxygen of Gly243 of C-GAT. The other is between the main chain amide nitrogen of Gln49 and the side chain carboxyl oxygen of Asp244 of C-GAT.

GGA3 C-GAT dimerization in the crystal

In addition to Site 1, which interacts with ubiquitin in the crystal, there is another large hydrophobic patch on the other side of the GGA3 C-GAT surface, which includes Leu276 and Leu280 residues of helix {alpha}3, together with Phe263 and Ala266 of helix {alpha}2. We hereafter refer to this hydrophobic patch as Site 2. The hydrophobic nature of this part of the protein is well conserved among GGAs and had been proposed to be involved in protein–protein interaction (Zhai et al. 2003; Mattera et al. 2004; Shiba et al. 2004). However, a major part of this hydrophobic patch is unfortunately used for dimerization of GGA3 C-GAT molecules in our crystal (Fig. 5). The dimerization interface of C-GAT is mainly hydrophobic and mediated by Arg260, Phe263, Ala266 and Ser267 in the C-terminal half of helix {alpha}2, and Leu276 and Leu280 in the N-terminal half of helix {alpha}3 (Fig. 5). The buried surface areas of the two GAT-GAT pairs are 1220 Å2 (AB) and 1197 Å2 (CD), comparable to those between C-GAT and ubiquitin.



View larger version (39K):
[in this window]
[in a new window]
  		Figure 5  Dimer interface of GGA3 C-GAT in the crystal. A ribbon diagram of the interface between C-GAT molecules A (blue) and B (green). Side chains of residues directly involved in the interaction are shown.

 
Mutational experiments

Indeed, our previous mutational analysis (Shiba et al. 2004) suggested that ubiquitin binds to the C-terminal half of helix {alpha}3 within the hydrophobic dimerization interface of C-GAT (Site 2). Therefore we were surprised to find that ubiquitin binds only Site 1 in the crystal. To investigate this discrepancy, we set out pull-down, surface plasmon resonance (SPR) and NMR titration experiments using various mutants on Sites 1 and 2.

First, we introduced mutations into Site 1 of GGA3 C-GAT and examined their effects on ubiquitin binding. The pull-down experiment using the GST-fused GAT Site 1 mutants M231S and E250N exhibited reduced affinities to both monoubiquitin and polyubiquitin, compared with the wild-type GAT domain (Fig. 6A). We also measured the affinity of monoubiquitin to the GST-fused GAT Site 1 mutants L227A, M231S and E250N by SPR analysis (Table 2 and Fig. 6B). The wild-type GGA3 GAT gave an equilibrium dissociation constant (Kd) of 0.23 mM, whereas the Site 1 mutations, L227A, M231S and E250N all increased the Kd values about threefold compared to the wild-type. Furthermore, the L227A and M231S mutations significantly decreased the Rmax value that represents the maximum ubiquitin binding capacity. These pull-down and SPR data suggest that the Site 1 of C-GAT contributes significantly to the ubiquitin binding, consistent with the crystal structure.



View larger version (26K):
[in this window]
[in a new window]
  		Figure 6  Binding of GGA3 GAT mutants to ubiquitin. (A) A mixture of polyubiquitin chains and monoubiquitin was pulled down with GST or GST fused to wild-type or mutant GGA3 GAT domain and subjected to immunoblotting with anti-ubiquitin antibody. On the input lane, 1/20 samples were loaded. Below is Coomassie blue staining of GST fusion proteins for loading control. (B) SPR data for the interaction between GGA3 GAT mutants and ubiquitin. Steady state surface plasmon resonance levels (Req) by ubiquitin binding to GST-fused GGA3 GAT mutants are plotted against ubiquitin concentration.

 

View this table:
[in this window]
[in a new window]
  		Table 2 SPR analysis of the interaction between GGA3 GAT mutants and ubiquitin
 
Second, we examined the ubiquitin binding affinity of helix {alpha}3 (Site 2) mutants: L276S, L280R and D284G. In our previous experiments, we could not detect any binding of GAT Site2 mutants L280R and D284G to monoubiquitin-agarose (Shiba et al. 2004). However, we observed decreased but detectable binding to monoubiquitin and polyubiquitin using GST-fused GAT mutants (Fig. 6A). This discrepancy may be due to the difference in sensitivity between the two pull-down experiments; soluble GST-fused GAT was loaded on to monoubiquitin-agarose in the former assay, whereas soluble monoubiquitin and polyubiquitin mixture was loaded on to GST-fused GAT which had been immobilized to the glutathione-Sepharose in this assay. To verify the pull-down experiments, we performed SPR analysis using GST-fused GAT Site 2 mutants and monoubiquitin. SPR analysis of these Site 2 mutants gave similar Kd values as the wild-type, but showed decreased amounts of maximum ubiquitin binding (Rmax) compared to the wild-type (Table 2 and Fig. 6B). These data show that defects in ubiquitin binding of the Site 2 mutants are substantial but less severe than those of Site 1 mutants. When the double mutations were introduced on both patches of GAT (E250N/D284G), the mutant showed an additional defect in ubiquitin binding (Table 2, Fig. 6A,B). Circular dichroism spectra of all these GAT mutants were identical to that of wild-type, indicating that the global secondary structure remained intact (data not shown). Taken together, these pull-down and SPR data suggest that ubiquitin can bind to two different sites of GAT domain. The first binding site found in the crystal (Site 1) has about three times higher affinity for ubiquitin than the second binding site, which is probably composed of helix {alpha}3 (Site 2).

NMR chemical shift perturbation mapping

Given the lack of the atomic details of the Site 2 interaction from the X-ray crystallography data, we then performed a nuclear magnetic resonance (NMR) spectroscopic analysis for the interaction in solution between GGA3 GAT and ubiquitin. Firstly, we attempted to identify ubiquitin-binding site of GGA3 C-GAT using chemical shift perturbation data of isotopically labeled GGA3 C-GAT titrated with ubiquitin. Progressive chemical shift changes were observed for the peaks originating from two distinct surface areas of the GGA3 C-GAT domain (Fig. 7A). One consists of the residues from {alpha}1 and the {alpha}1-{alpha}2 loop of GGA3 C-GAT (Site 1), and the other spans the C-terminal region of helix {alpha}2 and the {alpha}2-{alpha}3 loop (Site 2). These data indicate that GGA3 C-GAT possesses two ubiquitin binding sites (Sites 1 and 2) in solution. The ubiquitin-binding at Site 1 is consistent with the crystal structure, while Site 2 makes no contact with ubiquitin in the crystal.



View larger version (42K):
[in this window]
[in a new window]
  		Figure 7  NMR chemical shift perturbation mapping of the GGA3 C-GAT and ubiquitin. (A) Mapping of the perturbed residues of the GGA3 C-GAT upon binding to ubiquitin. The chemical shift changes are quantified for each residue according to the equation {gtc_865_mu2}, where {delta}N and {delta}H represent the change in nitrogen and proton chemical shifts upon addition of ubiquitin. The perturbed residues are mapped on to the GGA3 C-GAT crystal structure (molecule B in the GGA3 C-GAT/ubiquitin crystal). The residues showing chemical shift perturbation are colored in red with a gradient reflecting the strength of the perturbation. Significantly perturbed residues (more than 0.2 p.p.m.) are labeled. Residues whose peak became undetectable upon binding to ubiquitin are shown in purple and labeled. Residues whose HSQC peak was not assigned (Glu220 and Glu230) are shown in gray. Left figure is viewed in the same direction as Fig. 4D. (B) Mapping of GGA3-binding sites of ubiquitin. The perturbed residues of ubiquitin upon binding of GGA3 GAT wild-type (WT), M231S or L280R mutant are mapped on to the ubiquitin crystal structure (molecule F in the GGA3 C-GAT/ubiquitin crystal), viewed in the same direction as Fig. 4D. The chemical shift perturbations are quantified and displayed as (A). Three Pro residues (19, 37, and 38) and residues whose HSQC peak was not observable due to a high pH condition (Thr9, Ala46 and Gly53) are shown in gray. Although Arg74 and Gly75 were not observable in the spectra either, they are not shown because these residues were invisible in the crystal.

 
Next, we observed chemical shift perturbation of isotopically labeled ubiquitin induced upon binding to GGA3 C-GAT. In addition to the wild-type GGA3 C-GAT, we used two GGA3 C-GAT mutants in which a single amino acid residue was substituted at Site 1 or 2, i.e. M231S or L280R in order to identify amino acid residues involved in binding to Sites 1 and 2 of GGA3 C-GAT distinguishably. Binding of the wild-type GGA3 C-GAT induced chemical shift perturbations and/or broadening of NMR signals for the large surface of the ß-sheet (ß1-ß5), which is involved in binding to Site 1 of GGA3 C-GAT as shown in the crystal structure (Fig. 7B). Both of the M231S and L280R mutants perturb virtually the same surface area upon binding, although there are some differences in spectral perturbation among the wild-type and the mutants. This suggests that ubiquitin employs the same surface area surrounding Ile44 for binding to Site 1 and Site 2 of GGA3 C-GAT. In the crystal, there exist two modes of binding between ubiquitin and Site 1 of GGA3 C-GAT. Under the experimental conditions of the present NMR study, no residues gave double peaks corresponding to the two modes of interaction. Rather, significant line-broadening, presumably due to chemical exchange between the ‘Arg42-upright’ and ‘Arg42-inclined’ binding modes, was observed for the residues on or around the interface between ubiquitin and Site 1, i.e. Val225, Arg226, Leu227, Leu233, Ser236, and Ser241 of GGA3 C-GAT interacting with ubiquitin and Arg42, Phe45, Gly47, Lys48, Gln49, Leu69, Val70, and Leu71 of ubiquitin interacting with the L280R mutant.


   Discussion
Top
 Abstract
 Introduction
 Results
 Discussion
Experimental procedures
 References
 
Ubiquitin binding Site 1 of GGA3 C-GAT

The Ile44 surface of ubiquitin is commonly used as a binding core by a variety of differently folded ubiquitin-binding modules. The hydrophobic patch of the Ile44 ubiquitin surface in the GGA3 C-GAT/ubiquitin complex can be divided into three hydrophobic pockets, to which we refer hereafter as pockets I, II and III (Fig. 8A). Pocket I is a large cavity formed by the side chains of ubiquitin Leu8, Ile44, His68 and Val70. Pocket II is a shallow concave like the seat of a saddle, formed by the side chains of Ile44 and Gln49, and the main chains of Gly47 and Lys48. Pocket III is another large cavity formed by the side chains of Arg42, Ile44, Gln49 and Val70. Pockets I, II and III accommodate the side chains of the hydrophobic residues Leu227, Met231 and Leu247 of GGA3 C-GAT, respectively (Fig. 8B).



View larger version (61K):
[in this window]
[in a new window]
  		Figure 8  Three pockets of ubiquitin Ile44 surface. (A) Schematic representation of the Ile44 surface of ubiquitin. ß-Strands are represented as arrows. The residues that interact with C-GAT are shown. The hydrophobic patch of the Ile44 ubiquitin surface can be divided into three hydrophobic pockets I, II and III (pink circles). (B) Schematic representation of GGA3 C-GAT in complex with ubiquitin in the same view as (A). Helices {alpha}1 and {alpha}2 of GGA3 C-GAT are shown as yellow rectangles. Residues that interact with ubiquitin are shown in red. Pockets I, II and III of ubiquitin Ile44 surface (red circled) accommodate hydrophobic residues Leu227, Met231 and Leu247 of GGA3 C-GAT, respectively. (C) The molecular surface of ubiquitin (molecule E) is shown with a ribbon drawing of C-GAT (molecule A), viewed in the same direction as Fig. 4C. Three hydrophobic pockets of ubiquitin are circled in red. Negative potentials (–20 kT) are displayed in deep red, and positive potentials (20 kT) are displayed in deep blue. (D) The molecular surface of ubiquitin (molecule F) is shown with a ribbon drawing of C-GAT (molecule B), viewed in the same direction as Fig. 4D. Three hydrophobic pockets of ubiquitin are circled in red. Negative potentials (–20 kT) are displayed in deep red, and positive potentials (20 kT) are displayed in deep blue. (E) Close-up view of pocket III of (C). Helix {alpha}2 is omitted for clarity. (F) Close-up view of pocket III of (D). Helix {alpha}2 is omitted for clarity. Salt bridges are indicated by red dotted lines.

 
When four ubiquitins in the asymmetric unit of GGA3 C-GAT/ubiquitin complex are superimposed, the side chain of Arg42 shows substantial displacements, while the rest of the side chains on the Ile44 surface are well superimposed on each other (Fig. 4A). The displacements of ubiquitin Arg42 side chain causes (1) movements of side chains of Glu246 and Glu250 of C–GAT through electrostatic interactions, and (2) changes in size and shape of the pocket III which accommodates Leu247 of C-GAT, leading to two different modes of the C-GAT/ubiquitin interaction in the crystal. In the ‘Arg42-upright’ binding mode of the C-GAT/ubiquitin interaction, the Arg42 side chain stands perpendicular to the peptide backbone and lifts up helix {alpha}2 by making salt bridges with the side chains of Glu246 and Glu250 of C-GAT (Fig. 4A,D). These interactions keep the backbone of helix {alpha}2 away from the center of pocket III. The upright orientation of the Arg42 side chain narrows the pocket III (Fig. 8D,F), which effectively decreases the size of pocket III for Leu247. On the other hand, in the ‘Arg42-inclined’ binding mode of the C–GAT/ubiquitin interaction, the Arg42 side chain lying down on to the ß-sheet, points away from the side chains of Glu246 and Glu250 (Fig. 4A,C). This inclined conformation of Arg42 side chain widens the pocket III (Fig. 8C,E). Consequently, the backbone of helix {alpha}2 of C-GAT comes closer to the ubiquitin surface, and Leu247 side chain inserts itself more deeply into the pocket III (Fig. 8E). In summary, the inclined or upright conformation of Arg42 side chain dictates the size of pocket III and the relative orientation of helix {alpha}2 producing two distinct binding modes. We postulate that this provides a unified scheme of interpreting known ubiquitin binding motifs.

Ubiquitin binding modes of other ubiquitin-binding modules

In the crystal structure of free ubiquitin (Vijay-Kumar et al. 1987) (PDB entry 1UBQ [PDB] ), Arg42 is ‘upright’ and pocket III of the Ile44 surface is completely closed. The side chain of Arg42 exhibits relatively high temperature factor value compared to those of the surrounding residues, suggesting the high mobility of the Arg42 side chain. In the solution structure of free ubiquitin (Cornilescu et al. 1998) (PDB entry 1D3Z [PDB] ), the Arg42 side chain is not ordered well, also suggesting the intrinsic conformational flexibility. We then compared the known structures of the Ile44 surface of ubiquitin in complex with various ubiquitin-binding modules and found that the three hydrophobic pockets (I, II and III) of the Ile44 surface accommodate generally hydrophobic residues in all cases. As stated above, the interactions can be classified into two modes, depending on the orientation (‘upright’ or ‘inclined’) of the Arg42 side chain, which determines the size and the charge distribution of pocket III. In the Vps27 UIM/ubiquitin (Swanson et al. 2003) and Npl4 NZF/ubiquitin (Alam et al. 2004) complexes, the Arg42 side chain is ‘inclined’ to allow extensive interaction at pocket III. On the other hand, in the Vps9 CUE/ubiquitin (Prag et al. 2003), Cue2 CUE/ubiquitin (Kang et al. 2003) and Tsg101 UEV/ubiquitin (Sundquist et al. 2004) complexes, it takes the ‘upright’ conformation and electrostatically interacts with the partner proteins. Thus, Arg42 is indeed the key residue to enable interactions with a variety of structurally divergent ubiquitin-binding modules. Exceptionally, in the crystal structure of Vps23 UEV/ubiquitin complex (Teo et al. 2004), Arg72 side chain comes in to replace the position of Arg42 side chain, resulting in ‘Arg72-upright’ mode.

Comparison of GAT domains

The structure of GGA3 GAT in complex with ubiquitin provides clues to understanding the ubiquitin binding abilities of the other GAT domains when combined with a multiple sequence alignment of the GAT domains (Fig. 2A). We and others have previously shown that the GAT domains of GGA1, GGA3 and Tom1 bind ubiquitin (Yamakami et al. 2003; Katoh et al. 2004; Puertollano & Bonifacino 2004; Scott et al. 2004; Shiba et al. 2004). Tom1L1, a Tom1 homolog, also binds ubiquitin very weakly (Katoh et al. 2004), probably because ubiquitin-interacting residues are less conserved (Fig. 2A). On the other hand, we could not detect any binding of the GGA2 GAT domain to ubiquitin (Shiba et al. 2004). The multiple sequence alignment also shows that ubiquitin-interacting residues of GGA3 GAT in the crystal are identical or similar among GGAs and Tom1, except for Glu250. The residue corresponding to Glu250 is conserved in GGA1 (Glu251) and Tom1 (Glu256) but not in GGA2 (Val267) (Fig. 2A). Because Val267 of GGA2 is unable to make a salt bridge with Arg42 of ubiquitin, it is not suitable for ‘Arg42-upright’ binding mode in which Arg42 interacts with two acidic residues. In addition, Leu247 of GGA3, which is accommodated in pocket III of ubiquitin corresponds to Ala264 of GGA2. The small side chain of Ala264 of GGA2 may be unsuitable for ‘Arg42-inclined’ binding, because it is too small to fill pocket III of ubiquitin. Therefore the GGA2 GAT domain may be incompatible with either binding mode.

Ubiquitin binding Site 2 of GGA3 C-GAT

The fact that ubiquitin binding to Site 2 was not observed directly in the X-ray structure but confirmed by NMR and pull-down experiments was caused by several competing factors. We previously proposed that ubiquitin binds to helix {alpha}3 of GGA3 C-GAT, based on the reverse two-hybrid screening and pull-down assay which demonstrated that L276S, L280R and D284G mutations in helix {alpha}3 disrupted the interaction with ubiquitin (Shiba et al. 2004). Puertollano & Bonifacino (2004) also reported that the L276A mutation in helix {alpha}3 abolished binding of GGA3 GAT to ubiquitin. A mutational study by Mattera et al. (2004) demonstrated that Arg260, Ala267 and Leu277 of GGA1 GAT are important for binding of ubiquitin, which correspond to Arg259, Ala266 and Leu276 in GGA3 GAT, respectively. All these residues listed above are located distant from Site 1, but rather in or around Site 2, which includes Leu276 and Leu280 residues of helix {alpha}3, together with Phe263 and Ala266 in helix {alpha}2.

Site 2 of GGA1 C-GAT was proposed to constitute a protein–protein interaction site, by analogy with SNARE-motif binding sites of the syntaxin N-terminal domains, which show structural homology to the GAT domain (Suer et al. 2003). Site 2 of C-GAT domain of GGA1 and GGA2 had been shown to be involved in Rabaptin-5 binding by mutational analyses (Zhai et al. 2003; Mattera et al. 2004) which was confirmed by a recent report of the crystal structure of GGA1 C-GAT/Rabaptin-5 complex (Zhu et al. 2004). Interestingly, GGA3 GAT domain does not interact with Rabaptin-5 (Zhai et al. 2003; Mattera et al. 2004). This may be partly because the conformation of side chains around Phe263 of GGA3 does not fit well to the surface of Rabaptin-5. The Phe263 side chain of GGA3 clashes with the Met563 side chain of Rabaptin-5, when our GGA3 C-GAT structure was superimposed on the GGA1 C-GAT structure in complex with Rabaptin-5 (data not shown). In addition, the Arg260 side chain of GGA3 is too long to fit in a concave of Rabaptin-5 surface, where the Pro261 residue of GGA1 is accommodated.

Site 2 of GGA3 C-GAT is involved in dimerization of C-GAT in our crystal structure (Fig. 5). The dimerization is possibly a crystallization artifact because the GGA3 GAT domain is suggested to be monomeric in solution by gel filtration chromatography and small angle X-ray scattering (data not shown). The SPR measurement demonstrated that the mutations of Leu276, Leu280 and Asp284 of helix {alpha}3 decreased the ubiquitin-binding capacity, as measured by Rmax although Kd values were not significantly affected. On the other hand, the Site 1 mutants increased the Kd values about three times compared with the wild-type, indicating that the ubiquitin interaction of Site 1 is stronger than that of Site 2. In our NMR titration experiment, two distinct regions of GGA3 C-GAT are perturbed by ubiquitin, which correspond to Sites 1 and 2.

Very recently, Bilodeau et al. (2004) demonstrated that ubiquitin can indeed bind two sites of GGA3 C-GAT, using pull-down and NMR analyses. They identified ‘Site 1’ at the C-terminal half of C-GAT helix {alpha}1 (residues 218–232) and ‘Site 2’ at the N-terminal half of helix {alpha}3 (residues 272–286). These two binding sites identified by Bilodeau et al. (2004) partly overlap with ours. In our crystal, however, Site 1 consists of not only the C-terminal half of helix {alpha}1, but also the N-terminal half of helix {alpha}2, as described above. Furthermore, our NMR experiment using 15N-labeled GAT domain suggests that Site 2 includes hydrophobic residues of the C-terminal half of helix {alpha}2 (Phe263 and Ala266), in addition to the hydrophobic residues of the N-terminal half of helix {alpha}3 (Leu276 and Leu280). Their pull-down and NMR data show that Site 1 binds more strongly to ubiquitin than Site 2, which is consistent with our pull-down and SPR data. They analyzed the chemical shift changes of 15N-labeled ubiquitin upon binding of the GGA3 GAT domain, and concluded that both Site 1 and Site 2 interact with the Ile44 surface of ubiquitin. Our previous mutational analysis (Shiba et al. 2004) and NMR analysis using 15N-labeled ubiquitin in this study also support this, although the exact ubiquitin binding mode of Site 2 still awaits further X-ray or NMR analysis.

While this manuscript was in preparation, Prag et al. (2005) reported the crystal structure of the GGA3 C-GAT and ubiquitin complex. Although the packing of the complexes in their crystal is completely different from ours, ubiquitin binds to Site 1 of GGA3 C-GAT while Site 2 is involved in dimerization of C-GAT, similarly to our structure. Both two complexes in the asymmetric unit of their crystal are in the ‘Arg42-upright’ binding mode, which might be the reason for the different packing from ours that contains two different binding modes. They proposed a hypothetical model of GAT-Rabaptin-5-ubiquitin ternary complex (Prag et al. 2005). However, because Rabaptin-5 does not bind to GGA3 GAT (Zhai et al. 2003; Mattera et al. 2004), Site 2 of GGA3 GAT may serve as the platform for the second ubiquitin, rather than making a ternary GAT-Rabaptin-5-ubiquitin complex.

We previously observed that the binding of the GAT domain to ubiquitin was enhanced by GTP-bound ARF (Shiba et al. 2004). This is difficult to explain, because the N-terminal ARF-interacting N-GAT subdomain is structurally independent of the C-GAT subdomain, though two subdomains share long helix {alpha}1 (Collins et al. 2003; Shiba et al. 2003; Suer et al. 2003; Zhu et al. 2003). One possible explanation is that ARF stabilizes helix {alpha}1 which is intrinsically flexible (Shiba et al. 2003), thus stabilizing the overall structure of the GAT domain.

Biological implications of double-sided ubiquitin binding by GAT domain

The GGA3 GAT domain can bind both monoubiquitin and polyubiquitin (Fig. 6A). In many cases, the side chain of Lys48 can serve as an attachment site for subsequent ubiquitin molecules in polyubiquitin chains. In the C-GAT/ubiquitin complex, the side chain of Lys48 of ubiquitin is within a hydrogen bonding distance from the C-GAT molecule. Accordingly, when C-GAT binds monoubiquitin or the end of polyubiquitin chains, further ubiquitin conjugation could be prevented because C-GAT may prevent ubiquitin ligases from accessing the Lys48 side chain.

Several endocytic proteins contain multiple tandem copies of ubiquitin-binding modules (Di Fiore et al. 2003). One explanation for this could be that they overcome generally weak interaction between ubiquitin and ubiquitin-binding modules. Multiplication of the ubiquitin-binding modules within a protein may help to increase the probability of binding to monoubiquitin by elevating the local concentration of the ubiquitin-binding modules. The increased binding capacity of the multiple modules may also be advantageous to bind polyubiquitin chains and/or multiply ubiquitinated proteins. Another explanation for the multiplication of ubiquitin-binding modules in several endocytic proteins could be that they function as an adaptor protein to link multiple ubiquitinated proteins together.

Besides genetically copying ubiquitin-binding modules, there would be another strategy to multiply ubiquitin-binding sites. We and others found several examples of structures showing a two-site ubiquitin binding by a single module. In the crystal structure of the Vps9 CUE/ubiquitin complex, two CUE monomers dimerize by swapping {alpha}3 helices, and two distinct surfaces of the dimer interact with the ubiquitin Ile44 surface and Ile36/Leu73 region (Prag et al. 2003). We report here the structural basis of the dual ubiquitin recognition by GGA3 C-GAT using X-ray crystallography, NMR, SPR and pull-down experiments. A sequence alignment shows that the ubiquitin binding Site 1 of GGA3 C-GAT is fairly conserved in Tom1 GAT domain (Fig. 2A). Besides, our previous mutational analysis demonstrated that Site 2 of Tom1 GAT domain is directly involved in ubiquitin binding (Katoh et al. 2004). Hence it is plausible that Tom1 GAT domain has two ubiquitin binding sites analogous to GGA3 C-GAT domain. Thus, double sided ubiquitin interaction may be a recurring strategy of ubiquitin-binding motifs to multiply binding sites instead of tandem repeat of single sided ubiquitin-binding motifs.


   Experimental procedures
Top
 Abstract
 Introduction
 Results
 Discussion
 Experimental procedures
References
 
Protein expression, purification and crystallization

Residues 209–319 of human GGA3 (GGA3 C-GAT domain) were cloned into pGEX-4T-2 (Amersham) and expressed in Escherichia coli DL41 cells. For incorporation of selenomethionine (SeMet), cells were grown in LeMaster medium supplemented with 25 mg/L seleno-L-methionine (Wako pure chemical). Cells were lysed by sonication in PBS buffer. After centrifugation, the supernatant was loaded on to a glutathione Sepharose 4B (Amersham) column, and the GST fusion protein was eluted with glutathione. After cleavage of the fusion protein by thrombin, GGA3 C-GAT was purified by gel filtration chromatography. GGA3 C-GAT was dialyzed against a buffer of 1 mM Tris-HCl pH 8.0 and concentrated. The mixture of equimolar (1 mM each protein) amounts of SeMet-substituted GGA3 C-GAT domain and native bovine ubiquitin (Sigma) was crystallized using the hanging-drop vapor diffusion method. Crystals were obtained using 20% (w/v) PEG3350, 0.3 M ammonium formate and 50 mM MES pH 6.5 as a reservoir solution at 20 °C, and belong to the orthorhombic space group P212121.

X-ray data collection and structure determination

Crystals were cryoprotected in the reservoir solution supplemented with 15% ethylene glycol and frozen in liquid nitrogen. A three-wavelength data set for MAD phasing was collected to 2.9 Å resolution at PF beamline BL-6 A (Table 1). A data set at single wavelength with another crystal was collected to a higher resolution of 2.6 Å at PF-AR NW12 beamline (Table 1). Data were processed with HKL2000 (Otwinowski & Minor 1997). The selenium positions were located and phases were calculated with SOLVE (Terwilliger & Berendzen 1999). Density modification and initial model building was carried out with RESOLVE (Terwilliger 2003). Further model building was performed manually with TURBO-FRODO (Roussel & Cambilleau 1989) and CNS (Brünger et al. 1998) was used for refinement of the model. The crystal contains four C-GAT/ubiquitin complexes in an asymmetric unit. The final model structure of C-GAT consists of residues 211–300 (molecule A), 215–304 (B), 212–301 (C) and 215–302 (D), and the final model of ubiquitin consists of residues 1–72 (E), 1–73 (F), 1–72 (G) and 1–74 (H). Structure figures were created with MOLSCRIPT (Kraulis 1991), Raster3D (Merritt & Murphy 1994), GRASP (Nicholls et al. 1991) and POVScript+ (Fenn et al. 2003).

Pull-down experiments

A mixture of polyubiquitin chain (4 µg; Affinity Research Products) and monoubiquitin (2 µg; Sigma) in 200 µL assay buffer (25 mM HEPES-KOH, pH 7.4, 125 mM KOAc, 2.5 mM MgOAc, 5 mM EGTA, 0.1% NP-40, 25 µg/mL BSA) was precleared by centrifugation at 13 200 r.p.m. for 10 min. Resulting supernatant was mixed with 20 µL of Glutathione-Sepharose 4B beads (Amersham) preloaded with GST or GST fusion proteins of wild-type or mutant GGA3 GAT domain, and incubated at room temperature for 1 h. The beads were then pelleted and washed three times with an assay buffer. Proteins associated with the beads were subjected to immunoblotting with monoclonal anti-ubiquitin antibody P4D1 (Santa Cruz Biotechnology).

SPR measurements

SPR binding assay was performed at 25 °C using BIACORE 2000 (Biacore). HBS-EP buffer (10 mM HEPES pH 7.4, 150 mM NaCl, 3 mM EDTA, 0.005% (v/v) Surfactant P20) was used as a running buffer at 20 µL min–1 flow rate. The anti-GST antibody (Biacore) was immobilized on CM5 sensor chips (Biacore) according to the manufacturer's instructions. To determine the background response, GST was injected to the sensor surfaces to reach 1000 resonance units of binding. Then a range of concentrations (0.025–1.5 mM) of ubiquitin were injected over the surfaces. After GST was removed by 20 mM glycine, pH 2.0 and 0.005% SDS, GST-GGA3 GAT fusion protein was injected to bind 1000 resonance units over the same surface. The binding response was measured by injections of the same range of concentrations of ubiquitin. The net response was calculated by subtracting the background response from the binding response. Steady state responses (Req) were determined from the net response of sensorgrams using BIAevaluation 3.2 program (Biacore). The Req values were plotted against the ubiquitin concentrations, and the resulting curve was fitted to a simple 1 : 1 steady state binding model using the BIAevaluation 3.2 software:

Req = C · Rmax/(Kd + C)

Where C is the ubiquitin concentration, Rmax is the maximum binding response, and Kd is the equilibrium dissociation constant.

NMR spectroscopy

Isotopically labeled GGA3 GAT and ubiquitin were produced in E. coli grown in M9 minimal media containing [15N]NH4Cl (1 g/L) and [u-13C]glucose (2 g/L). For the mapping of ubiquitin-binding sites on GGA3 C-GAT (residues 209–304), [15N]GGA3 C-GAT was dissolved at a concentration of 0.3 mM in 30 mM sodium acetate, pH 6.0. To observe chemical shift perturbation, human ubiquitin was added to the [15N]GGA3 C-GAT solution. NMR spectra were acquired at 30 °C on Bruker DMX500 and Avance 600 MHz spectrometers. The 1H, 15N, and 13C resonances of the backbone of GGA3 C-GAT were assigned using a standard set of double and triple resonance experiments (Clore & Gronenborn 1994). For the mapping of GGA3 GAT-binding sites on ubiquitin, wild-type or mutated (M231S or L280R) GGA3 C-GAT (residues 140–319) was added to the 0.3 mM[15N]ubiquitin dissolved in 10 mM Tris-HCl, pH 7.4. Chemical shift change and/or line-broadening observed for 0.3 mM GGA C-GAT upon addition of 0.3 mM ubiquitin (or vice versa) were mapped. NMR spectra were acquired at 30 °C on a Bruker Avance 600 MHz spectrometer.

Coordinates

The atomic coordinates have been deposited in the Protein Data Bank (accession code 1WR6).


   Acknowledgements
 
We thank M. Kitabayashi for assistance with plasmid construction, protein purification and crystallization, Y. Yamada for preparing Fig. 8, Drs H. Kamikubo and M. Kataoka for their help with small angle X-ray scattering experiments. We also thank Y. Kito and K. Senda for their help in the preparation of recombinant proteins for NMR spectroscopy. This work was supported in part by Protein 3000 project and by Grants-in-Aid from the Ministry of Education, Culture, Sports, Science and Technology of Japan.


   Footnotes
 
Communicated by: Akihiko Nakano

{dagger}Present address: Institute for Protein Research, Osaka University, 3-2, Yamadaoka, Suita, Osaka 565–0871, Japan. Back

* Correspondence: E-mail: soichi.wakatsuki{at}kek.jp


   References
Top
 Abstract
 Introduction
 Results
 Discussion
 Experimental procedures
 References
 
Alam, S.L., Sun, J., Payne, M., et al. (2004) Ubiquitin interactions of NZF zinc fingers. EMBO J. 23, 1411–1421.[CrossRef][Medline]

Bilodeau, P.S., Winistorfer, S.C., Allaman, M.M., et al. (2004) The GAT domains of clathrin-associated GGA proteins have two ubiquitin-binding motifs. J. Biol. Chem. 279, 54808–54816.[Abstract/Free Full Text]

Boman, A.L. (2001) GGA proteins: new players in the sorting game. J. Cell Sci. 114, 3413–3418.[Abstract/Free Full Text]

Bonifacino, J.S. (2004) The GGA proteins: adaptors on the move. Nature Rev. Mol. Cell. Biol. 5, 23–32.[CrossRef][Medline]

Brünger, A.T., Adams, P.D., Clore, G.M., et al. (1998) Crystallography & NMR system: a new software suite for macromolecular structure determination. Acta Crystallogr. D 54, 905–921.[CrossRef][Medline]

Clore, G.M. & Gronenborn, A.M. (1994) Multidimensional heteronuclear nuclear magnetic resonance of proteins. Methods Enzymol. 239, 349–363.[Medline]

Collins, B.M., Watson, P.J. & Owen, D.J. (2003) The structure of the GGA1-GAT domain reveals the molecular basis for ARF binding and membrane association of GGAs. Dev. Cell 4, 321–332.[CrossRef][Medline]

Cornilescu, G., Marquardt, J.L., Ottiger, M. & Bax, A. (1998) Validation of protein structure from anisotropic carbonyl chemical shifts in a dilute liquid crystalline phase. J. Am. Chem. Soc. 120, 6836–6837.[CrossRef]

Di Fiore, P.P., Polo, S. & Hofmann, K. (2003) When ubiquitin meets ubiquitin receptors: a signalling connection. Nature Rev. Mol. Cell. Biol. 4, 491–497.[CrossRef][Medline]

Fenn, T.D., Ringe, D. & Petsko, G.A. (2003) POVScript+: a program for model and data visualization using persistence of vision ray-tracing. J. Appl. Cryst. 36, 944–947.[CrossRef]

Hicke, L. (2001) Protein regulation by monoubiquitin. Nature Rev. Mol. Cell. Biol. 2, 195–201.[CrossRef][Medline]

Hinners, I. & Tooze, S.A. (2003) Changing directions: clathrin-mediated transport between the Golgi and endosomes. J. Cell Sci. 116, 763–771.[Abstract/Free Full Text]

Kang, R.S., Daniels, C.M., Francis, S.A., et al. (2003) Solution structure of a CUE-ubiquitin complex reveals a conserved mode of ubiquitin binding. Cell 113, 621–630.[CrossRef][Medline]

Katoh, Y., Shiba, Y., Mitsuhashi, H., Yanagida, Y., Takatsu, H. & Nakayama, K. (2004) Tollip and Tom1 form a complex and recruit ubiquitin-conjugated proteins onto early endosomes. J. Biol. Chem. 279, 24435–24443.[Abstract/Free Full Text]

Katzmann, D.J., Babst, M. & Emr, S.D. (2001) Ubiquitin-dependent sorting into the multivesicular body pathway requires the function of a conserved endosomal protein sorting complex, ESCRT-I. Cell 106, 145–155.[CrossRef][Medline]

Kraulis, P.J. (1991) MOLSCRIPT: a program to produce both detailed and schematic plots of protein structures. J. Appl. Crystallogr. 24, 946–950.[CrossRef]

Mattera, R., Puertollano, R., Smith, W.J. & Bonifacino, J.S. (2004) The trihelical bundle subdomain of the GGA proteins interacts with multiple partners through overlapping but distinct sites. J. Biol. Chem. 279, 31409–31418.[Abstract/Free Full Text]

Merritt, E.A. & Murphy, M.E.P. (1994) Raster3D Version 2.0. A program for photorealistic molecular graphics. Acta Crystallogr. D 50, 869–873.[CrossRef][Medline]

Nakayama, K. & Wakatsuki, S. (2003) The structure and function of GGAs, the traffic controllers at the TGN sorting crossroads. Cell Struct. Funct. 28, 431–442.[CrossRef][Medline]

Nicholls, A., Sharp, K.A. & Honig, B. (1991) Protein folding and association: insights from the interfacial and thermodynamic properties of hydrocarbons. Proteins 11, 281–296.

Otwinowski, Z. & Minor, W. (1997) Processing of X-ray diffraction data collected in oscillation mode. Methods Enzymol. 276, 307–326.

Prag, G., Lee, S., Mattera, R., et al. (2005) Structural mechanism for ubiquitinated-cargo recognition by the Golgi-localized, {gamma}-ear-containing, ADP-ribosylation-factor-binding proteins. Proc. Natl. Acad. Sci. USA 102, 2334–2339.[Abstract/Free Full Text]

Prag, G., Misra, S., Jones, E.A., et al. (2003) Mechanism of ubiquitin recognition by the CUE domain of Vps9p. Cell 113, 609–620.[CrossRef][Medline]

Puertollano, R. & Bonifacino, J.S. (2004) Interactions of GGA3 with the ubiquitin sorting machinery. Nature Cell Biol. 6, 244–251.[Medline]

Raiborg, C., Rusten, T.E. & Stenmark, H. (2003) Protein sorting into multivesicular endosomes. Curr. Opin. Cell Biol. 15, 446–455.[CrossRef][Medline]

Roussel, A. & Cambilleau, C. (1989) Turbo-Frodo. In: Silicon Graphics Geometry, Partners Directory, pp. 77–79, Mountain View, CA: Silicon Graphics.

Schnell, J.D. & Hicke, L. (2003) Non-traditional functions of ubiquitin and ubiquitin-binding proteins. J. Biol. Chem. 278, 35857–35860.[Free Full Text]

Scott, P.M., Bilodeau, P.S., Zhdankina, O., et al. (2004) GGA proteins bind ubiquitin to facilitate sorting at the trans-Golgi network. Nature Cell Biol. 6, 252–259.[Medline]

Shiba, Y., Katoh, Y., Shiba, T., et al. (2004) GAT (GGA and Tom1) domain responsible for ubiquitin binding and ubiquitination. J. Biol. Chem. 279, 7105–7111.[Abstract/Free Full Text]

Shiba, T., Kawasaki, M., Takatsu, H., et al. (2003) Molecular mechanism of membrane recruitment of GGA by ARF in lysosomal protein transport. Nature Struct. Biol. 10, 386–393.[CrossRef][Medline]

Suer, S., Misra, S., Saidi, L.F. & Hurley, J.H. (2003) Structure of the GAT domain of human GGA1: a syntaxin amino-terminal domain fold in an endosomal trafficking adaptor. Proc. Natl. Acad. Sci. USA 100, 4451–4456.[Abstract/Free Full Text]

Sundquist, W.I., Schubert, H.L., Kelly, B.N., Hill, G.C., Holton, J.M. & Hill, C.P. (2004) Ubiquitin recognition by the human TSG101 protein. Mol. Cell 13, 783–789.[CrossRef][Medline]

Swanson, K.A., Kang, R.S., Stamenova, S.D., Hicke, L. & Radhakrishnan, I. (2003) Solution structure of Vps27 UIM-ubiquitin complex important for endosomal sorting and receptor downregulation. EMBO J. 22, 4597–4606.[CrossRef][Medline]

Teo, H., Veprintsev, D.B. & Williams, R.L. (2004) Structural insights into endosomal sorting complex required for transport (ESCRT-I) recognition of ubiquitinated proteins. J. Biol. Chem. 279, 28689–28696.[Abstract/Free Full Text]

Terwilliger, T.C. (2003) Automated main-chain model building by template matching and iterative fragment extension. Acta Crystallogr. D 59, 38–44.[CrossRef][Medline]

Terwilliger, T.C. & Berendzen, J. (1999) Automated MAD and MIR structure solution. Acta Crystallogr. D 55, 849–861.[CrossRef][Medline]

Vijay-Kumar, S., Bugg, C.E. & Cook, W.J. (1987) Structure of ubiquitin refined at 1.8 Å resolution. J. Mol. Biol. 194, 531–544.[CrossRef][Medline]

Yamakami, M., Yoshimori, T. & Yokosawa, H. (2003) Tom1, a VHS domain-containing protein, interacts with tollip, ubiquitin, and clathrin. J. Biol. Chem. 278, 52865–52872.[Abstract/Free Full Text]

Zhai, P., He, X., Liu, J., et al. (2003) The interaction of the human GGA1 GAT domain with rabaptin-5 is mediated by residues on its three-helix bundle. Biochemistry 42, 13901–13908.[CrossRef][Medline]

Zhu, G., Zhai, P., He, X., et al. (2003) Crystal structure of the human GGA1 GAT domain. Biochemistry 42, 6392–6399.[CrossRef][Medline]

Zhu, G., Zhai, P., He, X., et al. (2004) Crystal structure of human GGA1 GAT domain complexed with the GAT-binding domain of Rabaptin5. EMBO J. 23, 3909–3917.[CrossRef][Medline]

Received: 2 March 2005



This article has been cited by other articles:


Home page
 J. Biol. Chem.Home page
Y. Deng, Y. Guo, H. Watson, W.-C. Au, M. Shakoury-Elizeh, M. A. Basrai, J. S. Bonifacino, and C. C. Philpott
Gga2 Mediates Sequential Ubiquitin-independent and Ubiquitin-dependent Steps in the Trafficking of ARN1 from the trans-Golgi Network to the Vacuole
J. Biol. Chem., August 28, 2009; 284(35): 23830 - 23841.
[Abstract] [Full Text] [PDF]

Home page
 Mol. Biol. CellHome page
J. Wang, H.-Q. Sun, E. Macia, T. Kirchhausen, H. Watson, J. S. Bonifacino, and H. L. Yin
PI4P Promotes the Recruitment of the GGA Adaptor Proteins to the Trans-Golgi Network and Regulates Their Recognition of the Ubiquitin Sorting Signal
Mol. Biol. Cell, July 1, 2007; 18(7): 2646 - 2655.
[Abstract] [Full Text] [PDF]

Home page
 JCBHome page
Y. Pak, W. K. Glowacka, M. C. Bruce, N. Pham, and D. Rotin
Transport of LAPTM5 to lysosomes requires association with the ubiquitin ligase Nedd4, but not LAPTM5 ubiquitination
J. Cell Biol., November 20, 2006; 175(4): 631 - 645.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Kawasaki, M.
Right arrow Articles by Wakatsuki, S.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Kawasaki, M.
Right arrow Articles by Wakatsuki, S.

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE ADVANCED SEARCH TABLE OF CONTENTS