- Email: [email protected]
Crystal structure of cyclic Lys48-linked tetraubiquitin
Biochemical and Biophysical Research Communications 400 (2010) 329–333
Contents lists available at ScienceDirect
Biochemical and Biophysical Research Communications journal homepage: www.elsevier.com/locate/ybbrc
Crystal structure of cyclic Lys48-linked tetraubiquitin Tadashi Satoh a,b,1, Eri Sakata c,1,2, Shunsuke Yamamoto c,1, Yoshiki Yamaguchi b,c, Akira Sumiyoshi c, Soichi Wakatsuki a, Koichi Kato c,d,⇑ a
Structural Biology Research Center, Photon Factory, Institute of Materials Structure Science, High Energy Accelerator Research Organization (KEK), Tsukuba, Ibaraki 305-0801, Japan Structural Glycobiology Team, RIKEN Advanced Science Institute, 2-1 Hirosawa, Wako, Saitama 351-0198, Japan c 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 d Okazaki Institute for Integrative Bioscience and Institute for Molecular Science, National Institutes of Natural Sciences, 5-1 Higashiyama, Myodaiji, Okazaki 444-8787, Japan b
a r t i c l e
i n f o
Article history: Received 12 August 2010 Available online 20 August 2010 Keywords: Ubiquitin Tetraubiquitin Lys48-linked Cyclic Crystal structure
a b s t r a c t Lys48-linked polyubiquitin chains serve as a signal for protein degradation by 26S proteasomes through its Ile44 hydrophobic patches interactions. The individual ubiquitin units of each chain are conjugated through an isopeptide bond between Lys48 and the C-terminal Gly76 of the preceding units. The conformation of Lys48-linked tetraubiquitin has been shown to change dynamically depending on solution pH. Here we enzymatically synthesized a wild-type Lys48-linked tetraubiquitin for structural study. In the synthesis, cyclic and non-cyclic species were obtained as major and minor fractions, respectively. This enabled us to solve the crystal structure of tetraubiquitin exclusively with native Lys48-linkages at 1.85 Å resolution in low pH 4.6. The crystallographic data clearly showed that the C-terminus of the first ubiquitin is conjugated to the Lys48 residue of the fourth ubiquitin. The overall structure is quite similar to the closed form of engineered tetraubiquitin at near-neutral pH 6.7, previously reported, in which the Ile44 hydrophobic patches face each other. The structure of the second and the third ubiquitin units [Ub(2)–Ub(3)] connected through a native isopeptide bond is significantly different from the conformations of the corresponding linkage of the engineered tetraubiquitins, whereas the structures of Ub(1)– Ub(2) and Ub(3)–Ub(4) isopeptide bonds are almost identical to those of the previously reported structures. From these observations, we suggest that the flexible nature of the isopeptide linkage thus observed contributes to the structural arrangements of ubiquitin chains exemplified by the pHdependent closed-to-open conformational transition of tetraubiquitin. Ó 2010 Elsevier Inc. All rights reserved.
1. Introduction Ubiquitination is one of the most versatile protein modifications in eukaryotic cells to regulate a variety of cellular events [1,2]. The diversity of signals mediated by ubiquitin is derived from its ability to modify proteins in different manners. Ubiquitin conjugates to a protein substrate in a monomeric form (monoubiquitination), but Abbreviations: ABIN, A20 binding inhibitor of nuclear factor-jB protein; CUE, coupling of ubiquitin conjugation to endoplasmic reticulum degradation; DUB, deubiquitinating enzyme; NEMO, nuclear factor-jB essential modulator; Slz, thialysine; Ub, ubiquitin; UBA, ubiquitin-associated; UBAN, ubiquitin binding in ABIN and NEMO; UBZ, ubiquitin-binding ZnF; UIM, ubiquitin-interacting motif; ZnF UBP, zinc-finger ubiquitin-specific protease. ⇑ Corresponding author at: Okazaki Institute for Integrative Bioscience and Institute for Molecular Science, National Institutes of Natural Sciences, 5-1 Higashiyama, Myodaiji, Okazaki 444-8787, Japan. E-mail address: [email protected] (K. Kato). 1 These authors contributed equally to this work. 2 Present address: Department of Molecular Structural Biology, Max-PlanckInstitute of Biochemistry, Martinsried, Germany. 0006-291X/$ - see front matter Ó 2010 Elsevier Inc. All rights reserved. doi:10.1016/j.bbrc.2010.08.057
it is also conjugated by preceding ubiquitin units giving rise to polyubiquitin chain. All lysine residues of ubiquitin as well as the amino group of the N-terminal methionine are reported to be used for conjugation with the C-terminus of other ubiquitin units [3–7]. The fate of ubiquitinated proteins depends on the type of modification that they have been subjected to. Ubiquitin chains that form through their lysine (Lys) residues at position 48 are known to be involved in protein degradation by 26S proteasome [1,2]. The minimal chain required for efficient proteasomal recognition has been shown to be tetraubiquitin. Ubiquitination also has nonproteolytic roles as exemplified by Lys63-linked polyubiquitination, which functions in DNA repair and transcriptional regulation [1,2]. Structural studies have revealed that different types of linkages result in various conformations of the ubiquitin chains. These conformational varieties of ubiquitin chains mediated by specific linkage types are thought to create distinct molecular recognition mechanisms with different signals. Lys63-linked di-ubiquitin and tetraubiquitin chains exhibit extended structures in which these
330
T. Satoh et al. / Biochemical and Biophysical Research Communications 400 (2010) 329–333
canonical hydrophobic patches (Leu8, Ile44 and Val70) are exposed because of the lack of inter-molecular interactions [8–10]. Such extended structures of Lys63-linked and linear (head-to-tail) ubiquitin chains gain an advantage to recognize the UBAN domain of NEMO for example [11]. In addition, a recent study revealed that a Lys11-linked di-ubiquitin exhibits a different compact structure in which the hydrophobic patches are exposed [12]. On the other hand, Lys48-linked di-ubiquitin and tetraubiquitin chains adopt compact forms in which the conserved hydrophobic residues are shielded at the inter-molecular surface [13–15]. The cluster of the hydrophobic patch of Lys48-linked ubiquitin chains has an advantage to interact with multiple ubiquitin-interacting proteins: e.g. UBA, UIM and CUE [16–18]. A ‘‘closed form” of Lys48-linked diubiquitin units is transformed into an ‘‘open form” to recognize target proteins through the two proximate hydrophobic patches [19]. To date, 2.2–2.7 Å resolution crystal structures of engineered Lys48-linked tetraubiquitin with three distinct conformations have been solved [14,20,21] and the engineered thialysine 48–glycine 76 isopeptide linkage between the second and third ubiquitins was only observed in a crystal grown at near-neutral pH [20]. In these studies, ubiquitin chains contain engineered residues: Lys48 of the second and fourth ubiquitins are substituted by thialysine (Slz) and arginine, respectively; and they have mutations of Lys63Arg in the second and fourth ubiquitins and a deletion of Gly76 in the first ubiquitin. Hence we aimed to obtain the structure of the wild-type Lys48linked tetraubiquitin. E2-25 K is one of the E2 enzymes that have been demonstrated to form polyubiquitin chains in the absence of an E3 ubiquitin ligase in vitro. It possesses a UBA domain at its C-terminus which binds a Lys48-linked ubiquitin chain weakly [19]. Taking advantage of this remarkable property of E2-25 K, we synthesised a wild-type Lys48-linked tetraubiquitin for the structural study. In our experiment, cyclic and non-cyclic species were obtained as major and minor fractions, respectively. Hence, we solved a cyclic Lys48-linked tetraubiquitin crystal structure at 1.85 Å resolution.
crystal structure of tetraubiquitin was solved by the molecular replacement method using the program MOLREP [24] with apo ubiquitin (PDB code: 1UBQ) as a search model. The refinement procedures were carried out with REFMAC5 [25]. Model fitting to the electron density maps was performed manually using COOT [26]. The stereochemical quality of the final model was assessed by RAMPAGE [27]. The crystallographic parameters and final refinement statistics are summarized in Table 1. Molecular graphics were prepared using PyMOL [28]. 3. Results 3.1. Preparation of Lys48-linked tetraubiquitin To prepare a Lys48-linked tetraubiquitin, ubiquitin was incubated with E1 and E2-25 K in the presence of ATP and Mg2+, following the protocol previously reported [29]. After the reaction, tetraubiquitin was separated from the resultant mixture of ubiquitin chains by cation exchange chromatography. We found two different forms of tetraubiquitin in SDS–PAGE as previously reported [30]. It has been reported that E2-25 K catalyzes the formation of cyclic polyubiquitin chains as well as the non-cyclic variety in vitro [30]. Interestingly, in our SDS–PAGE analysis, the band with the higher molecular weight shift corresponding to the non-cyclic form was minor (Supplementary Fig. 1). Accordingly, we could obtain a large amount of cyclic tetraubiquitin, and the purified protein was subjected to the crystallization. 3.2. Comparison of overall structures of cyclic and non-cyclic Lys48linked tetraubiquitins The structure of tetraubiquitin was determined at 1.85 Å resolution by molecular replacement. All the C-termini of ubiquitin units of the tetraubiquitin are conjugated to Lys48 of the adjacent ubiquitin, forming a cyclic tetraubiquitin (Fig. 1). The final model contains residues 1–76 with no disordered residues, and refined to
2. Materials and methods 2.1. Enzymatic synthesis of Lys48-linked tetraubiquitin Human ubiquitin and ubiquitin E1 were expressed and purified as previously described [22]. cDNA of human E2-25 K was inserted into the pT7 vector to encode (His)6-E2-25 K. The E2-25 K protein was expressed in Escherichia coli BL21(DE3)CodonPlus and purified via Ni–NTA agarose (Qiagen) in a buffer, 50 mM Tris–HCl, pH 8.0 and 0.1 M NaCl. Samples were frozen with liquid nitrogen and stored at 80 °C until use. Ubiquitin was mixed with 0.2 lM E1, 10 lM E2-25 K, 1 mM dithiothreitol, 5 mM MgCl2, 10 mM ATP, 0.6 U/ml creatine phosphokinase, 1 mM creatine phosphate and incubated at 37 °C for 16 h. The tetraubiquitin was purified from polyubiquitin chain mixture through Resource S (GE Healthcare) cation exchange chromatography, and analyzed by SDS–PAGE. 2.2. Crystallization and structural determination of Lys48-linked tetraubiquitin Lys48-linked tetraubiquitin was crystallized by hanging drop vapor diffusion at 289 K against a well solution containing 20% PEG 3350, 0.2 M (NH4)2SO4, pH 4.6. A data set was collected using synchrotron radiation (1.0000 Å wavelength) at NW12A of Photon Factory (PF), High Energy Accelerator Research Organization (KEK). The diffraction data was processed using HKL2000 [23]. The tetraubiquitin crystal belongs to the hexagonal space group C2221 with four ubiquitin molecules per asymmetric unit. The
Table 1 Data collection and refinement statistics for cyclic Lys48-linked tetraubiquitin. Cyclic K48-linked Ub4 PDB ID
3ALB
Crystallographic data Space group Unit cell a/b/c (Å)
C2221 59.1/77.4/135.1
Data processing statistics Beam line Wavelength (Å) Resolution (Å) Total/unique reflections Completeness (%) Rmerge (%) I/r (I)
PF-AR NW12A 1.0000 50–1.85 (1.92–1.85) 189610/26663 99.8 (99.7) 8.3 (41.4) 40.2 (4.6)
Refinement statistics Resolution (Å) Rwork/Rfree (%)
20.0–1.85 19.4/23.9
Rms deviations from ideal Bond lengths (Å) Bond angles (°)
0.015 1.56
Ramachandran plot (%) Favored Allowed
99.3 0.7
Average B-factors (Å2) Protein (A/B/C/D chain) Water/sulfate
26.4/28.1/25.7/26.2 47.4/34.9
The highest-resolution shell is shown in parentheses.
T. Satoh et al. / Biochemical and Biophysical Research Communications 400 (2010) 329–333
1.85 Å resolution with an Rwork of 19.4% and Rfree of 23.9% and average B-factor of 26.6 Å2 (Table 1). The crystal belongs to space group C2221 with one set of all four molecules per asymmetric unit (Fig. 1A), and the areas of the electron density map of the isopeptide bonds between Lys48 and the C-terminal Gly76, as well as the core ubiquitin structure were very clear (Fig. 1B). As the biochemical data indicated, the crystallographic analysis also indicates that all Lys48 residues are conjugated to the C-terminal Gly76. Because we are unable to determine which ubiquitin unit is the proximal or the distal one in the cyclic tetraubiquitin structure, we assigned the numbering of the ubiquitin units according to our superimposition results with the non-cyclic form for descriptive purposes (Figs. 1 and 2). As the two hydrophobic patches, which face diubiquitin units in the cyclic and non-cyclic tetraubiquitin structures, are related by a 2-fold symmetry, the cyclic tetraubiquitin can be superimposed on the non-cyclic structure in two different ways. The superimposition to the non-cyclic form (chain A–D) as shown in Fig. 2A has a smaller root mean square (rms) deviation value of 0.47 Å for superimposed 304 Ca atoms than the other superimposition (0.97 Å, the superimposed figure not shown). We also compared our structure with the other tetraubiquitin (chain E–H, Fig. 2B) contained in the asymmetric unit, which gave in an rms deviation value of 0.66 Å (0.99 Å by the other superimposition). 3.3. Comparison of Lys48-linked isopeptide bonds between cyclic and non-cyclic tetraubiquitin structures We compared the structures of Lys48-linked isopeptide bonds between cyclic and non-cyclic tetraubiquitin, and focus particularly on a wild-type isopeptide bond conformation between the second and the third ubiquitins [Ub(2)–Ub(3)], which was not visible in previous reports [14,20,21]. The structures of Ub(1)–Ub(2)
331
Fig. 2. Comparison of overall structures of cyclic and non-cyclic tetraubiquitins. The cyclic tetraubiquitin are shown as in Fig. 1. Non-cyclic tetraubiquitin colored from proximal ubiquitin to distal ubiquitin is: tan–pink–orange–purple (1–2–3–4). Superimposed non-cyclic tetraubiquitin structures from chain A–D and E–H (PDB code: 2O6V) are shown in (A) and (B), respectively.
and Ub(3)–Ub(4) isopeptide bonds are very similar in cyclic and non-cyclic tetraubiquitins (Figs. 2A and B, upper and lower parts), suggesting that these linkages connecting the closed form of diubiquitin are well defined. In the non-cyclic engineered tetraubiquitin, the Ub(1)–Ub(2) and Ub(3)–Ub(4) linkages also have wild-type Lys48–Gly76 isopeptide bonds. On the other hand, there are significant structural differences at the Ub(2)–Ub(3) interface between cyclic and non-cyclic (chains A–D and E–H) forms. The configuration of the Lys48(2)–Gly76(3) native isopeptide bond is dissimilar to that of the thialysine Slz48(2)–Gly76(3) isopeptide bond (Figs. 3A and C). Since the chemical structures of lysine and thialysine are almost identical, the conformational flexibility of
Fig. 1. Three-dimensional structure of cyclic Lys48-linked tetraubiquitin. (A) The cyclic ubiquitin molecules are numbered according to the scheme of non-cyclic Lys48linked tetraubiquitin [20], and colored as yellow–cyan–green–marineblue, respectively. Ribbon models are also shown on the left and right, that are rotated by 45° in either direction around a vertical axis from the center model. (B) Omit Fo–Fc electron density map of Lys48–Gly76 isopeptide bonds and Gly75 of the cyclic tetraubiquitin contoured at 2.0r.
332
T. Satoh et al. / Biochemical and Biophysical Research Communications 400 (2010) 329–333
Fig. 3. Comparison of ubiquitins (2) vs. (3) and ubiquitins (1) vs. (4) interfaces of cyclic and non-cyclic tetraubiquitins. Residues involved in isopeptide bonds formation are shown in ball-and-stick models. The models of are colored as in Fig. 2. The superimposed non-cyclic tetraubiquitin structure from chain A–D are shown in (A) and (B), and structure from chain E–H are shown in (C) and (D). In the engineered tetraubiquitin, Lys48 of the second and fourth ubiquitins are substituted by thialysine (Slz) and arginine, respectively, and Gly76 is deleted in the first ubiquitin.
Ub(2)–Ub(3) linkage may result in its small interaction at the interface and due to the fact that the cyclic form was crystallized at lower pH 4.6. In the Ub(1)–Ub(4) interfaces, the structures are almost identical between cyclic and non-cyclic forms (Fig. 3B and D), even though the non-cyclic tetraubiquitin contains engineered residues including a mutation of Lys48Arg in Ub(4) and a deletion of Gly76 in Ub(1). This indicates that cyclization in itself does not induce a large conformational change of its structure and therefore the Lys48 of Ub(4) will be in close proximity to Gly76 of Ub(1) in the closed form of the wild-type tetraubiquitin chain. 4. Discussion In this study, we solved a crystal structure of cyclic Lys48linked tetraubiquitin (Fig. 1A). The orientations of ubiquitin units are quite similar to those of non-cyclic tetraubiquitin (Fig. 2), and the Ile44 hydrophobic surfaces are shielded with each other as previously reported [20]. The interaction modes through the Ile44 hydrophobic surface of the di-ubiquitin unit were described in detail previously [13,20]. To date, three crystal structures of non-cyclic Lys48-linked tetraubiquitin with distinct conformations have been reported, and it appears that their conformations are altered depending on the crystallization conditions with different pH values [14,20,21]. A closely packed conformation of the non-cyclic tetraubiquitin was obtained at near-neutral pH 6.7 [20], whereas an open conformation with exposed Ile44-surfaces prevailed at pH 5.0 [14]. Similar pH dependent conformational changes of tetraand di-ubiquitin were also revealed by NMR spectroscopy [15,20,31]. In our study, cyclic tetraubiquitin exhibits a closed form even though it was crystallized at pH 4.6, and a significant conformational change is observed around the Ub(2)–Ub(3) linkage. Ubiquitin functions as a signal molecule to interact with numerous proteins. Most ubiquitin-interacting proteins possess cognate ubiquitin interacting domain(s), e.g. UBA, UIM, CUE, UBZ, and bind to at least one canonical Ile44 hydrophobic surface on the ubiquitin chains [32]. An exception to this rule is seen with the variations in
ZnF domains including ZnF UBP domain of deubiquitinase (DUB) isopeptidase T [32,33]. These domains recognize monoubiquitin by binding to three different regions namely the canonical Ile44 surface, the C-terminal GG motif and a polar surface centered on Asp58. Hence, the ZnF UBP domain of cognate DUBs may be able to recognize cyclic Lys48-linked tetraubiquitin through the GG motifs of the second and the fourth ubiquitin units (Fig. 1A). On the other hand, it was shown that cyclic triubiquitin resisted hydrolysis by isopeptidase T in vitro, although it could be disassembled to monoubiquitin by isopeptidase(s) in a red blood cell extract [30]. Given the results and the dynamic properties of ubiquitin chains, it is suggested that cyclic Lys48-linked tetraubiquitin undergoes a dynamic conformational change in which the Ile44 hydrophobic patches are at least transiently exposed and thereby recognized by the deubiquitinating enzymes as observed in the Lys48-linked diubiquitin interacting with the UBA domain [34]. The Lys48-linked tetra- and di-ubiquitin chains exhibit the closed form at the physiological pH, in which the Ile44 hydrophobic surfaces of two consecutive ubiquitin units are shielded with each other. It seems conceivable that Lys48-linked chains composed of an even-number of ubiquitin units could generally exhibit closed conformations with all shielded hydrophobic surfaces, while ubiquitin chains with an odd number contain at least one unpaired ubiquitin with an exposed Ile44 surface. In the Lys48linked polyubiquitin synthesis, ubiquitin chains with an odd number of ubiquitin units leaving the proximal unit open can accept additional ubiquitin through the exposed Ile44 surfaces: Lys48 of the acceptor ubiquitin can be accommodated in the area around E2–ubiquitin thioester bond. In contrast, a conformational transition into an open form would be prerequisite for ubiquitin ligation on the Lys48-linked chains composed of even number of ubiquitin units. On the basis of the present structural data, we suggest that the flexible nature of the isopeptide linkage thus observed contributes to the structural arrangements of ubiquitin chains exemplified by the pH-dependent closed-to-open conformational transition of tetraubiquitin [14,15,20,21,31].
T. Satoh et al. / Biochemical and Biophysical Research Communications 400 (2010) 329–333
In general, the formation and elongation of polyubiquitin chains are performed by cooperation of E2 and E3 enzymes [1]. In contrast, E2-25 K is one of the E2 enzymes that has been demonstrated to form polyubiquitin chains in the absence of E3 ubiquitin ligase in vitro [29]. E2-25 K is a unique E2 enzyme that contains not only an E2 catalytic domain but also a UBA domain at its C-terminus [19,29]. The UBA domain is shown to be involved in polyubiquitin chain formation [35] and have inhibitory function in ubiquitin transfer from E1 [34]. The UBA domain is supposed to be involved in cyclization of polyubiquitin by connecting proximal ubiquitin to the distal end of ubiquitin chains. Possibly the unique UBA domain interacts with the Ile44 surface of one ubiquitin unit in longer Lys48-linked polyubiquitin chains unit and thereby cause their conformational changes to facilitate polyubiquitin chain elongation and cyclization in the absence of E3 ligase. 5. Concluding remarks In this study, we solved the crystal structure of cyclic Lys48linked tetraubiquitin at 1.85 Å resolution, which revealed the flexible nature of the wild-type isopeptide bond between the second and the third ubiquitins. It has been suggested that if cyclic polyubiquitins form in cells, they probably do not accumulate as dead-end products [30]. However, potential endoisopeptidase activity of DUBs remains largely unknown to date. Structural information of our cyclic Lys48-linked tetraubiquitin can be utilized for further investigation of the potential endoisopeptidase activity of DUBs and dynamics of Lys48-linked polyubiquitin chains. Acknowledgments cDNA encoding E2-25 K was kindly provided from Dr. Keiji Tanaka (The Tokyo Metropolitan Institute, Tokyo, Japan). We thank Mr. Takashi Hirano (Nagoya City University, Nagoya, Japan) for helpful comments on the manuscript. This work was supported by Grant-in-Aid for Scientific Research on Innovative Areas (20107004), Grant-in-Aid for Scientific Research (B) (21370050), and Targeted Proteins Research Program from the Ministry of Education, Culture, Sports, Science and Technology. Appendix A. Supplementary material Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.bbrc.2010.08.057. References [1] A. Hershko, A. Ciechanover, The ubiquitin system, Annu. Rev. Biochem. 67 (1998) 425–479. [2] C.M. Pickart, M.J. Eddins, Ubiquitin: structures, functions, mechanisms, Biochim. Biophys. Acta 1695 (2004) 55–72. [3] H.T. Kim, K.P. Kim, F. Lledias, A.F. Kisselev, K.M. Scaglione, D. Skowyra, S.P. Gygi, A.L. Goldberg, Certain pairs of ubiquitin-conjugating enzymes (E2s) and ubiquitin-protein ligases (E3s) synthesize nondegradable forked ubiquitin chains containing all possible isopeptide linkages, J. Biol. Chem. 282 (2007) 17375–17386. [4] J. Peng, D. Schwartz, J.E. Elias, C.C. Thoreen, D. Cheng, G. Marsischky, J. Roelofs, D. Finley, S.P. Gygi, A proteomics approach to understanding protein ubiquitination, Nat. Biotechnol. 21 (2003) 921–926. [5] F. Ikeda, I. Dikic, Atypical ubiquitin chains: new molecular signals. Protein modifications: beyond the usual suspects review series, EMBO Rep. 9 (2008) 536–542. [6] O.V. Baboshina, A.L. Haas, Novel multiubiquitin chain linkages catalyzed by the conjugating enzymes E2EPF and RAD6 are recognized by 26 S proteasome subunit 5, J. Biol. Chem. 271 (1996) 2823–2831.
333
[7] T. Kirisako, K. Kamei, S. Murata, M. Kato, H. Fukumoto, M. Kanie, S. Sano, F. Tokunaga, K. Tanaka, K. Iwai, A ubiquitin ligase complex assembles linear polyubiquitin chains, EMBO J. 25 (2006) 4877–4887. [8] R. Varadan, M. Assfalg, A. Haririnia, S. Raasi, C. Pickart, D. Fushman, Solution conformation of Lys63-linked di-ubiquitin chain provides clues to functional diversity of polyubiquitin signaling, J. Biol. Chem. 279 (2004) 7055–7063. [9] D. Komander, F. Reyes-Turcu, J.D. Licchesi, P. Odenwaelder, K.D. Wilkinson, D. Barford, Molecular discrimination of structurally equivalent Lys 63-linked and linear polyubiquitin chains, EMBO Rep. 10 (2009) 466–473. [10] A.B. Datta, G.L. Hura, C. Wolberger, The structure and conformation of Lys 63linked tetraubiquitin, J. Mol. Biol. 392 (2009) 1117–1124. [11] S. Rahighi, F. Ikeda, M. Kawasaki, M. Akutsu, N. Suzuki, R. Kato, T. Kensche, T. Uejima, S. Bloor, D. Komander, F. Randow, S. Wakatsuki, I. Dikic, Specific recognition of linear ubiquitin chains by NEMO is important for NF-kappaB activation, Cell 136 (2009) 1098–1109. [12] A. Bremm, S.M. Freund, D. Komander, Lys11-linked ubiquitin chains adopt compact conformations and are preferentially hydrolyzed by the deubiquitinase Cezanne, Nat. Struct. Mol. Biol. 17 (2010) 939–947. [13] W.J. Cook, L.C. Jeffrey, M. Carson, Z. Chen, C.M. Pickart, Structure of a diubiquitin conjugate and a model for interaction with ubiquitin conjugating enzyme (E2), J. Biol. Chem. 267 (1992) 16467–16471. [14] W.J. Cook, L.C. Jeffrey, E. Kasperek, C.M. Pickart, Structure of tetraubiquitin shows how multiubiquitin chains can be formed, J. Mol. Biol. 236 (1994) 601– 609. [15] R. Varadan, O. Walker, C. Pickart, D. Fushman, Structural properties of polyubiquitin chains in solution, J. Mol. Biol. 324 (2002) 637–647. [16] R. Varadan, M. Assfalg, S. Raasi, C. Pickart, D. Fushman, Structural determinants for selective recognition of a Lys48-linked polyubiquitin chain by a UBA domain, Mol. Cell 18 (2005) 687–698. [17] N. Zhang, Q. Wang, A. Ehlinger, L. Randles, J.W. Lary, Y. Kang, A. Haririnia, A.J. Storaska, J.L. Cole, D. Fushman, K.J. Walters, Structure of the s5a:k48-linked diubiquitin complex and its interactions with rpn13, Mol. Cell 35 (2009) 280– 290. [18] R.S. Kang, C.M. Daniels, S.A. Francis, S.C. Shih, W.J. Salerno, L. Hicke, I. Radhakrishnan, Solution structure of a CUE-ubiquitin complex reveals a conserved mode of ubiquitin binding, Cell 113 (2003) 621–630. [19] S. Raasi, R. Varadan, D. Fushman, C.M. Pickart, Diverse polyubiquitin interaction properties of ubiquitin-associated domains, Nat. Struct. Mol. Biol. 12 (2005) 708–714. [20] M.J. Eddins, R. Varadan, D. Fushman, C.M. Pickart, C. Wolberger, Crystal structure and solution NMR studies of Lys48-linked tetraubiquitin at neutral pH, J. Mol. Biol. 367 (2007) 204–211. [21] C.L. Phillips, J. Thrower, C.M. Pickart, C.P. Hill, Structure of a new crystal form of tetraubiquitin, Acta Crystallogr. D Biol. Crystallogr. 57 (2001) 341–344. [22] E. Sakata, T. Satoh, S. Yamamoto, Y. Yamaguchi, M. Yagi-Utsumi, E. Kurimoto, K. Tanaka, S. Wakatsuki, K. Kato, Crystal structure of UbcH5b–ubiquitin intermediate: insight into the formation of the self-assembled E2–Ub conjugates, Structure 18 (2010) 138–147. [23] Z. Otwinowski, W. Minor, Processing of X-ray diffraction data collected in oscillation mode, Methods Enzymol. 276 (1997) 307–326. [24] A. Vagin, A. Teplyakov, MOLREP: an automated program for molecular replacement, J. Appl. Crystallogr. 30 (1997) 1022–1025. [25] G.N. Murshudov, A.A. Vagin, E.J. Dodson, Refinement of macromolecular structures by the maximum-likelihood method, Acta Crystallogr. D Biol. Crystallogr. 53 (1997) 240–255. [26] P. Emsley, K. Cowtan, Coot: model-building tools for molecular graphics, Acta Crystallogr. D Biol. Crystallogr. 60 (2004) 2126–2132. [27] S.C. Lovell, I.W. Davis, W.B. Arendall 3rd, P.I. de Bakker, J.M. Word, M.G. Prisant, J.S. Richardson, D.C. Richardson, Structure validation by Calpha geometry: /, w and Cb deviation, Proteins 50 (2003) 437–450. [28] W. DeLano, Polym. Mol. Graph. Syst. (2002). [29] Z. Chen, C.M. Pickart, A 25-kilodalton ubiquitin carrier protein (E2) catalyzes multi-ubiquitin chain synthesis via lysine 48 of ubiquitin, J. Biol. Chem. 265 (1990) 21835–21842. [30] T. Yao, R.E. Cohen, Cyclization of polyubiquitin by the E2-25K ubiquitin conjugating enzyme, J. Biol. Chem. 275 (2000) 36862–36868. [31] Y. Ryabov, D. Fushman, Interdomain mobility in di-ubiquitin revealed by NMR, Proteins 63 (2006) 787–796. [32] I. Dikic, S. Wakatsuki, K.J. Walters, Ubiquitin-binding domains – from structures to functions, Nat. Rev. Mol. Cell Biol. 10 (2009) 659–671. [33] F.E. Reyes-Turcu, J.R. Shanks, D. Komander, K.D. Wilkinson, Recognition of polyubiquitin isoforms by the multiple ubiquitin binding modules of isopeptidase T, J. Biol. Chem. 283 (2008) 19581–19592. [34] A. Pichler, P. Knipscheer, E. Oberhofer, W.J. van Dijk, R. Korner, J.V. Olsen, S. Jentsch, F. Melchior, T.K. Sixma, SUMO modification of the ubiquitinconjugating enzyme E2-25K, Nat. Struct. Mol. Biol. 12 (2005) 264–269. [35] M.T. Haldeman, G. Xia, E.M. Kasperek, C.M. Pickart, Structure and function of ubiquitin conjugating enzyme E2-25K: the tail is a core-dependent activity element, Biochemistry 36 (1997) 10526–10537.
Contents lists available at ScienceDirect
Biochemical and Biophysical Research Communications journal homepage: www.elsevier.com/locate/ybbrc
Crystal structure of cyclic Lys48-linked tetraubiquitin Tadashi Satoh a,b,1, Eri Sakata c,1,2, Shunsuke Yamamoto c,1, Yoshiki Yamaguchi b,c, Akira Sumiyoshi c, Soichi Wakatsuki a, Koichi Kato c,d,⇑ a
Structural Biology Research Center, Photon Factory, Institute of Materials Structure Science, High Energy Accelerator Research Organization (KEK), Tsukuba, Ibaraki 305-0801, Japan Structural Glycobiology Team, RIKEN Advanced Science Institute, 2-1 Hirosawa, Wako, Saitama 351-0198, Japan c 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 d Okazaki Institute for Integrative Bioscience and Institute for Molecular Science, National Institutes of Natural Sciences, 5-1 Higashiyama, Myodaiji, Okazaki 444-8787, Japan b
a r t i c l e
i n f o
Article history: Received 12 August 2010 Available online 20 August 2010 Keywords: Ubiquitin Tetraubiquitin Lys48-linked Cyclic Crystal structure
a b s t r a c t Lys48-linked polyubiquitin chains serve as a signal for protein degradation by 26S proteasomes through its Ile44 hydrophobic patches interactions. The individual ubiquitin units of each chain are conjugated through an isopeptide bond between Lys48 and the C-terminal Gly76 of the preceding units. The conformation of Lys48-linked tetraubiquitin has been shown to change dynamically depending on solution pH. Here we enzymatically synthesized a wild-type Lys48-linked tetraubiquitin for structural study. In the synthesis, cyclic and non-cyclic species were obtained as major and minor fractions, respectively. This enabled us to solve the crystal structure of tetraubiquitin exclusively with native Lys48-linkages at 1.85 Å resolution in low pH 4.6. The crystallographic data clearly showed that the C-terminus of the first ubiquitin is conjugated to the Lys48 residue of the fourth ubiquitin. The overall structure is quite similar to the closed form of engineered tetraubiquitin at near-neutral pH 6.7, previously reported, in which the Ile44 hydrophobic patches face each other. The structure of the second and the third ubiquitin units [Ub(2)–Ub(3)] connected through a native isopeptide bond is significantly different from the conformations of the corresponding linkage of the engineered tetraubiquitins, whereas the structures of Ub(1)– Ub(2) and Ub(3)–Ub(4) isopeptide bonds are almost identical to those of the previously reported structures. From these observations, we suggest that the flexible nature of the isopeptide linkage thus observed contributes to the structural arrangements of ubiquitin chains exemplified by the pHdependent closed-to-open conformational transition of tetraubiquitin. Ó 2010 Elsevier Inc. All rights reserved.
1. Introduction Ubiquitination is one of the most versatile protein modifications in eukaryotic cells to regulate a variety of cellular events [1,2]. The diversity of signals mediated by ubiquitin is derived from its ability to modify proteins in different manners. Ubiquitin conjugates to a protein substrate in a monomeric form (monoubiquitination), but Abbreviations: ABIN, A20 binding inhibitor of nuclear factor-jB protein; CUE, coupling of ubiquitin conjugation to endoplasmic reticulum degradation; DUB, deubiquitinating enzyme; NEMO, nuclear factor-jB essential modulator; Slz, thialysine; Ub, ubiquitin; UBA, ubiquitin-associated; UBAN, ubiquitin binding in ABIN and NEMO; UBZ, ubiquitin-binding ZnF; UIM, ubiquitin-interacting motif; ZnF UBP, zinc-finger ubiquitin-specific protease. ⇑ Corresponding author at: Okazaki Institute for Integrative Bioscience and Institute for Molecular Science, National Institutes of Natural Sciences, 5-1 Higashiyama, Myodaiji, Okazaki 444-8787, Japan. E-mail address: [email protected] (K. Kato). 1 These authors contributed equally to this work. 2 Present address: Department of Molecular Structural Biology, Max-PlanckInstitute of Biochemistry, Martinsried, Germany. 0006-291X/$ - see front matter Ó 2010 Elsevier Inc. All rights reserved. doi:10.1016/j.bbrc.2010.08.057
it is also conjugated by preceding ubiquitin units giving rise to polyubiquitin chain. All lysine residues of ubiquitin as well as the amino group of the N-terminal methionine are reported to be used for conjugation with the C-terminus of other ubiquitin units [3–7]. The fate of ubiquitinated proteins depends on the type of modification that they have been subjected to. Ubiquitin chains that form through their lysine (Lys) residues at position 48 are known to be involved in protein degradation by 26S proteasome [1,2]. The minimal chain required for efficient proteasomal recognition has been shown to be tetraubiquitin. Ubiquitination also has nonproteolytic roles as exemplified by Lys63-linked polyubiquitination, which functions in DNA repair and transcriptional regulation [1,2]. Structural studies have revealed that different types of linkages result in various conformations of the ubiquitin chains. These conformational varieties of ubiquitin chains mediated by specific linkage types are thought to create distinct molecular recognition mechanisms with different signals. Lys63-linked di-ubiquitin and tetraubiquitin chains exhibit extended structures in which these
330
T. Satoh et al. / Biochemical and Biophysical Research Communications 400 (2010) 329–333
canonical hydrophobic patches (Leu8, Ile44 and Val70) are exposed because of the lack of inter-molecular interactions [8–10]. Such extended structures of Lys63-linked and linear (head-to-tail) ubiquitin chains gain an advantage to recognize the UBAN domain of NEMO for example [11]. In addition, a recent study revealed that a Lys11-linked di-ubiquitin exhibits a different compact structure in which the hydrophobic patches are exposed [12]. On the other hand, Lys48-linked di-ubiquitin and tetraubiquitin chains adopt compact forms in which the conserved hydrophobic residues are shielded at the inter-molecular surface [13–15]. The cluster of the hydrophobic patch of Lys48-linked ubiquitin chains has an advantage to interact with multiple ubiquitin-interacting proteins: e.g. UBA, UIM and CUE [16–18]. A ‘‘closed form” of Lys48-linked diubiquitin units is transformed into an ‘‘open form” to recognize target proteins through the two proximate hydrophobic patches [19]. To date, 2.2–2.7 Å resolution crystal structures of engineered Lys48-linked tetraubiquitin with three distinct conformations have been solved [14,20,21] and the engineered thialysine 48–glycine 76 isopeptide linkage between the second and third ubiquitins was only observed in a crystal grown at near-neutral pH [20]. In these studies, ubiquitin chains contain engineered residues: Lys48 of the second and fourth ubiquitins are substituted by thialysine (Slz) and arginine, respectively; and they have mutations of Lys63Arg in the second and fourth ubiquitins and a deletion of Gly76 in the first ubiquitin. Hence we aimed to obtain the structure of the wild-type Lys48linked tetraubiquitin. E2-25 K is one of the E2 enzymes that have been demonstrated to form polyubiquitin chains in the absence of an E3 ubiquitin ligase in vitro. It possesses a UBA domain at its C-terminus which binds a Lys48-linked ubiquitin chain weakly [19]. Taking advantage of this remarkable property of E2-25 K, we synthesised a wild-type Lys48-linked tetraubiquitin for the structural study. In our experiment, cyclic and non-cyclic species were obtained as major and minor fractions, respectively. Hence, we solved a cyclic Lys48-linked tetraubiquitin crystal structure at 1.85 Å resolution.
crystal structure of tetraubiquitin was solved by the molecular replacement method using the program MOLREP [24] with apo ubiquitin (PDB code: 1UBQ) as a search model. The refinement procedures were carried out with REFMAC5 [25]. Model fitting to the electron density maps was performed manually using COOT [26]. The stereochemical quality of the final model was assessed by RAMPAGE [27]. The crystallographic parameters and final refinement statistics are summarized in Table 1. Molecular graphics were prepared using PyMOL [28]. 3. Results 3.1. Preparation of Lys48-linked tetraubiquitin To prepare a Lys48-linked tetraubiquitin, ubiquitin was incubated with E1 and E2-25 K in the presence of ATP and Mg2+, following the protocol previously reported [29]. After the reaction, tetraubiquitin was separated from the resultant mixture of ubiquitin chains by cation exchange chromatography. We found two different forms of tetraubiquitin in SDS–PAGE as previously reported [30]. It has been reported that E2-25 K catalyzes the formation of cyclic polyubiquitin chains as well as the non-cyclic variety in vitro [30]. Interestingly, in our SDS–PAGE analysis, the band with the higher molecular weight shift corresponding to the non-cyclic form was minor (Supplementary Fig. 1). Accordingly, we could obtain a large amount of cyclic tetraubiquitin, and the purified protein was subjected to the crystallization. 3.2. Comparison of overall structures of cyclic and non-cyclic Lys48linked tetraubiquitins The structure of tetraubiquitin was determined at 1.85 Å resolution by molecular replacement. All the C-termini of ubiquitin units of the tetraubiquitin are conjugated to Lys48 of the adjacent ubiquitin, forming a cyclic tetraubiquitin (Fig. 1). The final model contains residues 1–76 with no disordered residues, and refined to
2. Materials and methods 2.1. Enzymatic synthesis of Lys48-linked tetraubiquitin Human ubiquitin and ubiquitin E1 were expressed and purified as previously described [22]. cDNA of human E2-25 K was inserted into the pT7 vector to encode (His)6-E2-25 K. The E2-25 K protein was expressed in Escherichia coli BL21(DE3)CodonPlus and purified via Ni–NTA agarose (Qiagen) in a buffer, 50 mM Tris–HCl, pH 8.0 and 0.1 M NaCl. Samples were frozen with liquid nitrogen and stored at 80 °C until use. Ubiquitin was mixed with 0.2 lM E1, 10 lM E2-25 K, 1 mM dithiothreitol, 5 mM MgCl2, 10 mM ATP, 0.6 U/ml creatine phosphokinase, 1 mM creatine phosphate and incubated at 37 °C for 16 h. The tetraubiquitin was purified from polyubiquitin chain mixture through Resource S (GE Healthcare) cation exchange chromatography, and analyzed by SDS–PAGE. 2.2. Crystallization and structural determination of Lys48-linked tetraubiquitin Lys48-linked tetraubiquitin was crystallized by hanging drop vapor diffusion at 289 K against a well solution containing 20% PEG 3350, 0.2 M (NH4)2SO4, pH 4.6. A data set was collected using synchrotron radiation (1.0000 Å wavelength) at NW12A of Photon Factory (PF), High Energy Accelerator Research Organization (KEK). The diffraction data was processed using HKL2000 [23]. The tetraubiquitin crystal belongs to the hexagonal space group C2221 with four ubiquitin molecules per asymmetric unit. The
Table 1 Data collection and refinement statistics for cyclic Lys48-linked tetraubiquitin. Cyclic K48-linked Ub4 PDB ID
3ALB
Crystallographic data Space group Unit cell a/b/c (Å)
C2221 59.1/77.4/135.1
Data processing statistics Beam line Wavelength (Å) Resolution (Å) Total/unique reflections Completeness (%) Rmerge (%) I/r (I)
PF-AR NW12A 1.0000 50–1.85 (1.92–1.85) 189610/26663 99.8 (99.7) 8.3 (41.4) 40.2 (4.6)
Refinement statistics Resolution (Å) Rwork/Rfree (%)
20.0–1.85 19.4/23.9
Rms deviations from ideal Bond lengths (Å) Bond angles (°)
0.015 1.56
Ramachandran plot (%) Favored Allowed
99.3 0.7
Average B-factors (Å2) Protein (A/B/C/D chain) Water/sulfate
26.4/28.1/25.7/26.2 47.4/34.9
The highest-resolution shell is shown in parentheses.
T. Satoh et al. / Biochemical and Biophysical Research Communications 400 (2010) 329–333
1.85 Å resolution with an Rwork of 19.4% and Rfree of 23.9% and average B-factor of 26.6 Å2 (Table 1). The crystal belongs to space group C2221 with one set of all four molecules per asymmetric unit (Fig. 1A), and the areas of the electron density map of the isopeptide bonds between Lys48 and the C-terminal Gly76, as well as the core ubiquitin structure were very clear (Fig. 1B). As the biochemical data indicated, the crystallographic analysis also indicates that all Lys48 residues are conjugated to the C-terminal Gly76. Because we are unable to determine which ubiquitin unit is the proximal or the distal one in the cyclic tetraubiquitin structure, we assigned the numbering of the ubiquitin units according to our superimposition results with the non-cyclic form for descriptive purposes (Figs. 1 and 2). As the two hydrophobic patches, which face diubiquitin units in the cyclic and non-cyclic tetraubiquitin structures, are related by a 2-fold symmetry, the cyclic tetraubiquitin can be superimposed on the non-cyclic structure in two different ways. The superimposition to the non-cyclic form (chain A–D) as shown in Fig. 2A has a smaller root mean square (rms) deviation value of 0.47 Å for superimposed 304 Ca atoms than the other superimposition (0.97 Å, the superimposed figure not shown). We also compared our structure with the other tetraubiquitin (chain E–H, Fig. 2B) contained in the asymmetric unit, which gave in an rms deviation value of 0.66 Å (0.99 Å by the other superimposition). 3.3. Comparison of Lys48-linked isopeptide bonds between cyclic and non-cyclic tetraubiquitin structures We compared the structures of Lys48-linked isopeptide bonds between cyclic and non-cyclic tetraubiquitin, and focus particularly on a wild-type isopeptide bond conformation between the second and the third ubiquitins [Ub(2)–Ub(3)], which was not visible in previous reports [14,20,21]. The structures of Ub(1)–Ub(2)
331
Fig. 2. Comparison of overall structures of cyclic and non-cyclic tetraubiquitins. The cyclic tetraubiquitin are shown as in Fig. 1. Non-cyclic tetraubiquitin colored from proximal ubiquitin to distal ubiquitin is: tan–pink–orange–purple (1–2–3–4). Superimposed non-cyclic tetraubiquitin structures from chain A–D and E–H (PDB code: 2O6V) are shown in (A) and (B), respectively.
and Ub(3)–Ub(4) isopeptide bonds are very similar in cyclic and non-cyclic tetraubiquitins (Figs. 2A and B, upper and lower parts), suggesting that these linkages connecting the closed form of diubiquitin are well defined. In the non-cyclic engineered tetraubiquitin, the Ub(1)–Ub(2) and Ub(3)–Ub(4) linkages also have wild-type Lys48–Gly76 isopeptide bonds. On the other hand, there are significant structural differences at the Ub(2)–Ub(3) interface between cyclic and non-cyclic (chains A–D and E–H) forms. The configuration of the Lys48(2)–Gly76(3) native isopeptide bond is dissimilar to that of the thialysine Slz48(2)–Gly76(3) isopeptide bond (Figs. 3A and C). Since the chemical structures of lysine and thialysine are almost identical, the conformational flexibility of
Fig. 1. Three-dimensional structure of cyclic Lys48-linked tetraubiquitin. (A) The cyclic ubiquitin molecules are numbered according to the scheme of non-cyclic Lys48linked tetraubiquitin [20], and colored as yellow–cyan–green–marineblue, respectively. Ribbon models are also shown on the left and right, that are rotated by 45° in either direction around a vertical axis from the center model. (B) Omit Fo–Fc electron density map of Lys48–Gly76 isopeptide bonds and Gly75 of the cyclic tetraubiquitin contoured at 2.0r.
332
T. Satoh et al. / Biochemical and Biophysical Research Communications 400 (2010) 329–333
Fig. 3. Comparison of ubiquitins (2) vs. (3) and ubiquitins (1) vs. (4) interfaces of cyclic and non-cyclic tetraubiquitins. Residues involved in isopeptide bonds formation are shown in ball-and-stick models. The models of are colored as in Fig. 2. The superimposed non-cyclic tetraubiquitin structure from chain A–D are shown in (A) and (B), and structure from chain E–H are shown in (C) and (D). In the engineered tetraubiquitin, Lys48 of the second and fourth ubiquitins are substituted by thialysine (Slz) and arginine, respectively, and Gly76 is deleted in the first ubiquitin.
Ub(2)–Ub(3) linkage may result in its small interaction at the interface and due to the fact that the cyclic form was crystallized at lower pH 4.6. In the Ub(1)–Ub(4) interfaces, the structures are almost identical between cyclic and non-cyclic forms (Fig. 3B and D), even though the non-cyclic tetraubiquitin contains engineered residues including a mutation of Lys48Arg in Ub(4) and a deletion of Gly76 in Ub(1). This indicates that cyclization in itself does not induce a large conformational change of its structure and therefore the Lys48 of Ub(4) will be in close proximity to Gly76 of Ub(1) in the closed form of the wild-type tetraubiquitin chain. 4. Discussion In this study, we solved a crystal structure of cyclic Lys48linked tetraubiquitin (Fig. 1A). The orientations of ubiquitin units are quite similar to those of non-cyclic tetraubiquitin (Fig. 2), and the Ile44 hydrophobic surfaces are shielded with each other as previously reported [20]. The interaction modes through the Ile44 hydrophobic surface of the di-ubiquitin unit were described in detail previously [13,20]. To date, three crystal structures of non-cyclic Lys48-linked tetraubiquitin with distinct conformations have been reported, and it appears that their conformations are altered depending on the crystallization conditions with different pH values [14,20,21]. A closely packed conformation of the non-cyclic tetraubiquitin was obtained at near-neutral pH 6.7 [20], whereas an open conformation with exposed Ile44-surfaces prevailed at pH 5.0 [14]. Similar pH dependent conformational changes of tetraand di-ubiquitin were also revealed by NMR spectroscopy [15,20,31]. In our study, cyclic tetraubiquitin exhibits a closed form even though it was crystallized at pH 4.6, and a significant conformational change is observed around the Ub(2)–Ub(3) linkage. Ubiquitin functions as a signal molecule to interact with numerous proteins. Most ubiquitin-interacting proteins possess cognate ubiquitin interacting domain(s), e.g. UBA, UIM, CUE, UBZ, and bind to at least one canonical Ile44 hydrophobic surface on the ubiquitin chains [32]. An exception to this rule is seen with the variations in
ZnF domains including ZnF UBP domain of deubiquitinase (DUB) isopeptidase T [32,33]. These domains recognize monoubiquitin by binding to three different regions namely the canonical Ile44 surface, the C-terminal GG motif and a polar surface centered on Asp58. Hence, the ZnF UBP domain of cognate DUBs may be able to recognize cyclic Lys48-linked tetraubiquitin through the GG motifs of the second and the fourth ubiquitin units (Fig. 1A). On the other hand, it was shown that cyclic triubiquitin resisted hydrolysis by isopeptidase T in vitro, although it could be disassembled to monoubiquitin by isopeptidase(s) in a red blood cell extract [30]. Given the results and the dynamic properties of ubiquitin chains, it is suggested that cyclic Lys48-linked tetraubiquitin undergoes a dynamic conformational change in which the Ile44 hydrophobic patches are at least transiently exposed and thereby recognized by the deubiquitinating enzymes as observed in the Lys48-linked diubiquitin interacting with the UBA domain [34]. The Lys48-linked tetra- and di-ubiquitin chains exhibit the closed form at the physiological pH, in which the Ile44 hydrophobic surfaces of two consecutive ubiquitin units are shielded with each other. It seems conceivable that Lys48-linked chains composed of an even-number of ubiquitin units could generally exhibit closed conformations with all shielded hydrophobic surfaces, while ubiquitin chains with an odd number contain at least one unpaired ubiquitin with an exposed Ile44 surface. In the Lys48linked polyubiquitin synthesis, ubiquitin chains with an odd number of ubiquitin units leaving the proximal unit open can accept additional ubiquitin through the exposed Ile44 surfaces: Lys48 of the acceptor ubiquitin can be accommodated in the area around E2–ubiquitin thioester bond. In contrast, a conformational transition into an open form would be prerequisite for ubiquitin ligation on the Lys48-linked chains composed of even number of ubiquitin units. On the basis of the present structural data, we suggest that the flexible nature of the isopeptide linkage thus observed contributes to the structural arrangements of ubiquitin chains exemplified by the pH-dependent closed-to-open conformational transition of tetraubiquitin [14,15,20,21,31].
T. Satoh et al. / Biochemical and Biophysical Research Communications 400 (2010) 329–333
In general, the formation and elongation of polyubiquitin chains are performed by cooperation of E2 and E3 enzymes [1]. In contrast, E2-25 K is one of the E2 enzymes that has been demonstrated to form polyubiquitin chains in the absence of E3 ubiquitin ligase in vitro [29]. E2-25 K is a unique E2 enzyme that contains not only an E2 catalytic domain but also a UBA domain at its C-terminus [19,29]. The UBA domain is shown to be involved in polyubiquitin chain formation [35] and have inhibitory function in ubiquitin transfer from E1 [34]. The UBA domain is supposed to be involved in cyclization of polyubiquitin by connecting proximal ubiquitin to the distal end of ubiquitin chains. Possibly the unique UBA domain interacts with the Ile44 surface of one ubiquitin unit in longer Lys48-linked polyubiquitin chains unit and thereby cause their conformational changes to facilitate polyubiquitin chain elongation and cyclization in the absence of E3 ligase. 5. Concluding remarks In this study, we solved the crystal structure of cyclic Lys48linked tetraubiquitin at 1.85 Å resolution, which revealed the flexible nature of the wild-type isopeptide bond between the second and the third ubiquitins. It has been suggested that if cyclic polyubiquitins form in cells, they probably do not accumulate as dead-end products [30]. However, potential endoisopeptidase activity of DUBs remains largely unknown to date. Structural information of our cyclic Lys48-linked tetraubiquitin can be utilized for further investigation of the potential endoisopeptidase activity of DUBs and dynamics of Lys48-linked polyubiquitin chains. Acknowledgments cDNA encoding E2-25 K was kindly provided from Dr. Keiji Tanaka (The Tokyo Metropolitan Institute, Tokyo, Japan). We thank Mr. Takashi Hirano (Nagoya City University, Nagoya, Japan) for helpful comments on the manuscript. This work was supported by Grant-in-Aid for Scientific Research on Innovative Areas (20107004), Grant-in-Aid for Scientific Research (B) (21370050), and Targeted Proteins Research Program from the Ministry of Education, Culture, Sports, Science and Technology. Appendix A. Supplementary material Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.bbrc.2010.08.057. References [1] A. Hershko, A. Ciechanover, The ubiquitin system, Annu. Rev. Biochem. 67 (1998) 425–479. [2] C.M. Pickart, M.J. Eddins, Ubiquitin: structures, functions, mechanisms, Biochim. Biophys. Acta 1695 (2004) 55–72. [3] H.T. Kim, K.P. Kim, F. Lledias, A.F. Kisselev, K.M. Scaglione, D. Skowyra, S.P. Gygi, A.L. Goldberg, Certain pairs of ubiquitin-conjugating enzymes (E2s) and ubiquitin-protein ligases (E3s) synthesize nondegradable forked ubiquitin chains containing all possible isopeptide linkages, J. Biol. Chem. 282 (2007) 17375–17386. [4] J. Peng, D. Schwartz, J.E. Elias, C.C. Thoreen, D. Cheng, G. Marsischky, J. Roelofs, D. Finley, S.P. Gygi, A proteomics approach to understanding protein ubiquitination, Nat. Biotechnol. 21 (2003) 921–926. [5] F. Ikeda, I. Dikic, Atypical ubiquitin chains: new molecular signals. Protein modifications: beyond the usual suspects review series, EMBO Rep. 9 (2008) 536–542. [6] O.V. Baboshina, A.L. Haas, Novel multiubiquitin chain linkages catalyzed by the conjugating enzymes E2EPF and RAD6 are recognized by 26 S proteasome subunit 5, J. Biol. Chem. 271 (1996) 2823–2831.
333
[7] T. Kirisako, K. Kamei, S. Murata, M. Kato, H. Fukumoto, M. Kanie, S. Sano, F. Tokunaga, K. Tanaka, K. Iwai, A ubiquitin ligase complex assembles linear polyubiquitin chains, EMBO J. 25 (2006) 4877–4887. [8] R. Varadan, M. Assfalg, A. Haririnia, S. Raasi, C. Pickart, D. Fushman, Solution conformation of Lys63-linked di-ubiquitin chain provides clues to functional diversity of polyubiquitin signaling, J. Biol. Chem. 279 (2004) 7055–7063. [9] D. Komander, F. Reyes-Turcu, J.D. Licchesi, P. Odenwaelder, K.D. Wilkinson, D. Barford, Molecular discrimination of structurally equivalent Lys 63-linked and linear polyubiquitin chains, EMBO Rep. 10 (2009) 466–473. [10] A.B. Datta, G.L. Hura, C. Wolberger, The structure and conformation of Lys 63linked tetraubiquitin, J. Mol. Biol. 392 (2009) 1117–1124. [11] S. Rahighi, F. Ikeda, M. Kawasaki, M. Akutsu, N. Suzuki, R. Kato, T. Kensche, T. Uejima, S. Bloor, D. Komander, F. Randow, S. Wakatsuki, I. Dikic, Specific recognition of linear ubiquitin chains by NEMO is important for NF-kappaB activation, Cell 136 (2009) 1098–1109. [12] A. Bremm, S.M. Freund, D. Komander, Lys11-linked ubiquitin chains adopt compact conformations and are preferentially hydrolyzed by the deubiquitinase Cezanne, Nat. Struct. Mol. Biol. 17 (2010) 939–947. [13] W.J. Cook, L.C. Jeffrey, M. Carson, Z. Chen, C.M. Pickart, Structure of a diubiquitin conjugate and a model for interaction with ubiquitin conjugating enzyme (E2), J. Biol. Chem. 267 (1992) 16467–16471. [14] W.J. Cook, L.C. Jeffrey, E. Kasperek, C.M. Pickart, Structure of tetraubiquitin shows how multiubiquitin chains can be formed, J. Mol. Biol. 236 (1994) 601– 609. [15] R. Varadan, O. Walker, C. Pickart, D. Fushman, Structural properties of polyubiquitin chains in solution, J. Mol. Biol. 324 (2002) 637–647. [16] R. Varadan, M. Assfalg, S. Raasi, C. Pickart, D. Fushman, Structural determinants for selective recognition of a Lys48-linked polyubiquitin chain by a UBA domain, Mol. Cell 18 (2005) 687–698. [17] N. Zhang, Q. Wang, A. Ehlinger, L. Randles, J.W. Lary, Y. Kang, A. Haririnia, A.J. Storaska, J.L. Cole, D. Fushman, K.J. Walters, Structure of the s5a:k48-linked diubiquitin complex and its interactions with rpn13, Mol. Cell 35 (2009) 280– 290. [18] R.S. Kang, C.M. Daniels, S.A. Francis, S.C. Shih, W.J. Salerno, L. Hicke, I. Radhakrishnan, Solution structure of a CUE-ubiquitin complex reveals a conserved mode of ubiquitin binding, Cell 113 (2003) 621–630. [19] S. Raasi, R. Varadan, D. Fushman, C.M. Pickart, Diverse polyubiquitin interaction properties of ubiquitin-associated domains, Nat. Struct. Mol. Biol. 12 (2005) 708–714. [20] M.J. Eddins, R. Varadan, D. Fushman, C.M. Pickart, C. Wolberger, Crystal structure and solution NMR studies of Lys48-linked tetraubiquitin at neutral pH, J. Mol. Biol. 367 (2007) 204–211. [21] C.L. Phillips, J. Thrower, C.M. Pickart, C.P. Hill, Structure of a new crystal form of tetraubiquitin, Acta Crystallogr. D Biol. Crystallogr. 57 (2001) 341–344. [22] E. Sakata, T. Satoh, S. Yamamoto, Y. Yamaguchi, M. Yagi-Utsumi, E. Kurimoto, K. Tanaka, S. Wakatsuki, K. Kato, Crystal structure of UbcH5b–ubiquitin intermediate: insight into the formation of the self-assembled E2–Ub conjugates, Structure 18 (2010) 138–147. [23] Z. Otwinowski, W. Minor, Processing of X-ray diffraction data collected in oscillation mode, Methods Enzymol. 276 (1997) 307–326. [24] A. Vagin, A. Teplyakov, MOLREP: an automated program for molecular replacement, J. Appl. Crystallogr. 30 (1997) 1022–1025. [25] G.N. Murshudov, A.A. Vagin, E.J. Dodson, Refinement of macromolecular structures by the maximum-likelihood method, Acta Crystallogr. D Biol. Crystallogr. 53 (1997) 240–255. [26] P. Emsley, K. Cowtan, Coot: model-building tools for molecular graphics, Acta Crystallogr. D Biol. Crystallogr. 60 (2004) 2126–2132. [27] S.C. Lovell, I.W. Davis, W.B. Arendall 3rd, P.I. de Bakker, J.M. Word, M.G. Prisant, J.S. Richardson, D.C. Richardson, Structure validation by Calpha geometry: /, w and Cb deviation, Proteins 50 (2003) 437–450. [28] W. DeLano, Polym. Mol. Graph. Syst. (2002). [29] Z. Chen, C.M. Pickart, A 25-kilodalton ubiquitin carrier protein (E2) catalyzes multi-ubiquitin chain synthesis via lysine 48 of ubiquitin, J. Biol. Chem. 265 (1990) 21835–21842. [30] T. Yao, R.E. Cohen, Cyclization of polyubiquitin by the E2-25K ubiquitin conjugating enzyme, J. Biol. Chem. 275 (2000) 36862–36868. [31] Y. Ryabov, D. Fushman, Interdomain mobility in di-ubiquitin revealed by NMR, Proteins 63 (2006) 787–796. [32] I. Dikic, S. Wakatsuki, K.J. Walters, Ubiquitin-binding domains – from structures to functions, Nat. Rev. Mol. Cell Biol. 10 (2009) 659–671. [33] F.E. Reyes-Turcu, J.R. Shanks, D. Komander, K.D. Wilkinson, Recognition of polyubiquitin isoforms by the multiple ubiquitin binding modules of isopeptidase T, J. Biol. Chem. 283 (2008) 19581–19592. [34] A. Pichler, P. Knipscheer, E. Oberhofer, W.J. van Dijk, R. Korner, J.V. Olsen, S. Jentsch, F. Melchior, T.K. Sixma, SUMO modification of the ubiquitinconjugating enzyme E2-25K, Nat. Struct. Mol. Biol. 12 (2005) 264–269. [35] M.T. Haldeman, G. Xia, E.M. Kasperek, C.M. Pickart, Structure and function of ubiquitin conjugating enzyme E2-25K: the tail is a core-dependent activity element, Biochemistry 36 (1997) 10526–10537.