- Email: [email protected]
Crystal Structure and Solution NMR Studies of Lys48-linked Tetraubiquitin at Neutral pH
doi:10.1016/j.jmb.2006.12.065
J. Mol. Biol. (2007) 367, 204–211
Crystal Structure and Solution NMR Studies of Lys48-linked Tetraubiquitin at Neutral pH Michael J. Eddins 1 , Ranjani Varadan 2 , David Fushman 2 Cecile M. Pickart 3 † and Cynthia Wolberger 1 ⁎ 1
Department of Biophysics and Biophysical Chemistry and the Howard Hughes Medical Institute, Johns Hopkins University School of Medicine, Baltimore, MD 21205, USA 2
Department of Chemistry and Biochemistry and Center for Biomolecular Structure and Organization, University of Maryland, College Park, MD 20742, USA 3
Department of Biochemistry and Molecular Biology, Bloomberg School of Public Health, Johns Hopkins University, Baltimore, MD 21205, USA *Corresponding author
Ubiquitin modification of proteins is used as a signal in many cellular processes. Lysine side-chains can be modified by a single ubiquitin or by a polyubiquitin chain, which is defined by an isopeptide bond between the C terminus of one ubiquitin and a specific lysine in a neighboring ubiquitin. Polyubiquitin conformations that result from different lysine linkages presumably differentiate their roles and ability to bind specific targets and enzymes. However, conflicting results have been obtained regarding the precise conformation of Lys48-linked tetraubiquitin. We report the crystal structure of Lys48-linked tetraubiquitin at near-neutral pH. The two tetraubiquitin complexes in the asymmetric unit show the complete connectivity of the chain and the molecular details of the interactions. This tetraubiquitin conformation is consistent with our NMR data as well as with previous studies of diubiquitin and tetraubiquitin in solution at neutral pH. The structure provides a basis for understanding Lys48-linked polyubiquitin recognition under physiological conditions. © 2007 Elsevier Ltd. All rights reserved.
Keywords: ubiquitin; polyubiquitin chains; tetraubiquitin; Lys48-linked; crystal structure
Introduction The covalent attachment of ubiquitin to substrate proteins helps to regulate numerous cellular pathways.1–3 Specific lysine residues on the substrate can be modified with a single ubiquitin (monoubiquitination) or with a polyubiquitin chain. Ubiquitin conjugation involves a series of enzymatic steps that end with the formation of an isopeptide bond between the C-terminal carboxylate of ubiquitin and the ε-amine of the substrate lysine. The 76 residue ubiquitin protein itself contains seven surface lysine residues that can each be modified by ubiquitin conjugation via Lys–Gly76 isopeptide bonds, leading to the formation of linear polyubi-
† Deceased. Abbreviations used: UIM, ubiquitin-interacting motif; UBA, ubiquitin-associated domain. E-mail address of the corresponding author: [email protected]
quitin chains.3 There are different types of polyubiquitin chains, each distinguished by the particular Lys residue to which each successive ubiquitin is conjugated. Although all seven surface lysine residues of ubiquitin can be used in the formation of polyubiquitin chains, Lys63 and Lys48-linked chains are more frequently observed and are the best characterized. Lys63-linked polyubiquitin chains play a role in DNA damage tolerance and in the inflammatory response through non-degradative signaling pathways.4 Lys48-linked polyubiquitin chains target their substrates to the 26 S proteasome for degradation,5 and the minimal chain needed for efficient proteasomal recognition and substrate degradation has been shown to consist of four ubiquitin molecules.6 The different types of polyubiquitin chains must have distinct structural characteristics that target their substrates to specific cellular pathways. A variety of protein domains called ubiquitinbinding domains recognize and bind ubiquitin, including UIMs (ubiquitin-interacting motifs), UBAs (ubiquitin-associated domains), and others.7
0022-2836/$ - see front matter © 2007 Elsevier Ltd. All rights reserved.
205
Lys48-linked Tetraubiquitin at Neutral pH
Some of these ubiquitin-binding domains bind monoubiquitin, while others recognize polyubiquitin chains.8 Ubiquitin-binding domains that recognize monoubiquitin all bind to a hydrophobic patch in ubiquitin, burying it at the complex interface.9,10 This hydrophobic patch is centered around three residues on the ubiquitin surface: Leu8, Ile44, and Val70.7 In the case of polyubiquitin binding, however, the way in which ubiquitinbinding domains distinguish among different chain linkages is not understood. In particular, there are UBA domains that are able to distinguish among the different types of polyubiquitin chains and bind preferentially to chains with a specific linkage.8 The specific topology of the polyubiquitin chain, or the surfaces of the polyubiquitin chain that are exposed upon formation of a particular lysine linkage, are possible structural elements that could be recognized.9,10 There are three crystal structures of Lys48-linked polyubiquitin that have been reported. The first crystal structure reported was of diubiquitin (1AAR),11 which adopts a conformation that buries the ubiquitins' hydrophobic patches at the interface. This arrangement, referred to as the closed conformation, obstructs the binding surface of ubiquitin that interacts with ubiquitin-binding domains.7 In the other conformations of polyubiquitin, referred to as open conformations, the hydrophobic patches of the ubiquitins are exposed and available for binding. The other two previously reported Lys48-linked polyubiquitin crystal structures are of tetraubiquitin (1TBE,12 1F9J13), and differ in overall organization and topology. In one case, the chain is in an open conformation, with the hydrophobic patch of each ubiquitin exposed instead of buried at an interface,12 while the other tetraubiquitin structure has a configuration that buries the hydrophobic patch of each ubiquitin.13 In addition, these tetraubiquitin structures differ from the diubiquitin crystal structure in the way in which the proximal ubiquitin (the first ubiquitin in the polyubiquitin chain that is conjugated to a substrate lysine residue, and which we denote as 1) and the next ubiquitin (which we denote as 2) are oriented. Different conformations are observed in each structure for the covalent linkages, which consist of the C-terminal ubiquitin residue, Gly76 (on 2), and the covalently linked Lys48 side-chain (on 1). A similar difference in the orientation and the linkage is observed for the other two ubiquitins, 3 and 4. Another complication affecting the interpretation of the previous tetraubiquitin crystal structures is that both were determined from crystals containing only half the tetraubiquitin in the asymmetric unit, making the linkage between ubiquitins 2 and 3 ambiguous. In one case (1F9J), this led to two possible conformations for tetraubiquitin.13 Distinguishing among the different possible conformations is important to understanding how Lys48-linked polyubiquitin chains are recognized in the cell. NMR studies of Lys48-linked diubiquitin suggest that its conformation is dynamic and pH sensitive.14
At neutral pH, diubiquitin adopts a predominantly closed conformation similar to that seen in the diubiquitin crystal structure,14,15 while at more acidic pH the diubiquitin takes on a more open conformation that exposes the hydrophobic patches. 14,16 A similar dependence on pH was observed for tetraubiquitin.14 Solution studies of Lys48-linked tetraubiquitin at neutral pH have yielded useful insights into tetraubiquitin topology, but ambiguities regarding pair-wise contacts between ubiquitins remain, and the overall chain topology is still not understood.14,17 In addition, the tetraubiquitin crystal structures that were crystallized at acidic pH do not account for the solution study data. This leaves the overall topology and inter-ubiquitin contacts for Lys48-linked tetraubiquitin at physiological pH still in question. We report here a new crystal structure of Lys48linked tetraubiquitin that differs in several respects from previously reported structures. The crystals were grown at near-neutral pH (6.7) and contain two complete tetraubiquitin chains in the asymmetric unit. The tetraubiquitin is in a conformation that differs from that in previously reported structures and clearly shows all the linkages between the ubiquitins, which were not seen in the previous structures. The overall conformation is a dimer of diubiquitins, consistent with results from solution studies at neutral pH. We present additional, site-directed spin-labeling studies that further support the closed compact structure seen in the crystal. This new structure provides molecular details on the relative orientation, and specific contacts of the ubiquitin monomers in the chain, lending insight into the overall shape and surface of Lys48-linked tetraubiquitin that may be used in its roles within the cell.
Results and Discussion The crystal structure of Lys48-linked tetraubiquitin at near-neutral pH Crystals of Lys48-linked tetraubiquitin formed at pH 6.7 in space group C2, and contain two tetraubiquitin chains in the asymmetric unit. Diffraction data were collected to 2.2 Å resolution and the structure was determined by molecular replacement using diubiquitin as a search model,11 with residues 74–76 omitted from both the proximal (1) and distal ubiquitins. Data collection and refinement statistics are summarized in Table 1. The two tetraubiquitin chains in the asymmetric unit are virtually identical, superimposing with a root-mean-square-deviation (r.m.s.d.) of 0.4 Å in Cα positions. The tetraubiquitin chain adopts a conformation consisting of a dimer of diubiquitins, with each diubiquitin subunit adopting the closed conformation (Figure 1). Starting from the proximal ubiquitin (1), the first diubiquitin subunit consists of ubiquitins 1 and 2, and the second diubiquitin subunit is made up of ubiquitins
206
Lys48-linked Tetraubiquitin at Neutral pH
Table 1. Data collection and refinement statistics K48-linked Ub4 A. Data collection Space group Cell dimensions a, b, c (Å) β (°) Resolution (Å) Rsym (%) I/σI Completeness (%) Redundancy B. Refinement Resolution (Å) No. reflections Rwork/Rfree No. atoms Protein Water Sulfate Mes B-factor Protein Water Sulfate Mes r.m.s. deviations Bond lengths (Å) Bond angles (°)
C2 59.10, 77.08, 139.36 90.32 50–2.2 (2.28–2.20) 10.0 (55.6) 22.5 (2.8) 93.4 (93.7) 3.2 (3.4) 2.2 31,476 22.2/26.2 4812 149 4 2 56.1 53.2 80.6 90.6 0.007 1.46
The highest-resolution shell is shown in parentheses.
3 and 4; 4 is the distal ubiquitin in the chain (Figure 1). The linkage between the two diubiquitin subunits is therefore between ubiquitins 2 and 3. Each diubiquitin subunit (1–2 and 3–4) is nearly identical in structure, superimposing with a r.m.s.d. value of 0.6 Å between the two subunits. Both are similar to the diubiquitin crystal structure (1AAR11), with an r.m.s.d. value of 0.6 Å for the 1–2 subunit, and 0.9 Å for the 3–4 subunit. The hydrophobic patch of each ubiquitin, comprising residues Leu8, Ile44, and Val70, is located at the diubiquitin interface. This inter-subunit interface has a total buried surface area of 1458 Å2 between ubiquitins 1 and 2 of the first diubiquitin subunit, and likewise between ubiquitins 3 and 4 of the second diubiquitin subunit. The
interface between the two diubiquitin subunits (1–2 and 3–4) is centered between ubiquitins 1 and 3 and has a total buried surface area of 1455 Å2 (Figure 1). The overall conformation of tetraubiquitin in the present structure is different from that seen in the two previous tetraubiquitin crystal structures (1TBE, and 1F9J).12,13 In particular, the Lys48–Gly76 linkage adopts a different conformation in each of the three structures (Figure 2), which places the ubiquitin hydrophobic patch residues, Leu8, Ile44, Val70, in different orientations within tetraubiquitin (Figure 3). Importantly, the two previous tetraubiquitin crystal structures contain only half a tetraubiquitin in the asymmetric unit, and the isopeptide bond between ubiquitins 2 and 3 was therefore not resolved in the electron density maps. In one case (1F9J), the overall linkage was ambiguous and two possible conformations were proposed. In the present structure, by contrast, all three linkages joining the C terminus of one ubiquitin to Lys48 of the next are well resolved in the electron density map. In addition, both previous tetraubiquitin crystal structures were crystallized at acidic pH, one at pH 5.0 (1TBE), and the other at pH 4.8 (1F9J), while the structure of tetraubiquitin reported here was determined from crystals grown at pH 6.7. As discussed below, the acidic pH at which the previous structures were determined may have favored a different conformation of tetraubiquitin. The conformation of tetraubiquitin in solution The present Lys48-linked tetraubiquitin crystal structure is supported by data from solution studies. NMR chemical shift perturbation data show that the conformation of diubiquitin is dynamic and pH dependent; at pH 4.5 the open conformation is fully populated, and at pH 7.5 the closed conformation is almost fully populated, with tetraubiquitin exhibiting a similar pH-dependent behavior.14 At pH 6.8, it is estimated that less than 15% of the diubiquitin chains are in the open conformation,14,16 with the inter-conversion between the closed and open conformations occurring on a 10 ns time scale. The chemical shift perturbation data indicate that
Figure 1. A new conformation of Lys48-linked tetraubiquitin. The tetraubiquitin coloring from proximal ubiquitin to distal ubiquitin is: yellow–cyan–green–blue (1–2–3–4). Rotating 45° in either direction about the x-axis from the center structure shows the interface of the two diubiquitin subunits (1–2 on the left, 3–4 on the right).
Lys48-linked Tetraubiquitin at Neutral pH
Figure 2. Overlays of the proximal ubiquitins 1–2 from the three different tetraubiquitin crystal structures. The alignment is done with ubiquitin 1; ubiquitin 2 appears in different positions due to the different Lys48–Gly76 linkages seen in the various crystal structures. Ubiquitins 1–2 from the present structure are shown in purple (which superimposes with the diubiquitin crystal structure 1AAR, not shown). Ubiquitins from 1TBE are colored orange, and ubiquitins from 1F9J are colored green.
ubiquitins 3 and 4 in tetraubiquitin are essentially identical to those in diubiquitin, thus suggesting that these two distal ubiquitins in the chain pack together in a manner similar to that seen in the closed form of diubiquitin.14 Moreover, additional NMR data indicate that the patterns of chemical shift perturbations in the proximal ubiquitin (1) in tetraubiquitin and in the proximal ubiquitin in the closed form of diubiquitin are nearly identical.17
207 Similar observations were made for ubiquitin 3 of tetraubiquitin (in agreement with Varadan et al.14), suggesting that all of these ubiquitins (1 and 3 of tetraubiquitin, and 1 of diubiquitin) are equivalent and form closed interfaces with another ubiquitin.17 This would suggest that ubiquitins 1 and 3 of tetraubiquitin are involved in closed diubiquitin interfaces with the other two ubiquitins in the chain (2 and 4). The tetraubiquitin structure reported here shows that ubiquitins 1 and 3 form closed diubiquitin interfaces with ubiquitins 2 and 4, respectively, with the interfaces between each ubiquitin moiety in good agreement with the chemical shift perturbations observed in solution14 (Figure 4). Site-directed spin labeling studies support the relevance of the present structure to tetraubiquitin's conformation in solution (Figure 5). A paramagnetic spin label, (1-oxyl-2,2,5,5-tetramethyl-3-pyrroline-3methyl)methanesulfonate (MTSL), was attached to Cys48 in the distal ubiquitin (4) in order to map residues in the other ubiquitins that are close in space to the spin label by using the distance dependence of the paramagnetic relaxation enhancement effect.18 Consistent with the present tetraubiquitin structure, strong signal attenuations were observed in all ubiquitin units, most notably in the proximal ubiquitin (1) (Supplementary Data, Figure 1). As seen in Figure 5, signal attenuations observed in ubiquitins 1–3 are consistent with the proximity of the affected sites to residue 48 in ubiquitin 4, where the spin label was attached. We note that, although diubiquitin (and therefore tetraubiquitin) in solution are in dynamic equilibrium between the closed and open conformations, the spin-labeling data primarily reflect the domain positioning in the compact, closed conformation. This is due to the sharp decay (∝ R− 6) of the paramagnetic relaxation effect as a function of the distance R from the unpaired electron of the spin label,19 such that the main contribution to the observed residue-specific signal attenuations comes from the predominant closed
Figure 3. Comparison of the three tetraubiquitin crystal structures. Coloring is the same as for Figure 1, with the proximal ubiquitin yellow and the distal ubiquitin blue. Hydrophobic patch residues, Leu8, Ile44, and Val70 are shown as red sticks. The structure on the left is the present structure (2O6V), with the hydrophobic patches buried as seen in the structure of diubiquitin (1AAR). The pH 5.0 structure in the center (1TBE) has all the hydrophobic patches exposed. The pH 4.8 structure on the right (1F9J) has the hydrophobic patches buried in a different arrangement than in the pH 6.7 structure (2O6V).
208
Lys48-linked Tetraubiquitin at Neutral pH
Figure 4. The tetraubiquitin conformation in the crystal structure is consistent with the interaction surfaces between ubiquitins mapped using NMR-detected chemical shift perturbations in solution at pH 6.8 as detailed.14 The orientation of tetraubiquitin is the same as for Figure 1; residues colored red show chemical shift perturbations (versus monoubiquitin) of >0.05 ppm. Perturbations in the linkage-involved residues, Lys48 (in 1 and 3), and 74–76 (in 2 and 4) are not colored.
conformation. The contribution from the less populated, open conformations is practically negligible, as shown for diubiquitin.16 Taken together, these NMR data indicate that the crystal structure of tetraubiquitin reported here reflects the Lys48linked tetraubiquitin conformation in solution at physiological pH. The compact tetraubiquitin conformation is also in agreement with NMR relaxation data in solution, which are sensitive to molecular tumbling, and hence to the size and shape of the molecule. The overall rotational correlation time of 16.6 ns (at 25 °C), predicted based on this structure using the HYDRONMR program,20 is a factor of 2 higher than the measured value for diubiquitin14 (∼ 8 ns) and
Figure 5. Validation of the tetraubiquitin structure by site-directed spin labeling data in solution. Sites in ubiquitins 1–3 with significant signal attenuation in the presence of spin label are colored purple for attenuation >20% and red for attenuation > 40%. Also shown is the side-chain of residue 48 in the distal ubiquitin (Arg48 in the present structure was replaced with a cysteine for spin labeling).
fourfold higher than that for monoubiqiuitin (4.3 ns). The experimental 1H T2 values, 13, 26, and 50 ms, measured at these conditions for backbone amides in tetra-, di-, and monoubiquitin, respectively,14 almost exactly reproduce the inverse proportionality dependence between the T2 and the rotational correlation time, expected for transverse spin relaxation in the case of protein tumbling in solution.21 There is also good agreement between the average experimental 15N T1 and T2 values for backbone amides in tetraubiquitin reported, 17 774(± 119) ms and 64.7(± 8.0) ms, respectively, and their predicted values, 862 ms of 59.5 ms, at the experimental conditions of the study (30 °C and 15N Larmor frequency of 50.7 MHz). The NMR data therefore suggest that the crystal structure of tetraubiquitin determined here indeed represents the predominant conformation of the chain in solution at near-physiological pH. In this conformation, the hydrophobic patches on ubiquitins are buried at the interface in each diubiquitin subunit. It should be emphasized, however, that the chain is not rigidly locked in this conformation in solution. This conclusion is suggested by the observation that the conformation of an isolated diubiquitin subunit is dynamic, with a relatively small (∼ 1 kcal/mol) difference in the free energy between the closed and open states.16 In addition, recent studies showed that some UBA domains10,22 as well as smaller molecules14,23 bind to diubiquitin directly at the inter-ubiquitin interface. The interconversion between the closed and open conformations of diubiquitin, and likely of longer chains, is essential for allowing access to the hydrophobic patch on ubiquitin monomers, which are targeted by almost all ubiquitin-binding domains. Locking the chain in the closed conformation could therefore block binding to this patch. It has been shown that cyclization of Lys48-linked diubiquitin by a crosslinker that restricted interface opening strongly impedes the ability of this chain to bind a UBA domain in a high-affinity sandwich-like mode.24 All this suggests that other states of the chain corresponding to more open conformations must exist
Lys48-linked Tetraubiquitin at Neutral pH
and be accessible to tetraubiquitin in solution. At neutral pH, however, these states are less populated than the conformation determined here. The observed conformation of polyubiquitin is dependent upon the type of isopeptide linkage that joins the individual ubiquitin molecules. The topology of the polyubiquitin chain and the interaction between individual monomers seen in the present Lys48-linked tetraubiquitin structure could not be adopted by a Lys63-linked chain. An isopeptide linkage between Lys63 and the neighboring ubiquitin would force a proximity of Lys63 with the C terminus of the neighboring ubiquitin that would not allow Lys63-linked tetraubiquitin to adopt the conformation seen in the present structure. Similarly, other types of tetraubiquitin chains containing different lysine linkages are unlikely to be able to adopt the precise conformation observed here. Polyubiquitin linkage type may therefore give rise to different recognition signals because of the constraints it places on the polyubiquitin conformation. Recognition surface of Lys48-linked tetraubiquitin The tetraubituitin structure described here suggests how the unique topology of Lys48-linked tetraubiquitin allows a key ubiquitin residue, Leu8, to play a central role in mediating the binding of Lys48-linked polyubiquitin to ubiquitin-binding domains. The hydrophobic patch on ubiquitin has been shown to be important for ubiquitin function.6,25–27 The residues comprising this patch on the surface of ubiquitin, Leu8, Ile44, and Val70, mediate interactions between mono- and polyubiquitin and numerous ubiquitin-binding domains.7 These resi-
209 dues have also been shown to be an important recognition element for the binding of Lys48-linked tetraubiquitin to the 26S proteasome.26,27 In particular, Leu8 is an important part of this recognition element, with the L8A mutation in the chain causing the largest defect in substrate degradation as compared to the I44A and V70A mutations.26,27 Because of this, a number of studies have examined the effects of Leu8 mutations in the context of Lys48linked polyubiquitin chains.6,25–27 Binding studies with S5a (which contain two UIM domains) or HHR23A (which contain two UBA domains) indicate that these two classes of ubiquitin-binding proteins behave differently in response to the position and number of L8A mutations in tetraubiquitin. In the case of S5a, only a subset of the Leu8 side-chains are needed for binding, 26,27 whereas in the case of HHR23A, all the Leu8 sidechains appear to play a role in the interaction.25 Mutation of two of the four Leu8 residues in tetraubiquitin is tolerated by S5a if the mutations occur only in ubiquitins 2 and 4; however, for HHR23A, substitutions of any two Leu8 residues have a significant effect on HHR23A binding. In the present tetraubiquitin crystal structure, the ubiquitin hydrophobic patches are buried at the diubiquitin intersubunit interface. However, Leu8 of each ubiquitin lies at the edge of the interface and is oriented in a way that gives rise to a hydrophobic stripe on the surface at the middle of each diubiquitin subunit (Figure 6). In the context of the tetraubiquitin structure, these hydrophobic surface stripes are related to each other by a 90° rotation. The presence of these two Leu8 hydrophobic stripes that are unique to this particular Lys48-linked tetraubiquitin structure may provide a potential recognition and interaction surface. It is also
Figure 6. The hydrophobic surface stripes on the Lys48-linked tetraubiquitin surface showing the orientation of all the ubiquitin Leu8 side-chains. The tetraubiquitin coloring is the same as for Figure 1, with the proximal ubiquitin 1 yellow, and the distal ubiquitin 4 being blue. The Leu8 side-chains are shown in red with the numbers corresponding to the specific ubiquitin in the chain a particular Leu8 is from.
210
Lys48-linked Tetraubiquitin at Neutral pH
possible that Leu8 plays an indirect role in binding to partner proteins by stabilizing the closed conformation of tetraubiquitin seen in the present crystal structure.10 In this case, contacts mediated by Leu8 could control access to the hydrophobic patches of the individual ubiquitins by regulating the opening and closing of the inter-ubiquitin interface in the diubiquitin subunits. In either case, the structure and conformation of Lys48-linked tetraubiquitin presented here is likely an important feature in its signaling role.
side-chains were adjusted using the program O31 and refined with CNS.32 The final model contains two tetraubiquitin complexes in the asymmetric unit, and includes residues 1–75 of ubiquitins A and E, residues 1–76 of ubiquitins B, C, D, F, and G, and residues 2–76 of ubiquitin H. Data collection and refinement statistics are summarized in Table 1.
Materials and Methods
NMR studies
Proteins and Lys48-linked tetraubiquitin chain formation Expression and purification of ubiquitin has been described.28 Synthesis of Lys48-linked tetraubiquitin followed published protocols,28 with the following substitutions. Two forms of Lys48-linked diubiquitin were synthesized with the distal ubiquitin carrying different mutations. The Lys48-linked diubiquitin used in the proximal end (1–2) of tetraubiquitin contained K48C, K63R mutations at ubiquitin 2. The Lys48-linked diubiquitin used in the distal end (3–4) of tetraubiquitin contained K48R, K63R mutations at ubiquitin 4, removing the need to use the K48C mutation to block K48 on ubiquitin 4 during diubiquitin formation. Tetraubiquitin constructs for NMR studies were synthesized using a segmental isotope-labeling strategy detailed elsewhere,14 such that a specific single ubiquitin unit per chain was 15N-enriched. All of these tetraubiquitin chains contained a K48C mutation on the distal ubiquitin (4), in order to provide a single site for spin-label attachment. The proximal ubiqutin (1) contained an additional Cterminal residue Asp77; no K63R mutation was used in ubiquitin 2. Crystallization and data collection Tetraubiquitin was concentrated to 8 mg ml− 1 in 10 mM Tris (pH 7.6), 10 mM NaCl, 0.1 mM EDTA, and 1 mM dithiothreitol. Crystals were grown by the method of hanging drop vapor diffusion at 20 °C, mixing equal volumes of the protein complex with a well solution containing 2 M (NH4)2SO4, 4%(w/v) PEG 400, and 100 mM Mes (pH 6.5). Crystals grew overnight. The measured pH of both the well solution and the drop after equilibration was pH 6.7. Crystals formed in space group C2 with unit cell dimensions a = 59.10 Å, b = 77.08 Å, c = 139.4 Å, and β = 90.32°, and contained two tetraubiquitin complexes in the asymmetric unit. Diffraction data were collected at APS on BioCARS beamline 14-BM-D at a wavelength of 0.9786 Å and processed using HKL2000.29 Structure determination and refinement The structure of tetraubiquitin was determined by molecular replacement with the program MOLREP30 using diubiquitin (1AAR)11 C-terminally truncated to residue 73 of both proximal and distal ubiquitins as the search model. Missing residues, including the C-terminal tail of ubiquitin and the covalent linkage, were built and
Protein Data Bank accession codes The atomic coordinates have been deposited with the RSCB Protein Data Bank, accession code 2O6V.
The NMR studies were performed at 25 °C on a Bruker Avance 600 spectrometer with the 1H Larmor frequency of 600 MHz. NMR samples were prepared in 20 mM phosphate buffer (pH 6.8) containing 7% (v/v)2H2O and 0.02% (w/v) NaN3. The tetraubiquitin concentration was 0.5–0.7 mM. The paramagnetic spin label (1-oxyl-2,2,5,5tetramethyl-3-pyrroline-3-methyl)methanesulfonate (MTSL) was covalently attached to a single cysteine site (Cys48) on the distal ubiquitin (4) as detailed.18 The paramagnetic relaxation enhancement due to close proximity to the spin label was monitored by signal attenuation observed in the 1H–15N TROSY spectra of the other three ubiquitins in the chain. The signal attenuation was quantified as the ratio of signal intensities in the spectra recorded with the spin label (in the oxidized, paramagnetic state) attached to the protein and in the absence of the spin label or after reducing it with a threefold molar excess of ascorbic acid.18
Acknowledgements This work was supported by the Howard Hughes Medical Institute (to C.W.) and by NIH grant GM065334 (to D.F.). Use of the Advanced Photon Source was supported by the US Department of Energy, Basic Energy Sciences, Office of Science, under contract no. W-31-109-Eng-38. Use of the BioCARS Sector 14 was supported by the National Institutes of Health, National Center for Research Resources, under grant number RR07707.
Supplementary Data Supplementary data associated with this article can be found, in the online version, at doi:10.1016/ j.jmb.2006.12.065
References 1. Sun, L. & Chen, Z. J. (2004). The novel functions of ubiquitination in signaling. Curr. Opin. Cell Biol. 16, 119–126. 2. Ulrich, H. D. (2002). Degradation or maintenance: actions of the ubiquitin system on eukaryotic chromatin. Eukaryotic Cell, 1, 1–10.
211
Lys48-linked Tetraubiquitin at Neutral pH 3. Pickart, C. M. & Fushman, D. (2004). Polyubiquitin chains: polymeric protein signals. Curr. Opin. Chem. Biol. 8, 610–616. 4. Chen, Z. J. (2005). Ubiquitin signalling in the NFkappaB pathway. Nature Cell Biol. 7, 758–765. 5. Pickart, C. M. & Cohen, R. E. (2004). Proteasomes and their kin: proteases in the machine age. Nature Rev. Mol. Cell Biol. 5, 177–187. 6. Thrower, J. S., Hoffman, L., Rechsteiner, M. & Pickart, C. M. (2000). Recognition of the polyubiquitin proteolytic signal. EMBO J. 19, 94–102. 7. Hicke, L., Schubert, H. L. & Hill, C. P. (2005). Ubiquitin-binding domains. Nature Rev. Mol. Cell Biol. 6, 610–621. 8. Raasi, S., Varadan, R., Fushman, D. & Pickart, C. M. (2005). Diverse polyubiquitin interaction properties of ubiquitin-associated domains. Nature Struct. Mol. Biol. 12, 708–714. 9. Varadan, R., Assfalg, M., Haririnia, A., Raasi, S., Pickart, C. & Fushman, D. (2004). Solution conformation of Lys63-linked di-ubiquitin chain provides clues to functional diversity of polyubiquitin signaling. J. Biol. Chem. 279, 7055–7063. 10. Varadan, R., Assfalg, M., Raasi, S., Pickart, C. & Fushman, D. (2005). Structural determinants for selective recognition of a Lys48-linked polyubiquitin chain by a UBA domain. Mol. Cell, 18, 687–698. 11. Cook, W. J., Jeffrey, L. C., Carson, M., Chen, Z. & Pickart, C. M. (1992). Structure of a diubiquitin conjugate and a model for interaction with ubiquitin conjugating enzyme (E2). J. Biol. Chem. 267, 16467–16471. 12. Cook, W. J., Jeffrey, L. C., Kasperek, E. & Pickart, C. M. (1994). Structure of tetraubiquitin shows how multiubiquitin chains can be formed. J. Mol. Biol. 236, 601–609. 13. Phillips, C. L., Thrower, J., Pickart, C. M. & Hill, C. P. (2001). Structure of a new crystal form of tetraubiquitin. Acta Crystallog. sect. D, 57, 341–344. 14. Varadan, R., Walker, O., Pickart, C. & Fushman, D. (2002). Structural properties of polyubiquitin chains in solution. J. Mol. Biol. 324, 637–647. 15. van Dijk, A. D., Fushman, D. & Bonvin, A. M. (2005). Various strategies of using residual dipolar couplings in NMR-driven protein docking: application to Lys48linked di-ubiquitin and validation against 15N-relaxation data. Proteins: Struct. Funct. Genet. 60, 367–381. 16. Ryabov, Y. & Fushman, D. (2006). Interdomain mobility in di-ubiquitin revealed by NMR. Proteins: Struct. Funct. Genet. 63, 787–796. 17. Tenno, T., Fujiwara, K., Tochio, H., Iwai, K., Morita, E. H., Hayashi, H. et al. (2004). Structural basis for distinct roles of Lys63- and Lys48-linked polyubiquitin chains. Genes Cells, 9, 865–875. 18. Varadan, R., Assfalg, M. & Fushman, D. (2005). Using NMR spectroscopy to monitor ubiquitin chain conformation and interactions with ubiquitin-binding
19. 20.
21. 22.
23.
24.
25.
26.
27.
28. 29. 30. 31.
32.
domains. In Ubiquitin and Protein Degradation, Methods in Enzymology (Deshaies, R. J., ed), vol. 399, part B, pp. 177–192, Wiley-VCH, Weinheim. Kosen, P. A. (1989). Spin labeling of proteins. Methods Enzymol. 177, 86–121. García de la Torre, J., Huertas, M. L. & Carrasco, B. (2000). HYDRONMR: prediction of NMR relaxation of globular proteins from atomic-level structures and hydrodynamic calculations. J. Magn. Reson. B147, 138–146. Wüthrich, K (1986). NMR of Proteins and Nucleic Acids. The George Fisher Baker non-resident lectureship in chemistry at Cornell University Wiley, New York. Trempe, J. F., Brown, N. R., Lowe, E. D., Gordon, C., Campbell, I. D., Noble, M. E. & Endicott, J. A. (2005). Mechanism of Lys48-linked polyubiquitin chain recognition by the Mud1 UBA domain. EMBO J. 24, 3178–3189. Verma, R., Peters, N. R., D'Onofrio, M., Tochtrop, G. P., Sakamoto, K. M., Varadan, R. et al. (2004). Ubistatins inhibit proteasome-dependent degradation by binding the ubiquitin chain. Science, 306, 117–120. Dickinson, B. Varadan, R. & Fushman, D. (2007). Effects of cyclization on conformational dynamics and binding properties of Lys48-linked di-ubiquitin. Protein Sci. In the press. Raasi, S., Orlov, I., Fleming, K. G. & Pickart, C. M. (2004). Binding of polyubiquitin chains to ubiquitinassociated (UBA) domains of HHR23A. J. Mol. Biol. 341, 1367–1379. Beal, R., Deveraux, Q., Xia, G., Rechsteiner, M. & Pickart, C. (1996). Surface hydrophobic residues of multiubiquitin chains essential for proteolytic targeting. Proc. Natl Acad. Sci. USA, 93, 861–866. Beal, R. E., Toscano-Cantaffa, D., Young, P., Rechsteiner, M. & Pickart, C. M. (1998). The hydrophobic effect contributes to polyubiquitin chain recognition. Biochemistry, 37, 2925–2934. Pickart, C. M. & Raasi, S. (2005). Controlled synthesis of polyubiquitin chains. Methods Enzymol. 399, 21–36. Otwinowski, Z. & Minor, W. (1997). Processing of X-ray diffraction data collected in oscillation mode. Methods Enzymol. 176, 307–326. Vagin, A. & Teplyakov, A. (1997). MOLREP: an automated program for molecular replacement. J. Appl. Crystallog. 30, 1022–1025. Jones, T. A., Zou, J. Y., Cowan, S. W. & Kjeldgaard, M. (1991). Improved methods for building protein models in electron density maps and the location of errors in these models. Acta Crystallog. sect. A, 47, 110–119. Brunger, A. T., Adams, P. D., Clore, G. M., DeLano, W. L., Gros, P., Grosse-Kunstleve, R. W. et al. (1998). Crystallography and NMR system: a new software suite for macromolecular structure determination. Acta Crystallog. sect. D, 54, 905–921.
Edited by P. Wright (Received 10 November 2006; received in revised form 22 December 2006; accepted 26 December 2006) Available online 29 December 2006
J. Mol. Biol. (2007) 367, 204–211
Crystal Structure and Solution NMR Studies of Lys48-linked Tetraubiquitin at Neutral pH Michael J. Eddins 1 , Ranjani Varadan 2 , David Fushman 2 Cecile M. Pickart 3 † and Cynthia Wolberger 1 ⁎ 1
Department of Biophysics and Biophysical Chemistry and the Howard Hughes Medical Institute, Johns Hopkins University School of Medicine, Baltimore, MD 21205, USA 2
Department of Chemistry and Biochemistry and Center for Biomolecular Structure and Organization, University of Maryland, College Park, MD 20742, USA 3
Department of Biochemistry and Molecular Biology, Bloomberg School of Public Health, Johns Hopkins University, Baltimore, MD 21205, USA *Corresponding author
Ubiquitin modification of proteins is used as a signal in many cellular processes. Lysine side-chains can be modified by a single ubiquitin or by a polyubiquitin chain, which is defined by an isopeptide bond between the C terminus of one ubiquitin and a specific lysine in a neighboring ubiquitin. Polyubiquitin conformations that result from different lysine linkages presumably differentiate their roles and ability to bind specific targets and enzymes. However, conflicting results have been obtained regarding the precise conformation of Lys48-linked tetraubiquitin. We report the crystal structure of Lys48-linked tetraubiquitin at near-neutral pH. The two tetraubiquitin complexes in the asymmetric unit show the complete connectivity of the chain and the molecular details of the interactions. This tetraubiquitin conformation is consistent with our NMR data as well as with previous studies of diubiquitin and tetraubiquitin in solution at neutral pH. The structure provides a basis for understanding Lys48-linked polyubiquitin recognition under physiological conditions. © 2007 Elsevier Ltd. All rights reserved.
Keywords: ubiquitin; polyubiquitin chains; tetraubiquitin; Lys48-linked; crystal structure
Introduction The covalent attachment of ubiquitin to substrate proteins helps to regulate numerous cellular pathways.1–3 Specific lysine residues on the substrate can be modified with a single ubiquitin (monoubiquitination) or with a polyubiquitin chain. Ubiquitin conjugation involves a series of enzymatic steps that end with the formation of an isopeptide bond between the C-terminal carboxylate of ubiquitin and the ε-amine of the substrate lysine. The 76 residue ubiquitin protein itself contains seven surface lysine residues that can each be modified by ubiquitin conjugation via Lys–Gly76 isopeptide bonds, leading to the formation of linear polyubi-
† Deceased. Abbreviations used: UIM, ubiquitin-interacting motif; UBA, ubiquitin-associated domain. E-mail address of the corresponding author: [email protected]
quitin chains.3 There are different types of polyubiquitin chains, each distinguished by the particular Lys residue to which each successive ubiquitin is conjugated. Although all seven surface lysine residues of ubiquitin can be used in the formation of polyubiquitin chains, Lys63 and Lys48-linked chains are more frequently observed and are the best characterized. Lys63-linked polyubiquitin chains play a role in DNA damage tolerance and in the inflammatory response through non-degradative signaling pathways.4 Lys48-linked polyubiquitin chains target their substrates to the 26 S proteasome for degradation,5 and the minimal chain needed for efficient proteasomal recognition and substrate degradation has been shown to consist of four ubiquitin molecules.6 The different types of polyubiquitin chains must have distinct structural characteristics that target their substrates to specific cellular pathways. A variety of protein domains called ubiquitinbinding domains recognize and bind ubiquitin, including UIMs (ubiquitin-interacting motifs), UBAs (ubiquitin-associated domains), and others.7
0022-2836/$ - see front matter © 2007 Elsevier Ltd. All rights reserved.
205
Lys48-linked Tetraubiquitin at Neutral pH
Some of these ubiquitin-binding domains bind monoubiquitin, while others recognize polyubiquitin chains.8 Ubiquitin-binding domains that recognize monoubiquitin all bind to a hydrophobic patch in ubiquitin, burying it at the complex interface.9,10 This hydrophobic patch is centered around three residues on the ubiquitin surface: Leu8, Ile44, and Val70.7 In the case of polyubiquitin binding, however, the way in which ubiquitinbinding domains distinguish among different chain linkages is not understood. In particular, there are UBA domains that are able to distinguish among the different types of polyubiquitin chains and bind preferentially to chains with a specific linkage.8 The specific topology of the polyubiquitin chain, or the surfaces of the polyubiquitin chain that are exposed upon formation of a particular lysine linkage, are possible structural elements that could be recognized.9,10 There are three crystal structures of Lys48-linked polyubiquitin that have been reported. The first crystal structure reported was of diubiquitin (1AAR),11 which adopts a conformation that buries the ubiquitins' hydrophobic patches at the interface. This arrangement, referred to as the closed conformation, obstructs the binding surface of ubiquitin that interacts with ubiquitin-binding domains.7 In the other conformations of polyubiquitin, referred to as open conformations, the hydrophobic patches of the ubiquitins are exposed and available for binding. The other two previously reported Lys48-linked polyubiquitin crystal structures are of tetraubiquitin (1TBE,12 1F9J13), and differ in overall organization and topology. In one case, the chain is in an open conformation, with the hydrophobic patch of each ubiquitin exposed instead of buried at an interface,12 while the other tetraubiquitin structure has a configuration that buries the hydrophobic patch of each ubiquitin.13 In addition, these tetraubiquitin structures differ from the diubiquitin crystal structure in the way in which the proximal ubiquitin (the first ubiquitin in the polyubiquitin chain that is conjugated to a substrate lysine residue, and which we denote as 1) and the next ubiquitin (which we denote as 2) are oriented. Different conformations are observed in each structure for the covalent linkages, which consist of the C-terminal ubiquitin residue, Gly76 (on 2), and the covalently linked Lys48 side-chain (on 1). A similar difference in the orientation and the linkage is observed for the other two ubiquitins, 3 and 4. Another complication affecting the interpretation of the previous tetraubiquitin crystal structures is that both were determined from crystals containing only half the tetraubiquitin in the asymmetric unit, making the linkage between ubiquitins 2 and 3 ambiguous. In one case (1F9J), this led to two possible conformations for tetraubiquitin.13 Distinguishing among the different possible conformations is important to understanding how Lys48-linked polyubiquitin chains are recognized in the cell. NMR studies of Lys48-linked diubiquitin suggest that its conformation is dynamic and pH sensitive.14
At neutral pH, diubiquitin adopts a predominantly closed conformation similar to that seen in the diubiquitin crystal structure,14,15 while at more acidic pH the diubiquitin takes on a more open conformation that exposes the hydrophobic patches. 14,16 A similar dependence on pH was observed for tetraubiquitin.14 Solution studies of Lys48-linked tetraubiquitin at neutral pH have yielded useful insights into tetraubiquitin topology, but ambiguities regarding pair-wise contacts between ubiquitins remain, and the overall chain topology is still not understood.14,17 In addition, the tetraubiquitin crystal structures that were crystallized at acidic pH do not account for the solution study data. This leaves the overall topology and inter-ubiquitin contacts for Lys48-linked tetraubiquitin at physiological pH still in question. We report here a new crystal structure of Lys48linked tetraubiquitin that differs in several respects from previously reported structures. The crystals were grown at near-neutral pH (6.7) and contain two complete tetraubiquitin chains in the asymmetric unit. The tetraubiquitin is in a conformation that differs from that in previously reported structures and clearly shows all the linkages between the ubiquitins, which were not seen in the previous structures. The overall conformation is a dimer of diubiquitins, consistent with results from solution studies at neutral pH. We present additional, site-directed spin-labeling studies that further support the closed compact structure seen in the crystal. This new structure provides molecular details on the relative orientation, and specific contacts of the ubiquitin monomers in the chain, lending insight into the overall shape and surface of Lys48-linked tetraubiquitin that may be used in its roles within the cell.
Results and Discussion The crystal structure of Lys48-linked tetraubiquitin at near-neutral pH Crystals of Lys48-linked tetraubiquitin formed at pH 6.7 in space group C2, and contain two tetraubiquitin chains in the asymmetric unit. Diffraction data were collected to 2.2 Å resolution and the structure was determined by molecular replacement using diubiquitin as a search model,11 with residues 74–76 omitted from both the proximal (1) and distal ubiquitins. Data collection and refinement statistics are summarized in Table 1. The two tetraubiquitin chains in the asymmetric unit are virtually identical, superimposing with a root-mean-square-deviation (r.m.s.d.) of 0.4 Å in Cα positions. The tetraubiquitin chain adopts a conformation consisting of a dimer of diubiquitins, with each diubiquitin subunit adopting the closed conformation (Figure 1). Starting from the proximal ubiquitin (1), the first diubiquitin subunit consists of ubiquitins 1 and 2, and the second diubiquitin subunit is made up of ubiquitins
206
Lys48-linked Tetraubiquitin at Neutral pH
Table 1. Data collection and refinement statistics K48-linked Ub4 A. Data collection Space group Cell dimensions a, b, c (Å) β (°) Resolution (Å) Rsym (%) I/σI Completeness (%) Redundancy B. Refinement Resolution (Å) No. reflections Rwork/Rfree No. atoms Protein Water Sulfate Mes B-factor Protein Water Sulfate Mes r.m.s. deviations Bond lengths (Å) Bond angles (°)
C2 59.10, 77.08, 139.36 90.32 50–2.2 (2.28–2.20) 10.0 (55.6) 22.5 (2.8) 93.4 (93.7) 3.2 (3.4) 2.2 31,476 22.2/26.2 4812 149 4 2 56.1 53.2 80.6 90.6 0.007 1.46
The highest-resolution shell is shown in parentheses.
3 and 4; 4 is the distal ubiquitin in the chain (Figure 1). The linkage between the two diubiquitin subunits is therefore between ubiquitins 2 and 3. Each diubiquitin subunit (1–2 and 3–4) is nearly identical in structure, superimposing with a r.m.s.d. value of 0.6 Å between the two subunits. Both are similar to the diubiquitin crystal structure (1AAR11), with an r.m.s.d. value of 0.6 Å for the 1–2 subunit, and 0.9 Å for the 3–4 subunit. The hydrophobic patch of each ubiquitin, comprising residues Leu8, Ile44, and Val70, is located at the diubiquitin interface. This inter-subunit interface has a total buried surface area of 1458 Å2 between ubiquitins 1 and 2 of the first diubiquitin subunit, and likewise between ubiquitins 3 and 4 of the second diubiquitin subunit. The
interface between the two diubiquitin subunits (1–2 and 3–4) is centered between ubiquitins 1 and 3 and has a total buried surface area of 1455 Å2 (Figure 1). The overall conformation of tetraubiquitin in the present structure is different from that seen in the two previous tetraubiquitin crystal structures (1TBE, and 1F9J).12,13 In particular, the Lys48–Gly76 linkage adopts a different conformation in each of the three structures (Figure 2), which places the ubiquitin hydrophobic patch residues, Leu8, Ile44, Val70, in different orientations within tetraubiquitin (Figure 3). Importantly, the two previous tetraubiquitin crystal structures contain only half a tetraubiquitin in the asymmetric unit, and the isopeptide bond between ubiquitins 2 and 3 was therefore not resolved in the electron density maps. In one case (1F9J), the overall linkage was ambiguous and two possible conformations were proposed. In the present structure, by contrast, all three linkages joining the C terminus of one ubiquitin to Lys48 of the next are well resolved in the electron density map. In addition, both previous tetraubiquitin crystal structures were crystallized at acidic pH, one at pH 5.0 (1TBE), and the other at pH 4.8 (1F9J), while the structure of tetraubiquitin reported here was determined from crystals grown at pH 6.7. As discussed below, the acidic pH at which the previous structures were determined may have favored a different conformation of tetraubiquitin. The conformation of tetraubiquitin in solution The present Lys48-linked tetraubiquitin crystal structure is supported by data from solution studies. NMR chemical shift perturbation data show that the conformation of diubiquitin is dynamic and pH dependent; at pH 4.5 the open conformation is fully populated, and at pH 7.5 the closed conformation is almost fully populated, with tetraubiquitin exhibiting a similar pH-dependent behavior.14 At pH 6.8, it is estimated that less than 15% of the diubiquitin chains are in the open conformation,14,16 with the inter-conversion between the closed and open conformations occurring on a 10 ns time scale. The chemical shift perturbation data indicate that
Figure 1. A new conformation of Lys48-linked tetraubiquitin. The tetraubiquitin coloring from proximal ubiquitin to distal ubiquitin is: yellow–cyan–green–blue (1–2–3–4). Rotating 45° in either direction about the x-axis from the center structure shows the interface of the two diubiquitin subunits (1–2 on the left, 3–4 on the right).
Lys48-linked Tetraubiquitin at Neutral pH
Figure 2. Overlays of the proximal ubiquitins 1–2 from the three different tetraubiquitin crystal structures. The alignment is done with ubiquitin 1; ubiquitin 2 appears in different positions due to the different Lys48–Gly76 linkages seen in the various crystal structures. Ubiquitins 1–2 from the present structure are shown in purple (which superimposes with the diubiquitin crystal structure 1AAR, not shown). Ubiquitins from 1TBE are colored orange, and ubiquitins from 1F9J are colored green.
ubiquitins 3 and 4 in tetraubiquitin are essentially identical to those in diubiquitin, thus suggesting that these two distal ubiquitins in the chain pack together in a manner similar to that seen in the closed form of diubiquitin.14 Moreover, additional NMR data indicate that the patterns of chemical shift perturbations in the proximal ubiquitin (1) in tetraubiquitin and in the proximal ubiquitin in the closed form of diubiquitin are nearly identical.17
207 Similar observations were made for ubiquitin 3 of tetraubiquitin (in agreement with Varadan et al.14), suggesting that all of these ubiquitins (1 and 3 of tetraubiquitin, and 1 of diubiquitin) are equivalent and form closed interfaces with another ubiquitin.17 This would suggest that ubiquitins 1 and 3 of tetraubiquitin are involved in closed diubiquitin interfaces with the other two ubiquitins in the chain (2 and 4). The tetraubiquitin structure reported here shows that ubiquitins 1 and 3 form closed diubiquitin interfaces with ubiquitins 2 and 4, respectively, with the interfaces between each ubiquitin moiety in good agreement with the chemical shift perturbations observed in solution14 (Figure 4). Site-directed spin labeling studies support the relevance of the present structure to tetraubiquitin's conformation in solution (Figure 5). A paramagnetic spin label, (1-oxyl-2,2,5,5-tetramethyl-3-pyrroline-3methyl)methanesulfonate (MTSL), was attached to Cys48 in the distal ubiquitin (4) in order to map residues in the other ubiquitins that are close in space to the spin label by using the distance dependence of the paramagnetic relaxation enhancement effect.18 Consistent with the present tetraubiquitin structure, strong signal attenuations were observed in all ubiquitin units, most notably in the proximal ubiquitin (1) (Supplementary Data, Figure 1). As seen in Figure 5, signal attenuations observed in ubiquitins 1–3 are consistent with the proximity of the affected sites to residue 48 in ubiquitin 4, where the spin label was attached. We note that, although diubiquitin (and therefore tetraubiquitin) in solution are in dynamic equilibrium between the closed and open conformations, the spin-labeling data primarily reflect the domain positioning in the compact, closed conformation. This is due to the sharp decay (∝ R− 6) of the paramagnetic relaxation effect as a function of the distance R from the unpaired electron of the spin label,19 such that the main contribution to the observed residue-specific signal attenuations comes from the predominant closed
Figure 3. Comparison of the three tetraubiquitin crystal structures. Coloring is the same as for Figure 1, with the proximal ubiquitin yellow and the distal ubiquitin blue. Hydrophobic patch residues, Leu8, Ile44, and Val70 are shown as red sticks. The structure on the left is the present structure (2O6V), with the hydrophobic patches buried as seen in the structure of diubiquitin (1AAR). The pH 5.0 structure in the center (1TBE) has all the hydrophobic patches exposed. The pH 4.8 structure on the right (1F9J) has the hydrophobic patches buried in a different arrangement than in the pH 6.7 structure (2O6V).
208
Lys48-linked Tetraubiquitin at Neutral pH
Figure 4. The tetraubiquitin conformation in the crystal structure is consistent with the interaction surfaces between ubiquitins mapped using NMR-detected chemical shift perturbations in solution at pH 6.8 as detailed.14 The orientation of tetraubiquitin is the same as for Figure 1; residues colored red show chemical shift perturbations (versus monoubiquitin) of >0.05 ppm. Perturbations in the linkage-involved residues, Lys48 (in 1 and 3), and 74–76 (in 2 and 4) are not colored.
conformation. The contribution from the less populated, open conformations is practically negligible, as shown for diubiquitin.16 Taken together, these NMR data indicate that the crystal structure of tetraubiquitin reported here reflects the Lys48linked tetraubiquitin conformation in solution at physiological pH. The compact tetraubiquitin conformation is also in agreement with NMR relaxation data in solution, which are sensitive to molecular tumbling, and hence to the size and shape of the molecule. The overall rotational correlation time of 16.6 ns (at 25 °C), predicted based on this structure using the HYDRONMR program,20 is a factor of 2 higher than the measured value for diubiquitin14 (∼ 8 ns) and
Figure 5. Validation of the tetraubiquitin structure by site-directed spin labeling data in solution. Sites in ubiquitins 1–3 with significant signal attenuation in the presence of spin label are colored purple for attenuation >20% and red for attenuation > 40%. Also shown is the side-chain of residue 48 in the distal ubiquitin (Arg48 in the present structure was replaced with a cysteine for spin labeling).
fourfold higher than that for monoubiqiuitin (4.3 ns). The experimental 1H T2 values, 13, 26, and 50 ms, measured at these conditions for backbone amides in tetra-, di-, and monoubiquitin, respectively,14 almost exactly reproduce the inverse proportionality dependence between the T2 and the rotational correlation time, expected for transverse spin relaxation in the case of protein tumbling in solution.21 There is also good agreement between the average experimental 15N T1 and T2 values for backbone amides in tetraubiquitin reported, 17 774(± 119) ms and 64.7(± 8.0) ms, respectively, and their predicted values, 862 ms of 59.5 ms, at the experimental conditions of the study (30 °C and 15N Larmor frequency of 50.7 MHz). The NMR data therefore suggest that the crystal structure of tetraubiquitin determined here indeed represents the predominant conformation of the chain in solution at near-physiological pH. In this conformation, the hydrophobic patches on ubiquitins are buried at the interface in each diubiquitin subunit. It should be emphasized, however, that the chain is not rigidly locked in this conformation in solution. This conclusion is suggested by the observation that the conformation of an isolated diubiquitin subunit is dynamic, with a relatively small (∼ 1 kcal/mol) difference in the free energy between the closed and open states.16 In addition, recent studies showed that some UBA domains10,22 as well as smaller molecules14,23 bind to diubiquitin directly at the inter-ubiquitin interface. The interconversion between the closed and open conformations of diubiquitin, and likely of longer chains, is essential for allowing access to the hydrophobic patch on ubiquitin monomers, which are targeted by almost all ubiquitin-binding domains. Locking the chain in the closed conformation could therefore block binding to this patch. It has been shown that cyclization of Lys48-linked diubiquitin by a crosslinker that restricted interface opening strongly impedes the ability of this chain to bind a UBA domain in a high-affinity sandwich-like mode.24 All this suggests that other states of the chain corresponding to more open conformations must exist
Lys48-linked Tetraubiquitin at Neutral pH
and be accessible to tetraubiquitin in solution. At neutral pH, however, these states are less populated than the conformation determined here. The observed conformation of polyubiquitin is dependent upon the type of isopeptide linkage that joins the individual ubiquitin molecules. The topology of the polyubiquitin chain and the interaction between individual monomers seen in the present Lys48-linked tetraubiquitin structure could not be adopted by a Lys63-linked chain. An isopeptide linkage between Lys63 and the neighboring ubiquitin would force a proximity of Lys63 with the C terminus of the neighboring ubiquitin that would not allow Lys63-linked tetraubiquitin to adopt the conformation seen in the present structure. Similarly, other types of tetraubiquitin chains containing different lysine linkages are unlikely to be able to adopt the precise conformation observed here. Polyubiquitin linkage type may therefore give rise to different recognition signals because of the constraints it places on the polyubiquitin conformation. Recognition surface of Lys48-linked tetraubiquitin The tetraubituitin structure described here suggests how the unique topology of Lys48-linked tetraubiquitin allows a key ubiquitin residue, Leu8, to play a central role in mediating the binding of Lys48-linked polyubiquitin to ubiquitin-binding domains. The hydrophobic patch on ubiquitin has been shown to be important for ubiquitin function.6,25–27 The residues comprising this patch on the surface of ubiquitin, Leu8, Ile44, and Val70, mediate interactions between mono- and polyubiquitin and numerous ubiquitin-binding domains.7 These resi-
209 dues have also been shown to be an important recognition element for the binding of Lys48-linked tetraubiquitin to the 26S proteasome.26,27 In particular, Leu8 is an important part of this recognition element, with the L8A mutation in the chain causing the largest defect in substrate degradation as compared to the I44A and V70A mutations.26,27 Because of this, a number of studies have examined the effects of Leu8 mutations in the context of Lys48linked polyubiquitin chains.6,25–27 Binding studies with S5a (which contain two UIM domains) or HHR23A (which contain two UBA domains) indicate that these two classes of ubiquitin-binding proteins behave differently in response to the position and number of L8A mutations in tetraubiquitin. In the case of S5a, only a subset of the Leu8 side-chains are needed for binding, 26,27 whereas in the case of HHR23A, all the Leu8 sidechains appear to play a role in the interaction.25 Mutation of two of the four Leu8 residues in tetraubiquitin is tolerated by S5a if the mutations occur only in ubiquitins 2 and 4; however, for HHR23A, substitutions of any two Leu8 residues have a significant effect on HHR23A binding. In the present tetraubiquitin crystal structure, the ubiquitin hydrophobic patches are buried at the diubiquitin intersubunit interface. However, Leu8 of each ubiquitin lies at the edge of the interface and is oriented in a way that gives rise to a hydrophobic stripe on the surface at the middle of each diubiquitin subunit (Figure 6). In the context of the tetraubiquitin structure, these hydrophobic surface stripes are related to each other by a 90° rotation. The presence of these two Leu8 hydrophobic stripes that are unique to this particular Lys48-linked tetraubiquitin structure may provide a potential recognition and interaction surface. It is also
Figure 6. The hydrophobic surface stripes on the Lys48-linked tetraubiquitin surface showing the orientation of all the ubiquitin Leu8 side-chains. The tetraubiquitin coloring is the same as for Figure 1, with the proximal ubiquitin 1 yellow, and the distal ubiquitin 4 being blue. The Leu8 side-chains are shown in red with the numbers corresponding to the specific ubiquitin in the chain a particular Leu8 is from.
210
Lys48-linked Tetraubiquitin at Neutral pH
possible that Leu8 plays an indirect role in binding to partner proteins by stabilizing the closed conformation of tetraubiquitin seen in the present crystal structure.10 In this case, contacts mediated by Leu8 could control access to the hydrophobic patches of the individual ubiquitins by regulating the opening and closing of the inter-ubiquitin interface in the diubiquitin subunits. In either case, the structure and conformation of Lys48-linked tetraubiquitin presented here is likely an important feature in its signaling role.
side-chains were adjusted using the program O31 and refined with CNS.32 The final model contains two tetraubiquitin complexes in the asymmetric unit, and includes residues 1–75 of ubiquitins A and E, residues 1–76 of ubiquitins B, C, D, F, and G, and residues 2–76 of ubiquitin H. Data collection and refinement statistics are summarized in Table 1.
Materials and Methods
NMR studies
Proteins and Lys48-linked tetraubiquitin chain formation Expression and purification of ubiquitin has been described.28 Synthesis of Lys48-linked tetraubiquitin followed published protocols,28 with the following substitutions. Two forms of Lys48-linked diubiquitin were synthesized with the distal ubiquitin carrying different mutations. The Lys48-linked diubiquitin used in the proximal end (1–2) of tetraubiquitin contained K48C, K63R mutations at ubiquitin 2. The Lys48-linked diubiquitin used in the distal end (3–4) of tetraubiquitin contained K48R, K63R mutations at ubiquitin 4, removing the need to use the K48C mutation to block K48 on ubiquitin 4 during diubiquitin formation. Tetraubiquitin constructs for NMR studies were synthesized using a segmental isotope-labeling strategy detailed elsewhere,14 such that a specific single ubiquitin unit per chain was 15N-enriched. All of these tetraubiquitin chains contained a K48C mutation on the distal ubiquitin (4), in order to provide a single site for spin-label attachment. The proximal ubiqutin (1) contained an additional Cterminal residue Asp77; no K63R mutation was used in ubiquitin 2. Crystallization and data collection Tetraubiquitin was concentrated to 8 mg ml− 1 in 10 mM Tris (pH 7.6), 10 mM NaCl, 0.1 mM EDTA, and 1 mM dithiothreitol. Crystals were grown by the method of hanging drop vapor diffusion at 20 °C, mixing equal volumes of the protein complex with a well solution containing 2 M (NH4)2SO4, 4%(w/v) PEG 400, and 100 mM Mes (pH 6.5). Crystals grew overnight. The measured pH of both the well solution and the drop after equilibration was pH 6.7. Crystals formed in space group C2 with unit cell dimensions a = 59.10 Å, b = 77.08 Å, c = 139.4 Å, and β = 90.32°, and contained two tetraubiquitin complexes in the asymmetric unit. Diffraction data were collected at APS on BioCARS beamline 14-BM-D at a wavelength of 0.9786 Å and processed using HKL2000.29 Structure determination and refinement The structure of tetraubiquitin was determined by molecular replacement with the program MOLREP30 using diubiquitin (1AAR)11 C-terminally truncated to residue 73 of both proximal and distal ubiquitins as the search model. Missing residues, including the C-terminal tail of ubiquitin and the covalent linkage, were built and
Protein Data Bank accession codes The atomic coordinates have been deposited with the RSCB Protein Data Bank, accession code 2O6V.
The NMR studies were performed at 25 °C on a Bruker Avance 600 spectrometer with the 1H Larmor frequency of 600 MHz. NMR samples were prepared in 20 mM phosphate buffer (pH 6.8) containing 7% (v/v)2H2O and 0.02% (w/v) NaN3. The tetraubiquitin concentration was 0.5–0.7 mM. The paramagnetic spin label (1-oxyl-2,2,5,5tetramethyl-3-pyrroline-3-methyl)methanesulfonate (MTSL) was covalently attached to a single cysteine site (Cys48) on the distal ubiquitin (4) as detailed.18 The paramagnetic relaxation enhancement due to close proximity to the spin label was monitored by signal attenuation observed in the 1H–15N TROSY spectra of the other three ubiquitins in the chain. The signal attenuation was quantified as the ratio of signal intensities in the spectra recorded with the spin label (in the oxidized, paramagnetic state) attached to the protein and in the absence of the spin label or after reducing it with a threefold molar excess of ascorbic acid.18
Acknowledgements This work was supported by the Howard Hughes Medical Institute (to C.W.) and by NIH grant GM065334 (to D.F.). Use of the Advanced Photon Source was supported by the US Department of Energy, Basic Energy Sciences, Office of Science, under contract no. W-31-109-Eng-38. Use of the BioCARS Sector 14 was supported by the National Institutes of Health, National Center for Research Resources, under grant number RR07707.
Supplementary Data Supplementary data associated with this article can be found, in the online version, at doi:10.1016/ j.jmb.2006.12.065
References 1. Sun, L. & Chen, Z. J. (2004). The novel functions of ubiquitination in signaling. Curr. Opin. Cell Biol. 16, 119–126. 2. Ulrich, H. D. (2002). Degradation or maintenance: actions of the ubiquitin system on eukaryotic chromatin. Eukaryotic Cell, 1, 1–10.
211
Lys48-linked Tetraubiquitin at Neutral pH 3. Pickart, C. M. & Fushman, D. (2004). Polyubiquitin chains: polymeric protein signals. Curr. Opin. Chem. Biol. 8, 610–616. 4. Chen, Z. J. (2005). Ubiquitin signalling in the NFkappaB pathway. Nature Cell Biol. 7, 758–765. 5. Pickart, C. M. & Cohen, R. E. (2004). Proteasomes and their kin: proteases in the machine age. Nature Rev. Mol. Cell Biol. 5, 177–187. 6. Thrower, J. S., Hoffman, L., Rechsteiner, M. & Pickart, C. M. (2000). Recognition of the polyubiquitin proteolytic signal. EMBO J. 19, 94–102. 7. Hicke, L., Schubert, H. L. & Hill, C. P. (2005). Ubiquitin-binding domains. Nature Rev. Mol. Cell Biol. 6, 610–621. 8. Raasi, S., Varadan, R., Fushman, D. & Pickart, C. M. (2005). Diverse polyubiquitin interaction properties of ubiquitin-associated domains. Nature Struct. Mol. Biol. 12, 708–714. 9. Varadan, R., Assfalg, M., Haririnia, A., Raasi, S., Pickart, C. & Fushman, D. (2004). Solution conformation of Lys63-linked di-ubiquitin chain provides clues to functional diversity of polyubiquitin signaling. J. Biol. Chem. 279, 7055–7063. 10. Varadan, R., Assfalg, M., Raasi, S., Pickart, C. & Fushman, D. (2005). Structural determinants for selective recognition of a Lys48-linked polyubiquitin chain by a UBA domain. Mol. Cell, 18, 687–698. 11. Cook, W. J., Jeffrey, L. C., Carson, M., Chen, Z. & Pickart, C. M. (1992). Structure of a diubiquitin conjugate and a model for interaction with ubiquitin conjugating enzyme (E2). J. Biol. Chem. 267, 16467–16471. 12. Cook, W. J., Jeffrey, L. C., Kasperek, E. & Pickart, C. M. (1994). Structure of tetraubiquitin shows how multiubiquitin chains can be formed. J. Mol. Biol. 236, 601–609. 13. Phillips, C. L., Thrower, J., Pickart, C. M. & Hill, C. P. (2001). Structure of a new crystal form of tetraubiquitin. Acta Crystallog. sect. D, 57, 341–344. 14. Varadan, R., Walker, O., Pickart, C. & Fushman, D. (2002). Structural properties of polyubiquitin chains in solution. J. Mol. Biol. 324, 637–647. 15. van Dijk, A. D., Fushman, D. & Bonvin, A. M. (2005). Various strategies of using residual dipolar couplings in NMR-driven protein docking: application to Lys48linked di-ubiquitin and validation against 15N-relaxation data. Proteins: Struct. Funct. Genet. 60, 367–381. 16. Ryabov, Y. & Fushman, D. (2006). Interdomain mobility in di-ubiquitin revealed by NMR. Proteins: Struct. Funct. Genet. 63, 787–796. 17. Tenno, T., Fujiwara, K., Tochio, H., Iwai, K., Morita, E. H., Hayashi, H. et al. (2004). Structural basis for distinct roles of Lys63- and Lys48-linked polyubiquitin chains. Genes Cells, 9, 865–875. 18. Varadan, R., Assfalg, M. & Fushman, D. (2005). Using NMR spectroscopy to monitor ubiquitin chain conformation and interactions with ubiquitin-binding
19. 20.
21. 22.
23.
24.
25.
26.
27.
28. 29. 30. 31.
32.
domains. In Ubiquitin and Protein Degradation, Methods in Enzymology (Deshaies, R. J., ed), vol. 399, part B, pp. 177–192, Wiley-VCH, Weinheim. Kosen, P. A. (1989). Spin labeling of proteins. Methods Enzymol. 177, 86–121. García de la Torre, J., Huertas, M. L. & Carrasco, B. (2000). HYDRONMR: prediction of NMR relaxation of globular proteins from atomic-level structures and hydrodynamic calculations. J. Magn. Reson. B147, 138–146. Wüthrich, K (1986). NMR of Proteins and Nucleic Acids. The George Fisher Baker non-resident lectureship in chemistry at Cornell University Wiley, New York. Trempe, J. F., Brown, N. R., Lowe, E. D., Gordon, C., Campbell, I. D., Noble, M. E. & Endicott, J. A. (2005). Mechanism of Lys48-linked polyubiquitin chain recognition by the Mud1 UBA domain. EMBO J. 24, 3178–3189. Verma, R., Peters, N. R., D'Onofrio, M., Tochtrop, G. P., Sakamoto, K. M., Varadan, R. et al. (2004). Ubistatins inhibit proteasome-dependent degradation by binding the ubiquitin chain. Science, 306, 117–120. Dickinson, B. Varadan, R. & Fushman, D. (2007). Effects of cyclization on conformational dynamics and binding properties of Lys48-linked di-ubiquitin. Protein Sci. In the press. Raasi, S., Orlov, I., Fleming, K. G. & Pickart, C. M. (2004). Binding of polyubiquitin chains to ubiquitinassociated (UBA) domains of HHR23A. J. Mol. Biol. 341, 1367–1379. Beal, R., Deveraux, Q., Xia, G., Rechsteiner, M. & Pickart, C. (1996). Surface hydrophobic residues of multiubiquitin chains essential for proteolytic targeting. Proc. Natl Acad. Sci. USA, 93, 861–866. Beal, R. E., Toscano-Cantaffa, D., Young, P., Rechsteiner, M. & Pickart, C. M. (1998). The hydrophobic effect contributes to polyubiquitin chain recognition. Biochemistry, 37, 2925–2934. Pickart, C. M. & Raasi, S. (2005). Controlled synthesis of polyubiquitin chains. Methods Enzymol. 399, 21–36. Otwinowski, Z. & Minor, W. (1997). Processing of X-ray diffraction data collected in oscillation mode. Methods Enzymol. 176, 307–326. Vagin, A. & Teplyakov, A. (1997). MOLREP: an automated program for molecular replacement. J. Appl. Crystallog. 30, 1022–1025. Jones, T. A., Zou, J. Y., Cowan, S. W. & Kjeldgaard, M. (1991). Improved methods for building protein models in electron density maps and the location of errors in these models. Acta Crystallog. sect. A, 47, 110–119. Brunger, A. T., Adams, P. D., Clore, G. M., DeLano, W. L., Gros, P., Grosse-Kunstleve, R. W. et al. (1998). Crystallography and NMR system: a new software suite for macromolecular structure determination. Acta Crystallog. sect. D, 54, 905–921.
Edited by P. Wright (Received 10 November 2006; received in revised form 22 December 2006; accepted 26 December 2006) Available online 29 December 2006