Sonja Lorenz
Sonja Stamp


Crystal Structure of a Ube2S-Ubiquitin Conjugate

Sonja Lorenz, Moitrayee Bhattacharyya, Christian Feiler, Michael Rape and John Kuriyan

PloS One. 2016 Feb 2;11(2):e0147550 doi: 10.1371/journal.pone.0147550     (local copy)

Abstract

Protein ubiquitination occurs through the sequential formation and reorganization of specific protein-protein interfaces. Ubiquitin-conjugating (E2) enzymes, such as Ube2S, catalyze the formation of an isopeptide linkage between the C-terminus of a “donor” ubiquitin and a primary amino group of an “acceptor” ubiquitin molecule. This reaction involves an intermediate, in which the C-terminus of the donor ubiquitin is thioester-bound to the active site cysteine of the E2 and a functionally important interface is formed between the two proteins. A docked model of a Ube2S-donor ubiquitin complex was generated previously, based on chemical shift mapping by NMR, and predicted contacts were validated in functional studies. We now present the crystal structure of a covalent Ube2S-ubiquitin complex. The structure contains an interface between Ube2S and ubiquitin in trans that resembles the earlier model in general terms, but differs in detail. The crystallographic interface is more hydrophobic than the earlier model and is stable in molecular dynamics (MD) simulations. Remarkably, the docked Ube2S-donor complex converges readily to the configuration seen in the crystal structure in 3 out of 8 MD trajectories. Since the crystallographic interface is fully consistent with mutational effects, this indicates that the structure provides an energetically favorable representation of the functionally critical Ube2S-donor interface.

Figures from the paper

(Click on the small image to get a higher-resolution version.)
Figure 1A from paper

Figure 1A - Schematic of the critical step in ubiquitin linkage formation by E2 enzymes.


The C-terminal carbonyl group of a donor ubiquitin that is thioesterified with the E2 active site cysteine undergoes a nucleophilic attack by a primary amino group of an acceptor ubiquitin. As a result, an isopeptide bond between donor and acceptor ubiquitin is formed.

Figure 1B from paper

Figure 1B - NMR derived docked model.


NMR-derived docked model of the Ube2S-donor ubiquitin complex [6]. The catalytic UBC domain of Ube2S (yellow) and ubiquitin (blue) are shown in cartoon representation. The C-terminal carbonyl group of ubiquitin and the catalytic cysteine side chain of Ube2S are displayed in ball-and-stick mode.

Figure 1C from paper

Figure 1C - Crystal structure of the Ube2S-ubiquitin conjugate.


Crystal structure of Ube2S C118M (orange) disulfide-linked to ubiquitin G76C (light green). The cysteine side chains forming the disulfide linkage are highlighted in ball-and-stick mode. Note that contacts between the two proteins (in cis) are limited to the vicinity of the disulfide linkage. In the crystal, however, Ube2S forms an interface with a second ubiquitin molecule (dark green) in trans.

Figure 1D from paper

Figure 1D - Overlay of crystal and docked model structures.


The crystallographic complex formed in trans between Ube2S (orange) and a neighboring ubiquitin molecule (green) is superposed with the NMR-derived docked model of Ube2S (yellow) and donor ubiquitin (blue) [6]. The catalytic cysteine side chains are shown in ball-and-stick rendition. The angle by which the axes of helix α1 of ubiquitin are pivoted with respect to each other in the two configurations is indicated.

Figure 2 from paper

Figure 2 - Comparison of donor recognition between the crystal structure and the NMR-derived docked model.


A. Open-book view of the interaction “footprint” on the surface of Ube2S and ubiquitin (residues 1–71) in the crystal structure (top) and in the docked model (bottom), respectively, as defined by residues that become ≥ 10% buried at the interface. The C-terminal tail of ubiquitin (residues 72–76) was omitted in this analysis and representation.
B. Details of the Ube2S-donor ubiquitin interfaces seen in the crystal structure (top) and the docked model (bottom), respectively. Key interfacial residues are displayed as balls-and-sticks. Note that Lys 117 of Ube2S and Thr 66 of ubiquitin do not make contacts in the docked model and are displayed for comparison only. Cys 118 is replaced by methionine in the crystal structure.

Figure 3A from paper

Figure 3A - Effect of Lys 117 of Ube2S on activity.


A. Detail of the structural superposition of the docked model and the crystal structure provided in Fig 1D. The Ube2S molecules are shown in yellow and orange, respectively; the ubiquitin molecules in blue and green, respectively. In the crystal structure the sidechain amino group of Lys 117 of Ube2S forms a hydrogen bond with the backbone oxygen atom of Leu 8 of ubiquitin (as indicated by the dotted line), while Lys 117 is removed from the donor interface in the docked model (see Fig 2B).
Figure 3B from paper

Figure 3B - Effect of Lys 117 of Ube2S on activity.


In vitro activity assays monitoring diubiquitin formation by full-length Ube2S (left) and the UBC domain of Ube2S (right). We compared the corresponding wildtype proteins with the K117A variants in the absence (-) and presence (+) of ATP. The K117A substitution results in decreased diubiquitin formation.

Figure 4 from paper

Figure 4 - Molecular dynamics simulations.


A. Five independent trajectories (colored differently) of the crystal structure over 100 ns each. For each simulation the Ube2S molecules were aligned and the Cα RMSD values for ubiquitin (in Å) with respect to the crystal structure are plotted over the time.
B. Eight independent trajectories (colored differently) of the docked model over 100 ns each. For each trajectory the Cα RMSD values for ubiquitin (in Å) with respect to the starting model after aligning the Ube2S molecules are plotted.
C. For each trajectory of the docked model (see B), the Cα RMSD values for ubiquitin (in Å) with respect to the crystal structure after aligning the Ube2S molecules are plotted. In three trajectories (colored in red, dark blue, and green) the model converges to a configuration that is very similar to the crystal structure.
D.Superposition of the docked model after 100 ns of simulation time (see B and C, green simulation run) with the crystal structure.


Supplemental figures from the paper

(Click on the small image to get a higher-resolution version.)
Supplemental Figure 1 from paper

Supplemental Figure 1 - Mutational effects mapped onto the interfaces seen in the crystal structure and the docked model.


The crystallographic (trans) and docked interfaces are displayed in open-book style, as in Fig 2A. Interfacial residues that were mutated and found to reduce activity in this study or previously [6] are encircled red. Residues whose substitution had no effect on activity are encircled white. Note that residues with ≥ 10% BSA (buried surface area; colored pink) do not necessarily make contacts that are energetically relevant for the interaction between Ube2S and donor ubiquitin. Lys 117 of Ube2S and Thr 66 of ubiquitin, both of which are important for activity, are interfacial residues in the crystal structure, but do not make contacts in the docked model (for details see text). Note that Arg 101 and Asp 102 of Ube2S, which are important for activity, are close to the active site and thus contact residues of the ubiquitin tail in the docked model. Since the crystallographic interface is formed in trans, i.e. the tail is oriented towards a neighboring E2 molecule, these Ube2S residues are not part of the interface in the crystal structure.

Supplemental Figure 2A from paper

Supplemental Figure 2A - Effects of Cys 118 substitutions in Ube2S on ubiquitin binding and activity.


NMR data were recorded at 25°C on a Bruker 800 MHz DRX spectrometer, equipped with a 1H/15N/13C cryoprobe and were processed with NMRPipe [50]. The binding experiments were performed as described previously [6]. In short, we prepared two samples (in 50 mM Tris, 100 mM NaCl, 7.5% D2O, 30 μM DSS, pH 7.4) containing 200 μM 15N-enriched ubiquitin and either no or a 5 x molar excess of the unlabeled Ube2S (residues 1–156) Cys 118 variant and recorded phase-sensitive gradient-enhanced 1H-15N HSQC spectra [51]. A weighted combined chemical shift difference, Δδ(1H15N), was calculated according to Δδ(1H15N)=[(δ(1H)-δ(1H)0)2+0.04(δ(15n)-δ(15n)0)2]0.5, where δ(1H) and δ(15N) denote the chemical shifts in the presence of Ube2S, and δ(1H)0 and δ(15N)0 denote the chemical shifts in the absence of Ube2S, respectively. The weighted combined chemical shift differences are plotted. Ubiquitin interacts with all tested Ube2S variants in a similar way.

Supplemental Figure 2B from paper

Supplemental Figure 2B - Effects of Cys 118 substitutions in Ube2S on ubiquitin binding and activity.


In vitro activity assays monitoring diubiquitin formation by Ube2S (residues 1–156). We compared reactions in the absence (-) and presence (+) of ATP. All three variants are active in diubiquitin formation, but display reduced activity compared to the wildtype; the activity of the C118M variant is closest to the wildtype level. The reduced activities of the Cys 118 variants are not due to a loss of donor binding.

Supplemental Figure 3A from paper

Supplemental Figure 3A - Analysis of lattice contacts of ubiquitin in the crystal structure of the Ube2S-ubiquitin conjugate.


Overview of all molecules that ubiquitin (green) contacts in the context of the crystal lattice. The molecules are shown in cartoon representation along with the symmetry operations that should be applied to ubiquitin in order to obtain the respective interface (specified in fractional space relative to the structure position given in the PDB file), the interface area (calculated as the difference in the total accessible surface areas of the isolated and interfacing structures and divided by 2), and the solvation free energy gain upon formation of the interface, ΔiG (calculated as the difference in the total solvation energies of the isolated and the interfacing structures), according to the PDBePISA server (www.ebi.ac.uk/pdbe/pisa). The Ube2S molecule to which ubiquitin (green) is linked covalently is shown in yellow, the Ube2S molecule, with which ubiquitin forms the hydrophobic, closed trans interface is shown in orange.

Supplemental Figure 3B from paper

Supplemental Figure 3B - Analysis of lattice contacts of ubiquitin in the crystal structure of the Ube2S-ubiquitin conjugate.


Detailed view of the crystallographic interface between ubiquitin (green) and a neighboring Ube2S molecule (grey). Contacting side chains are displayed in ball-and-stick representation.

Supplemental Figure 4 from paper

Supplemental Figure 4 - Comparison of the closed Ube2S-ubiquitin interface with other closed E2-donor complexes.


A.Superposition of the Ube2S-ubiquitin configuration seen in our crystal structure in trans with the closed UbcH5A-ubiquitin conjugate bound to the RING domain dimer of RNF4 (PDB ID: 4AP4) [10]. Note that the second RNF4-RING subunit is bound to another E2-conjugate in the crystal structure that is not displayed here.
B. Superposition of the Ube2S-ubiquitin configuration seen in our crystal structure (in trans) with a non-covalent, closed Cdc34-ubiquitin complex bound to an inhibitor (PDB ID: 4MDK; the inhibitor is not displayed) [12].
C. Interaction “footprints” on the surface of ubiquitin in the crystal structures of the three E2-donor complexes displayed in Figure A and Figure B, as defined by residues that become ≥ 10% buried at the interface. The contacting surface areas on ubiquitin are very similar. Note that we truncated the C-terminal tail of ubiquitin in our Ube2S-ubiquitin complex (PDB ID: 5BNB) for this representation due to the closed interface being formed in trans.