Skip to main page content
Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
, 36 (6), 1018-33

Together, Rpn10 and Dsk2 Can Serve as a Polyubiquitin Chain-Length Sensor


Together, Rpn10 and Dsk2 Can Serve as a Polyubiquitin Chain-Length Sensor

Daoning Zhang et al. Mol Cell.


As a signal for substrate targeting, polyubiquitin meets various layers of receptors upstream to the 26S proteasome. We obtained structural information on two receptors, Rpn10 and Dsk2, alone and in complex with (poly)ubiquitin or with each other. A hierarchy of affinities emerges with Dsk2 binding monoubiquitin tighter than Rpn10 does, whereas Rpn10 prefers the ubiquitin-like domain of Dsk2 to monoubiquitin, with increasing affinities for longer polyubiquitin chains. We demonstrated the formation of ternary complexes of both receptors simultaneously with (poly)ubiquitin and found that, depending on the ubiquitin chain length, the orientation of the resulting complex is entirely different, providing for alternate signals. Dynamic rearrangement provides a chain-length sensor, possibly explaining how accessibility of Dsk2 to the proteasome is limited unless it carries a properly tagged cargo. We propose a mechanism for a malleable ubiquitin signal that depends both on chain length and combination of receptors to produce tetraubiquitin as an efficient signal threshold.


Figure 1
Figure 1. Comparison of Ub-like domains binding to Rpn10
Ub-like domains of major proteasome-interacting polyUb-delivery proteins were added to UIM-containing segment of Rpn10 (Rpn10204-268) at equimolar ratio. 2D 1H-15N HSQC NMR spectra of 15N-labeled Rpn10204-268 were collected (Fig. S1), and a representative region is shown in panels A-E as overlays of NMR spectra of Rpn10204-268 alone (black) and in the presence (blue) of a molar equivalent of the UBL of (A) Dsk2, (B) Rad23, (D) Ddi1, (E) Ubp6, or Ub (as control, panel C). Shifts in specific UIM signals of Rpn10 indicate residues participating in binding. A reciprocal experiment similarly charts changes in NMR spectra of a 15N-labeled UBL upon addition of Rpn10204-268 (panels F,I and Fig. S2). Shown are overlays of NMR spectra of the UBLs of (F) Dsk2 or (H) Rad23 alone (black) and in the presence of a molar equivalent of Rpn10204-268 (blue). Panels (G,I) show the results of NMR competition assay comparing directly the affinities of Dsk2 and Rad23 for Rpn10. (G) The addition of a molar equivalent of Rad23 to prebound Dsk2-UBL/Rpn10204-268 (from panel F) did not perturb NMR spectra of Dsk2-UBL (the result shown in green), indicating that Dsk2 remains in the Rpn10-bound (blue) state. A reciprocal competition experiment (panel I) shows that upon addition of a molar equivalent of Dsk2-UBL to prebound Rad23-UBL/Rpn10204-268 (from panel H), Rad23 signals perturbed by Rpn10 (blue) returned (green) to their reference position (black), indicating that Dsk2 efficiently displaced Rad23 on Rpn10. Panels J–L emphasize the Dsk2/Rpn10 interaction by depicting overlay of representative regions of the spectrum of full-length 15N-Dsk2 alone (blue) and in the presence of a molar equivalent of Rpn10204-268 (green); also shown are positions of the corresponding signals of the isolated Dsk2-UBL free in solution (crosses) and in Rpn10-bound state (diamonds). Panel L shows signals (as 2D contours and 1D slices through peak maxima) of the indole NH group of W14 (Dsk2). In each panel, numbers represent the assigned residue for the corresponding NMR signal; to guide the eye, a shift in the peak position is shown by a red arrow. XUBA in panel K indicates a (unassigned) signal of Dsk2-UBA.
Figure 2
Figure 2. Mapping the Rpn10 surface of interaction with Ub, polyUb, and Dsk2-UBL
(Left) The magnitudes of chemical shift perturbations (CSPs) for backbone amides in Rpn10204-268 upon addition of ligand (Ub(n) or UBL) are shown as black bars for each residue in the Rpn10 sequence. Residues showing strong signal attenuation (> 80%) in the presence of the binding partner are indicated by grey vertical bars. The UIM of Rpn10 comprises an α-helix (E227-E244) and an N-terminal stretch (F218-P226) (see cartoon). Shown on the bottom is a fragment of 1H-15N HSQC spectrum of free Rpn10204-268 (black contours) superimposed with its spectra (blue contours) in complex with Dsk2-UBL (left) or Ub (right), to illustrate differences in perturbations in Rpn10 residues Q240-R242 upon binding to these proteins. (Right) Rpn10 residues perturbed by each ligand are colored red on the surface of Rpn10204-268 (CSPs > 0.07 ppm and/or signal attenuations > 80%). Strongly attenuated residues are marked in blue next to the surface drawings for each pair. Note the unique perturbations in residues 240-242 (indicated by red numbers) caused by Dsk2-UBL binding. The conserved LAMAL residues are highlighted in the cartoon at the bottom of the figure. Modeling of Rpn10-UIM structure is detailed in Supplemental Data, Fig. S5.
Figure 3
Figure 3. Mapping the Rpn10-interacting sites on Ub, Ub2, and Dsk2-UBL
(A–C) Magnitudes of CSPs for backbone amides in Ub, Ub2 (proximal Ub), or Dsk2-UBL at the endpoint of titration with Rpn10204-268 are shown as black bars for each residue. The grey bars indicate residues exhibiting strong signal attenuation (> 80%). (D–F) Maps of the perturbed residues (CSPs > 0.07 ppm and/or signal attenuations > 80%) on the surface of Ub, Ub2, or Dsk2-UBL. The location of severely attenuated residues is indicated by residue numbers (in white) on each molecule surface. Note the additional unique perturbations on the surface of Dsk2 (residues 61-64 and 72-74; marked in red) compared to the more limited hydrophobic interaction surface of Ub with Rpn10-UIM (F). (G–H) Structural cartoons of Ub and UBL point out the residues of this hydrophobic patch. (I, J) Complexes of Rpn10/Ub (I) and Rpn10/UBL (J) modeled by superimposing the structures of each protein in the pair (details in Supplemental Data) agree with the binding interface between Rpn10 and Ub or Dsk2-UBL mapped by NMR perturbation studies (Panels A–F). To guide the eye, Rpn10 is shown as a ribbon, while Ub and UBL are in surface representation. Coloring of the perturbed sites in both binding partners is the same as in D–E and Fig. 2. (K, L) Experimental validation of the models of (K) Ub/Rpn10 and (L) Dsk2/Rpn10 complexes using site-directed spin labeling of Rpn10 (details in Supplemental Data and Fig. S8). Shown are the same structures as in (I, J); painted blue are those residues in Ub or Dsk2 that were “illuminated” by the attachment of a spin label to Rpn10, as detected by strong attenuation (>54 %) in NMR signals of these residues. The spin label was attached through disulfide bond to the side chain of C247 in Rpn10 (R247C). The gold ball in panels K,L represents the position of the unpaired electron of MTSL reconstructed from the measured attenuations in NMR signals of Ub or Dsk2-UBL, respectively (see Fig. S8). To guide the eye, the approximate location of the backbone (nitrogen) of Rpn10’s residue 247, extrapolated from the orientation of the S5a UIM-1 α-helix, is shown as the cyan-colored ball.
Figure 4
Figure 4. Quantification of interaction equilibria
(A,B) Titration curves for Rpn10-binding were obtained by plotting normalized CSPs (averaged over 6-14 participating residues) as a function of ligand/protein molar ratio for proteins in complexes. Results in red represent titration of 15N-Ub (A) or 15N-Dsk2-UBL (B) with unlabeled Rpn10204-268. Superimposed in blue on the same graphs are results of the reverse titration of 15N-Rpn10204-268 with unlabeled Ub or Dsk2-UBL. The error bars represent standard deviations. The agreement between binding curves obtained in either the forward and reverse titrations indicates a 1:1 stoichiometry. (C) Illustration of a NMR titration experiment: gradual shifts in signals of selected 15N-Rpn10204-268 residues upon addition of increasing amounts of Dsk2-UBL. Various contours correspond to the following Dsk2/Rpn10 molar ratios: 0 (black), 0.21 (purple), 0.43 (light green), 0.64 (orange), 0.85 (blue), 1.07 (magenta), 1.45 (dark green), 1.83 (yellow), and 2.21 (red). (D–F) Similar titration curves obtained using Surface Plasmon Resonance (SPR) measurements (signal vs. ligand concentration) expose a clear hierarchy in the strengths of pairwise interactions between Rpn10204-268 and Ub, Ub2, Ub4, or Dsk2, both full length and Dsk2-UBL (WT or D64K mutant), and between Dsk2-UBA and Ub or Ub2. All curves are a nonlinear fit of data points to a 1:1 binding model (detailed in Supplemental Data); the results are in Table 1. Note the weaker binding resulting from D64K mutation in UBL (panel D). The corresponding differences in binding contacts are highlighted by superposition models of Rpn10 complexes with Ub (F) or Dsk2-UBL (G) (as in Fig 3I,J). Residues R242 (Rpn10) and D64 (UBL), uniquely perturbed in Rpn10/Dks2 binding, are colored red and magenta, respectively. Their side chains might form a salt bridge upon reorientation of the corresponding loop in UBL.
Figure 5
Figure 5. Competition assays reveal hierarchy in binding between Rpn10, Dsk2, and (poly)Ub
(A) Overlay of representative regions of 2D NMR spectra of 15N-Ub (left) or 15N-Ub2 (proximal Ub, right) free in solution (black contours), upon binding to Rpn10204-268 (blue), and after addition of Dsk2-UBL (green). (B) Overlay of representative regions of 2D NMR spectra of 15N-Dsk2-UBL free (black), upon binding to Rpn10204-268 (blue), and after subsequent addition (green) of Ub2 (left) or increasing amounts of Ub4, from one Ub unit per Dsk2 (middle) to one Ub4 chain per Dsk2 (right). To guide the eye, blue and green arrows show shifts in peak positions caused by the corresponding binding events. These results directly demonstrate that Dsk2-UBL can outcompete monoUb for binding to Rpn10 whereas Ub4 can outcompete Dsk2-UBL. Interestingly, Dsk2-UBL and Ub2 bind Rpn10 with comparable strength, as neither protein can fully outcompete the other. (C,D) Representative regions of 2D NMR spectra of 15N-Rpn10204-268 free (left) and bound (middle) to (C) Ub or (D) Ub2.UBA can outcompete Rpn10 for binding to both monoUb and Ub2 (right). Shifts in peak positions are indicated by red arrows. Underlined residue numbers indicate signals broadened beyond detection. The incomplete reversion of the Rpn10 spectra (right) reflects the fact that at the equimolar ratio of the proteins, there is still some fraction of Rpn10 molecules in the bound state. As the consequence of dynamic equilibrium between the free and bound states, during the time (~100 ms) relevant to NMR experiments each Rpn10 molecule has the chance to spend some fraction of time in complex with Ub2. This would result in signal broadening (due to chemical exchange and slower tumbling), which explains why some of the observed signals in are still somewhat attenuated compared to free Rpn10. (E) Pull-down assays demonstrate that the observed hierarchy in affinities is preserved at the level of full-length Rpn10: Dsk2-UBL binds Rpn10 stronger than monoUb does but weaker than polyUb. Rpn10 was crosslinked to activated Sepharose beads and mixed with either monoUb, polyUb, or recombinant purified Dsk2-UBL in PBS buffer. Samples of starting material are shown on the left resolved on 18% SDS-PAGE and protein content stained with Coomassie Blue. Following extensive washes at low and high salt, bound protein was eluted with 8M Urea and resolved by SDS PAGE to determine protein content. Sample of elution is shown in the middle panel. All monoUb was washed off and none was detected in the elution. By contrast, Dsk2-UBL and polyUb were retained on the Rpn10 column. For competition assays, UBL and Ub or UBL and polyUb were premixed at estimated 1:1 molar ratio and subjected to the same sequence of washes and elution. Elution samples of competition binding are shown on the right. MonoUb had no effect on UBL binding to the Rpn10-affinity column, however, polyUb is preferentially retained on the column, indicating a significantly lower affinity of UBL for Rpn10.
Figure 6
Figure 6. Together, Rpn10 and Dsk2 act as a Ub chain-length sensor
(A) Pull-down assays show that full-length Rpn10 forms a ternary complex with polyUb and Dsk2ΔUBL. Purified Dsk2ΔUBL, polyUb (n ≥ 4), or pre-mixed Dsk2ΔUBL and polyUb (in a 1:1 ratio) were applied to an affinity column generated from Rpn10 cross-linked to Sepharose (Supplemental Data). Bound proteins were eluted with 2M urea and assayed for presence of Dsk2ΔUBL or Ub by immunoblotting with specific antibodies. (B) Similar results were obtained for binding to the UIM-containing construct of Rpn10204-268. (C) Overlay of representative regions of 2D NMR spectra of full-length 15N-Dsk2 in a 1:1 complex with Rpn10204-268 (green, as in Fig. 1J-L) and upon subsequent addition of polyUb chains (n ≥ 4) in approximately 1:1:1 (red) or 2:1:1 (blue) molar ratio. PolyUb causes the UBL signals to return to their positions in free isolated UBL, determined independently (indicated by crosses). Signals of UBL residues that interact with UIM (W14, G48, Y60, and V71) shift upon addition of polyUb. The unshifted signal of A34 serves as a control for a residue that does not directly participate in Rpn10 binding (see Fig. 1K, Fig. 3). (D) Overlay of NMR spectra of full length 15N-Dsk2 in a 1:1 complex with Rpn10 (green) and upon subsequent addition of monoUb in 4:1 (red) or 8:1 (blue) molar ratios. In this case, the UBL signals remain essentially in the Rpn10-bound state. To guide the eye, positions of the corresponding NMR signals of free UBL (obtained from a separate experiment) are indicated as crosses or a dashed line for W14ε. UBA residues experience binding to both mono- and polyUb. For example, a signal indicated by XUBA (middle-column spectra) shifts upon addition of monoUb (D) but attenuates beyond detection in the presence of polyUb (C). These data demonstrate that by sharing a polyUb chain, the complex of Rpn10 and Dsk2 rearranges to unmask the UBL domain (as schematically shown in panel E (bottom right), see also Fig. S15A). However, the strong preference of UIM for UBL alongside the stronger affinity of UBA for monoUb result in a different ternary complex in which Dsk2 links Ub and Rpn10 (panel E (top right) and Fig. S15B). (E) A chain-length sensor. Possible ternary complexes formed by two ubiquitin receptors in mixture with ubiquitin chains of various lengths. The hierarchy of affinities of receptors for each other and for (poly)Ub provides a chain-length-sensitive mechanism able to shape Ub signaling. In a ternary complex with monoUb or short chains, Dsk2 mediates their interaction with Rpn10, and could enhance targeting of monoUb to downstream elements. By contrast, longer polyUb chains (n ≥ 4) can be shared by Rpn10 and Dsk2; the UBL domain is unmasked in the resulting ternary complex and available for interactions that recognize the “UBL signal”. Note that the arrows in E show a possible sequence of binding and rearrangement events; all these states are intrinsically at equilibrium.

Similar articles

See all similar articles

Cited by 53 PubMed Central articles

See all "Cited by" articles

Publication types

MeSH terms