Research ArticleBIOCHEMISTRY

Structural insights into ADP-ribosylation of ubiquitin by Deltex family E3 ubiquitin ligases

See allHide authors and affiliations

Science Advances  18 Sep 2020:
Vol. 6, no. 38, eabc0418
DOI: 10.1126/sciadv.abc0418
  • Fig. 1 All Deltex E3s ADP-ribosylate Ub via the RING-DTC domains.

    (A) Reduced autoradiogram showing Ub ADP-ribosylation reactions with DTX3L and PARP9 variants in the presence of E1, E2 UbcH5B, Mg2+-ATP, Ub, and 32P-NAD+. The reaction scheme is shown above, where E2~Ub was first generated before the addition of E3 and NAD+ to initiate the reaction. (B) Domain organization of the Deltex family of E3s. All members share adjoining C-terminal RD domains. DTX1, DTX2, and DTX4 have tandem WWE domains at the N terminus and C3H2C3-type RING fingers. DTX3 and DTX3L have C3HC4-type RING fingers. DTX3L has a unique N-terminal region that mediates its interaction with PARP9, while DTX1 to DTX4 contain a proline-rich region in domain D. (C) Western blot of in vitro Ub ADP-ribosylation reactions in which E1, E2 UbcH5B, Ub, Mg2+-ATP, or full-length DTX3L (DTX3L-FL) has been omitted as indicated in the presence of biotin-NAD+ and separated by SDS–polyacrylamide gel electrophoresis (PAGE) in reducing conditions. (D) As in (C), but with DTX3L-RD. (E) Western blot of in vitro reactions with E1, E2 UbcH5B, Mg2+-ATP, Ub, biotin-labeled NAD+, and the indicated Deltex protein RING-DTC domains, pCBL, RNF38, or BIRC7 separated by SDS-PAGE in reducing conditions. For (C) to (E), α-Ub is shown in red, and NeutrAvidin DyLight is shown in green.

  • Fig. 2 DTX1 DTC domain binds NAD+.

    (A) 1H-15N HSQC spectra of 15N-DTX1-RD (black) and after the addition of NAD+ (green). ppm, parts per million. (B) Cartoon representation of the structure of the DTX1-RD. The RING, linker, and DTC domains are colored green, cyan, and magenta, respectively. Zn2+ atoms are depicted as gray spheres. (C) Structure of DTX1-RD bound to NAD+. DTX1-RD is colored as in (B) and in the same orientation. NAD+ is shown in sticks with C atoms colored yellow, O atoms colored red, N atoms colored blue, and P atoms colored orange. (D) Close-up view of NAD+-binding site corresponding to the region outlined in (C) (left). A different close-up view of the NAD+-binding site is shown in the right panel. Key NAD+-binding residues are shown in sticks, and atoms are colored as in (C). Black dashes indicate hydrogen bonds. (E) The DTX1-RD is shown as a surface representation with identical and conserved residues from the sequence alignment in fig. S2C colored black and gray, respectively. The RING, linker, and DTC domains are colored light green, cyan, and light pink, respectively. (F) Western blot (top) and Ponceau stain (bottom) of in vitro ADP-ribosylation reactions with DTX1-RD wild-type (WT) or indicated variants in the presence of E1, UbcH5B, Mg2+-ATP, Ub, and 32P-NAD+. Reactions without ATP were used to show similar E3 loading.

  • Fig. 3 DTX1 RD relies on recruitment of E2~Ub for the formation of ADPr-Ub.

    (A) Reduced SDS-PAGE gels following autoubiquitination of GST-DTX1-RD over time. Ubn, Ub chains of length n. (B) Reduced autoradiogram of Ub ADP-ribosylation reactions with DTX1-RD and UbcH5B variants in the presence of E1, Mg2+-ATP, Ub, and 32P-NAD+. (C) Western blot of in vitro reactions with E1, Mg2+-ATP, DTX1-RD, and biotin-labeled NAD+ and the indicated UbcH5B or Ub variant, separated with SDS-PAGE under reducing conditions; α-Ub is shown in red, and NeutrAvidin DyLight is shown in green.

  • Fig. 4 Flexible arrangement of RING-DTC allows the formation of ADPr-Ub.

    (A) Reduced autoradiogram showing Ub ADP-ribosylation reactions with DTX1-RD or DTX1 RING domain (DTX1-RING) with increasing concentrations of DTX1-DTC domain (DTX1-DTC) in the presence of E1, E2 UbcH5B, Mg2+-ATP, Ub, and+ 32P-NAD+. (B) Model of UbcH5B~Ub complex bound to DTX1-RD obtained by aligning the RING domain of DTX1 to RNF38 bound to E2~Ub (PDB: 4V3L), with RMSD value of 2.072 Å over 76 Cα atoms. UbcH5B~Ub complex is shown as a cartoon representation with UbcH5B in wheat, and Ub in shown blue. DTX1-RD and NAD+ are colored as in Fig. 2C. The stable conjugate of E2~Ub (UbcH5B C85K–Ub) was used in the modeling, and the C85K-Ub isopeptide linkage is shown in sticks. (C) Population frequency versus Rg (top) and Dmax (bottom), respectively, obtained modeling the solution structure of DTX1-RD using an ensemble of conformations. The gray area represents the distribution of the initial pool of generated structures, and the red line shows the two different populations of conformers obtained with this analysis. (D) Western blot of in vitro reactions with E1, UbcH5B, Ub, Mg2+-ATP, biotin-labeled NAD+, and the indicated DTX1-RD linker variant, separated with SDS-PAGE under reducing conditions; α-His (~30 kDa) is shown in green, α-Ub (~10 kDa) is shown in red, and NeutrAvidin DyLight is shown in green. (E) Left: Western blot of the hydrolysis of UbcH5B C85S–Ub or UbcH5B C85K–Ub with biotin-labeled NAD+ in the presence and absence of DTX1-RD; α-Ub is shown in red, and NeutrAvidin DyLight is shown in green. Right: SDS-PAGE gel of reactions in the left panel. (F) Proposed mechanism of ADP-ribosylation of Ub’s C terminus by DTX1-RD.

  • Fig. 5 Ub ADP-ribosylation on Gly76 is reversible by USP2.

    (A) 1H-15N HSQC spectra of wild-type 15N-Ub (black), ADPr-15N-Ub (red), and ADPr-15N-Ub following treatment with USP2 (blue). Upon ADP-ribosylation, peaks from Gly75 and Gly76 undergo distinct shifts. (B) Changes in CSP (ΔCSP) determined by 1H-15N-HQSC for each residue of ADPr-15N-Ub (red) and ADPr-15N-Ub following treatment with USP2 (blue) compared to wild-type 15N-Ub. (C) Autoradiogram of 32P-ADPr-Ub before and after treatment with USP2. (D) Nonreduced SDS-PAGE gel following the formation of UbcH5B~Ub using purified ADPr-Ub, purified ADPr-Ub treated with USP2 (middle), or wild-type Ub (right). (E) Western blot of GST-DTX1-RD autoubiquitination in the presence of biotin-NAD+; α-GST is shown in red, and NeutrAvidin DyLight is shown in green and separated by SDS-PAGE in reducing conditions. (F) Western blot of GST-Af1521 pull-downs on whole-cell lysates of HEK293 cells transfected with empty vector (EV) or full-length Myc-tagged DTX2 variants as indicated. Probed using α-Ub, α-actin, and α-Myc-tag with 1% of input loaded as a loading control. (G) As in (F), but with Myc-tagged DTX2-RD variants.

  • Table 1 Data collection and refinement statistics.

    DTX1-RDDTX1-RD-NAD+
    PDB code6Y5N6Y5P
    Data collection
    Space groupP212121P212121
    Cell dimensions
    a, b, c (Å)67.49, 86.550, 129.83066.85, 85.25, 130.95
    α, β, γ (°)90.0, 90.0, 90.090, 90, 90
    Resolution (Å)36.01–1.88 (1.88–1.93)*66.85–1.74 (1.79–1.74)*
    Rmerge (%)7.3 (117.0)5.6 (133.4)
    II14.7 (1.4)16.0 (1.0)
    Completeness (%)100.0 (100.0)100 (100)
    Redundancy6.6 (6.8)6.4 (5.9)
    Refinement
    Resolution (Å)36.01–1.88 (1.88–1.93)*65.47–1.74 (1.79–1.74)*
    No. reflections119069148052
    Rwork/Rfree (%)18.71/21.0518.53/21.31
    No. atoms
    Protein71287144
    Ions44
    Ligand140
    Water526420
    B factors (Å2)
    Protein37.2635.96
    Ions57.6843.01
    Ligand56.44
    Water45.8946.42
    RMSDs
    Bond lengths (Å)0.0110.010
    Bond angles (°)1.0791.103

    *Values in parentheses are for highest-resolution shell.

    Supplementary Materials

    • Supplementary Materials

      Structural insights into ADP-ribosylation of ubiquitin by Deltex family E3 ubiquitin ligases

      Chatrin Chatrin, Mads Gabrielsen, Lori Buetow, Mark A. Nakasone, Syed F. Ahmed, David Sumpton, Gary J. Sibbet, Brian O. Smith, Danny T. Huang

      Download Supplement

      This PDF file includes:

      • Figs. S1 to S9

      Files in this Data Supplement:

    Stay Connected to Science Advances

    Navigate This Article