A distinct mammalian disome collision interface harbors K63-linked polyubiquitination of uS10 to trigger hRQT-mediated subunit dissociation

Abstract

Translational stalling events that result in ribosome collisions induce Ribosome-associated Quality Control (RQC) in order to degrade potentially toxic truncated nascent proteins. For RQC induction, the collided ribosomes are first marked by the Hel2/ZNF598 E3 ubiquitin ligase to recruit the RQT complex for subunit dissociation. In yeast, uS10 is polyubiquitinated by Hel2, whereas eS10 is preferentially monoubiquitinated by ZNF598 in human cells for an unknown reason. Here, we characterize the ubiquitination activity of ZNF598 and its importance for human RQT-mediated subunit dissociation using the endogenous XBP1u and poly(A) translation stallers. Cryo-EM analysis of a human collided disome reveals a distinct composite interface, with substantial differences to yeast collided disomes. Biochemical analysis of collided ribosomes shows that ZNF598 forms K63-linked polyubiquitin chains on uS10, which are decisive for mammalian RQC initiation. The human RQT (hRQT) complex composed only of ASCC3, ASCC2 and TRIP4 dissociates collided ribosomes dependent on the ATPase activity of ASCC3 and the ubiquitin-binding capacity of ASCC2. The hRQT-mediated subunit dissociation requires the K63-linked polyubiquitination of uS10, while monoubiquitination of eS10 or uS10 is not sufficient. Therefore, we conclude that ZNF598 functionally marks collided mammalian ribosomes by K63-linked polyubiquitination of uS10 for the trimeric hRQT complex-mediated subunit dissociation.

Fig. 1. In vitro reconstitution of translational arrest and ZNF598-mediated ubiquitination.

A Schematic representation of mRNA templates used in RRL in vitro translation reactions. All constructs encode the N-terminal part of XBP1u (1–193), which is appended (X) by either the XBP1u (194–261) stalling sequence, a poly(A) stretch, a stable RNA stem-loop sequence or a self-cleaving ribozyme (Rz). All constructs contain an N-terminal His₆ and PA-tag for affinity purification and antibody detection. B Schematic overview of the in vitro stalling and ubiquitination assays. C–E In vitro translation of staller mRNAs from A. The free peptide and the peptidyl-tRNA (pep-tRNA) arrest products were detected by Western blotting using an anti-PA antibody. We obtained essentially the same results as two independent experiments. C Purified RNCs were treated with RNase A to verify the peptidyl-tRNA arrest products. D, E RNCs purified from in vitro translation reactions using all mRNA templates (A) but XBP1u (1–235) is used as a control and were subjected to ultracentrifugation through sucrose density gradients and individual gradient fractions were visualized by Western blotting using the anti-PA antibody. F In vitro ubiquitination by ZNF598 of RNCs stalled on mRNAs from A. RNCs from in vitro translation reactions with or without recombinantly purified 3FLAG-ZNF598 were purified and eS10 and uS10 ubiquitination was visualized by Western blotting. We obtained essentially the same results in at least three independent experiments.

Fig. 2. Cryo-EM structure comparison of the human XBP1-stalled disome and the yeast disome.

A Composite cryo-EM density and the molecular model (B) of the human XBP1-stalled collided disome. C Cut-in view of the tRNA states. The leading ribosome is in a non-rotated POST-state with P/P E/E-tRNAs and an empty A-site. The colliding ribosome is in a rotated PRE-translocation state with A/P P/E hybrid tRNAs. D Cryo-EM structures of the human XBP1-stalled disome and the rabbit eRF1 mutant stalled disome (reconstructed from individual maps EMD-0192, 0194, 0195, and 0197) were superimposed. Little variance in the inter-ribosomal orientation is observed between the two disomes. E Overview of the inter-ribosomal interface between leading and colliding ribosome with orientation indicated in panel B. F–I Focused views on main contact areas with indicated rotation in reference to panel E. F Detail of 40S head-to-head contacts. G Detail of the 40S body-to-body contacts around 18S rRNA h16 of the colliding ribosome (h16-2). H Detail of the interactions between the 60S of the leading ribosome and the 40S of the colliding ribosome. I Detail of the contacts of the 18S rRNA expansion segment ES6c of the colliding ribosome (ES6c-2) with the 40S proteins of the leading ribosome. Focused view panels with indicated rotations refer to orientation in lower panel B .

Fig. 3. Structural comparison of the human and the yeast disome.

Cryo-EM densities and models are colored in shades of blue for the human disome and red for the yeast disome (EMD-4427 [https://www.ebi.ac.uk/emdb/EMD-14181], PDB ID: 6I7O [10.2210/pdb7QVP/pdb]) in all panels. Focused views with indicated rotations refer to panel A left. A Overlay of the cryo-EM densities of the human and the yeast disomes shows an apparent 18° rotation of the colliding ribosome with respect to the leading ribosome in yeast compared to the human disome. B, C Detailed view of the single human inter-ribosomal 40S–60S contact area. D, E Detail of the yeast-specific inter-ribosomal 40S–60S contact via the expansion segment ES6b of the colliding ribosome (ES6b-2). F, G Detail of differences in the head-to-head contact area.

Fig. 4. hRQT-mediated subunit dissociation of collided ribosomes.

A Schematic overview of the hRQT ribosome splitting assay. B Schematic representation of the XBP1u and poly(A) staller mRNAs. C–F Collided ribosomes derived from in vitro translation reactions of XBP1u (C, D) or poly(A) (E, F) staller mRNAs were isolated and ubiquitinated by ZNF598. The purified ubiquitinated RNCs were incubated with (D, F) or without (C, E) the hRQT complex. Ribosomal species resulting from the splitting reaction were separated by ultracentrifugation through sucrose density gradients. Gradient fractions were analysed by Western blotting using antibodies against ribosomal proteins uS10 and eS10 and against HA- and FLAG-tag. We obtained essentially the same results in at least three independent experiments.

Fig. 5. The hRQT complex activity requires ASCC2 binding to K63-linked polyubiquitin chains and ASCC3 ATPase activity.

A Schematic overview of the ZNF598-mediated in vitro ubiquitination assay using XBP1u-stalled collided ribosomes. B Purified XBP1u-RNCs were separated by sucrose density gradient ultracentrifugation. Indicated polysome fractions (di-, tri-, and tetrasomes) were collected for the ZNF598-mediated ubiquitination reaction. C The collected polysome fractions were ubiquitinated by ZNF598 using either His₆-tagged wild-type ubiquitin (WT-Ub) or the indicated mutants (K63only-Ub, K63R-Ub, and K0-Ub). Reactions were analysed by western blotting with anti-uS10 and -eS10 antibodies. We obtained essentially the same results in at least three independent experiments. D, E In vitro binding assay of K63-linked polyubiquitin chain with the hRQT complex. The hRQT complex composed of FLAG-ASCC3, HA-ASCC2, and HA-TRIP4 on the beads were incubated with monoubiquitin or K63-linked polyubiquitin chain (D), and monoubiquitin (Mono-Ub), K48-linked tetraubiquitin chains (Tetra-UB(K48)), or K63-linked tetraubiquitin chains (Tetra-Ub(K63)) (E). Samples were analysed by western blotting with anti-Ub, -ASCC3, -ASCC2, or -TRIP4 antibodies. We obtained essentially the same results in three independent experiments. F, G In vitro ribosome splitting assay with the mutant hRQT complex containing either the ubiquitin-binding impaired ASCC2-Ubm (F) or the ATPase-deficient ASCC3-K505R (G). RNCs generated from XBP1u staller mRNA were purified, ubiquitinated by ZNF598, and incubated with the particular mutant hRQT complex. These hRQT in vitro splitting reactions were then subjected to ultracentrifugation through sucrose density gradients and gradient fractions were analysed by western blotting using indicated antibodies. We obtained essentially the same results in at least three independent experiments.

Fig. 6. The hRQT complex dissociates collided ribosomes with K63-linked polyubiquitin chains on uS10.

A–D In vitro hRQT complex-mediated subunit disassociation assay of K63-linked polyubiquitinated ribosomes. Isolated RNCs generated by XBP1u staller mRNA were ubiquitinated by ZNF598 with either Ub-K63R (A, B) or Ub-K63only (C, D). The ubiquitinated RNCs were incubated with (B, D) or without (A, C) the hRQT complex. After hRQT-mediated splitting reactions, samples were subjected to ultracentrifugation through sucrose density gradients and gradient fractions were analyzed by western blotting using indicated antibodies. We obtained essentially the same results in at least three independent experiments.