Fig 2: CryoEM of the RUVBL1–RUVBL2-ZNHIT2?C complex. (A) SDS-PAGE of the purified RUVBL1–RUVBL2–ZNHIT2?C complex used for structural studies. (B) Representative 2D averages obtained from the cryoEM images of RUVBL1–RUVBL2- ZNHIT2?C. Grey arrowheads label some of the RUVBL1–RUVBL2 DII domains, while yellow arrowheads indicate the position of ZNHIT2?C. (C) Two views of the consensus cryoEM map of the RUVBL1–RUVBL2-ZNHIT2?C complex at 4.4 Å resolution. Density for ZNHIT2?C (yellow color) locates at the DII-face of the ring (grey color). (D) A view of one of the cryoEM maps obtained after classification using cryoDRGN (39), showing the DII-side of the RUVBL1–RUVBL2 ring. The density for ZNHIT2?C complex is divided in two lobules that correlate with the two domains predicted for ZNHIT2?C by AlphaFold (40). (E) Left panel, a tilted view of the cryoEM map in D showing the interaction of the large ZNHIT2?C domain. Both the atomic model of the RUVBL1–RUVBL2 complex and the ZNHIT2?C C-terminal domain are fitted within the map. Linked residues, identified by XL-MS, are shown as green dots. Each RUVBL subunit is labelled from 1 to 3 (subscript). Right panel, close-up view to highlight the proximity between the cross-linked residues in ZNHIT2 and the adjacent DII domains in the RUVBL1–RUVBL2 model.

Fig 3: Purification and assembly of RUVBL1–RUVBL2–ZNHIT2?C. (A) Left top panels, domain architecture of the components of R2TP (TPRs, tetratricopeptide repeats; PBD, PIH1D1 binding domain; RBD, RUVBL2 binding domain; PIH, protein interacting with HSP90 domain; CS, CHORD-containig proteins and SGT1 domain). Domains involved in interactions within R2TP are indicated as links with dashed lines. Left bottom panel, constructs of human ZNHIT2 used in this work (H, HIT domain). Right panel, cartoon of the structure of the R2TP complex. (B) Diagram of the network of interactions described in this work for R2TP, ZNHIT2, ECD, AAR2 and PRPF8. (C) SDS-PAGE of purified ZNHIT2?C, GST-ZNHIT2?C, GST-ZNHIT2-C and GST-ZNHIT2-N. (D) GST-pull-down experiments to test the interaction of GST-ZNHIT2?C, GST-ZNHIT2-C and GST-ZNHIT2-N with RUVBL1, RUVBL2 and RUVBL1–RUVBL2 complexes (RUVBL1/2 for simplicity in figure legends hereafter). RUVBL1/2-?DII corresponds to mutants lacking part of the DII domain. RUVBL1/2-WB corresponds to mutants in the Walker B domains, able to bind ATP but affected in ATP hydrolysis. Control lanes for these experiments are shown in Supplementary Figure S1B, C. (E) GST-pull-down experiment to evaluate the effect of the presence or absence of the nucleotides ATP, ADP or AMP–PNP in the interaction between RUVBL1–RUVBL2 and ZNHIT2?C. Control lanes for these experiments are shown in Supplementary Figure S1D. (F) Cartoon showing the crosslinks found in RUVBL1–RUVBL2–ZNHIT2 complex detected by XL-MS.

Fig 4: RUVBL1–RUVBL2 interacts with AAR2, ECD and ZNHIT2?C. (A) Pull-down experiments using the GST tag in ZNHIT2?C showing that ZNHIT2?C interacts with RUVBL1–RUVBL2 and ECD. ZNHIT2?C contaminants (*) coelute with AAR2 so the interaction between ZNHIT2?C and AAR2 could not be determined in this experiment. Electrophoretic protein mobility is shown with coloured arrowheads according to the protein labels. Main results are highlighted within rectangles. (B) Pull-down experiments using the Strep II tag in ECD confirmed its interaction with ZNHIT2?C and RUVBL1–RUVBL2. Electrophoretic protein mobility is shown with coloured arrowheads according to the protein labels. Main results are highlighted within rectangles. (C) Pull-downs using the His-tag in AAR2 confirmed its interaction with ECD and ZNHIT2?C. Albeit the His-tag been removed from RUVBL1–RUVBL2 some unspecific interaction was detected in the RUVBL1–RUVBL2 control lane (**). Electrophoretic protein mobility is shown with coloured arrowheads according to the protein labels. Main results are highlighted within rectangles. (D) Pull-down experiments using the Strep II tag in AAR2 probed that AAR2 interacts with RUVBL1–RUVBL2 both in the absence and in the presence of ZNHIT2?C. Main results are highlighted within rectangles.

Fig 5: Interactions between PRPF8 and RUVBL1–RUVBL2, RPAP3-PIH1D1, AAR2, ECD and ZNHIT2 (A) Flag-pull-down experiments showed a direct interaction between PRPF8 (bait) and RUVBL1–RUVBL2 (prey), which is not affected by the presence of ATP. Asterisks indicate contaminants that co-eluted with PRPF8. (B) Upper panel, pull-down experiments using Flag-PRPF8 (bait) and RPAP3-PIH1D1 in the presence or absence of RUVBL1–RUVBL2 were analyzed by western blot. Stained SDS-PAGE gels can be found in Supplementary Figure S8B. Lower panel, pull-down experiment using the GST tag in RPAP3 within the RPAP3-PIH1D1 complex, analyzed by western blot. MW, molecular weight markers. (C) In the presence of ECD, ZNHIT2?C, RUVBL1–RUVBL2, RPAP3-PIH1D1 and AAR2 (preys), PRPF8 (bait) interacted with RUVBL1–RUVBL2, RPAP3-PIH1D1, AAR2 and ECD as show in the Flag-pull-down assay. Control experiments using a cell extract lacking the overexpressed PRPF8, but overexpressing the Flag-tag were run in parallel SDS-PAGE for the experiments and they are shown in Supplementary Figure S8. Stained SDS-PAGE gels can be found in Supplementary Figure S8C.The presence of PRPF8, RUVBL1–RUVBL2, ECD, AAR2, RPAP3, PIH1D1 and ZNHIT2 in the Flag-pull-down assays was analyzed by western blot.