Fig 2: Recognition of partially assembled CCT subunits by the ZNRD2-HERC2 complex(A) The CCT4-ZNRD2-RLD3 model was docked into the complete CCT complex structure (PDB: 7lum) by aligning the CCT4 chains. The CCT4-ZNRD2-RLD3 model is colored, as in Figure 5A. The non-CCT4 subunits within the top ring of the CCT complex are shown as cartoons in khaki. The CCT subunits of the bottom ring are shown in transparent surface representation.(B) Cartoon depicting a hypothetical partial CCT assembly comprising three subunits (colored in light coral, yellow, and cyan). One of the end subunits is compatible with binding the ZNRD2-HERC2 complex. The interior chamber is indicated for reference.(C) HEK293T cells transiently overexpressing the indicated 3xFLAG-tagged proteins were subjected to non-denaturing anti-FLAG IP. The bound proteins were eluted with 3xFLAG peptide and subsequently analyzed by immunoblotting for the indicated proteins relative to serial 2-fold dilutions of the cytosolic fraction (cytosol) prepared from control transfected cells. 20-fold less of the 3xFLAG-PhLP1 sample was loaded on the gel relative to the control and ZNRD2-3xFLAG (WT and L42E) samples. Two exposures of the blots are shown for all CCT subunits.(D) ZNRD2-KO cells stably overexpressing ZNRD2-3xFLAG (or untagged ZNRD2 as a negative control) were subjected to non-denaturing anti-FLAG IP. The bound proteins were eluted with 3xFLAG peptide and subjected to a second round of non-denaturing IP with anti-CCT4 antibody or control IgG. Samples were analyzed by immunoblotting as indicated.See also Figure S7.

Fig 3: Additional characterization of in vitro ubiquitination assays, related to Figure 4(A) 35S-methionine-labeled CCT4-3xFLAG was translated in a completely purified translation system (the PURE system; see method details) consisting of E. coli translation factors. The translation reaction was separated on a 5%–25% sucrose gradient and analyzed by SDS-PAGE and autoradiography. Soluble CCT4-3xFLAG migrating in fractions 3–5 (arrowheads) was recovered and used as the substrate for in vitro ubiquitination assays shown in Figure 4. Under the centrifugation conditions used, these fractions of the gradient would correspond to monomeric CCT4. A population of CCT4 fails to fold and is aggregated (fraction 11).(B) Purified WT 3xFLAG-HERC2 and its mutants (C4762S and ?RLD3) were tested for auto-ubiquitination by incubation with E1, E2 (UBCH5), His-Ub, and ATP, as indicated. The reactions were analyzed by anti-FLAG immunoblotting, either directly (input; bottom) or after His-Ub pull-down under denaturing conditions (His-Ub PD; top). Auto-ubiquitination was observed for HERC2 WT and ?RLD3, but not for catalytically inactive C4762S mutant. As expected, ubiquitination was dependent on E1 enzyme.

Fig 4: Recognition of orphan CCT subunits by ZNRD2, related to Figures 5 and 6(A) The CCT1 (top), CCT4 (middle), and CCT7 (bottom) reporters were assessed for degradation by flow cytometry in WT cells, ZNRD2-KO cells (transfected with empty vector), and ZNRD2-KO cells transiently re-expressing WT or mutant ZNRD2 constructs, as indicated.(B) Cells were transfected with CCT2-targeting siRNA together with or without ZNRD2-targeting siRNA, as indicated. The following day, cells were transfected with GFP together with or without the indicated ZNRD2 constructs and then cultured for another 2 days. After sorting GFP-positive cells to enrich the transfected cells, total cell lysates were prepared and analyzed by immunoblotting for the indicated proteins.(C) Interface between CCT4 and ZNRD2 in the predicted CCT4-ZNRD2-RLD3 model. CCT4 is shown as black cartoon with transparent surface representation colored by hydrophobicity. ZNRD2 is shown in blue cartoon representation. Highlighted are the residues on the hydrophobic patch in CCT4 that forms part of the predicted ZNRD2 interaction surface.(D) The indicated CCT4-TST constructs were translated in RRL, affinity-purified under native conditions, and analyzed by immunoblotting as indicated. Three amounts of total RRL were analyzed in parallel. The AAA mutant contains I77A, M81A, and V83A mutations.(E) The indicated CCT4 reporter constructs were expressed in WT cells and assessed for degradation by flow cytometry.(F) Cartoon representation of CCT4 in the predicted CCT4-ZNRD2-RLD3 model. The inset shows a close-up view highlighting the residues deleted or mutated in the misfolding CCT4 mutants used in (G) and (H).(G) The indicated CCT4-TST constructs were translated in RRL, affinity-purified under native conditions, and analyzed by immunoblotting as indicated. Three amounts of total RRL were analyzed in parallel.(H) The indicated CCT4 reporter constructs were assessed for degradation by flow cytometry in WT and ZNRD2-KO cells.(I) AlphaFold-predicted model of the CCT4-ZNRD2-RLD3 complex was docked onto an experimentally determined structure of the CCT complex (PDB: 7lum) by aligning CCT4 in the prediction with one of the two CCT4 chains in the experimental structure. CCT1 (light green), CCT2 (light yellow), and CCT4 (light coral) from the experimentally determined structure, and ZNRD2 (blue) from the predicted model, are shown in cartoon representation. A close-up view shows that the N-terminal tail of CCT1 (magenta) might clash with ZNRD2 if it were to remain in the position seen in fully assembled CCT. Given the minimal interactions made by this tail (and the adjacent C-terminal tail), we posit that they could easily move out of the way to permit ZNRD2 binding to CCT4, even when CCT4 is flanked by adjacent CCT subunits.

Fig 5: HERC2-dependent degradation of unassembled chaperonin subunits, related to Figure 2(A) CCT4 reporter degradation (bottom) was assessed by flow cytometry in HEK293T cells knocked down for HERC1 or HERC2. Total cell lysates were also analyzed by immunoblotting to confirm depletion of the target gene products (top).(B) WT and HERC2-KO Flp-In T-REx 293 cells were transiently transfected with the CCT4 reporter along with empty vector, 3xFLAG-HERC2 (WT), or catalytically inactive HERC2 (C4762S). CCT4 reporter degradation was assessed by flow cytometry (right), and total cell lysates were analyzed by immunoblotting as indicated (left).(C) HEK293T cells were co-transfected with the CCT4 reporter and empty vector, 3xFLAG-HERC2 (WT), or catalytically inactive HERC2 (C4762S). Cells were analyzed by flow cytometry (right), and total cell lysates were analyzed by immunoblotting (left).(D) WT and HERC2-KO Flp-In T-REx 293 cells were transiently transfected with the indicated reporter constructs and analyzed by flow cytometry. The data for CCT4 is from (B) and is shown again for the purpose of comparison to the other CCT subunits.(E and F) Flp-In T-REx 293 cells stably integrated with an inducible 3xFLAG-HERC2 construct (WT), its catalytically inactive mutant (C4762S), or empty vector as a negative control (-) were treated with doxycycline to induce the expression of each HERC2 construct. Cytosolic extracts were separated by SDS-PAGE (E) or native PAGE (F) and analyzed by immunoblotting as indicated. In (F), two exposures of the blots are shown for both CCT7 and CCT8.(G) MCF7 cells treated with either control or HERC2-targeting siRNAs were radiolabeled with 35S-methionine for 30 min and chased for 4 h. The pulse-labeled cells and pulse-chased cells were lysed under denaturing conditions and analyzed directly (input) or subjected to IP using anti-CCT7 antibody. As a specificity control, a pulse-labeled sample was subjected to IP without antibody (no Ab). Input and IP samples were separated by SDS-PAGE and visualized by autoradiography.