Domain-selective folding of the tandem β-propeller protein Coronin 7 (Coro7) by the chaperonin CCT/TRiC

TL;DR

This study shows that the chaperonin CCT/TRiC folds only the first β-propeller domain of the large Coronin 7 protein, revealing a new mechanism for domain-selective folding.

Contribution

The paper identifies Coronin 7 as a novel CCT/TRiC substrate and reveals domain-selective folding of tandem β-propeller proteins.

Findings

CCT/TRiC interacts only with the first β-propeller domain of Coronin 7.

CCT/TRiC preferentially binds to the first β-propeller domain regardless of its position in the protein.

CCT/TRiC recognizes β-propeller domains based on sequence rather than topology.

Abstract

The Chaperonin containing tailless complex polypeptide 1 (CCT) or TCP-1 ring complex (TRiC) plays a central role in maintaining cellular homeostasis by supporting protein folding and damping protein aggregation. Besides the abundant cytoskeletal proteins, actin and tubulin, CCT/TRiC is emerging as an obligate chaperone for the β-propeller domain of WD40 proteins. To date, only WD40 proteins consisting of a single β-propeller domain have been described as CCT/TRiC substrates. Using a combination of biotin proximity ligation, co-immunoprecipitation, and knockdown studies, we here identify the tandem β-propeller protein, Coronin 7 (Coro7), as a novel substrate of CCT/TRiC. This raised the question how CCT/TRiC can fold a protein that is too large to fit into its folding chamber, but consists of two domains that require its folding. Surprisingly, co-immunoprecipitation of truncated Coro7 proteins or cleaved full length Coro7 demonstrated that CCT/TRiC only interacts with the first β-propeller domain of Coro7. Further experiments showed that CCT/TRiC preferentially binds to the first β-propeller, independently of whether this domain is situated at the N- or C-terminus of Coro7. This strongly suggests that CCT/TRiC does not identify β-propeller substrates by their topology, but instead developed specific ways to recognize β-propeller sequences that require folding.

Figures

Full text

Correct protein folding is critical to maintain tissue homeostasis (1, 2). It ensures sufficient levels of proteins that are essential to sustain cellular function as well as prevents the excessive accumulation of misfolded and aggregation-prone proteins that could lead to the development of neuropathies and myopathies (3, 4, 5, 6). The process of protein folding is initiated as soon as proteins emerge as simple strings of amino acids from ribosomes. However, whereas many smaller proteins will spontaneously assume their correct conformation, larger proteins with more complex folds will require the help of professional molecular folding machines consisting of a combination of chaperones, co-chaperones and chaperonins (7, 8). The latter comprise a family of proteins that require ATP hydrolysis to fold proteins inside of an enclosed chamber. In case of the group II TCP-1 ring complex (TRiC) also called Chaperonin containing tailless complex polypeptide 1 (CCT), eight paralogous subunits (CCT1-CCT8) assemble into a 1 MDa hexadecameric double-ring structure (9, 10, 11, 12). Each subunit consists of an apical domain that participates in substrate recognition, an equatorial domain that mediates ATP-hydrolysis, and an intermediate domain that translates equatorial ATP hydrolysis into conformational changes of the apical domain (13, 14). As such, ATP hydrolysis following substrate binding results in repositioning of a flexible loop in the apical domain, thereby closing off the TRiC chamber and releasing the substrate in the chamber for assisted folding (12, 15). After ATP hydrolysis, the chamber opens and if the protein has achieved its proper fold and lost its contacts within the chamber, it is released.

The diverse amino acid composition of the CCT/TRiC apical domains as well as the N- and C-terminal tails of the various CCT subunits create a plethora of distinct hydrophilic and hydrophobic patterns that can recognize many different substrates. The most well-known CCT/TRiC substrates are the cytoskeletal proteins, actin and tubulin (16, 17, 18). In addition, CCT/TRiC seems to have a propensity for folding of WD40 or β-propeller domain proteins, including the heterotrimeric G-protein β-subunit (19, 20, 21), mTORC subunits mLST8 and Raptor (22) and the anaphase promoting complex proteins Cdc20 and Cdh1 (23, 24). A recent cryo-EM study was able to capture the stepwise folding of G-protein β5 (GNB5) inside the folding chamber of CCT/TRiC and elucidated the TRiC surfaces that guide this process (21). However, the different amino acid composition of CCT/TRiC β-propeller substrates calls into question whether they share a common CCT/TRiC recognition motif and whether the folding mechanism established for GNB5 is generally applicable to other CCT/TRiC β-propeller substrates. In addition, WD40 proteins in higher vertebrates can contain up to 4 β-propeller domains, whereas the folding chamber of CCT/TRiC has been estimated to only hold proteins of up to ∼70 kDa (25, 26) and thus can barely accommodate two β-propeller domains (∼40–50 kDa each). Given that no multi-β-propeller protein has been identified as a CCT/TRiC substrate to date, this raises the question whether mammalian CCT/TRiC is the designated chaperonin that mediates the folding of these complex proteins.

We here identify the tandem β-propeller Coronin 7 (Coro7) as the first multi-β-propeller substrate of CCT/TRiC. In addition, we show that CCT/TRiC does not interact equally with both β-propeller domains, but preferentially binds to one of the β-propellers. We further show that this interaction does not depend on the order of the β-propellers, suggesting that CCT/TRiC specifically recognizes one β-propeller over the other. As such, this study not only expands the number of CCT/TRiC substrates, but supports the hypothesis that CCT/TRiC rather recognizes substrate sequence than the topology of its obligate substrates. Future work will have to delineate the differences between β-propeller domains that can fold autonomously and β-propeller domains that require chaperone-mediated folding.

Results

Identification of CCT/TRiC as a Coro7 interacting protein by miniTurboID

The extensive β-propeller family of the Coronins plays a prominent role in the remodeling of actin networks and hence is essential for many biological processes ranging from cell migration and intracellular transport to T-cell immune response and wound healing (27, 28, 29). In contrast to yeast, mammals have six different Coronin genes, which show ubiquitous or tissue-specific expression. In addition, Coronins underwent a gene duplication event, which led to the appearance of the class of the tandem Coronins as early as Dictyostelium discoideum (30). As the name suggests, these Coronins are comprised of two consecutive Coronin β-propeller domains (Fig. 1A). They further lack the coiled-coil region present in the conventional Coronins and thus have been postulated to function as monomers rather than oligomers. Whereas the tandem Coronins from Dictyostelium, C. Elegans and D. Melanogaster, called CorB, and Pod-1 respectively, colocalize with F-actin, and are involved in actin-driven processes such as neurite outgrowth, polarity, endocytosis, and phagocytosis (31, 32, 33, 34, 35), it remains quite controversial whether these functions are conserved in mammalian Coro7. Instead, Coro7 seems to rather play a role as a scaffolding protein that supports the assembly of specific protein complexes. In addition, several genetic studies have identified Coro7 as a factor related to longevity, circadian rhythm and appetite (36, 37, 38), however it remains elusive how the scaffolding function of Coro7 contributes to regulating these processes.Figure 1Biotin Proximity Ligation and mass spec analysis identify CCT/TRiC and Coro7 as novel interactors. A, domain map of V5-mTurbo-NES cytoplasmic control and N-terminal V5-mTurbo tagged full-length Coro7 (V5-mTurbo-Coro7). The predicted tandem β-propeller structure of Coro7, generated using AlphaFold, is shown below the domain maps. The Western blot on the right was probed with horseradish peroxidase -labelled streptavidin and demonstrates the increase in biotinylated proteins observed after addition of exogenous biotin to cells expressing V5-mTurbo-Coro7. B, volcano plot showing all Coro7 interacting proteins identified by mass spec after proximity ligation using V5-mTurbo-Coro7 compared to V5-mTurbo-NES control. To identify significantly enriched proteins, peptide counts of each protein were Log_2_ transformed and compared to peptide abundance from the NES control group. Significant enrichment was gated at 1.5 Log_2_ fold change and a Log_2_p-value of 4.322 (p ≥ 0.05). Data was analyzed from three independent experiments. Statistical significance was determined using an unpaired two-tailed Student’s t test. C, Mrakov Clustering algorithm analysis of proteins with >Log_2_ fold change of 1.0 over the NES control. CCT, chaperonin containing tailless complex polypeptide 1; Coro7, coronin 7; TRiC, TCP-1 ring complex.

To identify novel complexes supported by Coro7, we built a Coro7 interactome using biotin proximity ligation in combination with mass spec analysis. We opted for proximity ligation based on the lack of discernable accumulation of Coro7 to specific organelles or vesicles (Fig. S1), which strongly suggests that most of its interactions are very dynamic and transient. To circumvent this problem, we used miniTurboID, which is a third generation biotin ligase that can rapidly and covalently attach a biotin molecule to proteins that are in close proximity of the protein of interest under cellular conditions (39). Potential interactors can then be specifically isolated using Streptavidin affinity purification and identified by a combination of mass spectrometry and proteomic analysis. To perform this experiment in a Coro7-free background, we first generated HEK293T Coro7 CRISPR/Cas9 knockout cells (HEK293T Coro7^−/−^) (See Materials and Methods). In addition, we developed an N-terminal miniTurboID fusion of human Coro7 (V5-mTurbo-Coro7) as well as a construct consisting of miniTurboID fused to tandem nuclear export signal motifs (V5-mTurbo-NES) to control for promiscuous interactions with highly abundant cytosolic proteins (Fig. 1A). Before performing the actual experiment, we optimized transient expression of these fusion proteins in HEK293T Coro7^−/−^ cells to obtain equivalent expression of V5-mTurbo-NES and V5-mTurbo-Coro7 as well as determined the optimal incubation time with biotin to obtain good labeling (Figs. 1A and S2). Using this optimized protocol, we generated three independent samples for each fusion protein and analyzed the biotinylated proteins by quantitative label-free mass spectrometry (Fig. S2). Potential Coro7 interacting proteins were identified by calculating the enrichment of individual proteins, which was obtained by comparing their relative abundance in the V5-mTurbo-Coro7 samples to the V5-mTurbo-NES control samples (40).

As expected, we observed considerable self-labeling of V5-mTurbo-Coro7. In line with the controversial role of Coro7 as an actin-regulatory protein, GO-analysis did not show enrichment of actin or actin-binding proteins. Instead, we identified a small cluster of proteins that are linked to the cell cycle and centrosome maintenance (Fig. 1C). Common to these potential interactors and the processes they are involved in is that they are all placed near the centrosome. Whereas this is in line with previous studies that reported perinuclear localization of endogenous Coro7 (41, 42), we mostly observed diffuse cytoplasmic staining of endogenous and recombinant Coro7 and only occasionally detected cells with perinuclear enrichment of Coro7 (Fig. S1). Together with their lower enrichment score (less than 2-fold), we decided to not follow up on these targets any further. Instead, we focused on clusters of proteins that all showed at least a 4-fold enrichment (Log_2_ fold change >2, p < 0.05). Interestingly, this was observed for all eight subunits of the CCT/TRiC complex (Fig. 1, B and C), which is the obligate chaperonin complex that folds actin and tubulin as well as several well-known β-propeller proteins. This raised the question whether Coro7 is a novel CCT/TRiC substrate or whether Coro7 perhaps functions as a scaffold that supports the assembly and function of CCT/TRiC, thereby indirectly regulating the cytoskeleton through folding of actin and tubulin by CCT/TRiC.

Coro7 is a novel CCT/TRiC substrate

To further characterize the relation between Coro7 and CCT/TRiC, we first confirmed that recombinant and endogenous Coro7 interact with CCT/TRiC. First, transiently expressed enhanced green fluorescent protein (EGFP) or EGFP-Coro7 was pulled down from HEK293T cell lysates using EGFP-nanobody coated beads. When the resins were probed for the CCT2 and TCP1α subunits of CCT/TRiC by Western blot, we only observed a CCT/TRiC signal for beads carrying EGFP-Coro7, but not for beads decorated with EGFP (Fig. 2A). To test binding under endogenous conditions, Coro7 was immunodepleted from HEK293T cell lysates and analyzed for bound CCT/TRiC by Western blot. As shown in Figure 2B, only beads coated with an antibody directed against Coro7 displayed a signal for the CCT5 or TCP1α (CCT1) subunit of CCT/TRiC, whereas beads coated with rabbit immunoglubulins (RigG) did not. Given that the Coro7 and CCT2 antibodies were both raised in rabbit and that the size of the individual CCT subunits overlaps with that of the antibody heavy chain, we were unable to also probe for CCT2 in our endogenous Coro7 co-immunoprecipitation. To overcome this hurdle, we examined direct interaction of Coro7 and CCT/TRiC by BN-PAGE and sucrose gradient fractionation. Analysis of HEK293T lysates by BN-PAGE showed both a lower Coro7 band as well as a higher Coro7 band that co-migrated with CCT2 (Fig. 2C). For sucrose gradient analysis, HEK293T cell lysates were separated by centrifugation on a 5 to 40% sucrose gradient and collected as fractions spanning 3% sucrose increments. As expected, the majority of the CCT/TRiC complex, detected by using CCT2 as a proxy, was found in the higher density fractions (Fig. 2D). In addition, we observed a smaller amount of CCT2 signal in the lower density fractions, which is in line with the fact that CCT2 can also exist as a monomer (43, 44). When analyzing Coro7, we found that most of the protein was retained in the lower density fractions, however a faint, but clear band could also be seen in the higher density fractions (Fig. 2D). The distinct separation between these Coro7 populations and the lack of a Coro7 signal in the highest density fractions further indicated that the second Coro7 band does not correspond to aggregated Coro7. Together, these results from sucrose gradient fractionation and BN-PAGE analysis strongly support the idea that the higher density Coro7 band is part of a Coro7-CCT/TRiC supercomplex. Altogether, we conclude from these experiments that endogenous Coro7, like recombinant EGFP-Coro7, not only interacts with CCT1, CCT2 and CCT5, but can bind to the CCT/TRiC complex as a whole.Figure 2Coro7 is a substrate of CCT/TRiC. A, interaction between Coro7 and CCT/TRiC was confirmed by coimmunoprecipitation of CCT/TRiC subunits, CCT2 and TCP1α, with recombinant EGFP or EGFP-Coro7. B, coimmunoprecipitation of CCT/TRiC subunits, CCT5 and TCP1α, with anti-Coro7 or rabbit immunoglobulins as control. C, BN-PAGE analysis of HEK293T lysates. Western blot shows a lower Coro7 band and a higher Coro7 band that co-migrates with CCT2. D, Western blot showing Coro7 and CCT/TRiC after sucrose gradient fractionation (5–40%) of HEK293T cell lysate. Coro7 was present in two distinct populations in fractions 3 to 5 and fraction 13, whereas CCT/TRiC was mostly detected in higher sucrose fractions. Each fraction corresponds to a 3% increase in sucrose concentration. E, Western blot showing transient knockdown of CCT2 in MDA-MB-231 cells. Graphs show densitometry analysis of CCT2, CCT5 and Coro7 expression from scrambled and siCCT2 treated MDA-MB-231 cells. Data was analyzed from three independent biological experiments consisting of three technical replicates for each condition (N = 3 for CCT2 and Coro7, N = 2 for CCT5, n = 3). Technical replicates are indicated by similar colored data points. Statistical significance was determined using a Welch’s t test. Approximate p-values are shown in each graph. F, co-immunoprecipitation of CCT2 with recombinant EGFP-tagged Coro7 following incubation of IPs with 5 mM ATP. The graph shows the densitometry analysis of CCT2 precipitated by EGFP-Coro7 with and without exogenous ATP. Data was analyzed from three independent biological experiments. Statistical significance was determined using a Welch’s t test. Exact p-values are shown in the graph. CCT, chaperonin containing tailless complex polypeptide 1; Coro7, coronin 7; EGFP, enhanced green fluorescent protein; TRiC, TCP-1 ring complex.

Having established that CCT/TRiC and Coro7 are binding partners, we next investigated the biological role of this interaction. To examine whether Coro7 is a novel substrate of CCT/TRiC and requires the chaperone for its folding, we examined the effect of CCT/TRiC depletion on the protein expression level of Coro7. As it was previously shown that depletion of one subunit is sufficient to disrupt the formation of the CCT/TRiC complex, we worked with a mix of siRNAs directed against the CCT2 subunit, whereas control cells were transfected with a non-targeting scrambled siRNA (SCR). Western blot analysis showed a complete loss of CCT2 in CCT2 siRNA treated cells compared to control cells (Fig. 2E). In addition, these cells also showed a lower CCT5 signal, suggesting that the CCT/TRiC complex is indeed disrupted (Fig. 2E). Knockdown of CCT/TRiC further decreased Coro7 protein levels by about 50% (p < 0.0001) (Fig. 2E), suggesting that Coro7 folding is disrupted. As fully folded substrates are released from CCT/TRiC in an ATP-dependent manner, we additionally measured the interaction between Coro7 and CCT/TRiC in the presence of ATP. For this purpose, EGFP-Coro7 was immunoprecipitated from HEK293T cell lysates and incubated with or without excess ATP (5 mM). Subsequent analysis of bound CCT/TRiC by Western blot showed about 60% less CCT2 signal after incubation with ATP (Fig. 2F). Combined, the decrease in cellular Coro7 protein levels after depletion of CCT/TRiC and reduction of the Coro7-CCT/TRiC interaction by ATP suggest that Coro7 is indeed a novel CCT/TRiC substrate.

Only the first β-propeller of Coro7 interacts with CCT/TRiC

In contrast to other CCT/TRiC β-propeller substrates identified to date, Coro7 consists of two consecutive β-propeller domains and hence is also larger than the 70 kDa that can be accommodated by the CCT/TRiC folding chamber. This raised the question whether CCT/TRiC perhaps folds each β-propeller sequentially or whether CCT/TRiC has a different mechanism to induce the folding of either β-propeller domain. To address this, we generated a series of truncated Coro7 proteins. Due to the lack of a structural model of Coro7, truncation constructs were developed based on multiple sequence alignment of Coro7 with the canonical Coronins (Coro1-6) and by the predicted Coro7 structure in Alphafold (Fig. 1A). This enabled us to not only delineate both β-propeller domains, but also demonstrated that the proline-serine rich linker between the β-propellers as well as the C-terminal region of Coro7 are largely unstructured. Interestingly, the very C-terminal end of this sequence is not found in any of the other Coronins except for yeast Crn1, where it was recently shown to be an intrinsically disordered region (IDR) that regulates the function of Crn1 (45). In addition, the Crn1 IDR was shown to also contain a CA-like motif that regulates Arp2/3-mediated actin filament branching (46). In case of Coro7, this motif was found to be required for Arp2/3 regulation by human Coro7, but not by C. Elegans Pod-1 (34, 47). As analysis of the human Coro7 sequence by Metapredict (48) demonstrated that this C-terminal region does indeed classify as an IDR (Fig. 3B) and given that this region is conserved in Coro7 proteins from different species where this region does not necessarily function as a CA-like motif, we will refer to it as IDR.Figure 3CCT/TRiC interacts with the first propeller of Coro7 in the absence of other domains. A, domain maps of truncated Coro7 constructs. Amino acid residues are indicated for each construct. B, Metapredict v3.0 prediction of intrinsically disordered regions of full-length Coro7. Regions with disorder scores above 0.5 are predicted to be disordered with 50% confidence. The pLDDT score reflects the confidence of AlphaFold2 in the local structure prediction. C, interaction between CCT/TRiC and different Coro7 domains was examined by coimmunoprecipitation of CCT2 with recombinant EGFP or EGFP-tagged Coro7 truncation constructs. D, Densitometry analysis of CCT2 precipitated by the indicated truncated Coro7 protein. Data was analyzed from three independent experiments. Statistical significance was determined using a Welch’s t test comparing each truncated Coro7 with full length Coro7. Exact p-values are reported with exception of p > 0.05, which is considered not significant (ns). E, interaction between CCT/TRiC and the IDR of Coro7 was examined by co-immunoprecipiation of CCT2 with either EGFP or EGFP-tagged PropB-ΔIDR or IDR of Coro7. F, densitometry analysis of CCT2 precipitated by the indicated Coro7 truncations. Data was analyzed from three independent experiments. Statistical significance was determined using a Welch’s t test comparing PropB-ΔIDR and IDR to full length Coro7. Exact p-values are shown in the graph. CCT, chaperonin containing tailless complex polypeptide 1; Coro7, coronin 7; EGFP, enhanced green fluorescent protein; IDR, intrinsically disordered region; TRiC, TCP-1 ring complex.

Based on this information, we generated EGFP-tagged Coro7 truncation constructs consisting of either the first β-propeller (PropA), the second β-propeller together with the IDR (PropB) or both propellers without the C-terminal IDR (ΔIDR) (Fig. 3A). After transient expression, truncated Coro7 proteins were immunoprecipitated from HEK239 T cells using an EFGP-nanobody and analyzed by Western blot for bound CCT/TRiC. Surprisingly, we found that PropA and Coro7-ΔIDR pulled down CCT/TRiC, whereas PropB barely showed a CCT2 signal compared to the EGFP only control (Fig. 3C). On top of that, densitometric analysis further showed that Coro7-ΔIDR precipitated significantly more CCT/TRiC than full length Coro7 (Fig. 3D). To exclude that the lack of interaction of PropB with CCT/TRiC is due to the presence of the IDR, we generated a PropB plasmid without the IDR (Fig. 3A). Additionally, we generated a GFP-tagged IDR (Fig. 3A) to test whether the IDR directly binds to CCT/TRiC. However, transient expression of these plasmids in HEK293T cells followed by immunoprecipitation and Western blot analysis did not show any binding of PropB or IDR by itself to CCT/TRiC (Fig. 3, E and F). Altogether, these results unexpectedly elucidated that CCT/TRiC is only required for folding of the first β-propeller domain and that there might be an additional role for the C-terminal IDR to efficiently release Coro7 (PropA) from the CCT/TRiC folding chamber.

To examine whether this interaction pattern is similar for full length Coro7, we leveraged the short working distance (within 10 nm of the bait protein) of the miniTurboID tag and the large size of Coro7 to compare the interactome of PropA and PropB. For this purpose, biotin proximity ligation was repeated with an N-terminal as well as a C-terminal miniTurboID fusion of Coro7, referred to as V5-mTurbo-Coro7 and V5-Coro7-mTurbo, respectively (Fig. 4A). Next, we analyzed the enrichment of the CCT/TRiC subunits in either sample compared to the V5-mTurbo-NES control. CCT1-CCT8 were found in both the N- and C-terminal biotin ligation reaction, however whereas each subunit showed increased binding for V5-mTurbo-Coro7 compared to the control sample (Fig. 1B), only CCT1 and CCT3 were slightly enriched for interaction with V5-Coro7-mTurbo, while CCT2, CCT5 and CCT6 binding was downregulated (Fig. 4B, data points shown in red). This difference became even more clear when directly plotting the Log_2_ fold enrichment to the NES control of each CCT subunit in the N-terminal versus the C-terminal Coro7 interactome, which showed lower interaction of each subunit, except CCT3, for the C-terminal than the N-terminal miniTurboID Coro7 fusion (Fig. 4C). Proximity ligation analysis thus supports the observation that CCT/TRiC mainly interacts with the first β-propeller.Figure 4Domain-selectivity of CCT/TRiC for PropA is conserved in full length Coro7. A, domain maps of constructs used for N-and C-terminal proximity ligation of Coro7. B, volcano plot showing all Coro7 interacting proteins identified by mass spec after proximity ligation using V5-Coro7-mTurbo compared to the V5-mTurbo-NES control. Significantly enriched proteins were identified as in Figure 1B. CCT/TRiC subunits (CCT1-8) are indicated in red. Data was analyzed from three independent experiments. Statistical significance was determined using an unpaired two-tailed Student’s t test. C, comparison of Log_2_ fold enrichment to the NES control of CCT/TRiC subunits in the N-terminal versus the C-terminal Coro7 interactomes. D, schematic overview of EGFP-Coro7-PreScission-Protease cleavage and subsequent coimmunoprecipitation of CCT/TRiC using either EGFP-nanobody or anti-Coro7 coated resin. E, overview of sample acquisition workflow. One 10 cm dish of HEK293T Coro7 KO cells transfected with either pEGFP-C1 or pEGFP-Coro7-PreScission Protease is split into 4 samples and subjected to the outlined treatment. F, representative Western blot of three independent experiments showing co-immunoprecipitation of CCT/TRiC subunit CCT5 with EGFP and Coro7 in the presence and absence of PreScission protease. G, densitometry analysis of CCT5 precipitated by PropA (EGFP-IP) or PropB (Coro7-IP) after PreScission Protease cleavage. Data was analyzed from 3 independent experiments. Statistical significance was determined using a Welch’s t test. Exact p-values are shown on the graph. CCT, chaperonin containing tailless complex polypeptide 1; Coro7, coronin 7; EGFP, enhanced green fluorescent protein; TRiC, TCP-1 ring complex.

In addition, we generated full length EGFP-Coro7 with a PreScission protease cleavage site introduced in the flexible linker between the β-propeller domains to more directly assess CCT/TRiC binding to either propeller. For this purpose, co-immunoprecipitation was performed using beads coated with either EGFP-nanobody or an antibody raised against the C-terminal 50 amino acids of Coro7 to specifically pull down the first or second β-propeller domain (Fig. 4, D and E). In the absence of PreScission protease, only intact Coro7 was detected and a signal for CCT5 was observed after co-immunoprecipitation with either pulldown strategy. By contrast, in the presence of the protease, no signal was observed for intact Coro7 and bands corresponding to either EGFP-PropA or PropB could be detected instead (Fig. 4F). In line with our previous results, CCT5 only significantly co-immunoprecipitated with EGFP-PropA, but not PropB (Fig. 4G).

To exclude that PropA is bound by CCT/TRiC because it is the first complete set of β-strands of Coro7 to be generated upon translation, we decided to engineer a Coro7 protein in which the order of the β-propellers is swapped (Coro7-PropSwap) as well as designed Coro7 proteins consisting of two PropA (Coro7-AA) or PropB (Coro7-BB) domains (Fig. 5A). Alphafold analysis of these mutant Coro7 β-propeller proteins predicts them to fold normally into a tandem β-propeller structure (Fig. S3). In line with this, these Coro7 mutants showed a similar cellular localization and expression level as wild type Coro7 (Fig. 5, B and C). Analysis of these mutant Coro7 proteins by co-immunoprecipitation further demonstrated binding of CCT/TRiC with Coro7-PropSwap and Coro7-AA, but not Coro7-BB (Fig. 5, C and D). Like PropA by itself, both Coro7-PS and Coro7-AA showed stronger binding to CCT/TRiC than WT Coro7. Altogether, this supports the idea that only PropA of Coro7 interacts with CCT/TRiC, suggesting that CCT/TRiC specifically recognizes PropA over PropB.Figure 5CCT/TRiC binds PropA of Coro7 regardless of its N- or C-terminal position. A, domain maps of engineered Coro7 constructs. B, confocal live-cell images of ARPE19 cells transiently expressing BFP-LifeAct and EGFP-Coro7 PropSwap, AA, or BB. Scale bar 10 μm. C, representative Western blot of co-immunoprecipitation of EGFP or EGFP-Coro7 wild type and engineered constructs (PropSwap, AA, BB) and CCT2. D, densitometry of CCT2 precipitated by engineered Coro7 constructs. Data was analyzed from four independent biological experiments. Statistical significance was determined using a Welch’s t test comparing each construct to wild type Coro7. Exact p-values are shown on the graph. CCT, chaperonin containing tailless complex polypeptide 1; Coro7, coronin 7; EGFP, enhanced green fluorescent protein; TRiC, TCP-1 ring complex.

Discussion

The mammalian chaperonin, TRiC, has been estimated to fold >10% of the cytosolic proteome. This is mainly accounted for by its indispensable role in the folding of the abundant cytoskeletal proteins, actin and tubulin. In addition, CCT/TRiC is required for the folding of a growing list of proteins with a WD40 or β-propeller domain. Using a combination of biotin proximity ligation. mass spec analysis, coimmunoprecipitation and siRNA knockdown, we show here that the tandem β-propeller protein, Coro7, is a novel substrate of CCT/TRiC (Figure 1, Figure 2, Figure 3), elucidating for the first time that CCT/TRiC also participates in the folding of WD40 proteins that consist of multiple β-propeller domains.

β-propeller domains are generally comprised of 28 β-strands that arrange into seven blades arranged around a central axis like the blades of a propeller (49). β-propeller domains further display another unique feature where the N-terminus folds back to complete the final β-sheet of the last blade that closes the propeller, referred to as the molecular clasp or velcro (19, 50). It has been postulated that this specific structural organization is the reason why β-propeller proteins require CCT/TRiC to complete folding. In line with this, a recent cryo-EM study was able to capture snapshots of the folding trajectory of GNB5 by CCT/TRiC (21) and elucidated that CCT/TRiC actively directs the folding of the β-propeller structure by establishing specific intramolecular interactions that drive folding from the middle blade (blade 4) fanning out to the outer blades (blades 3 and 5, then blades 2, 6 and 7) to finally induce the first blade that will close the propeller. This is different in proteins that consist of multiple β-propeller domains, where the N-terminus folds back to complete the last blade in the neighboring propeller rather than close the first propeller and additional stabilization is obtained by interaction with other domains or by atypical blades (51, 52, 53, 54, 55). Our data suggests that PropA does not complete PropB as either propeller can be precipitated individually after PreScission protease cleavage (Fig. 4, F and G). This suggests that PropA assumes the CCT/TRiC-mediated β-propeller velcro-fold, whereas PropB might not show the same closed structure. This is quite surprising as Coro7 has been postulated to have evolved by gene duplication of the canonical Coronins, which display the typical β-propeller velcro-fold (50).

Given that some of the Coro7 targets identified by C-terminal proximity ligation are WD40 proteins, it is tempting to speculate that PropB is stabilized by a β-propeller domain from another WD40 protein. This is particularly interesting as the canonical Coronins (Coro1-6) consist of only one β-propeller domain, but have an additional coiled-coil domain that enables them to form trimers (56, 57, 58). What is more, Coronins need to form oligomers in order to bind to actin filaments and exert their cellular functions (59). As Coro7 already has two β-propeller domains and lacks a coiled-coil domain, it is possible that it fulfills this requirement by interacting with another WD40 protein. Taken together with the results obtained after deletion of the C-terminal IDR (Fig. 3), we further speculate that this additional Coro7 partner might be recruited by binding to the IDR and might help to release Coro7 from CCT/TRiC. As we detected WD40 proteins involved in different processes, this might explain the diffuse cytoplasmic localization of recombinantly expressed Coro7 in contrast to the reported Golgi localization of endogenous Coro7 (41, 42). Going forward, it will be interesting to test whether WD40 binding partners can induce localization of Coro7 to the Golgi or even increase colocalization with F-actin.

Our observation that only one β-propeller of Coro7 interacts with CCT/TRiC is further in line with previous work showing that CCT/TRiC handles substrates that do not fit its folding chamber by only enclosing the domain of the protein that requires assisted folding (60). In our study, this is made possible by the presence of a long, flexible linker that allows PropA to be folded by CCT/TRiC, unaffected by the presence of a second β-propeller domain. It does raise the question how CCT/TRiC recognizes β-propeller domains that need support during folding in contrast to β-propeller domains that can fold autonomously or with the help of other chaperones. Using G-protein β as a model protein, Kubota and colleagues (61) showed that hydrophobic residues in β-sheets are critical for recognition by CCT/TRiC. Detailed structural information on the folding trajectory of GNB5 expanded upon this observation by showing that folding is initiated by contact between an acidic patch on CCT5 and a prominent positively charged residue on GNB5 (21). By contrast, ribosome profiling of nascent WD40 domains suggested that CCT/TRiC rather recognizes folding intermediates like β-sheets than linear sequences (62). As Coro7 is assumed to have evolved through gene duplication, it might offer the unique possibility to determine which sequences in PropA versus PropB mediate interaction with CCT/TRiC and to begin to define the rules of substrate recognition by CCT/TRiC. That said, it should be noted that PropA and PropB only show 27% sequence similarity, precluding in silico prediction of a possible CCT/TRiC interaction surface in PropA.

Finally, our initial intent was to identify new interaction partners of Coro7 to gain insight into its enigmatic correlation with longevity, appetite and regulation of circadian rhythm. Instead, we found Coro7 to be a novel substrate of CCT/TRiC, whose expression and activity have been reported to decline with aging (7), leading to disruption of proteostasis and build-up of protein aggregates. The decrease in Coro7 expression upon transient depletion of CCT/TRiC (Fig. 2E) without concomitant decline in tubulin or actin levels (not shown), suggests that Coro7 and its potential WD40 interaction partners might be among the first CCT/TRiC substrates that get affected by even slightly reduced CCT/TRiC activity. As such, it will be interesting to determine if there is a correlation between CCT/TRiC and Coro7 expression in tissues of different ages and relate this to changes in metabolism, the onset of senescence and altered expression of circadian clock genes in cells derived from these tissues.

Experimental procedures

Primer sequences, CRISPR guides and reagents are listed in Table 1.Table 1. ReagentssgsgRNASequence HEK293T Coro7 Knockout5′ - ccatgaacgcttccagggtgtcc - 3′siRNASequence Scrambled (IDT)# 51–01–14–03 siCCT2 (IDT)GAUGGUGCCACUAUUCUAAAAAACA siCCT2 (IDT)UACUACCACGGUGAUAAGAUUUUUUGUPrimerSequencePlasmid F V5-Coro7- miniTurbo5′- agtagtgctagcaaccgcttcagggtgtccaagttcCoro7-V5-miniTurbo R V5-Coro7-miniTurbo5′- agtagtgctagcgtcccactcgtcctcgtccacCoro7-V5-miniTurbo F pEGFP-C1 Coro75′- agtagtctcgagctatgaaccgcttcagggtgtccpEGFP-C1 Coro7, pEGFP-C1 PropA, pEGFP-C1 ΔIDR R pEGFP-C1 Coro75′- agtagtgaattcgactagtcccactcgtcctcgtccapEGFP-C1 Coro7, pEGFP-C1 PropBpEGFP-C1-IDR R pEGFP-C1 ΔIDR5′- agtagtgaattcttacaccatggcattcagcagctccpEGFP-C1 ΔIDRpEGFP-C1 PropB- ΔIDR R pEGFP-C1 PropA5′- agtagtgaattcctagcttgcgtctgcatcacccacpEGFP-C1 PropA F pEGFP-C1 PropB5′- agtagtctcgagctagtttgaggtcgctgcagagcpEGFP-C1 PropBpEGFP-C1 PropB- ΔIDR F pEGFP-C1 IDR5′-agtagtctcgaggcaaaactggggaaccgggagpEGFP-C1 IDR F pEGFP-C1 Coro7 PreScission Protease5′- ttccagggccccagttccaagttcpEGFP-C1 Coro7 PreScission Protease R pEGFP-C1 Coro7 PreScission Protease5′- caacacctccagcaggctctgcagcgacpEGFP-C1 Coro7 PreScission Protease F pEGFP-C1 Coro7 PropSwap PropB5′ - ctcagatctcgagctagttccaagttccgccatgctcaggpEGFP-C1-Coro7 PropSwap R pEGFP-C1 Coro7 PropSwap PropB5′ - gaactccacagcctttcggggcagccggaaggcpEGFP-C1-Coro7 PropSwap F pEGFP-C1 Coro7 PropSwap PropA5′ - ttgttccagggccccatgaaccgcttcagggtgtccpEGFP-C1-Coro7 PropSwap R pEGFP-C1 Coro7 PropSwap PropA5′ - gaactctttccggacgcggggcacatggtagccpEGFP-C1-Coro7 PropSwap F pEGFP-C1 Coro7 PropSwap Linker5′ - ttccggctgccccgaaaggctgtggagttccacpEGFP-C1-Coro7 PropSwap R pEGFP-C1 Coro7 PropSwap Linker5′ - cctgaagcggttcatggggccctggaacaacacpEGFP-C1-Coro7 PropSwap F pEGFP-C1 Coro7 PropSwap Vector5′ - taccatgtgccccgcgtccggaaagagttcttcpEGFP-C1-Coro7 PropSwap R pEGFP-C1 Coro7 PropSwap Vector5′ - gcggaacttggaactagctcgagatctgagtccpEGFP-C1-Coro7 PropSwap F pEGFP-C1 Coro7 AA PropA-Linker5′ - ctcagatctcgagctatgaaccgcttcagggtgpEGFP-C1-Coro7 AA R pEGFP-C1 Coro7 AA PropA1-Linker5′ - cctgaagcggttcatggggccctggaacaacacpEGFP-C1-Coro7 AA F pEGFP-C1 Coro7 AA PropA25′ - ttgttccagggccccatgaaccgcttcagggtgtccpEGFP-C1-Coro7 AA R pEGFP-C1 Coro7 AA PropA25′ - gaactctttccggacgcggggcacatggtagccpEGFP-C1-Coro7 AA F pEGFP-C1 Coro7 AA Vector5′ - taccatgtgccccgcgtccggaaagagttcttcpEGFP-C1-Coro7 AA R pEGFP-C1 Coro7 AA Vector5′ - cctgaagcggttcatagctcgagatctgagtccpEGFP-C1-Coro7 AA F pEGFP-C1 Coro7 BB PropB1ctcagatctcgagctggccccagttccaagttccgcpEGFP-C1-Coro7 BB R pEGFP-C1 Coro7 BB PropB1gaactccacagcctttcggggcagccggaaggcpEGFP-C1-Coro7 BB F pEGFP-C1 Coro7 BB Linker-PropB2ttccggctgccccgaaaggctgtggagttccacgaggacpEGFP-C1-Coro7 BB R pEGFP-C1 Coro7 BB Linker-PropB2gaactctttccggactcggggcagccggaaggcpEGFP-C1-Coro7 BB F pEGFP-C1 Coro7 BB VectorttccggctgccccgagtccggaaagagttcttcpEGFP-C1-Coro7 BB R pEGFP-C1 Coro7 BB VectorcttggaactggggccagctcgagatctgagtccpEGFP-C1-Coro7 BBAntibodyCompany & CatalogDilutionSpeciesCoro7Novus #NBP3-184511:1000 (WB); 1:200 (IF)RabbitCoro7Rybakin et al., 2004N/A (IF); 1:100 (WB)MouseGFPSanta Cruz #sc-99961:1000 (WB)MouseTCP1αSanta Cruz #sc-534541:1000 (WB)RatCCT2CST #3561S1:1000 (WB)RabbitCCT5Novus #NBP2-436801:1000 (WB)MouseV5-tagCST #13202S1:1000 (WB)Rabbitα/β-TubulinCST #214851:1000 (WB)RabbitStreptavidin-HRPLeinco #S2071:1000 (WB)N/AGAPDHSanta Cruz #sc-3650621:1000 (WB)MouseRabbit IgG ControlInvitrogen #02–6102N/ARabbitAnti-mouse IgG-HRPCST #7076S1:10,000 (WB)HorseAnti-rabbit IgG-HRPCST #7074S1:10,000 (WB)HorseAnti-rat IgG-HRPCST #7077S1:10,000 (WB)GoatGoat-anti-rabbit, Oregon Green 488ThermoFisher #O110381:500 (IF)GoatDonkey-anti-mouse AlexaFluor 488ThermoFisher # A212021:500 (IF)DonkeyGoat anti-mouse IgG Secondary antibody, DyLight 680ThermoFisher #355181:20,000GoatGoat anti-rabbit IgG Secondary antibody, DyLight 800Invitrogen #SA5-355711:20,000Goat

Plasmid Cloning

A complete ORF encoding for human Coro7 was obtained from the DNASU plasmid repository (pDNR227) and cloned into V5-miniTurbo-NES (gift from David Kast). For V5-miniTurbo-Coro7, Coro7 was cut out of an existing construct with AgeI and XhoI and ligated into V5-miniTurbo-NES using KpnI and XhoI restriction sites, destroying the AgeI and KpnI sites after ligation. For Coro7-V5-miniTurbo, Coro7 was generated by PCR and cloned into the NheI restriction site. Insertion and correct orientation of Coro7 was determined by colony screen. EGFP-tagged Coro7 and Coro7 truncation constructs were generated by PCR and cloned into pEGFP-C1 (Clontech) using XhoI and EcoRI restriction sites. Insertion of a PreScission Protease recognition sequence was generated using Q5 mutagenesis. pEGFP-tagged Coro7 PropSwap, AA, and BB were generated using NEBuilder HiFi DNA Assembly. All constructs were confirmed by sequencing. Primers used to generate the plasmids for this study were obtained from IDT.

Mammalian cell culture

Authenticated HEK293T, MDA-MB231, and ARPE19 cells were obtained from ATCC. HEK293T and HEK293T Coro7 KO, MDA-MB231, and ARPE19 were cultured at 37 °C and 5% CO_2_ in Dulbecco modified Eagle’s medium (DMEM), high glucose, GlutaMax (Gibco) supplemented with 10% fetal bovine serum (Biowest #S1620) and 1% penicillin-streptomycin solution (Gibco). Cell lines were regularly tested for Mycoplasma contamination to ensure all experiments were performed with healthy cells.

Coro7 CRISPR/Cas9 KO cells

100 μM of tracrRNA (IDT) and crRNA (IDT) were mixed and heated at 95 °C for 5 min to create a tracrRNA-cRNA duplex. This duplex was cooled to room temperature and combined with Cas9 (IDT) in sterile PBS. The RNA and Cas9 were incubated together for 20 min at room temperature to form a ribonucleoprotein (RNP). To introduce the RNP to HEK293T cells were electroporated using a Lonza 4D Nucleofector. Briefly 2 x 10^5^ HEK293T WT cells were resuspended in 20uL of Nucleofector Solution SE. Next, 5 μl of RNP and 1 μl of Cas9 Electroporation Enhancer was then added to the resuspended cells. After mixing, the cells were then transferred to a Nucleocuvette strip and electroporated using the DG-130 program. After electroporation, cells were transferred to pre-warmed DMEM containing 10% fetal bovine serum. To generate monoclonal populations, electroporated HEK293Ts were seeded as single cells in 96-well plates, allowed to form colonies and then harvested for Next Generation Sequencing. Coro7 KO was additionally confirmed by western blotting.

Biotin proximity ligation and mass spec analysis

1.5 x 10^6^ HEK293T Coro7 KO cells were seeded onto 10 cm plates 1 day prior to transfection. To obtain similar levels of expression, HEK293T Coro7 KO cells were transfected with either 0.75 μg of V5-mTurbo-NES (cytoplasmic control) or 3 μg of V5-mTurbo-Coro7 plasmid using PEI at a 1:3 ratio (Polysciences). 48 h after transfection, cells were treated with 500 μM exogenous biotin for 10 min at 37 °C then lysed in 350 μl of 20 mM Tris-HCl pH 7.4, 150 mM NaCl, 1% Triton X-100, 1X protease inhibitor cocktail (APeXBIO) and 1 mM PMSF lysis buffer. Lysates were incubated on ice for 30 min with agitation every 10 min. Lysates were clarified by centrifugation at 21,000×g for 20 min and the supernatant was collected. 200 μl of Dynabeads MyOne Strepatvidin T1 beads (Thermo Fisher Scientifc) were equilibrated by washing three times in 500 μl of lysis buffer. Following equilibration, lysates were added to the equilibrated Dynabeads and incubated at room temperature for 2 h under gentle agitation. The beads were then washed twice with RIPA-LS (10 mM Tris-Cl pH 7.4, 1 mM EDTA, 140 mM NaCl, 0.1% SDS, 0.1% DOC, and 1% Triton X-100), once with 1M KCl, once with 1M Na_2_CO_3,_ and twice with 2M urea in 10 mM Tris-Cl pH 7.4. The final wash was removed, and the beads were resuspended in 30 μl of 3X Laemmli buffer. 5 μl of sample was separated by SDS-PAGE and stained with Coomassie to check for protein pulldown. The remaining 25 μl of sample was loaded onto a Mini-Protean TGX Stain-Free Gel, 4 to 20% (Bio-Rad) and run into the stacking gel. Following electrophoresis, the gel was briefly washed in water and samples were cut out of the gel in 1 cm × 0.5 cm pieces. Gel pieces were washed in water twice for 10 min each. Samples were processed for mass spectrometry by the Washington University in St Louis School of Medicine Mass Spectrometry Technology Access Center (MTAC). Protein fold change and p-values were calculated as previously described by Aguilan et al., 2020 (40). Briefly, raw spectrum counts for both the NES control and Coro7 experimental groups were subjected to a Log_2_ transformation to generate a normalized skew. Following transformation, the Log_2_ spectrum counts were first normalized by the average of all spectrum counts in each control group and further normalized by the distribution width or slope of each control group. Next, missing values were imputed to account for missing data due to detection limitations. Finally, the fold change of Log_2_ spectrum counts in the Coro7 experimental group was calculated compared to the average of the NES control spectrum counts. p-values were determined using a two-tailed, paired t test. Log_2_ fold changes were then plotted against the Log_2_ calculated p-values. A 1.5 Log_2_ fold change with a 4.322 Log_2_ p-value (0.05) were used as cutoffs for proteins to be considered significantly enriched. For Markov Clustering algorithm of enriched proteins (fold change > Log_2_ 1.0), protein-protein interactions were identified using STRING v12.0 (63) against the Homo sapiens protein database with a 0.400 confidence interaction score. Protein networks were then exported to Cytoscape v3.10.3 (64) where the network was clustered using Markov Clustering algorithm with a granularity parameter of 4.

Western Blotting and BN-PAGE

For western blotting experiments, 1.5 x 10^6^ HEK293T WT or KO cells were seeded onto 10 cm plates 1 day prior to transfection. Cells were transfected with indicated constructs (GFP, WT ΔIDR, PropA, PropB, PropSwap, AA, or BB) using PEI (Polysciences) at a 1:3 ratio. Transfections proceeded for 2 days prior to lysis. Cells were lysed in 20 mM Tris-HCl pH 7.4, 150 mM NaCl, 1% Triton X-100, 1X protease inhibitor cocktail (APeXBIO) and 1 mM PMSF. Lysates were incubated on ice for 30 min with agitation every 10 min, then clarified by centrifugation at 21,000×g for 20 min. The supernatant was collected and mixed with 3X Laemmli buffer and heated at 95 °C for 5 min. Samples were separated by 12% SDS-PAGE and proteins were transferred onto Immobilon-P polyvinylidene difluoride membranes (Millipore Sigma) in transfer buffer containing 10% ethanol. Transfers proceeded for 1 h at 100V. Membranes were allowed to dry and were blocked in 5% BSA in TBS plus 0.01% Tween-20 (TBS-T). Blots were incubated with indicated primary antibodies at 1:1000 dilution in TBS-T overnight at 4 °C with gentle agitation. Following primary incubation, blots were washed with TBS-T and incubated with horseradish peroxidase (HRP)-linked anti-mouse or anti-rabbit IgG secondary antibodies or anti-mouse DyLight 680 or anti-rabbit DyLight 800 (Thermo Fisher Scientific & Invitrogen) tagged secondaries for 1 h at room temperature with gentle agitation. Blots were washed with TBS-T and incubated in Clarity Western enhance chemiluminescence substrate (Bio-Rad) for at least 1 min before imaging. Blots were imaged using a BioRad ChemiDoc Imager or Odyssey M Imager.

For BN-PAGE analysis, cells were lysed in 10 mM Tris-HCl pH 7.4, 75 mM NaCl, 0.5% Triton X-100, 1X protease inhibitor cocktail (APeXBIO) and 1 mM PMSF. Lysates were incubated on ice for 30 min with agitation every 10 min, then clarified by centrifugation at 21,000×g for 15 min. The supernatant was collected and mixed with 3X native sample buffer (30% glycerol, 3X NativePAGE Running Buffer (Invitrogen), and 0.125% G-250) and 50 μg of total lysate was run on a NativePAGE 3 to 12% Bis-Tris gel (Invitrogen). Proteins were transferred onto Immobilon-P polyvinylidene difluoride membranes (Millipore Sigma) in transfer buffer containing 10% ethanol and 0.001% SDS. Membrane blocking and staining proceeded as described above.

Purification and conjugation of HIS-GFP nanobody

HIS-GFP nanobody was a gift from Dr Brett Collins at the University of Queensland in Brisbane Australia. HIS-GFP Nanobody cDNA was transformed into BL21 competent cells. Starter cultures were inoculated using single colonies and grown in 100 ml of LB at 37 °C at 215 rpm. The following day, 1 L TB cultures were inoculated and grown at 37 °C at 215 rpm for until the OD600 was ≥ 2, cooled to 18 °C and induced with 500 μM IPTG overnight. The following day, cultures were harvested by centrifugation at 3000×g for 10 min and lysed by microfluidization in 200 ml of 20 mM Tris-Cl pH7.4, 500 mM NaCl, 10 mM Imidazole, 4 mM Benzamidine, 1 mM PMSF, and 1X protease inhibitor cocktail (APeXBIO). Lysates were clarified by centrifugation at 12,000×g for 30 min. Supernatants were then passed over 4 ml bed volume of NiNTA equilibrated 3x in 20 mM Tris-Cl pH 7.4, 500 mM NaCl, and 10 mM Imidazole by gravity flow twice. The NiNTA was washed with 20CV of 20 mM Tris-Cl pH 7.4, 500 mM NaCl, and 10 mM Imidazole and bound protein was eluted in 20 mM Tris pH 7.4, 50 mM NaCl, and 250 mM Imidazole. Protein-containing fractions were determined by Coomassie stain, combined, and passed through a HiLoad 26/600 Superdex 75 pg size exclusion column equilibrated in 200 mM NaHCO_3_ and 500 mM NaCl. Proteins were eluted using an isocratic elution over 1CV at a flow rate of 1.5 ml/min. Peaks were checked by Coomassie stain and HIS-GFP Nanobody containing fractions were desalted to PBS using a PD-10 column (GE Healthcare). The HIS-GFP Nanobody was then diluted 2.5 mg/ml and added to 1 g of Pierce NHS-activated dry agarose (Thermo Fisher Scientific). NHS-nanobody conjugation proceeded overnight at 4 °C with gentle agitation. The following day, the NHS beads were washed 3 times with PBS and the reaction was quenched using 1M Tris-Cl pH 7.4 for 15 min.

GFP pulldowns

HEK293T Coro7 KO cells were transiently transfected with indicated GFP constructs using a 1:3 DNA to PEI ratio (Polysciences). Following transfection, cells were lysed in 20 mM Tris-HCl pH 7.4, 150 mM NaCl, 1% Triton X-100, 1X protease inhibitor cocktail (APeXBIO) and 1 mM PMSF. A 50 μl bed volume of GFP-Nanobody beads were equilibrated in lysis buffer. Lysates were passed over the GFP-Nanobody beads once and beads were subsequently washed three times in 20 mM Tris-HCl pH 7.4 and 150 mM NaCl. After the final wash, the GFP-Nanobody beads were pelleted and resuspended in 75 μl of 3X Laemmli sample buffer and heated to 95 °C. The GFP-nanobody beads were then pelleted by centrifugation at 21,000×g for 1 min and the supernatant was collected and used for western blotting. For experiments assessing ATP-induced release of Coro7 by CCT/TRiC (Fig. 2F), immunoprecipitation was performed as described above. After the final wash, GFP-nanobody beads were instead split in half and incubated with 5 mM ATP (GoldBio) for 4 h. The beads were washed twice in 20 mM Tris-HCl pH 7.4 and 150 mM NaCl. Following washes, samples were eluted in 3X Laemmli buffer and analyzed by western blotting.

Endogenous Coro7 Co-IPs

1.5 x 10^6^ HEK293T cells were seeded onto 10 cm dishes the day prior to Co-IP. Cells were lysed in 200 μl of 20 mM Tris-HCl pH 7.4, 150 mM NaCl, 1% Triton X-100, 1X protease inhibitor cocktail (APeXBIO) and 1 mM PMSF. Lysates were incubated on ice for 30 min with agitation every 10 min, then clarified by centrifugation at 21,000×g for 20 min. Following clarification, 2ug of either rabbit IgG or anti-Coro7 antibody were added to cell lysates and incubated at 4 °C for 1 h under gentle agitation. Protein-antibody complexes were isolated by the addition of a 25 μl bed volume of magnetic protein G beads (NEB) for 1 h under gentle agitation. The magnetic protein G beads were washed with 3x with 20 mM Tris-HCl, pH 7.4 and 150 mM NaCl and protein was eluted by the edition of 50 μl of 3X Laemmli buffer and heating at 95 °C for 5 min. The magnetic protein G beads were pelleted by centrifugation at 21,000×g for 1 min and the supernatant was collected and used for western blotting.

Sucrose gradients

1.5 x 10^6^ HEK293T cells were seeded onto 10 cm dishes the day prior to lysing. Cells were lysed in 350 μl of 20 mM Tris-HCl pH 7.4, 150 mM NaCl, 1% Triton, 1X protease inhibitor cocktail (APeXBIO) and 1 mM PMSF. Lysates were incubated on ice for 30 min with agitation every 10 min, then clarified by centrifugation at 21,000×g for 20 min. Sucrose gradients were formed by the mixing of 5% and 40% sucrose in 20 mM Tris-Cl, pH 7.4 and 150 mM NaCl. 11 ml gradients were deposited using a Labconco Auto Densi-Flow Gradient Fractionator. Lysates were gently layered on top of cold 5 to 40% sucrose gradients and centrifuged at 26,300 RPM at 4 °C for 16 h. Following separation, 0.5 ml fractions were taken and subsequently used for western blotting.

siRNA knockdown of CCT2

2.5 × 10^5^ MDA-MB-231 cells were seeded into a 6-well plate the day prior to siRNA transfection. MDA-MB-31 cells were transfected with either 20 nM of scrambled siRNA or siRNA targeting CCT2 (see table for siRNA sequences) using Lipofectamine RNAiMAX Transfection Reagent (Thermo Fisher Scientific) following manufacturer’s instructions. The following day, cell media was replaced with fresh complete media. siRNA knockdown proceeded for 72 h and cells were then lysed, and lysates were used for subsequent western blotting as previously described.

GFP-Coro7-PreScission Protease Cleavage and Co-IPs

1.5 x 10^6^ HEK293T Coro7 KO cells were seeded onto 10 cm dishes the day prior to transfection. Either an empty pEGFP-C1 vector or pEGFP-Coro7-PreScission-Protease (GFP-Coro7-PP) were transfected using PEI at a 1:3 DNA to PEI ratio (Polysciences). Following transfection, cells were lysed in 500 μl 20 mM Tris-HCl pH 7.4, 150 mM NaCl, 1% Triton X-100, 1X protease inhibitor cocktail (APeXBIO) and 1 mM PMSF. Lysates were incubated on ice for 30 minutes with agitation every 10 min. Lysates were clarified by centrifugation at 21,000×g for. 20 min. GFP-Coro7-PP was then cleaved by the addition of 50 μg of PreScission Protease for 1 h at 4 °C. Following cleavage, lysates were split in half with one half subjected to GFP pulldowns and the other half subjected to anti-Coro7 pulldowns as described above. Protein was eluted by the addition of 3X Laemmli sample buffer and heated to 95 °C and used for subsequent western blotting.

Immunofluorescence staining and (live cell) microscopy

For live cell imaging, 0.5 x 10^5^ MDA-MB-231 or ARPE19 cells were seeded onto glass bottom D35 imaging dishes. Next day, cells were transfected with GFP-Coro7 and BFP-LifeAct using Lipofectamine 3000 at a 1:4 ratio. For immunofluorescence, 0.5 x 10^5^ MDA-MB-231 or ARPE19 WT cells were seeded onto circular 12 mm #1.5 glass coverslips the day before fixation. Cells were briefly washed with warm PBS and fixed for 10 min in pre-warmed 4% PFA in PBS at room temperature. Following fixation, cells were washed in ice-cold PBS. Cells were then permeabilized by the addition of 0.3% Triton X-100 in PBS for 10 min at room temperature, followed by a PBS rinse. Coverslips were then blocked in 3% BSA in PBS + 0.1% Tween (PBS-T) for 1 h at room temperature. Primary antibodies were diluted in 3% BSA in PBS-T at indicated dilutions (see table) and incubated on coverslips for 1 h at room temperature. Following primary antibody incubation, coverslips were washed 3 times in PBS-T for 5 min each. Secondary antibodies and Alexa Fluor 647-Phallodin were diluted in PBS-T at indicated dilutions (see table) and incubated on coverslips for 1 h at room temperature in the dark. Coverslips were then washed 3 times in PBS-T for 5 min each. Finally, coverslips were mounted on to glass slides using ProLong Glass Antifade (Thermo Fisher Scientific) after briefly dipping in H_2_O to remove excess salts. Fixed slips and overexpressing live cells were imaged using a Nikon Ti2 inverted microscope with a x100 Plan-Apo oil immersion objective and a Yokogawa CSU-W1 spinning disk confocal attached to a Hamamatsu ORCA-FLASH4.0 CMOS camera. Images were at 16 bit in a 2048 x 2044 resolution with Z-stacks of 0.2 μm. Images were processed using Fiji (https://imagej.net/software/fiji/downloads) to create sum-intensity projections.

Statistical analysis, cartoons/schematics and structure prediction

Statistical analysis was performed using Prism 10.5.0 (htpps://prismjs.com). Each dataset was tested for normality to determine the correct analytical test. Cartoons and schematics in Figure 4 and Figure S2 were generated with BioRender. Alphafold structure predictions were accomplished using ColabFold v1.5.5: Alphafold2 using MMseqs2 (65) Default settings were used and the rank 1 model generated was imported into the Pymol 3.1.6.1 (https://colab.research.google.com/github/sokrypton/ColabFold/blob/main/AlphaFold2.ipybn) to highlight indicated domains.

Data availability

All data supporting the findings of this study can be found within this article and its supporting information. Raw MS/MS data can be accessed in the public ProteomeXchange repository (PDX073687).

Supporting information

This article contains supporting information.

Conflict of interest

The authors declare that they have no conflicts of interest with the contents of this article.