Molecular mechanisms of CAND2 in regulating SCF ubiquitin ligases

Wang, Kankan; Li, Lihong; Kenny, Sebastian; Gan, Dailin; Reitsma, Justin M.; Zhou, Yun; Das, Chittaranjan; Liu, Xing

doi:10.1038/s41467-025-57065-5

Download PDF

Article
Open access
Published: 26 February 2025

Molecular mechanisms of CAND2 in regulating SCF ubiquitin ligases

Nature Communications volume 16, Article number: 1998 (2025) Cite this article

4189 Accesses
1 Citations
1 Altmetric
Metrics details

Subjects

Abstract

Protein degradation orchestrated by SKP1·CUL1·F-box protein (SCF) ubiquitin ligases is a fundamental process essential for cellular and organismal function. The dynamic assembly of SCFs, facilitated by CAND1, ensures timely ubiquitination of diverse SCF target proteins. As a homolog of CAND1, CAND2 alone has been implicated in various human diseases, yet its functional mechanisms remain elusive. Here, we investigate the role of CAND2 in human cells and its distinct mode of action compared to CAND1. Using an array of quantitative assays, we demonstrate that CAND2 promotes SCF-mediated protein degradation as an F-box protein exchange factor. While CAND2 binds CUL1 with structure and affinity comparable to CAND1, it exhibits lower efficiency in exchanging F-box proteins. Kinetic measurements reveal a significantly higher K_M for CAND2-catalyzed SCF disassembly than CAND1, which explains the lower exchange efficiency of CAND2 and is likely due to conformations of the CAND2·SCF exchange intermediate complex being less favorable for F-box protein dissociation. Our study provides mechanistic insights into the biochemical and structural properties of CAND2, as well as its role in regulating cellular dynamics of SCFs, laying a foundation for understanding contributions of CAND2 to healthy and diseased human cells.

Cadmium binding by the F-box domain induces p97-mediated SCF complex disassembly to activate stress response programs

Article Open access 08 May 2024

CAND1 inhibits Cullin-2-RING ubiquitin ligases for enhanced substrate specificity

Article 04 January 2024

Mechanism of ASF1 engagement by CDAN1

Article Open access 16 March 2025

Introduction

As the largest family of ubiquitin E3 ligases in human cells, cullin-RING ligases (CRLs) play a vital role in catalyzing the transfer of ubiquitin from ubiquitin-conjugating enzymes (E2) to specific protein targets^1,2,3. The catalytic core of each CRL comprises a cullin scaffold, such as CUL1, and an E2-binding RING domain protein, such as RBX1. Each type of cullin-RING core associates with a unique class of substrate receptor modules, recruiting diverse substrates and enabling a broad yet specific range of substrate ubiquitination. Through the SKP1 adaptor protein, the CUL1·RBX1 core engages substrate receptor proteins characterized by an F-box motif, forming SKP1·CUL1·F-box protein (SCF) ubiquitin ligases^4,5,6. Up to 69 F-box proteins (FBPs) can be synthesized in human cells, some of which are well recognized for mediating the ubiquitination and degradation of key regulatory proteins⁷. Given the competition among various FBPs for access to CUL1, cells rely on the efficient and opportune assembly of diverse SCF complexes to dynamically handle a changing substrate pool.

The adaptive exchange mechanism, centered by the FBP exchange factor CAND1, enables SKP1·FBPs of diverse structures and cellular abundance to assemble active SCFs when their corresponding substrates emerge^8,9. When the U-shaped CAND1 binds CUL1 in an SCF, it creates a CAND1·CUL1·SKP1·FBP intermediate complex with an ensemble of conformations^8,10,11,12. In this complex, CAND1 clashes with SKP1, so either CAND1 or SKP1·FBP quickly dissociates to yield SCF or CAND1·CUL1, each of which by itself is a stable complex with a long half-life (t_1/2)^8,10. This feature of the CAND1-SCF interaction allows the rapid exchange of SKP1·FBPs on the common CUL1·RBX1 core, establishing a dynamic cellular repertoire of SCFs that ubiquitinate a wide spectrum of protein substrates in a timely manner^8,9. Through this molecular mechanism, CAND1 promotes SCF-mediated protein ubiquitination in vitro, in vivo, and ex vivo^{8,9,10,13,14,15,16,17,18,19,20}. The activity of CAND1 is regulated by NEDD8, a ubiquitin-like protein that modifies CUL1, triggers conformational changes that activate SCF and simultaneously prevents CAND1 from binding to CUL1²¹. Following the dissociation of CAND1 from CAND1·CUL1·SKP1·FBP, CUL1 is conjugated to NEDD8 (neddylation), thereby locking SCF in a stable and active state optimal for substrate ubiquitination. In the absence of a loaded substrate, the CSN deneddylase removes NEDD8, restoring SCF to a state that can be exchanged by CAND1^1,2,22,23,24. The cooperative effects of CAND1, NEDD8, and SCF substrates thus enable rapid cycling of CUL1·RBX1 throughout the entire pool of SKP1·FBPs, granting adaptive assembly of SCFs on demand for substrate ubiquitination^1,2,8. This adaptive exchange mechanism, powered by CAND1, is essential for the survival and fitness of cells and organisms^20,24,25.

With a sequence identity of 63% (Supplementary Fig. 1), CAND2 has been identified as a homolog of CAND1 in mammals^26,27. Multiple recent “omics” studies have revealed CAND2 as a risk factor for multiple pathological conditions, such as a few cardiovascular abnormalities^{28,29,30,31,32,33,34,35,36,37,38}, and upregulation of Cand2 by mTORC1 in mice was found to cause adverse cardiac remodeling³⁹. Unlike CAND1 which has been extensively studied for its role in regulating SCF, the biochemical basis through which CAND2 regulates SCF has yet to be defined.

Here, combining biochemical, biophysical, and cell biological approaches, we study the effect of CAND2 on SCF assembly and activity, the CAND2·CUL1 structure, and the dynamic interactions involving CAND2 and SCF. Our results reveal that CAND2 functions as an FBP exchange factor, akin to CAND1. However, it regulates SCFs with kinetic parameters and biochemical effects that differ from those associated with CAND1. These findings establish a foundation for understanding the regulatory roles of CAND2 in SCF-mediated ubiquitination and cell signaling in healthy and diseased cells.

Results

CAND2 binds CUL1 and promotes SCF-dependent protein degradation

The high sequence similarity between CAND1 and CAND2 suggests that CAND2 can directly bind CUL1 and prefers binding unneddylated CUL1, as confirmed by the in vitro pulldown assay (Supplementary Fig. 2A) and the co-immunoprecipitation (co-IP) assay (Supplementary Fig. 2B), respectively. We further investigated the CAND2–CUL1 interaction in lysates of HEK293 cells treated with the neddylation inhibitor MLN4924⁴⁰. CAND1/CAND2 double knockout (DKO) HEK293 cells expressing CAND1^HA or CAND2^HA at equal levels were used in this study. After IP with the anti-HA antibody, CUL1 co-immunoprecipitated with both CAND1^HA and CAND2^HA, but a much lower level of CUL1 was associated with CAND2^HA (Fig. 1A). The same assay also revealed that CAND1^HA and CAND2^HA bind multiple other cullins besides CUL1, with varying levels of binding detected (Supplementary Fig. 2C). Next, we performed IP experiments using DKO cells expressing endogenous ^3xFLAGCUL1 (DKO^3xFLAG-CUL1) and transgenic CAND1^HA or CAND2^HA. The cells were lysed in buffer-containing inhibitors, which preserve CUL1 neddylation status, and excess recombinant CUL1^split·RBX1 protein “sponge,” which suppresses the exchange of CUL1 ligands and thereby preserves high-affinity CUL1-containing complexes assembled prior to cell lysis⁹. Consistently, much less CAND2^HA than CAND1^HA was co-immunoprecipitated with ^3xFLAGCUL1, while similar levels of SKP1 were associated with ^3xFLAGCUL1 (Supplementary Fig. 2D). We then asked whether the CAND2–CUL1 interaction affects the degradation of SCF target proteins. In previous studies, we observed that in the presence of tumor necrosis factor alpha (TNF-α), the rate of SCF^β-TrCP-mediated degradation of phosphorylated IκBα (p-IκBα) was significantly reduced in DKO cells but the reduction was milder in the CAND1 single knockout (KO) cells⁸, indicating that CAND2 enhances SCF^β-TrCP activity. To test this hypothesis, we examined p-IκBα degradation in DKO cells supplemented with CAND1^HA or CAND2^HA at a level matching that of endogenous CAND1 (Fig. 1B)⁸. Consistent with findings from the previous study, p-IκBα degradation was defective in DKO cells (Supplementary Fig. 2E, F), and this defect was rescued by CAND1^HA supplementation (Fig. 1B, C). Furthermore, CAND2^HA also restored p-IκBα degradation (Fig. 1B, C), demonstrating that CAND2 can promote SCF activity in cells.

**Fig. 1: CAND2 promotes SCF-mediated protein degradation.**

Although CAND2^HA supplementation promotes IκBα degradation in DKO cells, IκBα degradation was not defective in CAND2 single KO cells⁸, suggesting that CAND1 is the dominant activator of the SCF^β-TrCP–IκBα axis in HEK293 cells. We thus wondered whether knocking out CAND2 alone is sufficient to affect the degradation of a target protein ubiquitinated by a different SCF. Driven by this question, we monitored the iron-induced degradation of IRP2, a substrate of the SCF^FBXL5 ubiquitin ligase^41,42, in cells lacking endogenous CAND1 or CAND2. Adding ferric ammonium citrate (FAC) to the cell culture media stabilized FBXL5 and elevated the FBXL5 level in all the experimental cell lines, including wild-type (WT), CAND1-KO, CAND2-KO, and DKO cells (Fig. 1D). As expected, IRP2 was degraded at a reduced rate in DKO cells, with its t_1/2 increased by 2.8-fold (Fig. 1D, E and Supplementary Fig. 2G). Notably, the t_1/2 of IRP2 was increased by 1.7-fold in CAND1-KO cells and 1.8-fold in CAND2-KO cells (Fig. 1D, E and Supplementary Fig. 2G), indicating that both CAND1 and CAND2 are required for optimal SCF^FBXL5 activity. To further define the molecular basis underlying defective IRP2 degradation in cells lacking CAND2, we examined the recruitment of IRP2 to active SCF^FBXL5 complexes. We lysed cells expressing ^FLAGIRP2 in buffer containing the CUL1^split·RBX1 protein “sponge” (Fig. 1F) and then analyzed the amount of neddylated CUL1 (^N8CUL1) co-immunoprecipitated with ^FLAGIRP2 on anti-FLAG beads. The ^N8CUL1/^FLAGIRP2 ratio revealed that IRP2 was less efficiently recruited to active SCF^FBXL5 in DKO, CAND1-KO, and CAND2-KO cells after FAC treatment (Fig. 1G, H). Taken together, these results demonstrated that CAND2 promotes SCF-mediated protein degradation.

Broad effects of CAND2 on SCF assembly and protein turnover

We next decided to assess the cellular effects of CAND2 on a larger scale. First, we studied how CAND2 affects the landscape of assembled SCFs in HEK293 cells expressing endogenous ^3xFLAGCUL1, with or without CAND2-KO (WT^3xFLAG-CUL1 and CAND2-KO^3xFLAG-CUL1 cells), using an established experimental procedure⁹. The cells were subjected to IP with anti-FLAG beads in the presence of the CUL1^split·RBX1 protein “sponge” that suppresses post-lysis exchange, and the relative levels of SCF assembly were determined by mass spectrometry. Of the 35 FBPs detected in this assay (Supplementary Data 1), 13 showed significant changes upon CAND2-KO, as well as the adaptor protein SKP1 (Fig. 2A). While the level of SCF assembly in CAND-KO cells does not necessarily predict the level of SCF activity⁸, this finding underscores the influence of CAND2 on various SCFs.

**Fig. 2: CAND2 affects the cellular landscape of SCFs and the degradation of SCF target proteins.**

We also asked whether the effect of CAND2 on SCF-mediated protein degradation observed in HEK293 cells applies in a different cellular context. According to the Human Protein Atlas (proteinatlas.org)⁴³, the ratio of CAND2 to CAND1 transcript levels in HEK293 is 0.28, and this ratio is higher in several cancer cell lines, such as TOV-112D, JMSU-1, and JHOS-4. We found that in JMSU-1, a bladder cancer cell line, the ratio of CAND2 to CAND1 protein levels was much higher than that in HEK293 (Supplementary Fig. 2H, I). We therefore generated CAND2-KO in JMSU-1 cells using CRISPR/Cas9 (Fig. 2B). Similar to HEK293 cells, CAND2-KO also impaired the iron-induced degradation of IRP2 in JMSU-1 cells, and the t_1/2 of IRP2 was increased by 1.7-fold (Fig. 2B, C and Supplementary Fig. 2J).

Using the parental and CAND2-KO JMSU-1 cells, we further assessed the impact of CAND2 on global protein turnover rates, using a previously established dynamic SILAC assay⁴⁴. By monitoring both the disappearance of pre-existing cellular proteins and the emergence of newly synthesized isotopically heavier proteins (Fig. 2D), while excluding effects caused by cell division (Supplementary Fig. 2K), the t_1/2 of 1874 proteins detected in both JMSU-1 cell lines were calculated, and fold changes in t_1/2 introduced by CAND2-KO were ranked (Supplementary Data 2). About 600 proteins showed altered t_1/2 (log2 fold change > 0.5) in CAND2-KO cells (Fig. 2E and Supplementary Data 2). These proteins were categorized into a broad spectrum of biological pathways based on GO terms (Fig. 2F and Supplementary Data 3). Among them, more than 400 proteins exhibited increased t_1/2 in CAND2-KO cells, some of which were previously characterized as SCF targets (Fig. 2E, right panel)^{45,46,47,48,49,50,51,52,53} or identified as putative SCF targets (Supplementary Fig. 2L) through global protein stability profiling⁵⁴. The turnover rates for the SCF targets listed in Fig. 2E were slow, with the exception of BCL10 (t_1/2 < 10 h), an SCF^β-TrCP target⁵². Using the cycloheximide (CHX) chase assay as an orthogonal approach, we quantified the t_1/2 of BCL10 in JMSU-1 cells with or without CAND2-KO (Fig. 2G, H). The result was consistent with that from the dynamic SILAC assay, showing that CAND2-KO increased the t_1/2 of BCL10 by 1.4-fold (Fig. 2H and Supplementary Fig. 2M). Taken together, these results reveal the impact of CAND2 on global protein turnover rates, partially through altering SCF activities.

CAND2 promotes the dynamic exchange of a variety of F-box proteins

The finding that CAND2 facilitates the degradation of SCF target proteins suggests that, like CAND1, CAND2 is an exchange factor for FBPs. This idea was first supported by the results of the quantitative in vitro pulldown assays (Supplementary Fig. 3). Shortly after incubating preassembled and immobilized CUL1·SKP2^TAMRA with excess CAND2 (Supplementary Fig. 3A), up to half of the CUL1·SKP2^TAMRA was disassembled, a rate significantly greater than that in the absence of CAND2 (Supplementary Fig. 3B, C). Reciprocally, ~75% of the CUL1·CAND2^TAMRA complex was disassembled shortly after incubating with SKP1·SKP2, a much greater rate of disassembly than that observed in the absence of FBP (Supplementary Fig. 3D–F). We then expanded our assessments and analyzed FBP exchange using DKO^3xFLAG-CUL1 cells⁹. After the cells were lysed in buffer containing SKP1^Sortase·SKP2^ΔLRR, a chase protein complex that disassembles CUL1·CAND1 as efficiently as SKP1·SKP2 (Supplementary Fig. 4A, B), ^3xFLAGCUL1 was immunoprecipitated with and without the addition of recombinant CAND1 or CAND2 (Fig. 3A). The presence of either CAND1 or CAND2 significantly reduced the level of CUL1-bound FBPs, such as SKP2 and FBXL5 (Fig. 3B), suggesting accelerated SCF disassembly. Furthermore, the addition of CAND1 or CAND2 resulted in a decrease in CUL1-bound SKP1 with a simultaneous increase in CUL1-bound SKP1^Sortase (Fig. 3B), suggesting that both CAND1 and CAND2 accelerate the assembly of new SCFs. These results align with the role of CAND2 in facilitating FBP exchange for CUL1.

**Fig. 3: CAND2 promotes the exchange of F-box proteins (FBPs) but less efficiently than CAND1 does.**

Building on the insights gained from these initial experiments, we further expanded our investigation and used a SILAC-based proteomics approach⁹ to systematically assess CAND2-mediated FBP exchange. We cultured DKO cells in isotopically “heavy” media and DKO^3xFLAG-CUL1 cells in normal “light” media. Cells were treated with MLN4924 to eliminate cullin neddylation, lysed, mixed, and immunoprecipitated with anti-FLAG beads. Mass spectrometry analyses of proteins recovered from the beads yielded heavy/light (H/L) ratios of ^3xFLAGCUL1-bound proteins, which provided an estimation of protein exchange levels (Fig. 3C, upper half). In a pilot experiment, adding CAND1 during the IP process increased the exchange level of almost all types of FBPs, as reported previously⁹. Unexpectedly, adding an equal concentration of CAND2 had a marginal effect on the FBP exchange level (Fig. 3C, lower half). We thus included samples with a higher level of CAND2 in our study (Supplementary Data 4). As expected, all the experimental groups showed negligible exchange of CUL1 and RBX1 (Fig. 3D). In the absence of CAND1/2, the exchange level of SKP1 was low, consistent with the low dissociation rate (k_off) of SCF¹⁰ (Fig. 3D). The exchange levels of all the CSN subunits were high, consistent with the high k_off of CSN·CUL1⁵⁵ (Fig. 3D). The majority of the detected FBPs exhibited low exchange levels in the DKO cell lysate, and with a few exceptions (Supplementary Fig. 4C), the exchange levels substantially increased when CAND1 was added at a ratio to ^3xFLAGCUL1 equivalent to the endogenous ratio (1x)⁹. In contrast, the addition of (1x) CAND2 had little effect on the exchange of these FBPs. Increasing CAND2 to 10x the CAND1 level accelerated the exchange of a large group of FBPs, reaching exchange levels lower than those achieved with (1x) CAND1 (Fig. 3E). In addition, the exchange levels of a few FBPs were not noticeably affected by (10x) CAND2 (Fig. 3F). We also observed that the addition of CAND1, but not CAND2, led to a sharp decrease in the level of assembled SCFs, resulting in fewer FBPs being detected by mass spectrometry (Fig. 3G and Supplementary Fig. 4D). Concurrently, considerably more CAND1 than CAND2 was associated with ^3xFLAGCUL1 (Supplementary Fig. 4D), consistent with our earlier findings (Fig. 1A and Supplementary Fig. 2D). Collectively, these results demonstrate that CAND2 can facilitate the dynamic exchange of various FBPs but less efficiently than CAND1.

Differential dissociation dynamics of SCF complexes

The SILAC-based proteomic assay also revealed notable variation in the exchange level of FBPs in DKO cell lysates, ranging from 2% to more than 38% (Fig. 3E–G). The differing exchange levels indicate that the k_off of SCF complexes differ. We therefore analyzed the dissociation dynamics of the SCFome. DKO^3xFLAG-CUL1 cells cultured in “heavy” media were treated with MLN4924, lysed, and IP with anti-FLAG beads in the presence of excess SKP1^Sortase·SKP2^ΔLRR chase. As a control for quantitation, another group of DKO^3xFLAG-CUL1 cells was cultured in “light” media, treated with MLN4924, lysed, and IP with anti-FLAG beads in the absence of the chase protein. The two IP samples were mixed and analyzed by mass spectrometry to determine the H/L ratio of each FBP (Fig. 4A). A decrease in the H/L ratio from 1 reflects the relative rate of dissociation from ^3xFLAGCUL1. As an internal control, H/L ratios for RBX1 and CSN were also determined and were found to be ~1, as expected (Fig. 4B). In contrast, a range of H/L ratios was observed for FBPs, indicating varying k_off values among different SCFs (Fig. 4C). We also found a correlation between the exchange level and the dissociation level for most FBPs (Fig. 4D). For example, with the highest levels of exchange in DKO cells, FBXL8 and FBXO33 exhibited the highest levels of dissociation (Fig. 4D). These FBPs were also not detected in samples with (1x) CAND1 and displayed increased exchange levels in response to (10x) CAND2 (Fig. 3G). On the other hand, FBXO7 and FBXO9 exhibited low levels of both exchange and dissociation in DKO cells (Fig. 4D). Together with FBXL14, FBXO5, FBXO17, and FBXO31, these FBPs displayed increased exchange rates only in response to CAND1 (Fig. 3F). The variation in the dissociation dynamics of SCF complexes may arise from their structural diversity and/or variations in F-box sequences¹².

**Fig. 4: Differential rates of FBP dissociation from CUL1.**

CAND1 and CAND2 bind CUL1 with comparable structures and affinities

In parallel with investigating the impact of CAND2 on FBP exchange, we studied the structure and binding affinity of CUL1·CAND2. Single-particle cryo-EM imaging of purified CAND2·CUL1·RBX1 yielded a final map with a resolution of 3.49 Å (Supplementary Fig. 5 and Supplementary Table 1), enabling the generation of a model for the CUL1·CAND2 complex (Fig. 5A). The overall structure of CUL1·CAND2 closely resembles that of CUL1·CAND1 (Fig. 5B). Like in CAND1, multiple helical repeats in CAND2 pack sequentially, forming a tortuous superhelical structure that wraps around the elongated CUL1 scaffold. The major mechanism for CAND1-mediated FBP exchange involves the structural clash between SKP1 and the CAND1 β hairpin^8,12,56. Similarly, in the CUL1·CAND2 complex, the highly conserved β hairpin of CAND2 (Supplementary Fig. 1) is buried into a hydrophobic groove located in the first cullin repeat of CUL1 (Fig. 5C, D). When the structures of CUL1·CAND2 and CAND1·SCF were superimposed (PDB: 7ZBW)¹², a distinct structural clash was observed between the CAND2 β hairpin and SKP1 (Fig. 5E), indicating a similar molecular basis for the CAND2-mediated exchange of FBPs. In addition, recent structural studies of the CAND1·SCF complex illustrated the direct interaction between CAND1 and SKP1^11,12, proposing a pincer-like mechanism in which the two jaws—the HEAT repeat 27 and the β hairpin of CAND1—attack SKP1 from two different trajectories¹¹. Here, we noted a similar pincer structure in CAND2, with the two jaws appearing slightly more distant, likely resulting in a gentler clash with SKP1 than with CAND1 (Fig. 5F). Nevertheless, it is important to note that the observations based on docking comparisons (Fig. 5E, F) are limited by the static nature of known complex structures. Elucidating the structural basis of the CAND2-mediated SCF disassembly process will require further studies that capture the CAND2·SCF intermediate complexes in action.

**Fig. 5: Architecture of the CUL1·CAND2 complex.**

We next compared the binding kinetics of CUL1·CAND2 with those of CUL1·CAND1. We previously determined the association rate constant (k_on) of CUL1·CAND1 through in vitro fluorescence resonance energy transfer (FRET) using recombinant CAND1 with an N-terminal truncation (CAND1^∆H1). However, CAND1^∆H1 dissociates from CUL1 with a higher k_off than that of full-length CAND1⁸. In addition, attempts to generate CAND2^∆H1 failed to yield viable protein. With the aid of the CUL1·CAND1 and CUL1·CAND2 structures (Fig. 5B), we designed new FRET assays using full-length CAND1 and CAND2, both labeled with the same fluorophore, in conjunction with the same fluorescently labeled CUL1. We fused the cyan fluorescent protein (CFP) to the N-terminus of CAND1 (^CFPCAND1) or CAND2 (^CFPCAND2), and we labeled the C-terminus of CUL1 with the FlAsH fluorophore (CUL1^FlAsH) via a tetracysteine tag. Adding CUL1^FlAsH to either ^CFPCAND1 or ^CFPCAND2 yielded FRET, with donor signals reduced by at least 30%, which was significant enough for kinetic assays (Fig. 6A, B). The addition of unlabeled chase protein abolished the FRET signal, confirming that FRET was dependent on ^CFPCAND-CUL1^FlAsH binding (Fig. 6A, B, orange lines). Next, we measured the k_off of ^CFPCAND·CUL1^FlAsH by monitoring the increase in the donor CFP signal over time after mixing the FRET pair with excess CUL1 chase. The k_off for CUL1·CAND1 was determined to be 8.4 × 10⁻⁶ s⁻¹ (Fig. 6C), which is consistent with our previous measurement using a pulldown assay¹⁰. The k_off for CUL1·CAND2 was 3.1 × 10⁻⁵ s⁻¹ (Fig. 6C), 3.7 times greater than that for CUL1·CAND1. Real-time monitoring of the reduction in the CFP signal upon mixing varying concentrations of CUL1^FlAsH with ^CFPCAND in a stopped-flow apparatus (Supplementary Fig. 6) yielded k_on of 3.2 × 10⁷ M⁻¹ s⁻¹ and 2.3 × 10⁷ M⁻¹ s⁻¹ for CUL1·CAND1 and CUL1·CAND2, respectively (Fig. 6D). Collectively, these findings revealed that, similar to CAND1, CAND2 forms a highly stable complex with CUL1, with a minor difference in the dissociation constant (K_D) between CUL1·CAND1 and CUL1·CAND2 (0.26 vs. 1.3 pM).

**Fig. 6: FRET-based kinetic characterization of CUL1·CAND2.**

SKP1·SKP2 accelerates the disassembly of CUL1·CAND2

The crucial property of CAND1 as an exchange factor for SCFs is its ability to both rapidly disassemble an SCF upon binding to CUL1 and promptly dissociate from CUL1 upon the formation of a new SCF. With the new CUL1·CAND FRET assay, we first quantitatively assessed the effect of SKP1·SKP2 on the assembly of CUL1·CAND2. When CUL1^FlAsH was preassembled with SKP1·SKP2, the k_on of ^CFPCAND1·CUL1^FlAsH was slightly reduced (fivefold), reaching 6.1 × 10⁶ M⁻¹ s⁻¹ (Fig. 7A and Supplementary Fig. 7A). Similarly, the preassembly of CUL1^FlAsH with SKP1·SKP2 led to a modest reduction in the k_on of ^CFPCAND2·CUL1^FlAsH, decreasing from 2.3 × 10⁷ M⁻¹ s⁻¹ to 3.0 × 10⁶ M⁻¹ s⁻¹ (Fig. 7A and Supplementary Fig. 7B).

**Fig. 7: FBP changes the k_off but not the k_on of CUL1·CAND2.**

We next investigated the effect of SKP1·SKP2 on the disassembly of CUL1·CAND2. In line with our earlier findings⁸, mixing SKP1·SKP2 with preassembled ^CFPCAND1·CUL1^FlAsH led to rapid dissociation of the CAND1·CUL1 complex, as evidenced by the rapid increase in the donor CFP signal (Fig. 7B and Supplementary Fig. 8A). The same effect was also observed for CAND2·CUL1 (Fig. 7C and Supplementary Fig. 8B), which was consistent with the results of the pulldown assay (Supplementary Fig. 3D–F). The plotted observed dissociation rate (k_obs) of ^CFPCAND1·CUL1^FlAsH vs. the SKP1·SKP2 concentration conforms to saturation kinetics, giving a maximum rate of 2.5 s⁻¹ and a half maximal concentration (K_M) of 355 nM (Fig. 7D). The same analysis for ^CFPCAND2·CUL1^FlAsH disassembly in the presence of SKP1·SKP2 yielded a maximum rate of 4.4 s⁻¹ and a K_M of 648 nM (Fig. 7D). Thus, CAND2 dissociates from the CAND2·CUL1·SKP1·SKP2 intermediate complex with a k_off of 4.4 s⁻¹, representing a more than 100,000-fold increase compared to the k_off of CAND2·CUL1.

CAND2 accelerates the disassembly of SCF less efficiently than CAND1

We established a new FRET assay for the CUL1-SKP1·SKP2 interaction, using CUL1 fused with an N-terminal CFP (^CFPCUL1) as the FRET donor and SKP1·SKP2 labeled with a FlAsH dye at the N-terminus of SKP1 (^FlAsHSKP1·SKP2) as the FRET acceptor. Mixing ^CFPCUL1 and ^FlAsHSKP1·SKP2 resulted in FRET with a substantial decrease in the donor signal, and the donor signal was restored when ^CFPCUL1 was preincubated with the unlabeled SKP1·SKP2 chase (Fig. 8A). We then evaluated the association of SCF^SKP2 in the absence and presence of CAND. Increasing concentrations of ^FlAsHSKP1·SKP2 were mixed with ^CFPCUL1 (Supplementary Fig. 9A), and the resulting k_obs were plotted against ^FlAsHSKP1·SKP2 concentrations, revealing a k_on of 1.4 × 10⁷ M⁻¹ s⁻¹ (Fig. 8B). We repeated these measurements with ^CFPCUL1 preassembled with CAND1 (Supplementary Fig. 9B), revealing a slightly reduced SCF^SKP2 k_on of 3.4 × 10⁶ M⁻¹ s⁻¹ (Fig. 8B). Similarly, adding ^FlAsHSKP1·SKP2 to ^CFPCUL1 preassembled with CAND2 (Supplementary Fig. 9C) revealed a k_on of 2.5 × 10⁶ M⁻¹ s⁻¹ (Fig. 8B). Thus, either CAND1 or CAND2 had negligible effects on the rate of SCF^SKP2 assembly.

**Fig. 8: CAND2 accelerates the disassembly of SCF^SKP2 with a k_off similar to that of CAND1 but a K_M greater than that of CAND1.**

The most intriguing insights emerged from the analysis of SCF dissociation. The k_off of an SCF has been determined to be extremely low (8.5 × 10⁻⁷s⁻¹)¹⁰. When preassembled ^CFPCUL1·^FlAsHSKP1·SKP2 was mixed with either CAND1 or CAND2 in the stopped-flow apparatus, the disassembly of SCF^SKP2 became very fast, with a notably greater increase in k_obs observed with CAND1 (Fig. 8C). The k_obs increased as the concentration of CAND1 increased (Supplementary Fig. 10A), with a maximum rate at 0.35 s⁻¹ and a K_M of 17 nM (Fig. 8D)—findings comparable to those of previous studies¹⁰. Increasing the CAND2 concentration also resulted in an increase in k_obs (Fig. 8D and Supplementary Fig. 10B), reaching a maximum rate at 0.29 s⁻¹, which is approximately the same as that of CAND1. However, K_M was determined to be 231 nM, ~14 times greater than that of CAND1. While the maximal dissociation rates imply an equivalent speed of SKP1·SKP2 dissociation when SKP1·SKP2 is in the intermediate complex with either CUL1·CAND1 or CUL1·CAND2, the >10-fold difference in K_M suggested that CAND2 is less efficient than CAND1 in disassembling an SCF. This difference aligns with the finding that at least a 10× higher concentration of CAND2 is needed to promote FBP exchange in human cell lysates (Fig. 3C–G) and is likely due to the unproductive assembly of the intermediate complex in a conformation where FBP has a chance to dissociate.

Discussion

Prior studies have extensively elucidated the molecular mechanism through which CAND1 regulates SCF E3 ligases. CAND2, the close homolog of CAND1, has emerged as a risk factor for multiple diseases in population genetics studies^{28,29,30,31,32,33,34,35,36,37,38}. However, the functional mechanism of CAND2 has yet to be defined. Here, with multi-pronged approaches, we investigated how CAND2 binds to CUL1 and affects SCF activity. Our quantitative analyses revealed both similarities and differences between CAND1- and CAND2-mediated SCF activation.

The SILAC-based proteomic assay, which assesses the exchange levels of diverse FBPs in human cell lysates, revealed that, similar to CAND1, CAND2 accelerates the exchange of FBPs associated with the common CUL1 core. However, a higher concentration of CAND2 is required for detectable FBP exchange in bulk solution, suggesting that CAND2 is less efficient at promoting FBP exchange than CAND1 is. Furthermore, while CAND1 promotes the exchange of almost all FBPs, there is a subset that does not show accelerated exchange in response to CAND2. These findings indicate that CAND2 selectively regulates specific SCF-associated pathways.

Measurements of CUL1-CAND and SCF-CAND binding kinetics both confirmed that CAND2 is an FBP exchange factor and provided insights into the differential exchange efficiencies of CAND1 and CAND2. Although CAND1 and CAND2 increase the k_off of SCF to similar levels, the K_M for CAND2-catalyzed SCF disassembly is ~14 times higher than that for CAND1-catalyzed SCF disassembly. The process of SCF destabilization by CAND1 was dissected and visualized in detail in recent structural studies of CAND1-SCF conformations^11,12. When encountering an unneddylated SCF complex, the N-terminal domain of CAND1 (CAND1^NTD) initially binds to the CUL1 C-terminal domain (CUL1^CTD). Subsequently, CAND1 undergoes a rolling motion around CUL1, enabling the binding of CAND1^CTD to CUL1^NTD and destabilizing the CUL1-SKP1·FBP interaction^11,12. Because the structures of CUL1·CAND1 and CUL1·CAND2 are highly similar (Fig. 5B), we expect that CAND2 destabilizes SCF in a similar fashion. The FRET for the CUL1-CAND interaction used in this study relies on the binding of CAND^NTD and CUL1^CTD, and thus, the similar k_on observed for CAND association with FBP-bound CUL1 suggests that both CAND1 and CAND2 efficiently form the initial CAND·SCF complex (Fig. 8E). Therefore, the requirement of a higher CAND2 concentration (or K_M) to disassemble an SCF suggests that the conformations of the CAND2·SCF intermediate complex have a lower probability of dissociating SKP1·FBP from CUL1. This hypothesis may be tested by future structural studies of CAND2·SCF conformations in comparison with CAND1·SCF conformations.

The SCF-CAND binding kinetics data also revealed that the K_M for FBP-mediated CUL1·CAND1 disassembly (355 nM) was 20 times higher than the K_M for CAND1-mediated SCF disassembly (17 nM) (Fig. 8E). Given the ratio of FBPs to CAND1 in HEK293 cells (~2:1⁹), CUL1 is expected to predominantly form CUL1·CAND1 rather than SCF in bulk cell lysate; for the CAND2-SCF interactions, the difference in K_M is much smaller (2.8-fold), so CUL1 is expected to form a higher level of SCF when CAND1 is replaced by CAND2. These predictions are consistent with the experimental data (Supplementary Fig. 4D). However, within HEK293 cells, ~40% of CUL1 is occupied by SKP1·FBP, and another ~40% of CUL1 is occupied by CAND1⁹, suggesting that additional factors counteract the K_M difference to favor SCF assembly in intact cells. Because ~40% of CUL1 in these cells is neddylated⁸, CUL1 neddylation, together with deneddylation, is likely the essential factor maintaining a functional SCF repertoire in cells containing CAND1^8,9. Interestingly, when CAND1 was replaced by CAND2 in HEK293 cells, much less CUL1 was occupied by CAND2, as expected, but the level of SCF complexes and CUL1 neddylation remained largely unchanged (Supplementary Fig. 2D). This finding suggested that despite exhibiting a weaker exchange efficiency, CAND2 is still effective at keeping unwanted SCF complexes disassembled, which are likely SCFs deneddylated due to the loss of loaded substrates. Thus, CAND2 cooperates with NEDD8 and SCF substrates to shape the cellular SCFome, echoing the adaptive exchange hypothesis^8,24,25.

A few questions arise from our study. First, CAND2 exhibits a lower exchange activity than CAND1 in bulk solution; thus, how does it activate SCF-mediated protein degradation as efficiently as CAND1 (Fig. 1)? Since CAND2 has a higher K_M than CAND1 for accelerating SCF disassembly, it is possible that cellular compartmentation creates a local environment and/or maintains a local concentration that allows CAND2 to be effective for FBP exchange. Subcellular localization data from the Human Protein Atlas⁴³, which were technically verified in our experiments (Supplementary Fig. 11), provide some support for this idea. They reported that CAND1 is localized in the nucleoplasm, Golgi apparatus, and cytosol, while CAND2 is localized in nuclear bodies and the cytosol. Second, how does the cell benefit from having CAND2, a weaker exchange factor for SCF? In a recent report, we highlighted that CAND1-mediated CRL2 disassembly inhibits the ubiquitination of low-affinity substrates⁵⁷. A similar scenario may also apply to SCFs, in which the less effective SCF disassembly conferred by CAND2 allows FBPs longer retention time on CUL1, increasing the chance for low-affinity substrates to engage SCFs and undergo ubiquitination. Additionally, both CAND1 and CAND2 bind multiple cullins (Supplementary Fig. 2C), and CAND1 is known to promote CRL3- and CRL4-mediated protein ubiquitination^58,59. Thus, CAND2 is expected to affect protein ubiquitination through regulating CRLs beyond SCF. Finally, why do some cell types require both CAND1 and CAND2, while others seem to rely solely on CAND1^60,61,62? The respective roles of CAND1 and CAND2 in cellular and developmental processes need further investigation in cell types and biological events in which CAND2 is highly enriched or upregulated.

Methods

Cell lines

For DNA plasmid amplification, the DH5α and SURE2 Escherichia coli (E. coli) strains were used. For protein expression, BL21 (DE3) and Rosetta E. coli cells were used. For human cell-based assays, Flp-In T-Rex 293 cells (Invitrogen, RRID: CVCL_U427), JMSU-1 cells (RIKEN, RRID: CVCL_2081), JHOS-4 cells (RIKEN, RRID: CVCL_4649), TOV-112D cells (ATCC, RRID: CVCL_3612) and their genetically modified versions were used. CAND1/CAND2 single KO or DKO cells were generated and confirmed previously⁸. WT, DKO, and CAND2-KO HEK293 cells expressing endogenous ^3xFLAGCUL1 (WT^3xFLAG-CUL1, DKO^3xFLAG-CUL1, and CAND2-KO^3xFLAG-CUL1) were generated and confirmed previously⁹.

Cells with tetracycline-inducible gene expression were generated by co-transfecting Flp-In T-REx 293 cells with pcDNA5/FRT/TO plasmids carrying the gene of interest (CAND1^HA, CAND2^HA, or ^3xFLAGIRP2) and pOG44 plasmids carrying the Flp recombinase using Lipofectamine 3000 (Invitrogen). Transgenic cells were selected using 100 µg/mL hygromycin and validated for expression of integrated genes in the presence of tetracycline. Desirable expression levels were achieved by titrating the concentration of tetracycline.

CAND2-KO JMSU-1 cells were generated via CRISPR/Cas9. The parental JMSU-1 cells were transfected with the PX459 construct (Addgene # 62988) encoding Cas9 and CAND2-targeting gRNA (GAAGATGACGTCCAGCGACA) using Lipofectamine 3000 (Invitrogen). Twenty-four hours after transfection, puromycin (1 µg/mL) was applied. Three days later, the puromycin selection was removed and the enriched cells were seeded onto 15-cm plates. Single colonies were isolated and screened for CAND2-KO by immunoblotting.

Antibodies

The following primary antibodies were used for western blotting: anti-IRP2 (Santa Cruz Biotechnology # 33682, 1:500; Cell Signaling # 37135, 1:500), anti-FBXL5 (NeoClone # N0036, 1:500; BioLegend # 672604, 1:500), anti-GAPDH (Cell Signaling # 2118, 1:1000; Santa Cruz Biotechnology # sc-47724, 1:5000), anti-CAND1 (Santa Cruz Biotechnology # 10672, 1:2000; Bethyl Laboratories # A302-901A, 1:1000), anti-CAND2 (Bethyl Laboratories # A304-046A, 1:1000), anti-CUL1 (Invitrogen # 32-2400, 1:500; Bethyl Laboratories # A303-373A, 1:500), anti-CUL2 (Thermo Fisher Scientific # 51-1800, 1:1000), anti-CUL3 (Cell Signaling # 2759, 1:500), anti-CUL4A (Cell Signaling # 2699, 1:1000), anti-BCL10 (Santa Cruz Biotechnology # sc-5273, 1:200), anti-SKP1 (Cell Signaling # 12248, 1:1000), anti-SKP2 (Cell Signaling # 2652, 1:1000), anti-HA (Cell Signaling # 3724, 1:1000), anti-FLAG (Sigma‒Aldrich # F1804, 1:5000), anti-StrepII (Abcam # ab76949, 1:500), and anti-IκBα (Abcam # ab32518, 1:1000). The secondary antibodies used for western blotting included Alexa Fluor 790-conjugated anti-mouse IgG (Thermo Fisher Scientific # ab186699, 1:10,000), Alexa Fluor 790-conjugated anti-rabbit IgG (Thermo Fisher Scientific # A11374, 1:10,000), Alexa Fluor 680-conjugated anti-rabbit IgG (Thermo Fisher Scientific # ab175772, 1:10,000), and Alexa Fluor 680-conjugated anti-mouse IgG (Abcam # A10038, 1:10,000).

Materials and plasmids

The following constructs were generated for expressing recombinant protein in bacterial cells. A codon-optimized CAND2 coding DNA fragment was synthesized (GeneScript) and integrated into the pGEX-4T vector (Novagen) via restriction digestion and ligation. The plasmids expressing CAND1, SKP1·SKP2, CUL1·RBX1, and ^CFPCUL1·RBX1 were generated previously¹⁰. The coding sequence of CFP was inserted into the 5′ end of the CAND1 gene using the HiFi DNA Assembly Cloning Kit (New England Biolabs) to express ^CFPCAND1, or inserted to the 5′ end of the CAND2 gene via restriction digestion and ligation to express ^CFPCAND2. The StrepII tag (TGGAGCCATCCGCAGTTTGAAAAA) or Sortase tag (TTACCAGAAACAGGAGGA) was added to the 3′ end of the CAND1/2 gene via the Site-Directed Mutagenesis Kit (New England Biolabs). The same method was used to insert the Lumio tag (TGTTGTCCAGGATGTTGT) or Sortase tag into the 5′ or 3′ end of the SKP1 or SKP2 gene, respectively, and to insert the Lumio tag into the 3′ end of the CUL1 gene. For the generation of the SKP1^Sortase·SKP2^ΔLRR chase protein, the LRR domain of SKP2 (SKP2^138-424) was deleted using the Site-Directed Mutagenesis Kit (New England Biolabs).

To generate stable cell lines using the Flp-In system, the coding sequences of CAND1^HA, CAND2^HA, and ^3xFLAGIRP2 were individually integrated into the pcDNA5/FRT/TO vector (Thermo Fisher Scientific) via restriction digestion and ligation. To express ^HAFBXL5 in human cells, the coding sequence of ^HAFBXL5 was integrated into the pcDNA 3.1(+) vector (Invitrogen) via restriction digestion and ligation.

Cellular degradation assay

For the IRP2 degradation assay, cells (0.6 million) of the desired genotypes were seeded on 6-well plates and allowed to grow overnight. On the 2nd day, 0.1 mg/mL FAC (Thermo Fisher Scientific) was added to the cell culture medium, and at different time points after the addition of FAC, the cells were washed with PBS and lysed with 2× SDS sample buffer added directly to the wells. Total cell lysates were collected, sonicated, and fractionated by SDS‒PAGE for western blot analyses. Fluorescence from secondary antibodies was detected by an Odyssey Imager (LI-COR Biosciences). The intensities of IRP2 signals were normalized to the intensities of GAPDH signals in the corresponding sample and were fit to a single exponential in GraphPad Prism to calculate half-lives.

The degradation of IκBα was studied as previously described⁸. DKO cells with CAND1^HA or CAND2^HA integrated through Flp-In were seeded on 6-well plates in the presence of 1000 or 20 ng/mL tetracycline, respectively. The next day, the cells were incubated with serum-free medium for 4 h. CHX (100 μg/mL) was added to the medium 10 min before IκBα degradation was induced with 25 ng/mL TNFα (Sigma-Aldrich). The cells were kept at room temperature during the assay, collected at different time points and analyzed by western blotting as described above.

Protein expression and purification

Recombinant CUL1·RBX1, ^CFPCUL1·RBX1, CUL1·^GSTRBX1, SKP1·SKP2, CAND1, and SKP1^Sortase·SKP2^ΔLRR were expressed and purified as previously described⁸. ^LumioSKP1·SKP2 and CUL1^Lumio·RBX1 were similarly prepared and labeled with FlAsH-EDT₂ (Cayman Chemical) as previously described⁶³. SKP1·SKP2^Sortase and CAND2^Sortase were labeled with the GGGGK^TAMRA peptide (New England Peptide) by sortase-mediated transpeptidation following a previous protocol⁶³. ^CFPCAND1 was prepared in the same way as CAND1. CAND2 and ^CFPCAND2 were expressed in Rosetta E. coli cells induced by IPTG overnight at 16 °C. The overexpressed protein was then extracted by glutathione-affinity chromatography, incubated with thrombin to remove the GST tag, and purified by tandem size exclusion chromatography. CAND1^StrepII and CAND2^StrepII were similarly prepared, except that StrepII-tag affinity chromatography⁶⁴ was performed prior to size exclusion chromatography.

Cryo-EM sample and grid preparation

The CUL1·CAND2 complex freshly purified by size exclusion chromatography was concentrated to 1 mg/mL and incubated with 0.03% glutaraldehyde for 40 min for crosslinking. The reaction was quenched with 50 mM Tris-HCl (pH 8.0) for 15 min, and the crosslinked sample was purified again by size exclusion chromatography to remove aggregates. Fractions containing the crosslinked CUL1·CAND2 complex were pooled and concentrated to 2.5 mg/mL.

For cryo-EM sample grid preparation, aliquots (3.5 μL) of the CUL1·CAND2 sample were applied onto glow-discharged holey carbon Cu R1.2/1.3 grids (Quantifoil, Germany). The grids were blotted for 3.5 s after incubating for 10 s at 100% humidity and 4 °C and quickly plunged into liquid ethane cooled by liquid nitrogen using a Vitrobot Mark IV (Thermo Fisher Scientific).

Cryo-EM data collection and data processing

Cryo-EM data for the CUL1·CAND2 complex were collected at the Stanford SLAC National Accelerator Laboratory S2C2 cryo-EM facility on a Titan Krios G3i (Thermo Fisher Scientific) at 300 kV with a Falcon 4 direct electron detector at a magnification of 105,000 corresponding to a pixel size of 0.95 Å using a Selectris energy filter operated with a slit width of 10 eV. Each movie was recorded for a total of 40 frames over 4.5 s with an exposure dose of 1.28 e⁻/Å² for each frame. The datasets were collected using EPU software (V3.0) with a defocus range of 1–3 µm. The data collection statistics are in Supplementary Table 1.

A total of 5333 movies were collected for the CUL1·CAND2 complex. RBX1 was co-purified but not modeled in the structure. The dataset was motion-corrected via MotionCor2 in the Appion before further processing via the cryoSPARC (V3.2.0). Contrast Transfer Function (CTF) parameters were estimated via patch CTF estimation in the cryoSPARC. The Blob Picker was implemented on the whole dataset to pick all possible particles with a 256-pixel box size. These particles were then subjected to seven rounds of 2D classification and an initial 3D ab initio reconstruction. A subset of templates generated from the best class was used in Template Picker, resulting in the extraction of 2,240,514 particles. Classes with a high number of complex (both CAND2 and CUL1) were selected, with classes that represent a majority of single protein components not chosen to narrow down the complex specifically. A total of 348,902 particles were selected after three rounds of 2D classification and were subsequently sorted into three 3D classes via Ab initio reconstruction. Particles from the best 3D class of the CUL1·CAND2 complex were subjected to further postprocessing steps, including non-uniform refinement, which yielded a 3.49 Å resolution map. Only one class with the highest number of particles was selected for the subsequent processing. This stringent selection was performed to avoid slight differences in structural conformation, which we find affects the resolution of the overall complex. The data processing workflow is shown in Supplementary Fig. 5.

Cryo-EM model building, refinement, and analysis

The CUL1·CAND2 models were constructed using the AlphaFold II model as a starting template. The predicted structure was docked into the density map using UCSF Chimera (v1.14) and manually adjusted and rebuilt by Coot (v0.9.3). The resulting model was refined using the real_space_refine module of PHENIX (v1.20.1-4487) with secondary structure and geometry restraints. The structures were validated with MolProbity (v4.5), and refinement statistics are in Supplementary Table 1. All the structure figures were prepared in PyMOL (V2.4.0) and ChimeraX (v1.5).

FRET-based assay

To obtain fluorescence emission spectra, fluorescent proteins were mixed in FRET buffer [30 mM Tris (pH 7.6), 100 mM NaCl, 0.5 mM DTT, and 1 mg/mL ovalbumin] and analyzed by a MOS-EMM-MONO emission monochromator coupled with a MOS-200/M spectrometer (BioLogic). For assays using ^CFPCAND1 and ^CFPCAND2 as FRET donors, samples were excited at 430 nm, and the emission signals were collected from 450 nm to 600 nm with 1 nm increments. For assays using ^CFPCUL1 as the FRET donor, samples were excited at 430 nm, and the emission signals were collected from 450 to 650 nm with 1 nm increments. The FRET-based pre-steady-state measurements were performed on an SFM-3000 stopped-flow mixer coupled with a MOS-200/M spectrometer (BioLogic), using Bio-Kine32 software (v4) for data collection. A bandpass filter AT480/30 m (Chroma) was used to monitor the CFP signals.

To determine the k_off of CUL1·CAND1 and CUL1·CAND2, 25 nM ^CFPCAND1 or ^CFPCAND2 was mixed with 25 nM CUL1^FlAsH in FRET buffer, and aliquots of this mixture were incubated with 250 nM unlabeled CUL1 chase for the desired time periods at room temperature. All the samples (150 µL each) were transferred to a black 96-well plate (COSTAR), and the intensity of the CFP emission signal was measured by a Clariostar microplate reader (BMG Labtech, Cary, NC). The data were normalized to the blank and to the sample with no chase protein, and the normalized data were fit into a one-phase decay model in GraphPad Prism (GraphPad software). Three independent replicates were included to obtain the k_off.

In vitro pulldown for protein exchange

Equal amounts of glutathione Sepharose 4B resin (Cytiva) were incubated with 700 μL of solution containing 0.2 μM CUL1·^GSTRBX1 and CAND1^TAMRA, or 0.2 μM CUL1·^GSTRBX1 and CAND2^TAMRA, or 0.2 μM CUL1·^GSTRBX1 and SKP1·SKP2^TAMRA, for 15 min. Aliquots of the resin slurry were transferred to Micro-Spin columns (Bio-Rad), and the beads were incubated with 100 μL of 2 μM SKP1·SKP2, or 2 μM CAND1, or 2 μM CAND2, as indicated in the figure, for 30 s or 5 min. Protein‒protein interactions were terminated by separating the resin and the supernatant via centrifugation. Equivalent portions of the resin and flowthrough were sampled and fractionated by SDS‒PAGE. Gels were scanned by a Typhoon scanner (Cytiva) to quantify the percentage of bound CAND1^TAMRA, CAND2^TAMRA, or SKP1·SKP2^TAMRA.

Co-immunoprecipitation assay

For the CUL1·CAND2 interaction, DKO cells and DKO cells expressing transgenic CAND1^HA or CAND2^HA were treated with or without 1 µM MLN4924 for 1 h, collected, and sonicated in lysis buffer [50 mM HEPES (pH 7.5), 5 mM Mg(OAc)₂, 70 mM KOAc, 1 mM DTT and a protease inhibitor cocktail (Roche)]. After centrifugation at 20,000 × g for 15 min, the clear cell lysate was incubated with anti-HA beads (Sigma-Aldrich) for 2 h at 4 °C. The beads were then washed with lysis buffer three times and eluted with 2× SDS sample buffer for western blot analysis.

To determine levels of net SCF assembly in cells, DKO^3xFLAG-CUL1 Flp-In cells were induced to express CAND1^HA or CAND2^HA. Cells were collected and lysed in Pierce IP lysis buffer (Thermo Fisher Scientific) supplemented with 1 mM DTT, 1× protease inhibitor cocktail, 50 μM quinoline-8-thiol (8TQ)⁶⁵, 1 μM MLN4924, and recombinant CUL1·RBX1 at ~100× the level of endogenous CUL1. The clear cell lysate was incubated with anti-FLAG beads (Sigma-Aldrich) for 30 min at 4 °C. Proteins eluted from the beads were immunoblotted with anti-FLAG, anti-SKP1, anti-SKP2, anti-HA, and anti-GAPDH antibodies.

To determine the effect of CAND2 on the assembly of individual SCF in the cellular SCFome⁹, WT^3xFLAG-CUL1 and CAND2-KO^3xFLAG-CUL1 cells were subjected to IP with anti-FLAG beads as described above. Proteins immobilized on the beads were eluted and processed for mass spectrometry analysis as described below in the “SILAC sample preparation” section.

To assess the levels of IRP2-bound SCF^FBXL5, Flp-In cells expressing ^3xFLAGIRP2 were transfected with the pcDNA 3.1(+) plasmid expressing ^HAFBXL5. One day after transfection, the cells were treated with 0.1 mg/mL FAC for 4 h, collected, and sonicated in lysis buffer [10 mM HEPES (pH 7.9), 1.5 mM MgCl₂, 10 mM KCl, 1 mM DTT, 0.6% IGEPAL CA-630, 1× protease inhibitor cocktail] supplemented with recombinant CUL1·RBX1 at ~100× the level of endogenous CUL1. The lysis buffer also contained 50 μM 8TQ and 1 μM MLN4924 to inhibit deneddylation and neddylation in the cell lysate. The clear cell lysate was incubated with anti-FLAG beads (Sigma-Aldrich) for 30 min at 4 °C. Proteins eluted from the beads were immunoblotted with anti-FLAG, anti-HA, and anti-CUL1 antibodies.

To investigate the effect of CAND1 and CAND2 on the disassembly of SCF complexes, DKO^3xFLAG-CUL1 cells were lysed in Pierce IP lysis buffer (Thermo Fisher Scientific) supplemented with 1 mM DTT, 1x protease inhibitor cocktail, 50 μM 8TQ, and 1 μM MLN4924. The clear cell lysate was then mixed with recombinant SKP1^Sortase·SKP2^ΔLRR (~100× of the endogenous SKP1 level) ± recombinant CAND1 or CAND2 (5× of the endogenous CAND1 level) and incubated with anti-FLAG beads (Sigma-Aldrich) for 30 min at room temperature. Proteins eluted from the beads were immunoblotted with anti-SKP1, anti-SKP2, anti-FBXL5, and anti-FLAG antibodies.

Immunofluorescence analysis

WT or DKO cells were seeded onto poly(l-lysine)-coated coverslips (Corning, cat. no. 354085) in 24-well plates. The next day, the cells were rinsed with PBS three times, fixed with 4% paraformaldehyde (Thermo Fisher Scientific) for 10 min at 37 °C, and washed with PBS three times. Permeabilization was performed by incubating the cells with 0.1% Triton X-100 at room temperature for 10 min, followed by an additional wash with PBS three times. The cells were then blocked with 10% goat serum (Abcam, cat. no. ab7481) for 60 min at room temperature. After removing the blocking buffer, the cells were incubated with primary antibodies (anti-CAND1, Sigma-Aldrich, cat. no. HPA055748, 1 μg/mL; anti-CAND2, Sigma-Aldrich, cat. no. HPA005777, 1 μg/mL) at room temperature for 2 h, washed three times with PBS-T, and incubated with the secondary antibody (Alexa Fluor 488-conjugated anti-rabbit IgG, Thermo Scientific, cat. no. A11008, 4 μg/mL) for 1 h. Coverslips were finally rinsed twice with PBS-T, twice with water, and embedded in the antifade mounting reagent containing DAPI (Invitrogen). Images were acquired on a Zeiss LSM 880 laser scanning confocal microscope and processed with ZEN 3.0 software (blue edition, Zeiss).

SILAC sample preparation

SILAC medium was prepared as Arg- and Lys-depleted DMEM (Thermo Scientific) containing 10% dialyzed fetal bovine serum (Cytiva) supplemented with 87 mg/L [¹³C₆, ¹⁵N₄] L-arginine (Arg10) (Sigma-Aldrich) and 153 mg/L [¹³C6, ¹⁵N₂] l-lysine (Lys8) (Sigma-Aldrich) (“heavy” medium), or the corresponding unlabeled Arg and Lys (Sigma-Aldrich) (“light” medium). To profile the exchange of FBPs, DKO^3xFLAG-CUL1 cells were grown in “light” or “heavy” media for 12 days and treated with 1 µM MLN4924 for 1 h. Equal numbers (4 × 10⁷) of “light-” and “heavy-” grown cells were separately collected and lysed in Pierce IP lysis buffer supplemented with 1 mM DTT, 1x protease inhibitor cocktail, 50 μM 8TQ and 1 μM MLN4924. The “light” and “heavy” clear cell lysates were then mixed and incubated with 100 µL of anti-FLAG resin for 30 min at room temperature in the presence and absence of recombinant CAND1 or CAND2. Subsequently, the resin was pelleted via centrifugation at 1000 × g for 2 min and washed twice with lysis buffer and 100 mM Tris-HCl (pH 8.0). The resin was subsequently transferred to a spin column (Bio-Rad, 7326204) and incubated twice with 150 µL of 10% ammonium hydroxide for 5 min to elute the bound proteins. The eluates were dried in a SpeedVac vacuum concentrator and resuspended in 200 µL of buffer supplemented with 4 M urea in 10 mM Tris-HCl (pH 8.0). At room temperature, the samples were further treated with 5 mM DTT for 30 min (for reduction) and 10 mM iodoacetamide for 20 min in the dark (for alkylation of cysteine residues), incubated with 1 µg of LysC protease (Wako Chemicals, 125–05061) for 3 h at room temperature, diluted with 50 mM ammonium bicarbonate to bring the urea concentration to 2 M, and incubated with 1 µg of trypsin (Sigma-Aldrich) overnight at 37 °C. The resulting peptides were loaded onto C18 silica microspin columns (The Nest Group, SS18V), prewashed with 100% acetonitrile, preequilibrated in Buffer A (0.1% formic acid in water), washed twice with Buffer A and then twice with Buffer B (3% acetonitrile, 97% H₂O and 0.1% formic acid), and eluted with 80% acetonitrile. Small aliquots of the eluates were used to determine the peptide concentration with a Pierce Quantitative Colorimetric Peptide Assay Kit (Thermo Scientific). The rest of the eluates were brought to complete dryness in a SpeedVac vacuum concentrator and stored at −80 °C until analysis by mass spectrometry.

For the analysis of global protein turnover rate using dynamic SILAC, WT, and CAND2-KO JMSU-1 cells in duplicate were grown in “light” media. At around 60% confluency, the cells were washed three times with DPBS and continue to be cultured in “heavy” media for 2, 6, or 18 h. At indicated time points, cells were washed with DPBS and dissolved in 4 M urea in 10 mM Tris-HCl (pH 8.0) and processed for mass spectrometry analysis as described above. To measure the cell doubling time (Supplementary Fig. 2I), JMSU-1 cells with or without CAND2-KO grown in the “light” media were seeded 25,000 cells per well in 6-well plates. The cells were then cultured in “heavy” media for 4 days, collected, counted, and the doubling time was calculated.

For the SILAC-based comparison of FBP dissociation from CUL1, DKO^3xFLAG-CUL1 cells were grown in “light” or “heavy” SILAC medium, treated with MLN4924, collected and lysed as described above. A large excess of SKP1^Sortase·SKP2^ΔLRR chase protein (~100x the endogenous SKP1 level) was added to the “heavy” clear cell lysate. Equal amounts of anti-FLAG beads were added to the “light” and “heavy” clear cell lysates, incubated at room temperature for 60 min, washed and eluted as described above. The “light” and “heavy” eluate samples were mixed and subjected to the same sample preparation steps (protease digestion, peptide desalting, and quantification) as described above. Label-swap replications were performed in parallel, and heavy/light (or light/heavy) ratios measured in all individual replicates are reported in Fig. 4B, C.

Mass spectrometry data collection and analysis

Mass spectra were acquired with an Orbitrap Exploris™ 480 Mass Spectrometer (Thermo Fisher Scientific) after the peptides were separated on a 25 cm C18 reverse-phase analytical column (Thermo Fisher EasySpray ES902 A). The raw data were processed in Proteome Discoverer 2.5 (Thermo Fisher Scientific). The data collection and analysis methods used were the same as those reported in a previous study⁵⁷. The SILAC H/L ratios for proteins were determined by calculating the median of all pairwise peptide ratios.

Quantitation of global protein turnover rate

Half-lives of proteins detected in samples from the dynamic SILAC assay (see the “SILAC sample preparation” section above) were determined using a previously established method⁴⁴. Proteins with H/L ratios (r) from at least three samples were kept for half-life calculation. Briefly, ln(r + 1) for each protein were linearly fitted to get the apparent degradation rates, which were further corrected for diluting effect resulting from cell division. As a quality screen for the quantitation, both goodness-of-fit of linear regressions and coefficient of variation of the slope analyzed by leave-one-out cross-validation were calculated, with ≥0.85 and ≤1% as the cutoff value, respectively. Gene Ontology analysis was performed using STRING⁶⁶.

Quantification and statistical analyses

Western blots with infrared fluorescent signals were scanned on an Odyssey Imager (LI-COR Biosciences) using Image Studio (v5.2) and quantified by Image Studio Lite (v5.2, LI-COR Biosciences). Fluorescence signals detected by the Typhoon scanner were quantified by GelQuantNET (v1.7, kindly provided by biochemlabsolutions.com). The statistical parameters used are reported in the figures and figure legends. Differences were considered statistically significant at p < 0.05 according to unpaired two-tailed Student’s t-tests. All the statistical analyses were performed in GraphPad Prism (v9.0). Alignments were performed in Jalview (v2.11.3.0) and visualized with the ESPript 3.0 web server⁶⁷.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.

Data availability

The proteomics data generated in this study have been deposited to the ProteomeXchange Consortium via PRIDE under accession codes PXD045616 [http://proteomecentral.proteomexchange.org/cgi/GetDataset?I D = PXD045616], PXD045618, PXD056462 [http://proteomecentral.proteomexchange.org/cgi/GetDataset?I D = PXD056462] and PXD056464 [http://proteomecentral.proteomexchange.org/cgi/GetDataset?I D = PXD056464]. Cryo-EM map and corresponding coordinate have been deposited under EMDB and PDB entries EMD-43572 and 8VVY. Source data are provided with this paper.

Code availability

The code used to analyze the global protein turnover data is available on GitHub [https://github.com/RavenGan/GlobalProteinTurnoverRates/tree/main].

References

Harper, J. W. & Schulman, B. A. Cullin-RING ubiquitin ligase regulatory circuits: a quarter century beyond the F-box hypothesis. Annu. Rev. Biochem. 90, 403–429 (2021).
Article CAS PubMed PubMed Central Google Scholar
Wang, K., Deshaies, R. J. & Liu, X. Assembly and regulation of CRL ubiquitin ligases. Adv. Exp. Med. Biol. 1217, 33–46 (2020).
Article CAS PubMed MATH Google Scholar
Deshaies, R. J. & Joazeiro, C. A. RING domain E3 ubiquitin ligases. Annu. Rev. Biochem. 78, 399–434 (2009).
Article CAS PubMed Google Scholar
Bai, C. et al. SKP1 connects cell cycle regulators to the ubiquitin proteolysis machinery through a novel motif, the F-box. Cell 86, 263–274 (1996).
Article CAS PubMed Google Scholar
Feldman, R. M., Correll, C. C., Kaplan, K. B. & Deshaies, R. J. A complex of Cdc4p, Skp1p, and Cdc53p/cullin catalyzes ubiquitination of the phosphorylated CDK inhibitor Sic1p. Cell 91, 221–230 (1997).
Article CAS PubMed Google Scholar
Zheng, N. et al. Structure of the Cul1-Rbx1-Skp1-F boxSkp2 SCF ubiquitin ligase complex. Nature 416, 703–709 (2002).
Article ADS CAS PubMed MATH Google Scholar
Nguyen, K. M. & Busino, L. The biology of F-box proteins: the SCF family of E3 ubiquitin ligases. Adv. Exp. Med. Biol. 1217, 111–122 (2020).
Article CAS PubMed Google Scholar
Liu, X. et al. Cand1-mediated adaptive exchange mechanism enables variation in F-box protein expression. Mol. Cell 69, 773–786 e776 (2018).
Article CAS PubMed PubMed Central MATH Google Scholar
Reitsma, J. M. et al. Composition and regulation of the cellular repertoire of SCF ubiquitin ligases. Cell 171, 1326–1339 e1314 (2017).
Article CAS PubMed PubMed Central MATH Google Scholar
Pierce, N. W. et al. Cand1 promotes assembly of new SCF Complexes through dynamic exchange of F box proteins. Cell 153, 206–215 (2013).
Article CAS PubMed PubMed Central MATH Google Scholar
Shaaban, M. et al. Structural and mechanistic insights into the CAND1-mediated SCF substrate receptor exchange. Mol. Cell 83, 2332–2346 e2338 (2023).
Article CAS PubMed MATH Google Scholar
Baek, K. et al. Systemwide disassembly and assembly of SCF ubiquitin ligase complexes. Cell 186, 1895–1911 e1821 (2023).
Article CAS PubMed PubMed Central MATH Google Scholar
Wu, S. et al. CAND1 controls in vivo dynamics of the cullin 1-RING ubiquitin ligase repertoire. Nat. Commun. 4, 1642 (2013).
Article ADS PubMed MATH Google Scholar
Zemla, A. et al. CSN- and CAND1-dependent remodelling of the budding yeast SCF complex. Nat. Commun. 4, 1641 (2013).
Article ADS PubMed MATH Google Scholar
Cheng, Y., Dai, X. & Zhao, Y. AtCAND1, a HEAT-repeat protein that participates in auxin signaling in Arabidopsis. Plant Physiol. 135, 1020–1026 (2004).
Article CAS PubMed PubMed Central Google Scholar
Chuang, H. W., Zhang, W. & Gray, W. M. Arabidopsis ETA2, an apparent ortholog of the human cullin-interacting protein CAND1, is required for auxin responses mediated by the SCF^TIR1 ubiquitin ligase. Plant Cell 16, 1883–1897 (2004).
Article CAS PubMed PubMed Central Google Scholar
Feng, S. et al. Arabidopsis CAND1, an unmodified CUL1-interacting protein, is involved in multiple developmental pathways controlled by ubiquitin/proteasome-mediated protein degradation. Plant Cell 16, 1870–1882 (2004).
Article CAS PubMed PubMed Central Google Scholar
Alonso-Peral, M. M. et al. The HVE/CAND1 gene is required for the early patterning of leaf venation in Arabidopsis. Development 133, 3755–3766 (2006).
Article CAS PubMed MATH Google Scholar
Zhang, W. et al. Genetic analysis of CAND1-CUL1 interactions in Arabidopsis supports a role for CAND1-mediated cycling of the SCF^TIR1 complex. Proc. Natl. Acad. Sci. USA 105, 8470–8475 (2008).
Article ADS CAS PubMed PubMed Central MATH Google Scholar
Li, L. et al. CAND1 is required for pollen viability in Arabidopsis thaliana-a test of the adaptive exchange hypothesis. Front. Plant Sci. 13, 866086 (2022).
Article PubMed PubMed Central Google Scholar
Baek, K. et al. NEDD8 nucleates a multivalent cullin-RING-UBE2D ubiquitin ligation assembly. Nature 578, 461–466 (2020).
Article ADS CAS PubMed PubMed Central MATH Google Scholar
Gutierrez, C. et al. Structural dynamics of the human COP9 signalosome revealed by cross-linking mass spectrometry and integrative modeling. Proc. Natl. Acad. Sci. USA 117, 4088–4098 (2020).
Article ADS CAS PubMed PubMed Central MATH Google Scholar
Schulze-Niemand, E. & Naumann, M. The COP9 signalosome: a versatile regulatory hub of Cullin-RING ligases. Trends Biochem. Sci. 48, 82–95 (2023).
Article CAS PubMed MATH Google Scholar
Zhang, Y. et al. Adaptive exchange sustains cullin-RING ubiquitin ligase networks and proper licensing of DNA replication. Proc. Natl. Acad. Sci. USA 119, e2205608119 (2022).
Article CAS PubMed PubMed Central Google Scholar
Li, L., Wang, K., Zhou, Y. & Liu, X. Review: a silent concert in developing plants: dynamic assembly of cullin-RING ubiquitin ligases. Plant Sci. 330, 111662 (2023).
Article CAS PubMed PubMed Central MATH Google Scholar
Aoki, T. et al. TIP120B: a novel TIP120-family protein that is expressed specifically in muscle tissues. Biochem. Biophys. Res. Commun. 261, 911–916 (1999).
Article CAS PubMed MATH Google Scholar
Shiraishi, S. et al. TBP-interacting protein 120B (TIP120B)/cullin-associated and neddylation-dissociated 2 (CAND2) inhibits SCF-dependent ubiquitination of myogenin and accelerates myogenic differentiation. J. Biol. Chem. 282, 9017–9028 (2007).
Article CAS PubMed MATH Google Scholar
Chen, S. et al. Identification of recurrent variants implicated in disease in bicuspid aortic valve patients through whole-exome sequencing. Hum. Genomics 16, 36 (2022).
Article CAS PubMed PubMed Central Google Scholar
Zhang, C., Qin, F., Li, X., Du, X. & Li, T. Identification of novel proteins for lacunar stroke by integrating genome-wide association data and human brain proteomes. BMC Med. 20, 211 (2022).
Article CAS PubMed PubMed Central MATH Google Scholar
Weng, L. C. et al. Genetic determinants of electrocardiographic P-wave duration and relation to atrial fibrillation. Circ. Genom. Precis. Med. 13, 387–395 (2020).
Article PubMed PubMed Central MATH Google Scholar
Zhu, X. et al. Hepatic transcriptome and proteome analyses provide new insights into the regulator mechanism of dietary avicularin in diabetic mice. Food Res. Int. 125, 108570 (2019).
Article CAS PubMed MATH Google Scholar
Koskimaki, J. et al. Transcriptome clarifies mechanisms of lesion genesis versus progression in models of Ccm3 cerebral cavernous malformations. Acta Neuropathol. Commun. 7, 132 (2019).
Article PubMed PubMed Central Google Scholar
Dong, S. S. et al. Detecting epistasis within chromatin regulatory circuitry reveals CAND2 as a novel susceptibility gene for obesity. Int J. Obes. 43, 450–456 (2019).
Article CAS MATH Google Scholar
Gregers, E. et al. Deep sequencing of atrial fibrillation patients with mitral valve regurgitation shows no evidence of mosaicism but reveals novel rare germline variants. Heart Rhythm 14, 1531–1538 (2017).
Article PubMed MATH Google Scholar
Wei, T. et al. NEURL rs6584555 and CAND2 rs4642101 contribute to postoperative atrial fibrillation: a prospective study among Chinese population. Oncotarget 7, 42617–42624 (2016).
Article PubMed PubMed Central Google Scholar
Dmitriev, A. A. et al. Identification of novel epigenetic markers of prostate cancer by NotI-microarray analysis. Dis. Markers 2015, 241301 (2015).
Article PubMed PubMed Central MATH Google Scholar
Sinner, M. F. et al. Integrating genetic, transcriptional, and functional analyses to identify 5 novel genes for atrial fibrillation. Circulation 130, 1225–1235 (2014).
Article CAS PubMed PubMed Central MATH Google Scholar
Christophersen, I. E. et al. Large-scale analyses of common and rare variants identify 12 new loci associated with atrial fibrillation. Nat. Genet. 49, 946–952 (2017).
Article CAS PubMed PubMed Central MATH Google Scholar
Gorska, A. A. et al. Muscle-specific Cand2 is translationally upregulated by mTORC1 and promotes adverse cardiac remodeling. EMBO Rep. 22, e52170 (2021).
Article CAS PubMed PubMed Central Google Scholar
Soucy, T. A. et al. An inhibitor of NEDD8-activating enzyme as a new approach to treat cancer. Nature 458, 732–736 (2009).
Article ADS CAS PubMed MATH Google Scholar
Salahudeen, A. A. et al. An E3 ligase possessing an iron-responsive hemerythrin domain is a regulator of iron homeostasis. Science 326, 722–726 (2009).
Article ADS CAS PubMed PubMed Central Google Scholar
Vashisht, A. A. et al. Control of iron homeostasis by an iron-regulated ubiquitin ligase. Science 326, 718–721 (2009).
Article ADS CAS PubMed PubMed Central MATH Google Scholar
Thul, P. J. et al. A subcellular map of the human proteome. Science 356, eaal3321 (2017).
Schwanhausser, B. et al. Global quantification of mammalian gene expression control. Nature 473, 337–342 (2011).
Article ADS PubMed MATH Google Scholar
Alfano, L. et al. NONO ubiquitination is mediated by FBW7 and GSK3 via a degron lost upon chromosomal rearrangement in cancer. J. Cell. Physiol. 233, 4338–4344 (2018).
Article CAS PubMed MATH Google Scholar
Lin, X. T. et al. Elevated FBXO45 promotes liver tumorigenesis through enhancing IGF2BP1 ubiquitination and subsequent PLK1 upregulation. Elife 10, e70715 (2021).
Tan, M. et al. SAG/ROC-SCF beta-TrCP E3 ubiquitin ligase promotes pro-caspase-3 degradation as a mechanism of apoptosis protection. Neoplasia 8, 1042–1054 (2006).
Article CAS PubMed PubMed Central MATH Google Scholar
Lai, Y., Li, J., Li, X. & Zou, C. Lipopolysaccharide modulates p300 and Sirt1 to promote PRMT1 stability via an SCF(Fbxl17)-recognized acetyldegron. J. Cell Sci. 130, 3578–3587 (2017).
Article CAS PubMed PubMed Central MATH Google Scholar
Luo, L. et al. FBXO7 ubiquitinates PRMT1 to suppress serine synthesis and tumor growth in hepatocellular carcinoma. Nat. Commun. 15, 4790 (2024).
Article ADS CAS PubMed PubMed Central MATH Google Scholar
Liu, L. et al. Ubiquitin ligase subunit FBXO9 inhibits V-ATPase assembly and impedes lung cancer metastasis. Exp. Hematol. Oncol. 13, 32 (2024).
Article CAS PubMed PubMed Central Google Scholar
Duan, S. et al. Loss of FBXO31-mediated degradation of DUSP6 dysregulates ERK and PI3K-AKT signaling and promotes prostate tumorigenesis. Cell Rep. 37, 109870 (2021).
Article CAS PubMed PubMed Central MATH Google Scholar
Lobry, C., Lopez, T., Israel, A. & Weil, R. Negative feedback loop in T cell activation through IkappaB kinase-induced phosphorylation and degradation of Bcl10. Proc. Natl. Acad. Sci. USA 104, 908–913 (2007).
Article ADS CAS PubMed PubMed Central MATH Google Scholar
Mouery, R. D. et al. Proteomic analysis reveals a PLK1-dependent G2/M degradation program and a role for AKAP2 in coordinating the mitotic cytoskeleton. Cell Rep. 43, 114510 (2024).
Article CAS PubMed PubMed Central MATH Google Scholar
Yen, H. C. & Elledge, S. J. Identification of SCF ubiquitin ligase substrates by global protein stability profiling. Science 322, 923–929 (2008).
Article ADS CAS PubMed MATH Google Scholar
Mosadeghi, R. et al. Structural and kinetic analysis of the COP9-Signalosome activation and the cullin-RING ubiquitin ligase deneddylation cycle. Elife 5, e12102 (2016).
Goldenberg, S. J. et al. Structure of the Cand1-Cul1-Roc1 complex reveals regulatory mechanisms for the assembly of the multisubunit cullin-dependent ubiquitin ligases. Cell 119, 517–528 (2004).
Article CAS PubMed Google Scholar
Wang, K., Diaz, S., Li, L., Lohman, J. R. & Liu, X. CAND1 inhibits Cullin-2-RING ubiquitin ligases for enhanced substrate specificity. Nat. Struct. Mol. Biol., 31, 323-335 (2024).
Lo, S. C. & Hannink, M. CAND1-mediated substrate adaptor recycling is required for efficient repression of Nrf2 by Keap1. Mol. Cell. Biol. 26, 1235–1244 (2006).
Article CAS PubMed PubMed Central Google Scholar
Reichermeier, K. M. et al. PIKES analysis reveals response to degraders and key regulatory mechanisms of the CRL4 network. Mol. Cell 77, 1092–1106 e1099 (2020).
Article CAS PubMed MATH Google Scholar
Karlsson, M. et al. A single-cell type transcriptomics map of human tissues. Sci. Adv. 7, eabh2169 (2021).
Sjostedt, E. et al. An atlas of the protein-coding genes in the human, pig, and mouse brain. Science 367, eaay5947 (2020).
Uhlen, M. et al. Proteomics. Tissue-based map of the human proteome. Science 347, 1260419 (2015).
Article PubMed MATH Google Scholar
Garsamo, M., Zhou, Y. & Liu, X. Using in vitro fluorescence resonance energy transfer to study the dynamics of protein complexes at a millisecond time scale. J. Vis. Exp., 145, e59038 (2019).
Wang, K. & Liu, X. Determining the effects of neddylation on Cullin-RING ligase-dependent protein ubiquitination. Curr. Protoc. 2, e401 (2022).
Article CAS PubMed PubMed Central MATH Google Scholar
Li, J. et al. Capzimin is a potent and specific inhibitor of proteasome isopeptidase Rpn11. Nat. Chem. Biol. 13, 486–493 (2017).
Article CAS PubMed PubMed Central MATH Google Scholar
Szklarczyk, D. et al. The STRING database in 2023: protein–protein association networks and functional enrichment analyses for any sequenced genome of interest. Nucleic Acids Res. 51, D638–D646 (2023).
Article CAS PubMed Google Scholar
Robert, X. & Gouet, P. Deciphering key features in protein structures with the new ENDscript server. Nucleic Acids Res. 42, W320–W324 (2014).
Article CAS PubMed PubMed Central MATH Google Scholar

Download references

Acknowledgements

The mass spectrometry work was performed by the Indiana University School of Medicine Center for Proteome Analysis (IUSM-CPA). Acquisition of the IUSM-CPA instrumentation used for this project was provided in part by the Indiana University Precision Health Initiative and the IU Simon Comprehensive Cancer Center. The proteomics work was supported, in part, by the Indiana Clinical and Translational Sciences Institute (Award Number UL1TR002529 from the National Institutes of Health, National Center for Advancing Translational Sciences, Clinical and Translational Sciences Award) and, in part, by the IU Simon Comprehensive Cancer Center Support Grant (Award Number P30CA082709 from the National Cancer Institute). Measurements of IRP2 t_1/2 (Fig. 1D, E) and the generation of CAND2-KO^3xFLAG-CUL1 cells (Fig. 2A) were performed in Dr. Raymond Deshaies's laboratory at the California Institute of Technology. Some of the Cryo-EM data were collected at the Stanford-SLAC Cryo-EM Center (S2C2), which is supported by the National Institutes of Health Common Fund Transformative High-Resolution Cryo-Electron Microscopy program (U24 GM129541). The authors would like to thank the S2C2 personnel Htet Khant and Yan (Amy) Liu for their invaluable support and assistance. This research was supported by the National Institutes of Health grants R35GM138016 (to X.L., primary funding source of this study) and R01GM143268 (to Y.Z.), the American Heart Association Career Development Award 20CDA35270030 (to X.L.), the NIGMS-funded predoctoral fellowship T32GM132024 (to S.K.), the SIRG Graduate Research Assistantships Award from the Purdue Institute for Cancer Research P30CA023168 (to S.K.), and the American Heart Association grant # 905924/SK/2021 (to S.K.). The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.

Author information

Justin M. Reitsma
Present address: AbbVie Inc., North Chicago, IL, USA

Authors and Affiliations

Department of Biochemistry, Purdue University, West Lafayette, IN, USA
Kankan Wang, Lihong Li & Xing Liu
Center for Plant Biology, Purdue University, West Lafayette, IN, USA
Lihong Li, Yun Zhou & Xing Liu
Department of Chemistry, Purdue University, West Lafayette, IN, USA
Sebastian Kenny & Chittaranjan Das
Department of Applied and Computational Mathematics and Statistics, University of Notre Dame, Notre Dame, IN, USA
Dailin Gan
Division of Biology and Biological Engineering, California Institute of Technology, Pasadena, CA, USA
Justin M. Reitsma
Department of Botany and Plant Pathology, Purdue University, West Lafayette, IN, USA
Yun Zhou

Authors

Kankan Wang
View author publications
Search author on:PubMed Google Scholar
Lihong Li
View author publications
Search author on:PubMed Google Scholar
Sebastian Kenny
View author publications
Search author on:PubMed Google Scholar
Dailin Gan
View author publications
Search author on:PubMed Google Scholar
Justin M. Reitsma
View author publications
Search author on:PubMed Google Scholar
Yun Zhou
View author publications
Search author on:PubMed Google Scholar
Chittaranjan Das
View author publications
Search author on:PubMed Google Scholar
Xing Liu
View author publications
Search author on:PubMed Google Scholar

Contributions

K.W. designed experiments, performed all experiments unless stated otherwise, and prepared the manuscript. L.L. contributed to the FRET-based assays. S.K. and C.D. performed the cryo-EM analyses and contributed to manuscript preparation. D.G. assisted with data analysis of global protein turnover rates. J.M.R generated the CAND2-KO^3xFLAG-CUL1 HEK293 cell line. Y.Z. contributed to the confocal microscopy and manuscript preparation. X.L. conceived the study, designed experiments, prepared the manuscript, and supervised the project. All authors critically reviewed and revised the manuscript.

Corresponding author

Correspondence to Xing Liu.

Ethics declarations

Competing interests

The authors declare no competing interests.

Peer review

Peer review information

Nature Communications thanks Peter Kaiser, Adán Pinto-Fernández, and the other, anonymous, reviewer(s) for their contribution to the peer review of this work. A peer review file is available.

Additional information

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary information

Supplementary Information

Reporting Summary

Description of Additional Supplementary Files

Supplementary Data 1

Supplementary Data 2

Supplementary Data 3

Supplementary Data 4

Transparent Peer Review file

Source data

Source Data

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.

Reprints and permissions

About this article

Cite this article

Wang, K., Li, L., Kenny, S. et al. Molecular mechanisms of CAND2 in regulating SCF ubiquitin ligases. Nat Commun 16, 1998 (2025). https://doi.org/10.1038/s41467-025-57065-5

Download citation

Received: 22 February 2024
Accepted: 11 February 2025
Published: 26 February 2025
DOI: https://doi.org/10.1038/s41467-025-57065-5

This article is cited by

Targeting NOTCH1-KEAP1 axis retards chronic liver injury and liver cancer progression via regulating stabilization of NRF2
- Chunxiao Chang
- Meng Wang
- Shengbin Shi
Journal of Experimental & Clinical Cancer Research (2025)