Article

Recognition of BACH1 quaternary structure degrons

by two F-box proteins under oxidative stress
Graphical abstract Authors
Shiyun Cao, Sheena Faye Garcia,
Huigang Shi, ..., Miklos Guttman,
Michele Pagano, Ning Zheng

Correspondence
[email protected] (M.P.),
[email protected] (N.Z.)

In brief
The stability of BACH1 is controlled by
two F-box proteins, FBXO22 and FBXL17.
There are two forms of the BACH1
quaternary structure degrons that are
encoded by its dimeric BTB domain.
FBXO22 recognizes the intact dimeric
BTB domain, whereas FBXL17 targets
and remodels the BTB domain dimer
when destabilized by S-nitrosylation.

Highlights
d FBXO22 and FBXL17 play non-redundant roles in regulating
the stability of BACH1

d FBXO22 binds a BACH1 quaternary structure degron

presented by its dimeric BTB fold

d S-nitrosylation destabilizes the BACH1 BTB domain dimer

and prevents FBXO22 binding

d Destabilized BACH1 BTB domain dimer is recognized and

remodeled by a pair of FBXL17

Cao et al., 2024, Cell 187, 1–17

December 26, 2024 ª 2024 The Author(s). Published by Elsevier Inc.
https://doi.org/10.1016/j.cell.2024.10.012 ll
 Please cite this article in press as: Cao et al., Recognition of BACH1 quaternary structure degrons by two F-box proteins under oxidative
stress, Cell (2024), https://doi.org/10.1016/j.cell.2024.10.012

ll
OPEN ACCESS

Article
Recognition of BACH1 quaternary structure degrons
by two F-box proteins under oxidative stress
Shiyun Cao,1,2,11 Sheena Faye Garcia,3,4,11 Huigang Shi,1,2,11 Ellie I. James,5,6 Yuki Kito,3,4 Hui Shi,1,2,12 Haibin Mao,1,2
Sharon Kaisari,3,4 Gergely Rona,3,4,7,8 Sophia Deng,3,4 Hailey V. Goldberg,3,4 Jackeline Ponce,3,4,9 Beatrix Ueberheide,3,4,9
Luca Lignitto,3,4,10 Miklos Guttman,5,6 Michele Pagano,3,4,7,* and Ning Zheng1,2,13,*
1Department of Pharmacology, University of Washington, Box 357280, Seattle, WA 98195, USA
2Howard Hughes Medical Institute, University of Washington, Seattle, WA 98195, USA
3Department of Biochemistry and Molecular Pharmacology, New York University Grossman School of Medicine, New York, NY, 10016, USA
4Laura and Isaac Perlmutter Cancer Center, New York University Grossman School of Medicine, New York, NY 10016, USA
5Department of Medicinal Chemistry, University of Washington, Seattle, WA 98195, USA
6Molecular Engineering & Science Institute, University of Washington, Seattle, WA 98195, USA
7Howard Hughes Medical Institute, New York University Grossman School of Medicine, New York, NY 10016, USA
8Institute of Molecular Life Sciences, HUN-REN Research Centre for Natural Sciences, Budapest, Hungary
9Proteomics Laboratory, Division of Advanced Research Technologies, New York University Grossman School of Medicine, New York,

NY 10016, USA
10Cancer Research Center of Marseille (CRCM), CNRS, Aix Marseille University, INSERM, Institut Paoli-Calmettes, Marseille, France
11These authors contributed equally
12Present address: Biortus Discovery Co., Ltd., 101 South Building, 99-3 Linhu Avenue, Xinwu District, Wuxi, Jiangsu, P.R. China
13Lead contact

*Correspondence: [email protected] (M.P.), [email protected] (N.Z.)

https://doi.org/10.1016/j.cell.2024.10.012

SUMMARY

Ubiquitin-dependent proteolysis regulates diverse cellular functions with high substrate specificity, which
hinges on the ability of ubiquitin E3 ligases to decode the targets’ degradation signals, i.e., degrons. Here,
we show that BACH1, a transcription repressor of antioxidant response genes, features two distinct uncon-
ventional degrons encrypted in the quaternary structure of its homodimeric BTB domain. These two degrons
are both functionalized by oxidative stress and are deciphered by two complementary E3s. FBXO22 recog-
nizes a degron constructed by the BACH1 BTB domain dimer interface, which is unmasked from transcrip-
tional co-repressors after oxidative stress releases BACH1 from chromatin. When this degron is impaired by
oxidation, a second BACH1 degron manifested by its destabilized BTB dimer is probed by a pair of FBXL17
proteins that remodels the substrate into E3-bound monomers for ubiquitination. Our findings highlight the
multidimensionality of protein degradation signals and the functional complementarity of different ubiquitin
ligases targeting the same substrate.

INTRODUCTION number of known degrons is still far below the predicted num-
ber of E3 ligases in most, if not all, eukaryotes.
Ubiquitin-dependent protein degradation is a widespread The cullin-RING ubiquitin ligase (CRL) complexes constitute
mechanism regulating virtually every cellular function.1 The the largest superfamily of multi-subunit E3s in humans with
selectivity of this proteolytic process is exquisitely controlled more than 200 family members classified in five subfamilies:
by the specific degradation signal, also known as ‘‘degron,’’ CRL1 (also known as SCF for SKP1, CUL1, F-box protein),
which is encoded in the substrate protein and perceived by CRL2, CRL3, CRL4A/B, and CRL5.6 Built with a modular archi-
its cognate E3 ubiquitin ligase.2,3 Early studies have mapped tecture, each CRL subfamily employs a battery of substrate re-
the degrons of many classical substrate proteins to various ceptor (SR) subunits to recognize the unique degrons of their
short linear ‘‘motifs’’ (SLiMs) with specific consensus se- targets. For example, the CRL1/SCF SRs, b-TrCP and
quences.4 Many of these degrons are recognized by E3 ligases FBXW7, are two F-box proteins well known to recognize phos-
in their native form, while others require modulation by either phorylated SLiM degrons.7,8 A panel of BC-box and DCAF pro-
post-translational modifications or hormonal signal to promote teins act as the CRL2 and CRL4 SRs to recognize various
E3 binding. Although recent proteome-wide studies have C-end degrons found at the extreme C terminus of substrate
further expanded the cellular repertoire of SLiM degrons,5 the polypeptides.9,10

Cell 187, 1–17, December 26, 2024 ª 2024 The Author(s). Published by Elsevier Inc. 1
This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).
 Please cite this article in press as: Cao et al., Recognition of BACH1 quaternary structure degrons by two F-box proteins under oxidative
stress, Cell (2024), https://doi.org/10.1016/j.cell.2024.10.012

ll
OPEN ACCESS Article

CRLs also closely participate in redox homeostasis regulation. nature of the BACH1-BTB domain by size-exclusion chromatog-
Oxidative stress results from an accumulation of oxygen and ni- raphy coupled with multi-angle light scattering (SEC-MALS) (Fig-
trogen-derived reactive species (ROS/RNS), which cause direct ure 1C). Despite its expected symmetric architecture, the
damage to cellular components and interfere with redox BACH1 BTB dimer appeared to be associated with only a single
signaling.11,12 To modulate the levels of prooxidant and anti- copy of FBXO22-SKP1.
oxidant species, mammalian cells have evolved two antago- To reveal the mechanism by which FBXO22 recognizes
nizing redox-sensitive transcription factors to dynamically con- BACH1-BTB, we used cryoelectron microscopy (cryo-EM) to
trol the expression of cytoprotective genes. BACH1, also known determine the structure of the SKP1-CUL1-RBX1-FBXO22 com-
as BTB (broad-complex, tramtrack, and bric-à-brac) and CNC plex (also known as SCFFBXO22) bound to the BACH1 BTB dimer
(cap ‘‘n’’ collar) homology 1, is a master transcriptional repressor at 3.9 Å resolution (Figure S1; Table S1). The SCFFBXO22-BACH1
in heme metabolism and redox homeostasis.13 As a member of complex adopts a canonical SCF architecture31 in which the
the basic leucine zipper (bZIP) protein family, BACH1 heterodi- FBXO22-SKP1 SR module, together with the BACH1 substrate,
merizes with small musculoaponeurotic fibrosarcoma (sMAF) is docked to the N-terminal half of the cullin scaffold (Figure 1D).
proteins and binds to the anti-oxidant response elements Unique among F-box proteins, FBXO22 features a C-terminal
(AREs) of many anti-oxidant genes, including heme oxygenase FIST (F-box and intracellular signal transduction proteins)
1 (HMOX1), to suppress their transcription under normal condi- domain responsible for substrate recruitment (Figure 1A).
tions.14–18 Under oxidative stress, a critical transcription acti- Consistent with our SEC-MALS analysis, the F-box protein forms
vator, NRF2 (nuclear factor erythroid-derived 2-like 2), accumu- an asymmetric complex with the BACH1 BTB dimer at a 1:1
lates and competes with BACH1 for binding sMAF and AREs. In molar ratio (Figure 1E).
doing so, NRF2 induces the expression of these genes, whose The FIST domain is found in eukaryotic F-box proteins and pro-
products go on to eliminate reactive species and restore redox karyotic proteins involved in signal transduction.32 It has been pro-
homeostasis.19–22 posed to function as a sensory domain, possibly binding small
A wealth of studies revealed a crucial role of the CRL3 SR molecular ligands, although its fold was previously uncharacter-
KEAP1 (Kelch-like ECH-associated protein 1) in regulating ized. Our structure reveals that the FBXO22 FIST domain is con-
NRF2 activity in a redox-sensitive manner.23 Under steady-state structed by three structural repeats, each consisting of a central
conditions, KEAP1 recognizes two SLiM degrons of NRF2 and four-stranded b-sheet sandwiched by two a-helices on one side
promotes its ubiquitination and degradation.24–26 When cells and a b-hairpin loop on the other (Figure 1F). The central b-sheet
are challenged by oxidative or electrophilic stress, covalent from each repeat packs against the other two repeats, giving rise
modification of highly reactive cysteine residues in KEAP1 im- to a compact globular fold with a pseudo 3-fold symmetry. A Dali
pairs its E3 ligase activity, enabling newly synthesized NRF2 to search identified two classes of unrelated metabolic enzymes,
rapidly accumulate and move to the nucleus to induce the chorismatases and cyanuric acid hydrolases, with a similar overall
expression of anti-oxidant genes.27,28 Recent studies have iden- fold, which could have evolved from monomers of the trimeric
tified two F-box proteins, FBXO22 and FBXL17, which mirror the YjgF superfamily (Figure S2A).33–35 Different from the metabolic
function of KEAP1 to regulate the stability of BACH1 in response enzymes, the FIST domain of FBXO22 does not feature any
to elevated heme and ROS/RNS levels.29,30 Both F-box proteins obvious ligand-binding pocket. We name the three structural re-
have been implicated in oxidative stress-induced degradation of peats of the FIST domain, FIST-1, FIST-2, and FIST-3. Noticeably,
BACH1. However, how FBXO22 and FBXL17 recognize BACH1 FIST-3 is distinguished from the other two repeats by having a
and together regulate its stability remains unknown. longer b-hairpin loop, whose two b-strands, b15 and b16,
In this study, we reveal that the stability of BACH1 is dictated perfectly align with the edge of the central b-sheet and extend it
by two mutually exclusive degrons, both of which are encoded in into a six-stranded sheet (Figures 1F and S2A).
the quaternary structures of its dimeric BTB domain. The two
BACH1 degrons are functionalized by different forms of oxida- BACH1 quaternary structure degron
tive stress and recognized by FBXO22 and FBXL17 via distinct Similar to most BTB dimer structures, the BACH1 BTB domain
and complementary mechanisms. By shedding light on the mo- dimer is characterized by a domain swapping architecture, in
lecular basis of BACH1 regulation, our results define a previously which the two protomers exchange their N-terminal
undescribed class of protein degradation signals, which exploit b-strands.36,37 In BACH1, the N-terminal b-strand (b1) of each
the structural complexity of protein assembly to achieve sub- protomer runs in antiparallel with the C-terminal a-helix (CTH)
strate specificity. (a6) of the opposite protomer, extending the dimer interface
beyond the core BTB domain (Figure 1G). Remarkably, one of
RESULTS these cross-protomer b1-a6 pairs within the BACH1-BTB dimer
makes up the entire interface with the F-box protein by clinging
Structure of the SCFFBXO22-BACH1-BTB complex to the edge of FBXO22 FIST-3 (Figure 2A).
BACH1 contains an N-terminal BTB domain, which mediates its The tri-molecular interface between FBXO22 and the BACH1-
homodimerization and is critical for FBXO22 interaction (Fig- BTB dimer is stabilized by a hybrid of polar and hydrophobic in-
ure 1A).30 In a GST pull-down assay with purified recombinant teractions. Specifically, the b1 strand of one BACH1-BTB chain
proteins, we validated the direct binding between the BACH1 (BTB-B), together with the b5 strand of the second BACH1 BTB
BTB domain and the full-length FBXO22 protein in complex chain, is juxtaposed with the b15-b16 hairpin of FBXO22 FIST-3,
with SKP1 (Figure 1B). We further confirmed the homodimeric creating an intermolecular anti-parallel b-sheet supported by a

2 Cell 187, 1–17, December 26, 2024

 Please cite this article in press as: Cao et al., Recognition of BACH1 quaternary structure degrons by two F-box proteins under oxidative
stress, Cell (2024), https://doi.org/10.1016/j.cell.2024.10.012

ll
Article OPEN ACCESS

A B C
BACH1 - + + GST-FBXO22
BTB LZ + + - + BACH1-BTB 0.08 FBXO22-SKP1
1 130 736
75 GST-FBXO22 62 kDa (± 4%)
FBXO22 63

UV (AU)
0.06
48 BACH1-BTB
F FIST 35
1 25 SKP1 0.04 FBXO22-SKP1 32 kDa (± 2%)
403 -BACH1-BTB
FBXL17 11
BACH1-BTB 0.02 88 kDa (± 6%)
F LRRs
1 701 0.00
GST pull down 5 10 15 20
Volume (ml)
D E
FBXO22
BACH1-BTB(A)
SKP1 FBXO22
BACH1-BTB(A)
BACH1-BTB(B)

RBX1 BACH1-BTB Homodimer

CUL1 BACH1-BTB(B)

F G
FIST-1
FIST
T1 C
N
FIST-2
o α6
90
9
β5
FIST-3 β1

FIST-3
α10 BACH1-BTB
Homodimer
β15

β16
β1 β5
FIST-1
FIST-2 α6
FBXO22 F
FIST Domain
omain C
N

Figure 1. Overall structure of the FBXO22-BACH1-BTB complex

(A) Domain composition of BACH1, FBXO22, and FBXL17.
(B) BACH1-BTB pull-down by GST-FBXO22.
(C) SEC-MALS analyses of BACH1-BTB, FBXO22-SKP1, and their complex with experimentally determined molecular weights. Monomeric BACH1-BTB and
FBXO22-SKP1 have a theoretical molecular weight of 17 and 63 kDa, respectively.
(D) Structure model of SCF-FBXO22-BACH1-BTB fitted in the electron microscopy map. The RBX1 RING domain is not shown due to its flexibility.
(E) Electron microscopy map fit with the FBXO22-BACH1-BTB complex.
(F) Orthogonal views of the FBXO22 FIST domain with three repeats (FIST-1, 2, and 3).
(G) Structure of the BACH1 BTB domain dimer with select secondary structure elements indicated.
See also Figures S1 and S2 and Table S1.

network of backbone hydrogen bonds (Figure 2B). These sec- by two salt bridges and intermolecular van der Waal packing
ondary structure interactions are cemented by an adjacent hy- made by additional residues from the F-box protein (Asp366,
drophobic core, which is nucleated by four amino acids from Arg308, Lys377, and Arg376) and the two protomers of the sub-
the BACH1-BTB b1-a6 pair (Phe9, Tyr11, Cys122, and Phe125) strate (Glu12 and Phe128).
and three residues in the FBXO22 b15-b16 hairpin (Ile368, In support of a critical role played by the N-terminal b1 strand
Phe373, and Leu375). Peripheral to the hydrophobic core, the of BACH1-BTB at the interface, previous studies have shown
FBXO22-BACH1-BTB dimer interface is further strengthened that individual alanine mutations of its two aromatic residues,

Cell 187, 1–17, December 26, 2024 3

 Please cite this article in press as: Cao et al., Recognition of BACH1 quaternary structure degrons by two F-box proteins under oxidative
stress, Cell (2024), https://doi.org/10.1016/j.cell.2024.10.012

ll
OPEN ACCESS Article

A B C

D G H

F I J

K L M

Figure 2. Recognition of the BACH1-BTB quaternary structure degron by FBXO22

(A) Recognition of the BACH1 cross-protomer b1-a6 degron by FBXO22. BACH1-BTB(A) and BACH1-BTB(B) are shown as cartoons and colored in cyan and
orange, respectively.
(B) Close-up view of the FBXO22-BACH1-BTB interface with key residues shown in sticks.
(C) HA immunoprecipitation of BACH1 a6 mutants. HEK293T cells were transfected with either empty vector (EV) or HA-BACH1 wild type (WT) or mutants. Cells
were then treated with MLN-4924 prior to HA immunoprecipitation. Proteins were immunoblotted as indicated.

(legend continued on next page)

4 Cell 187, 1–17, December 26, 2024

 Please cite this article in press as: Cao et al., Recognition of BACH1 quaternary structure degrons by two F-box proteins under oxidative
stress, Cell (2024), https://doi.org/10.1016/j.cell.2024.10.012

ll
Article OPEN ACCESS

Phe9 and Tyr11, are sufficient to abolish FBXO22-BACH1 inter- Regulation of FBXO22-BACH1 interaction
actions.30 Similarly, replacing Phe125 in the BACH1 a6 helix with Using biolayer interferometry (BLI), we determined the affinity of
alanine or aspartate effectively abrogates E3-substrate complex FBXO22 toward BACH1-BTB, which is
90 nM and on par with
formation (Figure 2C). Therefore, both secondary structure ele- other CRL-substrate interactions (Figure 2G).42 Consistent with
ments of the BACH1-BTB b1-a6 pair, which is shaped by the the lack of any obvious ligand-binding pocket in FBXO22 or
quaternary structure of the BTB dimer, are necessary to define BACH1-BTB, the binding between the two proteins is insensitive
the functional degron. In a second set of mutational analyses, to heme (Figure 2H). Heme, therefore, does not appear to be
we validated that the majority of the FBXO22 residues at the physically required for the productive interaction between
tri-molecular interface are indispensable for binding BACH1, FBXO22 and BACH1 (Figures S2C and S2D) and most likely pro-
suggesting that the edge of the third FBXO22 FIST repeat repre- motes BACH1 recognition by FBXO22 through an indirect mech-
sents the primary docking site for BACH1 (Figure 2D). Notice- anism. In previous studies, heme has been reported to block the
ably, although the symmetric BACH1 BTB domain dimer con- DNA-binding activity of BACH1-sMAF heterodimers indepen-
tains two cross-protomer b1-a6 pairs, two FBXO22 molecules dent of the BACH1 BTB domain.19 This raises the possibility
cannot simultaneously recognize the two identical BACH1 de- that the quaternary structure degron of BACH1 might not be
grons without a severe steric collision (Figure S2B). A single accessible to FBXO22 until the transcription factor is released
SCFFBXO22 E3 complex, therefore, appears responsible for ubiq- from the DNA.
uitinating both protomers of a BACH1 dimer via their BTB do- The BTB domain is widely found in CRL3 SRs and transcrip-
mains (Figures S2C and S2D). tion factors.37 Besides mediating protein oligomerization, the
Mammalian BACH1 has a close paralog, BACH2, which is pri- BTB domain of select transcription repressors, such as PLZF
marily expressed in the brain and spleen with well-documented and BCL6, has been reported to interact with co-repressors.43,44
roles in regulating immune cell differentiation.38–40 The BTB do- In the case of BCL6, distinct SLiMs of SMRT/NCOR2 and BCOR
mains of the two BACH proteins, including their terminal regions, co-repressors have been shown to form the same interface with
are overall 62% identical in sequence (Figure 2E). Their homo- the BCL6 BTB domain dimer, which is highlighted by an inter-
dimer structures can be superimposed with a root mean square molecular b-sheet involving the N-terminal b1-strand of BCL6-
deviation (RMSD) of 0.6 Å (Figure S2E).36,41 Despite their BTB (Figure 2I).45,46 Such a binding mode is remarkably analo-
sequence and structural similarity, only BACH1 but not BACH2 gous to the FBXO22-BACH1-BTB interface (Figure 2J). Similar
can be recognized by FBXO22 (Figure 2F). To identify key ele- to PLZF and BCL6, BACH1 has been previously shown to be
ments in BACH1 conferring this specificity, we made chimeric functionally associated with nuclear co-repressor NCOR1, a
BACH proteins and tested their interactions with the F-box pro- close paralog of SMRT/NCOR2.47 This prompted us to test
tein. Interestingly, replacing Phe9 in the N-terminal b1 strand of whether NCOR1 can compete with FBXO22 for binding the
BACH1 with its counterpart in BACH2 (Tyr12) had little effect BTB domain of BACH1. A central region of NCOR1 has been pre-
on FBXO22 binding. By contrast, swapping the C-terminal a6 he- viously identified to interact with the BTB domain of multiple
lices of the two BACH proteins effectively switched their ability to functionally unrelated transcriptional co-repressors.43 Sequence
engage the E3 (Figure 2F). A close examination of the BACH2 alignment and structure prediction analyses indicate that this re-
BTB domain structure reveals a non-helical nature of its C-termi- gion contains several potential SLiMs conserved among NCOR1
nal region (Figure S2E) distinct from the a6 helix of BACH1. This orthologs, including the one used by NCOR2 for binding BCL6-
structural difference provides a possible explanation for BTB (Figure S2F). When co-expressed and co-purified from
BACH2’s inability to bind FBXO22. E. coli, this
230 amino acid NCOR1 fragment was indeed suf-
Together, these results reinforce the importance of the struc- ficient to form a complex with the BACH1 BTB domain (Fig-
tural integrity of the b1-a6 degron of the BACH1 BTB dimer ure 2K). Importantly, the NCOR1 fragment was able to effectively
for FBXO22 recognition and underscore the capacity of the qua- block FBXO22-BACH1 binding in an AlphaScreen-based
ternary structure degron in conferring high E3-substrate selec- competition assay (Figure 2L), suggesting that the transcriptional
tivity, a previously undescribed characteristic of CRL substrates. co-repressor can occupy the quaternary structure degron of

(D) FLAG immunoprecipitation of FBXO22 mutants. HEK293T cells were transfected with either EV or FLAG-FBXO22 WT or mutants. Cells were then treated with
MLN-4924 prior to FLAG immunoprecipitation. Proteins were immunoblotted as indicated. Non-specific protein bands are labeled with asterisks.
(E) Sequence alignment of the N- and C-terminal regions of BACH1 and BACH2 BTB domains.
(F) FLAG immunoprecipitation of BACH1-BACH2 chimeras. HEK293T cells were transfected with either EV or FLAG-BACH1/BACH2 chimeras as indicated. Cells
were then treated with MG-132 prior to FLAG immunoprecipitation. Proteins were immunoblotted as indicated. Non-specific protein bands are labeled with
asterisks.
(G) Biolayer interferometry (BLI) measurements of FBXO22 and BACH1-BTB interaction. Kd-equilibrium, dissociation constant.
(H) BACH1-BTB pull-down by GST-FBXO22 in the presence and absence of 10 mM hemin. Non-specific protein bands are labeled with asterisks.
(I) Structure of BCL6-BTB dimer (green) in complex with the co-repressor SMRT (purple, PDB:1R2B) or BCOR (slate, PDB: 3BIM).
(J) Structure of BACH1-BTB dimer in complex with FBXO22 b15-b16 hairpin (purple).
(K) BACH1-BTB pull-down by GST-NCOR1.
(L) Inhibition of the FBXO22-BACH1-BTB interaction by NCOR1 fragments measured in AlphaLISA competition assays. Data are presented as mean ± SD. IC50,
half-maximum inhibitory concentration.
(M) Inhibition of SCFFBXO22-catalyzed BACH1 ubiquitination by an
230 amino acid NCOR1 fragment.
See also Figure S2.

Cell 187, 1–17, December 26, 2024 5

 Please cite this article in press as: Cao et al., Recognition of BACH1 quaternary structure degrons by two F-box proteins under oxidative
stress, Cell (2024), https://doi.org/10.1016/j.cell.2024.10.012

ll
OPEN ACCESS Article

BACH1. Interestingly, when the N-terminal and C-terminal the NOR3-S1PC-treated BACH1 BTB domain still retains its ho-
halves of the NCOR1 central fragment were individually tested, modimeric form in solution (Figure S3I), it can no longer interact
each sub-fragment was able to inhibit the interaction between with FBXO22 in a Ni2+ NTA pull-down assay (Figure 3F). By
FBXO22 and the BACH1 BTB dimer, albeit at a much lower po- contrast, FBXL17 showed robust activity in binding the NOR3-
tency (Figure 2L). Therefore, NCOR1 might employ more than S1PC-treated BACH1 BTB domain (Figure 3G). Likewise, NO
one SLiM to simultaneously engage the two identical FBXO22- treatment reduced the amount of FBXO22 that can be co-immu-
binding sites on BACH1-BTB to block the ubiquitination of the noprecipitated by BACH1, while the opposite is true for FBXL17
transcription repressor (Figure 2M). Together, these results imply (Figures 3H and 3I). These results suggest that FBXL17 might
that the NCOR1 co-repressor can shield the quaternary structure have been evolved to recognize and ubiquitinate BACH1
degron of chromatin-bound BACH1 from FBXO22 until BACH1 is when BACH1 is in a form incompatible with FBXO22 binding
released from DNA by heme. (Figures S3J and S3K).
Interestingly, treatment with either an NO donor or hemin
Degradation of S-nitrosylated BACH1 by FBXO22 and induced the release of BACH1 from chromatin (Figure 3J).
FBXL17 Accordingly, sNONO reduced the binding of BACH1 to MafF
FBXL17 is a leucine-rich repeats (LRRs)-containing F-box pro- and NCOR1 in the cell (Figure 3K). Similar to hemin,19 an NO donor
tein (Figure 1A) that, similarly to FBXO22, has been implicated can effectively abrogate the ability of the BACH1 bZIP domain to
in the degradation of BACH1 in response to heme-mediated super-shift sMAF-bound DNA in an electrophoretic mobility shift
oxidative stress.29 Furthermore, it has been proposed that assay (Figure S3L). These results implicate that, in addition to
FBXL17 can recognize non-functional heterodimers of BTB pro- its BTB domain, the bZIP DNA-binding domain of BACH1, which
teins and sequester their monomeric BTB domains to ubiquiti- is rich in cysteine residues, might also be susceptible to
nate and degrade the mismatched substrate proteins.48,49 S-nitrosylation. We speculate that a subpopulation of the
How FBXL17 captures BACH1 upon oxidative stress, presum- BACH1 proteins, depending on their cellular context, might be
ably by recognizing its homodimeric BTB domain, remains un- modified by S-nitrosylation within their bZIP domain but not the
clear. We first investigated the functional relationship between BTB domain, which could explain why BACH1-FBXO22 complex
FBXL17 and FBXO22 in regulating BACH1 stability in the context could still be detected in NO-treated cells (Figure 3H) and, in com-
of BACH1 S-nitrosylation. Nitric oxide (NO) is a free radical in the parison to the parental cells, BACH1 was partially stabilized in
cell that is either produced endogenously or exogenously.50 By FBXO22 KO cells despite NO treatment (Figures 3B–3D).
reacting with cysteine residues of selective targets, NO induces
protein S-nitrosylation and further disulfide formation, which FBXL17 preferentially remodels NO-modified BACH1-
contribute to redox signaling that protects the cell against oxida- BTB dimer
tive stress.51,52 NOR3, an NO donor, has recently been shown From the single time point (30 min) affinity pull-down assay, we
to induce BACH1 degradation, an effect that can be augmented noticed that FBXL17 preferentially interacts with the NO-modi-
by a garlic-derived sulfur amino acid, S-1-propenylcysteine fied BACH1 BTB domain over the native polypeptide (Figure 3G).
(S1PC).53 We first confirmed by biotin switch assay that To further dissect the activity of FBXL17 in BACH1-BTB recogni-
BACH1 can be modified by two additional NO donors, tion, we used BLI to monitor the kinetics of their interaction with
S-nitrosoglutathione (GSNO) and spermine NONOate (sNONO) or without treating BACH1-BTB with NOR3-S1PC. In agreement
(Figure 3A). Using FBXL17 and FBXO22 individual and double with the affinity pull-down results, compound-treated BACH1-
knockout (KO) HCT-116 cells (Figures S3A and S3B), we verified BTB interacted strongly with FBXL17 with an ‘‘apparent Kd’’ of
that both FBXO22 and FBXL17 contribute to BACH1 turnover at
110 nM, whereas the native BACH1-BTB protein only showed
steady state, although FBXO22 appears to play a predominant a detectable affinity toward the F-box protein with an apparent
role (Figures S3C–S3E). By contrast, in response to NO treat- Kd greater than 2.4 mM (Figures 4A and 4B). Noticeably, with
ment, BACH1 degradation is similarly slowed down when either compound-treated BACH1-BTB immobilized on the probe, the
FBXO22 or FBXL17 is missing (Figures 3B–3D, S3F, and S3G). maximum binding signal appeared far from plateauing within
The two F-box proteins, therefore, appear to play nonredundant the time window of the measurement, even though the amplitude
roles in regulating BACH1 stability in response to stress. of the association curves started to show saturation when the
The BTB domain of BACH1 contains five cysteine residues, FBXL17 concentration reached the mM range (Figure 4A). The
two of which, Cys107 and Cys122, are located at the b1-a6 de- same phenomenon held true when we lengthened the time win-
gron recognized by FBXO22 (Figure 3E). Consistent with a recent dow to 10 min (Figure S4A). This feature hinted at a minimally
proteome-wide S-nitrosylation site mapping study,56 mass biphasic nature of the interaction, which might involve complex
spectrometry (MS) analysis of the recombinant BACH1 BTB rearrangement in addition to simple binding equilibrium.
domain treated with the NO donor NOR3-S1PC identified To further characterize the interaction between FBXL17 and
Cys107 but not Cys55 as an S-nitrosylation site (Figure S3H). compound-treated BACH1-BTB, we used SEC-MALS to analyze
Due to incomplete sequence coverage, whether Cys34, samples of the two proteins mixed for different periods of time. In
Cys109, and Cys122 could be modified by NO remained unclear. congruence with the reported activity of FBXL17 in forming stable
As a residue on the a5 helix of BACH1-BTB making direct con- complexes with monomeric BTB domains,48 the F-box protein
tact with the a6 helix (Figure 3E), covalent modification of was found to bind a single copy of BACH1-BTB after the two pro-
Cys107 is expected to perturb the local structure of the b1-a6 teins were mixed for 24 h (Figure 4C). Interestingly, it took a rela-
degron, thereby impairing FBXO22 binding. Indeed, even though tively long incubation time for such a stable 1:1 complex to form,

6 Cell 187, 1–17, December 26, 2024

 Please cite this article in press as: Cao et al., Recognition of BACH1 quaternary structure degrons by two F-box proteins under oxidative
stress, Cell (2024), https://doi.org/10.1016/j.cell.2024.10.012

ll
Article OPEN ACCESS

A HCT-116 B C
sNONO

HCT-116
GSNO

FBXL17 + FBXO22 KO
FBXO22 KO
NT FBXL17 +
WT FBXL17 KO FBXO22 KO FBXO22 KO FBXL17 KO
0 1 2 3 4 5 0 1 2 3 4 5 0 1 2 3 4 5 0 1 2 3 4 5 sNONO (h)
BACH1
AP: biotin

BACH1
KEAP1 (short exp.)
BACH1
SKP1 (long exp.)
Parental
FBXO22
BACH1
WCE

KEAP1 Nrf2 sNONO (h)

SKP1 GAPDH
F Ni2+ pull down
- - + NOR3-S1PC
Parental FBXO22 KO
C
- + + BACH1-BTB
D FBXL17 KO FBXL17 + FBXO22 KO E N kDa + + + FBXO22-SKP1
α6 α5 BACH1-BTB(A) 100
C107 75
C109 63 GST-FBXO22
48
35
GST-BACH1
25
-BTB
C122 20 His-SKP1
β1
G - - + - NOR3-S1PC
- + + + BACH1-BTB
C55
C34
kDa + + + - FBXL17-SKP1
100
75 MBP-FBXL17
BACH1- 63 *
48 **GST-BACH1
BTB(B)
35
CHX (h) sNONO + CHX (h) -BTB
25
20 His-SKP1
HEK293T
Ni2+ pull down
H
(MLN4924)
EV + + FLAG-BACH1 I HEK293T J K
- + sNONO HCT-116 HEK293T
EV

- + + + HA-BACH1-BTB soluble fraction chromatin fraction

FBXO22 + - + + FLAG-FBXL17 WCE IP: α-HA
- - - + sNONO DMSO MG132 DMSO MG132
- + - - + - - + - - + - hemin + + - - + + - - EV
IP: BACH1 FBXL17 (α-FLAG) - - + + - - + + HA-BACH1
(α-FLAG) - - + - - + - - + - - + sNONO - + - + - + - + sNONO
α-FLAG IP: BACH1-BTB (α-HA)
GAPDH α-HA BACH1 BACH1
(α-HA)
actin
MafF *
NCoR1
FBXO22
FBXL17 (α-FLAG)
BACH1 tubulin MafF
WCE WCE
(α-FLAG) BACH1-BTB (α-HA)
* histone H3 PCNA
GAPDH actin

Figure 3. Degradation of BACH1 by FBXO22 and FBXL17

(A) Biotin switch assay to assess BACH1 S-nitrosylation. HCT-116 cells were treated with either DMSO (NT) or NO donors (GSNO and sNONO) before lysates
were subjected to S-nitrosylation biotin switch. KEAP1 and SKP1 were included as positive and negative controls, respectively.54,55 Biotinylated proteins were
purified and immunoblotted as indicated.
(B and C) Assessing the stability of endogenous BACH1 in parental, KO cells upon NO treatment. Proteins were immunoblotted as indicated. The graph in
(C) shows the quantification of BACH1 protein levels relative to that at 0 h.
(D) Quantification of BACH1 protein levels measured in CHX chase assays in HCT-116 parental KO cells with/without NO treatment. Data are presented as
mean ± SD. The graph is created based on data shown in Figures S3F and S3G.
(E) Five cysteine residues in BACH1-BTB with their side chains shown in dot spheres.
(F and G) Effects of BACH1-BTB S-nitrosylation on FBXO22 and FBXL17 binding investigated by Ni2+ pull-down assay. Non-specific protein bands are labeled
with asterisks.
(H) Co-immunoprecipitation of FBXO22 by FLAG-BACH1 with/without NO treatment.
(I) Co-immunoprecipitation of FBXL17 by HA-BACH1 with/without NO treatment.
(J) Western blots monitoring the amount of endogenous BACH1 in the soluble and chromatin fractions of HCT-116 cells upon hemin or NO treatment.
(K) HEK293T cells were transfected with empty vector (EV) or HA-BACH1. Co-immunoprecipitation of NCOR1 and MafF by HA-BACH1 with/without NO
treatment. Proteins were immunoblotted as indicated.
See also Figure S3.

Cell 187, 1–17, December 26, 2024 7

 Please cite this article in press as: Cao et al., Recognition of BACH1 quaternary structure degrons by two F-box proteins under oxidative
stress, Cell (2024), https://doi.org/10.1016/j.cell.2024.10.012

ll
OPEN ACCESS Article

A B
Apparent Probe: Biotin-BACH1-BTB Probe: Biotin-BACH1-BTB
0.5 (NOR3-S1PC treated) 0.5
Kd = 110 nM (Untreated)
Analyte: FBXL17 Analyte: FBXL17
0.4 0.4

Response (nm)
Response (nm)

[FBXL17]
2000 nM 0.3
0.3 Apparent Kd > 2.4 μM
666 nM
0.2 222 nM 0.2 [FBXL17]
74 nM 2000 nM
0.1 25 nM 0.1 666 nM
222 nM
0.0 0.0 74 nM
0 100 200 300 400 0 100 200 300 400
Time (s) Time (s)

C D Biotin pull down Biotin pull down

0.5 2 10 26 53 0.5 1 2 21 28 Incubation time (h)
FBXL17-SKP1- FBXL17-SKP1
+ + + + + + + + + + + + - Biotin-BACH1-BTB
BACH-BTB 105kDa BACH1-BTB treated + + + + + + + + + + + - + MBP-FBXL17-SKP1
128kDa (± 8%) (± 2%) with NOR3-S1PC kDa
100 MBP-FBXL17
0.03 FBXL17-SKP1 75
Mixed for 1 min 63
48
Mixed for 30 mins 35
25
Mixed for 24 hrs 20 SKP1
0.02 17
Biotin-BACH1-BTB
UV (AU)

BACH1-BTB 11
treated Untreated NOR3-S1PC treated BACH1-BTB
34kDa
Binding signal (%)

120

Binding signal (%)

120
0.01
(±25%) 100 100
80 80
60 60
0.00 40 t1/2 > 14 h 40 t1/2 = ~1 h
6 9 12 15 18 20 20
0 0
Volume (ml) 0 10 20 30 40 50 60 0 10 20 30
Time (h) Time (h)

E Probe: Biotin-BACH1-BTB
F
0.5
0.5 Probe: Biotin-BACH1-BTB
(Untreated)
0.4 (NOR3-S1PC treated)
Analyte: FBXL17 400nM
Response (nm)

0.4 Analyte: FBXL17 400nM

Response (nm)

WT
0.3 C55A 0.3 WT
0.2 C107A
0.2 C55A
C109A
0.1 C107A
C122A
0.1 C109A
0.0 C122A
0 100 200 300 400 0.0
Time (s) 0 100 200 300 400
Time (s)

Figure 4. FBXL17 preferentially remodels S-nitrosylated BACH1-BTB

(A and B) BLI measurements of FBXL17 and BACH1-BTB interaction with and without NOR3-S1PC treatment. The apparent Kd-equilibrium values are indicated.
(C) SEC-MALS analyses of NO-treated BACH1-BTB, FBXL17-SKP1, and their mixture at various incubation time with the experimentally determined molecular
weights. The theoretical molecular weights of MBP-FBXL17-SKP1 and monomeric BACH1-BTB are 108 and 17 kDa, respectively. A larger amount of protein
sample was applied to BACH1-BTB alone run (gray).
(D) The time-courses of the interactions between FBXL17 and BACH1-BTB with or without NOR3-S1PC treatment monitored by affinity pull-down. t1/2: time
needed for 50% complex formation.
(E and F) BLI measurements of the interaction between FBXL17 and BACH1-BTB (WT and mutants) with/without NOR3-S1PC treatment. The effect of compound
treatment was diminished by C107A and C122A but was enhanced by C109A, which by itself is insufficient to stimulate FBXL17 binding.
See also Figures S3 and S4.

even though BLI could instantaneously detect the interaction be- To fully assess the binding kinetics of FBXL17 and BACH1-
tween FBXL17 and BACH1-BTB. Given that compound-treated BTB, we returned to the affinity pull-down assay and monitored
BACH1-BTB remains a stable dimer even after being diluted to the interaction with a time window up to 1–2 days. As expected

3 mM (Figure S3I), we speculate that the biphasic binding curve from our BLI and SEC-MALS results, it took about 1 h for half of
detected by BLI reflects both the initial contact between FBXL17 the maximal amount of FBXL17 to become stably associated
and dimeric BACH1-BTB and the final complex formation be- with compound-treated BACH1-BTB under the tested experi-
tween FBXL17 and monomeric BACH1-BTB after the BACH- mental condition (Figure 4D). By contrast, the amount of
BTB dimer is remodeled by the F-box protein. FBXL17 captured by untreated BACH1-BTB did not saturate

8 Cell 187, 1–17, December 26, 2024

 Please cite this article in press as: Cao et al., Recognition of BACH1 quaternary structure degrons by two F-box proteins under oxidative
stress, Cell (2024), https://doi.org/10.1016/j.cell.2024.10.012

ll
Article OPEN ACCESS

A B 3 more no observable less 6 VTVKGFEPLIQ 8

RAHRSVL 1
protected change protected (79-89)
2 (46-52) 4
DMSO-treated BACH1-BTB
Tm = 63.2 + 0.5°C 1 2

0 0
NOR3-S1PC treated BACH1-BTB 1 10 102 103 104 105 1 10 102 103 104 105

Exchange (Da)
Tm = 54.9 + 0.2°C 1
4 LSLNDQRK 2
d(Fluorescence)/d(T)

8 2
10000 3 (22-29) 2
DMSO 2 4 7
5 1
5000 NOR3-S1PC (97-101)
1 ILSKE 7
3 6
0 0 0
1 10 102 103 104 105 1 10 102 103 104 105
3 4
-5000 6 (91-96)
30 40 50 60 70 80 90 4 3
2 AYTAKL 4
Temperature (°C) 4 2
2
2 (6-14) 1 (112-119) (123-128)
1
SSVFAYESS 3 FLSVHNIE 5 FQFLKF 6
0 0 0 0
1 10 102 103 104 105 1 10 102 103 104 105 1 10 102 103 104 105 1 10 102 103 104 105
Time (sec)
C 6NSVFAYESSVHSTNVL21 79 VTVKGFEPLIQ 89 113 LSVHNIEES121

Unmod. NO-treated
C
Unmod. NO-treated Unmod. NO-treated
Cent: 864.3895 m/z Cent:864.3907 m/z Cent: 616.1757 m/z Cent: 616.1787 m/z Cent:1028.004 m/z Cent:1028.006 m/z

Undeut.

Cent: 867.003 m/z Cent: 866.9031 m/z Cent: 616.4203 m/z Cent: 617.0101 m/z Cent: 1030.707 m/z Cent: 1030.862 m/z
F-test p = 1.23e-5 F-test p = 2.98e-6
Bimodal 1: Bimodal 1: 64%
21% Bimodal 2:: 36%
Bimodal 2:
79% 3 sec
Relative Intensity

Cent: 869.0092 m/z Cent: 868.7944 m/z Cent: 617.9536 m/z Cent: 618.5192m/z Cent:1032.555 m/z Cent:1032.202 m/z
F-test p = 4.55e-10 F-test p = 1.96e-3 F-test p = 6.21e-8
Bimodal 1: Bimodal 1:
Bimodal 1: 76% 42%
13% Bimodal 2: Bimodal 2:
Bimodal 2: 24% 58% 20 hr
87%

Cent: 869.2066 m/z Cent: 869.2045 m/z Cent: 619.1559 m/z Cent: 619.1489 m/z Cent:1032.766 m/z Cent:1032.892 m/z

Max
deut.

863 867 871 863 867 871 615 617 619 620 623 617 619 620 623 1025 1030 1035 1040 1030 1035 1040
m/z

Figure 5. S-nitrosylation-induced structural changes of BACH1-BTB dimer

(A) Thermostability of native and NO-treated BACH1-BTB determined by differential scanning fluorimetry. Tm is melting temperature and is presented as
mean ± SD.
(B) Summary of HDX data overlayed onto BACH1-BTB dimer. Red and orange regions indicate faster exchange observed in the NO-treated sample. Red regions
display correlated EX1 kinetics, likely reflective of dimer reconfiguration over longer time points (see Figures S4D and S4E). Blue regions show a subpopulation

(legend continued on next page)

Cell 187, 1–17, December 26, 2024 9

 Please cite this article in press as: Cao et al., Recognition of BACH1 quaternary structure degrons by two F-box proteins under oxidative
stress, Cell (2024), https://doi.org/10.1016/j.cell.2024.10.012

ll
OPEN ACCESS Article

even after 2 days. Such a strong preference of FBXL17 to cap- FBXO22 but increased its ability to bind FBXL17 (Figures S5E–
ture BACH1-BTB monomers from the compound-treated dimer S5G). This could explain the similar half-lives of the BACH1
over the intact non-treated dimer raised the possibility that NO mutant and the wild-type (WT) protein (Figure S5H), as the
treatment destabilizes BACH1 to promote FBXL17 binding and impairment of its FBXO22-mediated turnover is compensated
subsequent ubiquitination. by FBXL17. Based on these observations, we conclude that
Importantly, this effect of NO is attributable to the modification S-nitrosylation of the two cysteine residues in BACH1 reduces
of two specific cysteine residues on the BACH1 BTB domain. the structural stability of the BACH1-BTB dimer by reconfiguring
Without impacting the resistance of native BACH1-BTB dimer the dimer arrangement, thereby facilitating FBXL17 binding.
to FBXL17 binding (Figure 4E), individual alanine mutations of
Cys107 and Cys122 but not the other three cysteine residues Transient remodeling of BACH1-BTB dimer by a pair of
significantly reduced the ability of FBXL17 to release monomeric SCFFBXL17 proteins
BACH-BTB from its NO donor-treated dimeric form (Figures 4F, To reveal the mechanism by which FBXL17 recognizes and
S4B, and S4C). This result allows us to map Cys122, an amino remodels NO-modified homodimeric BACH1-BTB, we mixed
acid also residing within the b1-a6 degron, as another key resi- NO donor-treated BACH1-BTB with the SCFFBXL17 complex for
due sensitive to S-nitrosylation. different periods of time and subjected the samples to cryo-EM
analysis. 2D classification of the particles imaged at three incuba-
Global destabilization of BACH1-BTB dimer by tion times (30 s, 10 min, and 24 h) indicated that the predominant
S-nitrosylation observable species were either in the form of free SCFFBXL17 or the
To confirm that S-nitrosylation indeed destabilizes BACH1-BTB E3 complex bound to the monomeric BACH1 BTB domain (here-
dimerization, we first used differential scanning fluorometry after referred to as mSCFBTB) (Figures S6A–S6C). Strikingly, a
(DSF) to show that NO treatment reduced the thermostability small fraction of the particles appeared to contain a pair of parallel
of BACH1-BTB by almost 10 degrees Celsius (Figure 5A). Subse- SCFFBXL17 complexes with the FBXL17 LRR domains facing each
quent hydrogen-deuterium-exchange MS (HDX-MS) experi- other, likely sandwiching a BACH1-BTB dimer in the middle (Fig-
ments allowed us to map the secondary structure elements ure 6A). Although these minor classes, which possibly represent a
within the BACH1-BTB dimer that are structurally destabilized. transient intermediate species of the BACH1-BTB dimer remod-
Strikingly, while the a6 and b1 terminal regions are intrinsically eled by the E3 complex, yielded two 7.8 Å 3D reconstructions (Fig-
more susceptible to deuteration in both native and NO donor- ure S6C), the numbers of particles were too limited to determine
treated samples, multiple secondary structure elements within the structures at a higher resolution.
the BTB core domain, including those at the central dimer inter- To overcome the challenge of resolving these presumably un-
face, are impacted the most by NO treatment (Figures 5B, S4D, stable intermediates, we turned to the native BACH1-BTB sam-
and S4E). Because BACH1-BTB remains as a dimer upon NO ple, which is still susceptible to reconfiguration by FBXL17,
treatment (Figure S3I), these HDX results hint at an altered topol- except at a much slower rate (Figure 4D). We reasoned that the
ogy of the BTB dimer induced by S-nitrosylation. In support of higher resistance of the native BACH1-BTB sample to FBXL17-
this notion, several peptides detected in our HDX-MS analysis mediated monomerization might offer an experimental window
displayed a bimodal exchange profile in the NO-treated samples to capture the intermediate steps of the reaction. Remarkably,
(Figures 5C and S4F), indicating that the BACH1-BTB dimer cryo-EM analysis of a sample of SCFFBXL17 incubated with native
might exist in two distinct populations where one population un- BACH1-BTB unveiled four species, free SCFFBXL17, mSCFBTB,
derwent faster deuterium exchange than the other. For some and two distinct dimeric SCFFBXL17-BACH1-BTB super-assemblies
peptides, the two species were also detectable in the native (hereafter referred to as dSCFBTB-I and dSCFBTB-II) (Figure S7).
sample. NO treatment appeared to shift the equilibrium between The structure of the mSCFBTB complex resolved at 3.2 Å resolu-
the two populations, increasing the fast exchange species, tion is highly analogous to the previously reported crystal struc-
which is structurally less stable. These results are in full agree- ture of FBXL17 in complex with the monomeric mutant KEAP1
ment with the global destabilizing effect of NO treatment de- BTB domain,48 except that no clear density is present for the
tected by the DSF thermal shift assay (Figure 5A). b1 strand of the BACH1 BTB domain (Figure 6B). In both struc-
By directly detecting two S-nitrosylated peptides from our LC- tures, the canonical dimer interface of the BTB domains is
MS analysis of the NOR3-S1PC-treated sample, we further blocked by the FBXL17 CTH. Importantly, both Cys107 and
confirmed that C107 and C122 are the only two NO-modified Cys122 of BACH1-BTB are solvent-exposed and far away from
amino acids within the BACH1 BTB domain (Figures S5A– the BACH1-FBXL17 interface, indicating that they are not directly
S5C). Akin to NO treatment, a C122E mutation in BACH1 recognized by the F-box protein. Modification of these two resi-
mimicking the effect of S-nitrosylation reduced the thermosta- dues in the NO donor-treated BACH1 sample, therefore, likely
bility of the BTB domain (Figure S5D). Both in vitro and in cells, accelerated FBXL17 binding by destabilizing the BACH1-
the BACH1 C122E mutant was no longer recognized by BTB dimer.

that was slightly more protected in the NO-treated sample. White indicates no observable differences. Individual uptake plots show the exchange for unmodified
(blue) and NO modified (red). Data are presented as mean ± SD.
(C) Peptides 6–21 and 79–89 exhibit bimodal exchange behavior but with different extents in the unmodified and NO-modified states. Individual populations
within the spectra are shown in purple and green. By contrast, peptides 113–121 do not exhibit bimodal profiles.
See also Figures S4 and S5.

10 Cell 187, 1–17, December 26, 2024

 Please cite this article in press as: Cao et al., Recognition of BACH1 quaternary structure degrons by two F-box proteins under oxidative
stress, Cell (2024), https://doi.org/10.1016/j.cell.2024.10.012

ll
Article OPEN ACCESS

A S-nitrosylated BACH-BTB + SCFFBXL17 B C D

BACH1-BTB α6 FBXL17
Cys122 N LRR Domain
BACH1-
Cys107 BTB(A)
β5 (β1)

Native BACH1-BTB + SCFFBXL17 FBXL17 α5

LRR Domain β5
Parallel dimer
β5
CTH

BACH1-
Anti-parallel dimer BTB(B)
FBXL17
LRR Domain

C dSCFFBXL17-BTB-I dSCFFBXL17-BTB-II E
BACH1- Free form dSCFFBXL17-bound form
SKP1 BACH1-
BTB BTB Partially α6 BACH1- BACH1-
Melt BTB(A) (β1) (α6) BTB(A)
β1 1 2
β5 Disordered β5
BACH1- BACH1- α5 β5
FBXL17 BTB FBXL17 BTB Disordered
β1 α5
α6
β5 BACH1- (α6) (β1)
CUL1 CUL1 BTB(B)
BACH1- α5
RBX1 BTB(B) Partially Melt
RBX1

F HEK293T G H
WCE IP: α-FLAG 1.5 8 24 hours 1.5 8 24 hours
EV - - + + + EV - - + + + FLAG-FBXL17 + - + - + - Biotin-BACH2-BTB + - + - + - Biotin-BACH1-BACH2
+--+- +--+- HA-FBXL17 - + - + - + Biotin-BACH1-BTB - + - + - + Biotin-BACH1-BTB
- +- - + - +- - + HA-FBXL17 ΔCTH + + + + + + MBP-FBXL17-SKP1 + + + + + + MBP-FBXL17-SKP1
kDa kDa
HA-FBXL17 (α-HA) 75 75 MBP-FBXL17
MBP-FBXL17
HA-FBXL17 ΔCTH (α-HA) 63 63
48 48
35 35
FLAG-FBXL17 (α-FLAG) 25 SKP1 25 SKP1
17 17
SKP1 11
Biotin-BACH1/2-BTB Biotin-BACH1-BTB
11 BACH2-BTB
GAPDH Biotin pull down Biotin pull down

1.5 6 24 hours 1.5 6 24 hours

I J K
+ - + - + - His-Biotin-BACH1MT + - + - + - His-Biotin-BACH1WT-MBP-BACH1MT
- + + BACH1-BACH2 - + - + - + His-Biotin-BACH1WT - + - + - + His-Biotin-BACH1WT-MBP-BACH1WT
+ + - FBXO22-SKP1 + + + + + + MBP-FBXL17-SKP1 + + + + + + MBP-FBXL17-SKP1
kDa kDa
75 GST-FBXO22 75 MBP-FBXL17 75 MBP-FBXL17
63 63 63 MBP-BACH1 WT or Mutant
MBP-BACH2-BTB 48 48
48 MSB-BACH1-BTB 35 35
35 25 25 SKP1
SKP1 20
20 His-Biotin-BACH1-BTBWT
SKP1 17 His-Biotin-BACH1-BTB 17
20 11 WT or Mutant 11
MBP pull down Biotin pull down
Biotin pull down

Figure 6. Structural analysis of BACH1-BTB remodeling by SCFFBXL17

(A) Representative 2D classes of dimeric SCF-FBXL17-BACH1-BTB complexes found in the native and NOR3-S1PC-treated samples.
(B) Structure of FBXL17-LRR in complex with monomeric BACH1-BTB. Structurally disordered b1 strand of BACH1-BTB is shown in a dashed line.
(C) Two dimeric SCF-FBXL17-BACH1-BTB complexes fitted in their electron microscopy maps.
(D) Close-up views of BACH1-BTB dimer interface in the dimeric FBXL17-BACH1 complex. Local refinement-improved electron microscopy map is fitted with a
BACH1-BTB dimer and two copies of FBXL17-LRRs.
(E) Conformational changes of the BACH1-BTB dimer during the transition from the free to the dimeric SCF-FBXL17-bound form. Two arrowed dashed lines
indicate the two-step rotations of the protomer A (cyan) of BACH1-BTB.
(F) Co-immunoprecipitation of HA-FBXL17 by FLAG-FBXL17. FLAG-FBXL17 was co-transfected with full-length or C-terminally truncated HA-FBXL17 into
HEK293T cells. Proteins were immunoblotted as indicated. CTH, C-terminal helix.
(G) FBXL17 pull-down by biotinylated BACH1-BTB but not BACH2-BTB.

(legend continued on next page)

Cell 187, 1–17, December 26, 2024 11

 Please cite this article in press as: Cao et al., Recognition of BACH1 quaternary structure degrons by two F-box proteins under oxidative
stress, Cell (2024), https://doi.org/10.1016/j.cell.2024.10.012

ll
OPEN ACCESS Article

Although the 3D reconstructions of dSCFBTB-I and dSCFBTB-II BTB domain and a mutant BTB defective for binding FBXL17
suffered from a lower resolution (4.4 Å) (Figure 6C), local refine- (Figures 6J and 6K). These results strongly suggest that the
ment improved the density map and enabled us to resolve the homodimeric BACH1 BTB dimer is indeed inspected by a pair
secondary structures of the two BACH1 BTB domains in the mid- of SCFFBXL17 proteins, which is distinct from SCFFBXO22. If the
dle of the dimeric SCF (Figures 6D, S7C, and S7D). In dSCFBTB-I, structural integrity of the substrate dimer is compromised by
FBXL17 is engaged with BACH1-BTB through the same interface oxidative modifications, such as S-nitrosylation, SCFFBXL17 can
as found in mSCFBTB. However, in addition to the b1 strand, the a6 efficiently seize monomeric BACH1.
helix in each BACH1-BTB protomer is also structurally disordered.
Remarkably, the b5 strands of the two protomers join each other DISCUSSION
and form an intermolecular anti-parallel b sheet, structurally
mimicking the b1–b5 cross-protomer two-stranded b sheet found Degrons beyond SLiMs
in the isolated BACH1-BTB dimer (Figure 6E). This new dimer Ubiquitin ligases have been well known to recruit substrates by
interface is further substantiated by the cross-promoter packing recognizing their SLiM degrons.2 Although portability has long
of a coiled region transformed from the mostly melt a5 helices. been implicated as an inherent property of these degradation
A similar dimer interface, albeit being slightly twisted, is found in signals, it is not an essential feature originally used to define
dSCFBTB-II. As a result, the C-terminal lobes of the two CUL1- the term degron.3 Recent studies have shown that the targeting
RBX1 catalytic scaffolds become separate from each other (Fig- signals of a number of CRL-substrate proteins are constructed
ure 6C). The dimeric arrangement of these dSCFBTB complexes, by their tertiary folds.57,58 This study expands the degron space
therefore, is predominantly, if not solely, stabilized by the into the quaternary structures of protein oligomers, which allow
BACH1 BTB domains. Importantly, the two structure models, E3s to differentiate targets based on signature motifs encrypted
namely dSCFBTB-I and dSCFBTB-II, derived from the native only in high-order assemblies. In the case of the b1-a6 degron of
BACH1-BTB sample, fit snuggly into the 7.8 Å 3D reconstruction the dimeric BACH1-BTB recognized by FBXO22, substrate
maps obtained with the NO-treated sample, indicating that they specificity is dictated by the proper association of the two iden-
belong to common species found in the two specimens tical subunits of the substrate. By contrast, the degradation
(Figures S6D and S6E). Taken together, we postulate that these signal of the BACH1-BTB dimer decoded by FBXL17 is deter-
dimeric SCFFBXL17-BACH1-BTB super-assemblies represent inter- mined by the strength of protein-protein interaction within the
mediate states of BACH1-BTB dimer remodeling, which is simul- homodimeric assembly, a physical property that can be affected
taneously catalyzed by two SCFFBXL17 E3 complexes. This notion by post-translation modifications associated with oxidative
is further supported by co-immunoprecipitation analysis, which stress. These examples highlight the diverse nature of degrada-
revealed physical association between two differentially tagged tion signals recognized by ubiquitin ligases and underscore the
FBXL17 when co-expressed in the cell (Figures 6F and S5I). Their complex mechanisms governing E3-substrate specificity in pro-
complexation was largely abrogated when the CTH, which is tein degradation. We speculate that the number of degrons with
responsible for BTB binding, was removed from one of the high-order structural complexity beyond simple primary amino
FBXL17 constructs (Figure 6F), suggesting that the association acid sequence will continue growing, especially as more poten-
of the two FBXL17 proteins is dependent on BTB binding. tial CRL E3 substrates are deconvoluted.59–61
To assess whether efficient substrate remodeling indeed re-
quires the synergy of two copies of the E3 complexes, we iso- Substrate repertoire of FBXO22 and FBXL17
lated a BACH1-BACH2 BTB domain heterodimer with tandem FBXO22 has been suggested to ubiquitinate several substrate
affinity purification upon co-expression. Lacking the key resi- proteins with no structural similarity to BACH1.62 How the F-box
dues for FBXL17 binding,48 homodimeric BACH2-BTB is protein recognizes these targets, presumably also through its
completely resistant to remodeling by the F-box protein (Fig- FIST domain, remains to be determined. With a role in protein
ure 6G). Interestingly, even with an incubation time of about dimerization quality control, FBXL17 has been implicated in ubiq-
1 day, FBXL17 can barely capture BACH1-BTB from the uitinating a subset of BTB-domain-containing proteins when they
BACH1-BACH2 heterodimer, which is in stark contrast to the ho- form mismatched heterodimers through their BTB domains.48,49
modimeric BACH1-BTB (Figure 6H). Of note, because the However, only the structure of FBXL17 in complex with a mono-
BACH1-BACH2 heterodimer retains a functional b1-a6 quater- meric BTB domain has been captured so far, leading to a paradox
nary structure degron (Figure 2F), it can still be recognized by in the mechanism of action of this E3. We propose that it is not the
FBXO22 (Figure 6I). Similarly, FBXL17 can sequester WT mismatched nature of the BTB domain heterodimers but the
BACH1-BTB from a WT BACH1-BTB homodimer far more effi- strength of protein-protein interaction at the dimeric interface
ciently than a BACH1-BTB heterodimer consisting of a WT that is perceived by FBXL17. This distinction explains why the

(H) A comparison of the remodeling efficiency of BACH1-BTB homodimer and BACH1-BTB-BACH2-BTB heterodimer by FBXL17 as monitored by the time
course of BACH1-BTB-MBP-FBXL17 association.
(I) Binding of the BACH1-BTB-BACH2-BTB heterodimer to FBXO22 by MBP pull-down assay. MSB is a protein solubilizing tag unrelated to MBP.
(J) FBXL17 pull-down by biotinylated wild-type (WT) BACH1-BTB but not a mutant (MT) defective in FBXL17 binding.
(K) A comparison of the remodeling efficiency of WT BACH1-BTB homodimer and a WT and mutant BACH1-BTB heterodimer by FBXL17 as monitored by the
time course of BACH1-BTB-MBP-FBXL17 association.
See also Figures S6 and S7 and Table S1.

12 Cell 187, 1–17, December 26, 2024

 Please cite this article in press as: Cao et al., Recognition of BACH1 quaternary structure degrons by two F-box proteins under oxidative
stress, Cell (2024), https://doi.org/10.1016/j.cell.2024.10.012

ll
Article OPEN ACCESS

Figure 7. A schematic model of oxidative-

stress-induced BACH1 ubiquitination by
FBXO22 and FBXL17
Under homeostatic conditions, BACH1 homo-
dimerizes via its BTB domain and interacts with
small MAF proteins through its bZIP domain to
repress gene expression by binding to the anti-
oxidant response elements (AREs) of oxidative
stress response genes. The interaction between
BACH1 and components of the transcriptional
co-repressor complex (i.e., NCOR1 and other
histone modifiers) blocks the recognition of the b1-
a6 quaternary structure degron of BACH1 by
FBXO22. Under oxidative stress, reactive oxygen
species (ROS) covalently modify cysteine residues
on the bZIP domain of BACH1 and release it from
chromatin. If the BTB domain of BACH1 remains
intact, its b1-a6 degron is recognized by FBXO22,
which promotes its ubiquitination and degradation.
If the pro-oxidant also modifies the BACH1 BTB
domain and compromises the structural integrity
of the cross-protomer b1-a6 degron, thereby
weakening the BTB dimer, a pair of FBXL17 will
transiently associate with the BACH1 BTB dimer
and remodel it into stably bound monomer for
ubiquitination and degradation.

beyond simple recognition. Consistent

with this notion, a minor population
of dimeric SCFFBXL17-BACH1-BTB particles
configured in an antiparallel manner can
be found in the 2D class averages of the
native BACH1-BTB sample (Figure 6A).
Unlike conventional E3s, FBXL17 appears
to be able to enzymatically remodel and
promote the degradation of its substrates.

BACH1 as a sensor for oxidative

stress
As a key transcription factor regulating
heme metabolism and oxidative stress,
BACH1 is tightly regulated at its protein
level. FBXO22 and FBXL17 appear to
have been co-evolved to play comple-
mentary roles in promoting BACH1 ubiq-
uitination. Our studies not only reveal the
distinct structural basis of BACH1 recog-
nition by the two F-box proteins but also
shed light on the regulatory mechanisms
same F-box protein can also sequester protomers derived from by which different forms of oxidative stress make the transcrip-
the homodimeric BTB domain of BACH1 when given long enough tional repressor discernible to the E3s (Figure 7). With an unusu-
incubation time or when the stability of the homodimer is compro- ally large number of cysteine residues, BACH1 appears to func-
mised by oxidative stress. Moreover, the structures of the tion as a redox sensor, which can modulate its own stability in
FBXL17-BTB dimer intermediates indicate that efficient clearance response to pro-oxidants. While heme or pro-oxidants releases
of these substrate proteins requires the synergistic actions of a BACH1 from the chromatin and exposes its b1-a6 degron to
pair of SCFFBXL17, which can remodel its homodimeric BACH1 FBXO22, NO modifies two cysteine residues in the BACH1
substrate beyond a simple monophasic binding event. Interpreta- BTB domain and makes the BTB dimer susceptible to
tion of the degradation signal by FBXL17, therefore, involves at FBXL17-mediated remodeling by weakening the dimer interface.
least two, if not more, steps of action taken by the F-box protein Outside the BTB domain, a series of reactive cysteine residues of

Cell 187, 1–17, December 26, 2024 13

 Please cite this article in press as: Cao et al., Recognition of BACH1 quaternary structure degrons by two F-box proteins under oxidative
stress, Cell (2024), https://doi.org/10.1016/j.cell.2024.10.012

ll
OPEN ACCESS Article

BACH1, including several in the bZIP domain, have been identi- d Any additional information required to reanalyze the data reported in this
fied by proteome-wide analyses to be sensitive to modifications paper is available from the lead contact upon request.
by H2O2 and NO donors.56,63 It is plausible that these cysteine
residues of BACH1 could also serve as sensors for a variety of ACKNOWLEDGMENTS
pro-oxidants and possibly affect the function and stability of
the transcriptional repressor via unknown mechanisms.64,65 The authors would like to thank R. Yan, X. Zhao, J. Jung, and Z. Yu at the Cryo-
Beyond its role in the oxidative stress response pathway, EM Facility on the Janelia Research Campus of the Howard Hughes Medical
BACH1 also targets genes involved in cell cycle control, apoptosis, Institute and J.D. Quispe and S. Dickinson at the Arnold and Mabel Beckman
Cryo-EM Center at the University of Washington for their assistance in electron
and metabolism.16,47,66 By positively regulating the expression of
microscopy data acquisition. The authors thank Mengxi Liu for her contribution
multiple pro-metastatic genes, BACH1 has been shown to pro- to this work. We also thank the NYU Proteomics Lab (supported in part by the
mote cancer cell invasiveness.67–69 Our studies suggest that com- NYU School of Medicine and the Laura and Isaac Perlmutter Cancer Center
pounds that could modify the BTB or DNA-binding domains of Support grant P30CA016087 from the National Cancer Institute). In addition,
BACH1 have the potential to promote the ubiquitination and the mass spectrometric experiments were supported with a shared instrumen-
degradation of BACH1 by FBXL17 or FBXO22, thereby amelio- tation grant from the NIH (1S10OD010582-01A1 for the purchase of an Orbi-
trap Eclipse) for S-nitrosylation site-mapping experiments. The authors also
rating BACH1-mediated cancer metastasis. Similar strategies to
thank T.R. Hinds for help on BLI data analysis and members of the Zheng
reshape or re-surface domains and complexes of disease-pro- Lab and Pagano Lab for discussion. N.Z. and M.P. are Howard Hughes Med-
moting target proteins with small molecules hold the promise to ical Institute Investigators. This work is also supported by NIH grant
create ‘‘neo’’ tertiary or quaternary structure degrons for targeted GM136250 to M.P. and NIH NIGMS grant S10OD030237 and the National Sci-
protein degradation as an emerging therapeutic modality.70,71 ence Foundation award 2304707 to M.G.

AUTHOR CONTRIBUTIONS
Limitations of the study
The affinity of the central NCOR1 fragment toward the BACH1- S.C., S.F.G., M.P., and N.Z. conceived the project. S.C. constructed and pu-
BTB domain is much lower than that of FBXO22, raising the ques- rified all protein samples in this study with inputs from S.F.G. and H.M. S.C.
tion as to how the transcriptional co-repressor could compete performed BLI, AlphaScreen, EMSA, DSF, and in vitro pull-down and ubiquiti-
with the E3 ubiquitin ligase before BACH1 is released from chro- nation experiments and prepared the sample for cryo-EM analysis. S.F.G.
generated CRISPR KO cell lines, designed plasmid constructs, and performed
matin. We hypothesize that binding of NCOR1 to BACH1 might
all of the cell-based co-immunoprecipitations and protein stability assays, with
be reinforced by multivalent interactions between other subunits notable contributions from Y.K., H.V.G., S.D., S.K., G.R., and L.L. J.P. and B.U.
of the co-repressor complex and chromatin. The nature of these performed the mass spectrometric sample preparation, data acquisition, and
interactions remains to be elucidated. Future studies are also analysis. Huigang Shi performed the majority of cryo-EM grid preparation,
needed to dissect the sub-populations of BACH1, which might specimen screening, data collection, and processing tasks with early contribu-
be differentially modified by S-nitrosylation at different sites and tions from Hui Shi and S.C. E.I.J. and M.G. designed and performed the HDX-
at different times after induction of oxidative stress. Validation of MS experiments. The manuscript was written with inputs from all authors. The
order of the second and third co-first authors was chosen randomly.
this hypothesis could help explain why, in addition to FBXL17,
FBXO22 also plays a role in promoting the degradation of DECLARATION OF INTERESTS
BACH1 upon NO treatment, despite its inability to bind the
S-nitrosylated BACH1 BTB domain. Finally, future structural N.Z. and M.P. are scientific cofounders of SEED Therapeutics with financial in-
studies might reveal additional intermediate states of BACH1- terests. They also received research funding from and are shareholders in Ky-
BTB dimer recognized by FBXL17. mera Therapeutics. N.Z. serves on the scientific advisory boards of Synthex,
Molecular Glue Labs, and Differentiated Therapeutics with financial interests.
M.P. is a consultant and/or serves on the scientific advisory boards of CullGen,
RESOURCE AVAILABILITY
Sibylla Biotech, and Triana Biomedicines with financial interests. The findings
presented in this manuscript were not shared with these companies.
Lead contact
Requests for further information and resources should be directed to and will
STAR+METHODS
be fulfilled by the lead contact, Ning Zheng ([email protected]).

Detailed methods are provided in the online version of this paper and include
Materials availability the following:
All unique/stable reagents generated in this study are available upon request. d KEY RESOURCES TABLE
d EXPERIMENTAL MODEL AND STUDY PARTICIPANT DETAILS
Data and code availability B Cell lines
d Original western blot images have been deposited at Mendeley at B Source organism
https://dx.doi.org/10.17632/vf73rr8p5n.1 and are publicly available as d METHOD DETAILS
of the date of publication. B Protein preparation
d The atomic coordinates and EM maps of protein complexes have been B Affinity pull-down assay
deposited at the Protein Data Bank (PDB) and the Electron Microscopy B BioLayer interferometry
Data Bank (EMDB) with accession codes listed in the key resources ta- B AlphaLISA
ble. They are publicly available as of the date of publication. B SEC–MALS
d HDX-MS data have been deposited at ProteomeXchange as PXD053923. B Differential scanning fluorimetry (DSF)
Proteomics data have been deposited at ProteomeXchange as B In vitro ubiquitination assay
PXD056513. They are publicly available as of the date of publication. B Electrophoretic mobility shift assay (EMSA)
d This paper does not report original code. B Cryo-EM sample preparation and data collection

14 Cell 187, 1–17, December 26, 2024

 Please cite this article in press as: Cao et al., Recognition of BACH1 quaternary structure degrons by two F-box proteins under oxidative
stress, Cell (2024), https://doi.org/10.1016/j.cell.2024.10.012

ll
Article OPEN ACCESS

B Image processing and 3D reconstruction 13. Igarashi, K., Nishizawa, H., Saiki, Y., and Matsumoto, M. (2021). The tran-
B Model building and refinement scription factor BACH1 at the crossroads of cancer biology: From epithe-
B Plasmids and antibodies lial-mesenchymal transition to ferroptosis. J. Biol. Chem. 297, 101032.
B Drug Treatment Procedures https://doi.org/10.1016/j.jbc.2021.101032.
B Co-immunoprecipitation assay
14. Oyake, T., Itoh, K., Motohashi, H., Hayashi, N., Hoshino, H., Nishizawa, M.,
B Cell-based protein stability assay
Yamamoto, M., and Igarashi, K. (1996). Bach proteins belong to a novel
B CRISPR-Cas9 genome editing
family of BTB-basic leucine zipper transcription factors that interact with
B Chromatin Fractionation Assay
MafK and regulate transcription through the NF-E2 site. Mol. Cell. Biol.
B Biotin Switch Assay
16, 6083–6095. https://doi.org/10.1128/MCB.16.11.6083.
B S-nitrosylation site mapping
15. Sun, J., Hoshino, H., Takaku, K., Nakajima, O., Muto, A., Suzuki, H., Ta-
B HDX-MS analyses
shiro, S., Takahashi, S., Shibahara, S., Alam, J., et al. (2002). Hemoprotein
d QUANTIFICATION AND STATISTICAL ANALYSIS
Bach1 regulates enhancer availability of heme oxygenase-1 gene. EMBO
J. 21, 5216–5224. https://doi.org/10.1093/emboj/cdf516.
SUPPLEMENTAL INFORMATION
16. Warnatz, H.J., Schmidt, D., Manke, T., Piccini, I., Sultan, M., Borodina, T.,
Supplemental information can be found online at https://doi.org/10.1016/j.cell. Balzereit, D., Wruck, W., Soldatov, A., Vingron, M., et al. (2011). The BTB
2024.10.012. and CNC homology 1 (BACH1) target genes are involved in the oxidative
stress response and in control of the cell cycle. J. Biol. Chem. 286,
23521–23532. https://doi.org/10.1074/jbc.M111.220178.
Received: October 12, 2023
Revised: August 25, 2024 17. Hintze, K.J., Katoh, Y., Igarashi, K., and Theil, E.C. (2007). Bach1 repres-
Accepted: October 10, 2024 sion of ferritin and thioredoxin reductase1 is heme-sensitive in cells and
Published: November 5, 2024 in vitro and coordinates expression with heme oxygenase1, beta-globin,
and NADP(H) quinone (oxido) reductase1. J. Biol. Chem. 282, 34365–
REFERENCES 34371. https://doi.org/10.1074/jbc.M700254200.
18. Dhakshinamoorthy, S., Jain, A.K., Bloom, D.A., and Jaiswal, A.K. (2005).
1. Hershko, A., and Ciechanover, A. (1998). The ubiquitin system. Annu. Rev. Bach1 competes with Nrf2 leading to negative regulation of the antioxidant
Biochem. 67, 425–479. https://doi.org/10.1146/annurev.biochem.67. response element (ARE)-mediated NAD(P)H:quinone oxidoreductase 1
1.425. gene expression and induction in response to antioxidants. J. Biol.
2. Zheng, N., and Shabek, N. (2017). Ubiquitin Ligases: Structure, Function, Chem. 280, 16891–16900. https://doi.org/10.1074/jbc.M500166200.
and Regulation. Annu. Rev. Biochem. 86, 129–157. https://doi.org/10. 19. Ogawa, K., Sun, J., Taketani, S., Nakajima, O., Nishitani, C., Sassa, S.,
1146/annurev-biochem-060815-014922. Hayashi, N., Yamamoto, M., Shibahara, S., Fujita, H., and Igarashi, K.
3. Varshavsky, A. (1991). Naming a targeting signal. Cell 64, 13–15. https:// (2001). Heme mediates derepression of Maf recognition element through
doi.org/10.1016/0092-8674(91)90202-A. direct binding to transcription repressor Bach1. EMBO J. 20, 2835–
2843. https://doi.org/10.1093/emboj/20.11.2835.
4. Mészáros, B., Kumar, M., Gibson, T.J., Uyar, B., and Dosztányi, Z. (2017).
Degrons in cancer. Sci. Signal. 10, eaak9982. https://doi.org/10.1126/sci- 20. Reichard, J.F., Sartor, M.A., and Puga, A. (2008). BACH1 is a specific
signal.aak9982. repressor of HMOX1 that is inactivated by arsenite. J. Biol. Chem. 283,
5. Zhang, Z., Sie, B., Chang, A., Leng, Y., Nardone, C., Timms, R.T., and El- 22363–22370. https://doi.org/10.1074/jbc.M801784200.
ledge, S.J. (2023). Elucidation of E3 ubiquitin ligase specificity through 21. Reichard, J.F., Motz, G.T., and Puga, A. (2007). Heme oxygenase-1 induc-
proteome-wide internal degron mapping. Mol. Cell 83, 3377–3392.e6. tion by NRF2 requires inactivation of the transcriptional repressor BACH1.
https://doi.org/10.1016/j.molcel.2023.08.022. Nucleic Acids Res. 35, 7074–7086. https://doi.org/10.1093/nar/gkm638.
6. Harper, J.W., and Schulman, B.A. (2021). Cullin-RING Ubiquitin Ligase 22. Namgaladze, D., Fuhrmann, D.C., and Brüne, B. (2022). Interplay of Nrf2
Regulatory Circuits: A Quarter Century Beyond the F-Box Hypothesis. and BACH1 in inducing ferroportin expression and enhancing resistance
Annu. Rev. Biochem. 90, 403–429. https://doi.org/10.1146/annurev-bio- of human macrophages towards ferroptosis. Cell Death Discov. 8, 327.
chem-090120-013613. https://doi.org/10.1038/s41420-022-01117-y.
7. Welcker, M., and Clurman, B.E. (2008). FBW7 ubiquitin ligase: a tumour 23. Baird, L., and Yamamoto, M. (2020). The Molecular Mechanisms Regu-
suppressor at the crossroads of cell division, growth and differentiation. lating the KEAP1-NRF2 Pathway. Mol. Cell. Biol. 40, e00099-20. https://
Nat. Rev. Cancer 8, 83–93. https://doi.org/10.1038/nrc2290. doi.org/10.1128/MCB.00099-20.
8. Bi, Y., Cui, D., Xiong, X., and Zhao, Y. (2021). The characteristics and roles 24. Itoh, K., Wakabayashi, N., Katoh, Y., Ishii, T., O’Connor, T., and Yama-
of b-TrCP1/2 in carcinogenesis. FEBS Journal 288, 3351–3374. https:// moto, M. (2003). Keap1 regulates both cytoplasmic-nuclear shuttling
doi.org/10.1111/febs.15585. and degradation of Nrf2 in response to electrophiles. Genes Cells 8,
9. Koren, I., Timms, R.T., Kula, T., Xu, Q., Li, M.Z., and Elledge, S.J. (2018). 379–391. https://doi.org/10.1046/j.1365-2443.2003.00640.x.
The Eukaryotic Proteome Is Shaped by E3 Ubiquitin Ligases Targeting 25. Kobayashi, A., Kang, M.I., Okawa, H., Ohtsuji, M., Zenke, Y., Chiba, T.,
C-Terminal Degrons. Cell 173, 1622–1635.e14. https://doi.org/10.1016/j. Igarashi, K., and Yamamoto, M. (2004). Oxidative stress sensor Keap1
cell.2018.04.028. functions as an adaptor for Cul3-based E3 ligase to regulate proteasomal
10. Lin, H.C., Yeh, C.W., Chen, Y.F., Lee, T.T., Hsieh, P.Y., Rusnac, D.V., Lin, degradation of Nrf2. Mol. Cell. Biol. 24, 7130–7139. https://doi.org/10.
S.Y., Elledge, S.J., Zheng, N., and Yen, H.S. (2018). C-Terminal End- 1128/MCB.24.16.7130-7139.2004.
Directed Protein Elimination by CRL2 Ubiquitin Ligases. Mol. Cell 70, 26. Tong, K.I., Katoh, Y., Kusunoki, H., Itoh, K., Tanaka, T., and Yamamoto, M.
602–613.e603. https://doi.org/10.1016/j.molcel.2018.04.006. (2006). Keap1 recruits Neh2 through binding to ETGE and DLG motifs:
11. Hayes, J.D., Dinkova-Kostova, A.T., and Tew, K.D. (2020). Oxidative characterization of the two-site molecular recognition model. Mol. Cell.
Stress in Cancer. Cancer Cell 38, 167–197. https://doi.org/10.1016/j. Biol. 26, 2887–2900. https://doi.org/10.1128/MCB.26.8.2887-2900.2006.
ccell.2020.06.001. 27. McMahon, M., Lamont, D.J., Beattie, K.A., and Hayes, J.D. (2010). Keap1
12. Forman, H.J., and Zhang, H. (2021). Targeting oxidative stress in disease: perceives stress via three sensors for the endogenous signaling molecules
promise and limitations of antioxidant therapy. Nat. Rev. Drug Discov. 20, nitric oxide, zinc, and alkenals. Proc. Natl. Acad. Sci. USA 107, 18838–
689–709. https://doi.org/10.1038/s41573-021-00233-1. 18843. https://doi.org/10.1073/pnas.1007387107.

Cell 187, 1–17, December 26, 2024 15

 Please cite this article in press as: Cao et al., Recognition of BACH1 quaternary structure degrons by two F-box proteins under oxidative
stress, Cell (2024), https://doi.org/10.1016/j.cell.2024.10.012

ll
OPEN ACCESS Article
28. Kobayashi, A., Kang, M.I., Watai, Y., Tong, K.I., Shibata, T., Uchida, K., Mol. Cell. Biol. 22, 1804–1818. https://doi.org/10.1128/MCB.22.6.1804-
and Yamamoto, M. (2006). Oxidative and electrophilic stresses activate 1818.2002.
Nrf2 through inhibition of ubiquitination activity of Keap1. Mol. Cell. Biol. 44. Huynh, K.D., and Bardwell, V.J. (1998). The BCL-6 POZ domain and other
26, 221–229. https://doi.org/10.1128/MCB.26.1.221-229.2006. POZ domains interact with the co-repressors N-CoR and SMRT. Onco-
29. Tan, M.K., Lim, H.J., Bennett, E.J., Shi, Y., and Harper, J.W. (2013). Paral- gene 17, 2473–2484. https://doi.org/10.1038/sj.onc.1202197.
lel SCF adaptor capture proteomics reveals a role for SCFFBXL17 in NRF2 45. Ahmad, K.F., Melnick, A., Lax, S., Bouchard, D., Liu, J., Kiang, C.L., Mayer,
activation via BACH1 repressor turnover. Mol. Cell 52, 9–24. https://doi. S., Takahashi, S., Licht, J.D., and Privé, G.G. (2003). Mechanism of SMRT
org/10.1016/j.molcel.2013.08.018. corepressor recruitment by the BCL6 BTB domain. Mol. Cell 12, 1551–
30. Lignitto, L., LeBoeuf, S.E., Homer, H., Jiang, S., Askenazi, M., Karakousi, 1564. https://doi.org/10.1016/S1097-2765(03)00454-4.
T.R., Pass, H.I., Bhutkar, A.J., Tsirigos, A., Ueberheide, B., et al. (2019). 46. Ghetu, A.F., Corcoran, C.M., Cerchietti, L., Bardwell, V.J., Melnick, A., and
Nrf2 Activation Promotes Lung Cancer Metastasis by Inhibiting the Degra- Privé, G.G. (2008). Structure of a BCOR corepressor peptide in complex
dation of Bach1. Cell 178, 316–329.e18. https://doi.org/10.1016/j.cell. with the BCL6 BTB domain dimer. Mol. Cell 29, 384–391. https://doi.
2019.06.003. org/10.1016/j.molcel.2007.12.026.
31. Zheng, N., Schulman, B.A., Song, L., Miller, J.J., Jeffrey, P.D., Wang, P., 47. Dohi, Y., Ikura, T., Hoshikawa, Y., Katoh, Y., Ota, K., Nakanome, A., Muto,
Chu, C., Koepp, D.M., Elledge, S.J., Pagano, M., et al. (2002). Structure A., Omura, S., Ohta, T., Ito, A., et al. (2008). Bach1 inhibits oxidative stress-
of the Cul1-Rbx1-Skp1-F boxSkp2 SCF ubiquitin ligase complex. Nature induced cellular senescence by impeding p53 function on chromatin. Nat.
416, 703–709. https://doi.org/10.1038/416703a. Struct. Mol. Biol. 15, 1246–1254. https://doi.org/10.1038/nsmb.1516.
32. Borziak, K., and Zhulin, I.B. (2007). FIST: a sensory domain for diverse 48. Mena, E.L., Jevtic, P., Greber, B.J., Gee, C.L., Lew, B.G., Akopian, D., No-
signal transduction pathways in prokaryotes and ubiquitin signaling in eu- gales, E., Kuriyan, J., and Rape, M. (2020). Structural basis for dimerization
karyotes. Bioinformatics 23, 2518–2521. https://doi.org/10.1093/bioinfor- quality control. Nature 586, 452–456. https://doi.org/10.1038/s41586-
matics/btm384. 020-2636-7.
33. Sinha, S., Rappu, P., Lange, S.C., Mäntsälä, P., Zalkin, H., and Smith, J.L. 49. Mena, E.L., Kjolby, R.A.S., Saxton, R.A., Werner, A., Lew, B.G., Boyle,
(1999). Crystal structure of Bacillus subtilis YabJ, a purine regulatory pro- J.M., Harland, R., and Rape, M. (2018). Dimerization quality control en-
tein and member of the highly conserved YjgF family. Proc. Natl. Acad. Sci. sures neuronal development and survival. Science 362, eaap8236.
USA 96, 13074–13079. https://doi.org/10.1073/pnas.96.23.13074. https://doi.org/10.1126/science.aap8236.
34. Volz, K. (1999). A test case for structure-based functional assignment: the 50. Hess, D.T., Matsumoto, A., Kim, S.O., Marshall, H.E., and Stamler, J.S.
1.2 A crystal structure of the yjgF gene product from Escherichia coli. Pro- (2005). Protein S-nitrosylation: purview and parameters. Nat. Rev. Mol.
tein Sci. 8, 2428–2437. https://doi.org/10.1110/ps.8.11.2428. Cell Biol. 6, 150–166. https://doi.org/10.1038/nrm1569.
35. Burman, J.D., Stevenson, C.E., Sawers, R.G., and Lawson, D.M. (2007). 51. Fernando, V., Zheng, X., Walia, Y., Sharma, V., Letson, J., and Furuta, S.
The crystal structure of Escherichia coli TdcF, a member of the highly (2019). S-Nitrosylation: An Emerging Paradigm of Redox Signaling. Anti-
conserved YjgF/YER057c/UK114 family. BMC Struct. Biol. 7, 30. https:// oxidants 8, 404. https://doi.org/10.3390/antiox8090404.
doi.org/10.1186/1472-6807-7-30.
52. Wolhuter, K., Whitwell, H.J., Switzer, C.H., Burgoyne, J.R., Timms, J.F.,
36. Ito, N., Watanabe-Matsui, M., Igarashi, K., and Murayama, K. (2009). Crys- and Eaton, P. (2018). Evidence against Stable Protein S-Nitrosylation as
tal structure of the Bach1 BTB domain and its regulation of homodimeriza- a Widespread Mechanism of Post-translational Regulation. Mol. Cell 69,
tion. Genes Cells 14, 167–178. https://doi.org/10.1111/j.1365-2443.2008. 438–450.e5. https://doi.org/10.1016/j.molcel.2017.12.019.
01259.x.
53. Tsuneyoshi, T., Kunimura, K., and Morihara, N. (2019). S-1-Propenylcys-
37. Bonchuk, A., Balagurov, K., and Georgiev, P. (2023). BTB domains: A teine augments BACH1 degradation and heme oxygenase 1 expression
structural view of evolution, multimerization, and protein-protein interac- in a nitric oxide-dependent manner in endothelial cells. Nitric Oxide 84,
tions. BioEssays 45, e2200179. https://doi.org/10.1002/bies.202200179. 22–29. https://doi.org/10.1016/j.niox.2019.01.003.
38. Muto, A., Hoshino, H., Madisen, L., Yanai, N., Obinata, M., Karasuyama, 54. Fourquet, S., Guerois, R., Biard, D., and Toledano, M.B. (2010). Activation of
H., Hayashi, N., Nakauchi, H., Yamamoto, M., Groudine, M., and Igarashi, NRF2 by nitrosative agents and H2O2 involves KEAP1 disulfide formation.
K. (1998). Identification of Bach2 as a B-cell-specific partner for small maf J. Biol. Chem. 285, 8463–8471. https://doi.org/10.1074/jbc.M109.051714.
proteins that negatively regulate the immunoglobulin heavy chain gene 30 55. Li, C.Q., Kim, M.Y., Godoy, L.C., Thiantanawat, A., Trudel, L.J., and Wo-
enhancer. EMBO J. 17, 5734–5743. https://doi.org/10.1093/emboj/17. gan, G.N. (2009). Nitric oxide activation of Keap1/Nrf2 signaling in human
19.5734. colon carcinoma cells. Proc. Natl. Acad. Sci. USA 106, 14547–14551.
39. Muto, A., Tashiro, S., Nakajima, O., Hoshino, H., Takahashi, S., Sakoda, https://doi.org/10.1073/pnas.0907539106.
E., Ikebe, D., Yamamoto, M., and Igarashi, K. (2004). The transcriptional 56. Mnatsakanyan, R., Markoutsa, S., Walbrunn, K., Roos, A., Verhelst,
programme of antibody class switching involves the repressor Bach2. Na- S.H.L., and Zahedi, R.P. (2019). Proteome-wide detection of
ture 429, 566–571. https://doi.org/10.1038/nature02596. S-nitrosylation targets and motifs using bioorthogonal cleavable-linker-
40. Roychoudhuri, R., Hirahara, K., Mousavi, K., Clever, D., Klebanoff, C.A., based enrichment and switch technique. Nat. Commun. 10, 2195.
Bonelli, M., Sciumè, G., Zare, H., Vahedi, G., Dema, B., et al. (2013). https://doi.org/10.1038/s41467-019-10182-4.
BACH2 represses effector programs to stabilize T(reg)-mediated immune 57. Xing, W., Busino, L., Hinds, T.R., Marionni, S.T., Saifee, N.H., Bush, M.F.,
homeostasis. Nature 498, 506–510. https://doi.org/10.1038/nature12199. Pagano, M., and Zheng, N. (2013). SCF(FBXL3) ubiquitin ligase targets
41. Rosbrook, G.O., Stead, M.A., Carr, S.B., and Wright, S.C. (2012). The cryptochromes at their cofactor pocket. Nature 496, 64–68. https://doi.
structure of the Bach2 POZ-domain dimer reveals an intersubunit disulfide org/10.1038/nature11964.
bond. Acta Crystallogr. D Biol. Crystallogr. 68, 26–34. https://doi.org/10. 58. Wang, H., Shi, H., Rajan, M., Canarie, E.R., Hong, S., Simoneschi, D., Pa-
1107/S0907444911048335. gano, M., Bush, M.F., Stoll, S., Leibold, E.A., et al. (2020). FBXL5 Regu-
42. Cao, S., Kang, S., Mao, H., Yao, J., Gu, L., and Zheng, N. (2022). Defining lates IRP2 Stability in Iron Homeostasis via an Oxygen-Responsive
molecular glues with a dual-nanobody cannabidiol sensor. Nat. Commun. [2Fe2S] Cluster. Mol. Cell 78, 31-41.e35. https://doi.org/10.1016/j.mol-
13, 815. https://doi.org/10.1038/s41467-022-28507-1. cel.2020.02.011.
43. Melnick, A., Carlile, G., Ahmad, K.F., Kiang, C.L., Corcoran, C., Bardwell, 59. Lumpkin, R.J., Baker, R.W., Leschziner, A.E., and Komives, E.A. (2020).
V., Prive, G.G., and Licht, J.D. (2002). Critical residues within the BTB Structure and dynamics of the ASB9 CUL-RING E3 Ligase. Nat. Commun.
domain of PLZF and Bcl-6 modulate interaction with corepressors. 11, 2866. https://doi.org/10.1038/s41467-020-16499-9.

16 Cell 187, 1–17, December 26, 2024

 Please cite this article in press as: Cao et al., Recognition of BACH1 quaternary structure degrons by two F-box proteins under oxidative
stress, Cell (2024), https://doi.org/10.1016/j.cell.2024.10.012

ll
Article OPEN ACCESS

60. Sherpa, D., Chrustowicz, J., Qiao, S., Langlois, C.R., Hehl, L.A., Gotte- 75. Emsley, P., Lohkamp, B., Scott, W.G., and Cowtan, K. (2010). Features
mukkala, K.V., Hansen, F.M., Karayel, O., von Gronau, S., Prabu, J.R., and development of Coot. Acta Crystallogr. D Biol. Crystallogr. 66,
et al. (2021). GID E3 ligase supramolecular chelate assembly configures 486–501. https://doi.org/10.1107/S0907444910007493.
multipronged ubiquitin targeting of an oligomeric metabolic enzyme. 76. Adams, P.D., Grosse-Kunstleve, R.W., Hung, L.W., Ioerger, T.R., McCoy,
Mol. Cell 81, 2445–2459.e13. https://doi.org/10.1016/j.molcel.2021. A.J., Moriarty, N.W., Read, R.J., Sacchettini, J.C., Sauter, N.K., and Ter-
03.025. williger, T.C. (2002). PHENIX: building new software for automated crystal-
61. Teng, F., Wang, Y., Liu, M., Tian, S., Stjepanovic, G., and Su, M.Y. (2023). lographic structure determination. Acta Crystallogr. D Biol. Crystallogr. 58,
Cryo-EM structure of the KLHL22 E3 ligase bound to an oligomeric meta- 1948–1954. https://doi.org/10.1107/S0907444902016657.
bolic enzyme. Structure 31, 1431–1440.e5. https://doi.org/10.1016/j.str.
77. Afonine, P.V., Poon, B.K., Read, R.J., Sobolev, O.V., Terwilliger, T.C., Urz-
2023.09.002.
humtsev, A., and Adams, P.D. (2018). Real-space refinement in PHENIX
62. Cheng, J., Lin, M., Chu, M., Gong, L., Bi, Y., and Zhao, Y. (2020). Emerging for cryo-EM and crystallography. Acta Crystallogr. D Struct. Biol. 74,
role of FBXO22 in carcinogenesis. Cell Death Discov. 6, 66. https://doi. 531–544. https://doi.org/10.1107/S2059798318006551.
org/10.1038/s41420-020-00303-0.
78. Guttman, M., Weis, D.D., Engen, J.R., and Lee, K.K. (2013). Analysis of
63. Zhang, J., Simpson, C.M., Berner, J., Chong, H.B., Fang, J., Ordulu, Z., overlapped and noisy hydrogen/deuterium exchange mass spectra.
Weiss-Sadan, T., Possemato, A.P., Harry, S., Takahashi, M., et al. J. Am. Soc. Mass Spectrom. 24, 1906–1912. https://doi.org/10.1007/
(2023). Systematic identification of anticancer drug targets reveals a nu- s13361-013-0727-5.
cleus-to-mitochondria ROS-sensing pathway. Cell 186, 2361–2379.e25.
https://doi.org/10.1016/j.cell.2023.04.026. 79. Suloway, C., Pulokas, J., Fellmann, D., Cheng, A., Guerra, F., Quispe, J.,
Stagg, S., Potter, C.S., and Carragher, B. (2005). Automated molecular mi-
64. Zenke-Kawasaki, Y., Dohi, Y., Katoh, Y., Ikura, T., Ikura, M., Asahara, T.,
croscopy: the new Leginon system. J. Struct. Biol. 151, 41–60. https://doi.
Tokunaga, F., Iwai, K., and Igarashi, K. (2007). Heme induces ubiquitina-
org/10.1016/j.jsb.2005.03.010.
tion and degradation of the transcription factor Bach1. Mol. Cell. Biol.
27, 6962–6971. https://doi.org/10.1128/MCB.02415-06. 80. Mastronarde, D.N. (2005). Automated electron microscope tomography
using robust prediction of specimen movements. J. Struct. Biol. 152,
65. Hira, S., Tomita, T., Matsui, T., Igarashi, K., and Ikeda-Saito, M. (2007).
36–51. https://doi.org/10.1016/j.jsb.2005.07.007.
Bach1, a heme-dependent transcription factor, reveals presence of multi-
ple heme binding sites with distinct coordination structure. IUBMB Life 59, 81. Wu, C., Huang, X., Cheng, J., Zhu, D., and Zhang, X. (2019). High-quality,
542–551. https://doi.org/10.1080/15216540701225941. high-throughput cryo-electron microscopy data collection via beam tilt
66. Matsumoto, M., Kondo, K., Shiraki, T., Brydun, A., Funayama, R., Na- and astigmatism-free beam-image shift. J. Struct. Biol. 208, 107396.
kayama, K., Yaegashi, N., Katagiri, H., and Igarashi, K. (2016). Genome- https://doi.org/10.1016/j.jsb.2019.09.013.
wide approaches for BACH1 target genes in mouse embryonic fibroblasts 82. Guo, H., Franken, E., Deng, Y., Benlekbir, S., Singla Lezcano, G., Janssen,
showed BACH1-Pparg pathway in adipogenesis. Genes Cells 21, B., Yu, L., Ripstein, Z.A., Tan, Y.Z., and Rubinstein, J.L. (2020). Electron-
553–567. https://doi.org/10.1111/gtc.12365. event representation data enable efficient cryoEM file storage with full
67. Wiel, C., Le Gal, K., Ibrahim, M.X., Jahangir, C.A., Kashif, M., Yao, H., Zie- preservation of spatial and temporal resolution. IUCrJ 7, 860–869.
gler, D.V., Xu, X., Ghosh, T., Mondal, T., et al. (2019). BACH1 Stabilization https://doi.org/10.1107/S205225252000929X.
by Antioxidants Stimulates Lung Cancer Metastasis. Cell 178, 330– 83. Punjani, A., Rubinstein, J.L., Fleet, D.J., and Brubaker, M.A. (2017). cryo-
345.e22. https://doi.org/10.1016/j.cell.2019.06.005. SPARC: algorithms for rapid unsupervised cryo-EM structure determina-
68. Liang, Y., Wu, H., Lei, R., Chong, R.A., Wei, Y., Lu, X., Tagkopoulos, I., tion. Nat. Methods 14, 290–296. https://doi.org/10.1038/nmeth.4169.
Kung, S.Y., Yang, Q., Hu, G., and Kang, Y. (2012). Transcriptional network 84. Zhang, K. (2016). Gctf: real-time CTF determination and correction.
analysis identifies BACH1 as a master regulator of breast cancer bone J. Struct. Biol. 193, 1–12. https://doi.org/10.1016/j.jsb.2015.11.003.
metastasis. J. Biol. Chem. 287, 33533–33544. https://doi.org/10.1074/
85. Bepler, T., Morin, A., Rapp, M., Brasch, J., Shapiro, L., Noble, A.J., and
jbc.M112.392332.
Berger, B. (2019). Positive-unlabeled convolutional neural networks for
69. Yun, J., Frankenberger, C.A., Kuo, W.L., Boelens, M.C., Eves, E.M.,
particle picking in cryo-electron micrographs. Nat. Methods 16, 1153–
Cheng, N., Liang, H., Li, W.H., Ishwaran, H., Minn, A.J., and Rosner,
1160. https://doi.org/10.1038/s41592-019-0575-8.
M.R. (2011). Signalling pathway for RKIP and Let-7 regulates and predicts
metastatic breast cancer. EMBO J. 30, 4500–4514. https://doi.org/10. 86. Punjani, A., Zhang, H., and Fleet, D.J. (2020). Non-uniform refinement:
1038/emboj.2011.312. adaptive regularization improves single-particle cryo-EM reconstruction.
Nat. Methods 17, 1214–1221. https://doi.org/10.1038/s41592-020-
70. Verma, R., Mohl, D., and Deshaies, R.J. (2020). Harnessing the Power of
00990-8.
Proteolysis for Targeted Protein Inactivation. Mol. Cell 77, 446–460.
https://doi.org/10.1016/j.molcel.2020.01.010. 87. Punjani, A., and Fleet, D.J. (2021). 3D variability analysis: Resolving contin-
71. Teng, M., and Gray, N.S. (2023). The rise of degrader drugs. Cell Chem. uous flexibility and discrete heterogeneity from single particle cryo-EM.
Biol. 30, 864–878. https://doi.org/10.1016/j.chembiol.2023.06.020. J. Struct. Biol. 213, 107702. https://doi.org/10.1016/j.jsb.2021.107702.

72. Zheng, S.Q., Palovcak, E., Armache, J.P., Verba, K.A., Cheng, Y., and 88. Rohou, A., and Grigorieff, N. (2015). CTFFIND4: Fast and accurate defocus
Agard, D.A. (2017). MotionCor2: anisotropic correction of beam-induced estimation from electron micrographs. J. Struct. Biol. 192, 216–221.
motion for improved cryo-electron microscopy. Nat. Methods 14, https://doi.org/10.1016/j.jsb.2015.08.008.
331–332. https://doi.org/10.1038/nmeth.4193. 89. Murphree, T.A., Vorauer, C., Brzoska, M., and Guttman, M. (2020). Imida-
73. Mirdita, M., Schütze, K., Moriwaki, Y., Heo, L., Ovchinnikov, S., and Stei- zolium Compounds as Internal Exchange Reporters for Hydrogen/
negger, M. (2022). ColabFold: making protein folding accessible to all. Nat. Deuterium Exchange by Mass Spectrometry. Anal. Chem. 92, 9830–
Methods 19, 679–682. https://doi.org/10.1038/s41592-022-01488-1. 9837. https://doi.org/10.1021/acs.analchem.0c01328.
74. Pettersen, E.F., Goddard, T.D., Huang, C.C., Couch, G.S., Greenblatt, 90. Watson, M.J., Harkewicz, R., Hodge, E.A., Vorauer, C., Palmer, J., Lee,
D.M., Meng, E.C., and Ferrin, T.E. (2004). UCSF Chimera–a visualization K.K., and Guttman, M. (2021). Simple Platform for Automating Decoupled
system for exploratory research and analysis. J. Comput. Chem. 25, LC-MS Analysis of Hydrogen/Deuterium Exchange Samples. J. Am. Soc.
1605–1612. https://doi.org/10.1002/jcc.20084. Mass Spectrom. 32, 597–600. https://doi.org/10.1021/jasms.0c00341.

Cell 187, 1–17, December 26, 2024 17

 Please cite this article in press as: Cao et al., Recognition of BACH1 quaternary structure degrons by two F-box proteins under oxidative
stress, Cell (2024), https://doi.org/10.1016/j.cell.2024.10.012

ll
OPEN ACCESS Article

STAR+METHODS

KEY RESOURCES TABLE

REAGENT or RESOURCE SOURCE IDENTIFIER

Antibodies
BACH1 (1:1000) Bethyl Cat# A303-058A; RRID: AB_10894145
FBXO22 (1:1000) Proteintech Cat# 13606-1-AP; RRID: AB_2104403
NRF2 (1:1000) Cell Signaling Cat# 12721S; RRID: AB_2715528
KEAP1 (1:10 000) Proteintech Cat# 10503-2-AP; RRID: AB_2132625
HMOX1 (1:1000) Bethyl Cat# A303-662A; RRID: AB_11205464
SKP1 (1:5000) Michele Pagano’s lab N/A
CUL1 (1:1000) Thermo Fisher Scientific Cat# 718700; RRID: AB_2534002
HA (1:1000) Sigma-Aldrich Cat# H3663; RRID: AB_262051
FLAG (1:4000) Sigma-Aldrich Cat# F1804; RRID: AB_262044
Vinculin (1:1000) Bethyl Cat# A302-535A;
RRID: AB_1999080
NCOR1 (1:1000) Cell Signaling Technologies Cat# 5948S; RRID: AB_10834809
MAFF (1:1000) Proteintech Cat# 12771-1-AP; RRID: AB_2137677
PCNA (1:1000) Santa Cruz Biotechnologies Cat# sc-56; RRID: AB_628110
Cyclin D1 (1:1000) Thermo Fisher Scientific Cat# MA1-39546; RRID: AB_11000217
GAPDH (1:1000) Cell Signaling Cat# 97166; RRID: AB_2756824
b-actin (1:5000) Sigma-Aldrich Cat# A5441; RRID: AB_476744
Bacterial and virus strains
E. coli NEB5a New England Biolabs Cat# C2992H
E. coli NEB Stable New England Biolabs Cat# C3040H
E. coli BL21 (DE3) New England Biolabs Cat# C2527I
E. coli DH5a New England Biolabs Cat# C2987I
E. coli DH10Bac Life Technologies Cat# 10361012
Chemicals, peptides, and recombinant proteins
Lipofectamine 3000 Thermo Fisher Scientific Cat# L3000150
Cycloheximide Sigma-Aldrich Cat# C7698-1G
MG132 Peptides International Cat# IZL-3175v
MLN-4924 Active Biochem Cat# A-1139
Spermine NONOate Enzo Life Sciences Cat# ALX-430-013-5005
S-1-propenyl-L-cysteine MedChemExpress Cat# HY-111827
S-Nitrosoglutathione EMD Millipore Cat# 487920
NOR3 Sigma-Aldrich Cat# E2895
Hemin Sigma-Aldrich Cat# 51280-1G
Critical commercial assays
S-Nitrosylated Protein Detection Kit (Biotin Switch) Cayman Chemical Cat# 10006518
BCA Protein Assay Kit Thermo Fisher Scientific Cat# 23225
Mycostrip, Mycoplasma Detection Kit Invivogen Cat# rep-mys-50
Deposited data
FBXO22-BACH1-BTB structure This paper PDB: 8UA3
SCFFBXO22-BACH1-BTB structure This paper PDB: 8UA6
Monomeric FBXL17-BACH1-BTB structure This paper PDB: 8UAH
Monomeric SCFFBXL17-BACH1-BTB structure This paper PDB: 8UBT
Dimeric SCFFBXL17-BACH1-BTB structure This paper PDB: 8UBU
(Continued on next page)

e1 Cell 187, 1–17.e1–e10, December 26, 2024

 Please cite this article in press as: Cao et al., Recognition of BACH1 quaternary structure degrons by two F-box proteins under oxidative
stress, Cell (2024), https://doi.org/10.1016/j.cell.2024.10.012

ll
Article OPEN ACCESS

Continued
REAGENT or RESOURCE SOURCE IDENTIFIER
Dimeric FBXL17LRR-BACH1-BTB structure This paper PDB: 8UBV
EM map of FBXO22-BACH1-BTB This paper EMDB: EMD-42049
EM map of SCFFBXO22-BACH1-BTB This paper EMDB: EMD-42051
EM map of monomeric FBXL17-BACH1-BTB This paper EMDB: EMD-42064
EM map of monomeric SCFFBXL17-BACH1-BTB This paper EMDB: EMD-42102
EM map of dimeric SCFFBXL17-BACH1-BTB This paper EMDB: EMD-42105
close conformation
EM map of dimeric FBXL17LRR-BACH1-BTB This paper EMDB: EMD-42106
EM map of dimeric SCFFBXL17-BACH1-BTB open conformation This paper EMDB: EMD-42115
HDX-MS data This paper ProteomeXchange: PXD053923
Original images of western blot data This study; Mendeley Data https://dx.doi.org/10.17632/vf73rr8p5n.1
Proteomic data This study ProteomeXchange: PXD056513
Experimental models: Cell lines
HEK293T (female) ATCC Cat# CRL-3216; RRID: CVCL_0063
HCT-116 (male) ATCC Cat# CCL-247; RRID: CVCL_0291
FBXO22 -/- HCT-116 (male) This paper N/A
FBXL17 -/- HCT-116 (male) This paper N/A
FBXO22 -/-; FBXL17 -/- HCT-116 (male) This paper N/A
Spodoptera frugiperda Sf9 Life Technologies Cat# B825-01
Trichoplusia ni Hi5 Life Technologies Cat# B85502
Oligonucleotides
Full list of oligos See Table S2 N/A
Recombinant DNA
pcDNA3.1 2xFLAG-2xSTREP-BACH1 Lignitto et al.30 N/A
pcDNA3.1 HA-BACH1 This paper N/A
pcDNA3.1 2xFLAG-2xSTREP-BACH2 This paper N/A
pcDNA3.1 2xFLAG-2xSTREP-FBXO22 Lignitto et al.30 N/A
pcDNA3.1 HA-FBXO22 This paper N/A
pcDNA3.1 FLAG-FBXL17 This paper N/A
pcDNA3.1 HA-FBXL17 This paper N/A
pSpCas9(BB)-2A-GFP (px458) Addgene Plasmid# 48138
pcDNA3.1 2xFLAG-2xSTREP FBXO22 (D366A) This paper N/A
pcDNA3.1 2xFLAG-2xSTREP FBXO22 (D366R) This paper N/A
pcDNA3.1 2xFLAG-2xSTREP FBXO22 (I368A) This paper N/A
pcDNA3.1 2xFLAG-2xSTREP FBXO22 (N372A) This paper N/A
pcDNA3.1 2xFLAG-2xSTREP FBXO22 (F373D) This paper N/A
pcDNA3.1 2xFLAG-2xSTREP FBXO22 (K377A) This paper N/A
pcDNA3.1 2xFLAG-2xSTREP FBXO22 (K377D) This paper N/A
pcDNA3.1 2xFLAG-2xSTREP FBXO22 (L375A) This paper N/A
pcDNA3.1 2xFLAG-2xSTREP FBXO22 (L375D) This paper N/A
pcDNA3.1 2xFLAG-2xSTREP-BACH1 (F9Y) This paper N/A
pcDNA3.1 2xFLAG-2xSTREP-BACH1 (a6-BACH2) This paper N/A
pcDNA3.1 2xFLAG-2xSTREP-BACH2 (a6-BACH1) This paper N/A
pcDNA3.1 HA-BACH1 (E12A) This paper N/A
pcDNA3.1 HA-BACH1 (C122E) This paper N/A
pcDNA3.1 HA-BACH1 (F125A) This paper N/A
pcDNA3.1 HA-BACH1 (F125D) This paper N/A
pcDNA3.1 HA-BACH1 (F128A) This paper N/A
(Continued on next page)

Cell 187, 1–17.e1–e10, December 26, 2024 e2

 Please cite this article in press as: Cao et al., Recognition of BACH1 quaternary structure degrons by two F-box proteins under oxidative
stress, Cell (2024), https://doi.org/10.1016/j.cell.2024.10.012

ll
OPEN ACCESS Article

Continued
REAGENT or RESOURCE SOURCE IDENTIFIER
pcDNA3.1 HA-BACH1 (F128D) This paper N/A
pcDNA3.1 HA-BACH1 BTB (K129*), *: stop codon This paper N/A
pcDNA3.1 HA-FBXL17 DCTH (T675*), *: stop codon This paper N/A
pFB-GST-FBXO22 This paper N/A
pFB-His-SKP1 This paper N/A
pACE-His-MBP-FBXL17 This paper N/A
pAL-His-BACH1-BTB This paper N/A
pET-GST-BACH1-BTB This paper N/A
pET-GST-Avi-BACH1-BTB This paper N/A
pAL-His-Avi-BACH1-BTB This paper N/A
pAL-His-Avi-BACH1-BTB (C34A) This paper N/A
pAL-His-Avi-BACH1-BTB (C55A) This paper N/A
pAL-His-Avi-BACH1-BTB (C107A) This paper N/A
pAL-His-Avi-BACH1-BTB (C109A) This paper N/A
pAL-His-Avi-BACH1-BTB (C122A) This paper N/A
pAL-His-Avi-BACH1-BTB (C122E) This paper N/A
pET-His-Avi-BACH1-BTB This paper N/A
pET-His-Avi-BACH1-BTB (S17C/V20I/V32I/S50A/H60W/S61Q) This paper N/A
pAL-MBP-BACH1-BTB This paper N/A
pAL-MBP-BACH1-BTB (S17C/V20I/V32I/S50A/H60W/S61Q) This paper N/A
pET-GST-Avi-BACH2-BTB This paper N/A
pET-GST-BACH2-BTB This paper N/A
pAL-His-BACH2-BTB This paper N/A
pAL-MBP-BACH2-BTB This paper N/A
pET-MSB-BACH1-BTB This paper N/A
pACE-His-Venus-BACH1-FL This paper N/A
pET-MSB-FLAG-BACH1-BTB This paper N/A
pET-GST-MAF This paper N/A
pAL-His-MBP-BACH1-bZIP This paper N/A
pET-GST-NCOR1 This paper N/A
pET-GST-GLMN This paper N/A
pET-His-RBX1-His-CUL1 This paper N/A
Software and algorithms
ImageJ Software 2.0 NIH https://imagej.net/software/imagej2/
GraphPad Prism v9.5.1 Graphpad Software https://www.graphpad.com
MotionCor2 Zheng et al.72 https://emcore.ucsf.edu/ucsf-software
AlphaFold-Multimer in Google ColabFold Mirdita et al.73 https://colab.research.google.com/
github/sokrypton/ColabFold/blob/
main/AlphaFold2.ipynb
UCSF Chimera Pettersen et al.74 https://www.cgl.ucsf.edu/chimera/
CryoSPARC Structura Biotechnology Inc. https://cryosparc.com
75
Coot Emsley et al. https://www2.mrc-lmb.cam.ac.uk/
personal/pemsley/coot/
PHENIX Adams et al.76 https://phenix-online.org
Afonine et al.77
PyMOL Schrödinger https://www.pymol.org
HDExaminer v3 Sierra Analytics https://massspec.com/hdexaminer/
HXExpress v2 Guttman et al.78 https://www.hxms.com/HXExpress/
(Continued on next page)

e3 Cell 187, 1–17.e1–e10, December 26, 2024

 Please cite this article in press as: Cao et al., Recognition of BACH1 quaternary structure degrons by two F-box proteins under oxidative
stress, Cell (2024), https://doi.org/10.1016/j.cell.2024.10.012

ll
Article OPEN ACCESS

Continued
REAGENT or RESOURCE SOURCE IDENTIFIER
Other
HiTrap Q HP GE Healthcare Cat# 17115401
Superdex 200 Increase 10/300 GL GE Healthcare Cat# 28990944
Quantifoil R1.2/1.3 200 mesh copper grid Electron Microscopy Sciences Cat# Q250-CR1.3

EXPERIMENTAL MODEL AND STUDY PARTICIPANT DETAILS

Cell lines
Cell lines were purchased from ATCC. HCT-116 (male, ATCC CCL-247) cells were propagated in McCoy’s 5A medium (Gibco).
HEK293T (female, ATCC CRL-3216) cells were propagated in Dulbecco’s modified Eagle’s medium (DMEM) (Gibco). Tissue culture
media were supplemented with 10 % fetal bovine serum (FBS) (Corning Life Sciences) and 1 % penicillin-streptomycin (Corning Life

Sciences). Cells were maintained at 37 C and 5 % CO2 in a humidified atmosphere. Cells were routinely monitored for Mycoplasma
contamination using the MycoStrip - Mycoplasma Detection Kit (Invivogen). No cell lines used in this study were found in the data-
base of commonly misidentified cell lines that is maintained by ICLAC and NCBI Biosample. Specific details about cell lines used are
provided in the key resources table.

Source organism
For DNA extraction, E. coli DH5a, NEB5a, and NEB stable was used. For bacmid production, E. coli DH10Bac was used. For bacu-
lovirus production and amplification, Sf9 insect cells were used. For protein expression, both E. coli BL21 (DE3) and Hi5 insect cells
were used. E. coli grew at 37  C in Luria-Bertani medium (Fisher). Protein expression was induced by 0.2-0.4 mM Isopropyl b-D-1-
thiogalactopyranoside (IPTG) at 16  C overnight. Insect cells grew at 27  C in Grace’s Insect Medium (Gibco) supplemented with
10 % fetal bovine serum (Cytiva) and 1 % penicillin-streptomycin-glutamine (Gibco).

METHOD DETAILS

Protein preparation
Human FBXO22 (N-terminally GST-tagged) and SKP1 (N-terminally His-tagged) were co-expressed and purified from Hi5 insect cells
using glutathione affinity chromatography followed by TEV cleavage to remove the tags if necessary. The tag-free FBXO22-SKP1
complex was further purified by anion exchange and size exclusion chromatography. The wild-type and mutant BACH1-BTB domain
(amino acids 7-128) or BACH2-BTB domain (amino acids 9-131) constructs were expressed with an N-terminal His- or GST-tag in
E. coli BL21. The full-length BACH1 protein was constructed with an N-terminal His-Venus tag and expressed in Hi5 insect cells. His
or GST-tagged BACH1 or BACH2 was purified by Ni-NTA or glutathione resin with subsequent TEV protease treatment to remove the
tags if necessary. BACH1 or BACH2 was further purified by anion exchange and/or size exclusion chromatography. To prepare
biotin-tagged BACH1-BTB or BACH2-BTB, an additional Avi-tag was designed after the TEV cleavage site. After TEV cleavage of
the His- or GST-tag, Avi-BACH1-BTB or Avi-BACH2-BTB was generated and subsequently biotinylated in a biotin ligase BirA-cata-
lyzed biotinylation reaction. Biotinylated protein was further purified by size exclusion chromatography. To make the BACH1-BTB-
BACH2-BTB heterodimer, BACH1-BTB and BACH2-BTB with different tags were co-expressed in E. coli BL21 and isolated by
tandem purification. For example, MBP-tagged BACH2-BTB/MSB-tagged BACH1-BTB were purified by amylose resin (for MBP-
tagged protein) and anion exchange (for MSB-tagged protein, eluted at high salt
450 mM NaCl). In another case, GST-Avi-
BACH1-BTB-His-BACH2-BTB were tandemly purified, biotinylated and TEV treated to generate the biotin-BACH1-BTB-
tag-free BACH2-BTB heterodimer. The FBXL17-binding defective BACH1-BTBMT (S17C/V20I/V32I/S50A/H60W/S61Q) was used
to prepare the BACH1-BTBWT-BACH1-BTBMT heterodimer. MBP-tagged BACH1-BTBMT and His-Avi-tagged BACH1-BTBWT
were co-expressed in E. coli BL21, purified by Ni-NTA resin and subject to biotinylation. Amylose resin was subsequently used to
purify the MBP-BACH1-BTBMT-His-biotin-BACH1-BTBWT heterodimer and remove extra biotin. The MBP- BACH1-BTBWT-His-
biotin-BACH1-BTBWT heterodimer was made by the same method. GST-tagged NCOR1 was expressed in E. coli BL21 and purified
by glutathione resin and size exclusion chromatography. GST-MAF (254-361) and MBP-BACH1-bZIP (510-658) were co-expressed
in E. coli BL21 and subject to tandem affinity purification. GLMN with an N-terminal GST tag was expressed and purified from E. coli
BL21 using glutathione affinity chromatography followed by TEV cleavage to remove the tag. Tag-free GLMN was further purified by
anion exchange and size exclusion chromatography. RBX1 (amino acids 16-108) and CUL1 with two short unstructured segments
(amino acids 1-12 and 58-81) truncated were constructed with an N-terminal His-tag followed by a TEV cleavage site. After co-
expression in E. coli BL21(DE3), the CUL1-RBX1 complex was purified by a Ni-NTA affinity column followed by TEV cleavage to
remove the tag. The tag-free CUL1-RBX1 was further purified by gel filtration. Separately purified BACH1-BTB, FBXO22-SKP1,
CUL1-RBX1, and GLMN were mixed at the molar ratio of 3:1:1:1.5 with BACH1 and GLMN at extra to assemble and purify the
GLMN-SCFFBXO22-BACH1 complex on size exclusion chromatography for cryo-EM analysis. FBXL17 (amino acids 310–701) with an

Cell 187, 1–17.e1–e10, December 26, 2024 e4

 Please cite this article in press as: Cao et al., Recognition of BACH1 quaternary structure degrons by two F-box proteins under oxidative
stress, Cell (2024), https://doi.org/10.1016/j.cell.2024.10.012

ll
OPEN ACCESS Article

N-terminal His-MBP-tag and SKP1 with an N-terminal His-tag were co-expressed and purified from Hi5 insect cells using Ni-NTA
affinity chromatography followed by TEV cleavage to remove the tags if necessary. GST-BACH1-BTB without or with NOR3-
S1PC treatment and FBXL17-SKP1 were incubated for 16 hours or at various times, respectively, and purified by a glutathione affinity
column. Thereafter, CUL1-RBX1 and GLMN were loaded on the above column to form the GLMN- SCFFBXL17-BACH1-BTB complex.
After on-column cleavage by TEV, the tag-free complex was collected in the flow-through of the glutathione affinity column and
was further purified by size exclusion chromatography for cryo-EM analysis.

Affinity pull-down assay

The GST, Ni2+, and biotin pull-down assays were performed using
15-80 mg of purified GST-, His-, or biotin-tagged protein as the
bait. Their binding partners were applied in excess. BACH1-BTB was pre-treated with the saturating amount of NOR3 and S1PC at
1.8 mM overnight at 4 C if indicated. Reaction mixtures with various incubation times were incubated with 20 mL glutathione, Ni2+ or
Strep-Tactin beads at 4 C with gentle shaking. After extensive wash with the buffer containing 20 mM Tris, pH 8.0, 150 mM NaCl,
0.25 mM TCEP (additional 20 mM imidazole for the Ni2+ pull-down assay), the protein complexes on the beads were eluted by
40 ml 20 mM glutathione, 300 mM imidazole, or 50 mM biotin. The eluted samples were resolved by SDS-PAGE and analyzed by
Coomassie staining.

BioLayer interferometry
The binding affinity between FBXO22 (or FBXL17) and BACH1-BTB was measured using the Octet Red 96 (ForteBio, Pall Life Sci-
ences) following the manufacturer’s procedures. The optical probes were coated with streptavidin, loaded with 200 nM biotinylated
BACH1-BTB (without or with NOR3-S1PC treatment), or anti-GST probes loaded with 200 nM GST-FBXO22-SKP1. Subsequently,
the probes were quenched with 0.5 mM biocytin or 1 mM GST protein prior to kinetic binding analysis. The reactions were carried out
in black 96 well plates maintained at 30 C. The reaction volume was 200 mL in each well. The binding buffer contained 25 mM HEPES,
pH 7.4, 100 mM NaCl, 1 mM TCEP, 0.1% Tween-20, and 0.05 mg/mL Bovine Serum Albumin. As the analyte, BACH1-BTB or
FBXL17-SKP1 was tested at various concentrations. There was no binding of the analyte to the unloaded probes. Binding kinetics
of the analyte at different concentrations were measured simultaneously. The data were analyzed by the Octet data analysis soft-
ware. The dissociation constant (Kd) was determined from the steady-state equilibrium measurements. All BLI experiments have
been repeated at least 2-3 times.

AlphaLISA
Amplified luminescent proximity homogeneous assays for measuring protein-protein interactions were performed using EnSpire
reader (PerkinElmer). Biotinylated BACH1-BTB was immobilized to streptavidin-coated AlphaScreen donor beads. His-SKP1-
FBXO22 was bound to anti-His acceptor beads. The acceptor and donor beads were brought into proximity by the interactions be-
tween the E3 and its substrate. Excitation of the donor beads by a laser beam of 680 nm promotes the formation of singlet oxygen.
When an acceptor bead is in close proximity, the singlet oxygen reacts with thioxene derivatives in the acceptor beads and causes
the emission of 520-620 nm photons, which are detected as the binding signal. If the beads are not in close proximity to each other,
the oxygen will return to its ground state and the acceptor beads will not emit light. Competition assays were performed by titrating
GST-tagged NCOR1 fragments or GST alone at various concentrations. To establish the binding signal, the concentration of tagged
E3 and biotinylated BACH1-BTB were first titrated with fixed amount of donor and acceptor beads. The concentration of each
component was selected based on the rising AlphaScreen signal before the beads became saturated. 25 nM His-SKP1-FBXO22
and 8.3 nM biotinylated BACH1-BTB were used. The binding assay buffer contained 25 mM HEPES, pH 7.4, 100 mM NaCl,
1 mM TCEP, 0.1% Tween-20, and 0.05 mg/mL Bovine Serum Albumin. 5 mg/mL donor and acceptor beads were used in the assays.
The experiments were performed in three or four replicates. IC50 was determined using non-linear curve fitting of the dose response
curves generated with Prism (GraphPad).

SEC–MALS
Size-exclusion chromatography coupled with light scattering, refractive index, and ultraviolet absorption (SEC-LS-RI-UV) was done
using a SEC-MALS system consisting of an Agilent HPLC pump with a 1260 Infinity lamp, a Dawn light scattering instrument (Wyatt),
an OptiLab refractive index instrument (Wyatt), and the Superdex 200 Increase 10/300 GL (GE Healthcare). The column was equil-
ibrated with 20 mM Tris-HCl pH 8.0, 150 mM NaCl, and 0.25 mM TCEP. Tag-free FBXO22-SKP1, His-BACH1-BTB and their mixture
were analyzed at room temperature. Similarly, His-MBP-FBXL17-SKP1, NOR3-S1PC-treated His-BACH1-BTB and their mixture
with different incubation times were analyzed.

Differential scanning fluorimetry (DSF)

The Applied Biosystems StepOnePlus Real-Time PCR machine was used to perform the differential scanning fluorimetry thermosta-
bility assay. 5 mg BACH1 protein was pre-treated with the saturating amount of NOR3 and S1PC at 100 mM overnight at 4 C. For
negative control, 5 mg BACH1 was treated with the same amount of the compound solvent (DMSO). 10 ml protein was mixed with
10 ml 0.25% Protein Thermal Shift Dye (Applied Biosystems) in a MicroAmp Fast 96-well reaction plate (0.1 ml). The temperature
increment was 0.3  C, recording fluorescence from 30 to 90 C. To compare WT BACH1-BTB and the BACH1-BTB C122E mutant,

e5 Cell 187, 1–17.e1–e10, December 26, 2024

 Please cite this article in press as: Cao et al., Recognition of BACH1 quaternary structure degrons by two F-box proteins under oxidative
stress, Cell (2024), https://doi.org/10.1016/j.cell.2024.10.012

ll
Article OPEN ACCESS

2.5 mg of each protein was mixed with Protein Thermal Shift Dye for analysis. The temperature increment was 0.3  C, recording
fluorescence from 35 C to 90 C. The data was analyzed by Protein Thermal Shift software 1.4. The melting temperature (Tm) was
determined based on the peak of the first derivative of the fluorescence vs temperature plots. The experiments were performed in
three replicates.

In vitro ubiquitination assay

A reaction mixture containing 0.5-4 mM His-Venus-BACH1 (full length) or 1mM MSB-Flag-BACH1-BTB, 1-3 mM FBXO22-SKP1 or
FBXL17-SKP1, 0.5-1 mM NEDD8
CUL1-RBX1, 0.5-1 mM E2(s) (UBCH5 alone or together with CDC34), 0.4 mM UBE1 and
50-100 mM ubiquitin and 2 mM ATP/10 mM MgCl2 was incubated at 37  C for 0.5-1.5 h with multiple controls, each lacking one
component of the reaction. 100 mM NCOR1 was used and pre-mixed with BACH1 for 10 mins if indicated. To monitor the effect
of NO treatment on ubiquitination, BACH1 was pre-treated with 1.8 mM NOR3 and S1PC. The reaction mixtures were resolved
by a 10% SDS–PAGE gel. The BACH1 signal was monitored by detecting the Venus-fluorescence absorbance at 488 nm or anti-
FLAG-HRP western blots. The absorbance at 647 nm was used to detect the protein markers.

Electrophoretic mobility shift assay (EMSA)

1-3 mM GST-MAF or GST-MAF/MBP-BACH1-bZIP was incubated with 100 nM HS2-700 DNA probe (5’-5IRD700/AGCACAG
CAGTGCTGAGTCATGCTGAGTCATGCTGAGGCTT-3’) annealed with unlabeled complementary DNA strand in sub-stoichiometric
amount. The MAF/BACH1 complex was pre-treated with NOR3 and S1PC if indicated. To overcome the effect of 250 mM TCEP used
in the protein buffer, NOR3-S1PC was titrated at a concentration of 66 mM, 200 mM, 600 mM, and 1.8 mM. Samples were separated
using 5% Mini-PROTEAN TBE Gels (Bio-Rad) with running buffer (45 mM Tris, 45 mM boric acid, 1 mM EDTA, pH 8.2). The TBE gels
were run in the cold room (4  C). The absorbance at 680 nm was used to detect the DNA probes.

Cryo-EM sample preparation and data collection

The GLMN-SCFFBXO22-BACH1-BTB complex and the GLMN-SCFFBXL17-BACH1-BTB complex (without or with NOR3-S1PC treatment with
different incubation times) were concentrated to
2.0 mg/ml for cryo-EM sample preparation. For each complex,
3 mL sample was
applied to a glow-discharged Quantifoil R1.2/1.3 200 mesh copper grid. The grid was subsequently blotted for
3 s (blotting force
at -2, temperature at 10  C, and relative humidity at 100%), plunged and flash frozen into liquid ethane using a Vitrobot Mark IV sys-
tem (Thermo Fisher Scientific), and stored in liquid nitrogen for data collection. The grids were inspected and screened with Talos
Glacios equipped with a K3 camera. For the GLMN-SCFFBXO22-BACH1-BTB sample, data collection was carried out on a Titan Krios
transmission electron microscope (Thermo Fisher Scientific) operated at 300 kV at the University of Washington. The automation
scheme was implemented using the Leginon software79 at a nominal magnification of 105 K, resulting in a physical pixel size of
0.84 Å. The zero-loss-energy images were acquired on a Gatan K3 direct detector operated in super-resolution counting mode (pixel
size: 0.42 Å) with the slit width of post-column Gatan BioQuantum GIF energy filter set to be 20 eV. The dose rate was adjusted to 13.9
e-/Å2S, and a total dose of 59 e-/Å2 for each image fractionated into 50 frames. The images were recorded at a defocus range of
-0.8
3.5 mm. For the chemical-treated or untreated GLMN-SCFFBXL17-BACH1-BTB complex, data collection was carried out on a
300 kV FEI Titan Krios transmission electron microscope operated at 300 kV at the HHMI Janelia Research Campus. The automated
data collection was carried out with SerialEM software80 using beam-image shift81 at a nominal magnification of 165 K. The images
were acquired on Falcon 4i direct electron detector operated in electron counting mode (pixel size: 0.743 Å) with the slit width
of Selectris X (Thermo Fisher Scientific) energy filter set to be 6 eV. The dose rate was set to 15.39 e-/Å2S, and a total dose of
60 e-/Å2 for each image fractionated into EER (Electron-event representation) frames.82 The images were recorded at a defocus
range of 0.8
1.5 mm.

Image processing and 3D reconstruction

For the GLMN-SCFFBXO22-BACH1-BTB sample, 6,131 movies were acquired. The beam-induced motion of each micrograph stack was
corrected by MotionCor2.72 The motion-corrected micrographs were imported into CryoSPARC for the following processing.83 The
defocus parameter of each motion-corrected micrograph was determined by Gctf.84 4,991 micrographs were kept after filtering the
micrographs with CTF parameters and visual inspection. An ab-initio reconstruction from a previous data collection was used to
generate 2D projections as templates for particle picking. 2,888,035 particles were picked, extracted, and subjected to 2D classifi-
cation. After two rounds of 2D classification, 243,712 particles were selected as training particles for Topaz particle picking.85
3,645,675 particles were picked with Topaz. After data cleaning by 2D classification, 949,569 particles were kept and subjected
to ab-initio reconstruction. Subsequently, all the particles were used for heterogeneous refinement. 413,179 particles from good
reconstruction were selected for non-uniform refinement86 to generate a reconstruction with an overall resolution of 3.9 Å. To better
resolve the interface between the substrate and its receptor, a soft mask focused on BACH1-BTB and FBXO22-SKP1 was applied to
3D Variability Analysis.87 171,378 particles were chosen for local refinement. The BACH1-BTB-FBXO22-SKP1 subcomplex was
refined to 3.8 Å. More details about the data processing can be found in Figure S1.
For the native GLMN-SCFFBXL17-BACH1-BTB sample, 12,618 movies were collected. The beam-induced motion of micrographs was
corrected by Patch Motion Correction in CryoSPARC. The defocus of motion-corrected micrographs was estimated by CTFFIND4.88
12,423 micrographs were kept while the rest were rejected by CTF parameters. These micrographs were subjected to Blob particle

Cell 187, 1–17.e1–e10, December 26, 2024 e6

 Please cite this article in press as: Cao et al., Recognition of BACH1 quaternary structure degrons by two F-box proteins under oxidative
stress, Cell (2024), https://doi.org/10.1016/j.cell.2024.10.012

ll
OPEN ACCESS Article

picker and two rounds of 2D classification sequentially. 497,102 out of 3,087,124 particles were picked for heterogeneous refinement
using reconstructions from the previous dataset as templates. 267,523 good particles from the refinement were selected and further
cleaned by 2D classification. 84,154 particles were chosen as training particles for Topaz particle picking. 7,309,475 particles were
picked with Topaz and subjected to two rounds of 2D classification. 1,099,238 particles were selected for ab-initio reconstruction
and heterogeneous refinement sequentially. 364,786 particles from one good class were applied to non-uniform refinement. The
overall resolution of the complex in the monomeric form was refined to 3.1 Å. To improve the local density of FBXL17-BACH1-
BTB, all particles were subjected to 3D Variability. 193,908 particles were selected for local refinement to generate the SKP1-
FBXL17-BACH1-BTB subcomplex with a resolution of 3.2 Å. In addition to the above complex in the monomeric form, the dimeric
GLMN-SCFFBXL17-BACH1-BTB complex was detected as well. 30,123 particles of the dimer were selected as training particles for
Topaz particle picking. 244,189 out of 2,969,728 particles were chosen for heterogeneous refinement. Two good classes represent-
ing different conformations, named dSCFBTB-I and dSCFBTB-II, were kept with 54,546 particles and 47,336 particles, respectively. The
C2 symmetry was imposed for non-uniform refinement. The overall resolution of the two classes reached 4.4 Å and 4.5 Å, respec-
tively. To overcome the pseudo-symmetry of the dimeric reconstructions, all particles of dSCFBTB-I were subjected to symmetry
expansion in CryoSPARC. Afterward, a soft mask focused on the leucine-rich repeat domain of FBXL17 and BACH1-BTB was
applied to subtract the signal from raw particles. These subtracted particles were used for local refinement. The density of the
FBXL17-LRR-BACH1-BTB structure was refined to 4.1 Å. See Figure S7 for more details.
For the compound-treated GLMN-SCFFBXL17-BACH1-BTB sample, the collected data was processed with CryoSPARC in a similar
manner. The data processing workflow was shown in Figure S6. Briefly, 22,720 EER movies were collected and subjected to patch
motion correction and patch CTF estimation sequentially. After filtering with CTF parameters, 20,028 micrographs were kept for the
subsequent blob picker and 2D classification. Representative particles were selected as the training data for Topaz particle picking.
After heterogeneous refinement, a monomeric SCFFBXL17-BACH1 class yielded a 3D reconstruction at 5.7 Å resolution. A dimeric
SCFFBXL17-BACH1 with 461 particles was selected and employed for Topaz particle picking. After 2D classification and heterogeneous
refinement, two good classes of dimeric SCFFBXL17-BACH1 were kept. 10,173 particles and 7,944 particles from dimeric reconstruc-
tions were subjected to non-uniform refinement with C2 symmetry imposed. After refinement, the overall resolution of two classes
both reached 7.8 Å.

Model building and refinement

The initial structural model of FBXO22-SKP1-BACH1-BTB was predicted with AlphaFold-Multimer in Google ColabFold73 and fitted
into the 3.8 Å Cryo-EM map by using UCSF Chimera.74 Subsequently, the model was inspected and trimmed in Coot75 based on the
protein sequences and the EM density. The model was further improved by real-space refinement in PHENIX76,77 and manual
rebuilding in Coot. To generate an initial model of the monomeric SCFFBXL17-BACH1-BTB complex, BACH1-BTB (PDB: 2IHC) was
applied to superimpose with and replace the KEAP1-BTB subunit of the SCFFBXL17-KEAP1-BTB complex (PDB: 6WCQ) in PyMOL.
This model was fitted into the 3.2 Å cryo-EM map using UCSF Chimera. The BACH1 BTB domain was manually adjusted to fit
the density in Coot. The model was further improved by using real-space refinement in PHENIX and manual rebuilding in Coot.
For the SCFFBXL17-BACH1 dimer, two copies of the monomeric complexes were fitted into the 4.4 Å cryo-EM maps and subsequently
trimmed and adjusted in Coot. The dimeric interface of BACH1-BTB in the SCF complex was rebuilt in Coot since it is completely
different from that of the isolated BACH1-BTB dimer. The model of dimeric FBXL17-BACH1-BTB was refined in PHENIX by real-
space refinement. Cryo-EM data collection and refinement statistics for all structures in this paper were summarized in Table S1.

Plasmids and antibodies

Homo sapiens BACH1, FBXO22, and FBXL17 cDNAs were amplified by PCR using Q5 High Fidelity DNA Polymerase (New England
Biolabs) and sub-cloned into modified pcDNA3.1 vectors containing a variety of N-terminal tags, including 2xFLAG tag-2x Strep tag
and HA tag. Specific details will be provided upon request. Site-directed mutagenesis was performed using Q5 High-Fidelity DNA
polymerase (New England Biolabs) with primers generated using either QuikChange Primer Design (Agilent) or NEBaseChanger
(New England Biolabs) and were described in Table S2. Plasmids and antibodies used in the paper were listed in the key resources
table.

Drug Treatment Procedures

Where indicated, cells were treated with 2 mM MLN-4924 (Active Biochem) for 4 hours or 10 mM MG-132 (Peptides International) for 4
hours, spermine NONOate (sNONO) (Cayman Biochem) was prepared in 0.01N NaOH and used at 100 mM for the indicated times.
S-nitrosoglutathione (GSNO) (Sigma-Aldrich) was prepared in DMSO and used at 100 mM for the times indicated. S-1-propenyl-L-
cysteine (S1PC) (MedChemExpress) was prepared in DMSO and used at 25 mM for the times indicated. Cycloheximide (Sigma-
Aldrich) was prepared in DMSO and used at 100 mM for the times indicated.

Co-immunoprecipitation assay
HEK293T cells were transiently transfected using polyethyleneimine (Polysciences); HCT-116 cells were transiently transfected with
Lipofectamine 3000 (Thermo Fisher Scientific). Where indicated, 48 h after transfection, cells were treated with 2.5 mM MLN-4924 for
4 h before further treatment with either vehicle (DMSO) or the indicated NO donor. Cells were collected by scraping and washed with

e7 Cell 187, 1–17.e1–e10, December 26, 2024

 Please cite this article in press as: Cao et al., Recognition of BACH1 quaternary structure degrons by two F-box proteins under oxidative
stress, Cell (2024), https://doi.org/10.1016/j.cell.2024.10.012

ll
Article OPEN ACCESS

PBS prior to lysis in IP lysis buffer (50 mM Tris-HCl, pH 7.5, 150 mM NaCl, 1 mM EDTA, 1 mM EGTA, 5 mM MgCl2, 0.2% NP-40)
supplemented with protease inhibitor (Complete ULTRA, Roche) and phosphatase inhibitor (Phosphatase Inhibitor Cocktail 2,
Sigma-Aldrich) immediately before the experiment. The supernatant was clarified by centrifugation at 21, 000 x g for 10 minutes
 
at 4 C. For FLAG-tagged proteins, the supernatant was incubated with FLAG-M2 agarose beads (Sigma-Aldrich) for 2h at 4 C.
For HA-tagged proteins, the supernatant was incubated with Pierce Anti-HA magnetic beads (Thermo Fisher Scientific) for 2h at

4 C. The beads were then washed 5 times in lysis buffer before incubating the beads in NuPAGE LDS Sample Buffer (Thermo Fisher

Scientific) supplemented with b-mercaptoethanol (Sigma-Aldrich) at 95 C for 10 minutes. For Western blotting, samples of the whole
cell extract and immunoprecipitates were separated by SDS-PAGE and transferred onto 0.45 mm Immobilon-P PVDF membranes
(Millipore Sigma). Membranes were then blocked in 5% nonfat dried milk / PBST for 30 minutes at room temperature and incubated

with the indicated primary antibodies at 4 C overnight. After washing the membranes, secondary antibodies conjugated with horse-
radish peroxidase were applied. Immunoreactive bands were visualized via application of enhanced chemiluminescence (ECL) re-
agent (Thermo Fisher Scientific). Each Western blot was performed no less than three times to ensure results exhibit a high degree
of reproducibility. Representative Western blots have been selected for each figure.

Cell-based protein stability assay

HCT-116 cells were treated with 100 mM spermine NONOate (sNONO) with or without 100 mg/mL cycloheximide (CHX) for the times
indicated. Cells were then washed with ice-cold PBS supplemented with 100 mM phenylmethanesulfonyl fluoride (PMSF) (Sigma-
Aldrich) before they were collected by scraping. Cell pellets were lysed in RIPA buffer (Pierce) supplemented with protease inhibitor
(Complete ULTRA, Roche) and phosphatase inhibitor (Phosphatase Inhibitor Cocktail 2, Sigma-Aldrich) immediately before the

experiment. The supernatant was clarified by centrifugation at 21, 000 x g for 10 minutes at 4 C. Protein concentrations in cell
lysates were then normalized with the Pierce BCA Protein Assay Kit (Thermo Fisher Scientific), according to the manufacturers’
instructions. Samples of the cell lysates were then prepared with NuPAGE LDS Sample Buffer (Thermo Fisher Scientific) supple-
mented with b-mercaptoethanol (Sigma-Aldrich) and changes in protein levels were visualized via Western blot. All experiments
were performed at least 3 times. Quantification of Western blots were done using ImageJ and graphs were plotted using Prism
9 (Graphpad).

CRISPR-Cas9 genome editing

To generate FBXO22 knockout cells, optimal single guide RNA (sgRNA) target sequences were designed using the Benchling
CRISPR Genome Engineering tool (https://www.benchling.com). FBXO22 gRNA (CTTCGTGTTGAGTAACCTGG) was cloned into
pSpCas9(BB)-2A-GFP (px458), a gift from F. Zhang (Addgene plasmid #48138), using the following primers: CACCGCTTC
GTGTTGAGTAACCTGG and aaacCCAGGTTACTCAACACGAAGC. To generate the FBXL17 knockout and the FBXL17 / FBXO22
double knockout cells, sgRNA targeting FBXL17 (AAAACUGGAAGUCUAAACAA) designed by Synthego was used. Briefly, 5x106
HCT-116 cells were seeded into a 15 cm dish before being transfected with 10 mg sgRNA-containing px458 plasmid using Lipofect-
amine 3000 (Thermo Fisher Scientific). 48h after transfection, the top 5% of GFP-positive cells were sorted using the Beckman
Coulter MoFlo XDP cell sorter (100 mm nozzle). 10 000 cells were plated between two 15 cm dishes (5000 cells per plate) and grown
for about 10-14 days to allow single-cell clones to form colonies. The colonies were then picked and plated into individual wells of a
96-well plate. After 48 hours, the 96-well plate were split into two new 96-well plates, with one plate used for genotyping. For gen-
otyping, genomic DNA was extracted using QuickExtract (Epicentre) and genotyping PCRs were performed with MyTaq HS Red Mix
(Bioline) using the following genotyping primers for FBXO22: GCAACCCTATGCTGGCGTAA and AGATACCTCGCGGGTAGGG; and
for FBXL17: GCTTTACCTTGAAATCACATACACA and AGGTGATCTCTTACGGTGACT. The resulting PCR products were then puri-
fied and sequenced to determine the presence of a CRISPR-mediated frameshift event that results in an early stop codon. Positive
knockout candidates were further validated by Western blot. Five homozygous knockout clones were then pooled together to miti-
gate any clonal effects, and this collection of pooled knockout cells were propagated and subsequently used for experiments.

Chromatin Fractionation Assay

Cells were treated with DMSO, 10 mM hemin, or 100 mM sNONO + 25 mM S1PC for 1h before being washed once with ice-cold PBS
with PMSF, and then harvested by scraping. The cell pellet is then resuspended in CSK buffer (10 mM HEPES, pH 7.4, 100 mM NaCl,
300 mM sucrose, 3 mM MgCl2, 1 mM EGTA, 0.5% Triton X-100, 1 mM DTT, phosphatase inhibitor cocktail 2 (Sigma-Aldrich), and
complete ULTRA EDTA-free protease inhibitor cocktail (Roche)). The CSK suspension is incubated on ice for 5 minutes before being
spun down at 1300xg for 3 minutes at 4oC. The supernatant is transferred to a new tube (‘‘soluble’’ fraction). The pellet is washed once
with CSK buffer, spun down at 1300xg for 3 minutes at 4oC, before the supernatant is discarded. The pellet is then resuspended in
chromatin lysis buffer (50 mM Tris-HCl, pH 7.4, 250 mM NaCl, 50 mM NaF, 1 mM EDTA, 1 mM EGTA, 0.1% Triton X-100, 2 mM MgCl2,
2500 U benzonase nuclease (Sigma-Aldrich), phosphatase inhibitor cocktail 2 (Sigma-Aldrich), and complete ULTRA EDTA-free pro-
tease inhibitor cocktail (Roche)) by vortexing and incubated at 4oC for 15 minutes before sonicating the sample at 4oC for 10 minutes.
After sonication, the sample is then incubated once more at 4oC for 10 more minutes before the suspension is spun down at 21000xg
for 10 minutes at 4oC. The supernatant is transferred to a new tube (‘‘chromatin’’ fraction). The protein concentrations are normalized
using the Pierce BCA Protein Assay kit (Thermo Scientific) before immunoblot for analysis.

Cell 187, 1–17.e1–e10, December 26, 2024 e8

 Please cite this article in press as: Cao et al., Recognition of BACH1 quaternary structure degrons by two F-box proteins under oxidative
stress, Cell (2024), https://doi.org/10.1016/j.cell.2024.10.012

ll
OPEN ACCESS Article

Biotin Switch Assay

Cells were treated with either 100 mM GSNO or 100 mM sNONO for 1h prior to collection. Cells were then subjected to the
S-nitrosylation Protein Detection Kit (Cayman Chemical), according to the manufacturers’ instructions. Briefly, free reactive thiol
groups were blocked with N-ethylmaleimide (NEM) for 30 minutes at room temperature, then excess NEM was removed via acetone
precipitation. Any S-nitrosylated cysteines were subsequently reduced and labelled with a sulfhydryl-specific biotin. Excess reducing
and labelling were removed via acetone precipitation. The final protein pellet was then resuspended in IP lysis buffer (50 mM Tris-HCl,
pH 7.5, 150 mM NaCl, 1 mM EDTA, 1 mM EGTA, 5 mM MgCl2, 0.2% NP-40, 1 mM DTT) supplemented with protease inhibitor (Com-
plete ULTRA, Roche) and phosphatase inhibitor (Phosphatase Inhibitor Cocktail 2, Sigma-Aldrich). The resuspended protein was
incubated with streptavidin Dynabeads MyOne Streptavidin C1 beads (Thermo Fisher Scientific) for 2h at 4 C before the beads
were washed 4 times with IP lysis buffer. Samples of the resuspended protein and the streptavidin bead-bound protein were pre-
pared in NuPAGE LDS Sample Buffer (Thermo Fisher Scientific) supplemented with b-mercaptoethanol (Sigma-Aldrich) and proteins
were assessed via Western blot.

S-nitrosylation site mapping

10 mg of recombinantly purified BACH1 BTB was allowed to react in the presence of 100 mM NOR3 and 100 mM S1PC for 30 minutes
at room temperature before excessing nitric oxide donor was removed by Zeba desalting column pre-equilibrated with 50 mM Tris-
HCl (pH 8.0), 150 mM NaCl. Nitrosylated protein was subsequently treated with the S-nitrosylation Protein Detection Kit (Cayman
Chemical), according to the manufacturers’ instructions. Biotin switch-treated protein reduced with 2.5 mL of 0.2 M dithiothreitol
(Sigma) for 1 hr at 57 C and the samples cooled before adding 2.5 mL of 0.5M iodoacetamide and incubated in the dark for 45 minutes.
The sample was digested with 400 ng of trypsin overnight at room temperature and light agitation. The digestion was stopped by
adding 10% trifluoroacetic acid (TFA) to a final concentration of 0.5% TFA. The resulting peptides were desalted using equilibrated
PierceC18 Spin columns. The samples were washed 3 times with 0.1% TFA followed by a wash with 0.5% acetic acid. The desalted
peptides were eluted using 80% acetonitrile in 0.5% acetic acid. The peptides were dried using a SpeedVac concentrator. Samples
were reconstituted in 0.5% acetic acid and stored at -80 C until further analysis. 40 pmol of each sample was loaded onto an Acclaim
PepMap trap column (2 cm x 75 mm) in line with an EASY-Spray analytical column (50 cm x 75 mm ID PepMap C18, 2 mm bead size)
using the autosampler of an EASY-nLC 1200 HPLC (Thermo Fisher Scientific). Solvent A consists of 2% acetonitrile in 0.5% acetic
acid, and Solvent B of 80 % acetonitrile in 0.5% acetic acid. Peptides were gradient eluted with a flowrate of 200nL/min into an Orbi-
trap Eclipse Tribrid (Thermo Fisher Scientific) mass spectrometer. The gradient used was the following: in 5 min to 5% solvent B,
60 min to 35% solvent B, 10 min to 45% solvent B, and 10 min to 100% solvent B. High resolution full MS spectra were acquired
with a resolution of 120,00, an AGC target of 4e5, with a maximum ion time of 50 ms, and a scan range of 400 to 1500 m/z. All
HCD MS/MS spectra were collected with the following instrument parameters: resolution of 30,000, AGC target of 2e5, maximum
ion of 200ms, one microscan, 2 m/z isolation window, and NCE of 27. The resulting MS/MS spectra were searched against
BACH1 BTB and common contaminants using the search engine Byonic (Protein Metrics). Settings were as following: trypsin cleav-
age allowing for 2 missed cleavages, precursor, and fragment ion mass tolerance of 10 ppm, variable modifications of carbamido-
methyl, N-ethylmaleimide, and 3-(N-Maleimidopropionyl)-biocytin on cysteine, deamidation on asparagine and glutamine, and
oxidation on methionine. Spectra identified as containing modified cysteine residues were manually verified.

HDX-MS analyses
Stock concentrations of BACH1 (unmodified or NO-treated, 0.2 mg/mL) in Tris buffer (20 mM tris, 150 mM NaCl, 0.25 mM TCEP,
pH 8.0) were diluted into 90 mL of deuterated PBS buffer (20 mM phosphate, 150 mM NaCl, 0.02% sodium azide, 1 mM EDTA
pH* 7.77, 85%D final) containing 0.2 nM bradykinin and incubated for either 3 seconds, 1 minute, 30 minutes, or 20 hours at
21 C. Each starting stock also included a mixture of imidazolium compounds to serve as exchange reference standards.89 At the
desired time point the sample was rapidly mixed with an equal volume of ice cold 0.2% formic acid and 0.1% trifluoroacetic acid
(TFA) for a final pH of 2.5. Samples were then immediately frozen on ethanol/dry ice and stored at -80 C until LC-MS analysis.
Undeuterated samples were prepared the same way but with undeuterated buffer for each step. Maximally deuterated samples
were prepared by collecting the Nepenthesin II digested peptides after they were trapped and resolved as described below. These
samples were dried down, resuspended in 10 mL of undeuterated buffer, incubated in deuterated PBS for 2 hours at 21 C, then
quenched and frozen as described above. Some late-eluting peptides were outside of the collection window and therefore are absent
in the maximally deuterated data.
Samples were thawed at 5 C for 4.5 minutes and injected using a custom LEAP robot integrated with an LC-MS system.90 The
protein was first passed over a Nepenthesin II column (2.1 x 30 mm; AffiPro) at 400 mL/min for inline digestion with the protease col-
umn held at 20 C. Peptides were then trapped on a Waters XSelect CSH C18 trap cartridge column (2.1 x 5 mm 2.5 mm) and resolved
over a CSH C18 column (1 x 50 mm 1.7 mm 130Å) using linear gradient of 5 to 35% B (A: 0.1% FA, 0.025% TFA, 5% ACN; B: ACN with
0.1% FA) over 10 minutes and analyzed on a Thermo Orbitrap Ascend mass spectrometer at a resolution setting of 120,000. A series
of washes over the trap and Nepenthesin II columns was used between injections to minimize carry-over.90 Data dependent MS/MS
acquisition was performed on an undeuterated sample of either condition using rapid CID and HCD scans and processed in Byonic
(Protein Metrics) with a score cutoff of 150 to identify peptides. S-nitrosylation and cysteine mono- and di-oxidation were included as
potential variable modifications. Deuterium incorporation was analyzed using HDExaminer v3 (Sierra Analytics). Peptide spectra

e9 Cell 187, 1–17.e1–e10, December 26, 2024

 Please cite this article in press as: Cao et al., Recognition of BACH1 quaternary structure degrons by two F-box proteins under oxidative
stress, Cell (2024), https://doi.org/10.1016/j.cell.2024.10.012

ll
Article OPEN ACCESS

were exported from HDExaminer and screened for multimodal mass envelopes with PyHXExpress (https://github.com/tuttlelm/
pyHXExpress). Peptides of interest were then analyzed with HXExpress.78 Data are available via ProteomeXchange with identifier
PXD053923.

QUANTIFICATION AND STATISTICAL ANALYSIS

Resolution estimations of cryo-EM density maps are based on the Fourier Shell Correlation (FSC) curves at the gold-standard
threshold of 0.143. Protein quantification was done using Bio-rad Protein Assay Dye and comparing readings against a standard
curve of BSA. Quantification of protein bands in Western blots and SDS-PAGE gels were done using ImageJ. Data were analyzed
and plotted by Prism (GraphPad). Unless otherwise noted, data were presented as mean ± SD of n = 3 biologically independent
samples.

Cell 187, 1–17.e1–e10, December 26, 2024 e10

 ll
Article OPEN ACCESS

Supplemental figures
A B

C Template picker
MotionCor2 CTFFIND4 with cryoSPARC 2D classification
6,131 movies 6,131 micrographs 4,991 micrographs 2,888,035 particles 243,712 particles
Manually Curate
Topaz Picking
with cryoSPARC
with cryoSPARC
Remove duplicates 2D classification
949,569 selected particles 1,384,053 particles 3,645,675 particles
Ab-initio reconstruction
Hetergenous refinement
D

30.7 % 43.9 % 16.4 % 9%

291,541 particles 413,179 particles 155,525 particles 85,362 particles

Particle
3D Variability Analysis Particle
Substraction
with cryoSPARC Substraction
171,378 particles
Local Local refinement
refinement with cryoSPARC 3.8 Å
3.6 Å with cryoSPARC
3.9 Å
413,179 particles F

FBXO22 β-strand FBXO22 α-helix BACH1-BTB BACH1-BTB

(aa 92-98) (aa 87-95) β-strand (aa 7-14) α-helix (aa81-91)

Figure S1. The schematic workflow of single-particle reconstruction of the SCFFBXO22-BACH1complex, related to Figure 1
(A) A representative cryo-EM micrograph.
(B) Typical 2D averages of the cryo-EM dataset. Scale bar, 10 nm.

(legend continued on next page)

 ll
OPEN ACCESS Article

(C) The flowchart of single particle analysis of the SCFFBXO22-BACH1complex.

(D) The angular distribution of particles used in the final reconstruction.
(E) Fourier shell correlation (FSC) curves for SCFFBXO22-BACH1. At the gold-standard threshold of 0.143, the resolution is 3.9 Å.
(F) Representative density in local refined EM map.
 ll
Article OPEN ACCESS

B C

D E

(legend on next page)

 ll
OPEN ACCESS Article

Figure S2. Structural and biochemical analysis of FBXO22-BACH1 interaction, related to Figures 1 and 2
(A) A comparison of the three FIST domains of FBXO22 and homologous structure of YabJ from the YjgF superfamily (PDB: 1QD9).
(B) Steric hindrance prevents the formation of a BACH1-BTB-FBXO22 complex with a 2:2 ratio. The asymmetric complex formed between a BACH1 dimer
(protomer A, light blue; subunit B, orange) and FBXO22 (purple) is shown as cartoon diagram. A second copy of FBXO22 shown in slate surface representation is
modeled onto the BACH1-BTB dimer and is in clash with the other FBXO22 macromolecule.
(C and D) In vitro ubiquitination of BACH1 and BACH1-BTB domain by SCFFBXO22.
(E) Superposition of the crystal structures of BACH1-BTB (PDB: 2IHC) and BACH2-BTB (PDB: 3OHU). BACH2 C-terminal region is disordered and highlighted
in red.
(F) Sequence alignment of five NCOR1 vertebrate orthologs and human NCOR2. Highly conserved short linear motifs (SLiMs) are underlined with gray bars. The
BCL6 BTB-interacting motif found in PDB: 1R2B is underlined with a purple bar.
 ll
Article OPEN ACCESS

A B E
FBXL17 + FBXO22 KO
FBXO22 KO

HCT-116 HCT-116
FBXL17 KO
FBXL17 Parental
FBXO22 wild type
wild type
representative
representative FBXL17
FBXO22 knock-out CHX (h)
knock-out

C HCT-116 D HCT-116
FBXL17 + 0 3 6 CHX (h) 0 3 6 CHX (h)
Parental FBXL17 KO FBXO22 KO FBXO22 KO 1 2 3 1 2 3 1 2 3 replicate 1 2 3 1 2 3 1 2 3 replicate
0 3 6 0 3 6 0 3 6 0 3 6 CHX (h)

FBXO22 KO
BACH1 BACH1
Parental

BACH1 FBXO22 FBXO22

cyclin D1 cyclin D1
FBXO22
actin actin
BACH1 BACH1

FBXO22 KO
FBXL17 KO

cyclin D1

FBXL17 +
FBXO22 FBXO22

vinculin cyclin D1 cyclin D1

actin actin

F HCT-116 G HCT-116
FBXL17 + FBXL17
Parental FBXL17 KO FBXO22 KO FBXO22 KO Parental FBXL17 KO FBXO22 KO + FBXO22 KO
CHX (h) sNONO
0 6 0 6 0 6 0 6 0 6 0 6 0 6 0 6 + CHX (h)
1 2 3 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3 replicate 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3 replicate
BACH1
* BACH1
FBXO22
*
FBXO22
cyclin D1 cyclin D1

GAPDH GAPDH

H I
1.2e+04
8765 43 2 1 NOR3-S1PC-treated
ENVDEVC107K BACH1-BTB
1.0e+04 2.0
1234567 8 10-fold diluted
Relative Intensity

UV x 10-2(AU)

8.0e+03 1.5
BACH1-BTB
6.0e+03 34 kDa (± 25%)
1.0

4.0e+03
0.5

2.0e+03
0.0
6 9 12 15 18
0.0e+00 Volume (ml)
200 400 600 m/z 800 1000 1200

J Ubiquitination: K Ubiquitination: L - - + GST-MAF/MBP-BACH1-bZIP

E1. E2, ATP, Ubiquitin E1. E2, ATP, Ubiquitin NOR3-S1PC
- + - GST-MAF + + + + GST-MAF/MBP-BACH1-bZIP
Nedd8~CUL1RBX1 Nedd8~CUL1RBX1 + + + DNA (HS2-700) + + + + DNA (HS2-700)
FBXO22-SKP1 FBXL17-SKP1
+ - NOR3/S1PC MAF/BACH1-DNA MAF/BACH1-DNA
WB: Anti-Flag

+ - + FBXL17-SKP1 MAF-DNA
(Ub)n~MSB-Flag-BACH1-BTB kDa - + + Ubiquitin MAF-DNA
245 (Ub)n~BACH1-FL
180 dsDNA dsDNA
135 Venus-BACH1-FL ssDNA ssDNA
MSB-Flag-BACH1-BTB 100 NOR3-S1PC treated

(legend on next page)

 ll
OPEN ACCESS Article

Figure S3. Nonredundant roles of FBXO22 and FBXL17 in regulating BACH1 stability, related to Figures 3 and 4
(A and B) Characterization of FBXO22 KO, FBXL17 KO, and FBXL17/FBXO22 double KO cell lines.
(C and D) Cycloheximide (CHX) chase assays monitoring the stability of BACH1 in parental, FBXL17 KO, FBXO22 KO, and FBXL17/FBXO22 double KO HCT-116
cells.
(E) Quantification of the data in (D) with each data point representing mean densities of BACH1 relative to untreated at 0 h. Data are presented as mean ± SD.
(F and G) BACH1 CHX chase assays performed with and without sNONO treatment. The data were used to create the graph in Figure 3D.
(H) An MS/MS fragmentation spectrum assigned to the S-nitrosylated peptide from BACH1-BTB containing Cys107 together with the assignments of the
fragmentation series to the sequence of the modified peptide. The B and Y ions represent the two halves of the peptide formed by splitting the original peptides
between different amino acids. The split sites are indicated by the red and blue ‘‘L’’-shaped marks lying above and below the peptide sequence. The B ions are the
products when the charge is retained on the N terminus, and the Y ions are the product when the charge is retained at the C terminus. The peaks of the products
are labeled accordingly within the spectrum. The measured m/z value of the y2 ion is consistent with the modification of Cys107 by 3-(N-maleimidopropionyl)-
biocytin with a monoisotopic mass of 523 Da.
(I) SEC-MALS analyses of BACH1-BTB treated with NOR3-S1PC with experimentally determined molecular weight. The theoretical molecular weight of BACH1-
BTB is 17 kDa for monomer or 34 kDa for dimer.
(J) Inhibition of in vitro SCFFBXO22-catalyzed BACH1 ubiquitination by NO treatment.
(K) In vitro ubiquitination of NO-treated BACH1 by SCFFBXL17.
(L) NO treatment abrogates the DNA-binding activity of BACH1 bZIP but not sMAF as detected by electrophoresis mobility shift assay (EMSA).
 ll
Article OPEN ACCESS

A B C
0.5 Probe: Biotin-BACH1(NOR-3/S1PC) 0.5 Probe: Biotin-BACH1-BTB 1.0
Analyte: FBXL17 (Untreated)
0.4 Apparent Kd = 57 nM 0.4 WT
Analyte: FBXL17 400nM 0.8
Response (nm)

Response (nm)

Response (nm)
C34A
0.3 [FBXL17] 0.3 WT 0.6
1700 nM Probe: Biotin-BACH1-BTB
0.2 567 nM C34A (NOR3/S1PC treated)
0.2 0.4
189 nM Analyte: FBXL17 400nM
0.1 63 nM 0.1 0.2
21 nM
0.0
0.0 0.0
0 200 400 600 800 1000 1200 0 100 200 300 400 0 100 200 300 400
Time (s) Time (s) Time (s)
D Unmodified

6 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100 105 110 115 120 125

SSVFAYESSVHSTNVLLSLNDQRKKDVLCDVTIFVEGQRFRAHRSVLAACSSYFHSRIVGQADGELNITLPEEVTVKGFEPLIQFAYTAKLILSKENVDEVCKCVEFLSVHNIEESCFQFLKF
3s
1m
30m
20h

< 10%
< 20%

E NO-treated <
<
<
30%
40%
50%
< 60%
< 70%
10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100 105 110 115 120 125 < 80%
6 < 90%
SSVFAYESSVHSTNVLLSLNDQRKKDVLCDVTIFVEGQRFRAHRSVLAACSSYFHSRIVGQADGELNITLPEEVTVKGFEPLIQFAYTAKLILSKENVDEVCKCVEFLSVHNIEESCFQFLKF > 90%

3s
1m
30m
20h

F 22
LSLNDQRKK30 45
FRAHRSVL52 59
FHSRIVGQADGELN 72
Unmod. NO-treated Unmod. NO-treated Unmod. NO-treated.
1×10 8 1×10 8 2×10 9 8×10 8 2×10 8 1×10 8
Cent: 551.6003 m/z Cent: 493.5491 m/z Cent: 493.5464 m/z Cent: 772.2515 m/z z2 Cent: 772.3086 m/z
8×10 7 8×10 7
1.5×10 9 6×10 8 1.5×10 8 7.5×10 7

6×10 7 6×10 7
Not detected 1×10 9 4×10 8 1×10 8 5×10 7 Undeut.
4×10 7 4×10 7

5×10 8 2×10 8 5×10 7 2.5×10 7

2×10 7 2×10 7

0 0 0 0 0 0
551 553 555 557 559 551 553 555 557 559 492 494 496 498 500 492 494 496 498 500 770 772 774 776 778 780 770 772 774 776 778 780
1.5×10 8 5×10 7 8×10 8 3×10 7 1×10 8 5×10 7
Cent: 551.7519 m/z Cent: 552.3394 m/z Cent: 493.6878 m/z Cent: 493.7668 m/z Cent: 775.1324 m/z Cent: 775.1757 m/z
F-test p = 1.1e-5 F-test p = 0.0873
4×10 7 Bimodal 1: 72% Bimodal 1: 85%
6×10 8 7.5×10 7 3.75×10 7
Bimodal 2: 28% Bimodal 2: 15%
1×10 8 2×10 7
3×10 7

2×10 7
4×10 8 5×10 7 2.5×10 7
3 sec
5×10 7 1×10 7
2×10 8 2.5×10 7 1.25×10 7
1×10 7

0 0 0 0 0 0
551 553 555 557 559 551 553 555 557 559 492 494 496 498 500 492 494 496 498 500 770 772 774 776 778 780 770 772 774 776 778 780
1.5×10 8 4×10 7 5×10 8 1.5×10 8 1×10 8 6×10 7
Cent: 551.873 m/z Cent: 552.4223 m/z Cent: 493.7357 m/z Cent: 493.8935 m/z Cent: 775.6057 m/z Cent: 775.6356 m/z
F-test p = 2.19e-6 F-test p = 0.127
Bimodal 1: 73% 4×10 8 Bimodal 1: 80%
3×10 7 7.5×10 7 4.5×10 7
Bimodal 2: 27% Bimodal 2: 20%
1×10 8 1×10 8
3×10 8
2×10 7 5×10 7 3×10 7 60 sec
2×10 8
5×10 7 5×10 7
1×10 7 2.5×10 7 1.5×10 7
1×10 8
Intensity

0 0 0 0 0 0
551 553 555 557 559 551 553 555 557 559 492 494 496 498 500 492 494 496 498 500 770 772 774 776 778 780 770 772 774 776 778 780
8×10 7 3×10 7 5×10 8 2×10 7 5×10 7 5×10 7
Cent: 552.339 m/z Cent: 552.8369 m/z Cent: 494.0153 m/z Cent: 494.2374 m/z Cent: 776.0401 m/z Cent: 776.1068 m/z
F-test p = 1.31e-5 F-test p = 0.0066
Bimodal 1: 69% 4×10 8 Bimodal 1: 83%
6×10 7 1.5×10 7 3.75×10 7 3.75×10 7
Bimodal 2: 31% Bimodal 2: 17%
2×10 7
3×10 8
4×10 7 1×10 7 2.5×10 7 2.5×10 7 30 min
2×10 8
1×10 7
2×10 7 5×10 6 1.25×10 7 1.25×10 7
1×10 8

0 0 0 0 0 0
551 553 555 557 559 551 553 555 557 559 492 494 496 498 500 492 494 496 498 500 770 772 774 776 778 780 770 772 774 776 778 780
6×10 7 2.5×10 7 2.5×10 8 5×10 7 4×10 7 6×10 7
Cent: 553.0741 m/z Cent: 555.5091 m/z Cent: 494.4945 m/z Cent: 494.7657 m/z Cent: 776.1102 m/z Cent: 776.0953 m/z
F-test p = 9.53e-5 F-test p = 5.38e-7 F-test p = 0.0703 F-test p = 2.66e-4
Bimodal 1: 65% 2×10 7 Bimodal 1: 30% 2×10 8 Bimodal 1: 74% 4×10 7 Bimodal 1: 45% 3×10 7 4.5×10 7
Bimodal 2: 35% Bimodal 2: 70% Bimodal 2: 26% Bimodal 2: 55%
4×10 7
1.5×10 7 1.5×10 8 3×10 7
2×10 7 3×10 7 20 hr
1×10 7 1×10 8 2×10 7
2×10 7
1×10 7 1.5×10 7
5×10 6 5×10 7 1×10 7

0 0 0 0 0 0
551 553 555 557 559 551 553 555 557 559 492 494 496 498 500 492 494 496 498 500 770 772 774 776 778 780 770 772 774 776 778 780
4×10 7 4×10 7 2.5×10 8 4×10 7 1.6×10 7 8×10 6
Cent: 554.1212 m/z Cent: 554.1219 m/z Cent: 495.3677 m/z Cent: 495.2787 m/z Cent: 776.2465 m/z Cent: 776.2316 m/z
2×10 8
3×10 7 3×10 7 3×10 7 1.2×10 7 6×10 6

1.5×10 8
2×10 7 2×10 7 2×10 7 8×10 6 4×10 6 Max deut.
1×10 8

1×10 7 1×10 7 1×10 7 4×10 6 2×10 6

5×10 7

0 0 0 0 0 0
551 553 555 557 559 551 553 555 557 559 492 494 496 498 500 492 494 496 498 500 770 772 774 776 778 780 770 772 774 776 778 780

m/z

(legend on next page)

 ll
OPEN ACCESS Article

Figure S4. Analysis of non-treated and NO-treated BACH1-BTB, related to Figures 4 and 5
(A) BLI measurements of the binding between FBXL17 and NOR3-S1PC-treated BACH1-BTB with a 10-min association step. Kd, dissociation constant.
(B and C) BLI measurements of the interaction between FBXL17 and BACH1-BTB (WT and C34A mutant) untreated or treated with NOR3-S1PC. In the absence of
compound treatment, C34A enhanced FBXL17 binding. This effect is exaggerated upon compound treatment. The C34 residue, therefore, is not required for
S-nitrosylation.
(D and E) Heat map summary of HDX data for unmodified (D) and NO-treated (E) BACH1-BTB. Colors under the primary sequence show deuterium uptake
from <10% (blue) to >90% (red) for each time point. Bars above the protein sequence show peptide coverage; green bars represent high-confidence peptides,
yellow bars represent medium-confidence peptides as assigned in HDExaminer v3. Data for the NO-treated sample reflect the population of peptides containing
unmodified cysteines.
(F) Bimodal and unimodal hydrogen-deuterium exchange profiles of additional BACH1-BTB peptides. As observed with peptides 79–89, both unmodified and
NO-modified BACH1 exhibit bimodal exchange profiles at peptides 22–30 and 45–52. With the NO modification, the less-protected population (green) appears
much earlier and is the predominant population by the longest (20 h) time point. Note for peptides 45–52: the 3 and 60 s NO-modified time points do not pass the
F-test but have been plotted as two populations due to the presence of statistically significant bimodal fits for the 30 min and 20 h time points. The peptide-based
maximally deuterated peptides 45–52 control for the unmodified condition eluted late and were not collected for preparation of the maximally deuterated
standard. Peptides 59–72 appear unimodal with either condition throughout the HDX time course. Importantly, the maximally deuterated controls do not show
any bimodal profiles, confirming that the bimodality is not a spectral artifact.
 ll
Article OPEN ACCESS

A B

D E H

I
F G

(legend on next page)

 ll
OPEN ACCESS Article

Figure S5. Analysis of NO-treated BACH1-BTB and C122E mutant, related to Figure 5
(A and B) Peptide mapping experiments to identify and quantify levels of S-nitrosylation. Extracted ion chromatograms are shown for the monoisotopic peak of
the 2+ charge state of peptide LSVHNIEESC and ILSKENVDEVCK as unmodified (blue), nitrosylated (+29, orange), and deoxidized (+32, purple). For
LSVHNIEESC, an additional species that is trioxidized (+48, cyan) was detected and is also shown. The traces are shown for both the treated protein and the
untreated sample. * The signal at the unmodified m/z corresponds to some in-source loss of the nitroso group. ** The third isotopic peak from the nitrosylated
(+29) peptide overlaps with the deoxidized (+32) m/z peak. *** Unrelated peptide that is present in both treated and untreated samples.
(C) The annotated MS/MS spectra identifying unmodified and nitrosylated peptides extracted from Byonic (Protein Metrics Inc). Positions of peptide frag-
mentation resulting in b/y ions are shown on each sequence in the inset.
(D) Thermostability of WT BACH1-BTB and the C122E mutant determined by differential scanning fluorimetry. Tm is melting temperature and is presented as
mean ± SD.
(E) Affinity pull-down assay showing the loss of FBXO22 binding in BACH1-BTB C122E mutant.
(F) The time-courses of the interactions between FBXL17 and WT BACH1-BTB vs. C122E mutant as assessed by affinity pull-down assay.
(G) Co-immunoprecipitation analysis of WT BACH1 vs. C122E mutant for FBXO22 and FBXL17 binding.
(H) CHX chase assay assessing the stability of FLAG-tagged WT BACH1 vs. C122E mutant.
(I) FLAG co-immunoprecipitation assay detecting physical association between two differentially tagged FBXL17. Three unrelated F-box proteins (FBXL1,
FBXL11, and FBXL15) were included as negative controls.
 ll
Article OPEN ACCESS

A B

C
Patch Motion
Correction Manual curate Blob picker
22,720 movies 22,720 micrographs 20,028 micrographs 7,148,295 particles
Two rounds
Ab-initio 2D classification
Topaz picking reconstruction
6,019,230 particles 660,991 particles
2D classification 143,997 particles
Ab-initio
reconstruction
718,765 particles
Topaz
183,814 picking
484,750 particles particles 156,298 particles
Heterogeneous 2D classification
refinement
213,113 particles 461 particles
5.7 Å Two rounds
7,944 10,173 2D classification
particles particles Ab-initio
reconstruction
38,168 particles
Heterogeneous
Non uniform Non uniform refinement
refinement refinement
with C2 symmetry with C2 symmetry

7.8 Å 7.8 Å

dSCFFBXL17-BTB-I (NO treated)

D E dSCFFBXL17-BTB-II (NO treated)
SKP1
BACH1- BACH1-
BTB BTB

BACH1-
BACH1-
FBXL17 BTB
FBXL17 BTB

CUL1 CUL1

RBX1 RBX1

Figure S6. Schematic workflow of single-particle reconstruction of NO-treated SCFFBXL17-BACH1 complex, related to Figure 6
(A) A representative cryo-EM micrograph for the sample containing a mixture of SCFFBXL17 and BACH1-BTB treated with NOR3-S1PC.
(B) Typical 2D averages.
(C) The flowchart of single particle analysis of the sample containing a mixture of SCFFBXL17 and BACH1-BTB treated with NOR3-S1PC.
(D and E) Fitting of the WT dSCFFBXL17-BACH1-I and dSCFFBXL17-BACH1-II structures in the 3D reconstructions obtained with the NO-treated BACH1-BTB sample.
 ll
OPEN ACCESS Article

A B

C Patch Motion CTFFIND4 Two rounds Template from ab-initio

Correction Manual curate Blob picker 2D classification n
reconstruction
12,618 12,618 12,423 3,087,124 497,102
Raw movies micrographs micrographs picked particles Kepted Heterogenous
CryoSPARC us
particles refinement 267,523 particles
Template picker
Two rounds
Template picker 2D classification 2D classification
4,169,727 particles Topaz picking
Ab-initio reconstruction 7,309,475 particles 84,154 particles
70,723 selected particles
Two rounds
2,750,659 particles 2D classification
2D classification 1,099,238 particles
Two rounds
Heterogenous Ab-initio
91,639 selected particles
Topaz 3D templates refinement reconstruction
picking 2,969,728 244,189
particles particles Heterogenous
2D classification refinement
30,213 particles
with C1 symmetry
Non-uniform 364,786 particles
refinement
Expand Non-uni
Non-uniform refinement
symmetry C2 symmetry
19.4% 22.3% 3D variability
94,672 47,336 particles 54,546 particles
4.4 Å
expnad particles C2 symmetry Non-uniform
Overall 3.1 Å
O
refinement
Particle substraction
& Local refinement 193,908 paticles

4.5 Å Local
refinement 3.2 Å
4.1 Å
D E G

Figure S7. Cryo-EM single-particle analysis workflow of native SCFFBXL17-BACH1, related to Figure 6
(A) A representative cryo-EM micrograph.
(B) Typical 2D averages revealing the most populated monomeric complex.
(C) The flowchart of single particle analysis of the sample containing a mixture of SCFFBXL17 and BACH1-BTB.
(D) Local resolution map of the BACH1-BTB dimer flanked by the LLR domains of two FBXL17 ranging from 3 to 8 Å.
(E) The particle angular distribution and FSC curves of dSCFFBXL17-BACH1-I with local refinement.
(F) The particle angular distribution and FSC curves of dSCFFBXL17-BACH1-II.
(G) The angular distribution and FSC curves of the monomeric SCFFBXL17-BACH1 complex.