Regulatory mechanisms of NRF2-mediated transcription
NRF2 is a potent transcriptional activator belonging to the CNC
transcription factor group (a family of bZip-type transcription factors
homologous to the Drosophila transcription factor Cap’n’collar)
and has been originally identified as a related member of NFE2 p45 (Itoh
et al., 1995). NRF2 heterodimerizes with small MAF proteins and binds to
a consensus sequence, antioxidant response element (ARE)
(GCnnnG/CTCAC/T)
(Motohashi et al., 2000; Motohashi et al., 2004; Katsuoka et al., 2005;
Yamamoto et al., 2006). Six functional domains, NRF2-ECH homology (Neh)
1 to 6, have been identified in NRF2 (Figure 2). The Neh1 domain
contains a basic region required for DNA binding and a bZip structure
required for heterodimer formation with small MAF proteins (Kyo et al.,
2004; Kimura et al., 2007; Kurokawa et al., 2009; Sengoku et al., 2022).
The Neh2 domain mediates NRF2 binding to KEAP1, an inhibitory regulator
of NRF2, via two motifs, DLG and ETGE motifs. The two-site binding of
NRF2 to KEAP1 homodimer enables KEAP1-CUL3 ubiquitin E3 ligase complex
to ubiquitinate NRF2 for degradation in the proteasome (Figure 3)
(Kobayashi et al., 2004; Zhang et al., 2004; Tong et al., 2006). The DLG
motif utilizes only hydrogen bonding and forms a relatively weak binding
that exhibits rapid binding and dissociation. The ETGE motif, on the
other hand, relies on both hydrogen bonding and hydrophobic interactions
to bind to KEAP1 and is thought to exhibit two-step binding and
dissociation, resulting in strong binding (Fukutomi et al., 2014; Horie
et al., 2021). The Neh3 domain was shown to interact with CHD6 for
enhancing NRF2-mediated transcriptional activation (Nioi et al., 2005).
The Neh4 and Neh5 domains are known as transcriptional activation
domains, and the binding of transcriptional coactivators, such as CREB
binding protein (CBP) / p300 and chromatin remodeling factor BRG1,
promotes transcriptional activation by NRF2 (Katoh, et al., 2001; Zhang
et al., 2007). MED16, a subunit of Mediator complex, is another binding
partner of NRF2 to the Neh4 and Neh5 domains, conferring transcriptional
activation ability by recruiting the Mediator complex and subsequently
RNA polymerase II (Sekine et al., 2015). When glucocorticoid receptor
binds to the Neh4 and Neh5 domains, NRF2 activity is suppressed (Alam et
al., 2017). The Neh6 domain contains serine residues that can be
phosphorylated by GSK3β. NRF2 undergoes proteasome-mediated degradation
upon the phosphorylation of the Neh6 domain, indicating that the Neh6
domain mediates KEAP1-independent degradation of NRF2 (Rada et al.,
2012; Chowdhry et al., 2013).
Cooperativity with other transcription factors provides another layer of
regulation for NRF2 transcriptional activity (Figure 4). ATF4 and NRF2
cooperatively enhance expression of xCT, a cystine transporter (Ye et
al., 2014), and enzymes regulating de novo synthesis serine
(DeNicola et al., 2015). Cooperativity of CEBPB and NRF2, which is
uniquely observed in cancer cells with persistent activation of NRF2,
promotes the enhancer activity of canonical NRF2 target genes and also
generates novel enhancers at the loci that are not normally regulated by
transiently-activated NRF2 (Okazaki et al., 2020; Okazaki et al., 2022).
The NRF2-CEBPB cooperativity is likely to underly the emergence of a
novel enhancer in NOTCH3 locus for promoting cancer stemness, and
at the same time, it also activates canonical NRF2-dependent enhancers,
such as in AKR1C1 -AKR1C2 locus, leading to the increased
chemo-resistance of cancer cells (Figure 4). Compared to the persistent
activation of NRF2, the transient activation only temporarily induces
CEBPB expression, and results in a very short duration of the
coexistence of the two factors, which hardly open the enhancers that the
NRF2-CEBPB cooperativity does. In contract, the persistent activation of
NRF2 leads to continuous expression of CEBPB, leading to constitutive
coexistence of the two factors to create a new mode of transcriptional
regulation.
KEAP1 is a substrate recognition subunit of CUL3-based ubiquitin E3
ligase and mediates NRF2 ubiquitination, serving as a negative regulator
of NRF2 (Kobayashi et al., 2004; Kobayashi et al., 2006). The most
unique feature of KEAP1 is to possess highly reactive cysteine residues,
which enables KEAP1 to serve as a biosensor for electrophiles. KEAP1
consists of three functional domains: a broad complex
tramtrack-bric-à-brac (BTB) domain, an intervening region (IVR), a
double glycine repeat and a COOH-terminal region (DC) domain (Chauhan et
al., 2013) (Figure 2). KEAP1 forms homodimers via the BTB domain and
further forms a complex with Cullin3 (CUL3) and RING-box protein 1
(RBX1) to function as a ubiquitin E3 ligase (Figure 3). The DLG and ETGE
motifs of NRF2 interact with DC domains of the KEAP1 homodimer, which
allows ubiquitination of NRF2, resulting in proteasome-dependent
degradation of NRF2. Intriguingly, the homodimer formation of KEAP1 is
essential for the ubiquitin E3 ligase activity of KEAP1-CUL3 complex
(Suzuki et al., 2011). The enzymatic activity of KEAP1-CUL3 ubiquitin E3
ligase is inhibited when reactive cysteine residues of KEAP1, such as
Cys151, Cys273 and Cys288 in murine KEAP1, are directly modified by
electrophiles (Figure 2) (Yamamoto et al., 2008; Saito et al., 2015;
Suzuki et al., 2019). Different electrophiles target different cysteine
residues of KEAP1, which is regarded as a multimodal sensing system, and
the electrophilic signals converge on NRF2 to activate genes for
antioxidant response and cytoprotection. Thus, under normal conditions,
NRF2 is degraded and functionally repressed by KEAP1, and when
electrophiles attack KEAP1, NRF2 is de-repressed and activates
transcription. KEAP1 is an electrophilic biosensor regulating NRF2
pathway activity for the stress response.
KEAP1-independent regulation of NRF2 activity has been also described.
CUL1-βTrCP ubiquitin E3 ligase ubiquitinates NRF2 when NRF2 is
phosphorylated at the Neh6 domain (Figure 3) (Rada et al., 2012;
Chowdhry et al., 2013). GSK3 is responsible for the NRF2 phosphorylation
at Ser344 and Ser347 in murine NRF2 (Figure 2), which is suppressed
under the active proliferation signals mediated by AKT (Taguchi et al.,
2014; Shirasaki et al., 2014).