Celastrol binds to its target protein via specific noncovalent interactions and reversible covalent bonds

Article abstract of DOI:10.1039/c8cc06140h

Celastrol is one of the most studied natural products. Our studies show for the first time that celastrol can bind to its target protein via specific noncovalent interactions that position celastrol next to the thiol group of the reactive cysteine for reversible covalent bond formation. Such specific noncovalent interactions confer celastrol binding specificity and demonstrate the feasibility of improving the efficacy and selectivity of celastrol for therapeutic applications.

Full text of DOI:10.1039/c8cc06140h

ChemComm
View Article Online
View Journal | View Issue
COMMUNICATION
Celastrol binds to its target protein via specific
noncovalent interactions and reversible covalent
bonds†‡
Cite this: Chem. Commun., 2018,
4, 12871
5
Received 28th July 2018,
Accepted 18th October 2018
a
a
a
a
a
Duo Zhang,§ Ziwen Chen, § Chaochao Hu, Siwei Yan, Zhuoer Li,
a
a
a
a
ab
Baohuan Lian, Yang Xu, Rong Ding, Zhiping Zeng, Xiao-kun Zhang* and
DOI: 10.1039/c8cc06140h
ab
Ying Su
*
rsc.li/chemcomm
Celastrol is one of the most studied natural products. Our studies Nur77 with a Kd of 0.29 mM and showed that Nur77 was
show for the first time that celastrol can bind to its target protein required for the anti-inflammatory effects of celastrol in vitro
via specific noncovalent interactions that position celastrol next to and in inflammatory animal models. Celastrol binding to the
the thiol group of the reactive cysteine for reversible covalent bond ligand-binding domain (LBD) of Nur77 induces Nur77 mito-
formation. Such specific noncovalent interactions confer celastrol chondrial targeting, resulting not only in the suppression of
binding specificity and demonstrate the feasibility of improving the inflammation but also the induction of mitochondrial auto-
1
4
efficacy and selectivity of celastrol for therapeutic applications.
phagy. Because celastrol contains an electrophilic quinone
methide moiety (Fig. 1A, red), a known Michael acceptor, it is
Tripterygium wilfordii plant, the ‘‘Thunder God Vine’’ (lei gong widely speculated and assumed that celastrol acts through a
teng), has been used traditionally in Chinese medicine for the covalent binding mechanism in which an adduct is formed
1
5–17
treatment of rheumatoid arthritis, lupus and other autoimmune between celastrol and specific cysteines on the target proteins.
1
diseases. Celastrol, viewed as one of the most promising bio- However, few studies have conclusively demonstrated so.
active molecules of the ‘‘Thunder God Vine’’, has drawn much Furthermore, detailed characterization of how celastrol could
attention. An increasing number of studies have revealed the bind to the target protein has not been reported. Given the
potential of celastrol in treating diverse diseases and cancers due importance of celastrol and Nur77, we set out to study the
2–4
to its potent anti-inflammatory effect. Multiple pathways have binding mechanism of celastrol action.
been proposed to be regulated by celastrol, however, the direct
cellular targets for celastrol remain elusive.
We first used LC-MS to investigate if celastrol bound
covalently to Nur77-LBD. Fig. 1B shows that in the absence
Nur77 (also called TR3, NGFIB, or NR4A1), a unique member of celastrol the Nur77-LBD peak corresponded to a 29 672 Da
of the nuclear receptor superfamily, acts as an immediate early molecule, but in the presence of celastrol, the Nur77-LBD/
5
,6
or stress response gene. Nur77 has emerged as a critical celastrol solution displayed a second peak at 30 122 Da, corres-
regulator for the onset and progression of cancer and metabolic ponding to the Nur77-LBD/celastrol complex with an increase in
5
,7–12
and inflammatory diseases.
The potent anti-inflammatory MW of 450 Da compared to Nur77-LBD. This observed increase
role of Nur77 in inflammatory diseases and cancer has received in MW is equal to the expected MW of celastrol (Fig. 1B and
particular attention. An aberrant expression of Nur77 is observed Fig. S1A, ESI‡). These results confirm that celastrol covalently
in inflamed human synovial tissues, cancer cells, psoriasis, binds to Nur77. We then asked if the covalent bond formation
1
3
atherosclerotic lesions and multiple sclerosis. Thus, Nur77 was a result of Michael adduction between cysteines of Nur77-
represents a promising target for anti-inflammatory therapy. LBD and celastrol. There are 6 cysteines in Nur77-LBD: C465,
We recently identified Nur77 as a direct cellular protein target C475, C505, C534, C551 and C566. The crystal structures of
of celastrol. We found that celastrol was a potent binder of Nur77-LBD show that C475, C505 and C534 are buried (Fig. S2,
ESI‡), and C465, C551 and C566 are solvent exposed and acces-
a
sible for modification with celastrol. However, not all spatially
accessible cysteines have the potential to act as nucleophilic
thiols. To identify the cysteine(s) responsible for adduct forma-
tion, 3 individual mutations, C465A, C551A and C566A, were
generated for LC-MS analyses. The results showed that mutating
C551 to alanine abolished celastrol covalent modification
(Fig. 1C), and C465A and C566A had the same LC-MS spectra
School of Pharmaceutical Sciences, Fujian Provincial Key Laboratory of Innovative
Drug Target Research, Xiamen University, Xiamen 361102, P. R. China
Cancer Center, Sanford Burnham Prebys Medical Discovery Institute, La Jolla,
CA 92037, USA. E-mail: xzhang@sbpdiscovery.org, ysu@sbpdiscovery.org
Proteomics data were deposited with PeptideAtlas under the deposition number
b
†
PASS01273.
‡
§
Electronic supplementary information (ESI) available. See DOI: 10.1039/c8cc06140h
These authors contributed equally to this work.
This journal is ©The Royal Society of Chemistry 2018
Chem. Commun., 2018, 54, 12871--12874 | 12871

View Article Online
Communication
ChemComm
Fig. 1 C551 is the cysteine responsible for the adduct formation between Nur77-LBD and celastrol. (A) Structure of celastrol. (B) The deconvoluted mass
spectra before and after celastrol binding. (C) The deconvoluted mass spectra of mutant C551A in the absence and presence of celastrol. (D) The
deconvoluted mass spectra of mutants C465A and C566A, respectively, in the absence and presence of celastrol. (E) Structure of XS0419. (F) The
deconvoluted mass spectra of Nur77 in the absence and presence of XS0419. Supporting LC/MS spectra are presented in the ESI.‡
distribution as the wild type protein in the presence of celastrol demonstrated reversibility (Fig. S4D–F and S5, ESI‡). Satisfied that
(
Fig. 1D). Mutating the cysteine to alanine did not alter the this methodology demonstrates the covalent reversibility for
overall structure of Nur77-LBD as shown by circular dichroism celastrol-cysteine modification, we applied the same method to
CD) spectra and confirmed by the mutants’ ability to dimerize the Nur77/celastrol system. Fig. 2A shows that the absorption
(
with RXR as the wild-type Nur77 (Fig. S3, ESI‡). Taken together, peak of celastrol disappeared when it was incubated with Nur77-
these results identify C551 as the residue covalently interacting LBD and its absorption peak reappeared when the Nur77-LBD/
with celastrol.
celastrol mixture was denatured with SDS (Fig. 2A). These data
Celastrol’s susceptibility to reduction has been studied and it suggest that the covalent bond between Nur77-LBD and celastrol
has been demonstrated that among the 3 electrophilic carbons in is reversible. Furthermore, analysis of the Nur77-LBD/celastrol
the quinone methide motif of celastrol, C-6 is the most reactive
18
for nucleophilic addition with a remarkable stereospecificity.
To confirm whether the nucleophilic addition occurs at C-6 of
celastrol, we synthesized XS0419, where the C-6 carbon is reduced
and is not capable of nucleophilic addition (Fig. 1E). Fig. 1F
shows that the molecular weight for Nur77 did not change in the
presence of XS0419 (Fig. 1F and Fig. S1B, ESI‡), indicating the
involvement of C-6 in adduct formation.
We then asked whether celastrol was an irreversible or
reversible electrophile. To answer this question, we used a
UV-visible spectroscopy method originally developed for the
covalent reversibility studies of cyanoacrylamides as Michael
1
9
acceptors. We first investigated the nature of the covalent
bond resulting from the adduct formation between celastrol
and glutathione (GSH), a weak nucleophilic thiol. The visible
absorption spectrum (with a maximum visible spectrum at
450 nm) was used to monitor the reactivity of celastrol. When
celastrol solution was treated with GSH, the celastrol absorp-
tion band was reduced, indicating the occurrence of a reaction
between celastrol and GSH (Fig. S4A, ESI‡). When the reaction
Fig. 2 The covalent and noncovalent interactions between celastrol and
Nur77. (A) UV-visible spectra showing that the covalent addition of celastrol
to Nur77 is reversible. (B) Model of celastrol binding to Nur77: the protein is
mixture was diluted, we observed an increase in the absorption displayed as a surface representation and celastrol is shown as wheat sticks.
The contributing hydrophobic residues are colored in pale green. The polar
spectra at 450 nm, consistent with a reversible reaction (Fig. S4B,
ESI‡). Fitting titration data generated by treating a celastrol
solution with increasing concentration of GSH yielded an apparent
residues Q547 and D499 are colored in blue and in orange, respectively.
Docked celastrol displays a good shape complementary with Nur77 and
places the C-6 atom of celastrol close to C551. (C and D) UV-visible spectra
K
d
of 86 mM (Fig. S4C, ESI‡). Consistently, the adduct reaction
of Nur77-LBD/D499L and Nur77-LBD/Q547W mutants in the presence and
of celastrol with the reducing agent beta-mercaptoethanol also absence of celastrol.
1
2872 | Chem. Commun., 2018, 54, 12871--12874
This journal is ©The Royal Society of Chemistry 2018

View Article Online
ChemComm
Communication
complex trypsin digestion using Orbitrap mass spectrometry supporting our celastrol binding model. The observation that
showed that the derived peptide 537EHVAAVAGEPQPASCLSR554 unfolding Nur77 releases bound celastrol is also consistent
was the only peptide labelled by celastrol. This peptide contains with our binding model because the stability of the Nur77/
C551, therefore providing independent evidence that covalent celastrol complex requires the intact folded structure of Nur77
modification occurs at this cysteine (Fig. S6A–C, ESI‡).
Reversible covalent ligands require specific noncovalent close to nucleophile C551.
interactions with the target protein at a region near the targeted To explore the biological relevance of the identified binding
and noncovalent interactions are required to bring celastrol
2
0
nucleophilic residue. The fact that celastrol acts as a rever- mechanism of celastrol, we studied the roles of C551, D499 and
sible covalent ligand to Nur77 prompted us to further explore if Q547 in the Nur77-mediated activities of celastrol. Celastrol acts
celastrol binds to Nur77 in close proximity to C551 by using as an anti-inflammation agent by binding to Nur77 to prime
molecular modeling and the mutagenesis approach. The crystal inflamed mitochondria for autophagy through inducing Nur77
structure of Nur77-LBD (PDB code: 4KZI) reveals that C551 is interactions with tumor necrosis factor receptor-associated
1
4
located at the end of a disordered loop, 541AVAGEPQPAS550, and factor 2 (TRAF2) and the autophagic adaptor p62/SQSTM1.
at the start of helix 10. Given our results, the C6 atom of Co-IP assay showed that Nur77/C551A, but not Nur77/C566A,
celastrol should be in close proximity to C551 and the inter- markedly reduced its interaction with TRAF2 (Fig. 3A) and
action between celastrol and Nur77 would require the partici- p62/SQSTM1 (Fig. 3B). This impact of mutation on celastrol’s
pation of this disordered loop. The predicted binding mode for function demonstrates the importance of the covalent bond in
celastrol to Nur77-LBD was developed in two steps: first building celastrol’s potency and is consistent with the observed change
2
1–23
the disordered loop using Schr ¨o dinger’s Prime,
and then in the binding affinity between celastrol and Nur77/C551A, and
docking celastrol to a region partly composed of the constructed between celastrol and Nur77/C566A (Fig. S7B and C, ESI‡). The
2
4–26
loop using Glide.
The binding mode (Fig. 2B) from the role of the covalent bond in celastrol’s potency is also reflected
docking studies showed that celastrol fits well into this region in the biological function of XS0419 and the binding affinity of
with the C6 atom positioned next to the –SH group of C551: the XS0419 to Nur77 (Fig. S8A–D, ESI‡). Substitution of D499 with
polar groups, hydroxyl and carboxyl at the opposite ends of the Leu (Nur77/D499L) or Q547 with Trp (Nur77/Q547W) also
celastrol molecule forming H-bonds with D499 and Q547, reduced the ability of Nur77 to interact with p62/SQSTM1,
respectively, and the central hydrophobic portion of the celastrol while their simultaneous mutation (Nur77/D499L/Q547W) com-
molecule making extensive van der Waals interactions with the pletely abolished the interaction (Fig. 3C), which again is con-
protein involving residues L496, L497, V498, A502, V539, A540, sistent with the change in the binding affinity of celastrol to
V542, A543, A549 and L552 (Fig. 2B). To validate this model, the wild type and mutant Nur77. Therefore, these mutagenesis
Nur77-LBD/D499L and Nur77-LBD/Q547W mutants were made, studies confirm that celastrol binds to C551 in the Nur77
respectively, and tested for their binding to celastrol.
protein and the importance of the noncovalent interactions
The involvement of D499 and Q547 in the noncovalent between celastrol and the protein.
binding was first confirmed by binding affinity studies. In the
non-covalent binding
Ð
covalent binding
Ð
absence of the covalent bond, celastrol bound to Nur77/C551A-
LBD with a Kd of 1.17 mM, reflecting a noncovalent binding
between Nur77 and celastrol (Fig. S7B, ESI‡). Mutants D449L
and Q547W, respectively, displayed weaker binding to celastrol
E þ I
EꢀI
E ꢁ I
(1)
(2)
K1
K2
½
EꢀIꢂ þ ½E ꢁ Iꢂ 1 þ K2
Kd ¼
¼
½
Eꢂ½Iꢂ
K1K2
(Fig. S7D and E, ESI‡), indicating the contribution of D499 and
Q547 to the binding of celastrol. UV-visible absorption studies
showed that the absorption peak of celastrol was consistent inhibitor consists of two steps as shown in eqn (1).
A general mechanism for the binding of a reversible covalent
2
7
1 2
K and K ,
with the unbound compound spectrum when incubated with respectively, are the equilibrium dissociation constants of
either mutant Nur77-LBD/D499L (Fig. 2C) or Nur77-LBD/Q547W step 1 and step 2 and the overall K for the reversible covalent
Fig. 2D). These results indicate that both D499 and Q547 inhibitor can be defined by K and K using eqn (2). In the
d
2
7
(
1
2
residues contribute to the noncovalent binding of celastrol, case of celastrol binding to Nur77, K and K were determined
d
1
Fig. 3 Biological studies support the identified binding nature of celastrol. (A and B) Co-immunoprecipitation assays revealed the importance of C551 in
mediating the effect of celastrol on inducing Nur77 interactions with TRAF2 and p62/SQSTM1. (C) Co-immunoprecipitation assay revealed that both
D499 and Q547 played key roles in interacting with celastrol.
This journal is ©The Royal Society of Chemistry 2018
Chem. Commun., 2018, 54, 12871--12874 | 12873

View Article Online
Communication
ChemComm
to be 0.29 mM and 1.17 mM (Fig. S7B, ESI‡) respectively, therefore, 10 A. A. Hamers, R. N. Hanna, H. Nowyhed, C. C. Hedrick and C. J.
de Vries, Curr. Opin. Lipidol., 2013, 24, 381–385.
2
K is equal to 0.23 mM according to eqn (2). This derived result
showed that the reversible covalent bond formation increased the
binding affinity and rendered celastrol more potent.
In conclusion, using diverse approaches, we have demon-
strated and confirmed that celastrol binds to the target protein
Nur77 via a Michael addition reaction at C551 in Nur77.
Importantly, the formation of this covalent bond is reversible
and requires specific noncovalent interactions with Nur77 to
1
1
1 J. A. Beard, A. Tenga and T. Chen, Cell. Signalling, 2015, 27, 257–266.
2 P. Pinton and G. Kroemer, Nat. Chem. Biol., 2014, 10, 89–90.
13 H. M. Mohan, C. M. Aherne, A. C. Rogers, A. W. Baird, D. C. Winter
and E. P. Murphy, Clin. Cancer Res., 2012, 18, 3223–3228.
1
4 M. Hu, Q. Luo, G. Alitongbieke, S. Chong, C. Xu, L. Xie, X. Chen,
D. Zhang, Y. Zhou, Z. Wang, X. Ye, L. Cai, F. Zhang, H. Chen,
F. Jiang, H. Fang, S. Yang, J. Liu, M. T. Diaz-Meco, Y. Su, H. Zhou,
J. Moscat, X. Lin and X. K. Zhang, Mol. Cell, 2017, 66, 141–153.
5 S. Sreeramulu, S. L. Gande, M. Gobel and H. Schwalbe, Angew.
Chem., Int. Ed., 2009, 48, 5853–5855.
1
position celastrol in close proximity to the –SH group of C551. 16 A. Salminen, M. Lehtonen, T. Paimela and K. Kaarniranta, Biochem.
Biophys. Res. Commun., 2010, 394, 439–442.
7 R. Kannaiyan, M. K. Shanmugam and G. Sethi, Cancer Lett., 2011,
The results elucidate an unprecedented mechanism on how
celastrol can selectively bind to an intracellular target to exert
1
303, 9–20.
specificity, offering new strategies to improve the efficacy and 18 L. Klai ´c , P. C. Trippier, R. K. Mishra, R. I. Morimoto and R. B.
Silverman, J. Am. Chem. Soc., 2011, 133, 19634–19637.
selectivity of celastrol for therapeutic applications via modifying
its electrophilicity and/or noncovalent interaction properties.
1
9 I. M. Serafimova, M. A. Pufall, S. Krishnan, K. Duda, M. S. Cohen,
R. L. Maglathlin, J. M. McFarland, R. M. Miller, M. Frodin and
J. Taunton, Nat. Chem. Biol., 2012, 8, 471–476.
2
0 R. Lagoutte, R. Patouret and N. Winssinger, Curr. Opin. Chem. Biol.,
2017, 39, 54–63.
Conflicts of interest
2
1 M. M. Mysinger, D. R. Weiss, J. J. Ziarek, S. Gravel, A. K. Doak,
J. Karpiak, N. Heveker, B. K. Shoichet and B. F. Volkman, Proc. Natl.
Acad. Sci. U. S. A., 2012, 109, 5517–5522.
There are no conflicts to declare.
2
2 M. P. Jacobson, D. L. Pincus, C. S. Rapp, T. J. F. Day, T. B. Honig,
D. E. Shaw and R. A. Friesner, Proteins, 2004, 55, 351–367.
3 M. P. Jacobson, R. A. Friesner, Z. Xiang and B. Honig, J. Mol. Biol.,
2002, 320, 597–608.
Notes and references
2
1
2
T. W. Corson and C. M. Crews, Cell, 2007, 130, 769–774.
X. Ma, L. Xu, A. T. Alberobello, O. Gavrilova, A. Bagattin, M. Skarulis, 24 R. A. Friesner, R. B. Murphy, M. P. Repasky, L. L. Frye, J. R. Greenwood,
J. Liu, T. Finkel and E. Mueller, Cell Metab., 2015, 22, 695–708.
V. Narayan, K. C. Ravindra, C. Chiaro, D. Cary, B. B. Aggarwal, A. J.
Henderson and K. S. Prabhu, J. Mol. Biol., 2011, 410, 972–983.
H. J. Yang, D. Chen, Q. Z. C. Cui, X. Yuan and Q. P. Dou, Cancer Res.,
T. A. Halgren, P. C. Sanschagrin and D. T. Mainz, J. Med. Chem., 2006,
49, 6177–6196.
25 R. A. Friesner, J. L. Banks, R. B. Murphy, T. A. Halgren, J. J. Klicic,
D. T. Mainz, M. P. Repasky, E. H. Knoll, M. Shelley, J. K. Perry,
D. E. Shaw, P. Francis and P. S. Shenkin, J. Med. Chem., 2004, 47,
1739–1749.
3
4
2
006, 66, 4758–4765.
5
6
7
8
J. P. McMorrow and E. P. Murphy, Biochem. Soc., 2011, 39, 688–693.
M. R. Paillasse and P. de Medina, Med. Hypotheses, 2015, 84, 135–140. 26 T. A. Halgren, R. B. Murphy, R. A. Friesner, H. S. Beard, L. L.
P. C. Evans, Circ. Res., 2009, 104, 707–709.
S. O. Lee, X. Li, S. Khan and S. Safe, Expert Opin. Ther. Targets, 2011,
Frye, W. T. Pollard and J. L. Banks, J. Med. Chem., 2004, 47,
1750–1759.
1
5, 195–206.
U. M. Moll, N. Marchenko and X.-K. Zhang, Oncogene, 2006, 25,
725–4743.
27 P. Chatterjee, W. M. Botello-Smith, H. Zhang, L. Qian, A. Alsamarah,
D. Kent, J. J. Lacroix, M. Baudry and Y. Luo, J. Am. Chem. Soc., 2017,
139, 17945–17952.
9
4
1
2874 | Chem. Commun., 2018, 54, 12871--12874
This journal is ©The Royal Society of Chemistry 2018