Relative impact of 3- and 5-hydroxyl groups of cytosporone B on cancer cell viability

Article abstract of DOI:10.1039/c2md20243c

A novel and the shortest route, thus far, for preparing cytosporone B (Csn-B) is reported. Csn-B and two analogs were used to probe the importance of hydroxyl groups at the 3- and 5-positions of the Csn-B benzene ring in inhibiting the viability of human H460 lung cancer and LNCaP prostate cancer cells, inducing H460 cell apoptosis, and interacting with the NR4A1 (TR3) ligand-binding domain (LBD). These studies indicate that Csn-B and 5-Me-Csn-B, having a phenolic hydroxyl at the 3-position of their aromatic rings, had similar activities in inhibiting cancer cell viability and in inducing apoptosis, whereas 3,5-(Me)₂-Csn-B was unable to do so. These results are in agreement with ligand-binding experiments showing that the interaction with the NR4A1 LBD required the presence of the 3-hydroxyl group. The Royal Society of Chemistry.

Full text of DOI:10.1039/c2md20243c

MedChemComm
View Article Online
View Journal | View Issue
CONCISE ARTICLE
Relative impact of 3- and 5-hydroxyl groups of
cytosporone B on cancer cell viability†
Cite this: Med. Chem. Commun., 2013,
4, 332
Zebin Xia,‡ Xihua Cao,‡ Elizabeth Rico-Bautista, Jinghua Yu, Liqun Chen, Jiebo Chen,
Andrey Bobkov, Dieter A. Wolf, Xiao-Kun Zhang and Marcia I. Dawson
*
A novel and the shortest route, thus far, for preparing cytosporone B (Csn-B) is reported. Csn-B and two
analogs were used to probe the importance of hydroxyl groups at the 3- and 5-positions of the Csn-B
benzene ring in inhibiting the viability of human H460 lung cancer and LNCaP prostate cancer cells,
inducing H460 cell apoptosis, and interacting with the NR4A1 (TR3) ligand-binding domain (LBD). These
studies indicate that Csn-B and 5-Me-Csn-B, having a phenolic hydroxyl at the 3-position of their
aromatic rings, had similar activities in inhibiting cancer cell viability and in inducing apoptosis, whereas
3,5-(Me)2-Csn-B was unable to do so. These results are in agreement with ligand-binding experiments
showing that the interaction with the NR4A1 LBD required the presence of the 3-hydroxyl group.
Received 17th August 2012
Accepted 7th November 2012
DOI: 10.1039/c2md20243c
www.rsc.org/medchemcomm
receptor co-activator (SRC) 1 and SRC-2 proteins,8 which is
further evidence of NR4A1 transcriptional agonism. In contrast,
Introduction
The orphan nuclear receptor (NR) 4A1, also called testicular
receptor 3 (TR3) in humans, Nur77 in mice and nerve growth
factor-inducible protein B (NGFI-B) in rats, is an immediate-
early gene product. Its expression is induced in response to cell
stress, serum withdrawal or treatment with mitogenic factors or
apoptotic agents.1,2 In the nucleus, NR4A1 protein functions as
a transcription factor by regulating the expression of genes
having roles in cell survival, proliferation and apoptosis. NR4A1
is known to act constitutively (without a bound ligand) in a
variety of cell types.3 However, in response to apoptotic stimuli,
nuclear NR4A1 can translocate from the nucleus to the cyto-
plasm, where it can reverse the activity of the normally cyto-
protective mitochondrial protein Bcl-2 to facilitate
mitochondrial apoptosis (programmed cell death).4–6 Recently,
several functional NR4A1 ligands have been reported.7–9 The
identication of such ligands enables the exogenous modula-
tion of NR4A1 activity and thus could lead to agents for treating
such diseases as cancer. One such ligand8 is the fungal octa-
Csn-B was unable to activate the human heterodimer complex
of NR4A1 and another nuclear receptor, retinoid X receptor
(RXR), on the bRARE (a DR-5 response element) in a reporter
construct8 that is known to be activated by the RXR transcrip-
tional agonist-bound NR4A1–RXR complex.12,13 Csn-B at 5 or
10 mM was found to interact with the NR4A1 monomer bound to
the NBRE found in the NR4A1 promoter14 leading to enhance-
ment of NR4A1 mRNA expression. Moreover, Csn-B was also
found to inhibit expression of the anti-apoptotic brain and
reproductive organ-expressed (BRE) gene by inducing recruit-
ment of the corepressor protein N-CoR to NR4A1 when bound to
the NBRE in the BRE promoter.9 Treatment of BCG-823 gastric
cancer cells with Csn-B at concentrations $15.5 mM for 12 h led
to the translocation of NR4A1 from nucleus to mitochondria
followed by apoptosis (64% aer 48 h).8 Thus, the function of
NR4A1 appears to be dependent on the cellular context.
Here, we report the relative impact of the hydroxyl groups at
the 3- and 5-positions of the Csn-B benzene ring on the ability of
ketide-metabolite cytosporone
B (Csn-B, compound 1 in
Fig. 1).10 Wu and colleagues8 found that Csn-B bound to the
NR4A1 homodimer in the nucleus to induce the activation of a
Nur response element (NuRE)-driven reporter.11 At a concen-
tration as low as 0.1 mM Csn-B was able to induce the maximal
luciferase reporter response (12.5 fold) and to recruit the steroid
Cancer Center, Sanford-Burnham Medical Research Institute, 10901 North Torrey
Pines Rd., La Jolla, CA 92037, USA. E-mail: mdawson@sanfordburnham.org
† Electronic supplementary information (ESI) available: Synthetic methods for
preparing compounds 1–3 and their intermediates by the routes shown in Fig. 2
and their characterization data, and biological methods included. See DOI:
10.1039/c2md20243c
Fig. 1 Structures of cytosporone B (Csn-B, 1), 5-methyl-Csn-B (2), 3,5-dimethyl-
Csn-B (3), and di(1H-indol-3-yl)(4-trifluromethylphenyl)-methane (4).
‡ These authors contributed equally to this work.
332 | Med. Chem. Commun., 2013, 4, 332–339
This journal is ª The Royal Society of Chemistry 2013

View Article Online
Concise Article
MedChemComm
Csn-B to inhibit the viability of H460 non-small cell lung cancer Deprꢀotection of 3-benzyl ether 10 using boron tribromide at
(NSCLC) and LNCaP prostate cancer cells. Both cell lines have À78 C produced Csn-B (1). The overall yield for the ve steps
been reported to express NR4A1 in response to apoptotic or was 34.5%. The 3,5-dimethyl ether 6 was also acylated under
mitogenic stimuli.15,16
step d conditions to afford 3,5-(Me)2-Csn-B (3) (89% for two
steps). Low-temperature deprotection of 3 using boron tri-
bromide selectively cleaved its 3-methoxy group yielding 5-Me-
Csn-B (2) (82% for three steps). Schafer and Franck20 reported
that this type of ether cleavage was facilitated by the formation
of a stable six-membered ring complex that contained the
oxygen atoms of the ketone carbonyl and ether groups in the
substrate and the boron atom of the cleavage reagent. Hydro-
lysis of the complex produced the b-hydroxyketone. In the
present case, this complex would include the oxygen atom in
the n-octanoyl side chain and the 3-position hydroxyl oxygen in
the benzene ring of 5-Me-Csn-B (2) and the boron dibromide
moiety. The interaction between these two oxygens is supported
by the downeld position of the 3-hydroxyl proton that occurs in
the NMR spectrum of 5-Me-Csn-B (2), which is caused by its
hydrogen-bonding to the oxygen atom of the 2-n-octanoyl group.
Results and discussion
Synthesis of cytosporone B, its 5-methyl ether 2 and 3,5-
dimethyl ether 3
The routes used to synthesize Csn-B (1), 5-Me-Csn-B (2, pho-
mopsin C)17 and 3,5-(Me)2-Csn-B (3) are outlined in Fig. 2.
Although the synthesis of Csn-B has been reported by two
groups,9,17 the present ve-step approach shown in Fig. 2 is the
shortest thus far. The synthesis of 5-Me-Csn-B has also been
reported but in much lower yield17 than we obtained.
Converting 3,5-(Me)2-Csn-B (3) to Csn-B (1) was problematic
because typical methyl ether cleavage conditions produced only
5-Me-Csn-B (2), whereas attempted cleavage at higher temper-
atures ($0 ꢀC) led to mixtures. Therefore, the more readily
cleaved benzyl ether protecting group was used in the synthesis
of 1. This modication necessitated the transformation of 3,5-
dimethyl ether 6 to the 3,5-dibenzyl ether 8 (steps b and c in
Fig. 2). Introducing the n-octanoyl group at the 2-position of
Cytosporone B and its 5-methyl ether 2 and 3,5-dimethyl ether
3 dock to the NR4A1 ligand-binding domain
intermediate 8 to produce 9 (step d) employed the potent acyl- On the basis of the pose achieved by docking Csn-B (1) to the
ating agent octanoyl bis(triuoroacetyl)phosphate, which was crystal structure of the human NR4A1 LBD (PBD code 2QW4),
formed in situ from n-octanoic acid and triuoroacetic anhy- Zhan et al.8 hypothesized that Csn-B interacted with the
dride in phosphoric acid.18,19 Its potency may account for the phenolic hydroxyl group of Y122 in helix H5 (Y453 in NR4A1
debenzylation that occurred at the 5-position to give a 1 : 10 numbering) of the LBD. This interaction was demonstrated by
mixture of the 3,5-dibenzyl ether 9 and 3-benzyl ether 10. their binding affinity experiments and further supported by the
Fig. 2 Synthesis of cytosporone B (Csn-B, 1), 5-Me-Csn-B (2), and 3,5-(Me)2-Csn-B (3). Reagents and conditions: (a) Anhydrous ethanol, benzene, sulfuric acid, reflux,
20 h; yield: 93%. (b) Transformation of 6 to 7, the mixture of 9 plus 10 to 1, and 3 to 2: boron tribromide, dichloromethane, À78 ꢀC, 0.5 h, and 0 ꢀC, 4 h (6), À78 ꢀC, 4.5 h
(9 plus 10), and room temperature, 23 h (3); H2O; yields: 84% (7), 98% (1) and 92% (2). (c) Benzyl bromide, potassium carbonate, acetone, 68 ꢀC, 21.5 h; yield: 98%. (d)
Conversion of 8 to 9 plus 10, and 6 to 3: n-octanoic acid, trifluoroacetic anhydride/phosphoric acid (4/1), 23 h; yields: 4% (9) and 42% (10), and 96% (3).
This journal is ª The Royal Society of Chemistry 2013
Med. Chem. Commun., 2013, 4, 332–339 | 333

View Article Online
MedChemComm
Concise Article
observed loss of Csn-B-induced transactivation and trans-
localization activities by their NR4A1 (Y453A) mutant.
The low-energy conformation of Csn-B used for docking to
the same NR4A1 LBD structure had its 2-octanoyl side chain
held in an extended conformation, resembling that used by
Zhan et al.8 Our docking produced a similar pose (see Fig. 3A).
This pose suggested that the ethyl carboxylate, 2-n-octanoyl
carbonyl and 3- and 5-OH groups of Csn-B resided in a polar
cle on the LBD surface and were partially surrounded by
residues from helices H5 and H7, while its extended alkyl chain
resided in a hydrophobic pocket surrounded by residues from
helices H1, H5 and H8, and the H1–H3 loop. This pocket led
into the interior of the LBD. In our docking pose, Csn-B was
bounded by residues from helices H1, H3, H5, H7 and H8 and
the H1–H3, H5–S1 and H7–H8 loops. Our ligand-binding pocket
(LBP) differs from the canonical LBP found in the LBDs of other
NRs such as the retinoic acid and retinoid X nuclear receptors.
Their LBP is located deeper within the LBDs and bounded by
residues from helices H3, H5, H7 and H11 and the H5–H6
loop.21,22 Interestingly, the region comparable to the canonical
LBP in the rat NR4A1 (NGFI-B) LBD structure (PDB 1YJE) is lled
with aromatic and alkyl side chains of residues from helices
H3–H5.23 In summary, our docking to the 2QW4 structure
supports the reported docked Csn-B pose8 and suggests the
following potential H-bond stabilizing interactions (i) Csn-B
2-n-octanoyl carbonyl O atom with LBD H5 Y122 phenolic OH
˚
(4.8 A interatom distance between Os) that would need to be
mediated by a water molecule; (ii) Csn-B ester carboxylate O
˚
with H7 H163 imidazole ring amine H (3.93 A); (iii) Csn-B 3-OH
˚
with H5 Y122 OH (2.47 A) and H8 V167 backbone carbonyl O
˚
(2.96 A); and (iv) Csn-B 5-OH with LBD backbone carbonyl Os of
both H7 L162 and H163, with the former requiring mediation
˚
˚
by a water molecule (4.66 A) and the latter being direct (2.3 A),
˚
and with the H163 ring imine N (3.41 A), assuming that the
Fig. 3 Cytosporone B (Csn-B, 1), 5-Me-Csn-B (2) and 3,5-(Me)2-Csn-B (3) were
24
˚
docked to the human NR4A1 ligand-binding domain (PDB crystal structure code H-bond distance is 3 A.
2QW4) as described in the Methods (see ESI†). Ligand poses are shown in stick
Compared to our docked pose for Csn-B (1) (Fig. 3A), docking
format and residue side chains in line format. Carbon atoms in the pose for Csn-B
and in the accompanying NR4A1 binding-pocket side chains are shown in
magenta (A), those for 5-Me-Csn-B and accompanying side chains are in green
(B), and those for 3,5-(Me)2-Csn-B and accompanying side chains are in cyan (C).
of 5-Me-Csn-B (2) and 3,5-(Me)2-Csn-B (3) into the same LBP
produced poses (Fig. 3B and C, respectively) that were shied
farther from the helix H7 H163 backbone carbonyl O and
Nitrogen atoms are colored blue; oxygens, red; and sulfurs, yellow. Potential towards the LBD interior. As a result, the Cs at the benzene ring
hydrogen-bonds between O atoms are denoted by dashed red lines with
1-positions of 5-Me-Csn-B and 3,5-(Me)2-Csn-B were shied 3.7
˚
interatom distances given in A. H atoms have been omitted to enhance clarity.
˚
and 2.2 A away, respectively, from those of Csn-B. Despite this
Residue numbering is that used for the human NR4A1 LBD.
shi, 5-Me-Csn-B could maintain a H-bond between its 3-OH
Table 1 Dependence of cancer cell viability inhibition on the interatom distance between Y122 oxygen atom and oxygen atom at 3- or 5-position of ring of
compounds 1À3
Distance between Y122 residue
Concentration (mM) required to inhibit cell viability by
50% (IC50) (% inhibition at 20 mM)a
phenolic O atom and O atoms on
the benzene rings of 1–3 (A)
˚
Compound (groups at 3- and
5-ring positions)
H460 lung cancer
LNCaP prostate cancer
3-O atom
5-O atom
1 (3,5-dihydroxy)
2 (3-hydroxy, 5-methoxy)
3 (3,5-dimethoxy)
15.3 (88)
15.9 (75)
(11)
13.4 (76)
11.4 (80)
(14)
2.47
2.74
4.67
7.13
4.66
5.97
a
Cell lines were treated for 72 h.
334 | Med. Chem. Commun., 2013, 4, 332–339
This journal is ª The Royal Society of Chemistry 2013

View Article Online
Concise Article
MedChemComm
˚
˚
and the H5 Y122 OH(2.74 A). However, the distance (4.61 A) by androgens, some anti-androgens, progestins and estro-
between its ester carbonyl O and the Lys125 side chain N, while gens.29 This cancer cell line also expresses estrogen receptors a
smaller than that (5.47 A) of Csn-B would not permit any direct and b,30 constitutively activated casein kinase 2,30 and high
˚
H-bonding to the K125 amine H. The bulk of the 3-methyl group levels of ErbB2/HER-2/Neu,31 which up-regulates Akt2 kinase
of 3,5-(Me)2-Csn-B deected the Y122 side chain to suggest that activity to enhance cell proliferation,32 but does not express the
there would be no, or only weak, H-bond interactions between tumor suppressor gene PTEN.33 NR4A1 was found to be induced
the O at its 3-ring position and the Y122 phenolic OH. H-bond in this cell line on treatment with such cell stressors as
distances between the 3- and 5-Os in the low-energy docked apoptosis inducers, mitogens, anti-androgens, a calcium iono-
conformers of Csn-B and its methyl ethers (1–3) and the LBP phore, the anticancer drug etoposide15,34 and the apoptosis-
Y122 ring Os are listed in Table 1 in comparison with their inducer and small heterodimer partner ligand 3-Cl-AHPC.35
inhibitory effects on two cancer cell lines and suggest that the
LNCaP cell viability was determined aer a 72 h treatment
3-OHs of Csn-B and 5-Me-Csn-B would be responsible for their with increasing concentrations of Csn-B (1) and its methyl ether
higher activity.
These results led us to examine the impact of H-bond
acceptor or donor status of the 3- and 5-OHs of Csn-B on cancer
cell viability and NR4A1 function.
Cytosporone B is a more effective inhibitor of H460 lung and
LNCaP prostate cancer cell viability than its monomethyl and
dimethyl ethers
We rst examined the effects of Csn-B and its methyl ether
analogs 5-Me-Csn-B and 3,5-(Me)2-Csn-B on the growth of NCI-
H460 NSCLC cells in comparison with another reported NR4A1
agonist, di(1H-indol-3-yl)(4-triuromethylphenyl)methane (DIM-
Ph-4-CF3, 4), which was rst reported by the Safe group.7 The
H460 cell line was established from a human large-cell undif-
ferentiated lung carcinoma.25 It is reported to express the wild-
type tumor suppressor p53, have elevated levels of NAD(P)
H:quinone oxidoreductase (NQO1),26 which stabilizes p53 in the
nucleus, but lack the cyclin-dependent kinase 4 inhibitor
p16/INK4A,27 which functions as a G1 cell-cycle regulator. As
shown in Fig. 4A, Csn-B and its methyl ethers were less effective
than DIM-Ph-4-CF3 in inhibiting H460 cell viability. Aer 72 h of
treatment, the concentrations of Csn-B (1), 5-Me-Csn-B (2) and
3,5-(Me)2-Csn-B (3) required to inhibit cell viability by 50% (IC50
value) were 15.3, 15.9 and >20 mM, respectively (see Table 1),
whereas that for DIM-Ph-4-CF3 (4) was 4.6 mM. However, Csn-B
was more active than 5-Me-Csn-B and 3,5-(Me)2-Csn-B. In a
separate experiment, the activities of the four compounds (1–4)
at 10 mM were compared at 72 h. Their relative inhibitions of
H460 viability were 50%, 22%, 24% and 90%. Thus, both methyl
ethers had similar potencies at this concentration but these
values were half that of Csn-B. These results suggest that at
10 mM (i) hydrogen-bond donation by the Csn-B 3-OH was more
important as a determinant of cell viability inhibition than that
of the 5-OH; (ii) the ability of the 5-OH to function as a H-bond
donor did not have a role in reducing H460 cell viability; and (iii)
the interactions of Csn-B with NR4A1 had less impact on cell
viability than those of the DIM-Ph-4-CF3. These results are in
agreement with the report that a 24 h treatment of NR4A1-
Fig. 4 Effects of cytosporone B (Csn-B, 1), its monomethyl ether (5-Me-Csn-B, 2)
expressing BGC-823 gastric cancer cells with 10 mM 5-des- and dimethyl ether (3,5-(Me)2-Csn-B, 3) on cancer cell viability. (A) H460 lung
cancer cells in RPMI-1640 medium containing 10% fetal bovine serum were
hydroxy-Csn-B was unable to inhibit viability, whereas Csn-B
treated for 72 h with each compound or the positive control (DIM-Ph-4-CF3, 4) at
inhibited viability by 36%.9
0.1–20 mM or with vehicle alone (DMSO). (B) LNCaP prostate cancer cells were
Compounds 1–4 were next evaluated in androgen-depen-
dent,28 ATRA-sensitive LNCaP prostate cancer cells, which
similarly treated for 72 h at 1.0–20 mM compound or with vehicle alone. Cell
viability was measured using an MTT assay (see ESI†). Results shown are the
express a mutant androgen receptor capable of being activated means of six replicates Æ SD.
This journal is ª The Royal Society of Chemistry 2013
Med. Chem. Commun., 2013, 4, 332–339 | 335

View Article Online
MedChemComm
Concise Article
analogs 5-Me-Csn-B (2) and 3,5-(Me)2-Csn-B (3) in comparison independent of interaction with peroxisome proliferator-acti-
with DIM-Ph-4-CF3 (4) as a positive control (Fig. 4B). IC50 values vated nuclear receptor g, concentration dependent and
for viability inhibition were determined from the dose–response produced an IC50 value between 5 and 10 mM at 72 h,37 which is
curves and indicated that Csn-B and 5-Me-Csn-B had compa- comparable to that we attained (9.7 mM).
rable inhibitory activity (IC50 values of 13.4 and 11.4 mM,
respectively, and viability inhibitions of 76 Æ 7% and 80 Æ 8% at
Cytosporone B and its 5-methyl ether induce NR4A1 nuclear
export
20 mM, Table 1). Whereas an IC50 value for 3,5-(Me)2-Csn-B was
not reached at 20 mM, the highest concentration evaluated, at
which viability inhibition was only 14 Æ 7%. The positive Zhang and co-workers rst reported that several inducers of
control, DIM-Ph-4-CF3, was again the most active inhibitor (IC50 cancer cell apoptosis also induced the translocation of NR4A1
value of 9.7 mM and 97% growth inhibition at 20 mM).
from the nucleus to the cytoplasm in the same cell lines.4
In terms of the IC50 values for 50% inhibition of H460 and Translocation was accompanied by the colocalization of NR4A1
LNCaP cell viability at 72 h, the positive control, DIM-Ph-4-CF3 with mitochondrial Bcl-2.4,10 The interaction between NR4A1
(4), with IC50 values of 4.6 and 9.7 mM, respectively, was more and Bcl-2 caused the cytoprotective Bcl-2 protein to undergo a
active than Csn-B (1) and 5-methyl-Csn-B (2). Its IC50 value for conformational change that led to induction of apoptosis, a
inhibiting H460 lung cancer cell growth is in the same process involving loss of mitochondrial membrane potential,
concentration range as that reported by the Safe group (approx. release of mitochondrial cytochrome c and activation of cas-
7.7 mM).36 Moreover, as this group previously reported, the pases.4,6 Aer these reports, Wu and co-workers demonstrated
inhibition of LNCaP cell viability by DIM-Ph-4-CF3 was that a 12 h treatment of human BCG-8223 gastric cancer cells
Fig. 5 Cytosporone B (Csn-B, 1) and 5-Me-Csn-B (2) induce the export of NR4A1 from the cell nucleus to the cytoplasm. H460 cells were treated for 8 h with 10 mM
Csn-B (1), 5-Me-Csn-B (2), 3,5-(Me)2-Csn-B (3) or DIM-Ph-4-CF3 (4) or DMSO vehicle alone (control) for 8 h, fixed and then stained for NR4A1 using anti-rabbit NR4A1
monoclonal antibody followed by goat anti-rabbit antibody linked to Alexa Fluor 594 (red in the panels of column 1) and for nuclear double-stranded DNA using DAPI
(blue in panels of column 2) (see Methods (ESI†)). Cells were visualized by confocal microscopy and images overlaid (column 3). Localization of NR4A1 in the cell nuclei is
demonstrated by the magenta color of the nuclei in the overlays for the control and 3,5-(Me)2-Csn-B (3)-treated cells (column 3), whereas the nuclei remain blue and
the cytoplasm is red in cells treated with Csn-B, 5-Me-Csn-B or DIM-Ph-4-CF3.
336 | Med. Chem. Commun., 2013, 4, 332–339
This journal is ª The Royal Society of Chemistry 2013

View Article Online
Concise Article
MedChemComm
with 15.5 mM Csn-B (1) induced the export of NR4A1 from the condensation and/or DNA fragmentation by Csn-B, 5-Me-Csn-B
nucleus to the cytoplasm and NR4A1 colocalization with the and DIM-Ph-4-CF3 was at least two-fold higher than that
mitochondrial heat-shock protein (HSP) 60.8 Similarly, treat- induced by 3,5-(Me)2-Csn-B and 4.4, 5.1 and 3.7-fold, respec-
ment of H460 cells with 10 mM Csn-B, 5-Me-Csn-B (2) or DIM-Ph- tively, higher than that that observed for the DMSO control. In
4-CF3 (4) resulted in the translocation of NR4A1 from the addition to chromatin condensation, cleavage of the 116 kDa
nucleus to the cytoplasm (Fig. 5). In contrast and as observed DNA repair protein poly(ADP)ribose polymerase (PARP) to an 85
with the DMSO alone-treated control, treatment of the cells with kDa fragment by activated proteases is a marker of apoptosis.38
the 3,5-(Me)2-Csn-B (3) was unable to induce the export of Again, 3,5-(Me)2-Csn-B was less potent than the other three
NR4A1 from the nucleus.
compounds (Fig. 6B and C).
Cytosporone B and its 5-methyl ether induce apoptosis
Cytosporone B and its 5-methyl ether interact with the human
NR4A1 LBD
Wu and co-workers also reported that a 48 h treatment with
15.5 mM Csn-B induced 63% BCG-823 cell apoptosis as Csn-B (1) was reported to physically interact with the NR4A1
measured by nuclear DNA condensation, mitochondrial cyto- LBD by its quenching of LBD uorescence and circular
chrome c release and procaspase-9 cleavage.8 The apoptosis- dichroism, but was unable to interact with the LBD Y122A
inducing effects of Csn-B (1), its methyl ethers 5-Me-Csn-B (2) mutant as demonstrated by its inability to quench the uores-
and 3,5-(Me)2-Csn-B (3) and DIM-Ph-4-CF3 (4) on H460 cells cence of the mutant.8 On this basis, Csn-B was used as a positive
were examined at a lower concentration (10 mM) and at an control in binding studies using one-dimensional 1H NMR
earlier time point (8 h) to observe early apoptotic events. As spectroscopy to determine the interaction of methyl ethers
shown by nuclear staining in Fig. 6A, induction of chromatin 5-Me-Csn-B (2) and 3,5-(Me)2-Csn-B (3) with the recombinant
Fig. 6 Cytosporone B (Csn-B) and 5-Me-Csn-B induce H460 cell apoptosis as evidenced by chromatin condensation and/or DNA fragmentation and by PARP cleavage.
Cells were treated for 8 h with Csn-B (1), 5-Me-Csn-B (2), 3,5-(Me)2-Csn-B (3) or DIM-Ph-4-CF3 (4) at 10 mM or with DMSO vehicle alone (control). (A) Cells were fixed,
permeabilized and stained using DAPI for visualization of nuclei (blue). Fields of 300 cells were scored for nuclear fragmentation and/or chromatin condensation to
determine those undergoing apoptosis. Results shown are representative of three replicates. (B) Levels of chromatin condensation and/or DNA fragmentation shown
are averages of three replicates Æ SD and represented graphically. (C) Cells were lysed, and lysates separated using SDS-PAGE. Separated proteins were transferred to
nitrocellulose membranes and blots stained for PARP cleavage using anti-PARP antibody followed by a peroxidase-conjugated secondary antibody. Staining for b-actin
was used as a loading control. Results shown are representative of three replicates. Procedures are described in the ESI.†
This journal is ª The Royal Society of Chemistry 2013
Med. Chem. Commun., 2013, 4, 332–339 | 337

View Article Online
MedChemComm
Concise Article
human NR4A1 LBD. NMR can be used to establish the inter-
action of a small molecule with a much larger protein, because
aer their complexation the proton signal relaxation time of the
former increases leading to broadening of the proton signals in
its NMR spectrum. Proton signals for the aromatic regions from
5.85 to 6.35 ppm in the NMR spectra of Csn-B, 5-Me-Csn-B and
3,5-(Me)2-Csn-B (1–3) were selected for comparison purposes
because they did not overlap with proton signals from either the
protein or buffer. The two singlets for the 4- and 6-hydrogens on
the benzene rings of Csn-B and 5-Me-Csn-B (Fig. 7A and B,
respectively) were totally suppressed in the presence of the LBD
protein to indicate that both compounds interacted with the
protein. In contrast, the peak heights for these two aromatic
hydrogens in the spectrum of 3,5-(Me)2-Csn-B did not decrease
in the presence of the protein (Fig. 7C) to indicate that its
interaction with the LBD was absent or very weak. The absence
of any specic interaction between 3,5-(Me)2-Csn-B and the LBD
protein was also demonstrated using isothermal titration calo-
rimetry (data not shown).
The suppression of the aromatic hydrogen signals in the
NMR spectra of Csn-B and 5-Me-Csn-B in the presence of the
NR4A1 LBD protein (Fig. 7A and B, respectively) supports their
interaction with the NR4A1 LBD and is in agreement with
earlier reports on the interaction of Csn-B with the LBD.8,9 The
present results also support the importance of the unprotected
phenolic 3-OH groups at the 3-position in the benzene rings of
Csn-B and 5-Me-Csn-B and their H-bond donation to the LBD as
evidenced by (i) the LBD-mediated suppression of the same two
aromatic hydrogen signals in the NMR spectra of these
compounds; and (ii) the absence of signal suppression by the
LBD in the spectrum of 3,5-(Me)2-Csn-B (3), which is unable to
donate such a H. The importance of the H-bond interaction of
the 3-OH with the LBD Y122 phenolic OH was rst demon-
strated by loss of interaction of Csn-B with the Y453A mutant of
the full-length receptor8 and the retention of binding to NR4A1
by 5-deshydroxy-Csn-B (only 12% higher Kd value than Csn-B).9
Moreover, the similar abilities of Csn-B and 5-Me-Csn-B to
inhibit cancer cell viability and the inability of 3,5-(Me)2-Csn-B
to comparably do so support the role of NR4A1–ligand inter-
action in mediating NR4A1 activity. The present results also
suggest that mediating NR4A1 activity using more potent
analogs of Csn-B would be a feasible therapeutic approach for
treating cancer or other NR4A1-regulated diseases.
Fig. 7 Comparison of proton nuclear magnetic resonance spectra of cytospor-
one B (Csn-B, 1), 5-Me-Csn-B (2) or 3,5-(Me)2-Csn-B (3) alone and in the presence
of recombinant NR41A ligand-binding domain (LBD) protein indicates that Csn-B
and 5-Me-Csn-B interact with this protein, whereas 3,5-(Me)2-Csn-B does not.
Low-field regions of one-dimensional 1H NMR spectra of each compound alone
(upper spectrum in black) and combined with recombinant LBD protein (lower
spectrum in red) are shown. As described in the Methods (ESI†), spectra of
100 mM Csn-B (A), 50 mM 5-Me-Csn-B (B) and 50 mM 3,5-(Me)2-Csn-B (C) in the
absence and presence of the protein at 10 mM for Csn-B and at 5 mM for 5-Me-
Csn-B and 3,5-(Me)2-Csn-B were taken at 11 ꢀC to maintain protein stability.
Interaction between compound and protein is demonstrated by loss of
compound proton signals in the presence of protein (A and B). Spectral regions
shown for comparison are those lacking overlapping signals from the buffer,
protein and water.
Conclusions
In summary, cytosporone B (Csn-B, 1) and its two derivatives,
5-Me-Csn-B (2) and 3,5-(Me)2-Csn-B (3), were synthesized and
evaluated for their abilities to inhibit the viability of human
H460 lung and LNCaP prostate cancer cells, and to induce
NR4A1 nuclear export and apoptosis of H460 cancer cells. Both
Csn-B and 5-Me-Csn-B having a 3-OH on their aromatic rings
have similar bioactivities, whereas, 3,5-(Me)2-Csn-B was inactive
or showed weak activities. These results suggest that the 3-OH is
more important than the 5-OH in negatively impacting cancer
1
cell viability and inducing apoptosis, and agree with H NMR
ligand–receptor binding experiments, which indicated both
338 | Med. Chem. Commun., 2013, 4, 332–339
This journal is ª The Royal Society of Chemistry 2013

View Article Online
Concise Article
MedChemComm
Csn-B and 5-Me-Csn-B strongly interacted with NR4A1 LBD, 16 S. K. Kolluri, N. Bruey-Sedano, X. Cao, B. Lin, F. Lin,
whereas 3,5-(Me)2-Csn-B showed no or weak interaction. These
results were supported by the results of docking the three
Y. H. Han, M. I. Dawson and X.-K. Zhang, Mol. Cell. Biol.,
2003, 23, 8651–8667.
compounds to the NR4A1 LBD. However, the interactions of 17 H. Yoshida, T. Morishita and J. Ohshita, Chem. Lett., 2010,
Csn-B and 5-Me-Csn-B with the LBD led to weaker inhibition of
cell viability than that of the positive control DIM-Ph-4-CF3.
39, 508–509.
18 G. P. Luke, C. K. Seekamp, Z. Q. Wang and B. L. Chenard, J.
Org. Chem., 2008, 73, 6397–6400.
19 T. P. Smyth and B. W. Corby, J. Org. Chem., 1998, 63, 8946–
8951.
Acknowledgements
¨
20 W. Schaefer and B. Franck, Chem. Ber., 1966, 99, 160–164.
This work was supported in part by NIH Grant R01 CA107039
(M.I.D., X.-K.Z.) and by a pilot grant to D.A.W. from NIH Grant
P30 CA030199-30.
21 B. P. Klaholz, J. P. Renaud, A. Mitschler, C. Zusi,
P. Chambon, H. Gronemeyer and D. Moras, Nat. Struct.
Biol., 1998, 5, 199–202.
22 P. F. Egea, A. Mitschler and D. Moras, Mol. Endocrinol., 2002,
16, 987–997.
References
23 R. Flaig, H. Greschik, C. Peluso-Iltis and D. Moras, J. Biol.
Chem., 2005, 280, 19250–19258.
24 T. Steiner, Angew. Chem., Int. Ed., 2002, 41, 49–76.
25 T. J. Hegedus, M. Falzon, N. Margaretten, A. F. Gazdar and
H. M. Schuller, Biochem. Pharmacol., 1987, 36, 3339–
3343.
1 Q. Wu, S. Liu, X. F. Ye, Z. W. Huang and W. J. Su,
Carcinogenesis, 2002, 23, 1583–1592.
2 S. O. Lee, X. Li, S. Khan and S. Safe, Expert Opin. Ther. Targets,
2011, 15, 195–206.
3 R. E. Paulsen, C. A. Weaver, T. J. Fahrner and J. Milbrandt, J.
Biol. Chem., 1992, 267, 16491–16496.
4 X. Cao, W. Liu, F. Lin, H. Li, S. K. Kolluri, B. Lin, Y.-H. Han,
M. I. Dawson and X.-K. Zhang, Mol. Cell. Biol., 2004, 24,
9705–9725.
26 S. L. Winski, R. H. Hargreaves, J. Butler and D. Ross, Clin.
Cancer Res., 1998, 4, 3083–3088.
27 X. Y. Fu, S. W. Zhang, R. Q. Ran, Z. H. Shen, J. X. Gu and
S. L. Cao, J. Cancer Res. Clin. Oncol., 1998, 124, 621–626.
28 J. S. Horoszewicz, S. S. Leong, E. Kawinski, J. P. Karr,
H. Rosenthal, T. M. Chu, E. A. Mirand and G. P. Murphy,
Cancer Res., 1983, 43, 1809–1818.
5 Y.-H. Han, X. Cao, B. Lin, F. Lin, S. K. Kolluri, J. Stebbins,
J. C. Reed, M. I. Dawson and X.-K. Zhang, Oncogene, 2006,
25, 2974–2986.
6 F. Luciano, M. Krajewska, P. Ortiz-Rubio, S. Krajewski,
D. Zhai, B. Faustin, J. M. Bruey, B. Bailly-Maitre,
A. Lichtenstein, S. K. Kolluri, A. C. Satterthwait,
X.-K. Zhang and J. C. Reed, Blood, 2007, 109, 3849–3855.
7 S. Chintharlapalli, R. Burghardt, S. Papineni, S. Ramaiah,
K. Yoon and S. Safe, J. Biol. Chem., 2005, 280, 24903–24914.
8 Y. Zhan, X. Du, H. Chen, J. Liu, B. Zhao, D. Huang, G. Li,
Q. Xu, M. Zhang, B. C. Weimer, D. Chen, Z. Cheng,
L. Zhang, Q. Li, S. Li, Z. Zheng, S. Song, Y. Huang, Z. Ye,
W. Su, S. C. Lin, Y. Shen and Q. Wu, Nat. Chem. Biol.,
2008, 4, 548–556.
9 J. J. Liu, H. N. Zeng, L. R. Zhang, Y. Y. Zhan, Y. Chen,
Y. Wang, J. Wang, S. H. Xiang, W. J. Liu, W. J. Wang,
H. Z. Chen, Y. M. Shen, W. J. Su, P. Q. Huang, H. K. Zhang
and Q. Wu, Cancer Res., 2010, 70, 3628–3637.
10 S. F. Brady, M. P. Singh, J. E. Janso and J. Clardy, Org. Lett.,
2000, 2, 4047–4049.
11 A. Philips, S. Lesage, R. Gingras, M. H. Maira, Y. Gauthier,
P. Hugo and J. Drouin, Mol. Cell. Biol., 1997, 17, 5946–5951.
12 B. M. Forman, K. Umesono, J. Chen and R. M. Evans, Cell,
1995, 81, 541–550.
13 T. Perlmann and L. Jansson, Genes Dev., 1995, 9, 769–782.
14 R. F. Paulsen, K. Granas, H. Johnsen, V. Rolseth and S. Sterri,
J. Mol. Neurosci., 1995, 6, 249–255.
29 J. Veldscholte, C. Ris-Stalpers, G. G. Kuiper, G. Jenster,
C. Berrevoets, E. Claassen, H. C. van Rooij, J. Trapman,
A. O. Brinkmann and E. Mulder, Biochem. Biophys. Res.
Commun., 1990, 173, 534–540.
30 C. Le Page, I. H. Koumakpayi, L. Lessard, F. Saad and
A. M. Mes-Masson, Prostate, 2005, 65, 306–315.
31 C. Le Page, I. H. Koumakpayi, B. Peant, N. Delvoye, F. Saad
and A. M. Mes-Masson, Prostate, 2012, 72, 777–788.
32 N. Moriyama-Gonda, H. Shiina, M. Terashima, K. Satoh and
M. Igawa, BJU Int., 2008, 101, 485–491.
33 H. Huang, J. C. Cheville, Y. Pan, P. C. Roche, L. J. Schmidt
and D. J. Tindall, J. Biol. Chem., 2001, 276, 38830–
38836.
34 C. Y. Young, P. E. Murtha and J. Zhang, Oncol. Res., 1994, 6,
203–210.
35 M. I. Dawson, P. D. Hobbs, V. J. Peterson, M. Leid,
C. W. Lange, K. C. Feng, G. Chen, J. Gu, H. Li,
S. K. Kolluri, X. Zhang, Y. Zhang and J. A. Fontana, Cancer
Res., 2001, 61, 4723–4730.
36 N. Ichite, M. B. Chougule, T. Jackson, S. V. Fulzele, S. Safe
and M. Singh, Clin. Cancer Res., 2009, 15, 543–552.
37 S. Chintharlapalli, S. Papineni and S. Safe, Mol. Pharmacol.,
2007, 71, 558–569.
38 S. H. Kaufmann, S. Desnoyers, Y. Ottaviano, N. E. Davidson
and G. G. Poirier, Cancer Res., 1993, 53, 3976–3985.
15 H. Uemura and C. Chang, Endocrinology, 1998, 139, 2329–
2334.
This journal is ª The Royal Society of Chemistry 2013
Med. Chem. Commun., 2013, 4, 332–339 | 339