Structural elucidation and synthesis of eudistidine A: An unusual polycyclic marine alkaloid that blocks interaction of the protein binding domains of p300 and HIF-1α

Article abstract of DOI:10.1021/jacs.5b02156

Low oxygen environments are a hallmark of solid tumors, and transcription of many hypoxia-responsive genes needed for survival under these conditions is regulated by the transcription factor HIF-1 (hypoxia-inducible factor 1). Activation of HIF-1 requires binding of its α-subunit (HIF-1α) to the transcriptional coactivator protein p300. Inhibition of the p300/HIF-1α interaction can suppress HIF-1 activity. A screen for inhibitors of the protein binding domains of p300 (CH1) and HIF-1α (C-TAD) identified an extract of the marine ascidian Eudistoma sp. as active. Novel heterocyclic alkaloids eudistidines A (1) and B (2) were isolated from the extract, and their structures assigned by spectroscopic analyses. They contain an unprecedented tetracyclic core composed of two pyrimidine rings fused with an imidazole ring. Eudistidine A (1) was synthesized in a concise four-step sequence featuring a condensation/cyclization reaction cascade between 4-(2-aminophenyl)pyrimidin-2-amine (3) and 4-methoxy-phenylglyoxal (4), while eudistidine B (2) was synthesized in a similar fashion with glyoxylic acid (5) in place of 4. Naturally occurring eudistidine A (1) effectively inhibited CH1/C-TAD binding with an IC₅₀ of 75 μM, and synthetic 1 had similar activity. The eudistidine A (1) scaffold, which can be synthesized in a concise, scalable manner, may provide potential therapeutic lead compounds or molecular probes to study p300/HIF-1α interactions and the role these proteins play in tumor response to low oxygen conditions. The unique structural scaffolds and functional group arrays often found in natural products make these secondary metabolites a rich source of new compounds that can disrupt critical protein-protein binding events.

Full text of DOI:10.1021/jacs.5b02156

Article
pubs.acs.org/JACS
Structural Elucidation and Synthesis of Eudistidine A: An Unusual
Polycyclic Marine Alkaloid that Blocks Interaction of the Protein
Binding Domains of p300 and HIF-1α
Susanna T. S. Chan,† Paresma R. Patel,‡,§ Tanya R. Ransom,† Curtis J. Henrich,†,⊥ Tawnya C. McKee,†,#
Andrew K. L. Goey,∥ Kristina M. Cook,∥ William D. Figg,∥ James B. McMahon,† Martin J. Schnermann,*,‡
and Kirk R. Gustafson*,†
†Molecular Targets Laboratory, Center for Cancer Research, National Cancer Institute, Frederick, Maryland 21702-1201, United
States
‡Chemical Biology Laboratory, Center for Cancer Research, National Cancer Institute, Frederick, Maryland 21702-1201, United
States
§National Center for Advancing Translational Sciences, National Institutes of Health, Bethesda, Maryland 20892, United States
⊥Basic Science Program, Leidos Biomedical Research, Inc., Frederick National Laboratory for Cancer Research, Frederick, Maryland
21702-1201, United States
∥Genitourinary Malignancies Branch, Center for Cancer Research, National Cancer Institute, Bethesda, Maryland 20892, United
States
S
* Supporting Information
ABSTRACT: Low oxygen environments are a hallmark of
solid tumors, and transcription of many hypoxia-responsive
genes needed for survival under these conditions is regulated
by the transcription factor HIF-1 (hypoxia-inducible factor 1).
Activation of HIF-1 requires binding of its α-subunit (HIF-1α)
to the transcriptional coactivator protein p300. Inhibition of
the p300/HIF-1α interaction can suppress HIF-1 activity. A
screen for inhibitors of the protein binding domains of p300
(CH1) and HIF-1α (C-TAD) identified an extract of the
marine ascidian Eudistoma sp. as active. Novel heterocyclic
alkaloids eudistidines A (1) and B (2) were isolated from the
extract, and their structures assigned by spectroscopic analyses.
They contain an unprecedented tetracyclic core composed of
two pyrimidine rings fused with an imidazole ring. Eudistidine A (1) was synthesized in a concise four-step sequence featuring a
condensation/cyclization reaction cascade between 4-(2-aminophenyl)pyrimidin-2-amine (3) and 4-methoxy-phenylglyoxal (4),
while eudistidine B (2) was synthesized in a similar fashion with glyoxylic acid (5) in place of 4. Naturally occurring eudistidine A
(1) effectively inhibited CH1/C-TAD binding with an IC50 of 75 μM, and synthetic 1 had similar activity. The eudistidine A (1)
scaffold, which can be synthesized in a concise, scalable manner, may provide potential therapeutic lead compounds or molecular
probes to study p300/HIF-1α interactions and the role these proteins play in tumor response to low oxygen conditions. The
unique structural scaffolds and functional group arrays often found in natural products make these secondary metabolites a rich
source of new compounds that can disrupt critical protein−protein binding events.
implicated in tumor progression, metastasis, resistance to
chemotherapies, and poor clinical outcomes for a variety of
tumor types.3,4 While inhibition of HIF-1 function has long
been recognized as an attractive potential anticancer target,5
effective chemotherapeutic agents that can modulate or target
HIF-1 activity are still needed.6
INTRODUCTION
■
Hypoxia is a condition characteristic of many solid tumors that
results from accelerated cellular proliferation and inadequate
vascularization. The transcription factor HIF-1 (hypoxia-
inducible factor 1) is critical for initiating adaptive responses
to low oxygen environments and maintaining cellular homeo-
stasis. HIF-1 regulates the transcription of numerous hypoxia-
responsive genes, including ones that modulate glycolysis and
glucose flux, and those associated with vasodilation and
angiogenesis such as vascular endothelial growth factor.1,2
Overexpression or dysregulation of HIF-1 function has been
HIF-1 is a heterodimer of α- and β-subunits, and the
transcriptional activity of this complex is regulated by either the
Received: February 27, 2015
© XXXX American Chemical Society
A
DOI: 10.1021/jacs.5b02156
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
accumulation or turnover of the HIF-1α monomer.7 Under
normoxic conditions, HIF-1α is rapidly recycled in a process
that involves hydroxylation of specific proline residues, followed
by ubiquitination and proteasomal degradation of the
monomer.8 When oxygen tension is low, HIF-1α accumulates
in the nucleus where it dimerizes with the constitutively present
HIF-1β subunit. This allows recruitment and binding of the
transcriptional coactivator p300, a multidomain protein that not
only plays a crucial role in HIF-1 activation but also has
intrinsic histone acetyl transferase and polyubiquitin ligase
activities.9−12 The essential binding interaction between p300
and HIF-1 that facilitates hypoxia-induced transcription
involves the cysteine histidine-rich domain 1 (CH1) of p300
and the C-terminal transactivation domain (C-TAD) of HIF-
1α.13−15 With the goal of identifying small molecule inhibitors
of the p300/HIF-1α interaction, we developed a high-
throughput assay based on the methodologies of Kung et
al.16 to screen for compounds and natural product extracts that
can disrupt CH1/C-TAD binding as a means of blocking HIF-1
activation and potentially inhibiting tumor response to hypoxic
conditions.
carbon signals (δC 55.1 and 94.4) in the 13C NMR spectrum
(Table 1). Resonances attributable to 10 aromatic protons (δH
Table 1. NMR Spectroscopic Data for Eudistidine A (1) in
DMSO-d6
a
position
δC
δN
δH (mult, J in Hz)
HMBC
1
255.9
2
166.2
92.2
8.54 (d, 4.9)
6.95 (d, 4.8)
−
3, 3a, 3b, 8c
c
3
2, 3a, 3b, 8a
3a
144.5
117.0
125.5
128.7
135.0
128.6
145.7
−
−
3b
−
4
8.40 (d, 8.2)
7.71 (dt, 1.0, 7.6)
7.92 (dt, 1.1, 7.6)
7.85 (d, 8.0)
−
3a, 6, 7a
3b, 7
4, 7a
3b, 5
−
5
6
7
7a
8
254.2
164.0
8a
160.0
154.8
−
−
−
−
8b
8c
b
9
n.o.
10
94.4
−
133.4
127.5
113.3
159.1
55.1
−
−
RESULTS AND DISCUSSION
■
10-OH
11
6.86 (s)
8a, 10, 11
A high-throughput screening assay was developed that
employed the N-terminal biotinylated HIF-1α C-TAD domain
(aa 786−826) bound to streptavidin-coated 384 well plates and
a GST-labeled p300 CH1 domain (aa 323−423). Binding of
soluble CH1 to the immobilized C-TAD was measured by the
fluorescence of a europium-tagged anti-GST antibody. The
CH1/C-TAD binding interaction is very strong, as it forms a
highly ordered complex with 3393 Å2 of buried surface area.14
The hit rate from a screen for small molecule inhibitors of this
protein−protein interaction was anticipated to be very low.
Prior screening of a library of 600 000 pure compounds using a
similar CH1/C-TAD system provided only a single lead
compound.16 A primary focus of the current study was to mine
the vast chemical diversity of the NCI’s Natural Products
Repository, which contains extracts from taxonomically diverse
terrestrial plants, marine invertebrates, and microbial isolates.17
Our testing of >158 000 crude natural product extracts and
prefractionated extract samples for inhibition of CH1/C-TAD
binding identified 27 confirmed hits. One of these was a
relatively polar fraction from the organic solvent extract of the
colonial marine ascidian Eudistoma sp. collected in Palau that
exhibited significant activity; therefore, detailed chemical
studies were initiated. Bioassay-guided fractionation of the
crude extract employing a combination of diol SPE cartridges
and repeated Sephadex LH-20 column chromatography
provided two new compounds that were named eudistidine A
(1) and eudistidine B (2).
−
−
12/16
13/15
14
7.54 (d, 8.8)
6.91 (d, 8.9)
−
10, 11, 12/16, 14
13/15, 14
−
14
14-OMe
3.74 (s)
a
15N assignments were based on 1H−15N HMBC correlations. The δN
values were not calibrated to an external standard but were referenced
to neat NH3 (δ 0.00) using the standard Bruker parameters. Not
observed. Weak four-bond HMBC correlation.
b
c
8.54, 8.40, 7.92, 7.85, 7.71, 7.54, 6.95, 6.91), one methoxyl (δH
3.74, 3H), and one hydroxyl group (δH 6.86) were observed in
the 1H NMR spectrum (DMSO-d6). COSY NMR data
identified the presence of three proton−proton spin-systems:
a 1,2-disubstituted benzene ring [δH 8.40 (d, J = 8.2 Hz), 7.92
(dt, J = 1.1, 7.6 Hz), 7.85 (d, J = 8.0 Hz), 7.71 (dt, J = 1.0, 7.6
Hz)], a para-substituted benzene ring [(δH 7.54, 2H (d, J = 8.8
Hz), 6.91, 2H (d, J = 8.9 Hz)], and two mutually coupled
heteroaromatic protons [(δH 8.54 (d, J = 4.9 Hz), 6.95 (d, J =
4.8 Hz)]. HSQC and HMBC data confirmed both the 1,2-
disubstituted benzene ring (designated as ring B) and the para-
substituted benzene ring (ring E), and established the presence
of a methoxyl group on the latter ring (Figure 1).
Deshielded carbon and proton chemical shifts (δC 166.2/δH
8.54) observed for the C-2 aromatic methine indicated it was
situated directly adjacent to a nitrogen, while the highly
shielded signals for C-3 (δC 92.2/δH 6.95) suggested the
presence of additional nitrogen functionality in ring A. The
position of C-3a in the A ring and linkage to ring B were
revealed by HMBC correlations from H-2, H-3, and H-4. An
HMBC correlation from H-3 to C-3b confirmed the A-B ring
linkage, while the deshielded chemical shift of C-3a (δC 144.5)
indicated it was directly bound to a nitrogen atom. An HMBC
correlation from H-2 to a quaternary carbon resonance at δC
154.8 (C-8c) suggested this carbon was part of a conjugated
guanidino moiety. This functionality could be accommodated
by linking C-8c with the nitrogen (N-8b) that was also joined
to C-3a, to give a pyrimidine core for ring A. Observed
correlations in the HMBC spectrum from protons at δH 8.40
Eudistidine A (1) was isolated as an optically inactive green
oil with a molecular formula of C19H14N4O2 that was
established by HRESI mass spectrometry and required 15
degrees of unsaturation. The highly aromatic character of 1 was
apparent from the presence of 17 sp2 carbons and only two sp3
B
DOI: 10.1021/jacs.5b02156
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
1
1
Figure 1. Selected H−13C HMBC (black arrows), H−15N HMBC (red arrows), and COSY (bold blue lines) correlations for eudistidine A (1).
(H-4) and 7.92 (H-6) to a deshielded carbon resonance δC
145.7 (C-7a), and from δH 7.85 (H-7) and 7.71 (H-5) to δC
117.0 (C-3b) suggested that ring B was fused with another
nitrogen-containing heterocycle (ring C). A weak 4-bond
HMBC correlation from H-3 in ring A to a carbon resonance at
δC 160.0 was also observed. The deshielded nature of this
aromatic carbon suggested it was directly bound to more than
one nitrogen atom, and the fact that no other HMBC
correlations were observed indicated it was remote from all
other aromatic protons. This carbon (C-8a) was assigned to a
second fused pyrimidine ring (ring C) that incorporated
nitrogens N-8 and N-8b, and was situated between rings A and
1
B. H−15N HMBC correlations from H-2 and H-3 to N-1 (δN
255.9), H-3 to N-8b (δN 164.0), and H-7 to N-8 (δN 254.2)
supported both this assignment and the fused nature of rings A,
B, and C.
HMBC correlations observed from the methoxyl protons (δH
3.74) to a carbon resonance at δC 159.1 (C-14), as well as from
the aromatic protons at δH 7.54 (H-12/H-16) to C-14 and
from δH 6.91 (H-13/H-15) to both δC 133.4 (C-11) and 113.3
(C-13/C-15), defined the para-methoxy-substituted benzene
ring (E). A correlation from the H-12/H-16 protons to a
quaternary carbon resonance at δC 94.4 revealed this carbon
(C-10) was the connection point between ring E and the rest of
the molecule. The chemical shift of C-10 was appropriate for a
hemiaminal functionality, and this was supported by an HMBC
Figure 2. pH effects on the UV absorption spectra of compound 1 (A)
and 2 (B).
correlation from the hydroxyl proton at δH 6.86 to C-10. An
HMBC correlation from the OH proton to C-8a (δC 160.0)
hydroxyl proton signal (δH 6.86) and a hemiaminal carbon
1 can be envisioned (Figure 3), the presence of both a sharp
also revealed that C-8a and C-10 were adjacent. One additional
ring was needed to satisfy the unsaturation equivalents implicit
in the molecular formula of 1. This was accommodated by
linking together C-10 and N-9 to form a fused imidazole ring
(D), which then completed the planar structural assembly of
eudistidine A (1). The pentacyclic framework of eudistidine A
(1) contains several novel structural elements. Specifically, the
4-hydroxy, 4-aryl, 5-imino dihydroimadzole ring system and the
fused pyrimido-quinazoline heterocycle are unprecedented in
the natural product literature. These rare structural features are
suggestive of an unusual biogenesis of 1.
Figure 3. Cyclized (a) and ring-opened (b) tautomers of eudistidine
A.
The extended π-conjugation of eudistidine A (1) prompted
us to investigate its UV absorption properties. While only
sparingly soluble in methanol, 1 showed excellent solubility in
DMSO and good aqueous solubility. UV absorption maxima in
DMSO were observed at 327 and 450 nm. In aqueous solutions
the absorption profile was strongly pH dependent. Local
absorption maxima were observed at 393 and 440 nm in acidic
and basic conditions, respectively, and an isosbestic point at
407 nm (Figure 2). The pKa of the titratable proton in 1 was
5.1 0.1 (Supporting Information).
resonance (δC 94.4), and the lack of any signal for a ketone
carbon in either the 13C NMR spectrum or the IR spectrum of
1, indicated that the cyclized form is predominant under all
conditions examined to date.
Eudistidine B (2) was isolated as a yellow-green oil, and its
molecular formula of C12H6N4O was established by (+)-HRESI
mass spectrometry. The molecular formula was indicative of an
aromatic heterocyclic structure that incorporated 12 unsatura-
tion equivalents. Eudistidine B (2) had rather poor solubility in
MeOH so DMSO was used for the UV studies, which showed
λmax at 327 and 445 nm. Direct comparison of the NMR data
for 2 (Table 2) with those obtained for eudistidine A (1)
Eudistidine A (1) was isolated as an optically inactive
compound (no optical rotation [α]D or Cotton effects in the
CD spectrum when recorded in DMSO); thus, the natural
product is racemic. While a ring-opened tautomeric structure of
C
DOI: 10.1021/jacs.5b02156
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
Table 2. NMR Spectroscopic Data for Eudistidine B (2) in
DMSO-d6
Scheme 1. Retrosynthetic Analysis of Eudistidine A (1) and
B (2)
position
δC
δH (mult, J in Hz)
HMBC
2
163.8
102.2
144.2
117.3
126.5
131.6
136.6
130.9
145.6
142.5
163.3
172.0
9.10 (d, 5.6)
8.26 (d, 5.8)
3, 3a, 8c
2, 3a, 3b
3
3a
3b
4
9.00 (d, 7.9)
8.16 (t, 7.6)
8.33 (t, 7.6)
8.44 (d, 8.1)
3a, 6, 7a
3b, 7
5
6
4, 7a
7
3b, 5
7a
8a
8c
10
revealed that the hydroxyl moiety and the para-substituted
benzene ring in 1 were not present in 2. However, analysis of
1
the H, 13C, COSY, HSQC, and HMBC NMR data confirmed
that the same fused ring system substructure observed in
compound 1 (rings A, B, and C) was also present in 2. There
was a consistent downfield shift for all of the protons present in
eudistidine B (2), while the carbon resonances for C-3, C-8c,
and C-10 were shifted downfield relative to those in 1, and C-
8a was shifted upfield. The most significant change was the
absence of the C-10 hemiaminal carbon resonance at δC 94.4
and the appearance of a resonance at δC 172.0. This revealed
the presence of an amide carbonyl functionality that was bound
to nitrogen N-9, and it allowed assignment of the structure of
eudistidine B (2). In contrast to eudistidine A (1), compound 2
showed no pH dependence in its UV absorption profile over a
pH range of 3−11 (Figure 2), with similar absorption maxima
in DMSO and water (327 and 445 nm in DMSO; 317 and 420
nm in H2O).
While the structures of eudistidines A (1) and B (2) were
assigned based on a comprehensive analysis and interpretation
of their NMR spectroscopic data, the proton-deficient and
heteroatom-rich nature of rings A, C, and D made definitive
proof of their structures difficult. Given the unprecedented
constellation of rings and heteroatoms in the eudistidines,
spectral comparisons with known compounds were of limited
value. We reasoned a concise total synthesis of compounds 1
and 2 would serve to both confirm the assigned structures and
provide sufficient quantities of these compounds to support
more detailed biological and biochemical evaluations. The
general retrosynthetic approach is outlined in Scheme 1. The
tautomeric structure in which the D ring is opened could
feasibly be a minor component of 1, and it is a readily apparent
synthetic precursor to eudistidine A. It was envisaged that this
tautomer could be formed from 4-(2-aminophenyl)pyrimidin-
2-amine (3) and 4-methoxy-phenylglyoxal (4) through a
sequential condensation, cyclization, and oxidation reaction
process.18−22 A related sequence with 3 and glyoxylic acid (5)
was also anticipated to provide 2.
and all four nitrogen atoms in 1, was used in the subsequent
step without further purification after verification of the
structure by LCMS and H NMR. The aryl glyoxal reagent 4,
1
predominantly as its hydrate form, was generated via a
Kornblum oxidation of 2-bromo-4′-methoxyacetophenone in
DMSO/water and was used immediately without further
purification.22 Efficient condensation of reagents 3 and 4,
followed by oxidative conversion to eudistidine A (1) was
ultimately accomplished through a one-pot sequence consisting
of dropwise addition of in situ generated aryl glyoxal 4 to a
DMSO solution of 3, mild heating (1 h, 65 °C) presumably to
promote condensation and cyclization, and then finally in situ
oxidation with one equivalent of I2 (1 h, 65 °C). Purification of
the reaction product by reversed-phase C18 HPLC provided
compound 1 in excellent yield (70%), which was identical by
MS and NMR with the natural product eudistidine A.
Significant optimization was required to develop the conditions
used in the final key step. For example, precondensation of the
aryl glyoxal 4 with intermediate 3 prior to I2-mediated oxidation
was required, as the simultaneous addition of 4 and iodine to 3
provided 1 in only trace quantities. Moreover, evaluation of a
series of different oxidants [I2, DDQ, MnO2, Fe(III)] in the
reaction sequence revealed that iodine alone provided both a
high yield of 1 and the most reproducible results.
Using a similar strategy, eudistidine B (2) was synthesized by
heating a solution of glyoxylic acid (5) and 4-(2-aminophenyl)-
pyrimidin-2-amine (3) at 60 °C for 18 h in DMF to provide the
desired product 2 in 38% yield. Condensation proceeded with
autoxidation and it is likely that ambient O2 is involved in this
process. However, the addition of exogenous oxidants (I2,
DDQ, or MnO2) significantly hindered the productive reaction
pathway. The spectroscopic data for the synthetic material were
1
In brief, the synthesis of 1 was accomplished as follows. The
key precursor, 4-(2-aminophenyl)pyrimidin-2-amine (3), was
synthesized in three steps from 1-(2-nitrophenyl)ethanone as
reported previously,21 but with a slight modification at the last
step. By using Pd/C to hydrogenate 4-(2-nitrophenyl)-
pyrimidin-2-amine, instead of HCl and iron as described in
the literature, compound 3 could readily be prepared on a
multigram scale. Intermediate 3, which provided rings A and B
in complete agreement (MS, H NMR, and 13C NMR) with
those obtained with the natural product, eudistidine B (2). The
syntheses of eudistidines A (1) and B (2) were both
accomplished in a concise, scalable, and economic manner
with quite acceptable overall yields. These efforts help to
confirm the assigned structures of the natural products and they
open the way for large-scale production and analogue
development.
D
DOI: 10.1021/jacs.5b02156
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
a
The p300 protein is an attractive molecular target for the
identification of potential therapeutic agents, and a variety of
small molecule natural products have recently been reported to
interact with p300 and inhibit its acetyl transferase activity.
Scheme 2. Synthesis of Eudistidines A (1) and B (2)
These include the prenylated benzophenone garcinol,27
a
rearranged steroidal derivative from a marine sponge,28 and a
plant flavonoid related to quercitin.29 Naturally occurring small
molecules have also been identified that can disrupt the
formation or stability of the p300/HIF-1α complex. For
example, the fungal metabolite chetomin, a complex
epidithiodiketopiperazine dimer derived from two serine and
two tryptophan residues, was the only compound out of
600 000 screened that effectively blocked p300/HIF-1α
binding.16 Studies suggest that the redox potential of chetomin
and other compounds with disulfides result in Zn2+ loss from
p300 with a resultant alteration in its global fold.16,30 Zinc is
critical for maintaining the tertiary structure of p300 in a form
that can complex with HIF-1α.14,15 A similar mechanism
involving zinc abstraction from p300 has also been proposed
for the p300/HIF-1α inhibitory activity of a series of quinones
and indandiones.31 Any nonselective targeting of Zn2+
coordinating groups could impact many different proteins
and result in a wide range of undesirable off target effects that
would limit the utility of these types of agents. However, recent
studies show that care must be taken when making assumptions
about the selectivity and mode of action for a broad class of
redox active compounds. A series of dimeric epidithiodiketo-
piperazines was designed and synthesized,32 and the lead
compound was shown to selectively inhibit the p300/HIF-1α
interaction. In cellular assays, it downregulated the expression
of hypoxia-responsive genes, and in a murine model it
significantly reduced in vivo tumor growth.33 This study served
as a proof of principle that validated small molecule targeting of
the p300/HIF-1 interaction to inhibit tumors. The actino-
bacterial metabolite novobiocin, an aminocoumarin glycoside,
disrupts the p300/HIF-1α interaction by directly binding to the
HIF-1α C-TAD domain.34 None of these natural products or
other previously reported p300/HIF-1α inhibitors are structur-
ally related to the heterocyclic eudistidine scaffold.35−37
a
Reagents and conditions: (i) DMSO, H2O, 65 °C, 3 h; I2, 60 °C, 1 h;
(ii) DMF, 60 °C, 18 h.
In a dose−response format of the screening assay, naturally
occurring eudistidine A (1) effectively blocked the binding of
soluble CH1 (p300) to immobilized C-TAD (HIF-1α) with an
IC50 of approximately 75 μM, while eudistidine B (2) showed
no inhibition at a high dose of 250 μM. Eudistidine A (1) was
initially isolated from the ascidian extract and characterized as
its free base. Alkaloids are generally more stable as salts so
several different salts were prepared by treatment of 1 with
HCl, TFA, or acetic acid. While the different salt forms of 1 had
a measurable impact on its observed activity in the assay, TFA
salts of the natural product and synthetically prepared
eudistidine A (1) had consistent and virtually identical
inhibitory activity (Supporting Information). In an effort to
evaluate a broader range of potential activities of interest,
compound 1 was screened against a commercial panel of 456
protein kinase enzymes and 32 bromodomain proteins
(DiscoverRx, San Diego, CA), and it had no effect against
any of these targets. These results suggest that inhibition of the
CH1/C-TAD interaction by 1 is not due to a nonspecific
propensity to bind to proteins, but is more specific for these
two binding partners. A degree of protein target selectivity is
consistent with the lack of general cytotoxic effects observed for
1 when tested at the standard concentration of 40 μM in the
NCI-60 cell anticancer screen.
Protein−protein interactions play fundamental roles in a
wide range of aberrant biological processes associated with
disease development and progression. However, these inter-
actions have often been viewed as “undruggable” due to the
large surface area and varied topography of the protein regions
that interact.23,24 Effective small molecule inhibitors of
protein−protein interactions are rare and difficult to design
de novo, so efforts to disrupt protein−protein binding for
potential therapeutic applications have often focused on small
proteins or peptide mimetics.25,26 Many of the small molecules
that can inhibit protein−protein interactions have poor
specificity or potencies in the millimolar range. With its ability
to block CH1/C-TAD binding at micromolar concentrations,
eudistidine A (1) represents an intriguing new structural
scaffold for developing small molecule inhibitors of p300/HIF-
1α interactions.
CONCLUSION
■
The rich chemical diversity of the NCI’s library of natural
product extracts was assessed in a high-throughput screening
assay for inhibitors of p300/HIF-1α binding. We have isolated
and identified two new compounds with a novel heterocyclic
scaffold from an extract of the marine ascidian Eudistoma sp.,
one of which showed activity in the screen. Prior studies of
ascidians in the genus Eudistoma led to the discovery of a
variety of compound classes including β-carbolines (eudisto-
mins, eudistomindins),38,39 cyclic peptides (eudistomides),40
macrolides (iejimalide),41 and pyrrolopyrimidines (rigidins),42
among others. Eudistidines A (1) and B (2) are unlike any
other Eudistoma metabolites, and they represent a new
structural class of alkaloids in which two pyrimidine rings and
an imidazole ring are fused to generate an unprecedented
tetracyclic core. This core ring system contains embedded
guanidine and amidine functionalities, and in the case of
eudistidine A (1) an additional hemiaminal group. The novel
heterocyclic architecture of eudistidine A, originally assigned
from analysis of its spectroscopic data, was confirmed by a
concise four-step synthesis that employed 4-(2-aminophenyl)-
pyrimidin-2-amine (3) and 2-(4-methoxyphenyl)-2-oxoacetal-
dehyde (4) as key precursors. Eudistidine B (2) was
E
DOI: 10.1021/jacs.5b02156
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
(2) Yang, Y.; Sun, M.; Wang, L.; Jiao, B. J. Cell. Biochem. 2013, 114,
967−974.
successfully synthesized using a similar approach that used
glyoxylic acid (5) in place of compound 4.
(3) Tsai, Y.-P.; Wu, K.-J. J. Biomed. Sci. 2012, 19, 102−108.
(4) Harris, A. L. Nat. Rev. Cancer 2002, 2, 38−47.
(5) Semenza, G. L. Nat. Rev. Cancer 2003, 3, 721−732.
(6) Semenza, G. L. Drug Discovery Today 2007, 12, 853−859.
(7) Wang, G. L.; Jiang, B. H.; Rue, E. A.; Semenza, G. L. Proc. Natl.
Acad. Sci. U. S. A. 1995, 92, 5510−5514.
Small molecule inhibitors of specific protein−protein
interactions are notoriously difficult to design and synthesize.
However, the structures of many natural product secondary
metabolites have apparently evolved to interact with specific
biopolymer targets. Natural products therefore provide a rich
source of chemical diversity for discovery of agents that can
modulate protein interactions. In the present case, eudistidine
A (1) was identified as an effective inhibitor of CH1/C-TAD
binding, which is the key protein−protein interaction required
to form the transcriptionally active p300/HIF-1α complex.
Eudistidine A (1) can be readily synthesized, which should
facilitate the generation of a family of analogues to help define
the structural requirements for activity and provide adequate
material to support additional biological studies. Since the
activated p300/HIF-1α complex is required for the tran-
scription of hypoxia-responsive genes in tumors, more extensive
evaluations of eudistidine A (1) and related compounds are
warranted. They may provide agents with potential anticancer
therapeutic applications or small molecule probes to further
define the roles of p300 and HIF-1α in tumor adaptations to
low oxygen environments.
(8) Huang, L. E.; Gu, J.; Schau, M.; Bunn, H. F. Proc. Natl. Acad. Sci.
U. S. A. 1998, 95, 7987−7992.
(9) Chen, J.; Li, Q. Epigenetics 2011, 6, 957−961.
(10) Bowers, E. M.; Yan, G.; Mukherjee, C.; Orry, A.; Wang, L.;
Holbert, M. A.; Crump, N. T.; Hazzalin, C. A.; Liszczak, G.; Yuan, H.;
Larocca, C.; Saldanha, S. A.; Abagyan, R.; Sun, Y.; Meyers, D. J.;
Marmorstein, R.; Mahadevan, L. C.; Alani, R. M.; Cole, P. A. Chem.
Biol. 2010, 17, 471−482.
(11) Grossman, S. R.; Deato, M. E.; Brignone, C.; Chan, H. M.;
Kung, A. L.; Tagami, H.; Nakatani, Y.; Livingston, D. M. Science 2003,
300, 342−344.
(12) Shi, D.; Pop, M. S.; Kulikov, R.; Love, I. M.; Kung, A. L.;
Grossman, S. R. Proc. Natl. Acad. Sci. U. S. A. 2009, 106, 16275−
16280.
(13) Gu, J.; Milligan, J.; Huang, L. E. J. Biol. Chem. 2001, 276, 3550−
3554.
(14) Freedman, S. J.; Sun, Z.-Y. J.; Poy, F.; Kung, A. L.; Livingston,
D. M.; Wagner, G.; Eck, M. T. Proc. Natl. Acad. Sci. U. S. A. 2002, 99,
5367−5372.
(15) Dial, R.; Sun, Z.-Y. J.; Freedman, S. J. Biochemistry 2003, 42,
9937−9945.
ASSOCIATED CONTENT
* Supporting Information
■
S
General experimental, isolation, synthesis, 1H NMR, 13C NMR,
COSY, HSQC, HMBC, and ROESY, together with HRESIMS
and UV−vis data for compounds 1 and 2. p300/HIF-1α testing
data for compound 1. This material is available free of charge
via the Internet at http://pubs.acs.org.
(16) Kung, A. L.; Zabludoff, S. D.; France, D. S.; Freedman, S. J.;
Tanner, E. A.; Vieira, A.; Cornell-Kennon, S.; Lee, J.; Wang, B.; Wang,
J.; Memmert, K.; Naegeli, H.-U.; Petersen, F.; Eck, M. J.; Bair, K. W.;
Wood, A. W.; Livingston, D. M. Cancer Cell 2004, 6, 33−43.
(17) McCloud, T. G. Molecules 2010, 15, 4526−4563.
(18) Gao, Q.; Wu, X.; Jia, F.; Liu, M.; Zhu, Y.; Cai, Q.; Wu, A. J. Org.
Chem. 2013, 78, 2792−2797.
AUTHOR INFORMATION
■
(19) Liu, S.; Chen, R.; Chen, H.; Deng, G.-J. Tetrahedron Lett. 2013,
54, 3838−3841.
Corresponding Authors
*gustafki@mail.nih.gov
*martin.schnermann@nih.gov
(20) Zhu, Y.-P.; Fei, Z.; Liu, M.-C.; Jia, F.-C.; Wu, A.-X. Org. Lett.
2013, 15, 378−381.
(21) Agarwal, P. K.; Sharma, S. K.; Sawant, D.; Kundu, B.
Tetrahedron 2009, 65, 1153−1161.
Present Address
#T.C.M.: Diagnostic Biomarkers and Technology Branch,
Cancer Diagnosis Program, DCTD, National Cancer Institute,
Bethesda, Maryland 20850, United States.
(22) Watanabe, K.; Uehara, F.; Hiki, S.; Yokoshima, S.; Usui, Y.;
Okuyama, M.; Shoda, A.; Aritomo, K.; Kohara, T.; Fukunaga, K.
Preparation of 2,3,6-trisubstituted-4-pyrimidones as tau protein kinase
1 inhibitors. PCT Int. Appl., 2004085408, 07 Oct 2004.
(23) Wells, J. A.; McClendon, C. L. Nature 2007, 450, 1001−1009.
(24) Surade, S.; Blundell, T. L. Chem. Biol. 2012, 19, 42−50.
(25) Bullock, B. N.; Jochim, A. L.; Arora, P. S. J. Am. Chem. Soc. 2011,
133, 14220−14223.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
We gratefully acknowledge D. Newman (NCI) for organizing
and documenting the collection, the Natural Products Support
Group at NCI-Frederick for extraction, and S. Tarasov and M.
Dyba (Biophysics Resource Core, Structural Biophysics
Laboratory, CCR) and H. Bokesch (MTL) for assistance
with high-resolution mass spectrometry. This research was
supported in part by the Intramural Research Program of the
NIH, National Cancer Institute, Center for Cancer Research.
This project was also funded in part with Federal funds from
the National Cancer Institute, National Institutes of Health,
under contract HHSN261200800001E. The content of this
publication does not necessarily reflect the views or policies of
the Department of Health and Human Services, nor does
mention of trade names, commercial products, or organizations
imply endorsement by the U.S. Government.
(26) Azzarito, V.; Long, K.; Murphy, N. S.; Wilson, A. J. Nat. Chem.
2013, 5, 161−173.
(27) Balasubramanyam, K.; Altaf, M.; Varier, R. A.; Swaminathan, V.;
Ravindran, A.; Sadhale, P. P.; Kundu, T. K. J. Biol. Chem. 2004, 279,
33716−33726.
(28) Gong, J.; Sun, P.; Jiang, N.; Riccio, R.; Lauro, G.; Bifulco, G.; Li,
T.-J.; Gerwick, W. H.; Zhang, W. Org. Lett. 2014, 16, 2224−2227.
(29) Holzhauser, S.; Freiwald, A.; Weise, C.; Multhaup, G.; Han, C.-
T.; Sauer, S. Angew. Chem., Int. Ed. 2013, 52, 5171−5174.
́
(30) Cook, K. M.; Hilton, S. T.; Mecinovic, J.; Motherwell, W. B.;
Figg, W. D.; Schofield, C. J. J. Biol. Chem. 2009, 284, 26831−26838.
(31) Jayatunga, M. K. P.; Thompson, S.; McKee, T. C.; Chan, M. C.;
Reece, K. M.; Hardy, A. P.; Sekirnik, R.; Seden, P. T.; Cook, K. M.;
McMahon, J. B.; Figg, W. D.; Schofield, C. J.; Hamilton, A. D. Eur. J.
Med. Chem. 2015, 94, 509−516.
(32) Block, K. M.; Wang, H.; Szabo,
́
L. Z.; Polaske, N. W.; Henchey,
́
C. F.; Makhoul, J.; Song, Z.;
L. K.; Dubey, R.; Kushal, S.; Laszlo,
́
REFERENCES
(1) Semenza, G. L. J. Clin. Invest. 2013, 123, 3664−3671.
Meuillet, E. J.; Olenyuk, B. Z. J. Am. Chem. Soc. 2009, 131, 18078−
■
18088.
F
DOI: 10.1021/jacs.5b02156
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
(33) Dubey, R.; Levin, M. D.; Szabo, L. Z.; Laszlo, C. F.; Kushal, S.;
Singh, J. B.; Oh, P.; Schnitzer, J. E.; Olenyuk, B. Z. J. Am. Chem. Soc.
2013, 135, 4537−4549.
(34) Wu, D.; Zhang, R.; Zhao, R.; Chen, G.; Cai, Y.; Jin, J. PLoS One
2013, 8, e62014.
(35) Kwon, H. S.; Kim, D.-R.; Yang, E. G.; Park, Y. K.; Ahn, H.-C.;
Min, S.-J.; Ahn, D.-R. Bioorg. Med. Chem. Lett. 2012, 22, 5249−5252.
(36) Yang, H.; Pinello, C. E.; Luo, J.; Li, D.; Wang, Y.; Zhao, L. Y.;
Jahn, S. C.; Saldanha, S. A.; Planck, J.; Geary, K. R.; Ma, H.; Law, B. K.;
Roush, W. R.; Hodder, P.; Liao, D. Mol. Cancer Ther. 2013, 12, 610−
620.
(37) Na, Y.-R.; Han, K.-C.; Park, H.; Yang, E. G. Biochem. Biophys.
Res. Commun. 2013, 434, 879−884.
(38) Rinehart, K. L., Jr.; Kobayashi, J.; Harbour, G. C.; Gilmore, J.;
Mascal, M.; Holt, T. G.; Shield, L. S.; Lafargue, F. J. Am. Chem. Soc.
1987, 109, 3378−3387.
(39) Kobayashi, J.; Cheng, J.-F.; Ohta, T.; Nozoe, S.; Ohizumi, Y.;
Sasaki, T. J. Org. Chem. 1990, 55, 3666−3670.
(40) Whitson, E. L.; Ratnayake, A. S.; Bugni, T. S.; Harper, M. K.;
Ireland, C. M. J. Org. Chem. 2009, 74, 1156−1162.
(41) Kobayashi, J.; Cheng, J.-F.; Ohta, T.; Nakamura, H.; Nozoe, S.;
Hirata, Y.; Ohizumi, Y.; Sasaki, T. J. Org. Chem. 1988, 53, 6147−6150.
(42) Kobayashi, J.; Cheng, J.-F.; Kikuchi, Y.; Ishibashi, M.;
Yamamura, S.; Ohizumi, Y.; Ohta, T.; Nozoe, S. Tetrahedron Lett.
1990, 31, 4617−4620.
G
DOI: 10.1021/jacs.5b02156
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX