Plectosphaeroic acids A, B, and C, Lndoleamine 2,3-dioxygenase inhibitors produced in culture by a marine isolate of the fungus plectosphaerella cucumerina

Article abstract of DOI:10.1021/ol900972j

Laboratory cultures of the fungus Plectosphaerella cucumerina obtained from marine sediments collected in Barkley Sound, British Columbia, yielded the novel alkaloids plectosphaeroic acids A (1) to C (3). The alkaloids 1-3 are inhibitors of indoleamine 2,3-dioxygenase (IDO).

Full text of DOI:10.1021/ol900972j

ORGANIC
LETTERS
2009
Vol. 11, No. 14
2996-2999
Plectosphaeroic Acids A, B, and C,
Indoleamine 2,3-Dioxygenase Inhibitors
Produced in Culture by a Marine Isolate
of the Fungus Plectosphaerella
cucumerina
Gavin Carr,† Wendy Tay,†,‡ Helen Bottriell,† Sarah K. Andersen,‡
A. Grant Mauk,*,‡ and Raymond J. Andersen*,†
Departments of Chemistry and Earth & Ocean Sciences, UniVersity of British Columbia,
VancouVer, B.C., Canada, V6T 1Z1, and Department of Biochemistry and Molecular
Biology, UniVersity of British Columbia, VancouVer, B.C., Canada, V6T 1Z3
randersn@interchange.ubc.ca; mauk@interchange.ubc.ca
Received May 2, 2009
ABSTRACT
Laboratory cultures of the fungus Plectosphaerella cucumerina obtained from marine sediments collected in Barkley Sound, British Columbia,
yielded the novel alkaloids plectosphaeroic acids A (1) to C (3). The alkaloids 1-3 are inhibitors of indoleamine 2,3-dioxygenase (IDO).
Indoleamine 2,3-dioxygenase (IDO) has emerged as a
promising molecular target for the development of a new
class of therapeutic agents for treating cancer that work by
modulating an extrinsic property of tumor cells.1 The interest
in IDO arises from its putative role in the ability of tumors
to escape the T-lymphocyte-based immune response, a
characteristic trait of human cancer that plays a central role
in the progesssion of the disease from small dormant
localized tumors to the often lethal state of metastasis. IDO
catalyzes the cleavage of the 2,3-indole bond of L-tryptophan,
converting it to N-formylkynurenine in the first and rate-
limiting step of the kynurenine catabolic pathway. IDO-
expressing cancer cells use this transformation to reduce
tryptophan concentrations in their microenvironments to
levels that prevent T-lymphocyte activation and proliferation.
Evidence also suggests that other catabolites in the kynure-
nine pathway are toxic to T-cells and further inhibit their
action. A growing body of clinical data shows that many
primary tumor cell lines obtained from patients overexpress
IDO, and this strongly correlates with a poor prognosis for
survival.2
† Departments of Chemistry and EOS.
‡ Department of Biochemistry and Molecular Biology.
(1) (a) Prendergast, G. C. Oncogene 2008, 27, 3889–3900. (b) Muller,
A. J.; Prendergast, G. C. Curr. Cancer Drug Tar. 2007, 7, 31–40. (c)
Muller, A. J.; Prendergast, G. C. Cancer Res. 2005, 65, 8065–8068. (d)
Muller, A. J.; Malachowski, W. P.; Prendergast, G. C. Expert. Opin. Ther.
Tar. 2005, 9, 831–849. (e) Uyttenhove, C.; Pilotte, L.; Theate, I.; Stroobant,
V.; Colau, D.; Parmentier, N.; Boon, T.; van den Eynde, B. Nat. Med. 2003,
9, 1269–1274.
Muller, Prendergast, and co-workers have provided com-
pelling evidence from murine cancer models for the promise
10.1021/ol900972j CCC: $40.75
Published on Web 06/19/2009
 2009 American Chemical Society

of IDO inhibitors as anticancer drugs that can marshal the
immune system to help regress established tumors. Their
experiments have been conducted with the three modestly
active IDO inhibitors N-methyltryptophan,3 5-bromobrassi-
nin,4 and menadione.5 Each of these compounds failed to
regress murine tumors when used as a monotherapy, but they
all showed impressive regression of established tumors when
used in combination with the antimitotic drug paclitaxel, that
was also ineffective when used alone. Careful control
experiments confirmed that in each case the target IDO was
required for the positive in vivo results.
Although N-methyltryptophan, 5-bromobrassinin, and me-
nadione have provided important proof-of-principle demon-
strations for the use of IDO inhibitors in cancer chemo-
therapy, none of them are potent enough to be realistic drug
development candidates. In an attempt to discover more
potent IDO inhibitors belonging to new structural classes,
we have screened a library of crude extracts of marine
invertebrates and laboratory cultures of microorganisms
obtained from marine habitats with an in vitro assay for IDO
inhibition that uses purified recombinant human IDO. This
screening program identified the marine natural products
annulins A and B6 (Ki ) 120 nM) and exiguamines A (Ki )
40 nM)7 and B8 as potent IDO inhibitors. Annulin B inspired
a medicinal chemistry program that generated a series of
synthetic IDO inhibitory pyranonaphthoquinones with low
nanomolar potency,5 and it stimulated the discovery in our
laboratories that menadione was an IDO inhibitor.9 Similarly,
exiguamine A was the starting point for the preparation of
synthetic tryptaminequinone analogues that are also submi-
cromolar inhibitors of IDO.8,10
extract identified the novel alkaloids plectosphaeroic acids
A (1), B (2), and C (3) as in vitro inhibitors of IDO. Details
of the isolation, structure elucidation, and biological activities
of 1-3 are presented below.
Production cultures of P. cucumerina were grown as lawns
on solid agar (medium: ISP4) at rt for 14 days. The mature
cultures were sliced into small squares containing both fungal
biomass and media, and these were soaked in EtOAc.
Concentration of the EtOAc extracts in vacuo gave a residue
that was dissolved in MeOH and then fractionated using
Sephadex LH20 chromatography eluting with MeOH. The
IDO inhibitory fractions were pooled and purified further
using reversed-phase HPLC (Supporting Information) to give
pure samples of plectosphaeroic acids A (1), B (2), and C
(3) along with the known alkaloid T988 A (4), which was
identified by comparison of its spectroscopic data with
literature values.12
Plectosphaeroic acid A (1) was obtained as an optically
active red/orange solid that gave an [M - H]- peak at
m/z 807.1534 in the negative ion HRESIMS appropriate
for a molecular formula of C39H32N6O10S2 (calcd for
C39H31N6O10S2 807.1543) requiring 27 sites of unsaturation.
Thirty nine well-resolved carbon resonances could be identi-
fied in the 13C NMR spectrum of 1, and the HSQC spectrum
showed that there were 25 hydrogen atoms attached to carbon
(3 × CH3; 1 × CH2; 14 × CH). The LRESIMS recorded in
CD3OD gave an [M - D]- peak at m/z 813 and an [M - D
+ 2Na]+ ion at m/z 859 confirming that there were seven
exchangeable hydrogen atoms in 1, which together with the
25 hydrogens attached to carbon accounted for the 32
hydrogen atoms present in the molecular formula indicated
by the HRESIMS data.
(2) (a) Pan, K.; Wang, H.; Chen, M.; Zhang, H.; Weng, D.; Zhou, J.;
Huang, W.; Li, J.; Song, H.; Xia, J. J. Cancer Res. Clin. Oncol. 2008, 134,
1247–1253. (b) Okamoto, A.; Nikaido, T.; Ochiai, K.; Takakura, S.; Saito,
M.; Aoki, Y.; Ishii, N.; Yanaihara, N.; Yamada, K.; Takikawa, O.;
Kawaguchi, R.; Isonishi, S.; Tanaka, T.; Urashima, M. Clin. Cancer Res.
2005, 11, 6030–6039. (c) Brandacher, G.; Perathoner, A.; Ladurner, R.;
Schneeberger, S.; Obrist, P.; Winkler, C.; Werner, E. R.; Werner-Felmayer,
G.; Weiss, H. G.; Gobel, G.; Margreiter, R.; Konigsrainer, A.; Fuchs, D.;
Amberger, A. Clin. Cancer Res. 2006, 12, 1144–1151.
(3) Muller, A. J.; DuHadaway, J. B.; Donover, P. S.; Sutanto-Ward, E.;
Prendergast, G. C. Nat. Med. 2005, 11, 312–319.
(4) Banerjee, T.; DuHadaway, J. B.; Gaspari, P.; Sutanto-Ward, E.;
Munn, D. H.; Mellor, A. L.; Malachowski, W. P.; Prendergast, G. C.; Muller,
A. J. Oncogene 2008, 27, 2851–2857.
(5) Kumar, S.; Malachowski, W. P.; DuHadaway, J. B.; LaLonde, J. M.;
Carroll, P. J.; Jaller, D.; Metz, R.; Prendergast, G. C.; Muller, A. J. J. Med.
Chem. 2008, 51, 1706–1718.
(6) Pereira, A.; Vottero, E.; Roberge, M.; Mauk, A. G.; Andersen, R. J.
J. Nat. Prod. 2006, 69, 1496–1499.
(7) Brastianos, H. C.; Vottero, E.; Patrick, B. O.; Van Soest, R.;
Matainaho, T.; Mauk, A. G.; Andersen, R. J. J. Am. Chem. Soc. 2006, 128,
16046–16047.
(8) Volgraf, M.; Lumb, J.-P.; Brastianos, H. C.; Carr, G.; Chung,
M. K. W.; Mu¨nzel, M.; Mauk, A. G.; Andersen, R. J.; Trauner, D. Nat.
Chem. Biol. 2008, 4, 535–537.
(9) Andersen, R. J.; Pereira, A.; Huang, X.-H.; Mauk, G.; Vottero, E.;
Roberge, M.; Balgi, A. PCT Int. Appl. 2006, WO 2006005185.
(10) Carr, G.; Chung, M. K. W.; Mauk, A. G.; Andersen, R. J. J. Med.
Chem. 2008, 51, 2634–2637.
A third crude extract that showed promising IDO inhibition
in the in vitro screen came from laboratory cultures of the
fungus Plectosphaerella cucumerina11 obtained from marine
sediments collected at -100 m depth in Barkley Sound,
British Columbia. Bioassay guided fractionation of the crude
(11) Plectosphaerella cucumerina was identified by analysis of 26S
rDNA. It has been reported to have ultility as a biological control agent
against potato cyst nematodes. Atkins, S. D.; Clark, I. M.; Sosnowska, D.;
Hirsch, P. R.; Kerry, B. R. Appl. EnViron. Microbiol. 2003, 69, 4788–4793.
(12) Feng, Y.; Blunt, J. W.; Cole, A. L. J.; Munro, M. H. G. J. Nat.
Prod. 2004, 67, 2090–2092.
Org. Lett., Vol. 11, No. 14, 2009
2997

Analysis of the COSY, HSQC, and HMBC data recorded
for plectosphaeroic acid A (1) identified two fragments A
and B (Figure 1) that are also present in the co-occurring
(Table 1, Supporting Information). An HMBC correlation
between the N-methyl resonance at δ 2.94 and a carbon
resonance at δ 164.3 revealed the C-1 diketopiperazine
carbonyl, and the C-4 carbonyl resonance at δ 162.4 was
identified by an HMBC correlation from the H-14 resonance
at δ 4.02. The Me-16 thiomethyl resonance at δ 2.14 showed
an HMBC correlation to the C-12 carbon resonance at δ 74.1,
and both of the Me-15 thiomethyl and H-14 resonances at δ
1.89 and 3.56, respectively, showed HMBC correlations to
the C-3 resonance at δ 69.4.
A set of HMBC correlations from the H-5a methine
resonance at δ 6.78 to the aromatic carbon resonances at δ
149.1 (C-6a) and 133.3 (C-10a), to the C-4 carbonyl
resonance at δ 162.4, and to the C-10b and C-12 quaternary
carbon resonances at δ 58.5 and 74.1, respectively, and
additional HMBC correlations from the H-10 resonance at
δ 7.57 to the C-10b resonanace at δ 58.5, and from the H-11
resonanance at δ 5.45 to the C-5a resonance at δ 86.2, were
all in agreement with the proposed tetracyclic substructure
B containing phenyl, carbinol methine, and diketopiperazine
moieties. The 13C NMR assignments for the nonaromatic
carbons in substructure B in 1 (Figure 1) were again in very
good agreement with the literature assignments to their
counterparts in T988 A (4) and B.12
Figure 1. 2D NMR correlations for plectosphaeroic acid A (1).
Subtracting the atoms accounted for by fragments A and
B (C25H25N4O4S2) from the molecular formula of 1 showed
that the remaining fragment had to account for C14H7N2O6,
12 sites of unsaturation, and four exchangeable hydrogens.
Resonances at δ 167.2 (C-14′′) and 168.9 (C-13′′) in the
13C NMR spectrum were tentatively assigned to carboxylic
acid functionalities. This assignment was confirmed by
reacting 1 with methyl iodide and base to produce the
dimethyl ester 5. A carbon resonance at δ 178.1 (C-3′′) had
a chemcial shift appropriate for a quinone carbonyl, but since
there was not a second carbon with a similar shift, we
assumed this was part of an iminoquinone. The COSY
spectrum identified a pair of ortho protons (δ 7.68, d, J )
8.8 Hz, H-8′′; 7.72, d, J ) 8.6 Hz, H-7′′), and the HSQC
data showed that there was one additional proton (δ 6.67, s,
H-4′′) attached to carbon (δ 105.5, C-4′′).
metabolite T988 A (4).12 COSY correlations revealed the two
isolated 1H spin systems (H-4′ to H-7′ and H-2′/NH-1′) of the
C-3′ substituted indole fragment A, and HSQC/HMBC cor-
relations confirmed the carbon assignments. The carbon chemi-
cal shifts assigned to fragment A in 1 were nearly identical to
the corresponding carbon chemical shifts assigned to this
fragment in T988 A (4) (Figure 1 and Supporting Informa-
tion).12
The presence of fragment B in 1 was indicated by the
1
existence of H and 13C NMR resonances in the 1D spectra
recorded for 1 that showed chemical shifts and scalar coupling
patterns in the COSY and HMBC data nearly identical with
those observed for the corresponding resonances assigned to
the same fragment in 4.12 A series of COSY correlations
identified the resonances in the H-7 to H-10 spin system at δ
6.00 (d, J ) 8.0 Hz, H-7), 6.91 (t, J ) 7.6 Hz, H-8), 6.57 (t,
J) 7.6 Hz, H-9), and 7.57 (d, J ) 7.5 Hz, H-10). A singlet
resonance at δ 5.45 that showed an HSQC correlation to a
carbon resonance at δ 78.6 was assigned to the H-11 carbinol
Plectosphaeroic acid A (1) had a UV λmax at 428 nm, which
is similar to that reported for phenoxazinone dyes.13 Thus,
it appeared that the final fragment of plectosphaeroic acid
A (1) might be a substituted cinnabarinic acid moiety, which
would account for all of the remaining atoms in the structure,
the functionality identified from the NMR data, and the UV
1
methine proton. A second H singlet at δ 6.78, correlated in
the HSQC to a carbon resonance at δ 86.2, was assigned to the
H-5a proton. Methyl singlets at δ 1.89 (13C δ 12.9), 2.14 (13C
δ 15.7), and 2.94 (13C δ 27.8) in the 1H NMR spectrum could
be assigned to two thiomethyls (Me-15 and Me-16) and an
N-methyl (Me-13), respectively.
Three bond HMBC correlations observed between H-10
(δ 7.57) and a carbon resonance at δ 149.1 (C-6a) and
between H-7 (δ 6.00) and a carbon resonance at δ 133.3
(H-10a), along with the HSQC correlations from the aromatic
protons H-7 to H-10 to their attached carbons, assigned all
of the carbon resonances in the C-6a to C10a aromatic ring.
There was very close agreement between these assignments
in 1 and the assignments for the corresponding carbons in 4
λmax. To confirm this hypothesis, cinnabarinic acid (6) was
synthesized from 3-hydroxy anthranilic acid following
literature procedures,14 and its carbon and proton NMR
spectra were assigned as shown in Figure 1. There was
excellent agreement between the 1H and 13C NMR assigments
for fragment C in plectosphaeroic acid A (1) and the
synthetic cinnabarinic acid (Figure 1). It was also apparent
from these assignments that fragment C had to be linked to
(13) Giurg, M.; Piekielska, K.; Gebala, M.; Ditkowski, B.; Wolanski,
M.; Peczynska-Czoch, W.; Mlochowski, J. Synth. Commun. 2007, 37, 1779–
1789.
(14) Horva´th, T.; Kaizer, J.; Speier, G. J. Mol. Catal. A: Chem. 2004,
215, 9–15.
2998
Org. Lett., Vol. 11, No. 14, 2009

fragments A or B via C-9′′ since the resonance corresponding
diketopiperazine ring. Comparison of the CD spectra re-
corded for plectosphaeroic acids A (1), B (2), C (3), and
T988 A (4) (Supporting Information) with the literature
values for the structurally related leptosins15 revealed that
compounds 1-4 have the 3S, 5aR, 10bR, 11S, and 12S
absolute configurations shown.
1
to H-9′′ in the H NMR spectrum of cinnabarinic acid (6)
(δ 7.96), which is the most deshielded of the proton
resonances and shows an HMBC correlation to the carbon
of the carboxylic acid attached to C-10′′, was missing in the
1H NMR spectrum of 1. Synthetic cinnabarinic acid had UV
λmaxs at 231 and 429 nm.
The remaining unsatisfied valences in fragments A, B, and
C were at N-6, C-10b, C-3′ and C-9′′. ROESY correlations
between H-4′ (δ 8.05) and both H-10 (δ 7.57) and H-11 (δ
5.45) were consistent with a bond between C-10b and C-3′
as in T988 A (4), and a ROESY correlation between H-7 (δ
6.00) and H-8′′ (δ 7.68) identified a bond between N-6 and
C-9′′, completing the consitution of 1 (Figure 2). The
The plectospaeroic acids represent a new family of
complex fungal alkaloids. Biogenetically they appear to be
derived from 4 equiv of tryptophan and 1 equiv of either
serine or alanine. Remarkably, the tryptophan residues have
been processed in three distinctly different ways to produce
the final structure. The indole ring in fragment A presumably
arises from a tryptophan residue where the side chain has
been removed, possibly via a Grob fragmentation of a
diketopiperazine similar to substructure B.16 Fragment B
contains an intact tryptophan residue that has formed normal
peptide bonds with either serine or alanine to produce the
core of the modified diketopiperazine. The phenoxazinone
heterocycle found in fragment C is known to arise from
oxidative coupling of two 3-hydroxyanthranilic acid units17
that can in turn arise from catabolic degradation of tryptophan
via the kynurenine pathway.18
Plectosphaeroic acids A (1), B (2), and C (3) all inhibited
purified, recombinant human IDO assayed in vitro by the
method of Takikawa et al.19 and exhibited identical IC50s of
≈2 µM, while T988 A (4) was completely inactive. Synthetic
cinnabarinic acid (6) was also active in the assay with an
IC50 of ≈2 µM indicating that some portion of the phenox-
azinone fragment C represents a new IDO inhibitory phar-
macophore. The discovery of the plectosphaeroic acids
further highlights the ability of marine isolates of apparently
well-known microorganisms to produce novel bioactive
secondary metabolites that are potential drug leads.20
Figure 2. ROESY correlations for plectosphaeroic acid A (1).
observation of a ROESY correlation between H-5a (δ 6.78)
and H-2′ (δ 7.34) showed that H-5a and the indole ring were
cis to each other. ROESY correlations between H-11 (δ 5.45)
and both Me-16 (δ 2.14) and H-10 (δ 7.57) required that
H-11 and Me-16 were cis to each other and trans to H-5a as
shown (Figure 2).
Acknowledgment. Financial support was provided by
operating grants from NSERC (RJA) and the Canadian
Cancer Society (RJA and AGM), an infrastructure grant from
the Michael Smith Foundation for Health Research, a UBC
University Graduate Fellowship (GC), and a Canada Re-
search Chair (AGM).
Plectosphaeroic acid B (2) was isolated as an optically
active red/orange solid that gave an [M - H]- peak at m/z
791.1599 in the negative ion HRESIMS appropriate for a
molecular formula of C39H32N6O9S2 (calcd for C39H31N6O9S2
791.1594) that differed from the formula for 1 by loss of an
oxygen atom. Routine analysis of the NMR data for 2
(Supporting Information) showed that it was missing the
alcohol functionality found at C-14 in 1.
Plectosphaeroic acid C (3) was isolated as an optically
active red/orange solid that gave an [M - H]- peak at
m/z 809.0786 in the negative ion HRESIMS appropriate
for a molecular formula of C37H26N6O10S3 (calcd for
C37H25N6O10S3 809.0794) that differed from the formula for
1 by loss of C2H6 and addition of one sulfur atom. The NMR
data obtained for 3 were missing the 1H and 13C resonances
assigned to the C-15 and C-16 thiomethyl groups in the NMR
data for 1 but were in all other ways nearly identical with
the data for 1. Therefore, plectosphaeroic acid C was assigned
the structure 3, containing a trisulfide bridge across the
Supporting Information Available: Experimental details,
tables of NMR assignments, 1D and 2D NMR spectra, and
CD spectra. This material is available free of charge via the
Internet at http://pubs.acs.org.
OL900972J
(15) Takahashi, C.; Numata, A.; Ito, Y.; Matsumura, E.; Araki, H.; Iwaki,
H.; Kushida, K. J. Chem. Soc., Perkin Trans 1 1994, 1859–1864.
(16) Minato, H.; Matsumoto, M.; Katayama, T. J. Chem. Soc., Perkin
Trans 1 1973, 1819–1825.
(17) Bruyneel, F.; Enaud, E.; Billottet, L.; Vanhulle, S.; Marchand-
Brynaert, J. Eur. J. Org. Chem. 2008, 72–79.
(18) Lima, W. C.; Varani, A. M.; Menck, C. F. M. Mol. Biol. EVol.
2009, 26, 399–406.
(19) Takikawa, O.; Kuroiwa, T.; Yamazaki, F.; Kido, R. J. Biol. Chem.
1988, 263, 2041–2048.
(20) For other examples, see: (a) Barsby, T.; Warabi, K.; Sørensen, D.;
Zimmerman, W. T.; Kelly, M. T.; Andersen, R. J. J. Org. Chem. 2006, 71,
6031–6037. (b) Gu, W.; Cueto, M.; Jensen, P. R.; Fenical, W.; Silverman,
R. B. Tetrahedron 2007, 63, 6535–6541. (c) Kobayashi, M.; Uehara, H.;
Matsunami, K.; Aoki, S.; Kitagawa, I. Tetrahedron Lett. 1993, 34, 7925–
7928. (d) Amagata, T.; Rath, C.; Rigot, J. F.; Tarlov, N.; Tenney, K.;
Valeriote, F. A.; Crews, P. J. Med. Chem. 2003, 46, 4342–4350.
Org. Lett., Vol. 11, No. 14, 2009
2999