Gliotoxin analogues from a marine-derived fungus, penicillium sp., and their cytotoxic and histone methyltransferase inhibitory activities

Article abstract of DOI:10.1021/np200740e

Seven gliotoxin-related compounds were isolated from the fungus Penicillium sp. strain JMF034, obtained from deep sea sediments of Suruga Bay, Japan. These included two new metabolites, bis(dethio)-10a-methylthio-3a-deoxy-3,3a- didehydrogliotoxin (1) and 6-deoxy-5a,6-didehydrogliotoxin (2), and five known metabolites (3-7). The structures of the new compounds were elucidated by analysis of spectroscopic data and the application of the modified Moshers analysis. All of the compounds exhibited cytotoxic activity, whereas compounds containing a disulfide bond showed potent inhibitory activity against histone methyltransferase (HMT) G9a. None of them inhibited HMT SET7/9.

Full text of DOI:10.1021/np200740e

Note
pubs.acs.org/jnp
Gliotoxin Analogues from a Marine-Derived Fungus, Penicillium sp.,
and Their Cytotoxic and Histone Methyltransferase Inhibitory
Activities
Yi Sun,† Kentaro Takada,† Yasushi Takemoto,‡ Minoru Yoshida,‡ Yuichi Nogi,§ Shigeru Okada,†
and Shigeki Matsunaga*,†
†Laboratory of Aquatic Natural Products Chemistry, Graduate School of Agricultural and Life Sciences, The University of Tokyo,
Bunkyo-ku, Tokyo, 113-8657, Japan
‡Chemical Genomics Research Group, RIKEN Advanced Science Institute, Saitama, 351-0198, Japan
§Japan Agency for Marine-Earth Science and Technology (JAMSTEC), Natsushima, Yokosuka, Kanagawa, 237-0061, Japan
S
* Supporting Information
ABSTRACT: Seven gliotoxin-related compounds were isolated from the fungus
Penicillium sp. strain JMF034, obtained from deep sea sediments of Suruga Bay,
Japan. These included two new metabolites, bis(dethio)-10a-methylthio-3a-deoxy-
3,3a-didehydrogliotoxin (1) and 6-deoxy-5a,6-didehydrogliotoxin (2), and five
known metabolites (3−7). The structures of the new compounds were elucidated
by analysis of spectroscopic data and the application of the modified Mosher’s
analysis. All of the compounds exhibited cytotoxic activity, whereas compounds
containing a disulfide bond showed potent inhibitory activity against histone
methyltransferase (HMT) G9a. None of them inhibited HMT SET7/9.
fter intensive investigation of secondary metabolites of
A
marine macroorganisms for almost half a century,1 some
marine natural product chemists concluded that new structural
entities from these organisms were nearly exhausted. Therefore,
they turned their attention to marine microorganisms as an
untapped source of secondary metabolites.2,3 With this trend in
mind we have been searching for cytotoxic metabolites
produced by marine-derived fungi. We found that a Penicillium
sp. (strain JMF034) isolated from deep-sea sediments collected
in Suruga Bay exhibited potent cytotoxic activity. From the
culture medium we isolated a series of metabolites related to
gliotoxin4 and gliotoxin itself as the active constituents, among
which two metabolites are new. Gliotoxin is a representative
member of the epipolythiodioxopiperazine (ETP) class of
fungal metabolites.5 Due to its potent cytotoxicity toward
cancer cell lines, this compound has been considered as a lead
for anticancer agents.5 Recently, dimeric ETPs were shown to
inhibit histone methyltransferase (HMT).6 These observations
prompted us to examine the HMT-inhibitory activity of the
metabolites, some of which showed significant activity. Here we
describe structure elucidation of the new compounds.
The molecular formula of compound 1 was determined to be
C14H16N2O3S on the basis of the HRESIMS data. Analysis of
1
the H NMR spectrum (Table 1) in conjunction with the
HSQC data revealed two singlet methyls, a pair of non-
equivalent methylene protons, two oxygen- or nitrogen-
substituted methines, five olefinic protons, two of which were
a pair of exomethylene protons, and an exchangeable proton.
Even though the number of protons was small, interpretation of
the COSY spectrum was complex due to the overlap of H-5a
and H-6, the absence of coupling between H-6 and H-7, and
the observation of long-range couplings between H-5a and H-9
The culture medium of Penicillium sp. JMF034 was extracted
with EtOAc, and the combined EtOAc layers were fractionated
by solvent partitioning, gel permeation, and reversed-phase
column chromatographies followed by ODS-HPLC to yield
compounds 1−7. The known compounds were identified as
bis(dethio)bis(methylthio)gliotoxin (3),7 bis(dethio)bis-
(methylthio)-5a,6-didehydrogliotoxin (4),8 5a,6-didehydroglio-
toxin (5),9 gliotoxin (6),10 and gliotoxin G (7),8 on the basis of
Received: September 14, 2011
Published: December 12, 2011
1
ESIMS and H NMR data.
© 2011 American Chemical Society and
American Society of Pharmacognosy and
American Society of Pharmacognosy
111
dx.doi.org/10.1021/np200740e | J. Nat. Prod. 2012, 75, 111−114

Journal of Natural Products
Note
1
Table 1. H and 13C NMR Data for Compounds 1 and 2
HMBC data. HMBC correlations from H-10″ to C-1 and from
N-Me to C-1 implied the linkage of C-1 and C-10a, revealing
the other linkage of C-4 and N-5. These final linkages are
consistent with the established structures of gliotoxin-type
metabolites and with recent biosynthetic evidence11 for this
class of compounds.
(600 MHz, CDCl3)
1
2
a
a
δC,
δH (J in
Hz)
δC,
δH (J in
Hz)
position
type
HMBC
type
HMBC
1
164.5, C
138.1, C
166.1, C
77.0, C
The relative configuration of 1 was determined by
interpretation of the NOESY data (Figure 2). In the NOESY
3
3a
105.2,
CH2
HZ: 5.90, d 3, 4
(1.8)
61.0,
CH2
4.48, d
(12.7)
3, 4
3, 4
HE:5.09, d 3, 4
(1.8)
4.32, d
(12.7)
4
161.5, C
161.5, C
138.2, C
5a
69.2,
CH
4.90, d
(13.0)
6, 7
6
7
74.1,
CH
4.89, d
(13.0)
5a, 7,
9a
115.8,
CH
7.95, d
(8.3)
8, 9a
5a, 9
130.8,
CH
5.79, d
(9.7)
5a, 9
6, 9a
5a, 7
125.1,
CH
7.36, dd
(8.3,
7.9)
Figure 2. Key NOESY correlations of 1.
8
9
123.2,
CH
5.91, m
126.2,
CH
7.24, dd
(7.9,
7.5)
6, 9a
5a, 7
spectrum, 10a-S-Me was correlated to H-6 and H-10′ (δ 3.00),
whereas H-5a was correlated to H-10″ (δ 3.07) and OH-6.
Additionally, a coupling constant of 13.0 Hz between H-5a and
H-6 indicated that the two protons were antiperiplanar.
Therefore, H-6 and 10a-S-Me were on the same face of the
tetrahydroindole ring system, whereas H-5a was on the other
face.
The absolute configuration of 1 was determined by the
modified Mosher’s method.12 Esterification of 1 with (R)- and
(S)-MTPA-Cl afforded the (S)- and (R)-MTPA esters, 1a and
1b, respectively. The distribution of Δδ values indicated the 6S
configuration (Figure 3), the same configuration as is known
120.5,
CH
5.95, br s
128.8,
CH
7.37, d
(7.5)
9a
10
132.0, C
128.1, C
39.5,
CH2
3.07, d
(16.2)
1, 9, 9a, 36.7,
4.33, d
(18.4)
1, 9
10a
CH2
3.00, d
(16.2)
5a, 9,
9a,
10a
3.35, d
(18.4)
5a,
10a,
9
10a
73.5, C
74.0, C
S-Me
14.5,
CH3
2.13, s
3.29, s
6.07, s
10a
27.6,
CH3
3.27, s
1, 3
N-Me 30.5,
CH3
1, 3
6-OH
5a, 6, 7
a
Carbon chemical shifts were assigned from HSQC and HMBC
spectra.
and between H-5a and H-10″.11 Two partial structures (units a
and b, Figure 1) were assigned by interpretation of the 2D
Figure 1. Partial structures of 1.
NMR data (Table 1). In unit a, the connectivity of the three
olefinic protons (H-7, H-8, and H-9) was unambiguously
assigned from the COSY data. HMBC cross-peaks from OH-6
to C-5a, C-6, and C-7 showed that C-7 was connected to C-6,
which was linked to C-5a. HMBC cross-peaks from H2-10 to C-
9, C-5a, C-9a, and C-10a indicated that C-9 was connected to
C-9a. HMBC cross-peaks from H2-10 to C-10a and from the S-
Me to C-10a showed the connection between C-10 and C-10a.
NMR data of this portion matched well with those of
gliotoxin,10 suggesting that C-5a was attached to a nitrogen
atom, which was connected to C-10a. In unit b, an
exomethylene (C-3 and C-3a) was flanked by an amide
carbonyl (C-4) and the N-methyl group on the basis of the
Figure 3. Values of Δ δS − δR of the MTPA esters of 1.
for gliotoxin.10 Therefore, compound 1 is 5aS,6S,10aR-
bis(dethio)-10a-methylthio-3a-deoxy-3,3a-didehydrogliotoxin.
Compound 2 had the molecular formula C13H12N2O3S2 as
1
determined by HR ESIMS and NMR data (Table 1). The H
NMR data in conjunction with the HSQC data suggested the
presence of four aromatic protons, a pair of nonequivalent
methylene protons, an N-methyl, and a hydroxymethyl group.
The COSY data suggested that the benzene ring was ortho
disubstituted, whereas the remaining signals coincided well with
those of the corresponding parts of dehydrogliotoxin (5).9
112
dx.doi.org/10.1021/np200740e | J. Nat. Prod. 2012, 75, 111−114

Journal of Natural Products
Note
Table 2. Inhibitory Activity against HMT G9a and Cytotoxicity on P388 Cells of Compounds 1−7 (IC50)
1
2
3
4
5
6
7
HMT G9a (μM)
cytotoxicity (μM)
>100
3.4
55
0.058
>100
0.11
58
2.6
0.056
6.4
0.024
2.1
0.020
0.11
(1:1), followed by ODS-HPLC using gradient elution from 50% to
80% aqueous MeOH to yield five fractions (A−E). Fraction B was
then purified by semipreparative ODS-HPLC using a gradient of
MeCN in H2O (30−50%) to afford 3 (5.4 mg), 4 (2.6 mg), 5 (2.8
mg), and 6 (4.2 mg). Fractions C and D were combined and purified
by HPLC using gradient elution from 35% to 55% MeCN in H2O to
yield 1 (1.5 mg), 2 (2.6 mg), and 7 (3.0 mg).
Therefore, 2 was suggested to be the deoxy derivative of 5,
which was supported by the HMBC data (Table 1).13 Both
compound 1 and the co-isolated gliotoxin 614 have a 10aR
configuration, with the configuration of 6 being assigned by
comparison of specific rotation data. On the basis of biogenetic
considerations we also presume a 10aR configuration for 2.
The cytotoxic activities of compounds 1−7 were examined
against P388 murine leukemia cells. Gliotoxin (6) and gliotoxin
G (7) exhibited the most potent activity, whereas compounds
2−5 also showed significant activity (Table 2). However,
compound 1 had only marginal activity. We also examined the
inhibitory activity of 1−7 against HMT G9a and HMT Set7/9
(lysine-specific histone methyltransferase for lysine 4 in histone
H3). As expected from the previous report,4 compounds with a
disulfide or tetrasulfide bond (5, 6, and 7) exhibited potent
inhibitory activity. The weaker activity observed for 2, which
also has a disulfide bond, suggested that the C-6 hydroxy group
interfered with the G9a inhibitory activity. None of 1−7
inhibited HMT Set7/9 at 100 μM.
Bis(dethio)-10a-methylthio-3a-deoxy-3,3a-didehydrogliotoxin
(1): white solid; [α]25D −4.6 (c 0.03, MeOH); UV (MeOH) λmax (log
1
ε) 205 (4.20), 265 (4.07) nm; H NMR (CDCl3) and 13C NMR
(CDCl3), see Table 1; HRESIMS m/z 315.0759 [M + Na]+ (calcd for
C14H16N2O3S, 315.0779).
6-Deoxy-5a,6-didehydrogliotoxin (2): colorless oil; [α]25 −20 (c
D
0.05, MeOH); UV (MeOH) λmax (log ε) 206 (4.26), 265 (3.91) nm;
1H NMR (CDCl3) and 13C NMR (CDCl3), see Table 1; HRESIMS
m/z 331.0315 [M + Na]+ (calcd for C13H12N2O3S2, 331.0338).
Gliotoxin (6): [α]25 −440 (c 0.11, MeOH). The sign and
D
magnitude of the specific rotation value were comparable to those
of gliotoxin in the literature,14 which was [α]25D −290 (c 0.08, EtOH).
Preparation of MTPA Esters of 1a and 1b. Compound 1 (100 μg
for each) was reacted with either R-(−)- or S-(+)-MTPA Cl (5 μL) in
50 μL of CH2Cl2 and 50 μL of pyridine for 2 h. The reaction mixture
was diluted with H2O and extracted with CH2Cl2 three times. The
organic layers were combined and separated by ODS-HPLC (C18-
stationary phase, 10 × 50 mm; 40−60% MeCN in H2O) to afford the
S-(−)- or R-(+)-MTPA esters 1a and 1b.
EXPERIMENTAL SECTION
■
General Experimental Procedures. Optical rotations were
measured on a JASCO DIP-1000 digital polarimeter. UV spectra
were recorded on a Shimadzu Biospec 1600. NMR spectra were
recorded on a JEOL alpha 600 NMR spectrometer at 300 K. Chemical
shifts were referenced to solvent peaks: δH 7.27 and δC 77.2 for
CHCl3. ESI mass spectra were measured on a JEOL JMS-T 100LC.
HPLC was carried out on a Shimadzu LC 20AT with a SCL-10Avp
controller and a SPD-10Avp detector.
(S)-MTPA Ester of 1 (1a): 1H NMR (CDCl3) δ 6.32 (H-6), 5.99
(H-9), 5.94 (H-8), 5.67 (HZ-3a), 5.43 (H-7), 5.36 (H-5a), 4.94 (HE-
3a), 3.28 (N-CH3), 3.02 (H-10), 2.23 (S-CH3); ESIMS m/z 531 [M +
Na]+.
(R)-MTPA Ester of 1 (1b): 1H NMR (CDCl3) δ 6.41 (H-6), 6.04
(H-8), 6.03 (H-9), 5.69 (H-7), 5.50 (HZ-3a), 5.34 (H-5a), 4.91 (HE-
3a), 3.26 (N-CH3), 3.03 (H-10), 2.21 (S-CH3); ESIMS m/z 531 [M +
Na]+.
Fungal Material. Deep-sea sediments were collected by the
unmanned ROV KAIKO system from Fujikawa, Suruga-Bay, Japan, at
a depth of 1151 m, in July 1996. The sediment sample was stored in a
sterilized sampler and frozen with liquid nitrogen. Then the sample
was transported to the laboratory, where it was kept frozen until
processed. The Penicillium sp. JMF034 strain was isolated from this
sample. To investigate the taxonomic position of the strain, the 28S
rDNA-D1/D2 gene was amplified using the PCR method with primers
NL1 and NL4.15 The PCR product was sequenced with the
dideoxynucleotide chain-termination method, using a BigDye
Terminator v3.1 kit (Applied Biosystems) and ABI PRISM 3130xl
genetic analyzer system (Applied Biosystems). The 28S rDNA-D1/D2
gene sequence of the isolated fungus (DDBJ accession no. AB684325)
was compared with other sequences in the public database using the
BLAST program. Strain JMF034 showed the highest similarities to
strain Penicillium angulare NRRL28157T (sequence identity 99.1%), P.
adametzioides NRRL3405T (98.9%), and P. brocae NRRL31472T
(sequence identity 98.8%). Therefore JMF034 is included in the
Penicillium genus..
Fermentation, Extraction, and Isolation. The fungal strain was
cultured in 20 × 500 mL Erlenmeyer flasks each containing 100 mL of
production medium (0.3 g yeast extract, 0.3 g malt extract, 0.5 g
peptone, 1 g glucose, pH 7.2−7.4) at 27 °C. After 14 days of static
culture, the fermentation broth, including cells, was harvested and then
centrifuged to separate the mycelial mass from the aqueous layer.
The mycelial mass and the aqueous layer were exhaustively
extracted with acetone and EtOAc, respectively. Then each extract
was concentrated in vacuo. The EtOAc extract was subjected to C18
flash column chromatography (5 × 30 cm), eluting with a stepwise
gradient of 20%, 40%, 60%, 80%, and 100% (v/v) MeOH in H2O (2 L
each). The fraction that eluted with 60% MeOH was further
fractionated with a Sephadex LH-20 column using CHCl3/MeOH
Assay for Cytotoxicity against P388 Cells. P388 murine
leukemia cells were cultured in RPMI-1640 medium containing 10%
fetal bovine serum, 100 μg/mL kanamycin, and 10 μg/mL 2-
hydroxyethyl disulfide at 37 °C under an atmosphere of 5% CO2. To
each well of the 96-well microplate containing 100 μL of tumor cell
suspension (1 × 104 cells/mL) was added 100 μL of test solution
dissolved in RPMI-1640 medium, and the plate was incubated in a
CO2 incubator at 37 °C for 96 h. After the addition of 50 μL of 3-(4,5-
dimethyl-2-thiazolyl)-2,5-diphenyl-2H-tetrazolium bromide saline sol-
ution (1 mg/mL) to each well, the plate was incubated for 3 h under
the same conditions to stain live cells. After the incubation, the plate
was centrifuged, the supernatants were removed, and the cells were
dissolved in 150 μL of DMSO to determine the IC50 values.
Histone Methyltransferase Inhibitory Activity Assay. A
mixture of 1 μL of GST (glutathione S-transferase)-G9a (200 ng) or
His (histone)-Set7/9 (931 ng), 1 μL of BSA (150 ng), 25 μL of 2×
HMT activity buffer (100 mM Tris-HCl (pH 8.5), 20 mM MgCl2, 40
mM KCl, 20 mM 2-mercaptoethanol, 500 mM sucrose), 1 μL of a
DMSO solution of the compound of interest, and 20 μL of H2O was
incubated at room temperature (rt) for 1 h in a total volume of 48 μL/
well. After the addition of 1 μL of S-adenosylmethionine (50 ng) and 1
μL of biotinylated H3 peptide (50 ng) into this reaction mixture, the
resulting mixture was further incubated at 37 °C for 2 h followed by
boiling at 96 °C for 30 min. The supernatant of the mixture was
transferred to a streptavidin-coated plate and incubated at rt for 1 h.
The supernatant was removed, and the remaining plate was washed
with 300 μL of PBS containing 0.5% Tween20 (PBST) three times.
Then, the plate was treated with an antimethylated-lysine antibody16
for G9a inhibitory activity assay or an antidimethyl-Histone H3 (Lys4)
113
dx.doi.org/10.1021/np200740e | J. Nat. Prod. 2012, 75, 111−114

Journal of Natural Products
Note
antibody (Upstate, 05-790) for a Set7/9 inhibitory activity assay in
PBST. After incubation at rt for 1 h, the supernatant was removed, and
the remaining plate was washed three times with 300 μL of PBST.
Subsequently, the secondary antibody conjugated with mouse HRP
(horseradish peroxidase) for the G9a inhibitory activity assay or rabbit
HRP for the Set7/9 inhibitory activity assay in PBST (100 μL/well)
was added to the plate, and the resulting mixture was incubated at rt
for 1 h. After the removal of the supernatant, the plate was washed five
times with 300 μL of PBST. A substrate of 3,3′,5,5′-tetramethylbenzi-
dine peroxidase was added into the plate, and the reaction was run at rt
for 40 min. Finally, the inhibitory activity was evaluated by measuring
absorption of each well at 650 nm using a plate reader.
ASSOCIATED CONTENT
■
S
* Supporting Information
1H NMR, 1H−1H COSY, HSQC, and HMBC data for
1
compounds 1 and 2 and H NMR data for the MTPA esters
of 1. This material is available free of charge via the Internet at
http://pubs.acs.org.
AUTHOR INFORMATION
Corresponding Author
*Tel: +81-3-5841-5297. Fax: +81-3-5841-8166. E-mail:
assmats@mail.ecc.u-tokyo.ac.jp.
■
ACKNOWLEDGMENTS
This work was partly supported by a Grant-in-Aid for Scientific
Research on Innovative Areas (23102007) from MEXT.
■
REFERENCES
■
(1) Blunt, J. W.; Copp, B. R.; Munro, M. H. G.; Northcote, P. T.;
Prinsep, M. R. Nat. Prod. Rep. 2011, 28, 196−268.
́
(2) Olano, C.; Mendez, C.; Salas, J. A. Mar. Drugs 2009, 7, 210−248.
(3) Rateb, M. E.; Ebel, R. Nat. Prod. Rep. 2011, 28, 290−344.
(4) Gardiner, D. M.; Waring, P.; Howlett, B. J. Microbiology 2005,
151, 1021−1032.
(5) Vigushi, D. M.; Mirsaidi, N.; Brooke, G.; Sun, C.; Pace, P.;
Inman, L; Moody, C. J.; Coombers, R. C. Med. Oncol. 2004, 21, 21−
30.
(6) Iwasa, E.; Hamashima, Y.; Fujishiro, S.; Higuchi, E.; Ito, A.;
Yoshida, M.; Sodeoka, M. J. Am. Soc. Chem. 2010, 132, 4078−4079.
(7) Kirby, G. W.; Robins, D. J.; Sefton, M. A.; Talekar, R. R. J. Chem.
Soc., Perkin Trans. 1 1980, 119−121.
(8) Kirby, G. W.; Rao, G. V.; Robins, D. J. J. Chem. Soc., Perkin Trans.
1 1988, 301−304.
(9) Ayent, A. G.; Hanson, J. R.; Truneh, A. Phytochemistry 1993, 32,
197−198.
(10) Kaouadji, M.; Steiman, R.; Seigle-Murandi, F.; Krivobok, S.;
Sage, L. J. Nat. Prod. 1990, 53, 717−719.
(11) Forseth, R. R.; Fox, E. M.; Chung, D. W.; Howlett, B. J.; Keller,
N. P.; Schroeder, F. C. J. Am. Soc. Chem 2011, 133, 9678−9681.
(12) Ueoka, R.; Nakao, Y.; Kawatsu, S.; Yaegashi, J.; Matsumoto, Y.;
Matsunaga, S.; Furihata, K.; van Soest, R. W.; Fusetani, N. J. Org.
Chem. 2009, 74, 4203−4207.
(13) Compound 2 has been reported in a patent as a product of
deacetoxylation of gliotoxin (Mukai, C. JP Patent WO 9824926,
1998).
(14) Johnson, J. R.; Bruce, W. F.; Dutcher, J. D. J. Am. Chem. Soc.
1943, 65, 2005−2009.
(15) O’Donnell, K. In The Fungal Holomorph: Mitotic, Meiotic and
Pleomorphic Speciation in Fungal Systematics; Reynolds, D. R.; Taylor, J.
W., Eds.; CAB International: Wallingford, United Kingdom, 1993;
Fusarium and its near relatives, pp 225−233.
(16) Iwabata, H.; Yoshida, M.; Komatsu, Y. Proteomics 2005, 18,
4653−4664.
114
dx.doi.org/10.1021/np200740e | J. Nat. Prod. 2012, 75, 111−114