Characterization, synthesis and self-aggregation of (-)-alternarlactam: A new fungal cytotoxin with cyclopentenone and isoquinolinone scaffolds

Article abstract of DOI:10.1002/chem.201002205

(-)-Alternarlactam [(-)-1], a new promising cytotoxin against two human cancer cell lines, was isolated from an endophyte culture and synthesized (along with (+)-1) from readily available starting materials. The absolute configuration, chirality-activity relevance and self-aggregation of (-)-1 were assigned by a combination of synthetic, spectroscopic and computational approaches. The full characterization of the new fungal cytotoxin may provide valuable information in the discovery of new antitumor agents. Double the framework: The absolute configuration, chirality-activity relevance and self-aggregation of (-)-alternarlactam [(-)-1], a cytotoxin from Alternaria sp. HG1 culture with double antitumor pharmacophores in a new framework, were assigned by a combination of synthetic, spectroscopic and computational approaches. Copyright

Full text of DOI:10.1002/chem.201002205

FULL PAPER
DOI: 10.1002/chem.201002205
Characterization, Synthesis and Self-Aggregation of (ꢀ)-Alternarlactam:
A New Fungal Cytotoxin with Cyclopentenone and Isoquinolinone Scaffolds
Ai Hua Zhang,[a] Nan Jiang,[b] Wen Gu,[a] Jing Ma,[b] Yu Rong Wang,[a]
Yong Chun Song,[a] and Ren Xiang Tan*[a]
Abstract: (ꢀ)-Alternarlactam [(ꢀ)-1], a new promising cytotoxin against two
human cancer cell lines, was isolated from an endophyte culture and synthesized
(along with (+)-1) from readily available starting materials. The absolute configu-
ration, chirality–activity relevance and self-aggregation of (ꢀ)-1 were assigned by
a combination of synthetic, spectroscopic and computational approaches. The full
characterization of the new fungal cytotoxin may provide valuable information in
the discovery of new antitumor agents.
Keywords: alternarlactam · antifun-
gal agents · antitumor agents · total
synthesis
Introduction
organisms, are a rich source of functional biomolecules as
substantiated by our characterization of new symbiont me-
tabolites with antitumor,[6–8] antimicrobial,[9] and immuno-
suppressive activities.[10] In continuation of our bioassay for
the antitumor secondary metabolites from symbiont cul-
tures, the pronounced cytotoxicity against the test cancer
cell line was discerned with the extract derived from the cul-
ture of Alternaria sp. HG1, a fungus isolated from inside the
healthy leave of Carex aridula belonging to the Cyperaceae
family.
Cancer remains one of the most severe threats to the public
health worldwide, and unfortunately oncology retains one of
the most pressing records in terms of investigational drug in
clinical development with the success rate even below one-
third of that for cardiovascular diseases.[1] This is worsened
by the readily acquired resistance of carcinomas to the ma-
jority of the existing tumor-suppressing agents.[2] The iden-
tification of chemically unique and biologically promising
molecules valuable for the antitumor drug discovery is
therefore of increasing necessity.
Plants usually harbor in some organs or tissues a specific
community of microorganisms called technically “endo-
phyte”, some of which have been ascertained to have gene
flows with the host[3] and/or among different microbial spe-
cies residing in the same niche.[4–5] This observation funda-
mentalizes that endophytes, as a big special group of micro-
The subsequent bioassay-directed fractionation afforded a
new cytotoxic metabolite named trivially (ꢀ)-alternarlactam
[(ꢀ)-1] (see below). After purification, the cytotoxicity of
[a] A. H. Zhang, Dr. W. Gu, Y. R. Wang, Dr. Y. C. Song,
Prof. Dr. R. X. Tan
Institute of Functional Biomolecules
State Key Laboratory of Pharmaceutical Biotechnology
Nanjing University, Nanjing 210093 (P.R. China)
Fax : (+86)25-8330-2728
E-mail: rxtan@nju.edu.cn
[b] Dr. N. Jiang, Prof. Dr. J. Ma
(ꢀ)-1 against two human cancer cell lines (cervix HeLa ade-
nocarcinoma and hepatocellular carcinoma 7701) was con-
firmed to be comparable to those of doxorubicin (Table S1
in the Supporting Information), a marketed anticancer drug
co-assessed as a positive control in the study.
School of Chemistry and Chemical Engineering, Institute of Theoreti-
cal and Computational Chemistry, Lab of Mesoscopic Chemistry,
Nanjing University, Nanjing 210093 (P. R. China)
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.201002205.
Chem. Eur. J. 2010, 16, 14479 – 14485
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
14479

Table 1. 1H and 13C NMR data of (ꢀ)-1.[a]
Results and Discussion
Position
dC
DEPT
dH (J in Hz)
Key HMBC
(H!C)
AHCTUNGTRENNUNG
The structure determination of (ꢀ)-alternarlactam [(ꢀ)-1]
was accommodated by a combination of spectral methods
including IR (Figure S1), HR-ESI-MS (Figure S2), 1H and
13C NMR (Table 1; Figures S3 and S4), a set of 2D NMR ex-
periments (1H–1H COSY, HMBC and HMQC; Figures S5–
S7) and its NOE difference spectroscopy, coupled with ex-
perimental and computational electronic/vibrational circular
dichroism (ECD/VCD) approaches. Briefly, the molecular
formula (C14H13NO4) of (ꢀ)-1 was evidenced from the qua-
simolecular ion peak at m/z 282.0844 ([M+Na]+) in its ESI
mass spectrum, which was reinforced by a total of 13-proton
1
2a
30.2
44.1
CH
CH2
3.53 (m, 1H)
3.01 (dd, 1H,
J = 19.2, 6.6)
2.36 (dd, 1H,
J =19.2, 1.2)
C-3
C-9b, C-10
2b
C-3a, C-9b, C-10
C-5a, C-9b
3
3a
4
5
5a
6
7
8
9
9a
9b
10
6-OH
8-OCH3
197.8
134.2
C
C
8.63 (brs, 1H)
165.6
107.4
164.9
102.6
165.1
101.4
135.6
144.0
21.2
C
C
C
CH
C
CH
C
C
6.67 (d, J = 2.1, 1H)
6.71 (d, J =2.1, 1H)
C-9
1
integration and 14 resonance lines in its H and 13C NMR
C-5a, C-7, C-9b
spectra, respectively. An ortho-substituted 2-hydroxy-4-me-
thoxybenzoyl group was indicated by a hydroxyl singlet at d
13.18 shifted downfield presumably via hydrogen bonding
with carbonyl group, and from a pair of mutually meta-cou-
pling doublets (J=2.1 Hz) at dH 6.67 and 6.71 ppm, which
both showed NOE enhancements upon pre-saturation of the
three-proton singlet at d = 3.92 PPM. Furthermore, the
presence of an a,b-disubstituted 4-methylcyclopent-2-enone
moiety was demonstrated by the three quaternary carbon
resonance lines at dC = 197.8, 165.6 and 134.2 ppm, as well
as by a coupling sequence arising from a 2-substituted
propyl group signifying at dH = 1.46 (d, 3H, J=7.2 Hz),
3.53 (m, 1H), 3.01 (dd, 1H, J=19.2, 6.6 Hz) and 2.36 ppm
(dd, 1H, J=19.2, 1.2 Hz). These two observations, com-
pounding a nitrogen atom and nine degrees of unsaturation
implied by the molecular formula, demonstrated that 2-hy-
droxy-4-methoxybenzoyl and 4-methylcyclopent-2-enone
moieties had to combine through the d-lactam ring to form
the entire molecule. This elucidation was unambiguously
CH3
1.46 (d, J =7.2, 3H)
12.86 (brs, 1H)
3.92 (s, 3H)
C-2, C-9b
C-5a, C-7
C-8
55.8
CH3
[a] Acquired in CDCl3 at 300 and 75 MHz, respectively.
ing compounds were shown to be stereochemically sensitive
as summarized elsewhere.[14] It was therefore essential to un-
derstand the chirality of the cytotoxin. We thus acquired its
CD spectrum that displayed clear Cotton effects at 218 (pos-
itive), 251 (negative), and 365 nm (negative) (Figure 1a),
which depended solely on the absolute configuration of C-1,
the only chiral carbon in the molecule. However, it could
not be determined readily by the discerned CD bands since
no model compound could be compared or consulted. This
frustration was overcome by the quantum chemical ap-
proach. As illustrated in Figure 1, the experimentally record-
ed CD spectrum was well comparable with the theoretically
predicted ECD curve for (ꢀ)-1, which was totally opposite
to that calculated for its synthetic enantiomer (+)-1 (see
below). In particular, the two positive Cotton effects of (ꢀ)-
1 were calculated to center at 220 and 336 nm, which
matched well with the observed bands at 218 and 333 nm
(Figure 1a). Concerning the assignment of these distinctive
ECD bands, the excitations from n-type (lone pair orbital
on oxygen and nitrogen) and p-type (filled C=C orbital) mo-
lecular orbitals (designated as MO) to p*-type (antibonding
C=C and C=O orbitals) MO, MO 77!80 and MO 76!78
(Table S3 and Figure S19), contributed dominantly to these
absorbed bands. The first negative Cotton effect in the cal-
culated ECD curve was located at 252 nm, corresponding to
the broad band at 251 nm in the experimental spectrum.
1
confirmed collectively by H–1H COSY, HMBC and HMQC
experiments (Figures S5–S7), which allowed the exact as-
1
signment of all H and 13C NMR signals of (ꢀ)-1 (Table 1).
Particularly, the elucidated framework was in agreement
with the discerned HMBC correlation of H-1 with C-3, of
H-2 with C-3a, C-9b and C-10, and of the NH singlet with
C-5a and C-9b. This observation, along with the NOE en-
hancement of H-9 discerned upon irradiating H-10 signal,
required that (ꢀ)-1 was 6-hydroxy-8-methoxy-1-methyl-1H-
cyclopenta[c]isoquinoline-3,5ACTHUNTGRNEUNG(2H,4H)-dione, which pos-
sessed a hitherto unreported framework.
(ꢀ)-Alternarlactam [(ꢀ)-1] was unique in its possession of
an isoquinolinone moiety in conjunction with a cyclopente-
none residue, which might be an ideal combination of cyto-
toxic pharmacophores as highlighted below. The isoquinoli-
none/isoquinoline moiety was recognized as a scaffold im-
portant for some antitumor and/or antiinflammatory drug
candidates such as ecteinascidin 743 (see above).[11–13] On
the other hand, the a,b-unsaturated cyclopentone, a soft
electrophile, was recognized as an antitumor pharmaco-
phore,[14] which was included as well in other bioactive mole-
cules such as parthenin[15] and 1-deoxyrubralactone, a re-
cently characterized eukaryotic DNA polymerase inhibi-
tor.[16] And the biological functions of cyclopentenone-bear-
*
The pC¼C ! pC¼O (MO 76!79) played an important role.
The next calculated negative Cotton effect at 364 nm could
be assigned to the experimental Cotton effect at 365 nm.
Contributed to this band was the excitation of p!p* (MO
77!78), where p and p* were the filled C=C and antibond-
ing C=C and C=O orbitals, respectively.
On the basis of the above information, the optimized 3D
structure of (ꢀ)-1 (Figure S20) was obtained with the Gaus-
sian03 program[17] at B3LYP/6-31+G
ACTHNUTRGNEUNG(d,p) level within the
14480
www.chemeurj.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 14479 – 14485

A New Fungal Cytotoxin
FULL PAPER
(Figure 2). This observation
could be explained by assuming
that (ꢀ)-1 tends to pack
through intermolecular hydro-
gen bonds. Subsequent scrutiny
of the calculated IR and VCD
spectra of the aggregates
(Figure 2) showed that the ter-
nary (ꢀ)-1 involves various in-
termolecular non-covalent (C=
O···H-N) linkages. The detailed
geometries of the ternary (ꢀ)-1
are given in Table S6. In addi-
tion, the VCD bands in the
1800–1200 cmꢀ1 region could be
resolved by Fourier self-decon-
volution of the absorption spec-
trum. Aided by a comparison
between the theoretically pre-
dicted and experimentally ac-
quired VCD spectra of ternary
(ꢀ)-1, the assignment was per-
formed on the basis of the
visual observation with Gauss-
Figure 1. Attribution of the absolute configuration of (ꢀ)-1 (natural) and (+)-1 (synthetic) as 1S and 1R by
comparing between the experimental (c) and calculated (g) ECD spectra. The computation was carried
out at B3LYP/6-31+GACHTUNGTRENNUNG(d,p) level within the combined explicit solvent/PCM model.
combined explicit solvent/polarized continuum model
(PCM; CH3OH solvent: dielectric constant e=32.63). The
calculated distances between H-1/H-2a (2.348 ꢁ), H-2b/Ha-
10 (2.585 ꢁ), H-1/H-9 (2.564 ꢁ), Hb-10/H-9 (2.316 ꢁ) and
H-9/Ha-OCH3 (2.329 ꢁ) were all less than 4 ꢁ, which was
consistent with the NOEs observed for each of the proton
pairs. Thus, the stereochemical
view software (graphical interface for Gaussian 03, Gaussian
Inc.). Harmonic vibrational frequency, dipole strength, and
rotational strength as well as the assignment were given in
Tables S7 and S8. As shown in Figure 2, the VCD signals at
1713(+)/1702(ꢀ) cmꢀ1 was due to the C=O stretching vibra-
tions in the cyclopentenone substructure. This band correlat-
structure of the fungal metabo-
lite was established as (1S)-6-
hydroxy-8-methoxy-1-methyl-
1H-cyclopenta[c]isoquinoline-
3,5
N
In view of the dependence of
the biological performance on
its physical behavior of bioac-
tive compounds,[18] we subse-
quently probed the aggregation
of (ꢀ)-1 in the solid state by
utilizing the vibrational circular
dichroism (VCD) technique
that records the differential ab-
sorption between the left- and
right-handed circularly polar-
ized IR radiation via molecular
vibration transition to provide
the stereochemical and confor-
mational information of the
tested organic compounds.[19–25]
As anticipated, the IR and
VCD spectra acquired for (ꢀ)-1
in a KBr disc was shown to be
well comparable to that compu-
Figure 2. Comparisons of experimental IR (top) and VCD (bottom) spectra of (ꢀ)-1 with those calculated for
terized for its ternary aggregate the aggregates of (ꢀ)-1 at B3PW91/6-31+G
ACHTUNGTREN(NUNG d,p) level in vacuum.
Chem. Eur. J. 2010, 16, 14479 – 14485
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
14481

R. X. Tan et al.
ed to those observed experimentally at 1712(+)/
1704(ꢀ) cmꢀ1. The peaks at 1678(+) and 1645(ꢀ) cmꢀ1 in
the theoretical VCD curve corresponded to the experimen-
tally intense band at 1657(+) and 1639(ꢀ) cmꢀ1, which were
assigned to the C=O stretching vibrations in the d-lactam.
The signals at 1637(ꢀ)/1628(+) were ascribed to the C=C
stretching vibrations. They could be correlated to the experi-
mental bands at 1624(ꢀ) and 1608(+) cmꢀ1. On the other
ꢀ
ꢀ
ꢀ
hand, the O H, N H, and C H rocking vibrations gave rise
to the bands at 1578 and 1569 cmꢀ1. The VCD signals rang-
ing from 1517 to 1442 cmꢀ1 in the theoretically predicted
spectrum corresponded to the experimental bands around
1549–1449 cmꢀ1. They were attributed to the in-plane scis-
soring vibrations of CH3 group. Finally, the bands in 1352–
1200 cmꢀ1 region in the experimental curve were reproduced
by the theoretical calculations (in 1347–1200 cmꢀ1). In these
ꢀ
ꢀ
bands, the coupled in-plane rocking vibrations of C H, N
ꢀ
H, and O H groups played a dominant role. It seems inter-
esting that (ꢀ)-1 present in its ternary form looks “similar”
to ecteinascidin 743 (an anticancer drug at phase III clinical
trial[12]) composed of three tetrahydroisoquinolinone units
(see above).
Scheme 1. Total synthesis of (ꢀ)-1 and (+)-1. a) Oxalyl chloride, 1658C,
30 min (100%); b) 33%NaOH/30%H2O2 (59%); c) NaNO2, HCl, 08C
(100%); d) 508C, 12 h (48%); e) NH3·H2O, formamide, 808C, 3 h
(91%); f) BBr3, CH2Cl2, 08C, 1 h (92%).
The promising cytotoxicity of (ꢀ)-1 was diminished by its
low content in the fungal culture. It was apparently impera-
tive to develop a reliable protocol that could supply it hope-
fully in an unlimited manner. We were therefore obligated
to synthesize it from simple chemicals with an intention to
afford meanwhile its enantiomer (+)-1 that was highly de-
sired for a detailed enantiomeric comparison between the
functional and spectral features. For this reason, we pursued
tioned stereochemical assignment, the specific rotation and
ECD/VCD bands were opposite (Figures 1, 3 and S22) be-
tween the two enantiomers [(ꢀ)-1 and (+)-1] although both
1
signified equally in their IR, MS, H and 13C NMR spectra.
With the two single enantiomers of alternarlactam (1) in
hand, we were able to compare the antitumor magnitude
among the racemic and stereochemical isomers. The enan-
tiomer (ꢀ)-1 was shown to inhibit the growth of human
cervix HeLa adenocarcinoma cell (IC50 = 1.10 mgmLꢀ1) and
human hepatocellular carcinoma cell line QGY-7701 (IC50 =
1.52 mgmLꢀ1), with the magnitude close to those discerned
with doxorubicin co-assayed as a positive control in the
study (Figure 3). The enantiomer (+)-1, only moderately
active to the QGY-7701 cell line (IC50 = 27.54 mgmLꢀ1), ex-
hibited a weaker inhibition to the HeLa cell line with an
a
synthetic approach toward (ꢁ)-alternarlactam which
might give the corresponding single enantiomer after chiral
HPLC separation. Accordingly, retrosynthetic analyses sug-
gested that (ꢁ)-alternarlactam could be constructed by the
ꢀ
C C bondage between (ꢁ)-4-methyl-1,2-cyclopentanedione
and a 2,4-dioxygenated benzoyl units. Thus, a five-step syn-
thetic route was designed for its chemical generation from
the two commercially available chemicals (Scheme 1). Start-
ing from 3,5-dimethoxyaniline, the intermediate 2-amino-
4,6-dimethoxybenzoic acid (B) was prepared in two steps.[28]
After diazotizing B, the d-lactone C formed upon addition
of (ꢁ)-4-methyl-1,2-cyclopentanedione.[29] However, pre-
sumably owing to the stronger p–p conjugation of the ester
oxygen with a 3-phenyl-cyclopentenone, the d-lactone C
could not be transformed into the expected d-lactam D by
the reported addition of formamide or gaseous ammo-
nia.[11,26] This problem was solved eventually by a high-yield
(91%) conversion of d-lactone C to d-lactam D, which we
tried out by utilizing aqueous ammonia in the presence of
formamide. The subsequent de-O-methylation using boron
bromide in dichloromethane afforded (ꢁ)-alternarlactam
(E), which was separated into (ꢀ)-1 and (+)-1 by the chiral
HPLC procedure.
1
The IR, MS, H and 13C NMR data and specific rotation
Figure 3. Comparisons of the cytotoxicity (in IC50 value) of (ꢀ)-1 (blank),
(+)-1 (sparse), (ꢁ)-1 (dense) and doxorubicin (solid, co-assayed as a pos-
itive control) against HeLa (human cervix HeLa adenocarcinoma) and
QGY-7701 (human hepatocellular carcinoma QGY-7701) cell lines (see
also Table S1).
of the synthetic (ꢀ)-alternarlactam [(ꢀ)-1] were identical
with those acquired for the natural product isolated from
the fungal culture. Also as a confirmation for the aforemen-
14482
www.chemeurj.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 14479 – 14485

A New Fungal Cytotoxin
FULL PAPER
IC50 value of 7.19 mgmLꢀ1, and those of the racemate [(ꢁ)-
1] against the QGY-7701 and HeLa cell lines were 11.97 and
1.91 mgmLꢀ1, respectively. These data suggested that the cy-
totoxicity of alternarlactam is most likely chirality-sensitive
(Figure 3).
vacuo to give an oily residue (ca. 108 g), which was then subjected to
column chromatography (CC) on silica gel (1200 g, 200–300 mesh), elut-
ing with CHCl3/MeOH 1:0!0:1 to give nine fractions E1–E9. The bioac-
tive fraction E2 was rechromatographed over silica gel. CC (170 g, 200–
300 mesh) and then eluted again with CHCl3/MeOH 1:0!0:1 to turn out
five parts, one of which displayed bioactivity. The bioactive part was fur-
ther separated through gel filtration over Sephadex LH-20 with CHCl3/
MeOH 1:1 to give a white solid of (ꢀ)-1 (6.7 mg).
Spectral data of (ꢀ)-1: White solid; m.p. 234–2368C; [a]2D5 = ꢀ20.68 (c=
0.04, MeOH); UV (MeOH): lmax = 255, 358 nm; IR: n˜ = 3383, 3157,
Conclusion
3048, 1706, 1650 (s), 1625, 1559, 1503, 1466, 1284, 1207, 969, 841 cmꢀ1
;
In summary, Alternaria sp. HG1 from inside the normal
Carex aridula leave was shown to produce in a trace amount
the new antitumor polyketide (ꢀ)-alternarlactam [(ꢀ)-1]
whose structure was determined rigorously by spectral and
computational methods. (ꢀ)-Alternarlactam is unique in its
co-possession of cyclopentenone and isoquinolinone scaf-
folds, both being important antitumor-related pharmaco-
phores.[11–14] The usual structural character might be an im-
plication for the new mode of action. The formation of (ꢀ)-
1 was rendered renewable through the synthesis from the
commercially available chemicals. The lactamization proce-
dure established in the work is useful for the construction of
some lactam nuclei. The co-synthesis of both enantiomers
[(ꢀ)-1 and (+)-1] allowed both the disclosure of the chirality
relevance to the antitumor action, and the robust validation
of the computational ECD and VCD methods. Particularly,
the correlative interpretation of the experimental and com-
putational VCD spectra led to the recognition of the self-ag-
gregation of (ꢀ)-1 in KBr, which seems to be the first exem-
plification of the natural product-based supramolecule that
may be of wider interest in chemistry. Finally, the new
framework and biological profiles of both enantiomers of al-
ternarlactam (1), along with the established synthetic proto-
col and its aggregative properties, provided collectively at-
tractive and important information that may shed fresh light
onto the elusive path towards the discovery of new antitu-
mor agents, which has long been and continues to be one of
the most urgent tasks in medicine.
CD (MeOH): l (De)
=
355 (ꢀ0.84), 255 nm (ꢀ1.67); ESI-MS: m/z:
260.1021 [M+H]+, 282.0844 [M+Na]+, 298.0140 [M+K]+, 541.2416
[2M+Na]+, 800.3215 [3M+Na]+; H and 13C NMR data: see Table 1.
1
Synthesis of 4,6-dimethoxyanthranilic acid (B): 3,5-Dimethoxyaniline hy-
drochloride (12 g, 63 mmol) was dissolved in oxalyl chloride (20 mL, 230
mol) at room temperature, and the resultant solution was stirred at
1658C for 30 min. In vacuo removal of the excess oxalyl chloride gave a
green-yellow residue, which was refluxed for 10 min with methanol
(50 mL). Subsequent filtration gave 4,6-dimethoxyindole-2,3-dione
A
(13 g, 63 mmol, 100%), which was added into an oversized flask and dis-
solved with 33% NaOH solution (60 mL), followed by dropwise addition
of 30% solution of H2O2 (30 mL) with the reaction mixture maintained
at 1008C for an additional 10 min. The solution was then adjusted to
pH 3 with concentrated HCl. The solid that formed was filtered, washed
with water and dried to yield B (7.3 g, 37 mmol, 59%).[29] Compound A:
m.p. 302–3068C; 1H NMR (300 MHz, [D6]DMSO): d = 10.93 (brs, 1H),
6.18 (d, J=1.8 Hz, 1H), 6.00 (d, J=1.8 Hz, 1H), 3.89 (s, 3H), 3.87 nm (s,
3H); EI-MS: m/z: calcd for C10H9NO4: 207.1; found: 208.1 [M+H]+.
Compound B: m.p. 123–1278C; 1H NMR (300 MHz, CDCl3): d = 11.09
(brs, 1H), 6.45 (brs, 2H), 5.83 (d, J=2.1 Hz, 1H), 5.81 (d, J=2.1 Hz,
1H), 3.97 (s, 3H), 3.79 ppm (s, 3H); EI-MS: m/z: calcd for C9H11NO4:
197.1; found: 198.1 [M+H]+.
Synthesis of 6,8-dimethoxy-1-methyl-1,2-dihydrocyclopenta[c]isochro-
mene-3,5-dione (C): Similar to the procedure detailed elsewhere,[30] con-
centrated HCl (2 mL) was added to the mixture of B (394 mg, 2 mmol)
and water (2 mL), to which, sodium nitrite (144 mg, 2.1 mmol) in water
(1 mL) was added at 08C, followed by addition of 4-methyl-1,2-cyclopen-
tanedione (280 mg, 2.5 mmol). After stirred at 508C for 12 h, the reaction
mixture was diluted with water (5 mL), extracted with CH2Cl2 (30 mL)
and dried over Na2SO4. The solid residue afforded after evaporation was
purified over silica gel column chromatography eluted with CHCl3/
MeOH 100:1 to give C (263 mg, 0.96 mmol, 48%) as a yellowish solid.
1H NMR (300 MHz, CDCl3): d = 6.67 (d, J=2.1 Hz, 1H), 6.65 (d, J=
2.1 Hz, 1H), 4.00 (s, 3H), 3.99 (s, 3H,), 3.44 (m, 1H), 2.93 (dd, J=
19.2,6.3 Hz, 1H), 2.28 (dd, J=18.9,0.9 Hz, 1H), 1.44 ppm (d, J=7.2 Hz,
3H); 13C NMR (75 MHz, CDCl3): d = 195.9, 165.6, 164.4, 157.0, 149.1,
142.9, 137.1, 104.3, 100.8, 100.2, 56.4, 55.9, 42.8, 28.1, 21.1; ESI-MS: m/z:
calcd for C15H14O5: 274.10: 297.0834 [M+Na]+; found: 275.1001 [M+H]+.
Experimental Section
General reagents and instrumentation: As described earlier.[7,9,10]
Synthesis of 6,8-dimethoxy-1-methyl-1H-cyclopenta[c]isoquinoline-3,5-
AHCTUNGTREG(NNUN 2H,4H)-dione (D): A solution of C (274 mg, 1 mmol) in formamide
Strain and cultivation: The alternarlactam-producing fungus, identified
by Dr. Y. C. Song as Alternaria sp. HG1 according to its macro- and
micro-graphic morphological characteristics,[27] was isolated from inside
the fresh Carex aridula leaves collected in November 2001 from the
suburb of Nanjing, China. The fresh mycelia grown on PDA plates were
incubated into 1000 mL flasks preloaded with each 400 mL potato dex-
trose broth followed by shaking at 140 rpm at 258C for 5 d. After that,
the liquid culture (10 mL) was transferred into each of 400 bottles con-
taining millet medium,[28] before hatched quietly at 258C for 30 d. At the
end of the solid-substrate fermentation, a total of 40 L fermented materi-
al was harvested, dehydrated, and then pulverized into dry powder (ca.
2 kg).
(5 mL) was refluxed for 3 h at 808C with concentrated ammonia (2 mL)
added in every 30 min. The progress of the reaction was monitored by
TLC developed with petrol ether/acetone 2:1. After evaporated in vacuo,
the residue was subjected to gel filtration over Sephadex LH-20 with
CHCl3/MeOH 1:1 to give D (248 mg, 0.91 mmol, 91%) as a white solid.
1H NMR (500 MHz, CDCl3): d
= 8.74 (brs, 1H), 6.73 (d, J=2.0 Hz,
1H), 6.68 (d, J=2.0 Hz, 1H), 4.00 (s, 3H), 3.97 (s, 3H,), 3.52 (m, 1H),
3.00 (dd, J=19.1,6.5 Hz, 1H), 2.34 (dd, J=19.1,1.1 Hz, 1H), 1.45 ppm (d,
J=7.0 Hz, 3H); 13C NMR (125 MHz, CDCl3): d = 197.7, 163.8, 163.7,
160.4, 141.3, 138.1, 135.4, 112.1, 100.7, 99.5, 56.5, 55.7, 44.3, 30.1,
21.5 ppm; ESI-MS: m/z: calcd for C15H15NO4: 273.09; found: 296.0868
[M+Na]+, 274.1051 [M+H]+.
Fractionation of (ꢀ)-1: The powder derived from the solid substrate fer-
mentation with Alternaria sp. HG1 was extracted exhaustively with
CHCl3/MeOH 1:1. In vacuo evaporation of the solvent yielded an oily
crude extract (ca. 280 g), which was dissolved completely in the least
amount of methanol prior to being kept at ꢀ208C overnight to get rid off
the waxy material through precipitation. The filtrate was concentrated in
Synthesis of 6-hydroxy-8-methoxy-1-methyl-1H-cyclopenta[c]isoquino-
line-3,5
was added dropwise at 08C under a nitrogen atmosphere to a solution of
6,8-dimethoxy-1-methyl-1H-cyclopenta[c]isoquinoline-3,5(2H,4H)-dione
D (28 mg, 0.1 mmol) in dry dichloromethane (2 mL). One hour later, the
A
ACHTUNGTNER(NUNG III) bromide (150 mg, 0.6 mmol)
AHCTUNGTRENNUNG
Chem. Eur. J. 2010, 16, 14479 – 14485
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
14483

R. X. Tan et al.
reaction mixture was quenched by the addition of 1m HCl (3 mL). The
resulting solution was extracted with CH2Cl2 (20 mL), washed sequential-
ly with H2O (40 mL), dried over Na2SO4, and evaporated to give a resi-
due that was chromatographed over Sephadex LH20 with MeOH/H2O
B3PW91/6-31+GACTHUNGETRNNU(G d,p) is employed here. The IR and VCD calculations
are also performed with the Gaussian03 program.[17]
1:1 to furnish (ꢁ)-1 (24 mg, 0.092 mmol, 92%) as a white solid. [a]D25
=
08; 1H NMR (300 MHz, CDCl3): d = 12.85 (brs, 1H), 8.70 (brs, 1H),
6.71 (d, J=2.1 Hz, 1H), 6.67 (d, J=2.1 Hz, 1H), 3.92 (s, 3H), 3.52 (m,
1H), 3.00 (dd, J=19.2,6.6 Hz, 1H), 2.36 (dd, J=19.2,1.2 Hz, 1H),
1.45 ppm (d, J=7.2 Hz, 3H); 13C NMR (75 MHz, CDCl3): d = 197.8,
165.6, 165.1, 164.9, 144.0, 135.6, 134.3, 107.5, 102.6, 101.4, 55.8, 44.2, 30.2,
21.2 ppm; HR-ESI-MS: m/z: calcd for C14H13NO4: 259.08: found:
282.0751 [M+Na]+.
Acknowledgements
This work, to which A.H.Z. and N.J. contributed equally, was co-financed
by grants from NSFC (30821006
& 90813036) and from MOST
(2007AA09Z446), and a project for N.J. from CPSF (20100471313).
Chiral HPLC separation of (ꢁ)-1: The racemic alternarlactam [(ꢁ)-1]
was separated over a Chiralpak AS-H column (4.6ꢂ250 mm, Daicel
Chemical Ltd) at 408C using methanol/DEA (100:0.1, v/v) as mobile
phase at a flow rate of 1.0 mLminꢀ1. The retention time of the two single
enantiomers was 8.67 [(+)-1] and 11.81 min [(ꢀ)-1], respectively, with
the optical purity (>99% ee) checked also by the chiral HPLC. The ESI-
[1] A. Kamb, S. Wee, C. Lengauer, Nat. Rev. Drug Discov. 2007, 6, 115–
120.
[2] G. Szakꢃcs, J. K. Paterson, J. A. Ludwig, C. Booth-Genthe, M. M.
Gottesman, Nat. Rev. Drug Discov. 2006, 5, 219–234.
[3] T. J. Sullivan, S. H. Faeth, Mol. Ecol. 2004, 13, 649–656.
[4] S. Taghavi, T. Barac, B. Greenberg, B. Borremans, J. Vangronsveld,
D. van der Lelie, Appl. Environ. Microbiol. 2005, 71, 8500–8505.
[5] A. Gentile, M. S. Rossi, D. Cabral, K. D. Craven, C. L. Schardl, Mol.
Phylogenet. Evol. 2005, 35, 196–208.
[6] H. M. Ge, W. Y. Zhang, G. Ding, P. Saparpakorn, Y. C. Song, S.
Hannongbua, R. X. Tan, Chem. Commun. 2008, 5978–5980.
[7] L. Shen, R. H. Jiao, Y. H. Ye, X. T. Wang, C. Xu, Y. C. Song, H. L.
Zhu, R. X. Tan, Chem. Eur. J. 2006, 12, 5596–5602.
[8] R. H. Jiao, S. Xu, J. Y. Liu, H. M. Ge, H. Ding, C. Xu, H. L. Zhu,
R. X. Tan, Org. Lett. 2006, 8, 5709–5712.
[9] H. W. Zhang, W. Y. Huang, J. R. Chen, W. Z. Yan, D. Q. Xie, R. X.
Tan, Chem. Eur. J. 2008, 14, 10670–10674.
[10] Y. L. Zhang, H. M. Ge, W. Zhao, H. Dong, Q. Xu, S. H. Li, J. Li, J.
Zhang, Y. C. Song, R. X. Tan, Angew. Chem. 2008, 120, 5907–5910;
Angew. Chem. Int. Ed. 2008, 47, 5823–5826.
[11] P. G. Jagtap, E. Baloglu, G. J. Southan, J. G. Mabley, H. S. Li, J.
Zhou, J. van Duzer, A. L. Salzman, C. Szabꢄ, J. Med. Chem. 2005,
48, 5100–5103.
MS, H and 13C NMR spectra of (+)-1 ([a]2D5 =+20.38) was identical with
1
those of (ꢀ)-1.
Cytotoxicity evaluation: The cytotoxicity was tested on the HeLa (human
cervix HeLa adenocarcinoma) and QSG-7701 (human hepatocellular car-
cinoma 7701) cell lines by the MTT [3-(4,5-dimethylthiazol-2-yl)-2,5-di-
phenyltetrazolium bromide] colorimetric method.[31] Briefly, the test cells
at the exponential growth phase were collected and transferred into 96-
well plates. After incubation for 24 h, compound dilutions were dispensed
to the established culture plates. Two days (48 h) later, the MTT solution
(0.1 mg per well) was then added to each well. After further incubation
for 4 h, the supernatant was removed, the crystals were fully dissolved in
DMSO (150 mL), and the absorbance of each well was read at 570 nm.
The IC50 value was determined as the concentration, at which a half of
the test cell growth was inhibited. The experiment was performed in trip-
licate, and the data expressed as means ꢁSD.[32]
ECD calculations: Density functional theory (DFT) at B3LYP/6-31+G-
A
Subsequently, the corresponding TDDFT calculations with the same
basis set were performed on the basis of the ground-state optimized
structures. The ECD calculation was accomplished by utilizing the com-
bined explicit solvent/PCM (polarized continuum model; CH3OH sol-
vent: dielectric constant e=32.63), which was demonstrated to be effec-
tive in describing solvent effects.[33,34] The final ECD spectra were ob-
tained according to the following equations:[35]
[12] O. Koepler, S. Laschat, A. Baro, P. Fischer, B. Miehlich, M. Hotfild-
er, C. le Viseur, Eur. J. Org. Chem. 2004, 3611–3622.
[13] G. Germano, R. Frapolli, M. Simone, M. Tavecchio, E. Erba, S.
Pesce, F. Pasqualini, F. Grosso, R. Sanfilippo, P. G. Casali, A. Gron-
chi, E. Virdis, E. Tarantino, S. Pilotti, A. Greco, M. Nebuloni, C. M.
Galmarini, J. C. Tercero, A. Mantovani, M. D’Incalci, P. Allavena,
Cancer Res. 2010, 70, 2235–2244.
ꢂ
ꢀ
ꢁ ꢃ
[14] M. Conti, Anti-Cancer Drugs 2006, 17, 1017–1022.
[15] W. Herz, H. Watenabe, M. Miyazaki, Y. Kishida, J. Am. Chem. Soc.
1962, 84, 2601–2610.
2
X
l ꢀ ln
Dln
DeðlÞ ¼
Den exp ꢀ
ð1Þ
ð2Þ
n
[16] M. Naganuma, M. Nishida, K. Kuramochi, F. Sugawara, H. Yoshida,
Y. Mizushina, Bioorg. Med. Chem. 2008, 16, 2939–2344.
[17] Gaussian 03, Revision D.01, M. J. Frisch, G. W. Trucks, H. B. Schle-
gel, G. E. Scuseria, M. A. Robb, J. R. Cheeseman, J. A. Montgom-
ery, Jr., T. Vreven, K. N. Kudin, J. C. Burant, J. M. Millam, S. S.
Iyengar, J. Tomasi, V. Barone, B. Mennucci, M. Cossi, G. Scalmani,
N. Rega, G. A. Petersson, H. Nakatsuji, M. Hada, M. Ehara, K.
Toyota, R. Fukuda, J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda,
O. Kitao, H. Nakai, M. Klene, X. Li, J. E. Knox, H. P. Hratchian,
J. B. Cross, V. Bakken, C. Adamo, J. Jaramillo, R. Gomperts, R. E.
Stratmann, O. Yazyev, A. J. Austin, R. Cammi, C. Pomelli, J. W.
Ochterski, P. Y. Ayala, K. Morokuma, G. A. Voth, P. Salvador, J. J.
Dannenberg, V. G. Zakrzewski, S. Dapprich, A. D. Daniels, M. C.
Strain, O. Farkas, D. K. Malick, A. D. Rabuck, K. Raghavachari,
J. B. Foresman, J. V. Ortiz, Q. Cui, A. G. Baboul, S. Clifford, J. Cio-
slowski, B. B. Stefanov, G. Liu, A. Liashenko, P. Piskorz, I. Komaro-
mi, R. L. Martin, D. J. Fox, T. Keith, M. A. Al-Laham, C. Y. Peng,
A. Nanayakkara, M. Challacombe, P. M. W. Gill, B. Johnson, W.
Chen, M. W. Wong, C. Gonzalez, J. A. Pople, Gaussian, Inc., Wall-
ingford CT, 2004.
lnRn
pffiffiffi
Den
¼
ꢂ 1040
22:94 pDln
where Den was the peak intensity given in Lmolꢀ1 cmꢀ1, Rn was the rota-
tory strength in unit of 10ꢀ40 cgs, ln was the wavelength of the nth transi-
tion, and Dln was the half-width at 1/e of peak maximum. Here we used
a half-width Dln = l2nDn˜ with Dn˜ = 1200 cmꢀ1 for the studied systems.
All calculations were performed with the Gaussian03 program.[17]
VCD calculations: Since VCD spectrum is conformationally sensitive,
systematic QM conformation search for the ternary aggregates of (ꢀ)-1
are firstly carried out. Then the VCD calculations are performed to get
the frequencies and rotational strengths. For the vibrational frequencies,
an empirical scaling factor of 0.945 was used to achieve the better agree-
ment with the experimental results. Such a scaling factor is comparable
to the reported values.[36] The IR absorption and VCD intensity were
converted to the corresponding spectra using a Lorentzian function (with
the full line width at half maximum of 8 cmꢀ1). The calculated dipole
strengths and rotational strength are scaled 7.1ꢂ10ꢀ4 and 4.3ꢂ10ꢀ4 to get
better agreement with the experiments. Since the three-parameter hybrid
approach using the Perdew and Wang correction functional (B3PW91)
works well in predicting the IR and VCD spectra for many chiral mole-
cules with chainlike, cylic, and even spherical cage structures,[37,38] the
[18] C. Allen, N. Dos Santos, R. Gallagher, G. N. C. Chiu, Y. Shu, W. M.
Li, S. A. Johnstone, A. S. Janoff, L. D. Mayer, M. S. Webb, M. B.
Bally, Biosci. Rep. 2002, 22, 225–250.
14484
www.chemeurj.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 14479 – 14485

A New Fungal Cytotoxin
FULL PAPER
[19] L. A. Nafie, R. K. Dukor, T. B. Freedman, Handbook of Vibrational
Spectroscopy, Vol. 1 (Eds.: J. M. Chalmers, P. R. Griffiths), Wiley,
New York, 2002, pp. 731–744.
[20] A. Aamouche, F. J. Devlin, P. J. Stephens, J. Am. Chem. Soc. 2000,
122, 2346–2354.
[21] P. J. Stephens, F. J. Devlin, J.-J. Pan, Chirality 2008, 20, 643–663.
[22] E. Castiglioni, P. Biscarini, S. Abbate, Chirality 2009, 21, E28–E36.
[23] J. Autschbach, Chirality 2009, 21, E116–E152.
[24] G. Pescitelli, T. Kurtꢃn, U. Flçrke, K. Krohn, Chirality 2009, 21,
E181–E201.
[29] T. Gꢅngçr, Y. Chen, R. Golla, Z. Ma, J. Med. Chem. 2006, 49, 2440–
2455.
[30] C. Venturello, G. P. Chiusoli, J. Heterocycl. Chem. 1972, 9, 1155–
1157.
[31] R. Menicagli, S. Samaritani, G. Signore, F. Vaglini, L. D. Via, J. Med.
Chem. 2004, 47, 4649–4652.
[32] H. M. Ge, Z. G. Yu, J. Zhang, J. H. Wu, R. X. Tan, J. Nat. Prod.
2009, 72, 753–755.
[33] H. Cao, T. Fang, S. Li, J. Ma, Macromolecules 2007, 40, 4363–4369.
[34] S. Meng, J. Ma, J. Phys. Chem. B 2008, 112, 4313–4322.
[35] A. Jiemchooroj, P. Norman, J. Chem. Phys. 2007, 126, 134102/1–7.
[36] http://srdata.nist.gov/cccbdb/vibrations.asp.
[37] P. J. Stephens, F. J. Devlin, C. F. Chabalowski, M. J. Frisch, J. Phys.
Chem. 1994, 98, 11623–11627.
[25] S. Abbate, E. Castiglioni, F. Gangemi, R. Gangemi, G. Longhi, Chir-
ality 2009, 21, E242–E252.
[26] H. Shojaei, Z. Li-Bçhmer, P. von Zezschwitz, J. Org. Chem. 2007,
72, 5091–5097.
[27] F. M. Queiroz, M. de F. S. Muniz, M. Menezes, Mycopathologia
2001, 150, 61–65.
[38] T. Brotin, D. Cavagnat, J.-P. Dutasta, T. Buffeteau, J. Am. Chem.
Soc. 2006, 128, 5533–5540.
[28] W. S. Horn, M. S. J. Simmonds, R. E. Schwartz, W. M. Blaney, Myco-
logia 1996, 88, 588–595.
Received: August 1, 2010
Published online: October 29, 2010
Chem. Eur. J. 2010, 16, 14479 – 14485
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
14485