Lambertellols A and B, Novel 3,4-Dihydronaphthalen-1(2H)-ones with Spiro-butenolide Produced by Lambertella sp.1346

Article abstract of DOI:10.1021/ol035889d

(Matrix presented) Lambertella sp. 1346 was found to produce lambertellols A (1) and B (2) carrying a novel dihydronaphthalen-1(2H)-ones with spiro-1-furan-2(5′H)-one. The spiro-lactone ring moiety of both 1 and 2 were easily migrated to afford lambertell

Full text of DOI:10.1021/ol035889d

ORGANIC
LETTERS
2
004
Vol. 6, No. 2
57-160
Lambertellols A and B, Novel
,4-Dihydronaphthalen-1(2H)-ones with
1
3
Spiro-butenolide Produced by
Lambertella sp.1346
Takanori Murakami, Yoshifumi Morikawa, Masaru Hashimoto,*
Toshikatsu Okuno, and Yukio Harada
Faculty of Agriculture and Bioscience, Hirosaki UniVersity,
3-Bunkyo-cho, Hirosaki, 036-8561, Japan
hmasaru@cc.hirosaki-u.ac.jp
Received September 28, 2003
ABSTRACT
Lambertella sp. 1346 was found to produce lambertellols A (1) and B (2) carrying a novel dihydronaphthalen-1(2H)-ones with spiro-1-furan-
(5′H)-one. The spiro-lactone ring moiety of both 1 and 2 were easily migrated to afford lambertellin, a known metabolite of Lambertella
corni-maris. The absolute stereochemistry of these compounds was established on the basis of CD spectrum after chemical derivatization.
2
In 1990, Harada, one of the present authors, reported that
discomycete Monilinia fructigena infection on apple fruits
is sectored by Lambertella sp. 1346 under ambient atmo-
active fractions obtained right after elution from HPLC
[Waters µ-bondasphere C18, 150 × 19 mm i.d. CH CN/H O
3
2
(20:80, without TFA), 10 mL/min flow, t ) 18 min (for
R
1
sphere, which is recognized as so-called mycoparasitism.
1), 25 min (for 2)] were colorless solutions, but gradual
This observation led to a hypothesis that some chemical
substances of L. sp. 1346 may play an important roll in this
mycoparasitism. The ensuing chemical investigation discov-
ered lambertellols A (1) and B (2), a diastereomeric pair of
unique 4,8-dihydroxy-2,3,4-trihydronaphthalene-1-one car-
rying spiro-butenolide at the C3 position. Now, we would
like to report their structures, absolute stereochemistry,
chemical properties, and some basic knowledge on their
biosynthesis.
yellow colorization was observed. Lyophilization of each
fraction after removal of acetonitrile keeping below 25 °C
1
gave 1 and 2 with sufficient purity (>95% based on H
NMR) both as pale yellow amorphous powders. Recrystal-
lization from CHCl -hexane afforded pure 1 (mp 161 °C,
3
dec.) and 2 (mp 107 °C, dec.) both as colorless needles.
However, when the recrystallized products were dissolved
in MeOH, both solutions of 1 and 2 colorized gradually.
Similar colorization was observed by adsorbing 1 and 2 each
on silica gel. These conditions efficiently transformed them
into the same red needles, which were identical to lamber-
tellin (3) (mp, NMR, IR, and MS spectra). Lambertellin (3)
After L. sp. 1346 was cultured employing potato-sucrose
medium for 5 days at 26 °C with shaking, isolation was
performed with guidance of growth inhibition of spores using
2
Cochlibolus miyabeanus in place of Monilinia because of
3
was originally isolated from Lambertella corni-maris and
technical difficulty in preparing spores of Monilinia. The
4
another fungus, Ciboria gordonii Funk. The yields of 1 and
(
1) Harada, Y.; Sasaki, M. Rep. Tottori Mycol. Inst. 1990, 28, 275-
85.
2) Nihei, K.; Itoh, H.; Hashimoto, K.; Miyairi, K.; Okuno, T. Biosci.,
Biotechnol., Biochem. 1998, 62, 852-857.
2 varied by culture (1, 2-15 mg; 2, 5-20 mg from 4 L of
2
(
(3) Armstrong, J. J.; Turner, W. B. J. Chem. Soc. 1965, 5927-5930.
1
0.1021/ol035889d CCC: $27.50 © 2004 American Chemical Society
Published on Web 12/17/2003

Table 1. NMR Spectral Data for Lambertellols A (1) and B (2), Triol 5, and Dibenzoate 6 in CDCl
3
lambertellol A (1)
δH (J in Hz)
lambertellol B (2)
δH (J in Hz)
triol 6
dibenzoate 7
δH (J in Hz)
position
δC
HMBC
δC
HMBC
δH (J in Hz)a
1
2
198.79
43.64
2
199.16
43.7
2
4
5.16 (4.4, 5.8)
2.17 (3.9, 13.7)
2.26 (5.4, 13.9)
6.52 (2.9, 5.8)
2.47 (2.9, 14.7)
2.64 (5.8, 14.7)
2.93 (17.1)
2.94 (17.5)
3.19 (17.5)
3
.22 (17.1)
3
4
87.37
70.83
2, 4, 4′
2, 5
86.82
71.91
2, 4, 4′
2, 5
4.88
2.47
4.88
2.64
4.49
4.92 (6.4)b
2.89 (6.4)b
4
4
5
6
7
8
8
8
2
3
4
-OH
a
141.41
118.40
137.70
118.69
162.68
4, 6
4, 7
141.58
118.29
137.75
118.70
162.79
4, 6
4, 7
7.10 (7.3)
7.56 (7.3, 8.2)
7.01 (8.2)
7.06 (7.8)
7.55 (7.8, 8.3)
7.00 (8.3)
6.66 (7.8)
7.08 (7.8, 7.8)
6.93 (7.8)
not assigned
7.54 (7.8, 7.8)
not assigned
5, 8-OH
6, 8-OH
5, 8-OH
6, 8-OH
-OH
a
′
′
′
12.08
12.09
115.02
172.37
132.32
147.74
4, 5, 7, 8-OH
114.72
172.34
132.54
147.54
4,5,7,8-OH
4′, 6′
4′, 6′
4′, 6′
4′, 6′
6′
2.91 (m)
1.86 (2.9, 13.7),
3.06 (m)
6.97 (1.5)
1.97 (1.5)
7.09 (1.5)
1.96 (1.5)
2, 4, 6′
1.83 (2.9, 13.2),
2.79 (3.4, 13.2)
1.19 (7.3)
2
.56 (3.9, 13.7)
3
′-Me
10.76
4′
10.78
1.12 (7.3)
a
Observed in CD3OD. b Coupling with C4-OH and the signal at C4-OH disappeared upon addition of D2O.
medium) due to this type of decomposition. In some cases,
a considerable amount of 3 was obtained even though the
crude extract exhibited a trace amount of 3 on HPLC.
Fortunately, however, both 1 and 2 withstood the conditions
not provide crucial information as to whether this compound
was spiro-butenolide 2 or pyrone 4.
3
used for the NMR measurements in CDCl .
Figure 1. HMBC Correlation observed in 2.
The structural analysis of lambertellol B (2) was carried
out as follows. The molecular formula of 2 was established
as C14H O by EI-mass spectra. A characteristic fragment
12 5
The chemical shift of the benzylic methine proton on C4
4.88 ppm) suggested that this carbon was attached to alcohol
(
or ether but not to the ester function. This interpretation is
signal at m/z ) 242 (M-18) suggested the existence of an
1
supported by the H NMR of benzoate 7, which was prepared
alcohol function, which was also supported by a strong
1
for CD analysis (vide infra). The H NMR of 6 showed the
-
1
1
adsorption at 3490 cm in the IR spectrum. The H NMR
methine proton on C4 (attached to alcohol) at 4.92 ppm and
the methine proton on C1 (attached to ester) at 6.52 ppm.
Thus, pyrone 4 was eliminated, and we could reveal a novel
dihydronaphthalen-1(2H)-one framework with spiro-buteno-
lide 2. Benzo[g]chromene 5 was also a candidate since 2
was easily transformed to 3 under mild condition. However,
neither HMBC nor NOE experiments supported this structure
as shown in Figure 2.
spectrum displayed signals corresponding to a methyl (δ
1
7
.96), three aromatic (δ 7.00, 7.06, and 7.55), an olefinic (δ
.09), and a phenolic proton (δ 12.09) (see Table 1), which
suggested a structure similar to 3. In addition to these proton
signals, an alcoholic (δ 2.64), a methylene (δ 2.94 and 3.19,
AB), and a methine proton (δ 4.88, s) were observed. The
1
3
C NMR showed 14 carbons, the same number as 3;
The H and 13C NMR spectra of 1 were quite analogous
1
however, the spectral profiles were totally different. Detailed
HSQC and HMBC studies revealed almost of all correlations
between protons and carbons, but the HMBC spectrum did
to those of 2 as shown in Table 1. Each corresponding signal
1
13
appeared within differences of 0.17 ( H) and 0.55 ppm ( C).
Furthermore, HMBC correlations of 1 were also almost same
as those of 2. These results suggest that 1 and 2 are
diastereomeric to each other regarding the configurations at
(
4) Poulton, G. A.; Bushnel, G. W.; Yun-Long, L. Can. J. Chem. 1992,
7
0, 2688.
158
Org. Lett., Vol. 6, No. 2, 2004

Figure 4. NOEs observed in 6.
Figure 2. Eliminated structures 4 and 5.
derivatization. Benzoylation under usual conditions pro-
ceeded at C1 and C8 hydroxy groups, giving dibenzoate 7
selectively.
The CD spectrum of 7 in MeOH gave a pair of typical
exciton-split Cotton effects with a positive Cotton effect at
the C3 and C4 positions. Irradiation of a signal for the C4
proton of 1 induced NOEs at C2H, C5H, and C4′H. Similar
NOEs for C2HTC4H and C4HTC5H were detected in 2;
however, that for C4HTC4H′ was not detected. Accordingly,
the relative stereochemisties of 1 and 2 could be established
as (3S*,4S*) and (3R*,4S*), respectively (Figure 3.).
2
-
42 nm (∆ꢀ242 +7.9) and negative one at 226 nm (∆ꢀ226
6.9), suggesting a positive chirality. The tetrahydronaph-
thalene moiety of 7 was unlikely to interact with the benzoyl
group in the CD spectrum, since UV absorption of 6 around
230 nm was weak enough (log ꢀ ) 3.1 at 230 nm) compared
to the benzoate 7 (log ꢀ ) 4.5 at 230 nm). However, one
benzoyl group of 7 was attached to the C8 phenol function,
which leaves an ambiguity in the conformational profile of
this molecule. Thus, molecular modeling was performed.
6
Conformation search of 7 employing AM1 semiempirical
7
molecular orbital method found 12 local minimum confor-
mations within 5 kcal/mol steric energy from that of global
minimum conformation. All of the conformations provided
by these calculations took type A or type B conformations
Figure 3. Stereochemistry and NOEs of 1 and 2.
8
about the dibenzoate moiety (Figure 5). Accordingly, we
The absolute stereochemistries of 1 and 2 were studied
5
next. We supposed that the 3,4-dihydro-4-hydroxynaphtha-
len-1(2H)-one system caused the instability of the natural
products. Thus, reduction of the C1-carbonyl group was
attempted. Reduction of 2 proceeded to provide triol 6 in
2
55% yield by catalytic hydrogenation using PtO . The
stereochemistry of 6 was established on the basis of detailed
difference NOE experiments. Irradiation of a signal for C2H
induced NOEs at C1H and C4H signals. These results
revealed the cis relationship between C1 and C4 hydroxy
groups. The stereochemistry of the C3′-methyl group was
also established by observing NOEs for C2HTC4′H and
Figure 5. Typical conformations about the benzoate moiety of 7.
9
were able to apply the dibenzoate rule to assign the absolute
C4′HTC3′CH
3
. Triol 6 was stable enough for the next
stereochemistry of the newly introduced hydroxyl group at
C1 as the (S)-configuration. The stereochemistry of C3 and
C4 positions in the 2 was concluded to be (3R,4S)- by taking
Scheme 1
(
5) Unfortunately, single crystals suitable for X-ray crystarographic
analysis could not be obtained.
6) Conformation search was performed employing Pc Spartan Pro
version 1.08 (Wavefunction, Inc: Irvine, CA).
7) Dewar, M. J. S.; Zoebisch, E. G.; Hearly, E. F.; Stewart, J. J. P. J.
Am. Chem. Soc. 1985, 107, 3902.
8) When a MMFF molecular force field was employed, the geometries
(
(
(
of the two benzoate chromopores were similar to those obtained by AM1
in all of the stable conformations, but the planes of two benzene rings have
a tendency to be more parallelized.
(
9) Koreeda, M.; Harada, N.; Nakanishi, K. J. Chem. Soc., Chem.
Commun. 1969, 548-549.
Org. Lett., Vol. 6, No. 2, 2004
159

into account that the relationship of C1 and C4 alcohol
functions was cis.
Interestingly, when L. corni-maris, isolated in Aomori,
Japan, was cultivated for 5 days under the same conditions,
3 was detected as the major component along with small
amounts of 1 and 2. (1:2:3 ) 7:15:78). Thus, L. corni-maris
definitely produces 3 by another biosynthetic pathway. This
biosynthetic difference between L. sp. 1346 and Lambertella
c. has remained unclear and should be investigated in detail.
We observed that L. corni-maris also brings about similar
mycoparasitism against Monilinia fructigena on apple fruit.1
This difference might provide clues for this phenomenon.
It was found that 1-3 all exhibited remarkable growth
inhibition of spores of Cochlibolus miyabeanus (IC50 ) 0.5,
0.5, 3.0 µg/mL, respectively). They also showed weak
cytotoxity against P388 murine leukemia (IC50 ) 12 µg/mL
for 1, 15 µg/mL for 2, and 15 µg/mL for 3). However, it is
difficult to discuss their biological properties because of their
instabilities.
It is noteworthy that the two-dimensional silica gel TLC
analysis revealed an interconversion between 1 and 2. This
tautomerism can be explained by retro-Michael reaction of
the C3 carboxy group giving carboxylic acid 8 followed by
re-Michael addition of the produced hydroxy group from
the opposite site of the cyclohexane ring plane as shown in
Scheme 2. These transformations should be reversible, and
1
can be converted to 2 in the same manner. On the basis of
these considerations, we concluded that because the isomer-
ization occurred at the C3 position, the absolute stereochem-
istry of 1 is (3S,4S)-. Both 1 and 2 might be produced
nonenzymatically, and the real biosynthetic metabolite might
be 8.
Scheme 2
As described, we have succeeded in obtaining unique
metabolites 1 and 2 from mycoparasitic fungus L. sp. 1346
and found that both showed potent antifungal activity. Taking
the ready interconversion between 1 and 2 into account, we
propose that 8 might be the real genetic metabolite. Some
of the metabolites may play important roles in mycopara-
sitism of Lambertella species against Monilinia fructigena.
Acknowledgment. We are grateful to Dr. Kaoru Yamada
(Nagoya University) for cytotoxic assay. We would also like
to thank Professor Akira Kawamura (Hunter College) for
fruitful discussions about CD analysis and for reviewing the
manuscript.
Note Added after ASAP Posting. There was an error in
Figure 5 in the version posted ASAP December 17, 2003;
the corrected version was posted December 19, 2003.
Formation of lambertellin (3) from 1 and 2 is discussed
next. The intermediate 8 could be also susceptible to air
oxidation, giving quinone 9. Then, 9 is transformed into 3
by another intramolecular Michael addition at the C2
position, followed by air oxidation. So far, 8-10 have not
been detected in our experiment. Our studies may indicate
that 3 is an artifact in the case of L. sp. 1346. Actually,
Supporting Information Available: Experimental pro-
cedure, physical data of 1 and 2, and spectra of 1-3, 6, and
7
. This material is available free of charge via the Internet
at http://pubs.acs.org.
3
cultivation for long periods of time (22 days) provided 3 as
OL035889D
the major component (1:2:3 ) 7:14:79, based on the peak
intensity at 280 nm), while HPLC analysis after 5 days
indicated that 1 and 2 were the main metabolites involving
(
10) These experiments could be performed only using static conditions,
because cultutivation with shaking increased the viscosity of the media and
the shaker became ineffective after 5 or 6 days. We employed the shaking
conditions for production of 1 and 2.
1
0
trace amounts of 3 (1:2:3 ) 32:57:11) in contrast.
160
Org. Lett., Vol. 6, No. 2, 2004