Avinosol, a meroterpenoid-nucleoside conjugate with antiinvasion activity isolated from the marine sponge Dysidea sp.

Article abstract of DOI:10.1021/ol061333p

The new meroterpenoids avinosol (1), 3′-aminoavarone (2), and 3′-phenethylaminoavarone (3) have been isolated from the marine sponge Dysidea sp. collected in Papua New Guinea, and their structures were elucidated by analysis of spectroscopic data. Avinoso

Full text of DOI:10.1021/ol061333p

ORGANIC
LETTERS
2006
Vol. 8, No. 17
3749-3752
Avinosol, A Meroterpenoid-Nucleoside
Conjugate with Antiinvasion Activity
Isolated from the Marine Sponge
Dysidea sp.
|
Ana R. Diaz-Marrero,† Pamela Austin,‡,§ Rob Van Soest, Teatulohi Matainaho,
Calvin D. Roskelley,§ Michel Roberge,‡ and Raymond J. Andersen*,†
Departments of Chemistry, Earth & Ocean Sciences,
Biochemistry & Molecular Biology, and Cellular & Physiological Sciences,
UniVersity of British Columbia, VancouVer, British Columbia, Canada V6T 1Z1;
Institute for Systematics and Ecology, UniVersity of Amsterdam, 1090 GT Amsterdam,
The Netherlands; and Discipline of Pharmacology,
School of Medicine and Health Sciences, UniVersity of Papua New Guinea,
N.C.D., Papua New Guinea
randersn@interchange.ubc.ca
Received May 31, 2006
ABSTRACT
The new meroterpenoids avinosol (1), 3′-aminoavarone (2), and 3′-phenethylaminoavarone (3) have been isolated from the marine sponge
Dysidea sp. collected in Papua New Guinea, and their structures were elucidated by analysis of spectroscopic data. Avinosol (1), which is
apparently the first example of a naturally occurring meroterpenoid-nucleoside conjugate, showed antiinvasion activity in a cell-based assay.
Angiogenesis and metastasis are pivotal steps in the lethal
progression and spreading of solid tumors.1 Both processes
involve the invasion and migration of either vascular endo-
thelial cells or tumor cells through the extracellular matrix,
and as a consequence, inhibition of tissue invasion is an
attractive target for developing anticancer drugs. As part of an
ongoing program designed to find new antiangiogenic and
antimetastatic agents,2 we have screened a library of crude
marine invertebrate extracts in a cell-based assay3 that detects
compounds capable of inhibiting invasion of the extracellular
matrix but are not overtly cytotoxic. Crude extracts of the
sponge Dysidea sp. collected in Papua New Guinea showed
promising activity in the assay. Bioassay-guided fractionation
(2) (a) Williams, D. E.; Craig, K. S.; Patrick, B.; Mc Hardy, L. M.; van
Soest, R.; Roberge, M.; Andersen, R. J. J. Org. Chem. 2002, 67, 245-258.
(b) McHardy, L. M.; Sinotte, R.; Troussard, A.; Sheldon, C.; Church, J.;
Williams, D. E.; Andersen, R. J.; Dedhar, S.; Roberge, M.; Roskelley, C.
D. Cancer Res. 2004, 64, 1468-1474. (c) Warabi, K.; McHardy, L. M.;
Matainaho, L.; Van Soest, R.; Roskelley, C. D.; Roberge, M.; Andersen,
R. J. J. Nat. Prod. 2004, 67, 1387-1389. (d) McHardy, L. M.; Warabi, K.;
Andersen, R. J.; Roskelley, C. D.; Roberge, M. Mol. Cancer Therap. 2005,
4, 772-778. (e) Williams, D. E.; Austin, P.; Marrero, A. R.-D.; Van Soest,
R.; Matainaho, T.; Roskelley, C. D.; Roberge, M.; Andersen, R. J. Org.
Lett. 2005 7, 4173-4176.
† Depts. of Chemistry and Earth & Ocean Sciences, UBC.
‡ Dept. of Biochemistry and Molecular Biology, UBC.
§ Dept. of Cellular and Physiological Sciences, UBC.
| University of Amsterdam.
University of Papua New Guinea.
(1) (a) Sahai, E.; Marshall, C. J. Nat. Cell Biol. 2003, 5, 711-719. (b)
Friedl, P. Curr. Opin. Cell Biol. 2004, 16, 14-23. (c) Wolf, K.; Mazo, I.;
Leung, H.; Engelke, K.; von Andrian, U. H.; Deriyugina, E. I.; Strongin,
A. Y.; Brocker, E.-B.; Friedl, P. J. Cell Biol. 2003, 160, 267-277.
(3) Roskelley, C.; Williams, D. E.; McHardy, L. M.; Leong, K.; Karsan, K.;
Andersen, R. J.; Dedhar, S.; Roberge, M. Cancer Res. 2001, 61, 6788-6794.
10.1021/ol061333p CCC: $33.50
© 2006 American Chemical Society
Published on Web 07/20/2006

identified the new meroterpenoid-nucleoside conjugate avino-
sol (1) as the antiinvasive compound in the extract. Two addi-
tional new meroterpenoids, the selective cytotoxin 3′-amino-
avarone (2) and the phenethylamine analogue 3, were also
present in the Dysidea sp. extract along with the known
meroterpenoids avarone (4) and avarol (5).4 Details of the
isolation and structure elucidation of 1, 2, and 3 and the
biological activities of 1 and 2 are presented below.
Specimens of Dysidea sp. (113 g wet wt) were collected
by hand using SCUBA in Papua New Guinea, frozen on site,
and transported to Vancouver over dry ice. Thawed sponge
specimens were exhaustively extracted with MeOH (3 × 500
mL), and the combined MeOH extracts were concentrated
in vacuo to give a dark red residue (4.9 g). The residue was
chromatographed on a reversed-phase Amberchrom HP-20
column (step gradient from H2O to MeOH) to give six
fractions. A fraction eluted with 100% MeOH showed
antiinvasion activity. Sephadex LH-20 chromatography (elu-
ent MeOH) of the antinvasive material (622 mg) gave a small
group of active fractions that were combined (105 mg) and
further purified via reversed-phase flash chromatography
(Waters Sep-pak, 10 g, step gradient from 3:7 H2O/CH3CN
to CH3CN) to give pure avinosol (1, 2.0 mg) and a mixture
of other compounds. This mixture gave 3′-aminoavarone (2,
0.9 mg), the phenethylamine analogue 3, avarone (4), and
avarol (5) after further reversed-phase HPLC fractionation
(CSC-Inertsil 150A/ODS2; 7:3 CH3CN/H2O).
2.5, Hz; H-4′). HMBC correlations observed between δ 6.75
(H-6′), 6.57 (H-4′), and 6.55 (H-4′) and carbon resonances
at 146.0 (C-2′) and 151.2 (C-5′), in conjunction with the
upfield shift of the proton resonances, were consistent with
the presence of a hydroquinone ring in 1. The existence of
only meta coupling between the proton resonances assigned
to the hydroquinone (H-6′ and H-4′) required that the
sesquiterpenoid fragment (C-11) and one other substituent
were both ortho to one of the phenol funtionalities.
Avinosol (1) was obtained as a colorless glass that gave a
[M + Na]+ ion at m/z 587.2847 in the HRESIMS consis-
tent with a molecular formula of C31H40N4O6 (calcd for
C31H40N4O6Na, 587.2846) requiring 14 sites of unsaturation.
The 13C NMR spectrum of avinosol (1) showed more than the
31 resonances expected from the HRMS analysis, and many of
the resonances appeared to be twinned, suggesting the exis-
tence of two slowly interconverting forms of the molecule.
Similarly, a number of resonances in the 1H spectrum of 1 (δ
8.37, s, H-8′′; 8.36, s, H-8′′; 8.07, s, H-2′′; 8.14, s, H-2′′; 6.57,
d, J ) 2.5 Hz, H-4′; 6.55, d, J ) 2.5, Hz, H-4′) integrated
for one-half of a proton relative to an upfield methyl singlet
(δ 0.87, Me-12) used as an internal standard, and each of these
resonances was paired with a second resonance of identical
multiplicity that also integrated for one-half a proton.
Analysis of the 1D and 2D NMR data obtained for
avinosol (1) identified a sesquiterpenoid fragment (C-1 to
C-15) identical to that found in the co-occurring metabolites
avarone (4) and avarol (5). HMBC correlations observed
from the H-11 (δ 2.70 and 2.77) terpenoid resonances to
carbon resonances at δ 132.0 (C-1′), 146.0 (C-2′) and 122.4
(C-6′) demonstrated that the sesquiterpenoid fragment in 1
was linked to an aromatic fragment as in 5 (Figure 1 and
Supporting Information). The carbon resonance at δ 122.4
(C-6′) was correlated in the HSQC spectrum to a proton
resonance at δ 6.75 (d, J ) 2.5 Hz; H-6′) that showed meta
coupling in the COSY spectrum to the twinned proton
resonances at δ 6.57 (d, J ) 2.5 Hz; H-4′) and 6.55 (d, J )
An isolated 1H spin system extending from the deshielded
methine at δ 6.47 (H-1′′′) to the geminal methylene protons
at δ 3.75 (H-5′′′) and 3.82 (H-5′′′) was identified from the
COSY, HSQC, and HMBC data. This spin system was
assigned to a 2-deoxyfuranopentose fragment (C-1′′′ to C-5′′′)
on the basis of the 2D NMR data. Subtracting the atoms
present in the sesquiterpenoid (C15H25), hydroquinone
(C6H4O2), and 2-deoxyfuranose (C5H9O3) fragments from the
molecular formula of 1 showed that the remaining fragment
had to account for C5H2N4O and 6 sites of unsaturation.
The 1H resonances assigned to the remaining nitrogenous
fragment all integrated for 0.5 protons each, were very
deshielded (δ 8.37, H-8′′; 8.36, H-8′′; 8.07, H-2′′; 8.14,
H-2′′), and were correlated in the HSQC spectrum to
relatively deshielded carbon resonances (δ 8.37 and 8.36 to
141.0; 8.07 and 8.14 to 150.4 and 150.3, respectively), all
suggestive of a purine substructure. HMBC correlations were
(4) (a) Minale, L.; Riccio, R.; Sodano, G. Tetrahedron Lett. 1974, 15,
3401-3404. (b) deRosa, L.; Minale, L.; Riccio, R.; Sodano, G. J. Chem.
Soc., Perkin Trans. 1 1976, 1408-1414.
Figure 1. Selected HMBC correlations observed for 1.
3750
Org. Lett., Vol. 8, No. 17, 2006

observed from the 1H resonances at δ 8.36 and 8.37 (H-8′′)
to nonprotonated carbon resonances at δ 126.0 (C-5′′) and
148.7 (C-4′′) and to the furanose anomeric carbon at 86.6
(C-1′′′), and from the furanose anomeric proton at δ 6.47
(H-1′′′) to the nonprotonated carbon at 148.7 (C-4′′) and the
C-8′′ carbon at 141.0, consistent with attachment of the de-
oxypentose fragment via the anomeric carbon C-1′′′ to N-9′′
of the purine, typical of a purine nucleoside. The remaining
proton resonances assigned to the purine residue at δ 8.07
and 8.14 (H-2′′) showed HMBC correlations to carbon reson-
ances at δ 148.7 (C-4′′) and 158.7 (C-6′′), which completed
the 13C NMR assigment of the purine fragment. Comparison
of the 13C chemical shift assignments for this fragment to
literature values for the purine residues of known oxygenated-
purine nucleosides revealed a near identical resemblance to
N-1 alkylated inosines.5 This suggested that the constitution
of avinosol (1) involved a bond between N-1′′ of the purine
fragment and C-3′ of the hydroquinone fragment.
Scheme 1. Synthesis of Avinosol (1) and 3′-Aminoavarone
(2) from Avarol (5)
the C-6′′ carbonyl might play an important role in creating
the barrier to rotation about the N-1′′/C-3′ bond in avinosol
(1).
2D NOESY correlations were observed between H-1′′′ (δ
6.47) and both H-4′′′ (δ 4.04) and H-2R′′′ (δ 2.48), and be-
tween H-2â′′′ (δ 2.77) and H-3′′′ (δ 4.57), consistent with a
ribose linked â to N-9′′. The NOESY data also confirmed that
the relative configurations in the sesquiterpenoid fragment
were identical to those in avarol (5), thereby completing the
proposed structure for 1 as shown. It is assumed that twinning
of the proton resonances assigned to H-4′, H-2′′ and H-8′′,
and several of the resonances assigned to carbon atoms in this
region of the molecule, is due to the existence of atropisomers
resulting from hindered rotation around the N1′′-C-3′ bond.
The H-4′ (δ 6.57 and 6.55) and H-2′′ (δ 8.07 and 8.14) reson-
ances show the greatest difference in chemical shifts in the two
isomers, in agreement with this assumption. An absence of any
twinning of the 1H resonances assigned to the ribose fragment
supports the N-1′′ to C-3′ linkage rather than the alternative
N-3′′ to C-3′ linkage. Elevated temperature NMR experi-
ments undertaken in an attempt to measure the magnitude
of the N-3′′/C-3′ rotational barrier led to decomposition of
avinosol (1) before the coalescence temperature was reached.
3′-Aminoavarone (2) was isolated as a red oil that gave a
[M + Na]+ ion at m/z 350.2097 in the HRESIMS appropriate
for a molecular formula of C21H29NO2 (calcd for C21H29-
NO2Na, 350.2096). Analysis of the NMR data obtained for
2 (Supporting Information) readily identified the 3′ substi-
tuted avarone substructure found in the semisynthetic
compound 6. The molecular formula of 2 required that the
3′ substituent was an amino group. To confirm the structure
of 2, avarone (4) was reacted with Me3SiN3 in refluxing
EtOH for 24 h to give a mixture of 3′-aminoavarone (2) and
4′-aminoavarone, that could be separated by HPLC.7 The
synthetic 3′-aminoavarone (2) was identical by TLC, NMR,
and MS comparison with the natural product.
The 3′-phenethylaminoavarone analogue 3 was isolated
as a red oil that gave a [M + Na]+ ion at m/z 454.2726 in
the HRESIMS consistent with a molecular formula of C29H37-
NO2 (calcd for C29H37NO2Na, 454.2722). The NMR data for
3 (Supporting Information) also contained resonances that
could be assigned to the 3′-substituted avarone substructure
found in avarone (4) and 3′-aminoavarone (2). An isolated
1H spin system consisting of two scalar coupled methylene
resonances (δ 2.93, t, J ) 7 Hz, H-3′′; 3.35, q, J ) 7 Hz,
H-2′′) and a broad one proton resonance at δ 5.67 (NH-1′′)
was identified in the COSY data. The δ 5.67 resonance was
correlated to the H-2′′ resonance in the COSY spectrum but
not correlated to a carbon resonance in the HSQC spectrum.
A second isolated spin system in the 1H NMR spectrum could
be assigned to a monosubstituted phenyl ring (δ 7.21, d, J
) 7.3 Hz, H-5′′/H-9′′; 7.24, t, J ) 7.3 Hz, H-7′′; 7.32, t, J
) 7.3 Hz, H-6′′/H-8′′). In the HMBC spectrum, the H-2′′ (δ
3.35) and H-3′′ (δ 2.93) resonances were both correlated to
the C-4′′ resonance at δ 139.0, demonstrating that C-3′′ was
attached to the phenyl ring and, therefore, that N-1′′ was
attached to C-3′ of the avarone fragment as shown in 3.
The pure meroterpenoids 1, 2, 4, 5, and 6 were tested in
the antiinvasion assay against two human tumor cell lines
To confirm the proposed structure of avinosol (1), the nat-
ural product was synthesized from avarone (4) and 2′-deoxy-
inosine as shown in Scheme 1. Avarone (4) was prepared in
quantitative yield by oxidation of naturally occurring avarol
(5),6 obtained from the Dysidea sp. extract, with MnO2 in
Et2O at room temperature for 10 min. Reaction of avarone
(4) with 2′-deoxyinosine in DMF and K2CO3 at room temper-
ature for 30 min gave avinosol (1) in 22% yield.5 The syn-
thetic material was identical to the natural product by TLC,
NMR, and MS comparison, confirming the proposed constitu-
tion and the absolute configuration shown. Synthetic avinosol
(1) was determined to have [R]22D ) -26.9° (c 0.35, MeOH).
Avinosol (1) slowly oxidized to the corresponding quinone
avinosone (6) upon exposure to air. There was no evidence
for atropisomers in the NMR data obtained for 6. This
suggests that hydrogen bonding between the C-2′ phenol and
(5) Narukulla, R.; Shuker, D. E. G.; Xu, Y.-Z. Nucleic Acids Res. 2005,
33, 1767-1778.
(6) Ling, T.; Xiang, A. X.; Theodorakis, E. A. Angew. Chem., Int. Ed.
1999, 38, 3089-3091.
(7) Cozzolino, R.; De Giulio, A.; De Rosa, S.; Strazzullo, G.; Gasic, M.
J.; Sladic, D.; Zlatovic, J. Nat. Prod. 1990, 53, 699-702.
Org. Lett., Vol. 8, No. 17, 2006
3751

with distinct modes of invasion. MDA-MB-231 breast cancer
cells utilize the mesenchymal mode of invasion (path-
generating) whereas LS174T colon carcinoma cells invade
in an amoeboid manner (path-finding).1 Avinosol (1) had
an IC50 of ∼50 µg/mL in the antiinvasion assay against both
cell lines.8 Avarone (4), avarol (5), and avinosone (6) were
only active in the assay at concentrations of g100 µg/mL.
3′-Aminoavarone (2) did not show significant antiinvasion
activity but did show differential cytotoxic effects on the
cell lines used in the assay. It was found to be 10-fold more
toxic to the MDA-MB-231 cell line (IC50 ∼2 µg/mL)
compared to the LS174T cell line (IC50 ∼20 µg/mL). Further
studies on the antiinvasive activity of avinosol (1) and the
cytotoxic activity of 3′-aminoavarone (2) are ongoing in our
laboratories.
nucleoside. The putative biogenesis of avinosol represents
four distinct biosynthetic pathways: terpenoid, acetate/or
shikimate, purine, and carbohydrate. It appears that nature
has constructed avinosol (1) in an efficient convergent
manner by first utilizing the common secondary metabolite
pathway to the meroterpenoid fragment and a primary
metabolite pathway to a nucleoside and then taking advantage
of the intrinsic reactivities of the p-quinone in avarone and
the purine ring in deoxyinosine to link them together.
Acknowledgment. Financial support was provided by the
National Cancer Institute of Canada (R.J.A., C.D.R.), the
Canadian Institutes of Health Research (M.R.), and the
Michael Smith Foundation for Health Research. The authors
thank M. LeBlanc (UBC-EOS) for assistance collecting
Dysidea sp. and Kathryn Westendorf (UBC-Cell Biology)
for technical assistance. A.R.D.-M. was supported by a
postdoctoral fellowship from the Spanish MEC (Ministerio
de Educacio´n y Ciencia; Secretar´ıa de Estado de Univer-
sidades e Investigacio´n).
There are a number of naturally occurring derivatives of
avarone and other meroterpenoids where the aromatic/
quinone fragment is substituted with simple amines and
amino acids.9 However, to the best of our knowledge,
avinosol (1) is the first example of a natural product from
any source that is a conjugate of a meroterpenoid and a
Supporting Information Available: Experimental sec-
tion and tabulated NMR data and NMR spectra for avinosol
(1), 3′-aminoavarone (2), 3′-phenethylaminoavarone (3), and
avinosone (6). This material is available free of charge via
the Internet at http://pubs.acs.org.
(8) The meroterpenoid strongylophorine-26 has an IC50 of ∼1 µg/mL in
the same assay. See ref 2c.
(9) (a) Alvi, K. A.; Diaz, M. C.; Crews, P.; Slate, D. L.; Lee, R. H.;
Moretti, R. J. Org. Chem. 1992, 57, 6604-6607. (b) Rodriguez, J.; Quinoa,
E.; Riguera, R.; Peters, B. M.; Abrell, L. M.; Crews, P. Tetrahedron 1992,
48, 6667-6680. (c) Kobayashi, J.; Madono, T.; Shigemori, H. Tetrahedron
1995, 51, 10867-10874, and d) Cimino, G.; De Rosa, S.; Cariello, L.;
Zanetti, L. Experientia 1982, 38, 896.
OL061333P
3752
Org. Lett., Vol. 8, No. 17, 2006