Structural incongruities of coleophomone natural products: Insights by total synthesis of a semi-synthetic derivative

Article abstract of DOI:10.1016/S0040-4039(03)00615-4

An unexpected structural incongruity between coleophomones 1 and the reported structure of recently isolated natural product 2 was confirmed by total synthesis of key semi-synthetic derivative 3 and its positional isomer 4. In addition, possible mechanism

Full text of DOI:10.1016/S0040-4039(03)00615-4

TETRAHEDRON
LETTERS
Pergamon
Tetrahedron Letters 44 (2003) 3559–3563
Structural incongruities of coleophomone natural products:
insights by total synthesis of a semi-synthetic derivative
Jeffrey W. Bode and Keisuke Suzuki*
Department of Chemistry, Tokyo Institute of Technology and CREST, Japan Science and Technology Corporation O-okayama,
Meguro-Ku, Tokyo 152-8551, Japan
Received 6 February 2003; revised 27 February 2003; accepted 28 February 2003
Abstract—An unexpected structural incongruity between coleophomones 1 and the reported structure of recently isolated natural
product 2 was confirmed by total synthesis of key semi-synthetic derivative 3 and its positional isomer 4. In addition, possible
mechanisms for the interconversion of expected structure 22 to observed compound 2 are postulated, some of which may have
an important role in the biosynthetic origin and chemistry of these structurally unique natural products. © 2003 Elsevier Science
Ltd. All rights reserved.
The coleophomones (1) are structurally unique biologi-
cally active natural products isolated from two different
fungal species by researchers at Shionogi1 and at
Merck.2,3 The related fungal metabolite 2 was recently
reported as a dynamic mixture of at least four constitu-
tional isomers by rapid interconversion via a facile
aldol–retro-aldol reaction, a feature which confounds
spectral analysis.4,5 To address this complication, the
Shionogi researchers prepared the reportedly stable
derivative, ammonia adduct 3, by treatment of natural
2 with NH4HCO3. The structure of this compound was
determined using standard methods and the identity of
2 was inferred from these findings. Intriguingly, the
reported structure of 2 differs from 1 not only in the
absence of cyclization between the prenyl side chain
and the aromatic moiety, but also in the substitution
pattern about the aromatic ring (Scheme 1). Given that
this discrepancy was both surprising and inconsistent
with the probable biosynthesis of these compounds
Scheme 1.
Keywords: coleophomone; natural product; structure confirmation;
cyclocondensation; vinylogous amide.
* Corresponding author. Tel./fax: +81
3 5734 2228; e-mail:
ksuzuki@chem.titech.ac.jp
0040-4039/03/$ - see front matter © 2003 Elsevier Science Ltd. All rights reserved.
doi:10.1016/S0040-4039(03)00615-4

3560
J. W. Bode, K. Suzuki / Tetrahedron Letters 44 (2003) 3559–3563
(vide infra),6 we embarked on a structure determination
via total synthesis.
We now report the total synthesis of both reported
structure 3 and its positional isomer 4, the biosyntheti-
cally expected structure,7 which serves to confirm the
unusual structure of semi-synthetic 3. As well we have
postulated mechanisms for the origin of this unusual
structure.
In order to avoid potential structural rearrangements
during our synthetic efforts, we desired a route allowing
direct access to cyclic aminals 3 and 4 without recourse
to derivation of the tricarbonyl structures. In this
regard, we believed our recently developed synthesis of
fused isoxazoles would offer an ideal entry to the
vinylogous amide functionality under mild conditions
(Scheme 2).8
The development of a flexible, viable route to the target
aminals began with a synthetic approach to A, the
positional isomer corresponding to the expected
coleophomone pattern (Scheme 3). Following precedent
on related systems,9 aldehyde 5 was converted to the
corresponding dioxane acetal, which was treated with
tert-butyllithium in Et2O and the resulting lithium spe-
cies trapped with DMF to afford aldehyde 6. Conver-
sion to the oxime and nitrile oxide formation (NCS,
pyridine)8 set the stage for the critical cyclocondensa-
tion reaction.
Scheme 2. Retrosynthetic analysis of reported structure 3 and
biosynthetically expected isomer 4.
Scheme 3. Reagents and conditions: (a) 1,3-propanediol, CSA, C6H6, reflux; (b) t-BuLi, Et2O, −78°C, then DMF; (c)
NH2OH·HCl, K2CO3, EtOH–H2O; (d) NCS, pyridine, CH2Cl2; (e) NEt3, EtOH, 50°C; (f) LDA, THF–HMPA (6:1), 10, −78°C;
(g) Repeat f, 8:1 regioselectivity; (h) LDA, THF–HMPA (6:1), PhSeBr, −78°C; (i) Mo(CO)6, CH3CN–H2O; reflux, 15 min; (j) aq.
H2O2, CH2Cl2; (k) aq. HCl (1 M), rt, 2 h, THF.

J. W. Bode, K. Suzuki / Tetrahedron Letters 44 (2003) 3559–3563
3561
Despite the sterically encumbered and highly function-
alized nature of 7, we were pleased to find that conden-
sation with diketone 8 proceeded exceptionally well to
afford the desired fused isoxazole 9 in 88% yield.8
Notably, this compound is obtained as a 1:1 mixture of
diastereomers arising from the methyl stereogenic cen-
ter and hindered rotation about the biaryl bond.
In addition to masking the latent vinylogous amide
functionality, the isoxazole serves as a convenient han-
dle for the elaboration of the cyclohexanedione ring.
Although two possible alkylation modes were possible,
selective functionalization was achieved simply by
choice of the reaction conditions. Thus, treatment of 9
with LDA and trapping of the resulting enolate with
prenyl bromide provided 11 in good yield. A second
alkylation, in the presence of HMPA, provided the
desired quaternary bis-prenyl compound 12 in 8:1
regioselectivity for the desired product. In contrast, the
second alkylation performed in the absence of HMPA
afforded positional isomer 13 almost exclusively. The
ability to alkylate either position of the cyclohexane
ring permitted facile introduction of the benzenesele-
nenyl group, affording 14.
Scheme 4. Reagents and conditions: (a) 1,3-propanediol, CSA,
C6H6, reflux; (b) n-BuLi, Et2O, −78°C, then DMF; (c)
NH2OH·HCl, K2CO3, EtOH–H2O; (d) NCS, pyridine,
Selective reduction of the isoxazole to the correspond-
ing vinylogous amide in the presence of the selenenyl
functionality was achieved by brief (0.25 h) exposure of
14 to Mo(CO)6 in refluxing CH3CN–H2O (5:1), which
served to reduce the NꢀO bond of the isoxazole without
effecting CꢀSe bond cleavage. Further treatment of the
unpurified reaction mixture with H2O2 afforded the
unsaturated vinylogous amide 15. Upon deprotection
of the acetal with aqueous acid, targeted cyclic aminal
A was obtained in excellent yield.
i
CH2Cl2; (e) 1.4 equiv. NaOiPr, 1.5 equiv. 8, PrOH, 50°C; (f)
LDA, THF–HMPA, 10, −78°C; (g) Repeat f, 6:1 regioselec-
tivity; (h) LDA, THF–HMPA, PhSeBr, −78°C; (i) Mo(CO)6,
CH3CN–H2O; reflux, 15 min; (j) aq. H2O2, CH2Cl2; (k) aq.
HCl (1 M), rt, 5 min, THF.
contrast, A existed as a complex mixture of stereoiso-
meric amide and imine forms in aprotic solvents
(Scheme 5), but as predominantly one isomer in the
presence of a proton source (i.e. CD3OH). Our initial
difficulties in the characterization of A served as a
reminder of the care which must be taken in assigning
the structure of molecules interconverting between
structurally distinct constitutional isomers.
The general approach to coleophomone structures, thus
established, was directly applicable to the synthesis of
isomer B without modification or optimization of the
reaction conditions (Scheme 4). Known bromide 1710
served as a convenient starting material for the prepara-
tion of aldehyde 18. Following further transformation
to stable nitrile oxide 19, cyclocondensation with 5-
methylcyclohexan-1,3-dione (8) in the presence of
sodium isopropoxide provided key isoxazole 20 in good
yield.8 In contrast to 9, this compound did not show
evidence of atropisomerism.
An identical sequence of reactions employed for the
preparation of A was utilized to convert 20 to cyclic
aminal B, with a marginal decrease in the regioselectiv-
ity of the second alkylation with prenyl bromide as the
only deviation.
Scheme 5. Compound A, a mixture of imines, vinylogous
amides, atropisomers, and geometrical isomers in aprotic
solvents.
Comparison of the H and 13C NMR spectra of syn-
1
These results served to confirm, but not explain, the
aberrant structure of compound 3 and, by association,
naturally occurring 2. The biosynthesis of 1 and 2,
which are fungal metabolites, likely proceeds via oxida-
tive cleavage of the corresponding anthracene (Scheme
6). This analysis predicts the resulting benzophenone-
like core to have the ortho,ortho% disubstitution pattern
seen in 1.
thetic aminals A and B with the data reported for
semi-synthetic 3 confirmed that compound B was identi-
cal to 3, confirming that the reported, unusual structure is
indeed correct.11 In fact, despite the structural similar-
ity, the spectral properties of A and B differ signifi-
cantly. The spectral data of synthetic B showed a single
compound in numerous solvents and exactly matched
the spectra of a semi-synthetic, authentic sample.12 In

3562
J. W. Bode, K. Suzuki / Tetrahedron Letters 44 (2003) 3559–3563
In conclusion, we have described a novel synthetic
entry to the core structures of the coleophomone natu-
ral products, cumulating in the structure confirmation
of semi-synthetic derivative 3 by total synthesis. These
studies contribute to the growing interest in these struc-
turally unusual molecules and provide
a novel
Scheme 6. Postulated biosynthesis of coleophomone natural
products.
approach likely to be useful for continued synthetic
studies and for the preparation of designed analogs.
Further studies to employ this chemistry towards con-
clusive structure proof of compound 2 and the origin of
the structural incongruity are underway.
Although the incongruent structure of 2 could indeed
arise from specific oxidative enzymes which introduce
hydroxyl functionalities at this position, we have postu-
lated several mechanisms for the structural interconver-
sion of the biosynthetically expected structure 22 and
the reported structure 2 (Schemes 7 and 8).
Acknowledgements
We are grateful to the Shionogi & Co. for authentic
samples of 2 and 3, and to Professor Shosuke Yama-
mura and Professor Katsumi Kakinuma for helpful
discussions. Yoshifumi Hachisu is thanked for experi-
mental assistance. J.W.B. thanks the Japan Society for
the Promotion of Science for a postdoctoral fellowship.
This research was partially supported by the 21st Cen-
tury COE program.
References
1. (a) Kamigauchi, T.; Nakajima, M.; Hiroyoshi, T. Jpn.
Kokai Tokkyo Koho 1998, 10101666; (b) Human chymase
inhibitors, IC50=0.7 mM.
2. (a) Wilson, K. E.; Tsou, N. N.; Guan, Z.; Ruby, C. L.;
Pelaez, F.; Gorrochategui, J.; Vicente, F.; Onishi, H. R.
Tetrahedron Lett. 2000, 41, 8705–8708; (b) Bacterial cell
wall transglycosylase inhibitors, IC50=62 mM.
3. For the first total synthesis of coleophomone B and its
naturally occurring (Z)-isomer, see: Nicolaou, K. C.;
Vassilikogiannakis, G.; Montagnon, T. Angew. Chem.,
Int. Ed. 2002, 41, 3276–3279.
Scheme 7. Postulated origin of reported structure 2 from
expected compound 22 via pseudo-symmetric enol 23.
Scheme 7 illustrates a potential structural interconver-
sion via pseudo-symmetrical enol 23, which could be
formed from expected isomer 22 under acidic, basic, or
photochemical conditions. Indeed, precedent exists for
similar isomerization under a variety of conditions.13
Tautomerization of 23 would either regenerate isomer
22 or, alternatively, give rise to reported isomer 2.
4. (a) Kamigauchi, T.; Nakajima, M.; Hiroyoshi, T. Jpn.
Kokai Tokkyo Koho 1999, 11158109; (b) Human chymase
inhibitors, IC50=1.4 mM.
5. During the preparation of this manuscript, Nicolaou has
reported a total synthesis of 2 and has designated it as
Coleophomone D: Nicolaou, K. C.; Montagnon, T.;
Vassilikogiannakis, G. J. Chem. Soc., Chem. Commun.
2002, 2478–2479.
Likewise, a similar rearrangement via a transannular
hydride shift (Scheme 8) of 22a would also result in a
structural interconversion affording reported com-
pound 2a.14
6. Natori, S. In Natural Products Chemistry; Nakanishi, K.;
Goto, T.; Ito, S.; Natori, S.; Nozoe, S., Eds. Carboaro-
matic and Related Compounds. Kodansha Academic
Press: Tokyo, 1975; Vol. 2, pp. 131–254.
7. Portions of this work have been presented as a poster at
the IUPAC 14th International Conference on Organic
Synthesis (ICOS) at Christchurch, New Zealand, July
14–19, 2002.
8. (a) Bode, J. W.; Hachisu, Y.; Matsuura, T.; Suzuki, K.
Org. Lett. 2003, 5, 391–394; (b) Bode, J. W.; Hachisu, Y.;
Matsuura, T.; Suzuki, K. Tetrahedron Lett. 2003, 44,
3555–3558.
9. Hollinshead, S. P.; Nichols, J. B.; Wilson, J. W. J. Org.
Chem. 1994, 59, 6703.
Scheme 8. Potential structural interconversion via hydride
shift.
Given that these or related processes may be quite
facile, the question of whether the unusual structure of
3 is biosynthetically determined, the result of a sponta-
neous rearrangement, or an artifact of the isolation and
purification procedure remains to be established.
10. Couture, A.; Deniau, E.; Grandclaudon, P.; Hoarau, C.
J. Org. Chem. 1998, 63, 3128–3132.

J. W. Bode, K. Suzuki / Tetrahedron Letters 44 (2003) 3559–3563
3563
1
11. Spectral data for fully synthetic 3 and 4: (a) 3: H NMR
(400 MHz, CD3OD, 50°C) l 7.42–7.34 (m, 2H), 7.06 (dd,
1H, J=7.5, 1.0 Hz), 6.06 (s, 1H), 5.98 (s, 1H), 4.83–4.79
(m, 1H), 4.74–4.71 (m, 1H), 3.78 (s, 3H), 2.59–2.52 (m, 2H),
2.43–2.35 (m, 2H), 1.89 (d, 3H, J=1.2 Hz), 1.63 (s, 3H),
1.61 (s, 3H), 1.60 (s, 3H), 1.55 (s, 3H); 13C NMR (100 MHz,
CD3OD) l 201.8, 189.1, 167.9, 158.7, 157.0, 136.7, 135.0,
133.7, 132.1, 130.2, 123.0, 120.1, 116.3, 108.4, 84.0, 57.8,
56.3, 38.0, 25.9, 19.4, 18.0; IR (thin film) 3198, 2920, 1653,
1.57 (br s, 6H), 1.50 (br s, 6H); 13C NMR (100 MHz,
CD3OD) l 173.9, 159.3, 156.0, 137.0, 135.5, 134.4, 130.1,
119.7, 115.9, 114.3, 56.2, 51.1, 25.9, 19.3, 18.1 (some
carbons not visible due to molecular motions).
12. Authentic 3, prepared by researchers at Shionogi, was used
for comparison. We are grateful to Dr. Toshiyuki Kami-
gauchi for a generous sample of this compound.
13. (a) Acidic conditions: Moskalev, N. V.; Gribble, G. W.
Tetrahedron Lett. 2002, 43, 197–202; (b) Basic conditions:
Colson, J. G.; Lansbury, P. T.; Saeva, F. D. J. Am. Chem.
Soc. 1967, 89, 4987–4990.
14. For an example of base-induced 1,4-transannular hydride
shift in a hydroxy-cyclohexanone, see: Warnhoff, E. W. J.
Chem. Soc., Chem. Commun. 1976, 517–518.
1
1610, 1523, 1424, 1269, 1227, 1051 cm−1. (b) 4: H NMR
(400 MHz, CD3OD, 50°C) l 8.23–8.20 (m, 1H), 7.46–7.41
(m, 1H), 7.23–7.41 (m, 1H), 7.23–7.19 (m, 1H), 6.25–6.22
(m, 1H), 6.18 (s, 1H), 4.81–4.77 (m, 2H), 3.93–3.91 (m, 3H),
2.83–2.77 (m, 2H), 2.38–2.33 (m, 2H), 1.97–1.95 (m, 3H),