Stereoselective synthesis and structural correction of the naturally occurring lactone stagonolide G

Article abstract of DOI:10.1021/ol102599e

A convergent synthesis of the structure proposed for the naturally occurring lactone stagonolide G is described. All three stereocenters were created with the aid of asymmetric Brown allylations. The lactone ring was built by means of a ring-closing metat

Full text of DOI:10.1021/ol102599e

ORGANIC
LETTERS
2010
Vol. 12, No. 24
5752-5755
Stereoselective Synthesis and Structural
Correction of the Naturally Occurring
Lactone Stagonolide G§
Ce´sar A. Angulo-Pacho´n,† Santiago D´ıaz-Oltra,† Juan Murga,† Miguel Carda,*,†
and J. Alberto Marco*,‡
Departamento de Química Inorga´nica y Orga´nica, UniVersitat Jaume I, Castello´n,
E-12080 Castello´n, Spain, and Departamento de Química Orga´nica, UniVersitat de
Valencia, E-46100 Burjassot, Valencia, Spain
alberto.marco@uV.es
Received October 26, 2010
ABSTRACT
A convergent synthesis of the structure proposed for the naturally occurring lactone stagonolide G is described. All three stereocenters were
created with the aid of asymmetric Brown allylations. The lactone ring was built by means of a ring-closing metathesis (RCM). The synthetic
and the natural compound differed in their spectral properties. A new structure is now proposed for stagonolide G and demonstrated by
means of a chemical transformation.
Stagonospora cirsii Davis is a pathogenic fungus which
grows on the terrestrial plant Cirsium arVense, a noxious
weed, and causes necrotic lesions on its leaves. The
fungus, which has aroused interest as a potential myco-
herbicidal agent, has been found to produce various toxic
metabolites when grown in liquid culture. The first of these
metabolites was reported in 2007 and named stagonolide
(later called stagonolide A). The compound was assigned
the structure depicted in Figure 1 on the basis of its NMR
and other spectroscopic data.1 Stagonolide A was shown
to be a non-host-specific but selective phytotoxin.2
§ Dedicated to the memory of Prof. R. Suau (deceased November 11,
2010).
† Universitat Jaume I.
Figure 1
. Structures of some stagonolides and other structurally
‡ Universitat de Valencia.
related, naturally occurring 10-membered lactones.
(1) Yuzikhin, O.; Mitina, G.; Berestetskiy, A. J. Agric. Food Chem.
2007, 55, 7707–7711.
10.1021/ol102599e  2010 American Chemical Society
Published on Web 11/24/2010

A short time later, further metabolites were isolated from
the same fungus and named stagonolides B-I.3 They also
showed phytotoxic activity against C. arVense, although
except for stagonolide H, the activity was much weaker than
that of stagonolide A. Most particularly, stagonolide G was
essentially inactive. Figure 1 shows the structures of three
of the stagonolides.
Scheme 1. Retrosynthetic Analysis of Structure 1
The similarity of their structures with those of other
bioactive 10-membered lactones of natural origin (see
Figure 1)4 has attracted the attention of several groups,
who have devised stereoselective syntheses for lactones
of this type. As a matter of fact, stagonolides A,5 B,5b,6
C,7 F,8 and G9 have succumbed to total synthesis in the last
2 years.
Homoallyl alcohol 2 was prepared as reported12 by means
of asymmetric Brown allylation of a silylated 4-hydroxy-butanal
(Scheme 1, C with P3 ) TPS). Protection of the hydroxyl group
as its MOM derivative13 was followed by desilylation with
TBAF to yield primary alcohol 4.14 All attempts at direct
oxidation of 4 to acid 6 failed. However, PCC oxidation of
alcohol 4 to aldehyde 5 followed by sodium chlorite oxidation15
gave 6 in a fair overall yield (Scheme 2).
In recent years, we have been involved in the total synthesis
of several medium- and large-ring lactones.10 In the present
paper, we are disclosing our own synthesis of compound 1,
which has the structure proposed for stagonolide G. The
retrosynthetic concept followed is presented in Scheme 1.
Retrosynthetic cleavage of the lactone C-O bond and the
olefinic CdC bond (via ring-closing metathesis, RCM) in 1
leads to fragments A and B. Acid A can be referred via
inverse allylation and functional group interchange to the
protected 4-hydroxybutanal C. The monoprotected diol B
should be amenable to preparation by means of an asym-
metric R-alkoxyallylation11 of acetaldehyde with the chiral
(Z)-γ-alkoxyallylborane D.
Scheme 2
.
Synthesis of Acid 6 (≡ A, P1 ) MOM)a
(2) Berestetskiy, A.; Dmitriev, A.; Mitina, G.; Lisker, I.; Andolfi, A.;
Evidente, A. Phytochemistry 2008, 69, 953–960.
(3) (a) Evidente, A.; Cimmino, A.; Berestetskiy, A.; Mitina, G.; Andolfi,
A.; Motta, A. J. Nat. Prod. 2008, 71, 31–34. (b) Evidente, A.; Cimmino,
A.; Berestetskiy, A.; Andolfi, A.; Motta, A. J. Nat. Prod. 2008, 71, 1897–
1901.
(4) For reviews on various synthetic, biosynthetic, or pharmacological
aspects of medium-ring (8-11 membered) lactones, see: (a) Dra¨ger, G.;
Kirschning, A.; Thiericke, R.; Zerlin, M. Nat. Prod. Rep. 1996, 13, 365–
375. (b) Shiina, I. Chem. ReV. 2007, 107, 239–273. (c) Ferraz, H. M. C.;
Bombonato, F. I.; Longo, L. S., Jr. Synthesis 2007, 3261–3285. (d) Ferraz,
H. M. C.; Bombonato, F. I.; Sano, M. K.; Longo, L. S., Jr. Quim. NoVa
2008, 31, 885–900. (e) Ishigami, K. Biosci. Biotechnol. Biochem. 2009,
73, 971–979.
a Acronyms and abbreviations: TPS, tert-butyldiphenylsilyl; MOM
methoxymethyl; TBAF, tetra-n-butylammonium fluoride hydrate; PCC,
pyridinium chlorochromate; DIPEA, N,N-diisopropylethylamine.
(5) (a) Srihari, P.; Kumaraswamy, B.; Rao, G. M.; Yadav, J. S.
Tetrahedron: Asymmetry 2010, 21, 106–111. (b) Prabhakar, P.; Rajaram,
S.; Reddy, D. K.; Shekar, V.; Venkateswarlu, Y. Tetrahedron: Asymmetry
2010, 21, 216–221. (c) Mohapatra, D. K.; Somaiah, R.; Rao, M. M.; Caijo,
F.; Mauduit, M.; Yadav, J. S. Synlett 2010, 1223–1226. (d) Srihari, P.; Rao,
G. M.; Rao, R. S.; Yadav, J. S. Synthesis 2010, 2407–2412.
(6) (a) Srihari, P.; Kumaraswamy, B.; Somaiah, R.; Yadav, J. S. Synthesis
2010, 1039–1045. (b) Giri, A. G.; Mondal, M. A.; Puranik, V. G.; Ramana,
C. V. Org. Biomol. Chem. 2010, 8, 398–406.
In the only other, very recently published synthesis of
stagonolide G,9 a fragment equivalent to B (P2 ) Bn) was
prepared through a nine-step sequence from D-glucose
diacetonide. In the present synthesis, a shorter and more
efficient sequence has been utilized (Scheme 3). Our first
idea was to lithiate the MOM derivative of allyl alcohol and
then treat the organolithium thus formed with B-methoxy-
diisopinocampheylborane (Ipc2BOMe). The resulting chiral
(7) (a) Mohapatra, D. K.; Dash, U.; Naidu, P. R.; Yadav, J. S. Synlett
2009, 2129–2132. (b) Jana, N.; Mahapatra, T.; Nanda, S. Tetrahedron:
Asymmetry 2009, 20, 2622–2628.
(8) Perepogu, A. K.; Raman, D.; Murty, U. S. N.; Rao, V. J. Bioorg.
Chem. 2009, 37, 46–51.
(9) Srihari, P.; Kumaraswamy, B.; Bhunia, D. C.; Yadav, J. S.
Tetrahedron Lett. 2010, 51, 2903–2905.
(10) (a) Murga, J.; Falomir, E.; Garc´ıa-Fortanet, J.; Carda, M.; Marco,
J. A. Org. Lett. 2002, 4, 3447–3449. (b) Garc´ıa-Fortanet, J.; Murga, J.;
Falomir, E.; Carda, M.; Marco, J. A. J. Org. Chem. 2005, 70, 9822–9827.
(c) Garc´ıa-Fortanet, J.; Carda, M.; Marco, J. A. Tetrahedron 2007, 63,
12131–12137. (d) D´ıaz-Oltra, S.; Angulo-Pacho´n, C. A.; Kneeteman, M. N.;
Murga, J.; Carda, M.; Marco, J. A. Tetrahedron Lett. 2009, 50, 3783–3785.
(e) D´ıaz-Oltra, S.; Angulo-Pacho´n, C. A.; Murga, J.; Carda, M.; Marco,
J. A. J. Org. Chem. 2010, 75, 1775–1778. (f) D´ıaz-Oltra, S.; Angulo-Pacho´n,
C. A.; Murga, J.; Falomir, E.; Carda, M.; Marco, J. A. Chem.-Eur. J., in
press.
(12) Brimble, M. A.; Bryant, C. J. Org. Biomol. Chem. 2007, 5, 2858–
2866. In our hands, 2 was obtained with an enantiomeric purity above 99%
(see the Supporting Information).
(13) Wuts, P. G. M.; Greene, T. W. Greene’s ProtectiVe Groups in
Organic Synthesis; 4th ed.; John Wiley and Sons: New York, 2007; pp
30-38.
(14) Compound 4 has previously been prepared with a lower optical
purity (84%) alongside a different reaction sequence: Takahata, H.; Kubota,
M.; Momose, T. Tetrahedron: Asymmetry 1997, 8, 2801–2810.
(15) The reaction conditions for this oxidation were taken from
ref 6b.
(11) For a review of this specific allylation type, see: Lombardo, M.;
Trombini, C. Chem. ReV. 2007, 107, 3843–3879.
Org. Lett., Vol. 12, No. 24, 2010
5753

Scheme 3
.
Synthesis of Alcohol 10 (≡ B)a
Scheme 4. Fragment Coupling and Last Steps of the Synthesis
a Acronyms and abbreviations: Ipc, isopinocampheyl; EOM, ethoxy-
methyl.
allylborane would then be treated with acetaldehyde.16
However, we faced considerable practical difficulties in
obtaining the MOM derivative of allyl alcohol with an
adequate purity.17 In view of this drawback, we replaced
the MOM group with the structurally similar EOM
group.18,19 Thus, protection of allyl alcohol gave compound
7, which was then lithiated with sec-butyl lithium at low
temperature. The resulting organolithium reagent 8 was
treated at the same temperature with Ipc2BOMe to yield the
chiral (Z)-γ-alkoxyallylborane 9, which was then allowed
to react with acetaldehyde. This gave homoallyl alcohol 10
in 82% yield and 94% ee.20
however, the spectral data of synthetic 1 were found different
from those reported for stagonolide G.3b This was a most
unexpected result in view of the fact that structure 1 was
assumedly confirmed with a total synthesis.9
As an alternative structural possibility, we prepared lactone
15, a stereoisomer of 1, as described in Scheme 5. First, ent-10
Scheme 5. Synthesis of 15, a Stereoisomer of 1
Acid 6 and alcohol 10 were joined using the Yamaguchi
procedure21 (Scheme 4). The resulting ester 11 was then
subjected to RCM22 catalyzed by the second-generation
ruthenium complex Ru-II.23 This gave the 10-membered
lactone 12 in 84% yield as a single Z stereoisomer.24
Simultaneous cleavage of the two protecting groups was
performed under the same mild conditions used previously
for MOM groups10a and yielded lactone 1. Surprisingly,
(16) This reaction has previously been performed with an achiral
allylborane to yield the corresponding homoallyl alcohol in racemic form.
See: Hoffmann, R. W.; Kemper, B.; Metternich, R.; Lehmeier, T. Liebigs
Ann. Chem. 1985, 2246–2260.
(17) This was due to the considerable volatility of this compound, which
made it difficult to eliminate trace amounts of the solvent used in its
preparation, at least at the small scale in which we were working. These
solvent traces led to failure of the lithiation step.
(18) The lower volatility of the EOM derivative of allyl alcohol, as
compared with the MOM derivative, permitted its preparation with an
appropriate purity.
(19) For examples of use of the EOM protecting group in synthesis,
see: (a) Boudier, A.; Hupe, E.; Knochel, P. Angew. Chem., Int. Ed. 2000,
39, 2294–2297. (b) Adrio, J.; Rodr´ıguez-Rivero, M.; Carretero, J. C.
Chem.sEur. J. 2001, 7, 2435–2448. (c) Dakas, P.-Y.; Jogireddy, R.; Valot,
G.; Barluenga, S.; Winssinger, N. Chem.sEur. J. 2009, 15, 11490–11497.
(d) Harris, D. A.; Powers, M. E.; Romesberg, F. E. Bioorg. Med. Chem.
(the enantiomer of 10) was prepared in the same way as the
latter (Scheme 3) but using (+)-Ipc2BOMe in the asymmetric
allylation step. Then, 6 and ent-10 were connected as above
by means of the Yamaguchi procedure to yield ester 13. This
was followed by RCM to 14 and cleavage of the two protecting
groups to yield 15, epimeric of 1 at C-8 and C-9 (for numbering
see Figure 2). Again, lactone 15 proved different from natural
stagonolide G.
We found the fact very intriguing that the only previous,
target-directed synthesis of stagonolide G reported the successful
preparation of the natural compound.9 Our synthetic structure
being secured, we could only explain this result by assuming
that some undetected mistake of the same type was made in
both the initial structure assignment3b and the aforementioned
synthesis. Since the C-H connectivities of the natural com-
pound had been based on 2D NMR one-bond (HSQC) and also
long-range (HMBC) heteronuclear correlations, the carbon
framework should be correctly established.3b However, no
HMBC three-bond through-oxygen correlations (O)C-
Lett. 2009, 19, 3787–3790
.
(20) The experimental procedure followed was adapted from that
reported in: Wang, X.; Porco, J. A., Jr. J. Am. Chem. Soc. 2003, 125, 6040–
6041.
(21) Inanaga, J.; Hirata, K.; Saeki, H.; Katsuki, T.; Yamaguchi, M. Bull.
Chem. Soc. Jpn. 1979, 52, 1989–1993.
(22) For reviews on the uses of this reaction for the synthesis of
macrocycles, see: (a) Prunet, J. Angew. Chem., Int. Ed. 2003, 42, 2826–
2830. (b) Gradillas, A.; Pe´rez-Castells, J. Angew. Chem., Int. Ed. 2006, 45,
6086–6101. (c) Majumdar, K. C.; Rahaman, H.; Roy, B. Curr. Org. Chem.
2007, 11, 1339–1365.
(23) With the first-generation ruthenium catalyst PhCHdRuCl2(PCy3)2,
no reaction took place after 16 h of reflux in CH2Cl2.
(24) It is noteworthy that, except for stagonolide G, in all syntheses of
other stagonolides where RCM was used, the E isomer was the predominant
or the only isomer formed.
5754
Org. Lett., Vol. 12, No. 24, 2010

O-C-H) were presented in the original publication, such
correlations being very useful to confirm the position of lactone
carbonyl groups. Thus, we performed an HMBC experiment
with our synthetic 1 and found such a correlation between the
carbonyl 13C NMR peak at 173.9 ppm and the H-9 signal at δ
5.06 (see the Supporting Information).
Scheme 6. Acid-Catalyzed Isomerization of the 10-Membered
a
Lactone 1 to the Five-Membered Lactone 16
In view of these findings, we wondered whether the ring
closure in the putative seco-acid precursor of stagonolide G
(Figure 2) took place through the hydroxyl at C-8 (i.e., the
a Acronyms and abbreviations: CSA, camphorsulfonic acid.
whereas compound 16, which has the actual structure, has
[R]D +21.8 (c 0.1, CHCl3). A value of [R]D +96 (c 0.1,
CHCl3) was given for stagonolide G in the original
publication,3b and [R]D +7 (c 0.3, CHCl3) was the value
reported for a putative synthetic stagonolide G.9 Our
compound 16 has a high optical purity, as demonstrated by
the chiral HPLC analyses of its two precursors 6 (>99% ee)
and 10 (94% ee). We believe therefore that the aforemen-
tioned literature values are affected by errors likely due to
an inadequate purity.
While an adventitious translactonization in the course of
the chromatographic isolation process may be invoked to
explain the coexistence of a five-membered lactone with the
other 10-membered lactones in S. cirsii,3,26 there is still the
open question that a previous synthesis of stagonolide G was
claimed to give a product identical to the natural product.9
The only reasonable explanation for this is that a translac-
tonization inadvertently took place in the last step of the
synthetic sequence. Indeed, this was a deprotection step in
which two benzyl groups were cleaved under the influence
of TiCl4, a strong Lewis acid which can induce such kinds
of processes, as previously observed.29 It seems that the
translactonization 1 f 16 under such conditions is competi-
tive with, or even faster than, deprotection.
Figure 2. Alternative lactone ring closures available to stagonolide
G seco-acid.
compound would be a nine-membered rather than a ten-
membered lactone) or, perhaps more likely, through the
hydroxyl at C-4 (i.e., a five-membered lactone). This is not
without precedent, as five-membered lactones of this struc-
tural class have been found to coexist in the same organism
with lactones of higher ring sizes.25,26
Although the NMR data of synthetic 1 were different from
those reported for stagonolide G,3b an aged sample of the
synthetic compound displayed additional NMR signals which
were coincident with those of the natural product. Since
CDCl3 may contain traces of HCl, a solution of 1 was stirred
at room temperature in the presence of a catalytic amount
of camphorsulfonic acid. A new lactone 16 was then isolated
in 93% yield with spectral data identical to those reported
for the natural compound (Scheme 6). This indicates that a
translactonization has taken place with formation of the
thermodynamically more stable γ-lactone system.27 The
carbonyl band at 1765 cm-1 in the IR spectrum of 16 is also
diagnostic of a five-membered lactone.28
In summary, the structure of the stagonolide G isolated
from S. cirsii turns out to be markedly different from the
remaining lactones found in the same fungus. Most likely,
the accordingly different molecular shape explains why
stagonolide G, in contrast to the other lactones found in the
fungus, is practically devoid of phytotoxic activity.3b
Visible differences are observed between our optical
rotation values and those reported in the literature. Thus, as
indicated in this paper, compound 1 having the reported
stagonolide G structure has [R]D -24.2 (c 0.55, CHCl3),
Acknowledgment. Financial support has been granted by
the Spanish Ministry of Science and Technology by the
BANCAJA-UJI foundation (Project No. P1-1B-2008-14) and
by the Generalitat Valenciana (Project Nos. ACOMP/2009/
113 and ACOMP/2010/246).
(25) Smith, C. J.; Abbanat, D.; Bernan, V. S.; Maiese, W. M.; Greenstein,
M.; Jompa, J.; Tahir, A.; Ireland, C. M. J. Nat. Prod. 2000, 63, 142–145
.
(26) For natural lactones containing additional free hydroxyl groups,
the question remains open as to whether they all are true natural products
or some of them are the result of translactonization process during the
isolation. In the present case, it is not unlikely that the natural product
actually has structure 1, which then isomerized during the isolation to the
more stable γ-lactone 16.
Supporting Information Available: Experimental pro-
cedures for the preparation and tabulated spectral data of all
new compounds. Graphical NMR spectra of all new com-
pounds. This material is available free of charge via the
Internet at http://pubs.acs.org.
(27) Freshly dissolved samples of 1 in CDCl3 (taken from a bottle opened
a few weeks prior to use) already showed incipient NMR signals of 16
after several minutes at room temperature. After 15 h, about 50% of 1 was
converted into 16. Interestingly, lactone 15 did not show this marked
proclivity to acid-catalyzed translactonization.
(28) Lambert, J. B.; Shurvell, H. F.; Lightner, D. A.; Cooks, R. G.
Organic Structural Spectroscopy; Prentice Hall: Upper Saddle River, 1998;
p 193. An HMBC experiment with 16 further confirmed the presence of
the γ-lactone ring (see the Supporting Information).
OL102599E
(29) Sharma, G. V. M.; Reddy, J. J.; Reddy, K. L. Tetrahedron Lett.
2006, 47, 6531–6535.
Org. Lett., Vol. 12, No. 24, 2010
5755