Structural reassignment of cytosporolides A-C via Biomimetic synthetic studies and reinterpretation of NMR Data

Article abstract of DOI:10.1021/ol202181g

A structure revision for the recently isolated fungal meroterpenoids, cytosporolides A-C, is suggested based on biosynthetic speculation and reinterpretation of existing spectroscopic data. The structure revision is supported by a biomimetic synthetic stu

Full text of DOI:10.1021/ol202181g

ORGANIC
LETTERS
2011
Vol. 13, No. 19
5318–5321
Structural Reassignment of
Cytosporolides AꢀC via Biomimetic
Synthetic Studies and Reinterpretation of
NMR Data
Justin T. J. Spence and Jonathan H. George*
Discipline of Chemistry, School of Chemistry & Physics, University of Adelaide,
North Terrace, Adelaide SA 5005, Australia
jonathan.george@adelaide.edu.au
Received August 11, 2011
ABSTRACT
A structure revision for the recently isolated fungal meroterpenoids, cytosporolides AꢀC, is suggested based on biosynthetic speculation and
reinterpretation of existing spectroscopic data. The structure revision is supported by a biomimetic synthetic study, featuring a [4 þ 2]
cycloaddition reaction between a presumed o-quinone methide intermediate and β-caryophyllene.
Cytosporolides AꢀC (1ꢀ3, Figure 1) are three meroter-
penoids isolated from Cytospora sp., a fungus found in a
soil sample collected at high altitude on the Tibetan
plateau.1 1ꢀ3 Showed modest antibiotic activity against
the Gram-positive bacteria Staphylococcus aureus and
Streptococcus pneumonia. The stuctures 1ꢀ3 were pro-
posed on the basis of NMR studies by Che et al. The most
striking feature of the proposed structures is a highly
unusual 9-membered peroxylactone ring, which is fused to
a caryophyllene-derived bicyclo[7.2.0]undec-4-ene system.
Furthermore, an aromatic ring is embedded within the
9-membered peroxylactone ring of the proposed structures
1ꢀ3. The molecular framework of the proposed structures
is thus analogous to a [6]metacyclophane. [6]Metacyclo-
phanes are relatively unstable and difficult to synthesize
due to ring strain, which forces the aromatic ring toadopt a
nonplanar conformation. The proposed cytosporolide struc-
tures 1ꢀ3 would thereforebe expected tobehighlystrained
molecules. Although several cyclophanes have been isolated
as natural products, such as cavicularin2 and haouamine
A,3 we believed that the highly strained peroxylactone ring
system of the proposed structures of the cytosporolides
warranted further investigation and, possibly, structural
reassignment.4
Cytosporolides AꢀC were coisolated from cytospora sp.
with the known caryophyllene derived natural product
fuscoatrol5 (4), which has been previously isolated from a
marine fungus and whose structure has been unambigu-
ously assigned via X-ray crystallography. Che et al. assigned
the structures 1ꢀ3 for the cytosporolides AꢀC based on
NMR studies and proposeda biosynthesisinvolving cyclo-
addition of a polyketide derived peroxy radical cation spe-
cies to fuscoatrol. In our view, this biosynthetic proposal is
(2) Toyota, M.; Yoshida, T.; Kan, Y.; Takaoka, S.; Asakawa, Y.
Tetrahedron Lett. 1996, 37, 4745.
(3) Garrido, L.; Zubia, E.; Ortega, M. J.; Salva, J. J. Org. Chem. 2003,
68, 293.
(4) For a review of structural reassignment of natural products, see:
Nicolaou, K. C.; Snyder, S. A. Angew. Chem., Int. Ed. 2005, 44, 1012.
(5) Smetanina, O. F.; Kuznetsova, T. A.; Gerasimenko, A. V.;
Kalinovsky, A. I.; Pivkin, M. V.; Dmitrenok, P. C.; Elyakov, G. B.
Russ. Chem. Bull. 2004, 53, 2643.
(1) Li, Y; Niu, S; Sun, B; Liu, S; Liu, X; Che, Y Org. Lett. 2010, 12,
3144.
r
10.1021/ol202181g
Published on Web 09/02/2011
2011 American Chemical Society

Figure 2. Some related caryophyllene-derived natural products.
Figure 1. Originally proposed structures of cytosprolides AꢀC
and fuscoatrol.
require an extra CH2 group in the long alkyl chain instead
of the C-24 Me group of 1ꢀ3. This fits the NMR data
of Che et al., who assigned the C-24 Me group of 1 to a
broad singlet at δH 1.28 in the 1H NMR spectrum. This
resonance is very low for an aryl Me group. Furthermore,
the δH 1.28ꢀ1.30 region of the 1H NMR spectrum is heavily
congested due to overlapped CH2 resonances of the akyl
side chain, so an extra CH2 group in this region, rather
than an aryl Me group, is a more plausible assignment.
Close inspection of the 2D NMR spectra of Che et al. for
cytosporolide A reveals that the claimed HMBC correla-
tion between the proposed C-24 Me group at 1.28 ppm
and C-22 at 151.0 ppm has been erroneously reported.
However, the same HMBC spectrum does show a distinct
correlation between the C-7 proton at 3.08 ppm and C-22,
which would only be possible for our reassigned struc-
ture 8. This HMBC correlation would be impossible for
Che’s originally proposed structure 1. The remaining struc-
tural assignments for cytosporolide A made by Che et al.
appear to be sound on the basis of their 1D and 2D NMR
data, with the overall bond connectivity elucidated by COSY,
HMQC, and HMBC spectra, and with the relative stereo-
chemistry assigned by NOESY spectra.
The revised structures 8ꢀ10 are further supported by
the very close similarity between the aromatic regions of
the 13C NMR spectra of cytosporolides AꢀC and berkelic
acid9 (11), a spiroketal natural product isolated from
an extremophilic Penicillium fungus found in the Berkeley
Pit Lake.
The originally proposed cytosporolide structures were
speculated to arise in Nature via an unusual cycloaddi-
tion of a peroxy radical cation to fuscoatrol derivatives to
generate the 9-membered peroxylactone ring. The revised
implausible, and the proposed structures 1ꢀ3 are unlikely
to be correct due to the high degree of ring strain present in
the peroxylactone ring system.
A key part of the assignment by Che et al. involved the
chemical shift of C-8 (δC 87.5), which was observed to be
significantly downfield compared to C-8 (δC 74.2) of the
related caryophyllene-derived terpenoid 6-hydroxypunc-
taporonin(5, Figure2). Thisdifferenceinchemicalshift led
Che et al. to propose a peroxide bond between C-8 and
C-23. However, no further evidence for the peroxylactone
ring was given. Based in part on biosynthetic speculation,
we considered that a more likely structure for the cytos-
porolides AꢀC might contain a 6-membered aryl ether
ring, similar to those found in the caryophyllene-derived
natural products guajadial6 (6) and psidial A7 (7), rather
than the 9-membered peroxylactone ring proposed by Che
at al. Indeed, the 13C NMR chemical shifts of the oxyge-
nated C-4 carbon atoms of guajadial (δC 84.3) and psidial
A (δC 88.0) are very similar to that of C-8 of cytosporolide
A (δC 87.5). The biosynthesis of guajadial (6) and psidial
A (7) was proposed to involve a [4 þ 2] cycloaddition of an
o-quinone methide to β-caryophyllene, a pathway that has
recently been supported by an elegant biomimetic synth-
esis by Lee et al.8
Our suggested revised structures of cytosporolides AꢀC
(8ꢀ10), with 6-membered aryl ether rings replacing the
9-membered peroxylactone rings of 1ꢀ3, are shown in
Figure 3. These structures correlate well with the NMR
data of Che et al. Importantly, the revised structures
(6) Yang, X.-L.; Hsieh, K.-L.; Liu, J.-K. Org. Lett. 2007, 9, 5135.
(7) Fu, H.-Z.; Luo, Y.-M.; Li, C.-J.; Yang, J.-Z.; Zhang, D.-M. Org.
Lett. 2010, 12, 656.
(8) Lawrence, A. L.; Adlington, R. M.; Baldwin, J. E.; Lee, V.;
Kershaw, J. A.; Thompson, A. L. Org. Lett. 2010, 12, 1676.
(9) Stierle, A. A.; Stierle, D. B.; Kelly, K. J. Org. Chem. 2006, 71,
5357.
Org. Lett., Vol. 13, No. 19, 2011
5319

Scheme 1. Proposed Biosynthesis of Cytosporolide A
Figure 3. Revised structures of cytosporolides AꢀC and berkelic
acid.
addition of β-caryophyllene (22), the reaction mixture
was stirred at 100 °C for 18 h, which generated the cyclo-
adduct 23 as a single diasteroisomer in 53% yield. Pre-
sumably this three-component cascade reaction proceeds
via initial formation of the cyclic acetal 20, followed by eli-
mination of EtOH to give the o-quinone methide inter-
mediate 21. [4 þ 2] cycloaddition of 21 to 22 would then
generate 23 via either a pericyclic DielsꢀAlder reaction or
perhaps a stepwise ionic process. Overall this reaction
forms two new 6-membered rings and three new stereo-
centers in one step, with complete stereoselectivity. Basic
hydrolysis of the methyl ester of 23 then gave carboxylic
acid 24 in 57% yield.
Importantly, the 13C NMR chemical shift of C-8 of 24
was found to be 88.6 ppm, which is in good agreement with
the value recorded for C-8 of the cytosporolides. Further-
more, the aromatic region of the 13C NMR spectrum of 24
shows an excellent correlation with the cytosporolide
spectra (see Supporting Information for a full comparison).
The relative configuration of 24 was established via X-ray
crystallography.14 The stereochemical configurations at
C-8, C-9, and C-16 of 23 were found to be opposite to that
found in the cytosporolides. This fact can be rationalized
by invoking addition of o-quinone methide 21 to either
the RR or βR conformations of β-caryophyllene,15 whereas
in the biosynthesis of cytosporolide A the o-quinone
methide 13 might add to the favored ββ conformation of
fuscoatrol.5 The IR spectrum of the carboxylic acid 24
structures of the cytosporolides could be biosynthesized by
a far more feasible pathway (Scheme 1). In this case,
dehydration of the known fungal metabolite CJ-12,37310
(12) could generate the o-quinone methide11 13. Cycload-
dition of this reactive intermediate to fuscoatrol would
then generate the revised structure of cytosporolide A (8).
A related biosynthesis of berkelic acid (11), involving
cycloaddition of an o-quinone methide to an enol ether,
has been the subject of a recent biomimetic synthesis by
De Brabander et al.12
As part of our continuing interest in biomimetic reac-
tions of o-quinone methides,13 and to lend support to the
proposed biosynthesis of the revised cytosporolide struc-
tures, we conducted a brief biomimetic synthetic study
(Scheme 2). Aryl triflate 15 was prepared via reaction of
methyl 2,4,6-trihydroxy benzoate with Tf2O according to a
published procedure.12 Stille coupling of 15 with tributyl-
vinyltin gave 16 in good yield (75%), which was then
benzylated under standard conditions to give 17 (91%).
Epoxidation of 17 with mCPBA in CH2Cl2 gave 18 in 55%
yield. Hydrogenation of 18 then simultaneously cleaved the
benzyl ethers and ring-opened the epoxide to give 19 (86%).
Alcohol 19 was then treated with TFA and an excess
of HC(OEt)3 at room temperature (Scheme 3). Following
(10) Inagaki, T.; Kaneda, K.; Suzuki, Y.; Hirai, H.; Nomura, E.;
Sakakibara, T.; Yamauchi, Y.; Huang, L.; Norcia, M.; Wondrack,
L. M.; Sutcliffe, J. A.; Kojima, N. J. Antibiot. 1998, 51, 112.
(11) For a review of o-quinone methide chemistry, see: Van De
Water, R. W.; Pettus, T. R. Tetrahedron 2002, 58, 5367.
(12) Bender, C. F.; Yoshimoto, F. K.; Paradise, C. L.; De Brabander,
J. K. J. Am. Chem. Soc. 2009, 131, 11350.
(13) (a) George, J. H.; Baldwin, J. E.; Adlington, R. M. Org. Lett.
2010, 12, 2394. (b) George, J. H.; Hesse, M. D.; Baldwin, J. E.;
Adlington, R. M. Org. Lett. 2010, 12, 3532.
(14) Crystallographic data for the structure of 24 have been deposited
with the Cambridge Crystallographic Data Centre (CCDC 829815).
Copies of these data can be obtained free of charge via www.ccdc.cam.
ac.uk/data_request/cif.
(15) Clericuzio, M.; Alagona, G.; Ghio, C.; Toma, L. J. Org. Chem.
2000, 65, 6910.
5320
Org. Lett., Vol. 13, No. 19, 2011

Scheme 2. Synthesis of o-Quinone Methide Precursor 19
Scheme 3. Synthesis of 24 via a Biomimetic [4 þ 2] Cycloaddi-
tion
shows a CdO stretch at 1694 cmꢀ1. This is in good agree-
ment with the IR data for the cytosporolides AꢀC, which
were reported by Che et al. to have CdO stretches in the
range 1690ꢀ1694 cmꢀ1. Furthermore, berkelic acid (11)
has an identical carboxylic acid functional group to 24 and
the revised cytosporolide structures (8ꢀ10), and it has a
CdO stretch of 1693 cmꢀ1. This excellent correlation in IR
data between the model compound 24, berkelic acid, and
the cytosporolides is good evidence that all of these com-
pounds contain an aryl carboxylic acid with an ortho-
hydroxy substituent.
compounds guajadial (6) and berkelic acid (11). Finally, a
biomimetic cycloaddition of an o-quinone methide onto
β-caryophyllene generated a simplified cytosporolide ana-
logue via a three-component cascade reaction.
Acknowledgment. We thank Dr. Chris Sumby and Mr.
Witold Bloch (University of Adelaide) for assistance with
X-ray crystallography studies and Prof. Yongsheng Che
(Institute of Microbiology, Chinese Academy of Sciences)
for the provision of NMR spectra.
In conclusion, we have proposed that the structures of
the cytosporolides AꢀC have been incorrectly assigned,
and we suggest that the revised structures 8ꢀ10 more
accurately represent these natural products. In part, this
reassignment was inspired by biosynthetic considerations.
Re-evaluation of NMR data supported the argument, with
Supporting Information Available. Synthetic proce-
dures and analytical data for compounds 14ꢀ19 and 23.
This material is available free of charge via the Internet at
http://pubs.acs.org.
1
instructive comparisons made between the H and 13C
chemical shift data for the cytosporolides with the related
Org. Lett., Vol. 13, No. 19, 2011
5321