Synthesis and structure of two new mycolactones isolated from M. ulcerans subsp. shinshuense

Article abstract of DOI:10.1021/ol302072b

Two new mycolactones, mycolactones S1 and S2, were isolated from culture agar of Mycobacterium ulcerans subsp. shinshuense. Their structures were established in a three-step procedure: (1) probable structures were speculated from MS analysis; (2) candidates were synthesized; (3) HPLC profiles were established for identification of the natural products. Newly isolated mycolactones correspond to the oxidized forms of mycolactone A/B, the causative toxin of Buruli ulcer, isolated from Mycobacterium ulcerans.

Full text of DOI:10.1021/ol302072b

ORGANIC
LETTERS
2012
Vol. 14, No. 17
4618–4621
Synthesis and Structure of Two New
Mycolactones Isolated from M. ulcerans
subsp. shinshuense
Sudhir M. Hande,† Yuko Kazumi,‡ W. George Lai,§ Katrina L. Jackson,† Shinji Maeda,‡
and Yoshito Kishi*,†
Department of Chemistry and Chemical Biology, Harvard University, 12 Oxford Street,
Cambridge, Massachusetts 02138, United States, Department of Mycobacterium
Reference and Research, The Research Institute of Tuberculosis, 3-1-24 Matsuyama,
Kiyose, Tokyo, Japan, and Eisai, Inc., Andover, Massachsetts 01810, United States
kishi@chemistry.harvard.edu
Received July 26, 2012
ABSTRACT
Two new mycolactones, mycolactones S1 and S2, were isolated from culture agar of Mycobacterium ulcerans subsp. shinshuense. Their
structures were established in a three-step procedure: (1) probable structures were speculated from MS analysis; (2) candidates were
synthesized; (3) HPLC profiles were established for identification of the natural products. Newly isolated mycolactones correspond to the
“oxidized forms” of mycolactone A/B, the causative toxin of Buruli ulcer, isolated from Mycobacterium ulcerans.
Buruli ulcer is a severe and devastating skin disease
caused by Mycobacterium ulcerans infection, yet it is one
of the most neglected diseases.1 Infection with M. ulcerans
results in painless necrotic lesions that, if untreated, can
extend to15% of a patient’s skin surface. Surgical excision/
skin grafting has been the only method for treatment of
Buruli ulcer. Encouragingly, most patients have been
reported to respond to combination treatments with ri-
fampin and streptomycin.2
strains of M. ulcerans.3 Various in vitro and in vivo
studies demonstrated that mycolactones A and B are
the causative toxins of Buruli ulcer.4
The gross structure of mycolactones A and B was
elucidated with use of spectroscopic methods, and the
stereochemistry was predicted via the NMR database
approach and confirmed by total synthesis.5ꢀ7 Under
standard laboratory conditions, mycolactones A and B
exist as a 3:2 equilibrating mixture and are referred to as
mycolactone A/B in this paper. Mycolactones C and D
In 1999, Small and co-workers isolated polyketide-
derived macrolide mycolactones A and B from West African
(3) George, K. M.; Chatterjee, D.; Gunawardana, G.; Welty, D.;
Hayman, J.; Lee, R.; Small, P. L. C. Science 1999, 283, 854.
(4) George, K. M.; Pascopella, L.; Welty, D. M.; Small, P. L. C.
Infect. Immun. 2000, 68, 877.
† Harvard University.
‡ The Research Institute of Tuberculosis.
§ Eisai, Inc.
(5) Gunawardana, G.; Chatterjee, D.; George, K. M.; Brennan, P.;
Whittern, D.; Small, P. L. C. J. Am. Chem. Soc. 1999, 121, 6092.
(6) (a) Benewoitz, A. B.; Fidanze, S.; Small, P. L. C.; Kishi, Y. J. Am.
Chem. Soc. 2001, 123, 5128. (b) Fidanze, S.; Song, F.; Szlosek-Pinaud,
M.; Small, P. L. C.; Kishi, Y. J. Am. Chem. Soc. 2001, 123, 10117.
(7) (a) Song, F.; Fidanze, S.; Benowitz, A. B.; Kishi, Y. Org. Lett.
2002, 4, 647. Tetrahedron 2007, 63, 5739. (b) Jackson, K. L.; Li, W.;
Chen, C.-L.; Kishi, Y. Tetrahedron 2010, 66, 2263.
(1) For general reviews on Buruli ulcer and mycolactones, see: (a)
Asiedu, K., Scherpbier, R., Raviglione, M., Ed. Buruli ulcer: Mycobac-
terium ulcerans infection; World Health Organization: Geneva, Switzerland,
2000. (b) Hong, H.; Demangel, C.; Pidot, S. J.; Leadlay, P. F.; Stinear, T. Nat.
Prod. Rep. 2008, 25, 447. (c) Kishi, Y. Proc. Natl. Acad. Sci. U.S.A. 2011,
108, 6703.
(2) For example, see a review: Converse, P. J.; Nuermberger, E. L.;
Almeida, D. V.; Grosset, J. H. Future Microbiol. 2011, 6, 1185.
r
10.1021/ol302072b
Published on Web 08/27/2012
2012 American Chemical Society

were isolated from different strains of human M. ulcerans
(Figure 1).8,9 Fish and frog pathogenic mycobacteria are
also known to produce mycolactone-like metabolites.
30 min at rt and centrifuged (3500 rpm). The organic layer
was separated and concentrated under reduced pressure to
give a lipid extract.
On fluorogenic TLC analysis of the lipid extract, two
distinct greenish-blue fluorescent spots, characteristic to
mycolactone A/B, were detected; one of them, designated
as mycolactone S1, had an Rf value higher than mycolac-
tone A/B, whereas the other, designated as mycolatone S2,
had an Rf value very similar to mycolactone A/B.12 Two
metabolites were isolated by preparative TLC (approxi-
mately 0.6 and 0.4 μg from one Φ = 9.0 cm culture dish,
respectively) and subjected to the structure study.
Figure 1. Structures of the mycolactones isolated from different
strains of human pathogen M. ulcerans.
Structurally, all of the mycolactones reported to date are
composed of a 12-membered macrolactone, referred to as
mycolatone core, and a highly unsaturated fatty acid side
chain. Interestingly, the macrolactone core is conserved in
all of the members in the mycolactone class of natural
products. On the other hand, there is a remarkable struc-
tural diversity observed in the unsaturated fatty acid
portion, including the length of fatty acid backbone, the
degree of unsaturation, the degree of hydroxylation, the
stereochemistry of hydroxylation, and others.1c
We recently reported a fluorogenic chemosensor that
allows us to detect as low as 2 ng of mycolactone A/B in a
semiquantitative manner.10 We are interested in devel-
oping this method as (1) a cost and time effective kit to
detect M. ulcerans infection and (2) a simple tool for
mycolactone-based chemotaxonomy of mycobacteria. As
the first case study of exploring the second possibility, we
chose M. ulcerans subsp. shinshuense. This strain was iso-
lated from a skin ulcer patient, identified as M. ulcerans
subsp. M. shinshuense, registered as JATA 753, maintained
on Middlebrook 7H9 broth at ꢀ80 °C at Mycobacterium
Reference Center, Research Institute of Tuberculosis, Tokyo,
Japan, and used for the present study.11
Liquid inocula were prepared with Middlebrook 7H10
medium (OADC) and used to seed soft agar in a culture
dish (Φ = 9.0 cm). The mycobacteria was grown at 28 °C
for 6ꢀ8 weeks and harvested. After colonies were put into
distilled water, harvested bacteria were warmed at 98 °C
for 15 min and cooled. After a 2:1 mixture of chloroform
and methanol was added, the mixture was stirred for
Figure 2. Distinct ions found in the MS/MS spectrum of myco-
lactones S1 and S2 and probable structures.12
High-resolution mass spectroscopy has revealed the
molecular formula of mycolactones S1 and S2 to be
C44H68O9 and C44H68O10, which correspond to “myco-
lactone A/B ꢀ 2H” and “mycolactone A/B ꢀ 2H þ O”,
respectively.12 In the MS/MS analysis of [molecular ion þ
Na], both mycolactones S1 and S2 gave m/z = 429 ion as
the base peak, indicating that the mycolactone core is
conserved in these metabolites (Figure 2). Consistent with
this conclusion, the ions representing the unsaturated
fatty acid moieties changed from m/z = 359 (mycolactone
A/B) to m/z = 357 (mycolactone S1) and m/z = 373
(mycolactone S2), respectively.
The MS/MS analysis gave additional information to
speculate further the structure of mycolactones S1 and S2.
Unlike mycolactone A/B, both metabolites gave a distinct
ion at m/z = 705. We postulated that it was formed via a
McLafferty rearrangement of the C150-ketone with a
H-shift from the C130-OH (Figure 2). From these analyses,
C150-ketone ofmycolactoneA/B (S1) emerged asthe likely
(8) For mycolactone C, see the following. (a) Isolation and gross
structure: Mve-Obiang, A.; Lee, R. E.; Portaels, F.; Small, P. L. C.
Infect. Immun. 2003, 71, 774. (b) Stereochemistry and total synthesis:
Judd, T. C.; Bischoff, A.; Kishi, Y.; Adusumilli, S.; Small, P. L. C. Org.
Lett. 2004, 6, 4901.
(9) For the structure of mycolactone D, see: Hong, H.; Spencer, J. B.;
Porter, J. L.; Leadlay, P. F.; Stinear, T. ChemBioChem 2005, 6, 643.
(10) Spangenberg, T.; Kishi, Y. Chem Commun 2010, 46, 1410.
(11) Kazumi, Y.; Ohtomo, K.; Takahashi, M.; Mitarai, S.; Sugawara,
I.; Izumi, J.; Andoh, A.; Hasegawa, H. Kekkaku 2004, 79, 437.
(12) For the details, see the Supporting Information.
Org. Lett., Vol. 14, No. 17, 2012
4619

candidate for the structure of mycolactone S1. Similarly,
C160-hydroxylated C150-ketone of mycolactone A/B
(S2-140r or S2-140β) appeared to be a probable candidate
for mycolactone S2 (Figure 2).
At this stage, we opted to rely on the synthesis to
establish the structure of mycolactones S1 and S2. Scheme
1 outlines our synthesis of mycolactone A/B.13,14 In this
synthesis, mycolactone A/B was assembled from mycolac-
tone core 2 and unsaturated acid 3, which was, in turn,
prepared via Horner-Emmons olefination of aldehyde 5.
With this precedence, we recognize that the synthesis of
candidates S1 and S2-140r/-140β is reduced to the synthesis
of their C90ꢀC160 building blocks.
thereby showing the C140 stereochemistry kept intact
throughout the synthesis.
The synthesis of the C90ꢀC160 building block in the S1
series was achieved with a slight modification of the
previous synthesis.13 In this synthesis, p-methoxyphenyl-
methyl (MPM) was used as the C150 protecting group to
introduce the C150-ketone, cf., 14 f 15. As demonstrated
previously, the asymmetric dihydroxylation gave a ca. 8:1
diastereomeric mixture, but the undesired diastereomer
was chromatographically removed after the DIBAL step.
Scheme 2. Synthesis of the C90ꢀC160 Building Blocks12
Scheme 1. Summary of the Synthesis of Mycolactone A/B
Scheme 2 summarizes the synthesis of the C90ꢀC160
building blocks of candidates. For the S2-140r series, the
synthesis started with D-xylose, which was subjected to a
coupling reaction with ethyl bromopropionate under the
conditions reported by Demailly,15 followed by protection
of resultant primary and secondary alcohols, to give the
unsaturated ester 7, which was then converted to the
desired ketoaldehyde 9 via 8. Using the same synthetic
route, the C90ꢀC160 building block 11 in the S2-140β series
was synthesized from L-arabinose. The 1H and 13C NMR
analysis demonstrated that9 and 11are distinctly different,
With these three C90ꢀC160 building blocks in hand, we
studied the final assembly. Using the previous protocol, we
were able to selectively transfer 9, 11, and 15 to 16, 17, and
18, respectively (Scheme 3).
Our plan was to remove the TBS-protecting groups
under standard TBAF conditions, which was effective
for the mycolactone A/B series.13b For the present series,
however, we had a concern about the stability of S1,
S2ꢀ140r, and S2ꢀ140β; one could imagine possible side
reactions such as a retro-aldol reaction to form a conju-
gated enediol anion.
To address this question, we first tested the deprotection
of 16. On treatment with nonbuffered TBAF, 16 gave a
complex mixture of products. After numerous attempts,
we eventually found that TBAF buffered with 0.5 equiv
(13) (a) Song, F.; Fidanze, S.; Benowitz, A. B.; Kishi, Y Org. Lett.
2002, 4, 647. (b) Song, F.; Fidanze, S.; Benowitz, A. B.; Kishi, Y.
Tetrahedron 2007, 63, 5739. (c) Jackson, K. L.; Li, W.; Chen, C.-L.;
Kishi, Y. Tetrahedron 2010, 66, 2263.
(14) For synthetic studies of mycolactone A/B from other groups,
see: (a) Yin, N.; Wang, G.; Qian, M.; Negishi, E. Angew. Chem., Int. Ed.
2006, 45, 2916. Wang, G.; Yin, N.; Negishi, E. Chem.;Eur. J. 2011, 17,
4118. (b) Feyen, F.; Jantsch, A.; Altmann, K.-H. Synlett 2007, 415.
Gersbach, P.; Jantsch, A.; Feyen, F.; Scherr, N.; Dangy, J.-P.; Pluschke,
G.; Altmann, K.-H. Chem.;Eur. J. 2011, 17, 13017. Chany, A.-C.;
Casarotto, V.; Schmitt, M.; Tarnus, C.; Guenin-Mace, L.; Demangel,
C.; Mirguet, O.; Eustache, J.; Blanchard, N. Chem.;Eur. J. 2011, 17,
14413. (c) Alexander, M. D.; Fontaine, S. D.; La Clair, J. J.; DiPasquale,
A. G.; Rheingold, A. L.; Burkart, M. D. Chem. Commun. 2006, 4602. (d)
van Summeren, R. P.; Feringa, B. L.; Minnaard, A. J. Org Biomol.
Chem. 2005, 3, 2524.
(15) Le Mignot, V.; Lievre, C.; Frechon, C.; Demailly, G. Tetrahe-
dron lett. 1998, 39, 983.
4620
Org. Lett., Vol. 14, No. 17, 2012

imidazole hydrochloride gave the desired product 19.
Although it was slow (5 days at rt), the deprotection under
this condition was amazingly clean and furnished 19 in
95% isolated yield. With use of buffered TBAF, 17 and 18
were successfully deprotected to give 20 and 21, respec-
tively. Syntheticmycolactoneswerefully characterized and
used to identify the HPLC profiles useful for the structure
determination.
identification of the natural products. Newly isolated
mycolactones correspond to the “oxidized forms” of
mycolactone A/B, thereby hinting probable metabolic
and catabolic pathways of human mycolactones.
Scheme 3. Completion of the Synthesis of Structure Candidates
for Mycolactones S1 and S212
With synthetic mycolactones, we began to search for an
analytical method to establish the structure of mycolactones
S1 and S2. Given the fact that only a very minute amount of
natural metabolites was available, we needed an analytical
method with high sensitivity. In conjunction with the struc-
ture elucidation of fish and frog mycolactones, we demon-
strated that the HPLC profile, composed of the peaks
corresponding to the three major geometric isomers of the
unsaturated fatty acid moiety, is a powerful analytical tool.16
Like mycolactone A/B, the HPLC profile of mycolactones
S1 and S2 should be0composed of two peaks corresponding
Figure 3. HPLC analysis of synthetic and natural mycolactones
S1 and S2. Detection: UV absorption at 365 nm. Mycolactone
S1 series: Panel A-1 (column: Keystone Scientific, Hypersil silica
˚
(3 μm silica, 120 A) 250 ꢁ 4.6 mm; solvent (isocratic): CHCl3/
i-PrOH/H2O/Et3N = 93.3/6/0.2/0.5; flow rate: 1 mL/min. Key:
(a) natural S1; (b) synthetic 21; (c) a mixture of natural S1 and
21. Panel A-2 (column: Chiral Tech, Chiralpak IA (5 μm) 250 ꢁ
4.6 mm; solvent (isocratic): i-PrOH/toluene = 6/94; flow rate:
0.9 mL/min. Key: (a) natural S1; (b) synthetic 21; (c) a mixture of
natural S1 and 21. Mycolactone S2 series: Panel B-1 (column:
0
0
0
to Z-Δ4 ,5 - and E-Δ4 ,5 -geometric isomers (Figure 1).
With use of synthetic materials, we first searched for,
and identified, two orthogonal HPLC profiles of myco-
lactones S1 and S2, one with an achiral column and the
other with a chiral column (Figure 3). We should note that
the former column is sensitive to the gross structure of a
given molecule, whereas the latter is to the absolute stereo-
chemistry. We then subjected natural mycolactones S1 and
S2 to the HPLC analysis under the identified conditions,
thereby establishing that mycolactones S1 and S2 are 21 and
19, respectively.
In conclusion, we reported the structure for two new
mycolactones S1 and S2, isolated from the culture agar of
M. ulcerans subsp. shinshuense. The structure elucidation
was carried out in a three-step procedure: (1) probable
structures were speculated from MS analysis; (2) candidates
were synthesized; (3) HPLC profiles were established for
˚
Phenomex, Luna (5 μm silica, 100 A) 250 ꢁ 4.6 mm; solvent
(isocratic): CHCl3/i-PrOH/H2O/Et3N = 93.3/6/0.2/0.5; flow
rate: 1 mL/min. Key: (a) natural S2; (b) synthetic 19; (c) synthetic
20; (d) a mixture of natural S2 and 19; (e) a mixture of natural S2
and 20. Panel B-2 (column: Regis Tech, (S,S)-Whelk-01, 250 mm ꢁ
4.6 mm; solvent (isocratic): hexanes/CH2Cl2/EtOH/Et3N = 65/
27.5/7/0.5; flow rate: 0.6 mL/min. Key: (a) natural S2; (b) synthetic
19; (c) synthetic 20; (d) a mixture of natural S2 and 19; (e) a mixture
of natural S2 and 20.
Acknowledgment. We thank Dr. Y. Nancy Wong at
Eisai, Inc. for arrangement of the MS experiments.
Financial support from Eisai USA Foundation is greatly
acknowledged.
Supporting Information Available. Experimental details
and 1H and 13C NMR spectra. This material is available free
of charge via the Internet at http://pubs.acs.org.
(16) (a) Kim, H.-J.; Kishi, Y. J. Am. Chem. Soc. 2008, 130, 1842. (b)
Kim, H.-J.; Jackson, K. L.; Kishi, Y.; Williamson, H. R.; Mosi, L.;
Small, P. L. C. Chem. Commun. 2009, 7402. (c) Spangenberg, T.; Aubry,
S.; Kishi, Y. Tetrahedron Lett. 2010, 51, 1782.
The authors declare no competing financial interest.
Org. Lett., Vol. 14, No. 17, 2012
4621