Synthesis and stereochemical assignment of Brasilibactin A

Article abstract of DOI:10.1021/ol070355o

Brasilibactin A, a naturally occurring siderophore related to the mycobactins, has been synthesized in six steps. Use of asymmetric titanium-mediated aldol reactions allowed the preparation of both diastereomers from a common synthetic intermediate, thus

Full text of DOI:10.1021/ol070355o

ORGANIC
LETTERS
2007
Vol. 9, No. 9
1679-1681
Synthesis and Stereochemical
Assignment of Brasilibactin A
Judith M. Mitchell† and Jared T. Shaw*,‡
Broad Institute of HarVard and MIT Chemical Biology Program, 7 Cambridge Center,
Cambridge, Massachusetts 02142
shaw@chem.ucdaVis.edu
Received February 10, 2007
ABSTRACT
Brasilibactin A, a naturally occurring siderophore related to the mycobactins, has been synthesized in six steps. Use of asymmetric titanium-
mediated aldol reactions allowed the preparation of both diastereomers from a common synthetic intermediate, thus allowing the relative
stereochemistry of the natural product to be assigned. Brasilibactin A exhibits no inhibition of histone deacetylases (HDACs) in spite of the
N-formyl-N-hydroxy lysine moiety that is expected to affect the activity of these metal-dependent lysine-modifying enzymes.
Natural and synthetic compounds that mediate epigenetic
events, including histone lysine acetylation and methylation,
have emerged as promising new leads for controlling cancer.1
After the initial disclosure of histone deacetylase (HDAC)
inhibition by trapoxin B and trichostatin (Figure 1),2 many
studies demonstrating control of HDACs with small mol-
ecules have emerged.3 Suberoylanilide hydroxamic acid
(SAHA) is the first HDAC inhibitor approved for the
treatment of cancer.4 Control of histone methylation by small
molecules is still limited. Inhibitors of metal-dependent5
(jumonji domain-containing, JHMD/JMJD) and metal-
independent6 (BHC110/LSD1) histone demethylases have
recently been reported. Furthermore, there is evidence that
HDAC inhibitors also inhibit histone demethylation.7 Control
of both of these epigenetic events is important for further
understanding the genetic events that lead to cancer.8
We have undertaken the synthesis of brasilibactin A to
study its effects on histone modification. Brasilibactin A
contains a modified lysine residue for which analogues have
been described as inhibitors of HDACs.9 Although the
† Current address: Helicos BioSciences Corporation, One Kendall Square,
Building 700, Cambridge, Massachusetts 02139.
‡ Current address: Department of Chemistry, One Shields Avenue,
University of California, Davis, California 95616 (shaw@chem.ucd.edu).
(1) Suzuki, T.; Miyata, N. Curr. Med. Chem. 2006, 13, 935-958.
(2) Taunton, J.; Hassig, C. A.; Schreiber, S. L. Science 1996, 272, 408-
411.
(3) Miller, T. A.; Witter, D. J.; Belvedere, S. J. Med. Chem. 2003, 46,
5097-5116.
(4) Collins, I.; Workman, P. Nat. Chem. Biol. 2006, 2, 689-700.
(5) Cloos, P. A. C.; Christensen, J.; Agger, K.; Maiolica, A.; Rappsilber,
J.; Antal, T.; Hansen, K. H.; Helin, K. Nature 2006, 442, 307-311.
Figure 1. Inhibitors of metal-dependent enzymes that modify lysine
residues of histones.
10.1021/ol070355o CCC: $37.00
© 2007 American Chemical Society
Published on Web 03/31/2007

retrohydroxamic acid has been calculated to be less effective
at binding zinc,10 the fact that one class of HDMs (JHDM)
and potentially some HDACS employ iron in their active
sites suggests that retrohydroxamic acids might still offer
unique potencies and/or selectivities. At the outset of our
studies, the potential for mycobactin-like structures to affect
histone modification had not been explored.
The absolute stereochemistry of brasilibactin A was
previously assigned by total hydrolysis and amino acid
analysis, but the absolute stereochemistry of the â-alkoxy
acid residue has not been assigned. Analysis of the coupling
constants suggested a syn relationship. On the basis of the
need to prepare both enantiomers of this segment with either
of the two absolute configurations, we chose to employ the
titanium-mediated aldol reaction reported by Crimmins,13 in
which a single isomer of the thiazolidinethione auxiliary can
be used to form either of the two possible syn-aldol products.
In addition, this auxiliary allows for the direct cleavage of
the auxiliary by amines to form amides.
Aldol reaction of thiazolidinethione 2 with palmitaldehyde
proceeded with high diastereoselection (Scheme 1). Use of
Scheme 1. Synthesis of â-Hydroxy Amide Segments from
Common Propionyl Thiazolidinethione Auxiliary 2
Figure 2. Summary of mycobactins and related siderophore
structures.
Brasilibactin A (1, Figure 2) is a siderophore recently
isolated from Nocardia brasiliensis IFM 0995 that exhibits
potent cytotoxicity against murine leukemia L1210 (25 nM)
and human epidermoid carcinoma KB cells (50 nM).11
Although this compound exhibits a structural similarity to
many mycobactins in the backbone region, it is one of only
six compounds in this series reported with an N-hydroxy
formamide group.12 Most mycobactins contain long-chain
acyl groups at this position. We recognized that this
N-hydroxyformamide, or “retrohydroxamate”, singled out
these compounds as potential inhibitors of metal-dependent
histone-modifying enzymes such as HDACs and JHDMs.
1.0 equiv of base provided syn product 3a, whereas addition
of 2.5 equiv of base provided the complementary isomer 3b.
Each isomer was subsequently reacted with amine 4 to
provide amides 5a and 5b. Matched/mismatched reactivity
was observed in this reaction, as a slower reaction was
observed in the reaction of 5b.
The hydroxyl groups of amides 5a and 5b were next
acylated with protected N-Cbz-N′-hydroxyleucine (6).14
These reactions exhibited differential rates in their formations
of diastereomeric products 7a and 7b. Amide 7a was formed
in 60% yield, whereas the reaction to form 7b was slower
and lower yielding (Scheme 2).
(6) (a) Lee, M. G.; Wynder, C.; Schmidt, D. M.; McCafferty, D. G.;
Shiekhattar, R. Chem. Biol. 2006, 13, 563-567. (b) Culhane, J. C.;
Szewczuk, L. M.; Liu, X.; Da, G.; Marmorstein, R.; Cole, P. A. J. Am.
Chem. Soc. 2006, 128, 4536-4537.
(7) Lee, M. G.; Wynder, C.; Bochar, D. A.; Hakimi, M.-A.; Cooch, N.;
Shiekhattar, R. Mol. Cell. Biol. 2006, 26, 6395-6402.
(8) (a) Bolden, J. E.; Peart, M. J.; Johnstone, R. W. Nat. ReV. Drug
DiscoVery 2006, 5, 769-784. (b) Jenuwein, T. FEBS J. 2006, 273, 3121-
3135. (c) Oki, Y.; Issa, J.-P. J. ReV. Recent Clin. Trials 2006, 1, 169-182.
(d) Fraga, M. F.; Esteller, M. Cell Cycle 2005, 4, 1377-1381.
(9) (a) Wu, T. Y. H.; Hassig, C.; Wu, Y.; Ding, S.; Schultz, P. G. Bioorg.
Med. Chem. Lett. 2004, 14, 449-453. (b) Nishino, N.; Yoshikawa, D.;
Watanabe, L. A.; Kato, T.; Jose, B.; Komatsu, Y.; Sumida, Y.; Yoshida,
M. Bioorg. Med. Chem. Lett. 2004, 14, 2427-2431.
(10) Vanommeslaeghe, K.; Loverix, S.; Geerlings, P.; Tourwe, D. Bioorg.
Med. Chem. 2005, 13, 6070-6082.
(11) Tsuda, M.; Yamakawa, M.; Oka, S.; Tanaka, Y.; Hoshino, Y.;
Mikami, Y.; Sato, A.; Fujiwara, H.; Ohizumi, Y.; Kobayashi, J. I. J. Nat.
Prod. 2005, 68, 462-464.
(12) (a) Vergne, A. F.; Walz, A. J.; Miller, M. J. Nat. Prod. Rep. 2000,
17, 99-116. (b) A related compound (oxachelin) featuring a terminal
δ-lactam was recently reported: Sontag, B.; Gerlitz, M.; Paululat, T.; Rasser,
H.-F.; Gruen-Wollny, I.; Hansske, F. G. J. Antibiot. 2006, 59, 659-663.
Diastereomers 7a and 7b were converted to isomers of
brasilibactin A (Scheme 3). Hydrogenation to remove the
Cbz group and amide coupling to attach the D-serine-derived
(13) Crimmins, M. T.; King, B. W.; Tabet, E. A.; Chaudhary, K. J. Org.
Chem. 2001, 66, 894-902.
(14) (a) Miller, M. J.; Hu, J. J. Org. Chem. 1994, 59, 4858-4861. (b)
Yokokawa, F.; Izumi, K.; Omata, J.; Shioiri, T. Tetrahedron 2000, 56,
3027-3034.
1680
Org. Lett., Vol. 9, No. 9, 2007

Scheme 2. Attachment of an N-Formyl-N-hydroxy Lysine
Subunit to 5a and 5b
Scheme 3. Synthesis of Brasilibactin A (1) and
17,18-Bis-epi-brasilibactin A (11)
2-aryloxazoline15 proceeded in good yield. Although removal
of protecting groups proceeded without incident, isolation
of pure, iron-free material proved challenging. Deferration
by the method of Snow16 was successful at removing the
orange color associated with the iron complex, but we
observed partial decomposition of the product. Deferration
using EDTA17 was also unsuccessful. Ultimately, an aqueous
wash with HCl (3.0 M aq) followed by HPLC purification
provided pure samples of each diastereomer. Comparison
of the 1H NMR spectra with the published signals for
brasilibactin A demonstrated that the 17-(R),18-(S) config-
uration was consistent with the natural material.18
A pure sample of synthetic brasilibactin A was employed
in two assays for HDAC inhibition. HeLa cells were treated
with brasilibactin A (0.005-5.0 µM) followed by antibodies
which reveal nuclear and cytosolic hyperacetylation, which
is indicative of HDAC inhibition.19 No inhibition was
observed when compared to SAHA, an unselective HDAC
inhibitor. To determine if the inactivity resulted from the
fact that brasilibactin was not entering the cells, a second
assay using cellular extracts was also conducted. Treatment
of cell lysate with brasilibactin A (0.25-1.0 µM) revealed
no hyperacetylation of cellular proteins. These results suggest
that the cytotoxic activity observed for brasilibactin A, and
presumably amamistatin and other N-formyl mycobactins,
does not emanate from HDAC inhibition. On the basis of
these results, our future studies will focus on other cellular
processes, including histone methylation.
We have completed an efficient synthesis of brasilibactin
A in a total of six steps (longest linear sequence) from known
precursors.20 The use of the Crimmins aldol technology
facilitated the straightforward synthesis of both enantiomers
of the central â-hydroxy acid residue and allowed the
assignment of the configuration of these stereogenic centers
in the natural product. The potential for this compound to
modulate epigenetic events is under investigation.
Acknowledgment. This research was supported by the
NIGMS under the aegis of the Center for Methodology and
Library Development (CMLD). J.T.S. thanks Amgen Phar-
maceuticals for financial support. The authors thank Prof.
Stuart Schreiber (Broad Institute, Harvard University, and
the Howard Hughes Medical Institute), Dr. James E. Bradner
(Dana Farber Cancer Institute, Broad Institute), and Dr.
Stephen J. Haggerty (Broad Institute) for insightful discus-
sions. Prof. Marvin J. Miller and Ms. Leslie Patterson
(University of Notre Dame) are acknowledged for helpful
advice regarding siderophore deferration. Ms. Jessica Gross
(Harvard Medical School, Clardy Research group) is ac-
knowledged for assistance with HPLC purification of the
final product. Mr. Edward F. Greenberg (Broad Institute) is
acknowledged for performing HDAC assays. The project has
been funded in part with federal funds from the NCI’s
Initiative for Chemical Genetics, NIH, under Contract No.
N01-CO-12400. The content of this publication does not
necessarily reflect the views or policies of the Department
of Health and Human Services, nor does the mention of trade
names, commercial products, or organizations imply en-
dorsement by the U.S. Government.
(15) Miller, M. J.; Hu, J. J. Am. Chem. Soc. 1997, 119, 3462-3468.
(16) Snow, G. Biochem. J. 1965, 94, 160-165.
(17) Jalal, M. A. F.; Galles, J. L.; Van der Helm, D. J. Org. Chem. 1985,
50, 5642-5645.
(18) Attempts to contact Prof. Kobayashi (ref 11) in an effort to obtain
authentic material for direct comparison were unsuccessful.
(19) The details of this assay will be described elsewhere: Kwiatkowski,
N.; Bradner, J. E.; Greenberg, E. F.; Shaw, J. T.; Schreiber, S. L.;
Mazitschek, R. Angew. Chem., Int. Ed., manuscript in preparation.
(20) Compound 4 is prepared in four steps from protected lysine, making
the longest linear sequence ten steps from commercially available materials.
Supporting Information Available: Experimental proce-
dures and NMR spectra for all new compounds. This material
is available free of charge via the Internet at http://pubs.acs.org.
OL070355O
Org. Lett., Vol. 9, No. 9, 2007
1681