Further Studies on the Biosynthesis of the Manumycin-Type Antibiotic, Asukamycin, and the Chemical Synthesis of Protoasukamycin

Article abstract of DOI:10.1021/ja039336+

Asukamycin (2), a metabolite of Streptomyces nodosus ssp. asukaensis ATCC 29757 and a member of the manumycin family of antibiotics, is assembled from three components, an "upper" polyketide chain initiated by cyclohexanecarboxylic acid, a "lower" polyket

Full text of DOI:10.1021/ja039336+

Published on Web 03/05/2004
Further Studies on the Biosynthesis of the Manumycin-Type
Antibiotic, Asukamycin, and the Chemical Synthesis of
Protoasukamycin
Yiding Hu† and Heinz G. Floss*
Contribution from the Department of Chemistry, UniVersity of Washington, Box 351700,
Seattle, Washington 98195-1700
Received October 30, 2003; E-mail: floss@chem.washington.edu
Abstract: Asukamycin (2), a metabolite of Streptomyces nodosus ssp. asukaensis ATCC 29757 and a
member of the manumycin family of antibiotics, is assembled from three components, an “upper” polyketide
chain initiated by cyclohexanecarboxylic acid, a “lower” polyketide chain initiated by the novel starter unit,
3-amino-4-hydroxybenzoic acid (3,4-AHBA), and a cyclized 5-aminolevulinic acid moiety, 2-amino-3-
hydroxycyclopent-2-enone (C5N unit). To shed light on the order in which these components are assembled,
we synthesized in labeled form various potential intermediates and evaluated their incorporation into 2.
The assembly of the molecular framework of 2 from 3,4-AHBA and cyclohexanecarboxylic acid apparently
does not involve free, unactivated intermediates. However, protoasukamycin (12), the total synthesis of
which is reported, was efficiently converted into 2, demonstrating that the modification of the aromatic ring
to the epoxyquinol structure is the terminal step in the biosynthesis. The results suggest that the two
polyketide chains are synthesized separately and that the “upper” chain must be connected to the “lower”
polyketide chain before the C5N unit.
Introduction
2-amino-3-hydroxycyclopent-2-enone moiety (the C5N unit).
They only differ in the structure of the “upper” chain which is
Manumycin A (1a) and asukamycin (2) were the first two
members of the manumycin family of antibiotics1 discovered.
Manumycin A was isolated in 1963 from the bacterium
Streptomyces parVulus Tu¨ 64 by Za¨hner and co-workers,2 its
structure and stereochemistry were reported by the Zeeck group,3
and its stereochemistry was revised by Taylor and co-workers.4
The isolation5 and structure6 of asukamycin from S. nodosus
subsp. asukaensis was reported by the group of Omura, with a
subsequent revision of its stereochemistry.7 Both compounds
share the structural elements of a central 2-amino-4-hydroxy-
5,6-epoxycyclohex-2-enone moiety (the epoxyquinol moiety or
mC7N unit) and a “lower” chain extending from it, an all-trans
triene terminating in a carboxyl group amide-linked to a
amide-linked to the nitrogen of the mC7N unit. In manumycin
A, the “upper” chain consists of a methyl branched unsaturated
fatty acid, whereas in asukamycin it is a linear trienoic acid
terminating in a cyclohexane ring. Since the initial discovery
of 1a and 2, some 30 structurally closely related members of
the manumycin family have been isolated and structurally
characterized from natural sources,1 in addition to numerous
unnatural analogues generated by precursor-directed biosyn-
thesis.8,9 Compounds of the manumycin family exhibit a broad
range of biological activities. Most of these, such as their
antibacterial (gram-positive), anticoccidial, antifungal, and
insecticidal activities, are modest and probably not of clinical
interest.1 However, the discoveries that manumycin A is a potent
and selective inhibitor of RAS farnesyltransferase in vitro and
in vivo,10 and that several manumycins act as inhibitors of
interleukin-1â converting enzyme,11 making them potential lead
structures for the development of anticancer and antiinflamma-
† Present address: ICN Pharmaceuticals, 3300 Hyland Avenue, Costa
Mesa, CA 92629.
(1) Sattler, I.; Thiericke, R.; Zeeck, A. Nat. Prod. Rep. 1998, 15, 221-240.
(2) Buzzetti, F.; Ga¨umann, E.; Hu¨tter, R.; Keller-Schierlein, W.; Neipp, L.;
Prelog, V.; Za¨hner, H. Pharm. Acta HelV. 1963, 38, 871-874.
(3) (a) Schro¨der, K.; Zeeck, A. Tetrahedron Lett. 1973, 14, 4995-4998. (b)
Zeeck, A.; Schro¨der, K.; Frobel, K.; Grote, R.; Thiericke, R. J. Antibiot.
1987, 40, 1530-1540. (c) Thiericke, R.; Stellwaag, M.; Zeeck, A.; Snatzke,
G. J. Antibiot. 1987, 40, 1549-1554.
(8) Thiericke, R.; Zeeck, A. J. Chem. Soc., Perkin Trans. I 1988, 2123-2127.
(9) Thiericke, R.; Langer, H. J.; Zeeck, A. J. Chem. Soc., Perkin Trans. I 1989,
851-855.
(4) Alcarz, L.; Macdonald, G.; Ragot, J. P.; Lewis, N.; Taylor, R. J. K. J.
Org. Chem. 1998, 63, 3526-3527
(5) Omura, S.; Kitao, C.; Tanaka, H.; Oiwa, R.; Takahashi, Y.; Nakagawa,
A.; Shimida, M. J. Antibiot. 1976, 29, 876-881.
(10) (a) Hara, M.; Akasaka, K.; Akinaga, S.; Okabe, M.; Nakano, H.; Gomez,
R.; Wood, D.; Uh, M.; Tamanoi, F. Proc. Natl. Acad. Sci. U.S.A. 1993,
90, 2281-2285. (b) Hara, M.; Han, M. Proc. Natl. Acad. Sci. U.S.A. 1995,
92, 3333-3337.
(6) Kakinuma, K.; Ikegawa, N.; Nakagawa, A.; Omura, S. J. Am. Chem. Soc.
1979, 101, 3402-3404.
(7) Cho, H.; Sattler, I.; Beale, J. M.; Zeeck, A.; Floss, H. G. J. Org. Chem.
1993, 58, 7925-7928.
(11) Tanaka, T.; Tsukuda, E.; Uosaki, Y.; Matsuda, Y. J. Antibiot. 1996, 49,
1085-1090.
9
10.1021/ja039336+ CCC: $27.50 © 2004 American Chemical Society
J. AM. CHEM. SOC. 2004, 126, 3837-3844
3837

A R T I C L E S
Hu and Floss
tory agents, respectively, has generated considerable interest in
this class of compounds.
In the present paper, we report on the synthesis of a number
of labeled potential biosynthetic intermediates, including the
total synthesis of protoasukamycin, and their evaluation for
incorporation into 2. On the basis of these results, a hypothetical
pathway for the biosynthesis of the manumycin-type antibiotics
is proposed. Some of the results have been published in
preliminary form.18
Results
Nature of the mC7N Unit. The labeling pattern of the mC7N
unit of 1 and 2 from acetate, succinate, and glycerol or glucose
is duplicated in a metabolite, 4-hydroxy-3-nitrosobenzamide,
an iron chelator isolated from mutant strains of S. murayamaen-
sis.19 This compound represents a mC7N unit, and 3-amino-4-
hydroxybenzoic acid (3,4-AHBA, 3) was shown to be its specific
precursor.19a These findings led to the suggestion19b that 3,4-
AHBA may also be the precursor of the mC7N unit in
manumycin-type compounds; that is, it serves as the polyketide
starter unit for the synthesis of the “lower” chain. A preliminary
feeding experiment with [2-2H]-3,4-AHBA supported the pro-
posal by showing 16.5% incorporation of deuterium into 2.18
To test this hypothesis further, we synthesized [7-13C]-3,4-
AHBA (99 atom % 13C) from 4-hydroxy-[7-13C]benzoic acid14
by nitration and subsequent reduction using procedures devel-
oped by Kondo et al.20 and Balcom and Fu¨rst,21 respectively.
Feeding of this compound (10 mg/100 mL) to liquid cultures
of S. nodosus ssp. asukaensis ATCC 29757 and S. parVulus Tu¨
64 under previously described conditions12 gave samples of 2
and 1a, respectively, which were analyzed by 13C NMR
spectroscopy. Both compounds showed enriched signals at δ
136 ppm (in CDCl3, C-7), with16.5 atom % excess 13C in 2
and 5.4 atom % excess 13C in 1a. The fermentation of S.
parVulus also yielded manumycin B (1b) which was also
enriched (5.7 atom % excess 13C) at C-7. This result demon-
strates the intact incorporation of 3 into the manumycins and
shows that 3,4-AHBA indeed serves as the polyketide starter
unit for the formation of the “lower” chain.
Isolation of a Shunt Metabolite. The 13C NMR spectra of
crude extracts from early stage (day 2 after feeding) fermenta-
tions of S. nodosus ssp. asukaensis with [7-13C]-3,4-AHBA
showed two enriched signals in the olefinic region, one at δ
137 ppm and the other at δ 139 ppm (in acetone-d6). The former
belongs to C-7 of 2, but the latter represents a new compound
formed from 3. Because the chemical shift suggests that this
new compound is formed from 3 by polyketide chain extension
at C-7, its structure might shed light on the biosynthetic pathway.
The new compound (4a, Scheme 1) was therefore isolated from
the fermentation and purified by flash column chromatography
on RP18 silica gel followed by HPLC. FAB-MS indicated a
molecular weight of 273 Da; methylation with CH2N2 gave a
dimethyl derivative 4b of molecular formula C17H19NO4 (by
HR-MS), establishing the molecular formula of the parent
compound as C15H15NO4. The general similarity of the NMR
spectra of 4a and 4b to those of the synthetic 7-(3-amino-4-
Biosynthetic studies of the manumycins have mainly focused
on asukamycin and manumycin A and have involved feeding
experiments with radioactive and stable isotope-labeled precur-
sors to identify the biosynthetic building blocks. These com-
pounds appear to be assembled from several different compo-
nents, each of which has its own biosynthetic origin. Both the
“upper” and the “lower” chain are of polyketide origin. In 1a,
the “upper” chain is built up from acetyl-CoA as the starter
unit through chain elongation by one malonyl-CoA and three
methylmalonyl-CoA extender units. The “upper” chain in 2
arises from cyclohexanecarboxylic acid as the starter unit by
chain extension with three molecules of malonyl-CoA.12 The
cyclohexanecarboxylic acid in turn is derived from shikimic acid
through a series of dehydrations and reductions.13 The assembly
of the “lower” chain in both compounds is initiated by the mC7N
unit, with chain extension by three malonyl-CoA. The C5N unit
terminating the “lower” chain, as in other antibiotics,14 arises
by an intramolecular cyclization from δ-aminolevulinic acid,12
the common precursor of porphyrins and corrins. The mode of
formation of the mC7N unit is not clear yet. Contrary to an
earlier suggestion,15 its origin is different from that of the mC7N
units found in ansamycin and mitomycin antibiotics, which arise
from 3-amino-5-hydroxybenzoic acid formed by the aminoshiki-
mate pathway.16 Rather, its seven carbon atoms originate from
one molecule of a 4-carbon dicarboxylic acid, such as succinate
or oxalacetate, and one molecule of a glycerol-derived metabo-
lite, such as a triose phosphate.12 The epoxide and the hydroxy,
but not the carbonyl oxygen of the epoxyquinol moiety, were
shown to be derived from molecular oxygen,17 and some features
of the mode of incorporation of glycerol have been estab-
lished,7,12 but the mechanism of the assembly of the mC7N unit
from its building blocks is not known.
(12) Thiericke, R.; Zeeck, A.; Nakagawa, A.; Omura, S.; Herrold, R. E.; Wu,
S. T. S.; Beale, J. M.; Floss, H. G. J. Am. Chem. Soc. 1990, 112, 3979-
3987.
(13) (a) Moore, B. S.; Cho, H.; Casati, R.; Kennedy, E.; Reynolds, K. A.; Mocek,
U.; Beale, J. M.; Floss, H. G. J. Am. Chem. Soc. 1993, 115, 5254-5266.
(b) Moore, B. S.; Poralla, K.; Floss, H. G. J. Am. Chem. Soc. 1993, 115,
5267-5274.
(14) Cho, H.; Beale, J. M.; Graff, C.; Mocek, U.; Nakagawa, A.; Omura, S.;
Floss, H. G. J. Am. Chem. Soc. 1993, 115, 12296-12304.
(15) Becker, A, M.; Rickards, R. W.; Brown, R. F. C. Tetrahedron 1983, 39,
4189-4192.
(16) (a) Kim, C.-G.; Kirschning, A.; Bergon, P.; Zhou, P.; Su, E.; Sauerbrei,
B.; Ning, S.; Ahn, Y.; Breuer, M.; Leistner, E.; Floss, H. G. J. Am. Chem.
Soc. 1996, 118, 7486-7491. (b) Yu, T.-W.; Mu¨ller, R.; Mu¨ller, M.; Zhang,
H.; Draeger, G.; Kim, C.-G.; Leistner, E.; Floss, H. G. J. Biol. Chem. 2001,
276, 12546-12555.
(17) Thiericke, R.; Zeeck, A.; Robinson, J. A.; Beale, J. M.; Floss, H. G. J.
Chem. Soc., Chem. Commun. 1989, 402-403.
(18) Hu, Y.; Melville, C. R.; Gould, S. J.; Floss, H. G. J. Am. Chem. Soc. 1997,
119, 4301-4302.
(19) (a) Cone, M. C.; Melville, C. R.; Carney, J. R.; Gore, M. P.; Gould, S. J.
Tetrahedron 1995, 51, 3095-3102. (b) Gould, S. J.; Melville, C.; Cone,
M. C. J. Am. Chem. Soc. 1996, 118, 9228-9232.
(20) Kondo, S.; Murase, K.; Kuzuya, M. Chem. Pharm. Bull. 1994, 42, 768-
773.
(21) Balcom, D.; Fu¨rst, A. J. Am. Chem. Soc. 1953, 75, 4334-4335.
9
3838 J. AM. CHEM. SOC. VOL. 126, NO. 12, 2004

Studies on the Manumycin-Type Antibiotic, Asukamycin
A R T I C L E S
Scheme 1
upon feeding of this material to S. nodosus ssp. asukaensis.
These results indicate that 4a is a shunt metabolite of the
asukamycin pathway, which cannot be channeled back into the
biosynthetic manifold. The formation of 4a may be a reflection
of the fact that formation of the cyclohexanecarboxylic acid
starter unit, and hence of the “upper chain”, appears to be rate
limiting for the biosynthesis of 2. Supplementation experiments
with S. nodosus ssp. asukaensis showed that whereas feeding
of 0.65 mM 3,4-AHBA increased the ratio of 4a to 2 by a
modest 70%, feeding of 4 mM cyclohexanecarboxylic acid
increased the ratio of 2:4a 10-fold, that is, almost completely
suppressed 4a formation.
Synthesis of Potential Biosynthetic Intermediates. The
architecture of the manumycins embodies several different
building blocks. It was logical to expect that these are
synthesized separately from their respective precursors and are
then assembled into the complete molecular framework. To try
to elucidate in which order these building blocks are assembled,
we synthesized, in labeled form, the various components and
partial assemblies of components and determined which ones
are incorporated into the final product, asukamycin. In choosing
the targets, we made the assumption, borne out by the
subsequent experiments, that the modification of the aromatic
ring to the epoxyquinol structure is probably a late step in the
biosynthesis. Consequently, we targeted the synthesis of 3,4-
AHBA chain-extended by one (5), two (6), and three (7) acetate
units, 3,4-AHBA chain-extended to carry the complete “lower”
chain including the C5N unit (10), the complete “upper” chain
(8), 3,4-AHBA N-acylated with the complete “upper” chain (9),
3,4-AHBA with both the “upper” and the “lower” chain, but
without the C5N unit (11), and protoasukamycin (12), which
consists of the entire framework of asukamycin but lacks the
modification of the aromatic ring to the epoxyquinol structure.
[1,2-13C2]-5 was synthesized in 43.3% overall yield from
4-methoxybenzaldehyde by nitration with NH4NO3/trifluoro-
acetic anhydride (TFAA)22 to 3-nitro-4-methoxybenzaldehyde,
followed by a Knoevenagel condensation with [U-13C3]malonic
acid,23 demethylation with LiI,24and reduction of the nitro group
Scheme 2
25
with SnCl2 (Scheme 2A). Preparation of the corresponding
hydroxyphenyl)hepta-2,4,6-trienoic acid (7, Scheme 2), except
for the presence of signals corresponding to an acetyl group
and, in 4b, two additional O-methyl groups, suggested that 4a
might be the N-acetyl derivative of 7. This was confirmed by
synthesizing an authentic sample of 4b by acetylation of the
ethyl ester of 7 followed by ester hydrolysis with LiOH and
methylation with diazomethane (Scheme 1), which was in all
respects identical to the material obtained from the isolated 4a.
The specific incorporation of [7-13C]-3 into 4a was determined
by mass spectrometry as 19.5%. The yield of 4a depended very
much on the culture conditions and ranged up to 4-5 mg/L.
To define the role of 4a in asukamycin biosynthesis, we
prepared a sample of [1,2-13C2]-4a by acetylation of [1,2-13C2]-
7 (synthesis: Scheme 2) and subsequent mild base hydrolysis
of the resulting diacetate. Feeding of this compound to S.
nodosus ssp. asukaensis and analysis by MS and 13C NMR
showed no incorporation into 2. To determine whether polyketide
chain extension or N-acetylation occurs first in the formation
of 4a, a sample of [7-13C, acetyl-2H3]-N-acetyl-3,4-AHBA was
prepared by acetylation of [7-13C]-3 with [2H3]-acetyl chloride
and hydrolysis of the N,O-diacetate to the N-acetate. No
incorporation of either 13C or deuterium into 4a was observed
N-acetylcysteamine (SNAC) thioester 5a proved problematic
because of interference of the phenolic hydroxy group and
failure of common protecting groups (e.g., TBDMS, Boc), but
was finally accomplished in poor yield (15%) by direct
esterification of 5 with N-acetylcysteamine catalyzed by DCC
and DMAP.26 The 5-(3-amino-4-hydroxyphenyl)-(2E,4E)-penta-
2,4-dienoic acid 6 was synthesized analogously from 4-meth-
oxycinnamaldehyde, but using a Wadsworth-Emmons reaction
with triethyl phosphonoacetate27 for the two-carbon chain
extension. For the preparation of [1,2-13C2]-6, triethyl phosphono-
[1,2-13C2]acetate was used in this step. The subsequent cleavage
of the methyl ether with LiI also removed the ethyl ester function
and, after SnCl2 reduction of the nitro group, gave 6 in 40.9%
overall yield (Scheme 2B). For the synthesis of 7-(3-amino-4-
hydroxyphenyl)-(2E,4E,6E)-hepta-2,4,6-trienoic acid 7, a second
(22) Crivello, J. V. J. Org. Chem. 1981, 46, 3056-3060.
(23) Ressler, C.; Goodman, F. J.; Tsutsui, R.; Tsutsumi, M. J. Org. Chem. 1979,
44, 2027-2029.
(24) Harrison, I. T. J. Chem. Soc., Chem. Commun. 1969, 616.
(25) Bellamy, F. D.; Ou, K. Tetrahedron Lett. 1984, 25, 839-842
(26) Witter, D. J.; Vederas, J. C. J. Org. Chem. 1996, 61, 2613-2623
(27) (a) Williams, D. R.; White, F. H. J. Org. Chem. 1987, 52, 5067-5079. (b)
Seguineau, P.; Villieras, J. Tetrahedron Lett. 1988, 29, 477-480.
9
J. AM. CHEM. SOC. VOL. 126, NO. 12, 2004 3839

A R T I C L E S
Scheme 3
Hu and Floss
chain extension with triethyl phosphonoacetate was carried out
on 5-(3-nitro-4-methoxyphenyl)-(2E,4E)-penta-2,4-dienal, which
was obtained from ethyl 5-(3-nitro-4-methoxyphenyl)-(2E,4E)-
penta-2,4-dienoate, the intermediate in the synthesis of 6, by
DIBAL reduction and MnO2 oxidation28 of the resulting alcohol.
Following dealkylation with LiI and reduction with SnCl2, the
overall yield of 7 starting from 3-nitro-4-methoxycinnamalde-
hyde was 33.2%. [1,2-13C2]-7 was prepared by using triethyl
phosphono-[1,2-13C2]acetate in the second Wadsworth-Em-
mons reaction (Scheme 2C). All attempts to prepare the SNAC
thioesters of 6 or 7, either by direct esterification or by use of
protecting groups, were unsuccessful.
hepta-2,4,6-trienoyl chloride, obtained from [1,2-13C2]-8 with
oxalyl chloride. Initial demethylation with BBr3, surprisingly,
removed only the ester function, but a second BBr3 treatment
of the methyl ether gave 3-{7-cyclohexyl-(2E,4E,6E)-[1,2-13C2]-
hepta-2,4,6-trienoyl}-[15N]amino-4-hydroxybenzoic acid 9 in a
modest 12.5% overall yield (Scheme 3B). Labeled 10, 11, and
12 were all prepared from ethyl 7-(4-methoxy-3-nitrophenyl)-
(2E,4E,6E)-[1,2-13C2]hepta-2,4,6-trienoate ([1,2-13C2]-13), the
intermediate in the synthesis of [1,2-13C2]-7 (Scheme 2C), using
more or less the same standard reactions but in different orders,
as shown in Scheme 4. Thus, ester hydrolysis of 13, conversion
to the acid chloride with oxalyl chloride, and condensation with
2-amino-3-hydroxycyclopent-2-enone hydrochloride gave N1-
(2-hydroxy-5-oxo-cyclopent-1-enyl)-7-(4-methoxy-3-nitrophen-
yl)-(2E,4E,6E)-[1,2-13C2]hepta-2,4,6-trienamide ([1,2-13C2]-14)
in 66% yield. Demethylation of 14 with BBr3 and nitro group
reduction with SnCl2 proceeded in very poor yield (6%) to give
[1,2-13C2]-10. On the other hand, when the nitro group of 14
was reduced with SnCl2 first, and then the product was coupled
with 7-cyclohexyl-(2E,4E,6E)-hepta-2,4,6-trienoyl chloride and
demethylated with BBr3, [1,2-13C2]protoasukamycin ([1,2-13C2]-
12) was obtained in 12% yield from 14. Similarly, SnCl2
reduction of [1,2-13C2]-13, coupling with 7-cyclohexyl-(2E,4E,6E)-
hepta-2,4,6-trienoyl chloride, followed by demethylation with
BBr3 and ester hydrolysis with LiOH gave [1,2-13C2]-11 in 8%
yield. The last two steps in particular accounted for the low
overall yield.
The synthesis of the “upper” chain, 7-cyclohexyl-(2E,4E,6E)-
hepta-2,4,6-trienoic acid 8, proceeded from pyridinium-1-
sulfonate via glutaconaldehyde using chemistry developed by
the Taylor and Wipf groups.29 Reaction of the TBDMS-
protected enol of glutaconaldehyde with cyclohexylmagnesium
bromide and acid hydrolysis gave 5-cyclohexylpenta-2,4-dienal
(32.4% from pyridinium-1-sulfonate), which was subjected to
Wadsworth-Emmons chain extension with triethyl phospho-
noacetate followed by ester hydrolysis with LiOH to give 8
(22.5% from dienal). Again, [1,2-13C2]-8 was prepared by using
triethyl phosphono-[1,2-13C2]acetate in the Wadsworth-Em-
mons reaction (Scheme 3A). A third building block, 2-amino-
3-hydroxycyclopent-2-enone, was prepared from cyclopentane-
1,3-dione as described by Ebenezer.30
The advanced potential precursors 9-12 were prepared from
7 and/or 8 or intermediates from their synthesis. For the
synthesis of [1,2-13C2,15N]-9, methyl 4-methoxybenzoate was
nitrated with NH415NO3/TFAA and the nitro group was reduced
with SnCl2 to give methyl 4-methoxy-3-[15N]aminobenzoate.
The latter was coupled with 7-cyclohexyl-(2E,4E,6E)-[1,2-13C2]-
Feeding of Potential Biosynthetic Intermediates. The
labeled compounds 5 and its SNAC derivative 5a, 6, 7, and
9-12 were each fed to a 100 mL (5, 5a, 6, 7, 9) or 60 mL
(10-12) fermentation culture of S. nodosus ssp. asukaensis in
500 or 250 mL baffled Erlenmeyer flasks 24 h after inoculation.
After incubation for 48 h with shaking, the cultures were
harvested, and asukamycin and any other relevant compounds,
such as 4a or metabolites of the administered precursors, were
extracted from the culture broth and from the mycelium and
purified by flash column chromatography and reverse-phase
(28) (a) Rao, A. V. R.; Reddy, S. P.; Reddy, E. R. J. Org. Chem. 1986, 51,
4158-4159. (b) Hoeger, C. A.; Johnston, A. D.; Okamura, W. H. J. Am.
Chem. Soc. 1987, 109, 4690-4698.
(29) (a) Lewis, N.; McKen, P. W.; Taylor, R. J. K. Synlett 1991, 898-900. (b)
Wipf, P.; Coish, D. G. Tetrahedron Lett. 1997, 38, 5073-5076.
(30) Ebenezer, W. Synth. Commun. 1991, 21, 351-358.
9
3840 J. AM. CHEM. SOC. VOL. 126, NO. 12, 2004

Studies on the Manumycin-Type Antibiotic, Asukamycin
A R T I C L E S
Scheme 4
HPLC. The degree of isotope enrichment was measured by
electrospray mass spectrometry (ES-MS), using selective ion
monitoring, and the location of the isotope was determined by
13C NMR spectroscopy. In the feeding experiment with [1,2-
13C2,15N]-9, no production of either asukamycin or the shunt
metabolite 4a was observed. Feeding of [1,2-13C2]-5, [1,2-13C2]-
5a, [1,2-13C2]-6, or [1,2-13C2]-7 allowed the production of 2 at
normal yields, but no incorporation of isotope into the product
was detected. Likewise, no incorporation of 13C from [1,2-13C2]-
10 and [1,2-13C2]-11 into 2 was seen, and neither compound
underwent modification of the aromatic ring to the epoxyquinol
structure. However, the amount of 2 produced in the fermenta-
tion with [1,2-13C2]-11 was greatly reduced relative to that of
unfed cultures. In contrast, feeding of [1,2-13C2]-12 resulted in
an unambiguous incorporation of 13C into 2, with 13C NMR
showing labeling of the expected positions C-12 and C-13 (δ
122.9 and 167.1 ppm, d, J ) 65.9 Hz) and coupling between
the two labeled carbon atoms. The specific incorporation was
determined by ES-MS as 30.5%, showing that protoasukamycin
is a very efficient biosynthetic precursor of asukamycin.
the assembly of the lower polyketide chain. However, the mode
of its formation from the established building blocks, a glycerol-
derived three-carbon fragment and a four-carbon dicarboxylic
acid, is still unclear. A plausible pathway has been proposed
by Gould et al.,19b but conclusive evidence to support or disprove
this hypothetical sequence of reactions is still lacking.
We had assumed as our working hypothesis that the modi-
fication of the aromatic ring to the epoxyquinol structure of
the mC7N unit would be a late step in the biosynthesis of the
manumycins. We therefore synthesized for the first time a
putative late intermediate in asukamycin biosynthesis, proto-
asukamycin (12), which carries the unmodified aromatic ring
in place of the epoxyquinol structure of 2. The efficient and
specific incorporation of this compound into 2 leaves no doubt
that the entire molecular framework of the manumycins is
assembled first before the aromatic ring is modified to the final
epoxyquinol structure of the type I34 manumycins. The latter
transformation is presumably catalyzed by a dioxygenase, and
a likely mechanism36 is shown in Scheme 5. Such a mechanism
has been demonstrated for the oxidation of dihydrovitamin
K37-39 and for the oxidation of dihydroxyacetanilid in the
formation of antibiotics LL-C10037 and MPP3051.40 Consistent
with this mechanism, both the epoxide and the hydroxy oxygen
Discussion
Previous work12 had shown that the mC7N unit of the
manumycins is not derived from 3-amino-5-hydroxybenzoic
acid, as is that of the ansamycin and mitomycin antibiotics.31
Instead, the feeding experiments presented here and elsewhere18
show that it comes from a very similar compound, the
regioisomeric 3-amino-4-hydroxybenzoic acid (3,4-AHBA). 3,4-
AHBA, following conversion either to the CoA thioester or as
the free acid which is activated on the loading domain of a
polyketide synthase,32,33 evidently serves as the starter unit for
(33) Lowden, P. A. S.; Wilkinson, B.; Bo¨hm, G. A.; Handa, S.; Floss, H. G.;
Leadlay, P. F.; Staunton, J. Angew. Chem., Int. Ed. 2001, 40, 777-779.
(34) The type I manumycins with the epoxyquinol moiety in turn are the
precursors of the type II manumycins which carry a diol structure instead
and are derived from the former by a transformation late in the fermenta-
tion.35
(35) Hu, Y.; Floss, H. G. J. Antibiot. 2001, 54, 340-348.
(36) Ham, S. W.; Dowd, P. J. Am. Chem. Soc. 1990, 112, 1660-1661.
(37) Kuliopulos, A.; Hubbard, B. R.; Lam, Z.; Koski, I. J.; Furie, B.; Furie, B.
C.; Walsh, C. T. Biochemistry 1992, 31, 7722-7728.
(38) Naganathan, S.; Hershline, R.; Ham, S. W.; Dowd, P. J. Am. Chem. Soc.
1994, 116, 9831-9839.
(39) Dowd, P.; Hershline, R.; Ham, S. W.; Naganathan, S. Science 1995, 269,
1684-1691.
(40) Gould, S. J.; Kirchmeier, M. J.; LaFever, R. E. J. Am. Chem. Soc. 1996,
118, 7663-7666.
(31) Floss, H. G. Nat. Prod. Rep. 1997, 14, 433-452.
(32) Admiraal, S. J.; Walsh, C. T.; Khosla, C. Biochemistry 2001, 40, 6116-
6123.
9
J. AM. CHEM. SOC. VOL. 126, NO. 12, 2004 3841

A R T I C L E S
Scheme 5
Hu and Floss
of the epoxyquinol are derived from molecular oxygen,17 and
these two oxygens in all naturally occurring manumycins are
syn oriented.1,4 The availability of protoasukamycin as substrate
opens the way for the discovery and characterization of the
enzyme(s) catalyzing this interesting reaction.
corporation of 7 as well as 10, the PKS-bound compound 7
may have to be acylated by the complete “upper” chain before
it can be released from the enzyme and coupled to the C5N
unit. However, if this is the case, the product cannot be released
from the enzyme as the free acid, because compound 11,
representing protoasukamycin without the C5N unit, is also not
incorporated into 2.
Negative results from feeding experiments always have to
be interpreted with caution, as permeability barriers may prevent
the administered labeled precursors from entering the cells and
reaching the site of synthesis. In the present case, the efficient
incorporation of 12, a molecule similar in structure to and larger
in size than the other compounds fed, argues against the
possibility that the nonincorporation of the potential intermedi-
ates fed is due to permeability problems. However, it must be
kept in mind that 12 is the only molecule among the compounds
fed that does not contain an easily ionizable (at neutral pH)
group. Nevertheless, the results most likely indicate that the
assembly of 12 from 3,4-AHBA involves no free, unactivated
intermediates. It may involve some free activated intermediates,
such as CoA thioesters of the acids 7 and/or 8, but it is also
possible that all intermediates between the starter unit, 3,4-
AHBA, and the first identified product, protoasukamycin, remain
enzyme-bound during the entire assembly process.
The above results leave open the question in which order
the different building blocks of 2 are attached to each other.
Because compound 9 is not incorporated, attachment of the
“upper” chain to the nitrogen of the mC7N unit is probably not
the first step. It follows that the assembly of the “lower” chain
must be the first reaction which 3,4-AHBA undergoes. However,
because neither 10 nor 11 are incorporated, it is not clear
whether the C5N unit or the “upper” chain is attached first to
an activated version of 7. Several lines of evidence, however,
point toward the latter scenario, that is, transfer of the “upper”
chain to the nitrogen of an activated 7 before formation of the
amide bond to the aminocyclopentenolone moiety. (i) The fact
that 9 and 11, but not 10, inhibit asukamycin formation suggests
that 9 and 11, but not 10, resemble intermediates in the
biosynthesis and thus act as competitive inhibitors. (ii) The shunt
metabolite 4a must arise by acetylation of free, activated, or
enzyme-bound 7. The N-acetylation of arylamines is a com-
monly observed reaction in Actinomycetes.42 The fact that no
formation of an acetylated shunt metabolite equivalent to 10
Surprisingly, none of the potential intermediates in the
assembly of protoasukamycin from 3,4-AHBA tested were
incorporated into 2. The nonincorporation of the two partially
chain-extended versions of 3,4-AHBA, the carboxylic acids 5
and 6, which should be enzyme-bound intermediates in the
assembly of the “lower” polyketide chain, is not unexpected.
In general, the free acids corresponding to such intermediates
cannot be activated and can therefore not be loaded onto the
PKS. More surprising, however, is the failure of compound 5a,
the SNAC thioester derivative of the diketide 5, to label 2. It
has been shown in a number of systems that polyketide
synthases, particularly modular type I enzymes, can accept
intermediates presented to them in the form of their SNAC
thioesters, load them onto the cognitive module, and process
them through the remainder of the assembly steps to generate
the expected endproducts.26,41 Possibly, this paradigm does not
hold for the enzyme assembling the asukamycin “lower” chain,
which therefore cannot utilize compound 5a in this way. The
nonincorporation into 2 of [1,2-13C2]-7, representing the fully
assembled “lower” chain polyketide, suggests that 7 is not
released from the PKS as the free acid. It may be released as
an activated species, such as the CoA thioester, or may be
transferred directly from the PKS to the nitrogen of the C5N
unit. Yet, the product of this reaction, compound 10 representing
the starter unit with the entire “lower” chain, is also not
incorporated. Another possible explanation for nonincorporation
of 5, 5a, 6, 7, and 10 would be a scenario in which 3,4-AHBA
is first acylated on the nitrogen by the complete “upper” chain
to give compound 9 before the “lower” chain is built up. 15N-
and 13C-double labeled 9 was, however, not converted into 2,
ruling out this version of a biosynthetic pathway. Interestingly,
this compound completely shuts off the production of both 2
and the shunt metabolite 4a in the fermentation. Therefore, 9,
based on its similarity to an intermediate in the biosynthesis of
2, probably acts as an inhibitor of one of the biosynthetic
enzymes. As another scenario that would explain the nonin-
(41) (a) Yue, S.; Duncan, J. S.; Yamamoto, Y.; Hutchinson, C. R. J. Am. Chem.
Soc. 1987, 109, 1253-1255. (b) Cane, D. E.; Yang, C. J. Am. Chem. Soc.
1987, 109, 1255-1257. (c) Cane, D. E.; Lambalot, R. H.; Prabhakaran, P.
C.; Ott, W. R. J. Am. Chem. Soc. 1993, 115, 522-526. (d) Cane, D. E.;
Tan, W.; Ott, W. R. J. Am. Chem. Soc. 1993, 115, 527-535. (e) Tsantrizos,
Y. S.; Zhou, F.; Famili, P.; Yang, X. J. Org. Chem. 1995, 60, 6922-6929.
(42) Yu, T.-W.; Shen, Y.; Doi-Katayama, Y.; Tang, L.; Park, C.; Moore, B. S.;
Hutchinson, C. R.; Floss, H. G. Proc. Natl. Acad. Sci. U.S.A. 1999, 96,
9051-9056.
9
3842 J. AM. CHEM. SOC. VOL. 126, NO. 12, 2004

Studies on the Manumycin-Type Antibiotic, Asukamycin
A R T I C L E S
Scheme 6
NMR spectra were obtained on Bruker AF 300 and AM 500
spectrometers. Electron-impact mass spectrometry (EI-MS) was carried
out on a Kratos Profile HV-3 mass spectrometer, and electrospray mass
spectra (ES-MS) and tandem mass spectra (ES-MS/MS) were recorded
on a Bruker Esquire ion trap mass spectrometer. High-resolution mass
spectra (HR-MS) were obtained on a Micromass 70SEQ tandem hydrid
mass spectrometer. The isotope distributions in 2, 1a, 1b, and 4a were
determined by selected ion monitoring (SIM) on a Micromass Quattro
II tandem quadrupole mass spectrometer. Fermentations were carried
out in a New Brunswick G25 controlled environment incubator shaker
or in an Adolf Ku¨hner ISF-4-V rotary shaker cabinet. Analytical TLC
was performed on precoated silica gel plates (aluminum backing, 0.25
mm layer, UV-254 fluorescence), and preparative TLC was performed
on precoated silica gel plates (glass backing, 2.0 mm layer, UV-254
fluorescence), both from EM Science. Flash column chromatography
was performed on 230-400 mesh silica gel from Aldrich, and reverse-
phase chromatography was performed on C18 silica gel from Whatman.
HPLC was conducted with a Beckman model 116 isocratic pump and
a model 166 absorbance detector using C18 reverse-phase analytical or
semipreparative columns.
was observed suggests that acylation of the arylamine function
with the “upper” chain occurs before attachment of the C5N
unit. In the absence of a sufficient supply of the limiting “upper”
chain precursor, the biosynthesis stalls, leading to aberrant
release and acetylation of 7. On the basis of these considerations,
we propose a sequence for the assembly of the components of
protasukamycin as shown in Scheme 6, involving first chain
extension of 3,4-AHBA to the activated “lower” polyketide
chain, in parallel chain extension of cyclohexanecarboxylic to
the activated “upper” chain, followed by transfer of the “upper”
chain to the arylamine nitrogen, and finally amide bond
formation between the activated carboxyl group of the resulting
assembly and the nitrogen of the aminocyclopentenolone moiety
to give protoasukamycin. The genetic analysis of the asukamycin
biosynthetic gene cluster, currently underway in this and a
collaborating laboratory, will hopefully show whether this
proposed sequence and other aspects of the biosynthesis of 2
deduced from isotopic tracer experiments are indeed valid.
Experimental Section
Feeding Experiments. Streptomyces nodosus ssp. asukaensis was
grown on yeast extract-malt (YM) agar plates at 28 °C for 4 days, and
was then stored at 0 °C. A loop of mycelium was transferred into 100
mL of culture medium in a 500 mL baffled Erlenmeyer flask and was
grown on a rotary shaker at 300 rpm for 2 days at 28 °C (seed culture).
Ten (or six) milliliters of seed culture was used to inoculate each 100
(or 60) mL of culture medium in 500 (or 250) mL baffled Erlenmeyer
flasks, which were incubated on a rotary shaker at 300 rpm for 3 days
at 28 °C. Both culture media consisted of glucose, 20 g; Bacto Peptone,
Materials and General Methods. Streptomyces nodosus ssp.
asukaensis ATCC 29757 was obtained from the American Type Culture
Collection, and Streptomyces parVulus Tu¨ 64 was obtained from
Professor Axel Zeeck, University of Go¨ttingen, Germany. Fermentation
ingredients were purchased from Difco and Sigma, chemicals were from
Aldrich and Lancaster, and nitrogen dioxide was from Sigma. 4-Hy-
droxy-[7-13C]benzoic acid was purchased from Cambridge Isotope
Laboratories, triethyl phosphono-[1,2-13C2]acetate was from Sigma, and
[U-13C3]malonic acid had been synthesized previously in this laboratory.
All chemicals and solvents were of reagent or HPLC grade and were
used without further purification unless otherwise noted.
5 g; K2HPO4, 0.25 g; MgSO4, 0.25 g; trace elements (NH4)6Mo7O24
×
4H2O, 5 mg; FeSO4 × 7H2O, 50 mg; CuSO4 × 5H2O, 5 mg; ZnSO4 ×
7H2O, 5 mg; MnCl2 × 4H2O, 10 mg; deionized water, 1000 mL; pH
9
J. AM. CHEM. SOC. VOL. 126, NO. 12, 2004 3843

A R T I C L E S
Hu and Floss
7.0, and were sterilized for 20 min at 121 °C in an autoclave.
Fermentations with S. parVulus Tu¨ 64 were carried out as described
earlier.12
methylene chloride/methanol 100:3. The asukamycin was further
purified by semipreparative C18 reverse-phase HPLC eluting with
methanol/water or acetonitrile/water. Yields of asukamycins averaged
about 20-25 mg/L.
Labeled compounds were administered to the production cultures
24 h after inoculation as single doses of filter-sterilized solutions in
the amounts indicated per culture volume: 3,4-[7-13C]-AHBA ([7-13C]-
3) dissolved in 5% K2CO3, 10 mg/100 mL (0.649 mM); 3-(7-
cyclohexyl-(2E,4E,6E)-[1,2-13C2]hepta-2,4,6-trienoyl)-[15N]amino-4-
methoxybenzoic acid ([1,2-13C2,15N]-9) dissolved in 5% K2CO3, 12 mg/
100 mL (0.348 mM); 3-(3-amino-4-hydroxyphenyl)-(E)-[1,2-13C2]prop-
2-enoic acid ([1,2-13C2]-5) dissolved in 5% K2CO3, 15 mg/100 mL
(0.829 mM); 5-(3-amino-4-hydroxyphenyl)-(2E,4E)-[1,2-13C2]penta-2,4-
dienoic acid ([1,2-13C2]-6) dissolved in 5% K2CO3, 15 mg/100 mL
(0.725 mM); 7-(3-amino-4-hydroxyphenyl)-(2E,4E,6E)-[1,2-13C2]hepta-
2,4,6-trienoic acid ([1,2-13C2]-7) dissolved in 5% K2CO3, 20 mg/100
mL (0.858 mM); 3-(3-amino-4-hydroxyphenyl)-(E)-[1,2-13C2]prop-2-
enoic acid N-acetylcysteamine thioester ([1,2-13C2]-5a) dissolved in
DMSO, 20 mg/100 mL (0.707 mM); N1-(2-hydroxy-5-oxocyclopent-
1-enyl)-7-(3-amino-4-hydroxyphenyl)-(2E,4E,6E)-[1,2-13C2]hepta-2,4,6-
trienamide ([1,2-13C2]-10) dissolved in 5% K2CO3, 7 mg/60 mL (0.356
mM); 7-[3-{7-cyclohexyl-(2E,4E,6E)-hepta-2,4,6-trienoyl}-amino-4-
hydroxyphenyl]-(2E,4E,6E)-[1,2-13C2]hepta-2,4,6-trienoic acid (11) dis-
solved in 5% K2CO3, 9 mg/60 mL (0.356 mM); proto-asukamycin ([1,2-
13C2]-12) dissolved in 5% K2CO3, 11 mg/60 mL (0.356 mM).
After fermentation for 72 h, cultures were centrifuged at 9000 rpm
for 25 min. The supernatant was saturated with NaCl and extracted
three times with ethyl acetate. The mycelium was extracted with
acetone, and the acetone extracts were concentrated. The residue was
extracted with ethyl acetate. The combined ethyl acetate extracts were
dried, and the solvent was evaporated in vacuo. The crude product was
purified on 2.0 mm preparative TLC plates developed three times with
chloroform/methanol 9:1, or on a silica gel column eluting with
Isolation and Identification of Shunt Metabolite 4a. S. nodosus
ssp. asukaensis cultures fed with [7-13C]-3 were harvested after 2 days
of fermentation. The supernatant extracts were fractionated on an RP18
silica gel column eluted with acetone/water 20:80, and the compound
responsible for an enriched NMR signal at 137 ppm was further purified
on HPLC with water/2-propanol 72.5/27.5 as eluent. The structure of
this unknown compound was determined as 7-(3-N-acetylamino-4-
hydroxyphenyl)-(2E,4E,6E)-hepta-2,4,6-trienoic acid (4a) from its 1H,
13C NMR and FAB-MS spectra. The compound was methylated with
diazomethane, and the molecular formula of the resulting derivative
(4b) was determined by HR-MS. NMR comparison of the derivative
to an authentic sample prepared from synthetic 13 showed them to be
identical. 4a: Rf 0.21 (silica gel, CH2Cl2/MeOH 10:1); UV absorption
1
maximum 368 nm (MeOH); H NMR (300 MHz, DMSO-d6) δH 7.90
(d, 1H, J ) 2.0 Hz, H-2′), 7.23 (dd, 1H, J ) 15.1, 11.2 Hz, H-3), 7.12
(dd, 1H, J ) 8.3, 2.0 Hz, H-6′), 6.84 (d, 1H, J ) 8.3 Hz, H-5′), 6.80-
6.68 (m, 3H, H-5, H-6, H-7), 6.49 (dd, 1H, J ) 14.2, 11.2 Hz, H-4),
5.86 (d, 1H, J ) 15.1 Hz, H-2), 2.09 (s, 3H, CH3); 13C NMR (125
MHz, DMSO-d6) δC 169.0, 167.7, 148.6, 144.3, 141.2, 137.1 (enriched,
C-7), 129.0, 127.6, 126.7, 126.6, 125.6, 123.6, 120.8, 115.9, 23.7; FAB-
MS m/z 274 [M + H]+. HR-MS for 4b: [M]+ C17H19NO4, calculated
301.1322, found 301.1314.
Supporting Information Available: Synthesis of labeled
compounds (PDF). This material is available free of charge via
the Internet at http://pubs.acs.org.
JA039336+
9
3844 J. AM. CHEM. SOC. VOL. 126, NO. 12, 2004