Biosynthesis of aurachins A-L in Stigmatella aurantiaca: A feeding study

Article abstract of DOI:10.1021/np8003084

The isolation of aurachins A-L (1-11) from Stigmatella aurantiaca strain Sg a15 is described. Their structures and relative configurations were deduced from spectroscopic data, in particular NMR. Three structural types were identified: A-type aurachins (1, 2, 6) are C-3 oxygen-substituted quinolines carrying a farnesyl residue on C-4, C-type aurachins (3, 4, 7-11) are C-4 oxygen-substituted quinolines carrying a farnesyl residue on C-3, and C-type aurachin E (5) has a [1,1a,8,d]imidazoloquinoline structure. Feeding of ¹³C-labeled precursors showed that the quinoline ring is constructed from anthranihc acid and acetate, and the farnesyl residue from acetate by both the mevalonate and nonmevalonate pathways. Further, feeding of labeled aurachin C (3) indicated the A-type aurachins are derived by a novel intramolecular 3,4-migration of the farnesyl residue that is induced by a 2,3-epoxidation and terminated by a reduction step. ¹⁸O-Labeling experiments indicated the new oxygen substituents originate from atomospheric oxygen. On the basis of these results a biosynthetic scheme covering all aurachins is proposed. It is further proposed that quinolones with an unorthodox substitution pattern, such as the 2-geranylquinolones from Pseudonocardia sp. and the 3-heptylquinolones from Pseudomonas sp., are formed by related rearrangement mechanisms.

Full text of DOI:10.1021/np8003084

J. Nat. Prod. 2008, 71, 1843–1849
1843
Biosynthesis of Aurachins A-L in Stigmatella aurantiaca: A Feeding Study
Gerhard Ho¨fle* and Brigitte Kunze
Helmholtz Centre for Infection Research (preViously GBF, Gesellschaft fu¨r Biotechnologische Forschung),
Inhoffenstrasse 7, 38124 Braunschweig, Germany
ReceiVed May 22, 2008
The isolation of aurachins A-L (1-11) from Stigmatella aurantiaca strain Sg a15 is described. Their structures and
relative configurations were deduced from spectroscopic data, in particular NMR. Three structural types were identified:
A-type aurachins (1, 2, 6) are C-3 oxygen-substituted quinolines carrying a farnesyl residue on C-4, C-type aurachins
(3, 4, 7-11) are C-4 oxygen-substituted quinolines carrying a farnesyl residue on C-3, and C-type aurachin E (5) has
a [1,1a,8,d]imidazoloquinoline structure. Feeding of 13C-labeled precursors showed that the quinoline ring is constructed
from anthranilic acid and acetate, and the farnesyl residue from acetate by both the mevalonate and nonmevalonate
pathways. Further, feeding of labeled aurachin C (3) indicated the A-type aurachins are derived by a novel intramolecular
3,4-migration of the farnesyl residue that is induced by a 2,3-epoxidation and terminated by a reduction step. 18O-
Labeling experiments indicated the new oxygen substituents originate from atomospheric oxygen. On the basis of these
results a biosynthetic scheme covering all aurachins is proposed. It is further proposed that quinolones with an unorthodox
substitution pattern, such as the 2-geranylquinolones from Pseudonocardia sp. and the 3-heptylquinolones from
Pseudomonas sp., are formed by related rearrangement mechanisms.
During our screening of myxobacteria for antifungal activity
Stigmatella aurantiaca strain Sg a15 was identified as a particularly
prolific producer of structurally unrelated metabolites. In addition
to the myxalamids,1 known from Myxococcus xanthus,2 and the
iron chelator myxochelin,3 known from Angiococcus disciformis,4
the novel chromones stigmatellins A and B1 and the quinoline
alkaloids aurachins A-L (1-11)5,6 were isolated. Remarkably three
of these groups of compounds turned out to be inhibitors of
mitochondrial respiration and photosynthesis: the myxalamids
inhibit complex I,2b the stigmatellins inhibit complex III,7 photo-
system II, and the cytochrome b6/f-complex,8 and the various
aurachins inhibit complexes I9,10 and III,9 as well as the photosystem
II and the cytochrome b6/f-complex.11 Even though the structures
of these inhibitors are not related to one another, there is an overall
structural resemblance in that a relatively polar headgroup carries
a lipophilic side chain about 15 carbon atoms in length, thus
mimicking the electron carrier ubiquinone/ubiquinol in the respira-
tory chain. It is intriguing that Stigmatella apparently has invented
three different types of inhibitors of the respiratory chain that confer
antifungal activity. It may be speculated that their co-occurrence
in Sg a15 is a natural archetype of today’s antibiotic combination
therapy to counteract development of resistance in the target
organism.
Results and Discussion
The aurachins are a group of unique isoprenoid quinoline
alkaloids characterized by a farnesyl residue in either the 3- or
4-position of the quinoline nucleus. Conversely, the 4- or 3-position
is occupied by an oxygen substituent. In addition, the aurachins
may occur as quinolines, 4-quinolones, or quinoline N-oxides. The
basic aurachins A-D (1-4) are formed by S. aurantiaca strain Sg
a15 at various rates. On large-scale fermentation aurachin C (3)
predominates during the growth phase, but is consumed during the
stationary and decline phases to yield aurachin A (1), which
accumulates up to 8.8 mg/L, whereas aurachins B and D (2 and 4)
always remain as minor components5 (see Supporting Information,
Figure S1a). Batchwise feeding of anthranilic acid, a presumed
precursor of the aurachins, increased the yields of aurachin A (1)
to 12 mg/L and aurachin C (3) to 14 mg/L.
Whereas the biosynthesis of myxochelin, myxalamid, and
stigmatellin has been investigated in great detail by feeding studies12
and cloning of the biosynthesis genes,12,13 there is only limited
information on the biosynthesis of aurachins. On the basis of an
early feeding study a biosynthetic sequence starting with anthranilic
acid via aurachins D, C, and B to A has been proposed.14 Recently,
parts of the biosynthetic gene cluster were cloned that are
responsible for the synthesis of anthranilic acid and its conversion
to aurachin D in the early steps of aurachin biosynthesis.15 However,
those for the following oxidative and rearrangement steps to
aurachins A, B, C, and F-L could not be identified.
Herein we report the isolation and structure elucidation of
aurachins A-L (1-11) isolated from S. aurantiaca strain Sg a15.
Further, results of the incorporation of basic and advanced
precursors in aurachins A, B, and C are presented, and a biosynthetic
scheme covering all aurachins including aurachin P (12)16 is
proposed.
In the presence of XAD adsorber resin, and continuous feeding
of anthranilic acid, the yield of aurachin C (3) increased dramatically
up to 130 mg/L, while that of aurachin A remained unchanged at
* To whom correspondence should be addressed. Tel: int. 49-531-
61811040. Fax: int. 49-531-61812697. E-mail: gho@helmholtz-hzi.de.
10.1021/np8003084 CCC: $40.75
2008 American Chemical Society and American Society of Pharmacognosy
Published on Web 11/07/2008

1844 Journal of Natural Products, 2008, Vol. 71, No. 11
Ho¨fle and Kunze
(2) and C (3) with an extended chromophore shifting the UV bands
by 10-20 nm to longer wavelengths. Aurachin K (10) is an isomer
of L (11) with the 1′,2′ double bond moved to the 4′,5′-position,
giving rise to an additional strong UV band at 238 nm for the
conjugate diene system. Its 4′E,6′E configuration followed from
NOE data. Aurachin H (8) is an isomer of A (1) with the farnesyl
residue at C-3 and oxygen at C-4. It is further modified by a 4′,5′
double bond to give aurachin G (7) or by a methoxyl group in the
1′-position to give aurachin I (9). On the basis of the assumption
that aurachins B (2) and C (3) are epoxidized during biosynthesis
by one and the same monooxygenase, the 2′S,3′R configuration was
assigned to aurachins G-I (7-9). The relative 1′S configuration
of aurachin I (9) followed from the NOE between the 13′-Me and
1′-H and the 3 Hz coupling between 1′-H and 2′-H. Aurachin E
(5) is related to aurachin D (4), but differs from all other aurachins
by an extra nitrogen atom and a carbonyl group. From detailed 2D
NMR measurements the nitrogen is located at C-8 of the quinoline
ring, forming an imidazolone ring with N-1.
Figure 1. Selected NOEs and preferred conformation of aurachin
A (1) in CDCl3.
12 mg/L (see Supporting Information, Figure S1b). Aurachins E-L
(5-11) are always very minor components. It is noteworthy that
the related species Stigmatella eracta strain Pd e32 produces
predominantly the new aurachin P (12) and small amounts of
aurachin C (3).16
The biosynthesis of the basic aurachins A-D (1-4) was
investigated by feeding of 13C- and 18O-labeled anthranilic acid,
C-1 and C-2 13C-enriched acetate, and oxygen 18O2, as well as
labeled intermediates isolated from previous feedings (Tables 1 and
2). The significant increase of aurachin C production up to 130
mg/L on feeding of anthranilic acid already indicated that it is a
biosynthetic precursor and early bottleneck in the multistep bio-
synthesis. Indeed, on feeding [1′-13C]anthranilic acid, C-4-labeled
aurachins A and C were formed (Table 1, expt 1). The remaining
carbon atoms C-2, C-3, and C-9 of the quinoline moiety as well as
all carbon atoms of the side chain of aurachins A and C were labeled
by [1,2-13C2]acetate (Tables 1 and 2, expt 2). The incorporation
pattern of intact acetate units and isolated carbons in the 4′-, 8′-,
and 12′-position of the farnesyl side chain was deduced from the
magnitude of the C,C coupling constants (Table 2, expt 2). This
indicates the farnesyl residue is constructed by the classical
terpenoid pathway via mevalonate, as demonstrated previously for
the myxobacterial menaquinone MK-8.17 However, on closer
inspection of the signal intensities, conspicuous discrepancies were
observed. In general, the 13C isotope abundance in the terpenoid
side chain was only 50% and less than that of the acetate-derived
carbon atoms of the quinoline ring (Tables 1 and 2, expt 2). This
can be explained, at least in part, by a shunt mechanism channeling
a C5 unit from leucin to IPP and DMAPP. According to a feeding
experiment with Sg a15, about 10% of the farnesyl side chain of
the aurachins was labeled by [D10]leucine.18 To account for the
remaining major portion of label, a third pathway has to be
considered. The almost equal distribution of label from [2-13C]ac-
etate (Table 2, expt 3) clearly indicated involvement of the
The structural assignment of the aurachins was accomplished
straightforwardly from elemental composition, 1D and 2D NMR,
and UV data (see Supporting Information, Tables S1 and S2). Thus,
a nonmodified farnesyl residue was identified in aurachins B-E
(2-5) by comparison of the chemical shifts with literature values.
Similarly, the quinoline nucleus of aurachins A and B (1 and 2)
was derived from a comparison of 13C NMR data with literature
values, whereas the presence of four adjacent protons in the benzene
1
ring was determined from the H NMR spectra. From the HMBC
spectra the farnesyl residue in 1 and 2 had to be placed at C-4 of
the pyridine ring, a methyl group at C-2, and the two oxygen atoms
at the remaining C-3 and on the nitrogen. In aurachin A (1) the
2′,3′ double bond is oxidized and a 2′,3 ether ring is formed. On
the basis of the assumption that this transformation is brought about
by epoxidation of the 2′E double bond followed by cyclization with
the 3-OH, aurachin A (1) was assigned the relative 2′S,3′R
configuration. This was readily confirmed by the observation of
the NOEs shown in Figure 1. In addition, a strong NOE between
the 1′-protons and 5-H in the benzene ring confirmed the attachment
of the farnesyl residue at C-4 in aurachin A (1), as well as in B (2)
and F (6).
All other aurachins, C-E (3-5) and G-L (7-11), showed a
strong NOE between the 1′-proton(s) and the methyl group at C-2,
indicating the attachment of the farnesyl residue at C-3. Conversely,
in these compounds the oxygen is found at C-4, forming a quinolone
system in aurachins C-E (3-5) (C-4 around δ 170 ppm) or an
ether in aurachins G-L (7-11) (C-4 around δ 150 ppm). Aurachins
F (6) and L (11) are cyclo-didehydro derivatives of aurachins B

Biosynthesis of Aurachins in Stigmatella aurantiaca
Journal of Natural Products, 2008, Vol. 71, No. 11 1845
Table 1. 13C NMR Data of the Quinoline Moiety of Labeled Aurachins A1-A4 (13-16), B1, B2 (17, 18), and C1-C4 (19-22) in
CD3OD
13C isotope abundance (%)
expt no.
1
precursor fed
[1-13C]anthranilate
isotope abundance
90%
aurachins isolated
C-2
C-3
C-4
C-9
C-1′
A1 (13)
C1 (19)
A2 (14)a
B1 (17)
C2 (20)
B2 (18)
1.1
1.1
14
1.5
1.5
n.d.
1.1
1.1
14
14
19
18
80
66
1.1
78
78
74
1.1
1.1
15
13
16
14
1.1
1.1
5
n.d.
1.5
2
2
3
[1,2-13C2]acetate
[1-13C]anthranilate
[2-13C]acetate
90%
90%
90%
C-3 19%
C-4 78%
7.5% each
7% each
4
aurachin C2 (20)
5
6
[1-13C18O2]anthranilate
C3 (21)
A3 (15)
C4 (22)
A4 (16)e
1.1
15
1.1
16
7b
1.1
16
1.1
4
[4-13C18O]aurachin C3 (21)
5c
3.5d
n.d.
7
[1-13C]acetate
90%
90%
53%
[2-13C]acetate
18O2 on the second day
b
a Coupling constants for C-2 JCC ) 76.17 Hz; C-3 JCC ) 76.25 Hz; C-9 singlet. 18O isotope abundance 7%. c Isotope abundances determined by
e
EIMS; <0.3% 18O. d Isotope abundances determined by EIMS; 18O 3.5%. 18O isotope abundance at C-3 14%, 18O isotope shift ∆δ ) 0.020 ppm; at
C-2′ 16%, ∆δ ) 0.035 ppm; at C-3′ 20%, ∆δ ) 0.030 ppm.
Scheme 1. Incorporation of 13C- and 18O-Labeled Precursors and 18O-Labeled Molecular Oxygen in Aurachins A-C (1-3) (expts 1-7)a
a For the terpenoid moiety only labeling by the mevalonate pathway is shown. The figures in aurachins A1, A2, A3, etc., designate isotopomers.

1846 Journal of Natural Products, 2008, Vol. 71, No. 11
Ho¨fle and Kunze
Table 2. 13C NMR Data of the Terpene Moiety of Labeled
Aurachins A2 (14), C2 (20), and A4 (16) in CDCl3
spectroscopy. Incorporation of 18O at C-3/C-2′, as well as at C-3′,
was clearly seen by upfield isotope shifts of the expected order of
magnitude for C-3 (0.020 ppm), C-2′ (0.035 ppm), and C-3′ (0.030
ppm).22 In summary, the biosynthetic sequence depicted in Scheme
2 is proposed.
expt 2
expt 3
expt 7
precursor
[1,2-13C]acetate
[2-13C]acetate
[1-13C]acetate
[1-13C]anthranilate
[2-13C]acetate
aurachin13C isotope abundance (%) JC,C (Hz)
Starting from anthranilic acid, the first defined intermediate is
2-methyl-(1,4-dihydroquinoline-4-one) (II) according to feeding of
a blocked mutant. It is the substrate for a farnesyl transferase leading
to aurachin D (4),15b which is not accumulated but rapidly oxidized
to the N-oxide aurachin C (3). Epoxidation of its C-2, C-3 double
bond to III sets the stage for migration of the farnesyl residue to
C-4, which is brought about by a push-pull effect between the
nitrogen and the carbonyl group. On reduction of the 3-carbonyl
group in IV, the system aromatizes to aurachin B (2). Its
intermediate concentration is apparently kept low by rapid epoxi-
dation of the 2′,3′ double bond followed by spontaneous interception
of the epoxide by the 3-hydroxy group to give aurachin A (1), the
final product in S. aurantiaca.
Final and major product in Stigmatella erecta, aurachin P (12),
is formed by a subsequent hydroxylation of 1 at C-1′.16 Similarly
to aurachin B (2), side-chain epoxidation of aurachin C (3) leads
to aurachin H (8), which is hydroxylated and O-methylated to
aurachin I (9). Further, both aurachins B and C may be deydro-
genated to VII and VI, which by 3,3-electrocyclization give
aurachins F (6) and L (11). Finally, dehydrogenation of 8 gives
aurachin G (7), and double-bond transposition in 11 leads to
aurachin K (10).
The structurally aberrant aurachin E (5) has no reasonable place
in the biosynthetic scheme. As it was isolated only in minute
amounts, it was suspected that it may be derived from 2,3-
diaminobenzoic acid, which was introduced as an impurity with
the anthranilic acid fed during cultivation. Thus, 2,3-diaminobenzoic
acid was synthesized from 2-amino-3-nitrotoluene23 and fed as free
acid and ethyl ester to S. aurantiaca. As this had no influence on
the formation of aurachin E, its biosynthesis presumably branches
off at a later stage, e.g., from aurachin C or D. Alternatively
aurachin E might be an artifact formed from aurachin C or D,
ammonia, and a reactive C1 or a “CN” species during cultivation
or isolation. The latter possibility could be demonstrated by the
semisynthesis of aurachin E from aurachin C.24
C atom
A2 (14)
C2 (20)
A4 (16)
C-1′
C-2′
C-3′
C-4′
C-5′
C-6′
C-7′
C-8′
5
4
5
8
5
4
3
8
4
4
35.05
2
4
n.d
6
2
3
n.d.
6
2
4
n.d
5
2
2
2
4
3
3
5
5
3
3
6
4
3
3
5
6
6
6
35.01
39.67
singlet
43.65
43.92
42.49
singlet
44.11
44.12
n.d.
C-9′
C-10′
C-11′
C-12′
C-13′
C-14′
C-15′
6
5
4
5
singlet
39.89
42.29
37.09
glyceraldehydes 3-phosphate/pyruvate pathway to IPP and DMAPP
as described by Rohmer et al.17,19 This is corroborated by the low
isotope abundance at C atoms 2′, 3′, 6′, 7′, 10′, and 11′ on feeding
a mixture of [1-13C]- and [2-13C]acetate (Table 2, expt 7), as these
C atoms receive their label only through the mevalonate pathway.
The simultaneous operation of the mevalonate and non-mevalonate
pathways in bacteria is well documented for instance in Strepto-
mycetes20 and Listeria.21
Labeling of C-2 and C-3 by an intact acetate unit and of C-9 by
a cleaved acetate unit suggested a 4-quinolone-2-acetic acid
intermediate I (Scheme 2). As feeding of synthetic I or its ethyl
ester had no influence on the production of aurachins, it is not the
presumed substrate for the farnesyl transferase. This is rather its
decarboxylation product 2-methylquinolone II, which after feeding
to the knockout mutant Sga-D_AS restored aurachin synthesis.15b
However, its presence in cultures of strain Sg a15 after feeding of
anthranilic acid could not be demonstrated by HPLC/MS analysis.
Also, in contrast to feeding of anthranilic acid, feeding of synthetic
2-methylquinolone II had no influence on the production of
aurachins.
Conclusion
The most unusual step in aurachin biosynthesis is the isomer-
ization of aurachin C (3) to aurachin B (2), which, in a formal
sense, interchanges the 3-farnesyl and 4-oxygen substituents. Its
mechanism was first investigated by combined feeding of [1′-
13C]anthranilic acid and [2-13C]acetate (Table 1, expt 3). The
outcome was aurachin B1 (17) (the figure in B1 identifies
the isotopomer present) and aurachin C2 (20) with the label in the
expected positions. Feeding of aurachin C2 (20) with a 13C isotope
abundance of 19% at C-3 and 78% at C-4 to a fresh culture resulted
in aurachin B2 (18) with an unchanged labeling pattern for the
quinoline ring and side chain (Table 1, expt 4). This indicates that
an intramolecular migration of the farnesyl residue occurs. The
fate of the carbonyl oxygen in the rearrangement was investigated
by first feeding of [1′-13C18O2]anthranilic acid to produce
[4-13C18O]aurachin C3 (21) with isotope abundances of 7% (Table
1, expt 5). On feeding of this material to a fresh culture, aurachin
A3 (15) and in addition some aurachin C4 (22) were recovered
(Table 1, expt 6). Whereas the isotope ratio in the recovered material
was unchanged, the oxygen label in aurachin A3 was lost during
the rearrangement. In order to identify the origin of the newly
introduced oxygen atom including that on C-3′ of aurachin A, a
mixture of [1-13C] and [2-13C]acetate was fed. After 2 days, when
the titer of aurachin C was expected to be high, the atmosphere in
the culture bottle was replaced by a 4:1 mixture of nitrogen and
18O2 oxygen and cultivation was continued for 2 days (Table 1,
expt 7). Aurachin A4 (16) was isolated and analyzed by NMR
The C-type aurachins (3-5, 7-11) are characterized by an
oxygen substituent in the 4-position of the quinoline ring and a
farnesyl derived residue at C-3. From their structures they are related
to 2,3-substituted 4-quinolones isolated from plants25 and bacteria,26,27
although their biosynthesis is different. Most remarkably, the
farnesyl residue is constructed in parallel via the mavalonate, the
nonmevalonate pathway, and leucine degradation. The relative
contribution of these pathways may vary with the physiological
state of the organism and has to be investigated in detail by further
feeding studies with appropriate precursors. The A-type aurachins
(1, 2, 6, 12) are derived from aurachin C (3) by a unique
transposition of the farnesyl residue from C-3 to C-4 by consecutive
epoxidation/reduction steps. With this mechanism in mind, the
biosynthesis of 2-geranyl-substituted 4-quinolones in Pseudono-
cardia sp.27 appears in a new light. As direct prenylation of a
4-quinolone at C-2 is not favorable for electronic reasons, we
propose a 3-geranyl intermediate related to aurachin D (4). This,
instead of prior N-oxidation, is directly rearranged by similar
epoxidation/reduction steps to the 2-geranyl isomer (see Supporting
Information, Scheme S1).28 By contrast, the 4-quinolones produced
by Pseudomonas aeruginosa26 receive their 2-heptyl (and higher
alkyl) residues in the course of quinoline ring formation.29 Like
aurachin D (4) they are transformed to either the N-oxide or the
2,3-epoxide, which isomerizes to 2-heptyl-3-hydroxyquinolinone.
The latter represents an important quorum sensing signal (PQS) in

Biosynthesis of Aurachins in Stigmatella aurantiaca
Journal of Natural Products, 2008, Vol. 71, No. 11 1847
Scheme 2. Proposed Biosynthetic Relationships of Aurachins A-D (1-4) and F-P (6-12)
P. aeruginosa.26 Interestingly, in addition to the various 2-alky-
lquinolones a rearranged 3-heptyl-3-hydroxyquinolone has also been
isolated from P. aeruginosa30 and P. methanica.31 We suggest that
this completely neglected compound, in recent work,26 is formed
from PQS by another 2,3-epoxidation followed by rearrangement
(see Supporting Information, Scheme S2) and, moreover, that this
is the mechanism by which the quorum sensing signal is turned
off.
equipped with a Nucleodur C18 column, 5 µm, 2 × 125 mm
(Macherey-Nagel) (0.3 mL/min acetonitrile/5 mM ammonium acetate
buffer pH 5.5, gradient 5:95 to 95:5 in 30 min). Large-scale separations
were performed on a Prep LC-500 from Merck. Preparative HPLC was
performed using a Nucleosil column (250 × 21 mm, 7 µm, flow 18
mL/min) with a solvent of acetonitrile/water (gradients), and diode array
detection. Analytical TLC used TLC aluminum sheets of silica gel Si
60 F254 (Merck), detection UV absorption at 254 nm. Red to violet
spots were formed on staining with vanillin/sulfuric acid followed by
heating to 120 °C. Precoated silica gel Si 60 F254 plates of 0.25 and
0.5 mm layer thickness were used for preparative TLC.
Experimental Section
Cultivation of Stigmatella aurantiaca Strain Sg a15. (a) A 100 L
bioreactor with a circulating pump stirrer system (Giovanola Fre`res,
Monthey, Switzerland) was charged with 60 L of modified Zein liquid
medium and inoculated with 5 L of well-grown shake cultures as
described previously.5 The aeration rate was adjusted to 300 NL/h, and
the agitation was 500 rpm. During the cultivation 14 g of anthranilic
acid dissolved in 1000 mL of water and adjusted with KHCO3 to pH
7.0 was pumped in at different intervals; in particular, during the
logarithmic growth phase 220 mg was added every hour. After 45 h
the aeration rate was reduced to 250 NL/h and the agitation to 450
rpm. During the following 60 h the oxygen saturation dropped to 20%
General Experimental Procedures. Optical rotations were deter-
mined on a Perkin-Elmer 241 instrument. UV spectra were recorded
on a Shimadzu UV-2102 PC scanning spectrometer. IR spectra were
measured with a Nicolet 20DXB FT-IR spectrometer. NMR spectra
were recorded in CDCl3 on Bruker ARX-400 and DMX-600 spec-
trometers. For structural analogues only a selection of significant NMR
data is given. EI and DCI mass spectra (reactant gas ammonia) were
obtained on a Finnigan MAT 95 spectrometer, with high-resolution
data acquired using peak matching (M/DM ) 10 000). ESI mass spectra
were obtained with a Q-Tof II spectrometer from Micromass. HPLC/
DAD/ESI-MS were performed on a PE Sciex Api-2000 LC/MS

1848 Journal of Natural Products, 2008, Vol. 71, No. 11
Ho¨fle and Kunze
Table 3. Feeding of Labeled Precursors of Aurachins A-C (1-3) to Stigmatella aurantiaca
expt
1
precursor fed
[1-13C]anthranilate
isotope abundance
90%
amount fed mg/L culture
75 mg/1 L
metabolites isolated
4.5 mg 13
10 mg 19
3 mg 14
6 mg 17
10 mg 20
2 mg 18
2
3
[1,2-13C2]acetate
[1-13C]anthranilate
[2-13C]acetate
90%
90%
0.42 g/1 L
107 mg/1.6 L
0.6 g/1.6 L
6 mg/0.1 L
4
aurachin C2 (20)
C-3 19%
C-4 78%
7.5% each
7% each
5
6
[1-13C18O2]anthranilate
88 mg/1.2 L
12 mg/0.35 L
13 mg 21
3 mg 15
2 mg 22
3 mg 16
[4-13C18O]aurachin C3 (21)
7
[1-13C]acetate
90%
90%
53%
0.2 g/0.2 L
0.2 g/0.2 L
0.22 L
[2-13C]acetate
18O2 on the second day
in spite of increasing the aeration rate to 300 NL/h and agitation to
520 rpm. After 170 h, toward the end of cultivation, the oxygen
saturation increased to 80% and pH to 8.0. HPLC analysis indicated
the presence of aurachin A (12 mg/L), aurachin B (2 mg/L), and
aurachin C (14 mg/L). Three further peaks with identical characteristic
UV spectra were identified as 5-nitroresorcinol and its mono- and
dimethyl ether.16 However, these compounds were lost during the
standard workup and isolation procedure.
(b) A 100 L bioreactor was charged with 60 L of modified Zein
liquid medium as described above, and 2% XAD adsorber resin was
added before sterilization. Starting immediately after inoculation the
solution of anthranilic acid was continuously pumped in at a rate of 60
mg/h. After 70 h at the end of the cultivation, the pH was regulated
with acetic acid at pH 7.6. Samples of culture including the XAD were
collected in 5 to 15 h intervals and analyzed by HPLC as described.5
After 160 h, cell mass and XAD were collected by centrifugation and
extracted with acetone as described below for procedure a.
Isolation of Aurachins A-L (1-11). Two 65 L fermentation
batches from procedure a were harvested by centrifugation to give 1.6
kg of wet cell mass. This was extracted with 2 L of acetone and then
twice with 1 L each of acetone. The combined extracts were
concentrated in vacuo to an aqueous slurry (1.5 L). Sodium chloride
(150 g) was added followed by extraction with ethyl acetate (4 × 0.4
L). The organic extract was dried with MgSO4 and evaporated to give
a dark brown oil (37 g). For gross separation this was dissolved in 100
mL of dichloromethane and applied to an open column (4.5 × 16 cm)
packed with Florisil in dichloromethane. After elution with dichlo-
romethane/petroleum ether (8:2, 0.5 L), fraction 1 was obtained by
elution with 1.5 L of dichloromethane and dichloromethane/ethyl acetate
(8:2), fraction 2 with 0.5 L of ethyl acetate, and fraction 3 with 1 L of
ethyl acetate/methanol (8:2). Fraction 1 was further separated by MPLC
on Si 60 (Merck) with a dichloromethane/methanol gradient (10:0 to
8:2) to give consecutively aurachins E (5, 10 mg), F (6, 22 mg), D (4,
40 mg), K (10, 13 mg), L (11, 12 mg), and B (2, 10 mg). Fraction 2
was separated by MPLC on silica gel RP-18 with methanol/water/
triethyl amine (80:19:1) to give consecutively aurachins C (3, 830 mg),
B (2, 25 mg), and A (1, 560 mg). Fraction 3 was further separated by
MPLC on Si 60 (Merck) with a dichloromethane/methanol gradient
(10:0 to 7:3) to give consecutively aurachins G (7, 50 mg), H (8, 5
mg), and I (9, 7 mg). Physical data of the aurachins A-D (1-4) are
given in ref 5.
Aurachin H (8): colorless solid; Rf 0.12 (dichloromethane/acetone/
methanol, 75:20:5); UV (MeOH) λmax (log ε) 216 (4.62), 249 (4.57),
339 (4.07) nm; IR (CHCl3) νmax 3445, 1752, 1735, 1613 cm-1; EIMS
m/z 395 [M]+ (22), 379 (22), 184 (100); HREIMS m/z 395.2463 (calcd
for C25H33NO3, 395.2460).
Aurachin I (9): colorless solid; Rf 0.18 (dichloromethane/acetone/
methanol, 75:20:5); UV (MeOH) λmax (log ε) 216 (4.62), 249 (4.56),
350 (3.95) nm; 1H NMR, NOEs H3-13/H-1′, H3-9/H-1′, OMe-1′/H-1′,
H-2′; EIMS m/z 425 [M]+ (46), 409 (15), 214 (100); HREIMS m/z
425.2559 (calcd for C26H35NO4, 425.2566).
Aurachin K (10): colorless solid; Rf 0.50 (dichloromethane/acetone/
methanol, 75:20:5); [R]D20 ca. -1.0 (c 2.3, MeOH); UV (MeOH) λmax
(log ε) 221 (4.58), 242 (4.60), 338 (3.93) nm; EIMS m/z 377 [M]+
(40), 361 (100), 293 (48), 292 (47).
Aurachin L (11): colorless solid; Rf 0.40 (dichloromethane/acetone/
methanol, 75:20:5); [R]D20 ca. -1.0 (c 1.3, MeOH); UV (MeOH) λmax
(log ε) 225 (4.60), 273 (4.49), 280 (4.45), 352 (3.91) nm; EIMS m/z
377 [M]+ (37), 361 (100), 293 (37), 292 (45).
Labeling Studies with Stigmatella aurantiaca Strain Sg a15.
Labeled Precursors. Sodium [1-13C]acetate, sodium [2-13C]acetate,
sodium [1,2-13C2]acetate (97 at %), [13C6]glucose (90 at %), 15NH4Cl
(99 at %), and 18O2 (53 at %) were purchased from MSD Isotopes,
Ba13CO3 (99 at %) from Promochem, and 13C18O2 (99/95 at %),
[2-15N]glutamine and [5-15N]glutamine (99 at %) from Cambridge
Isotope Laboratories.
[13CO2H]Anthranilic Acid. o-Bromoaniline (1.75 g, 1.1 mmol) in
diethyl ether was lithiated with n-butyllithium in hexane (11.5 mL,
2.6 M).32 After 15 min stirring at rt the flask was cooled with liquid
nitrogen and connected by a vacuum line to a second flask containing
Ba13CO3 (2.0 g). By addition of concentrated sulfuric acid 13CO2 was
liberated and condensed on the lithio compound. The reaction mixture
was slowly warmed to rt, and water was added and extracted with
diethyl ether. The anthranilic acid was isolated from the aqueous phase
and purified via salt formation with sodium bicarbonate. The p-isomer
byproduct (ca. 10%) was eliminated by repeated recrystallization from
methanol. Yield: 220 mg (15%); 1H NMR (CD3OD) δ 7.81 (1H, ddd,
J ) 8.0, 1.8, and JC,H ) 4.2 Hz, H-5).33
3
[13C18O2H]Anthranilic Acid. o-Bromoaniline (0.85 g, 0.54 mmol)
was lithiated as described above, and the lithio compound quenched
with 13C18O2 (100 mL, ca. 4 mmol). Due to the excessive formation of
(labeled) pentanoic acid, the yield of anthranilic acid was very low
and isolation by crystallization was complicated. Two batches were
obtained, a fully labeled anthranilic acid (22 mg) and, after addition of
nonlabeled anthranilic acid and crystallization, a diluted labeled material
(88 mg). EIMS m/z 142 [M]+ (13C18O2, 5.6%), 137 M+ (65.4), 122 [M
- H218O]+ (8.4), 119 (M - H2O)+ (100).
Aurachin E (5): colorless needles (diethyl ether); mp 168-169 °C;
Rf 0.79 (dichloromethane/acetone/methanol, 80:18:2); UV (MeOH) λmax
(log ε) 241 (4.43), 249 (4.49), 300sh (3.68), 339 (4.12) nm; IR (CHCl3)
νmax 3445, 1752, 1735, 1654 cm-1; EIMS m/z 404 [M]+ (2), 5 (32),
267 (100); HREIMS m/z 404.2445 (calcd for C26H32N2O2, 404.2464),
335.1757 (calcd for C21H23N2O2, 335.1760), 267.1140 (calcd for
C16H15N2O2, 267.1134).
Aurachin F (6): colorless solid; Rf 0.59 (dichloromethane/acetone/
methanol, 80:18:2); UV (MeOH) λmax (log ε) 241sh (4.27), 253 (4.43),
267sh (4.10), 279sh (3.89), 305 (3.49), 320 (3.62), 336 (3.75), 373
(3.64), 390 (3.69) nm; EIMS m/z 377 [M]+ (4), 226 (20), 210 (100);
HREIMS m/z 377.2363 (calcd for C25H31NO2, 377.2355).
General Feeding Procedure for Strain Sg a15. Shake cultures with
100 to 400 mL of culture medium containing Zein (1%), peptone
(0.1%), and Mg2SO4 ·7H2O (0.1%) at pH 7.3 were inoculated with 5%
of a well-grown culture of Sg a15. During 3 to 4 days the labeled
precursors were added in 4 to 6 portions. The cell mass was harvested
by centrifugation and extracted with acetone. The extract was concen-
trated in vacuo and partitioned between water and ethyl acetate, and
the organic phase was evaporated to give a dark brown oil (1 g/L
culture). The aurachins were isolated by chromatography on Florisil,
silica gel, and RP-18 silica gel, followed, if necessary, by preparative
Aurachin G (7): colorless solid; Rf 0.18 (dichloromethane/acetone/
methanol, 75:20:5); UV (MeOH) λmax (log ε) 210 (4.47), 238 (4.52),
247 (4.55), 350sh (4.02) nm; EIMS m/z 393 [M]+ (4), 377 (8), 172
(100); HREIMS m/z 393.2299 (calcd for C25H31NO3, 393.2304).

Biosynthesis of Aurachins in Stigmatella aurantiaca
Journal of Natural Products, 2008, Vol. 71, No. 11 1849
(14) Ho¨fle, G.; Reichenbach, H. The Biosynthetic Potential of the Myxo-
bacteria. In Sekunda¨rmetabolismus bei Mikroorganismen; Kuhn, W.,
Fiedler, H.-P., Eds.; Attempto Verlag Tu¨bingen, 1995; pp 61-78.
(15) (a) Silakowski, B.; Kunze, B.; Mu¨ller, R Arch. Microbiol. 2000, 173,
403–411. (b) Sandmann, A.; Dickschat, J.; Jenke-Kodama, H.; Kunze,
B.; Dittmann, E.; Mu¨ller, R. Angew. Chem., Int. Ed. 2007, 46, 2712–
2716.
(16) Ho¨fle, G.; Irschik, H. J. Nat. Prod. 2008, in press.
(17) Putra, S. R.; Disch, A.; Bravo, J.-M.; Rohmer, M. FEMS Microbiol.
Lett. 1998, 164, 169–175.
TLC. Due to separation problems, the isolated yields were lower than
expected from HPLC analysis of the crude extracts.
Acknowledgment. We thank S. Reinecke, C. Zorzin, H. Liebler, S.
Ru¨he, and O. Ebelmann for technical assistance, and C. Kakoschke,
B. Jaschok-Kentner, R. Christ, and U. Felgentra¨ger for recording NMR
and mass spectra. We further thank Prof. M. Rohmer and Dr. V. Wray
for helpful discussions.
Supporting Information Available: Time course of aurachin
production, 13C and 1H NMR data of aurachins A-L (1-11), and
biosynthesis schemes for Pseudonocardia and Pseudomonas quinolones
are available free of charge via the Internet at http://pubs.acs.org.
(18) Mahmud, T.; Wenzel, S. C.; Wan, E.; Wen, K. W.; Bode, H. B.;
Gaitatzis, N.; Mu¨ller, R. ChemBioChem 2005, 6, 322–330.
(19) Rohmer, M.; Kani, M.; Simonin, P.; Sutter, B.; Sahm, H. Biochem. J.
1993, 295, 517–524.
(20) (a) Seto, H.; Orihara, N.; Furihata, K. Tetrahedron Lett. 1998, 39,
9497–9500. (b) Seto, H.; Watanabe, H.; Furihata, K. Tetrahedron Lett.
1996, 37, 7979–7982.
References and Notes
(21) Begley, M.; Gahan, C. G.M.; Kollas, A.-K.; Hintz, M.; Hill, C.; Jomaa,
(1) (a) Kunze, B.; Kemmer, T.; Ho¨fle, G.; Reichenbach, H. J. Antibiot.
1984, 37, 454–461. (b) Ho¨fle, G.; Kunze, B.; Zorzin, C.; Reichenbach,
H. Liebigs Ann. Chem. 1984, 1883–1904.
(2) (a) Jansen, R.; Reifenstahl, G.; Gerth, K.; Reichenbach, H.; Ho¨fle, G.
Liebigs Ann. Chem. 1983, 1081–1095. (b) Gerth, K.; Jansen, R.;
Reifenstahl, G.; Ho¨fle, G.; Irschik, H.; Kunze, B.; Reichenbach, H.;
Thierbach, G. J. Antibiot. 1983, 36, 1150–1059.
(3) Silakowski, B.; Kunze, B.; Nordsiek, G.; Blo¨cker, H.; Ho¨fle, G.;
Mu¨ller, R. Eur. J. Biochem. 2000, 267, 6476–6485.
(4) Kunze, B.; Bedorf, N.; Kohl, W.; Ho¨fle, G.; Reichenbach, H. J.
Antibiot. 1989, 42, 14–17.
(5) Kunze, B.; Ho¨fle, G.; Reichenbach, H. J. Antibiot. 1987, 40, 258–
265.
(6) Augustiniak, H.; et al. (GBF) German Patent Appl. DE 3520229 A1,
1986.
(7) Thierbach, G.; Kunze, B.; Reichenbach, H.; Ho¨fle, G. Biochim.
Biophys. Acta 1984, 765, 227–235.
(8) Oettmeier, W.; Godde, D.; Kunze, B; Ho¨fle, G. Biochim. Biophys.
Acta 1985, 807, 216–219.
H.; Eberl, M. FEBS Lett. 2004, 561, 99–104.
(22) See e.g.: (a) Cane, D. E.; Hasler, H.; Liang, T.-C. J. Am. Chem. Soc.
1981, 103, 5960–5962. (b) Lane, P. M.; Nakashima, T. T.; Vederas,
J. C. J. Am. Chem. Soc. 1982, 104, 913–915.
(23) James, C. W.; Kenner, J.; Stubbings, W. V. J. Chem. Soc 1920, 117,
773–776. (a) Schmidt, A.; Shilabin, A. G.; Nieger, M. Org. Biomol.
Chem. 2003, 1, 4324–4350.
(24) Ho¨fle, G.; Bo¨hlendorf, B.; Kunze, B.; Sasse, F. J. Nat. Prod. 2008, in
press.
(25) For reviews see e.g.: (a) Michael, J. P. Nat. Prod. Rep. 2007, 24, 223–
246, and previous articles in this series.
(26) For a review see e.g.: (a) Diggle, S. P.; Cornelis, P.; Eilliams, P.;
Ca´mara, M. Int. J. Med. Microbiol. 2006, 296, 83–91.
(27) Decker, K. A.; Inagaki, T.; Gootz, T. D.; Huang, L. H.; Kojima, Y.;
Kohlbrenner, W. E.; Matsunaga, Y.; McGuirk, P. R.; Nomura, E.;
Sakakibara, Y.; Sakemi, S.; Suzuki, Y.; Yamauchi, Y.; Kojima, N. J.
Antibiot. 1998, 51, 145–152.
(28) In principle, the aurachin intermediate III (Scheme 2) could also
rearrange to the 2-farnesyl isomer. However, this appears to be
disfavored by the 2-methyl and N-oxide groups.
(29) (a) Calfee, M. W.; Coleman, J. P.; Pesci, E. C. Proc. Natl. Acad. Sci.
U.S.A. 2001, 89, 11633–11637. (b) Bredenbruch, F.; Nimtz, M.; Wray,
V.; Morr, M.; Mu¨ller, R.; Ha¨ussler, S. J. Bacteriol. 2005, 187, 3630–
3635.
(30) Neuenhaus, W.; Budzikiewicz, H.; Korth, H.; Pulverer, G. Z. Natur-
forsch. 1979, 34b, 313–315.
(9) Kunze, B.; Ho¨fle, G.; Reichenbach H. Unpublished results.
(10) Friedrich, T.; Van Heek, P.; Leif, H.; Ohnishi, T.; Forche, E.; Kunze,
B.; Jansen, R.; Trowitzsch-Kienast, W.; Ho¨fle, G.; Reichenbach, H.;
Weiss, H. Eur. J. Biochem. 1994, 219, 691–698.
(11) Oettmeier, W.; Dostatni, R.; Majewski, C.; Ho¨fle, G.; Fecker, T.;
Kunze, B.; Reichenbach, H. Z. Naturforsch. 1990, 45c, 322–328.
(12) Gaitatzis, N.; Silakowski, B.; Kunze, B.; Nordsiek, G.; Blo¨cker, H.;
Ho¨fle, G.; Mu¨ller, R. J. Biol. Chem. 2002, 267, 13082–1390.
(13) (a) Silakowski, B.; Kunze, B.; Nordsiek, G.; Blo¨cker, H.; Ho¨fle, G.;
Mu¨ller, R. Eur. J. Biochem. 2000, 267, 6476–6485. (b) Gaitatzis, N.;
Kunze, B.; Mu¨ller, R. Proc. Natl. Acad. Sci. U.S.A. 2001, 98, 11136–
11141. (c) Gaitatzis, N.; Kunze, B.; Mu¨ller, R. ChemBioChem 2005,
6, 365–374. (d) Beyer, S.; Kunze, B.; Silakowski, B.; Mu¨ller, R.
Biochim. Biophys. Acta 1999, 1445, 185–195. (e) Silakowski, B.;
Nordsiek, G.; Kunze, B.; Blo¨cker, H.; Mu¨ller, R. Chem. Biol. 2001,
8, 59–69.
(31) Kitamura, S.; Hashizume, K.; Iida, T.; Miyashita, E.; Shirahata, K.;
Kase, H. J. Antibiot. 1986, 39, 1160–1165.
(32) Gilman, H.; Stuckwisch, C. G. J. Am. Chem. Soc. 1949, 71, 2933–
2934.
(33) For [1′-13C]benzoic acid 3JC-1′, H-2 ) 4.1 Hz; see: Mashall, J. L.; Seiwell,
R. J. Org. Magn. Reson. 1976, 8, 419–425.
NP8003084