Structure and biosynthesis of xenoamicins from entomopathogenic xenorhabdus

Article abstract of DOI:10.1002/chem.201302481

During the search for novel natural products from entomopathogenic Xenorhabdus doucetiae DSM17909 and X. mauleonii DSM17908 novel peptides named xenoamicins were identified in addition to the already known antibiotics xenocoumacin and xenorhabdin. Xenoamicins are acylated tridecadepsipeptides consisting of mainly hydrophobic amino acids. The main derivative xenoamicin A (1) was isolated from X. mauleonii DSM17908, and its structure elucidated by detailed 1 D and 2 D NMR experiments. Detailed MS experiments, also in combination with labeling experiments, confirmed the determined structure and allowed structure elucidation of additional derivatives. Moreover, the xenoamicin biosynthesis gene cluster was identified and analyzed in X. doucetiae DSM17909, and its participation in xenoamicin biosynthesis was confirmed by mutagenesis. Advanced Marfey's analysis of 1 showed that the absolute configuration of the amino acids is in agreement with the predicted stereochemistry deduced from the nonribosomal peptide synthetase XabABCD. Biological testing revealed activity of 1 against Plasmodium falciparum and other neglected tropical diseases but no antibacterial activity.

Full text of DOI:10.1002/chem.201302481

DOI: 10.1002/chem.201302481
Structure and Biosynthesis of Xenoamicins from Entomopathogenic
Xenorhabdus
Qiuqin Zhou,[a] Florian Grundmann,[a] Marcel Kaiser,[b] Matthias Schiell,[c]
Sophie Gaudriault,[d] Andreas Batzer,[c] Michael Kurz,[c] and Helge B. Bode*[a]
Abstract: During the search for novel
natural products from entomopathogenic
Xenorhabdus doucetiae DSM17909 and
X. mauleonii DSM17908 novel pep-
tides named xenoamicins were identi-
fied in addition to the already known
antibiotics xenocoumacin and xenor-
habdin. Xenoamicins are acylated tri-
decadepsipeptides consisting of mainly
hydrophobic amino acids. The main de-
rivative xenoamicin A (1) was isolated
from X. mauleonii DSM17908, and its
structure elucidated by detailed 1D
and 2D NMR experiments. Detailed
MS experiments, also in combination
with labeling experiments, confirmed
the determined structure and allowed
structure elucidation of additional de-
rivatives. Moreover, the xenoamicin
biosynthesis gene cluster was identified
and analyzed in X. doucetiae
DSM17909, and its participation in
xenoamicin biosynthesis was confirmed
by mutagenesis. Advanced Marfeyꢀs
analysis of 1 showed that the absolute
configuration of the amino acids is in
agreement with the predicted stereo-
chemistry deduced from the nonriboso-
mal peptide synthetase XabABCD.
Biological testing revealed activity of
1 against Plasmodium falciparum and
other neglected tropical diseases but
no antibacterial activity.
Keywords: biosynthesis
·
natural
products · peptides · structure eluci-
dation
Introduction
orhabdus have shown promise.[4–6] The bacteria live in sym-
biosis with nematodes of the genus Steinernema and are re-
leased from the nematode gut when the nematode infects an
insect. Besides insecticidal protein toxins Xenorhabdus also
produces small molecules that are involved in insect viru-
lence or act as antibiotics. Examples are the benzylideneace-
tone,[7] rhabduscin,[8] xenocoumacins,[9] xenorhabdins,[10] xen-
ofuranones,[11] nematophin,[12] PAX-Peptides,[13] GameXPep-
tides[14] and even depsipeptides, such as xenematides,[15]
szentiamides,[15] and xentrival peptides,[16] which show antibi-
otic, antifungal, or cytotoxic activity. As several of the com-
pounds may be released to protect the insect cadaver from
food competitors living in the soil, they may also be poten-
tial anti-infectives. Here we describe the structure and bio-
synthesis of a novel class of large hydrophobic depsipeptides
named xenoamicins from Xenorhabdus showing activity
against Plasmodium and Trypanosoma.
Natural products have always been a great source of novel
bioactive compounds, in particular in the field of anti-infec-
tives. As the number of multiresistant pathogens is increas-
ing, it is necessary to continue research in the field of natu-
ral products,[1] and thus novel approaches for the identifica-
tion of natural products from well-established or novel natu-
ral-product producers are required. Additionally, novel mo-
lecular approaches also allow the activation of previously
silent gene clusters or heterologous expression.[2,3] Among
novel natural-product producers, bacteria of the genus Xen-
[a] Q. Zhou,+ F. Grundmann,+ Prof. Dr. H. B. Bode
Goethe-Universitꢁt Frankfurt
Department of Molecular Biotechnology
Max-von-Laue-Strasse 9, 60438 Frankfurt am Main (Germany)
E-mail: h.bode@bio.uni-frankfurt.de
[b] M. Kaiser
Swiss Tropical and Public Health Institute, Parasite Chemotherapy
Socinstrasse 57, P.O. Box,
Results and Discussion
4002 Basel (Switzerland)
During our screening for novel natural products by UPLC-
ESI-MS of XAD-16 extracts[17] from Xenorhabdus cultures
in standard LB medium we identified a new family of natu-
ral products. According to characteristic neutral losses for
amino acids such as valine, leucine, and isoleucine in their
MS2 fragmentation patterns, these compounds were identi-
fied as peptides.[18] Their molecular formulas were deter-
mined on the basis of MALDI-Orbitrap-MS data (Table S1
of the Supporting Information) revealing C64H109N13O15 for
xenoamicin A (1; m/z 1300.822 [M+H]+, Dppm 0.5) as the
[c] M. Schiell, Dr. A. Batzer, Dr. M. Kurz
Sanofi R&D, Industriepark Hçchst
65926 Frankfurt am Main (Germany)
[d] Dr. S. Gaudriault
INRA, UMR 1333 Laboratoire DGIMI, 34095 Montpellier (France)
and
Universitꢂ Montpellier 2, UMR 1333 Laboratoire DGIMI
34095 Montpellier (France)
[+] These authors contributed equally to this work.
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.201302481.
16772
ꢃ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2013, 19, 16772 – 16779

FULL PAPER
main product in X. mauleonii DSM17908. The determina-
tion of the molecular formula was facilitated and confirmed
by the identification of the correct number of carbon and ni-
trogen atoms from labeling experiments with strain DSM
17908 in standard growth medium, 13C-labeled medium, and
15N-labeled medium (Figure S1 of the Supporting Informa-
tion), as described previously.[14,16] Compound 1 (29.2 mg)
was subsequently isolated from strain DSM17908 by prepa-
rative chromatography with HPLC-UV/ELSD in a three
step protocol (Figures S2–S4 of the Supporting Informa-
tion). The first two chromatographic steps were performed
on C18 columns with different particle sizes under acidic con-
ditions. In the third step the selectivity was changed through
a higher pH value of the eluent. Due to the lack of a chro-
mophore, 1 shows only weak absorption in the UV spectrum
even at 215 nm. Despite several peaks with a strong absorp-
tion at 215 nm in the extract of X. mauleonii DSM17908,
1 could easily be identified with the universal signal of the
evaporative light scattering detector (ELSD). Due to the
relatively large amount of 1 in the extract of DSM17908
a large signal in the ELSD but not in the UV spectra could
be observed, whereas other compounds in this extract with
a chromophore gave large signals in both spectra (Figure S2
of the Supporting Information). Thus, the ELSD signal gives
a very good indication of the actual amount of the individu-
al compounds in the extract, independent of the physico-
chemical properties, and therefore ELSD is especially valua-
ble for the detection of peptides without any chromo-
phore.[19]
The structure of 1 was determined by comprehensive
NMR studies. In most standard solvents, such as
[D6]DMSO, CD3OD, or CDCl3, NMR spectra of poor quali-
ty were obtained. The resonances of the amide protons in
particular appeared as very broad signals (Figures S5, S6
and Annex 1 of the Supporting Information). However, in
[D6]benzene at 293 K well-resolved NMR spectra with rea-
sonable dispersion of the amide signals were obtained (Fig-
ure S6 of the Supporting Information). The assignment of
all proton and carbon resonances was carried out with vari-
ous 2D NMR techniques, including DQF-COSY, TOCSY,
ROESY, multiplicity-edited HSQC, HSQC-TOCSY, and
HMBC spectra (Table 1, Annexes 1–20 in the Supporting In-
formation). The assignment was hampered by the presence
of two sets of signals in a ratio of 1:1 resulting from proline
cis/trans conformers (see below). Primary analysis of the
spin systems in the TOCSY and HSQC-TOCSY spectra re-
vealed the presence of 13 amino acid residues: five valines,
one leucine, one isoleucine, two alanines, two prolines, one
threonine, and one b-alanine residue. In addition a butyric
acid moiety was observed. The two sets of signals are caused
by different orientations of the amide bond between the b-
alanine residue and one of the proline residues. In the cis
conformer an intensive ROE correlation between proline
Ha and b-alanine Hb is observed. In the the trans confor-
mer such an intensive ROE should appear between proline
Hd and b-alanine Hb. Unfortunately, this ROE cannot be
assigned, because of peak overlaps between proline Hd and
the Hb of b-alanine. The presence of a cis and a trans con-
former is also confirmed by the difference in 13C chemical
13
shifts of Cb and Cg of proline [D(bg)=d(13CbÀ Cg)].[20]
The differences for the cis and trans conformers are D(bg)=
8.83 ppm and D(bg)=5.17 ppm, respectively, which are in
agreement with literature data.[20] Due to overlap of the
carbon resonances in the carbonyl region and line broaden-
ing of some amide protons, the HMBC spectrum was not
sufficient for the sequential assignment (see Annex 3 of the
Supporting Information). Therefore, the peptide sequence
was determined by analysis of ROE correlations which re-
sulted in a depsipeptide structure with ring formation be-
tween the C-terminal valine residue and the threonine side
chain at position 7 (Figure 1). The ring closure could be con-
firmed by an ROE between threonine Hb and valine 14 Ha
in both conformers, whereby the ROE of the cis conformer
is only weak (Annex 13 of the Supporting Information). All
other sequential NH/Ha ROE correlations of both conform-
ers were observed with the exception of the correlation be-
tween valine 10 and alanine 11 in both conformers. Never-
theless, the connection of valine 10 and alanine 11 was in ac-
cordance with the molecular formula and was shown by MS
fragmentation data (see below). The assignment of the
single spin systems to the corresponding conformer was con-
firmed by additional long-range ROE correlations between
valine 10 NH and isoleucine 6 Ha. Moreover, a long-range
ROE correlation between threonine Hb and alanine 11 Ha
confirmed the assignment of the trans conformer. The N-ter-
minal proline residue is acylated with butyric acid, as could
be concluded from an ROE correlation between Ha of the
acid and Ha of proline in position 2. Unfortunately, the ala-
nine NH in position 11 of the trans conformer could not be
detected due to very broad signals.
The configuration of the amino acids in 1 was elucidated
by using the advanced Marfeyꢀs method[16,21,22] leading to the
identification of the l configuration for isoleucine and both
proline residues. The d configuration was revealed for leu-
cine, allo-threonine, and both alanine residues. The ratio of
l to d valine was found to be 1:4. Therefore, four valine res-
idues in 1 may have the d configuration and one valine resi-
due has the l configuration (Figure 2, Table S2 of the Sup-
porting Information). However, the stereochemical assign-
ment of the valine residues was not possible at this stage.
To verify and support the structure elucidation by NMR
spectroscopy and to enable the structural elucidation of fur-
ther derivatives in X. doucetiae DSM17909, MS-based struc-
ture elucidation was additionally performed for
1 in
DSM1908. To confirm the building blocks, standard 12C-l-
amino acids were added to bacterial cultures growing in a re-
versed labeling experiment in fully 13C labeled medium (Fig-
ure S7 of the Supporting Information), as described previ-
ously.[14,16] Incorporation could be detected by HPLC-ESI-
MS as shifts to lower masses of m/z 1.5 for b-alanine, 2 for
threonine, 2.5 for proline, 3 for isoleucine, 3 for leucine, 1.5
for alanine, and 2.5 for valine, related to half of the numbers
of carbon atoms incorporated due to the presence of doubly
charged ions.
Chem. Eur. J. 2013, 19, 16772 – 16779
ꢃ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
16773

H. B. Bode et al.
Table 1. NMR spectroscopic data of conformer mixture (600 MHz (1H),
150 MHz (13C) in [D6]benzene) of xenoamicin A (1), d in ppm. Spin
system numbering from acylated N- to C-terminus. Only key sequential
ROESY correlations are listed. Pos.=Position.
Table 1. (Continued)
Amino acid Pos.
trans Conformer
cis Conformer
dC
dH
ROESY
dC
dH
ROESY
13-Pro
a
b
63.19
30.21
4.47
2.17
1.73
1.40
61.47
31.39
4.26 12-a, 12-b
1.76
1.66
1.26
1.11
3.54 14-NH
3.22 14-NH
6.65 13-d
4.51 7-b
2.17
Amino acid Pos.
trans Conformer
cis Conformer
dC
dH
ROESY
dC
dH
ROESY
b
1-BA
2-Pro
a
36.46
36.46
18.63
2.34 2-a
1.95
1.90
36.46
36.46
18.63
2.34 2-a
1.95
1.90
1.85
0.98
4.58 1-a, 3-NH
1.82 3-NH
1.47 3-NH
2.08
1.29
3.28
g
25.04
46.57
22.56
46.48
a
g
1.68
b
d
3.20 14-NH
2.81 14-NH
6.60 7-b, 13-d
4.69 7-b
2.27
b
1.85
d
g
14.23
60.30
29.54
0.98
14.23
14-Val
NH
a
b
a
4.58 1-a, 3-NH 60.30
57.70
32.77
18.83
18.33
56.14
32.05
18.81
16.98
b
1.82 3-NH
1.47 3-NH
2.08
1.29
3.28
29.54
24.96
47.13
b
g
0.97
0.96
0.66
0.63
g
24.96
47.13
g
g
[a] Interchangeable spin systems in the different conformers at the same
position. [b] Interchangeable positions in the spin systems of the different
conformers at the same position.
d
d
2.73
2.73
3-Ala[a]
4-Val[a]
NH
a
8.71 2-a, 2-b
8.71 2-a, 2-b
48.06[b] 5.01 4-NH
48.01[b] 5.03 4-NH
b
16.56
1.69
16.56
1.67
NH
a
7.82 3-a
4.84 5-NH
2.25
1.10
1.04
7.82 3-a
4.78 5-NH
2.25
1.06
1.00
60.04
30.70
19.63
19.45
60.08
30.70
19.64
19.38
b
g
g
5-Leu
6-Ile
NH
a
9.22 4-a
5.14 6-NH
2.03
9.23 4-a
5.13 6-NH
2.00
50.76
40.34
50.93
40.34
b
g
25.56[b] 1.97
25.50[b] 1.97
d
22.59
23.45
1.09
1.10
22.59
23.45
1.09
1.10
d
Figure 1. Connectivities in xenoamicin A (1), as determined by COSY
and TOCSY (bold lines). Arrows indicate ROESY correlations used for
the sequential assignment. Additionally the numeration of the spin sys-
tems is shown.
NH
a
8.70 5-a
5.00 7-NH, 7-b, 58.01
10-NH
2.34
0.96
1.72
1.52
1.14
8.74 5-a
4.92 7-NH, 7-b,
10-NH
2.34
0.94
1.72
1.52
1.12
58.05
b
34.91
9.75
25.44
34.83
9.75
25.39
g1
g2
g2
d
15.91
15.87
7-Thr
8-Val
9-Val
10-Val
NH
a
9.60 6-a
5.18 8-NH
6.05 6-a, 11-a,
14-a
9.54 6-a
5.15 8-NH
5.82 6-a, 14-a
56.71
70.03
56.14
71.82
b
g
17.66
1.47
17.24
1.49
NH
a
9.03 7-a, 7-b
4.62 9-NH
2.57
1.37
1.27
8.84 7-a, 7-b
4.56 9-NH
2.56
1.33
1.28
61.88
29.36
19.72
19.55
61.54
29.54
19.71
19.93
b
g
g
NH
a
8.97 8-a
4.76 10-NH
8.85 8-a
4.79 10-NH
60.82
60.57
b
30.45[b] 2.55
30.49[b] 2.56
g
19.99
19.99
1.24
1.24
9.15 6-a, 9-a
4.90
19.91
19.91
1.25
1.25
9.01 6-a, 9-a
5.49
g
NH
a
Figure 2. Structures of xenoamicins A–H (1–8) indicating b2–b5 and y2–
y5 fragmentation ions of the MS2 spectra. No differentiation between
leucine or isoleucine at position 9 (R4) could be obtained for 4–8.
60.30
29.68
58.64
31.07
b
2.56
2.38
g
19.44[b] 1.23
19.50[b] 1.23
g
19.36
1.18
19.09
1.17
11-Ala
NH
a
9.35
5.16 12-NH
1.58
6.70 11-a
4.42 13-a
4.26
The more intense doubly charged ions were used for frag-
mentation and showed mainly b2–b5 and y2–y6 ions
(Figure 2, Figure S8 and Table S3 of the Supporting Infor-
mation), as xenoamicins exhibited a characteristic collision-
induced dissociation (CID) MS2 fragmentation pattern.
However, as the ring stayed intact in the CID experiment,
the sequence of the ring could not be determined. There-
49.93
17.24
5.45 7-b, 12-NH 50.31
b
1.75
7.97 11-a
4.20
3.38
2.84
19.29
35.79
35.13
12-b-Ala
NH
a
35.66
34.80
a
b
3.00
2.35 13-a
b
2.18
16774
www.chemeurj.org
ꢃ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2013, 19, 16772 – 16779

Structure and Biosynthesis of Xenoamicins
FULL PAPER
Figure 3. Structures of linear xenoamicins A–C (1–3) after reaction with NH3. Described b1–b11 ions could be observed in MSn fragmentations. For R1–
R4 see Figure 2.
porting Information). Both
show m/z 1314.8 [M+H]+ indi-
cating the presence of an addi-
tional methyl group in both
compounds. The location of this
additional methyl group in 2
and 3 could be identified as po-
sition 9 due to the difference
between b7 and b8 ions (Fig-
ures 3 and 4; Figure S11 of the
Supporting Information). In
1 the difference between the b7
and b8 ions of 99 Da represents
the neutral loss of a valine
building block, whereas the dif-
ference in 2 and 3 of 113 Da in-
dicates leucine or isoleucine, re-
spectively. Isobaric leucine and
isoleucine could be distinguish-
ed by analyzing the results from
inverse feeding experiments
(Figure S7 of the Supporting In-
formation). By feeding 12C-leu-
cine to a culture in 13C-enriched
media two shifts of 3 Da from
Figure 4. MS3 and MS4 spectra of linear xenoamicin B (2) after reaction with NH3. According to MS3 and MS4
spectra, b2–b5 and b8 ions could be assigned in MS2 spectra. Assignment of b9–b11 ions were proposed due to
complex MS3 and MS4 fragmentation.
the double charged species
could be seen for 2 and show
the incorporation of two leu-
cine residues instead of one leu-
cine residue in 1 and 3. For 3
fore, xenoamicin-containing extracts were hydrolyzed with
28% aqueous NH3 solution, which resulted in positive mass
shifts of 8.5 Da and 9 Da in doubly charged species for the
addition of NH3 (17 Da; resulting in the linear amide) or for
the addition of water (18 Da; resulting in the linear acid),
respectively (Figure 3, Table S4 of the Supporting Informa-
tion).[23] Based on CID MS2 of the linearized peptide (Fig-
ure S9 of the Supporting Information) and MSn experiments
(Figure S10 of the Supporting Information), b1–b8 ions
could be identified and also used for determination of the
amino acid sequence of the ring, which confirmed the se-
quence of 1 as determined by NMR spectroscopy. The same
analyses were performed for compounds 2 and 3, which
appear in X. mauleonii DSM17908 and X. doucetiae
DSM17909 (Figures S8 and S9, Tables S3 and S4 of the Sup-
two shifts for isoleucine could be observed instead of one
shift in 1 and 2 after the feeding of 12C-isoleucine. The ex-
tracts of the inverse feeding experiments were also hydro-
lyzed with 28% aqueous NH3 solution. However, the inten-
sity of the b6 and b7 ions of the 13C-peptides with incorpo-
rated 12C-leucine or 12C-isoleucine was too low for detection
(Table S4 of the Supporting Information). In summary, MS4
fragmentation and CID MS2 fragmentation of the linearized
peptides together with inverse feeding experiments were
sufficient for the complete sequential structural elucidation
of 1–3.
Additionally, xenoamicins D–H (4–8) could be detected in
DSM17909, but not in DSM17908, as doubly charged spe-
cies ([M+2H]2+, m/z 644.0, 651.0, 651.0, 665.0), although
these were only produced in trace amounts (Table S1 of the
Chem. Eur. J. 2013, 19, 16772 – 16779
ꢃ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
16775

H. B. Bode et al.
Supporting Information). Their structure could only partial-
ly be determined from their b2–b5 and y2–y6 ions in MS2
spectra (Figure S12 and Table S5 of the Supporting Informa-
tion). Determination of the ring amino acids of 4–8 was not
possible. However, as variability in the ring was only ob-
served at position 9 in compounds 1–3, this may also be the
variable position for compounds 4–8, although no differen-
tiation between leucine and isoleucine is possible. As de-
scribed above, the differentiation of leucine and isoleucine
can only be achieved by reversed feeding experiments and
subsequent linearization if the structure elucidation is per-
formed by MS. However, the amounts of 4–8 were not suffi-
cient for further characterization.
Nevertheless, 5 and 6 show variability in positions 6 and
3, respectively. Furthermore, two additional N-terminal acyl
moieties, acetate and pentanoate, were observed in 4 and 8.
Experimentally proven characterization of the ring amino
acids is not essential for elucidation of these variabilities.
The b2 ion of 4 with a mass of 211 Da revealed that ace-
tate is acylated with the N-terminal proline residue. The
neutral loss of proline is 97 Da and that of alanine 71 Da.
Therefore the remaining component of the b2 ions had
a mass of 43 Da indicating acetate. By comparison of the b2
ions with 211 Da and the y2 ions with 1147 Da, alanine in
position 3 is confirmed. Valine in position 4, leucine in posi-
tion 5, and isoleucine in position 6 could be predicted by the
mass shifts of 99, 113, and 113 Da between the y3, y4, y5,
and y6 ions, respectively. Leucine and isoleucine cannot be
distinguished by their mass, but it can be assumed that leu-
cine is incorporated at position 5 and isoleucine at position
6, as in 1–3. However, CID MS2 fragmentation was not suf-
ficient for the analysis of the ring. Therefore, it had to be as-
sumed that the ring consists of the same or very similar
building blocks as in 1–3. However, because the y6 ions of 4
showed a mass of 751 Da, like for the y6 ions of 2 and 3, we
propose that leucine or isoleucine was incorporated at posi-
tion 9, as in 2 and 3, and that all other amino acids in the
ring were as in 1–3.
ever, the b2 ions with a mass of 253 Da showed that 7 con-
tained a pentanoyl moiety as the difference of the y2 and
the y3 ions indicated an alanine in position 3. When the
mass of alanine and proline were subtracted from the mass
of the b2 ions the mass of a pentanoyl moiety is left.
The MS2 spectrum of 8 is more complex. Therefore the
MS2 spectrum of the sodium adducts of 8 is additionally pre-
sented (Figure S12 of the Supporting Information). The y2–
y6 ions showed that 8 exhibited the same amino acid se-
quence and ring size as 2–4 and 7. Taking the molecular
mass of 1244 Da into account, it can be assumed that the N-
terminus of 8 is not acylated.
To fully assign the absolute configuration of xenoamicin,
which was not completely possible from the NMR data and
amino acid analysis, the biosynthesis gene cluster was identi-
fied and analyzed. As the genome of X. mauleonii
DSM17908 is not available yet, the already finished genome
sequence of strain DSM17909 (originally strain FRM16)
(http://www.genoscope.cns.fr/agc/microscope/home/in-
dex.php) was searched for a biosynthesis gene cluster encod-
ing nonribosomal peptide synthetases (NRPS) involved in
xenoamicin biosynthesis (xab) by using the antismash soft-
ware tool.[24–26] The only candidate that fits the predicted
biosynthesis gene cluster encodes five nonribosomal pepti-
dases XabABCD and an aspartic acid decarboxylase XabE
putatively involved in the formation of b-alanine.[14,27,28]
To prove that this gene cluster is indeed involved in xen-
oamicin biosynthesis, the gene xabB encoding the second
NRPS was disrupted by plasmid insertion. Comparison be-
tween the xabB:cat mutant and the wild type showed com-
plete loss of 1–8 in the mutant, whereas all other natural
products such as such as xenorhabdins and xenocoumacins
were still produced (Figures S13 and S14 of the Supporting
Information).
Overall 13 modules and all domains for the biosynthesis
of the xenoamicins could be identified (Figure 5, Table S6 of
the Supporting Information). The presence of an adenyla-
tion (A) domain specific for b-alanine was readily detecta-
ble due to the differences in the specificity conferring code
(Table S7 of the Supporting Information).[29–31] Analysis of
the condensation (C) domains[32] revealed the presence of
dual condensation/epimerization (C/E) domains[33] in mod-
ules 3, 5, 7, 9, and 11. Thus, it can be assumed that the
amino acids incorporated in the previous modules 2, 4, 6, 8,
and 10 are epimerized, which is in accordance with the re-
sults from the advanced Marfeyꢀs analysis (Table S2 of the
Supporting Information). For example, exclusively l-proline
and d-alanine were detected in the Marfeyꢀs analysis and
only C and C/E domains were identified following the mod-
ules responsible for the incorporation of proline (modules
1 and 12) and alanine (modules 2 and 10), respectively. Fur-
thermore, 1–3 were identical in both strains according to the
MS analysis of the linearized peptides as well as the reten-
tion times of cyclic or linear peptides. Thus, the absolute
configuration of all amino acids in xenoamicins can be pre-
dicted accordingly and this also allows the stereochemical
assignment for the valine moieties to be made (Figure 2).
The CID MS2 fragmentation revealed that compound 5
has alanine in position 3, valine in position 4, leucine in po-
sition 5, and valine in position 6 due to the shifts of 71, 99,
113, and 99 Da between the y2, y3, y4, y5, and y6 ions, re-
spectively. Therefore this compound shows variability in po-
sition 6 due to the incorporation of valine instead of isoleu-
cine. The b2 ions with a mass of 239 Da revealed a butanoyl
acid unit in position 1. The obtained y6 ions for compound 5
exhibited the same size as for compound 4, 6, 7, and 8, thus
the structure shown in Figure 2 resulted.
In compound 6 glycine instead of alanine is incorporated
in position 3, as revealed by the mass shift of 57 Da between
the y2 and y3 ions. According to the shifts between y3, y4,
y5, and y6 ions, valine, leucine, and isoleucine are incorpo-
rated in positions 4, 5, and 6, respectively. The loss of 14 Da
of the b2 ions in comparison with the b2 ions of 1, 2, 3, and
5 resulted from incorporation of glycine instead of alanine.
Compound 7 showed the same mass shifts as 2–4. There-
fore, the same amino acid sequence can be assumed. How-
16776
www.chemeurj.org
ꢃ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2013, 19, 16772 – 16779

Structure and Biosynthesis of Xenoamicins
FULL PAPER
Figure 5. Proposed biosynthesis of xenoamicin A (1) in X. doucetiae DSM 17909, showing the nonribosomal peptide synthetases XabABCD involved in
its biosynthesisis. Domains and modules are labeled: adenylation domain (A), peptidyl carrier protein domain (PCP), condensation domain (C), conden-
sation/epimerization domain (C/E), thioesterase domain (Te).
With alternating l and d configuration in the peptide
chain, xenoamicins belong to the class of peptides which can
form b helices in membranes and hence ion channels. The
best-studied example is the antibiotic gramicidin D,[34] which
is also formed nonribosomally and requires two molecules
to span a membrane. Recently, a novel class of ribosomally
made alternating d/l peptides, namely, proteusins, which has
strong cytotoxic activity and polytheonamide as most promi-
nent member, has been identified and its biosynthesis eluci-
dated.[35]
In contrast to these d/l peptides, 1 most likely cannot
form such a b-helix due to the depsipeptide structure. Nev-
ertheless, the presence of such a large number of hydropho-
bic amino acids and a N-terminal acyl moiety in 1 may indi-
cate an interaction with the membrane. Hence, the bioactiv-
ity of xenoamicin A (1) was investigated. Whereas no anti-
bacterial or antifungal activity could be observed in an agar
diffusion assay with E. coli DH10B, B. subtilis, M. luteus, P.
aeruginosa, S. cerevisiae, and C. albicans,[36] good activity
against Plasmodium falciparum NF 54 (IC50 of 2.35 mgmLÀ1)
and Trypanosoma brucei rhodesiense STIB900 (IC50 of
6.41 mgmLÀ1) was observed. Much weaker activity against
Trypanosoma cruci Tulahuen C4 (IC50 of 30.5 mgmLÀ1), L.
donnovanni MHOM-ET-67L82 (IC50 of 50.1 mgmLÀ1), and
mammalian L6 cells (IC50 of 68.5 mgmLÀ1) indicated a specif-
ic target.[37] However, further analysis is needed to identify
this target and also to elucidate the natural function of this
novel class of compounds in the complex life cycle of Xenor-
habdus with its nematode host and insect prey.
Experimental Section
Xenoamicin A (1): White powder; 1H NMR (600 MHz, [D6]benzene,
258C): see Table 1; 13C NMR (150 MHz, [D6]benzene, 258C, see Table 1;
MS/MS (70 eV) and HRMS (MALDI): see Table S1 of the Supporting
Information.
Cultivation and extraction: X. doucetiae DSM17909 and X. mauleonii
DSM17908 were always cultivated at 308C, 200 rpm. For inoculation, 1%
overnight preculture and 2% XAD-16 in LB medium were used. For an-
alytical and preparative HPLC, LB medium was used (20 mL and 12 L,
respectively). XAD-16 resin beads were collected with a sieve after 72 h,
washed with a small amount of water, and extracted with methanol to
yield the crude extract (9.5 g) after evaporation from the large-scale culti-
vation.
Feeding experiments: The strains DSM17909 and DSM17908 were culti-
vated at 308C overnight. The cell pellet of the overnight culture was
washed once with ISOGRO-13C medium (1 mL) and dissolved again in
13C medium (1 mL). ISOGRO-13C medium was prepared with ISO-
GRO-13C powder (1 g), K2HPO4 (1.8 gLÀ1), KH2PO4 (1.4 gLÀ1), and
CaCl2·xH2O (11.1 mgLÀ1) dissolved in H2O (100 mL). The feeding cul-
ture in 13C medium (5 mL) was inoculated with washed overnight culture
(50 mL). After incubation for 6 h at 308C, a stock solution (50 mL,
100 mm) of the respective amino acid (b-alanine, l-a-threonine, l-a-pro-
line, l-a-isoleucine, l-a-leucine, l-a-alanine, or l-a-valine) was added.
Another two additions were carried out at 24 and 48 h. The cultures were
extracted with ethyl acetate (5 mL) after incubation for 72 h. The extract
was then prepared for HPLC-MS analyses.[14]
Hydrolysis of the extract: The extract was treated with ammonium hy-
droxide solution (Sigma-Aldrich, 28%) to open the ring in the depsipep-
tides.[23] Hydrolysis was started with methanol extract solution (50 mL)
and ammonium hydroxide solution (500 mL). After incubation for 2 h at
408C, the reaction mixture was neutralized with HCl (3m) solution. After
evaporation of the solvent, the sample was then prepared for HPLC-MS
analyses.
Chem. Eur. J. 2013, 19, 16772 – 16779
ꢃ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
16777

H. B. Bode et al.
HPLC and mass spectrometric analysis of the extract: Analysis of the ex-
tracts was carried out by using an Ultimate 3000 LC system from
Dionex, coupled to an amaZon X electrospray ionization mass spectrom-
eter from Bruker Daltonics. Peptides were separated on a C18 Column
dation and complete assignment of proton and carbon resonances, 1D
1H, 1D 13C, DQF-COSY, ROESY (mixing time 150 ms, spinlock field
1
2 kHz), multiplicity-edited HSQC, and HMBC spectra were acquired. H
chemical shifts were referenced to the solvent signals (1H: 7.15 ppm, 13C:
128.0 ppm).
(ACQITY UPLC BEH, 1.7 mm, 2.1ꢄ50 mm, flow rate 0.6 mLminÀ1
,
Waters). Acetonitrile/water containing 0.1% HCOOH was used as
mobile phase under a linear gradient from 40–50% acetonitrile over
17.5 min for the separation. Collision-induced dissociation (CID) was
performed on the ion trap in the amaZon X in positive mode with a scan
range of m/z 100–1500.[17] HR-MALDI-MS data were obtained with
a MALDI LTQ Orbitrap XL from Thermo Fisher Scientific equipped
with a 337 nm nitrogen laser. The extract was diluted in acetonitrile for
MALDI-MS analysis. 0.2 mL of the extract solution and 0.5 mL of
a 20 mm 4-cloro-a-cyanocinnamic acid (ClCCA) solution (70% acetoni-
trile) were spotted on the MALDI target and air-dried. The following in-
strument parameters were used: laser energy, 25 mJ; automatic gain con-
trol, on; auto spectrum filter, off; resolution, 100000; plate motion,
survey CPS. Mass spectra were internally calibrated with calibration mix-
ture 2 (Applied Biosystems, Sequazyme peptide mass standard kits). The
molecular formulas were calculated by using the Qual Browser software
V. 2.0.7 (Thermo Fisher Scientific).[18]
Two-dimensional homonuclear experiments (DQF-COSY, TOCSY, and
ROESY) were performed with a spectral width of 11 ppm. Spectra were
recorded with 1024 increments in t1 and 4096 complex data points in t2.
For each t1 value 4 (DQF-COSY), 8 (TOCSY), or 16 (ROESY) transi-
ents were averaged.
For the HSQC spectrum 1024 increments with 3072 complex data points
in t2 were collected with sweep widths of 8 ppm in the proton and 90 ppm
in the carbon dimension. For each t1 value four transients were averaged.
For the HSQC-TOCSY spectrum 1024 increments with 3072 complex
data points in t2 were collected with a sweep width of 11 ppm in the
proton and 80 ppm in the carbon dimension. For each t1 value 32 transi-
ents were averaged, and a mixing time of 80 ms was used for the TOCSY
transfer. The HMBC spectrum was acquired with a sweep width of
11 ppm in the proton and 200 ppm in the carbon dimension by using a de-
focusing delay of 62 ms (optimized for coupling constants of 8 Hz). A
total of 32 transients were averaged for each of 1024 increments in t1,
and 4096 complex points in t2.
Advanced Marfeyꢀs method: Compound 1 (ca. 0.5 mg) was hydrolyzed
with HCl (6m, 0.8 mL) in a high-pressure Ace tube at 1108C for 16 h.
The hydrolysate was evaporated to dryness and resuspended in H2O
(100 mL). To 50 mL of this solution NaHCO3 (10 mL, 1m) and Na-(5-
fluoro-2,4-dinitrophenyl)-l-leucinamide or -d-leucinamide (l-FDLA or
d-FDLA, 100 mL of a 1% solution in acetone) were added. The reaction
vials were covered with aluminum foil and placed in a water bath at
408C for 1 h. Then the reaction mixtures were cooled to room tempera-
ture, quenched with HCl (10 mL, 1m), and evaporated to dryness.[21,22]
The residue was dissolved in methanol (400 mL). The analysis of l- and
d-FDLA-modified amino acids was carried out by LC-MS, as mentioned
above for the analysis of the extract. Acetonitrile/water containing 0.1%
HCOOH was used as mobile phase under a linear gradient from 20–60%
acetonitrile over 34 min for the separation. At same time, the commercial
standard amino acids were prepared as reference.
Identification and verification of the gene cluster: The proposed gene
cluster responsible for the biosynthesis of xenoamicins from Xenorhab-
dus doucetiae DSM17909 (originally strain FRM16) was analyzed by
using PKS/NRPS Analysis Web-site (http://nrps.igs.umaryland.edu/
nrps/).[24] To confirm the gene cluster, gene xabB was disrupted by inser-
tion of a plasmid. Xenorhabdus doucetiae DSM17909 was naturally resist-
ant to ampicillin, which was used to select Xenorhabdus after conjuga-
tion. Primer pair 5’-GAACTGGCATGCGGAAATTGAGGCGCAAC
and 5’-ACAAGAGAGCTCCCTTCCTGAAGCGGTG were used to am-
plify parts of gene xabB, which were then digested and cloned into
pDS132 vector and transformed into E. coli S17 (lpir). After conjuga-
tion, primer pair 5’-GCCCGATATCCTGTCATTG and 5’-ACATGTG-
GAATTGTGAGCGG was used to verify the insertion of the plasmid in
the mutants. The detailed method was used by us before.[38] The extract
from the mutant was prepared as for the wild type, except for chloram-
phenicol in the LB medium for the mutant.
Isolation: For isolation of 1, the XAD extract (9.5 g) of X. doucetiae
DSM 17909 was dissolved in 50 mL of DMSO/MeOH/iPrOH (7:2:1). The
resulting suspension was centrifuged at 13.3ꢄ103 rpm for 4 min and the
pellet was discarded after further analysis. The following preparative
HPLC setup was used: 4 mL per injection by using a Gilson 231 XL
sample injector with a sandwich injection method, a Gilson 402 syringe
pump, a Gilson 215 liquid handler and two Varian Prep Star SD-1 pumps
with a flow rate of 95 mLminÀ1 and an acetonitrile/water gradient from
0–1 min 5%, 1–16 min 5–95%, 16–19 min 95%, a Gilson 306 pump for
Acknowledgements
Work in the Bode lab was supported by the excellence initiative of the
Hessian Ministry of Science and Art via the LOEWE research focus
“Insect Biotechnology” and a European Research Council starting grant
under grant agreement no. 311477. The authors are grateful to Geno-
scope (Evry, France) for access to the X. doucetiae genome prior to pub-
lication.
the addition of 10% HCOOH buffer with a flow rate of 2.5 mLminÀ1
,
a Dionex P580 pump for the makeup flow of 1 mLminÀ1 acetonitrile,
a Jasco UV 975 recording at 215 nm, an MRA Active Splitter, a Varian
380LC ELSD, and a Luna C18 10 mm 50ꢄ50 column from Phenomenex
(for the chromatogram, see Figure S2 of the Supporting Information).
The obtained fractions were freeze-dried. The fractions containing
1 were combined and purified in a second step on the same instrument
setup, but with a flow rate of 75 mLminÀ1 and an acetonitrile/water gra-
dient from 0–1 min 45%, 1–16 min 45–65%, 16–19 min 65–95% and
a Luna C18 5 mm 30ꢄ75 column. The buffer for the second step was
changed to 3.33% trifluoroacetic acid buffer, which was added at a flow
rate of 2.5 mLminÀ1 (Figure S3 of the Supporting Information). In a third
purification step the selectivity of the chromatographic system was
changed through the pH value. Therefore, an aqueous solution of ammo-
nium acetate (50 gLÀ1) was used as buffer. A flow rate of 75 mLminÀ1
with an acetonitrile/water gradient from 0–1 min 5%, 1–16 min 5–95%,
16–19 min 95% and a Luna C18 5 mm 30ꢄ75 column (Figure S4 of the
Supporting Information) were used.
[1] H. B. Bode, Angew. Chem. 2009, 121, 6512–6514; Angew. Chem. Int.
Ed. 2009, 48, 6394–6396.
[2] C. Corre, G. Challis, Chem. Biol. 2007, 14, 7.
[3] L. Song, F. Barona-Gomez, C. Corre, J. Am. Chem. Soc. 2006, 128,
14754–14755.
[4] H. B. Bode, Curr. Opin. Chem. Biol. 2009, 13, 224–230.
[5] H. Goodrich-Blair, Curr. Opin. Microbiol. 2007, 10, 225–230.
[6] E. E. Herbert, H. Goodrich-Blair, Nat. Rev. Microbiol. 2007, 5, 634–
646.
[7] D. Ji, Y. Yi, G. H. Kang, Y. H. Choi, P. Kim, N. I. Baek, Y. Kim,
FEMS Microbiol Lett. 2004, 239, 241–248.
[8] J. Crawford, C. Portmann, X. Zhang, Proc. Natl. Acad. Sci. USA
2012, 109, 10821–10826.
[9] B. McInerney, W. Taylor, M. Lacey, R. Akhurst, R. Gregson, J. Nat.
Prod. 1991, 54, 785–795.
[10] T. Hjelmgaard, M. Givskov, J. Nielsen, Org. Biomol. Chem. 2007, 5,
344–348.
NMR spectroscopy: NMR spectra were recorded on a Bruker AVANCE
600 spectrometer operating at a 1H frequency of 600.2 MHz and a 13C
frequency of 150.9 MHz. The instrument was equipped with a 5 mm TCI
probe head. The final experiments were carried out with a sample of
1 (10 mg) dissolved in [D6]benzene (600 mL) at 293 K. For structure eluci-
16778
www.chemeurj.org
ꢃ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2013, 19, 16772 – 16779

Structure and Biosynthesis of Xenoamicins
FULL PAPER
[11] A. O. Brachmann, S. Forst, G. M. Furgani, A. Fodor, H. B. Bode, J.
Nat. Prod. 2006, 69, 1830–1832.
[12] J. Li, G. Chen, J. M. Webster, Can. J. Microbiol. 1997, 43, 770–773.
[13] S. W. Fuchs, A. Proschak, T. W. Jaskolla, M. Karas, H. B. Bode, Org.
Biomol. Chem. 2011, 9, 3130–3132.
[14] H. B. Bode, D. Reimer, S. W. Fuchs, F. Kirchner, C. Dauth, C.
Kegler, W. Lorenzen, A. O. Brachmann, P. Grꢅn, Chem. Eur. J.
2012, 18, 2342–2348.
[15] G. Lang, T. Kalvelage, A. Peters, J. Wiese, J. F. Imhoff, J. Nat. Prod.
2008, 71, 1074–1077.
[16] Q. Zhou, A. Dowling, H. Heide, J. Wçhnert, U. Brandt, J. Baum, R.
ffrench-Constant, H. B. Bode, J. Nat. Prod. 2012, 75, 1717–1722.
[17] F. Grundmann, V. Dill, A. Dowling, A. Thanwisai, E. Bode, N.
Chantratita, R. ffrench-Constant, H. B. Bode, Beilstein J. Org.
Chem. 2012, 8, 749–752.
[18] S. W. Fuchs, C. C. Sachs, C. Kegler, F. I. Nollmann, M. Karas, H. B.
Bode, Anal. Chem. 2012, 84, 6948–6955.
[19] C. R. Mitchell, Y. Bao, N. J. Benz, S. Zhang, J. Chromatogr. B 2009,
877, 4133–4139.
[26] K. Blin, M. H. Medema, D. Kazempour, M. A. Fischbach, R. Brei-
tling, E. Takano, T. Weber, Nucleic Acids Res. 2013, 41, W204–
W212.
[27] S. F. Altschul, T. L. Madden, A. A. Schꢁffer, J. Zhang, Z. Zhang, W.
Miller, D. J. Lipman, Nucleic Acids Res. 1997, 25, 3389–3402.
[28] T. Weber, K. J. Laiple, E. K. Pross, A. Textor, S. Grond, K. Welzel,
S. Pelzer, A. Vente, W. Wohlleben, Chem. Biol. 2008, 15, 175–188.
[29] M. Rottig, M. H. Medema, K. Blin, T. Weber, C. Rausch, O. Kohl-
bacher, Nucleic Acids Res. 2011, 39, W362–W367.
[30] C. Rausch, T. Weber, O. Kohlbacher, W. Wohlleben, D. H. Huson,
Nucleic Acids Res. 2005, 33, 5799–5808.
[31] T. Stachelhaus, H. D. Mootz, M. A. Marahiel, Chem. Biol. 1999, 6,
493–505.
[32] D. Konz, M. A. Marahiel, Chem. Biol. 1999, 6, R39–48.
[33] C. J. Balibar, F. H. Vaillancourt, C. T. Walsh, Chem. Biol. 2005, 12,
1189–1200.
[34] B. M. Burkhart, R. M. Gassman, D. A. Langs, W. A. Pangborn,
W. L. Duax, V. Pletnev, Biopolymers 1999, 51, 129–144.
[35] M. F. Freeman, C. Gurgui, M. J. Helf, B. I. Morinaka, A. R. Uria,
N. J. Oldham, H. Sahl, S. Matsunaga, J. Piel, Science 2012, 338, 387–
390.
[20] M. Schubert, D. Labudde, H. Oschkinat, P. Schmieder, J. Biomol.
NMR 2002, 24, 149–154.
[21] K. Fujii, Y. Ikai, H. Oka, M. Suzuki, K. Harada, Anal. Chem. 1997,
69, 5146–5151.
[22] K. Fujii, T. Shimoya, Y. Ikai, H. Oka, K. Harada, Tetrahedron Lett.
1998, 39, 2579–2582.
[23] I. Vallet-Gely, A. Novikov, L. Augusto, P. Liehl, G. Bolbach, M.
Pechy-Tarr, P. Cosson, C. Keel, M. Caroff, B. Lemaitre, Appl. Envi-
ron. Microbiol. 2010, 76, 910–921.
[24] B. O. Bachmann, J. Ravel, Meth. Enzymol. 2009, 458, 181–217.
[25] M. H. Medema, K. Blin, P. Cimermancic, V. de Jager, P. Zakrzewski,
M. A. Fischbach, T. Weber, E. Takano, R. Breitling, Nucleic Acids
Res. 2011, 39, W339–W346.
[36] D. Velasco, M. M. del Tomas, M. Cartelle, A. Beceiro, A. Perez, F.
Molina, R. Moure, R. Villanueva, G. Bou, J. Antimicrob. Chemother.
2005, 55, 379–382.
[37] F. I. Nollmann, F. I. Nollmann, A. Dowling, M. Kaiser, K. Deck-
mann, S. Grçsch, R. ffrench-Constant, H. B. Bode, H. B. Bode, Beil-
stein J. Org. Chem. 2012, 8, 528–533.
[38] A. O. Brachmann, S. A. Joyce, H. Jenke-Kodama, G. Schwꢁr, D. J.
Clarke, H. B. Bode, ChemBioChem 2007, 8, 1721–1728.
Received: June 27, 2013
Published online: November 7, 2013
Chem. Eur. J. 2013, 19, 16772 – 16779
ꢃ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
16779