Solid-Phase Enrichment and Analysis of Azide-Labeled Natural Products: Fishing Downstream of Biochemical Pathways

Article abstract of DOI:10.1002/chem.201503781

Many methods have been devised over the decades to trace precursors of specific molecules in cellular environments as, for example, in biosynthesis studies. The advent of click chemistry has facilitated the powerful combination of tracing and at the same time sieving the highly complex metabolome for compounds derived from simple or complex starting materials, especially when the click reaction takes place on a solid support. While the principle of solid-phase click reactions has already been successfully applied for selective protein and peptide enrichment, the successful enrichment of much smaller primary and secondary metabolites, showing great structural diversity and undergoing many different biosynthetic steps, has seen only little development. For bacterial secondary metabolism, a far broader tolerance for "clickable" precursors was observed than in ribosomal proteinogenesis, thus making this method a surprisingly valuable tool for the tracking and discovery of compounds within the cellular biochemical network. The implementation of this method has led to the identification of several new compounds from the bacterial genera Photorhabdus and Xenorhabdus, clearly proving its power.

Full text of DOI:10.1002/chem.201503781

DOI: 10.1002/chem.201503781
Full Paper
&
Click Chemistry
Solid-Phase Enrichment and Analysis of Azide-Labeled Natural
Products: Fishing Downstream of Biochemical Pathways
[a]
[a]
[a]
[a, b]
Alexander J. PØrez, Frank Wesche, HØl›ne Adihou, and Helge B. Bode*
Abstract: Many methods have been devised over the de-
cades to trace precursors of specific molecules in cellular en-
vironments as, for example, in biosynthesis studies. The
advent of click chemistry has facilitated the powerful combi-
nation of tracing and at the same time sieving the highly
complex metabolome for compounds derived from simple
or complex starting materials, especially when the click reac-
tion takes place on a solid support. While the principle of
solid-phase click reactions has already been successfully ap-
plied for selective protein and peptide enrichment, the suc-
cessful enrichment of much smaller primary and secondary
metabolites, showing great structural diversity and undergo-
ing many different biosynthetic steps, has seen only little de-
velopment. For bacterial secondary metabolism, a far broad-
er tolerance for “clickable” precursors was observed than in
ribosomal proteinogenesis, thus making this method a sur-
prisingly valuable tool for the tracking and discovery of com-
pounds within the cellular biochemical network. The imple-
mentation of this method has led to the identification of
several new compounds from the bacterial genera Photo-
rhabdus and Xenorhabdus, clearly proving its power.
Introduction
the bioorthogonal ring-strain-promoted cycloaddition as devel-
[
8]
oped by Bertozzi et al., a door has been opened towards
new approaches for the tracing and identification of potential-
Specialized metabolites are small molecules with many differ-
[
1]
[9,10]
ent roles, ranging from nutrient acquisition (siderophores)
and cell-to-cell signaling to weapons of interspecies warfare
ly interesting compounds.
We describe herein an easily synthesized cleavable azide re-
active resin (CARR, 1, Figure 1, top) capable of reacting with
and thus specifically absorbing a broad range of azides derived
from in vivo metabolization and incorporation of previously
fed precursor molecules. Methanolic Amberlite XAD-16 extracts
of azide-supplemented bacterial cultures were incubated with
the resin, and subsequent cleavage of the cycloaddition prod-
ucts under mild conditions resulted in the rapid release of sev-
eral natural products showing the incorporated azides. Subse-
quent HPLC-MS analysis allowed rapid identification of these
natural products, which included several that might otherwise
have been overlooked (Figure 1, bottom).
[
2]
(
antibiotics) and beyond. The structural and biosynthetic
nature of such compounds is just as diverse as the roles they
fulfill, with several metabolites being peptides (ribosomally or
non-ribosomally produced), polyketides, or peptide–polyketide
[
3]
hybrids. While the precise roles of several of these molecules
are still unknown, identification of new compounds and their
subsequent characterization are central to the assignment of
their function and the subsequent development of new thera-
[
4]
peutics.
Over the decades, numerous tools have been developed
and applied in the fast and reliable identification of new me-
[
5]
tabolites. These include MS methods, which may also be cou-
pled with stable isotope feeding experiments to unravel the
[
6]
structures of unknown compounds. With all of these meth- Results and Discussion
ods having their own limitations, it is highly favorable to have
Synthesis of the CARR resin
as many different methods as possible at our disposal. With
[7]
the emergence of easily applicable click chemistry, especially
For the synthesis of the cleavable azide reactive resin (CARR, 1,
Figure 1), a acrylamide–polyethylenglycol resin (PEGA) was
used as the basis since it is compatible with both polar mole-
cules such as peptides as well as small hydrophobic com-
[
a] A. J. PØrez, F. Wesche, Dr. H. Adihou, Prof. Dr. H. B. Bode
Merck Stiftungsprofessur für Molekulare Biotechnologie
Fachbereich Biowissenschaften, Goethe Universität Frankfurt, Max-von-
Laue-Strasse 9, 60438 Frankfurt (Germany)
[
11]
pounds.
Furthermore, functional group accessibility is re-
tained in polar and amphiphilic solvents such as water, metha-
nol, and acetonitrile. The amino moiety of the resin was succi-
nylated and then derivatized with a cystamine linker, allowing
for subsequent reductive cleavage of the disulfide bond under
mild conditions. Finally, a reactive carbonate ester of bicyclono-
[
b] Prof. Dr. H. B. Bode
Buchmann Institute for Molecular Life Sciences (BMLS)
Goethe Universität Frankfurt, Max-von-Laue-Strasse 15, 60438 Frankfurt
(
Germany)
Supporting information and ORCID(s) from the author(s) for this article is
available on the WWW under http://dx.doi.org/10.1002/chem.201503781
[
12]
nyne (BCN, 2, Scheme S1) was used to attach the alkyne to
Chem. Eur. J. 2016, 22, 639 – 645
639
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
[
14]
possess a complex lipidome. As expected, after feeding of
5-azidopentadecanoic acid (3i) over three days, b-oxidation
1
products could be directly detected in the base-peak chroma-
togram of the sample, and furthermore elongation products
up to 19-azidononadecanoic acid (3l) were also observed (Fig-
ures S1 and S2). The presence of this unusual FA was sugges-
tive of the activity of a fatty acid elongase in M. xanthus.
To further pursue this, azido fatty acids (AFAs) of varying
chain length were fed to strains that are known for a rich pro-
duction of secondary metabolites, such as the genera Xenor-
habdus and Photorhabdus. For example, Xenorhabdus doucetiae
is known to produce simple amides formed from phenethyl-
[
15]
amine (PEA) and a variety of FAs, indicating a great substrate
tolerance in terms of the fatty acyl residue. Thus, a feeding ex-
periment with three different AFAs (3b, 3c, and 3d) was con-
ducted, which yielded several phenethylamides (Figures 2 and
S3). Interestingly, not only was C -AFA (3d) degraded in steps
10
corresponding to C H to yield 5 f, 5d, and 5b, but the shorter
2
4
AFAs were also elongated, including unsaturated variants 7a,
b, and 7c with chain lengths of 10, 12, and 14 carbon atoms
7
in the case of feeding of 3c (Figure S4). In spite of the already
very short chain lengths, b-oxidation products of 3b and 3c
were also observed, leading to 5a and 5b, respectively. This
was confirmed by cultivation of X. doucetiae with C -AFA (3b)
6
13
in C-medium, which led to the incorporation of 8 carbon
1
3
Figure 1. Principle of azide enrichment and product identification. After an
azide (R-N ) is covalently bound to the resin through azide–alkyne cycloaddi-
atoms for the direct amidation product 5d and of 8+2m
C
3
atoms (two additional carbon atoms for each elongation cycle
m) for AFA elongation products 5 f (m=1), 5h (m=2), and 5j
(m=3), where m is the number of elongation cycles performed
tion, the compound is cleaved from the resin and analyzed by HPLC-MS.
The characteristic fragmentation pattern of the derivatized azide allows the
differentiation of true cycloaddition products from possible artifacts. The
representative base-peak chromatograms (grey) at the bottom demonstrate
the production of a large number of metabolites (*) from a single precursor
on the AFA prior to its incorporation (Figure S3). Additionally,
13
elongation products were also observed with 10+2m
C
(
+). The strength of this method to selectively enrich azide-bearing com-
atoms, requiring four elongation cycles (m=4, Figure S5). This
might be explained by an initial degradation step of 3c to 3a
(C -AFA to C -AFA, m=1) prior to elongation and final amida-
pounds that would otherwise be easily overlooked is made clear by the fact
that almost each peak in the base-peak chromatogram (BPC, grey) corre-
sponds to the extracted ion chromatogram (EIC, black, dotted lines) of a me-
tabolite of the precursor after enrichment, whereas most of the metabolites
were almost undetectable in the crude extract. A detailed description of
these particular chromatograms can be found in the Supporting Informa-
tion.
6
4
tion, resulting in the same final length of the observed amides
(
m=4À1=3). FA biosynthesis and thus elongation usually re-
quires ligation of the starter acyl chain to the phosphopante-
theinyl moiety of an acyl carrier protein (ACP), whereas degra-
[
16]
dation is performed with the FA activated as a CoA-thioester.
13
the resin (Scheme S2). The BCN was used as an endo/exo mix-
ture, as it has previously been shown that the chemical proper-
ties and retention times of each stereoisomer are almost indis-
The finding of 10+2m C atoms in elongated amides thus
shows that a transesterification of the CoA-thioester of 3a to
ACP must have occurred via a FabD-like protein. Since iso-
branched fatty acids, such as those produced by X. doucetiae
and M. xanthus, require special starting units such as isovaleryl-
[
13]
tinguishable. The resin was incubated with XAD extracts of
cultures that had been grown for three days after feeding each
azide compound at a final concentration of 5 mm and the
resin was then extensively washed with dichloromethane and
methanol. For disulfide bond cleavage, a solution of tris(2-car-
boxyethyl)phosphine (TCEP) in a 1:5:10 mixture of phosphate
buffered saline (PBS)/chloroform/methanol at pH 7 was used
as a reducing agent, affording high reactivity and a broad solu-
bility margin for both peptides and lipids.
[
17]
CoA, short-chain AFAs might also be activated as CoA-thio-
esters before entering the elongation cycle, ultimately yielding
respective ACP-bound intermediates that might be used in the
biosynthesis of phenethylamides.
When P. luminescens was fed with a long-chain AFA 3j, sev-
eral b-oxidation products were detected, along with a signal
[
18]
that could be assigned to an azido phurealipid derivative
10c, Figures 2 and S6). Upon feeding the shorter AFA 3i, simi-
(
lar signals for 9b and 9d could be observed, whereas feeding
of the short AFA 3c resulted in the formation of the elongated
FA 4h along with phurealipids 9a and 9c (Table S1). N-Methyl-
ation of these lipids was confirmed by the MS fragmentation
pattern, which showed neutral losses of methylamine and
CARR enrichment of azido fatty acids and natural products
derived therefrom
The resin was first tested with a fatty acid (FA) extract from
a culture of Myxococcus xanthus, a bacterium that is known to
Chem. Eur. J. 2016, 22, 639 – 645
www.chemeurj.org
640
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
1
2
Figure 2. Metabolization of AFAs (framed) to various compounds (R ) and their derivatized analogues (R ) as found in the bacterial strains Xenorhabdus douce-
tiae (amides), Photorhabdus luminescens (phurealipids), and Myxococcus xanthus (long fatty acids), showing important degradation (FAD) and elongation (FAB)
2 4
cycles (m=number of C H addition or removal cycles) as well as the natural products they are eventually incorporated into. The dashed arrow indicates the
case in which degradation was proven to occur prior to elongation (Figure S5). Cycloaddition products that were verified by HR-MS (Table S1) are highlighted
in grey in the table to the right. A detailed account of each feeding experiment can be found in the Supporting Information, Figures S1, S3, and S6.
1
3
methyl isocyanate (Figure S7). These results show that the
chain length tolerance for the incorporation of AFAs is narrow-
er than that for phenethylamides, which correlates well with
grown in C-medium led to a mass shift of 16 relative to the
+
mass of the natural compound of m/z 420 [M+H] , as expect-
13
ed for the incorporation of 16 C atoms into 13c. Addition of
PAA to this culture led to the incorporation of eight C atoms,
[
18]
12
previous findings
alkyl chains in phurealipids. Since in the present work 10a and
0d were only detected in trace amounts, it can be concluded
that the preferred chain length for AFAs in this case is C or C10
showing the incorporation of C9 to C14
whereas the addition of PEA led to the incorporation of 8 and
12
1
16 C atoms, showing that PEA can be easily converted into
PAA, but not vice versa (Figure S9). Thus, it can be concluded
that azido-PEA is not a suitable substrate for the unknown
amide synthase, but is readily converted into azido-PAA
through transamination and oxidation (Figure 3).
9
(
9b and 9c), which is in accordance with the main natural
products if the azide group is regarded as being equivalent in
size to an ethyl group rather than a propyl group due to its
greater rigidity.
Azido-PEA was also fed to cultures of Xenorhabdus nemato-
phila, which is known to incorporate PEA into rhabdopepti-
[
20]
[21]
des and xenortides. While it quickly became evident that
azido-PEA was not incorporated into any rhabdopeptide, the
expected adduct mass for the dimethylated tripeptide xenor-
tide A (14a) was indeed found and confirmed by HR ESI-MS
and its fragmentation pattern (Figures 3, S8B, and S10). Azido-
PAA was again found as a by-product of PEA metabolism, ap-
parently without further incorporation into other compounds.
Two additional signals could be linked to a derivative of nema-
tophin (15b, Figure 3), a branched a-ketoisocaproyl trypta-
mide and a homologous a-ketoisovaleryl derivative thereof
(16b, Figures 3, S8B, and S11). Although no PEA derivatives of
nematophin are known, the presence of the azido function
might allow the incorporation of this slightly larger PEA ana-
CARR enrichment of p-azidophenethylamine and natural
products derived therefrom
Following the good results achieved with phenethylamides
from X. doucetiae, p-azidophenethylamine (azido-PEA, 11 a) was
[
19]
synthesized and fed to the same strain. We anticipated the
generation of similar products but labeled on the other end of
the linear molecules. Instead, however, large amounts of azido-
phenylacetic acid (12a) were found, along with the corre-
sponding phenethylamide 13a (Figure S8A). To further elabo-
rate the biochemical pathways involved in the conversion of
PEA to phenylacetic acid (PAA, 12c), an isotope labeling ex-
periment was conducted (Figure S9). Extracts of cultures
[
22]
Chem. Eur. J. 2016, 22, 639 – 645
www.chemeurj.org
641
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
iscule amounts), followed by decarboxylation and eventually
oxidation of a putative aldehyde (Figure S13A). Since the for-
mation of 11 a was never observed, we deem it unlikely that
an alternative pathway involving transformation of 18a to
11 a, transamination, and subsequent oxidation took place. In-
terestingly, most enzymes showed a good substrate tolerance
of the azido group. However, consecutive degradations to
acetyl-CoA and succinyl-CoA derivatives, as have been shown
to follow in the catabolic pathways of PAA in many bacteria,
[
25]
have not been detected. In order to differentiate between
the two possible pathways of azido-Phe degradation, Escheri-
chia coli was incubated with azido-Phe and the culture broth
was analyzed for azido-phenylpyruvic acid and azidophenyla-
cetaldehyde. While azido-phenylpyruvic acid could be found
and identified in notable amounts within 24 h of feeding with
azido-Phe at 5 mm, the aldehyde was not detected in this ex-
periment (Figure S13B). When 18a was fed to X. doucetiae, the
formation of phenethylamide 13a could be observed (Fig-
ure S13C), just as after the feeding of 11 a previously, suggest-
ing that 18a was first converted to 12a via 19a, as has been
shown to occur in other strains. When azido-Phe was fed to
Xenorhabdus szentirmaii, many azido derivatives of the phenyl-
[
25]
alanine catabolism described above were found (19a, 12a;
Figure S14). Additionally, a new signal was identified that could
be attributed to azidobenzaldehyde (20a, detected as 20b in
Figure S15). The immediate precursor of 20a may have been
azidophenylpyruvic acid (19a), as has been described for Lac-
Figure 3. Metabolic pathways as observed in feeding experiments in various
Xenorhabdus strains and one Photorhabdus strain. The compounds that were
fed (11 a, 16a, and 24a) are framed. Conversions seen in all accordingly fed
strains are shown with thick black arrows, while other arrow types show spe-
cific conversions occurring only in one particular strain: P. luminescens (grey),
X. nematophila (thin dotted arrows), X. doucetiae (thin arrow), X. szentirmaii
[
26]
(
thick dashed arrows). X. indica only took part in general pathways and is
tobacillus plantarum. Eluting at a similar retention time as
12b, the oxidized form of aldehyde 20a, azido-benzoic acid
thus not shown explicitly. A detailed account of each feeding experiment
can be found in the Supporting Information, Figures S8, S13, S14, and S20.
(
21a), could also be detected (Figures 3 and S14). Although
1
2
For R (azide) and R (derivatized azide), see Figure 1.
produced in only small amounts after the regular three days of
incubation, the amounts of 21a were significantly higher after
five days (not shown), consistent with its direct formation from
20a. While it is known that Streptomyces maritimus produces
benzoyl-CoA from phenylalanine via cinnamic acid, it is un-
likely that this pathway takes place in X. szentirmaii, since cin-
namic acid was not detected, suggesting a direct oxidation of
the abundant aldehyde 20a instead (Figure 3). Considering the
versatility of phenylalanine metabolism in X. szentirmaii, azido-
benzoic acid (21a) was fed to the strain in a separate experi-
ment to search for possible downstream metabolites by the
CARR incubation method. A single product was observed,
which could be assigned as the ethyl ester of the said acid
logue. After feeding of azido-PEA to X. doucetiae, a compound
with a retention time and fragmentation pattern very similar
to those of 15b was found. However, careful HPLC-MS analysis
and HR-MS allowed identification of the compound as heptan-
oylated 11 b (17b, Figures S8A and S12A), with the alkyl chain
assumed to be iso-branched as is the case for the natural tryp-
[
27]
[15]
tamide in X. doucetiae. The respective natural products were
observed as also having the same retention time (Figure S12B),
making HR-MS important for correct assignment.
(
22b, Figures 3 and S14). This was confirmed by chemical syn-
CARR enrichment of p-azidophenylalanine and natural prod-
ucts derived therefrom
thesis followed by CARR enrichment (Figure S16A) and by its
fragmentation pattern (Figure S16B). The formation of a benzo-
ic acid ester is interesting insofar as the production of such
Next, the tolerance for the incorporation of azido-phenylala-
nine (azido-Phe, 18a) was investigated. A certain degree of
ambiguity/promiscuity in the incorporation of aromatic amino
[
28]
compounds is primarily a speciality of plants.
At retention times of 9.8–11.1 min, four additional new prod-
ucts could be seen (23b–26b, Figures 3, S14, and S15) after
the addition of azido-Phe (18a). Fragmentation patterns
showed the formation of positively charged 18a from all of
these compounds in their underivatized forms, and the sum
formulas obtained from HR-MS data (Table S1) indicated that
the first two derivatized products (23b and 24b) were N-acyla-
tion products of 18a bearing octanoyl (23b) and isononanoyl
(24b) side chains, as was confirmed by growth experiments in
[
23]
acids is known for compounds such as the taxlllaid peptides
[24]
and GameXPeptides. Taxlllaid-producing Xenorhabdus indica
was thus incubated with commercially available 18a at 5 mm
for three days, after which a single product was detected that
was identified as azido-PAA (12a), not showing the expected
incorporation into peptide natural products. However, addi-
tional metabolic steps must have occurred: transamination of
azido-Phe to azidophenylpyruvic acid (19a, found only in min-
Chem. Eur. J. 2016, 22, 639 – 645
www.chemeurj.org
642
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
1
3
C-medium (Figure S17B). The incorporation of an iso-
branched FA was confirmed through the feeding of
D ]leucine, which led to the expected mass shifts for 24a and
Conclusion
[
The combination of feeding of azide-labeled precursors with
CARR can be regarded as a special form of precursor-directed
9
2
6a in the crude extract, whereas feeding of [D ]methionine
3
[
33]
did not lead to a mass shift (Figure S17A). This ruled out the
possibility of S-adenosylmethionine-derived N- or C-methyla-
tion for these compounds. To confirm the general structure of
these compounds, 23a was synthesized on a solid support, re-
acted with CARR, and analyzed under the same conditions as
the natural products. This led to an identical retention time
biosynthesis, akin to the established procedure conducted
[
34]
with alkynylated precursors. Its advantage is the enrichment
step, which allows the detection of even minute amounts of
[
35a]
labeled compounds from complex mixtures.
Thus, it pro-
vides an easy means of investigating the fate of small mole-
cules in complex mixtures, but has not hitherto been applied
in the context of secondary metabolites with successful enrich-
2
and MS fragmentation pattern as seen for 23b (Figure S17C).
[
35]
The mass increase of 16 suggested the incorporation of an ad-
ditional oxygen atom in 25b and 26b. An in-depth investiga-
tion of the fragmentation patterns of compounds 23b and
ment. The use of a bioorthogonal label offers high selectivity
in the process, as opposed to enrichment methods that target
biologically ubiquitous functional groups such as amines and
[
36]
[10,13,37]
2
5b revealed that the mass shifts were caused by a heavier
thiols. As was previously shown for FA analysis,
derivat-
acyl chain (Figure S18), suggesting the presence of b-hydroxy
acyl or other oxygen-bearing moieties in these two structures;
Figure 3 shows one possible structure for compounds 25a and
ized azides show an excellent propensity for protonation. This
makes LC/MS analysis especially simple, since otherwise neutral
or negatively charged compounds become readily discernible
within a single analytical run (Figure S14). Our results also
show that azido groups introduced into amines or fatty acids
readily partake in primary and secondary metabolism, allowing
for an easy recovery of downstream products and their subse-
quent analysis. Similarly, the described approach may be ex-
panded to other functional groups, allowing the detection and
2
6a. At a retention time of 10.6 min, a small peak attributable
[29]
to the expected szentiamide derivative (27a, Figure 3) was
also detected, and the compound was identified through HR-
MS (Table S1), its fragmentation pattern (Figure S19), and
chemical synthesis (Figure S15). As could be deduced from the
m/z ratio of 27b and its fragmentation pattern, 18a was incor-
porated instead of tyrosine, indicating that this is favorable
when the natural amino acid to be substituted is larger than
phenylalanine, as was the case with nematophin (11 a instead
of tryptamine), due to the steric demand of the azido group.
When 18a was fed to P. luminescens, a small amount of PAA
could be detected, along with a large amount of p-azidocin-
namic acid (28b, Figures 3, S20, and S21A). P. luminescens is
known to produce cinnamic acid for incorporation into isopro-
[
34,38]
enrichment of complementary functions.
An additional
benefit of the described procedure might be possible scale-up,
thus allowing the collection of sufficient amounts for a prepara-
tive purification of as yet unidentified compounds. In this
work, it has been shown that even small-scale analytical proce-
dures such as liquid chromatography, high-resolution mass
2
spectrometry, and MS fragmentation are adequate for eluci-
dating the metabolic fates of pertinent compounds.
[
30]
pylstilbene (IPS) and possibly other secondary metabolites.
However, no azido-labeled IPS was detected, and as previous
experiments have shown that p-chlorocinnamic acid is not in-
corporated into this compound class, whereas m-chlorocin-
Material and methods
General experimental procedures
[
31]
namic acid is, it appears likely that no p-substituted cinnamic
Solvents and reagents were obtained from Sigma-Aldrich (Mün-
chen, Germany). Silica gel based chromatographic purification was
performed on a Biotage SP1ꢁ flash purification system (Biotage,
Uppsala, Sweden), using 40+M KP-Sil cartridges (Biotage, Uppsala,
Sweden) in combination with a UV detector. H and C NMR spec-
tra of the synthetic products were recorded on Bruker AV500
acid derivatives can be used as substrates for the required CoA
[30]
activation by the CoA ligase StlB. The second largest peak
showed a fragmentation pattern similar to those of acylated
phenylalanine structures found in X. szentirmaii. HR-MS and
the fragmentation pattern (29b, Table S1, Figure S21B) re-
vealed that this compound was indeed azido-phenylalanine
acylated with cinnamic acid, as was further confirmed by solid-
phase peptide synthesis of 29a and by repetition of the ex-
periment with a DstlB mutant lacking the CoA-ligase needed
1
13
(
500 MHz), AV400 (400 MHz), or AV300 (300 MHz) spectrometers
1
using CDCl or [D ]DMSO as solvent and internal standard. H NMR:
3
6
13
CHCl : d=7.27 ppm or [D ]DMSO: d=2.50 ppm; C NMR: CDCl3:
3
5
d=77.00 ppm or [D ]DMSO: d=39.51 ppm. ESI HPLC MS analysis
6
was performed with a Dionex UltiMate 3000 system coupled to
a Bruker AmaZon X mass spectrometer using an MeCN/0.1%
[30]
for activation of cinnamic acid (Figure S21A). Finally, a small
signal was observed at a longer retention time and assigned
formic acid in H O gradient ranging from 5 to 95% in 16 min at
2
[
6,24]
À1
to GameXPeptide A
(30b), with 18a being incorporated in-
a flow rate of 0.6 mLmin . High-resolution mass spectra were ob-
stead of phenylalanine. The structure was confirmed by HR-MS
data (Table S1), the fragmentation pattern (Figure S21C), and
repetition of the feeding experiment with a gxpS-overexpress-
tained on a Dionex Ultimate 3000 RSLC coupled to a Bruker micro-
TOF-Q II equipped with an ESI source set to positive ionization
mode. HR-HPLC settings were identical to those described above,
although the column bed was 100 mm in length and the gradient
length was set to 20 min. The mass spectrometer was calibrated
using a sodium formate calibrant solution (10 mm). MS data were
[32]
ing E. coli strain as described previously, whereby 30b was
detected as a major product (Figure S21A).
[12a]
acquired within the mass range m/z 100–1500.
Alternatively, in
some cases, high-resolution measurements were carried out using
a MALDI LTQ Orbitrap XL (Thermo Fisher Scientific, Inc., Waltham,
Chem. Eur. J. 2016, 22, 639 – 645
www.chemeurj.org
643
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
MA), equipped with a laser at 337 nm and a resolution of 100000.
A 4-chloro-a-cyanocinnamic acid matrix was used, and the sum
formulas and according masses were internally calibrated using
[2] J. Clardy, M. A. Fischbach, C. T. Walsh, Nat. Biotechnol. 2006, 24, 1541–
1
550.
3] S. C. Wenzel, R. Müller, Curr. Opin. Chem. Biol. 2005, 9, 447–458.
[4] D. J. Newman, G. M. Cragg, J. Nat. Prod. 2007, 70, 461–477.
[
1
+
des-Arg -Bradykinin (m/z: 904.46811224 [M+H] ) from Calibration
[5] R. D. Kersten, Y. Yang, Y. Xu, P. Cimermancic, S. Nam, W. Fenical, M. A.
Mixture 1 (SCIEX, Peptide Mass Standards Kit for Calibration of AB
Fischbach, B. S. Moore, P. C. Dorrestein, Nat. Chem. Biol. 2011, 7, 794–
[6b]
SCIEX MALDI-TOF Instruments) as a standard.
802.
[
6] a) H. B. Bode, D. Reimer, S. W. Fuchs, F. Kirchner, C. Dauth, C. Kegler, W.
Lorenzen, A. O. Brachmann, P. Grün, Chem. Eur. J. 2012, 18, 2342–2348;
b) S. W. Fuchs, C. C. Sachs, C. Kegler, F. I. Nollmann, M. Karas, H. B. Bode,
Anal. Chem. 2012, 84, 6948–6955.
Strains and cultivation conditions
All strains used in this work with the exception of M. xanthus were
grown on solid and liquid Luria–Bertani (LB, pH 7.0) medium at
[
[
[
7] J. E. Moses, A. D. Moorhouse, Chem. Soc. Rev. 2007, 36, 1249–1262.
8] J. C. Jewett, C. R. Bertozzi, Chem. Soc. Rev. 2010, 39, 1272–1279.
9] X. Zhu, W. Zhang, Front. Chem. 2015, 3, 1–5.
3
08C and 180 rpm on a rotary shaker. The Photorhabdus DstlB
mutant was prepared according to the procedure described by
[
[
10] M. Broncel, R. A. Serwa, P. Ciepla, E. Krause, M. J. Dallman, A. I. Magee,
[28]
Joyce et al.
and needed no selection marker. The E. coli
E. W. Tate, Angew. Chem. Int. Ed. 2015, 54, 5948–5951; Angew. Chem.
DH10B::mtaA gxpS overexpression mutant was grown with the ad-
2015, 127, 6046–6049.
À1
[31]
dition of 100 mgL ampicilin. M. xanthus was grown on solid
11] M. Meldal, F. Auzanneau, O. Hindsgaul, M. M. Palcic, J. Chem. Soc.,
Chem. Commun. 1994, 1849–1850.
[12] J. Dommerholt, S. Schmidt, R. Temming, L. J. A. Hendriks, F. P. J. T. Rutjes,
J. C. M. van Hest, D. J. Lefeber, P. Friedl, F. L. van Delft, Angew. Chem. Int.
Ed. 2010, 49, 9422–9425; Angew. Chem. 2010, 122, 9612–9615.
13] A. J. PØrez, H. B. Bode, ChemBioChem 2015, 16, 1588–1591.
14] a) W. Lorenzen, K. A. Bozhüyük, N. S. Cortina, H. B. Bode, J. Lipid Res. J.
Lipid. Res. 2014, 55, 2620–2633; b) T. Ahrendt, H. Wolff, H. B. Bode, App.
Environ. Microbiol. 2015, 81, 6538–6547.
15] A. Proschak, K. Schulz, J. Herrmann, A. J. Dowling, A. O. Brachmann, R.
ffrench-Constant, R. Müller, H. B. Bode, ChemBioChem 2011, 12, 2011–
2015.
[16] J. B. Parsons, C. O. Rock, Prog. Lipid Res. 2013, 52, 249–276.
17] J. S. Dickschat, H. B. Bode, R. M. Kroppenstedt, R. Müller, S. Schulz, Org.
Biomol. Chem. 2005, 3, 2824–2831.
18] F. I. Nollmann, A. K. Heinrich, A. O. Brachmann, C. Morisseau, K. Mukher-
jee, A. M. Casanova-Torres, F. Strobl, D. Kleinhans, S. Kinski, K. Schultz,
M. L. Beeton, M. Kaiser, Y. Chu, L. Phan Ke, A. Thanwisai, K. A. J. Boz-
hüyük, N. Chantratita, F. Gçtz, N. R. Waterfield, A. Vilcinskas, E. H. K.
Stelzer, H. Goodrich-Blair, B. D. Hammock, H. B. Bode, ChemBioChem
[39]
and liquid CTT medium at 308C and 180 rpm on a rotary shaker.
The strains used were Myxococcus xanthus DK1622, Escherichia coli
DH10B, Xenorhabdus doucetiae DSM17909, Xenorhabdus nemato-
phila HGB081, Xenorhabdus indica DSM17382, Xenorhabdus szentir-
maii DSM16338, and Photorhabdus luminescens TTO1.
[
[
Preparation of XAD extracts
[
Aliquots (10 mL) of the corresponding cultures were inoculated at
an optical density (OD600) of 0.5 from overnight pre-cultures and
fed with the respective compounds at a final concentration of
[
5
mm. Amberlite XAD-16 beads (2 vol.%) were added and the cul-
tures were cultivated at 308C on a rotary shaker at 180 rpm for
2 h unless stated otherwise. The supernatant was then removed
[
7
and the XAD was eluted with MeOH (10 mL) for 1 h. The solution
was filtered, the solvent was evaporated, and the residue was re-
dissolved in MeOH (1 mL).
2015, 16, 766–771.
[
19] a) N. E. Hausler, S. M. Devine, F. M. McRobb, L. Warfe, C. W. Pouton, J. M.
Azide enrichment procedure
Haynes, S. E. Bottle, P. J. White, P. J. Scammells, J. Med. Chem. 2012, 55,
3
3
521–3534; b) E. D. Goddard-Borger, R. V. Stick, Org. Lett. 2007, 9,
797–3800.
DMF-wetted CARR (2; 3 mg) was placed inside a syringe for solid-
phase peptide synthesis and thoroughly washed with MeOH and
CH Cl . The respective methanolic XAD extract (500 mL) was then
[
[
20] D. Reimer, K. N. Cowles, A. Proschak, F. I. Nollmann, A. J. Dowling, M.
2
2
Kaiser, R. ffrench-Constant, H. Goodrich-Blair, H. B. Bode, ChemBioChem
2
21] D. Reimer, F. I. Nollmann, K. Schultz, M. Kaiser, H. B. Bode, J. Nat. Prod.
2014, 77, 1976–1980.
added. The suspension was heated to 558C for 1 h and then stirred
overnight. Excess extract was then removed, and the resin was ex-
tensively washed with MeOH and CH Cl and then incubated with
013, 14, 1991–1997.
2
2
1
mL of a phosphate-buffered (pH 7) 5 mm solution of tris(2-car-
[22] J. Li, G. Chen, J. M. Webster, Can. J. Microbiol. 1997, 43, 770–773.
[23] M. Kronenwerth, K. A. J. Bozhüzük, A. S. Kahnt, D. Steinhilber, S. Gau-
driault, M. Kaiser, H. B. Bode, Chem. Eur. J. 2014, 20, 17478–17487.
boxyethyl)phosphine (TCEP) in PBS/CHCl /MeOH (1:5:10) for 1 h.
3
The resulting solution was collected and concentrated to dryness.
The cleavage products were redissolved in MeCN (200 mL) and ana-
lyzed directly by HPLC/MS.
[
[
[
[
24] F. I. Nollmann, C. Dauth, G. Mulley, C. Kegler, M. Kaiser, N. R. Waterfield,
H. B. Bode, ChemBioChem 2015, 16, 205–208.
25] R. Teufel, V. Mascaraque, W. Ismail, J. M. Voss, J. Perera, W. Eisenreich, W.
Haehnel, G. Fuchs, PNAS 2010, 107, 14390–14395.
26] M. N. Nierop Groot, J. A. M. de Bont, Appl. Environ. Microbiol. 1998, 64,
Acknowledgement
3009–3013.
27] B. S. Moore, C. Hertweck, J. N. Hopke, M. Izumikawa, J. A. Kalaitzis, G.
Nilsen, T. O’Hare, J. Piel, P. R. Shipley, L. Xiang, M. B. Austin, J. P. Noel, J.
Nat. Prod. 2002, 65, 1956–1962.
Work in the Bode lab was supported by a European Research
Council starting grant under grant agreement number 311477.
Special thanks are expressed towards our former interns
Martin Brinek and Gina Grammbitter for helping with the resin
synthesis and its first evaluation.
[
28] a) J. R. Widhalm, N. Dudareva, Mol. Plant 2015, 8, 83–97; b) N. Dudare-
va, L. M. Murfitt, C. J. Mann, N. Gorenstein, N. Kolosova, C. M. Kish, C.
Bonham, K. Wood, Plant Cell 2000, 12, 949–961.
[
[
29] B. Ohlendorf, S. Simon, J. Wiese, J. F. Imhoff, Nat. Prod. Commun. 2011,
6, 1247–1250.
30] S. A. Joyce, A. O. Brachmann, I. Glazer, L. Lango, G. Schwär, D. J. Clarke,
H. B. Bode, Angew. Chem. Int. Ed. 2008, 47, 1942–1945; Angew. Chem.
Keywords: azides · click chemistry · fatty acids · metabolic
labeling · natural products
2
008, 120, 1968–1971.
31] M. Kronenwert, A. O. Brachmann, M. Kaiser, H. B. Bode, ChemBioChem
014, 15, 2689–2691.
[
2
[
1] R. Saha, N. Saha, R. S. Donofrio, L. L. Bestervelt, J. Basic Microbiol. 2013,
3, 303–317.
[32] O. Schimming, F. Fleischhacker, F. I. Nollmann, H. B. Bode, ChemBioChem
2014, 15, 1290–1294.
5
Chem. Eur. J. 2016, 22, 639 – 645
www.chemeurj.org
644
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
[
[
[
33] H. Sun, Z. Liu, H. Zhao, E. L. Ang, Drug. Des. Devel. Ther. 2015, 9, 823–
33.
34] C. J. B. Harvey, J. D. Puglisi, V. S. Pande, D. E. Cane, C. Khosla, J. Am.
Chem. Soc. 2012, 134, 12259–12265.
35] a) M. A. Nessen, G. Kramer, J. Back, J. M. Baskin, L. E. J. Smeenk, L. J. de
Koning, J. H. van Maarseveen, L. de Jong, C. R. Bertozzi, H. Hiemstra,
C. G. de Koster, J. Proteome Res. 2009, 8, 3702–3711; b) R. P. Temming,
M. van Scherpenzeel, E. te Brinke, S. Schoffelen, J. Gloerich, D. J. Lefeber,
F. L. van Delft, Bioorg. Med. Chem. 2012, 20, 655–661.
[36] E. E. Carlson, B. F. Cravatt, Nature Methods 2007, 4, 429–435.
[37] A. J. PØrez, H. B. Bode, J. Lipid Res. 2014, 55, 1897–1901.
[38] A. Deb Roy, S. Grüschow, N. Cairns, R. J. Goss, J. Am. Chem. Soc. 2010,
132, 12243–12245.
8
[39] J. Hodgkin, D. Kaiser, Proc. Natl. Acad. Sci. USA 1977, 74, 2938–2942.
Received: September 21, 2015
Published online on December 2, 2015
Chem. Eur. J. 2016, 22, 639 – 645
www.chemeurj.org
645
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim