Structure and biosynthetic assembly of cupriachelin, a photoreactive siderophore from the bioplastic producer cupriavidus necator H16

Article abstract of DOI:10.1021/ja300620z

The bacterium Cupriavidus necator H16 produces a family of linear lipopeptides when grown under low iron conditions. The structural composition of these molecules, exemplified by the main metabolite cupriachelin, is reminiscent of siderophores that are excreted by marine bacteria. Comparable to marine siderophores, the ferric form of cupriachelin exhibits photoreactive properties. Exposure to UV light induces an oxidation of its peptidic backbone and a concomitant reduction of the coordinated Fe(III). Here, we report the genomics-inspired isolation and structural characterization of cupriachelin as well as its encoding gene cluster, which was identified by insertional mutagenesis. Based upon the functional characterization of adenylation domain specificity, a model for cupriachelin biosynthesis is proposed.

Full text of DOI:10.1021/ja300620z

Article
pubs.acs.org/JACS
Structure and Biosynthetic Assembly of Cupriachelin, a
Photoreactive Siderophore from the Bioplastic Producer Cupriavidus
necator H16
Martin F. Kreutzer, Hirokazu Kage, and Markus Nett*
Junior Research Group “Secondary Metabolism of Predatory Bacteria”, Leibniz Institute for Natural Product Research and Infection
Biology e.V., Hans-Knoll-Institute, Beutenbergstrasse 11a, 07745 Jena, Germany
̈
S
* Supporting Information
ABSTRACT: The bacterium Cupriavidus necator H16 pro-
duces a family of linear lipopeptides when grown under low
iron conditions. The structural composition of these
molecules, exemplified by the main metabolite cupriachelin,
is reminiscent of siderophores that are excreted by marine
bacteria. Comparable to marine siderophores, the ferric form
of cupriachelin exhibits photoreactive properties. Exposure to
UV light induces an oxidation of its peptidic backbone and a
concomitant reduction of the coordinated Fe(III). Here, we
report the genomics-inspired isolation and structural character-
ization of cupriachelin as well as its encoding gene cluster, which was identified by insertional mutagenesis. Based upon the
functional characterization of adenylation domain specificity, a model for cupriachelin biosynthesis is proposed.
open reading frames encoding lipoprotein receptors and ABC-
type iron transporters often provides ample evidence for the
presence of a siderophore locus. The linkage of the different
building blocks in siderophore biosynthesis is catalyzed by two
distinct enzyme classes, namely nonribosomal peptide
synthetases (NRPS)7 and NRPS-independent siderophore
synthetases (NIS).8 Exploiting the inherent modular logic of
the former,9 it is possible to predict the structure of the
encoded siderophore and to design selective protocols for its
isolation. The first siderophore to be discovered applying an in
silico strategy was coelichelin from the model actinomycete
Streptomyces coelicolor A3(2) in 2005.10 Since then, further
previously unrecognized siderophores have been identified by
genome mining approaches.11−14 Notably, all these studies
focused on actinobacteria, which are well-known for their
secondary metabolic proficiency,15 including siderophore
biosynthesis, but not on other groups of microorganisms.
This is surprising considering the increasing availability of
microbial genome sequence data and the ensuing identification
of neglected organisms as alternative natural product
resources.16
INTRODUCTION
■
Iron is indispensable for the vast majority of organisms due to
its involvement in cellular respiration and metabolic processes.1
Notwithstanding its abundance in the Earth’s crust, the
transition metal is not readily available in aerobic environments,
including soil, freshwater, and marine habitats. At physiological
pH and in the presence of water the biologically accessible form
of the element, Fe(II), spontaneously oxidizes to give insoluble
ferric oxide hydrate complexes. In order to overcome the
environmental scarcity of ferrous iron, bacteria and fungi have
thus evolved specific iron uptake systems.1 One strategy
involves the secretion of low molecular weight compounds, so-
called siderophores, that coordinate Fe(III) with high affinity.
Once a siderophore ligand has bound the metal ion, it is
actively transported back into the cell, where the iron is
released by a reductive (or hydrolytic) mechanism.2 It has been
shown that siderophore production may exert significant
biological effects on the environment of an organism, be it
the shaping of microbial communities3 or the suppression of
host defense mechanisms in case of a pathogenic siderophore
producer.4 Therefore, there is continued interest in the
identification and functional characterization of these metabo-
lites.
In the present study, we investigated the potential of the
industrially relevant bacterium Cupriavidus necator H16 (syn.
Ralstonia eutropha H16) for siderophore biosynthesis. C. necator
H16 is known to accumulate organic carbon in the form of
polyhydroxyalkanoates, which have found wide application as
raw materials for the production of medical devices.17
Furthermore, the oxygen-tolerant [NiFe]-hydrogenases of the
Genome mining approaches have proven to be particularly
useful for the discovery of novel siderophores, as the high metal
affinity of such compounds is associated with structural features
that can be predicted by sequence analyses from their
respective gene clusters; examples include catecholate or
hydroxamate groups, which are commonly found in side-
rophores.2 The assembly of these moieties typically requires a
select set of biosynthetic genes,5,6 and physical proximity to
Received: January 19, 2012
Published: March 1, 2012
© 2012 American Chemical Society
5415
dx.doi.org/10.1021/ja300620z | J. Am. Chem. Soc. 2012, 134, 5415−5422

Journal of the American Chemical Society
Article
Figure 1. Organization of the cuc biosynthetic gene cluster. Genes are color-coded according to their proposed functions.
bacterium have attracted much interest, as they may serve as
catalysts in photoelectrochemical biofuel cells.18 In contrast,
virtually nothing is known about the secondary metabolism of
C. necator H16. Analysis of its genome sequence revealed the
presence of a locus putatively involved in the biosynthesis of
the siderophore vibrioferrin19 as well as a chromosomally
encoded NRPS gene cluster of unknown function.20 The latter
appears to be unique to the strain H16, lacking homologues
even in phylogenetically related bacteria. The gene architecture
of the orphan NRPS locus suggested a role in siderophore
biosynthesis; a functional link with the annotated vibrioferrin-
like cluster could not be excluded. Using a genome mining
strategy that involved targeted gene disruption and comparative
metabolic profiling of mutant and wild-type strains, we tracked
the metabolites that derive from the NRPS gene cluster. The
structure of the predominating compound, cupriachelin, was
fully characterized by spectroscopic methods, and its stereo-
chemistry was solved following chemical derivatization.
Cupriachelin is unique in that it possesses characteristic
physicochemical properties of marine siderophores, albeit the
natural product is made by a terrestrial bacterium. The origin of
the β-hydroxyaspartate moieties, which contribute to the
photoreactivity of ferric cupriachelin, was biochemically
interrogated.
pantetheinyl transferase and is thus essential to convert carrier
protein domains in NRPSs from their apo into the active holo
forms.7 A similar role is conceivable for the MbtH-like protein
CucD, which might contribute to the activation of amino acid
building blocks as a cofactor.22 The four NRPSs that derive
from cucF, cucG, cucH, and cucJ are of particular interest, as they
account for the peptidic backbone of the encoded metabolite.
The corresponding proteins harbor five modules, each
including the complete set of condensation (C), adenylation
(A) and peptidyl carrier protein (PCP) domains, but they lack
a typical loading module and a terminal thioesterase (TE)
domain. While the offload from the assembly line might be
accomplished by the discrete hydrolase CucC, no possible
candidate gene for the activation of a starter unit was found
within the cluster. The fact that all NRPSs include an N-
terminal C domain is reminiscent, to a degree, to a situation
known from the biosynthesis of lipopeptide antibiotics, in
which a starter C domain acylates the first amino acid with a
fatty acid.23 To identify the substrates of the A domains in every
NRPS module, we analyzed their putative binding pockets both
manually and using computational specificity predictions.24,25
Clear results were obtained only for three out of the five A
domains (Table 1). The failure to predict substrate preferences
Table 1. Substrate Specificity Predictions for the
Adenylation Domains of the NRPSs Encoded in the cuc
Gene Cluster
RESULTS AND DISCUSSION
■
Architecture and Features of the NRPS Gene Cluster
of Strain H16. A putative siderophore locus was identified on
chromosome 2 of the C. necator H16 genome due to the local
concentration of genes associated with iron-siderophore
transport as well as siderophore assembly. Cluster boundaries
were assigned on the basis of functional annotations, on the
one hand, and operon predictions, on the other.21 According to
these analyses, the proposed cluster includes 17 ORFs between
h16_B1676 and h16_B1692 covering 36.4 kb of contiguous
DNA (Figure 1); the genes have been renamed in this study
from cucA to cucQ.
A domain
signature sequence
substrate according to prediction
CucF-A1
CucF-A2
CucG
DLTKVGHVGK
DIWELTADDK
DLTKIGHIGK
DGEGSGGVTK
DILQLGVVWK
L-Asp
unknown
L-Asp
CucH
unknown
Gly
CucJ
for the second A domain in CucF and for the A domain of
CucH hence suggested the priming of uncommon building
blocks.
A total of six genes (cucI, cucK, cucN, cucO, cucP, and cucQ)
are involved in siderophore export and uptake, while four
additional genes (cucA, cucE, cucL, and cucM) have likely roles
in tailoring reactions. The latter two, cucM and cucL, were
annotated to encode a lysine/ornithine N-monooxygenase and
an acyltransferase, respectively. The concerted action of both
enzymes would provide a hydroxamate moiety via an
intermediary hydroxylamine group, analogous to other side-
rophore biosynthetic pathways.8 In contrast, the functions of
the putative aminotransferase CucA and of the dioxygenase
CucE were less obvious. Among the seven remaining proteins
encoded on the gene cluster, CucB represents a phospho-
A complete overview of the organization of the cuc gene
cluster is provided in Table S1, Supporting Information.
Discovery and Isolation of Cupriachelin. We first noted
the secretion of iron-chelating metabolites by C. necator H16 in
the modified chrome azurol S (CAS) agar assay, in which the
siderophore detection is spatially separated from the growth
area of the bacterium.26,27 The distinctive color change from
blue to orange that can be observed in the presence of
siderophores is often used to guide the isolation of these
compounds after cultivation in a liquid medium. In case of C.
necator H16, however, this methodology was not applicable,
5416
dx.doi.org/10.1021/ja300620z | J. Am. Chem. Soc. 2012, 134, 5415−5422

Journal of the American Chemical Society
Article
since the required growth medium gave a false positive
response in the liquid CAS assay, possibly due to an
interference of media components.27 We therefore set out to
identify the potential siderophores from strain H16 by
comparative metabolic profiling of cultures grown in the
presence or absence of exogenous Fe(III), according to a
previously established protocol.28 While no additional peak
(with significant signal intensity) showed up in the UV
chromatogram of the iron-deficient culture, liquid chromatog-
raphy electrospray ionization mass spectrometry (LC-ESI-MS)
revealed the presence of distinctive pseudomolecular ions at m/
z 808 [M + H]+ and m/z 806 [M − H]−, respectively. These
ions disappeared once the culture was supplemented with 190
μM of Fe(NH4) citrate. To probe whether the annotated cuc
cluster is involved in the biosynthesis of a putative siderophore
with the corresponding mass, we disrupted the NRPS gene
cucH by insertional mutagenesis (Figure S21, Supporting
Information). The ensuing mutant strain was no longer capable
to induce a color change in the modified CAS agar assay.
Furthermore, it lacked the target ions at m/z 808 [M + H]+ and
m/z 806 [M − H]− when grown under iron limitation, thus
confirming the involvement of cucH in siderophore assembly.
The observation that the cucH mutant sustained its growth at
low iron conditions suggests the presence of an alternative
mechanism for iron acquisition. In compliance with the
modified CAS assay, we did not detect any evidence for the
production of vibrioferrin or another complementing side-
rophore by LC-MS and 1H NMR analysis. We noticed,
however, that C. necator H16 possesses an operon on
chromosome 2, which encodes a ferrous iron uptake system
(H16_B0083−H16_B0085).29 The latter might, in conjunc-
tion with secreted Fe(III) reductases,30,31 safeguard the iron
supply of the bacterium.
Figure 2. (A) Structure of cupriachelin. The stereochemistry was
assigned following chemical derivatization of a cupriachelin acid
hydrolysate by Marfey’s method and chiral GC-MS, respectively. (B)
Metabolic profiles of the C. necator H16 wild-type (profile a) and the
cucH mutant strain (profile b) grown under iron-deficient conditions
as well as from the wild-type strain in the presence of 190 μM
Fe(NH4) citrate (profile c). The chemical structures of the metabolites
(1) and (2) are given in Figure S1, Supporting Information.
acylated with the 3-hydroxybutanoyl moiety, giving rise to a
hydroxamate function. In accordance with the A domain
specificity analysis, a glycine moiety could be deduced, while
the two predicted Asp residues turned out to be β-hydroxylated
in cupriachelin. The remaining amino acid, for which no
prediction was available, could be identified as 2,4-diaminobu-
tyric acid (Dab). The sequence of the single residues was finally
determined exploiting NOE and HMBC correlations of the
amide protons. The ensuing structure was subsequently
confirmed by matrix-assisted laser desorption ionization time-
of-flight (MALDI-TOF)/TOF fragmentation (Figure S2,
Supporting Information).
During the structure elucidation of cupriachelin, we tested its
preference for the coordination of different metal ions. MS
studies showed that the natural product is forming monomeric
1:1 complexes with Fe3+ and Ga3+ (Figure S12, Supporting
Information) but does not chelate Mn2+, Co2+, Ni2+, Cu2+, and
Zn2+ (data not shown). The observed discrimination between
divalent and trivalent cations, together with the Fe(III)-
responsive production of cupriachelin, corroborates a potential
siderophore role.
Absolute Configuration of Cupriachelin. To resolve the
configuration of the amino acids in cupriachelin, we applied
Marfey’s method.32 Acidic cleavage using 6 N HCl is a common
strategy to liberate the amino acid constituents from a peptide
and is usually successful when the latter is composed of
proteinogenic amino acids only. In the case of cupriachelin,
HCl-promoted hydrolysis yielded Dab, but the masses of the
other amino acids could not be detected in the hydrolysate. To
release the two β-hydroxyaspartic acid residues as well as the
Orn moiety, a reductive HI cleavage was hence carried out.33
Following the derivatization with Marfey’s reagent (1-fluoro-
2,4-dinitrophenyl-5-L-alanine amide, L-FDAA), the Dab and
Orn residues of cupriachelin were both determined to be in L
For the isolation of the siderophore that derives from the cuc
gene cluster, we evaluated different extraction conditions. The
recovery of the water-soluble metabolite, which we tentatively
named cupriachelin, was significantly increased by adsorption
onto XAD-2 resin and subsequent elution with methanol. This
approach did not only give a yield (6.5 mg/L) sufficient for UV
detection but also revealed some minor metabolites in the same
mass range, co-occurring with cupriachelin (Figure 2). Cleanup
of cupriachelin was accomplished via reversed-phase chroma-
tography.
Planar Structure and Complexing Properties of
Cupriachelin. The empirical formula of cupriachelin was
assigned to be C33H57O16N7 by high-resolution (HR)-ESI-MS
(Figure S11, Supporting Information), which corresponds to 9
degrees of unsaturation. An inspection of the 13C NMR
spectrum immediately revealed that all double-bond equivalents
could be ascribed to carbonyl moieties, five of which were
present as amides according to heteronuclear long-range
correlations with exchangeable protons at 8.65, 8.37, 8.33,
1
8.20, and 8.15 ppm. First-order multiplet analysis of the H
NMR spectrum and correlation spectroscopy (COSY) data
allowed the discrimination of seven spin systems, including a
decanoyl, a 3-hydroxybutanoyl, and five amino acid moieties
(Figure S2, Supporting Information), which is consistent with
the number of NRPS modules encoded on the cuc gene cluster.
The preceding bioinformatic analyses facilitated the spectro-
scopic identification of the amino acid residues. As expected
from the preliminary annotation of CucL and CucM, a discrete
set of signals could be assigned to an Nδ-hydroxyornithine
residue. HMBC interactions revealed that the latter was N-
5417
dx.doi.org/10.1021/ja300620z | J. Am. Chem. Soc. 2012, 134, 5415−5422

Journal of the American Chemical Society
Article
Figure 3. Proposed reaction scheme for the UV photolysis of Fe(III)-cupriachelin. The ‘thunderbolt’ indicates the position of cleavage. The depicted
cleavage product was detected by HR-ESI-MS, and its structure was deduced by tandem mass spectrometry. Fe(III) is likely to be reduced via ligand-
to-metal charge transfer.
Figure 4. Relative activity of CucG with selected amino acids as judged by the ATP/PPi exchange assay for adenylation function.
configuration (Figures S15 and S16, Supporting Information) .
Furthermore, Marfey’s method indicated the exclusive presence
of threo-β-hydroxyaspartic acid in cupriachelin. It has previously
been demonstrated that the elution order of the diastereomeric
pairs of L-FDAA-derivatized threo-β-hydroxyaspartic acid is D→
L under reversed-phase conditions.34 The cupriachelin hydro-
lysate contained only one single peak of the correct mass upon
conversion with L-FDAA, which was found to elute at the same
retention time as the second peak of commercial D,L-threo-β-
hydroxyaspartic acid (Figure S16, Supporting Information).
This substantiates that, out of the four possible stereoisomers of
β-hydroxyaspartic acid, only the L-threo form is present in
cupriachelin.
To determine the configuration of the 3-hydroxybutyric acid
(Hbu) moiety, we conducted a chiral separation on a GC
column. Since the treatment of cupriachelin with neither
concentrated HI nor 6 N HCl yielded intact Hbu, we resorted
to a mild hydrolysis using 0.3 N HCl for 4 days at room
temperature.35 Upon conversion to its trifluoroacetyl ester, we
identified L-Hbu by comparison with authentic standards
(Figures S17 and S18, Supporting Information), thus establish-
ing the complete stereochemistry of cupriachelin.
photolysis when coordinated to Fe(III), resulting in the
reduction of the chelated ferric iron.37 This redox cycling
appears to have important ecological implications in the oceans.
By releasing photoreactive siderophores, marine bacteria
promote algal assimilation of iron and thus boost algal
growth.38 In return, they expand their pool of accessible
organic carbon.38 To evaluate whether cupriachelin is likewise
prone to photochemical reactions, we exposed an aqueous
solution of its ferric-ion complex to sunlight and analyzed the
reaction via HR-ESI-MS (Figures S19 and S20, Supporting
Information). As expected on the basis of its molecular
structure, photolytic cleavage occurred at a β-hydroxyaspartate
residue (Figure 3). The concomitant reduction of Fe(III) to
Fe(II) was demonstrated in a second assay by trapping the
latter with the specific chelating agent bathophenanthrolinedi-
sulfonate (BPDS). Samples of Fe(III)-cupriachelin that were
exposed to sunlight gradually turned red in the presence of
BPDS, whereas light-protected control reactions stayed color-
less in the same period. In the former case, the absorption at
535 nm increased from 0.002 0.001 (prior to exposure) to
0.078 0.002 (after exposure). The reactions that were kept in
the dark showed a mean absorption of 0.006 0.002 after the
incubation, starting from absorption values of 0.005 0.003.
The absorbance of the ferrous tris(BPDS) complex was stable
at room temperature for at least six hours.
Photoreactivity of Cupriachelin. Cupriachelin exhibits
structural features that are generally associated with side-
rophores from marine bacteria, including a fatty acid moiety
and α-hydroxy acid residues, namely two β-hydroxyaspar-
tates.2,36 The latter were previously shown to mediate
CucG ATP/PPi Exchange Assay. Considering the role of
β-hydroxyaspartate in the photoreactivity of cupriachelin and
5418
dx.doi.org/10.1021/ja300620z | J. Am. Chem. Soc. 2012, 134, 5415−5422

Journal of the American Chemical Society
Article
Figure 5. Proposed biosynthetic pathway for cupriachelin assembly. Domain notation: C, condensation; A, adenylation; ACP, acyl carrier protein;
PCP, peptidyl carrier protein; and TauD, hydroxylase.
many other siderophores,36 we were interested whether the
formation of this nonproteinogenic amino acid occurs prior to
the tethering to the assembly line or at a later biosynthetic
stage. Both the manual and the computational specificity
prediction of the responsible A domains (CucF-A1 and CucG-
A) suggested a preference for L-Asp (Table 1). On the other
hand, a signature motif for β-hydroxyaspartate has not been
reported to date. Therefore, we overexpressed and purified
CucG from Escherichia coli BL21(DE3) as an N-terminal His-
tag fusion via the expression plasmid pET28a(+) (Figure S22,
Supporting Information). Biochemical characterization of the
recombinant protein in the ATP/PPi exchange assay revealed a
distinct preference for Asp, confirming the bioinformatic
prediction.24,25 The results also showed affinity toward
asparagine and β-hydroxyaspartate, albeit at significantly
reduced levels, with the former being a somewhat better
amino acid substrate than the latter (Figure 4). Despite
intensive screening of C. necator H16, we did not find any
evidence for the production of a cupriachelin derivative
harboring an Asn residue. This suggests that the observed
activity with Asn and β-hydroxyaspartate in the ATP/PPi
exchange assay is likely to be negligible.
Model for the Biosynthesis of Cupriachelin-Type
Siderophores. Based upon the aforementioned bioinformatic
and experimental results, a model for cupriachelin biosynthesis
was deduced (Figure 5). Decanoic acid is hence proposed to act
as the biosynthetic starter unit, but it can also be replaced by
other fatty acids, e.g. octanoic acid or 3-hydroxydecanoic acid,
as seen in some byproducts of the major metabolite
cupriachelin (Figure S1, Supporting Information). Decanoic
acid is expected to originate from fatty acid metabolism, where
it is already present in its activated form as decanoyl-CoA or as
its biosynthetic predecessor β-hydroxydecanoyl-CoA, respec-
tively. Similar to the lipoinitiation reaction in surfactin
biosynthesis,39 the fatty acyl-CoA thioester is supposedly
loaded onto the acyl carrier protein domain of module 1 and
then transferred to the N-terminal C domain of CucF, which
catalyzes the acylation of PCP-bound L-Asp. In the next two
steps, the C domain of module 2 attaches the L-Dab moiety,
and CucG (representing module 3) incorporates the following
L-Asp residue. From the structure of cupriachelin it is clear that
the CucJ-catalyzed condensation of Gly must precede the
attachment of L-Nδ-hydroxy-Nδ-(3-hydroxybutyryl)-ornithine,
which is carried out by the C domain of CucH. The fully
assembled pentapeptidyl thioester is eventually released upon
action of CucC.
Although the biosynthesis largely follows the classical NRPS
enzymatic logic (maybe except for the sequence of reactions
catalyzed by CucH and CucJ), the structure of cupriachelin
could hardly be predicted from the analysis of its gene cluster
due to the high fraction of functionalized, nonproteinogenic
amino acids. By analyzing the substrate specificity of CucG, we
have demonstrated that the hydroxylation of Asp must takes
place after the initiating adenylation step, and we assume the
same to be true for the CucF-derived Asp moiety. The discrete
dioxygenase CucE as well as the C-terminal TauD domain of
CucF are possible candidates to catalyze the required
hydroxylations, which probably occur with the Asp residues
tethered as protein-bound S-pantetheinyl thioesters, as shown
in syringomycin biosynthesis.40 In contrast, the preparation of
L-Dab likely takes place prior to its adenylation, considering the
specificity-determining amino acid residues in CucF-A2. The
aminotransferase CucA might be involved in the conversion of
L-Asn to L-Dab, but this still needs to be proven. While the
presence of an N-hydroxylated and N-acylated ornithine or
lysine moiety was somewhat expected due to the annotation of
5419
dx.doi.org/10.1021/ja300620z | J. Am. Chem. Soc. 2012, 134, 5415−5422

Journal of the American Chemical Society
Article
300 K on a Bruker Avance III 500 MHz spectrometer with D2O or
H2O/D2O (4:1) as solvent and internal standard.
CucL and CucM, none of the A domains in the cupriachelin
NRPS enzymes exhibited a motif indicating the priming of L-
Nδ-formyl-Nδ-hydroxyornithine10 or L-Nδ-acetyl-Nδ-hydroxyor-
nithine.12,41 As evidenced by the literature, the actually
observed L-Nδ-hydroxy-Nδ-(3-hydroxybutyryl)-ornithine is not
an uncommon constituent of siderophores,2 but A domains
that are responsible for the tethering of this amino acid have
not been dissected to date.
Bacterial Strains and Growth Conditions. C. necator H16 was
grown in H-3 mineral medium lacking ammonium ferric citrate (1 g/L
aspartic acid, 2.3 g/L KH2PO4, 2.57 g/L Na2HPO4, 1 g/L NH4Cl, 0.5
g/L MgSO4 × 7 H2O, 0.5 g/L NaHCO3, 0.01 g/L CaCl2 × 2 H2O,
and 5 mL/L SL-6 trace element solution).46 The influence of Fe3+ on
cupriachelin production was tested in the same medium containing
190 μM ammonium ferric citrate. E. coli strains were cultured in liquid
or solidified LB medium. When required, antibiotics were added to the
following concentrations: 100 μg/mL ampicillin, 25 μg/mL
chloramphenicol, and 50 μg/mL kanamycin, respectively.
In summary, we isolated and structurally characterized a new
lipopeptide siderophore, cupriachelin, from the bioplastic
producer C. necator H16. The secondary metabolite was
discovered by using a genome mining strategy and represents
the first natural product from the genus Cupriavidus. The ferric
form of cupriachelin was shown to degrade when exposed to
sunlight. At the same time, the coordinated Fe(III) is reduced,
and ferrous iron is released, paralleling observations that have
been made with siderophores from oceanic bacteria.36 It is
evident that a biological role for cupriachelin’s photoreactivity is
hardly imaginable in a soil environment. Yet, while most
representatives of the species Cupriavidus necator are pertained
to soil,42 the investigated strain H16 was originally isolated
from sludge of a creek.43 The natural occurrence of the
bacterium in a riverine habitat suggests a function for
cupriachelin in freshwater ecology that is analogous to that of
marine siderophores, i.e., supplying ferrous iron for uptake by
planktonic assemblages.37,38 The evolutionary destined release
of the siderophore in an aquatic environment would also
provide a plausible explanation for its amphiphilic structure. We
assume that the inherent surface activity of cupriachelin will
slow its diffusion away from the producing bacterium and
secure a relatively high concentration around the cell.44
Comparative genomics lends support toward our hypothesis
on the biological role of cupriachelin: The majority of genes of
C. necator H16 have orthologs in the genome of C. necator
strain N-1,45 but notwithstanding the great degree of global
synteny between the two genomes, the cuc gene cluster is
completely absent in the soil-derived N-1 strain.
Siderophore Screening. Chrome azurol S (CAS) plates were
prepared as previously reported.26,27 Half of each CAS agar layer was
cut out, and the gap was filled with H-3 mineral medium agar. C.
necator wild-type and mutant strains were plated on the H-3 half of the
plates. The secretion of iron-scavenging molecules was detected by a
color change from blue to orange after overnight incubation at 30 °C.
Construction of the cucH Disruption Mutant. A fragment of
the gene-disruption target cucH was amplified by PCR from C. necator
H16 genomic DNA with the primers P1 (5′-GAATT-
CAAGGGGCTGCGCGTTACGTC-3′) and P2 (5′-GAATT-
CACTGGTTGAAGCGGTTGCGGC-3′). The PCR product was
cloned into pJET1.2 (Fermentas) to give pHiK003, which was
subsequently transformed into E. coli BW25113/pIJ790 by electro-
poration.47 Insertion of a kanamycin resistance cassette into the 2.2 kb
gene fragment of cchH was achieved by λ red-mediated recombina-
tion.47 For this purpose, sequences homologous to cchH were
appended at either end of the resistance gene by PCR using primers
P3 (5′-CCCGAACACGATGCCGAAATTGCCGTGCACCCGGCG-
CAGTGACTAACTAGGAGGAATAA-3′) and P4 (5′-AAACG-
GATCGGGAATAAAGCGGTCTGCCGTCATGCCGGC-
TATTCCTTCCAGTACTAAAC-3′) as well as pAphA-3 as the PCR
template.48 After the induction of λ red genes in E. coli BW25113/
pIJ790/pHiK003 by addition of 10 mM L-arabinose at 30 °C, the cells
were electrotransformed with the PCR product. Integration of the
marker in the plasmid DNA by homologous recombination yielded the
cucH disruption vector pHiK004. The latter was introduced into C.
necator H16 by electroporation. For electroporation, the wild type of
C. necator H16 was cultured in LB medium at 30 °C to an optical
density of 0.4 at 600 nm. After centrifugation, the cells were washed
with ice-cold washing buffer containing 1 mM N-(2-hydroxyethyl)-
piperazine-N′-ethanesulfonic acid (pH 6.8) and 10% glycerol. The
ensuing electrocompetent cells were transformed with pHiK004 in a
0.2 cm gap cuvette using a Bio-Rad GenePulser II set to 200 Ω, 25 μF,
and 2.5 kV. SOC medium was added to the shocked cells, which were
then cultured at 30 °C for 3 h. The regenerated culture was spread
onto LB agar containing 350 μg/mL of kanamycin and incubated at 30
°C. Positive transformants were identified by colony PCR.
Isolation and Purification of Cupriachelin. C. necator H16 was
cultivated in 5 L Erlenmeyer flasks filled with 2.5 L of H-3 mineral
medium. The strain was shaken (150 rpm) at 30 °C for 3 days. At the
end of cultivation, the culture supernatant was separated from the cells
by centrifugation at 9500 × g and extracted with 150 g/L Amberlite
XAD-2 (Supelco) for 24 h. After two washing steps with distilled
water, the adsorbed metabolites were eluted from the resin with
methanol. The ensuing eluate was concentrated under vacuum to
dryness prior to resuspension in 1 mL of H2O. Initial fractionation of
the extract was accomplished by flash column chromatography over
Polygoprep 60−50 C18 (Macherey-Nagel) using an increasing
concentration of methanol in water. Fractions that showed the target
ion at m/z 808 [M + H]+ were further purified on a Shimadzu UFLC
liquid chromatography system equipped with a Nucleodur C18 HTec
column (VP 250 × 10 mm, 5 μm; Macherey-Nagel) using an isocratic
flow of 30% acetonitrile in water + 0.1% trifluoroacetic acid at a flow
rate of 4 mL min−1 with wavelength monitoring at 210 nm.
EXPERIMENTAL PROCEDURES
■
Sequence Analyses. The genome of C. necator H16 consists of
three circular replicons (GenBank accession numbers: NC_008313,
NC_008314, and NC_005241, respectively), which were screened for
genes putatively involved in siderophore biosynthesis using homology-
based alignments. To identify potential biosynthetic gene clusters,
initial hits were grouped by physical proximity and mapped onto the
genome using Vector NTI (Invitrogen). Identified loci were further
analyzed using Web-based bioinformatic software, e.g., NRPSpredic-
tor2.25
General Experimental Procedures. LC-MS experiments were
conducted on an Agilent 1100 series HPLC-DAD system coupled with
an MSD trap (Agilent) operating in alternating ionization mode and
an Antek 8060 HPLC-CLN detector (Antek Instruments GmbH)
using a C8 column (Zorbax Eclipse XDB C8, 150 × 4.6 mm, 5 μm;
Agilent). A linear gradient of methanol in water + 0.1% trifluoracetic
acid (10% → 90% methanol within 15 min; flow rate 1 mL min−1) was
used for metabolic profiling. Analytical HPLC for Marfey’s analysis
was conducted on a Shimadzu UFLC liquid chromatography system
equipped with a Nucleosil 100 C18 column (250 × 4.6 mm, 5 μm; CS
Chromatographie service). The separation was accomplished by using
a linear gradient from solvent A (10% acetonitrile in 25 mM aqueous
KH2PO4) to solvent B (50% acetonitrile + 0.05% trifluoroacetic acid)
over 40 min with wavelength monitoring at 365 nm. High-resolution
mass determination was carried out using an Exactive Mass
Spectrometer (Thermo-Scientific). NMR spectra were recorded at
Amino Acid Analysis by Marfey’s Method. Configuration of
the amino acids present in cupriachelin was determined following acid
hydrolysis and derivatization with Marfey’s reagent32 (1-fluoro-2,4-
dinitrophenyl-5-L-alanine amide, L-FDAA, Sigma-Aldrich) by coelution
5420
dx.doi.org/10.1021/ja300620z | J. Am. Chem. Soc. 2012, 134, 5415−5422

Journal of the American Chemical Society
Article
experiments with FDAA-derivatized amino acids. To this end, 500 μg
purified cupriachelin was dissolved in 400 μL concentrated HI and
heated at 110 °C for 3 h. The solution was lyophilized, and the dried
hydrolysate was resuspended in 10 μL of water and 20 μL of 1 M
NaHCO3. Derivatization was carried out with 170 μL of 1% L-FDAA
in acetone at 37 °C for 1 h. The products were lyophilized and
prepared for LC-MS analysis by dissolving in 1 mL of 50% acetonitrile.
Standards for cochromatography were prepared by reacting 50 μL of
50 mM aqueous amino acid solution with 20 μL of 1 M NaHCO3 and
100 μL of 1% L-FDAA in acetone at 37 °C for 1 h. The lyophilized
products were then dissolved in 1 mL of 50% acetonitrile. LC-MS
analysis was performed under the aforementioned conditions with a
detection wavelength of 365 nm. For the analysis of 2,4-
diaminobutyric acid, we applied the traditional cleavage method with
6 N HCl for 24 h. The derivatization with L-FDAA was conducted as
described for the HI cleavage products.
Configuration of the 3-Hydroxybutyrate Moiety. A 1 mL
aliquot of 0.3 N HCl was added to 1.3 mg of cupriachelin in an
Eppendorf vial. The vial was kept at room temperature for 4 days after
which excess reagent was removed under vacuum. The hydrolysate
was resuspended in water and purified over a Sep-Pak RP-18 cartridge.
Following the removal of the eluent, 100 μL of trifluoroacetic
anhydride/methylene chloride (1:1, v/v) was added, and the mixture
heated at 100 °C for 10 min. Excess reagents were removed with a
stream of argon. Methylene chloride, 100 μL, was added, and the
resulting trifluoroacetyl ester analyzed using a Trace GC Ultra
(Thermo) coupled in parallel to an FID detector and a Polaris-Q ion
trap mass detector (Thermo). Separation was established on a chiral
silica column (25 m × 0.25 mm ID) coated with Chir-D-Val (Varian)
using a continuous He gas flow of 1.5 mL min−1 and a temperature
gradient from 40 to 200 °C at a rate of 5 °C min−1. Co-
chromatography of the derivatized cupriachelin hydrolysate was
conducted against a derivatized Hbu standard in D-configuration as
well as a mixture containing both enantiomers.
NaH2PO4, 300 mM NaCl, 10 mM imidazole, 20 mM β-
mercaptoethanol, 10% glycerol) and subjected to sonication. The
cell debris was subsequently removed by centrifugation, and the
supernatant was loaded onto a Ni-NTA column. The recombinant
protein was eluted using lysis buffer containing increasing concen-
trations of imidazole. Fractions that contained the His-tagged CucG
were identified via sodium dodecyl sulfate polyacrylamide gel
electrophoresis (SDS-PAGE), desalted with PD-10 columns (GE
Healthcare), and concentrated using Amicon ultracentrifugation
devices (Millipore). The identity of the purified protein was confirmed
by MALDI-TOF MS.
ATP/PPi Exchange Assay. The substrate preference of the
adenylation domain of CucG was determined by amino acid-
dependent exchange of the radiolabel from [32P]-pyrophosphate
(PPi) to ATP according to established protocols.49 A 100 μL reaction
contained 80 mM Tris-HCl (pH 7.5), 5 mM MgCl2, 5 mM ATP, 100
nM purified enzyme, 0.1 mM [32P]-pyrophosphate, and 1 mM amino
acid. Reactions were initiated by addition of enzyme and incubated at
37 °C for 30 min prior to quenching with 500 μL of charcoal
suspension (1% charcoal and 4.5% Na4P2O7 in 3.5% perchloric acid).
Precipitate was collected with paper filter discs under vacuum.
Following three consecutive washing steps with 40 mM Na4P2O7 in
1.4% perchloric acid (200 mL), water (200 mL), and ethanol (200
mL), the filter papers were added to 2.5 mL of scintillation fluid and
read by a Beckman Coulter LS6500 multipurpose scintillation counter.
All reactions were run in triplicate.
ASSOCIATED CONTENT
* Supporting Information
■
S
A complete annotation of the cuc gene cluster, including
GenBank accession numbers, NMR data and spectra of
cupriachelin, MS data of cupriachelin and its minor congeners,
HPLC-MS and GC-MS of cupriachelin stereochemistry
assignment, photoreactivity test of Fe(III)-cupriachelin, PCR-
based confirmation of cucH gene disruption, and SDS-PAGE of
purified recombinant CucG. This material is available free of
charge via the Internet at http://pubs.acs.org.
Photoreactivity Test of Cupriachelin: Identification of
Degradation Products. A 2 mM solution of ferric cupriachelin in
PBS buffer (pH 7.5) was prepared and exposed to natural sunlight for
6 h. An identical solution that was shielded from sunlight served as a
negative control. After photoexposure, both samples were dried in
vacuo. The samples were then taken up in 100 μL of 50% methanol
and subjected to LC coupled with HR-MS analysis. For this purpose,
an Accela UPLC-system (Thermo Scientific) equipped with a Betasil
C18 column (2.1 × 150 mm, 3 μm; Thermo Scientific) and an
Exactive mass spectrometer was used. HPLC conditions were as
follows: 5% → 98% acetonitrile in water + 0.1% formic acid within 15
min; flow rate 250 μL min−1.
AUTHOR INFORMATION
Corresponding Author
Markus.Nett@hki-jena.de.
■
Notes
The authors declare no competing financial interest.
Photoreactivity Test of Cupriachelin: BPDS assay. Reduction
of the complexed ferric iron to ferrous iron was examined using a
modified photolysis assay as described for vibrioferrin.38 Each reaction
contained 100 μM cupriachelin, 10 μM FeCl3, and 40 μM of the
ferrous trapping agent BPDS (Fluka) in PBS buffer (pH 7.5). The
reactions were either exposed to sunlight or stored in the dark for 4 h.
ACKNOWLEDGMENTS
We gratefully acknowledge financial support from the
■
Bundesministerium fur Bildung und Forschung (BMBF
̈
GenoMik-Transfer programme; grant no. 0315591A) and
from the Jena School for Microbial Communication (JSMC).
2+
The formation of Fe(BPDS)3 was recorded before and after
We thank A. Perner and M. Poetsch (Hans-Knoll-Institute
Jena) for numerous HR-ESI-MS and MALDI-TOF/TOF
measurements, respectively. We are further indebted to D.
̈
exposure to sunlight/darkness by measuring the absorption at 535
nm using a Genesys 10 UV spectrophotometer (Thermo). All
reactions were run in duplicate.
Cloning, Expression, and Purification of CucG. The entire
ORF of cucG including a 30 nt sequence downstream of its stop codon
was amplified by PCR with primers P5 (5′-C ATATGCCTGAGCTG-
CAGTCCGCCTG-3′) and P6 (5′-GTATCACGGTCAATGGA-
TATCGGGGAG-3′). The PCR product was cloned into pJET1.2 to
give pHiK005. After digestion with NheI and HindIII, the resulting
fragment was cloned as an N-terminal His-tag fusion into pET28a(+)
(Novagen). The resulting vector pHiK006 was transformed into E. coli
BL21(DE3) as an expression host. The overexpression strain was
grown at 37 °C in terrific broth (TB) containing 50 μg/mL kanamycin
to an OD of 0.6 at 600 nm. To induce the protein expression, 1 mM of
isopropyl 1-thio-β-D-galactopyranoside (IPTG) was added to the
culture, which was afterward incubated at 16 °C for 20 h. After
centrifugation the cell pellet was resuspended in lysis buffer (50 mM
Hoffmeister (Friedrich-Schiller-Universitat Jena) and his
̈
research group for valuable discussions as well as advice on
the experimental setup of the ATP/PPi exchange assay. We also
̈
thank N. Uberschaar (Hans-Knoll-Institute Jena, Department
̈
of Biomolecular Chemistry) for help with the GC analysis.
REFERENCES
■
(1) Andrews, S. C.; Robinson, A. K.; Rodriguez-Quinones, F. FEMS
Microbiol. Rev. 2003, 27, 215−237.
(2) (a) Hider, R. C.; Kong, X. Nat. Prod. Rep. 2010, 27, 637−657.
(b) Sandy, M.; Butler, A. Chem. Rev. 2009, 109, 4580−4595.
(c) Miethke, M.; Marahiel, M. A. Microbiol. Mol. Biol. Rev. 2007, 71,
413−451.
5421
dx.doi.org/10.1021/ja300620z | J. Am. Chem. Soc. 2012, 134, 5415−5422

Journal of the American Chemical Society
Article
(3) D’Onofrio, A.; Crawford, J. M.; Stewart, E. J.; Witt, K.; Gavrish,
E.; Epstein, S.; Clardy, J.; Lewis, K. Chem. Biol. 2010, 17, 254−264.
(4) Paauw, A.; Leverstein-van Hall, M. A.; van Kessel, K. P. M;
Verhoef, J.; Fluit, A. C. PLoS ONE 2009, 4, e8240.
(5) Walsh, C. T.; Liu, J.; Rusnak, R.; Sakaitani, M. Chem. Rev. 1990,
90, 1105−1129.
(36) Butler, A.; Theisen, R. M. Coord. Chem. Rev. 2010, 254, 288−
296.
(37) Barbeau, K.; Rue, E. L.; Bruland, K. W.; Butler, A. Nature 2001,
413, 409−413.
(38) Amin, S. A.; Green, D. H.; Hart, M. C.; Kupper, F. C.; Sunda,
̈
W. G.; Carrano, C. J. Proc. Natl. Acad. Sci. U.S.A. 2009, 106, 17071−
17076.
(6) de Lorenzo, V.; Bindereif, A.; Paw, B. H.; Neilands, J. B. J.
Bacteriol. 1986, 165, 570−578.
(39) Kraas, F. I.; Helmetag, V.; Wittmann, M.; Strieker, M.; Marahiel,
M. A. Chem. Biol. 2010, 17, 872−880.
(7) Crosa, J. H.; Walsh, C. T. Microbiol. Mol. Biol. Rev. 2002, 66,
223−249.
(8) Challis, G. L. ChemBioChem 2005, 6, 601−611.
(9) Fischbach, M. A.; Walsh, C. T. Chem. Rev. 2006, 106, 3468−
3496.
(40) Singh, G. M.; Fortin, P. D.; Koglin, A.; Walsh, C. T. Biochemistry
2008, 47, 11310−11320.
(41) Schwecke, T.; Gottling, K.; Durek, P.; Duenas, I.; Kaufer, N. F.;
̈
̈
Zock-Emmenthal, S.; Staub, E.; Neuhof, T.; Dieckmann, R.; von
Dohren, H. ChemBioChem 2006, 7, 612−622.
̈
(10) Lautru, S.; Deeth, R. J.; Bailey, L. M.; Challis, G. L. Nature
(42) Balkwill, D. L. In Bergey’s Manual of Systematic Bacteriology, 2nd
ed. Boone, D. R., Castenholz, R. W., Garrity, G. M., Brenner, D. J.,
Krieg, N. R., Staley, J. T., Eds.; Springer: New York, 2005; Vol. 2, pp
600−604.
Chem. Biol. 2005, 1, 265−269.
(11) Dimise, E. J.; Widboom, P. F.; Bruner, S. D. Proc. Natl. Acad. Sci.
U.S.A. 2008, 105, 15311−15316.
(12) (a) Lazos, O.; Tosin, M.; Slusarczyk, A. L.; Boakes, S.; Cortes, J.;
Sidebottom, P. J.; Leadlay, P. F. Chem. Biol. 2010, 17, 160−173.
(b) Robbel, L.; Knappe, T. A.; Linne, U.; Xie, X.; Marahiel, M. A.
FEBS J. 2010, 277, 663−676.
(13) Bosello, M.; Robbel, L.; Linne, U.; Xie, X.; Marahiel, M. A. J.
Am. Chem. Soc. 2011, 133, 4587−4595.
(43) Wilde, E. Arch. Mikrobiol. 1962, 43, 109−137.
(44) Martinez, J. S.; Carter-Franklin, J. N.; Mann, E. L.; Martin, J. D.;
Haygood, M. G.; Butler, A. Proc. Natl. Acad. Sci. U.S.A. 2003, 100,
3754−3759.
(45) Poehlein, A.; Kusian, B.; Friedrich, B.; Daniel, R.; Bowien, B. J.
Bacteriol. 2011, 193, 5017.
(14) Seyedsayamdost, M. R.; Traxler, M. F.; Zheng, S.-L.; Kolter, R.;
Clardy, J. J. Am. Chem. Soc. 2011, 133, 11434−11437.
(15) Nett, M.; Ikeda, H.; Moore, B. S. Nat. Prod. Rep. 2009, 26,
1362−1384.
(46) Pfennig, N. Arch. Microbiol. 1974, 100, 197−206.
(47) Gust, B.; Challis, G. L.; Fowler, K.; Kieser, T.; Chater, K. F. Proc.
Natl. Acad. Sci. U.S.A. 2003, 100, 1541−1546.
́
(48) Menard, R.; Sansonetti, P. J.; Parsot, C. J. Bacteriol. 1993, 175,
(16) Winter, J. M.; Behnken, S.; Hertweck, C. Curr. Opin. Chem. Biol.
2011, 15, 22−31.
5899−5906.
(49) Wackler, B.; Schneider, P.; Jacobs, J. M.; Pauly, J.; Allen, C.;
Nett, M.; Hoffmeister, D. Chem. Biol. 2011, 18, 354−360.
(17) Steinbuchel, A. Macromol. Biosci. 2001, 1, 1−24.
̈
(18) Vincent, K. A.; Cracknell, J. A.; Lenz, O.; Zebger, I.; Friedrich,
B.; Armstrong, F. A. Proc. Natl. Acad. Sci. U.S.A. 2005, 102, 16951−
16954.
(19) Schwartz, E.; Henne, A.; Cramm, R.; Eitinger, T.; Friedrich, B.;
Gottschalk, G. J. Mol. Biol. 2003, 332, 369−383.
(20) Pohlmann, A.; Fricke, W. F.; Reinecke, F.; Kusian, B.; Liesegang,
H.; Cramm, R.; Eitinger, T.; Ewering, C.; Potter, M.; Schwartz, E.;
̈
Strittmatter, A.; Voß, I.; Gottschalk, G.; Steinbuchel, A.; Friedrich, B.;
̈
Bowien, B. Nat. Biotechnol. 2006, 24, 1257−1262.
(21) Price, M. N.; Huang, K. H.; Alm, E. J.; Arkin, A. P. Nucleic Acids
Res. 2005, 33, 880−892.
(22) Felnagle, E. A.; Barkei, J. J.; Park, H.; Podevels, A. M.;
McMahon, M. D.; Drott, D. W.; Thomas, M. G. Biochemistry 2010, 49,
8815−8817.
(23) Rausch, C.; Hoof, I.; Weber, T.; Wohlleben, W.; Huson, D. H.
BMC Evol. Biol. 2007, 7, 78.
(24) Stachelhaus, T.; Mootz, H. D.; Marahiel, M. A. Chem. Biol.
1999, 6, 493−505.
(25) Rottig, M.; Medema, M. H.; Blin, K.; Weber, T.; Rausch, C.;
̈
Kohlbacher, O. Nucleic Acids Res. 2011, 39, W362−W367.
(26) Milagres, A. M.; Machuca, A.; Napoleao, D. J. Microbiol. Methods
1999, 37, 1−6.
(27) Schwyn, B.; Neilands, J. B. Anal. Biochem. 1987, 160, 47−56.
(28) Kreutzer, M. F.; Kage, H.; Gebhardt, P.; Wackler, B.; Saluz, H.
P.; Hoffmeister, D.; Nett, M. Appl. Environ. Microbiol. 2011, 77, 6117−
6124.
(29) Cartron, M. L.; Maddocks, S.; Gillingham, P.; Craven, C. J.;
Andrews, S. C. BioMetals 2006, 19, 143−157.
(30) Cowart, R. E. Arch. Biochem. Biophys. 2002, 400, 273−281.
(31) Schroder, I.; Johnson, E.; de Vries, S. FEMS Microbiol. Rev.
̈
2003, 27, 427−447.
(32) Marfey, P. Carlsberg Res. Commun. 1984, 49, 591−596.
(33) Stephan, H.; Freund, S.; Meyer, J.-M.; Winkelmann, G.; Jung, G.
Liebigs Ann. Chem. 1993, 43−48.
(34) Fujii, K.; Ikai, Y.; Mayumi, T.; Oka, H.; Suzuki, M.; Harada, K.
Anal. Chem. 1997, 69, 3346−3352.
(35) Schaffner, E. M.; Hartmann, R.; Taraz, K.; Budzikiewicz, H. Z.
Naturforsch., C 1996, 51, 139−150.
5422
dx.doi.org/10.1021/ja300620z | J. Am. Chem. Soc. 2012, 134, 5415−5422