Concerted action of P450 plus helper protein to form the amino-hydroxy-piperidone moiety of the potent protease inhibitor crocapeptin

Article abstract of DOI:10.1021/ja4047153

The crocapeptins are described here as cyclic depsipeptides, isolated from cultures of the myxobacterium Chondromyces crocatus. Structure elucidation of the compounds revealed a cyanopeptolin-like skeleton, containing the characteristic amino-hydroxy-piperidone (Ahp)-heterocycle. Like the cyanopeptolins, the myxobacterial crocapeptins proved to be serine protease inhibitors. The nonribosomal origin of the peptide was confirmed by mutagenesis experiments, and the biosynthesis gene cluster was sequenced. It could be shown that the Ahp-heterocycle originates from a proline residue in the precursor molecule precrocapeptin, which is converted to crocapeptin by the tailoring enzymes CpnE and CpnF. Conversion of precrocapeptin isolated from a cpnF mutant into crocapeptin was achieved using recombinant CpnF, a cytochrome P450 enzyme responsible for hydroxylation of the proline residue in precrocapeptin. Addition of protein CpnE resulted in strongly increased conversion rates toward Ahp containing product. A mutant with 10-fold increased production of crocapeptin A was created through insertion of the Pnpt-promotor in front of the NRPS gene.

Full text of DOI:10.1021/ja4047153

Article
pubs.acs.org/JACS
Concerted Action of P450 Plus Helper Protein To Form the Amino-
hydroxy-piperidone Moiety of the Potent Protease Inhibitor
Crocapeptin
Konrad Viehrig, Frank Surup, Kirsten Harmrolfs, Rolf Jansen, Brigitte Kunze, and Rolf Mu
̈
ller*,†
†
‡
†
‡
§
†
Helmholtz Institute for Pharmaceutical Research Saarland, Helmholtz Centre for Infection Research, and Department of
cken, Germany
Pharmaceutical Biotechnology, Saarland University, Building C 2.3, D-66123, Saarbru
̈
‡
Department of Microbial Drugs, Helmholtz Centre for Infection Research, Inhoffenstrasse 7, D-38124, Braunschweig, Germany
§
Helmholtz Centre for Infection Research, Department of Microbial Communication, Inhoffenstrasse 7, D-38124, Braunschweig,
Germany
*
S Supporting Information
ABSTRACT: The crocapeptins are described here as cyclic depsipep-
tides, isolated from cultures of the myxobacterium Chondromyces crocatus.
Structure elucidation of the compounds revealed a cyanopeptolin-like
skeleton, containing the characteristic amino-hydroxy-piperidone (Ahp)-
heterocycle. Like the cyanopeptolins, the myxobacterial crocapeptins
proved to be serine protease inhibitors. The nonribosomal origin of the
peptide was confirmed by mutagenesis experiments, and the biosynthesis
gene cluster was sequenced. It could be shown that the Ahp-heterocycle
originates from a proline residue in the precursor molecule precrocapeptin, which is converted to crocapeptin by the tailoring
enzymes CpnE and CpnF. Conversion of precrocapeptin isolated from a cpnF mutant into crocapeptin was achieved using
recombinant CpnF, a cytochrome P450 enzyme responsible for hydroxylation of the proline residue in precrocapeptin. Addition
of protein CpnE resulted in strongly increased conversion rates toward Ahp containing product. A mutant with 10-fold increased
production of crocapeptin A was created through insertion of the Pnpt-promotor in front of the NRPS gene.
INTRODUCTION
sponge, they represent unique compound families. Partial
genome sequencing of strain Cm c5 revealed the presence of
more putative secondary metabolite gene clusters, which were
not yet associated with a product. Thus, further bioactive
compounds are expected to be produced by this strain under
currently unknown conditions. In a combined genome and
metabolite mining approach, noticeable UV peaks of Cm c5
extracts were investigated and a number of putative NRPS and
PKS gene clusters were inactivated, to correlate compound
■
Natural products from microorganisms continue to be
important sources for bioactive compounds with potential for
clinical, agricultural, or biotechnological use. Compounds with
antimicrobial, antiviral, cytotoxic, immunosuppressive, and
other biological activities are produced by bacteria and fungi.
Myxobacteria are an established source of secondary metabo-
1
lites, and a large variety of novel chemical structures have been
2
isolated, many of which possess novel modes of action.
9
peaks to genes and isolate novel bioactive metabolites.
Myxobacteria typically produce polyketides and nonribosomal
peptides or hybrid molecules thereof, which are assembled in
the cells by large multienzyme complexes, polyketide
The crocapeptins presented in this Article belong to the
cyanopeptolin-like compound family, cyanobacterial cyclic
depsipeptides with over 100 congeners already isolated from
synthetases (PKS), or nonribosomal peptide synthethases
10
3
various cyanobacteria. Their unique structural feature is an
amino-hydroxy-piperidone (Ahp)-heterocycle, which is crucial
for the bioactivity as potent serine protease inhibitors. Different
congeners show preference to different target enzymes, based
on the amino acids incorporated at key positions within the
(
NRPS). As in other prokaryotic producers, the genes
encoding the enzymes necessary for biosynthesis of a
compound family are usually found clustered in close vicinity
in the genome of the producing organism, often together with
genes necessary for product export, regulation, and self-
resistance.
11
peptide. Cocrystallization of the elastase with scyptolin
revealed that the compound fits into the active site of the
The myxobacterium Chondromyces crocatus Cm c5 has been
found to produce an outstanding compound spectrum,
12
enzyme and directly blocks the serine of the catalytic triad.
4
5
6
The rigidity of the compound, crucial for blocking the enzyme
activity, is provided by the Ahp-heterocycle. Promising
comprising the ajudazols, chondramids, crocacins, chondro-
7
8
chlorens, and thuggacins, which have been detected in
activity-based screenings because of their antifungal, anti-
bacterial, or cytotoxic qualities. Except for the chondramides,
which are highly similar to the jaspamides isolated from a Jaspis
Received: May 17, 2013
Published: October 8, 2013
©
2013 American Chemical Society
16885
dx.doi.org/10.1021/ja4047153 | J. Am. Chem. Soc. 2013, 135, 16885−16894

Journal of the American Chemical Society
Article
pharmaceutical potential may arise from this compound family
as highly specific inhibitors for serine proteases, obtained either
from a natural source or through modification of an existing
scaffold. Because of the absence of toxicity, the cyanopeptolins
are seen as lead structures with high potential in therapy, yet
their development is hampered by the low amounts typically
obtained from cyanobacteria. As an alternative approach, total
1
3
synthesis of some congeners has been pursued.
14
Since their first isolation, the pharmaceutical potential of
various cyanopeptolins has been investigated, because serine
proteases are involved in a large variety of diseases, such as virus
infections, chronic inflammations, and cancer. Scyptolin is a
specific elastase inhibitor, while the ichthyopeptins were
found to block influenza virus replication in infected cells and
to increase survival rates in infected mice. Intriguingly, several
cyanopeptolin and crocapeptin congeners were recently
patented for specific inhibition of human kallikreins to treat
12
15
29
chronic skin diseases such as psoriasis and atopic dermatitis.
Cyanopeptolins are produced by NRPS enzymes in
cyanobacteria, and four cyanopeptolin NRPS clusters from
different cyanobacterial strains have been sequenced and
16
published to date. Despite the numerous variants isolated
and studied, the origin and biosynthesis of the characteristic
Ahp-heterocycle have not yet been elucidated. Because genetic
tools for Cm c5 had been established previously, over-
production of the crocapeptins and the elucidation of the
Ahp formation in Chondromyces were of particular interest.
Detailed knowledge of the biosynthetic pathway and the
responsible enzymes is often crucial for overproduction and
further development of a promising compound and may
provide options to generate non-natural derivatives by means of
biotechnology. In this Article, the isolation and structure
elucidation of the crocapeptins, the identification of the
corresponding biosynthetic gene cluster, the isolation and
structure elucidation of a precrocapeptin from a P450 mutant,
and finally Ahp formation within this precursor molecule in
vitro using purified P450 plus an unprecedented helper protein
are presented.
Figure 1. Crocapeptins A1−A3 (1−3) isolated from strain Cm c5 and
crocapeptin B (4) isolated from strain Cm c2.
structures accounting for all atoms of the molecule were
detected, that is, the amino acids glutamine, threonine, leucine,
phenylalanine, N-methyltyrosine, and valine together with 3-
amino-6-hydroxy-2-piperidone (Ahp) and isobutyric acid (Ibu)
moieties. Their linear sequence in 2 was determined from
HMBC correlations as Ibu-Gln-Thr-Leu-Ahp-Phe-NMeTyr-Val
(
Supporting Information Table S1, Figure S1). The high field
shift of the oxymethine H3 of Thr (δ 5.39) indicated an ester
linkage at this position, which was confirmed by the HMBC
correlation between H3 of Thr (δ 5.39) and C1 of Val (δ
1
71.9), thus completing the planar depsipeptide structure of 2.
Comparison of the NMR data of 1, 2, and 3 revealed the
presence of different starter units: the isobutyrate moiety of 2 is
replaced by propionate in the case of 1 and by 2-methylbutyrate
in 3. The absolute configurations of L-Gln, L-Thr, L-Leu, L-Phe,
L-NMeTyr, and L-Val residues were assigned by Marfey’s
RESULTS
■
Analysis and Isolation of Crocapeptins. The crocapep-
tins 1−3 were detected in extracts of C. crocatus Cm c5 by
HPLC−UV−HR−ESI−MS. The molecular formulas of
C H N O for crocapeptin A (1), C H N O for
17
method. However, initially an oxidation with CrO3 was
performed liberating glutamate from the unusual 3-amino-6-
18
hydroxy-2-piperidone (Ahp) moiety. Because no D-glutamate
was detected, the 3S-configuration was deduced for Ahp.
Furthermore, ROESY correlations within the Ahp residue
between NH and H4a, H4a, and OH suggested a 6R-Ahp
configuration (Supporting Information Figure S3). A strong
NOE between α-protons of Phe and NMeTyr showed a cis-
amide bond between these amino acids characteristic for
structurally related compounds. Because 2-methylbutyrate
biosynthetically derives from isoleucine, an S-configuration
can be proposed for the branched-chain starter unit in 3. The
NMR spectra of 4 were very similar to those of 1−3. They
revealed the presence of isoleucine, N-methyltyrosine, phenyl-
alanine, 3-amino-6-hydroxy-2-piperidone, leucin, threonin,
citrullin (Cit), and isobutyric acid moieties. Their linear
sequence in 4 was also determined by key HMBC correlations
as Ibu-Cit-Thr-Leu-Ahp-Phe-NMeTyr-Ile (Supporting Infor-
mation Table S2, Figure S2). Using Marfey’s method, L-
configurations were determined for all amino acids. The Ahp-
configuration of 4 was deduced as 3S,6R similarly to 2.
Database searches resulted in one match for 4, which refers to a
4
7
66
8
12
1
48 68
8
12
crocapeptin A (2), and C H N O for crocapeptin A (3)
2
49 70
8
12
3
+
were calculated from the [M + H] ion clusters m/z 935.4868,
49.5027, and 963.5186. Thus, 1 and 3 differ from 2 by the
9
formal abstraction and addition of a methyl group, respectively.
The molecular formula of crocapeptin B (4), which is
produced by C. crocatus Cm c2, was determined as
C H N O by [M + H]+ ion cluster at m/z 992.5450.
50
73
9
12
Additionally, in the extract of strain Cm c2, minor compounds
with molecular formulas C H N O and C H N O were
4
9
71
9
12
51 75
9
12
observed.
Structure Elucidation of Crocapeptins. The basic
structure of the crocapeptins (Figure 1) was primarily
elucidated with main component 2. Using HMQC and
DEPT NMR spectra, all protons were assigned to their
respective carbon atoms. The remaining 8 protons were
recognized from their chemical shifts as 7 amide protons and
1
3
one hydroxyl proton. The C NMR spectrum revealed nine
carbon atoms within the amide/ester range. Following the
interpretation of COSY, TOCSY, and HMBC data, eight partial
1
6886
dx.doi.org/10.1021/ja4047153 | J. Am. Chem. Soc. 2013, 135, 16885−16894

Journal of the American Chemical Society
Article
Figure 2. The crocapeptin A biosynthetic gene cluster from Chondromyces crocatus Cm c5. The genes responsible for biosynthesis are indicated as
dark blue arrows. The closest homologue found by BLAST sequence search against the nr database of the encoded protein sequence is shown below
for each ORF.
Figure 3. Biosynthesis of crocapeptin A2. The first condensation reaction defines the crocapeptin A derivative. The cyclic depsipeptide is released as
precrocapeptin, which is then hydroxylated at the proline residue by the cytochrome P450 enzyme CpnF. Rearrangement of hydroxy-proline to Ahp
is catalyzed by CpnE.
member of a depsipeptide patented because of submicromolar
biosynthetic gene cluster. Using prediction of substrate
specificity of adenylation domains from sequence data
2
9,30
19
Kallikrein 7 inhibitory activity,
metabolites.
whereas 1−3 comprise new
allowed one to reduce the number of target sequences to two
contigs, containing an NRPS gene fragment with expected
adenylation domain specificity for Thr-Leu-Pro-Phe and Tyr.
Inactivation mutants for each contig were found unable to
produce crocapeptin. Thus, both identified NRPS gene
fragments were proven to be part of the crocapeptin
biosynthetic gene cluster. To complete the cluster sequence,
Protease Inhibition Assays. Ahp-containing cyclic dep-
sipeptides frequently possess serine protease inhibitory activity;
the amino acids adjunct to Ahp determine the enzyme
1
2
specificity. With Phe and Leu in these positions, the
crocapeptins belong to the class of chymotrypsin inhibitors.
Consequently, their inhibitory activity against Chymotrypsin
was determined and resulted in IC50 values of 0.1 μM for 1−3
and 0.2 μM for 4.
20
a cosmid gene bank covering the Cm c5 genome was
screened with DNA probes directed against parts of the two
crocapeptin NRPS gene fragments, which are present on
cosmids B:E7, C:O19, and D:J18. Cosmid end-sequencing,
primer walking, and inverse PCR were applied to complete the
gene cluster sequence. In Figure 2, a schematic drawing of the
gene cluster is shown, with the closest homologues to the
Analysis of the Crocapeptin A Gene Cluster and
Proposed Biosynthesis. Because of the apparent similarity to
the nonribosomally synthesized cyanopeptolins, partial genome
data of crocapeptin A producing strain Cm c5 were screened
for the presence of NRPS genes to identify the corresponding
1
6887
dx.doi.org/10.1021/ja4047153 | J. Am. Chem. Soc. 2013, 135, 16885−16894

Journal of the American Chemical Society
Article
Figure 4. (A) Effect of gene inactivation in Cm c5 by insertional mutagenesis. For consistency, relative crocapeptin A content of extracts from Cm c5
wild type and mutants is given relative to the chondramide content in extracts of cells and growth medium. Crocapeptin A content was related to
chondramid B content in each extract by division of the integrated MS-signal intensities of crocapeptins A by chondramid intensities, respectively.
(
B) Comparison of crocapeptin production in Cm c5 wild type and Pnpt-cpnD mutant cells. Extracts of cells and medium showed a 10-fold increased
crocapeptin content in the mutant.
+
+
+
identified ORFs listed below. The NRPS gene cpnD is 24.5 kb
in size and forms an operon with the hypothetical protein gene
cpnE and cytochrome P450 encoding gene cpnF. Upstream of
the NRPS gene, a putative serine protease gene cpnA and two
ABC transporter genes cpnB and cpnC are located. Downstream
of cpnF, a protein kinase is encoded by gene cpnG on the
opposite strand. Because of their arrangement and predicted
gene product function, ORFs cpnA to cpnG were chosen for
knockout experiments to test their involvement in crocapeptin
biosynthesis, transport, or regulation.
by HR−MS. [M + H] , [M + NH ] , and [M − H O + H]
4
2
molecular ion clusters of crocapeptins A −A exhibiting a m/z
1
3
that increased by 6 units were clearly detected, however only at
a low percentage, making quantification hardly possible. In the
negative control experiment, these ions were not detected.
Thus, proline is indeed incorporated into the crocapeptins, and
all proline N and C atoms remain in the metabolite after
conversion of this residue to Ahp.
Gene Inactivation Studies in Cm c5. To investigate the
function of the genes identified, targeted insertional muta-
genesis was applied. Mutant Cm c5 clones exhibiting
hygromycin B resistance were grown in liquid culture for
isolation of genomic DNA, and insertion was verified by PCR
experiments (for genetic verification, see Supporting Informa-
tion section 3). At least three independent verified mutants of
genes cpnA, cpnB, cpnC, cpnD, and cpnF were obtained and
grown in liquid culture to enable secondary metabolite analysis.
For cpnE and cpnG, no mutants could be obtained. As two of
the genes that were inactivated are predicted to encode ABC-
transporters, mutant cells and culture supernatant were
extracted separately, to observe differences in crocapeptin
content between cells and growth medium. Quantification of
crocapeptin production was performed by dividing the
The catalytic domain sequence of CpnD and the substrate
specificity of its adenylation domains were obtained from the
prediction based on the amino acid sequence, according to the
1
9
NRPS predictor. As depicted in Figure 3, biosynthesis is
expected to start with incorporation of glutamine by domain
A1, to which one of the observed short chain fatty acids is
bound by domain C1, creating the crocapeptins A −A . The
1
3
lipopeptide chain is then elongated by incorporation of
threonine, leucine, proline, phenylalanine, tyrosine (which
becomes N-methylated in trans), and valine. The TE domain
finally releases precrocapeptin by intramolecular lactonization
via the hydroxyl group of threonine and the carboxy-terminus
of valine. Free precrocapeptin is hydroxylated by CpnF at the
proline residue, which is then transformed to the bioactive Ahp-
containing crocapeptin. This biosynthesis scheme implies that
each crocapeptin derivative must have its distinct precursor, as
the variable short fatty acid chain is introduced during the start
of biosynthesis. No function could be assigned to the gene
product of cpnE, which has high sequence similarity to ORFs
annotated as “hypothetical proteins” in various prokaryotic
genomes. The closest sequence homologue with an assigned
protein function is an Aha1 activator protein of the Hsp90
+
summarized [M + H] signal peak areas from MS analysis
+
regarding crocapeptins A −A by the chondramide B [M+H]
signal peak area in each chromatogram, because chondramid B
production was known to be very stable in Cm c5 by
experience. UV data were not used because of overlapping
peaks with other compounds. The average ratio of crocapeptins
to chondramide was normalized to wild-type levels and is
shown in Figure 4A.
As mentioned above, inactivation of the NRPS gene cpnD
was found to abolish crocapeptin production. Inactivation of
the putative serine protease encoding gene cpnA led to a 2-fold
increase of crocapeptin levels in extracts from cells and
medium, but high variation was observed within the individual
cultures. Inactivation of the ABC transporter genes cpnB and
cpnC led to cell extracts devoid of crocapeptins, while these
metabolites were still present in the medium in similar amounts
as compared to the wild type. Inactivation of cpnF did not
abolish crocapeptin production completely, but reduced
1
3
23
2
1
ATPase. The function of CnpE in crocapeptin biosynthesis is
described below. The adenylation domain A4 of the
crocapeptin assembly line, responsible for incorporation of
the amino acid that would be converted to Ahp, is predicted to
incorporate proline, as deduced from the motif “DVQFAAH-
22
VAK” by using the nonribosomal code. To verify this
prediction, 13C5 N-proline was amended to Cm c5 wild-type
liquid culture, and cultivation was continued for 2 days. After
extraction of cells and supernatant, crocapeptins were detected
15
1
6888
dx.doi.org/10.1021/ja4047153 | J. Am. Chem. Soc. 2013, 135, 16885−16894

Journal of the American Chemical Society
Article
Figure 5. Detection and structure elucidation of precrocapeptin A (5). Extracted ion chromatogram for crocapeptin masses (red areas) and
2
−
precrocapeptin masses (blue areas). The putative crocapeptin precursors were found in extracts of cpnF mutants of strain Cm c5. Subsequent
isolation and NMR analysis of 5 confirmed the structure shown on the right, a cyclic depsipeptide exhibiting L-proline incorporated instead of Ahp.
Figure 6. (A) SDS-PAGE of purified proteins CpnE and CpnF. The target proteins are indicated by black arrows. (B) CO difference spectrum of
purified CpnF. The peak characteristic for cytochrome P450 enzymes at 450 nm can be observed. (C) Difference spectra of CpnF with increasing
concentrations of precrocapeptin A show type I binding.
2
+
abundance to less than one-half of wild-type levels in the
respective mutants. The effects of the inactivation experiments
are discussed below.
Indeed, new [M + H] mass signals were detected in the
extracts of the growth medium from cpnF− mutants. These
masses were later found to be detectable in the supernatant of
wild-type cultures, but only in trace amounts. Because of their
retention times and m/z values, the new compounds were
designated precrocapeptins pA m/z = 919.49, pA m/z =
Overproduction of Crocapeptins in Cm c5 by
Promotor Insertion. The tn5-derived npt-promotor Pnpt
had been shown to enable overexpression of genes for
secondary metabolite production of strain Cm c5 previously,
where replacing the native thuggacin promotor by Pnpt led to
1
2
933.52, and pA m/z = 947.52. To verify the identity of the
newly observed mass signals, Cm c5 cpnF was grown in liquid
culture, and the putative precursors were isolated from culture
supernatant. The purified precrocapeptin A (pA ) with a mass
3
−
24
significantly increased levels of thuggacins. To test whether a
similar effect could be induced with crocapeptins, the promotor
was placed in front of the NRPS gene cpnD by insertional
mutagenesis, leading to a disrupted short copy of cpnD and a
Pnpt-driven intact cpnD in the mutant. Extracts of mutant cells
and medium were prepared and analyzed as described
previously, and crocapeptin content was measured as described
above. An increase of approximately 10-fold crocapeptin
content was found in the cells and 15-fold in the medium, as
shown in Figure 4B. As compared to the wild type, the
overproducer mutants did not show altered growth or
swarming behavior in liquid culture or on agar plate.
2
2
of 932.3 Da was analyzed by NMR, which confirmed the
expected lipopeptide lactone, corresponding to crocapeptin A2
with L-proline replacing the Ahp heterocycle (Figure 5).
Establishing the Role of CpnE and CpnF in the
Conversion of Precrocapeptin to Crocapeptin in Vitro.
To elucidate the reaction mechanism and the role of the gene
product of cpnE and cpnF, corresponding proteins CpnE and
CpnF were overexpressed in E. coli and purified, as shown in
Figure 6A. CO-difference spectrometry was applied to verify
that CpnF is expressed as active Cytochrome P450, but
revealed that only a fraction of the purified CpnF is active
holoenzyme, while a larger fraction is apoenzyme, as indicated
by the peak at 420 nm (Figure 6B). Difference spectrometry of
−
Detection of Precrocapeptins in cpnF Mutants of Cm
c5. The proposed crocapeptin biosynthetis scheme implies that
the Ahp-heterocycle is derived from a proline residue, which is
converted to Ahp either in trans or after release of a
depsipeptide precursor molecules by the tailoring cytochrome
P450 oxidase CpnF. These crocapeptin precursors would be
expected to have a L-proline residue instead of Ahp in the
CpnF coincubated with precrocapeptin A led to concen-
tration-dependent type I binding, as indicated by the reduction
of the Soret absorption band at 420 nm, as shown in Figure 6C.
2
This confirms that that free pA is indeed a substrate for CpnF.
2
−
respective position. Extracts of Cm c5 cpnF mutants, in which
Cytochrome P450 enzymes catalyze electron-dependent
redox reactions, typically hydroxylations and epoxidations,
and thus depend on electron carrier proteins to deliver the
the tailoring P450 was inactivated, were therefore screened for
masses that would correspond to the postulated precursors.
1
6889
dx.doi.org/10.1021/ja4047153 | J. Am. Chem. Soc. 2013, 135, 16885−16894

Journal of the American Chemical Society
Article
necessary electrons. Because the corresponding native electron
carriers for CpnF are not known, the versatile recombinant
ferredoxin Adx 4-108 was utilized instead, as it had been shown
ually. From this analysis, formation of a 5-OH proline moiety is
strongly indicated, as discussed below. Using the same type of
analysis, also nonenzymatic conversion from 5-OH-proline to
Ahp could be proven, as shown in Supporting Information
Figure S7-4.
2
5,26
to be able to interact with several myxobacterial CYPs.
To
transfer electrons to CpnF, Adx was used together with its
native electron donor, adrenodoxin reductase, a flavoenzyme
that reduces oxidized adrenodoxin by oxidizing NADPH to
DISCUSSION
■
+
NADP . To keep NADPH levels elevated during the reaction, a
The Myxobacterium Chondromyces crocatus Produces
Potent Protease Inhibitors: The Crocapeptins. The
crocapeptins, isolated from the myxobacterium Chondromyces
crocatus, were found to belong to the diverse compound class of
cyanopeptolin-like peptides, and mutagenesis experiments in
strain Cm c5 confirmed their nonribosomal origin. The Ahp-
heterocycle is a structural feature crucial for protease inhibiting
bioactivity of the cyanopeptolins. Here, the previously
unknown biosynthesis of this structural part was elucidated
by mutagenesis and biochemical studies. These investigations
led to the identification of the precrocapeptins as precursor
molecules released from the NRPS assembly line, featuring an
L-proline instead of the Ahp heterocycle. These precursors are
then converted into the bioactive crocapeptins by the
hydroxylation via the tailoring P450 CpnF and the Aha1-
domain containing helper protein CpnE. Comparison of the
crocapeptin gene cluster with published cyanopeptolin gene
clusters from cyanobacteria shows that the cyanobacterial gene
clusters contain neither a predicted P450 gene nor a gene with
homology to cpnE. Furthermore, in the cyanopeptolin NRPS
genes, the predicted substrate for the A-domain corresponding
to the Ahp-position is not proline. Instead, all cyanopeptolin A-
domains at the Ahp-position have the same DVENAGVV-
motif, of which the substrate is not known yet. Consequently,
further studies are required to unravel the biosynthesis of the
Ahp herterocycle in cyanobacteria.
NADPH regeneration system consisting of glucose-6-phos-
phate and glucose-6-phosphate dehydrogenase was used, which
+
restores NADPH from NADP by conversion of glucose-6-
27
phosphate to 6-phosphoglucono-δ-lactone. Purified CpnE
was also included in this assay, to test its involvement in this
conversion reaction. Enzymes CpnE and CpnF were tested
separately and in combination in three separate reactions A, B,
and C, each containing precrocapeptin A as substrate, as
2
shown in Figure 7. Reaction A contained only CpnE, B
Figure 7. In vitro conversion of precrocapeptin A2 (peak 5) to
crocapeptin A (peak 2). The result of incubation of precrocapeptin
2
The function of the ORFs in the crocapeptin gene cluster
was investigated by knockout experiments. The putative serine
protease gene cpnA is considered to be part of the gene cluster,
possibly involved in regulatory mechanisms or otherwise linked
to crocapeptin production, because of elevated crocapeptin
content in the respective mutant. The results of cpnB and cpnC
inactivation are contrary to the initially expected function as
crocapeptin export proteins and point toward a function as
importer proteins. Another possibility is a rather complex
export process involving the bacterial secretory pathway, which
cannot be investigated by simple separation of cell clumps and
media. We suggest that crocapeptins are actively secreted into
the extracellular slime of the Chondromyces biofilm by CpnB
and CpnC to serve as feeding deterrent, whereas crocapeptin
dissolves in the medium by diffusion. CpnF mutants are still
able to produce crocapeptins, so either the disrupted cpnF gene
is still translated into a protein with residual activity, or another
P450 in the genome is complementing its function.
CpnE Assists in Ahp Formation after Hydroxylation
by Cytochrome P450 Enzyme CpnF. Because of the small
size of the gene and its location, inactivation of cpnE via
insertional mutagenesis was not possible, and therefore the
function of a hypothetical CpnE could not be elucidated by
mutagenesis. A cpnE mutant is expected to exhibit a phenotype
similar to a cpnF-mutant, because transcription of cpnF would
be impaired due to the pSUP_hyg insertion. However, because
cpnE was found encoded between cpnD and cpnF, it was
supposed to play a role in crocapeptin biosynthesis. To shed
light on this hypothesis, cpnE and cpnF were heterologously
expressed in E. coli and purified.
A2 (933.5 Da) with different enzyme combinations A, B, and C is
compared to a reference of crocapeptin A2 (949.5 Da). Extracted ion
chromatograms for the masses 933.5 Da (blue) and 949.5 Da (red) are
shown. Peak 6 is a crocapeptin isomer that forms mostly in the
absence of CpnE.
contained only CpnF, and in C both proteins were present.
After incubation for 30 min, the extracts were subjected to MS
analysis and compared to a reference sample of crocapeptin A2
(
2). Unaltered substrate precrocapeptin A was detected in
2
reaction A. Complete hydroxylation of precrocapeptin A2 took
place in reactions B and C where two isomeric products were
detected with m/z 949.5 by ESI MS with retention times of 7.2
and 7.8 min. Comparison with the reference confirmed the
identity of the peak at 7.8 min as crocapeptin A . The role of
2
CpnE in the conversion is shown by comparing the ratio of
crocapeptin A (peak 2) and the presumed intermediate (peak
2
6
) in reactions B and C. Without CpnE, mainly the isomer with
a retention time of 7.2 min was formed, while in the presence of
CpnE a much more efficient conversion into crocapeptin A2
was observed. Nevertheless, in the course of several days, a slow
rearrangement of the supposed intermediate into crocapeptin
A2 could also be observed.
Using pA as substrate for in vitro hydroxylation leads to the
3
same effects, that is, one intermediate peak corresponding to
peak 6 in Figure 7 and one product peak. For corroboration of
the identity of the rather instable intermediate, the in vitro
hydroxylation reaction was repeated in a larger scale, using 400
μg of pA as substrate. The reaction products were separated
3
chromatographically (HPLC) and analyzed by NMR individ-
1
6890
dx.doi.org/10.1021/ja4047153 | J. Am. Chem. Soc. 2013, 135, 16885−16894

Journal of the American Chemical Society
Article
Scheme 1. Proposed Formation of the Amino-hydroxy-piperidone (Ahp) Heterocycle
Ahp formation is expected to occur as shown in Scheme 1.
Hydroxylation at C5 of the proline residue in precrocapeptins
gives rise to a cyclic hemiaminal corresponding to peak 6 in
Figure 7. This hydroxy-proline derivative is then converted to
Ahp via a glutamate-semialdehyde intermediate, which is highly
reactive and forms the more stable six-membered heterocycle
via the amide bond of the adjacent phenylalanine residue. From
Figure 8. 1H NMR shifts in cyclic hemiaminals after in vitro
synthetic efforts of the cyanobacterial congener symplocamide,
spontaneous Ahp formation has been reported starting from a
hydroxylation of proline in precrocapeptin A : Ahp-containing
3
13
L-glutamic acid semialdehyde residue, which can be obtained
by hydroxylation of proline at C5 and subsequent ring-opening.
Indeed, a hydroxylated precrocapeptin with either hydroxypro-
line or glutamic-acid-semialdehyde does exhibit a mass identical
to that of the respective Ahp-containing crocapeptin.
Hydroxylation of proline at positions other than C5 is not
expected, as this would not lead to a hemiaminal able to
undergo ring-opening. Proline residues with hydroxylations at
C3 and C4 have been described for the lantibiotic micro-
biosporin, and hydroxylation at these positions does not lead to
crocapeptin A (right) and its unstable isomeric intermediate
3
containing 5-OH proline instead (left).
and δ 2.44, 2.05 (Supporting Information Figure S7-3).
Correlation to the CHOH group at δ 5.30 cannot be detected
with the applied parameters, but integrity of the spin system
can be assigned via correlations of the signal of CHOH with
both CH groups and the α-proton. A strong indication for the
2
presence of the hydroxy-proline is the missing correlation to an
NH proton in this spin system in addition to the existence of
the NH-signal belonging to the phenylalanine spinsystem at δ
2
8
Ahp formation.
Activity of Purified Recombinant Proteins. CpnE and
CpnF were tested with isolated precursor precrocapeptin A2
8
.86. Furthermore, the shifts of N-Me-Tyr aromatic protons in
the intermediate (δ 7.09, 6.50) are exactly the same as for
precrocapeptin (Supporting Information Table S4-4), whereas
these proton signals in the product are significantly shifted to δ
(5) as substrate. Hydroxylation was achieved in vitro with
CpnF, as shown by HPLC−MS analyses of the reaction
extracts, as shown in Figure 7. Incubation with CpnF alone
mostly leads to an unstable isomer (peak 6) exhibiting shorter
7
.14 and 6.85, indicating a conformational change of the
structure probably due to rearrangement from a five- to a six-
membered ring. Thus, Ahp formation is likely initiated by
opening of the five-membered ring hemiaminal to glutamic-acid
semialdehyde (Scheme 1). Efforts to indirectly detect the
glutamic-acid semialdehyde intermediate were carried out by
incubating the reaction mixture with an excess of ethanolamine
to capture the aldehyde. A fraction of the short-lived
semialdehyde would then favor nucleophilic addition to the
primary amino group of the ethanolamine over the adjacent
peptide bond. A mass signal corresponding to the imine form of
the ethanolamine adduct could indeed be detected, but is also
formed when pure crocapeptin is treated with ethanolamine.
We therefore propose that the crocapeptins exist in equilibrium
between the six-membered Ahp and the five-membered 5-OH-
proline heterocycles, with Ahp being the more preferred state.
Incubation of precrocapeptin with ethanolamine did not result
in any detectable products. Details for the aldehyde trapping
experiments are provided in Supporting Information section S6.
Because the glutamic acid intermediate in the crocapeptin
conversion is highly reactive, a function of CpnE, which is
shown here to significantly enhance Ahp formation, could be to
promote formation of Ahp and shield the reactive glutamic-acid
semilaldehyde intermediate from unfavorable cross-reactions.
retention time, which slowly converts to crocapeptin A (peak
2
2
). Incubation of precrocapeptin A with CpnE alone does not
2
indicate any reaction. However, coincubation of precrocapeptin
with CpnF and CpnE results mostly in crocapeptin formation,
while formation of the unstable isomer (peak 6) is detected in
much smaller quantities.
In vitro reactions of CpnF with precrocapeptin A3 as
substrate were scaled up to isolate the presumed intermediate
5
-hydroxy-proline form of crocapeptin A3 for structure
elucidation. First, the mixture of both isomers was used for
NMR spectroscopic analysis; afterward, the isomers were
separated and analyzed individually (Supporting Information
Figure S7-1). The NMR spectrum of isomeric crocapeptin A3
1
clearly shows shifted H signals in the area of interest as
compared to crocapeptin A (Ahp-ring, Figure 8). A chemical
3
shift of δ 3.81 can be detected for the α-proton in Ahp. The
signal of the α-proton exhibits TOCSY correlations with two
CH -groups at δ 2.61, 1.69 and δ 1.81, 1.68, respectively,
2
additional to correlating with a CHOH-group at δ 5.19 and a
NH at δ 7.58 (Supporting Information Figure S7-2). In
contrast, the α-proton of the isomer shows a chemical shift of δ
4
.30 with TOCSY correlations to CH groups at δ 1.71, 1.61
2
1
6891
dx.doi.org/10.1021/ja4047153 | J. Am. Chem. Soc. 2013, 135, 16885−16894

Journal of the American Chemical Society
Article
Whether CpnE interacts directly with CpnF or with the free
substrate remains to be elucidated.
Isolation and Purification of Precrocapeptins. Cm c5 mutant
cpnF was grown in shake flasks with Pol 0.3 medium containing Hyg.
−
Medium was replaced regularly, and used medium was extracted with
Amberlite XAD-16 adsorber resin under sterile conditions. Pooled
XAD batches were extracted with ethyl acetate, dried in vacuo, and
dissolved in methanol. The crude extract was separated on a Sephadex
DAE20 column (80 cm length, 2.5 cm diameter) using methanol, and
precrocapeptin containing fractions were identified by LC−MS.
Precrocapeptin was purified by RP-HPLC [HPLC system (Dionex)
Famos autosampler, P680 pump, TCC100 thermostat, and PDA100
detector; Phenomenex Luna C18 column, 250 × 4.6 mm, 5 μm dp;
The ability of CpnF to receive electrons from the broad-
range ferredoxin Adx4-108 also disclosed a route for
biotechnological production of Ahp-containing molecules for
targeted protease inhibitor design. From the cyanobacterial
cyanopeptolins, a large variety of structures with different target
specificities is known, depending mostly on the amino acids
incorporated at the positions adjacent to Ahp. Further studies
should investigate the substrate range of CpnF and CpnE to
investigate whether this enzyme pair could be used to generate
Ahp-containing peptides of a specific and even non-natural
structure, to create highly specific/selective protease inhibitors.
For heterologous expression of the compounds, coexpression of
Adx4-108 in the same host might be an option to obtain a
functional cytochrome in a host that naturally has no
compatible ferredoxin for CpnF.
11
solvent A = H O, solvent B = ACN, gradient from 20% B to 60% B in
2
3
0 min; flow rate 2.5 mL/min; temp 30 °C; UV detection 254 and 280
nm]. The sample was injected by μL-pick-up technology with a water/
methanol (50:50 v/v) mixture as supporting solvent. Up to 50 μL of
sample was injected for manual fraction collection. For precrocapeptin
A2 NMR data, see Supporting Information Table S3. Fractions
containing 0.3 mg of precrocapeptin A1, 0.9 mg of precrocapeptin A2,
and 0.3 mg of precrpcapeptin A were obtained.
3
General Experimental Procedures. Optical rotations were
determined with a Perkin-Elmer 241 instrument; UV spectra were
recorded with a Shimadzu UV−vis spectrophotometer UV-2450; IR
spectra were measured with a Spectrum 100 FTIR spectrometer
EXPERIMENTAL PROCEDURES
■
Bacterial Strains and Growth Conditions. C. crocatus Cm c5
was cultivated in 70 L of Pol 0.3 medium (Probion 3 g/L, soluble
starch 3 g/L, MgSO × 7 H O 2 g/L, CaCl × 2 H O 0.5 g/L) in a
1
(Perkin-Elmer); NMR spectra were recorded with Bruker AM 300 ( H
300 MHz, 13C 75 MHz), Bruker ARX 600 ( H 600 MHz, C 150
1
13
4
2
2
2
1
biofermenter Biostat UE100 (Braun Melsungen, Melsungen, D) at 30
C for 11 days; C. crocatus Cm c2 was cultivated in 100 L of pol
medium in a biofermenter P150 (Bioenineering AG, Wald, CH) at 30
C for 10 days. After 3 days, 1% Amberlite XAD-16 was added to the
fermentation broths. For metabolite analysis of Cm c5 mutants, 1 mL
of cell clumps was used to inoculate 20 mL of Pol 0.3 medium with 50
mg/L Hyg. At least three verified clones from each inactivation
construct were used. Cm c5 wild-type cells were cultivated without
Hyg. After 4 days, cell clumps and medium supernatant were
separated. Medium supernatant was extracted with 25 mL of ethyl
acetate, and cell clumps were extracted with 10 mL of ethyl acetate and
MHz), and Bruker Ascend 700 with a 5 mm TXI cryoprobe ( H 700
MHz, 13C 175 MHz) spectrometers; HRESIMS mass spectra were
obtained with an Agilent 1200 series HPLC−UV system combined
with an ESI−TOF−MS (Maxis, Bruker) [column 2.1 × 50 mm, 1.7
°
°
μm, C18 Acquity UPLC BEH (Waters), solvent A, H O + 0.1% formic
2
acid; solvent B, AcCN + 0.1% formic acid, gradient, 5% B for 0.5 min
increasing to 100% B in 19.5 min, maintaining 100% B for 5 min, FR =
−1
0.6 mL min , UV detection 200−600 nm].
Crocapeptin A (1). Colorless, amorphous powder, [α] −34° (c
1
D
0.5, MeOH); UV (MeOH) λ (log ε) 230 (sh), 277 nm (3.49); IR
max
(neat) 3380.9, 2962,2, 2934.1, 1736.0, 1655.4, 1534.4, 1518.1, 1445.4,
−
1
1
13
4
0 mL of methanol. Extracts were dried in vacuo and dissolved in 1
mL of methanol.
1384.5 cm ; H NMR, C NMR data (DMSO-d ), see Supporting
6
+
Information Table S3; HR−ESI−MS m/z 935.4868 [M + H] (calcd
Isolation and Purification of Crocapeptins. XAD-16 and cell
mass of C. crocatus Cm c5 were harvested by centrifugation. Cells were
separated from XAD-16 by flotation and discarded. The XAD was
washed with water and 50% methanol, and subsequently eluted by
methanol and acetone. All fractions were analyzed by HPLC−UV−MS
for the presence of target compounds. Consequently, the methanol
extract was subjected to solvent partition using 85% aqueous
methanol, which was extracted twice with heptane. Subsequently,
the aqueous methanol was adjusted to 70% methanol and extracted
twice with dichloromethane. The dichloromethane fraction containing
the main amount of 1−3 was fractionated by RP MPLC [column 480
for C H N O , 935.4873).
47
67
8
12
Crocapeptin A (2). Colorless, amorphous powder, [α] −31° (c
2
D
0.5, MeOH); UV (MeOH) λ (log ε) 229 (sh), 277 nm (3.45); IR
max
(neat) 3401.0, 2961,5, 2931.4, 1735.0, 1659.7, 1540.4, 1517.4, 1453.1,
−1
1
13
1384.5 cm ; H NMR, C NMR, COSY, HMBC, ROESY data
(DMSO-d ), see Supporting Information Table S1; HR−ESI−MS m/z
6
+
949.5032 [M + H] (calcd for C H N O , 949.5027).
48
69
8
12
Crocapeptin A (3). Colorless, amorphous powder, [α] −25° (c
3
D
0.3, MeOH); UV (MeOH) λ (log ε) 228 (sh), 275 nm (3.43); IR
max
(neat) 3376.6, 2961.7, 2932.6, 1735.4, 1663.7, 1535.1, 1518.2, 1452.3,
−
1
1
13
1384.4 cm ; H NMR, C NMR data (DMSO-d ), see Supporting
6
+
×
6
30 mm, ODS/AQ C18 (Kronlab), gradient 30−100% methanol in
0 min, flow 30 mL/min, UV peak detection at 210 nm]. Fractions
Information Table S3; HR−ESI−MS m/z 963.5186 [M + H] (calcd
for C H N O , 963.5186).
49
71
8
12
containing 1−3 were combined and further purified by preparative RP
HPLC (column 250 × 21 mm, VP Nucleodur 100-10 C18 EC,
gradient 40−80% methanol in 25 min, 50 mM sodium acetate, flow 20
mL/min). Finally, preparative RP HPLC (column 250 × 21 mm, VP
Nucleodur C18 Gravity 5 μm, gradient 45−60% methanol in 25 min,
Crocapeptin B (4). Colorless, amorphous powder, [α]D −44° (c
0.5, MeOH); UV (MeOH) λmax (log ε) 229 (sh), 276 nm (3.68); IR
(neat) 3377.0, 2964,1, 2933.8, 1735.2, 1648.2, 1536.1, 1518.1, 1452.4,
−1
1
13
1384.4 cm ; H NMR, C NMR, COSY, HMBC, ROESY data
(DMSO-d ), see Supporting Information Table S2; HR−ESI−MS m/z
6
+
0
.2% acetic acid, flow 20 mL/min) provided 6.2 mg of 1, 8.9 mg of 2,
992.5450 [M + H] (calcd for C H N O , 992.5451), 990.5305 [M
5
0
74
9
12
−
and 6.7 mg of 3.
− H] (calcd for C H N O , 990.5306).
50
72
9
12
Precrocapeptin A (5). H NMR, 13C NMR, COSY, HMBC data
1
XAD-16 and cells of C. crocatus Cm c2 were washed stepwise with
water (2 L) and 30% methanol (4 L), and extracted with methanol (8
L). The crude methanol extract was evaporated and partitioned three
times between ethyl acetate and water. After evaporation of the ethyl
acetate, the residue was dissolved in 85% aqueous methanol and
extracted twice with heptane. The methanol extract was fractionated in
two batches by RP MPLC (column 480 × 30 mm, ODS/AQ C18
2
(MeOH-d ), see Supporting Information Table S4; HR−ESI−MS m/z
4
+
933.5070 [M + H] (calcd for C H N O , 933.508).
48
69
8
11
Amino Acid Stereochemistry Determination. Crocapeptin
A1−A3 and B were hydrolyzed in 6 N HCl at 90 °C for 17 h,
conditions that resulted in the conversion of Gln to Glu. The
hydrolysate was evaporated to dryness and dissolved in H O (100 μL).
2
(
Kronlab), gradient 37−100% methanol in 60 min, flow 30 mL/min,
1 N NaHCO3 (20 μL) and 1% 1-fluoro-2,4-dinitrophenyl-5-L-
alaninamide (100 μL in acetone) were added, and the mixture was
heated at 40 °C for 40 min. After being cooled to room temperature,
the solutions were neutralized with 2 N HCl (20 μL) and evaporated
to dryness. The residues were dissolved in MeOH and analyzed by
UV detection 210 nm). The fraction containing 4 was separated by
preparative RP HPLC (column 250 × 21 mm, VP Nucleodur C18
Gravity 5 μm, gradient 45−60% methanol/0.2% acetic acid in 25 min,
flow 20 mL/min) providing 10.8 mg of 4.
1
6892
dx.doi.org/10.1021/ja4047153 | J. Am. Chem. Soc. 2013, 135, 16885−16894

Journal of the American Chemical Society
Article
LC−MS. Retention times in minutes of FDAA-derivatized amino acids
were Glu 6.6; Thr 5.5; Leu 8.5; Phe 8.4; NMeTyr 5.8; Val 7.6 for 2
and Cit 5.5; Thr 5.5; Leu 8.5; Glu 6.6; Phe 8.4; NMeTyr 5.8; Ile 8.3
for 4. Retention times of the FDAA-derivatized authentic standards
and used as substrate for purified CpnF. As electron transporters,
recombinant bovine Adx 4-108 and adrenodoxin reductase were used.
In vitro conversion was performed in 10 mM potassium phosphate
buffer pH 7.4 or 50 mM Tris/HCl buffer at pH 8.0 in 0.5 mL volumes.
Substrate concentration was 100 μM, and enzyme concentration was
0.25 μM for CpnF and CpnE, 2 μM for AdR, and 20 μM for Adx. To
−
were L-Val 7.5, D-Val 8.5 m/z 368.123 [M − H] ; L-Thr 5.4, D-Thr 6.2
−
m/z 370.100 [M − H] ; L-allo-Thr 5.3, D-allo-Thr 5.7 m/z 370.100
−
−
[
M − H] ; L-NMeTyr 5.8, D-NMeTyr 8.4 m/z 446.131 [M − H] ; L-
regenerate NADPH, 5 mM glucose-6-phosphate, 1 mM MgCl , and 2
2
−
Phe 8.5, D-Phe 9.2 m/z 416.120 [M − H] ; L-Glu 6.7, D-Glu 7.4 m/z
u/mL glucose-6-phosphate dehydrogenase were in the reaction
mixture. All ingredients were mixed on ice; the reaction was started
by adding 200 μM NADPH and incubated at 30 °C for 30 min. The
reaction was stopped by adding 500 μL of ethyl acetate and strong
mixing. After brief centrifugation, 50 μL of the organic phase was
subjected to ESI−MS for detection of the reaction products.
−
−
4
8
3
12.110 [M − H] ; L-Leu 8.5, D-Leu 9.4 m/z 382.138 [M − H] ; L-Ile
−
.3, D-Ile 9.3 m/z 382.138 [M − H] ; L-allo-Ile 8.4, D-allo-Ile 9.4 m/z
−
−
18
82.138 [M − H] ; L-Cit 5.5, D-Cit 5.6 m/z 426.138 [M − H] .
Oxidation of the compounds was performed as previously described.
2
and 4 (0.3 mg) were dissolved in glacial acetic acid (0.3 mL) and
mixed with CrO (2.0 mg). After being stirred at room temperature for
3
2
0 h, the reaction mixture was separated using a C18 SPE cartridge
CONCLUSIONS
■
eluting with H O followed by MeOH. The resulting oxidized material
from the MeOH fraction was treated with FDAA as described above;
2
The crocapeptins, cyanopeptolin-like cyclic depsipeptides with
kallikrein and protease inhibiting bioactivity, have been isolated
from the terrestrial myxobacterium Chondromyces crocatus. The
previously unknown biosynthesis of the Ahp-heterocycle, the
main structural feature crucial for bioactivity of this compound
class, was elucidated biochemically. The formation of Ahp in
crocapeptin A was achieved in vitro using purified precroca-
peptin and recombinant tailoring enzymes CpnE and CpnF.
The established ability of cytochrome P450 enzyme CpnF to
receive electrons from the broad-range ferredoxin Adx4-108
opens a route for biotechnological production of Ahp-
containing molecules for selective protease inhibitor design. A
function could be assigned to the previously uncharacterized
protein CpnE, encoding a SBPRCC-domain superfamily
protein that catalyzes the rearrangement of a hydroxyproline
residue to Ahp.
L-Glu was detected at t 6.6 min.
Mutagenesis of Chondromyces crocatus Cm c5. An internal
R
fragment of the target gene was amplified by PCR and cloned into the
20
suicide plasmid pSUP_hyg, yielding the respective pSUP_hyg
derivative. After restriction analysis, the plasmid was introduced into
the donor strain E. coli ET12567/pUB307. E. coli ET12567/pUB307
clones containing a pSUP_hyg plasmid derivative were selected on LB-
Agar (low salt) at 37 °C supplemented with 50 mg/L Kanamycin
sulfate (Kan), 30 mg/L Chloramphenicol (Cm), and 50 mg/L Hyg.
Clones were grown in liquid LB medium (low salt), supplemented
with the same antibiotic concentrations at 37 °C. Introduction of the
plasmid into Cm c5 was done through biparental conjugation with E.
6
coli strain ET12567/pUB307 as described elsewhere. Mutant clones
of Cm c5 were selected on Pol03-Agar containing 100 mg/L
Hygromycin B (Hyg) and 60 mg/L Spectinomycin, followed by
cultivation in liquid Pol03 medium containing 50 mg/L Hyg. Hyg-
resistant clones of Cm c5 were verified genetically by PCR as
described in Supporting Information S1.
Heterologous Expression and Purification of CpnE and
CpnF. CpnE was expressed in E. coli BL21 DE3 as an N-terminal his-
tagged protein using the pET28b plasmid (Novagen). Cultivation,
induction of expression, and protein purification were done according
to standard procedures, and CpnE purity was confirmed by SDS-
PAGE. Protein concentration was determined 10 mg/mL using
Bradford measurement. CpnF was expressed in E. coli BL21 DE3 as a
SUMO-Fusion protein using pSUMO plasmid (TaKaRa). Expression
and purification were done according to protocols adapted for
ASSOCIATED CONTENT
Supporting Information
■
*
S
Spectroscopic data, NMR spectra of crocapeptins, details for a
chymotrypsin inhibition assay, genetic verification of Cm c5
mutants, and aldehyde-trapping experiments. This material is
available free of charge via the Internet at http://pubs.acs.org.
AUTHOR INFORMATION
Corresponding Author
rom@helmholtz-hzi.de
■
2
9
optimized heterologous expression of cytochromes. E. coli BL21
DE3/pSUMO_CYP clone was grown in 200 mL of TB medium
supplemented with 50 mg/L Kan. As the culture reached OD600 = 0.6,
Notes
5-amino-levulinic acid was added to a final concentration of 1 mM,
The authors declare no competing financial interest.
FeCl was added to 0.5 mM, and the culture was incubated for 30 min
3
at 26 °C before addition of IPTG to 0.2 mM. The culture was then
ACKNOWLEDGMENTS
■
further incubated for 63 h until cell harvest. Cleared E. coli lysate was
We thank Prof. Rita Bernhardt for supplying purified
adrenodoxin and adrenodoxin-reductase, Prof. Uli Kazmaier
for advice and discussion, Dr. Alberto Plaza for recording NMR
spectra of precrocapeptin, Christel Kakoschke for NMR
̈
separated with the AKTA HPLC system using a 5 mL HisTrap Nickel
IMAC column. A linear gradient from 25 to 500 mM Imidazol was
applied, and absorption was monitored at 280 and 430 nm. Fractions
containing the absorption peak at 430 nm were subjected to SDS-
PAGE, and purity of SUMO-CpnF was judged by Coomassie staining.
Fractions containing the 55 kDa SUMO-CpnF bands were pooled,
concentrated, and cleaved with 20 u SUMO2-protease. The cleaved
SUMO-tag was separated from CpnF in a second IMAC step. Buffer
was exchanged four times with 10 mM potassium phosphate buffer pH
́
measurements, Dr. Giovanna Salamanca-Pinzon for help with
difference spectrometry, Aileen Teichmann and Thomas
Hoffmann for ESI−MS measurements. We thank Diana
Telkemeyer for strain maintenance and Wolfgang Kessler and
his team for cultivation. This work was supported by a return
grant to F.S. from the German Academic Exchange service.
Research in the laboratory of R.M. was supported by the
Deutsche Forschungsgemeinschaft and the Bundesministerium
7.4, and protein was concentrated to a volume of 0.5 mL. Protein
concentration was determined 7.5 mg/mL using Bradford measure-
ment. SDS-PAGE of the protein fraction revealed a strong band at 40
kDa, and two distinct bands at 55 and 70 kDa. These bands were
excised, and subjected to in-gel-digestion and MALDI-MS analysis
according to standard protocols. The signal at 40 kDa is CpnF, the
signal at 55 kDa is SUMO-CpnF, and the signal at 70 kDa is the E. coli
chaperone DnaK.
fu
̈
r Bildung and Forschung.
REFERENCES
■
(1) Wenzel, S. C.; Mu
(2) Weissman, K. J.; Mu
(3) Wenzel, S. C.; Muller, R. Nat. Prod. Rep. 2007, 24, 1211.
̈
ller, R. Mol. BioSyst. 2009, 5, 567.
In Vitro Conversion of Precrocapeptin to Crocapeptin.
̈
ller, R. Nat. Prod. Rep. 2010, 27, 1276.
−
Precrocapeptin was isolated from cpnF mutants as described above
̈
1
6893
dx.doi.org/10.1021/ja4047153 | J. Am. Chem. Soc. 2013, 135, 16885−16894

Journal of the American Chemical Society
Article
(
4) Kunze, B.; Jansen, R.; Ho
004, 57, 151.
5) Kunze, B.; Jansen, R.; Sasse, F.; Ho
Antibiot. 1995, 48, 1262.
6) Kunze, B.; Jansen, R.; Ho
994, 47, 881.
7) Jansen, R.; Kunze, B.; Reichenbach, H.; Ho
Chem. 2003, 2684.
8) Steinmetz, H.; Irschik, H.; Kunze, B.; Reichenbach, H.; Ho
Jansen, R. Chemistry 2007, 13, 5822.
9) Cortina, N. S.; Krug, D.; Plaza, A.; Revermann, O.; Mu
Angew. Chem., Int. Ed. 2012, 51, 811.
10) Welker, M.; von Dohren, H. FEMS Microbiol. Rev. 2006, 30,
30.
11) Yamaki, H.; Sitachitta, N.; Sano, T.; Kaya, K. J. Nat. Prod. 2005,
8, 14.
12) Matern, U.; Schleberger, C.; Jelakovic, S.; Weckesser, J.; Schulz,
G. E. Chem. Biol. 2003, 10, 997.
13) Stolze, S. C.; Meltzer, M.; Ehrmann, M.; Kaiser, M. Chem.
Commun. (Cambridge, U.K.) 2010, 46, 8857.
14) Martin, C.; Oberer, L.; Ino, T.; Konig, W. A.; Busch, M.;
Weckesser, J. J. Antibiot. 1993, 46, 1550.
15) Zainuddin, E. N.; Mentel, R.; Wray, V.; Jansen, R.; Nimtz, M.;
Lalk, M.; Mundt, S. J. Nat. Prod. 2007, 70, 1084.
16) Rounge, T. B.; Rohrlack, T.; Kristensen, T.; Jakobsen, K. S.
BMC Microbiol. 2008, 8, 141.
̈
fle, G.; Reichenbach, H. J. Antibiot.
2
(
̈
fle, G.; Reichenbach, H. J.
(
̈
fle, G.; Reichenbach, H. J. Antibiot.
1
(
̈
fle, G. Eur. J. Org.
fle, G.;
ller, R.
(
̈
(
̈
(
̈
5
(
6
(
(
(
(
(
(
(
17) Marfey, P. Carlsberg Res. Commun. 1984, 49, 591.
18) Itou, Y.; Ishida, K.; Shin, S. J.; Murakami, M. Tetrahedron 1999,
55, 6871.
(
19) Bachmann, B. O.; Ravel, J. Methods Enzymol. 2009, 458, 181.
(
20) Rachid, S.; Krug, D.; Kunze, B.; Kochems, I.; Scharfe, M.;
Zabriskie, T. M.; Blocker, H.; Mu
̈
ller, R. Chem. Biol. 2006, 13, 667.
(
21) Koulov, A. V.; LaPointe, P.; Lu, B.; Razvi, A.; Coppinger, J.;
Dong, M. Q.; Matteson, J.; Laister, R.; Arrowsmith, C.; Yates, J. R., III;
Balch, W. E. Mol. Biol. Cell 2010, 21, 871.
(
22) Stachelhaus, T.; Mootz, H. D.; Marahiel, M. A. Chem. Biol.
999, 6, 493.
23) Buntin, K.; Irschik, H.; Weissman, K. J.; Luxenburger, E.;
Blocker, H.; Muller, R. Chem. Biol. 2010, 17, 342.
24) Khatri, Y.; Hannemann, F.; Ewen, K. M.; Pistorius, D.; Perlova,
O.; Kagawa, N.; Brachmann, A. O.; Muller, R.; Bernhardt, R. Chem.
Biol. 2010, 17, 1295.
25) Khatri, Y.; Girhard, M.; Romankiewicz, A.; Ringle, M.;
1
(
̈
(
̈
(
Hannemann, F.; Urlacher, V. B.; Hutter, M. C.; Bernhardt, R. Appl.
Microbiol. Biotechnol. 2010, 88, 485.
(
26) Ewen, K. M.; Hannemann, F.; Khatri, Y.; Perlova, O.; Kappl, R.;
Krug, D.; Huttermann, J.; Muller, R.; Bernhardt, R. J. Biol. Chem. 2009,
84, 28590.
27) Foulston, L. C.; Bibb, M. J. Proc. Natl. Acad. Sci. U.S.A. 2010,
07, 13461.
28) Hussain, H. A.; Ward, J. M. Appl. Environ. Microbiol. 2003, 69,
73.
29) Krastel, P.; Liechty, B.-M.; Meingassner, J. G.; Schmitt, E.;
̈
2
(
1
(
3
(
Schreiner, E. P. Cyclic Depsipeptides. WO Patent WO 2009/024527
A1, February 26, 2009.
(
30) Krastel, P.; Liechty, B.-M.; Meingassner, J. G.; Moore, C.;
Schmitt, E. Nucleic Acid of a Biosynthetic Cluster Encoding Non
Ribosomal Peptide Synthases and Uses Therof. WO Patent WO 2010/
092109 A2, August 19, 2010.
1
6894
dx.doi.org/10.1021/ja4047153 | J. Am. Chem. Soc. 2013, 135, 16885−16894