Nitro versus hydroxamate in siderophores of pathogenic bacteria: Effect of missing hydroxylamine protection in malleobactin biosynthesis

Article abstract of DOI:10.1002/anie.201303196

The elusive structure of malleobactin, a virulence factor of pathogens belonging to the Burkholderia mallei family, was finally unveiled by genetic and chemical analyses. The novel nitro-substituted siderophore is derived from an unusual, unprotected hydroxylamine, which undergoes spontaneous oxidation, as shown by in vitro assays and detection of analogues featuring hydroxylamino, nitroso, and azoxide groups. Copyright

Full text of DOI:10.1002/anie.201303196

Angewandte
Chemie
DOI: 10.1002/anie.201303196
Natural Products
Nitro versus Hydroxamate in Siderophores of Pathogenic Bacteria:
Effect of Missing Hydroxylamine Protection in Malleobactin
Biosynthesis**
Jakob Franke, Keishi Ishida, Mie Ishida-Ito, and Christian Hertweck*
Siderophores play a major role as virulence factors of
pathogenic bacteria.[1] Known as the strongest Fe3+-binding
agents, these structurally diverse compounds are used to
scavenge scarcely soluble iron from the human or animal
hosts. Strikingly, infamous human pathogenic bacteria like
Yersinia pestis, the causative agent of devastating black death
epidemics, completely lose their virulence in the absence of
a siderophore.[2] Thus, knowledge of a pathogenꢀs siderophore
structures and of the corresponding biosynthetic machineries
is considered a prerequisite for new therapeutic approaches
such as targeting pathogenicity factors and Trojan horse
strategies using siderophore–drug conjugates.[3] Over the past
two decades, much research has been devoted to elucidating
the siderophores of Burkholderia mallei and Burkholderia
pseudomallei,[4–7] the causative agents of the infectious
diseases glanders[8] and melioidosis.[9] Infections with these
b-proteobacteria are often lethal, even with the best treat-
ment available.[10] Both species have thus been categorized as
potential biological warfare agents, and indeed B. mallei has
been abused in World War I to kill enemy horses and
mules.[11,12] Although it has long been known that B. pseudo-
mallei and B. mallei produce two types of siderophores,
pyochelin (1 and 2’’-epi-1, Figure 1) and malleobactin (2),[4]
surprisingly, the structure of the latter has remained obscure.
It was only a matter of speculation that malleobactin could be
related to the ornibactins (3–5, Scheme 1), hydroxamate
siderophores produced by the Burkholderia cepacia complex
(Bcc).[5,6] Here we disclose the unusual structure and absolute
configuration of malleobactin, the siderophore of the human
pathogenic B. mallei family and reveal the biogenetic origin
of an unprecedented aliphatic nitro amino acid.
Figure 1. Comparison of malleobactin and ornibactin biosynthesis
gene clusters (mba and orb, respectively) of Burkholderia spp., and
analysis of siderophore production of B. thailandensis. A) Malleobactin
biosynthesis gene clusters: Bm: B. mallei; Bp: B. pseudomallei; Bt:
B. thailandensis. Ornibactin biosynthesis gene clusters: Bc: B. cenocepa-
cia; Ba: B. ambifaria. Deduced functions of mba biosynthesis genes,
see Table S2 in the Supporting Information. pvd genes refer to
orthologues from the pyoverdin biosynthesis gene cluster in Pseudomo-
nas spp. B) CAS agar plate assay of B. thailandensis wild type (a),
DpchE mutant (b, pch=pyochelin biosynthesis gene cluster), DmbaA
mutant (c), DpchEDmbaA mutant (d). C) LC–MS profiles monitoring
malleobactin (2, m/z 637) and pyochelin (1 and 2’’-epi-1, m/z 325)
production of mutant strains (as in (B)).
To gain first insights into the structural deviations of the
encoded siderophores, we compared the gene loci for
ornibactin and malleobactin biosynthesis.[6,13] Both types of
gene clusters share genes for a tetramodular nonribosomal
peptide synthetase (NRPS) and accessory enzymes. Further-
more, genes for putative amino acid tailoring enzymes
(ornithine monooxygenase,[14] aspartic acid b-hydroxylase,[15]
N-formyltransferase[16]), siderophore receptors and transport-
ers are present in both types of gene clusters. However, we
noted that the gene loci in bacteria belonging to the B. mallei
family harbor an additional gene for a hypothetical protein
(mbaM), but lack orthologues of orbK and orbL (Figure 1A,
yellow open reading frames (orfs)). The latter genes code for
acyltransferases, and although their biochemical function has
not yet been studied, one may assume that they are required
for loading acyl units onto the N-terminal ornithine residue.[13]
We concluded that ornibactins and malleobactin differ in
their substitution patterns.
[*] J. Franke,[+] Dr. K. Ishida,[+] Dr. M. Ishida-Ito, Prof. Dr. C. Hertweck
Dept. of Biomolecular Chemistry, Leibniz Institute for Natural
Product Research and Infection Biology, HKI
Beutenbergstr. 11a, 07745 Jena (Germany)
E-mail: Christian.Hertweck@hki-jena.de
Prof. Dr. C. Hertweck
Friedrich Schiller University
Jena (Germany)
[+] These authors contributed equally to this work.
[**] We thank A. Perner and H. Heinecke for MS and NMR measure-
ments. This work was supported by the “Pakt fꢀr Forschung und
Innovation” of the Free State of Thuringia and the BMBF.
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.201303196.
Angew. Chem. Int. Ed. 2013, 52, 8271 –8275
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
8271

.
Angewandte
Communications
and solved its structure by extensive 2D NMR spectroscopy
(Scheme 2).
In D2O five spin systems were identified by using COSY
that corroborated the predicted relationship with the orni-
Scheme 2. Structure elucidation of malleobactin A (2). Key COSY and
HMBC correlations in D2O, and results from amino acid analysis.
bactins. Four spin systems readily identified putrescine, Nd-
hydroxy-Nd-formylornithine (hfOrn), serine, and b-hydroxy-
aspartic acid (OH-Asp) as peptide building blocks. The
remaining spin system, however, did not fit to the N-terminal
ornibactin residue, Nd-hydroxy-Nd-acylornithine (haOrn). A
proton resonating at 4.37 ppm, which was assigned as the a-
proton of the remaining amino acid, was used as the starting
point of this spin system. From 4JHH and HMBC correlations
we concluded that a formyl group at dH = 8.01 ppm is
connected to this amino group and that this amino acid is
located at the N terminus. Continuing from the CHa group,
two neighboring methylene groups (dH = 1.73 and 1.91 ppm,
respectively) were detected. Whereas no additional coupling
protons could be detected in the COSY spectrum, the HMBC
spectrum showed a correlation of the outer methylene group
to a carbon resonating at 74.1 ppm, which could not be
detected in the 13C spectrum in D2O. MS measurement of the
sample incubated in D2O showed a mass shift of 2 Da,
indicating proton–deuterium exchange owing to the presence
of a highly acidic functional group. Since the resulting CD2
group corresponds to a quaternary carbon atom split into
a 1:2:3:2:1 multiplet, it could not be detected in the 13C
spectrum. Indeed, further NMR measurements in [D6]DMSO
revealed a C-H acidic methylene group (dH = 4.55 ppm, dC =
74.9 ppm). In light of the unusual chemical shifts, the high
acidity, the molecular weight, and the biosynthetic connection
to the ornibactins, it was evident that the remaining amino
acid features a terminal nitro group. This interpretation was
backed up by the IR spectrum of 1, which shows bands for NO
valence vibrations at 1530 and 1359 cmÀ1.[18] Thus, the missing
amino acid represents 2-amino-5-nitropentanoic acid
(ANPA). Notably, this unusual nitro amino acid has only
been known as a synthetic compound and is here presented
for the first time as part of a natural product. Finally, we
Scheme 1. Structures of ornibactins (3–5) and malleobactin A (2), B
(6), C (7), and D (8).
For functional analyses of the gene cluster and to
elucidate the structure of the encoded siderophore, we
chose Burkholderia thailandensis, the least virulent member
of the B. mallei complex as a model.[17] To reveal the potential
of B. thailandensis to produce B. mallei-type siderophores we
analyzed metabolic profiles of the wild type and engineered
strains deficient in functional pyochelin and malleobactin
biosynthesis gene clusters. Therefore, we specifically dis-
rupted genes coding for the putative pyochelin and malleo-
bactin NRPS modules by homologous recombination. By
means of a CAS (chrome azurol S) assay to detect Fe3+
binding capacities of secreted compounds we corroborated
that both gene clusters are sufficient and necessary for
siderophore activity in B. thailandensis (Figure 1B). LC–MS
monitoring of cultures grown under iron limitation proved the
correct assignment of the pyochelin and malleobactin bio-
synthesis gene clusters (Figure 1C). The highly polar com-
pound 2 with a molecular weight of 636 Da, which corre-
sponds to the mystery siderophore reported earlier,[6] is only
detectable in the wild type, but not in the mutant strain. As we
detected several putative congeners with similar retention
time and molecular weight in the crude extracts, we named
the main compound malleobactin A. To elucidate the struc-
ture of malleobactin A (2), crude extracts of B. thailandensis
cultures grown in iron-deficient media were subjected to size-
exclusion chromatography.[6] After multiple rounds of prep-
arative HPLC we obtained a sample of pure malleobactin A
8272
www.angewandte.org
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2013, 52, 8271 –8275

Angewandte
Chemie
assigned the sequence of the amino acid residues by HMBC
correlations in [D6]DMSO and MS/MS fragmentation pat-
terns (Figures S4 and S5 in the Supporting Information).
The relative configuration of OH-Asp was determined to
comparison of MbaC with the well-characterized ornithine
monooxygenase PvdA[29] by means of homology modeling
and sequence alignment (Figures S12 and S13 in the Support-
ing Information) revealed any differences that could explain
over-oxidation of the substrate. Thus, to test the function of
MbaC, we heterologously produced the enzyme in E. coli and
performed an in vitro assay[30] using ornithine as a substrate.
In full agreement with the predicted function of MbaC, N-
hydroxyornithine represented the predominant oxygenation
product (Figure 2B, trace b). Furthermore, we detected small
amounts of ANPA as well as traces of the putative nitroso
intermediate and the azoxide shunt product. The results from
the MbaC assay indicated that the hydroxylamine 6 is the
actual product of the mba NRPS (Figure 2D). Derivatives of
higher oxidation states are more likely formed by autoxida-
tion. This model would be in full accord with the reported
reactivity of aliphatic hydroxylamines.[31] To test this hypoth-
esis we incubated the hydroxylamine malleobactin B (6)
under physiological conditions. Within several hours, 6 was
readily transformed nonenzymatically into malleobactin C
(7), malleobactin A (2), and the side product malleobactin D
(8; Figure 2C). We also observed the formation of a yet
uncharacterized isomer of malleobactin B, which can also be
detected in small amounts in culture extracts. As to the timing
of N-formylation, nonformylated precursors could not be
detected that would provide evidence for a post-NRPS
tailoring reaction. However, the adenylation domain A1
contains a conserved aspartate residue[32] that is required for
interactions with free a-amino groups. The biosynthesis is
completed by nucleophilic release of the nascent peptide
chain with putrescine, catalyzed by condensation domain C4
in lieu of a thioesterase domain.[33]
be threo by J-based analysis (3JHH = 2.8 Hz, JH b = À3 Hz,
2
a
C
2JH a = À2 Hz).[19] The absolute configuration of 2 was
b
C
elucidated using Marfeyꢀs method. Thus, we found that 2 is
composed of l-serine, d-threo-b-hydroxyaspartic acid and l-
hfOrn. ANPA was synthesized as a reference compound as
reported.[20] However, to our surprise neither l- nor d-ANPA
could be detected in the hydrolysate, but l-glutamate, which
proved to be a degradation of product of l-ANPA. We
rationalized this unexpected finding by a nitro–nitrite rear-
rangement, loss of NO, and subsequent oxidation with in situ
formed nitrosyl chloride (Figure S6 in the Supporting Infor-
mation).[21] Thus, we solved the absolute configuration of all
residues and showed that malleobactin A differs from orni-
bactin in an unusual a-formylated l-ANPA in lieu of a l-
haOrn unit.
Aliphatic nitro compounds such as ANPA are extremely
scarce in nature, and only little is known about their
biosynthesis.[22] The most plausible mechanism involves
a full 6-electron oxidation of the corresponding amine.
Whereas it has been proposed that during the biosynthesis
of nitrosugars[23] such as kijanose a single enzyme would
catalyze the complete oxidation to the nitro group,[24] in vitro
enzyme assays only showed the production of the corre-
sponding hydroxylamine.[25] For evernitrose it has been
demonstrated that a nitrososynthase catalyzes the sequential
4-electron oxidation to the nitroso sugar, which is oxidized
spontaneously to the nitro compound.[26,27] Only for aromatic
nitro groups, an enzymatic full 6-electron oxidation of the
corresponding amine has been reported.[28]
From these genetic, biochemical, and chemical data we
concluded that N-acylation in hydroxamate siderophore
pathways has not only evolved to provide a bidentate ligand
for ferric iron coordination, but is also a prerequisite to
protect the N-hydroxy group from further oxidation. Inter-
estingly, diverse natural product structures suggest that
similar scenarios take place in various other biosynthetic
pathways. For example, the siderophores IC202B and
IC202C[34] could originate from a truncated desferrioxamine
pathway that lacks acylation of the terminal hydroxylamine
(Figure S16 in the Supporting Information). On the other
hand, in the absence of N-acyl groups, N-hydroxylation may
set the stage for intriguing downstream reactions. For
example, it has been implicated that the piperazic acid like
amino acids of the nonribosomal peptides kutznerides are
derived from N-hydroxyornithine.[35] Our findings support
a model according to which dehydropiperazic acid is formed
by spontaneous oxidation to the nitroso intermediate fol-
lowed by cyclization to a stable six-membered ring (Fig-
ure S17 in the Supporting Information). Nonetheless, to our
knowledge the formation of nitroso and nitro derivatives has
not been reported to date in the context of in vitro assays with
ornithine and lysine N-monooxygenases.[35,36]
To gain insights into the biosynthesis of the unusual nitro-
substituted siderophore 2 we examined the metabolic profile
of B. thailandensis more closely. Indeed, in crude extracts we
detected compounds with MS and MS/MS data corresponding
to the hydroxylamine 6 and the nitroso derivative 7, which we
termed malleobactin B and C (Scheme 1). Isolation and full
characterization of 6 confirmed the identity of the hydroxyl-
amine moiety. We also detected a compound (8) with
a molecular weight of 1224 Da, suggesting that it could
result from a fusion of two malleobactin units. In fact, MS/MS
as well as NMR data of a pure sample of 8 revealed that it
represents the azoxy-linked dimer malleobactin D. As we
have shown earlier, aromatic azoxy natural products may be
formed from hydroxylamine and nitroso intermediates during
the stepwise enzymatic oxidation of aromatic amino
groups.[28] In light of these findings, we concluded that
a sequential 6-electron transfer leads to the aliphatic nitro
group of malleobactin.
To shed more light on the biosynthesis of the nitro group
we re-examined the mba gene cluster. Yet, we gathered only
one candidate oxygenase gene, mbaC, which could be
involved in nitro group formation. The deduced gene product,
MbaC, belongs to the well-studied class of flavin-dependent
ornithine monooxygenases involved in N-hydroxylation.[14]
Interestingly, neither phylogenetic analyses (Figure 2A)
including known ornithine and lysine monooxygenases nor
In conclusion, we have elucidated the structure of
malleobactin A, the long-sought-after siderophore of animal
and human pathogenic bacteria of the B. mallei complex.
Through gene cluster analyses and targeted knock-outs, we
Angew. Chem. Int. Ed. 2013, 52, 8271 –8275
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
8273

.
Angewandte
Communications
extraordinary siderophore variant featur-
ing a nitro group, we report the unusual w-
nitro amino acid ANPA for the first time as
part of a natural product. The biosynthesis
of the aliphatic nitro group was inferred
from in vitro studies with the N-hydroxy-
lase MbaC and the detection of the
hydroxylamine and nitroso congeners mal-
leobactin B and C, respectively, as well as
the azoxy shunt product malleobactin D.
To our knowledge, natural products con-
taining N-hydroxyornithine without acyla-
tion are unprecedented. Moreover, we
demonstrate that the unprotected hydrox-
ylamine malleobactin B oxidizes spontane-
ously to give the nitroso and nitro deriva-
tives. Our in vitro studies with ornithine
and MbaC fully support this new scheme
for aliphatic nitro group formation. A
prerequisite for generating the reactive
precursor is the lack of genes coding for
dedicated N-acyltransferases (such as
OrbK and OrbL from the ornibactin bio-
synthesis gene cluster). These results imply
that hydroxamate formation in other side-
rophore pathways serves as a natural func-
tional group protection to prevent over-
oxidation. Overall, our study not only
unveils the chemical structure of a virulence
factor of an important pathogenic bacte-
rium, but also showcases a novel route
towards rare aliphatic nitro compounds.
Received: April 16, 2013
Published online: July 2, 2013
Keywords: Burkholderia pseudomallei ·
.
Burkholderia thailandensis · genome mining ·
siderophores · virulence factor
[1] M. Miethke, M. A. Marahiel, Microbiol.
Mol. Biol. Rev. 2007, 71, 413 – 451.
[2] S. Schubert, A. Rakin, J. Heesemann, Int. J.
Med. Microbiol. 2004, 294, 83 – 94.
[3] M. G. P. Page, Ann. N. Y. Acad. Sci. 2013,
1277, 115 – 126.
[4] B. H. Kvitko, A. Goodyear, K. L. Propst,
S. W. Dow, H. P. Schweizer, PLoS
Neglected Trop. Dis. 2012, 6, e1715.
[5] H. M. Yang, W. Chaowagul, P. A. Sokol,
Infect. Immun. 1991, 59, 776 – 780.
[6] A. F. Alice, C. S. Lꢁpez, C. A. Lowe, M. A.
Ledesma, J. H. Crosa, J. Bacteriol. 2006,
188, 1551 – 1566.
Figure 2. Bioinformatics, biochemical, and chemical analyses of the biosynthesis of malleo-
bactins A–D. A) Cladogram of known ornithine and lysine monooxygenases. Clade I: PvdA
orthologues from Burkholderia spp. Clade II: PvdA orthologues from the pyoverdin pathway
in Pseudomonas spp. Clade III: SidA orthologues from siderophore pathways in Aspergillus
fumigatus. Clade IV: MbtG orthologues from the mycobactin pathway in Mycobacterium spp.
The scale bar indicates amino acid substitutions per site. B) LC–MS profiles for MbaC-
catalyzed N-hydroxylation activity assay (negative ion mode). The traces show benzyloxycar-
bonyl (Cbz)-derivatized products of heat-denatured MbaC+ornithine (a), MbaC+ornithine
(b), reference compounds ornithine (9), N-hydroxyornithine (10), ANPA (11) (c). C) LC–MS
profiles showing nonenzymatic degradation of hydroxylamine 6 (m/z 623) to nitroso, azoxy,
and nitro derivatives 7 (m/z 621), 8 (m/z 613), and 2 (m/z 637), respectively, in minimal
medium at 308C after 0 h (a) and after 48 h (b) (positive ion mode). D) Biosynthetic model.
A, adenylation; T, thiolation; C, condensation; E: epimerase; TE, thioesterase; MbaC:
ornithine monooxygenase; MbaE: formyltransferase; MbaH: hydroxylase.
[7] H. Yang, C. D. Kooi, P. A. Sokol, Infect.
Immun. 1993, 61, 656 – 662.
[8] I. Khan, L. H. Wieler, F. Melzer, M. C.
Elschner, G. Muhammad, S. Ali, L. D.
Sprague, H. Neubauer, M. Saqib, Transbound. Emerg. Dis.
2013, 60, 204 – 221.
have unequivocally correlated siderophore activity to the mba
gene cluster in B. thailandensis, the less-virulent model
organism of the B. mallei family. In addition to disclosing an
8274 www.angewandte.org
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2013, 52, 8271 –8275

Angewandte
Chemie
[9] W. J. Wiersinga, T. van der Poll, N. J. White, N. P. Day, S. J.
Peacock, Nat. Rev. Microbiol. 2006, 4, 272 – 282.
[23] S. C. Timmons, J. S. Thorson, Curr. Opin. Chem. Biol. 2008, 12,
297 – 305.
[10] U. Parameswaran, R. W. Baird, L. M. Ward, B. J. Currie, Med. J.
Aust. 2012, 196, 345 – 348.
[11] J. Gilad, I. Harary, T. Dushnitsky, D. Schwartz, Y. Amsalem, Isr.
Med. Assoc. J. 2007, 9, 499 – 503.
[24] H. Zhang, J. A. White-Phillip, C. E. MelanÅon, H. Kwon, W. Yu,
H. Liu, J. Am. Chem. Soc. 2007, 129, 14670 – 14683.
[25] N. A. Bruender, J. B. Thoden, H. M. Holden, Biochemistry 2010,
49, 3517 – 3524.
[12] M. Wheelis, Nature 1998, 395, 213 – 213.
[13] K. Agnoli, C. A. Lowe, K. L. Farmer, S. I. Husnain, M. S.
Thomas, J. Bacteriol. 2006, 188, 3631 – 3644.
[14] J. Olucha, A. L. Lamb, Bioorg. Chem. 2011, 39, 171 – 177.
[15] G. M. Singh, P. D. Fortin, A. Koglin, C. T. Walsh, Biochemistry
2008, 47, 11310 – 11320.
[26] Y. Hu, A. Al-Mestarihi, C. L. Grimes, D. Kahne, B. O. Bach-
mann, J. Am. Chem. Soc. 2008, 130, 15756 – 15757.
[27] J. L. Vey, A. Al-Mestarihi, Y. Hu, M. A. Funk, B. O. Bachmann,
T. M. Iverson, Biochemistry 2010, 49, 9306 – 9317.
[28] R. Winkler, C. Hertweck, Angew. Chem. 2005, 117, 4152 – 4155;
Angew. Chem. Int. Ed. 2005, 44, 4083 – 4087.
[16] B. J. McMorran, H. M. C. S. Kumara, K. Sullivan, I. L. Lamont,
Microbiology 2001, 147, 1517 – 1524.
[29] J. Olucha, K. M. Meneely, A. S. Chilton, A. L. Lamb, J. Biol.
Chem. 2011, 286, 31789 – 31798.
[17] a) P. J. Brett, D. DeShazer, D. E. Woods, Int. J. Syst. Bacteriol.
1998, 48 Pt 1, 317 – 320; b) J. B. Biggins, M. A. Ternei, S. F. Brady,
J. Am. Chem. Soc. 2012, 134, 13192 – 13195; c) J. Franke, K.
Ishida, C. Hertweck, Angew. Chem. 2012, 124, 11779 – 11783;
Angew. Chem. Int. Ed. 2012, 51, 11611 – 11615.
[30] M. Bosello, A. Mielcarek, T. W. Giessen, M. A. Marahiel,
Biochemistry 2012, 51, 3059 – 3066.
[31] B. Lindeke, Drug Metab. Rev. 1982, 13, 71 – 121.
[32] T. Stachelhaus, H. D. Mootz, M. A. Marahiel, Chem. Biol. 1999,
6, 493 – 505.
[18] M. Hesse, H. Meier, B. Zeeh, Spektroskopische Methoden in der
Organischen Chemie, Thieme, Stuttgart, 2005.
[33] T. A. Keating, C. G. Marshall, C. T. Walsh, Biochemistry 2000,
39, 15513 – 15521.
[19] N. Matsumori, D. Kaneno, M. Murata, H. Nakamura, K.
Tachibana, J. Org. Chem. 1999, 64, 866 – 876.
[34] M. Iijima, T. Someno, M. Amemiya, M. Ishizuka, H. Naganawa,
T. Takeuchi, J. Antibiot. 2000, 53, 1411 – 1415.
[20] B. D. Zlatopolskiy, M. Radzom, A. Zeeck, A. de Meijere, J. Org.
Chem. 2006, 71, 1525 – 1534.
[21] B. G. Gowenlock, G. B. Richter-Addo, Chem. Rev. 2004, 104,
3315 – 3340.
[35] C. S. Neumann, W. Jiang, J. R. Heemstra, E. A. Gontang, R.
Kolter, C. T. Walsh, ChemBioChem 2012, 13, 972 – 976.
[36] L. Robbel, V. Helmetag, T. A. Knappe, M. A. Marahiel,
Biochemistry 2011, 50, 6073 – 6080.
[22] R. Winkler, C. Hertweck, ChemBioChem 2007, 8, 973 – 977.
Angew. Chem. Int. Ed. 2013, 52, 8271 –8275
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
8275