Antiterminator-mediated unveiling of cryptic polythioamides in an anaerobic bacterium

Article abstract of DOI:10.1002/anie.201108214

Anti-Terminator: Rise of the Molecules: Overexpression of an antiterminator gene (nusG) in Clostridium cellulolyticum induced the biosynthesis of the novel antibiotic closthioamide and related thioamides. This is the first successful genetic engineering of an anaerobe to trigger a cryptic pathway. Furthermore, synthetic probes provide valuable insights into the biogenetic relationship of the rare thioamide metabolites. Copyright

Full text of DOI:10.1002/anie.201108214

Angewandte
Chemie
DOI: 10.1002/anie.201108214
Natural Products
Antiterminator-Mediated Unveiling of Cryptic Polythioamides in an
Anaerobic Bacterium**
Swantje Behnken, Thorger Lincke, Florian Kloss, Keishi Ishida, and Christian Hertweck*
The genus Clostridium comprises a highly heterogeneous
array of obligate anaerobic organisms that inhabit diverse
ecological niches, ranging from soil to human intestines.
These microorganisms have been extensively studied, not
only because of their ability to produce useful solvents, but
also because some species produce harmful protein toxins
such as botulinum toxin.[1,2] However, despite the large body
of knowledge on clostridia, until recently no secondary
metabolites have been isolated from these or any other
strict anaerobes. Yet, mining the sequenced genomes of
Clostridium species has revealed a widespread occurrence of
secondary metabolism genes, in particular polyketide syn-
thase and non-ribosomal peptide synthetase genes.[3,4] We
have shown that the encoded pathways remain silent under
standard laboratory conditions.[4] Obviously, such cryptic
biosynthesis genes are only activated in the presence of
particular stimuli.[5,6] This is plausible, because secondary
metabolite pathways require much ATP-bound energy, a rare
commodity in the anaerobic world. We recently discovered
that the addition of aqueous soil extract to a C. cellulolyticum
culture activates an otherwise silent metabolic pathway,
leading to a wholly unprecedented type of polythioamide
named closthioamide (1; Scheme 1).[7] Apart from its excep-
tional structural features, closthioamide is highly active
against a variety of bacteria, such as vancomycin-resistant
enterococci (VRE) and methicillin-resistant Staphylococcus
aureus (MRSA), and is the first antibiotic from a strictly
anaerobic bacterium.[7] However, the variable nature of soil
constrains constant production rates and hampers the com-
plicated fermentation of these organisms. As a result, various
minor closthioamide congeners evaded isolation and full
characterization, and biosynthetic studies were unapproach-
able. Herein we present a new strategy using an antitermi-
nator gene to trigger a cryptic biosynthetic pathway and
disclose the structures and antibacterial activities of seven
novel polythioamide congeners. Besides revealing structure–
activity relationships, we used synthetic closthioamide ana-
logues as probes to gain an initial insight into the biogenetic
relationships of the natural polythioamides.
An established approach for the targeted induction of
secondary metabolite biosynthesis is the manipulation of
pathway-specific regulatory genes.[5,8–10] However, as the
closthioamide biosynthesis gene locus still remains obscure,
this approach was out of reach. As an alternative, we focused
on regulatory elements that could be involved in a more
global activation of secondary metabolism.[6–11] When analyz-
ing the C. cellulolyticum genome, we noted the presence of
a putative antiterminator gene, nusG. N-utilizing factor G
(NusG) is an essential protein in E. coli that can increase the
overall rate of transcription. NusG has the ability to decrease
the occupancy of some of the paused RNA-polymerase
(RNAP) complexes by promoting the forward translocation
of RNAP.[12] NusG also enables RNAP to read through
transcription-terminating 1-dependent sites (Figure 1A),
especially in rrn operons, and plays a role in translation
owing to the presence of the Kyprides–Onzonis–Woese
(KOW) motif it shares with ribosomal protein families.[13]
Increasing the processivity of RNA polymerase aids in the
efficient synthesis of the corresponding transcripts and
produces polycistronic mRNA, which results in elevated
protein production. Thus, increasing the nusG expression rate
could potentially activate secondary metabolism in C. cellu-
lolyticum.
Scheme 1. Structure of closthioamide (1), the first antibiotic from
a strictly anaerobic bacterium, Clostridium cellulolyticum.[7]
[*] Dipl.-Biochem. S. Behnken, Dr. T. Lincke, Dipl.-Chem. F. Kloss,
Dr. K. Ishida, Prof. Dr. C. Hertweck
Leibniz Institute for Natural Product Research and Infection
Biology, HKI, Dept. of Biomolecular Chemistry, and Bio Pilot Plant
Beutenbergstrasse 11a, 07745 Jena (Germany)
E-mail: christian.hertweck@hki-jena.de
To generate a C. cellulolyticum mutant overexpressing the
antiterminator gene, nusG was PCR-amplified from genomic
DNA, subcloned and sequenced. It was then cloned down-
stream of the strong and constitutive promoter Pthl into the
high-copy vector pSOS95,[14] yielding E. coli–Clostridium
shuttle plasmid pSB050, which was introduced into C. cellu-
lolyticum. The successful transformation was verified by PCR
and plasmid re-isolation and restriction analysis.[15] To test the
ability of pSB050 to promote overexpression of the nusG
gene, reverse-transcription-quantitative polymerase chain
reaction (RT-qPCR) experiments were performed using
total RNA isolated from logarithmically growing cultures of
Prof. Dr. C. Hertweck
Friedrich Schiller University, Jena (Germany)
[**] We thank A. Perner for MS measurements and E. T. Papoutsakis for
providing plasmid pSOS95. We are grateful to U. Knꢀpfer for
support in fermentations. This work was supported by the “Pakt fꢀr
Forschung und Innovation” of the Free State of Thuringia and the
BMBF, and the International Leibniz Research School for Microbial
and Biomolecular Interactions (ILRS), as part of the excellence
graduate school “Jena School for Microbial Communication”
(JSMC).
Supporting information for this article, including full experimental
details, is available on the WWW under http://dx.doi.org/10.1002/
anie.201108214.
Angew. Chem. Int. Ed. 2012, 51, 2425 –2428
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
2425

.
Angewandte
Communications
homologue triggers an otherwise silent
biosynthetic pathway (Figure 1C, trace a).
In addition to the known closthioamide
and ample amounts of hydroxybenzoate (9),
HPLC-MS and HR-MS analyses of the
mutant broth revealed the presence of
closthioamide congeners with varying num-
bers of sulfur atoms. With this constitutive
producer strain in hand, we were able to
explore the nature of these minor compo-
nents. The crude extracts of 8 L (10 ꢀ 0.8 L)
of the fermentation broth containing the
NusG overproducing strain were subjected
to open column chromatography followed
by a final purification by preparative HPLC
to provide closthioamide 1 (18.73 mg) as
well as seven novel compounds, 2 (3.58 mg),
3 (3.51 mg), 4 (1.49 mg), 5 (0.85 mg), 6
(1.24 mg), 7 (0.96 mg), and 8 (0.77 mg).
HRMS data provided the molecular
formula for 2 and 3 (C29H38O3N6S5), indicat-
ing that one sulfur atom is replaced by
oxygen in these compounds. The 1H and
13C NMR spectra of 2 and 3 are very similar
to those of 1 except for the presence of one
signal for an amide carbonyl carbon at C11
(d169.9) in 2 and at C8 (d173.9) in 3. The
HMBC correlations of H9 (d3.65), H10
(d2.41), and H12 (d3.05) to the carbonyl
carbon C11 (d169.9) in 2 and the correla-
tions of H9 (d3.57), H7 (d2.64), and H6
(d4.00) to C8 (d173.9) in 3 reveal the
Figure 1. A) Simplified model of NusG antiterminator function. RNAP=RNA polymerase.
B) Relative quantity of the mRNA level determined by RT-qPCR of nusG in C. cellulolyticum/
pSB050 (d) vs. wild-type (a). Constitutive expression of 16S rRNA was used for normal-
ization; data of wild-type were set as 1. C) HPLC profiles of extracts from cultures of
a) C. cellulolyticum wild-type; b) C. cellulolyticum/pSOSzero (negative control); c) wild-type
induced with soil extract; d) C. cellulolyticum/pSB050 overexpressing nusG. UV detection at
270 nm.
both wild-type and C. cellulolyticum/pSB050. One-step RT-
qPCRs were performed using a 16S rRNA gene as an internal
standard for the calculation of expression levels. We obtained
relative expression levels for each mRNA sample using the
2ÀDDCq method with the wild-type culture as a calibration
culture (Figure 1).[16] Indeed, the expression level of nusG was
substantially higher in the C. cellulolyticum/pSB050 strain, in
comparison to the wild-type (Figure 1B).
To investigate the effect of nusG expression, both the
wild-type and the mutant were grown in strictly anaerobic
batch cultures in 1 L fermenters with 400 mL of modified
CM3 medium[7] and metabolite production was monitored by
HPLC-HRMS. Intriguingly, the metabolic profile of the
mutant was significantly altered (Figure 1C, trace d). We
observed a peak pattern comparable to the wild-type treated
with soil extract (Figure 1C, trace c),[7] and HR ESI-MS
measurements corroborated the successful production of the
antibiotic closthioamide 1 in the mutant. To exclude the
possibility of any secondary effects from the plasmid, we
constructed a strain containing an empty vector (C. cellulo-
lyticum/pSOSzero) by deleting the SalI expression cassette in
pSOS95.[17] The absence of any thioamides in the broth of this
strain (Figure 1C, trace b) showed that the induction of
closthioamide biosynthesis correlates with the overexpression
of nusG. Notably, whereas a NusG-like antiterminator has
been shown to be essential for myxovirescin (TA) biosynth-
esis,[11] this is the first case where the overexpression of a nusG
positions of the amide functions (Scheme 2). Thus, 2 and 3
represent regioisomeric oxa analogues of closthioamide
(Scheme 2).
The molecular formulas of
4
(C18H25N3O3S3),
6
(C16H23N3O2S3), and 7 (C16H21N3O3S3) suggested that these
metabolites are truncated non-symmetrical congeners. The
13C NMR spectra of 4, 6, and 7 are similar to the spectrum
obtained for 1, indicating that these compounds share the
same substructures, and 2D NMR spectral analysis (1H–1H
COSY, HSQC, and HMBC) indeed revealed the expected p-
hydroxythiobenzoyl-b-thioalanyl-b-thioalanyl sequence for
these three compounds. However, according to 1H and
13C NMR spectra, thioamides 4, 6, and 7 feature different
residues in place of a diaminopropane unit. In the 13C NMR
spectrum of 6, a signal for a hydroxymethylene carbon at
1
58.5 ppm (C1) was observed. The H–1H COSY (H1 to C3-
NH) and HMBC (H3 to C1, C2, and C4) correlations showed
that 6 bears a 3-aminopropanol unit. The 13C NMR data for 4
differed from those for 6 in two signals for methyl and
carbonyl carbons (d173.0; C2), and the HMBC correlations
of C2 to methyl protons (H1) and hydroxymethylene protons
(H3) revealed that 4 is an acetyl ester of 6. In contrast,
compound 7 has a carbonyl carbon (C1) resonating at
1
175.4 ppm, which is part of a b-Ala moiety, according to H–
1H COSYand HMBC correlations. HRMS and 13C NMR data
of 5 and 8 indicated that these two metabolites feature only
two thioamide groups. As for the other polythioamides, p-
2426 www.angewandte.org
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2012, 51, 2425 –2428

Angewandte
Chemie
Scheme 2. Structures of closthioamides B–H (2–8).
(hydroxy)thiobenzoyl and b-thioalanyl units were deduced
from 2D NMR data. Compound 8 proved to be the shorter
analogue of 7, composed of only two b-Ala building blocks.
The NMR data for 5 were similar to those for 8, yet the
quaternary carbon (C1) resonating at 119.3 ppm, the HRMS
data, and a characteristic band at 2328 cmÀ1 in the IR
spectrum revealed that 5 represents a dithioamide variant
with a terminal nitrile moiety (Scheme 2).
Considering the scarcity of currently known natural
products featuring thioamide groups (5 out of > 170000),
the prevalence of thioamides produced by this anaerobic
bacterium is impressive. The structural kinship of these
Clostridium-derived thioamides is highly suggestive of
a common biosynthetic origin. A particularly intriguing
question is whether the oxa analogues 2 and 3 were precursors
of the hexathioamide 1 or vice versa. To pinpoint the timing of
thionation and to elucidate the course of metabolite forma-
tion, we synthesized chemical polyamide and polythioamide
probes equipped with fluoro and deuterium labels (11–13,
Scheme 3). In brief, a convergent peptide coupling strategy
enabled the efficient assembly of the hexaamide backbone of
difluoroclosamide (11) from bis-amide 14. In conjunction
with the terminal b-alanine moieties, the p-fluorobenzamide
residues were introduced to give 11 in good overall yield
(84%). Thionation of the hexaamide 11 by Lawessonꢁs
reagent in pyridine afforded the corresponding difluoroclos-
thioamide 12 in good yield (65%). For the synthesis of
deuterated closamide 13, a different approach was chosen
that allows the late incorporation of the costly [D4]-p-
hydroxybenzoic acid. First, bis(amide) 14 was symmetrically
Scheme 3. A) Outline of the synthesis of labeled probes 11–13.
B) Summary of labeling experiments and model for the biosynthesis of
closthioamides. a) EDC, HOBt, Hꢀnig’s Base, DMF, RT, 19.5 h, 86%;
b) Lawesson’s reagent, pyridine, 1008C, 22 h, 65%; c) 1. Boc-b-Ala-
OH, EDC, HOBt, Hꢀnig’s Base, DMF, RT, 42 h, 80%; 2. MeOH,
dioxane, HCl (3.0m), RT, sonication, 30 min; directly converted;
d) 1. EDC, HOBt, Hꢀnig’s Base, DMF, 408C, 15 h, 91%; 2. H2, Pd/C
(12 mol%), acetic acid, 40–508C, sonication, 3.5 h, 96%. EDC=
1-ethyl-3-(3-dimethylaminopropyl)carbodiimide hydrochloride,
HOBt=hydroxybenzotriazole.
elongated by two b-alanine units, deprotection then gave
tetraamide 15. Finally, amide coupling was used to introduce
the benzyl-protected [D4]-p-hydroxybenzoyl residue, and
ultrasound-mediated hydrogenolysis afforded [D8]-closamide
(13) in excellent yield (96%).
We then monitored growing cultures of C. cellulolyticum/
pSB050 by HPLC-HRMS (Orbitrap) with and without
supplementation of fluorinated or deuterated probes. When
administering [D8]-closamide, neither single nor multiple
thionation could be observed. Conversely, after addition of
Angew. Chem. Int. Ed. 2012, 51, 2425 –2428
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
2427

.
Angewandte
Communications
fluoro-labeled hexathioamide 12 to the C. cellulolyticum/
pSB050 culture, we observed the slow exchange of sulfur for
oxygen, resulting in the formation of oxa analogues. These
results indicate that thionation does not take place after the
polyamide backbone has been assembled, but rather con-
comitant with chain elongation. As to the truncated thio-
amides, no fluorinated analogues of 4, 6, or 7 could be
detected. Thus, we could rule out that these compounds are
degradation products, but rather result from a diverging
pathway. A plausible scenario would be that 7 results from an
erratic third elongation of the (thio)amide backbone. The
carboxyl of the third b-alanine unit would then undergo
reduction and acetylation to yield 6 and 4. The fusion to an
aminopropanol linker in lieu of the diaminopropane unit is
likewise conceivable, and it is also possible that the diamino-
propane unit could undergo transamination to form the
hydroxy and carboxy substituents. The metabolite production
curve (see the Supporting Information) suggests that dithio-
amide 8 is a biosynthetic intermediate that was hydrolyzed
instead of being ligated to the central diaminopropane unit of
1. As traces of fluorinated 8 could be detected, it appears that
the acid might also result from hydrolytic cleavage of
closthioamide. We also observed the formation of the fluoro
analogue of 5, which indicates that the nitrile moiety derives
from transformation of the thioamide group. Such trans-
formations, through a biosynthetic mechanism involving
consecutive oxygenations of the thioamide sulfur atom to
form the corresponding nitrile and sulfur dioxide, have been
reported.[18,19] HRMS data also point to another nitrile,
corresponding to the truncated monothioamide (see the
Supporting Information). Taken together, our results strongly
suggest that closthioamide is assembled stepwise from
hydroxybenzoate (9), and one, two, or three b-alanine units
with concomitant thionation of the intermediates. The final
product would be obtained by the fusion of two of these
precursors with a diaminopropane linker (Scheme 3B).
Finally, all new compounds were subjected to antibacterial
assays.[20] Compared to hexathioamide 1, which is highly
active against pathogens such as MRSA and VRE,[7] the
activities of oxa analogues 2 and 3 are significantly less
pronounced. This observation is in line with the complete loss
of activity of closamide (10).[7] We also observed dramatically
reduced activities of the truncated closthioamide congeners 4,
5, 6, and 8, implying that the symmetrical polythioamide
backbone is crucial for antibiotic action against E. coli,
MRSA, and VRE (Supporting Information, Table S3).
probes we have gained an initial insight into the biogenetic
relationship of the polythioamides. Biological profiling of the
new di- and polythioamides gave valuable information on the
structure–activity relationships of this new type of antibiotic
and revealed the importance of the symmetric hexathioamide
backbone for biological activity. To the best of our knowledge,
this is the first report of both the genetic manipulation of an
anaerobic microorganism to stimulate secondary metabolism
and the use of an antiterminator to induce the production of
cryptic natural products. We believe that this approach is
generally applicable and of high value for the discovery and
sustainable production of new biologically active, yet cryptic,
natural products. In particular, this method may represent
another avenue to explore the untapped biosynthetic poten-
tial of neglected and difficult to handle bacteria, such as strict
anaerobes.
Received: November 22, 2011
Published online: January 27, 2012
Keywords: anaerobe · closthioamide · Clostridium cellulolyticum ·
.
nusG antiterminator · thioamide
[1] H. Bahl, P. Dꢂrre, Clostridia: Biotechnology and Medical
Applications, Wiley-VCH, Weinheim, 2001.
[2] C. J. Paredes, K. V. Alasker, E. T. Papoutsakis, Nat. Rev. Micro-
biol. 2005, 3, 969.
[3] H. Seedorf, W. F. Fricke, B. Veith, H. Brꢂggemann, H. Liese-
gang, A. Strittmatter, W. Buckel, J. Hinderberger, C. Hagemeier,
R. K. Thauer, G. Gottschalk, Proc. Natl. Acad. Sci. USA 2008,
105, 2128.
[4] S. Behnken, C. Hertweck, PlosOne 2012, 7(1), e29609.
[5] J. M. Winter, S. Behnken, C. Hertweck, Curr. Opin. Chem. Biol.
2011, 15, 22.
[6] K. Scherlach, C. Hertweck, Org. Biomol. Chem. 2009, 7, 1753.
[7] T. Lincke, S. Behnken, K. Ishida, M. Roth, C. Hertweck, Angew.
Chem. 2010, 122, 2055; Angew. Chem. Int. Ed. 2010, 49, 2011.
[8] S. Bergmann, J. Schꢂmann, K. Scherlach, C. Lange, A. A.
Brakhage, C. Hertweck, Nat. Chem. Biol. 2007, 3, 213.
[9] C. Corre, G. L. Challis, Nat. Prod. Rep. 2009, 26, 977.
[10] K. Ishida, T. Lincke, S. Behnken, C. Hertweck, J. Am. Chem.
Soc. 2010, 132, 13966.
[11] Y. Paitan, E. Orr, E. Z. Ron, E. Rosenberg, FEMS Microbiol.
Lett. 1999, 170, 221.
[12] K. M. Herbert, J. Zhou, R. A. Mooney, A. L. Porta, R. Landick,
S. M. Block, J. Mol. Biol. 2010, 399, 17.
[13] J. Liu, H. Pei, S. Mei, J. Li, L. Zhou, H. Xiang, Biochem. Biophys.
Res. Commun. 2008, 371, 573.
[14] S. B. Tummala, N. E. Welker, E. T. Papoutsakis, Appl. Environ.
Microbiol. 1999, 65, 3793.
[15] C. Tardif, H. Maamar, M. Balfin, J. P. Belaich, J. Ind. Microbiol.
Biotechnol. 2001, 27, 271.
[16] K. J. Livak, T. D. Schmittgen, Methods 2001, 25, 402.
[17] S. Perret, H. Maamar, J. P. Belaich, C. Tardif, Mol. Microbiol.
2004, 51, 599.
[18] A. G. Dodge, J. E. Richman, G. Johnson, L. P. Wackett, Appl.
Environ. Microbiol. 2006, 72, 7468.
[19] T. A. Vannelli, A. Dykman, P. R. Ortiz de Montellano, J. Biol.
Chem. 2002, 277, 12824.
[20] F. Kloss, T. Lincke, C. Hertweck, Eur. J. Org. Chem. 2011, 1429 –
1431.
In summary, we have successfully employed an antitermi-
nator gene, nusG, for the induction of a cryptic biosynthetic
pathway. We thus generated a mutant that can be harnessed
for the sustainable production of a potent polythioamide
antibiotic. This mutant allowed the isolation of seven novel
thioamide metabolites (2–8), which more than doubles the
number of currently known natural products featuring this
rare functional group. Notably, closthioamides B–H (2–8) are
also the second to eighth known secondary metabolites from
a strictly anaerobic bacterium. Furthermore, through syn-
thesis and application of deuterium- and fluoro-labeled
2428 www.angewandte.org
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2012, 51, 2425 –2428