Assembly and absolute configuration of short-lived polyketides from burkholderia thailandensis

Article abstract of DOI:10.1002/anie.201200067

Decoded before decay: Cryptic and highly unstable polyketide metabolites, thailandamidesA and B, were isolated from Burkholderia thailandensis, and their absolute configurations fully elucidated using a combination of chemical degradation, bioinformatics, NMR spectroscopy, precursor-directed biosynthesis, and analysis of pathway intermediates (see scheme). Copyright

Full text of DOI:10.1002/anie.201200067

Angewandte
Chemie
DOI: 10.1002/anie.201200067
Polyketides
Assembly and Absolute Configuration of Short-Lived Polyketides from
Burkholderia thailandensis**
Keishi Ishida, Thorger Lincke, and Christian Hertweck*
Complex polyketides from bacteria comprise a broad range of
macrolides, polyenes, and polyethers, many of which play
eminent roles for the development of therapeutics.[1] Typi-
cally, these compounds feature multiple chiral centers that
render them challenging benchmarks for organic synthesis. In
light of this, it is impressive to learn how swiftly complex
polyketides are assembled in microorganisms, utilizing only
simple building blocks such as activated acyl and malonyl
units.[2,3] The types, number, and processing of these basic
blocks are programmed by the architecture of modular
polyketide synthases (PKS) which are encoded in the micro-
bial genomes. One most remarkable discovery was that the
order of catalytic domains involved in chain elongation and
side-chain processing is typically colinear with the final
product.[4] This principle has not only opened the door to
rationally engineering polyketides, but also enabled forecast-
ing PKS architectures from polyketide structures and vice
versa.[5] However, through recent full genome sequencing
projects it has become more and more obvious that most of
the genes coding for biosynthetic assembly lines remain silent
under standard laboratory conditions.[6] Furthermore, PKS
expression and polyketide formation may only occur tran-
siently, thus suggesting that the metabolites function as
temporary signals.[7] This remarkable case has been observed
for Burkholderia thailandensis, a bacterium that serves as
a model for the pathogenic relatives B. mallei and B.
pseudomallei. Mining the genome of B. thailandensis revealed
a giant trans-acyltransferase (trans-AT) PKS[8] gene cluster,
Scheme 1. Structure of thailandamide A (1) and experiments revealing
and a bioinformatic structure prediction aided in the discov-
ery of a hitherto unknown polyketide, thailandamide A (1,
Scheme 1A).[7] However, the absolute configuration of the
labile polyene 1 and the structure of a short-lived congener
(thailandamide B; 2; see Figure 2 for structure), have
remained elusive. Yet this very structural information would
serve as a stepping-stone to clarify the biological function of
this enigmatic, quorum-sensing-regulated metabolite. Herein
we report a multidisciplinary approach involving bioinfor-
the absolute configuration of 1. A) Degradation and derivatization.
a) O3, À788C; b) 35% H2O2, AcOH, 608C; c) 6m HCl, 1058C;
d) S-PGME, PyBOP; e) l-FDAA; f) (S,R)-PGME, PyBOP. B) Synthesis of
deoxythailandamide. g) B. thailandensis DPthaA, precursor-directed
biosynthesis; h) (S,R)-MTPA chloride, pyridine. B. DdSR values in ppm.
matics, chemical degradation, and precursor-directed biosyn-
thesis to rigorously determine the absolute configuration of
thailandamide A. On the basis of the stereochemical assign-
ments, in silico predictions, and mutational pathway dissec-
tion, we also unveil the structure of the immediate, thermo-
labile PKS product thailandamide B.
[*] Dr. K. Ishida, T. Lincke, Prof. Dr. C. Hertweck
Leibniz Institute for Natural Product Research and Infection
Biology, HKI, Dept. of Biomolecular Chemistry, and Bio Pilot Plant
Beutenbergstr. 11a, 07745 Jena (Germany)
Thailandamide features six chiral centers, at C2, C3, C13,
C22, C24, and C29. We initiated the stereochemical analysis
by chemical degradation and derivatization. For the isolation
of preparative amounts of thailandamide A (1) we employed
a genetically engineered strain, the quorum sensing mutant B.
thailandensis DPthaA, which produces thailandamide consti-
tutively.[9] From an upscaled fermentation (10 L) of the
mutant we succeeded in isolating sufficient amounts of
1 (36.9 mg) for several degradation and derivatization experi-
E-mail: christian.hertweck@hki-jena.de
Prof. Dr. C. Hertweck
Chair for Natural Product Chemistry, Friedrich Schiller University
Jena (Germany)
[**] We thank A. Perner, H. Heinecke, and F. Rhein for MS and NMR
measurements. This research was financially supported by the Pakt
fꢀr Wissenschaft und Innovation.
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.201200067.
Angew. Chem. Int. Ed. 2012, 51, 1 – 6
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1
These are not the final page numbers!

.
Angewandte
Communications
ments. First, to elucidate the absolute configuration at
positions C22 and C24, the polyene was subjected to
ozonolysis (Scheme 1A). The resulting oxidative degradation
product (3) was hydrolyzed using 6m HCl and derivatized
using Marfeyꢀs reagent for Ala, and (S)-PGME (phenyl-
glycine methylester) for (R)-methylsuccinate. Since the RP-
HPLC analysis of the l-FDAA (1-fluoro-2,4-dinitrophenyl-5-
l-alanine amide) derivative of the hydrolysate revealed the
identity of d-Ala, we could unequivocally assign an R confi-
guration to position C22. The di-(S)-PGME amide of hydro-
lyzed 3 showed the same retention time as the (R)-methyl-
succinate di-(S)-PGME amide on the HPLC column, thus
indicating the R configuration at C24.
To elucidate the absolute configuration at C2, again the
formation of a PGME amide provided valuable insights. From
the DdSR values of the thailandamide A-derived (S)- and (R)-
PGME amides 4a and 4b, respectively, we deduced the
R configuration of position C2. The configuration of C3,
however, could not be determined in this way. Therefore, we
inspected the NOE correlations of the g-butyrolactone ring in
the thailandamide lactone 5, a congener and derivative of
1 produced by B. thailandensis DPthaA. From these correla-
tions we were able to unequivocally deduce the S configura-
tion at C3. The determination of the absolute configurations
at positions C13 and C29 proved to be most challenging, since
no reference compounds were available for fragments
obtained by ozonolysis. Furthermore, treatment with Mosh-
erꢀs reagent (MTPA; a-methoxy-a-(trifluoromethyl)phenyl-
acetic acid) chloride gave mixed products, including the
mono-MTPA ester at the phenolic position of 1. Unexpect-
edly, attempts to selectively protect the phenolic group
yielded several by-products because of the temperature
sensitivity of 1. To circumvent these limitations we aimed at
generating a thailandamide variant lacking the phenolic OH
group. Since a synthetic method for achieving this was out of
reach, we attempted producing this derivative by precursor-
directed biosynthesis. Analysis of the polyketide backbone
suggested that p-hydroxphenylacetate (PHPA) served as
a starter unit which is loaded onto the first PKS module.
The deoxy variant of 1 would in principle result from the
incorporation of phenylacetate (PA). After optimization of
production parameters, we eventually achieved this goal by
supplementing a B. thailandensis DPthaA culture with the
nonnatural starter unit in minimal medium (Scheme 1B).
HPLC-HRMS monitoring of the fermentation indicated that
the mutant indeed incorporated phenylacetate (Figure 1,
trace c).
Figure 1. HPLC profiles of a sample of purified thailandamide B [2;
(a)], extracts from cultures of B. thailandensis DPthaA (b), extracts
from cultures of B. thailandensis DPthaA supplemented with phenyl-
acetate in minimal medium (c), and d) extracts from cultures of B.
thailandensis DPthA/DTE double mutant. UV absorbance at l=317 nm
(a–c) and l=220 nm (d). 6’: Putative deoxythailandamide B;
*: unidentified unstable (deoxy)thailandamides.
Gly, Ala, and Val. Since the chemical analyses showed that 1 is
derived from d-Ala, it is surprising that the condensation (C)
domain does not have a dual condensation/epimerization
function.[11] Consequently, this seems to be a rare scenario
where an A domain incorporates a d-amino acid, as for
example in the cyclosporin synthetase.[12] Next, we predicted
all ketoreductase (KR) domain specificities[13–15] in the
thailandamide PKS. For the hydroxy-substituted positions,
the KR finger printing fully matched with the detected
configurations at C29(l/R) and C13 (d/S; see Table S7 in the
Supporting Information). Interestingly, the S configuration of
the methoxy-substituted carbon atom (C3) could not be
predicted since the product is not colinear with the architec-
ture of module 13, which lacks a KR domain. The domain
organization could be suggestive for a route involving
enolization and methylation, followed by hydrogenation.[8]
A likely candidate is the trans-enoylreductase (trans-ER)
domain in ThaF, which controls the absolute configuration at
C24. While stereocontrol of enoylreductases has been
unveiled for cis-AT PKS systems,[16–18] the predictions are
not adaptable for the trans-ER in trans-AT PKS systems,
which are more similar to PfaD in polyunsaturated fatty acid
(PUFA) biosynthesis.[19] However, one may assume that the
stereochemical course of enoyl reduction would be compara-
ble for C24 and C2/3, but no clear mechanistic analogies could
be drawn. Thus, a more plausible alternative would be the
iterative use of a KR domain upstream of module 13. Indeed
all vicinal KR domains produce d-hydroxy groups, which is in
accord with the observed S configuration at C3. Furthermore,
by using HPLC-HRMS we were able to detect the exact mass
for predicted hydroxy-substituted intermediate in a DTE-
mutant (see below, and the Supporting Information).
An isolation/purification workflow as established for
1 yielded sufficient amounts of deoxythailandamide A (6;
(Scheme 1B), which could be readily esterified by MTPA
chloride without the formation of any side products. Finally,
the DdSR values of the deoxythailandamide A-derived di-(S)-
and di-(R)-MTPA esters 7a and 7b, respectively, revealed an
S configuration for C13 and an R configuration for C29. To
complement the stereochemical analysis, we analyzed the
enzymatic domains encoded in the tha gene locus (Scheme 2).
According to NRPSPredictor2,[10] the adenylation (A)
domain of ThaH has the amino acid code DMPQLGMVWK,
which indicates a preference for small amino acids such as
As to the double-bond architectures, previous work
suggested that the dehydratase (DH) domain mediates an
2
www.angewandte.org
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2012, 51, 1 – 6
These are not the final page numbers!

Angewandte
Chemie
Scheme 2. Model for thailandamide assembly on the modular tha trans-AT PKS, based on domain analysis and metabolic profiling of a DTE
mutant. KS: ketosynthase, AT: acyl transferase, ACP: acyl carrier protein, MT: methyltransferase, KR: ketoreductase, DH: dehydratase, ER: enoyl
reductase, TE: thioesterase. Insert: mechanism of b branching (green circles highlight methyl groups derived from b-branching). ECH: enoyl-CoA
hydratase/crotonase, HCS: HMG-CoA synthase. Structure numbers are intermediates detected by HR-LCMS.
anti elimination of water in d-3-hydroxyacyls to yield trans
double bonds, whereas cis double bonds result from the
corresponding l-isomers.[2,20] However, this prediction did not
match with the observed E configuration at C16/17. This
inconsistency implied that thailandamide A (1) might not
represent the true primary product of the tha PKS, and we
assumed that the yet enigmatic, unstable congener of 1,
thailandamide B (2) would be the product released from the
megasynthase. However, we repeatedly experienced difficul-
ties in isolating the isomer as it readily degraded and
isomerized in the course of purification, even in the absence
of light. The resulting mixtures gave complex NMR data
which evaded interpretation because of overlapping signals of
various stereoisomers. Finally, we established a low-temper-
ature isolation and purification protocol, thus carefully
maintaining temperatures below room temperature to
obtain pure 2. NMR measurements were also conducted
below 108C to provide high quality 1H-1H COSY, HSQC, and
HMBC NMR data (see the Supporting Information) which
allowed the full structural elucidation of 2.
at C16/17 was Z (cis) in 2 compared to that of E (trans) in 1.
The diagnostic coupling constant 11.2 Hz between H16 and
H17 provided further support for the Z configuration at C16–
C17. According to coupling constants and ROESY data, all
other geometrical constitutions were identical with those in
1 (Figure 2 and Figure S11 in the Supporting Information).
The architecture of 2 is in full agreement with the deduced
biosynthetic code, thus implying that 2 would be the
immediate product released from the modular assembly
line. The cis double bond could be prone to rapid isomer-
ization to give 1. Alternatively, both cis and trans isomers
could form concomitantly because of some degree of
flexibility of the DH domain. To address this, we first tested
whether 2 would readily isomerize in vitro to yield 1.
Interestingly, the 1H and 13C NMR spectra of 2 were
almost identical with those of 1 except for positions C16 to
C18 (position 16, 1; dH = 6.24 ppm, dC = 130.8 ppm, 2; dH =
6.11 ppm, dC = 125.0 ppm, position 17, 1; dH = 7.62 ppm, dC =
140.5 ppm, 2; dH = 6.93 ppm, dC = 139.3 ppm, position 18, 1;
dH = 6.19 ppm, dC = 129.9 ppm, 2; dH = 7.42 ppm, dC =
128.6 ppm). This deviation suggested that the double bond
Figure 2. Structure of thailandamide B (2) and selected correlations
observed by 2D NMR spectroscopy. Bold line: 1H-1H COSY correla-
tions; arrows: HMBC correlations.
Angew. Chem. Int. Ed. 2012, 51, 1 – 6
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
3
These are not the final page numbers!

.
Angewandte
Communications
However, only small amounts of 1 are formed by incubating 2
at room temperature with or without light, with formation of
many yet uncharacterized isomerization or degradation
products, a scenario that does not reflect co-occurrence of
1 and 2 found in the culture broth. Thus, to pinpoint the
stereochemical aspects of polyketide biosynthesis, but also
endorse the power of synergizing chemical and biological
methods in natural product research.
timing of cis/trans double-bond formation, we next aimed at Experimental Section
See the Supporting Information.
dissecting the biosynthetic pathway in vivo. For this purpose,
we deleted the gene region coding for the off-loading
thioesterase (TE) domain at the C-terminus of ThaQ. This
genetic manipulation was believed to result in a DTE-mutant
accumulating and releasing premature polyketide intermedi-
ates as in the bacillaene[21,22] and rhizoxin[23,24] pathways.
Indeed, by HPLC-MS and HRMS (Exactive) measurements
we were able to detect the masses for the entire range of
predicted structures. Through SIM (selected ion monitoring)
we obtained snapshots of the entire pathway (Scheme 2). In
this way, it was not only possible to deduce the course chain
propagation and branching events (Scheme 2, inset), but also
to trace the onset of Z/E isomerization. We found that up to
module 5 only single isomers are formed, whereas products
generated by module 6 and all additional downstream mod-
ules appeared as two main isomers (e.g. 15a,b, 16a,b;
Scheme 2). It is highly unlikely that an as yet unidentified
isomerase would be promiscuous enough to act equally
efficiently on all of these different substrates. Thus, the results
taken together, we reason that the DH domain of split
module 6 unexpectedly yields cis and trans double bonds,
either through catalyzing both anti and syn eliminations or
through isomerization of the double bond of the tethered
substrate.
Received: January 4, 2012
Revised: March 5, 2012
Published online: && &&, &&&&
Keywords: bioinformatics · biosynthesis · polyketides ·
.
stereochemistry · structure elucidation
[1] D. OꢀHagan, The Polyketide Metabolites, Ellis Horwood, Chi-
chester, 1991.
[2] C. Hertweck, Angew. Chem. 2009, 121, 4782 – 4811; Angew.
Chem. Int. Ed. 2009, 48, 4688 – 4716.
[3] C. Khosla, J. Org. Chem. 2009, 74, 6416 – 6420.
[4] J. Staunton, K. J. Weissman, Nat. Prod. Rep. 2001, 18, 380 – 416.
[5] K. J. Weissman, P. F. Leadlay, Nat. Rev. Microbiol. 2005, 3, 925 –
936.
[6] J. M. Winter, S. Behnken, C. Hertweck, Curr. Opin. Chem. Biol.
2011, 15, 22 – 31.
[7] T. Nguyen, K. Ishida, H. Jenke-Kodama, E. Dittmann, C.
Gurgui, T. Hochmuth, S. Taudien, M. Platzer, C. Hertweck, J.
Piel, Nat. Biotechnol. 2008, 26, 225 – 233.
[8] J. Piel, Nat. Prod. Rep. 2010, 27, 996 – 1047.
[9] K. Ishida, T. Lincke, S. Behnken, C. Hertweck, J. Am. Chem.
Soc. 2010, 132, 13966 – 13968.
[10] M. Rçttig, M. H. Medema, K. Blin, T. Weber, C. Rausch, O.
Kohlbacher, Nucleic Acids Res. 2011, 39, W362 – W367.
[11] C. J. Balibar, F. H. Vaillancourt, C. T. Walsh, Chem. Biol. 2005,
12, 1189 – 1200.
[12] K. Hoffmann, E. Schneider-Scherzer, H. Kleinkauf, R. Zocher, J.
Biol. Chem. 1994, 269, 12710 – 12714.
[13] P. Caffrey, ChemBioChem 2003, 4, 649 – 662.
[14] A. Baerga-Ortiz, B. Popovic, A. P. Siskos, H. M. OꢀHare, D.
Spiteller, M. G. Williams, N. Campillo, J. B. Spencer, P. F.
Leadlay, Chem. Biol. 2006, 13, 277 – 285.
[15] R. Castonguay, W. He, A. Y. Chen, C. Khosla, D. E. Cane, J. Am.
Chem. Soc. 2007, 129, 13758 – 13769.
[16] D. H. Kwan, Y. Sun, F. Schulz, H. Hong, B. Popovic, J. C. C. Sim-
Stark, S. F. Haydock, P. F. Leadlay, Chem. Biol. 2008, 15, 1231 –
1240.
[17] D. H. Kwan, P. F. Leadlay, ACS Chem. Biol. 2010, 5 829 – 838.
[18] D. H. Kwan, F. Schulz, Molecules 2011, 16, 6092 – 6115.
[19] S. B. Bumpus, N. A. Magarvey, N. L. Kelleher, C. T. Walsh, C. T.
Calderone, J. Am. Chem. Soc. 2008, 130, 11614 – 11616.
[20] J. Wu, T. J. Zaleski, C. Valenzano, C. Khosla, D. E. Cane, J. Am.
Chem. Soc. 2005, 127, 17393 – 17404.
[21] J. Moldenhauer, X.-H. Chen, R. Borriss, J. Piel, Angew. Chem.
2007, 119, 8343 – 8345; Angew. Chem. Int. Ed. 2007, 46, 8195 –
8197.
[22] J. Moldenhauer, D. C. G. Gçtz, C. R. Albert, S. K. Bischof, K.
Schneider, R. Sꢁßmuth, M. Engeser, H. Groß, G. Bringmann, J.
Piel, Angew. Chem. 2010, 122, 1507 – 1509; Angew. Chem. Int.
Ed. 2010, 49, 1465 – 1467.
[23] B. Kusebauch, B. Busch, K. Scherlach, M. Roth, C. Hertweck,
Angew. Chem. 2009, 121, 5101 – 5104; Angew. Chem. Int. Ed.
2009, 48, 5001 – 5004.
[24] B. Kusebauch, B. Busch, K. Scherlach, M. Roth, C. Hertweck,
Angew. Chem. 2010, 122, 1502 – 1506; Angew. Chem. Int. Ed.
2010, 49, 1460 – 1464.
Elucidating the absolute configuration of complex poly-
ketides with multiple chirality centers is often an ambitious
undertaking
which
may
require
multidisciplinary
approaches.[26–29] In this work we describe a particularly
challenging case of a stereochemical analysis to unravel the
structure of a short-lived polyene, thailandamide A (1). We
succeeded in the unequivocal determination of all six
stereocenters of this fragile molecule (2R, 3S, 13S, 22R, 24R,
29R), which was only feasible by a combination of chemical
degradation, derivatization, 1D/2D NMR analyses, and pre-
cursor-directed biosynthesis to yield the deoxy variant (6) of
1. The stereochemical elucidation in conjunction with com-
plementary bioinformatic analyses implied the unexpected
and quite unusual direct incorporation of d-Ala into the
thailandamide backbone. Furthermore, PKS domain analyses
(KR fingerprinting) guided us to the presumed immediate
product of the tha PKS. Using a low temperature isolation
protocol we indeed succeeded in isolating and fully character-
izing the highly unstable thailandamide B (2). However,
monitoring pathway dissection by HRMS and using an
engineered DTE mutant eventually revealed that E/Z isomers
are generated online by module 6, which seems to lack rigid
stereocontrol. This finding is particularly noteworthy since
the predicted stereochemical course of b-keto processing in
many other cases could not be correlated with the ultimate
double bond geometry in trans-AT PKS products. Finally, the
stereochemical analysis also provides strong support for
a model involving an iterative use of a KR domain. In sum,
these unexpected results not only point out intriguing
4
www.angewandte.org
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2012, 51, 1 – 6
These are not the final page numbers!

Angewandte
Chemie
[25] C. T. Calderone, Nat. Prod. Rep. 2008, 25, 845 – 853.
[26] D. Menche, Nat. Prod. Rep. 2008, 25, 905 – 918.
[27] M. Bock, K. Buntin, R. Mꢁller, A. Kirschning, Angew. Chem.
2008, 120, 2341 – 2344; Angew. Chem. Int. Ed. 2008, 47, 2308 –
2311.
[28] D. Menche, F. Arikan, O. Perlova, N. Horstmann, W. Ahlbrecht,
S. C. Wenzel, R. Jansen, H. Irschik, R. Mꢁller, J. Am. Chem. Soc.
2008, 130, 14234 – 14243.
[29] D. Janssen, D. Albert, R. Jansen, R. Mꢁller, M. Kalesse, Angew.
Chem. Int. Ed. 2007, 46, 4898 – 48901.
Angew. Chem. Int. Ed. 2012, 51, 1 – 6
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
5
These are not the final page numbers!

.
Angewandte
Communications
Communications
Polyketides
Decoded before decay: Cryptic and highly
unstable polyketide metabolites, thailan-
damides A and B, were isolated from
Burkholderia thailandensis, and their
absolute configurations fully elucidated
using a combination of chemical degra-
dation, bioinformatics, NMR spectrosco-
py, precursor-directed biosynthesis, and
analysis of pathway intermediates (see
scheme).
K. Ishida, T. Lincke,
C. Hertweck*
&&&& — &&&&
Assembly and Absolute Configuration of
Short-Lived Polyketides from Burkholderia
thailandensis
6
www.angewandte.org
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2012, 51, 1 – 6
These are not the final page numbers!