Biosynthesis of 2-Alkyl-4(1H)-Quinolones in Pseudomonas aeruginosa: Potential for Therapeutic Interference with Pathogenicity

Full text of DOI:10.1002/cbic.201100014

DOI: 10.1002/cbic.201100014
Biosynthesis of 2-Alkyl-4(1H)-Quinolones in Pseudomonas aeruginosa:
Potential for Therapeutic Interference with Pathogenicity
[
a]
[b]
[a]
[a]
[b]
[a]
Dominik Pistorius, Angelika Ullrich, Simon Lucas, Rolf W. Hartmann, Uli Kazmaier, and Rolf Mꢀller*
Pseudomonas aeruginosa is a ubiquitous Gram-negative bacte-
rium capable of surviving in a broad range of natural environ-
ments and known to be involved in infectious diseases of vari-
ous hosts. In humans, the opportunistic pathogen is one of
the leading causes for nosocomial infections in immuno-com-
promised patients and is responsible for chronic lung infec-
[
1]
tions in the majority of cystic fibrosis patients. The ability of
P. aeruginosa to adapt to different environments and lifestyles
is closely related to its ability to coordinate the survival strat-
egy of a population by so-called quorum-sensing (QS) systems.
QS is based on the production and release of small signaling
molecules, called autoinducers, that increase in concentration
as a function of cell density and activate corresponding tran-
scriptional regulators after a threshold concentration has been
Scheme 1. Biosynthetic pathways to DHQ, HHQ, and PQS. The nature of the
accepted b-ketodecanoyl moiety is unknown to date. Possible substrates are
b-ketodecanoic acid (1), b-ketodecanoyl-CoA (2) and b-ketodecanoyl-ACP.
[
2]
reached. Three different QS systems are known from P. aeru-
ginosa. The las and rhl systems use acyl-homoserine lactone
(
AHL) autoinducers and belong to the LuxI/LuxR-type systems
[3]
that are widespread among Gram-negative bacteria. The
third QS system is rather unique and restricted to particular
Pseudomonas and Burkholderia strains. Therein 2-alkyl-4(1H)-
quinolones (AQ) autoinducers such as 2-heptyl-3-hydroxy-
inhibit PQS biosynthesis. This treatment led to a significant
[11]
increase in survival in comparison to the control group.
To further explore this target it is necessary to understand
the details of PQS formation in the pathogen. It is known that
4(1H)-quinolone (the Pseudomonas quinolone signal: PQS) and
[12]
its direct precursor 2-heptyl-4(1H)-quinolone (HHQ) are used
HHQ biosynthesis absolutely requires the genes pqsA–D, en-
coding an anthranilate:coenzyme A (CoA) ligase (pqsA) and
three b-ketoacyl-acyl carrier protein synthase III (KAS III) homo-
[
4]
(
see Scheme 1). The pqs system is involved in the regulation
[5]
of P. aeruginosa virulence such as pyocyanin biosynthesis,
[
6]
[13]
biofilm formation and maturation, the production of exo-
products like elastase, alkaline proteases, rhamnolipids, and
logues. An additional gene (pqsH), located apart from pqsA–
D, is responsible for the hydroxylation of HHQ to form PQS.
Feeding studies in vivo have demonstrated that HHQ most
likely arises from “head-to-head” condensation of an anthrani-
[
7]
[8]
hydrogen cyanide, and the expression of efflux pumps. Fur-
ther, PQS itself can down-regulate the host-innate immune
[
9]
[14]
response. The sum of these effects makes the pqs system a
highly attractive target for drug development to interfere with
P. aeruginosa pathogenicity and biofilm formation. The general
validity of this approach is supported by the results of several
infection models in which PQS-deficient mutants show a re-
loyl precursor and a b-keto fatty acid derivative (Scheme 1).
However, the details of the enzymatic mechanism of this reac-
tion and the nature of the b-keto fatty acid remained elusive.
Furthermore, it has been shown in vitro that PqsA and PqsD
catalyze the formation of 2,4-dihydroxyquinoline (DHQ), anoth-
er secondary metabolite of P. aeruginosa. In DHQ biosynthesis
PqsA activates anthranilic acid to anthraniloyl-CoA which is
loaded to the active-site cysteine (C112) of PqsD, which itself
catalyzes the decarboxylative Claisen condensation with ma-
[7,10]
duced pathogenicity compared to P. aeruginosa wild type.
A reduced pathogenicity was also observed in a mouse infec-
tion model when animals infected with the wild-type strain
were treated with halogenated anthranilic acid derivatives that
[15]
lonyl-CoA. Based on the structural similarity between DHQ
and HHQ, we reasoned that PqsD might be involved in a simi-
lar condensation reaction in HHQ biosynthesis. To prove this
hypothesis, we heterologously expressed PqsD from strain
PA14 in Escherichia coli and purified the enzyme for biochemi-
cal characterization in vitro. Anthraniloyl-CoA and three poten-
tial b-keto acid derivatives were chemically synthesized as sub-
strates, including b-ketodecanoic acid (1), b-ketodecanoyl-CoA
[
a] D. Pistorius, Dr. S. Lucas, Prof. Dr. R. W. Hartmann, Prof. Dr. R. Mꢀller
Helmholtz-Institut fꢀr Pharmazeutische Forschung Saarland
Helmholtz Zentrum fꢀr Infektionsforschung und
Pharmazeutische Biotechnologie, Universitꢁt des Saarlandes
Postfach 151150, 66041 Saarbrꢀcken (Germany)
Fax: (+49)681-302-70202
E-mail: rom@mx.uni-saarland.de
[b] Dr. A. Ullrich, Prof. Dr. U. Kazmaier
Institut fꢀr Organische Chemie, Universitꢁt des Saarlandes
Postfach 151150, 66041 Saarbrꢀcken (Germany)
Fax: (+49)0681-302-2409
(
2), and b-ketodecanoyl-N-acetylcysteamine thioester (3) as
mimics of the hypothetical ACP-bound substrate (for details
see the Supporting Information). The in vitro reaction con-
tained recombinant PqsD, anthraniloyl-CoA, and one of the
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/cbic.201100014.
8
50
ꢁ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemBioChem 2011, 12, 850 – 853

three b-keto acid substrates, and formation of the product
HHQ was analyzed by high-performance liquid chromatogra-
phy coupled to high-resolution mass spectrometry (HPLC-
HRMS). HPLC-HRMS analysis confirmed HHQ formation in all
three reactions by comparison of retention time, accurate
mass, and fragmentation pattern to the synthetic reference
compound (Figure 1), though high concentrations and longer
ure S1. These results challenge our previous hypothesis of
HHQ formation wherein ACP or CoA-bound b-ketodecanoyl
[14]
derivatives were discussed as substrates for HHQ formation.
Recent structural analysis of PqsD by crystallography has lent
further support to our findings, as the CoA or ACP-bound b-ke-
todecanoyl substrates could not be modeled into the active-
[16]
site cavity. However, the deduced argumentation that PqsD
might not be involved in HHQ formation has been clearly dis-
proven by our results. The initially observed activity employing
2
or 3 most likely arises from the nonenzymatic hydrolysis of
the thioester upon longer incubation times to generate 1 as a
direct substrate. Indeed, we could show spontaneous hydroly-
sis of the thioester under the assay conditions (for more infor-
mation see the Supporting Information). Based on the signifi-
cant difference in k /K
for the substrates malonyl-CoA
cat m app.
and 1, which is not reflected by the amounts of DHQ and HHQ
found in P. aeruginosa cultures, and the fact that PqsB and
PqsC are absolutely required for HHQ biosynthesis, it seems
likely that these two enzymes produce a somehow activated
form of 1. This so-far-unknown intermediate probably exhibits
a higher affinity to PqsD than 1.
P. aeruginosa produces a vast array of AQs in addition to
HHQ and PQS. In total more than 50 different AQs have been
reported that vary in the substitution at position 3 and the
length and saturation of their alkyl side chain, and most are
[17]
present only in small amounts.
To address the question
whether PqsD is promiscuous concerning the alkyl chain
length of its b-keto fatty acid substrates and thereby responsi-
ble for the diversity of AQs in P. aeruginosa, we tested several
short-chain b-keto fatty acids in the assay. b-Ketobutanoic acid
did not serve as a substrate of PqsD but incubation with b-ke-
topentanoic and b-ketohexanoic acid yielded the respective
short chain AQs, 2-ethyl-4(1H)-quinolone and 2-propyl-4(1H)-
quinolone, which were identified by HPLC-HRMS based on
their accurate masses and characteristic fragmentation patterns
Figure 1. Identification of HHQ formed in vitro. A) HPLC–HRMS analysis (ex-
tracted ion chromatogram (EIC) 244.16-244.18) of synthetic HHQ reference.
B) HRMS analysis of synthetic HHQ reference at retention time 4.87 min.
2
C) HRMS fragmentation pattern of synthetic HHQ reference at retention
time 4.87 min. D–F) Same analyses for in vitro assay containing anthraniloyl-
CoA, PqsD, and b-ketodecanoic acid. All data were obtained on a LTQ Orbi-
trap hybrid FT mass spectrometer, (Thermo Scientific, Dreieich, Germany).
[17]
(Figure S2). These data imply that the broad substrate specif-
icity of PqsD is indeed responsible for the variation in length
(
and likely also for the differences in saturation) of the alkyl
side chain observed for AQs in P. aeruginosa.
incubation times were required for compounds 2 and 3 in
comparison to the free acid. Thus, we provide direct experi-
mental proof that PqsD catalyzes the condensation reaction
between the anthraniloyl and the b-ketodecanoyl moieties in
HHQ biosynthesis.
Concerning the molecular mechanism of HHQ formation by
PqsD, two possible scenarios are proposed that differ concern-
ing the sequence of reactions (Scheme 2). One scenario ex-
pects the reaction to start with the formation of an imine inter-
mediate of anthranilate and b-ketodecanoate followed by de-
carboxylative Claisen condensation to form the heterocycle
and release the final product from the enzyme. The second
possibility assumes the sequence of reactions to start with the
decarboxylative “head-to-head” condensation succeeded by
heterocycle formation via intramolecular condensation of the
aromatic amine with the aliphatic carbonyl group under elimi-
nation of water, presumably assisted by PqsD in anology to
Kinetic parameters of the PqsD reactions were determined
by quantifying the product HHQ by HPLC–HRMS. We focused
on varying the concentrations of 1 with a fixed concentration
of the second substrate anthraniloyl-CoA. Under these condi-
tions PqsD exhibited an apparent K of 598.5Æ106 mm and a
m
À1 À1
k /K
cat m app.
of 14 s
m
for compound 1. Kinetic parameters for
2
and 3 could not be obtained under identical conditions. To
[18]
validate that the PqsD used in this study was fully functional,
we also determined kinetic parameters of DHQ formation cata-
the reactions catalyzed by chalcone synthases.
To discern
these two possibilities, we designed experimental setups to
stop the reactions after the first step of the respective mecha-
nism. In order to trap the putative enzyme-bound imine inter-
mediate, a reaction mixture containing PqsD, anthraniloyl-CoA,
lyzed by PqsD. The apparent K and k /K
of malonyl-CoA
m
4
cat m app.
À1 À1
were 48.7Æ7 mm and 6.9ꢂ10 s m , respectively, which is
consistent with previous findings by Zhang et al. (K =104Æ
m
[
15]
3
7 mm). For more details of the kinetic parameters see Fig-
and 1 was quenched by the addition of NaBH , thereby con-
4
ChemBioChem 2011, 12, 850 – 853
ꢁ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chembiochem.org
851

ry concentration (IC ) of each
50
tested compound can be deter-
mined. Based on the structural
and functional homologies of
PqsD and bacterial FabH, well-
known FabH inhibitors have
[19]
been resynthesized to experi-
mentally validate our in vitro ap-
proach. The obtained IC50 values
of 65 mm for 2-(4-bromo-3-dieth-
ylsulfamoyl-benzoylamino)-ben-
zoic acid (4) and 35 mm for 2-[(2-
phenoxybiphenyl-4-carbonyl)a-
mino]benzoic acid (5; Figure 2)
provide evidence that the
chosen
compounds
show
modest inhibitory activity to-
wards PqsD mediated HHQ bio-
synthesis. However, they clearly
demonstrate the feasibility of
the in vitro screening for PqsD
inhibitors.
Scheme 2. Proposed reaction mechanisms of HHQ biosynthesis. A) Starting with imine formation. B) Starting with
decarboxylative Claisen condensation.
verting the imine intermediate to the stable secondary amine
and presumably preventing subsequent reaction steps. When
PqsD was analyzed for the expected change in mass by HPLC-
MS (38.439 kDa for native PqsD and 38.728 kDa for the re-
duced secondary amine intermediate) no evidence for the for-
mation of such an intermediate could be found. To test for the
second mechanism, benzoyl-CoA instead of anthraniloyl-CoA
was incubated with PqsD and 1 (or alternatively malonyl-CoA)
to prevent heterocycle formation after decarboxylative Claisen
condensation. Indeed, HPLC–HRMS analysis revealed masses of
247.1690 Da and 165.0543 Da to be present in the assay, which
correspond to the expected products 1-phenyldecane-1,3-
dione for the reaction with 1 and b-keto-3-phenylpropanoic
acid for the reaction with malonyl-CoA, respectively. However,
the amounts of these products were too little to be isolated
for structural determination by NMR spectroscopy. Therefore
we synthesized the postulated products as reference com-
pounds for comparison, and unambiguously confirmed the
identities of the intermediates observed in the in vitro reac-
tions by comparison of retention time, accurate mass, and
fragmentation pattern in HPLC–HRMS (Figure S3). This result
provides first experimental evidence for the hypothesis that
DHQ and HHQ biosynthesis proceed via decarboxylative Clais-
en condensation of the b-keto fatty acid and anthraniloyl-CoA,
followed by heterocycle formation via intramolecular conden-
sation and elimination of the leaving group. In order to target
AQ biosynthesis by inhibition of PqsD function, we next devel-
oped an assay to screen for PqsD inhibitors in vitro. For this
purpose PqsD, anthraniloyl-CoA, 1, and increasing concentra-
tions of the potential inhibitor were mixed and incubated for a
defined time period. HHQ production in this assay was quanti-
fied and compared to that observed in control reactions with-
out inhibitor. By using this approach the half-maximal inhibito-
Figure 2. Chemical structures and respective IC50 values (indicated by
dashed lines) of the FabH inhibitors A) 4 IC50 =65 mm and B) 5 IC50 =35 mm,
determined in our in vitro PqsD-inhibitor assay. Data represent average
values taken from three independent experiments.
The advantages of this assay are its unambiguity, because it
bypasses the complex and interconnected QS networks in
P. aeruginosa, the possibility to work without the opportunistic
pathogen, the fast rate with which results are obtained and
the potential to adapt it to high-throughput screening. Further,
it is possible to test compounds with inherent antibiotic activi-
ty, for example, FabH inhibitors, for inhibition of HHQ forma-
tion which is neither possible in P. aeruginosa, nor in the estab-
[20]
lished in vivo reporter-gene-coupled assays. Even though an
antibiotic effect of a quorum-sensing inhibitor might not be
desirable regarding the abetting development of bacterial re-
sistance, it is highly beneficial to have an assay available that
allows screening of compounds with inherent antibiotic activi-
8
52
www.chembiochem.org
ꢁ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemBioChem 2011, 12, 850 – 853

ty in the initial hit-finding phase of compounds with inhibitory
activity against PqsD.
[7] E. Deziel, S. Gopalan, A. P. Tampakaki, F. Lepine, K. E. Padfield, M. Saucier,
G. P. Xiao, L. G. Rahme, Mol. Microbiol. 2005, 55, 998–1014.
[
8] S. P. Diggle, K. Winzer, S. R. Chhabra, K. E. Worrall, M. Camara, P. Williams,
In conclusion, we have revealed the role of PqsD in HHQ
biosynthesis and gained valuable information concerning the
reaction mechanism and the nature of the processed sub-
strates. By reconstitution of HHQ biosynthesis we were able to
establish an assay that allows for in vitro screening of inhibi-
tors of PqsD mediated HHQ biosynthesis that might become a
valuable tool in the discovery of drugs targeting HHQ biosyn-
thesis in P. aeruginosa.
Mol. Microbiol. 2003, 50, 29–43.
[9] K. Kim, Y. U. Kim, B. H. Koh, S. S. Hwang, S. H. Kim, F. Lepine, Y. H. Cho,
G. R. Lee, Immunology 2010, 129, 578–588.
10] H. Cao, G. Krishnan, B. Goumnerov, J. Tsongalis, R. Tompkins, L. G.
Rahme, Proc. Natl. Acad. Sci. USA 2001, 98, 14613–14618.
11] B. Lesic, F. Lepine, E. Deziel, J. Zhang, Q. Zhang, K. Padfield, M. H. Cas-
tonguay, S. Milot, S. Stachel, A. A. Tzika, R. G. Tompkins, L. G. Rahme,
PLoS Pathog. 2007, 3, 1229–1239.
[
[
[
[
[
12] L. A. Gallagher, S. L. McKnight, M. S. Kuznetsova, E. C. Pesci, C. Manoil, J.
Bacteriol. 2002, 184, 6472–6480.
13] J. P. Coleman, L. L. Hudson, S. L. McKnight, J. M. Farrow, M. W. Calfee,
C. A. Lindsey, E. C. Pesci, J. Bacteriol. 2008, 190, 1247–1255.
14] F. Bredenbruch, M. Nimtz, V. Wray, M. Morr, R. Mꢀller, S. Haussler, J. Bac-
teriol. 2005, 187, 3630–3635.
Acknowledgements
The authors would like to thank Dr. Yanyan Li, Dr. Anke Stein-
bach and Profs. Susanne Hꢁussler and Wulf Blankenfeldt for val-
uable comments on this manuscript and fruitful discussions.
[15] Y. M. Zhang, M. W. Frank, K. Zhu, A. Mayasundari, C. O. Rock, J. Biol.
Chem. 2008, 283, 28788–28794.
[
16] A. K. Bera, V. Atanasova, H. Robinson, E. Eisenstein, J. P. Coleman, E. C.
Pesci, J. F. Parsons, Biochemistry 2009, 48, 8644–8655.
[
17] F. Lepine, S. Milot, E. Deziel, J. X. He, L. G. Rahme, J. Am. Soc. Mass Spec-
trom. 2004, 15, 862–869.
Keywords: biosynthesis inhibitors · drug discovery · HHQ
formation · Pseudomonas aeruginosa · quorum sensing
[
[
18] I. Abe, H. Morita, Nat. Prod. Rep. 2010, 27, 809–838.
19] Z. Nie, C. Perretta, J. Lu, Y. Su, S. Margosiak, K. S. Gajiwala, J. Cortez, V.
Nikulin, K. M. Yager, K. Appelt, S. Chu, J. Med. Chem. 2005, 48, 1596–
[
[
[
[
1] C. Winstanley, J. L. Fothergill, FEMS Microbiol. Lett. 2009, 290, 1–9.
2] M. B. Miller, B. L. Bassler, Annu. Rev. Microbiol. 2001, 55, 165–199.
3] J. S. Dickschat, Nat. Prod. Rep. 2010, 27, 343–369.
4] E. C. Pesci, J. B. Milbank, J. P. Pearson, S. McKnight, A. S. Kende, E. P.
Greenberg, B. H. Iglewski, Proc. Natl. Acad. Sci. USA 1999, 96, 11229–
1
609.
20] M. P. Fletcher, S. P. Diggle, M. Camara, P. Williams, Nat. Protoc. 2007, 2,
254–1262.
[
1
11234.
[5] J. F. Dubern, S. P. Diggle, Mol. Biosyst. 2008, 4, 882–888.
[6] M. Allesen-Holm, K. B. Barken, L. Yang, M. Klausen, J. S. Webb, S. Kjelle-
berg, S. Molin, M. Givskov, T. Tolker-Nielsen, Mol. Microbiol. 2006, 59,
Received: January 7, 2011
1114–1128.
Published online on March 18, 2011
ChemBioChem 2011, 12, 850 – 853
ꢁ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chembiochem.org
853