Inhibition of quorum sensing and biofilm formation in Vibrio harveyi by 4-fluoro-DPD; A novel potent inhibitor of AI-2 signalling

Article abstract of DOI:10.1039/c3cc49678c

(S)-4,5-Dihydroxypentane-2,3-dione [(S)-DPD, (1)] is a precursor for AI-2, a quorum sensing signalling molecule for inter- and intra-species bacterial communication. The synthesis of its fluoro-analogue, 4-fluoro-5-hydroxypentane- 2,3-dione (2) is reported. An intermediate in this route also enables a new, shorter synthesis of the native (S)-DPD. 4-Fluoro-DPD (2) completely inhibited bioluminescence and bacterial growth of Vibrio harveyi BB170 strain at 12.5 μM and 100 μM, respectively.

Full text of DOI:10.1039/c3cc49678c

ChemComm
View Article Online
View Journal
COMMUNICATION
Inhibition of quorum sensing and biofilm
formation in Vibrio harveyi by 4-fluoro-DPD;
a novel potent inhibitor of AI-2 signalling†
Cite this:DOI: 10.1039/c3cc49678c
Received 20th December 2013,
Accepted 7th March 2014
Manikandan Kadirvel,‡ab Fariba Fanimarvasti,‡ac Sarah Forbes,a Andrew McBain,a
John M. Gardiner,*d Gavin D. Brown*b and Sally Freeman*a
DOI: 10.1039/c3cc49678c
www.rsc.org/chemcomm
(S)-4,5-Dihydroxypentane-2,3-dione [(S)-DPD, (1)] is a precursor for AI-2, a
quorum sensing signalling molecule for inter- and intra-species bacterial
communication. The synthesis of its fluoro-analogue, 4-fluoro-5-
hydroxypentane-2,3-dione (2) is reported. An intermediate in this route
also enables a new, shorter synthesis of the native (S)-DPD. 4-Fluoro-DPD
(2) completely inhibited bioluminescence and bacterial growth of Vibrio
harveyi BB170 strain at 12.5 lM and 100 lM, respectively.
Biofilms are a common cause of persistent bacterial infections, are
often recalcitrant to antimicrobial therapy and are rarely resolved via
host immune defence mechanisms.1–4 Biofilm formation is regulated
through cell-to-cell communication between bacteria via the release
of autoinducer signalling molecules, this process is referred to as
quorum sensing (QS).5,6 AI-2 acts as a universal signalling molecule
and is present in more than 70 species of bacteria.7 Regulation of AI-2
has been shown to play a significant role in biofilm formation in
many bacterial species.8–11 The ability to disrupt this signalling
process and possibly prevent bacterial biofilm formation may there-
fore be advantageous in the treatment or prevention of infectious
disease. Previous synthetic AI-2 analogues have shown to have an
inhibitory effect on biofilm formation in Vibrio harveyi and E. coli11
as well as QS associated pyocyanin production in Pseudomonas
aeruginosa, due to alterations in gene expression after exposure to
extracellular AI-2.12 The structures of (S)-DPD (1), and its boronate
complex exist in equilibria of hydrated and cyclised forms in solution
(Scheme 1).13,14 Vibrio harveyi, an indicator bacterium which forms
Scheme 1 Autoinducer AI-2: acyclic and cyclic forms of (S)-DPD (1),16
and its borate complex; F-DPD (2).
the 2,3-borate diester of the hydrated a-anomer of DPD, exhibits
bioluminescence properties.15–17
Compounds that interfere with QS may provide a strategy for
novel antibacterials.18 Previously we reported a new synthesis and
bioluminescence effect of the parent DPD.19 Here, the synthesis of
the novel 4-fluoro analogue of DPD (2, F-DPD) is reported, and
shown to act as a powerful suppressor of bioluminescence and
displays potent antibacterial activity. Fluorine is a common isosteric
and isoelectronic substitution for a hydroxyl group, the differences
being that F is only a H bond acceptor and the F–H bond is weaker
than O–H. F-DPD (2) may be helpful in understanding the mole-
cular mechanisms of AI-2 based quorum sensing. We aimed to
investigate whether (2) inhibits the bioluminescence, growth and
biofilm formation of Vibrio harveyi.
The synthesis of the novel F-DPD, (2) is shown in Scheme 2. The key
intermediates, (R)-1-(benzyloxy)-pent-3-yn-2-ol (8) and (R)-1-(4-methoxy-
benzyloxy)-pent-3-yn-2-ol (9), were prepared from (R)-glycidol (3)
(Scheme 2a).20–22 In addition, this intermediate (9) also enabled us to
develop an improved, shorter synthesis of the native (S)-DPD, wherein
conversion of (9) into isopropylidene (10), followed by oxidation
provides dione (11) which we previously reported was trivially converted
into (S)-DPD under mild-acid hydrolysis19 (Scheme 2b).
a Manchester Pharmacy School, University of Manchester, M13 9PT, UK.
E-mail: sally.freeman@manchester.ac.uk; Tel: +44 (0)161 275 2366
b Wolfson Molecular Imaging Centre, University of Manchester, M20 3LJ, UK.
E-mail: gavin.d.brown@manchester.ac.uk
c Bahai Institute for Higher Education, Iran
d School of Chemistry, MIB, University of Manchester, M1 7DN, UK.
E-mail: john.m.gardiner@manchester.ac.uk
† Electronic supplementary information (ESI) available: Experimental details
including synthesis, NMR spectra, Mass and GC chromatogram and microbiolo-
gical methods. See DOI: 10.1039/c3cc49678c
Two approaches towards the synthesis of F-DPD (2) were pursued,
the key step in each being the use of the deoxyfluorinating agent,
Xtal-Fluor, to replace an hydroxyl group with a fluorine atom. In the
‡ Equal contribution to the research; MK & FF completed the synthesis and S
Forbes the microbiology.
This journal is ©The Royal Society of Chemistry 2014
Chem. Commun.

View Article Online
Communication
ChemComm
To circumvent the vinylic migration, a second approach moved
the fluorination step to the alkyne precursors (8) and (9). Fluorina-
tion of (ꢀ)-(8) or (ꢀ)-(9) with Xtal-Fluor-E, triethylamine trihydro-
fluoride and triethylamine at ꢁ72 1C gave the desired propargylic
fluoro compounds (ꢀ)-(16) and (ꢀ)-(17) in 73% and 61% yield,
respectively. Subsequent oxidative cleavage of (ꢀ)-(17) with NaIO4
and RuO2 gave a-diketone, 5-(4-methoxybenzyloxy)-4-fluoropentane-
1
2,3-dione (ꢀ)-(19) in 82% (Scheme 2d). The H NMR was readily
assigned,23 however further structural confirmation was provided
by conversion into the quinoxaline derivative (ꢀ)-(21) by reaction
of diketone (ꢀ)-(19) with 1,2-phenylenediamine, and subsequent
deprotection of (ꢀ)-(20) with DDQ (Scheme 2e). Quinoxaline (ꢀ)-(21)
was characterised by 1H, 13C and 19F NMR spectroscopy.24 On
attempting to complete an enantiospecific synthesis of diketone
(18), chiral GC of (18) showed racemisation of the product, attributed
to facile enolisation, therefore (2) could only be made as a racemate.
Thus, rationalising the selection of a route from racemic (8) or (9).
The oxidative deprotection of (ꢀ)-(19) with DDQ gave F-DPD
(ꢀ)-(2) as an equilibrium mixture of non-hydrated (cyclic and acyclic)
and hydrated (cyclic) compounds (Scheme 2c). GCMS showed two
peaks for F-DPD (2): retention time (Rt) 6.77 min, m/z [134.0371]. and
(Rt) 10.25 min, m/z 152.0 in a ratio of 1 : 3, consistent with the non-
hydrated and hydrated forms of F-DPD (ꢀ)-(2). Despite being a low
1
molecular weight compound, the H NMR spectrum of (ꢀ)-(2) was
very complex, attributed also to the cyclic forms existing as diastereo-
isomers. The 1H NMR was similar to that for (S)-DPD (1),16,17,19
however it was further complicated by H–F coupling. For acyclic
F-DPD (ꢀ)-(2), the 1H NMR spectrum included a ddd at 5.61 ppm for
the CHF group with coupling constants of 2JH–F 55.6, 3JH4–H5 7.0 and
3JH4–H5 4.0 Hz. For the diastereoisomeric non-hydrated cyclic forms
of F-DPD (ꢀ)-(2), the CHF groups appeared as a ddd at 5.09 ppm
3
3
(2JF–H 54.7, JH4–H5 5.3, JH4–H5 2.3 Hz) and a dt at 4.47 ppm
3
(2JF–H 46.5, JF–H5 5.1 Hz). For non-hydrated F-DPD (ꢀ)-(2), the
1H-coupled 19F NMR spectrum showed peaks at ꢁ202.7 ppm
(acyclic), ꢁ192.5 ppm (cyclic) and ꢁ186.4 ppm (cyclic). In the
1H-decoupled 19F NMR spectrum, the diastereoisomeric hydrated
cyclic forms were observed as ddd at ꢁ175.1 ppm (2JF–H 58.5, 3JF–H5
31.1, 3JF–H5 27.1 Hz) and at ꢁ183.2 ppm (2JF–H 57.2, 3JF–H5 35.4, 3JF–H5
21.8 Hz) in a ratio of 1 : 3. The spectroscopic data for each form of
(ꢀ)-(2), is given in the ESI.†
Scheme 2 Synthesis of F-DPD (2): (a) BnBr, NaH, dry DMF; (b) 4-MPMCl,
NaH, TBAI, dry DMF; (c) lithium acetylide ethylenediamine complex (90%),
dry DMSO; (d) potassium-t-butoxide, dry DMSO: (e) DDQ, DCM : H2O
[10 : 1]; (f) 2,2-dimethoxypropane, dry DMF, con. H2SO4 (cat); (g) RuO2ꢂ
H2O, NaIO4, CCl4 : MeCN: H2O [2 : 2 : 3]; (h) LiAlH4, dry THF; (i) XtalFluor-E,
(Et)3Nꢂ3HF, TEA, ꢁ72 1C; (j) 1,2-phenylenediamine, MeOH.
Concentrations equal or greater than 12.5 mM F-DPD (2) resulted
in greater than a 10 000-fold decrease in luminescence production in
Vibrio harveyi, which corresponds to greater than 99.99% reduction
(Fig. 1). Inhibition of luminescence production was also evident at
lower F-DPD (2) concentrations albeit to a lesser extent. At the lowest
test concentration (0.8 mM), luminescence was shown to be reduced
first pathway (Scheme 2b), the benzyl analogue (8) was reduced to by 37%. Vibrio harveyi planktonic growth was completely inhibited at
the alkene (12), but on reaction with Xtal-Fluor-E, the product was 100 mM and 200 mM F-DPD (ꢀ)-(2) (Fig. 2). Slower growth of the
not the desired (E)-1-(benzyloxy)-2-fluoropent-3-ene (13) (Scheme 2c). bacterium, resulting in an increase in lag phase and delay in
The 19F NMR spectrum confirmed the formation of a fluorinated stationary phase, was evident at 12.5 mM, 25 mM and 50 mM F-DPD
compound, but the 1H NMR data supported a double bond migra- (ꢀ)-(2). When compared to the untreated culture (0 mM) changes in
tion, confirmed by the presence of a dd for the methyl group (C-5) at bacterial growth kinetics became more pronounced at higher F-DPD
1.34 ppm (3JF–H5 23.4 Hz, 3JH4–H5 6.6 Hz). This is consistent with the (ꢀ)-(2) concentrations. Inhibition of biofilm formation in Vibrio
i
formation of (15), attributed to an SN intramolecular substitution harveyi showed a clear dose response to increasing concentrations
reaction via intermediate (14), although by 1H NMR the compound of F-DPD (ꢀ)-(2). At 200 mM, biofilm formation was reduced by over
was not pure (Scheme 2c).
90% in comparison to the untreated control (Fig. 3).
Chem. Commun.
This journal is ©The Royal Society of Chemistry 2014

View Article Online
ChemComm
Communication
activity at 100 mM and showed a pronounced ability to disrupt
luminescence production in Vibrio harveyi, which is an AI-2
signalling-mediated event. Furthermore F-DPD (ꢀ)-(2) has been
shown to directly disrupt biofilm formation in the bacterium,
possibly due to interference in the quorum sensing process. These
data thus show F-DPD (ꢀ)-(2) to be an effective novel antibacterial
and anti-biofilm agent in Vibrio harveyi.
FF thanks the Bahai Institute for Higher Education, Iran.
Prof. Nicola Tirelli and Arianna Gennari are thanked for their
valuable assistance with the luminescence plate reader. Elena
Bichenkova is thanked for NMR support. Soraya Alnabulsi and
Fig. 1 Fold decrease in luminescence in V. harveyi BB170 grown in the
presence of F-DPD (2) in comparison to an untreated control at peak
luminescence. F-DPD (2) concentrations ranging from 0.2 mM to 6.25 mM Biljana Arsic for valuable assistance. Mass spectra were
did not affect specific bacterial growth rate or productivity in batch culture
recorded at the School of Chemistry, University of Manchester.
(see Fig. 2). Error bars show standard deviation of biological replicates, n = 3.
Notes and references
1 J. W. Costerton, Trends Microbiol., 2001, 9, 50.
2 R. O. Darouiche, New Engl. J. Med., 2004, 350, 1422.
3 P. Gilbert and A. J. McBain, Am. J. Infect. Control, 2001, 29, 252.
4 P. Gilbert, T. Maira-Litran, A. J. McBain, A. H. Rickard and
F. W. Whyte, Adv. Microb. Physiol., 2002, 46, 203.
5 D. G. Davies, M. R. Parsek, J. P. Pearson, B. H. Iglewski,
J. W. Costerton and E. P. Greenberg, Science, 1998, 280, 295.
6 K. F. Kong, C. Vuong and M. Otto, Int. J. Med. Microbiol., 2006, 296, 133.
7 J. B. Sun, R. Daniel, I. Wagner-Dobler and A. P. Zeng, Evol. Biol., 2004, 4.
8 A. F. G. Barrios, R. J. Zuo, Y. Hashimoto, L. Yang, W. E. Bentley and
T. K. Wood, J. Bacteriol., 2006, 188, 305.
9 J. Li, C. Attila, L. Wang, T. K. Wood, J. J. Valdes and W. E. Bentley,
J. Bacteriol., 2007, 189, 6011.
10 S. A. Rice, K. S. Koh, S. Y. Queck, M. Labbate, K. W. Lam and
S. Kjelleberg, J. Bacteriol., 2005, 187, 3477.
11 V. Roy, M. T. Meyer, J. A. I. Smith, S. Gamby, H. O. Sintim, R. Ghodssi
Fig. 2 Planktonic growth of V. harveyi BB170 grown in the presence of
F-DPD (2). F-DPD (2) concentrations ranging between 0.2 mM and 200 mM
were tested. For clarity the following selected concentrations have been
included in this figure: 0 mM (black diamond), 12.5 mM (grey circle), 25 mM
(white circle), 50 mM (grey triangle), 100 mM (black square) and 200 mM (black
cross). F-DPD (2) concentrations ranging between 0.2 mM and 6.25 mM did not
affect specific bacterial growth rate or productivity in batch culture (data not
shown). Error bars show standard deviation of biological replicates, n = 3.
and W. E. Bentley, Appl. Microbiol. Biotechnol., 2013, 97, 2627.
12 K. M. Duan, C. Dammel, J. Stein, H. Rabin and M. G. Surette, Mol.
Microbiol., 2003, 50, 1477.
13 S. T. Miller, K. B. Xavier, S. R. Campagna, M. E. Taga,
M. F. Semmelhack, B. L. Bassler and F. M. Hughson, Mol. Cells,
2004, 15, 677.
14 X. Chen, S. Schauder, N. Potier, A. Van Dorsselaer, I. Pelczer,
B. L. Bassler and F. M. Hughson, Nature, 2002, 415, 545–549.
15 M. F. Semmelhack, S. R. Campagna, C. Hwa, M. J. Federle and
B. L. Bassler, Org. Lett., 2004, 6, 2635.
16 M. F. Semmelhack, S. R. Campagna, M. J. Federle and B. L. Bassler,
Org. Lett., 2005, 7, 569.
17 M. M. Meijler, L. G. Hom, G. F. Kaufmann, K. M. McKenzie, C. Sun,
J. A. Moss, M. Matsushita and K. D. Janda, Angew. Chem., Int. Ed.,
2004, 43, 2106.
18 M. Guo, S. Gamby, Y. Zheng and H. O. Sintim, Int. J. Mol. Sci., 2013,
14, 17694.
19 M. Kadirvel, W. T. Stimpson, S. Moumene-Afifi, B. Arsic, N. Glynn,
N. Halliday, P. Williams, P. Gilbert, A. J. McBain, S. Freeman and
J. M. Gardiner, Bioorg. Med. Chem. Lett., 2010, 20, 2625.
20 L. F. Walker, A. Bourghida, S. Connolly and M. Wills, J. Chem. Soc.,
Perkin Trans. 1, 2002, 965.
21 K. C. Nicolaou, Y. Li, K. Sugita, H. Monenschein, P. Guntupalli,
H. J. Mitchell, K. C. Fylaktakidou, D. Vourloumis, P. Giannakakou
and A. O’Brate, J. Am. Chem. Soc., 2003, 125, 15443.
22 Y. Kiyotsuka and Y. Kobayashi, J. Org. Chem., 2009, 74, 7489.
Fig. 3 Crystal violet biofilm assay of V. harveyi BB170 in the presence of
F-DPD (2). F-DPD (2) concentrations ranging from 0.2 mM and 6.25 mM did not
affect specific bacterial growth rate or productivity in batch culture (see Fig. 2).
Error bars show standard deviation of biological replicates, n = 3.
1
23 H NMR (400 MHz, CDCl3) of 5-(40-methoxybenzyloxy)-4-fluoro-
pentane-2,3-dione (19); d 7.19 (2H, d, J 8.8 Hz, H–Ar), 6.88 (2H,
dm, J 8.8 Hz, H–Ar), 5.72 (1H, ddd, 2J4-F 55.6, 3J4–5a 4.1, 3J4–5b 2.0 Hz,
H-4), 4.52 (d, 1H, J 11.5 Hz, CH2-Ar), 4.40 (d, 1H, J 11.5 Hz, CH2-Ar),
4.11 (1H, ddd, 3J5a–F 34.9, 2J5a3–5b 11.7, 3J5a–4 4.1 Hz, H-5a), 3.90 (1H,
3
2
ddd, J5b–F 21.2, J5b–5a 11.7, J5b–4 2.0 Hz, H-5b), 3.81 (3H, s, OMe),
2.35 (3H, s, Me). 19F NMR (375 MHz, CDCl3) d ꢁ201.4.
A short synthesis of F-DPD (ꢀ)-(2) (alongside divergence to an
improved synthesis of the native (S)-DPD) an isosteric analogue of 24 1H NMR (400 MHz, CDCl3) of 2-fluoro-2-(3-methylquinoxalin-2-yl)-
ethanol (21); d 8.04 (2H, dd, 2JH–H 8.4, 4JH–H 1.2 Hz, H–Ar), 7.81–7.71
quorum sensing DPD, is reported. Its quorum sensing properties to
(2H, m, H–Ar), 5.84 (1H, dt, 2JH–F 46.5, 3JH–H 4.7 Hz, H-2), 4.48–4.32
(m, 2H, H-1), 2.87 (3H, d, J 2.3 Hz, Ar–Me). 19F NMR (375 MHz,
CDCl3) d ꢁ188.1. See ESI† for 13C and 19F NMR data.
affect bioluminescence and bacterial growth of Vibrio harveyi BB170
strain were evaluated. F-DPD (ꢀ)-(2) displayed direct antibacterial
This journal is ©The Royal Society of Chemistry 2014
Chem. Commun.