Development of quinoline-based disruptors of biofilm formation against Vibrio cholerae

Article abstract of DOI:10.1021/ol400150z

Biofilm formation is a major cause of bacterial persistence in nosocomial infections, leading to extended treatment times and increased rates of morbidity and mortality. Despite this, there are currently no biofilm inhibitors approved for clinical use. Th

Full text of DOI:10.1021/ol400150z

ORGANIC
LETTERS
2013
Vol. 15, No. 6
1234–1237
Development of Quinoline-Based
Disruptors of Biofilm Formation Against
Vibrio cholerae
†
‡
†
†
‡
ꢀ
Brian Leon, Jiunn C. N. Fong, Kelly C. Peach, Weng Ruh Wong, Fitnat H. Yildiz,
and Roger G. Linington*,†
Department of Chemistry and Biochemistry and Department of Microbiology and
Environmental Toxicology, University of California, Santa Cruz, 1156 High Street,
Santa Cruz, California 95064, United States
rliningt@ucsc.edu
Received January 17, 2013
ABSTRACT
Biofilm formation is a major cause of bacterial persistence in nosocomial infections, leading to extended treatment times and increased rates of
morbidity and mortality. Despite this, there are currently no biofilm inhibitors approved for clinical use. The synthesis and biological evaluation of
a library of amino alcohol quinolines as lead compounds for the disruption of biofilm formation in Vibrio cholerae is now reported. Application of
selective metalꢀhalogen exchange chemistry installed both stereocenters in one step, to afford a simpler scaffold than the initial lead molecule,
with an EC50 < 10 μM.
Bacterial biofilms are surface-associated bacterial as-
semblages containing microbial cells encased in an extra-
cellular matrix of exopolysaccharides, proteins and DNA.
Biofilm formation is increasingly being recognized as a
major component of bacterial pathogenesis, with current
estimates suggesting that up to 75% of pathogenic micro-
bial infections are biofilm mediated.1 This is considered to
be a significant contributor to the development and persis-
tence of nosocomial (hospital acquired) infections. Biofilm-
associated infections are particularly difficult to clear
because cells within the matrix can enter a latent state that
reduces their susceptibility to traditional antibiotics. This
results in a significant increase in bacterial persistence,
leading to a concomitant increase in antibiotic treatment
times, and a subsequent acceleration in the development of
drug-resistance.2 Despite detailed investigation of several
compound series in recent years,3 there are still no com-
mercially available biofilm inhibitors that are approved for
clinical use.
To discover new therapeutic options for biofilm-associated
infections, we recently developed an image-based high
content screening platform for the discovery of bio-
film inhibitors in the model organism Vibrio cholerae.4
Previous screening of a 3080-member small molecule
library from the National Cancer Institute (NCI) revealed
29compoundswiththe ability toselectively disrupt biofilm
formation. Of these, two (NSC 13480, 3, EC50 = 11.1 μM;
NSC 305787, 4, EC50 = 10.6 μM) were closely related,
possessing a 2,4-disubstituted quinoline core and a
β-amino alcohol motif (Figure 1). Many FDA approved
drugs contain this privileged quinoline ring structure,5
† Department of Chemistry and Biochemistry.
‡ Department of Microbiology and Environmental Toxicology.
(1) Richards, J.; Melander, C. ChemBioChem 2009, 10, 2287–2294.
(2) Davies, D. Nat. Rev. Drug Discovery 2003, 2, 114–122.
(3) Worthington, R. J.; Richards, J. J.; Melander, C. Org. Biomol.
Chem. 2012, 10, 7457–7474.
(4) Peach, K. C.; Bray, W. M.; Shikuma, N. J.; Gassner, N. C.;
Lokey, R. S.; Yildiz, F. H.; Linington, R. G. Mol. BioSyst. 2011, 7, 1176–
1184.
(5) Solomon, V. R.; Lee, H. Curr. Med. Chem. 2011, 18, 1488–1508.
r
10.1021/ol400150z
Published on Web 03/05/2013
2013 American Chemical Society

suggesting that this scaffold could be a valuable starting
point for the development of new biofilm inhibitors.6
concise access to either racemic or enantiomerically pure
forms of the β-amino alcohol scaffold using the same set of
building blocks and chemical transformations, to allow for
future development of either racemic or chiral approaches
using the same synthetic methodology.
Scheme 1. Retrosynthetic Analysis
Retrosynthetically, we envisioned a CꢀC bond discon-
nection at the 2-position of the quinoline13 (derived from
Suzuki cross coupling chemistry) and a subsequent CꢀC
bond disconnection at the 4-position, which could be
installed via the regiospecific Grignard chemistry recently
developed by Knochel and co-workers14 (Scheme 1). It
was proposed that this methodology could be employed
to install the key β-amino alcohol motif in one step
from 2,4-dibromoquinoline and an appropriately protected
aryl/alkyl aldehyde. Because the predicted approach of the
two substrates in the Grignard step follows the Felkin Ahn
model, we hypothesized that the resulting products would
be created with full stereochemical control at these two new
stereogenic centers.
In this initial study, we elected to explore the effects of
perturbation of the β-amino alcohol (ring size, aromati-
city, N excision) and/or modification of the aryl group at
the 2-position (H-donors/acceptors and stereoelectronic
effects) on biofilm formation. To this end, we selected four
aldehyde substrates and four boronic acids from which to
create a 16-member screening library that examined the
effects of systematic variation of these factors at the two
key positions on the scaffold.
All library members were synthesized using the same
general synthetic scheme. As a representative example,
compound 5 was treated with i-PrMgCl•LiCl at ꢀ78 °C
and underwent metalꢀhalogen exchange exclusively at the
4-position (Figure 2). Nucleophilic addition to protected
aldehyde 6 proceeded in 67% yield, to afford alcohol 10 as
a single pair of enantiomers.15 Subsequent Suzuki cross-
coupling with boronic acid 14using the Pd(PPh3)4 catalyst,
followed by trityl deprotection with HCl, gave the desired
amino alcohol 26 in good overall yield. Determination of
the relative and absolute configurations of final com-
pounds was accomplished via a combination of X-ray
crystallographic analysis for compound 33 and comparisons
Figure 1. Initial lead compounds and related compounds.
The β-amino alcohol quinoline series are a well-studied
class of compounds originally developed for antimalarial
therapy by the Walter Reed Army Institute of Research as
synthetic analogues of the natural product quinine (1).7
This work ultimately lead to the development of meflo-
quine (2) for both prophylactic and therapeutic use against
malaria. Recent studies have shown that mefloquine is
effective against a number of other targets, including
Gram-positive bacteria. However, previous work has re-
ported that mefloquine and its derivatives are inactive
against Gram-negative bacteria.8
The original racemic synthesis of mefloquine in 1971
afforded the drug in five steps.9 More recently, attention
has shifted to the effects of individual enantiomers on both
antimalarial and neurotoxic activities,10 culminating in two
reports presenting enantioselective syntheses of the meflo-
quine scaffold in 2011.11 Each of these studies employed a
series of oxidation state manipulations and rearrange-
ments to control the installation of the contiguous stereo-
centers of the β-amino alcohol motif. Given that recent
studies indicate that the (ꢀ) form of the drug causes
pronounced neurological side effects,12 while the (þ) form
is responsible for the antimalarial activity, we were moti-
vated to design a synthetic approach that could provide
(6) Sintim, H. O.; Smith, J. A. I.; Wang, J.; Nakayama, S.; Yan, L.
Future Med. Chem. 2010, 2, 1005–1035.
(7) Croft, A. M. J. R. Soc. Med. 2007, 100, 170–174.
(8) Kunin, C. M.; Ellis, W. Y. Antimicrob. Agents Ch. 2000, 44, 848–
852.
(9) Ohnmacht, C. J.; Patel, A. R.; Lutz, R. E. J. Med. Chem. 1971, 14,
926–928.
(10) Chen, L. H.; Wilson, M. E.; Schlagenhauf, P. J. Amer. Med.
Assoc. 2007, 297, 2251–2263.
(11) (a) Jonet, A.; Dassonville-Klimpt, A.; Da Nascimento, S.; Leger,
J. M.; Guillon, J.; Sonnet, P. Tetrahedron:Asymmetry 2011, 22, 138–148.
(b) Knight, J. D.; Sauer, S. J.; Coltart, D. M. Org. Lett. 2011, 13, 3118–
3121.
(13) Lord, A.-M.; Mahon, M. F.; Lloyd, M. D.; Threadgill, M. D.
J. Med. Chem. 2009, 52, 868–877.
(14) Boudet, N.; Lachs, J.; Knochel, P. Org. Lett. 2007, 9, 5525–5528.
(15) Bejjani, J.; Chemla, F.; Audouin, M. J. Org. Chem. 2003, 68,
9747–9752.
(12) Fletcher, E. A. Uses of (þ)-Mefloquine for the Treatment of
Malaria. International PCT patent application PCT/GB98/00675, 1998.
Org. Lett., Vol. 15, No. 6, 2013
1235

of key 3JHH coupling constants between the hydroxy- and
amino-methine protons for compounds 26ꢀ33.
Figure 2. Synthetic approach to compounds 26ꢀ41.
Figure 3. Biofilm formation EC50 values for library members.
Values in μM.
Screening of both this 16-member library and 11 struc-
turally related analogues from the NCI and commercial
sources was performed using our standard in-house image-
based biofilm screen. Briefly, this screen examines the
ability of test compounds to inhibit the formation of
biofilms on solid surfaces by treatment of test wells with
compounds at the time of innoculation, followed by in-
cubation for 4.5 h, washing, and imaging of biofilms formed
on the surface of screening wells. The results of this analysis
revealed three main SAR conclusions:
(1) The β-amino alcohol moiety is essential for reducing
biofilm formation. This is evidenced by the lack of activity
for compounds 34ꢀ37 and 38ꢀ41, which contain respec-
tively the pyridinyl and phenyl rings in place of the
piperidinyl ring system, and the lack of the secondary
amine from the original lead compounds.
(2) The quinoline motif is required for activity but is
not in and of itself sufficient to impact biofilm formation.
This conclusion is supported by the lack of activity for
ephedrine (42), which contains the β-amino alcohol but lacks
the quinoline motif, and compounds 45, 49, and 50, which
contain the quinoline but lack the β-amino alcohol (Figure 3).
Chelation of free iron has been shown to directly
impact biofilm formation in Pseudomonas aeruginosa.16
These SAR results indicate that, in Vibrio cholerae, biden-
tate chelators such as compounds 42 and 50 have no direct
impact on biofilm formation.
(3) Substitution on the aryl ring at the 2-position of the
quinoline impacts biofilm formation. In particular, inclu-
sion of a pyridine ring eliminated activity in all cases, while
inclusion of either phenyl or 4-(methoxy)phenyl substituents
afforded compounds with weaker EC50 values than those
containing 4-(trifluoromethyl)phenyl substituents.
To be valuable in a clinical setting, lead compounds
should be capable of disrupting existing preformed bio-
films, in addition to inhibiting initial bacterial attach-
ment and biofilm formation on surfaces. To examine the
effectiveness of compound 26 at eliminating preformed
biofilms, we treated static cultures of V. cholerae with 26 at
both t = 0 (initial bacterial inoculation) and t = 5 h
(preformed biofilms). Figure 4 demonstrates that com-
pound 26 is capable of both preventing initial bacterial
attachment (Figure 4A) and inducing detachment of pre-
formed biofilms (Figure 4B).
There are few current examples of compounds capable
of inducing bacterial detachment once biofilm formation
is established,3 making this result particularly valuable
(16) Banin, E.; Vasil, M. L.; Greenberg, E. P. Proc. Natl. Acad. Sci.
U.S.A. 2005, 102, 11076–81.
1236
Org. Lett., Vol. 15, No. 6, 2013

challenging.17 In many cases, the only therapeutic options
are irrigation (for open wounds) and repeated courses of
antibiotic treatments, which are only marginally effective
and provide a fertile environment for the development of
antimicrobial resistance. The discovery that 26 is effective
at reducing bacterial cell densities for both initial and
preformed biofilms is therefore significant, as it offers the
promise for the development of new therapeutic options
for biofilm-mediated bacterial infections based on this
compound scaffold.
In addition to evaluation of these compounds on bio-
film formation and detachment, compounds 26ꢀ41 were
screened for antibiotic activity against a diverse panel of
clinically relevant bacterial pathogens using our recently
reported BioMAP screening platform.18 These results were
in line with previously published studies of quinoline
amino alcohols as antimicrobial agents, with activities in
the low micromolar range against Gram-positive strains
and little activity against Gram-negative organisms (Table S1).
In addition, growth curves were obtained for 26 at
50 μM, using both OD595 and colony forming unit (cfu)
counts as independent readouts for cell survival (Figure S2).
Under these experimental conditions, compound 26 was
shown to inhibit growth of V. cholerae, suggesting that this
compound hastheabilitytokillV. choleraecellsinboth the
biofilm and planktonic states.19
In conclusion, we report the synthetic development
of a new class of disruptors of biofilm formation in
V. cholerae. Application of recently established regio-
selective metalꢀhalogen exchange methodology led to
the creation of a library of synthetic analogues, one of
which possesses a simpler structure and improved
potency over the original lead compounds. This com-
pound (26) is able to disrupt biofilm formation at low
micromolar concentrations and to eliminate preformed
biofilms under static conditions. This quinoline-based
scaffold therefore represents a new avenue in the search
for therapeutics capable of eliminating biofilm-mediated
infections.
Figure 4. Confocal laser scanning microscopy images of top-
down (large panels) and side (small panels) projections of
biofilm structures formed by V. cholerae wild type at 5 or 7
and 24 h postinoculation. Compound 26 (50 μM) and DMSO
control (0.4%) were added either during inoculation (A) or 5 h
postinoculation (B). White bars represent 40 μm. Assays were
repeated with two biological replicates.
Acknowledgment. We thank R. S. Lokey and W. Bray
for assistance with biological screening through the UCSC
Chemical Screening Center and H. Fei and A. Oliver for
assistance with X-ray crystallographic analysis. Funding
was provided by the QB3 Rogers Family Fund Bridging
the Gap Award (R.L. and F.Y.), the Malaysia Biotechnol-
ogy Corporation (W.R.W.), and an NSF Bridges to the
Doctorate Fellowship (B.L.). The X-ray diffractometer
was provided under NSF CHE-0521569.
from anapplications perspective. Inparticular, few current
FDA-approved antibiotics are effective against the bio-
film forms of bacterial pathogens, meaning that clearance
of established biofilm-mediated infections is particularly
Supporting Information Available. Experimental details
for all synthetic procedures and biological screening, 1H and
13C spectra for all compounds, ORTEP and cif files for the
X-ray crystal structure for compound 33. This material is
available free of charge via the Internet at http://pubs.acs.org.
(17) Hall-Stoodley, L.; Costerton, J. W.; Stoodley, P. Nat. Rev.
Microbiol. 2004, 2, 95–108.
(18) Wong, W. R.; Oliver, A. G.; Linington, R. G. Chem. Biol. 2012,
19, 1483–1495.
(19) Differences in microbicidal activities between BioMAP and
standard growth curve measurements can be attributed to the use of
different variants of V. cholerae in these two screens.
The authors declare no competing financial interest.
Org. Lett., Vol. 15, No. 6, 2013
1237