Development of benzo[1,4]oxazines as biofilm inhibitors and dispersal agents against Vibrio cholerae

Article abstract of DOI:10.1039/c4cc07003h

Bacterial biofilms are estimated to be associated with over 65 percent of all nosocomial infections. However, no therapeutics have been approved by the FDA which directly mediate biofilm formation or persistence. Herein we report oxazine 25 as a highly po

Full text of DOI:10.1039/c4cc07003h

ChemComm
View Article Online
View Journal | View Issue
COMMUNICATION
Development of benzo[1,4]oxazines as biofilm
inhibitors and dispersal agents against
Vibrio cholerae†
Cite this: Chem. Commun., 2015,
51, 1305
Received 10th September 2014,
Accepted 27th November 2014
Christopher J. A. Warner,a Andrew T. Cheng,b Fitnat H. Yildizb and
Roger G. Linington*a
DOI: 10.1039/c4cc07003h
www.rsc.org/chemcomm
Bacterial biofilms are estimated to be associated with over 65 percent against the Gram-negative pathogens V. cholerae and Pseudomonas
of all nosocomial infections. However, no therapeutics have been aeruginosa.8–10 Screening of our natural product library, compromis-
approved by the FDA which directly mediate biofilm formation or ing of over 6000 prefractions, identified the aureomycin chromo-
persistence. Herein we report oxazine 25 as a highly potent inhibitor, phore 1 as a moderate inhibitor of V. cholerae biofilms (biofilm
disperser and in the presence of the appropriate antibiotic eradicator inhibitory concentration (BIC50) = 63 mM). Given the structural
of V. cholerae biofilms.
novelty of this scaffold compared with other biofilm inhibitors,
and the unusual biofilm inhibitory phenotype observed in the
Bacterial biofilms are surface associated bacterial communities primary screening images, we elected to develop the benzo[1,4]-
of sessile cells encased in a matrix of polysaccharides, extra- oxazine scaffold through medicinal chemistry optimization in
cellular DNA and proteins.1 Such biofilms are of significant order to identify key elements of the required pharmacophore,
concern in nosocomial infections, where it is attributed to over and generate analogues with improved potency and pharmaco-
$1 billion in increased hospital costs per annum in the US alone.2 logical properties.11 Key to this approach was the formation of the
Unlike cells in the planktonic state, bacterial biofilms do not exert a-keto-amide 7 and its subsequent application in a debenzylation–
their antimicrobial resistance through mutation or acquisition of cyclization strategy to form hemi-acetal 8. Gratifyingly, treatment of
resistance functions by horizontal gene transfer.3 Instead, resistance the a-ketoamide 7 (formed in 5 steps from the commercially
is largely driven by the formation of latent cells within the biofilm available ester 2) with 2% Pd(OH)2 on charcoal and four equivalents
matrix that reduce cellular turnover and therefore remove the of 1,4-cyclohexadiene in ethanol at 50 1C enabled formation of
susceptibility of targets associated with traditional antimicrobials.4
the cyclic hemi-acetal 8 in excellent yield on a multi-gram scale
V. cholerae is a diarrheal pathogen that naturally inhabits with reaction times of less than 5 minutes. Dehydration of the
both fresh and saltwater environments.5 In spite of its prevalence, acetal afforded the target molecule in 7 steps on a multigram
no clinical therapeutics have been approved for use in the US or scale (Scheme 1).
elsewhere that directly target biofilm formation and persistence.
To identify key elements of the pharmacophore for this
A limited number of small molecule inhibitors of V. cholerae compound class, a library of 41 derivatives were prepared using
biofilms have been reported in the literature, both from natural a divergent strategy to diversify both the exocyclic alkene and
product screening campaigns and medicinal chemistry develop- the methyl ester moiety of the original scaffold (Fig. 1, see ESI†
ment efforts.6,7 However in the majority of cases these com- for a comprehensive list of synthesized compounds). Initial
pounds have been shown to impact quorum sensing (QS) rather examination of the lead compound 1 highlighted the a,b
than directly targeting processes involved with biofilm matrix unsaturated carbonyl (C9–C11) as a Michael acceptor with potential
production or regulation.
involvement in the mechanism of action. In line with the behavior
We recently reported the development of two high throughout of other Michael acceptors in the literature, modification of the
image-based screens capable of identifying biofilm inhibitors exocyclic methylene unit (compounds 8–12) eliminated activity in
all cases.12 Introduction of any substituent onto the double bond
a Department of Chemistry and Biochemistry, University of California Santa Cruz,
(compounds 13–17) also resulted in a pronounced decrease in
biofilm inhibition, indicating a tight steric limitation for modifica-
tions at this position.
To probe whether the increase in steric size of the Michael
acceptor directly correlated with the ability of the compound to
undergo Michael addition, both the original oxazine 1 and phenyl
California, 95064, USA. E-mail: rliningt@ucsc.edu
b Department of Microbiology and Environmental Toxicology, University of
California Santa Cruz, California, 95064, USA
† Electronic supplementary information (ESI) available: Experimental procedures
and analytical data along with protocols for all biological experiments. See DOI:
10.1039/c4cc07003h
This journal is ©The Royal Society of Chemistry 2015
Chem. Commun., 2015, 51, 1305--1308 | 1305

View Article Online
Communication
ChemComm
Scheme 1 The total synthesis of the benzo[1,4]oxazine biofilm inhibitor 1.
Reagents and conditions: (i) HNO3, Ac2O, 0 1C to rt, 1 hour, 94% yield;
(ii) 4 eq. AlCl3, DCM, 0 1C to reflux, 3 hours, 89% yield; (iii) 4 eq. BnBr, 4 eq.
K2CO3, 1 : 1 DCM/MeOH, reflux, 3 hours, 97% yield; (iv) 4 eq. SnCl2, 3 : 1
EtOH/6N HCl, rt to reflux, 45 minutes, 81% yield; (v) 1.1 eq. pyruvoyl
chloride, 1.5 eq. pyridine, DCM, 0 1C to rt, 90 minutes, 65% yield; (vi) 2%
Pd(OH)2/C, 4 eq. 1,4-cyclohexadiene, EtOH, 50 1C, 2 minutes, 81% yield;
(vii) 1.2 eq. MsCl, 1.5 eq. NEt3, DCM, 0 1C to rt, 90 minutes, 82% yield.
substituted compound 16 were added to either N-acetyl cysteine
methyl ester or a simple polypeptide (Ac-Ala-Cys-Ala-Gly-OH) and
the formation of Michael addition adducts monitored by LC-MS.
Complete formation of the expected Michael adduct was observed
with compound 1 for both substrates. By contrast very little (B5%)
of the expected Michael adducts were observed in either case with
compound 16, suggesting that the tight steric limitation around
the alkene may be due to the inability of sterically encumbered
compounds to act as Michael acceptors.
A similar loss in activity was observed upon conversion of
the endocyclic amide to the corresponding N-methyl derivative
(compound 18). This suggests that a hydrogen bond donor may
be required to correctly orientate the compound in the active site, or
alternatively that this portion of the molecule is present in a
congested region of the active site that is not tolerant of increased
steric bulk. In contrast to modifications on the oxazine ring,
alteration of the ester substituent at C2 resulted in a pronounced
increase in biofilm inhibition. Although conversion of the methyl
Fig. 1 Screening of a selection of the oxazine inhibitors against V. cholerae
biofilms. A selection of the oxazine derivatives screened as inhibitors of
V. cholerae biofilms. BIC50 and BDC50 determined with 3 biological repli-
cates each consisting of two technical replicates, see ESI† for full BIC50 dose
response curves and complete list of all compounds screened in the assay.
ester to the corresponding free acid 19 eliminated activity, introduc- against a panel of 15 clinically relevant bacterial pathogens up
tion of the para-methoxy aromatic residue 25 resulted in a 10-fold to 200 mM.12 Screening in our in-house image-based cytological
increase in compound efficacy, with a BIC50 value of 6 mM. Colony profiling assay also revealed that compound 25 had no measurable
forming unit analysis demonstrated that compound 25 possessed cytotoxicity to HeLa cells, up to 200 mM (see ESI†).13 Interestingly,
no bactericidal activity up to 200 mM (see ESI†). This was confirmed the parent phenyl compound 21 showed only a marginal increase
by BioMAP screening, with compound 25 displaying no activity in activity (45 mM vs. 62 mM), while the corresponding substituted
1306 | Chem. Commun., 2015, 51, 1305--1308
This journal is ©The Royal Society of Chemistry 2015

View Article Online
ChemComm
Communication
phenyl ester 22 exhibited a 4-fold increase in activity over
the parent methyl ester 1, demonstrating the importance of
substituent effects on the aromatic ring.
The observation that the carboxylic acid 19 was inactive as a
biofilm inhibitor led us to probe whether hydrolysis of the phenolic
ester could be masking the true potency of this compound class.
Incubation of oxazine 25 in either LB media or PBS buffer at 37 1C
for 24 hours failed to result in any measurable hydrolysis and
suggested that this was not a limiting factor for compound activity.
Interestingly, formation of either the phenyl amide 23 or the
analogous para-methoxy amide, two compounds that would be
expected to be far more resistant to hydrolysis resulted in a
complete loss in compound activity.
One potential application for biofilm inhibitors is as dispersal
agents to eliminate biofilms in established infections, which
otherwise persist during antibiotic treatment and can lead to
recrudescence of infection. We have recently disclosed a novel
biofilm dispersal model (BDM) in P. aeruginosa that can be
used to examine the capabilities of small molecules to induce
dispersion of pre-existing biofilms.10 To investigate whether
compound 25 was capable of the dispersal of pre-formed
V. cholerae biofilms, a similar procedure was employed in the
V. cholerae system. In brief, cultures of V. cholerae were allowed
to pre-form biofilms in screening plates for 2 hours prior to
compound addition. Following standard incubation conditions
(4 hours at 30 1C) macrocolonies were imaged and quantified as
previously described to determine the percentage of remaining
biofilm coverage. Compound 25 displayed strong biofilm dispersal
activity and no bactericidal activity, with a biofilm dispersal
concentration (BDC50) value of 13 mM and optical density readings
indicating good bacterial growth. To our knowledge this represents
the first example of a small molecule capable of both inhibiting
and inducing dispersal of V. cholerae biofilms, and places it among
just a handful of compounds capable of inducing the dispersal of
mature surface-associated biofilms.14
Fig. 2 Static well images of V. cholerae biofilms. Static well GFP images of
V. cholerae biofilms and normalized OD600 readings. In all instances the
antibiotic, dispersal agent or a combination of the two were introduced
after two hours of incubation and incubated for a further 4 hours before
being washed and analyzed. (A) DMSO control; (B) 50 mM ciprofloxacin; (C)
50 mM erythromycin; (D) 20 mM compound 25; (E) 20 mM compound 25
and 50 mM ciprofloxacin; (F) 20 mM compound 25 and 50 mM erythromycin.
White bars represent 50 mm.
A major challenge surrounding the treatment of biofilm-
mediated infections is that bacterial cells within the biofilm
have the potential to enter a latent state that renders them the cellular viability, as determined by OD600 analysis, indicating that
much less susceptible to traditional antibiotics.4 One potential these drug combinations possess the ability to both clear and kill
application for biofilm dispersal agents is as combination established V. cholerae infections (Fig. 2). One limitation of image-
therapies with existing antibiotics to both clear and eliminate based high content screening is that samples must be incubated
otherwise persistent infections. To examine whether our biofilm under static culture conditions. The disadvantage of this method is
dispersal model could recapitulate this antibiotic resistance for that static culture allows the accumulation of signaling factors and
V. cholerae we screened five FDA-approved antibiotics (ciprofloxacin, quorum sensing molecules including V. cholerae autoinducer-1
erythromycin, azithromycin, doxycycline and furazilidinone) in the (CAI-1), autoinducer-2 (AI-2) and indole which all impact the
dispersal assay. Interestingly, without the presence of compound 25, rate and degree of biofilm formation.15,16
erythromycin, ciprofloxacin and furazilidinone (all therapeutics pre-
To examine whether oxazine 25 was capable of disrupting
scribed as first stage treatments for V. cholerae infection) failed to biofilm formation and persistence under more physiologically
induce biofilm dispersal, with confocal microscopy indicating the relevant conditions, we examined its anti-biofilm properties
presence of large biofilm macrocolonies and very few background under flow cell conditions. At all concentrations tested (see ESI†),
planktonic cells. Optical density readings confirmed this observation strong biofilm inhibition was observed, with a marked decrease in
and suggested that these antibiotics have the capacity to eradicate the size, thickness and morphology of the biofilm. COMSTAT
cells in the planktonic state, but not significantly impact biofilm analysis17 indicated a 7-fold reduction in biomass upon treatment
coverage. By contrast, addition of 20 mM of compound 25 in addition of the V. cholerae biofilms with compound 25 under flow condi-
to 50 mM of either erythromycin or ciprofloxacin resulted in near tions (Fig. 3B). These data indicate that compound 25 is capable
quantitative elimination of biofilm coverage, and a lowering in of disrupting biofilms under both static and flow cell conditions,
This journal is ©The Royal Society of Chemistry 2015
Chem. Commun., 2015, 51, 1305--1308 | 1307

View Article Online
Communication
ChemComm
the tolerance of biofilms toward traditional antibiotics, and
reduce the potential for the development of bacterial resistance
often observed with antibiotic mono-therapies.
Notes and references
1 C. A. Fux, J. W. Costerton, P. S. Stewart and P. Stoodley, Trends
Microbiol., 2005, 13, 34–40.
2 M. Wilson, Sci. Prog., 2001, 84, 235–254.
3 T.-F. Mah, Future Microbiol., 2012, 7, 1061–1072.
4 K. Lewis, Curr. Top. Microbiol. Immunol., 2008, 322, 107–131.
5 V. Berk, J. C. N. Fong, G. T. Dempsey, O. N. Develioglu, X. Zhuang,
J. Liphardt, F. H. Yildiz and S. Chu, Science, 2012, 337, 236–239.
6 N. Augustine, P. A. Wilson, S. Kerkar and S. Thomas, Curr. Microbiol.,
2012, 64, 338–342.
7 (a) V. L. Schramm, J. A. Gutierrez, G. Cordovano, I. Basu, C. Guha,
T. J. Belbin, G. B. Evans, P. C. Tyler and R. H. Furneaux, Nucleic Acids
Symp. Ser., 2008, 75–76; (b) R. C. Kelly, M. E. Bolitho, D. A. Higgins,
W. Lu, W.-L. Ng, P. D. Jeffrey, J. D. Rabinowitz, M. F. Semmelhack,
F. M. Hughson and B. L. Bassler, Nat. Chem. Biol., 2009, 5, 891–895;
(c) K. Sambanthamoorthy, A. A. Gokhale, W. Lao, V. Parashar, M. B.
Neiditch, M. F. Semmelhack, I. Lee and C. M. Waters, Antimicrob.
Agents Chemother., 2011, 55, 4369–4378; (d) W.-L. Ng, L. Perez,
J. Cong, M. F. Semmelhack and B. L. Bassler, PLoS Pathog., 2012,
8, e1002767; (e) S. T. Rutherford and B. L. Bassler, Cold Spring
Harbor Perspect. Med., 2012, 2, a012427.
Fig. 3 Static and confocal flow cell images of compound 25 against
V. cholerae biofilms. (A) static culture conditions, from left to right, DMSO
control, compound 25 at 15 mM. In both instances OD600 data suggested no
bactericidal activity; (B) confocal flow cell images of V. cholerae, see ESI† for
further details. From left to right, DMSO control and lead compound 25 at
250 mM. COMSTAT analysis of mean biomass (mm3/mm2) indicated a 7-fold
reduction in the presence of lead compared 25 compared to DMSO control.
In both cases white bars indicate 50 mm.
and suggest that this compound has potential value as a biofilm
inhibitor for the clearance of established biofilm-mediated
infections in vivo.
In conclusion, a medicinal chemistry approach is reported
for the development of the V. cholerae biofilm inhibitor 25 that
involves a 7 step, multi-gram, purification-free route to primary
lead 1. An SAR study on the oxazine scaffold identified the
importance of both the exocyclic alkene and amidic proton for
compound activity. Modifications at either position led to complete
8 K. C. Peach, A. T. Cheng, A. G. Oliver, F. H. Yildiz and
R. G. Linington, ChemBioChem, 2013, 14, 2209–2215.
9 (a) K. C. Peach, W. M. Bray, N. J. Shikuma, N. C. Gassner, R. S. Lokey,
F. H. Yildiz and R. G. Linington, Mol. BioSyst., 2011, 7, 1176–1184;
´
(b) B. Leon, J. C. N. Fong, K. C. Peach, W. R. Wong, F. H. Yildiz and
R. G. Linington, Org. Lett., 2013, 15, 1234–1237.
10 G. Navarro, A. T. Cheng, K. C. Peach, W. M. Bray, V. S. Bernan,
F. H. Yildiz and R. G. Linington, Antimicrob. Agents Chemother.,
2014, 58, 1092–1099.
11 M. Shibuya, H. Sakurai, T. Maeda, E. Nishiwaki and M. Saito,
Tetrahedron Lett., 1986, 27, 1351–1354.
loss of activity for all tested compounds. However, expansion 12 W. R. Wong, A. G. Oliver and R. G. Linington, Chem. Biol., 2012, 19,
1483–1495.
around the methyl ester identified phenyl ester derivative 25 as
an advanced lead compound with one of the highest potencies
13 C. J. Schulze, W. M. Bray, M. H. Woerhmann, J. Stuart, R. S. Lokey
and R. G. Linington, Chem. Biol., 2013, 20, 285–295.
reported to date for biofilm inhibition against V. cholerae (BIC50
=
14 J. Singh, R. C. Petter, T. A. Baillie and A. Whitty, Nat. Rev.
Drug Discovery, 2011, 10, 307–317.
15 (a) R. Frei, A. S. Breitbach and H. E. Blackwell, Angew. Chem., Int. Ed.,
2012, 51, 5226–5229; (b) R. E. Furlani, A. A. Yeagley and C. Melander,
Eur. J. Med. Chem., 2013, 62, 59–70.
16 M. J. Minvielle, K. Eguren and C. Melander, Chemistry, 2013, 19,
17595–17602.
17 A. Heydorn, A. T. Nielsen, M. Hentzer, C. Sternberg, M. Givskov,
B. K. Ersbøll and S. Molin, Microbiology, 2000, 146, 2395–2407.
6 mM) and the capacity to induce the dispersal of preformed
biofilms (BDC50 = 13 mM). Co-dosing of compound 25 with
50 mM erythromycin or azithromycin demonstrated the potential
of the synergistic action of a dispersal reagent and antibiotic in
inducing detachment and subsequent clearance of preformed
biofilms. Such a strategy may offer a means of overcoming both
1308 | Chem. Commun., 2015, 51, 1305--1308
This journal is ©The Royal Society of Chemistry 2015