A small molecule for controlled generation of reactive oxygen species (ROS)

Article abstract of DOI:10.1021/ol403300a

Due to the short half-life of reactive oxygen species (ROS) such as a superoxide radical, controlled and localized generation of ROS is challenging. Here, we report a rationally designed small-molecule 1c that generates ROS only when triggered by a bacterial enzyme. We provide evidence for 1c predictably enhancing the intracellular superoxide radical in a model bacterium. Spatiotemporal control over ROS generation offered by 1c should help better understand stress responses in bacteria to increased ROS.

Full text of DOI:10.1021/ol403300a

Letter
pubs.acs.org/OrgLett
A Small Molecule for Controlled Generation of Reactive Oxygen
Species (ROS)
Allimuthu T. Dharmaraja and Harinath Chakrapani*
Indian Institute of Science Education and Research Pune, Dr. Homi Bhabha Road Pashan, Pune 411 008, Maharashtra, India
S
* Supporting Information
ABSTRACT: Due to the short half-life of reactive oxygen species (ROS) such as a superoxide radical, controlled and localized
generation of ROS is challenging. Here, we report a rationally designed small-molecule 1c that generates ROS only when
triggered by a bacterial enzyme. We provide evidence for 1c predictably enhancing the intracellular superoxide radical in a model
bacterium. Spatiotemporal control over ROS generation offered by 1c should help better understand stress responses in bacteria
to increased ROS.
•−
generate O2 that is well suited to simulate increased ROS in
bacteria.
early all organisms inadvertently produce superoxide
O2•−, by 1-electron transfer to oxygen during respira-
N
1
•−
5-(4-Nitrobenzyloxy)-1,4,4a,9a-tetrahydro-1,4-ethanoanthra-
tion. O2 is subsequently converted to hydrogen peroxide
H2O2, which through the Fenton reaction generates the highly
•−
cene-9,10-dione (1c) was considered as a cell-permeable O2
generator (Scheme 1).6 The 4-nitrobenzyl group is a known
substrate for E. coli nitroreductase (NTR), a commonly
expressed oxygen-insensitive bacterial enzyme that reduces a
broad range of aromatic nitro compounds to amines.7
Reduction of the nitro group might lead to a rearrangement
leading to the departure of 1a, which has been previously
1
•
reactive OH. Together, these reactive oxygen species (ROS)
can damage vital cellular components and are hence deployed
by the immune system to counter infectious pathogens. Several
recent studies have shown that ROS can sensitize infectious
bacteria to clinical antibiotics suggesting the possible
therapeutic utility for ROS. Due to a weak pipeline of
antibiotics in preclinical development and the global emergence
of antibiotic resistance, methodologies for selectively enhancing
intracellular ROS including O2•− might help better understand2
and evaluate the therapeutic potential of ROS.3
•−
reported to generate O2 in ambient aerobic buffer through a
keto−enol tautomerism as the first step (Scheme 1).
Enolization of the carbonyls in 1a is promoted by a proximal
H-bonding hydroxyl group, and as a consequence, ROS
generation by this compound was enhanced in comparison
with an analogous benzylated derivative 1b (Scheme 1). Hence,
similarly in the case of 1c, where intramolecular H-bonding is
not possible, enolization is disfavored and this compound is
•−
Due to its short life, O2 must be produced in situ by
reaction with oxygen for use in biochemical studies. Hence,
either small organic molecules that spontaneously generate
O2••−− or enzymatic methods that process a substrate to generate
O2 are used.4 For example, a combination of hypoxanthine
and xanthine oxidase (X + XO) where hypoxanthine is
•−
predicted to be a poor O2 generator in buffer.
Compound 1c was synthesized from 1a by reaction with 4-
nitrobenzylbromide (Supporting Information, Scheme S1).
During the incubation of 1c in ambient aerobic buffer,
•−
metabolized by XO predominantly produces O2 in the
4
•−
proximity of cells. Any O2 that is produced must diffuse
•−
•−
increased O2 was observed only in the presence of NTR as
across a lipid bilayer to exert its effects. However, O2 is not
determined by a luminol-based chemiluminescence assay
highly permeable at neutral pH and such a method may not be
(Figure 1a and 1b).8 An HPLC-based dihydroethidium
useful for enhancing intracellular ROS.5 Small molecules such
•−
•−
(DHE) assay was used to independently assess O2
as paraquat or menadione, which require bioactivation for O2
production (Figure 1c). During incubation of 1c in the
presence of NTR, the nearly complete disappearance of DHE
with concomitant formation of 2-hydroxyethidium (2-OH-E+)
production, have often been used but at elevated concen-
trations that can potentially complicate mechanistic interpre-
tations (see Supporting Information, Chart S1).3 Thus, a cell
permeable small molecule that can predictably and exclusively
increase O2 within cells is not available. Herein, we report a
small molecule that is activated by a bacterial enzyme to
•−
Received: November 15, 2013
Published: December 27, 2013
© 2013 American Chemical Society
398
dx.doi.org/10.1021/ol403300a | Org. Lett. 2014, 16, 398−401

Organic Letters
Letter
a
Scheme 1. Design of a Cell Permeable ROS Generator
a
The 4-nitrobenzyl derivative 1c is a substrate for nitro reduction, which would convert 1c to the active ROS producing 1a. When this
transformation is catalyzed by E. coli nitroreductase (NTR), exclusive intracellular accumulation of ROS is predicted.
Figure 1. Estimation of ROS in ambient aerobic aqueous buffer. + NTR implies E. coli nitroreductase and NADPH; X + XO: hypoxanthine and
xanthine oxidase were used. (a) Chemiluminescence measurement upon reaction with luminol as a measure of O2•− production was carried out with
various compounds (10 μM) and combinations in pH 8.0 buffer. (b) Results of the above assay conducted after 30 min; (c) A HPLC-based
•−
dihydroethidium (DHE) assay was used to infer generation of O2 after incubation of compounds (100 μM) in pH 8.0 buffer for 3 h. 2-
Hydroxyethidium (2-OH-E+), which is exclusively formed by the reaction of O2•− with DHE, elutes at 30.9 min, and ethidium E+, which is formed
by nonspecific oxidation of DHE, elutes at 31.7 min. (d) The amount of hydrogen peroxide generated during incubation of 10 μM of each
compound in pH 7.4 was estimated using an Amplex Red fluorescence assay.
Figure 2. ROS generation during incubation with E. coli: (a) HPLC traces of assay for intracellular O2•− production using a hydroethidine (DHE)
assay. Incubation with 1c (250 μM) was for 30 min; DHE levels indicate unoxidized dye while 2-OH-E+ formed is an indicator for O2•− production
and E+ is indicative of an increase in oxidative species. Ctrl: untreated bacteria. (b) Extracellular H2O2 generated during incubation of E. coli with
compounds (100 μM) for 1 h was estimated using an Amplex Red fluorescence assay. Ctrl: untreated bacteria. (c) H2O2 generation during
incubation of E. coli with 1c (100 μM) was recorded during 6 h as described above. A first-order rate constant for ROS production was calculated as
0.96 h−1. Ctrl: untreated bacteria. (d) E. coli treated with 1c (100 μM) for 1 h; centrifugation and removal of the supernatant followed by
resuspension of bacteria in fresh media. H2O2 generated after incubation of bacteria for 3 h was recorded as described above. Ctrl: untreated bacteria.
(e) A 2,7-dichlorodihydrofluorescein-diacetate (DCFH2-DA)-based fluorescence assay was used to estimate oxidative species generated
intracellularly in E. coli. Increase in relative fluorescence intensity (RFI) with respect to DMSO (0.5%).
•−
was observed, thus, confirming the intermediacy of O2
(Figure 1c).9 H2O2 is produced by 1e− transfer to O2•−, and
we estimated H2O2 using an Amplex Red fluorescence assay.10
In the absence of NTR, negligible H2O2 was produced during
incubation of 1c (10 μM), but in the presence of NTR, a yield
of 3.65 μM H2O2 was recorded after 1 h (Figure 1d). 4-
Nitrobenzyl benzoate 3 (Chart S1), which is a substrate for
NTR but should not produce ROS during its metabolism by
this enzyme, was synthesized using a reported procedure.11
Compound 3 (100 μM) was metabolized by NTR, and as
and 3 (10 equiv) were coincubated in the presence of NTR,
diminished O2 and H2O2 were produced possibly due to
•−
inhibition of turnover of 1c to 1a (Figure 1). Under similar
experimental conditions, the amounts of O2 and H2O2
•−
generated by 1a and 1b were consistent with previously
reported data (Figure 1b and 1d).6
Next, the possibility of the use of 1c as a tool for predictably
•−
enhancing intracellular O2 in bacteria was examined. The
production of O2•− in E. coli was estimated using a DHE assay
that is selective for intracellular O2•− (Figure 2a).9 The bacterial
control showed unreacted DHE with negligible oxidized DHE
(Figure 2a), but in the presence of 1c, the formation of 2-OH-
E+ was observed with concomitant loss of DHE suggestive of
•−
predicted no evidence for O2 or H2O2 was found (Figure 1)
indicating that possible products of metabolism of the 4-
nitrobenzyl group were incapable of generating ROS. When 1c
399
dx.doi.org/10.1021/ol403300a | Org. Lett. 2014, 16, 398−401

Organic Letters
Letter
•−
O2 production intracellularly (Figure 2a). Due to the
Scheme 2. Proposed Mechanisms for ROS Generation
during Incubation of 1c in Buffer in the Presence of NTR
•−
difficulty in accurately quantifying O2 using the DHE assay,
measurement of H2O2, which provides a quantitative basis for
assessing ROS production, was carried out. During incubation
of E. coli with 1c (100 μM), a yield of 10.5 μM H2O2 after 1 h
was recorded (Figure 2b). The formation of H2O2 was nearly
completely abrogated when 1c was cotreated with 3 (10 equiv),
a substrate for intracellular nitroreductases including NTR
(Figure 2b).
Treatment of E. coli with 3 did not increase H2O2 and this
compound did not inhibit ROS generation by 1a (Figure 2b),
suggesting that 3 inhibited the metabolism of 1c to 1a. A time
course for ROS generation during incubation of E. coli with 1c
showed an increase in H2O2 levels during 6 h, which is
consistent with temporally controlled enhancement in intra-
cellular ROS (Figure 2c).
Next, a series of experiments to test if 1c was metabolized
only intracellularly in bacteria to enhance levels of ROS were
carried out. E. coli was incubated with 1c for 1 h, which was
followed by removal of media, and cells were resuspended in
fresh media: under these conditions, any 1c that does not
permeate cells would be removed and any H2O2 produced
would be due to intracellular activation. The difference in yields
of H2O2 obtained during this experiment (Figure 2d) and
during a bolus treatment of 1c (Figure 2b) was minimal, which
supports intracellular metabolism to produce ROS as the major
pathway. Next, E. coli was grown overnight and the cell-free
media was incubated with 1c: any reductases secreted might
activate 1c to produce ROS. However, no significant increase in
H2O2 was found even after 3 h suggesting that extracellular
activation of 1c was not important (see Supporting
Information, Figure S3).
1c with concomitant formation of 1a as the major product was
observed (Figure S5).
Mass spectrometric analysis of this reaction mixture (Figure
S6) revealed the formation of an amine and hydroxylamine
byproducts (Scheme 1), which is consistent with a reduction−
rearrangement cascade.15 When 1c was treated with NTR and
NADPH, mass spectrometric analysis revealed the formation of
2a, again consistent with the proposed mechanism.
Incubation of 1a in aerobic buffer generated O2•− (Figure 1b
and 1c) and produced 2a as the major organic product (Figure
S7). The role of intramolecular H-bonding is apparent in the
large difference in the rate of conversion of 1c to 2c (<20%
conversion after 18 h) and the oxidation of 1a to 2a, which was
nearly complete in the same time period (Figure S7). The
quinone 2a is also a candidate for ROS generation in the
presence of bioreductive enzymes including NTR. However,
our cellular assays showed that 2a did not produce ROS
(Figure 2b and 2d) perhaps due to the diminished rates of
redox cycling of this compound intracellularly.16 This
observation also implies that once 1a converts to 2a, no
further ROS is produced suggesting that 1c might produce up
Next, the intermediacy of hydroxyl radical •OH was assessed
by a supercoiled plasmid DNA cleavage assay. When cotreated
with Fe(II), the ROS generator 1c produced nicks only in the
presence of NTR, supporting the use of this protocol for
enhancement of ROS (Figure S4). For cellular assays, the
ability of 1c to increase intracellular oxidative species including
•OH was studied.12 Incubation of E. coli with dichlorofluor-
escein-diacetate, DCFH2-DA, followed by treatment with 1c
resulted in an increased intracellular fluorescence attributable to
oxidative species (Figure 2e).13 In addition, formation of
ethidium (E+) in the DHE assay is suggestive of an increased
oxidative intracellular environment caused by 1c (Figure 2a).
Taken together, our data suggest that the use of 1c
predictably increases ROS in bacteria. Compound 1c was
next compared with the X + XO protocol. The time course of X
to 2 mol of O2 per mol of compound (Scheme 2).17 An
•−
alternate ROS producing pathway occurred during conversion
of 1c to 2c, which in turn can be cleaved in the presence of
NTR to produce 2a (Scheme 2). If this was a major pathway,
ROS generation from 1c neither would depend on NTR nor
should be inhibited by 3. However, ROS generation by 1c is
significant only in the presence of NTR and was nearly
completely inhibited by 3 in both test tube and cellular assays
(Figures 1 and 2). In addition, the nitrobenzyl derivative 2c was
a poor ROS generator (Figures 1 and 2). Thus, ROS generation
by direct conversion of 1c to 2c appears to be minor. Although
biocatalytic reduction of 1c by NTR requires NADPH as a
cofactor, our observation that treatment of E. coli with 3
resulted in no significant increase of H2O2 (Figure 2b) suggests
that such perturbations of NADPH levels (estimated millimolar
concentrations) do not result in ROS generation.18
•−
+ XO indicates that a burst of O2 is produced while 1c +
NTR showed a gradual increase in ROS during this time period
(Figure 1a). Thus, for biochemical studies, the use of 1c + NTR
might simulate gradual ROS production. In cell-based assays,
the ability of X + XO to increase intracellular ROS was
diminished in comparison with 1c (see Supporting Informa-
tion, Figure S1).14 Thus, 1c complements the existing
repertoire of ROS generators while offering distinct advantages
such as cell permeability and temporal control.
Recently, compounds that potentiate ROS in bacteria were
proposed as adjuvants, which are small molecules that may not
have significant antimicrobial properties themselves but when
cotreated with antibiotics can reduce effective dosages and
minimize side effects.4,19 As few drug candidates with novel
mechanisms are in the pipeline, the development of adjuvants
assume importance.18 The ROS generator 1c did not inhibit E.
coli growth (Table S5), but this compound significantly
sensitized bacteria to aminoglycosides supporting possible
applications of ROS as an adjuvant to this class of antibiotics
(Table S6).20
Two possible pathways for metabolism of 1c to generate
•−
O2
were considered (Scheme 2). First, NTR-induced
reduction of the nitroaryl group should produce an amine or
hydroxylamine, which might initiate a deprotection cascade to
produce 1a (Scheme 1). Under chemoreductive conditions
(Zn/HCOONH4) in pH 7.4 buffer, complete consumption of
In conclusion, we report that 1c is capable of predictably
increasing intracellular levels of ROS in a model bacterium. To
our knowledge, this is the first report of a rationally designed
400
dx.doi.org/10.1021/ol403300a | Org. Lett. 2014, 16, 398−401

Organic Letters
Letter
•−
(11) Sharma, K.; Sengupta, K.; Chakrapani, H. Bioorg. Med. Chem.
Lett. 2013, 23, 5964.
(12) Due to its extremely short half-life and indiscriminate reactivity,
enzyme activated O2 generator that is suitable for enhancing
ROS in bacteria. It is anticipated that 1c would facilitate the
understanding of bacterial responses to ROS, mechanisms of
antibiotic action, and the therapeutic potential of ROS. The use
of 1c to predictably enhance ROS in bacteria other than E. coli
is expected to depend on the expression levels of NTR and on
the uptake of this small molecule. These studies are underway
in our laboratory.
in vivo detection of the hydroxyl radical is challenging and the
•
specificity of previously used fluorescence assays for OH has been
questioned (see ref 1c).
(13) Wardman, P. Free Radical Biol. Med. 2007, 43, 995.
(14) In cellular assays at comparable concentrations with menadione,
despite several attempts, nonspecific decomposition of DHE without
formation of 2-OH-E+ or E+ was observed (Figure S1).
(15) Attempts to isolate these amine byproducts were unsuccessful
and instead produced 1a or 2a as the major product.
(16) The precise rationale for this observation is unclear, but steric
hindrance might hamper the accessibility of the quinone ring to
reductive enzymes.
(17) Intracellular ROS generation yield and rate are dictated by the
efficiency and kinetics of metabolism of 1c by NTR in E. coli.
(18) Kumar, S. R.; Imlay, J. A. J. Bacteriol. 2013, 2, 2.
(19) Ruben Morones-Ramirez, J.; Winkler, J. A.; Spina, C. S.; Collins,
J. J. Sci. Transl. Med. 2013, 5, 190ra81.
(20) A Clinical and Laboratory Standards Institute (CLSI)
recommended protocol was used to determine the minimum
inhibitory concentration (MIC) against E. coli (Table S5). No
significant inhibition of bacterial growth by 1c was observed at
elevated concentrations (MIC >100 μM) suggesting that perhaps ROS
alone is well tolerated by E. coli. However, when gentamycin was
cotreated with 1c (50 μM), an 8-fold decrease in MIC for this
antibiotic was observed suggesting that a lower antibiotic dosage was
required to obtain similar efficacy.
ASSOCIATED CONTENT
* Supporting Information
Synthesis, characterization data, assay protocols, and associated
data. This material is available free of charge via the Internet at
http://pubs.acs.org.
■
S
AUTHOR INFORMATION
Corresponding Author
■
*E-mail: harinath@iiserpune.ac.in.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
The authors thank IISER Pune and the Department of
Biotechnology, Ministry of Science and Technology, India
(BT/PR6798/MED/29/636/2012; BT/05/IYBA/2011) for
financial support. A.T.D. acknowledges a research fellowship
from the Council for Scientific and Industrial Research (CSIR).
REFERENCES
■
(1) (a) Fang, F. C. Nat. Rev. Microbiol. 2004, 2, 820. (b) Imlay, J. A.
Nat. Rev. Microbiol. 2013, 11, 443. (c) Fang, F. C. Nat. Biotechnol.
2013, 31, 415.
(2) (a) Keren, I.; Wu, Y.; Inocencio, J.; Mulcahy, L. R.; Lewis, K.
Science 2013, 339, 1213. (b) Liu, Y.; Imlay, J. A. Science 2013, 339,
1210. (c) Ezraty, B.; Vergnes, A.; Banzhaf, M.; Duverger, Y.;
Huguenot, A.; Brochado, A. R.; Su, S. Y.; Espinosa, L.; Loiseau, L.;
Py, B.; Typas, A.; Barras, F. Science 2013, 340, 1583.
(3) (a) Kohanski, M. A.; Dwyer, D. J.; Hayete, B.; Lawrence, C. A.;
Collins, J. J. Cell 2007, 130, 797. (b) Foti, J. J.; Devadoss, B.; Winkler,
J. A.; Collins, J. J.; Walker, G. C. Science 2012, 336, 315. (c) Brynildsen,
M. P.; Winkler, J. A.; Spina, C. S.; MacDonald, I. C.; Collins, J. J. Nat.
Biotechnol. 2013, 31, 160.
(4) (a) McCord, J. M.; Fridovich, I. J. Biol. Chem. 1968, 243, 5753.
(b) Lynch, R. E.; Fridovich, I. J. Biol. Chem. 1978, 253, 4697.
(c) Ingold, K. U.; Paul, T.; Young, M. J.; Doiron, L. J. Am. Chem. Soc.
1997, 119, 12364.
(5) (a) Korshunov, S. S.; Imlay, J. A. Mol. Microbiol. 2002, 43, 95.
(b) Zenno, S.; Koike, H.; Tanokura, M.; Saigo, K. J. Biochem. 1996,
120, 736.
(6) Dharmaraja, A. T.; Alvala, M.; Sriram, D.; Yogeeswari, P.;
Chakrapani, H. Chem. Commun. 2012, 48, 10325.
(7) (a) Zenno, S.; Koike, H.; Tanokura, M.; Saigo, K. J. Biochem.
1996, 120, 736. (b) Race, P. R.; Lovering, A. L.; Green, R. M.; Ossor,
A.; White, S. A.; Searle, P. F.; Wrighton, C. J.; Hyde, E. I. J. Biol. Chem.
2005, 280, 13256.
(8) Trung Pham, H.; Marquetty, C.; Pasquier, C.; Hakim, J. Anal.
Biochem. 1984, 142, 467.
(9) (a) Zhao, H.; Joseph, J.; Fales, H. M.; Sokoloski, E. A.; Levine, R.
L.; Vasquez-Vivar, J.; Kalyanaraman, B. Proc. Natl. Acad. Sci. U.S.A.
2005, 102, 5727. (b) Zielonka, J.; Vasquez-Vivar, J.; Kalyanaraman, B.
Nat. Protoc. 2008, 3, 8. (c) Georgiou, C. D.; Papapostolou, I.;
Grintzalis, K. Nat. Protoc. 2008, 3, 1679.
(10) (a) Zhou, M.; Diwu, Z.; Panchuk-Voloshina, N.; Haugland, R. P.
Anal. Biochem. 1997, 253, 162. (b) Khodade, V. S.; Dharmaraja, A. T.;
Chakrapani, H. Bioorg. Med. Chem. Lett. 2012, 22, 3766−3769.
401
dx.doi.org/10.1021/ol403300a | Org. Lett. 2014, 16, 398−401