Antimicrobial photodynamic efficacy of side-chain functionalized benzo[a]phenothiazinium dyes

Article abstract of DOI:10.1111/j.1751-1097.2008.00403.x

5-(Ethylamino)-9-diethylaminobenzo[a]phenothiazinium chloride (EtNBS) is a photosensitizer (PS) with broad antimicrobial photodynamic activity. The objective of this study was to determine the antimicrobial photodynamic effect of side chain/end group modifications of EtNBS on two representative bacterial Gram-type-specific strains. Two EtNBS derivatives were synthesized, each functionalized with a different side-chain end-group, alcohol or carboxylic acid. In solution, both exhibited photochemical properties consistent with those of the EtNBS parent molecule. In vitro photodynamic therapy experiments revealed an initial Gram-type-specificity with two representative strains; both derivatives were phototoxic to Staphylococcus aureus 29213 but the carboxylic acid derivative was nontoxic to Escherichia coli 25922. This difference in photodynamic efficacy was not due to a difference in the binding of the two molecules to the bacteria as the amount of both derivatives bound by bacteria was identical. Interestingly, the carboxylic acid derivative produced no fluorescence emission when observed in cultures of E. coli via fluorescence microscopy. These early findings suggest that the addition of small functional groups could achieve Gram-type-specific phototoxicity through altering the photodynamic activity of PSs and deserve further exploration in a larger number of representative strains of each Gram type.

Full text of DOI:10.1111/j.1751-1097.2008.00403.x

Photochemistry and Photobiology, 2009, 85: 111–118
Antimicrobial Photodynamic Efficacy of Side-chain Functionalized
Benzo[a]phenothiazinium Dyes
Sarika Verma1, Ulysses W. Sallum1, Humra Athar1, Lauren Rosenblum1, James W. Foley2 and
Tayyaba Hasan*1
1Wellman Center for Photomedicine, Massachusetts General Hospital; and Department of Dermatology,
Harvard Medical School, Boston, MA
2The Rowland Institute at Harvard, Cambridge, MA
Received 15 April 2008, accepted 6 May 2008, DOI: 10.1111/j.1751-1097.2008.00403.x
photodynamic antimicrobial activity in the PDT of numerous
microbial infections (6–9). The size and charge of the side
ABSTRACT
5-(Ethylamino)-9-diethylaminobenzo[a]phenothiazinium chlo-
ride (EtNBS) is a photosensitizer (PS) with broad antimicrobial
photodynamic activity. The objective of this study was to
determine the antimicrobial photodynamic effect of side
chain ⁄ end group modifications of EtNBS on two representative
bacterial Gram-type-specific strains. Two EtNBS derivatives
were synthesized, each functionalized with a different side-chain
end-group, alcohol or carboxylic acid. In solution, both exhibited
photochemical properties consistent with those of the EtNBS
parent molecule. In vitro photodynamic therapy experiments
revealed an initial Gram-type-specificity with two representative
strains; both derivatives were phototoxic to Staphylococcus
aureus 29213 but the carboxylic acid derivative was nontoxic to
Escherichia coli 25922. This difference in photodynamic efficacy
was not due to a difference in the binding of the two molecules to
the bacteria as the amount of both derivatives bound by bacteria
was identical. Interestingly, the carboxylic acid derivative
produced no fluorescence emission when observed in cultures of
E. coli via fluorescence microscopy. These early findings suggest
that the addition of small functional groups could achieve Gram-
type-specific phototoxicity through altering the photodynamic
activity of PSs and deserve further exploration in a larger
number of representative strains of each Gram type.
chains on these tetra- and tricyclic heteroaromatic PSs are
major determinants of their phototoxicity and specificity (10).
A
member of this group, 5-(ethyamino)-9-diethylamino-
benzo[a]phenothiazinium chloride (EtNBS), is particu-
larly effective in destroying a broad range of micro-organisms
(7).
With respect to commensal bacteria, a broad-spectrum
antibiotic is a double-edged sword, eradicating pathogenic
bacteria as well as beneficial microflora. Depletion of the
microflora makes their once occupied environmental niche
available for colonization by opportunistic pathogens. For
example, the distal portion of the human intestine contains the
most abundant microfloral population and the majority of the
bacterial inhabitants are Gram-negative anaerobes (11). The
colonization of the colon by opportunistic pathogens can lead
to severe and sometimes fatal colitis (12). The intracolonic
administration of antibiotics has been demonstrated to be
effective in the treatment of such infections (13). The use of
antimicrobials that effectively treat Gram-positive infections
and do not damage Gram-negative commensal organisms
could help in maintaining healthy microflora and enhance the
efficacy of the intracolonic treatment of bacterial colitis while
reducing antibiotic-associated diarrhea.
In a recent review by Wainwright et al. (9), the authors
described the structural alterations of Methylene Blue and Nile
Blue A to yield improved PSs. In the present study, a similar
approach was taken in an effort to change the bacterial
specificity of EtNBS. The ethyl side chain of EtNBS was
altered through the addition of two free functional groups
(alcohol or carboxylic acid) that have different charges at
physiological pH. Cationic PSs are expected to be more
effective antimicrobials in comparison with noncharged or
anionic PSs due to a mutual electrostatic affinity with the
largely anionic, bacterial phospholipid membrane (6). The
electrostatic affinity is thought to result in an increased
accumulation of cationic PSs at their microbial targets (6).
The structural differences between a Gram-positive and a
Gram-negative bacteria cause differences in the ability of
various antimicrobials to reach the bacterial cell membrane as
a target site for antibacterial PDT (3,6,14). We hypothesized
that the difference in charge would create differences in
the specificity of the derivatives for Gram-positive and
INTRODUCTION
The emergence of strains of human bacterial pathogens with
resistance to conventional therapies requires the development
of novel and effective therapeutic technologies. Photodynamic
therapy (PDT) has the potential to be one such technology (1–
3). PDT utilizes light to activate a photodynamic chemical
(photosensitizer [PS]). When illuminated, the PS absorbs
photons and the increase in energy results in the excitation
of electrons that, upon return to the ground state, transfer the
energy to an acceptor molecule (4,5). This results in the
generation of singlet oxygen and other reactive molecular
species, thus achieving an antimicrobial phototoxic effect.
The cationic, water soluble benzo[a]phenothiazinium and
phenothiazinium chloride PSs are recognized for their potent
*Corresponding author email: thasan@partners.org (Tayyaba Hasan)
ꢀ 2008 The Authors. JournalCompilation. The AmericanSocietyofPhotobiology 0031-8655/09
111

112 Sarika Verma et al.
Synthesis of 5-(naphthalen-1-ylamino)pentanoic acid, 2c. A stirred
mixture of 1-naphthylamine (2 g, 13.90 mmol) and ethyl 5-bromo
valerate (3.00 g, 14.35 mmol) was refluxed for nearly 15 h in ethanol
(7 mL), and continuously monitored by TLC. Upon the completion of
the reaction, the solvent was evaporated under reduced pressure and
the crude product was purified by column chromatography (silica gel,
hexane: dichloromethane 1:1 as an eluent). Ethyl 5-(naphthalen-1-
ylamino)pentanoate (2b) was obtained as a pale yellow oil in 67%
yield. ES-MS (m ⁄ z) for C17H22NO2 [M+]: calculated for 272.17;
found: 272.20.
Ethyl 5-(naphthalen-1-ylamino)pentanoate, 2b (220 mg, 0.81
mmol) was dissolved in 1,4-dioxane (4 mL). Sodium hydroxide (1 M,
1.5 mL) was added to this solution and the resultant mixture was
stirred at room temperature for 5 h. The reaction mixture was
acidified to pH 3 with 7% sulphuric acid and the product was
extracted in dichloromethane (20 mL). The organic layer was washed
twice with water and dried over magnesium sulfate. The solvent was
removed under reduced pressure to obtain ethyl 5-(naphthalen-
1-ylamino)pentanoic acid (2c, 110 mg) as a pale green solid in 55%
yield. ES-MS (m ⁄ z) for C15H18NO2 [M+]: calculated for 244.13;
found: 244.10.
General synthesis of side-chain functionalized benzo[a]
phenothiazinium chloride, 3a or 3b. Bunte salt (276 mg, 1.0 mmol),
naphthyl derivative (2a or 2c, 1.8 mmol) and 25 mL methanol were
placed in a round bottom flask equipped with a magnetic stirrer and
reflux condenser. This reaction mixture was heated to reflux temper-
ature. Silver carbonate (606 mg, 2.20 mmol) was slowly added to the
refluxing reaction mixture and an intense color change was observed
within 0–5 min of complete addition. The deep blue solution obtained
was checked for the characteristic absorption of benzo[a]phenothia-
zinium derivatives at 655 nm on a UV-visible spectrophotometer.
After heating for 30 min, the reaction flask was cooled to room
temperature. The reaction mixture was filtered and evaporated to
obtain a dark blue crude product. The crude product was redissolved
in 25 mL dichloromethane, washed with saturated sodium carbonate
solution and dried over sodium sulfate. The organic phase was filtered
and acidified with 0.4 mL of concentrated hydrochloric acid. The
solution was mixed gently and left overnight in the fume hood to dry.
The product was obtained as a dark blue solid and was purified by
column chromatography using 2–10% methanol in dichloromethane
as an eluent.
Gram-negative bacteria and that the zwitterionic, carboxylic
acid derivative would exhibit reduced photodynamic antimi-
crobial activity against Gram-negative bacteria.
Two EtNBS derivatives, 5-(3¢-hydroxypropylamino)-9-diet-
hylaminobenzo [a]phenothiazinium chloride (EtNBS-OH, 3a)
and 5-(4¢-carboxybutylamino)-9-diethylaminobenzo[a]pheno-
thiazinium chloride (EtNBS-COOH, 3b) were synthesized.
Their photophysical properties and in vitro antibacterial
photodynamic activity were investigated with respect to
Staphylococcus aureus 29213 as a representative of Gram-
positive strain and Escherichia coli 25922 as a representative of
Gram-negative strain. Here, we report that the addition of a
carboxylic acid group to EtNBS results in the abolition of its
photodynamic antimicrobial activity for E. coli, but not for
S. aureus; whereas no such observation was made by addition
of alcohol group to EtNBS. The main purpose of this article
is to present initial findings to demonstrate that the binding of
PSs to their bacterial targets is not the major determinant of
their photodynamic antibacterial activity and that subtle
structural differences could produce Gram-type-specific anti-
bacterial PSs.
MATERIALS AND METHODS
Chemical synthesis. Commercially available chemicals and reagent
grade solvents were purchased from Aldrich and used as received.
Nuclear magnetic resonance (1H-NMR) spectroscopy was measured
on a Varian 400 MHz instrument, using 15–20 mg of material in
chloroform-d and methanol-d4 as solvents. UV-visible absorption
spectra were measured using a Hewlett Packard 8453 spectrophotom-
eter equipped with a diode array detector system. Fluorescence spectra
were obtained in aqueous ⁄ organic solutions of products using a Jobin
Yvon Horiba FluoroMax-3 fluorometer. High-pressure liquid chro-
matography (HPLC) analysis was performed using a Shimadzu vp
series SCL-10A controller, SPD-M10A diode array detector, LC-
10AD pumps, DGU-14A degasser and C18 reverse phase column
controlled by Class VP software. HPLC method: the eluents were
A = 0.1% TFA in water and B = 0.1% TFA in acetonitrile, and the
gradient method used for all injections was from a relative concentra-
tion of 98.0% of A and 2.0% of B to 100% of B over a time of 20 min
with a flow rate of 4.0 mL min)1. ESI (electrospray ionization positive
mode) mass spectra were recorded with a Bruker Daltonics Esquire
3000 plus spectrometer, equipped with Esquire control version 5.2 and
Data analysis version 3.2 software for processing data.
Synthesis of sodium 2-amino-5-diethylaminophenylthiosulfuric acid
(Bunte salt) 1. The synthesis of Bunte salt is previously reported
elsewhere (15). Briefly, N,N,-diethyl-p-phenylenediamine (1 g,
6.09 mmol) was added to a stirring solution of aluminum sulfate
(4.11 g, 6.52 mmol) in water (10 mL). Sodium thiosulfate (2.21 g,
14 mmol) and zinc chloride (0.872 g, 6.39 mmol) were sequentially
added to this reaction mixture. The reaction mixture was cooled in an
ice bath, and an aqueous solution (4 mL) of potassium dichromate
(0.49 g, 1.68 mmol) was slowly added over a period of 20 min. The
solution was stirred for 2 h in an ice bath and thick precipitates that
were observed were filtered and washed with acetone. The crude
product was refluxed in methanol (12 mL) and filtered to obtain a dark
gray solid (0.79 g) in 47% yield. The product mass peak of 276 amu
was confirmed by mass spectroscopy. This product was used without
any further purification or characterization.
Characterization 3a, EtNBS-OH. 1H NMR (Fig. S2A, CDCl3,
CD3OD, 400 MHz): d = 1.34 (t, J = 7.2 Hz, 6H, 2 · NHCH2CH3),
2.05 (br s, 2H, NHCH2CH2CH2), 3.60–3.70 (q, J = 7.2 Hz, 4H,
2 · NHCH2CH3), 3.82 (br t, 2H, NHCH2CH2CH2), 3.91 (br t, 2H,
NHCH2CH2CH2), 6.88 (s, 1H, Ar-6CH), 7.04 (s, 1H, Ar-8CH), 7.12
(d, J = 9.2 Hz, 1H, Ar-10CH), 7.76–7.80 (m, 2H, Ar-2CH and 3CH),
7.92 (d, J = 9.2 Hz, 1H, Ar-11CH), 8.77 (m, 1H, Ar-4CH), 8.93 (m,
1H, Ar-1CH). ES-MS (m ⁄ z) (Fig. S1A) calculated for C23H26N3OS
[M+]: 392.20; found: 392.30. HPLC (Fig. S3A, C18 column): retention
time of 13.2 min with 25–100% acetonitrile in water (0.1% TFA)
within 2–20 min.
Characterization 3b, EtNBS-COOH. 1H NMR (Fig. S2B, CDCl3,
CD3OD, 400 MHz): d = 1.37 (br t, 6H, 2 · NHCH2CH3), 1.82 (br s,
2H, NHCH2CH2CH2CH2), 1.93 (br s, 2H, NHCH2CH2CH2CH2),
2.47 (s, 2H, NHCH2CH2CH2CH2), 3.62–3.72 (m, 4H,
2 · NHCH2CH3), 3.79 (br t, 2H, NHCH2CH2CH2CH2), 6.98 (s, 1H,
Ar-6CH), 7.09 (s, 1H, Ar-8CH), 7.12 (d, J = 8.8 Hz, 1H, Ar-10CH),
7.83 (m, 2 H, Ar-2CH and 3CH), 8.01 (d, J = 8.8 Hz, 1H, Ar-11CH),
8.45 (m, 1H, Ar-4CH), 9.03 (m, 1H, Ar-1CH). ES-MS (m ⁄ z) (Fig. S1B)
calculated for C25H28N3O2S [M+]: 434.19; found: 434.20. HPLC (Fig.
S3B, C18 column): retention time of 16.2 min with 25–100% aceto-
nitrile in water (0.1% TFA) within 2–20 min.
Fluorescence quantum yield determination. The fluorescence quan-
tum yield of each PS was determined through comparison with the
standard, cresyl violet perchlorate (FF = 0.54) (16). The fluorescence
and absorption spectra of five, sequentially diluted, optically dilute
solutions (OD < 0.05) of each PS and of cresyl violet perchlorate were
measured (17). Additional fluorescence measurements of cresyl violet,
excited at 500 nm and measured from 550–750 nm, were obtained to
verify that the shape of the fluorescence curve was maintained. The
fluorescence area was plotted versus absorption, and the yield was
determined by comparison with the standard cresyl violet, according to
Eq. (1):
Synthesis of 3-(naphthalen-1-ylamino)propan-1-ol, 2a. A stirred
mixture of 1-naphthylamine (1 g, 6.99 mmol) and 3-bromo-1-propanol
(1.02 g, 7.30 mmol) were refluxed for nearly 18 h in ethanol (5 mL),
with intermittent monitoring by thin layer chromatography (TLC).
When the reaction was completed, the solvent was evaporated under
reduced pressure and the crude product was purified by column
chromatography (silica gel, 0–5% methanol in dichloromethane as an
eluent). The product was obtained as pale pink-colored oil (960 mg) in
68.5% yield. Electron spray mass spectroscopy (ES-MS) (m ⁄ z)
calculated for C13H16NO [M+]: 202.12; found: 202.00.

Photochemistry and Photobiology, 2009, 85 113
Bacterial PS binding assay. Overnight cultures of E. coli 25922
and S. aureus 29213 were inoculated from single colonies into the
appropriate culture medium. The following day overnight cultures
were diluted 1:1000 in fresh media and grown to mid-exponential
phase, and the value of cells per milliliter of culture was observed
through turbidometric analysis and compared with a standard curve
for cell concentration versus optical density. Two to four microliter
volumes of PSs were added to 1.5 mL tubes to produce a final
concentration of 7 lM upon the addition of 1 mL of mid-exponential
phase culture. Cultures were incubated with the PSs for various time
intervals and subsequently centrifuged for 4 min at 3000 g. The
supernatant was removed and the remaining pellet was washed with
fresh culture medium and subsequently centrifuged again for 4 min
at 3000 g twice. The washed pellets were then incubated overnight in
1 mL of 50% EtOH ⁄ 10% SDS at room temperature. The next day
the mixtures were centrifuged for 4 min at 3000 g and the
supernatant was removed for fluorescence analysis. Fluorescence
was determined using a fluorometer with the excitation wavelength
set to 650 nm and the emission wavelength set to 700 nm. The
integration time was 2 s and the slits were set at 2. All readings were
dFDye=dBDye
dFCV=dBCV
UFðDyeÞ
¼
UFðCVÞ
;
ð1Þ
where FF are the fluorescence quantum yields for the PS (Dye) and the
cresyl violet (CV) reference. The dF ⁄ dB are the slopes from the plot of
the integrated fluorescence intensity (measured from 615–900 nm with
excitation at 610 nm) versus absorbance at 610 nm.
This procedure was repeated multiple times in three solutions:
(1) methanol, (2) ethanol and (3) ethanol with 1% acetic acid.
Singlet oxygen quantum yield determination. The singlet oxygen
quantum yield was determined by the 1,3-diphenylisobenzofuran
(DPBF) bleaching method and was evaluated using Eq. (2) (18).
Equation 2 was derived from the equation for the bleaching reaction in
concert with the equation for singlet oxygen formation (19). Methylene
blue (FD = 0.50) was used as a reference (20). Solutions of DPBF and
each PS, in ethanol with 1% acetic acid, were adjusted to an optical
density of 1.0, for each, at their respective absorbance peaks of 411 and
655 nm. The solution was irradiated with a 670 nm laser light, from
the intensity of 0.005 to 0.05 W cm)2, until the optical density of the
DPBF was ca. 0.5. Equation 2 was then employed to determine the
singlet oxygen quantum yield of each PS:
taken using
a quartz cuvette. The concentration of PS was
determined through the use of standard curves of fluorescence
versus concentration of PS for each PS tested. This value was
then divided by the concentration of cells in culture in order to
determine the value of molecules of PS per bacterial cell at various
time points.
DCDPBFðSÞ=BStS
DCDPBFðRÞ=BRtR
UDS
¼
UDR
;
ð2Þ
where FD are the quantum yields and DC are the changes in DPBF
concentration as measured by the absorbance at 411 nm according to
Beer’s Law, A = ꢀcl. The B are the photons absorbed by each sensi-
tizer at 670 nm (the wavelength of excitation) and t is the time elapsed
between the measured concentrations.
In-culture fluorescence assay. Overnight cultures of E. coli 25922
and S. aureus 29213 were inoculated from single colonies into the
appropriate culture medium. The following day, overnight cultures
were diluted 1:1000 in fresh media and grown to mid-exponential
phase, and the value of cells per milliliter of culture was observed
through turbidometric analysis and comparison with a standard curve
for cell concentration versus optical density. Two to four microliter
volumes of PSs were added to 1.5 mL tubes to produce a final
concentration of 7 lM upon the addition of 1 mL of mid-exponential
phase culture. Following a 5 min incubation period, 100 mL of the cell
suspension was spotted onto 0.1% lysine coated glass slides and
incubated in the dark for 10 min. The slides were rinsed gently,
through immersion in phosphate-buffered saline solution, pH 7.4 and
subsequently covered with a cover slip. Cultures were viewed under oil
at an objective magnification of 100· using an Olympus fluorescence
microscope integrated with a Nuance hyperspectral camera. The
incident beam of the microscope was filtered for a wavelength of
633 nm, and the power of the beam was metered at 7.87 mW cm)2 at
an objective magnification of 4·. Images were captured every 0.75 s
until the change in fluorescence reached a plateau. The intensity of the
emission values of the images was determined as a function of time
using Image J v1.37. Kinetic images were arranged as stacks and
measured using the measure stack plug-in. Seven arbitrary areas were
selected for measure per field. Data were plotted as arbitrary
fluorescence units versus time.
Bacterial strains, media and culture conditions. Staphylococcus
aureus 29213 and E. coli 25922 were purchased from the ATCC.
Single colonies were routinely streaked on brain–heart infusion (BHI)
agar. Cultures were incubated at 37ꢁC and broth cultures were
incubated with shaking. Overnight cultures were routinely grown in
BHI broth to aid in synchronization of growth, in subsequent culture,
to mid-exponential phase. Mid-exponential cultures and susceptibility
test of E. coli 25922 were grown in Mueller–Hinton broth and
corresponding cultures of S. aureus 29213 were grown in BHI broth.
Susceptibility testing. The susceptibility of bacterial strains to PDT
with various PSs was determined through microdilution broth,
turbidometric growth inhibition assays in 96-well tissue culture plates.
Single colonies were picked and used to inoculate overnight cultures.
The next day overnight cultures were diluted 1:1000 in fresh media and
subsequently cultured to mid-exponential growth for use in MIC
assays. The cultures were diluted to
a
concentration of
1 · 107 CFU mL)1, and 50 lL aliquots were added to 96-well plates
containing 50 lL of PS in growth medium for a final concentration of
5 · 106 CFU mL)1. The mixtures were incubated at room temperature
for 5 min prior to irradiation with 670 nm laser light with a power of
15 mW cm)2. Following PDT, the optical density of cultures was
measured as a function of time. Growth was monitored until the
control culture, with no PS and no light exposure, reached the end of
the exponential growth phase. Inhibition profiles were generated as
described (21). Briefly, the area under the growth curve (AUC) of each
culture was determined using Simpson’s rule. The areas of cultures
grown in the presence of PS and with irradiation were divided by the
AUC of the culture grown without PS or irradiation resulting in the
fractional growth of the cultures. The fractional growth of cultures at
various fluences was plotted versus the concentration of PS in order to
generate the inhibition profiles shown in the figures.
RESULTS
Synthesis of EtNBS derivatives
The synthesis of the functionalized EtNBS derivatives
(Fig. 1b,c) mirrored methods typical for the generation of
benzo[a]phenothiazinium chlorides (7) and functionalized
Figure 1. Reported PS (a) EtNBS and new side-chain modified PSs, (b) EtNBS-OH and (c) EtNBS-COOH.

114 Sarika Verma et al.
Figure 2. (A) (a) Br(CH2)3OH, EtOH, reflux, 18 h; (b) Br(CH2)4COOC2H5, EtOH, reflux, 15 h; (c) 1 M NaOH, 1,4-dioxane, rt, 5 h. (B) (a)
Al2(SO4)3, Na2S2O3, ZnCl2, K2Cr2O7, H2O, 0ꢁC to rt, 2.5 h; (b) 2a or 2b, Ag2CO3, MeOH, reflux, 30 min.
benzo[a]phenoxazine PSs (22,23). Figure 2A shows the
Photophysical properties of EtNBS derivatives
synthesis of 3-(naphthalen-1-ylamino)propan-1-ol (2a) and
The physical and photophysical properties of the PS deriva-
tives, 3a and 3b, were compared with those previously reported
for EtNBS and are presented in Table 1. We note that for
photophysical evaluation, ethanol was used as a solvent to
prevent aggregation. However, both the PSs were highly
soluble in water and in vitro measurements were recorded in
phosphate-buffered saline. Absorbance and emission measure-
ments were recorded at very dilute but slightly different
concentrations of 1.5 lM for EtNBS-OH and 1.6 lM for
EtNBS-COOH (Fig. 3). Both PSs were found to have an
absorption maximum of 655 nm and an emission maximum of
685 nm. Side-chain modifications did not impart much change
ethyl 5-(naphthalen-1-ylamino)pentanoate (2b) from naph-
thalene-1-amine and the subsequent hydrolysis of 2b with
aqueous sodium hydroxide to produce the caroboxylic acid
derivative (2c). The yields of 2a, 2b and 2c were 68.5%, 67%
and 55%, respectively, and their predicted molecular weights
were confirmed through mass spectroscopy. This step incor-
porated the functional groups to be analyzed in future PDT
specificity against bacteria. In the final synthetic step
(Fig. 2B), the products from the first step (2a, 2c) were each
condensed with Bunte salt (sodium N-[2-amino-5-(diethyla-
mino)phenyl] sulfanesulfiniperoxoate) (15) in the presence of
silver carbonate, as an oxidizing agent, to obtain the side-
chain functionalized PSs, EtNBS-OH (3a) and EtNBS-
COOH (3b). After column chromatography, cationic PSs
3a and 3b were isolated as solid material in yields of 60%
and 55%, respectively (Fig. 2B).
Table 1. Photophysical properties of pertinent dyes.
Photosensitizer kabs (nm)* kfl (nm)†
Ffl‡
1O2§
Log P–
The two benzo[a]phenothiazinium chloride derivatives (3a
1
and 3b) were characterized by ES-MS, HPLC and H-NMR
EtNBS
EtNBS-OH
EtNBS-COOH
654(+)
655
655
693(+)
685
685
0.21(+) 0.03(+) 2.76(++)
spectroscopy. The HPLC run produced single peaks at the
retention times of 13.2 (3a) and 16.3 (3b) min for purified
EtNBS-OH and EtNBS-COOH, respectively. Mass spectro-
scopic analysis revealed the respective molecular weights of
393.3 (3a) and 434.2 (3b) for EtNBS-OH and EtNBS-COOH.
Structural confirmations were obtained using 1H-NMR and
demonstrated a complete integration accord between aromatic
and aliphatic protons with eight aromatic protons between 6.8
and 9.2 ppm for both 3a and 3b (Fig. S2A,B).
0.12
0.12
0.026
0.032
2.03
1.51
*Absorption maximum measured in ethanol containing 0.1% acetic
acid. †Fluorescence emission maximum measured in ethanol. ‡Abso-
lute fluorescence quantum yield measured with cresyl violet perchlo-
rate as a standard. §Absolute quantum yield of singlet oxygen
formation, using 1,3-diphenylisobenzofuran bleaching method.
–Partition coefficient between 2-octanol and phosphate-buffered saline
at pH 7.4. (+)Value from citation 6. (++)Value from citation 7.

Photochemistry and Photobiology, 2009, 85 115
Figure 3. Absorption ⁄ emission profile of EtNBS-OH and EtNBS-
COOH in EtOH. The slight deviation observed in absorbance and
fluorescence emission between EtNBS-OH and EtNBS-COOH is due
to the slight difference in the concentrations of the solutions used.
Figure 4. The phototoxicity of the EtNBS derivatives to Staphylococ-
cus aureus 29213 (A) and Escherichia coli 25922 (B). The susceptibility
of representative bacteria to photodynamic therapy at various con-
centrations of PS and at various light doses was assayed through a
microscale, turbidometric, broth susceptibility assay. All experiments
were performed three separate times and each consisted of three
replicates (bottom: EtNBS-COOH; top: EtNBS-OH).
to photophysical properties as both the derivatives had
similarly shaped spectral profiles and an identical fluorescence
quantum yield of 0.12.
The singlet oxygen quantum yield (18,19) was measured to
gauge the photoactivity of both the derivatives and DPBF was
used as a singlet oxygen chemical quencher with methylene
blue as a standard (20). As shown in Table 1, the singlet
oxygen quantum yields of PSs were measured to be 0.026
(EtNBS-OH) and 0.032 (EtNBS-COOH), values close to that
reported for EtNBS (0.025 [7] and 0.03 [10]).
The log P-value for the PSs was measured using conven-
tional shake-flask technique with 2-octanol as an oil phase and
phosphate-buffered saline as an aqueous phase (at pH 7.4),
and the results are shown in Table 1. Visually in log P studies,
EtNBS-OH shows a nearly colorless blue aqueous layer with a
dark-blue organic phase, whereas EtNBS-COOH, being more
hydrophilic, shows a light-blue aqueous phase and less-dark
organic phase. The data indicated that both EtNBS-OH and
EtNBS-COOH were more hydrophilic than EtNBS
(log P = 2.76).
assayed the uptake of both PSs by the bacteria. We hypoth-
esized that EtNBS-COOH was taken up by E. coli to a lesser
extent than was EtNBS-OH and that this difference was
responsible for the observed difference in phototoxicity.
However, the amount of PS associated with both E. coli and
S. aureus was identical for both of the EtNBS derivatives, and
the maximum level of molecules of dye per cell was achieved as
early as 5 min of incubation (Fig. 5). The binding was
consistent for extended incubation of 10 and 15 min and was
higher in E. coli than S. aureus. This demonstrated that the
binding of both PSs was rapid and that the cause of the
difference in phototoxicity of the two dyes observed in E. coli
was not due to a decrease in the amount of EtNBS-COOH
taken up by the bacteria. This finding suggested a difference in
the subcellular localization of the two PSs. The difference in
phototoxicity of the two PSs for E. coli suggested that EtNBS-
COOH was located at a position where it was unable to
inactivate the bacteria with illumination. Further investigation
of this possibility required an assay in which the association of
Photodynamic antibacterial activity of EtNBS derivatives
Both EtNBS derivatives were evaluated for their phototoxicity
in cultures of S. aureus and E. coli via turbidometric
susceptibility testing. Both PSs exhibited a light–dose-depen-
dent phototoxic effect in cultures of S. aureus (Fig. 4A).
EtNBS-OH produced greater growth inhibition than EtNBS-
COOH, but both PSs produced 100% growth inhibition at a
concentration of 1.75 lM and a fluence of 15 J cm)2. In
cultures of E. coli, EtNBS-OH exhibited near complete
inhibition of growth at a concentration of 14 lM and a fluence
of 12.5 J cm)2 (Fig. 4B, top). Contrary wise, EtNBS-COOH
produced no significant inhibition of growth across all
concentrations of PS and fluences (Fig. 4B, bottom).
Escherichia coli 25922 exhibited a greater level of resistance
to PDT with EtNBS-OH than S. aureus 29213.
Figure 5. The binding of the PSs by both Staphylococcus aureus 29213
and Escherichia coli 25922. The number of molecules of PS associated
with each bacterial cell was determined through measuring the
fluorescence of the lysate of washed bacterial pellets following
incubation with PS for various times. All experiments were performed
three separate times and each consisted of three replicates.
Bacterial binding of EtNBS derivatives
Considering the predicted difference in charge between the two
EtNBS derivatives at the pH of the culture medium, we

116 Sarika Verma et al.
EtNBS, indicating that side-chain modification had little effect
on the photophysical properties of the PS derivatives. The
difference in value of fluorescence quantum yield for the
derivatives (0.12) in comparison with parent EtNBS (0.21) (7)
could be attributed to a difference in the sensitivity of devices
used in the two studies. The fluorescence quantum yield
determined for EtNBS on our device was also 0.12 (data not
shown). The low singlet oxygen quantum yields of both the
derivatives (EtNBS-OH: 0.026; EtNBS-COOH: 0.032) were
similar to the reported yield for the parent EtNBS (0.03).
Although these values are low compared with the typical
singlet oxygen quantum yield reported (60–70%) for PSs used
in clinical PDT, EtNBS has been demonstrated to be highly
effective in the PDT of cancer (25), producing complete cure
for naturally occurring cancers in cats and dogs (26), suggest-
ing different photochemical mechanisms for cell killing.
Consistent with the log P-values, both the PSs readily dis-
solved in isotonic aqueous solutions in the absence of an
organic phase, indicating the potential for the aqueous
application of the PSs in PDT. Taken together, the fluores-
cence quantum yield, singlet oxygen and log P-values support
the view that the side-chain modification of the PSs to
incorporate functionalizable end groups does not adversely
affect their inherent physical ⁄ photophysical properties.
Interestingly, the antimicrobial phototoxic effects appeared
to be bacterial-type specific within the limitation of the strains
tested. While both the EtNBS functionalized derivatives were
demonstrated to be phototoxic to S. aureus, only the alcohol
derivative exhibited phototoxicity for E. coli. These observa-
tions are significant in that the modification of EtNBS through
the addition of small, nontoxic, functional groups created a
significant difference in the Gram-type-specificity of the PS, at
least in the two strains used in the present study. This is in
contrast to the previous findings from several groups, includ-
ing ours, where macromolecular conjugates or complexes
(14,24,27,28) were needed to disrupt the barrier function of the
Gram-negative bacteria. Potential problems associated with
macromolecules such as delivery or hemolytic activity (24)
might be circumvented by small functional group modifica-
tions. The outer layer of Gram-positive bacteria is composed
of a porous matrix of peptidoglycan that allows molecules as
large as 25 kDa to reach the cell membrane whereas Gram-
negative organisms possess a selectively permeable outer
membrane (OM) that stands as a physical barrier for the PS
in reaching the inner membrane, a site of lethal action (29). It
is believed that cationic PSs have a greater affinity for the
negatively charged phospholipids of Gram-negative bacteria
and enter more easily via self-promoted uptake in comparison
with anionic and neutral PSs, a notion in accord with the
dependence of antimicrobial peptide action on cationic charge
(27,28). Although some anionic or neutral PSs do bind to the
Gram-negative OM, they are not always effective in photody-
namic bacterial inactivation (29).
Figure 6. Fluorescence decay of PSs in bacterial cultures. (A) Escher-
ichia coli ⁄ EtNBS-OH. (B) Staphylococcus aureus ⁄ EtNBS-OH. (C)
S. aureus ⁄ EtNBS-COOH. Seven arbitrary points were selected per field
and the maximum fluorescence value was determined using Image J
v1.37. Each experiment was repeated twice by two different techni-
cians.
the PS with the bacteria was preserved. For this reason, we
observed the fluorescence of cells bound by PS in live bacterial
cultures via fluorescence microscopy.
In-culture fluorescence of EtNBS derivatives
Cultures of E. coli and S. aureus were incubated with
concentrations of PS identical to that used in the binding assay
and were subsequently spotted onto poly-L-lysine coated slides
for viewing via fluorescence microscopy. During our initial
observations, a rapid, light-induced decay of fluorescence was
exhibited by both PSs when in cultures of S. aureus. In order to
further characterize this phenomenon, the fluorescence emitted
from the bacteria in culture was recorded over time. Under
illumination, both PSs exhibited a reduction of nearly 80%
fluorescence over the course of 50 s in cultures of S. aureus
(Fig. 6B,C; Movies S1–S4). The fluorescence emissions of the
two PSs in cultures of E. coli were quite different. Interestingly,
only EtNBS-OH produced fluorescence in cultures of E. coli
with a similar near 80% reduction in fluorescence over the 50 s
interval, while EtNBS-COOH produced no observable fluo-
rescence (Fig. 6A, NA; Movies S1–S4).
DISCUSSION
Treatment efficacy depends in part on the specificity of the
therapeutic compound for the target disease. Typically, spec-
ificity to target is increased by using macromolecular conju-
gates (24). In this study, we explored the possibility of small
functional group modifications to enhance specificity. Accord-
ingly, the functionalized EtNBS derivatives were synthesized
employing two frequently used functional groups, alcohol and
carboxylic acid, and were evaluated for their photodynamic
activity. The absorption and emission spectra (Fig. 3) indi-
cated that both EtNBS-OH and EtNBS-COOH absorbed red
light, suggesting their utility in deeper tissue infection; both
were efficient fluorophores. These values are similar to that of
Fluorescence analysis of PS bound to bacteria revealed no
significant difference in the amount of either EtNBS derivative
that was associated by each bacteria type, consistent with the
difference in membrane structure between the Gram-positive
and Gram-negative organisms. The outer leaflet of the OM of
Gram-negative bacteria is composed of anionic phospholipids
and lipopolysaccharide and contributes to the OM’s negative
charge (30). Chelators and cationic antimicrobial peptides

Photochemistry and Photobiology, 2009, 85 117
sequester or displace, respectively, the normally stabilizing
divalent cations, thereby disrupting the OM and making it
permeable to otherwise ineffective antimicrobials (31). In light
of the difference in pKa between the alcohol (typically >9.0)
and carboxylic acid groups (typically <6.0), one would predict
a difference in interaction sites of the bacteria types with the
PS at pH 7.4 (bacterial culture pH).
Figure S2. 1H-NMR spectra of (A) EtNBS-OH (3a) and (B)
EtNBS-COOH (3b).
Figure S3. HPLC profile of (A) EtNBS-OH (3a) and (B)
EtNBS-COOH (3b).
Movie S1. In-culture fluorescence decay of EtNBS-COOH
in Escherichia coli 25922 (EcoliEtNBSCOOH).
Movie S2. In-culture fluorescence decay of EtNBS-OH in
Escherichia coli 25922 (EcoliEtNBSOH).
The reason for lack of fluorescence emission by EtNBS-
COOH in cultures of E. coli (Movies S1–S4) is nonobvious.
One potential cause could be the aggregation of EtNBS-
COOH molecules in culture. However, the fluorescence of both
EtNBS derivatives was stable in media over the time period of
4 h and 200 J cm)2 illumination with 670 nm laser light (data
not shown), indicating a lack of aggregation of EtNBS-COOH
under these conditions. Alternatively, illumination of the PSs
could lead to the consumption of available oxygen. In the
absence of sufficient oxygen, the PSs would undergo redox
reactions with other molecules and yield nonfluorescent PS.
The observation that the fluorescence emission decreases over
time with both EtNBS-OH and EtNBS-COOH in S. aureus
cultures and with EtNBS-OH in E. coli cultures could be
explained by oxygen depletion, but it does not explain the
complete lack of fluorescence emission observed for EtNBS-
COOH in cultures of E. coli. Perhaps EtNBS-COOH is
localized in proximity to an endogenous quencher within the
E. coli, explaining the observed differences in phototoxicity
and fluorescence emission. However, this is speculative and
requires elucidation of the interaction between the EtNBS-
COOH and various bacterial components.
In summary, this study demonstrates the complexities and
promise of an emerging approach to antibacterial therapy and
offers insight for the development of antimicrobials with
greater pathogen specificity in the host environment. Under-
standing the effect of small modifications on the phototoxicity
of these PSs builds a scaffold for their future application in the
design of disease-specific PSs. The addition of targeted func-
tional groups to PSs can be accomplished without disrupting
their photochemical and photophysical properties; however,
these results demonstrate that such modifications can create
observable differences in the photochemistry and photophysics
of PSs when associated with target cells. As significant level of
bacterial discrimination was observed due to the addition of a
carboxylic acid group, possibly minor modifications could
achieve similar discrimination between the host cell and the
target cell. Although these initial findings are encouraging,
further studies using several different strains of each bacteria
type are required to establish and explore such a strategy.
Movie S3. In-culture fluorescence decay of EtNBS-COOH
in Staphylococcus aureus 29213 (SaureusEtNBSCOOH).
Movie S4. In-culture fluorescence decay of EtNBS-COOH
in Staphylococcus aureus 29213 (SaureusEtNBSOH).
This material is available as part of the online article from:
http://www.blackwell-synergy.com/doi/full/10.1111/j.1751-1097.
2008.00403.x
Please note: Blackwell Publishing are not responsible for the
content or functionality of any supplementary materials sup-
plied by the authors. Any queries (other than missing material)
should be directed to the corresponding author for the article.
REFERENCES
1. Jori, G. (2006) Photodynamic therapy of microbial infections:
State of the art and perspectives. J. Environ. Pathol. Toxicol.
Oncol. 25, 505–519.
2. Jori, G., C. Fabris, M. Soncin, S. Ferro, O. Coppellotti, D. Dei,
L. Fantetti, G. Chiti and G. Roncucci (2006) Photodynamic
therapy in the treatment of microbial infections: Basic principles
and perspective applications. Lasers Surg. Med. 38, 468–481.
3. Hamblin, M. R. and T. Hasan (2004) Photodynamic therapy:
A new antimicrobial approach to infectious disease? Photochem.
Photobiol. Sci. 3, 436–450.
4. Dougherty, T. J. and S. L. Marcus (1992) Photodynamic therapy.
Eur. J. Cancer 28A, 1734–1742.
5. Dougherty, T. J., C. J. Gomer, B. W. Henderson, G. Jori,
D. Kessel, M. Korbelik, J. Moan and Q. Peng (1998) Photody-
namic therapy. J. Natl Cancer Inst. 90, 889–905.
6. Akilov, O. E., S. Kosaka, K. O’Riordan, X. Song, M. Sherwood,
T. J. Flotte, J. W. Foley and T. Hasan (2006) The role of pho-
tosensitizer molecular charge and structure on the efficacy of
photodynamic therapy against Leishmania parasites. Chem. Biol.
3, 839–847.
7. Foley, J. W., X. Song, T. N. Demidova, F. Jalil and
M. R. Hamblin (2006) Synthesis and properties of benzo[a]
phenoxazinium chalcogen analogues as novel broad-spectrum
antimicrobial photosensitizers. J. Med. Chem. 49, 5291–5299.
8. Wainwright, M., M. N. Byrne and M. A. Gattrell (2006) Pheno-
thiazinium-based photobactericidal materials. J. Photochem.
Photobiol. B 84, 227–230.
9. Wainwright, M., H. Mohr and W. H. Walker (2007) Phenothi-
azinium derivatives for pathogen inactivation in blood products.
J. Photochem. Photobiol. B 86, 45–58.
10. Wainwright, M. and R. M. Giddens (2003) Phenothiazinium
photosensitizers: Choices in synthesis and application. Dyes
Pigments 57, 245–257.
11. Socha, J. (1995) Intestinal microflora and antibiotic therapy.
Pediatr. Pol. 17, 547–552.
12. Pasic, M., R. Jost, T. Carrel, L. Von Segesser and M. Turina
(1993) Intracolonic vancomycin for pseudomembranous colitis.
N. Engl. J. Med. 329, 583.
Acknowledgements—We would like to acknowledge Gregory Watt,
PhD, for his helpful discussions during the initial synthetic design and
Xiang Zheng, PhD, for the review of this work and insightful critiques.
This work was funded by the U.S. Air Force, Medical Free Electron
Laser Program (MFEL) (grant FA9550-04-1-0079).
13. Apisarnthanarak, A., H. Khoury, W. R. Reinus, J. S. Crippin and
L. M. Mundy (2002) Severe Clostridium difficile colitis: The role
of intracolonic vancomycin? Am. J. Med. 112, 328–329.
14. Grewer, C., G. Schermann, R. Schmidt, A. Volcker, H. D. Brauer,
A. Meier and F. P. Montforts (1991) Potential photosensitizers for
photodynamic therapy. III. Photophysical properties of a lipo-
philic chlorin and its zinc and tin chelates. J. Photochem. Photo-
biol. B 11, 285–293.
SUPPLEMENTAL MATERIALS
The following supplemental materials are available for this
article:
Figure S1. ES-MS spectra of (A) EtNBS-OH (3a) and (B)
EtNBS-COOH (3b).

118 Sarika Verma et al.
15. Clapp, R. C., J. P. English, C. E. Fellows, J. Forsythe, R. E. Grotz
and R. G. Shepherd (1952) Chemotherapeutic dyes. V. Benzo
[a]phenothiazines and benzo [a] phenazines. J. Am. Chem. Soc. 74,
1994–1996.
16. Magde, D., J. H. Brannon, T. L. Cremers and J. Olmsted III
(1979) Absolute luminescence yield of cresyl violet. A standard for
the red. J. Phys. Chem. 83, 696–699.
as long-wavelength fluorescent probes for amino acids. Tetrahe-
dron 63, 1654–1663.
24. Hamblin, M. R., D. A. O’Donnell, N. Murthy, K. Rajagopalan,
N. Michaud, M. E. Sherwood and T. Hasan (2002) Polycationic
photosensitizer conjugates: Effects of chain length and Gram
classification on the photodynamic inactivation of bacteria.
J. Antimicrob. Chemother. 49, 941–951.
17. Morris, J. V., M. A. Mahaney and J. R. Huber (1976) Fluores-
cence quantum yield determinations. 9,10-Diphenylanthracene as
a reference standard in different solvents. J. Phys. Chem. 80, 969–
974.
25. Cincotta, L., J. W. Foley, T. MacEachern, E. Lampros and
A. H. Cincotta (1994) Novel photodynamic effects of a benzo-
phenothiazine on two different murine sarcomas. Cancer Res. 54,
1249–1258.
18. Cincotta, L., J. W. Foley and A. H. Cincotta (1987) Novel red
absorbing benzo[a]phenoxazinium and benzo[a]phenothiazinium
photosensitizers: In vitro evaluation. Photochem. Photobiol. 46,
751–758.
26. Frimberger, A. E., A. S. Moore, L. Cincotta, S. M. Cotter and
J. W. Foley (1998) Photodynamic therapy of naturally occurring
tumors in animals using a novel benzophenothiazine photosensi-
tizer. Clin. Cancer Res. 4, 2207–2218.
19. Foote, C. S. (1979) Quenching of singlet oxygen. In Singlet
Oxygen (Edited by H. H. Wasserman and R. W. Murray), pp.
139–171. Academic Press, New York, NY.
20. Belfield, K. D., C. C Corredor, A. R. Morales, M. A. Dessources
and F. E. Hernandez (2006) Synthesis and characterization of new
fluorene-based singlet oxygen sensitizers. J. Fluoresc. 16, 105–110.
21. Lambert, R. J. and J. Pearson (2000) Susceptibility testing:
Accurate and reproducible minimum inhibitory concentration
(MIC) and non-inhibitory concentration (NIC) values. J. Appl.
Microbiol. 88, 784–790.
27. Jiang, Z., A. I. Vasil, J. Hale, R. E. Hancock, M. L. Vasil and
R. S. Hodges (2008) Effects of net charge and the number of posi-
tively charged residues on the biological activity of amphipathic
alpha-helical cationic antimicrobial peptides. Biopolymers 90, 369–
383.
28. Taheri-Araghi, S. and B. Y. Ha (2007) Physical basis for
membrane-charge selectivity of cationic antimicrobial peptides.
Phys. Rev. Lett. 98, 168101.
29. King, B. F., M. L. Biel and B. J. Wilkinson (1980) Facile pene-
tration of the Staphylococcus aureus capsule by lysostaphin. Infect.
Immun. 29, 892–896.
30. White, D. (2000) The Physiology and Biochemistry of Prokaryotes.
Oxford University Press Inc., New York, NY.
22. Frade, V. H. J., M. T. Goncalves and J. C. V. P. Moura (2006)
Synthesis
of
fluorescent
water-soluble
functionalised
benzo[a]phenoxazinium salts. Tetrahedron Lett. 47, 8567–8570.
23. Frade, V. H. J., P. J. G. Coutinho, J. C. V. P. Moura and
M. T. Goncalves (2007) Functionalized benzo[a]phenoxazine dyes
31. Vaara, M. (1992) Agents that increase the permeability of the
outer membrane. Microbiol. Rev. 3, 395–411.