1-s2.0-S0038092X0800128X-main

Available online at www.sciencedirect.com

www.elsevier.com/locate/solener

Solar Energy 82 (2008) 1107–1117

Involvement of both Type I and Type II mechanismsin Gram-positive and Gram-negative bacteria photosensitization

by a meso-substituted cationic porphyrin

Karim Ergaieg a,*, Martine Chevanne b, Josiane Cillard b, Rene Seux a

a Laboratoire d’Etude et de Recherche en Environnement et Sante, National School of Public Health, Av. Pr. Leon Bernard, CS 74312, Rennes 35043, Franceb Laboratoire de Biologie Cellulaire et Vegetale, UPRES 3891, UFR des Sciences Pharmaceutiques et Biologiques,

University of Rennes 1, 2 Av. Pr. Leon Bernard, CS 34317, Rennes 35043, France

Received 28 April 2007; accepted 16 May 2008Available online 17 June 2008

Communicated by: Associate Editor Gion Calzaferri

Abstract

A meso-substituted cationic porphyrin (TMPyP) showed a photocytotoxicity against Gram-positive and Gram-negative bacteria. Inorder to determine the mechanism involved in the phototoxicity of this photosensitizer, electron paramagnetic resonance (EPR) exper-iments with 2,2,6,6-tetramethyl-4-piperidone (TEMP), a specific probe for singlet oxygen, and the spin-trap 5,5-dimethyl-1-pyrroline-N-oxide (DMPO) were carried out with illuminated TMPyP. An EPR signal characteristic of TEMP-singlet oxygen (TEMPO) adductformation was observed, which could be ascribed to singlet oxygen (1O2) generated by TMPyP photosensitization. The signal for theDMPO spin adduct of superoxide anion (DMPO-OOH) was observed in DMSO solution but not in aqueous conditions. However,an EPR spectrum characteristic of the DMPO-hydroxyl radical spin adduct (DMPO-OH) was observed in aqueous conditions. Theobtained results testify a primary hydroxyl radical (�OH) generation probably from superoxide anion ðO��2 Þ via the Fenton reactionand/or via Haber-Weiss reaction. Gram-positive and Gram-negative bacteria inactivation by TMPyP photosensitization predominantlyinvolved Type II reactions mediated by the formation of 1O2, as demonstrated by the effect of quenchers for 1O2 and scavengers for �OH(sodium azide, thiourea, and dimethylsulphoxide). Participation of other active oxygen species cannot however be neglected since Type Ireactions also had a significant effect, particularly for Gram-negative bacteria. For Gram-negative bacteria the photoinactivation ratewas lower in the presence of superoxide dismutase, a specific O��2 scavenger, and/or catalase, an enzyme which specifically eliminatesH2O2, but was unchanged for Gram-positive bacteria. The generation of 1O2, O��2 and �OH by TMPyP photosensitization indicated thatTMPyP maintained a photodynamic activity in terms of Type I and Type II mechanisms.� 2008 Elsevier Ltd. All rights reserved.

Keywords: TMPyP; Photosensitization; Active oxygen; EPR; Photoinactivation

1. Introduction

Sensitized photo-oxidation, also referred as photosensi-tization or photodynamic action, is an indirect photochem-ical reaction very similar to photocatalytic oxidation(Cooper and Goswami, 1998). This process may offer an

0038-092X/$ - see front matter � 2008 Elsevier Ltd. All rights reserved.

doi:10.1016/j.solener.2008.05.008

* Corresponding author. Tel.: +33 2 9902 2918; fax: +33 2 9902 2929.E-mail address: [email protected] (K. Ergaieg).

advantage over the photocatalytic process because the sen-sitizers can absorb light in the visible spectrum, allowingfor use of a greater percentage of available sunlight. Photo-sensitization can represent a useful approach for the killingof microbial cells since it has been shown that several por-phyrins and related compounds display phototoxicityagainst bacteria (Merchat et al., 1996a,b; Minnock et al.,1996, 2000), yeast (Lambrechts et al., 2005), and Helmintheggs (Alouini and Jemli, 2001). Photosensitization seems to

mailto:[email protected]

1108 K. Ergaieg et al. / Solar Energy 82 (2008) 1107–1117

be a promising ecologically-friendly water disinfectiontechnique (Jimenez-Hernandez et al., 2006) which employsthe visible light (or even sunlight) as energy source and theoxygen dissolved in water as the oxidizing agent. This pro-cedure is based on the exposure of cells to a photosensitis-ing agent which is activated by irradiation with visible lightof wavelength compatible with its absorption spectrum.Upon light activation, the photosensitizer generates activeoxygen species which can modify many biological mole-cules and eventually lead to cell death (Moan and Peng,2003). The excited photosensitizer may undergo a photoin-duced electron transfer (reductive or oxidative) and/orexchange an hydrogen atom, producing radicals and activeoxygen species such as the superoxide ðO��2 Þ and the hydro-xyl (�OH) radicals. This pathway is termed as a Type I pro-cess. The excited state can also transfer energy to dioxygen,in a so-called Type II mechanism generating singlet oxygen(1O2) (Foote, 1991). Singlet oxygen is thought to play a keyrole in photosensitized inactivation of bacteria by porphy-rins (Henderson and Dougherty, 1992). However, even incases where Type I reactions may occur for these com-pounds, the Type II reactions will usually take place in tan-dem and it is difficult to differentiate the photobiologicaleffects that are exclusively due to radical species. Thus,the question of whether a singlet oxygen or a free radicalmechanism is involved in the photodamage is still unsettled(Tanielian et al., 2000).

Regarding bacterial inactivation, it has been shown thatGram-positive bacteria are sensitive to photosensitizationby many different dyes, while Gram-negative bacteria aremore resistant, being destroyed only after increasing thepermeability of the outer membrane either by pre-treat-ment with different chemical (Bertoloni et al., 1990) or bio-logical agents (Nitzan et al., 1992) or by employing cationicphotosensitizers (Merchat et al., 1996a). Jori and Brown(2004) suggest that the positive charge favours the bindingof the photosensitizer molecule at critical cellular sites thatonce damaged by exposure to light cause the loss of cellviability. In this context, a meso-substituted cationic por-phyrin (TMPyP), whose photocytotoxicity against bacteriahas been proved previously, has retained our attention.

Over the last decade, photosensitized inactivation mech-anisms of Gram-positive and Gram-negative bacteria bycationic meso-substituted porphyrins have been well stud-ied but none of these studies gave importance to the mech-anisms of formation of active oxygen species (Merchat etal., 1996b; Reddi et al., 2002; Salmon-Divon et al., 2004).In this paper, we have attempted to determine the natureof the oxygenated intermediates which occur during thephotosensitization of a meso-substituted cationic porphyrin(TMPyP) in either air-saturated buffered and aqueous con-ditions, or DMSO solution. We used Electron Paramag-netic Resonance (EPR) spectroscopy to detect and/orstudy short-lived intermediates such as 1O2 and the freeradicals O��2 and �OH. In order to determine the mechanisminvolved in the phototoxicity of TMPyP, cell photosensiti-zation studies on Gram-positive and Gram-negative bacte-

ria were performed in the presence of specific scavengersand quenchers of active oxygen species.

2. Materials and methods

2.1. Chemicals

Meso-tetra (N-methyl-4-pyridyl) porphyrin tetra-tosyl-ate (TMPyP) was obtained from Frontier Scientific Inc.(Logan, UT, USA). 5,5-Dimethyl-1-pyrroline-N-oxide(DMPO), 2,2,6,6-tetramethyl-4-piperidone (TEMP), deute-rium oxide (D2O), sodium azide, thiourea, b-carotene,superoxide dismutase (SOD), catalase (CAT), bovineserum albumin, xanthine, xanthine-oxidase, dimethyl-sulphoxide (DMSO), ethanol, mannitol, desferrioxamine(DFO), and hydrogen peroxide (H2O2) were purchasedfrom Sigma–Aldrich (Saint-Quentin-Fallavier, France).Tetrahydrofuran (THF) was provided by Carlo Erba-SDS (Val de Reuil, France). Histidine was provided byHoffmann-La Roche Ltd. (Bale, Switzerland). All theseproducts were used without purification, with the exceptionof DMPO which was purified before use by filteringthrough activated charcoal and was stored according tothe method described by Floyd et al. (1984).

2.2. EPR studies of the photogeneration of 1O2 and free

radicals (O��2 , �OH)

2.2.1. Irradiation procedure

The photosensitizer (TMPyP) was exposed to a 300 Whigh-pressure arc xenon lamp (ILLUM 4000, Eurosep,Cergy Pontoise, France) for different durations andstrengths of illumination. The spectral emission of thislamp mimics sunlight. This is obtained by using a coloredglass filter (Melles Griot Inc., Irvine, California, USA) tocut the wavelengths below 305 nm. The distance betweenthe illuminated samples (300 lL) placed in an EPR spec-trometer cavity and the light source was �20 cm. For thedetection of DMPO-OH adducts, irradiation of the sam-ples was carried out outside the microwave cavity in a300 lL aqueous flat cell as described previously (Viola etal., 1996; Hadjur et al., 1997). The illuminated photosensi-tizer was placed at 10 cm from the source of illuminationwith a light intensity ranging from 0.095 to 0.285 W cm�2,and then immediately transferred into quartz capillaries forEPR analysis. Samples were illuminated for periods rangefrom 1 to 15 min. The intensity of illumination was mea-sured by a luxmeter (TES-1339, TES Electrical ElectronicCorp., Taipei, Taiwan).

2.2.2. Measurement by EPR

EPR spectra were recorded with a Bruker Model ESP106 spectrometer operating at room temperature (22–24 �C). Two types of EPR spectrometer cavities were used.The first was specially designed for direct irradiation of thesamples (300 lL) inside the microwave cavity. EPR wasperformed under the following conditions: magnetic field,

K. Ergaieg et al. / Solar Energy 82 (2008) 1107–1117 1109

3478 G; microwave power, 20 mW; modulation frequency,100 kHz; modulation amplitude, 2.02 G; sweep width,80 G; conversion time, 82 ms. In some experiments,another EPR microwave cavity (significantly more sensi-tive), which is not designed for in situ direct irradiation,was also used for DMPO-OH detection. After irradiation,photoinduced EPR spectra were obtained from the samples(300 lL) immediately transferred into quartz capillariesspecially designed for EPR analysis. The instruments con-ditions were the same as above except: magnetic field,3497 G; microwave power, 10 mW; modulation amplitude,1.99 G. The results were reported either as arbitrary unitsgiven by computer double integration of the first derivativeof the EPR signal, or given by percentage of the EPR signalinhibition or enhancement.

2.2.3. EPR determination of 1O2 generation

The standard reaction mixture (300 lL) consisted of0.01 M air-saturated phosphate buffer (pH 7.4), 50 mMTEMP, a specific probe for singlet oxygen, and TMPyPat concentration range from 0.073 to 0.73 lM. In someexperiments, various agents were also added to the reactionmixture. Except for b-carotene, all the agents were dis-solved in phosphate buffer solution (pH 7.4). b-Carotenewas dissolved in tetrahydrofuran (THF). Samples were illu-minated for periods range from 1 to 15 min and then ana-lysed by EPR.

2.2.4. Detection of the free radicals �OH and O��2 by spin

trapping

The spin-trapping technique using DMPO as a spin trapwas used to determine the generation of free radicals (�OH,O��2 Þ by the photosensitizer. The standard reaction mixtureconsisted of air-saturated aqueous solution containingTMPyP (73 lM or 114 lM) and DMPO (40 or 120 mM)in a total volume of 300 lL. For the detection of superox-ide radical anion, DMPO (120 mM) and TMPyP (114 lM)were dissolved in DMSO. Superoxide anion was also gen-erated by 750 lM xanthine and 3 U mL�1 xanthine-oxi-dase in an aqueous solution consisting of 30% DMSO.Samples were illuminated as above and then analysed byEPR.

2.3. Cell photosensitization studies

2.3.1. Bacterial strains

We worked with two strains: Gram-positive E. hiraeATCC 10541, and Gram-negative Escherichia coli ATCC25922. Both strains were grown on nutritive agar (BiokarDiagnostic, Beauvais, France). After incubation, bacteriain stationary growth phase were suspended in a sterilephosphate buffer solution (PBS) at pH 7.2 containing2.7 mM KCl and 0.14 M NaCl. The bacterial pellet wasseparated from the nutritive medium by centrifugation(3000g for 10 min) followed by two washings in the sameconditions then resuspended to a final concentration of105 cells mL�1. The microplate counting system (BIO-

RAD, Marnes-la-Coquette, France) was retained for thisstudy to determine the most probable number (MPN) ofbacteria present in 100 mL water with a 95% confidenceinterval.

2.3.2. Sunlight simulator

A 300 W high-pressure arc xenon lamp was the lightsource. The spectral emission of this lamp mimics sunlight.This is obtained by using silver-coated mirror, as well as acolored glass filter to cut the wavelengths below 305 nm.The distance between the lamp and the irradiated waterslide was 10 cm so that the light beam was the same diam-eter as the plastic Petri dish (Fig. 1). The light intensity atthis distance was measured to be 0.285 W cm�2.

2.3.3. Irradiation procedure of cells

In all experiments the cell suspensions, having a densityvarying little around 105 cells mL�1 were added with suit-able volumes of the various chemicals to yield the concen-tration sought in the reaction medium. Before starting theirradiation, cell suspensions were maintained in the darkfor 10 min under moderate stirring to obtain a homoge-neous reaction medium (dissolved oxygen content to theorder of 8 mg mL�1). Experiments were run without stir-ring or supplementary aeration in plastic Petri dishes(diameter 54 mm) containing a 1 cm thick layer of water.Dark controls were used to check the toxicity of TMPyPor the other reagents without light.

2.3.4. Reactive oxygen species

The predominance of Type II reactions which involvethe generation of singlet oxygen compared to other reactiveoxygen species which intervene in Type I reactions, partic-ularly the hydroxyl radical, were evaluated by judiciouschoice of antioxidants. The choice of antioxidants dependson their bimolecular rate constants for reaction with thehydroxyl radical and singlet oxygen. The correspondingrate constants are nearly the equivalent for the hydroxylradical and very different for singlet oxygen (sodiumazide� thiourea > dimethylsulphoxide). We give theirvalue in Table 1. To estimate the respective contributionsof superoxide anion and hydrogen peroxide to Type I reac-tions, we used specific enzymes to eliminate these reactiveentities. Bovine superoxide dismutase (SOD) was used ata concentration of 55 U mL�1 and bovine catalase (CAT)at a concentration of 220 U mL�1 to eliminate superoxideanion and hydrogen peroxide, respectively, in the reactionmedium as described by Gourmelon et al. (1994).

2.3.5. Data processing

The inactivation of micro-organisms by light and otherprocessing methods has been traditionally assumed to fol-low first-order kinetics. The model assumes a linear rela-tionship between the decline in the logarithm of thenumber of survivors over treatment time (Peleg, 2003):

Fig. 1. Sunlight simulation setup.

Table 1Bimolecular rate constants for reaction between antioxidants and thehydroxyl radical (Dorfman and Adams, 1973) and singlet oxygen(Wasserman and Murray, 1978; Ranby and Rabek, 1978)

Antioxidants Scavenger rate constant(M�1 s�1)

Quencher rate constant(M�1 s�1)

Hydroxyl radical Singlet oxygen

Sodium azide 7.5 � 109 5 � 108

Thiourea 5 � 109 8 � 105

Dimethylsulphoxide 7 � 109 �103


log10

NðtÞN 0

� �¼ � t

Dð1Þ

where N0 is the initial number of cells (MPN per 100 mL),N the number of survivors after an exposure time t (MPNper 100 mL), D (decimal reduction time) the time requiredto destroy 90% of the organisms (min), and t is the treat-ment time (min).

In order to model the time course of bacterial countsduring exposure to light, we evaluated and compared thegoodness-of-fit of this linear model to a non-linear model(Weibull) (Peleg and Cole, 1998). The Weibull modelassumes that cells in a population have different resis-tances, and a survival curve is just the cumulative formof a distribution of lethal agents. The cumulative form ofthe Weibull distribution is as follows:

log10

NðtÞN 0

� �¼ �btn ð2Þ

where b and n are the scale and shape factors.

The mean square error (MSE) was used to compareboth models (3). The smaller the MSE values, the betterthe fit of the model to the data (Neter et al., 1996)

MSE ¼Pðpredicted-observedÞ2

n� pð3Þ

The letter n stands for the number of observations and p isthe number of parameters to be estimated.

3. Results and discussion

3.1. Detection of 1O2 and free radicals (O��2 , �OH)

3.1.1. Production of 1O2 by photosensitization of TMPyP

It has been previously reported that TEMPO, a nitrox-ide radical detectable by EPR, was generated from TEMPand singlet oxygen (4) (Lion et al., 1976; Moan and Wold,1979)

TEMPþ 1O2 ! TEMPO ð4ÞThe standard reaction mixture containing TEMP andTMPyP was exposed to light (0.095–0.285 W cm�2) forperiods of time ranging from 1 to 15 min. The EPR spectraobtained showed three equal intensity lines characteristic ofTEMPO nitroxide radical (Fig. 2a), with a g valueof 2.0064 ± 0.0002 and a superfine splitting constant (aN)of 16.03 G, which might suggest that singlet oxygen wasproduced by photosensitization of TMPyP. No EPR signalwas observed when the reaction mixture was measuredwithout light irradiation or when TMPyP was absent fromthe reaction mixture.

0

2

4

6

8

10

12

14

0 2 4 6 8 10 12 14 16Duration of illumination (min)

10 G (a) (b)

0

2

4

6

8

10

12

14

0 2 4 6 8 10 12 14 16Duration of illumination (min)

10 G (a) (b)

0

2

4

6

8

10

12

14

0 2 4 6 8 10 12 14 16

Duration of illumination (min)

10 G (a) (c)

0

2

4

6

8

10

12

14

0 2 4 6 8 10 12 14 16


10 G (a) (c)

0

2

4

6

8

10

12

14

0 2 4 6 8 10 12 14 16


10 G (a) (c)

TE

MP

O E

PR

sig

nal i

nten

sity

(a.

u.)

TE

MP

O E

PR

sig

nal i

nten

sity

(a.

u.)

Fig. 2. (a) EPR signal of TEMPO, (b) Effect of TMPyP concentration on TEMPO production. The samples were composed of 0.01 M phosphate buffer(pH 7.4), 50 mM TEMP, and TMPyP at various concentrations: (h) 0.073 lM, (}) 0.36 lM, and (�) 0.73 lM, and were exposed to light at an intensity of0.285 W cm�2, (c) Effect of light intensity on TEMPO production. The samples were composed of 0.01 M phosphate buffer (pH 7.4), 50 mM TEMP, and0.73 lM TMPyP, and were exposed to light at various intensities: (h) 0.095 W cm�2, (}) 0.190 W cm�2, and (�) 0.285 W cm�2.


3.1.2. Relationship between TMPyP concentration, light

intensity and TEMPO production

The intensity of the TEMPO EPR signal was measuredas function of increasing concentrations of TMPyP (Fig.2b). A linear relationship was observed between the pro-duction of TEMPO at a given TMPyP concentration andthe duration of illumination up to 15 min (Fig. 2b). Linear-ity was observed for TMPyP concentrations ranging from0.073 to 0.73 lM. In addition, the production of TEMPOshowed a light intensity-dependent increase (Fig. 2c). Ateach of the light intensities tested (0.095–0.285 W cm�2),a linear relationship was observed between the productionof TEMPO by TMPyP (0.73 lM) photosensitization andthe duration of illumination up to 15 min. Ando et al.(1997) reported that the production of TEMPO was pro-portional to the concentration of an hematoporphyrinderivative (HpD) up to 1 mM. In their experiments, theseauthors have used a light source having intensity 10-foldlower than that applied in our tests.

3.1.3. Effect of 1O2 quenchers or enhancer on TEMPO

production

In order to determine if the TEMPO EPR signal arisethrough the photosensitized production of singlet oxygenby TMPyP, we tested several agents known to quench

1O2 or to enhance the lifetime of 1O2. The quenchers used(sodium azide, histidine, or b-carotene) were found toreduce significantly the TEMPO EPR signal intensity(Fig. 3a). b-Carotene were found to give the best quenchingefficiency. Conversely, the EPR signal of TEMPO was sig-nificantly increased in the presence of 30% or 50% D2O(Fig. 3a). D2O is well known to increase the lifetime of1O2. These phenomena suggest that TEMPO is derivedfrom the reaction of TEMP with 1O2 formed by energy-transfer from the lowest excited triplet state of the photo-sensitizer TMPyP to ground state oxygen

3TMPyP� þ 3O2 ! TMPyP þ 1O2 ð5Þ

3.1.4. Contribution of O��2 and H2O2 to the production of 1O2

The effects of specific scavengers of O��2 , H2O2, and �OHon the production of TEMPO adduct during TMPyP illu-mination were examined in EPR experiments. The additionof superoxide dismutase (SOD, 300 U mL�1), a specificsuperoxide anion scavenger, or/and catalase (CAT,300 U mL�1), an enzyme which eliminates specificallyH2O2, significantly reduced the intensity of the EPR signalof TEMPO (Fig. 3b). In the presence of a powerful scaven-ger of hydroxyl radicals such as DMSO (10 mM), theintensity of the EPR signal of TEMPO was not signifi-

H

Histidine -carotene D2O

h

Control Sodium azide

0

150

0.01 mM 0.1 mM 0.1 m 1 mM 0.01 mM 0.05 mM 30% 50%H

Histidine -carotene D2O

h

Control Sodium azide

50

100T

EM

PO

EP

R s

igna

l int

ensi

ty (

%)

0

1 2 3 4 5 6300 U mL-1 300 U mL-1 10 mM 44 mM

Control SOD CAT SOD + CAT DMSO H2O2

0

20

40

60

80

100

120

140

160

1 2 3 4 5 6300 U mL-1 300 U mL-1 10 mM 44 mM

Control SOD CAT SOD + CAT DMSO H2O2

TE

MP

O E

PR

sig

nal i

nten

sity

(%

)

Fig. 3. Effect of different agents on TEMPO production. Samples were composed of 0.01 M phosphate buffer (pH 7.4), TEMP (50 mM), TMPyP (0.73 lM),and added with: (a) sodium azide (0.01 or 0.1 mM), histidine (0.1 or 1 mM), b-carotene (0.01 or 0.05 mM), D2O (30% or 50%) and (b) SOD (300 U mL�1)and/or CAT (300 U mL�1), DMSO (10 mM), H2O2 (44 mM). All the samples were directly irradiated in the resonant cavity at light intensity of0.095 W cm�2 for 2 min. The values are the average of three experiments, and the bars indicate standard deviation.


cantly different, whereas the addition of H2O2 (44 mM)markedly increased the TEMPO EPR signal intensity(Fig. 3b). These results suggested that H2O2 and O��2 couldbe involved in the production of 1O2. It has been previouslyreported that the production of 1O2 could originate fromthe Haber-Weiss reaction (6) (Khan and Kasha, 1994).These authors also proposed a possible production of 1O2

by electron transfer from O��2 to �OH (7). This reaction willprobably not occur in our system since a scavenger ofhydroxyl radical, DMSO, did not decrease significantlythe EPR signal of TEMPO.

O��2 þH2O2 ! 1O2ð1DgÞ þ �OHþ �OH ð6Þ�OHþO��2 ! 1O2ð1DgÞ þ �OH ð7Þ

3.1.5. Generation of O��2 by TMPyP

To assess the possible contribution of a Type I mecha-nism in the photosensitization process of TMPyP, theresulting generation of superoxide anion radical ðO��2 Þand/or hydroxyl radical (�OH) was investigated. The reac-tion between molecular oxygen and TMPyP was studiedin the presence of the spin-trap DMPO. This yields aDMPO-OOH adduct with O��2 and a DMPO-OH adductwith �OH with characteristic EPR spectra (Buettner andOberley, 1978; Finkelstein et al., 1980a,b). The signal forthe DMPO spin adduct of superoxide anion was notobserved in aqueous solutions. In this condition, the disap-pearance of superoxide is very rapid and is therefore diffi-cult to detect. The generation of superoxide in aproticsolvents, such as DMSO is extremely useful because it sta-

Fig. 4. (a) An example of the superoxide spin adduct of DMPO generated by an aqueous solution consisting of DMSO (30%), xanthine (750 lM),xanthine-oxidase (3 U mL�1). (b) Control run in the dark with TMPyP (114 lM) and DMPO (120 mM) in air-saturated DMSO solution. (c) Example ofthe superoxide spin adducts (DMPO-OOH) produced upon irradiation of the system TMPyP/DMPO in DMSO, directly in the resonant cavity at lightintensity of 0.095 W cm�2 for 10 s, (d) the same as (c) except the scan was started �2 min later, (e) the same as (c) except the scan was started 20 min later.


bilized the molecule for several days unlike seconds thatwas shown in aqueous solutions. The EPR spectrumreported in Fig. 4a showed the DMPO-OOH adduct gener-ated by the biological system xanthine/xantine-oxidase.The EPR spectra of air-saturated DMSO solution contain-ing TMPyP and the spin-trap DMPO showed only tracesof radicals when kept in the dark; these signals could beassigned to DMPO impurities and radicals formed due todim light exposure during sample preparation (Fig. 4b).After 10 s of TMPyP illumination inside the EPR micro-wave cavity in the presence of DMPO, a marked EPR sig-nal was observed (Fig. 4c). This EPR spectrum showed thecharacteristic lines of the superoxide adduct of DMPO sim-ilar to that reported in Fig. 4a and was assigned to theDMPO-OOH adduct (hyperfine splittings aN = 12.7 G,aH

b ¼ 10:3 G, and aHc ¼ 1:5 G) formed by TMPyP photo-

sensitization. A similar spectrum to that reported in Fig.4c was observed by Buettner and Oberley (1979) during aprotoporphyrin illumination. These authors identified thesignal as the superoxide spin adduct of DMPO. Theseobservations showed that the irradiation of TMPyP resultsin the generation of superoxide radicals. Fig. 4d is the sameas in Fig. 4c, but the scan was started �2 min after the scanof Fig. 4c. Superoxide spin adduct was still detectable,however new EPR lines could be observed. After 20 minof irradiation (Fig. 4e), we no longer see the superoxidespin adduct.

3.1.6. Characterization of �OH production

The EPR spectra of oxygen air-saturated aqueous solu-tions at pH 7.0 containing the photosensitizer TMPyP(73 lM) and the spin-trap DMPO (40 mM) were recordedas function of illumination time (from 1 to 15 min). Theillumination of TMPyP leads to the formation of two dif-ferent radical species (Fig. 5). Initially the characteristicfour-line signal of the hydroxyl radical adduct of DMPO(DMPO-OH) was observed. The EPR spectra are charac-terized by hyperfine coupling constants of aN = aH =14.6 G, showing that the radical species is the DMPO-OH adduct (Finkelstein et al.,1980a,b). The second trappedradical is characterized with six-line signal; from the split-

ting constants this radical could be identified as a trappedcarbon-centered alkyl (DMPO-R; hyperfine splittingsaN = 15.92 G, aH = 23.13 G). According to the literature,the DMPO-R adduct could be the DMPO-CH3 adduct(Buettner, 1987). Samples kept in the dark showed noEPR-detectable radical signal. When the free radicalsbecame detectable, the amount of the two DMPO adductsincreased as function of the TMPyP concentration (datanot shown), but decreased as function of the illuminationtime (Fig. 5), which might be ascribed to the photodegrada-tion of DMPO.

3.1.7. Sources of the DMPO-OH signal

What is the origin of the DMPO-OH? The detection ofDMPO-OH does not necessarily mean that the photosensi-tization process involves hydroxyl radicals directly (Type Ireactions); apart from free �OH trapped by DMPO, EPRsignal of DMPO-OH may be formed as a result ofDMPO-OOH decomposition as shown in Scheme 1 (Buett-ner, 1993; Finkelstein et al., 1979).

When various �OH scavengers (Mannitol or DMSO,100 mM) were added to samples containing the photosen-sitizer TMPyP and the spin-trap DMPO, the DMPO-OHsignal was either diminished or completely abolished(Fig. 6). The production of �OH in our experiment modelwas also verified by adding ethanol (100 mM) to samplesbefore irradiation. The addition of ethanol should bothinhibit the production of DMPO-OH and result in theappearance of the a-hydroxyethyl radical as shown inScheme 2 (Finkelstein et al., 1980a).

Indeed, we observed that in presence of ethanol, theDMPO-OH signal was decreased (Fig. 6). A concomitantincrease of the signal of the DMPO-CH3 adduct was alsoobserved, which could be probably due to the superposi-tion of the EPR signal of DMPO-hydroxyethyl with thatof DMPO-CH3 since the hyperfine constants coupling forthe two adducts were found to be the same (aN =15.92 G, aH = 23.13 G).

In order to determine the origin of �OH production,SOD (300 U mL�1) and/or CAT (300 U mL�1) were addedto samples before irradiation. Furthermore, since �OH

Fig. 5. EPR spectra of spin-trapped radicals generated during illumination (0.285 W cm�2) of an air-saturated aqueous solution consisting of TMPyP(73 lM) and DMPO (40 mM). The symbols indicate line components belonging to DMPO-OH (h), DMPO-CH3 (�), and degraded DMPO (}).

DMPO

O2•– DMPO-OOH

•OH DMPO-OH

H+

e- e- H+

Scheme 1.

•OH DMPO-OH DMPO

H

•

CH3CH2OH

DMPO DMPO-hydroxyethyl

H3C-C-OH

Scheme 2.


could arise from the decomposition of H2O2 with iron,DFO a well-known iron chelator was also tested. TheEPR signal of DMPO-OH was decreased by addition ofSOD and/or CAT and was completely inhibited by DFO(Fig. 6). The decreases of the magnitude of the DMPO-OH signal by SOD and CAT were specific of theseenzymes, since addition of bovine serum albumin withequivalent protein content did not exhibit a significantdecrease in the DMPO-OH intensity (data not shown).

These results testified a primary �OH generation proba-bly from O��2 . We suggested that the TMPyP/O2 systemwould be able to generate �OH via the Fenton reaction

1 2 3 4

Control

100 mM

DMSO

100 mM

Ethanol

100 mM

Mannitol

0

20

40

60

80

100

120

1 2 3 4

Control

100 mM

DMSO

100 mM

Ethanol

100 mM

Mannitol

DM

PO

-OH

EP

R s

igna

l int

ensi

ty (

%)

Fig. 6. EPR signal intensity of DMPO-OH generated during TMPyP illuminatiand 114 lM TMPyP with added 100 mM mannitol, 100 mM ethanol, 100 mM Dthe samples were irradiated at light intensity of 0.285 W cm�2 for 3 min. The vdeviation.

and/or via Haber-Weiss reaction(8)–(10). H2O2 would orig-inate from O��2 as in reaction (11) or (12)

O��2 þ Fe3þ ! O2 þ Fe2þ ð8Þ

Fe2þ þH2O2 ! Fe3þ þ �OHþ �OH Fenton reaction ð9Þ

O��2 þH2O2 !metal=catalystO2 þ �OHþ �OH

Haber-Weiss reaction ð10Þ

O��2 !Hþ

HO�2!e�

HO�2 !Hþ

H2O2 ð11ÞO��2 þO��2 þ 2Hþ ! H2O2 þO2 ð12Þ

5 6 7 8300 U mL-1

SOD

300 U mL-1

CAT

50 mM

DFOSOD + CAT

5 6 7 8300 U mL-1

SOD

300 U mL-1

CAT

50 mM

DFOSOD + CAT

on in air-saturated aqueous solution. The latter contained 120 mM DMPOMSO, 300 U mL�1 SOD and/or 300 U mL�1 CAT, and 50 mM DFO. All

alues are the average of three experiments, and the bars indicate standard

Table 2Photoinactivation rate constants (k ± standard deviation) for E. hirae andEscherichia coli in the presence of different antioxidants and TMPyP at0.73 lM; dark controls show no toxic effect

Antioxidants Concentration �kE. hirae

(min�1)�kE. coli

(min�1)

None – 0.258 ± 0.011 0.100 ± 0.006Sodium azide (M) 10�2 0.089 ± 0.025 0.022 ± 0.002

10�1 0.001 ± 0.0005 0.003 ± 0.001Thiourea (M) 10�1 0.041 ± 0.004 0.005 ± 0.001DMSO (M) 10�1 0.254 ± 0.044 0.042 ± 0.005

1 0.098 ± 0.024 0.015 ± 0.002SOD (U mL�1) 55 0.258 ± 0.031 0.064 ± 0.003CAT (U mL�1) 220 0.256 ± 0.025 0.076 ± 0.003SOD + CAT

(U mL�1)55 + 220 0.247 ± 0.035 0.062 ± 0.004


It has been reported in the literature that �OH could alsoarise from the reaction of 1O2 with the spin-trap DMPO(Nishizawa et al., 2004). In order to verify if this reactioncould be involved in �OH production, we measured theEPR signal intensity of DMPO-OH generated during illu-mination of TMPyP in presence of D2O. In these condi-tions, no effect on the DMPO-OH signal intensity wasobserved, implying that �OH is not derived from singletoxygen.

3.2. Photosensitized inactivation of bacteria cells

In the presence of TMPyP, the overall survival curves ofboth E. hirae and E. coli indicated that linear and non-lin-ear (Weibull) models were appropriate in describing thedata. The obtained results showed that the non-linearmodel was appreciably better in describing the survivalcurves particularly in the first minutes of irradiation (datanot shown). However, the differences in the square of thecoefficient of correlation (R2) between both models are neg-ligible. We thus considered that simple linear regressionwas sufficient to modelize the MPN time course, becauseone does not obtain more information than with the regres-sion line (13):

log10ðMPNðtÞÞ ¼ �kt þ log10ðMPNð0ÞÞ ð13Þwhere k is the photoinactivation rate constant (min�1),MPN(0) and MPN(t) are the bacterial concentrations ex-pressed in log units before and after exposure to simulatedsunlight for a given time.

In order to avoid any ambiguity, the term photoinacti-vation of bacteria consists in the loss of bacterial cultivabil-ity as the result of the combined action of exposure to lightand porphyrin.

3.2.1. Inactivation of Gram-positive bacteria

In order to determine the contribution of different reac-tive oxygen species in the TMPyP-induced photosensitiza-tion process, we conducted a set of experiments usingseveral scavengers of superoxide anion ðO��2 Þ, hydrogenperoxide (H2O2), hydroxyl radical (�OH) (Type I reactions),and quenchers of singlet oxygen (1O2) (Type II reactions).

Results with E. hirae (Table 2) show in the presence ofsodium azide (NaN3), a quencher of singlet oxygen, therate of inactivation being slowed down starting from a con-centration of 10�2 M. This is due to the diminution(10�2 M NaN3) or elimination (10�1 M NaN3) of singletoxygen and hydroxyl radicals in the medium. In the pres-ence of thiourea (10�1 M), the rate of photoinactivationof E. hirae is less decreased compared to that with sodiumazide, probably because of the weaker bimolecular rateconstants for reaction of thiourea compared to sodiumazide with singlet oxygen (8 � 105 and 5 � 108 M�1 s�1,respectively) and the hydroxyl radical (5 � 109 and7.5 � 109 M�1 s�1, respectively). In the presence of dimeth-ylsulphoxide (10�1 M), which reacts more slowly with sin-glet oxygen (�103 M�1 s�1) but rapidly with the hydroxyl

radical (7 � 109 M�1 s�1), the photosensitization of E. hirae

is comparable with that obtained without DMSO. Forhigher concentrations (1 M), the photoinactivation rateconstant was lower.

At equivalent concentrations (10�1 M), sodium azideprotects E. hirae from reactive oxygen species generatedin the system. Thiourea provides less protection andDMSO has a relatively insignificant effect. It can berecalled that the bimolecular rate constants for reactionof these reagents are very similar for the hydroxyl radicaland very different for singlet oxygen (sodium azide� thio-urea > DMSO). These results confirm the importance ofsinglet oxygen in the photoinactivation process, i.e. the pre-dominance of Type II reactions. It has been suggested thatthe cytoplasmic membrane is the site of lethal damage tobacteria via photosensitization and singlet oxygen (Nitzanet al., 1992). Singlet oxygen probably diffuses readilythrough the relatively open structure of the peptidoglycanlayer of the cell wall to react with the vital target (Dahlet al., 1988). The difference measured between the photoin-activation rate constants with and without DMSO can beconsidered as representative of the hydroxyl radical effect.Under our conditions, this action appears to be limited.

The presence of superoxide dismutase at 55 U mL�1

and/or catalase at 220 U mL�1 did not modify the rate ofE. hirae photoinactivation compared with the absence ofthese two enzymes, demonstrating that superoxide anionand hydrogen peroxide have a negligible effect on the reac-tion. For E. hirae photoinactivation, Type I reactionsappear to play a minor role compared with Type IIreactions.

3.2.2. Inactivation of Gram-negative bacteria

With E. coli the results obtained with each of the antiox-idants (Table 2) show that, considering the photoinactiva-tion rate constants, the presence of sodium azideconsiderably lowers the photoinactivation rate constantat a concentration of 10�2 M. In the presence of sodiumazide or thiourea at 10�1 M, there was no significant bacte-ria inactivation observed. On the other hand, the photoin-activation rate constants of E. coli were reduced in thepresence of the DMSO (1 or 10�1 M). These results show


that singlet oxygen seems to play a significant role in thephotoinactivation process of E. coli, but to a lesser extentthan for E. hirae. This is in agreement with data in the lit-erature showing that Gram-positive bacteria are more sen-sitive to photosensitization and singlet oxygen than Gram-negative bacteria (Dahl et al., 1989; Valduga et al., 1993).However, the contribution of Type I reactions are not neg-ligible as it is for E. hirae. Indeed, the measured variationbetween the photoinactivation rates with and withoutDMSO, representative of the hydroxyl radical effect, wassignificant. Moreover, the results obtained with superoxidedismutase at 55 U mL�1 and/or catalase at 220 U mL�1

show the respective contributions of superoxide anionand hydrogen peroxide (Type I reactions) in the mecha-nisms of E. coli photoinactivation (Table 2).

The Gram-negative bacterial cell wall lipopolysaccha-ride coat (LPS) may offer some protection from the toxiceffects of exogenous agents. The LPS has previously beenshown to present a physical or chemical barrier throughwhich singlet oxygen generated outside of cells must passto interact with a vital target, such as membrane or cyto-plasmic components (Dahl et al., 1989). Most Gram-posi-tive bacteria lack a protective structure analogous to theGram-negative LPS and the outer membrane in which itis anchored. It was mentioned previously that reaction ofsinglet oxygen reaching the outer membrane componentscould lead to generation of reactive secondary productsaccording to a cascade of secondary processes, such as per-oxy radicals, which may in turn be able to cause lethaldamage to the vital target (Dahl et al., 1989). Tanielianet al. (2000) postulated that in both Type I and Type IIreactions, the initial products are often peroxides that canbreak down and induce free radical reactions. The totaltoxicity, then, would be the sum of the singlet oxygenreaching the inner membrane and the lethal effects of sec-ondary reaction products from the outer membrane. Therate of Gram-positive killing should depend then only ondirect singlet oxygen–vital target interactions, withoutany need to invoke secondary reaction mechanisms suchas in the Gram-negative outer membrane. According todata in the literature, the two mechanisms can occur simul-taneously. The relative predominance between the twodepends mainly on the sensitizer, substrate, and the natureof the medium (Ochsner, 1997).

4. Conclusions

The results of our EPR experiments described aboveshowed that TMPyP, which is well known to exhibit poten-tialities as photosensitizer for the photoinactivation of awide range of microorganisms, would be able to operateby Type I (radical mediated) and Type II (singlet oxygen-mediated) mechanisms. Irradiation of Gram-positive orGram-negative bacteria in simultaneous presence ofTMPyP and scavengers/quenchers of active oxygen speciessuggested the implication of singlet oxygen (Type II reac-tions), one of the key reactive species generated by the pho-

tosensitization process. The participation of other activeoxygen species such as superoxide anion, hydroxyl radicaland hydrogen peroxide could not be neglected (Type I reac-tions) since they play a significant role in photoinactivationmechanisms, particularly for Gram-negative bacteria.

Acknowledgements

The authors wish to thank Dr. Olivier Briand (FrenchAgency for Environmental and Occupational HealthSafety, France) and Dr. Odile Sergent (University ofRennes 1, France) for their contribution to this work.

References

Alouini, Z., Jemli, M., 2001. Destruction of helminth eggs by photosen-sitized porphyrin. J. Environ. Monit. 3, 548–551.

Ando, T., Yoshikawa, T., Tanigawa, T., Kohno, M., Yoshida, N., Kondo,M., 1997. Quantification of singlet oxygen from hematoporphyrinderivative by electron spin resonance. Life Science 61, 1953–1959.

Bertoloni, G., Rossi, F., Valduga, G., Jori, G., van Lier, J., 1990.Photosensitizing activity of water- and lipid-soluble phthalocyanineson Escherichia coli. FEMS Microbiol. Lett. 71, 149–156.

Buettner, G.R., 1987. Spin trapping: ESR parameters of spin adducts.Free Rad. Biol. Med. 3, 259–303.

Buettner, G.R., 1993. The spin trapping of superoxide and hydroxyl freeradicals with DMPO (5,5-dimethylpyrroline-N-oxide): more aboutiron. Free Rad. Res. Commun. 19, S79–S87.

Buettner, G.R., Oberley, L.W., 1978. Considerations in the spin trappingof superoxide and hydroxyl radical in aqueous system using 5,5-dimethyl-1-pyrroline-1-oxide. Biochem. Biophys. Res. Commun. 83,69–74.

Buettner, R., Oberley, L.W., 1979. Superoxide formation by protopor-phyrin as seen by spin trapping. FEBS Lett. 98, 18–20.

Cooper, A.T., Goswami, D.Y., 1998. A survey of solar based drinkingwater treatment. solar engineering 1998. In: Proceedings of the ASMEInternational Solar Energy Conference, Albuquerque, New Mexico,pp. 265–275.

Dahl, T.A., Midden, W.R., Neckers, D.C., 1988. Comparison ofphotodynamic action by Rose Bengal in gram-positive and gram-negative bacteria. Photochem. Photobiol. 48, 607–612.

Dahl, T.A., Midden, W.R., Hartman, P.E., 1989. Comparison of killing ofGram-negative and Gram-positive bacteria by pure singlet oxygen. J.Bacteriol. 171, 2188–2194.

Dorfman, L.M., Adams, G.E., 1973. Reactivity of the hydroxyl radical inaqueous solutions. In: National Standard Reference Data Service,National Bureau of Standards, C13.48, p. 46.

Finkelstein, E., Rosen, G.M., Rauckman, E.J., 1979. Spin trapping ofsuperoxide. Mol. Pharmacol. 16, 676–685.

Finkelstein, E., Rosen, G.M., Rauckman, E.J., 1980a. Spin trapping ofsuperoxide and hydroxyl radical: practical aspects. Arch. Biochem.Biophys. 200, 1–16.

Finkelstein, E., Rosen, G.M., Rauckman, E.J., 1980b. Spin trapping.Kinetics of the reaction of superoxide and hydroxyl radicals withnitrones. J. Am. Chem. Soc. 102, 4994–4999.

Floyd, R.A., Lewis, C.A., Wong, P.K., 1984. High-pressure liquidchromatography-electrochemical detection of oxygen free radicals.In: Packer, L. (Ed.), . In: Methods in Enzymology, vol. 105. AcademicPress, New York, pp. 231–237.

Foote, S., 1991. Definition of type I and type II photosensitized oxidation.Photochem. Photobiol. 54, 659.

Gourmelon, M.J., Cillard, J., Pommepuy, M., 1994. Visible light damageto Escherichia coli in seawater: oxidative stress hypothesis. J. Appl.Bacteriol. 77, 105–112.


Hadjur, C., Wagnieres, G., Ihringer, F., Monnier, Ph., van den Bergh, H.,1997. Production of the free radicals O��2 and �OH by irradiation of thephotosensitizer zinc(II) phthalocynanine. J. Photochem. Photobiol., B:Biol. 38, 196–202.

Henderson, W., Dougherty, T.J., 1992. How does photodynamic therapywork? Photochem. Photobiol. 55, 145–157.

Jimenez-Hernandez, M.E., Manjon, F., Garcıa-Fresnadillo, D., Orellana,G., 2006. Solar water disinfection by singlet oxygen photogeneratedwith polymer-supported Ru(II) sensitizers. Solar Energy 80, 1382–1387.

Jori, G., Brown, S.B., 2004. Photosensitized inactivation of micro-organisms. Photochem. Photobiol. Sci. 3, 403–405.

Khan, U., Kasha, M., 1994. Singlet molecular oxygen in the Haber-Weissreaction. Proc. Natl. Acad. Sci. USA 91, 12365–12367.

Lambrechts, S.A.G., Aalders, M.C.G., van Marle, J., 2005. A mechanisticstudy of the photodynamic inactivation of Candida Albicans by acationic porphyrin. Antimicrob. Agents Chemother. 49, 2026–2034.

Lion, Y., Delmelle, M., van de Vorst, A., 1976. New method of detectingsinglet oxygen production. Nature 263, 442–443.

Merchat, M., Bertoloni, G., Giacomini, P., Villanueva, A., Jori, G., 1996a.Meso-substituted cationic porphyrins as efficient photosensitizers ofGram-positive and Gram-negative bacteria. J. Photochem. Photobiol.,B: Biol. 32, 153–157.

Merchat, M., Spikes, J.D., Bertoloni, G., Jori, G., 1996b. Studies on themechanism of bacteria photosensitization by meso-substituted cationicporphyrins. J. Photochem. Photobiol., B: Biol. 35, 149–157.

Minnock, A., Vernon, D.I., Schofield, J., Griffiths, J., Parish, J.H., Brown,S.B., 1996. Photoinactivation of bacteria. Use of a cationic water-soluble zinc phthalocyanine to photoinactivate both Gram-negativeand Gram-positive bacteria. J. Photochem. Photobiol., B: Biol. 32,159–164.

Minnock, A., Vernon, D.I., Schofield, J., Griffiths, J., Parish, J.H., Brown,S.B., 2000. Mechanism of uptake of a cationic water-soluble pyridi-nium zinc phthalocyanine across the outer membrane of Escherichia

coli. Antimicrob. Agents Chemother. 44, 522–527.Moan, J., Peng, Q., 2003. An outline of the hundred-Year History of

PDT. Anticancer Res. 23, 3591–3600.Moan, J., Wold, E., 1979. Detection of singlet oxygen by ESR. Nature

279, 450–451.

Neter, J., Kutner, M.H., Nachtshheim, C.J., Wasserman, W., 1996.Applied Linear Regression Models. McGraw-Hill Co. Inc., Chicago,IL.

Nitzan, Y., Gutterman, M., Malik, Z., Ehrenberg, B., 1992. Inactivationof Gram-negative bacteria by photosensitized porphyrins. Photochem.Photobiol. 55, 89–96.

Nishizawa, C., Takeshita, K., Ueda, J., Mizuno, M., Suzuki, K.T., Ozawa,T., 2004. Hydroxyl radical generation caused by the reaction of singletoxygen with a spin trap, DMPO, increases significantly in the presenceof biological reductants. Free Rad. Res. 38, 385–392.

Ochsner, M., 1997. Photophysical and photobiological processes in thephotodynamic therapy of tumours. J. Photochem. Photobiol., B: Biol.39, 1–18.

Peleg, M., 2003. Microbial survival curves: interpretation, mathematicalmodelling and utilization. Commun. Theor. Biol. 8, 357–387.

Peleg, M., Cole, M.B., 1998. Reinterpretation of microbial survival curves.Crit. Rev. Food Sci. 38, 353–380.

Ranby, B., Rabek, J.F., 1978. Singlet Oxygen Reactions with OrganicCompounds and Polymers. John Wiley, New York.

Reddi, E., Ceccon, M., Valduga, G., Jori, G., Bommer, J.C., Elisei, F.,Latterini, F., Mazzucato, U., 2002. Photophysical properties andantibacterial activity of meso-substituted cationic porphyrins. Photo-chem. Photobiol. 75, 462–470.

Salmon-Divon, M., Nitzan, Y., Malik, Z., 2004. Mechanistic aspects ofEscherichia coli photodynamic inactivation by cationic tetra-meso(N-methylpyridyl) porphine. Photochem. Photobiol. Sci. 3, 423–429.

Tanielian, C., Mechin, R., Seghrouchni, R., Schweitzer, C., 2000.Mechanistic and kinetic aspects of photosensitization in the presenceof oxygen. Photochem. Photobiol. 71, 12–19.

Valduga, G., Bertoloni, G., Reddi, E., Jori, G., 1993. Effect of extracel-lularly generated singlet oxygen on Gram-positive and Gram-negativebacteria. J. Photochem. Photobiol., B: Biol. 21, 81–86.

Viola, A., Hadjur, C., Jeunet, A., Julliard, M., 1996. Electron paramag-netic resonance evidence of the generation of superoxide ðO��2 Þ andhydroxyl (�OH) radicals by irradiation of a new photodynamic therapyphotosensitizer, Victoria Blue BO. J. Photochem. Photobiol., B: Biol.32, 49–58.

Wasserman, H.H., Murray, R.W., 1978. Singlet Oxygen. Academic Press,New York.

Documents

1-s2.0-S0038092X0800128X-main