Photoinactivation of Escherichia coli (SURE2) without intracellular uptake of the photosensitizer



Annegret Preuß, Institut für Physik, Humboldt – Universität zu Berlin, Newtonstraße 15, 12489 Berlin, Germany.



This study was performed to investigate the possibility to photodynamically inactivate Gram-negative bacteria without intracellular uptake of the photosensitizer. The efficiency of the photodynamic growth inhibition of Escherichia coli (SURE2) was proved in a comparative study of a neutral and a cationic photosensitizer.

Methods and Results

We used confocal laser scanning microscopy (CLSM) to investigate the uptake of the photosensitizer by the bacteria to show that both chlorin e6 and TMPyP are not accumulated in the cells. Fluorescence lifetime imaging (FLIM) and phototoxicity experiments were used to investigate the photodynamic inactivation of the Gram-negative bacteria. The phototoxicity experiments were carried out using a white light LED-setup to irradiate the bacterial suspensions. The viability of the bacteria was obtained by 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT)-assay. For the cationic TMPyP, photodynamic inactivation without intracellular uptake was observed, whereas for chlorin e6 such behaviour was not found.


It was proven that in general, it is possible to photodynamically inactivate Gram-negative bacteria without photosensitizer accumulation in the bacterial cells. This fact is especially interesting, considering that the development of resistances may be prevented, leaving the active components outside the bacterium.

Significance and Impact of the Study

In a world with bacteria that gain the ability to withstand the effects of antibiotics and are able to transmit on these resistances to the next generation, it becomes necessary to develop new approaches to inhibit the growth of multi-resistant bacteria. The photodynamic inactivation of bacteria is based on a three-component system by which the growth of the bacterial cells is inhibited. The well-directed oxidative damage is initiated by visible light of a certain wavelength, which excites a nontoxic photoactive molecule, called photosensitizer. Its reaction with oxygen causes the generation of cytotoxic species like singlet oxygen, which is highly reactive and causes the inactivation of the growth of bacteria.


At the end of the 1920s when Alexander Fleming discovered penicillin (Fleming 1929), which is nowadays known as an antibiotic, a new age of medicine was initiated. Since then plenty of antibiotics were developed with which a huge number of diseases may be treated and cured. Due to the widespread utilization of antibiotics in our world and the high adaptability of bacteria, the resistances against these antibiotics spread faster than new antibiotics are developed and produced. Nowadays, multi-resistant pathogenic microbes like methicillin-resistant Staphylococcus aureus (MRSA) display a threat to the health of many patients at hospitals worldwide and pass on their resistances to the next generations. The age of antibiotics, as dominant agents against harmful bacteria, has come to an end (Rice 2008; Arias and Murray 2009), and hence, new strategies for the control of diseases are indispensable.

The photodynamic inactivation of bacteria (PIB) is very promising as an alternative way of antimicrobial therapy and is attracting the interest of many research groups worldwide (Demidova and Hamblin 2004; Hamblin and Hasan 2004; Jori et al. 2006; Maisch 2007; Maisch et al. 2007, 2011; Carvalho et al. 2009; Dai et al. 2009). The photodynamic treatment is already known and investigated since the early 20th century and was used successfully in different fields of medicine, like oncology, ophthalmology and dermatology (Moan and Peng 2003; Calzavara-Pinton et al. 2005).

Photosensitizer (PS) molecules are nontoxic and photoactive molecules, which can be excited by light of a certain wavelength. Usually, they are nonreactive, but after the absorption of light, they can interact with other molecules in their direct vicinity. They perform interactions with molecular oxygen and by transferring energy to it, and singlet oxygen (inline image) is generated. inline image is highly reactive and is known to induce cytostatic or even cytotoxic reactions directly or indirectly. The generation of inline image is a reversible process, which means, that after the release of the absorbed energy, the PS molecule can excite again a new oxygen molecule acting as a catalyst for the photosensitization. The production of the highly reactive inline image is possible until the PS molecule is destroyed by oxidative processes (‘bleached out’).

It is well known that it is more difficult to inactivate the growth of Gram-negative bacteria than Gram-positive strains. This behaviour is explained by the specific differences in the structure of the cell walls of bacterial strains. The cell wall of Gram-negative bacteria has a far more nuanced structure than the one of Gram-positive bacteria (Scheffers and Pinho 2005)]. The lipopolysaccharides on the outer membrane of Gram-negative bacteria allow only hydrophilic dissolved particles to pass through it. Due to the differences in the wall structure, it is necessary to adjust the choice of the PS molecule to the micro-organism to be inactivated. By this reason, natural porphyrins and their anionic derivatives are not suitable for experiments with Gram-negative bacteria, because they do not bind properly to the outer membrane (Hamblin and Hasan 2004).

For the inactivation of Gram-negative bacteria, it is more useful to use PS molecules, which are cationic (Malik et al. 1992; Salmon-Divon et al. 2004; Almeida et al. 2011), like porphyrins or phthalocyanines that carry a positive charge. Another approach describes that the former treatment of the bacterial cell wall with polymyxin B nonapeptide (PMBN) or ethylenediaminetetraacetic acid (EDTA) results in the loss of approximately 50% of the lipopolysaccharides of the outer membrane and by this facilitating the penetration of the PS molecules through the cell wall, which gets more permeable even for hydrophobic compounds. Bacteria, which were treated by one of these membrane disruptor substances, can even be photosensitized by phthalocyanines, especially zinc-phthalocyanines and natural porphyrins (Bertolini et al. 1990; Malik et al. 1992).

Former research demanded that the PS molecules have to penetrate the cell wall of bacteria to have a certain toxic effect (Schastak et al. 2010). Most recent research showed that it is not necessarily required that the PS molecules pass through the outer membrane of the Gram-negative bacteria, since the diffusion length of inline image in pure water is approximately 1 μm (Maisch et al. 2011), which is about in accordance with the half length of an Escherichia coli (E. coli) bacterium. It gives measurable effects if the PS only acts in the vicinity of the bacteria (Hamblin and Hasan 2004).

Our experiments show that it is possible to photodynamically inactivate Gram-negative bacteria without intracellular uptake of the PS molecules. This is especially interesting because micro-organisms are able to develop methods to direct inhibiting components out of the cell and by this to become resistant to these components. Previous research showed that bacteria do not have yet strategies to develop resistances to photodynamic approach (Hamblin and Hasan 2004; Tavares et al. 2010), and furthermore, the development of resistances can be prevented by leaving the active components outside of the bacterium.

Materials and methods

Photosentizers: TMPyP and Clorine e6

The 5,10,15,20-tetrakis(1-methylpyridinium-4-yl)porphyrin tetra-iodide (TMPyP) used in this work was prepared in two steps according to the procedure described in the literature (Tomé et al. 2004). In the first step, the neutral porphyrin was synthesized by condensation of pyrrole and 4-pyridinecarboxaldehyde at reflux in acetic acid and nitrobenzene. The resulting porphyrin was purified from the crude mixture by column chromatography (silica), and the pyridyl groups were quaternized by reacting with methyl iodide. TMPyP was purified by crystallization from chloroform/methanol/petroleum ether and its purity was confirmed by thin layer chromatography and by 1H NMR spectroscopy. The inline image-quantum yield of TMPyP in water is Φ∆ = 0·74 (Wilkinson et al. 1993).

The chlorin e6 (Ce6) was supplied by APOCARE Pharma GmbH, (Bielefeld, Germany) and used without further purification. Stock solutions of 1 mmol l−1 of each PS were prepared, in dimethyl sulfoxide (DMSO) for TMPyP and in phosphate buffered saline (PBS) for Ce6 and then stored at 4°C. The inline image-quantum yield of Ce6 in water is Φ∆ = 0·65. Molecular structures of both photosensitizers are given in Fig. 1.

Figure 1.

Chemical structure of the TMPyP and Ce6.

E. coli SURE2 culture and bacterial growth conditions

For the experiments, a bacterial SURE2 strain of E. coli was used (genotype: endA1 glnV44 thi-1 gyrA96 relA1 lac recB recJ sbcC umuC::Tn5 uvrC e14-∆ (mcrCB-hsdSMR-mrr)171 F'[proAB+ lacIqlacZ∆ M15 Tn10 AmyCmR] – obtained from Stratagene), which is resistant to four different antibiotics: nalidixic acid, kanamycin, tetracycline and partly to chloramphenicol (resistant for <40 μg ml−1, sensitive for >100 μg ml−1).

The bacteria were grown at room temperature in lysogeny broth (Luria/Miller), which contained additional 50 μg ml−1 of kanamycin to prevent the growth of any different bacteria than the SURE2 strain. The suspension was shaken (350 rev min−1) overnight and then diluted to an OD600 between 0·05 and 0·35. This was carried out to ensure that the bacteria would not reach the stationary phase during the first five hours of an experiment.

Bacterial cell viability

The viability of the bacteria during their growth was measured with the 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT)-assay. This method was suggested by Wang et al. (2010). This assay gives either the opportunity to analyse a high amount of samples at the same time or to measure only the cells, which are still living.

Adapting the MTT-assay to our goal, the following procedure was established to evaluate the dark toxicity and the phototoxicity: The viability tests are performed on 96-well plates. At the beginning of an experiment, two 96-well plates for each incubation time (i.e. 1, 3, 5, and 24 h) were prepared. For each experimental time, one of the plates was covered with aluminium foil (dark reference) and the other was irradiated during all the incubation time. For the first incubation time (0 h), just one plate was prepared because no illumination was required. For each experiment, different samples were prepared: only with medium as a blank value; containing only bacteria as a reference; and containing bacteria with different concentrations of PS (3 and 5 μmol l−1 of TMPyP or Ce6). The OD600 of the bacterial suspension at the beginning of the experiment is below 0·1. The bacterial cell concentration should not be higher to avoid any starving of the cells at an early stage of the experiment.

Per sample, at least eight wells were analysed, each one containing 100 μl of bacterial suspension. At each incubation time, 10 μl of MTT-reagent (5 mg ml−1) were added to each well in both irradiated and nonirradiated 96-well plates. Then the 96-well plates were incubated for 30 min at room temperature in the dark. During that time, the active dehydrogenases in the living cells reduce the yellow MTT-dye to purple formazan crystals. These crystals were dissolved in DMSO (200 μl per well), and the quantification of the formazan was determined by measuring the absorbance (λ = 535 nm) directly on the 96-well plates (with Platereader Victor3; Perkin Elmer, Waltham, MA).

To enhance the permeability of the bacterial cell walls for the photosensitizers, the membrane disruptor PMBN was used. PMBN is a modified antibiotic, which binds selectively to the lipopolysaccharides of Gram-negative bacteria, such as E. coli, but instead of killing the bacterial cells PMBN only increases the permeability of the bacterial cell wall. PMBN disrupts the membrane and enables the diffusion of the PS into the cell. To check the toxicity of PMBN, the bacteria were incubated in the presence of TMPyP at the concentration of 3 μmol l−1 either with or without 10 μg ml−1 PMBN. The viability of the bacterial cells was measured using MTT-assay directly after the addition of PMBN and TMPyP (0 h) and after 4 h of incubation.

Irradiation conditions

To irradiate the bacteria four irradiation setups with white light LEDs (Osram, Munich, Germany) were used. Each setup contains eight LEDs with a colour temperature of 3500 K and eight LEDs with a colour temperature of 6500 K alternating in a square arrangement. The LEDs illuminate an opal glass plate from below, resulting in scattering of the light guaranteeing a uniform irradiation of the cell samples in the well plates at a fluence rate of 9 mW cm−2. Each 96-well plate was placed on one illumination setup.

CLSM and FLIM intracellular uptake

The association of the PSs with the E. coli cells was analysed by confocal laser scanning microscopy (CLSM). The fluorescence lifetime of the PSs depends on their microenvironment. Thus, fluorescence lifetime imaging (FLIM) was used to obtain the differences between PSs associated with the E. coli cells on the outside of their cell wall, and PSs taken up into the E. coli cells.

For incubation, 1 ml of E. coli cell suspension for each sample was placed on a 24-well plate. The E. coli suspensions were incubated for 4 h in the presence of 5 μmol l−1 of each PS in the dark at room temperature. To increase the membrane permeability and enhance the intracellular uptake of the PS, E. coli cells were incubated with and without PMBN (10 μg ml−1) in comparison.

The cell images were observed with a CLSM (FluoViewTM FV1000, Olympus, Hamburg, Germany) with FLIM-extension (Pico Quant, Berlin, Germany). The E. coli cell suspensions were excited with a laser diode at λex = 405 nm for imaging and excited with a nanosecond pulse laser at λex = 440 nm for fluorescence lifetime measurements.


Monitoring the intracellular uptake of TMPyP and Ce6 in E. coli SURE2 via CLSM and FLIM

To perform these studies, two tetrapyrrolic molecules have been used as PS: the synthetic tetra-cationic derivative TMPyP and the neutral natural derivative known as chlorin Ce6. The confocal laser scanning microscopy (CLSM) was used to investigate the PSs uptake by bacterial cells. In these experiments, the membrane disruptor PMBN was added to enhance the PS uptake (see Fig. 2). The results obtained show that without the addition of PMBN, no significant amount of the PSs was found inside of the cells.

Figure 2.

Confocal laser scanning microscopy (CLSM) images of Esherichia coli (SURE2) incubated for 24 h in the presence of photosensitizer (PS) [TMPyP or Ce6] at concentration of 5 μmol l−1, without polymyxin B nonapeptide (PMBN) (upper) and with PMBN (bottom). The left row shows the scattered light image, the row in the middle shows the green fluorescence of PS, the row on the right shows the merged images. The excitation wavelength is λex = 405 nm, detection of the fluorescence between 650 and 750 nm.

It is well known that during the uptake of the PS molecules by the cells – due to changes in the microenvironment – the fluorescence lifetime is changing. Therefore, it is highly convenient to choose fluorescence lifetime imaging to follow the intracellular uptake of the PS (see Fig. 3). After 3 h of incubation, the results show that, when the E. coli cells were incubated only with the PS, the fluorescence lifetime was approximately 4 ns in both cases (TMPyP and Ce6).

Figure 3.

Fluorescence lifetime imaging (FLIM) images of Escherichia coli (SURE2) incubated with photosensitizer (PS) [TMPyP or Ce6] at concentration of 5 μmol l−1, without PMBN (left) and with PMBN (right). The histogram (left) shows the normalized frequency of fluorescence lifetimes in the images. Excitation wavelength with a nanosecond pulse laser was λex = 440 nm.

Two major differences can be observed in the fluorescence behaviour after concomitant incubation of bacterial cells with PMBN and PS (TMPyP or Ce6). First, an enhanced intensity of the fluorescence in the CLSM can be noted (see Fig. 2). Second, fluorescence lifetimes of both PSs increase. Nevertheless, also the fluorescence lifetime of 4 ns was found (see shoulder of histogram), as observed for incubation without PMBN. Interpreting the 4 ns lifetime as associated to extracellular PSs and the longer lifetimes to intracellular PSs, these images (Figs 2 and 3) reveal that after the incubation with PMBN both TMPyP and Ce6 are taken up by the bacteria and only a small amount of molecules remains outside, associated with the cell wall.

The weak PS fluorescence, which was detected in bacterial suspensions not treated with PMBN, derives solely from PS molecules that were located outside of the bacterial cells. So, microscopic analysis revealed that both PS cannot be taken up by the cell without the use of a permeation enhancer like PMBN.

In difference to our FLIM measurement, Ràgas et al. (2010) measured the fluorescence lifetimes of TMPyP in E. coli suspensions by TCSPC (time correlated single photon counting).

They obtained three different lifetimes in the bacterial suspension (2·0, 4·6 and 10·5 ns) and determined the 4·6 ns as originating from TMPyP associated to cell walls and the other two as originating from TMPyP associated to intracellular nucleic acids.

The affinity of TMPyP to nucleic acids is in fact very high, and the fluorescence intensity of TMPyP attached to nucleic acids is much stronger than the fluorescence of TMPyP attached to other biological structures. Thus, a small amount of bacterial cells with disrupted cell walls will affect a small amount of TMPyP molecules attached to nucleic acids, and a high amount of fluorescence originating from TMPyP attached to nucleic acids.

The advantage of using a confocal microscopy technique like FLIM for investigations of intramolecular localization of fluorophors is the possibility of spatially (x-, y-, z-direction) well-defined measurements inside the cells. For the interpretation of our measurements, only fluorescence lifetimes originating from TMPyP attached to bacterial cells with intact cell walls have been treated. This is not possible in a spectroscopic experiment in cell suspension. In such ensemble measurements, we will obtain an average picture; moreover, the preparation of the sample is even not very gentle (see Ràgas et al. 2010). This may result in bacterial cells with different stages of cell wall damages. We measured TCSPC in E. coli suspension as well. The only difference in our protocol compared with that of Ràgas et al. (2010) was the reduction of the centrifuge power used in the process of cell washing from 4000 to 800 g Ràgas et al. (2010). This value is far enough to get a solid cell pellet. After such gentle washing procedure, we could not verify any ‘intracellular’ TMPyP by TCSPC measurement. Otherwise, the very short lifetimes reported by Ràgas et al. (2010) and discussed as originating from an ‘intracellular’ side were detectable in LB-Medium (bacterial growth medium) without bacteria as well.

In our experiments, carried out without the addition of a permeation enhancer like PMBN, we could not observe fluorescence lifetimes which could be identified as originating from intracellular TMPyP. The cell wall of a Gram-negative bacterium like E. coli seems to be an efficient barrier for the photosensitizer TMPyP. Thus, in conclusion, we can state that TMPyP does not penetrate living bacterial cells with intact cell walls.

Toxicity of PMBN

A preliminary test of toxicity was performed to evaluate the applicability of PMBN as a membrane disruptor without any toxicity induced to the cells. The bacteria were incubated in the presence of TMPyP at the concentration of 3 μmol l−1 either with or without 10 μg ml−1 PMBN. The viability of the bacterial cells was measured using MTT-assay directly after the addition of PMBN and TMPyP (0 h) and after 4 h of incubation (See Fig. 4).

Figure 4.

Viability of Escherichia coli (SURE2) incubated with TMPyP at concentration of 5 μmol l−1 in the dark and polymyxin B nonapeptide. The viability was measured via MTT assay; the normalized formazan absorbance is shown. The error bars show the standard deviation of six measurements (n = 6). (■) Ref; (image) Ref + PMBN; (image) TMPyP; (image) TMPyP + PMBN.

The results show that at the beginning of the experiment, the MTT absorbance of the bacteria incubated with PMBN seems to be significantly higher compared with the samples incubated without it, even if the cell number at the beginning of the experiment should be the same. There are two possible explanations: either the cell proliferation of the bacteria treated with PMBN is higher or the permeability of the cell walls triggered by the PMBN effects a better accessibility of the dehydrogenases to the purple formazan in the MTT-assay.

After 4 h of incubation in the dark, it can be seen that PMBN at the tested concentration has a significant toxic effect on the cells. This behaviour can be explained as a result of the strongly increased permeability of the cell walls affected by the PMBN. Based on these results, the phototoxicity evaluation of the analysed PS was solely measured without the addition of PMBN.

Phototoxicity of TMPyP and Ce6

The cytotoxicity evaluation was carried out to analyse if the PSs show phototoxic effect on the bacteria after the confirmation that both, TMPyP and Ce6, remain outside the cell. The dark toxicity or phototoxicity evaluations were measured by MTT-assays on 96-well plates. The results for the whole time-scale (24 h) are shown in Fig. 5. As we observed, PSs, TMPyP and Ce6 exhibited no dark toxicity. At the same time, no significant variations in growth could be discovered between the differently treated bacteria in the dark. The reference contained additional 0·5% of DMSO (solvent of TMPyP), to ensure the same conditions of all samples, and we could prove that the solvent of the PS itself had no inhibiting effects to the growth of the bacteria in the used concentrations (data not shown).

Figure 5.

Viability of Escherichia coli (SURE2) treated with 0·5% dimethyl sulfoxide (Ref) 3 μmol l−1, or 5 μmol l−1 of TMPyP or Ce6 in darkness (left) or illuminated (right) with white light at fluence rate of 9 mW cm−2 during the intervals of 0, 1, 3, 5 and 24 h. The viability of the bacteria was measured by MTT assay; the normalized formazan absorbance is shown. The error bars show the standard deviation of 10 measurements (n = 10). (■) Ref; (image) TMPyP 1 μmol l−1; (image) TMPyP 3 μmol l−1 (image) TMPyP 5 μmol l−1; (image) Ce6 1 μmol l−1; (image) Ce6 3 μmol l−1; (image) Ce6 5 μmol l−1.

The effect of phototoxicity needs time to become especially evident, but the samples with TMPyP containing a concentration of 3 and 5 μmol l−1 in the suspension show a significantly lower growth than the others (ref and samples incubated with Ce6) after incubation time of 5 h, and this is far more evident after 24 h. In contrast to TMPyP, the neutral chlorin Ce6 has no phototoxic effect on E. coli.


The present study shows that bacteria can be efficiently inactivated without intracellular uptake of the PS. This means, it is not necessary to generate inline image inside the bacterial cells to cause severe cell damage once the inline image generation in the immediate proximity to the bacterial cell walls allows an efficient photoinactivation. Due to the short inline image lifetime, one can assume that inline image generated outside the bacteria affects cell wall structures and thus interferes with the growth of the bacteria. The inline image will not reach the intracellular space within the time frame of the short inline image lifetime. Thus, intracellular antioxidants and enzymes that are in position to protect the cells against oxidative stress cannot act against the inline image. The fact that the studied PSs showed a phototoxic effect from the extracellular side reduces the possibility of developing drug resistance.

Although Ce6 and TMPyP are able to generate inline image efficiently and both are not taken up by the bacterial cells without the use of an uptake enhancer such as PMBN, only TMPyP shows significant phototoxicity in E. coli cells. This result can be explained by the different molecular structure of the photosensitizers. The most relevant difference is that TMPyP contains four positive charges, whereas Ce6 is a neutral molecule resulting in different interactions with the bacterial cell wall structures. It is well known that cationic molecules in contrast to uncharged ones interact with the bacterial cell wall structures far better (Carvalho et al. 2007; Huang et al. 2009; Almeida et al. 2011).

Ultimately, it is possible to inhibit Gram-negative bacteria such as E. coli by photodynamic inactivation without an intracellular uptake. As it was shown in our experiments, the TMPyP is attached to the bacteria's outer cell wall inducing photodynamic growth inhibition.

It is still not clear if this affinity is necessary in general for phototoxicity against bacteria. Further investigations will be focused on the question if inline image may affect bacteria from outside without attachment of the PS to the bacterial cell walls.


The authors thank Lutz Jäger of the Department of Physics, Humboldt-Universität zu Berlin, for technical assistance and Dr Korte (Department of Biology, Humboldt-Universität zu Berlin) for technical support on CLSM and FLIM. Thanks are due to FCT and FEDER for funding the QOPNA unit (project PEst-C/QUI/UI0062/2011) and to the transnational cooperation FCT-DAAD programme.