![[Feedback]](/icons/banner/feedbackACT.gif){kind=icon}
![[For Subscribers]](/icons/banner/subscriberACT.gif){kind=icon}
![[Archive]](/icons/banner/archiveACT.gif){kind=icon}
![[Search]](/icons/banner/searchACT.gif){kind=icon}
![[Contents]](/icons/banner/tocACT.gif){kind=icon}
Previous Article | Next Article 
Antimicrobial Agents and Chemotherapy, June 2005, p. 2329-2335, Vol. 49, No. 6
0066-4804/05/$08.00+0 doi:10.1128/AAC.49.6.2329-2335.2005
Copyright © 2005, American Society for Microbiology. All Rights Reserved.
Wellman Center for Photomedicine, Massachusetts General Hospital, Boston, Massachusetts 02114,1 Department of Molecular, Cellular and Developmental Biology, Tufts University, Boston, Massachusetts 02111,2 Department of Dermatology, Harvard Medical School, Boston, Massachusetts 02115,3 Harvard-MIT Division of Health Sciences and Technology, Cambridge, Massachusetts 021394
Received 6 July 2004/ Returned for modification 12 January 2005/ Accepted 18 February 2005
 |
ABSTRACT |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
INTRODUCTION |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
It is known that gram-positive bacteria are generally more susceptible to PDT as compared to gram-negative species (17, 24). This difference is explained by the structural differences in the cell walls. Gram-negative cells have a complex outer barrier structure including two lipid bilayers, while gram-positive cells have only one lipid bilayer and a relatively permeable outer coat. The yeasts such as Candida albicans are even more resistant to photodynamic inactivation (PDI) due to the presence of a nuclear membrane that may present an additional barrier for PS penetration (29); therefore, higher doses of PS and light have to be used. It was shown previously (12, 26) that a positive charge on the PS molecule (such as a poly-L-lysine-PS conjugate or a cationic substitution-containing porphyrin) allows it to bind to, and in some cases penetrate, the microbial permeability barrier. Therefore, positively charged PS are generally more efficient and can act at lower concentrations than neutral and anionic PS molecules. Negatively charged PS are not able to penetrate this gram-negative barrier but may still be effective (although at higher concentrations); in this case, singlet oxygen generated during the irradiation at the outer surface or in solution in close proximity to the cell is thought to diffuse into bacteria and produce fatal damage to lipids and proteins in the inner membrane (4, 5).
We formed the hypothesis that there are three groups of antimicrobial PS: those that are tightly bound and penetrate into microorganisms, those that are only loosely bound, and those that do not demonstrate binding. In order to confirm this hypothesis, we employed three PS thought to be representative of these PS classes—a poly-L-lysine chlorin(e6) conjugate (pL-ce6), toluidine blue O (TBO), and rose bengal (RB)—with representative species from different microbial groups: gram-negative bacterium Escherichia coli, gram-positive bacterium Staphylococcus aureus, and yeast Candida albicans. Fluence-dependent cell survival after illumination with the appropriate wavelength of light before and after extracellular PS had been washed out from the cell suspension was compared. Using the same microbial strains, we also tested the hypothesis that microbial cells compete for available dye and reactive oxygen, and therefore the cell density is an important parameter for defining the effectiveness of antimicrobial PDT.
|
MATERIALS AND METHODS |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Exponential cultures obtained by reculturing stationary overnight precultures were used for all experiments. E. coli and S. aureus were grown in fresh medium for approximately 1 h to a density of 108 cells/ml; the optical density (OD) values at 650 nm were 0.6 and 0.8, respectively. C. albicans was grown for approximately 4 h to an approximate density of 107 cells/ml, corresponding to an OD of 6 at 650 nm (10-fold dilution measured). Cells of all microorganisms then were diluted with phosphate-buffered saline (PBS) without Ca2+ or Mg2+ or concentrated 10-fold by centrifugation at 1,550 g for 6 min and then resuspension in PBS. Concentrations of 109, 108, and 107 cells/ml for E. coli and S. aureus and 108, 107, and 10 6 cells/ml for C. albicans were used for each experiment. Dry weights of microbial pellets were measured. Microbial suspensions in distilled water (10 ml of 108 cells/ml for E. coli and S. aureus and 107/ml for C. albicans) were centrifuged in preweighed tubes and freeze-dried (Freezone 4.5; Labconco, Kansas City, MO). Exact cell numbers were confirmed by counting CFU obtained after serial dilutions were streaked horizontally on square brain heart infusion agar plates as described by Jett et al. (14). Gram-stained slides were prepared and examined by bright-field microscopy and a digital camera. This allowed a determination of the relative size of microorganisms studied.
Photosensitizers and light sources. The structures of the three PS used (RB, TBO, and pL-ce6) are shown in Fig. 1, and their optical properties are given in Table 1. RB (89% purity) and TBO (84%) were purchased from Sigma (St. Louis, MO); pL-ce6 was prepared as described previously (11). The conjugate had an average length of 164 lysine residues, to which were attached an average of six (range, four to eight) chlorin(e6) molecules. Stock solutions (TBO and RB at 2 mM, pL-ce6 at 3.3 mM ce6 equivalent) were prepared in water and stored at 4°C in the dark for no longer than 2 weeks before use. Spectra of stock solutions diluted 140-fold in methanol were recorded on a UV-visible spectroscopy system (model 8453; Agilent Technologies Deutschland GmbH, Waldbronn, Germany). A noncoherent light source with interchangeable fiber bundles (LumaCare, London, United Kingdom) was employed. Thirty-nanometer bandpass filters at ranges given in Table 1 allowed total powers of roughly 1 W to be obtained.
|
|
PDI studies. Illumination was performed either after or before excess dye was washed out. Wash allowed the adjustment of the cell densities to 108 (for E. coli and S. aureus) and 107 (for C. albicans) before illumination; then aliquots of 200 µl were placed in 96-well plates and illuminated with the appropriate light at room temperature. Fluences ranged from 0 to 200 J cm–2 at an irradiance of 50 to 400 mW cm–2. During illumination after defined fluences had been delivered, aliquots of 20 µl were taken to determine the CFU. The contents of the wells were mixed before sampling. The aliquots were serially diluted 10-fold in PBS without Ca2+ or Mg2+ to give dilutions of 10–1 to 10–6 times the original concentrations, and colonies were grown overnight at 37°C as described. Colonies were counted, and the survival fraction was determined compared to that in the untreated control. Dye in the absence of light and light alone were also used as controls. PS were generally not toxic for microorganisms in the dark (except for pL-ce6 and S. aureus at 107/ml), and light alone did not cause cell destruction. All experiments were performed in triplicate. Effective killing was taken to be >99.9% or 3 logs.
Statistics. Values are expressed as means ± standard errors. The difference between two means was compared by a two-tailed unpaired Student's t test assuming equal or unequal variances as appropriate. For PDI experiments, statistical significance between killing curves was determined at the highest fluence at which comparisons could be made. P values of <0.05 were considered significant.
|
RESULTS |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
|
|
|
The phototoxicity of TBO was much lower than that of pL-ce6 (compare Fig. 2B, 3B, and 4B). Both the PS concentration and the light dose necessary to kill S. aureus (10 µM and 32 J/cm2), E. coli (35 µM and 32 J/cm2), and C. albicans (50 µM and 40 J/cm2) were all higher than those required for pL-ce6.
The phototoxicity of RB was roughly comparable in degree to that of TBO (compare Fig. 2C, 3C, and 4C). RB was much more phototoxic than TBO toward S. aureus (0.25 µM and 4 J/cm2) and somewhat more phototoxic toward E. coli (35 µM and 8 J/cm2) but less phototoxic than TBO toward C. albicans (200 µM and 40 to 80 J/cm2).
The order of susceptibility of the different species to PDI mediated by pL-ce6 was E. coli > S. aureus > C. albicans. For TBO and RB, it was S. aureus > E. coli > C. albicans.
Effect of washing excess PS from cell suspension. The effect that washing excess PS from the cell suspension had on the PDI of the microbial species was strongly dependent on the chemical nature of the PS. For pL-ce6 (compare Fig. 2A, 3A, and 4A), the killing curves with the pL-ce6 washed from the cell suspension were identical to those with the pL-ce6 still present for all three microbial species. In fact there was a hint that PDI with pL-ce6 could be slightly more effective with a wash than without (Fig. 2A and 4A). For experiments using TBO, the effectiveness of PDI was dramatically reduced when TBO was washed out of the cell suspensions (Fig. 2B, 3B, and 4B). This applied to all three microbial species but was particularly evident for S. aureus and E. coli. For RB-mediated PDI, there was a very significant drop in the phototoxicity after washing in the case of E. coli (Fig. 3C) and C. albicans (Fig. 4C) but not in the case of S. aureus (Fig. 2C).
PDI of microorganisms depends on cell density. The effectiveness of PDI increased dramatically with decrease of cell density. This effect of cell density on PDI was observed for all microorganisms studied and with all PS used. The dependence of PDI on cell density was most pronounced with pL-ce6. S. aureus (Fig. 2A) showed 2 log10 of killing if 109 cells/ml was used; however, killing increased to 6 log10 when 108 cells/ml was used; this pL-ce6 concentration was toxic in the dark for cells at 107 cells/ml. E. coli (Fig. 3A) at a cell density of 109 was resistant to PDI with 0.75 µM pL-ce6; 1 log10 of killing was seen under the same conditions when the cell density was reduced to 108; finally, killing of 6 log10 was obtained with less light when the cell density was 107 cells/ml. This effect was seen if the dye was either left in the solution or washed out. C. albicans (Fig. 4A) showed only 0.3 log10 of killing with 108 cells/ml, 3 log10 of killing at 107 cells/ml, and 6 log10 of killing when the cells were at 106/ml.
Experiments with TBO showed less effect of cell density on PDI, and this was mainly apparent when TBO was left in the solution. S. aureus (Fig. 2B) cells at 109/ml were resistant to PDI, while 2 log10 of killing was observed at 108 cells/ml and 4 log10 of killing was observed at 107 cells/ml. E. coli (Fig. 3B) was resistant at 109/ml; however, 5 to 6 log10 of killing was seen after 32 J/cm2 at 108, and similar killing was seen after only 8 J/cm2 if 107 cells per ml was used. In experiments with C. albicans (Fig. 4B), the effect of cell density on PDI was seen both if TBO was left in the solution or washed, but it was less prominent in the latter case.
The effect of cell density on PDI was the least in experiments with RB. Experiments with S. aureus (Fig. 2C) gave no significant killing at 109 cells/ml, while if 108 and 107 cells/ml were used 4 to 6 log10 of killing was achieved even if excess PS was washed out. With E. coli (Fig. 3C) if RB was left in the solution, 2, 3, and 4 log10 of killing was seen at cell densities of 109, 108, and 107 cells/ml, respectively. The dependence of the killing on cell density with C. albicans (Fig. 4C) was only seen if RB was left in the solution. A total of 108 cells/ml were resistant to PDI, while 4 and 6 log10 of killing were achieved if 107 and 106 cells/ml were used.
Uptake studies. In experiments where the excess PS was washed out of the cell suspensions, it was possible to measure the uptake of the PS by the microbial cells (except for the case of S. aureus and RB, where the low concentration of RB necessary to achieve killing was below the detection limits of the fluorescence assay). The data are presented in Table 2. The uptake of pL-ce6 by all three microbial cells was much higher than the other PS used by 1 to 3 orders of magnitude, and in addition, the concentration of pL-ce6 used was much lower than that needed for the other PS. The uptake of pL-ce6 showed a marked inverse dependence on cell density. For all three cell types, decreasing the cell density 10-fold gave a corresponding 10-fold increase in cell uptake of pL-ce6. The uptake of TBO by the cells was much lower than that of pL-ce6 despite higher concentrations being used in the incubations. There was again an inverse dependence of uptake on cell density but not as marked as that seen with pL-ce6. When the cell density was reduced 10-fold, the uptake of TBO by the cells increased roughly 3- to 5-fold. The uptake of RB by the cells was of comparable magnitude to that of TBO. However, in marked contrast to TBO, the uptake of RB did not show any dependence on cell density. For E. coli and for C. albicans, there was the same uptake per cell even though the cell densities were varied 100-fold.
|
DISCUSSION |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
TBO showed an intermediate degree of binding toward the microbes in our study but was overall much less effective than the conjugate in killing all three microbial species, requiring 10 to 40 times higher concentrations of PS in the incubation and 2 to 30 times more light. TBO was much more effective in mediating killing when it remained in the suspension than when it was removed by washing for all three microbes. This implies that the TBO in solution is playing a highly active role in the light-mediated killing of the cells. There are two possible explanations for this observation. First, it may be due to the light-mediated extracellular generation of reactive oxygen species (ROS) (particularly singlet oxygen) that can diffuse into the cell producing damage cell death. The second explanation is that the first part of the PDI process produces some damage at the cell surface that allows more TBO molecules present in solution to bind and/or penetrate the cell and therefore increasing the potency of the PDI process. The finding that there was a significant dependence of killing on cell density (although somewhat less than found with pL-ce6) also implies that there is binding or interaction of the cells with a limited concentration of some species and that this occurs whether or not the PS is washed out from the cell suspension.
RB was able to mediate killing of all three microbial species with roughly comparable efficiency to that of TBO. For S. aureus, it was equally effective whether washed out of the cell suspension or left in solution. However for both E. coli and C. albicans, it was almost ineffective if washed out while mediating significant killing if left in solution. The facts that RB was effective at a very low concentration (0.25 µM) against S. aureus together with the equivalence of washing and no washing and the lack of cell density dependence all suggest that RB kills S. aureus by entering the cell but that this occurs by a diffusion-controlled process rather than by active binding to the cell as in the case of pL-ce6. Much higher concentrations of RB were needed to kill both E. coli and C. albicans, and RB was not effective against these species after a wash. There was a cell-density-dependent effect on the killing, which was particularly pronounced with C. albicans. These data suggest that it is the extracellular singlet oxygen (or other ROS) that is responsible for killing these species.
This finding is in agreement with a report of experiments in which the bacteria were separated from RB by a layer of moist air and singlet oxygen in the gas phase was generated and allowed to diffuse across the gap before contacting the bacteria (4). Gram-negative species were harder to kill than gram-positive species, and the intracellular content of carotenoids was found to protect the bacteria from photoinactivation. In another study, RB was covalently bound to small polystyrene beads that were allowed to mix with the bacteria in suspension (1).
It is interesting to observe the uptake of RB by the various microbial species at different cell densities. In the two species that could be measured (E. coli and C. albicans), the uptake was the same at each cell density, although the cell densities varied by 2 orders of magnitude. This is presumably explained by entry into or attachment of the RB molecules to the cells by some diffusion-controlled process rather than affinity-mediated binding. However, this uptake by the cells is apparently insufficient to cause any light-mediated killing after excess RB is washed from the suspension. Presumably this lack of killing reflects the subcellular location of the PS in a relatively superficial layer, as RB lacks any positive charges needed to gain deeper penetration. The cell density dependence of the killing of these two species by RB when still present in the cell suspension implies that the extracellular singlet oxygen is in limited concentration and reacts rapidly with the cells, so that the more cells that compete for available ROS then each cell gets less damage.
The great difference in size between the bacterial and fungal cells means that the amount of singlet oxygen needed to kill a Candida cell is much greater than the amount necessary to kill a bacterial cell. This is the case regardless of whether we use the pL-ce6 conjugate that penetrates the cells or TBO or RB, which work mainly from outside the cell. Although many workers have shown that Candida species are susceptible to PDI (2, 23, 29), it has not previously been possible to draw comparisons about relative susceptibilities to different PS between fungal and bacterial cells. Now we have shown that C. albicans is the hardest to kill with all three PS and that TBO comes closest to having equivalent efficacy for Candida as it has for bacteria. In agreement with a report by Polo et al. (25), we have shown that a polycationic PS conjugate with pL is also highly effective against fungal cells as well as bacteria.
The dependence of the effectiveness of PDI on the cell density in the suspension, which was found to some degree for all species and for all PS, may be intuitively obvious, but this is the first time it has been shown. The finding is important to remember when comparing literature reports of antimicrobial PDI as the cell densities used by various groups can vary. In conclusion, the data reported here will be important in designing experiments concerning antimicrobial PDI and PDT for infections. It should be remembered that both TBO and RB have a long history of medical applications, particularly involving local application to stain various tissues, and are generally accepted as safe (3, 16, 18, 26). Their significant activity in mediating PDI may allow them to progress quickly to clinical trials.
|
ACKNOWLEDGMENTS |
|---|
We thank John Demirs for assistance with microscopy and determination of relative sizes of microorganisms.
|
FOOTNOTES |
|---|
|
REFERENCES |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| Clin. Vaccine Immunol. | Clin. Microbiol. Rev. |
|---|---|
| J. Clin. Microbiol. | ALL ASM JOURNALS |
, 109 cells/ml; 
and 
, 108 cells/ml; 
and 
, 107 cells/ml; •, 
, P < 0.05, between no wash and corresponding wash.
