Glutathione levels in vivo The oxidizable substrates found in the in vitro studies described in earlier chapters to be capable of reducing the dye to the active semireduced form that reacts with oxygen were NADH, glutathione, cysteine, tryptophan, and tyrosine. All of these are of potential in vivo relevance, as typical in vivo levels are similar to those used in the in vitro assays. One of these substrates, glutathione, was the subject of in vivo assays to determine whether this reducing potential translates to free radical generating activity in the cell. Cells exposed to dyes were assayed for their glutathione content, to determine whether the cell's supply of reduced glutathione were depleted by the action of the dye.
Factors that affect survival of E. coli in vivo Studies of the photodynamic
effect in vivo in E. coli suggest that the mediator of phototoxicity is OH.. Among the
results obtained in our laboratory that support this suggestion are the induction of
the protective enzymes SOD and catalase by exposure to the dyes, and of
protection against the light and oxygen dependent lethality of dyes by pre-induction
of these enzymes prior to
,183 In addition, it was determined that cell-permeant OH. scavengers protect against lethality in vivo. Martin and Logsdon found micromolar levels of the dye toluidine blue in combination with illumination halted growth of E. coli cells in NB medium, but that growth could begin again after a lag period, as though resistance were conferred by the induction of protective enzymes. Greater concentrations of the dye required correspondingly greater lag periods before growth could occur.306 Under the conditions of the growth experiments the dyes toluidine blue and proflavin, when illuminated, induced superoxide dismutase and catalase, and the enzyme levels induced increased as dye concentration was increased. 306
When the protective enzymes superoxide dismutase and catalase were induced by pre-growth in rich, glucose-free media, or by pre-growth in the presence of small concentrations of paraquat in combination with manganese supplementation in a glucose-containing minimal medium, dye lethality was dramatically reduced.200 Similarly, lethality was reduced in the presence of the cell-permeant OH. scavengers thiourea, sodium benzoate, and DMSO. All of this data supports the theory that the lethality of illuminated dyes is due to OH., which is produced by reactions requiring hydrogen peroxide and O2-, and presumably mediated by illuminated dyes in vivo.
One treatment used to induce superoxide dismutase and catalase, pre-growth in the presence of small concentrations of paraquat, has been found to also induce endonuclease IV, which repairs DNA damage by removing apurinic or apyrimidinic sites and blocked 3' termini.133 The fact that superoxide dismutase, catalase, and endonuclease IV are all coinduced by the same oxidative stress is highly suggestive of what the mediators of damage are and what a likely target of damage is, namely DNA, in the presence of oxidative stress. However, the possibility of multiple enzymes being induced by the same protective pre-treatment implies that protection by pre-induction does not definitively prove that superoxide dismutase and catalase are the only enzymatic agents that protect against lethality. In this study, the question of whether kill is ameliorated by supplementation with plasmids coding for the protective enzymes catalase and SOD was examined. In this way, unambiguous evidence was obtained that it is possible that not only the enzymes that protect against DNA damaging agents are involved in the protection of the cell against the lethal effects of illuminated dyes, but also that enzymes that protect against toxic oxygen species play a major role in protecting DNA and in defending the cell against oxygen toxicity. Therefore the toxic oxygen species, O2- and H2O2, and species derived from them, play a part in mediating photodynamic damage.
DNA damage in vivo The DNA damage that occurs in vivo as a result of photodynamic action has been studied in our laboratory by examining the varying survival of strains deficient in several different DNA repair enzymes, in the presence of illuminated dyes. Differences in survival among these strains reflect different forms of DNA damage that require the action of distinct DNA repair systems for their correction. Strains deficient in certain DNA repair enzymes were much more sensitive to photodynamic killing than their otherwise isogenic parent strains (fig. 70.). An E. coli K12 strain, xthA, which is deficient in that organism's major AP endonuclease, exonuclease III, showed great sensitivity to photodynamic kill, compared to otherwise identical K12 strains. The deficiency in exonuclease III rendered the strain as sensitive to kill as E. coli B, while addition of thiourea or preinduction of the protective enzymes SOD and catalase lent it complete protection against photodynamic kill. E. coli B is not deficient in SOD or catalase, but preinduction of these enzymes protect it against kill in just the same way. It seems an interesting possibility that the difference between K12 and B strains that renders the latter so much more sensitive to photodynamic kill may be a deficiency in DNA repair enzymes.
The SOS response, which is induced by DNA damage, was studied in an earlier investigation by Carla Bull of our laboratory with photodynamic action as a response trigger. Induction of the SOS response was seen in a din(damage-inducible)/ß-galactosidase fusion strain when it was exposed to azure c plus light (fig. 71.). The fusion strain contains the structural gene for ß-galactosidase, an enzyme which is convenient to assay, fused to a promotor which is activated by DNA damage. Exposure of the cells containing the gene fusion expressed ß-galactosidase in response to a drug known to induce the SOS responcse, mitomycin c; it also expressed high levels of ß-galactosidase in response to exposure to azure c and visible light. Light in the absence of dye and dye in the absence of light each induced expression more moderately.
Figure 70. Survival of E. coli xthA and E. coli B following incubation with 2 mM azure
c. E. coli strains grown to the mid-log phase in M9 salts + 0.4% glucose were
harvested and washed twice with M9 salts. The cells were diluted into three milliliter
cuvettes containing M9 salts, pH 7.4, 0.4% glucose, 80 mg/ml chloramphenicol, 2
mM azure c and scavengers as indicated. Cuvettes were exposed to visible light at
2.1 mW/cm2. Line 1, xthA grown in medium containing 100 mM Mn and 10 mM
paraquat; Line 2, xthA, grown as in line 1 plus 100 mM thiourea; Line 3, AB 1157
(xthA+); Line 4, xthA plus 100 mM thiourea; Line 5, E. coli B; Line 6,
The genotype of AB1157 is leuB6 D(gpt-proA2) hisG4 argE3 lacY1 galK2 ara-14 mtl-1 xyl-5 thi-1 tsx-33 rpsL31 supE44 rac ; the genotype of xthA is the same except for the mutation xthA.
Figure 71. Induction of the SOS response in E. coli GW 1040 by azure c and visible light. Cells were grown in M9 medium supplemented with 0.5% Casamino acids at 305C while exposed to 1.95 mW/cm2 visible light. Samples of bacteria were removed from exponential phase cultures at intervals and assayed for ß-galactosidase activity.319 Mit. C, cells exposed to mitomycin c plus light; Azure C Lt, cells exposed to azure c plus light; Lt, cells exposed to light but neither azure c nor mitomycin c; Azure C Drk, cells exposed to dye in the absence of illumination.
In this study, DNA damage caused by illuminated dyes was examined, using rapid isolation of plasmid DNA, to determine whether it involved single-stranded nicking. Protection from nicking was looked for in the presence of cell-permeant OH. scavengers, to determine whether damage was reduced by such treatment, giving information as to the mediator of the DNA damage.
In the hot ethanol extraction procedure, 1.5 mls of culture was centrifuged in an Eppendorf microcentrifuge for 3 minutes. The supernatant was removed and the pellet was resuspended in 5 ml of a 150 mM potassium phosphate buffer, pH 7.6. 57 ml of 86% prewarmed ethanol were added and heated for 3 minutes at 705C. The supernatant after centrifuging the extract as before was removed to another tube and the ethanol dried off in a Savant Speed Vac Concentrator. The same phosphate buffer was added to bring the total volume up to 1 ml, then DTNB (6.67 ml of a 3.92 mg/ml solution in 150 mM potassium phosphate buffer, pH 7.0) added. After about two minutes, absorbance at 412 nm was determined on a Gilson spectrophotometer. Concentration was determined using the molar extinction coefficient of 13,600 M-1cm-1.320 The levels were compared to cell survival by plating 50 ml samples of the cells on LB medium (1% NaCl, 0.5% yeast extract (Difco), 1% Bacto-tryptone (Difco), 1.5% agar (Difco)) prior to the extraction process. Cells were grown 24 hours at 375 and counted for survival.
Methods used in determining protective factors involved in survival
Preinduction of enzymes in cells In the kill experiments, some cells were subjected to preinduction treatments to increase their intracellular levels of SOD and catalase. Cells were grown overnight in minimal media, then inoculated into minimal media with a 2 to 4% inoculum. The cells were allowed to grow for one hour before the addition of 100 mM MnCl2 and 10 mM paraquat (methyl viologen, Sigma). The preinduced cells were then grown for the same amount of time as the non-preinduced cells, until the non-preinduced cells had reached a turbidity of 100 to 150 Klett units and the preinduced cells had reached a turbidity of 20 to 50 Klett units.
Kill experiments Fifty mls of medium are inoculated with a 2% inoculum from
a saturated overnight culture. Cells are grown to a Klett value of 100 to 125,
equivalent to an O.D.600 of 0.20 to 0.25. The cells are harvested by centrifuging at
6000 G for ten minutes, washed by resuspending in 40 mls phosphate buffer (50
mM, pH 7.6, with 0.1 mM EDTA), recentrifuging, and resuspending again in 40 mls
phosphate buffer. Each of a number of sterile cuvettes is filled with 3 mls of sterile M9
minimal salts (6 g Na2HPO4, 3 g KH2PO4, 0.5 g NaCl, and 1 g NH4Cl, per liter), 60 ml
sterile glucose (20%), 60 ml chloramphenicol (2 mg/ml in sterile water), plus the dye
to be examined, the amount of which to use is predetermined by trial and error to
provide a significant amount of kill. The chloramphenical is added to prevent
induction of protective enzymes during the dye exposure; the glucose, to keep the
cells metabolizing and maintaining cellular stores of oxidizable substrates. Cells
were harvested in mid-log phase after inoculation from an overnight culture. About
75 ml of resuspended bacteria are used per cuvette: an amount in each cuvette
calculated from the Klett absorbance to contain the same number of bacteria. The
cuvettes are then exposed to visible light to cause photochemical activation. The
light was provided by standing the cuvettes up in direct contact with the surface of a
General Electric Warm White fluorescent bulb. The light intensity was 1.2
Samples are withdrawn, diluted, and plated out on LB agar (10 g Bacto Tryptone (Difco), 5 g yeast extract (Difco), 10 g NaCl, and 15 g Bacto Agar (Difco) per liter) to determine survival after fifteen, thirty-five, and sixty-five minutes. Plates were incubated for 24 to 48 hours at 375C in the dark before counting surviving colonies. All dilutions were plated in duplicate or triplicate.
The E. coli B strains were prepared by isolating the specific plasmids from strains that were obtained from the investigators who constructed them (Table IV) and inserting them into E. coli B 23226.
Plasmid isolation For small scale plasmid preparation, five mls of LB plus ampicillin (50 mg/L) were inoculated with the strain containing the plasmid of interest and grown overnight. The cells were harvested by spinning in an Eppindorf microcentrifuge for two minutes and resuspended in 0.35 mls Triton
lysis buffer (0.5% Triton X-100, 8% sucrose, 50 mM EDTA, 10 mM Tris.Cl, pH 8.0). 25 ml of 10 mg/ml lysozyme in 10 mm Tris-HCl, pH 8.0, was then added,the microfuge tubes were vortexed for three seconds and then boiled for forty seconds. Each tube was spun for twenty minutes in the Eppindorf microcentrifuge, and the pellet removed with a sterile toothpick. Each tube was precipitated with 450 ml isopropanol and 25 ml 3 m sodium acetate, pH 5.2, at -205C for several hours. After twenty minutes of centrifugation at 45C, the pellet was dried in air with the tubes inverted, and resuspended in 50 ml TE.
Table IV. Plasmids Used in Kill Experiments
plasmid relevant genotype reference
pDT1-5 sodA+ 122
pBT22 cat+ 119
pBT28 cat + "
pWB21 nfo+ 135
Transformation Cells were grown in LB from a 1% inoculum of a saturated overnight culture to a Klett of no more than 80, and 45 mls were centrifuged for ten minutes at 6000 G. They were then resuspended in ten mls of PC (50 mm CaCl2, 1 mm PIPES, pH 6.8), held on ice 20 minutes, spun down again for ten minutes at 6000 G, and resuspended in ten mls PC. The DNA was diluted to a concentration of only 0.03 to 1.0 mg/ml. The competent cells (100 ml) were added to 50 ml of the DNA solution, shaken gently, held undisturbed on ice for 15 minutes, heat-shocked at 425C for two minutes, held on ice for five minutes, and held at room temperature for ten minutes; 0.35 ml of prewarmed LB was then added to each tube, and the cells incubated for 45 minutes at 375C before plating on LB + ampicillin. The plates were incubated at 375C for 24 hours.
Preparation of crude protein extracts Assays on crude protein extracts were used to aid in interpretation of results from kill experiments. The assays for the enzymes superoxide dismutase, catalase, and peroxidase were performed on crude cell-free extracts of E. coli cultures. The extracts were kept cold after harvesting, dialysis was carried out at 45C, and dialyzed extracts were used immediately or frozen; rethawing and refreezing was kept to a minimum. Frozen cells, in our experience, retain their SOD, catalase, and peroxidase levels indefinitely. Plasmid-containing strains were grown in 50 mls inducing or non-inducing minimal media (M9), at 375C, to late log phase (O.D.600 = 0.2 or 0.25), harvested by centrifuging ten minutes at 6000 G in 50 ml tubes, washed, sonicated at 35 watts, on ice, in 6 bursts of 45 seconds each with a Heat Systems-Ultrasonics, Inc., Sonifier Cell Disruptor, model W185, and dialysed against three 1 liter changes of a phosphate-EDTA buffer, then subjected to enzyme assays.
Catalase assay The crude protein extract is added to 10 mm H2O2 in 3 ml of
0.05 m phosphate buffer, pH 7.5, 0.1 mm EDTA. Amounts to use are determined by
trial and error; about 10 ml is a good amount to start with. The rate of decline in
absorbance of the hydrogen peroxide is followed spectrophotometrically on a
Gilford 2000 spectrophotometer at 240 nm. The molar extinction coefficient used is
43.6 m-1cm-1 One unit is the amount of enzyme required to break down 1 mmole of
O-dianisidine assay for peroxidase The crude protein extract is added to 3 ml. of 10 mm phosphate buffer, pH 6.5, 0.1 mm EDTA, containing 0.6 mg o-dianisidine, and 5 mm H2O2 . Change in absorbance is followed at 460 nm, the absorption maximum of the colored product of o-dianisidine oxidation. The molar extinction coefficient of oxidized o-dianisidine is 11,300 m-1cm-1.322
Cytochrome c assay for superoxide dismutase The crude protein extract was added to 3 mls of a cocktail consisting of 0.05 m phosphate buffer, pH 7.8, 0.1 mm EDTA, 50 mm xanthine, 10 mm oxidized cytochrome c, and xanthine oxidase. An amount of xanthine oxidase was used that gave a reduction rate of 0.025 A550 units per minute. One unit of superoxide dismutase is defined as the amount of enzyme that results in a 50% inhibition of the reduction of cytochrome c. Different amounts of each extract were assayed until an amount was found that would inhibit this reduction rate by between 40 and 60%, and the amount of extract that would yield one unit was calculated from this degree of inhibition.104
Methods used in determining strand scission in vivo
Plasmid-containing E. coli B 23226 was subjected to the presence of dyes while exposed to light. The plasmid used was either pBR322 or pJC7, a pBR322 derivative containg 2-3 kb of clostridial DNA. The cells were grown in minimal media M9 with ampicillin to a density of 100 Klett units, then centrifuged 10 minutes at 6000 G to harvest and resuspended in 6 mls (fig. 75) or 3 mls (fig.76) M9 salts, placed in small (60 mm) sterile disposable petri dishes (Falcon) with dye at the indicated concentrations, 20 ml 20% glucose per ml, 20 ml of a 2 mg/ml aqueous solution of chloramphenicol per ml, and, if indicated, 1.4 m DMSO or 100 mM thiourea. The plates were placed, with lids on, directly under either a Warm White fluorescent bulb, with minimal clearance between lid and lamp. Dark treatments were placed under an inverted stainless steel pan. A Thomas slow rotating platform was used to keep the cells continuously suspended.
After light treatment, cells were placed in Beckman amber microfuge tubes and harvested by spinning for one minute; the supernatant was removed using a pasteur pipette, and 43 ml of Triton-sucrose lysis buffer (0.5% Triton X-100, 8% sucrose, 50 mM EDTA, 10 mM Tris.Cl, pH 8.0) was added. The cells were resuspended by vortexing, two tubes were combined, DMSO added to stop DNA damage that might occur while the DNA was being isolated from the cell debris, and 10 ml of 10 mg/ml lysozyme were added. The lysozyme was mixed in by vortexing for 3 seconds, and the mixture then boiled for 40 seconds in a boiling water bath. Up to thirty minutes of microcentrifugation served to clarify the plasmid solution. 20 ml of this solution was mixed with 5 ml `blue juice' (50% glycerol, 0.5% bromphenol blue) and loaded onto one well of a 0.8% agarose minigel. Electrophoresis separated the bands of supercoiled, nicked, or linearized plasmid DNA. The gel was photographed using Polaroid type 667 film, which was then digitized using a Thunderware[ scanning device on a MacIntosh[ computer. Scans of the digitized bands were printed out using Scanning Analysis[ software315 and integrated by cutting out the peaks and weighing them on a Mettler balance.
Reduced glutathione, as detected by assay with Ellman's reagent, was depleted in vivo after exposure of the E. coli B cells to different concentrations of azure c for 65 minutes (fig. 72). Increasing concentrations caused greater depletion of glutathione. 65 minutes exposure to 1 mM azure c reduced glutathione levels by 18%, while the same exposure to 5 and 25 mM azure c caused reductions of 68% and 83%, respectively. Ellman's reagent is not entirely specific, in that reduced cysteine also may react with it, but this in no way detracts from the conclusion that oxidizable substrates are depleted by reaction with azure c. Hydroxyl radical assays and DNA strand scission assays showed that oxidizable substrates such as glutathione react with dyes in vitro. In addition Martin and Logsdon demonstrated that the dye sensitized oxidation of glutathione led to oxygen consumption and H2O2 formation which was augmented by addition of SOD. This experiment shows that this reaction actually occurs in vivo. While assays on other oxidizable substrates of biological importance, such as NADH and tryptophan and other amino acids, were not performed, one would expect that they also are depleted after in vivo exposure to illuminated dyes.
Figure 72. Depletion of reduced glutathione in E. coli B after exposure to dye, corrected for survival. After growth and preparation as described above under Materials and Methods, cells were exposed to no azure c or to 1, 5, or 25 mM azure c for an illumination period of sixty-five minutes.
Kill experiment results
After sixty-five minutes of exposure to 1 mM azure c, there was too much kill to make significant distinctions between the plasmid-free E. coli B 23226 and the strains containing the control plasmid pJC7, the SOD plasmid pDT1-5,and the catalase plasmids pBT22 and pBT28 (fig. 73). The endo IV plasmid, however, conferred a significant degree of protection on the strain which contained it, while both preinduction of protective enzymes and the presence of the hydroxyl radical scavenger thiourea conferred enough protection that with their help E. coli B could survive considerably better than even the E. coli K12 strain AB1157. E. coli K12 strains such as AB1157 have a considerably greater resistance to dyes than E. coli B strains such as E. coli B23226, usually around two to three orders of magnitude greater.
After thirty-five minutes of exposure in the same experiment, kill levels are sufficiently lower than after sixty-five minutes that the relative levels of protection can be distinguished. Preinduction of protective enzymes, addition of thiourea, and the presence of the endo IV plasmid all gave essentially complete protection, equivalent to the control K12 strain AB1157. Following this group of survivors, both catalase plasmid strains conferred a significant degree of protection, showing a drop in survival of about twenty-fold compared to starting values. The plasmid-free and control plasmid strains came in with much lower survival rates, around 0.1% of starting values, as did the SOD plasmid strain.
The survival rates after fifteen minutes were too high to allow much distinction between classes of survivors.
These rates are exactly what was predicted from in vitro experiments. Catalase showed a noticeable degree of protection, while superoxide dismutase did not; this makes sense, since even if all of the O2- is broken down into
hydrogen peroxide, there is still excited semi-reduced dye to take its place in the reactions that lead to the formation of OH., reducing ferric iron to ferrous iron so that it can participate in the Fenton reaction, producing OH.. Preinduction treatments, which increase catalase and superoxide dismutase approximately ten-fold, and endonuclease IV probably two-fold (estimating from other results; this value was not determined),133 give a significant degree of protection. Table V lists the levels of the protective enzymes SOD, catalase, and peroxidase in extracts made from the cells that were used in this experiment. 100 mM thiourea, a cell-permeant hydroxyl radical scavenger, gave a great deal of protection as well.
LEVELS OF SUPEROXIDE DISMUTASE, CATALASE, AND
PEROXIDASE IN CELLS USED IN PLASMID PROTECTION
STRAIN CATALASE PEROXIDASE SOD
(units/mg protein) (mmoles/min-mg protein) (units/mg protein)
B 23226 0.729 0.116 20
B 23226/pBT28 20.9 2.74 15
K12 AB1157 0.126 0.064 5.4
B23226 (preind.) 7.55 0.701 141
B23226/pBT22 21.2 3.02 17
B23226/pJC7 0.53 0.132 15
B23226/pWB21 0.39 0.842 15
B23226/pDT1-5 1.41 0.21 57
A higher level of azure c, 3 mM, was used to distinguish the protective effects of the preinduction of protective enzymes in a control plasmid strain, E. coli B/pJC7, with the effects of a similar preinduction in a catalase plasmid strain, E. coli B/pBT22 (fig. 74). After 65 minutes of exposure, the higher concentration of dye resulted in complete killing of the strains that were not preinduced, and even in the complete killing of the preinduced control plasmid strain. Thiourea provided complete protection again at 65 minutes, and a dark control (3 mM azure c in the absence of light) was run this time, resulting in very little kill. The only other strain showing any protection at all was the preinduced catalase plasmid. At thirty-five minutes, unpreinduced treatments of both plasmids showed complete kill, but the preinduced control plasmid showed some survival, though not as much as the preinduced catalase plasmid. At fifteen minutes of exposure to the 3 mM azure c, survival levels were again too high to make meaningful distinctions between the treatments, except for the unpreinduced treatment of the control strain. Table VI shows the levels of the protective enzymes SOD, catalase, and peroxidase in extracts made from the cells that were used in this experiment; the levels are similar to but not the same as the levels shown for the previous experiment in Table V, showing the necessity of testing the enzyme levels in each experiment due to inevitable variation, although similar methods of cell growth and induction were followed as described under Materials and Methods.
It is noteworthy that, even against a background of preinduced repair enzymes, an increased HPI level can still have a protective effect against photodynamic killing. While only one dye, azure c, was used in these experiments, the effectiveness of this particular dye in producing oxygen radicals, penetrating
LEVELS OF SUPEROXIDE DISMUTASE, CATALASE, AND PEROXIDASE IN CELLS USED IN PREINDUCTION KILL EXPERIMENT
STRAIN CATALASE PEROXIDASE SOD
(units/mg protein) (mmoles/min-mg protein) (units/mg protein)
B 23226/pBT22 9.04 1.95 15
(preinduction) 14.9 1.61 158
B23226/pJC7 0.083 0.099 15
(preinduction) 2.59 0.27 159
E. coli, intercalating into DNA, damaging DNA, and killing cells has been repeatedly demonstrated; moreover, previous experiments have amply shown that the radical generation mechanism of all the dyes is essentially similar.
DNA strand scission results
E. coli strain xthA showed significantly increased plasmid nicking after exposure to 20 mm acridine orange (fig. 75) in the light, as compared to the same exposure in the light and as compared to no acridine orange exposure. After treatment with 5 mM azure c, E. coli B/pJC7 showed noticeably greater nicking after illumination in the presence rather than the absence of dye. The hydroxyl radical scavenger thiourea, however, lent a significant degree of protection form the dye-induced damage (fig. 76). This result further underscores the significance of OH. in the single stranded scission of DNA in vivo.
Figure 75. In vivo DNA damage by acridine orange. Cells were exposed to dye as described in Materials and Methods for 65 minutes with or without light.
Figure 76. In vivo DNA damage by azure c is ameliorated by the OH. scavenger thiourea. Cells were exposed to light as described in Materials and Methods for sixty-five minutes with or without 5 mM azure c, or with 5 mM azure c and 50 mM thiourea.