Lucigenin chemiluminescence as a probe for measuring reactive oxygen species production in Escherichia coli

186,316319

ANALYTICALBIOCHEMISTRY

(1990)

Lucigenin Chemiluminescence as a Probe for Measuring Reactive Oxygen Species Production in Escherichia COG’ R. Peters,

Timothy department

Received

Jeffrey

of M~rob~ology,

November

M. Tosk,

Loma Linda

and Eric A. Goulbourne,

~~i~ers~ty,

Loma Linda,

Inc.

Priestley, the discoverer of oxygen, noted 200 years ago that aerobic organisms had developed the means to live within a narrow range of oxygen concentrations (1). During aerobic metabolism reactive oxygen species, such as superoxide radical anion (O;), arise as a result of the incomplete reduction of molecular oxygen. Superoxide radicals themselves can readily undergo further re* This work was supported by National Science Foundation grant Rll-8507560 to E.A.G. and by Public Health Service grant GM29481 from the National Institute of General Medical Sciences to B. L. Taylor. ’ To whom correspondence should be addressed. Present address: The Proctor and Gamble Company Miami Valley Laboratories, P.O. Box 398707, Cincinnati, Ohio 45239-8707. 316

Jr.2 9.2350

16,1989

Addition of oxygen to whole cells of Escherichia coli suspended in the presence of the chemiluminescent probe bis-N-methylacridinium nitrate (lucigenin) resulted in a light emission increase of 200% of control. Addition of air to cells showed a chemilumin~ent response far less than the response to oxygen. The redox cycling agents paraquat and menadione, which are known to increase intracellular production of 0, and HzOz, were also found to cause a measurable increase in lucigenin chemiluminescenee in E. eoli cells when added at concentrations of 1 and 0.1 mM, respectively. The oxygen-induced chemiluminescent response was not suppressed by extracellularly added superoxide dismutase or catalase. Further, the lucigenin-dependent chemiluminescent response of aerobically grown E. coli to oxygen was significantly greater than that of cells grown anaerobically. Heat-killed cells showed no increase in chemiluminescence on the addition of either oxygen, paraquat, or menadione. These results show that lucigenin may be used as a ehemiluminescent probe to demonstrate continuous intracellular production of reactive oxygen metabolites in E. coli. o l~eo Academic Press,

California

duction to form hydrogen peroxide (H,O,) and the hydroxyl radical (. OH) (2). Today, the toxicity of reactive species produced when oxygen is utilized is well known (3,4). Aerobic and facultative anaerobic microorganisms possessdetoxifying mechanisms against reactive oxygen species, notably superoxide dismutase (SOD)3 and catalase. In addition, motile bacteria can minimize oxygen toxicity by moving away from high oxygen concentrations (negative aerotaxis) (5). To elucidate the mechanisms involved in these protective phenomena it was necessary to develop techniques to measure the kinetics of reactive oxygen species production in bacteria on addition of oxygen. Lucigenin, a water-soluble acridinium salt, reacts with oxygen metabolites to yield an electronically excited Mmethylacridone which in turn relaxes by photon emission (6,7). Chemiluminescence caused by the reductive dioxygenation of lucigenin has been used to measure the respiratory burst metabolism that follows phagocyte stimulation. Furthermore, it has been reported that lucigenin-dependent chemiluminescence (CL) of human polymorphonuclear leukocytes reflects superoxide production (8). The specificity of lucigenin CL can be assessedby examining the effect of specific oxygen species scavengers on chemiluminescence. Catalase, a hydrogen peroxide scavenger, and SOD, a superoxide scavenger, are frequently employed (9). In this study we have adapted these methods for use in measuring the production of reactive oxygen metabolites in bacteria. MATERIALS

Microorganisms

AND

METHODS

and Culture

Conditions

Escherichia coli wild-type AN386 (10) was grown aerobically with stirring in LB medium (11) supplemented with 0.3% sodium succinate and 0.5% glycerol, and LB 3 Abbreviations nescence; GTB,

used: SOD, glycerol taxis

superoxide buffer.

dismutase;

CL, chemilumi-

0003-2697/90 $3.00 Copyright 0 1990 by Academic Press, Inc. All rights of reproduction in any form reserved.

OXYGEN

METAROLITE

medium supplemented with 0.3% (sterile) xylose. AN386 cells were grown anaerabically in a Coy anaerobic chamber in xylose-supplemented LB medium. A cytochrome oxidase mutant (RG99) (10) was grown anaerobically in LB medium supplemented with 0.4% glycerol and 0.5% potassium nitrate. Cells were harvested in midexpo~ential growth phase and suspended to an ODBoo nm of 0.35 in glycerol taxis buffer (GTB) which contained 100 PM EDTA, 10 mM potassium phosphate buffer, 1 mM magnesium sulfate, 1 m&l ammonium sulfate, and 10 mM glycerol (final pH 7.0). To obtain heat-killed cells bacteria suspended in GTE) were immersed in a boiling water bath for 10 min.

A 20 m&l stock solution of lucigenin (bis-~-methylacridinium nitrate, Aldrich) was prepared in dimethyl sulfoxide (Sigma) and stored frozen in the dark until use. Menadione (2-methyl-1~4-naphthoquinone sodium bisulfite, Sigma) was prepared shortly before use as an aqueous 10 mM solution. A 100 mM aqueous solution of paraquat (l,l’-dimethyl-4,4’-bypyridinium dichloride, Sigma) was prepared in advance and kept frozen in the dark until use. Catalase (9000 pg/ml, Sigma) or superoxide dismutase (l~hily~ed powder, Sigma) in GTB was added to cell suspensions in various amounts. To confirm the activity of SOD in GTB, superoxide was generated enzymatically by the reaction of xanthine with xanthine oxidase (Sigma) (12) and the effect of SOD on lucigenin CL was measured.

PRODUCTION

IN

317

E. COLZ

0

io

io

$0

40

5.0

so

io

io

so

100

Time (s) FIG. 1. Lucigenin ehemiluminescent response of aerobically grown wild-type E. colt to oxygen and air. Cell samples treated with oxygen (r) are compared with ceils treated with air (0) and oxygen addition to uninoculated GTE3 (a). Air or oxygen was introduced into each sample during the 10th s. Each point represents three averaged samples, with error bars representing SE. (Error bars not shown are smaller than the symbol.)

ringe (Hamilton) through a 24-gauge needle. The final concentration of paraquat in each sample was 1 mM, and that of menadione was 100 FM. To samples treated with menadione 5 ~1 of a 10 mM CuC12 solution was added (50 j&M final concentration} as a catalyst for menadione action (13) shortly before counting. E. coli and GTB samples were treated with various amounts of SOD and catalase before and during CL measurements by injection as described above.

Chemiluminescence Measurements one-milliliter aliquots of cells suspended in GTB were placed in 3.5ml plastic tubes (Sarstedt) and treated with ‘2 ,el of stock lucigenin, yielding a final lucigenin concentration of 40 PM. In addition, cell-free control samples contained 40 f&M lucigenin in GTB, Samples were placed in an automated Packard Picolite Model 6500 luminometer where they were maintained with stirring at 30°C. The CL response of each sample to the addition of oxygen or air was measured by determining the total photon emission during each l-s interval of a 99-s counting period. Baseline emissions were recorded for the first 9 s of this counting period. During the 10th s a 25 ml/min stream of oxygen (99.6%, Big Three Industries) or air as indicated was introduced into each sample through a 24gauge needle. The CL response of cells to the addition of paraquat or menadione was measured by determining the total photon emission during each 5-s interval of a 495-s counting period. Baseline emissions were counted for the first 20 s of this counting period, after which 10 ~1 of a stock solution of either paraquat or menadione was introduced to each sample by injection from a 100-~1 sy-

RESULTS In the presence of lucigenin the addition of O2 to a lml suspension of wild-type E. colt resulted in an increase in CL to a level 10 times higher than that of heat-killed cells (Fig. 1) or GTB alone (Fig. 3). Light emission increased for 30 to 50 s after O2 addition to a maximum CL level which was sustained for the duration of the counting period. The lucigenin-dependent chemiluminescent response of E”. coli to the addition of air was significantly less than that of O2f with air-treated cells showing only a slight increase in CL from baseline (Fig. 1). The baseline CL of heat-killed cells and uninoculated GTB was less than half that of respiring cells, and neither heatkilled nor uninoculated GTB samples showed any increase in chemiluminescence on addition of O2 (Figs. 1 and 3). The addition of 1 mM paraquat to wild-type cells treated with lucigenin resulted in an increase in CL of more than 100% of baseline levels within 475 s (Fig. 2). Photon emissions increased most rapidly during the first 75 s after paraquat addition, and a more gradual CL increase was maintained for the remainder of the 495-s counting period. Heat-killed cells treated with lucigenin

318

PETERS,

TOSK,

AND

GOULBOURNE

duced by the reaction of xanthine with xanthine oxidase (data not shown).

z 180 x 160 $140

DISCUSSION

;

These results show that the chemiluminescent probe lucigenin may be used to detect the production of reactive oxygen species in metabolically active E. coli cells. The claim that increases in lucigenin-dependent CL in E. coli cells are caused by increases in the production of one or more reactive oxygen species is first supported by the CL increase observed in lucigenin-treated E. coli cells when exposed to pure O2 (Fig. 1). It is known that light production results from the reaction of 0; and Hz02 with the chemiluminescent probe lucigenin (1). It is also known that aerobically respiring cells produce reactive oxygen intermediates as a product of 0, reduction (3,16). These intermediates include O,, HzOz, and . OH which are thought to be responsible for the toxic effect of O2 on E. coli (16-18) and may play a primary role in the growth-inhibiting effect of pure O2 on E. coli (17). It is therefore very likely that it is the increase in intracellular metabolite species production resulting from the introduction of pure oxygen that causes the observed increase in lucigenin-dependent CL. The conclusion that lucigenin-dependent CL reflects the production of reactive oxygen species in E. coli is further supported by the increase in CL of lucigenin-trea~d cells observed on addition of paraquat and menadione. It has been reported that both paraquat and menadione, when added to E. cob, cause an increase in the intracellular production of 0; and HzOz (19). That CL increases when these compounds are added to E. coli (Fig. 2) strongly indicates that it is the cellular production of reactive oxygen metabolites that cause lucigenin chemiluminescence. The observation that heat-killed cells show

120

g 100

;

80 8 60 5 g 40 2 20 u 0

0

50

100

150

200

250

300

350

400

450

500

Time (s) FIG. 2.

Lucigenin chemiluminescent response of viable (0) and heat-killed (0) E. co& to addition of 1 mM paraquat. Paraquat was injected during the fifth counting period. Each point represents three averaged samples, with error bars representing SE.

showed no CL response to the addition of 1 mM paraquat (Fig. 2). The addition of 0.1 mM menadione to wild-type E. coli, in the presence of CuC&, showed a lucigenin-dependent CL increase of more than 50% of baseline levels within 475 s (data not shown). As with paraquat-treated samples, heat-killed cells showed no increase in CL on addition of 0.1 mM menadione. To investigate the role of O2 metabolism in the lucigenin CL response of wild-type E. coli to oxygen we examined the CL response of cells grown under fermentative conditions on LB xylose medium. Under these conditions the levels of many enzymes required for aerobic respiration and electron transport are decreased (14). We found that cells grown under these conditions showed a lucigenin CL increase to 0, addition far less than that of cells grown aerobically in the same medium (Fig. 3). Aerobically grown cells showed peak counts over 200% higher than those of baseline after 90 s, a response significantly greater than the CL increase of less than 50% of baseline shown by anaerobically grown cells (Fig. 3). In addition to examining the CL response of wild-type cells to Oz, we employed a respiratory deficient mutant E. coli strain in this study. The respiratory chain of E. coli branches to two terminal oxidases, cytochrome o and cytochrome d, the products of the cyo and cyd genes, respectively (15). A strain that was deficient in both oxidases (RG99) showed a lucigenin-mediated chemiluminescent response to oxygen no greater than that of anaerobically grown wild-type cells and far less than that of wild-type cells grown aerobically (data not shown). No suppression of the lucigenin-mediated chemiluminescent response to O2 was measured on extracellular addition of either SOD or catalase. The activity of SOD in the absence of cells was confirmed by the complete suppression of the lucigenin CL detection of 0; pro-

.=I

K 50 c $ 40 i

30 a !j 20 e c2 10 d 0

0

10

20

30

40

50

60

70

80

90

100

Time (s) FIG. 3.

Lucigenin chemiluminescent response of E. coli grown aerobically (@) and anaerobically under fermentative conditions (O), and uninoculated GTB (m). Oxygen wasintroduced into each aerobically maintained sample during the 10th s. Each point represents three averaged samples, with error bars representing SE.

OXYGEN

METABOLITE

baseline CL levels no higher than those of GTB and lack a lucigenin CL response to the addition of 02, paraquat, or menadione also supports this conclusion, as cells not respiring do not produce oxygen metabolites. Furthermore, cell samples into which a stream of air (20% 0,) is introduced would not be expected to experience an increase in oxygen metabolite production as great as the increase caused by the introduction of pure OZ. It was therefore expected that pure oxygen would cause a lucigenin CL increase in E. coli far greater than that of air (Fig. 1). We see additional evidence that it is the intracellular production of O2 metabolites that is being detected by lucigenin CL in the inability of extracellularly added SOD or catalase to suppress the CL response of E. coli to OZ. Suppression of the CL response by these oxygen metabolite scavengers would indicate the participation of extracellular oxygen metabolites in the response for it is unlikely that these enzymes are able to cross the E. coli cell envelope. This suggests that the lucigenin probe reacts with intracellular oxygen metabolites to emit light. That we observe no suppression by SOD is not surprising as it has been reported that superoxide is not able to cross the E. coli cell envelope in either direction (20) and would remain within the cell where it is produced. Therefore it may be that lucigenin-dependent CL represents cellular production of 0, as it does in human polymorphonuclear leukocytes (18) and rat liver microsomes (21). However, there is no definite evidence that the lucigenin probe is specific for a particular O2 metabolite species in E. coli. It has been reported that the levels of many enzymes associated with aerobic metabolism are decreased during anaerobic growth in E. coli (14,22,23). A decrease in cellular redox components would cause a decrease in the ability of E. coli to metabolize O2 and presumably a decrease in the production of oxygen metabolites on 0, exposure. We therefore examined the lucigenin CL response to oxygen of cells grown anaerobically under fermentative conditions, noting that in the short period before and during the CL assay that the cells would be exposed to air the induction of transcription could not cause an increase in anaerobically repressed redox components to levels characteristic of aerobic growth. That anaerobically grown E. coli showed a lucigenin CL response to O2 far lower than that of E. coli grown aerobically (Fig. 3) supported our conclusion that it is the production of oxygen metabolites that causes lucigenin CL in E. coli. The repressed lucigenin CL response to O2 of the respiratory deficient cyo cyd E. coEi strain (RG99) also supported this conclusion. The use of the chemiluminescent probe lucigenin with bacterial cells is a promising new method for measuring

PRODUCTION

IN

319

E. COLI

reactive oxygen metabolite production in bacteria. It offers the ability to measure continuous production on a time scale not previously possible. ACKNOWLEDGMENT We thank

M. S. Cohen

for helpful

comments.

REFERENCES P. M., and Smith, G. (1972) Anasthesiology 37,210-225. 1. Winter, 2. Brunori, M., and Radilio, G. (1984) in Methods in Enzymology (Packer, L., Ed.), Vol. 105, pp. 22-35, Academic Press, San Diego. 3. Hassan, H. M., and Moody, C. S. (1984) in Methods in Enzymology (Packer, L., Ed.), Vol. 105, pp. 254-263, Academic Press, San Diego. 4. Southorn, P. A., and Powis, G. (1988) Mayo Clin. Proc. 63, 390408. 5. Shioi, J., Dang, C. V., and Taylor, B. L. (1987) J. Bacterial. 169, 3118-3123. 6. Allen, R. C. (1986) in Methods in Enzymology (DeLuca, M., and McElroy, W. D., Eds.), Vol. 133. pp. 449-493, Academic Press, San Diego. 7. Campbell, A. K., Hallett, M. B., and Weeks, I. (1985) Methods Biochem. Anal. 31,317-416. 8. Stevens, P., and Hong, D. (1984) Micro&em. J. 30,135-146. 9. Rao, P. S., Luber, J. M., Millinowicz, J., Lalezari, P., and Mueller, H. S. (1988) Biochem. Biophys. Res. Commun. 150,39-44. 10. Au, D. C. T., Lorence, R. M., and Gennis, R. B. (1985) J. Bacterial.

161,123-127. 11. Carlton, B. C., and Brown, B. J. (1981) in Manual of Methods for General Bacteriology (Gerhardt, P., Ed.), p. 239, American Society for Microbiology, Washington, DC. 12. McCord, J. M., and Fridovich, I. (1969) J. Biol. C&m. 244,60496055. 13. Greenberg, J. T., and Demple, B. (1989) J. Bacterial. 1'71, 3933-

3939. 14. Lin, E. C. C., and Kuritzkes, D. R. (1987) in Escherichia coli and Salmonella typhimurium: Cellular and Molecular Biology (Neidhardt, R. C., Ingraham, J. L., Low, K. B., Magasanik, B., Schaechter, M., and Umbarger, H. E., Eds.), pp. 201-221, American Society for Microbiology, Washington, DC. 15. Anraku, Y., and Gennis, R. B. (1987) Trends Biochem. Sci. 12, 262-266. 16. Fridovich, I. (1978) Science 201,875-880. 17. Halliwell, B. (1982) Trends Biochem. Sci. 7,270-272. 18. McCormick, J. P., Fischer, J. R., Pachlatko, J. P., and Eisenstark, A. (1976) Science 191,468-469. 19. Hassan, H. M., and Fridovich, I. (1979) Arch. Biochem. Biophys.

196,385-395. 20. Hassan, H. M., and Fridovich, I. (1979) J. Biol. Chem. 254, 10,846-10,852. 21. Ischiropoulos, H., Kumae, T., and Kikkawa, Y. (1989) Biophys. Biochem.Res.Commun. 161,1042-1048. 22. Iuchi, S., and Lin, E. C. C. (1988) Proc. Natl. Acad. Sci. USA 85, 1888-1892. 23. Poole, R. K., and Ingledew, W. J. (1987) in Escherichia coli and Salmonella typhimurium: Cellular and Molecular Biology (Neidhardt, R. C., Ingraham, J. L., Low, K. B., Magasanik, B., Schaechter, M., and Umbarger, H. E., Eds.), pp. 170-200, American Society for Microbiology, Washington, DC.