Vol. 258, No. 6, issue of March 25, pp. 36284631, 1983
Printed in U.S.A.
Spontaneous Chemiluminescenceof Human Breath
SPECTRUM, LIFETIME, T E M P O R A L DISTRIBUTION, AND CORRELATION WITH PEROXIDE*
(Received for publication, November 10, 1982)
Martin D. Williams+ and Britton Chance
From the Johnson Foundation, 0-501Richards G/4, University of Pennsyluania, Philadelphia, Pennsylvania 19104
Human breath spontaneously emits photons at a rate peroxide and the effects of breathing pure oxygen have been
of approximately 7,00O/liter-s. Theemission has a peak observed, as have the effects of cigarette smoking(12). In the
intheredpart of thespectrumandanultraviolet present study, results of optical spectral, radiative lifetime,
contribution. The emission count rate correlates with peroxide concentration correlation with luminescence inten-
peroxide concentration in a saturating manner under sity, and temporal distribution measurements reported. are
normal breathing conditions. When trapped in a bal-
loon, the breath luminescence count rate has a half- MATERIALS AND METHODS
decay time of approximately min and exhibits more
20 The photon counter employed as detectoran EMI-Gencom (Plain-
than one mode of decay. The photomultiplier pulses view, NY) 9658 A/R photomultiplier with cathode sensitivity of 322
generated by breath luminescence arrive bursts. The UA/L operated at -952kV with a dynode-resistor chain wired for
chemiluminescence process appears by these criteria photon counting. The tube was selected at the factory for low dark
Downloaded from www.jbc.org by guest, on December 1, 2011
to include chain reactions, long-lived emitters, or both.count and high red response. This unit was maintained a t -25 “C by
a thermoelectric cooler type Fact 50 Mark I1 (EMI-Gencom). A fused
silica windowwith an evacuated air space provided a thermal barrier
between the photomultiplier/cooler combination and the breath col-
lector. A polyvinyl chloride cylinder was attached by means of a
CL’ of phagocytosing cells(l), large populations these bayonet lock to the cooler to serve as adaptor for an off-axis parabolic
cells in lungs (Z), the association of peroxides with the CL of polished aluminum breath-light collector (13).The adaptor was bored
these cells (3), and the membrane-damaging properties of to provide two inlet/outlet ports, fitted with 3/8-inch pipe nipples, to
peroxidesandperoxyradicals (4) havebeencombined in exhaust exhaled breath. A mask consisting of double-thick rubberized
recent studies of pulmonary pathology and oxygen toxicity darkroom drapery cloth fitted with Velcro ties was employed to
(5). T o p u tthese results in the context of the lung, studies of maintain darkness around the subject’s mouth and cheeks. This was
perfused rat lung andthe roles of peroxides and edema in CL secured to the adaptor by a wire “drawstring” seated in a channel
milled in the annulus of the adaptor. The room was also darkened.
count rates have also been made ( 6 ) .Concomitant studies of High voltage was provided to the photomultiplier by a Hewlett-
the productionof malondialdehyde as a function of CL inten- Packard (Santa Clara, CA) 1600 A supply. The pulse train from the
sity, and of the CL-optical spectrum have suggested roles of photomultiplier was amplifiedand passed through awindow discrim-
lipidperoxidationandsingletoxygenin the CL of these inator (-2 to -7 mV) (Princeton Applied Research, Model 1121) to
reactions. Products of lipid peroxidation, such pentane, are
as a frequency counter (Hewlett Packard Model 5300/5308-A) and to a
Digital Equipment Corporation PDP 11/10 computer. A histogram
found in breath after chronic alcohol administration in rats
depicting the frequency distribution of photomultiplier counts was
(7) and lipid peroxide radicals as well as other possible emit- constructed for each experiment. The frequency of occurrence of
ters, including singlet oxygen and hydroxyl radical,are suffi- numbers of counts, ranging from 0 to 256, recorded in fixed counting
ciently volatile to be found in breath(8). intervals ranging from 7 to 25 ms, was recorded. The peroxide corre-
Breath is a voluminous and readily accessible waste prod- lation, intensity decay, and optical spectrum experiments each em-
uct. Thoughnot as complex as urine or blood, breath is known ployed25-ms intervals and the temporal distribution experiment
employed 10-ms intervals. The temporal distribution experiment re-
to contain at least 100 complex molecules present at nano-
quired recording of 1,024intervals and display of every fourth interval.
gram/liter quantities Most components breath originate The other experiments employed 10,240 intervals. The temporal
in the atmosphere and the CL of the atmosphere is well distribution experiment employed an EMI-9789 QB photomultiplier.
known (10). H u m a n breath also emitsCLspontaneously, The combination of this tube and associated apparatus had a quantum
without addition of any substance or perturbation. The pos- of
efficiency2.32 X by luminol calibration (14) when the high
sibility of determining pulmonary condition by measurement voltage was -1100 V and the discriminator passed pulses between
of breath luminescence (SBL) is being investigated by
-0.21 and mV.
this experiments-7 The count ratesthus obtained in breathing
were approximately equal to those obtained with the
laboratory. Initial experiments have been done to determine 9658 tube as were dark counts. The dispersion in count rate (noise)
aspects of subjects’ daily lives which produce physiological was 4 times greater with the 9789 (probably due to its IO-year age).
conditions detectable by these measurements. A preliminary The subject inhaled ambient air nasally and exhaled orally into the
spectral study has also been reported (11).The presence of breath/light collector. A typical experiment lasted 5 min and consisted
of approximately 20 full (inhale and exhale) breaths. Counting pro-
ceeded continuously during the experiment so that both inhalation
* This work was supported by Public Health Service Grant HL- and exhalation count rates were recorded except in the temporal
SCOR-15061from the National Heart and Lung Institute. The costs distribution experiment which consisted of a single exhalation.
of publication of this article were defrayed in part by the payment of The correlation of breath luminescence intensity and peroxide
page charges. This article must therefore be hereby marked “aduer- concentration was determined from the results of four series of
tisernent” in accordance with 18 U.S.C.Section 1734 solely to indicate simultaneous measurements of these parameters andseveral control
this fact. experiments described under “Results.” Dr. Takehashi Yonetani
+ Recipient of the Henry M. Chance, I1 Award. To whom corre- kindly provided cytochrome c peroxidase for use in a combined
spondence should be addressed. spectrophotometric assay with cytochrome c (15). The results of a
’ The abbreviations used are: CL, chemiluminescence;SBL, spon- different test for peroxides employingscopoletin have been presented
taneous breath luminescence. (12).
Breath Luminescence 3629
The optical spectrum measurements were performed using a com- In Fig. 2 is shown an optical spectrum of SBL CL. The red
bination of the apparatus employed in the other experiments and the region of the spectrum is especially represented in this and a
apparatus described by Cadenas et al. (1). This consisted of the replicate experiment not shown. This red peak is transmitted
addition of aWratten filter wheel in a light-tight box mounted by a Wrattenfilter 26 but not by filters 25 or 29. The bandpass
between the breath/light collector and the detector. Position of 633-
nm band and transmittance of 633-nm light outside the band were thus defined is centered at 610 nm. Experiments employing
determined using a helium-neon laser. Wratten 2B filters, which cut off the near UV at approximately
RESULTS 140 r
When breath peroxide concentrations are measured from
the same exhalate as SBL count rates thecorrelation of these
two variables is described by a curve indicating a saturation
effect as is shownin Fig. 1.The point at which the count rate
ceases to increase as a function of peroxide concentration
increases is at 2 x lo-’ M. The curve shown in Fig. 1 gives
results that would be obtained if the reaction were bimolecular
in peroxide as in the Hz02 disproportionation reaction de-
scribed by Smith and Kulig (16). The result for the unimo-
lecular case is essentially the same but more gently sloping.
The concentration values are those measured in the peroxide
assay cuvette. To correct them to moles/liter of breath, the
square root must be found and this value multiplied by 3.33
X (to account for the volumedifference between cuvette
and 20 exhalations). This produces results in the range of lo-’
Downloaded from www.jbc.org by guest, on December 1, 2011
M. Various control experiments were performedsuch as meas-
urement of the concentration of peroxide in laboratory air
(- M on the day measured), determining the possibility of
FIG. 2. The red region optical spectrum of SBL obtained
contamination of the assay reaction with ambient peroxide by using a Wratten filterwheel. Each bandwidth represents a 5-min
pumping argon through the breath luminescence collector and count of SBL through the respective filter. Position of the 633-nm
through the assay mixture in the same way as breath (none transmittance and transmittance of 633-nm light out of band were
found), determining the trapping efficiency of the assay by determined with a helium-neon laser. Results indicated here were not
running it in tandem, determining breath volume using both averaged with very similar results obtained in a replicate experiment
a gas flow valve with timed exhalations and a manometric because of the probable real variations in breath spectra. In both
technique to determine exhalation size explicitly. The amount cases,peak wavelength was the same and relative peak intensity
varied less than a factor of 2.
of peroxide lost in apparatus was determined by breathing
single breaths directly into separate aliquots of assay mixture
through a short piece of “Tygon” tubing. 65
&& 2 6
[\4]* I 10‘‘
8 IO 2304
FIG. 1. Correlation of peroxide concentration a n d SBL count 25 4
rate. Ovals indicate the error in both count rate and peroxide 0 IO 20 30
measurement. Filled squares and circles represent two different (minutes)
types of measurement. Count rates have dark count subtracted. FIG. 3. Decay of SBL intensity. Breath trapped in an aluminized
Concentrations are those found in the cuvette and are not corrected Mylar balloon generates photomultiplier counts for at least 30 min.
to breath volume. The peroxide assay mixture described in the The decrease of count rate with time is at least biphasic. Filled circles
methods section was prepared in 10-mlaliquots in 5 - 1 1 polyethylene indicate the decay rate obtained when the balloon is f i s t fiied. Open
graduated cylinders and breath exhausted from the photon counter circles indicate two intervening experiments. The decrease of inten-
bubbled through them during the respective photon-counting meas- sity after vigorous pulmonary exercise is a standard feature of SBL.
urements. Corrections based on control experiments described under Closed squares represent SBL count rate decay obtained when the
“Results” yield breath peroxide concentration in the IO-’ M range. balloon is filled 5 min after smoking a cigarette.
3630 Breath Luminescence
piratory burst accompanying CL and phagocytosis in many
cases (3). Lung microsomes also express hydrogen peroxide
when subject to hyperoxia (21). Hydrogen peroxide is also
liberated by mitochondria under a variety of conditions be-
sides hyperoxic stress (22). If these hydrogen peroxide mole-
cules attack lipids as is suggested by the measurements of
malondialdehyde referred to in the introduction, the resulting
alkyl peroxides and their known atmospheric reactions (8)
0 2 4 6 8 1 0 and CL emission (4) require consideration of their possible
Time (sec) of
roles in SBL. Furthermore, many reactions peroxides pro-
FIG. 4. Temporal distribution of SBL photomultiplier duce hydroxyl radicals and/or singlet oxygen (8, 17). The
counts. The counts generated by a single breath were collected in saturation characteristics of the correlation of peroxide con-
1024 10-ms wide counting intervals. Every fourth intervalwas used in centration and count rate may indicate the operation of a
the representation here. This permits visualization of the fact that branched chainreaction with adark pathway competing with
fewer than 20 of the 256 bins represented had any counts. The total the radiative process under concentration control. It is also
count rate was 12O/s uersus 33 for dark count (not subtracted).
possible that the emitter hasdark mode of reaction as the in
400 nm, resulted in definite reductions of luminescence inten- physical quenching of singlet oxygen (19).
sity. Threemeasurements of each caseyielded 147 k 2.2 The known sources producing atmospheric luminescence
counts/s for the unfiltered count rate versus 125 k 2.3 for the must beconsidered as possible sources of SBL since the
fitered case. When a sheet of cheesecloth is loosely wadded atmosphere constitutes a major portion of exhaled breath.
into a ball and stuffed into the entrance of the light/breath The atmospheric sources have a red emission band except
collector thecountrate is increased. Inoneset of three for nitrogen which emits yellow. These red sources include
experiments each with and without gauze, the average count singlet oxygen (6324 A) (23), hydroxyl radical (6329 A) (24),
Downloaded from www.jbc.org by guest, on December 1, 2011
rates were 27 -t 1.7 without gauze and 147 f 2.2 with gauze. oxygen atom ions (6300 A) (25), and the water cation HzO+
Pumping air through the gauze produced no signal in excess (6140 and 6190 A) (26). The reaction of NO + O:] or NO 0 +
of dark count rate. yields several bands in the red region (27). The presence of a
The decay of SBL count ratewhen breath is trappedin an red peak in the SBL optical emission spectrumthushas
aluminized Mylar balloon is shownin Fig. 3. It shouldbe several possible interpretations. The multitude of possible
noted that the lifetime is very long before as well as after energy transfer acceptors and re-emitters (9) must also be
cigarette smoking. The open circlesrepresent intervening taken into account. The presence of a near-UV component is
series of experiments between the first inflation of the balloon equally nonspecific. Various forms and reactionsof CO, emit
(filled squares) and theballoon filled 5 min after smoking a in this region (28, 29). There are also singlet oxygen (23) and
cigarette (filled circles). The decreaseof luminescence after hydroxyl radical bands (24) in the near UV.
filling the balloon the fist time is a typical response of healthy The alterationof the atmosphere'sluminescence count rate
subjects to vigorous pulmonary exercise. A control experiment by breathing requires consideration of the atmospheric com-
employing USP oxygen to fill the balloon produced a verylow ponents which are known to react with biological molecules
at substantialrates. Singlet oxygen ispresent at M (30).
When SBL counts are collected from a single breath and Ozone is present in air at lo-' to IO-" M concentrations, H202
plotted versus time, as shown in Fig. 4, it may be seen that at lo-" to 10"" M, OH. at to M and H 0 2 . at lo-"'
the luminescence is produced in bursts. Countswere collected to 10"' M (31). Singlet oxygen and hydroxyl radicalare formed
continuouslyfora single exhalation. The data string was in the excited state and may therefore luminesce spontane-
divided into 1024 10-ms intervals and every fourth segment ously whereas the otherbioreactive atmospheric species serve
as substrates in reactions leading to excited state products.
employed in the result. Use of every fourth interval instead of
Ozone reacts with linoleate and other unsaturatedmolecules
all intervals facilitatesvisualization of the result. Fewer than
yielding trioxides which decompose to aldehydes and peroxy
20 of the 256 illustrated segments have any counts in them.
radicals (32,33).Hydrogen peroxide (or HOO .) oxidizes lipids
The maximum counts/interval was 20. The total count rate/
to their hydroperoxy products of cleavage at unsaturated
s was 120 uersus 33 for the dark count.
bonds and the reaction is propagated a chain mechanism.
This reaction and thephysiological role of hydrogen peroxide
have recently been reviewed (34). Hydroxyl radical partici-
Peroxides are known to be present in air at M concen- pation in lipid peroxidation appears to proceed via a Fenton
trations even inthe absence industrial or urban waste gases reaction, the exact mechanism of which is currently debated
(17). Blowing ambient air into the breath collector of the (35, 36). In sum, any of these bioreactive atmospheric mole-
apparatus employed here, however, resulted in no signal dis- cules could be substrates in the reactions leading to SBL.
tinguishable fromthe dark count the detector. The
of peroxide Singlet oxygen and hydroxylradical give up their role as
content of breath is altered from that air with respect to its emitters if they serve as substrates.
luminescence properties. This is apparent from the peroxide- Bowen and Lloyd (37) have found CL accompanying the
correlated count rates of SBL. The intensification of this thermal decomposition of several hydroperoxy radicals. Rus-
signal by filtering breath through cheesecloth may be due to sell has found that the self-reaction of secondary peroxy
the removal of part of the liquid phase of breath. Protic radicals yields excited state ketones and ground state triplet
environments quench freeradicals and singlet oxygen, two of O2 or ground state ketones and excited state singlet oxygen
the possible emitters in this system(18, 19). There are many (38). Kellogg (39) also has found that secondary peroxy radi-
different peroxide reactions (20) and they are considered to cals yield CL on decomposition. Several workers have found
be among the most complex, kinetically, of all chemical reac- that cerium oxidation of secondary peroxides yields CL which
tions (8).Hydrogen peroxide is known to be excreted by lung has been attributed to singlet oxygen by various tests includ-
mitochondria under conditions of hyperoxic stress (21) and by ing optical spectra and deuterium oxide enhancement (40,41).
macrophages and other phagocytes as a function of the res- In their study of tertiary butyl peroxide perfused rat lung
Breath Luminescence 3631
luminescence Cadenas et al. (6) have also found peaks in the national Union of Pure and Applied Biophysics, Mexico City
red part of the spectrum. Their analysis suggested that these 12. Williams, M. D., Leigh, J . S., and Chance, B. (1982) Ann. N. Y .
Acad. Sci. 386,478-483
peaks could be interpreted as singlet oxygen emission. They
13. Hinterberger, H., and Winston, R. (1966) Reu. Sci. Znstrum. 37,
have found a similar emission from pulmonary macrophage 1094-1095
when activated (1). Biggley et al. (42), however, did not find 14. Lee, J., and Seliger, H. H. (1972) Photochem. Photobiol. 15, 227-
a red peak in the emission of polymorphonuclear leukocytes. 237
The chemiluminescence of polymorphonuclear leukocytes 15. Yonetani, T., and Ray,G. S. (1965) J. Biol. Chem. 240,4503-4508
phagocytosing opsonized zymosan granules is inhibited by 16. Smith, L. L., and Kulig, M. J. (1976) J . Amer. Chem. Soc. 98,
catalase and superoxide dismutase (41). Ground state oxygen 1027-1029
17. Graedel, T. E. (1979) J. Geophys. Res. 84(C1), 273-286
may serve as a substrate in additions to ethylenic bonds to 18. Faulkner, L. R. (1978) Methods Enzymol. 51, 494-511
form dioxetanes (43). 19. Wilkinson, F. (1978) in Singlet Oxygen (Ranby, B., and Rabek, J .
The correlationof peroxide concentration and photocounts F., eds)pp. 27-35, John Wiley and Sons, New York .
may have an overall character of being bimolecularin peroxide 20. Bors, W., Saran, M., and Czapski, G. (1980) in Biological and
since lo-’ M peroxidemolecules/liter)yields lo7photons. Chemical Aspects of Superoxide and Superoxide Dismutase
The appearanceof a plateau in that plot and the multimodal (Bannister, W. H., and Bannister, J. V., eds) pp. 1-31, Elsevier-
North Holland, New York
decay of the luminescence intensity are both indicative the 21. Turrens, J . F., Freeman, B. A., andCrapo, J . D. (1982) Arch.
operation of more than one process leading to photon emis- Biochem. Biophys. 217,411-421
sion. The very slow decay of SBL-count rates may indica- 22. Boveris, A., Oshino, N., and Chance, B. (1972) Biochem. J . 128,
tive of chain reactions and/or generation long lived species 617-630
such as singlet oxygen which has a lifetime of 0.5 s a t atmos- 23. Khan, A. U., and Kasha, M. (1970) J. Amer. Chem. SOC. 92,3293-
pheric conditions (30). The multiphasicdecay also provides a 3300
24. Becker, K. H., Lippmann, H., andSchurath, U. (1977) Ber.
mechanism for generation of light with a high quantum yield Bunsen-Ges. Phys. Chem. 81, 567-572
process since a chain reaction provides multiple oportunities 25. Link, R., McConnell, J . C., and Shepherd, L. R. (1981) Planet
Downloaded from www.jbc.org by guest, on December 1, 2011
for emission from a limited number of reactants. The appear- Space Sci. 29, 589-594
ance of counts in bursts is consistent with such a radical chain 26. Herzberg, G. (1980) Ann. Geophys. 36, 605
mechanism. The initiatorof these processes inthe pulmonary 27. Fontijn, A,, Golomb, D., and Hodgeson, J . (1973) in Chemilumi-
tissue could arise from the tissue itself or from the atmosphere nescence and Bioluminescence (Cormier, M. J., Hercules, D.
M., and Lee, J., eds) pp.393-424, Plenum, New York
if it is peroxide, hydroperoxide radical, or hydroxyl radical. 28. Stauff, J., Sander, U., and Jaeschke, W. (1973) in Chemilumines-
Ozone initiation originates in the inhaled gas based on the cence and Bioluminescence (Cormier, M. J., Hercules, D. M.,
present lack of evidence for ozone generation by tissue. and Lee, J., eds) pp. 131-141, Plenum, New York
Thedioxetanereaction (43) has been implicated in CL 29. Dixon, R. N. (1963) Discuss. Faraday SOC. 105-112 35,
studies of peroxidase activity (44). Carbonate reactions have 30. Kummler, R. H., Bortner, M. H., and Baurer, T. (1969) Enuiron.
alsobeenidentified in biological CL reactions (45, 46) so Sci. Technol. 248-250
neither of these species may be excluded as possible emitters 31. Chameides, W. L., and Davis, D. D. (1982) Chem. Eng. News 60,
in breath. Therole of singlet oxygen has notbeen established 32. Su, F., Calvert, J. G., and Shaw, J. H. (1980) J . Phys. Chem. 84,
in SBL. An excellent review of attempts to identify this species 239-246
in biological systems has been published by Krinsky (47). 33. Pryor, W.A. (1981) in Oxygen and Oxy-radicals in Chemistry
and Biology (Rodgers, M. A. J., and Powers, E. L., eds) pp.
Acknowledgments-Daisy Ann Rodriguez provided excellent as- 133-139, Academic, New York
sistance with peroxide assays the cytochromec peroxidase for which 34. Chance, B., Sies,H., and Boveris, A. (1979) Physiol. Rev. 59,527-
was generously furnished by Dr. Takehashi Yonetani. The blackout 605
mask was designed and constructed by Jane K. Williams. William 35. Borg, D. C., Schaich, K. M., and Elmore, J. J . (1981) in Oxygen
Pennie and John Sorge provided design and construction assistance and Oxy-radicals (Rodgers, M. A. J., and Powers, E. L., eds)
for the breath-collector adaptor and computer interface, pp. 177-186, Academic, New York
Dr. Scott Eleff provided the computer program. We wish to thank 36. Tien, M., Svingen, B. A,, and Aust, S.D. (1981) in Oxygen and
Dr. J. S . Leigh Jr. for many helpful conversations. Oxy-radicals (Rodgers, M. A. J., and Powers, E. L., eds) pp.
147-152, Academic, New York
REFERENCES 37. Bowen, E. J., and Lloyd, R. A. (1963) Proc. Roy. SOC.Lond. Ser.
1. Cadenas, E., Daniele, R. P., and Chance, B. (1981) FEBS Lett. A 275,465-568
123,225-228 38. Russell, G. A. (1957) J. Amer. Chem. SOC. 3871-3877
2. Dauber, J. H., Rossman, M. D., andDaniele, R. P. (1982) Enuiron. 39. Kellogg, R. E. (1969) J. Amer. Chem. SOC. 5433-5436
Res. 27, 226-236 40. Nakano, M., Takayama, K., Shimizu, Y., Tsuji, Y., Inaba, H., and
3. Papermaster-Bender, G., Whitcomb, M. E., and Sagone,A. L., Jr. Migita, T . (1976) J . Amer. Chem. Soc. 98, 1974-1975
(1980) J . Reticuloendothel. SOC. 129-139
28, 41. Cheson, B. D., Christensen,R. L., Sperling, R., Kohler, B. E., and
4. Howes, R. M., and Steele, R. H. (1971) Res. Com. Chem. Pathol. Babior, B. M. (1976) J. Clin. Inuest. 58, 789-796
Pharmacol. 2, 619-626 42. Biggley, W. H., Seliger, H. H., and Allen, R. C. (1982) in 10th
5. Cadenas, E., Arad, I. D., Boveris, A,, Fisher, A. B., and Chance, Annual Meeting Program and Abstracts of the American
B. (1980) FEBS Lett. 111,413-418 Society of Photochemistry and Photobiology p. 183, American
6. Cadenas, E., Arad, I. D., Fisher, A. B., Boveris, A,, and Chance, Society of Photochemistry and Vancouver,
B. (1980) Biochem. J. 192, 303-309 British Columbia
7. Litov, R. E. D., Gee, D. L., Downey, J. E., and Tappel, A. L. 43. Williams, M. D., and Fox, J. L. (1981) in Oxygen and Oxy-radicals
(1981) Lipids 16, 52-57 in Chemistry and Biology (Rodgers, M. A. J., and Powers, E.
8. Baldwin, A. C. (1981) in Oxygen and Oxy-radicals in Chemistry L., eds) pp. 771-778, Academic, New York
and Biology (Rodgers, M. A. J., and Powers, E. L., eds) pp. 44. Duran, N., DeBaptista, R. C., and Cilento, G. (1979) Photochem.
543-557, Academic, New York Photobiol. 30, 101-110
9. Krotoszynski, B., Gabriel, G., O’Neill, H., and Claudio, M. P. A. 45. Vladimirov, Yu. A., Gavrilov, V. B., Los’ev, G. M., Azizova, 0. A,,
(1977) J . Chromatogr. Sci. 15,239-244 and Olenev, V. I. (1980) Zh. Fiz. Khim. 54, 504-506
10. Wraight, P. C. (1976) Nature (Lond.) 263, 310-311 46. Hodgson, E. K., and Fridovich, I. (1976) Arch. Biochem. Biophys.
11. Williams, M. D., Leigh, J . S., and Chance, B. (1981) Proceedings 172,202-205
of the Seventh International Biophysics Congress and Third 47. Krinsky, N. I. (1979) in Singlet Oxygen (Wasserman, H. H., and
Pan American Biochemistry Congress. p. 104, (abstr.), Inter- Murray, H. W., eds) pp. 597-641, Academic, New York