1073 (1991} 329 335 :o 1991 Els~wmrScience Publishers BV. 0304-4165/91/$03.50 A D O N I S 0304aI6~9100099G

Biochimica et Biophy~ica Acta.

BBAGEN 2345,*

Cellular metabolism of proxyl nitroxides and hydroxylamines M . S e t x j u r c * , S. P e c a r * * , K . C h e n , M . W u a n d H . S w a x t z Utllt,ersltv o f Illinois, College o f 3,tedicine at Urbaaa Champaign Urbana I L (U S.A )

(Received 19 Malch 1990) (Revised m~uscript r~xeivcd 3 October. 1990)

Previous data from model systems indicated that the proxyl nitroxides should be especially resistant to bioredacfion and therefore could be an effective solutinn to this often pt'oblematic characteristic of nitroxides. Therefore, we investigated the rate of reduction by cells and by the usual model system, aseorhote, of four proxyl nitvexides and three reference nitroxides. We found that, while the rate of reduction by ascor~ate of the proxyl nitroxidas was slower than the rate Of a prototypic pyrrolidino nitroxide (PCA), the reverse was tnJe for reduction by cells. We also studied the rate of oxidation of the c0reesponding bydroxylaminas. The rate of nxia~non by cells of the proxyl hydi'oxylamiues was relatively f~at, especially for the most l i F ~ i l i e derivative. These ~e~lts indicate that:. (i) pro>:yl nitroxliles may not be unusually resistant to bioreduefion by functional biological systems: (ii) accurate knowledge o l relative rates of metabolism of nttroxidee and hydroxylamines in cells and tissues will require direct studies in these systems because the tales may not closely parallel those observed in model (chemical) systems: and (iii) proxy] nitmxides show potential value as agents to measure oxygen concentrations by the rates of oxidation of their corresponding byd|roxylaminus.

Tile nitroxide spin labels, which initially were used to study physical systems and model biological systems, increasingly have been applied to living biological systems [1-9]. These applic.ations initially wore as biophysical probes of membrane properties of cells but recently they have been used for increasingly complex studies at the cellular level and, with the development of EPR imaging and in vivo spectroscopic techniques, in tissues and intact animMs [7,8,10,11]. Nitroxidas also have been suggested as possible contrast agents for N M R , especially in conjunction with their metabolism by cells and tissues [4-6,8,9]. All of these uses have led to an increased need for ultroxides with differing properties, especially in regard to their biodistribution and rate of biological reduction. Although a number of different factors affect the rate of reduction, one of the most important is the nature of the ring on which the

• On leave from J. Stefan lnstUUt,Ljublj.n~ Yugoslavia. * * Schoolof pharmacy, E. Kardelj Onlve.r~itv of Ljubljana. Ljubljan~ Yugoslavia. Cortespond©n~'¢: H. Swart~ Department of Pi~ysiologyand BiOphysics. I9O Medical Sci~ces Building,5O6S. Mathe~ Ave., Urban. ]L 61801. U.S.A.

nitroxide may be located. 11 has been suggested that the "proxyl" nitsoxidas (Fig. 1) should be especially resistant to biological reduction [12-15]. It also has been reported recently that lipophilic nitroxides, especially nitroxides based on five-membered rings, may be useful probes for measuring oxygen concentrations by observing the rate of oxidation of the hydroxylamiucs to the paramagnetic nitroxides [6,16,17]. On the basis of the above considerations, we have carried out a systematic study on a set of proxyl nitrotides with different lipid solubilities. We studied reduction of the nitroxides and oxidation of the corresponding hydroxylamines of four proxyl nitroxidas and three additional nitroxides as reference compounds. Studies were carried out vAth a cell line, TB cells, tor which we have extensive data on interactions with nitroxides [9,16-18]. We also measured their partitioning in cells and their rates of reactions with ascorbatc. Although much remains to be discovered, many of the general principles on the metabolism of the nitroxides has been elucidated recently [20,21]. The predominant, perhaps almost exclusive metabolism is a reversible reduction to the corresponding hydroxytamines; both the reduction and the oxidation are principally enzymatic. The metabolism occurs intraceUula.rly. The mechat~isms appear to differ for water soluble and lipopMlic nitroxidas, with the latter undergoing facile

330 enzymatic oxidation of the corresponding hydroxylamines. There is some evidence that different cells have some significant differences in their rates of metabolism of nitroxides, perhaps reflecting different levels of mitochondr;al function. Although not completely conclusive, the b~:;!, of the available evidence indicates that the predominant site of reduction of the nitroxides is in the mitochondria; there is some evidence that the mechanisms of reduction are not the same for water soluble and lipid soluble nitroxides. Although ascorbate usually does not play a major role in the reduction of nitroxides by cells and tissues, it has been used frequently as a convenient model system to investigate the relationships between the structure of nitroxides and their reduction. While in general such data are useful, they are subject to some obvious limitations such as the effect of charge on their reaction rates. More importantly, because the predominant mechanisms of metabolism of nitroxides in cells do not involve aseorbate, there may be 'uneXpected' differences in metabolism in cells if one relies on the data from the experiments with ascorbate. The data in t.he present paper provide a specific indication of this problem, and we conclude that it is essential to include studies with intact functioning cells in attempts to assess structure-metabniism relationships for the nitroxides. Material and Methods Fig. ] shows the structural formulas and alibrevia. tions of the nitroxides used in these experiments, The proxyl nitroxides were synthesized in the Faculty of Science and Technology, School of Pharmacy, E. Kardelj University of Ljubljana, Yuzoslavia, by the procedure described here. The other mtroxides were obtained from Molecular Probes, Junction City, OR. Synthesis o/proxyl nitroxides. Fig. 2 summarizes the general procedure for preparation of the proxyls. A n-2

n:3 n=4 n= S

El Pr Pr Pr Su Pr Hx Pr

.ooo PCA 4rip 2N4 Fig. 1. StI~'turcs of the nitroxJdes studied in this ~port. Their full chemical names are: 2-ethyl-2,5,5-tfimelhylpyrrolidine-l-oxyl(EtPr); 2-propyl-2.5,5-trimethylpyrrolidine.l.oxyl (PrPr); 2-butyl-2,5,5-tdmethylpynolidine-l-oxyl (BuPr); 2-hexyl-2,2,5-tllrnethylpyrrolidine1-oxyl (HxPr); 3-carboxyl-2,2,5,S-tetramethylpyrrolidlne-l-oxyl (PCA); 4-doxyl pentanoic acid (4DP); and 2.4,4-tfimethyl-2-ethyl-3oxazolidinyloxyl(2N4).

~o.3

~_

~A?,~

~. .

Fig. 2. Outline of route of synthesisof proxyl nitroxides. R = -C2H~. C3H~,C4Hg,CdHI3. solution of the Grignard reagent (prepared from 8.1 m M alkylbromide and 7.9 m M magnesium turnings in 30-50 mi dry diethyl ether) was added dropwlse over 15 rain to a continuously stirred solution containing 5.74 mmol (830 mg) of the nitrone 2.5,5-trimethyM-pyrroline-N-oxide in 30 ml of dry diethyl ether. The reaction mixture was stirred for another 15 rain and then 8 ml of a saturated solution of NH4CI in water was added. The organic phase was separated and the water phase was extracted twice with 3 ml of diethyl ether. The oily residue obtained after the diethyl ether was evaporated from the combined diethyl ether washes was dissolved in 30 ml of methanol containing 12 nd of coneealtrated N H 4 O H and 50 mg cupric acetate monohydrate per 1130 ml. The mixture was stirred for 3 h and then evaporated to dryness. 50 ml of diethyl ether was added to the residue and the diethyl ether layer was washed with brine and dried with sodium sulphate. The proxyls were purified on a silica gel column (Merck, KJeselgel 60, 70-230 mesh) using hexane/diethyl ether = 2 : 1 as the dutent. The central part of a yellow-orange zone was collected. The compounds were viscous, volatile yellow-orange fluids, which had a single spot on T L C (Merck. DC-Alufolien, Kieselgel 60 F254, h e x a n e / diethyl ether = 1 : 1) plates and the expected molecular weight. Otc,~,r data are given in Table i. Hydroxylamines. The hydroxylamines were obtained by cellular reduction of the corresponding nitroxides. The nitroxides were added to a suspension of TB cells in N2 at 37°C for the time necessary for the nitroxides to be completely reduced (i.e., loss of detectable EFR spectrum) to the corresponding hydroxylamines (usually

30-60 mill). A~corbate experiment. The rate of reduction of nitroxides by ascorbate was measured in PBS buffer ( p H 7.4, 310 m o s m o l / l ) with final concentrations of 0.5 raM nitroxide and 5 m M aecorbate. For lipid soluble proxyls the nitroxides were dissolved in EtOH and then diluted in PBS. Pilot experiments were carried out with the addition of E U I ' A as a metal chelator. There was no change in the rates of reduction. These results indicated TABLE I

Synthesis of proxyt nitro:tides Compound Yield afler purificalion Rf (hexane/diethyl ether = I : 1)

EtPr

PrPr

auPr

HxPr

31%

29%

29,~

32%

0.29

0.31

0.33

0.38

tha . . . . . unlikel . . . . . . . . . eduction b . . . . . . bate invo,l'ves metala ions. Cell experiments. Passages 9-25 of a mouse thymus bone marrow cell line (TB cells) were prepared as described previously [9,18]. The cells were removed from monolayer cultures by treatment ,'.~lh 0.25% trypsin for 10 rain. The cells (108 eclls/ml) were resuspcnded in McCoy's medium with fetal calf serum and incubated with 0.1 m M nitroxide. Several control experiments were carried out to determine the role of ascorbat¢ a n d / o r other non-enzymatic mechanisrm of reduction cf nitroxides in TB cells. To determine whether there is significant ascothate in the TB cells, 107 intact or freeze/thawed cells were incubated with 10 units of ascorbate oxidase and the rate of reduction of a nitroxide, "Catt'(4-trimethylamino-2,2,f,6-tetramethylpiperidine-l-oxyl), that is very susceptible to reduction by aseorbate. The results (Table ll) indicate that these cells have no detectable ascorbate. To determine the amount of reduction of proxyl nitroxides that was due to non-enzymatic mechanisms, the initial rate of reduction was measured in cells heated to 78°C for 10 rain. These results (Table llI) indicate that up to 25% of the reduction was not inhibited by heating the cells. To measure the partitioning of nitroxides imo cells, integrals of EPR spectra were obtained with and without a paramagnetlc broadening agent, potassium trioxalatochromiate (CrOx) (60 raM, 310 mosmol/1), which does not cross the cell membrane. The CrOx broadened away the spectra from |fitroxides in the extracellular space. The permeability of the cells to aitroxides also was studied by nicasaring the rates of reduction of nitroxides after damaging the integrity of the membranes by fl',ceze-thawing the ceils three tL-nes. For the EPR measurements the sample.~ were drawn into gas perraeable Teflon tubes (Zeus Industries, Rafttan, NJ) and their EPR spectra taken on a Varian E-109 EPR spectrometer at 37°C with a Varian variable temperature accessory. The perfusing gas used to regulate the temperature was either N2, air, or a combination of oxygen and nitrogen. To measure the oxygen concentra-

TABLEIlL Elicit o/ :,earmg on rea'uc~io, ol n~t,oxides by r s ~#s Imtial rate of reduction (meleeales/cea per rain) in TB eelIsin N2 at 37°C( × 10- 6) PrPr HxPr EtPr BuPr PCA 4DP

7.1_+1.3" 11.2_+0.49 4.9-k0.10 7.3_+3.3 2.5±0.84 26.7~ 0.49

2.5 _+2.2 4.2 _+0.73 LI ±1.13 2O -~0.90 0.S2:k0.39 2.36± [.4

All values are the average of two or thr~ ~ p e n ~ n t s (±S.D.),

tion of the gas, a portion of the gas stream was bubbled into a stirred beaker and the oxygen concentration measured with a Yellow Springs, (YSI) model 54 APB oxygen mctec and standard YSI prohe. Changes in the intensity of the EPR spectra of nltroxldes induced by ascorbat¢ or TB cells were measured by setting the magnetic field of the spectrometer at the peak of the m I ~ 1 line of the nitroxide at zero field sweep and obtaining data as plots of peak height vs. time, f h e time inte4"val bctwecn addition of a nitroxide and the beginning of measurement was less than 2 rnin. To check the initial rate o! reduction, in the first set of experiments the measurements were performed with a stop flow apparatus which was constructed in our laboratory. Since the reduction of the proxyl nitroxides was found to he relatively slow, the stop flow apparatus was not used routinely for most experiments. Results and Discussion In preliminary experiments the toxicity of the nitroxides under the conditions of our experiments was tested in the TB cells by measuring effects on: the ability of the cells to exclude Trypan blue (Tables IV and V), the rate of oxygen consumption, and the total number of intact cells. N o toxic effects on the nitroxides were noted by these measures; this lack of acute

TABLEtl

TAIdLEIV

Role of a~corbaw in reduction of Carl by TB cells

Effecl of different nilroxMes on exclusion of trypan blue by TB cells after 3 h of exposure to the nilroxide

Treatment

Cat~ Catt + ascorbateoxidase Cat1 ~ t i + ascorbateoxidase

Status of cell

intact intact freeze/thawed fn~-'ze/thawed

Initial rate of reducli~ ~n air at 37°C (molecul~/ccB p~r ~ ) x tO 6 2.6 * 2.6 42.1 42.2

• All values are the average of two experiments.

Nitroaid~ (0.1 r a M ) PrPr HxPr EtPr BuPr PCA

i$ of ~l]s ~eludlns dye (10/calls/co) 77 88 s7 86 83

Cab

84

Control

ss

~ of control 91 t03 IO2 tot 98 99

332 TABLEVI

TABLEV

First artier rare constants (k) for reduction of proxy/ nttroxtdes and ~fe~nce nitroxides (0.5 raM) by ~corbate (5 raM) in PBS buffer in air and narogen atmospheres

Effect of confentratio, of nltroxides on exel~ion of trypan blue b), TB re/IJ after 30 rain of incubation ~,i,h the nltroxide

Nitroxide(raM) 4-DP O.l 0.2 2.0 4.0

Pr~t •

% of cellsexcludeto dye

%of control

ss 87 89 87

t~

87

0.2

8~

2.0 4.0

89 89

Control

88

101 99 99 1~ 101 10l

EtPr PrPr

2.3 t0.03) 1.S (0.1)

auPr HxPr

2.4 (0.I)

2.8 (0.2)

1.7 (0.1) 1.7 (0.1l

1.6 (0.2}

PCA 4DP 21q4

5.7 (0.2) 91 (2) 59 ~3)

4.9 (0.l) 89 (2) 59 O)

1.7 (OA)

Values are the mean values of three measurements, standard devialions are in parentheses.

between 0.96-0.99). The rates of reduction of nitroxides in PBS solution by ascorbate are summarized in Table VI for samples in air and N2. In N 2 the rates of reduction were slowest for the proxyls and the pyrrolidine derivative, PCA, and up to 50-tim¢s faster for the doxyl ring nitroxides, 4DP and 2N4. The rates ~f reduction of the proxyl nitroxides were slower when oxygen

effects of nitroxides at these concentrations is consistent with other reports [22-24], Reduction by ascorbate

The EPR signals of the nitroxides decreased by firstorder kinetics (the: correlation between the rates observed experimentally, with an exponential rate, was

20

1

Gaus.

Et pr

ti:" ' ,;: ...... U Fig. 3. EPR spectra of proxyls {0,I mI~ ni suspensions of TB cells(]0s cdls/rM) in McCoy's medium at 37~C. The hypertine splining for peak l is similar to that obse~ed for this nitroxlde in organic media and is cons, d=red to Be due to nilroxides in the ]ipophilic membr~e compartment: peak 2 is a component of the spectrum ~ssoc;~*¢ a~with nltroxides in an aqueous environment.

TABLEVII Distribution of nitroxides {0.1 mA~) between the memhrene~ of TB cel~ (tOg celts/toll and the aqueo~ compartmems (Im/l~) and the corre. sponding partuion coefficienls 1~ and IA are the inlensitlesof the EPR spectra gramthe membranes and from the aqueous compartments, respectively. K I C m / C ~ partition e~enicient, which has been c~hiulated taking into account the volume of the cdl membranes which is ~limated to be 7.70 ~ nam J (frona the tolal phospholipid content of TB cells: 3.2 nag/10s cells (as measured in our laboratory) and taking into account that the ~lume of one phosphollpid is al,-rox. 0.31 nm3). Nitroxide EtPr

VrPr BuPr

HxPr 4DP

2N4 PCA

I~/I A ooo 0.05 1.02 3.70

X

71 1450 5287

o a 0

was present (air);this may be due to the oxidation of hydroxylamines back to nitroxides, which proceeds at a moderate rate in the presence of oxygen. Quallmtive E P R spectra a n d partitioning into membranes o f T B cells The EPR spectra of proxyls in cells are illustrated in Fig. 3. For the ultroxides with longer alkyl chains, BuPr and especially HxPr, a component of the nitroxide in the membrane is resolved (peak 1 in Fig. 3). Table VII indicates the partition coefficients between the membrane and the aqueous compartments for the nitroxides. The partition coefficients were calculated from the EPR spectra by subtracting the aqueous fraction of a spectrum (peak 2 in Fig. 3) from the total spectrum and then carrying out appropriate double integrations t,~,king into account the volume of lipids, as noted in ]'able VII. The partition coefficients of the proxyls, as expected, increased with chain length. The methods used do not resolve small amounts of nitroxide in membranes and, therefore, we could calculate K only if the partition coefficient was substantially greater than 1,

were similar and modestly faster than for PCA ( P < 0.05). The rate of reduction of one of the doxyl nitroxides, 2N4. was significantly faster than the rates of the other nitroxides. The rates of reduction of all seven nitroxides were slower when the cells were perfused with air: this effect is similar to that oo~rved previously with other nitroxides [9,16-17]. In the presence of air, the rates of reduction of the two most lipophilic proxyls were significantly faster than the rates of the two more hydrophilic proxyls. The quamitative effects of the concentration of oxygen on the initial rates of reduction are shown in Fig. 4. (Note that the oxygen concentrations in this figure are those of the perfusing gas and the actual intraeelhilar oxygen concentrations are expected to be lower [9,171.) The relationship between the concentration of oxygen and the initial rates of reduction were similar to those found for other nitroxides [9,17]. The effects of breaking down membrane barriers by three freeze/thaw cycles was as expected for the experiments carried out with perfusion of air, only the nitroxide that has an appreciable charge (PICA) and therefore does not readily cross the cell membranes had an increased ~ate of reduction upon hx~-,.ze/thawing, lit the absence of air, freeze/thawing resulted in a signlfican~ increase in the rate of reduction of the two proxyls that were tested (EtPr and HxPr) as web as P e A ; a similar result has been noted previously for other nitroxides that readily enter cells [18|. R a t e s o f oxidation o f hydroxylamines bY cells The hydroxylamines were produced by redaction of the nitroxides by the ceils in the absence of oxygen [16,17]. Confirmation that the product was the hydroxylamine was provided by the observation that the initial concentration of the nitroxide could he restored by adding a mild oxidizing agent, 2 mM ferri-

i 14 ~la ~7

R a t e s o f reduction in T B cells To obtain relative rates during the period of optimum physiological conditions of ceils, we made measurements in the first few minutes after mixing. The changes in the peak heights of the EPR spectra were essentially linear for the first few minutes and, therefore, the rate constants were calculated as zero order rate constants (all calculations axe for studies with the same initial conditions: 0.1 mM nitroxlde and l 0 s ceUs/ml). O',er longer periods of time the rates of reduction were first order. In ir, tact cells peffused with N 2 the rates of reduction of the four proxyl nltroxides

• • •

to

,

8 8% °° ~ g

.g

o

1

o

2

3

.

4

5

6

7

8

9

lO 11 12

[0 2] [Mg/O Fig. 4. Reduction rate constants of EIPr and HxPr in the TB cells suspension (IO s c¢lls/ml) ~ a r~cdou of oxygen concenua~oo hi the

per~ns gas at 370C.The ~lla ~a~ ~ obt~n,~l by ~eona o,d~ reSression analysis.

A

B

t s m,.ut.. ~ to m J . . t . . o2 Fig. 5. EPR peak heights vs. lime for EtPr (A) and HxPr, (B), initial nilroxJdeconcentration 0.1 mM in TB cell suspensions(10s c¢lls/ml) at 370C. in N2 and after introduction of oxygen by perfusion of air at the times indicated by arrows.

cyanide, to the suspension after the reduction was completed [9,18,19]. When air was introduced after the cells had reduced the nitroxides, the intensities of the EPR spectra of the nitroxides increased (Fig. 5). The rate constants for this process are shown in Table VIlI. Significant oxidation of hydroxy!amines was observed only for the proxyls and PeA, which is consistent with previous studies which indicated that the pyrrolidine ring hydroxylamines can be more readily oxidized than those of the other commonly used nitroxides [16]. The rate of oxidation of HxPr decreased by a factor of four when cells were heated to 65°C for 10 rain prior to the introduction of air at 37°C indicating that the principal oxidation process was enzymatic [9] and that autooxidation of this nitroxide also is a moderately effective process. Previous studies [16] of oxidation of other hydroxylamines by cells also indicate that the observed rapid rates of oxidation of lipophilic hydroxylamines are due to enzymatic reactions, perhaps involving a hydrophobic site on cytochrome c.

The rate of oxidation of the hydroxylamines of the lipophilic hexyl proxyl was especially high and comparable or greater than the rate of reduction of the parent nitroxide, which is consistent with studies on other lipophilic nitroxides [16]. This was reflected in the equifibrium value of nitroxide observed in air for hexyl proxyl (Fig. 5). These resulls indicate that the initial rates of reduction in air of the proxyl nitroxides may have been underestimated because of the competing back reaction of oxidation of the hydroxylamines, although the shape of the kinetic curve did not indicate this effect was very pronounced. Discussion and Conclusions These results indicate clearly that it is unlikely that proxyl nitroxides will have significantly increased resistance to bioreduction compared to other five-membered ring nirroxides. Overall these results do not contain arty ma!'~r s~,~rlses; the qualitative behavior of the proxyl

TABLE Vl[i Reduction rates (zero order) of ntt~xides and oxldation rates of hydroxylamines (0.1 raM) in intact and frozen/th~ed (F/T~ TB cell s~p~mio~ (10s c~lls/m]) in McCoy's methum ( T - 37°C) in air and nitrogen

k ( × 10~ molvcules.¢~ll- t. min - ] ) Nit roaide or reduction (nltroxide) hydroxylamln¢ ~'r intact Etl~r erPr BuPr HxPr

2.1 (0.5) 1,2 (0,7) 4,0(0.7) 3.3 (o.s)

PCA 4DP

1.4 (0.5) 6,7 (0.5)

2N4

13.6(0.7)

oxidation (hydroxylamine)air

N2 F/T 2.0 (0.2)

intact I0 (2) 9 (2)

FfF ~- -(2)

tl (3)

3.6 (0.8)

16 (3)

30 (5)

2.5 (0.4)

6.7 (0.7) 16 (3)

15,0 (0.7)

36 (5)

Values are the avenge of at le~t three experiments (wilh standard devlatlons).

intact 2.1 (0.2) 4,0 (0.4) 7.0 (0.7) 22.0 (2.0) 2,5 (0.5) 0

0

335 nitroxides and the corresponding hydroxylamines is consistent with that of other nitroxides that have been studied under similar conditions. There are, however, at least two significant additional general conclusions that can be drawn from this study. First, these results provide a clear indication that one cannot reach quantitative conclusions on reduction rates of nitroxides in cells, from data on relative rates of reduction by asenrbate. As predicted [15], the proxyls were significantly more resistant than PCA to reduction by ascorbate, but they were reduced by cells at a faster rate than PCA. By the, most simplistic calculation the proxyls were reduced by hypoxic cells about 4 5-times fAStLCthan expected from the rates of reduction found with ascorbate. This result indicates the need to determine more precisely the types of mechanisms of reduction of nitroxides by cells and also to obtain data in the particular cells and tissues of interest on the relative contributions of the various types of mechanisms. The second general finding is that the proxyl nitroxides, especially the lipophilic derivatives, may be good candidates for the measurement of oxygen concentrations by following the rates of conversion of the hydroxylamines to the introaldes. The relative resistance to reduction of the lipophilic proxy|s, combined with the relatively rapid rate of oxidation of the corresponding hydroxylamines, makes them appear to be esp.-.cially suitable agents for this use. The conversion of hydroxylamines can be followed by conventional EPR spectroscopy for cell systems and by EPR imaging and in vivo EPR spectroscopy for tissues and intact animals [7,8,10,11]. It also may be used in conjunction with in vivo N M R techuiqaes [4-6,9] if the proxyl nitroxides prove to be adequate N M R contrast agents (such measurements are in progress). Acknowledgements This work was supported by N I H grants G M 35534, and G M 34250 and used the facilities of the University of Illinois EPR Center which is supported by grant R R 01811 from NIH.

References i Berlincr. LJ. (19761 Spin Labeling Theory and Application. Acadermc Press. New York. 2 Berliner. L.J. 0979) Spin Labeling Theory and Application II. Academic Press. New York. 3 Swartz. H.M. and SwartY~S. (19831in Methods of Biochemical Analysis. (Gbek. D.. ed.). pp. 207-323.John Wiley& Sons. New York. 4 Brasch. R.C, (1983)Radiology147.781-78g. 5 Brasch. R.C.. London. D.A.. Wesbey.O.F-. T~er. T.N.. Nit~ki. D.E.. Williams.R.D.. Doemeny. J.. Tuck, L.D. eund Lallc~md, D.P. (1983) Radiology147."/73-779. 6 Swan.~ H.M.0989) in Advancesin MagneticResonance Imaging (Fcig. E.. ¢d.I. pp. 47-71. Abk:x Publishing. Norcco~d. NJ. 7 gerliner. LJ. and Fujii. H. (19~51Scie~ 227. 5!7-519. 8 Bacic. G.. Demsar. F.. Zolnal. Z. and SwartT~H.M. (1988) J. Mag~ Reson. Meal.Biol. l, 55 65. 0 Swartz. HM.. Chen. K.. Pals. M.. Sentjulc. M. and Morse. P.D.. II. (1986)Magn. Resort. Med. 3.169-174. 10 8aclc. G.. Walczak. T.. Demsat. F. and Sx~anr H.M. (1988l Mash. R ~ n . Med. 8. 209-219. 11 Berliner.LJ.. FujiL H.. Wan. X. and Lukiewicz~S~L (19871Mash. Reson. Med. 4, 380-384. 12 8elkin. S.. Mehlhorr~R.J.. HideR.K.. Hanko~sky.O. and i'acker. L. (19871Arch. Uiochem.Bhiphys.2S6. 232-243. 13 Couet. W.R.. Btasch.R-C.. Somovsky. G.. Lukezo.J.. Prakash. L. Gnewach, C,T. and Tozer.T.N. (1985)Telrahedwn 41,1165-1172. 14 Couel, W,R,, Bra.sch,R.C., Sosnovsky, G. a~d Tozer, T.N, (1985) Mash, Re~on.Imaging3, 83-88. 23 K~'~.a.J.F.W.. Pou, S. and Rosen. GM. ~t987~ MaRn. Reson. Meal.5, 525-536. 16 ChelL K. and S'.varlz. H.M. (19881Biochim. Biophys. Acta 92O, 270-277. t7 Chen, K., Glcckner, J.F., Morse, P.D.. I1 and S~tT~ H.M.(t989l Biochemistry28, 2496-2.501. tg Swartz, H.M,, Senljurc, M. and Morse, P,D,. 11 [19~6l Biochim. Biophys. Acta 888, 82-90. 19 Chen. K. and SwaTtz.H.M. (19891Biochem. Biophys. Acta 992. 131-133. 20 lannone. A.. Hu. H.. Tomasi. A.. Vannini. V. and Swarl~ H.M. (1989) Biochin~BiophygActa 991.90-96, 21 Swarm H.M. (19861Bull. Mash. Reson. 8.172-175. 22 Ank¢l, FG,, Lai, C-S., Hopwood, LE. and Zivkovl¢, Z. (1987) Lite Sc~ qo, 49s-ag. 23 Afza], V., Brasdl, R,C., Nitecki, D. and Wolff, S. (19841Invcg[. Radiol. 1% 549-552. 2a Gordon, D.G,, Brasch, R.C., Ogan, M,D. and Deen, D. (t9881 Invest. RadloL23. 616-620.