112
Biochimica et Biophysica Acta. 1074(1991) 112-117 © 1991 ElsevierSciencePublishersB.V.0304-4165/91/$03.50 ADONIS 0304416591001625
Hypoxia-stimulated reduction of doxyl stearic acids in human red blood cells. Role of hemoglobin Maurizio Minetti and Giuseppe Scorza Laboratorio di Biologia Cellulare, Istituto Superiore di SanitY, Roma (Italy)
(Received 15 November1990)
Key words: Nitroxidemetabolism;Red blood cell; Doxyl stearic acid; Free radical; Hemoglobin;Hypoxia Nitroxide free radicals are under active investigation for their potential use as metabolically responsive contrast agents in electron paramagnetic resonance and nuclear magnetic resonance imaging. The metabolism in human red blood cells of lipid-soluble nitroxides, doxyl stearie acids (DSA), has been investigated. W e observed that under normoxia DSA were stable in red blood cells for at least 2 h, but hypoxia stimulated spin label reduction. Complete signal recovery after air or ferricyanide oxidation suggested the formation of hydroxylamine during hypoxia. DSA reduction was found to be dependent upon the position of the nitroxide ring in the fatty acid chain with the reduction rate higher when the - N O ° of the doxyl ring was closer to the fatty acid carboxylic end. The reduction kinetics of DSA with the doxyl ring nearest to the carboxylic end (5DSA) was bifasic. A rapid reduction of about half of the 5DSA was observed in the first hour and, thereafter, a slow reduction process become predominant. The slope of the slow reduction abruptly decreased below 5 p M, thus suggesting a concentration-dependent membrane-cytoplasm translucation of 5DSA. The reducing activity of the red blood cell (RBC) was completely recovered in the cell lysate. Under hypoxia, purified hemoglobin and myoglobin reduced 5DSA and a complete recovery of the signal was obtained after air reoxidation. Globin did not reduce 5DSA, while methemoglobin showed only a small reduction of 5DSA, thus suggesting that ferrous-berne was involved in the hypoxic reduction of DSA. Both DSA localization and the characteristics of intraeellular reductant (hemoglobin) are responsible for the high stability of DSA in the RBC.
Introduction Paramagnetic spin labels have recently been introduced as imaging agents for in vivo electron paramagnetic resonance (EPR) [1-4] and as contrast media for in vivo nuclear magnetic resonance (NMR) imaging [5,6]. Most contrast media function by shortening the relaxation times of tissue water protons [7] and, consequently, its tissue concentration determines the degree of contrast in N M R imaging. In this respect, a preferential uptake of the paramagnetic drug a n d / o r its metabolism to an unparamagnetic species may be of great importance.
Abbreviations: DSA, doxyl stearic acids (5-, 7-, 10- and 12- indicate the position of the doxyl ring from the carboxylic end); RBC, red blood cell; EPR, electron paramagnetic resonance; NMR, nuclear magnetic resonance. Correspondence: M. Mineui, Laboratorio Biologia Cellulare, lstituto Superiore di SanitY,299 Viale Regina Elena. 00161 Roma, Italy.
Nitroxide free radicals are under active investigation as paramagnetic contrast media due to their relative chemical stability at physiologic temperature and pH. Furthermore, the chemical flexibility of these free radicals allows the design of compounds for special diagnostic purposes [6]. Interestingly, nitroxides in several tissues and cell suspensions are reduced to their corresponding unparamagnetic hydroxylamines with a rate depending on oxygen tension [8,9]. These characteristics of nitroxides suggest their potential use as metabolically responsive contrast agents in EPR and N M R i m a ~ n g [101. The understanding of metabolic pathways of nitroxides in the cell is crucial to their use in in vivo studies. The metabolism of nitroxides has been found to depend upon several parameters including their redox potentials, the type of cell or tissue investigated, the oxygen concentration as well as their solubility w,hich is responsible for their intracdlular distribution. The metabolism of lipid-soluble nitroxides such as doxyl stearic acids (DSA) in h u m a n red blood cell
113 (RBC) suspensions has, to our knowledge, not been investigated in spite of their wide utilization in probing ;,;embr:,e dyna:r.2cs. In general, DSA are considered to be more stable in RBCs than in any other cell types. The aim of this work was to investigate parameters that determine DSA metabolism in RBCs. Chen et al. [11] showed that in nucleated mammalian cells, DSA were reduced preferentially in the mitochondria, a result that indirectly explains the enhanced stability of DSA in RBCs. Nevertheless, we observed that in normoxic RBCs, DSA spin labels remained nearly stable for hours, whereas hypoxia accelerated the reduction, a finding that could be important to clarify the mechanisms responsible for nitroxide metabolism. Hemoglobin was the intraceUular component responsible for DSA reduction to the hydroxylamines. Parameters involved in the high stability of DSA in RBC suspension are discussed. Materials and Methods
Chemicals. 5-, 7-, 10- and 12DSA were purchased from Syva (Palo Alto, CA). Myoglobin (equine skeletal muscle), K3Fe(CN)6, and Sodium Hydrosulfite were obtained from Sigma (St. Louis, MO). Ascorbate oxidase was obtained from US Biochemical Corporation (Cleveland, OH). RBC lysate and protein purification. H u m a n RBCs were washed in isotonic phosphate-buffered saline (150 m M NaCI, 5 m M sodium phosphate, pH 7.5) and the buffy coat was accurately removed by aspiration. To prepare RBC lysate, RBCs were washed at 0°C with 75 m M NaCI, 5 m M sodium phosphate buffer (pH 7.4) to decrease the salt concentration without producing hemolysis [12] and then lysed with an equal volume of 5 m M sodium phosphate buffer. RBC lysis was also obtained by freezing and thawing RBCs and water at 1 : 1 cell to volume ratio. A membrane-free cell lysate was obtained by centrifugation for 15 rain at 40000 x g and filtration through 0.2 p m disposable filter holders (Scleicher and Schuell, Dassei, F.R.O.). All the lysis procedures produced a RBC lysate at pH 7.4 and with about half the hemoglobin concentration of intact cells. Hemoglobin purification was performed essentially as described by Winterbourn [13] except that the ionicexchange column was a TSK DEAE-3SW (LKB, Bromma, Sweden). Conversion of oxyhemoglobin to methemoglobin was achieved by incubation of the former with potassium ferricyanide in a 1.5-fold excess over heine, followed by passage through a Sephadex G-25 column. Hemoglobin and methemoglobin concentrations were measured spectrophotometricaily [13]. Globin preparation was performed by cold acidicacetone as described by Antonini and Brunori [14]. Hemoglobin was passed through a Chelex 100 column (Bio-Rad, Richmond, CA) to reduce possible free iron contamination. Phosphate buffers were treated with
Chelex 100 to remove iron contamination [15]. Myoglobin was dissolved in isotonic phosphate buffer and reduced witl'~ sodium hydrosulfite followed by passage through a Sephadex (3-25 column. Determination of ascorbic acid in RBC lysate was performed by HPLC chromatography as described by Patriarca et al. [16]. EPR Measurements. A Bruker ESP 300 equipped with ESP 1600 data system and a Varian E4 spectrometers were used with the following settings: incident microwave power 20 roW; time constant 1.0 s; frequency, 9.4 GHz for ESP 300 and 9.1 GHz for E4; 100 kHz field modulation; scan range, 100 gauss or 8 gauss when only the midfield line was measured; scan time, 0.4 m i n / G ; field set, 3360 gauss for the ESP 300 or 3260 gauss for the E4. All measurements were carried out at 37°C. The gas flow was 100% N 2 or air as indicated. 100 ttL of packed RBC were diluted 1 : 1 with isotonic phosphate buffer containing an appropriate amount of DSA dissolved in ethanol (0.5% final concentration of ethanol). To minimize membrane perturbation and cell lysis induced by spin labels, concentration range of DSA was 0.1-15 p.M. After 1 - 2 min at rocrn temperature cells were centrifuged and the supernatant removed. The spin label was found to be completely associated with the cells. Samples were drawn into a gas-permeable Teflon tube, 0.81 m m internal diameter, 0.05 m m wall thickness (Zeus Industrial Products, Raritan, N J) and inserted into a quartz EPR tube open at each end. To measure the loss of DSA intensity, the area and the peak height of the midfield line were measured. A modulation amplitude of 5 gauss was chosen to minimize line-width changes during the measurements [11]. Under those conditions, differences in the reduction kinetics of DSA observed by plotting peak area or peak intensity as a function of time were within + 5%. The intensity of DSA under hypoxia at time = 0 was rite intensity of the midfield line under normoxic conditions. Results
Hypoxic reduction of doxyl stearic acids in red blood cells Doxyl stearie acids (5-, 7-, 10- and 12 DSA), when added to normoxic RBCs, produced a stable EPR signal. The concentration of 5DSA, monitored as time-dependent changes in the intensity of the EPR spectrum, was reduced only by 20 + 5% after 2 h at 37°C under normoxia. The substitution of sample atmosphere with 100% pure nitrogen strongly increased the signal loss (Fig. 1). As reported by Chen et al. [11], cell-dependent reduction of DSA could be due to the diamagnetic transformation of the nitroxide group to the corresponding hydroxylamine. Both air and ferricyanide reoxidation of nitroxides are considered evidence in favor of the hydroxylamine transformation of nitroxides [8,9]. In RBCs, we obtained complete signal recovery of 5DSA
114 1
0
0
~ 400"
40
300.
-~ e~
200'
100
20
0
A
0
30
60
90
120
50
mlrl Fig. l. Hypoxic reduction of doxyl stearic acids in human red blood cells at 37°C. 12DSA (A), 10DSA (zx),7DSA (e), 5DSA (o). 10 vM DSA was added to normoxic RBCs. After a control spectrum recorded in air, the sample was exposed to 100~ N2 in the EPR cavity. Signal intensity (H%) was calculated as percent of control. Results are representative of five different RBC donors. The equation for the best line of fit for the slow reduction of 5DSA was y=61.2-0.199× (correlation coefficient 0.996).
either after reoxygenation with air or after treatment with 1 m M K3Fe(CN)6. As shown in Fig. 1, the rate of DSA reduction was found to be strongly dependent upon the position of the nitroxidc ring in the fatty acid chain, i.e., reduction rate was higher when the - N O ° of doxyl ring was close to the fatty acid carboxylic end. In RBCs, DSA showed complex reduction kinetics, with the exception of 10DSA which showed a pseudo-linear reduction (Fig. 1). The 7DSA reduction kinetics showed a rapid loss of a b o u t 10~ of the signal in the first 1 0 - 1 5 min and, thereafter, a pseudo-linear reduction was observed (Fig. 1). As shown in Fig. 1 also the reduction process of 5DSA was clearly bifasie. A rapid reduction of a b o u t 5 0 ~ of the signal was observed in the first 60 rain followed b y a slow linear reduction (referred thereafter as the rapid and slow reduction of 5DSA). Interestingly, the slow reduction was found to be dependent upon probe conceutration, whereas the rapid reduction was not. A s shown in Fig. 2, the half-life ( i t / 2 ) of the slow reduction abruptly increased below the critical concentration of 5 vM. In these experiments, the tl/2 of slow reduction was measured by best fit analysis as in Fig. 1 or when the rapid reduction was exhausted (i.e., the control spectrum was taken after 1 h of hypoxic reduction) without appreciable differences.
Hypoxic reduction of doxyl stearic acids in the RBC lysate and identification of intracellular reductant In order to ;_nvestigate RBC component(s) responsible for hypoxic reduction of DSA, the spin label was
0 ~" , /" = 0
• 5
" 1'0 5DSA (/UM)
•
15
Fig. 2. 5DSA half-life in red blood cells under hypoxia as s function of spin label concentration. Half-life (tt/:z) of 5DSA was measured for the fast (4)) and slow (o) reduction processesof 5DSA in RBC from
the equations for the best line fit as shown in Fig. l.
a d d e d to cell lysate or to purified ghosts. R e d u c i n g activity was recovered o n l y in the cell lysate (Fig. 3). The spectra of 5DSA in R B C s a n d in the R B C lysate are shown in Fig. 4. The b r o a d e n i n g of low- a n d highfield p e a k s observed in c o n c e n t r a t e d (Fig. 4b) a n d diluted cell lysate (Fig. 4c) suggests the presence of 5 D S A in a micellar organization. Furthermore, 5 D S A spectrum in the concentrated cell lysate (Fig. 4b) shows a spin label strongly (arrows) a n d w e a k l y i m m o b i l i z e d (represented by the large line w i d t h of all lines). This weakly immobilized c o m p o n e n t of 5 D S A was not observed in the diluted lysate (Fig. 4c) and, therefore, could be due to p r o t e i n - S D S A interactions.
2O
0
o
2'o
4'o
60
rain Fig. 3. Hypoxic reduction of 5-, 7- and 12DSA in ghosts and red
blood cell lysate at 37°C. 10/iM DSA was added to packed ghosts (closed symbol) or to red blood cell lysate (open symbols), 5-, 7-, or 12DSA (Q) were added to ghosts (all the spin labels gave comparable results). 5DSA (o); 7DSA (~.) and 12DSA (n) in the cell lysate. The cell lysate was obtained with a cell buffer ratio of 1 : 1, as described under Materials and Methods. Hemoglobin concentration of cell lysate was 2 mM. H~ as in Fig, I.
115 In lipid bilayers, the ionization of carboxylic group of DSA strongly affects the location of DSA in the lipid bilayer [17] and, therefore, their partition coefficient between m e m b r a n e and solution. In addition the p H may also affect the rate of DSA hypoxic reduction. Due to the buffering capacity of the RBC intracellular components, we observed that the cell lysate obtained with a cell to buffer ratio of 1 : 1 produced a p H , , J u e of 7.40 + 0.08 (n = 5) irrespectively from the pH of the lysis buffer (5 m M phosphate in the 5 - 9 p H range). Even if the cell lysis was performed in water or according to the procedure reported by Wessels and Veerkamp [18] we measured a p H 7.4. Thus, we assumed that, when cells are suspended in phosphate-buffered saline [18], the p H value of cell lysate was c o m p a r a b l e to that found by DSA inside the RBC. C o m p a r i s o n of Figs. 1 and 3 shows that the rate of reduction of 5DSA in the cell lysate, was similar to the fast reduction in RBCs and faster if c o m p a r e d with the slow reduction of RBCs. T o test the effects of p H on the rate of DSA reduction, the RBC t.ysate was diluted five times with a 20 m M p h o s p h a t e buffer at p H 6.1 or at p H 7.4 (resulting p H of cell iysate were 6.4 and 7.4, respectively). The lowering of p H value by one unit did not affect the rate of 5DSA reduction in the cell lysate. This result indicates that p H changes, in a range that strongly affect the DSA m o t i o n [17], d o not affec! the intracellular rate of hypoxic reduction. The reduction kinetics of 5DSA and 7DSA in the cell lysate were virtually superimposable, thus suggesting a similar accessibility to the reductant (Fig. 3). This was
Fig. 4. EPR spectra of 5DSA at 37°C in red blood cells (a), in concentrated ( l : 1) red blood cell lysate (b), and in diluted (l : 5) red blood cell lysate (c). 5DSA concentration was l0 FM. Both cell lysates were at pH ?A. The arrows indicate the strongly immobilized components of 5DSA spectra in concentraWd RBC lysate. EPR spectrometer settings were as described under Materials and Methods with the exception of modulation amplitude that was 1.0 gauss. Spectra were obtained after 16 cumulative scans. The scale of spectrum (b) was twofold that of spectrum (a) and (c).
so 60 i
40 2O
0
20
40
60
rain Fig. 5. Hypoxic reduction of 5DSA in protein solutions at 37°C. Chymotrypsinogen A (A), Myoglobin (e) and Hemoglobin (0). Proteins were dissolved at 0.4 mM in isotonic phosphate-buffered saline pH 7.4. The concentration of SDSA was 10 taM. H% as in Fig. 1.
not the case in the RBC, in which reduction kinetics of 5DSA and 7DSA were quite different (Fig. 1). The h/z of reduction of DSA in RBC lysate (i) was not modified by dialysis (cut-off 10000); (ii) was unaffected by ascorbate oxidase treatment (50 U / m l for 30 min at 25°C completely effaced intracellular ascorbate); (iii) was temperature-dependent (rapidly increased above 20°C), and (iii) was doubled by thermal treatment (100 ° C, 10 min). These d a t a suggest the involvement of an enzymatic process. To identify the cellular component(s) responsible for the hypoxic-dependent reduction of DSA in RBCs, we tested the reducing behavior of purified hemoglobin. W e noted that 0.4 m M h e m o g l o b i n in a p h o s p h a t e buffer at p H 7.4 caused a small reduction of 5 D S A under normoxia (10% after 1 h), but a strong reduction was observed under hypoxic conditions (Fig. 5). The tt/2 of reduction of 5DSA in the h e m o g l o b i n solution was 45 min, a value that is slightly shorter than tP,at measured for a cell lysate with a similar h e m o g l o b i n content (h/2 = 55 rain). The tt/z of reduction of the concentrated cell lysate was 65 min, thus suggesting the presence of an intracellular component(s) that may lessen the rate of 5DSA hypoxic reduction. The involvement of thiol g r o u p s of proteins as possible DSA reducing sites [19] was excluded b y the observation that myoglobin, which lacks - S H groups, also reduced 5DSA u n d e r hypoxic conditions, whereas chymotrypsinogen A ( - S H present) did not (Fig. 5). The reduction of 5 D S A b:,, h e m o g l o b i n solution was most likely due to h y d r o x y l a m i n e transformation, as suggested by the 100~ recovery of the signal intensity after air reoxidation (Fig. 6). T o test the role of home in hypoxic reduction of DSA, we measured the reducing properties of methemoglobin and globin. A s shown in Fig. 6, after 1 h under
116
loo~
= =.-.~
8 0
_ _
_
40
2O o I
0
,
30
.
~'?
•
60 90 rain Fig. 6. Hypoxicreductionand normoxierecoveryof 5DSA at 37"C in hemoglobin,methemoglobinand globin solutions.Hemoglobin(A, zx), methemoglobin(11,ra) and globin (O) were exposed to 100% N2 for 1 h and. thereafter, the gas flow was changed into air (arrow). The 5DSA concentration was 10 #M. Proteins were dissolvedin iso'.onie phosphatebuffered saline (pH 7.4). Hemoglobinand methemoglobin concentration was 0.4 mM and that of globin was 0.1 raM. The methemoglobinsample contained 90% methemoglobin,4% oxyhemoglobin and 6% hemichromes(determinedaccordingto Ref. 13). H% as in Fig. 1.
hypoxia, methemoglobin produced only 23% of 5DSA reduction to hydroxylamine. It is to be noted that methemoglobin samples contained also some oxyhemoglobin (3-7%) that could be responsible for the residual 5DSA reduction. The extraction of heine, produce.; a complete loss of hemoglobin reducing ability. Globin, in fact, reduced 5DSA by only 10% after 1 h of hypoxia (Fig. 6) and this residual reduction may be due to the 4-5% of residual heine present in globin preparations. Discussion
Our data on DSA reduction in hypoxic RBCs support previous studies that the major variables affecting the reduction of nitroxides in call suspensions include (i) the chemical characteristics of the spin label; (ii) the accessibility of nitroxides to intraeeUular reductants; (iii) the cell type and (iv) the environmental conditions with particular emphasis to oxygen concentration [8,9,11,20,21]. Since mitochondria have been found to be the major reducing site of living cells [11], this study was undertaken to explain the unexpected reduction of DSA in hypoxic RBCs. We present evidences that DSA metabolism in RBCs may be due to both DSA cellular distribution and to the characteristics of intracellular reductant.
DSA cellular distribution DSA are lipid soluble, but in nucleated cells can freely and very rapidly diffuse from the plasma mem-
brane to the intracellular membranes [22]. In RBCs, this characteristic may be reflected in a membrane-cytoplasm-membrane translocation of a fraction of the spin label. Our data are consistent with this hypothesis. We found that 12DSA, with its doxyl ring near to the center of the bilayer, was reduced under hypoxia. Although lipid soluble spin labels could also be reduced through a mechanism involving the diffusion of reducing equivalents from the surface into the center of the bilayer [11], we believe that in the RBC the hypothesis of DSA translocation into the cytoplasm is more likely. In fact, free fatty acids do not seem to require a membrane protein to mediate their transfer across a membrane, and studies with an albumin-liposome model system are consistent with DSA diffusion through the water phase to reach a lipid bilayer and :hen cross the lipid bilayer to the opposite aqueous phase [23]. In favor of a possible DSA membrane-cytoplasm transloeation is the reproducible finding in DSA-labeled RBC spectra of a small signal with a characteristic rapid isotropic motion [22,24]. This signal could be due to +,he presence of a small fraction of DSA in the extracellular and intracellular milieu in equilibrium with the probe dispersed in the membrane. Also, the finding that the slow reduction of 5DSA was concentration-dependent seems to be in accordance with a membrane-cytoplasm translocation of DSA in RBC. Our results showed that below 5 pM, the rate of slow reduction of 5DSA decreases abruptly (Fig. 2). It is conceivable that, below this critical concentration, the fraction of 5DSA that can leave the membrane is extremely low and, accordingly, the rate of reduction decreases. By using a completely different approach, Gordon et al. [26] found that spectra of 5DSA-labeled RBCs were indicative of a membrane clustering of the probe above a 5DSA/lipid ratio < 1 : 2250 (i.e., clustering should be detected above 4.3 #M). This 5DSA concentration is very similar to the 5/LM critical concentration found in this study. The clustered DSA could be a fraction of the spin label in micellar structures in equilibrium with the cell cytoplasm where reduction can occur. Moreover, 5DSA is also dispersed into the lipid bilayer with the doxyl ring located at the membrane extraeellular and intracellular surface [25]. Thus, the fast reduction of about 50% of 5DSA signal, may be due to the accessibility of the probe to the intracelhilar reductant (hemoglobin). The other spin labels, with the doxyl ring deeper in the membrane and, therefore, less directly accessible to hemoglobin, seem to be reduced mainly through the postulated membrane translocation process.
lntracelhdar rednctant The identity of the cellular components responsible for the reduction of nitrorddes has been the subject of much debate [9,11,19,21,27-30]. Ascorbic acid and
117 glutathione have been suggested to be possible intracellular reductants of nitroxides. However, our data showed that these reducing agents were not the major reductants of DSA in RBCs. In fact, extensive dialysis as well as ascorbate oxidase treatment of cell lysate did not change the hypoxic reduction of DSA. Conversely purified hemoglobin efficiently reduced DSA under hypoxia. If the high intracellular concentration of hemoglobin is considered, our data indicate that this hemoprotein may he the c o m p o n e n t responsible for DSA hypoxic reduction in RBCs. The finding that heme-containing proteins such as hemoglobin and myoglobin are reductant of DSA was not completely unexpected. In fact, between the enzymatic processes hypothesized to be involved in o x i d a t i o n / r e d u c t i o n of nitroxides, hemoproteins were often suggested [9,11,28,31]. Our data show that the heme group of hemoglobin in the ferrous form seems to be necessary for nitroxide reduction to hydroxylamine. Conversion of h e m o g l o b i n to methemoglobin reduced the hypoxia-induced hyd r o x y l a m i n e transformation of DSA (Fig. 6), thus suggesting that ferric-hemoglobin was unable to reduce DSA. The finding that h e m o g l o b i n may be the intracellular reductant of RBCs may be one a d d i t i o n a l reason for the high stability in vitro of DSA in these cells. In fact, the reducing m e c h a n i s m of RBC not only shows reducing equivalents of D S A less effective than that of other living cells [11], b u t also high levels of oxygen are m a i n t a i n e d in R B C s by the hemoglobin. Due to its high oxygen affinity, h e m o g l o b i n c a n n o t b e rapidly deoxygenated b y lowering oxygen tension [14]. Furthermore, RBCs are characterized by a low oxygen c o n s u m p t i o n even in buffers w i t h o u t an energy source. The latter characteristic m a k e s this cell quite different from rapidly metabolizing nucleated m a m m a l i a n cells that can easily consume the oxygen present in the e n v i r o n m e n t especially if closed s t a n d a r d quartz E P R cuvettes are employed. The stability of D S A in RBCs (11/2 of hours even u n d e r hypoxic conditions instead of minutes as found in other cells) m a y be of potential relevance for in r i v e N M R and E P R i m a g i n g in p r o d u c i n g persistent signals arising essentially from circulating oxygenated RBC [32]. The metabolic responsiveness of DSA to oxygen tension in RBC may be potentially useful to detect in r i v e hypoxic regions. Acknowledgements W e would like to t h a n k Drs. M. Ferrari, and A. T o m a s i for helpful discussion a n d suggestions. W e are also indebted to Drs. T.C. Petrucci, a n d M.T. Santini for critical reading of the manuscript. W e are indebted to Dr. A n t o n i o M e n d i t t o who performed ascorbic acid measurements. This research was partially supported b y C N R contract No. 89.02565.04.
References I Berliner, L.J.. Fujii. H., Wang. X. and Lukiewicz, S.J. (1987) Magn. Reson. Med. 4, 380-384. 2 Bacic. G., Walczak, T. and Swartz, H.M. (1988) Magn. Reson. Med 8, 209-219. 3 Ishida, S.. Kumashiro, H.. Tsuchihashi, N., Ogata, T., One, M., Kamada, H. and Yoshida, E. (1989) Phys. Med. Biol. 34, 13171323. 4 Ferrari, A_ Colacicchi, S., Guahieri, G., Santini. M.T. and Sotgiu, A. (1990) Biochem. Biophys. Res. Commun. 166,168-173. 5 Paajannen, W.G.H, Eriksson, U.G., Revel, D., Terrier, F. and Brasch, R.C. (1987) Acta Radiologica 28, 593-600. 6 Rosen, G.M., Griffeth, LK., Brown, M.A. and Drayer, B.P. (1987) Radiology 163, 239-243. 7 Bacic, G., Niesman, M.R., Bennett, H.F., Magin, R.L. and Swanz, H.M.(1988) Magn. Reson. Meal. 6, 445-458. 8 Swartz, H.M., Chen, K.. Pals. M., Sentjure, M. and Morse I! ~".D (1986) Magn. Reson. Med. 3, 169-174. 9 Chen, K., Glockner, J.F., Morse I1, P.D. and Swartz H.M. ~q~9) Biochemistry 28, 2496-2501. 10 Swartz, H.M. (1987) 3. Chem. Soc. Faraday Trans. 83,191-202. 11 Chen, K., Morse II, P.D. and Swartz, H,M. (19881 Biochim. Biophys. Acta 943, 477-484. 12 Minetti, M., Ceccarini, M. and Di Stasi, A.M.M. (1984) J. Cell. Biochem. 25, 61-72. 13 Winterhourn, C.C. (1985) in Handbook of Methods for Oxygen Radical Research (Greenwald, R.A., ed.), pp. 137-141, CRC Press, Boca Raton, FL. 14 Antonini, E, and Brunori, M. (1971) Hemoglobin and Myoglobin in their Reactions with Ligands, pp. 123-125, North-Holland, Amsterdam. 15 Buetmer, G.R. (1988) J. Biochera. Biophys. Methods 16, 27-40. 16 Patfiarca, M., Menditto, A. and Morisi, G. (1991) J. Liq. Chromatogr., in press. 17 Barran, M.D, and Laggner, P, (1974) Biochim. Biophys, Acta 363, 127-133. 18 Wessels, J.M.C. and Veerkamp, J.H., (1973) Biochim. Biophys. Acta 291,178-189. 19 Perussi, J.R., Tinto, M.H., Nascimento, O.R. ae.d Tabak, M. (1988) Anal. Biochem. 173, 289-295. 20 Swartz, H.M, Semjurc, M. and Morse II, P.D. (1986) Biochim. Biophys. Acta 888, 82-90. 21 Belkin, S., Mehlhorn. R.J., Hideg, K,, Hankovsky, O. and Packer L. (1987) Arch. Biochem. Biophys. 256, 232-243. 22 Nettleton, D.O., Morse 11, P.D., Dobrucki, J.W., Swartz, H.M. and Dodd, N.J.F. (1988) Biochim. Biophys, Acta 944, 315-320. 23 Benga, G. (1988) Moloc. Aspects Med. 10, 201-222. 24 Minetti, M. and Di Staff, A.M.M. (1987) Biochemistry 26, 81338137. 25 Gaffney, B.J. (1974) Methods Enzymol. 328,161-198. 26 Gordon. L.M., Looney, F.D. and Curtain, C.C. (1985) J. Membr. Biol. 84, 81-95. 27 Branca, M., Denurra, T. and Turrini, F, (1988) Free Rad. Biol. Med. 5, 7-11. 28 Couet, W.R.. Eriksson, M.G., Tozer, T.N., Tuck, L.D. Wesbey, G.E.. Nitecki, D. and Brasch, R.C. (1984) Pharmacol. Res, 1, 203-209. 29 lannone, A., Bini, A., Swartz, H.M., Tomasi, A. and Vannini. V. (1989) Biochem. Pharmacol. 38, 2581-2586. 30 Samuni, A., Samuni, A. and Swartz, H.M. (1989) Free Rad. Biol. Med. 6, 179-183. 31 Quintanilha. A.T. and Packer, L. (1977) Proc. Natl. Acad. Sci. USA 74, 570-574. 32 Keana, J.F. and Pou, S. (1985) Physiol. Chem. Phys. Med. NMR 17, 235-240.
Biochimica et Biophysica Acta. 1074(1991) 112-117 © 1991 ElsevierSciencePublishersB.V.0304-4165/91/$03.50 ADONIS 0304416591001625
Hypoxia-stimulated reduction of doxyl stearic acids in human red blood cells. Role of hemoglobin Maurizio Minetti and Giuseppe Scorza Laboratorio di Biologia Cellulare, Istituto Superiore di SanitY, Roma (Italy)
(Received 15 November1990)
Key words: Nitroxidemetabolism;Red blood cell; Doxyl stearic acid; Free radical; Hemoglobin;Hypoxia Nitroxide free radicals are under active investigation for their potential use as metabolically responsive contrast agents in electron paramagnetic resonance and nuclear magnetic resonance imaging. The metabolism in human red blood cells of lipid-soluble nitroxides, doxyl stearie acids (DSA), has been investigated. W e observed that under normoxia DSA were stable in red blood cells for at least 2 h, but hypoxia stimulated spin label reduction. Complete signal recovery after air or ferricyanide oxidation suggested the formation of hydroxylamine during hypoxia. DSA reduction was found to be dependent upon the position of the nitroxide ring in the fatty acid chain with the reduction rate higher when the - N O ° of the doxyl ring was closer to the fatty acid carboxylic end. The reduction kinetics of DSA with the doxyl ring nearest to the carboxylic end (5DSA) was bifasic. A rapid reduction of about half of the 5DSA was observed in the first hour and, thereafter, a slow reduction process become predominant. The slope of the slow reduction abruptly decreased below 5 p M, thus suggesting a concentration-dependent membrane-cytoplasm translucation of 5DSA. The reducing activity of the red blood cell (RBC) was completely recovered in the cell lysate. Under hypoxia, purified hemoglobin and myoglobin reduced 5DSA and a complete recovery of the signal was obtained after air reoxidation. Globin did not reduce 5DSA, while methemoglobin showed only a small reduction of 5DSA, thus suggesting that ferrous-berne was involved in the hypoxic reduction of DSA. Both DSA localization and the characteristics of intraeellular reductant (hemoglobin) are responsible for the high stability of DSA in the RBC.
Introduction Paramagnetic spin labels have recently been introduced as imaging agents for in vivo electron paramagnetic resonance (EPR) [1-4] and as contrast media for in vivo nuclear magnetic resonance (NMR) imaging [5,6]. Most contrast media function by shortening the relaxation times of tissue water protons [7] and, consequently, its tissue concentration determines the degree of contrast in N M R imaging. In this respect, a preferential uptake of the paramagnetic drug a n d / o r its metabolism to an unparamagnetic species may be of great importance.
Abbreviations: DSA, doxyl stearic acids (5-, 7-, 10- and 12- indicate the position of the doxyl ring from the carboxylic end); RBC, red blood cell; EPR, electron paramagnetic resonance; NMR, nuclear magnetic resonance. Correspondence: M. Mineui, Laboratorio Biologia Cellulare, lstituto Superiore di SanitY,299 Viale Regina Elena. 00161 Roma, Italy.
Nitroxide free radicals are under active investigation as paramagnetic contrast media due to their relative chemical stability at physiologic temperature and pH. Furthermore, the chemical flexibility of these free radicals allows the design of compounds for special diagnostic purposes [6]. Interestingly, nitroxides in several tissues and cell suspensions are reduced to their corresponding unparamagnetic hydroxylamines with a rate depending on oxygen tension [8,9]. These characteristics of nitroxides suggest their potential use as metabolically responsive contrast agents in EPR and N M R i m a ~ n g [101. The understanding of metabolic pathways of nitroxides in the cell is crucial to their use in in vivo studies. The metabolism of nitroxides has been found to depend upon several parameters including their redox potentials, the type of cell or tissue investigated, the oxygen concentration as well as their solubility w,hich is responsible for their intracdlular distribution. The metabolism of lipid-soluble nitroxides such as doxyl stearic acids (DSA) in h u m a n red blood cell
113 (RBC) suspensions has, to our knowledge, not been investigated in spite of their wide utilization in probing ;,;embr:,e dyna:r.2cs. In general, DSA are considered to be more stable in RBCs than in any other cell types. The aim of this work was to investigate parameters that determine DSA metabolism in RBCs. Chen et al. [11] showed that in nucleated mammalian cells, DSA were reduced preferentially in the mitochondria, a result that indirectly explains the enhanced stability of DSA in RBCs. Nevertheless, we observed that in normoxic RBCs, DSA spin labels remained nearly stable for hours, whereas hypoxia accelerated the reduction, a finding that could be important to clarify the mechanisms responsible for nitroxide metabolism. Hemoglobin was the intraceUular component responsible for DSA reduction to the hydroxylamines. Parameters involved in the high stability of DSA in RBC suspension are discussed. Materials and Methods
Chemicals. 5-, 7-, 10- and 12DSA were purchased from Syva (Palo Alto, CA). Myoglobin (equine skeletal muscle), K3Fe(CN)6, and Sodium Hydrosulfite were obtained from Sigma (St. Louis, MO). Ascorbate oxidase was obtained from US Biochemical Corporation (Cleveland, OH). RBC lysate and protein purification. H u m a n RBCs were washed in isotonic phosphate-buffered saline (150 m M NaCI, 5 m M sodium phosphate, pH 7.5) and the buffy coat was accurately removed by aspiration. To prepare RBC lysate, RBCs were washed at 0°C with 75 m M NaCI, 5 m M sodium phosphate buffer (pH 7.4) to decrease the salt concentration without producing hemolysis [12] and then lysed with an equal volume of 5 m M sodium phosphate buffer. RBC lysis was also obtained by freezing and thawing RBCs and water at 1 : 1 cell to volume ratio. A membrane-free cell lysate was obtained by centrifugation for 15 rain at 40000 x g and filtration through 0.2 p m disposable filter holders (Scleicher and Schuell, Dassei, F.R.O.). All the lysis procedures produced a RBC lysate at pH 7.4 and with about half the hemoglobin concentration of intact cells. Hemoglobin purification was performed essentially as described by Winterbourn [13] except that the ionicexchange column was a TSK DEAE-3SW (LKB, Bromma, Sweden). Conversion of oxyhemoglobin to methemoglobin was achieved by incubation of the former with potassium ferricyanide in a 1.5-fold excess over heine, followed by passage through a Sephadex G-25 column. Hemoglobin and methemoglobin concentrations were measured spectrophotometricaily [13]. Globin preparation was performed by cold acidicacetone as described by Antonini and Brunori [14]. Hemoglobin was passed through a Chelex 100 column (Bio-Rad, Richmond, CA) to reduce possible free iron contamination. Phosphate buffers were treated with
Chelex 100 to remove iron contamination [15]. Myoglobin was dissolved in isotonic phosphate buffer and reduced witl'~ sodium hydrosulfite followed by passage through a Sephadex (3-25 column. Determination of ascorbic acid in RBC lysate was performed by HPLC chromatography as described by Patriarca et al. [16]. EPR Measurements. A Bruker ESP 300 equipped with ESP 1600 data system and a Varian E4 spectrometers were used with the following settings: incident microwave power 20 roW; time constant 1.0 s; frequency, 9.4 GHz for ESP 300 and 9.1 GHz for E4; 100 kHz field modulation; scan range, 100 gauss or 8 gauss when only the midfield line was measured; scan time, 0.4 m i n / G ; field set, 3360 gauss for the ESP 300 or 3260 gauss for the E4. All measurements were carried out at 37°C. The gas flow was 100% N 2 or air as indicated. 100 ttL of packed RBC were diluted 1 : 1 with isotonic phosphate buffer containing an appropriate amount of DSA dissolved in ethanol (0.5% final concentration of ethanol). To minimize membrane perturbation and cell lysis induced by spin labels, concentration range of DSA was 0.1-15 p.M. After 1 - 2 min at rocrn temperature cells were centrifuged and the supernatant removed. The spin label was found to be completely associated with the cells. Samples were drawn into a gas-permeable Teflon tube, 0.81 m m internal diameter, 0.05 m m wall thickness (Zeus Industrial Products, Raritan, N J) and inserted into a quartz EPR tube open at each end. To measure the loss of DSA intensity, the area and the peak height of the midfield line were measured. A modulation amplitude of 5 gauss was chosen to minimize line-width changes during the measurements [11]. Under those conditions, differences in the reduction kinetics of DSA observed by plotting peak area or peak intensity as a function of time were within + 5%. The intensity of DSA under hypoxia at time = 0 was rite intensity of the midfield line under normoxic conditions. Results
Hypoxic reduction of doxyl stearic acids in red blood cells Doxyl stearie acids (5-, 7-, 10- and 12 DSA), when added to normoxic RBCs, produced a stable EPR signal. The concentration of 5DSA, monitored as time-dependent changes in the intensity of the EPR spectrum, was reduced only by 20 + 5% after 2 h at 37°C under normoxia. The substitution of sample atmosphere with 100% pure nitrogen strongly increased the signal loss (Fig. 1). As reported by Chen et al. [11], cell-dependent reduction of DSA could be due to the diamagnetic transformation of the nitroxide group to the corresponding hydroxylamine. Both air and ferricyanide reoxidation of nitroxides are considered evidence in favor of the hydroxylamine transformation of nitroxides [8,9]. In RBCs, we obtained complete signal recovery of 5DSA
114 1
0
0
~ 400"
40
300.
-~ e~
200'
100
20
0
A
0
30
60
90
120
50
mlrl Fig. l. Hypoxic reduction of doxyl stearic acids in human red blood cells at 37°C. 12DSA (A), 10DSA (zx),7DSA (e), 5DSA (o). 10 vM DSA was added to normoxic RBCs. After a control spectrum recorded in air, the sample was exposed to 100~ N2 in the EPR cavity. Signal intensity (H%) was calculated as percent of control. Results are representative of five different RBC donors. The equation for the best line of fit for the slow reduction of 5DSA was y=61.2-0.199× (correlation coefficient 0.996).
either after reoxygenation with air or after treatment with 1 m M K3Fe(CN)6. As shown in Fig. 1, the rate of DSA reduction was found to be strongly dependent upon the position of the nitroxidc ring in the fatty acid chain, i.e., reduction rate was higher when the - N O ° of doxyl ring was close to the fatty acid carboxylic end. In RBCs, DSA showed complex reduction kinetics, with the exception of 10DSA which showed a pseudo-linear reduction (Fig. 1). The 7DSA reduction kinetics showed a rapid loss of a b o u t 10~ of the signal in the first 1 0 - 1 5 min and, thereafter, a pseudo-linear reduction was observed (Fig. 1). As shown in Fig. 1 also the reduction process of 5DSA was clearly bifasie. A rapid reduction of a b o u t 5 0 ~ of the signal was observed in the first 60 rain followed b y a slow linear reduction (referred thereafter as the rapid and slow reduction of 5DSA). Interestingly, the slow reduction was found to be dependent upon probe conceutration, whereas the rapid reduction was not. A s shown in Fig. 2, the half-life ( i t / 2 ) of the slow reduction abruptly increased below the critical concentration of 5 vM. In these experiments, the tl/2 of slow reduction was measured by best fit analysis as in Fig. 1 or when the rapid reduction was exhausted (i.e., the control spectrum was taken after 1 h of hypoxic reduction) without appreciable differences.
Hypoxic reduction of doxyl stearic acids in the RBC lysate and identification of intracellular reductant In order to ;_nvestigate RBC component(s) responsible for hypoxic reduction of DSA, the spin label was
0 ~" , /" = 0
• 5
" 1'0 5DSA (/UM)
•
15
Fig. 2. 5DSA half-life in red blood cells under hypoxia as s function of spin label concentration. Half-life (tt/:z) of 5DSA was measured for the fast (4)) and slow (o) reduction processesof 5DSA in RBC from
the equations for the best line fit as shown in Fig. l.
a d d e d to cell lysate or to purified ghosts. R e d u c i n g activity was recovered o n l y in the cell lysate (Fig. 3). The spectra of 5DSA in R B C s a n d in the R B C lysate are shown in Fig. 4. The b r o a d e n i n g of low- a n d highfield p e a k s observed in c o n c e n t r a t e d (Fig. 4b) a n d diluted cell lysate (Fig. 4c) suggests the presence of 5 D S A in a micellar organization. Furthermore, 5 D S A spectrum in the concentrated cell lysate (Fig. 4b) shows a spin label strongly (arrows) a n d w e a k l y i m m o b i l i z e d (represented by the large line w i d t h of all lines). This weakly immobilized c o m p o n e n t of 5 D S A was not observed in the diluted lysate (Fig. 4c) and, therefore, could be due to p r o t e i n - S D S A interactions.
2O
0
o
2'o
4'o
60
rain Fig. 3. Hypoxic reduction of 5-, 7- and 12DSA in ghosts and red
blood cell lysate at 37°C. 10/iM DSA was added to packed ghosts (closed symbol) or to red blood cell lysate (open symbols), 5-, 7-, or 12DSA (Q) were added to ghosts (all the spin labels gave comparable results). 5DSA (o); 7DSA (~.) and 12DSA (n) in the cell lysate. The cell lysate was obtained with a cell buffer ratio of 1 : 1, as described under Materials and Methods. Hemoglobin concentration of cell lysate was 2 mM. H~ as in Fig, I.
115 In lipid bilayers, the ionization of carboxylic group of DSA strongly affects the location of DSA in the lipid bilayer [17] and, therefore, their partition coefficient between m e m b r a n e and solution. In addition the p H may also affect the rate of DSA hypoxic reduction. Due to the buffering capacity of the RBC intracellular components, we observed that the cell lysate obtained with a cell to buffer ratio of 1 : 1 produced a p H , , J u e of 7.40 + 0.08 (n = 5) irrespectively from the pH of the lysis buffer (5 m M phosphate in the 5 - 9 p H range). Even if the cell lysis was performed in water or according to the procedure reported by Wessels and Veerkamp [18] we measured a p H 7.4. Thus, we assumed that, when cells are suspended in phosphate-buffered saline [18], the p H value of cell lysate was c o m p a r a b l e to that found by DSA inside the RBC. C o m p a r i s o n of Figs. 1 and 3 shows that the rate of reduction of 5DSA in the cell lysate, was similar to the fast reduction in RBCs and faster if c o m p a r e d with the slow reduction of RBCs. T o test the effects of p H on the rate of DSA reduction, the RBC t.ysate was diluted five times with a 20 m M p h o s p h a t e buffer at p H 6.1 or at p H 7.4 (resulting p H of cell iysate were 6.4 and 7.4, respectively). The lowering of p H value by one unit did not affect the rate of 5DSA reduction in the cell lysate. This result indicates that p H changes, in a range that strongly affect the DSA m o t i o n [17], d o not affec! the intracellular rate of hypoxic reduction. The reduction kinetics of 5DSA and 7DSA in the cell lysate were virtually superimposable, thus suggesting a similar accessibility to the reductant (Fig. 3). This was
Fig. 4. EPR spectra of 5DSA at 37°C in red blood cells (a), in concentrated ( l : 1) red blood cell lysate (b), and in diluted (l : 5) red blood cell lysate (c). 5DSA concentration was l0 FM. Both cell lysates were at pH ?A. The arrows indicate the strongly immobilized components of 5DSA spectra in concentraWd RBC lysate. EPR spectrometer settings were as described under Materials and Methods with the exception of modulation amplitude that was 1.0 gauss. Spectra were obtained after 16 cumulative scans. The scale of spectrum (b) was twofold that of spectrum (a) and (c).
so 60 i
40 2O
0
20
40
60
rain Fig. 5. Hypoxic reduction of 5DSA in protein solutions at 37°C. Chymotrypsinogen A (A), Myoglobin (e) and Hemoglobin (0). Proteins were dissolved at 0.4 mM in isotonic phosphate-buffered saline pH 7.4. The concentration of SDSA was 10 taM. H% as in Fig. 1.
not the case in the RBC, in which reduction kinetics of 5DSA and 7DSA were quite different (Fig. 1). The h/z of reduction of DSA in RBC lysate (i) was not modified by dialysis (cut-off 10000); (ii) was unaffected by ascorbate oxidase treatment (50 U / m l for 30 min at 25°C completely effaced intracellular ascorbate); (iii) was temperature-dependent (rapidly increased above 20°C), and (iii) was doubled by thermal treatment (100 ° C, 10 min). These d a t a suggest the involvement of an enzymatic process. To identify the cellular component(s) responsible for the hypoxic-dependent reduction of DSA in RBCs, we tested the reducing behavior of purified hemoglobin. W e noted that 0.4 m M h e m o g l o b i n in a p h o s p h a t e buffer at p H 7.4 caused a small reduction of 5 D S A under normoxia (10% after 1 h), but a strong reduction was observed under hypoxic conditions (Fig. 5). The tt/2 of reduction of 5DSA in the h e m o g l o b i n solution was 45 min, a value that is slightly shorter than tP,at measured for a cell lysate with a similar h e m o g l o b i n content (h/2 = 55 rain). The tt/z of reduction of the concentrated cell lysate was 65 min, thus suggesting the presence of an intracellular component(s) that may lessen the rate of 5DSA hypoxic reduction. The involvement of thiol g r o u p s of proteins as possible DSA reducing sites [19] was excluded b y the observation that myoglobin, which lacks - S H groups, also reduced 5DSA u n d e r hypoxic conditions, whereas chymotrypsinogen A ( - S H present) did not (Fig. 5). The reduction of 5 D S A b:,, h e m o g l o b i n solution was most likely due to h y d r o x y l a m i n e transformation, as suggested by the 100~ recovery of the signal intensity after air reoxidation (Fig. 6). T o test the role of home in hypoxic reduction of DSA, we measured the reducing properties of methemoglobin and globin. A s shown in Fig. 6, after 1 h under
116
loo~
= =.-.~
8 0
_ _
_
40
2O o I
0
,
30
.
~'?
•
60 90 rain Fig. 6. Hypoxicreductionand normoxierecoveryof 5DSA at 37"C in hemoglobin,methemoglobinand globin solutions.Hemoglobin(A, zx), methemoglobin(11,ra) and globin (O) were exposed to 100% N2 for 1 h and. thereafter, the gas flow was changed into air (arrow). The 5DSA concentration was 10 #M. Proteins were dissolvedin iso'.onie phosphatebuffered saline (pH 7.4). Hemoglobinand methemoglobin concentration was 0.4 mM and that of globin was 0.1 raM. The methemoglobinsample contained 90% methemoglobin,4% oxyhemoglobin and 6% hemichromes(determinedaccordingto Ref. 13). H% as in Fig. 1.
hypoxia, methemoglobin produced only 23% of 5DSA reduction to hydroxylamine. It is to be noted that methemoglobin samples contained also some oxyhemoglobin (3-7%) that could be responsible for the residual 5DSA reduction. The extraction of heine, produce.; a complete loss of hemoglobin reducing ability. Globin, in fact, reduced 5DSA by only 10% after 1 h of hypoxia (Fig. 6) and this residual reduction may be due to the 4-5% of residual heine present in globin preparations. Discussion
Our data on DSA reduction in hypoxic RBCs support previous studies that the major variables affecting the reduction of nitroxides in call suspensions include (i) the chemical characteristics of the spin label; (ii) the accessibility of nitroxides to intraeeUular reductants; (iii) the cell type and (iv) the environmental conditions with particular emphasis to oxygen concentration [8,9,11,20,21]. Since mitochondria have been found to be the major reducing site of living cells [11], this study was undertaken to explain the unexpected reduction of DSA in hypoxic RBCs. We present evidences that DSA metabolism in RBCs may be due to both DSA cellular distribution and to the characteristics of intracellular reductant.
DSA cellular distribution DSA are lipid soluble, but in nucleated cells can freely and very rapidly diffuse from the plasma mem-
brane to the intracellular membranes [22]. In RBCs, this characteristic may be reflected in a membrane-cytoplasm-membrane translocation of a fraction of the spin label. Our data are consistent with this hypothesis. We found that 12DSA, with its doxyl ring near to the center of the bilayer, was reduced under hypoxia. Although lipid soluble spin labels could also be reduced through a mechanism involving the diffusion of reducing equivalents from the surface into the center of the bilayer [11], we believe that in the RBC the hypothesis of DSA translocation into the cytoplasm is more likely. In fact, free fatty acids do not seem to require a membrane protein to mediate their transfer across a membrane, and studies with an albumin-liposome model system are consistent with DSA diffusion through the water phase to reach a lipid bilayer and :hen cross the lipid bilayer to the opposite aqueous phase [23]. In favor of a possible DSA membrane-cytoplasm transloeation is the reproducible finding in DSA-labeled RBC spectra of a small signal with a characteristic rapid isotropic motion [22,24]. This signal could be due to +,he presence of a small fraction of DSA in the extracellular and intracellular milieu in equilibrium with the probe dispersed in the membrane. Also, the finding that the slow reduction of 5DSA was concentration-dependent seems to be in accordance with a membrane-cytoplasm translocation of DSA in RBC. Our results showed that below 5 pM, the rate of slow reduction of 5DSA decreases abruptly (Fig. 2). It is conceivable that, below this critical concentration, the fraction of 5DSA that can leave the membrane is extremely low and, accordingly, the rate of reduction decreases. By using a completely different approach, Gordon et al. [26] found that spectra of 5DSA-labeled RBCs were indicative of a membrane clustering of the probe above a 5DSA/lipid ratio < 1 : 2250 (i.e., clustering should be detected above 4.3 #M). This 5DSA concentration is very similar to the 5/LM critical concentration found in this study. The clustered DSA could be a fraction of the spin label in micellar structures in equilibrium with the cell cytoplasm where reduction can occur. Moreover, 5DSA is also dispersed into the lipid bilayer with the doxyl ring located at the membrane extraeellular and intracellular surface [25]. Thus, the fast reduction of about 50% of 5DSA signal, may be due to the accessibility of the probe to the intracelhilar reductant (hemoglobin). The other spin labels, with the doxyl ring deeper in the membrane and, therefore, less directly accessible to hemoglobin, seem to be reduced mainly through the postulated membrane translocation process.
lntracelhdar rednctant The identity of the cellular components responsible for the reduction of nitrorddes has been the subject of much debate [9,11,19,21,27-30]. Ascorbic acid and
117 glutathione have been suggested to be possible intracellular reductants of nitroxides. However, our data showed that these reducing agents were not the major reductants of DSA in RBCs. In fact, extensive dialysis as well as ascorbate oxidase treatment of cell lysate did not change the hypoxic reduction of DSA. Conversely purified hemoglobin efficiently reduced DSA under hypoxia. If the high intracellular concentration of hemoglobin is considered, our data indicate that this hemoprotein may he the c o m p o n e n t responsible for DSA hypoxic reduction in RBCs. The finding that heme-containing proteins such as hemoglobin and myoglobin are reductant of DSA was not completely unexpected. In fact, between the enzymatic processes hypothesized to be involved in o x i d a t i o n / r e d u c t i o n of nitroxides, hemoproteins were often suggested [9,11,28,31]. Our data show that the heme group of hemoglobin in the ferrous form seems to be necessary for nitroxide reduction to hydroxylamine. Conversion of h e m o g l o b i n to methemoglobin reduced the hypoxia-induced hyd r o x y l a m i n e transformation of DSA (Fig. 6), thus suggesting that ferric-hemoglobin was unable to reduce DSA. The finding that h e m o g l o b i n may be the intracellular reductant of RBCs may be one a d d i t i o n a l reason for the high stability in vitro of DSA in these cells. In fact, the reducing m e c h a n i s m of RBC not only shows reducing equivalents of D S A less effective than that of other living cells [11], b u t also high levels of oxygen are m a i n t a i n e d in R B C s by the hemoglobin. Due to its high oxygen affinity, h e m o g l o b i n c a n n o t b e rapidly deoxygenated b y lowering oxygen tension [14]. Furthermore, RBCs are characterized by a low oxygen c o n s u m p t i o n even in buffers w i t h o u t an energy source. The latter characteristic m a k e s this cell quite different from rapidly metabolizing nucleated m a m m a l i a n cells that can easily consume the oxygen present in the e n v i r o n m e n t especially if closed s t a n d a r d quartz E P R cuvettes are employed. The stability of D S A in RBCs (11/2 of hours even u n d e r hypoxic conditions instead of minutes as found in other cells) m a y be of potential relevance for in r i v e N M R and E P R i m a g i n g in p r o d u c i n g persistent signals arising essentially from circulating oxygenated RBC [32]. The metabolic responsiveness of DSA to oxygen tension in RBC may be potentially useful to detect in r i v e hypoxic regions. Acknowledgements W e would like to t h a n k Drs. M. Ferrari, and A. T o m a s i for helpful discussion a n d suggestions. W e are also indebted to Drs. T.C. Petrucci, a n d M.T. Santini for critical reading of the manuscript. W e are indebted to Dr. A n t o n i o M e n d i t t o who performed ascorbic acid measurements. This research was partially supported b y C N R contract No. 89.02565.04.
References I Berliner, L.J.. Fujii. H., Wang. X. and Lukiewicz, S.J. (1987) Magn. Reson. Med. 4, 380-384. 2 Bacic. G., Walczak, T. and Swartz, H.M. (1988) Magn. Reson. Med 8, 209-219. 3 Ishida, S.. Kumashiro, H.. Tsuchihashi, N., Ogata, T., One, M., Kamada, H. and Yoshida, E. (1989) Phys. Med. Biol. 34, 13171323. 4 Ferrari, A_ Colacicchi, S., Guahieri, G., Santini. M.T. and Sotgiu, A. (1990) Biochem. Biophys. Res. Commun. 166,168-173. 5 Paajannen, W.G.H, Eriksson, U.G., Revel, D., Terrier, F. and Brasch, R.C. (1987) Acta Radiologica 28, 593-600. 6 Rosen, G.M., Griffeth, LK., Brown, M.A. and Drayer, B.P. (1987) Radiology 163, 239-243. 7 Bacic, G., Niesman, M.R., Bennett, H.F., Magin, R.L. and Swanz, H.M.(1988) Magn. Reson. Meal. 6, 445-458. 8 Swartz, H.M., Chen, K.. Pals. M., Sentjure, M. and Morse I! ~".D (1986) Magn. Reson. Med. 3, 169-174. 9 Chen, K., Glockner, J.F., Morse I1, P.D. and Swartz H.M. ~q~9) Biochemistry 28, 2496-2501. 10 Swartz, H.M. (1987) 3. Chem. Soc. Faraday Trans. 83,191-202. 11 Chen, K., Morse II, P.D. and Swartz, H,M. (19881 Biochim. Biophys. Acta 943, 477-484. 12 Minetti, M., Ceccarini, M. and Di Stasi, A.M.M. (1984) J. Cell. Biochem. 25, 61-72. 13 Winterhourn, C.C. (1985) in Handbook of Methods for Oxygen Radical Research (Greenwald, R.A., ed.), pp. 137-141, CRC Press, Boca Raton, FL. 14 Antonini, E, and Brunori, M. (1971) Hemoglobin and Myoglobin in their Reactions with Ligands, pp. 123-125, North-Holland, Amsterdam. 15 Buetmer, G.R. (1988) J. Biochera. Biophys. Methods 16, 27-40. 16 Patfiarca, M., Menditto, A. and Morisi, G. (1991) J. Liq. Chromatogr., in press. 17 Barran, M.D, and Laggner, P, (1974) Biochim. Biophys, Acta 363, 127-133. 18 Wessels, J.M.C. and Veerkamp, J.H., (1973) Biochim. Biophys. Acta 291,178-189. 19 Perussi, J.R., Tinto, M.H., Nascimento, O.R. ae.d Tabak, M. (1988) Anal. Biochem. 173, 289-295. 20 Swartz, H.M, Semjurc, M. and Morse II, P.D. (1986) Biochim. Biophys. Acta 888, 82-90. 21 Belkin, S., Mehlhorn. R.J., Hideg, K,, Hankovsky, O. and Packer L. (1987) Arch. Biochem. Biophys. 256, 232-243. 22 Nettleton, D.O., Morse 11, P.D., Dobrucki, J.W., Swartz, H.M. and Dodd, N.J.F. (1988) Biochim. Biophys, Acta 944, 315-320. 23 Benga, G. (1988) Moloc. Aspects Med. 10, 201-222. 24 Minetti, M. and Di Staff, A.M.M. (1987) Biochemistry 26, 81338137. 25 Gaffney, B.J. (1974) Methods Enzymol. 328,161-198. 26 Gordon. L.M., Looney, F.D. and Curtain, C.C. (1985) J. Membr. Biol. 84, 81-95. 27 Branca, M., Denurra, T. and Turrini, F, (1988) Free Rad. Biol. Med. 5, 7-11. 28 Couet, W.R.. Eriksson, M.G., Tozer, T.N., Tuck, L.D. Wesbey, G.E.. Nitecki, D. and Brasch, R.C. (1984) Pharmacol. Res, 1, 203-209. 29 lannone, A., Bini, A., Swartz, H.M., Tomasi, A. and Vannini. V. (1989) Biochem. Pharmacol. 38, 2581-2586. 30 Samuni, A., Samuni, A. and Swartz, H.M. (1989) Free Rad. Biol. Med. 6, 179-183. 31 Quintanilha. A.T. and Packer, L. (1977) Proc. Natl. Acad. Sci. USA 74, 570-574. 32 Keana, J.F. and Pou, S. (1985) Physiol. Chem. Phys. Med. NMR 17, 235-240.
