h~ the presence of oxygen, ceils can oxidize hydroxylamines, which are the products of the reduction of nitroxides in cells, back to nitroxides. Lipid-soluble hydroxylamines are oxidized much more rapidly than water-soluble ones, and most of this oxidation is inactivated by heat or ~chloroacefic acid, indicating that the principal mechanism is enzyme.linked. The rates of oxidation of some lipophilic hydroxylamines are comparable to the rates of reduction of the corresponding nitroxides. Hydroxylamines formed by reduction of aqueous soluble nitroxides are not oxidized by cells, except for slight oxidation of some pyrrolidine derivatives. The latter is due to autoxidation. The kinetics of oxidation of reduced lipid-soluble nitroxides are all first-order with respect to hydroxylamines, regardless of the position of the nitroxide group along the carbon backbone, indicating that the oxidation occurs within the membrane. The oxidation of hydroxyl- amines in cells is inhibited by cyanide but not by anfimycin A or SIOF-525A. We also deseril:e an effective method to oxidize hydroxylamines and follow this reach'on; the method is based on the use o! perdeuterated [lSN]Tempone.
Introduction
The use of nitroxides in complex biological systems as biophysical tools to study membrane dynamics and as contrast agents for in vivo NMR has led to a need to understand thoroughly the reduction and oxidation of nitroxides in living cells. As shown in this report, the predominant
Abbreviations: PDT, 4-oxo-,,,2,6,6-tetramethyipiperadine-1- oxyl. See Table I for further nomenclature.
Correspondence: H.M. Swartz, Department of Physiology and Biophysics, University of Illinois, 190 Medical Sciences Build- ing, 506 S. Mathews Avenue, Urbane, 1L 61801, U.S.A.
reaction of nitroxides in cells is their reversible reduction to hydroxylamines (Eqn. 1)
Whereas a large number of studies on the re- duction of nitroxides have been published [1-4], there have been only a few studies on oxidation of hydroxylamines in liver microsomes [5-9], and none in cells. Oxidation rates of some hydroxyl° amines, however, can be comparable to rates of reduction of nltroxides and, therefore, the full and
informed use of nitroxides in biological systems requires an understanding of the oxidation of iay- droxylamines.
In addition, there are some specific uses of oxidation that may be quite valuable in experi- mental studies. Reduction of nitroxide spin labels is a frequent difficulty confronted by electron spin resonance (ESR) spectroscopists who wish to study active biological systems such as subcellular frac- tions or living cells. Oxidation of reduced nitrox. ides provides a method to overcome this difficulty in some biological systems. Because oxidation of hydroxylamines to nitroxides depends on the con- centration of oxygen inside the cell (Chen et al., unpublished data), the rate of reoxidation of hy- droxylamines could allow the measurement of lo- cal oxygen concentrations and/or redox metabo- lism by administering hydroxylamines. The rates of conversion of hydroxylamines to nitroxides could be followed by ESR in vitro or in vivo [10] and, using their effects on relaxation times, by in vivo NMR techniques as well [11].
Materials and Methods
Carboxymethyl-2,5,5-trimethylpyrrolidine-l-oxyl and 2.earboxymethyI-2-tridecyl.5,5-dimethylpyr- rolidine-l-oxyl were kind gifts from Professor J. Keana of the University of Oregon. 3-(Isothio- cyanatomethyl)-2,2,5,5-tetramethylpyrrolidine-l-ox- yl was purchased from Sigma (St. Louis, MO). 4-Oxo-2,2,6,6-tetramethylpiperidine-d16-1[1- 15N]oxyl ([15N]PDT) was purchased from MSD lsotop- (St. Lotus, MO). Tempo sulfate was a kind sift from Dr. A. Keith, Pennsylvania State University. 2N4, 2N14 and TN14 were synthesized as previously described [13,14], and purified by flash chromatography in hexane/ethyl acetate (16:1)..All labels gave single spots on sili~a gel plates in hexane/ethyl acetate (16:1). The acronyms, full chemical names, and structures of the nitroxides used in these studies are shown in Table I.
Inorganic chemicals were of analytical grade. SKF-525A was a gift from Smith-Kline French laboratories, Philadelphia, PA. Malonate was purchased from Eastman Kodak Co. (Rochester, NY); rotenone and antimycin A were purchased from Sigma.
Cells Mouse thymus-bone marrow (TB) cells were
obtained from Professor P. Wong of the Univer- sity of Illinois, cloned from a single cell. TB cells were prepared as described previously [4]. Chinese hamster ovary (CHO) cells were obtained from Dr. L. Hopwood, Medical College of Wisconsin, Milwaukee, CHO stock cultures were maintained as monolayers and subcultured three times a week. 24 h before the experiment, CHO cells were trans- ferred to a spinner flask at a density of 10 s cells/mL Alpha-modification of minimum essen- tial medium supplemented w;'th !0%. bovine serum and penicillin/streptomycin was used for mono- layer and suspension cultures. Cells were washed wi, h serum.free medium before use. Hepatoeytes were isolated from large male Sprague-Dawley rats by the collagenase/hyaluronidase perfusion method [12].
Chemicals All nitroxides except those described below were
purchased from Molecular Probes (Junction City, OR) and used without further purification. 2-
Labeling Cells (10 s cells per nd or approx. 4.5 m 8 of
protein per ml) were labeled with lipid-soluble nitroxides as follows. An appropriate aliquot of a lipid-soluble nitroxide in ethanol was pipetted into a 6 × 50 mm glass culture tube and dried to make a uniform film of nitroxide on the side of the tube; 100 ~tl of the cell suspension was then added, vortexed intermittantly for 0.5-2 min, and removed for ESR studies. Samples were labeled with water-soluble nitroxides by adding nitroxide stock solutions to the cell suspension. These pro- cedures did not alter cell viability, as measured by exclusion of Trypan blue.
ESR All spectra were taken on a Varian E-109E
ESR spectrometer. The cavity (Varian E-238 TM) was mounted so the sample was horizontal, to prevent cells from settling out of the sensitive volume of the cavity. Samples were drawn into a gas-permeable Teflon tube (Zeus Industries, Rari- tan, N J) and inserted into a quartz ESR tube open at each end. Experiments were performed at 37°C
at different oxygen concentrations using a Varian gas flow temperature controller. The time-depen- dent changes in the concentration of nitroxides were monitored by setting the magnetic field of the spectrometer at the peak of the midfield line of the ESR spectrum of the nitroxide and turning the field sweep to zero. After the experiments were performed, cell viability was measured by 0.4~ Trypan blue exclusion. Typically 95-99~ of the cells were able to exclude the dye. Rates of reduc- tion and oxidation were calculated from the initial slopes using a Houston HI-PAD digitizing board connected to a Zenith Z-148 computer.
Measurement of concentrations of hydroxylamines The concentrations of hydroxylamines can be
measured by oxidizing them back to the corre- sponding nitroxides by femcyanide [15] or [15N]PDT and following this reaction using ESR. It has been shown previously that relatively low concentrations of ferricyanide can oxidize hydrox- ylamines to nitroxides [16]. Ferricyanide works well for some nitroxides, but it has two limita- tions: (1) ferricyanide does not penetrate mem- branes and therefore, is an inefficient oxidizer of membrane soluble hydroxylamines; (2) even low concentrations of ferricyanide (1-2 raM) may cause significant line.broadening effects on posi- tively charged nitroxides [17]. In order to over- come these difficulties, [1sN]PDT, an isotopic variant of perdenterated Tempone, was employed as an oxidizing agent. [1sN]PDT is membrane-per- meable and can oxidize both lipid- and water- soluble hydroxy!amines by hydrogen atom ex- change mechanisms [18]. The ESR spectrum of [~SN]PDT has two lines and, therefore, the middle line of conventional [14N]nitroxides can be re- solved easily from the spectrum of [lsN]PDT. The broadening caused by spin.spin exchange between nitroxides is negligible under our experimental conditions.
Results and Discussion
Description of the phenomena in ceils Nitroxides can be reduced and hydroxylamines
can be oxidized by cells under different condi- tions, as shown in Fig. 1. When 5-doxyl stearate is added the ESR signal decreases to zero in about
~ i |_.± . LI...ig
Fig. 1. Reduction of 5-doxyl stearate and oxidation of its corresponding hydroxylamine by cells. The sample consisted of 10 7 TB cells with approx. 4 nmol of 5-doxyl stearate in a gas-permeable tube Mid at 370 C. Time.dependent changes in the concentration of the nitroxi& were obtained by setting the recorder at the peak of tl~ middle (m=0) line of the ESR spectrum. The scan time was 30 aria for the reduction while nitrogen was ~-~'fused, and 120 rain for the oxidation while oxygen was perfuscd. The modulation amplitude is 2.0 (3.
initial ESR spectrum i~ also shown.
10 min. If oxygen is then introduced the signal increases, reaching a steady state after about 1 h. At low concentrations of nitroxides the reduction and oxidation are pseado-fn'st-order reactions with respect to nitroxides. The steady-state concentra- tions of nitroxides and hydroxylamines are func- tions of the initial concentrations of nitroxide and the rate c~mtants of reduction and oxidation, which themselves are affected by oxygen con- centration (our unpublished data):
d x= _k~,X_k~=~X dt
+ kB=(x(o) - x ) + k 8=°(X(O) - X) (2~
where X is the concentration of nitroxide as a function of time and X(0) is the initial concentra- tion of nitroxide; and k ~ , k~ m, k~ '~, k~ t°, are enzymatic reduction, chemical reduction, en- ~jmatic oxidation, and autoxidation rate con- stants, respectively. The solution to this equation in the presence of oxygen is:
X(t) -- X ( 0 ) ~ ( 1 -e -(km+ko)') (3)
where kg is the sum of the rate constants for reduction at a specific oxygen concentration and ko is the sum of the rate constants of oxidation at
274
TABLE II
OXIDATION OF REDUCED NITROXIDES BY [15N]PDT
10 nmol nitroxides were reduced by 10 7 TB ceils before adding 50 nmol [lsN]PDT.
that oxygen concentration. The details of this solution ate to be discussed elsewhere (unpub- lislied data).
I ~, .~ 6.0 g ¢: o . -
" ~ 4"° t
.~_ ~ "0 0 • ~ _~ 2,0 o o o%
o ~ 0 , 0 ,
0
o
. , o
J~ ." ~ ,
, ~ ,
20 40 6'0 Temperature (OC)
e o
Fig, 2. Effect of temperature on the rate of oxidation of hydroxylamine by cells. The sample consisting of 10 7 TB cells with 4 nmol of hydroxylamine from 5-doxyl stearate in 100/~1 volume was incubated at the indicated temperature for 15 rain under nitrogen before introducing oxygen and measuring the oxidation rates. The rates for temperatures greater than 40 °C
were measured at 37" C.
The products of reduction in these cells are princi- pally or only hydroxylamines
Table II shows that some nitroxides, after re. duction in ceils, can be completely oxidized back to nitroxides by 0.5 mM [15N]PDT, indicating that the products are the corresponding hydroxyl- amines. There is no evidence for the formation of amines or any other products from these nitrox. ides. So,..e other reduced nitroxides are not com- pletely oxidized by [15NIPDT, due to the redox potentials, nor by ferricyanide, due to the diffi- culties discussed in the Materials and Methods. Preliminary experimental results indicate that in isolated hepatocytes, not all of the metabolic products of 5-doxyl stearate ate oxidizable by ferricyanide or [~sN]PDT. Therefore, there may be products other than hydroxylamines with other ceil lines.
The rapid oxidation of hydroxylamines in these cells is principally an enzyme-mediated process
The plot of oxidation rate as a function of temperature is shown in Fig. 2. The rates for temperatures greater than 40 °C were measured at 37 * C using cells that previously were incubated at the indicated temperature for 15 rain. The rate of oxidation is highest at 37 ° C. The oxidation pro- cess can be largely inactivated by heating to 65 °C (Fig. 2) or by treatment with trichloroaeetic acid, or can be inhibl 'o'~' cyanide (Fig. 3, Table IV). These data ind~ , that the rapid oxidation of
hydroxylamines is an enzymatic or enzyme-media- ted process. The residual low rate of the oxidation by cells that were incubated at 65°C for 15 rain is presumably due to autoxidation.
Oxidation of hydroxylamines occurs in a membrane As shown in Table Ill, lipid-soluble piperidine,
pyrrolidine and oxazolidine hydroxylamines ate oxidized in living cells, but the analogous water- soluble hydroxylamines are not, except for slight oxidation seen for some pyrtolidine nitroxides. The rates of oxidation of the 7N14 and 10-doxyl stearate hydroxylamines are comparable to the rates of reduction of the corresponding nitroxides.
Fig. 3. Effect of cyanide on the rate of oxidation of hydroxyl- amine by cells. After reduction of 5-doxyl stearate was com- pleted in TB ceils under nitrogen, oxygen is introduced with addition of 5 mM NaCN. Other experimental conditions axe
same as in Fig. 1.
275
TABLE Iii
REDUCTION OF NITROXIDES AND OXIDATION OF CORRESPONDING HYDROXYLAMINES IN TB CELLS
The units of the rates are 106 molecules, cell- t. min- 1. The hydroxylamines were produced by reduction of nitroxides by cells, The initial concentration of aqueous sohbte nitroxides was 0.1 mM. The initial amount of lipid-soluble nitro,6des was 10 nnml per 100/~I ex~pt for 5- and 10-doxyl stearates, which was 4 nmol per 100/tt.
Nitroxides Rates of
reduction oxidation
Water-soluble Piperidines
Tempone 44 0.0 Tempol 53 0.0 Tempol 119 0.0 Catt 2.2 - Cat t (after three cycles of freeze-thawing) 64 0.0 Tempo sulfate 71 0.0 Tempo choline 16 0.0
The evidence that only lipid-soluble hydrox- ylamines are subject to enzymatic or enzyme- mediated oxidation suggests that the oxidation of hydroxylamines to nitroxides in living cells occurs in a membrane. Kinetic analysis also supports this condusion. The reduction of 5-doxyl stearate, whose doxyl moiety is dose to the surface of the membrane, is pseudo-first-order with respect to the nitroxide, whereas reduction ":s zero-order for the 10- and 12-doxyl stearate:;, whose doxyl moie- ties are deep in the membrar~e [20]. This suggests that the reduction of doxyl stearates occurs near the surface of a membrane, with the kinetics of reduction of doxyl moieties controlled by the
accessibility to reducing equivalents. Unlike reduction, oxidation of lipid-soluble hydroxyl- amines located at different positions within the membrane are all pseudo-first-order with respect to hydroxylamines, which is consistent with oxida- tion occurring in the membrane,
Pyrrolidine hydroxylamines autoxidize at a low rate The reactions of the pyrrolidine nitroxides PCA,
5-Tempamine and 3.(Isothiocyanatomethyl)- 2,2,5,5-tetramethylpyrrolidine-l-oxyl share the fol- lowing common properties (Table III): (1) rates of reduction are very low; (2) rates of oxidation are unaffected by prior heating or treatment with
276
trichloroacetic acid; (3) the medium from the cell suspensions can oxidize these pyrrolidine nitrox- ides. These data indicate that the oxidation of these water-soluble pyrrolidine hydroxylamines is due to autoxidation. This autoxidation can be- come significant when the nitroxides are reduced slowly or reduction activity is inactivated by heat, by trichloroacetic acid or by removal of the cells.
The major enzyme responsible for oxidation of hydroxylamines in these cells may be cytochrome oxidase
The extent and type of enzymatic participation in the oxidation of amines or hydroxylamines to nitroxides in liver microsomes is controversial. It has been suggested that cytochrome/'-450 [5,6,9], or, alternatively, a mixed-function amine oxidase [7,8], is involved in the oxidation of the reduced water-soluble nitroxide, Tempo. The oxidation of reduced Tempo catalyzed by cytochrome P-450 was so slow that a modulation amplitude as high as 6.25 G was required to observe the weak ESR signal [6]. We found (Table IV), that SKF-525A,
an inhibitor of cytochrome/'-450, has little effect on the oxidation of hydroxylamines, indicating that cytochrome P-450 is not responsible for the oxidation of lipid-soIuble hydroxylamines observed in these cells. The mixed-function amine oxidase also does not appear to be responsible for the oxidation, since its oxidase activity of the erzyme is completely insensitive to cyanide and azide [19] and oxidation of hydroxylamines in cells can be inhibited by either cyanide or azide (Table ,W).
We have shown that oxidation of hydroxyl- amines in cells is an enzymatic or enzyme-media- tc~ pro",,ess. It occurs in the membrane, requires oxygen, and can be inhibited by cyanide (Fig. 3). Our preliminary cell fractionation experiments indicate that the mitochondrial fraction is able to oxidize reduced 5-doxyl stearate with a rate equiv- alent to that of whole cells. As shown in Table IV, either cyanide or azide inhibits the oxidation, whereas antimycin A does not. All of these data are consistent with cytochrome oxidase in mitochondria being the principal enzyme responsi- ble for oxidation of hydroxylamines in TB cells under our experimental conditions.
TABLE IV
EFFECTS OF VARIOUS INHIBITORS ON RATE OF OXIDATION OF A HYDROXYLAMINE IN CELLS
The hydroxylamine was produced by reduction of 5-doxyl stearate by cells. The units of the rates of oxidation are 10 6 molecules, cell- 1. rain - !.
Our results show that oxidation of hydroxyl- amines in the cell systems studies in these experi- ments is due primarily to an enzymatic process and can occur at rates comparable to rates of reduction of nitroxides. The mechanism of the oxidation of hydroxylamines is different from that of reduction of nitroxides. Oxidation of hydroxyl- amines in these cells occurs in the membrane, is oxygen-dependent, and is inhibited by cyanide, whereas reduction of lipid-soluble nitroxides in these cells occurs near the surface of the mem- brane and is not inhibited by cyanide or azide [4,19]. This suggests that cytochrome oxidase in mitochondria is involved in much of the observed oxidation of hydroxylamine.
Acknowledgement
This work was supported by N1H grants GM 35534, and GM 34250 and used the facilities of
277
the University of Illinois ESR Center, which ';s supported by grant RR 01811 from NIH.
