Biochimica et Biophysica Acta, 1116 (1992) 183-191 © 1992 Elsevier Science Publishers B.V. All rights reserved 0304-4165/92/$05.00
183
BBAGEN 23664
Metal-ion-directed site-specificity of hydroxyl radical detection Dan Gelvan, Veronica Moreno, Walter Gassmann, Jack Hegenauer and Paul Saltman Department of Biology, University of California San Diego, La Jolla, CA (USA)
(Received 29 August 1991) (Revised manuscript received 31 December 1991)
Key words: Hydroxyl radical; Site specificity; Free radical; Salicylate; Dihydroxybenzoate; Phenylalanine A wide variety of • OH detectors are in use for determination of biological - OH production. The chemical generation of •OH is site-specific with respect to the metal-binding site, and thus .OH detectors with metal-binding properties may affect the biological damage and bias -OH detection. The present study shows that both salicylate and phenylalanine, added as low molecular weight •OH indicators, decreased Cu(II) binding to erythrocyte ghosts. In a cell-free system, Cu(II) complexed to both salicylate and phenylalanine. Phenylalanine is a stronger Cu(II) chelator than salicylate, both when competing for Cu(II) bound to ghosts and when competing directly with each other. When OH radicals were generated by ascorbate and Cu(II), the amount of • OH detected as dihydroxybenzoates was proportional to the amount of •OH produced. However, when phenylalanine was added to this system, the efficiency of •OH detection by salicylate strongly decreased, concomitant with the transfer of Cu(II) binding from salicylate to the amino acid. This decrease was larger than that predicted by calculations for random competition of the two detectors for •OH. Deoxyribose and mannitol, which do not bind copper appreciably, competed poorly with salicylate for the •OH. Hydroxylation of phenylalanine, on the other hand, was only slightly affected by the presence of salicylate and unaffected by deoxyribose and mannitol. These results suggest that the detection of .OH by low molecular weight "OH indicators was related to the relative affinity of the detectors for the catalyzing metal, and thus partially site-specific. Furthermore, glutamate, which does not contain an aromatic ring but binds Cu(II) with considerable affinity, competed strongly with salicylate for the •OH, indicating that metal-binding properties rather than the presence of an aromatic ring were the cause of the deviation from random competition. The results indicate that .OH indicators with metal-binding properties affect the distribution of catalytic metal ions in a biological system, causing a shift of free radical damage and localizing a site-specific reaction of "OH on these detectors, with a resulting positive bias in the apparent .OH production.
Introduction Free radical d a m a g e in biological systems is largely attributed to • O H generation from hydrogen peroxide. This reaction invariably requires the presence of transition metals capable of one-electron cycling, such as F e ( I I ) - F e ( I I I ) or Cu(I)-Cu(II) (as reviewed in refs. 1, 2). Coordination chemistry dictates that these metals are present in a b o u n d form rather than as free ions. T h e distribution of the metal ions between available macromolecular and, small-molecular-weight ligands and their specific associations is d e t e r m i n e d by individual
Abbreviations: DHBA, dihydroxybenzoate; TBA, 2-thiobarbituric acid; TBARS, thiobarbituric acid reactive substances; Hct, hematocrit. Correspondence (present address): D. Gelvan, Department of Nutrition, Hebrew University - Hadassah Medical School, P.O.B. 1172, Jerusalem 91010, Israel.
concentrations, affinities and speciation with potential ligands. O H radicals are highly reactive and react with most molecules at rates near the limits set by diffusion. Consequently, a high incidence of . O H 'hits' is expected at or near the metal-binding site [3]. This 'sitespecificity' of • O H d a m a g e has b e e n d e m o n s t r a t e d in macromolecules [3-5]. Alternative oxidant species have b e e n suggested to account for a p p a r e n t site-specific reaction mechanisms. T h e p r o p o s e d oxidants include metal ions in the higher oxidation states, such as C u ( I I I ) and Fe(IV), or transitory m e t a l - H 2 0 2 complexes [6-10]. T h e m e t a l - H 2 0 2 intermediate is considered to be p r o d u c e d in all cases, and this complex subsequently decays either by heterolytic cleavage of the H 2 0 2 to p r o d u c e . O H , or by homolytic cleavage which yields the hypervalent metal ions [9]. Thus, the alternative oxidizing species may be p r o d u c e d concomitantly with the • O H [7,9], or in specific cases as the sole oxidants [11], and bring about the oxidative d a m a g e c o m m o n l y ascribed to • O H [10]. Fac-
184 tors such as metal coordination [7,11,12], the presence of oxygen [6] and the availability of reducing and oxidizing species in the system determine which species will be produced and to what extent [7,9,12]. An important feature of the activity of the alternative oxidizing species is their localized reaction at the site of the metal coordination. This makes these species difficult to differentiate from site specifically acting .OH. This p a p e r will discuss the reactions in terms of •OH, while it is recognized that other oxidizing species may be involved. The phenomenon of site-specificity, whether it involves • O H or other coordinated oxidants, vastly complicates the detection and quantitation of free radicals formed in biological systems, since a fraction of the O H radicals reacts locally with the biomolecule and never reaches the free solution where it can react with the detector molecule• Detection is thus blind to the site-specific fraction of the O H radicals which is the biologically important fraction. The use and choice of . O H detectors in biology (reviewed in ref. [13]) is further complicated by considerations of toxicity, metabolism and their recovery or traceability in biological systems. The most suitable detectors that have been introduced for in vivo use arc aromatic compounds that are readily hydroxylated to stable derivatives by . O H and may be recovered and analyzed. The use of the aromatic hydroxylation of salicylate as an indicator of • O H production has a long history (ref. [14] and references therein), but the colorimetric assay of the products is quite insensitive. The quantification of salicylate hydroxylation products by H P L C and electrochemical detection was introduced by Floyd et al., who showed that the conversion of salicylate to 2,3- and 2,5-dihydroxybenzoates ( D H B A ) adequately reflected . O H generation [15]. Salicylate, when used at low concentrations, is well tolerated by cells, organs and whole organisms and seems to be a reliable reporter of • O H formation [16-19]. Halliwell et al. subsequently proposed using the formation of tyrosine isomers from phenylalanine as a superior indicator in biological systems, since phenylalanine is also non-toxic and its D-isomer is not metabolized by most systems [20]. The importance of the choice of an appropriate reporting system has recently been emphasized by the suggestion of Halliwell et at. that metabolic conversion of salicylate to 2,5-DHBA may be mistaken for genuine . O H detection [21]. While the data quoted in support of this contention are incomplete and the contribution of the metabolic conversion may be insignificant and readily corrected for in a controlled study, it serves to stress the complexity of the choice of appropriate . O H detectors for the task at hand. A more serious and often overlooked factor is that indicator molecules may possess metal-binding properties that will cause a redistribution of the metal ions in
the system. A shift in metal-binding will ultimately alter the site-specificity of the free radical attack, thus perturbing the patterns of biological damage. Furthermore, by competing metal ions away from macr~molecules there will be an increase in the fraction of - O H escaping into solution, causing an increase m • O H detection which does not reflect increased production. The binding of metal ions to the . O H detectors may, in turn, result in a site-specific rcaction of • O H with the detector. The effects will combine to produce a positive bias in . O H detection by detectors with metal-binding properties. The possibility of site specific reactions with small molecules has been suggested by Gutteridge, who observed that the mode of metal binding affected the ability of certain -Ott scavengers to protect small molecules from free radical damage [22,23]. The present study investigates metal-binding to various • O H detectors, its effects on - O H measurement and scavenging, and the occurrence of site-specificity in the reaction of •O H with small molecules. We show that some aromatic - O H detectors perturb metal binding in biological systems and cause an augmentation of • O H measurement that is related to site-specificity. Materials and Methods
Materials All materials used were of the highest purity awdlable. Chelex-100 (200-400 mesh) was from Bio-Rad. 2-Thiobarbituric acid (TBA) was crystallized from hot water before use. All other chemicals were used without further purification. Glass-distilled water was used throughout.
Reagents A stock of 0.1 M CuSO~ was prepared in 10 mM H2SO 4. Ascorbic acid, 0.25 M, was prepared fresh in water before each experiment and kept on ice. TBA, 1%, was dissolved i, 60 mM NaOH.
Preparation of erythrocyte ghosts Venous blood samples were collected in heparin from healthy human volunteers, and erythrocyte membrane ghosts were prepared as previously described [24]. The ghosts were suspended in 10 mM Tris buffer (pH 7.5) to a concentration corresponding to 30% hematocrit (Hct).
Metal binding to ghosts A suspension of ghosts corresponding to 5% Hot was mixed with 0.1 mM Cu(II) in 1 mM Tris-20 mM Mops (pH 7.4) on ice. Varying concentrations of salicylate or phenylalanine were added and the suspensions were incubated 10 min. The suspensions were ultrafiltered through Amicon 10 ultrafilters, and the filtrates
185 were assayed for Cu(II) content by the bathocoproine disulfonate method [25], using a Uvikon 930 spectrophotometer (Kontron instruments)• Metal binding to the ghosts was calculated from the difference between the Cu(II) concentration in the original sample and the filtrate.
Metal binding to salicylate and phenylalanine Mixtures containing 0.1 mM Cu(II), 5 mM salicylate and various concentrations of phenylalanine in a 1 mM Tris-20 mM Mops buffer (pH 7.4) were assayed spectrophotometrically. Copper-salicylate binding was calculated from the absorbance at 393 nm, using e = 200 M-1 cm-1 for salicylate bound to copper (experimentally determined in this study; data not shown)•
mobile phase containing 2% methanol (v/v) in an aqueous solution of 30 mM sodium citrate, 27 mM EDTA and 27.7 mM sodium acetate (pH 4.5) at a flow rate of 1.2 ml/min. Tyrosine content was determined by electrochemical detection, using the following settings: guard cell, + 0.40 V; detector 1, + 0.70 V. Residual ascorbate. Samples were diluted 1:1 in 0.8% metaphosphoric acid and 1% dimethyl sulfoxide to terminate the reaction• Residual ascorbate was determined by HPLC. The mobile phase contained: 1% acetic acid and 1.8 mM tetrabutylammonium hydroxide in 20% methanol/80% water (v/v). Data were acquired for 12 rain at a flow of 1 ml/min and ascorbate was detected by ultraviolet absorbance at 245 nm.
TBA assay Reaction systems and competitive •OH scavenging Oxygen-derived free radicals were generated chemically using 5 mM freshly prepared ascorbate and 0.1 mM Cu(I1) in 1 mM Tris-20 mM Mops buffer (pH 7.4). The reactions were initiated by the addition of ascorbate and performed with continuous vigorous shaking at 37°C. Reaction systems contained various combinations and concentrations of salicylate, phenylalanine, deoxyribose and mannitol with the Cu(II)/ascorbate free radical generating system, and were incubated at 37°C until the ascorbate was > 99% depleted• -OH was monitored as the conversion of salicylate to dihydroxybenzoates or of phenylalanine to tyrosine isomers, and by formation of thiobarbituric acid reactive substances (TBARS) from deoxyribose. Reaction products were quantitated by HPLC and by the thiobarbituric acid (TBA) assay.
To determine TBARS formation from deoxyribose, 1 ml samples were added to 0.5 ml 28% trichloroacetic acid-0.1 M NaAsO 2, mixed and treated with Chelex as previously described [24]. 670/~1 of sample were mixed with 330 tzl of 20% trichloroacetic acid, 0.5 ml of 1% TBA and 30/~1 of 10 mM Fe(III). The solutions were heated 12 min at 100°C, cooled on ice, and the absorbance measured against matched reagent blanks at 532 and 600 nm. The 600 nm absorbance was considered to be a nonspecific baseline drift and was subtracted from As32 [26].
Data presentation The results are the mean of at least three separate experiments with a typical coefficient of variation of less than 5%. Results
HPLC detection of dihydroxybenzoates, tyrosines and residual ascorbate Dihydroxybenzoates, tyrosines and residual ascorbate were determined by HPLC on a Rainin Rabbit-HP HPLC system with a Microsorb C18 column (5 lzm spherical bead diameter, 4.6 mm i.d. × 15 cm length), using a Knauer variable wavelength ultraviolet monitor, and an ESA Coulochem Model 5100A electrochemical detector equipped with an ESA Model 5021 guard cell and a 5011 analytical cell• Dihydroxybenzoates. Samples from reaction systems were diluted in a 0•8% metaphosphoric a c i d / l % dimethyl sulfoxide solution, containing 20 IzM omethoxybenzoic acid as internal standard. 2,3- and 2,5-dihydroxybenzoate (DHBA) were separated using a mobile phase containing 1% acetic acid in a gradient of 40-47•5% methanol in water (v/v) at pH 2•8• Data were acquired for 15 min at a flow of 1 ml/min and the DHBAs were detected by ultraviolet absorption at 313 nm. Tyrosines. Samples were diluted 1: 1 in 0•8% metaphosphoric acid and assayed by HPLC with a
Competition of •OH detectors with erythrocyte ghosts for Cu(II) binding • OH detectors with metal-binding properties may compete with biological material for metal ions• Thus, redistribution of Cu(II) in an erythrocyte ghost suspension upon the introduction of salicylate or phenylalanine was investigated. Erythrocyte ghosts were mixed with CuSO 4 and with various concentrations of salicylate or phenylalanine, ultrafiltered and assayed for Cu content• Both compounds decreased Cu(II) binding to ghosts in a concentration-dependent fashion (Table I). Phenylalanine, however, was more efficient than salicylate in the competition for the Cu(II).
Competition of phenylalanine and salicylate for Cu(II) The relative affinities of phenylalanine and salicylate for Cu(II) were determined by light spectroscopy in a system containing salicylate, Cu(II) and varying concentrations of phenylalanine. When the phenylalanine concentration was increased, Cu(II) binding to salicylate decreased (Fig. 1). Relatively low concentra-
186 TABLE I
Ghosts corresponding to 5% Hct were mixed wilh 0.1 mM Cu(ll) in 1 mM Tris-20 mM Mops buffer (pH 7.4) at 4°C and the detectors were added at the stated concentrations. The samples were ultrafiltered and the Cu(ll) content of the filtrates was determined by the bathocuproine disulfonate method. Cu(ll) bound to ghosts ( n m o l / m l ) ~'
[Detector] (raM)
0 0.2 1.0 5.(l "
salicylate
phenylalanine
46.3 39.8 38.4 32.6
46.3 25.7 I1.8 8.1
0
100 -
The effect of • OH detectors on Cu(ll) binding to erythrocyte ghosts
80
£
60
40 •
Ascorbate
o
Dihydroxybenzoates
20 r
i
0
Calculated as the difference between the Cu(ll) content before ultrafiltration (100 n m o l / m l ) and the concentration in the filtrate.
tions of phenylalanine were required for the shift of Cu(lI) binding, confirming that phenylalanine is a stronger Cu(II) chelator than salicylate. The observed binding was in good agreement with the calculated ligand distribution under competitive conditions (Fig. 1), although in apparent conflict with published binding constants [27] (see Discussion and Table II). Available data were insufficient for the calculation of ternary complex formation, which may readily account for the minor difference between the observed and calculated salicylate binding to Cu(ll).
Salicylate hydroxylation and ascorbate oxidation Before investigating the competition of various detectors and scavengers for .OH, it was necessary to establish that hydroxylation product formation was proportional to ascorbate oxidation. A C u ( I I ) / a s -
1'0
20
30
Time (min) Fig. 2. The relationship between ascorbatc oxidation and salicylate hydroxylation. 0.1 mM C t l S O 4 and 5 mM ascorbate were incubated with 5 mM salicylate in I mM Tris-20 mM Mops (pH 7.4) at 370( `. Samples were analyzed by HPLC for ascorbate and DHBA content. The data for the first 311 rain are presented as percentage of the initial ascorbate concentration (5 raM) and the maximal D t l B A concentration obtained (0.190 mM). Line was fit to DIqBA data points.
corbate free-radical generating system was incubated with 5 mM salicylate for various times to detect hydroxyl radicals. Determination of D H B A formation and residual ascorbate showed that the hydroxylation of salicylate was proportional to the oxidation of ascorbate (Fig. 2). The completion of ascorbate oxidation, defined as > 99% ascorbate oxidized, was achieved in 15 min under optimal shaking conditions. Slower kinetics and decreased yield were observed when oxygen availability was limited.
Competition of salicylate and phenylalanine for •OH "
200
j j,,o-
- -
,=
=E
i
15o.
I
"0
•
"-t
o
100
"0 C ¢8
50
,.Q
Sal. - o b s e r v e d
...... o. ..... ---o--
'
Sal. - c a l c u l a t e d Phe. - c a l c u l a t e d
,..,,I
0 0.0
-
,
.
1.0
,
•
2.0
Phenylalanine
,
"
3.0
, 4.0
•
; 5.0
(mM)
Fig. I. Cu(ll)-binding to salicylate and phenylalanine. 0.1 mM CuSO4, 5 mM salicylate and varying concentrations of phenylalanine were mixed in 1 mM Tris-20 mM Mops (pH 7.4) and absorption spectra were recorded. Salicylate binding to Cu(II) (solid line) was determined using a 393-nm extinction coefficient for copper-bound salicylate of 200 M - 1 cm 1. The binding of salicylate and phenylalanine to Cu(II) was also calculated by computer simulation (broken lines) using published binding constants [27].
When two or more - O H detectors or scavengers arc present in an assay system they will compete for available • OH. This is expected to follow statistical principles governing random reactions. However, it is conceivable that the binding of metal ions to the detectors may affect the efficiency of . O H detection. This was studied under conditions where detectors competed for both metal-binding and .OH. Salicylate and phenylalanine were incubated with the C u ( l I ) / a s c o r b a t e free-radical generating system until ascorbate was depleted, and the hydroxylation products were determined. The amount of D H B A formed decreased with increasing concentrations of added phenylalanine, and the production of tyrosine isomers increased concomitantly (Fig. 3). Assuming that competition is random, the effect of a detector or scavenger on the efficiency of a second detector may be calculated, using the concentration of the detectors and their second order reaction constants
187 300 A --t
•
Dihydroxybenzoates
[]
Tyrosines
100 O
u
I
Calculated
I
Observed
80
200 ' "O
g
Q.
a0
"O
to 100
E
40
'~
20
N, x
T O
'o
0
None
0.2
1.0
0
5.0
0.2
1
5
[Phenylalanine] (raM)
Phenylalanine (raM) Fig. 3. Competitive hydroxylation of salicylate and phenylalanine. 0.1 mM CuSO4, 5 mM ascorbate, 5 mM salicylate and varying concentrations of phenylalanine were incubated in 1 mM Tris-20 mM Mops (pH 7.4) at 37°C until depletion of ascorbate. Samples were analyzed for the formation of DHBA and tyrosine isomers by reversed phase HPLC.
Fig. 4. Salicylate hydroxylation in competition with phenylalanine. The production of DHBA from salicylate was assayed as described in Fig. 3. The relative DHBA production expected under competitive conditions was calculated as described in the text. The results are presented as percentage of the DHBA produced in the absence of phenylalanine (0.267 mM).
with -OH (k t and k2, respectively)as follows:
expected formation of dihydroxybenzoates was calculated using [28]: k(mannitol+.OH
p
k , . [detector 1]
P
k I • [detector 1] + k 2 • [detector 2]
where P and p are the amount of products formed from detector 1 in the absence and presence, respectively, of detector 2. The competition of salicylate and phenylalanine for • O H may thus be calculated from their respective reaction constants [28]:
=
10 9 M -1
s -I
The amount of D H B A formed was higher than expected from the calculated values (Fig. 5). The inability of the non-binding scavengers to intercept . O H to the
tO
1 . 6 . 1 0 ~° M -~ s -
1.7'
k(deoxyribose++OH) = 2.5 • 10 9 M - ] s-
o a,=
k(salicylate +. OH)
) =
1 O0
Jtated
u
k(phenylalanine+. OH) = 6.5 • 1 0 9 M - t
~
80
~
60
rved
s-
When phenylalanine competed with salicylate, D H B A production was consistently lower than calculated for random competition, Fig. 4. This suggested that other factors may be involved in the competition for O H radicals.
c O ~
E
2
183
BBAGEN 23664
Metal-ion-directed site-specificity of hydroxyl radical detection Dan Gelvan, Veronica Moreno, Walter Gassmann, Jack Hegenauer and Paul Saltman Department of Biology, University of California San Diego, La Jolla, CA (USA)
(Received 29 August 1991) (Revised manuscript received 31 December 1991)
Key words: Hydroxyl radical; Site specificity; Free radical; Salicylate; Dihydroxybenzoate; Phenylalanine A wide variety of • OH detectors are in use for determination of biological - OH production. The chemical generation of •OH is site-specific with respect to the metal-binding site, and thus .OH detectors with metal-binding properties may affect the biological damage and bias -OH detection. The present study shows that both salicylate and phenylalanine, added as low molecular weight •OH indicators, decreased Cu(II) binding to erythrocyte ghosts. In a cell-free system, Cu(II) complexed to both salicylate and phenylalanine. Phenylalanine is a stronger Cu(II) chelator than salicylate, both when competing for Cu(II) bound to ghosts and when competing directly with each other. When OH radicals were generated by ascorbate and Cu(II), the amount of • OH detected as dihydroxybenzoates was proportional to the amount of •OH produced. However, when phenylalanine was added to this system, the efficiency of •OH detection by salicylate strongly decreased, concomitant with the transfer of Cu(II) binding from salicylate to the amino acid. This decrease was larger than that predicted by calculations for random competition of the two detectors for •OH. Deoxyribose and mannitol, which do not bind copper appreciably, competed poorly with salicylate for the •OH. Hydroxylation of phenylalanine, on the other hand, was only slightly affected by the presence of salicylate and unaffected by deoxyribose and mannitol. These results suggest that the detection of .OH by low molecular weight "OH indicators was related to the relative affinity of the detectors for the catalyzing metal, and thus partially site-specific. Furthermore, glutamate, which does not contain an aromatic ring but binds Cu(II) with considerable affinity, competed strongly with salicylate for the •OH, indicating that metal-binding properties rather than the presence of an aromatic ring were the cause of the deviation from random competition. The results indicate that .OH indicators with metal-binding properties affect the distribution of catalytic metal ions in a biological system, causing a shift of free radical damage and localizing a site-specific reaction of "OH on these detectors, with a resulting positive bias in the apparent .OH production.
Introduction Free radical d a m a g e in biological systems is largely attributed to • O H generation from hydrogen peroxide. This reaction invariably requires the presence of transition metals capable of one-electron cycling, such as F e ( I I ) - F e ( I I I ) or Cu(I)-Cu(II) (as reviewed in refs. 1, 2). Coordination chemistry dictates that these metals are present in a b o u n d form rather than as free ions. T h e distribution of the metal ions between available macromolecular and, small-molecular-weight ligands and their specific associations is d e t e r m i n e d by individual
Abbreviations: DHBA, dihydroxybenzoate; TBA, 2-thiobarbituric acid; TBARS, thiobarbituric acid reactive substances; Hct, hematocrit. Correspondence (present address): D. Gelvan, Department of Nutrition, Hebrew University - Hadassah Medical School, P.O.B. 1172, Jerusalem 91010, Israel.
concentrations, affinities and speciation with potential ligands. O H radicals are highly reactive and react with most molecules at rates near the limits set by diffusion. Consequently, a high incidence of . O H 'hits' is expected at or near the metal-binding site [3]. This 'sitespecificity' of • O H d a m a g e has b e e n d e m o n s t r a t e d in macromolecules [3-5]. Alternative oxidant species have b e e n suggested to account for a p p a r e n t site-specific reaction mechanisms. T h e p r o p o s e d oxidants include metal ions in the higher oxidation states, such as C u ( I I I ) and Fe(IV), or transitory m e t a l - H 2 0 2 complexes [6-10]. T h e m e t a l - H 2 0 2 intermediate is considered to be p r o d u c e d in all cases, and this complex subsequently decays either by heterolytic cleavage of the H 2 0 2 to p r o d u c e . O H , or by homolytic cleavage which yields the hypervalent metal ions [9]. Thus, the alternative oxidizing species may be p r o d u c e d concomitantly with the • O H [7,9], or in specific cases as the sole oxidants [11], and bring about the oxidative d a m a g e c o m m o n l y ascribed to • O H [10]. Fac-
184 tors such as metal coordination [7,11,12], the presence of oxygen [6] and the availability of reducing and oxidizing species in the system determine which species will be produced and to what extent [7,9,12]. An important feature of the activity of the alternative oxidizing species is their localized reaction at the site of the metal coordination. This makes these species difficult to differentiate from site specifically acting .OH. This p a p e r will discuss the reactions in terms of •OH, while it is recognized that other oxidizing species may be involved. The phenomenon of site-specificity, whether it involves • O H or other coordinated oxidants, vastly complicates the detection and quantitation of free radicals formed in biological systems, since a fraction of the O H radicals reacts locally with the biomolecule and never reaches the free solution where it can react with the detector molecule• Detection is thus blind to the site-specific fraction of the O H radicals which is the biologically important fraction. The use and choice of . O H detectors in biology (reviewed in ref. [13]) is further complicated by considerations of toxicity, metabolism and their recovery or traceability in biological systems. The most suitable detectors that have been introduced for in vivo use arc aromatic compounds that are readily hydroxylated to stable derivatives by . O H and may be recovered and analyzed. The use of the aromatic hydroxylation of salicylate as an indicator of • O H production has a long history (ref. [14] and references therein), but the colorimetric assay of the products is quite insensitive. The quantification of salicylate hydroxylation products by H P L C and electrochemical detection was introduced by Floyd et al., who showed that the conversion of salicylate to 2,3- and 2,5-dihydroxybenzoates ( D H B A ) adequately reflected . O H generation [15]. Salicylate, when used at low concentrations, is well tolerated by cells, organs and whole organisms and seems to be a reliable reporter of • O H formation [16-19]. Halliwell et al. subsequently proposed using the formation of tyrosine isomers from phenylalanine as a superior indicator in biological systems, since phenylalanine is also non-toxic and its D-isomer is not metabolized by most systems [20]. The importance of the choice of an appropriate reporting system has recently been emphasized by the suggestion of Halliwell et at. that metabolic conversion of salicylate to 2,5-DHBA may be mistaken for genuine . O H detection [21]. While the data quoted in support of this contention are incomplete and the contribution of the metabolic conversion may be insignificant and readily corrected for in a controlled study, it serves to stress the complexity of the choice of appropriate . O H detectors for the task at hand. A more serious and often overlooked factor is that indicator molecules may possess metal-binding properties that will cause a redistribution of the metal ions in
the system. A shift in metal-binding will ultimately alter the site-specificity of the free radical attack, thus perturbing the patterns of biological damage. Furthermore, by competing metal ions away from macr~molecules there will be an increase in the fraction of - O H escaping into solution, causing an increase m • O H detection which does not reflect increased production. The binding of metal ions to the . O H detectors may, in turn, result in a site-specific rcaction of • O H with the detector. The effects will combine to produce a positive bias in . O H detection by detectors with metal-binding properties. The possibility of site specific reactions with small molecules has been suggested by Gutteridge, who observed that the mode of metal binding affected the ability of certain -Ott scavengers to protect small molecules from free radical damage [22,23]. The present study investigates metal-binding to various • O H detectors, its effects on - O H measurement and scavenging, and the occurrence of site-specificity in the reaction of •O H with small molecules. We show that some aromatic - O H detectors perturb metal binding in biological systems and cause an augmentation of • O H measurement that is related to site-specificity. Materials and Methods
Materials All materials used were of the highest purity awdlable. Chelex-100 (200-400 mesh) was from Bio-Rad. 2-Thiobarbituric acid (TBA) was crystallized from hot water before use. All other chemicals were used without further purification. Glass-distilled water was used throughout.
Reagents A stock of 0.1 M CuSO~ was prepared in 10 mM H2SO 4. Ascorbic acid, 0.25 M, was prepared fresh in water before each experiment and kept on ice. TBA, 1%, was dissolved i, 60 mM NaOH.
Preparation of erythrocyte ghosts Venous blood samples were collected in heparin from healthy human volunteers, and erythrocyte membrane ghosts were prepared as previously described [24]. The ghosts were suspended in 10 mM Tris buffer (pH 7.5) to a concentration corresponding to 30% hematocrit (Hct).
Metal binding to ghosts A suspension of ghosts corresponding to 5% Hot was mixed with 0.1 mM Cu(II) in 1 mM Tris-20 mM Mops (pH 7.4) on ice. Varying concentrations of salicylate or phenylalanine were added and the suspensions were incubated 10 min. The suspensions were ultrafiltered through Amicon 10 ultrafilters, and the filtrates
185 were assayed for Cu(II) content by the bathocoproine disulfonate method [25], using a Uvikon 930 spectrophotometer (Kontron instruments)• Metal binding to the ghosts was calculated from the difference between the Cu(II) concentration in the original sample and the filtrate.
Metal binding to salicylate and phenylalanine Mixtures containing 0.1 mM Cu(II), 5 mM salicylate and various concentrations of phenylalanine in a 1 mM Tris-20 mM Mops buffer (pH 7.4) were assayed spectrophotometrically. Copper-salicylate binding was calculated from the absorbance at 393 nm, using e = 200 M-1 cm-1 for salicylate bound to copper (experimentally determined in this study; data not shown)•
mobile phase containing 2% methanol (v/v) in an aqueous solution of 30 mM sodium citrate, 27 mM EDTA and 27.7 mM sodium acetate (pH 4.5) at a flow rate of 1.2 ml/min. Tyrosine content was determined by electrochemical detection, using the following settings: guard cell, + 0.40 V; detector 1, + 0.70 V. Residual ascorbate. Samples were diluted 1:1 in 0.8% metaphosphoric acid and 1% dimethyl sulfoxide to terminate the reaction• Residual ascorbate was determined by HPLC. The mobile phase contained: 1% acetic acid and 1.8 mM tetrabutylammonium hydroxide in 20% methanol/80% water (v/v). Data were acquired for 12 rain at a flow of 1 ml/min and ascorbate was detected by ultraviolet absorbance at 245 nm.
TBA assay Reaction systems and competitive •OH scavenging Oxygen-derived free radicals were generated chemically using 5 mM freshly prepared ascorbate and 0.1 mM Cu(I1) in 1 mM Tris-20 mM Mops buffer (pH 7.4). The reactions were initiated by the addition of ascorbate and performed with continuous vigorous shaking at 37°C. Reaction systems contained various combinations and concentrations of salicylate, phenylalanine, deoxyribose and mannitol with the Cu(II)/ascorbate free radical generating system, and were incubated at 37°C until the ascorbate was > 99% depleted• -OH was monitored as the conversion of salicylate to dihydroxybenzoates or of phenylalanine to tyrosine isomers, and by formation of thiobarbituric acid reactive substances (TBARS) from deoxyribose. Reaction products were quantitated by HPLC and by the thiobarbituric acid (TBA) assay.
To determine TBARS formation from deoxyribose, 1 ml samples were added to 0.5 ml 28% trichloroacetic acid-0.1 M NaAsO 2, mixed and treated with Chelex as previously described [24]. 670/~1 of sample were mixed with 330 tzl of 20% trichloroacetic acid, 0.5 ml of 1% TBA and 30/~1 of 10 mM Fe(III). The solutions were heated 12 min at 100°C, cooled on ice, and the absorbance measured against matched reagent blanks at 532 and 600 nm. The 600 nm absorbance was considered to be a nonspecific baseline drift and was subtracted from As32 [26].
Data presentation The results are the mean of at least three separate experiments with a typical coefficient of variation of less than 5%. Results
HPLC detection of dihydroxybenzoates, tyrosines and residual ascorbate Dihydroxybenzoates, tyrosines and residual ascorbate were determined by HPLC on a Rainin Rabbit-HP HPLC system with a Microsorb C18 column (5 lzm spherical bead diameter, 4.6 mm i.d. × 15 cm length), using a Knauer variable wavelength ultraviolet monitor, and an ESA Coulochem Model 5100A electrochemical detector equipped with an ESA Model 5021 guard cell and a 5011 analytical cell• Dihydroxybenzoates. Samples from reaction systems were diluted in a 0•8% metaphosphoric a c i d / l % dimethyl sulfoxide solution, containing 20 IzM omethoxybenzoic acid as internal standard. 2,3- and 2,5-dihydroxybenzoate (DHBA) were separated using a mobile phase containing 1% acetic acid in a gradient of 40-47•5% methanol in water (v/v) at pH 2•8• Data were acquired for 15 min at a flow of 1 ml/min and the DHBAs were detected by ultraviolet absorption at 313 nm. Tyrosines. Samples were diluted 1: 1 in 0•8% metaphosphoric acid and assayed by HPLC with a
Competition of •OH detectors with erythrocyte ghosts for Cu(II) binding • OH detectors with metal-binding properties may compete with biological material for metal ions• Thus, redistribution of Cu(II) in an erythrocyte ghost suspension upon the introduction of salicylate or phenylalanine was investigated. Erythrocyte ghosts were mixed with CuSO 4 and with various concentrations of salicylate or phenylalanine, ultrafiltered and assayed for Cu content• Both compounds decreased Cu(II) binding to ghosts in a concentration-dependent fashion (Table I). Phenylalanine, however, was more efficient than salicylate in the competition for the Cu(II).
Competition of phenylalanine and salicylate for Cu(II) The relative affinities of phenylalanine and salicylate for Cu(II) were determined by light spectroscopy in a system containing salicylate, Cu(II) and varying concentrations of phenylalanine. When the phenylalanine concentration was increased, Cu(II) binding to salicylate decreased (Fig. 1). Relatively low concentra-
186 TABLE I
Ghosts corresponding to 5% Hct were mixed wilh 0.1 mM Cu(ll) in 1 mM Tris-20 mM Mops buffer (pH 7.4) at 4°C and the detectors were added at the stated concentrations. The samples were ultrafiltered and the Cu(ll) content of the filtrates was determined by the bathocuproine disulfonate method. Cu(ll) bound to ghosts ( n m o l / m l ) ~'
[Detector] (raM)
0 0.2 1.0 5.(l "
salicylate
phenylalanine
46.3 39.8 38.4 32.6
46.3 25.7 I1.8 8.1
0
100 -
The effect of • OH detectors on Cu(ll) binding to erythrocyte ghosts
80
£
60
40 •
Ascorbate
o
Dihydroxybenzoates
20 r
i
0
Calculated as the difference between the Cu(ll) content before ultrafiltration (100 n m o l / m l ) and the concentration in the filtrate.
tions of phenylalanine were required for the shift of Cu(lI) binding, confirming that phenylalanine is a stronger Cu(II) chelator than salicylate. The observed binding was in good agreement with the calculated ligand distribution under competitive conditions (Fig. 1), although in apparent conflict with published binding constants [27] (see Discussion and Table II). Available data were insufficient for the calculation of ternary complex formation, which may readily account for the minor difference between the observed and calculated salicylate binding to Cu(ll).
Salicylate hydroxylation and ascorbate oxidation Before investigating the competition of various detectors and scavengers for .OH, it was necessary to establish that hydroxylation product formation was proportional to ascorbate oxidation. A C u ( I I ) / a s -
1'0
20
30
Time (min) Fig. 2. The relationship between ascorbatc oxidation and salicylate hydroxylation. 0.1 mM C t l S O 4 and 5 mM ascorbate were incubated with 5 mM salicylate in I mM Tris-20 mM Mops (pH 7.4) at 370( `. Samples were analyzed by HPLC for ascorbate and DHBA content. The data for the first 311 rain are presented as percentage of the initial ascorbate concentration (5 raM) and the maximal D t l B A concentration obtained (0.190 mM). Line was fit to DIqBA data points.
corbate free-radical generating system was incubated with 5 mM salicylate for various times to detect hydroxyl radicals. Determination of D H B A formation and residual ascorbate showed that the hydroxylation of salicylate was proportional to the oxidation of ascorbate (Fig. 2). The completion of ascorbate oxidation, defined as > 99% ascorbate oxidized, was achieved in 15 min under optimal shaking conditions. Slower kinetics and decreased yield were observed when oxygen availability was limited.
Competition of salicylate and phenylalanine for •OH "
200
j j,,o-
- -
,=
=E
i
15o.
I
"0
•
"-t
o
100
"0 C ¢8
50
,.Q
Sal. - o b s e r v e d
...... o. ..... ---o--
'
Sal. - c a l c u l a t e d Phe. - c a l c u l a t e d
,..,,I
0 0.0
-
,
.
1.0
,
•
2.0
Phenylalanine
,
"
3.0
, 4.0
•
; 5.0
(mM)
Fig. I. Cu(ll)-binding to salicylate and phenylalanine. 0.1 mM CuSO4, 5 mM salicylate and varying concentrations of phenylalanine were mixed in 1 mM Tris-20 mM Mops (pH 7.4) and absorption spectra were recorded. Salicylate binding to Cu(II) (solid line) was determined using a 393-nm extinction coefficient for copper-bound salicylate of 200 M - 1 cm 1. The binding of salicylate and phenylalanine to Cu(II) was also calculated by computer simulation (broken lines) using published binding constants [27].
When two or more - O H detectors or scavengers arc present in an assay system they will compete for available • OH. This is expected to follow statistical principles governing random reactions. However, it is conceivable that the binding of metal ions to the detectors may affect the efficiency of . O H detection. This was studied under conditions where detectors competed for both metal-binding and .OH. Salicylate and phenylalanine were incubated with the C u ( l I ) / a s c o r b a t e free-radical generating system until ascorbate was depleted, and the hydroxylation products were determined. The amount of D H B A formed decreased with increasing concentrations of added phenylalanine, and the production of tyrosine isomers increased concomitantly (Fig. 3). Assuming that competition is random, the effect of a detector or scavenger on the efficiency of a second detector may be calculated, using the concentration of the detectors and their second order reaction constants
187 300 A --t
•
Dihydroxybenzoates
[]
Tyrosines
100 O
u
I
Calculated
I
Observed
80
200 ' "O
g
Q.
a0
"O
to 100
E
40
'~
20
N, x
T O
'o
0
None
0.2
1.0
0
5.0
0.2
1
5
[Phenylalanine] (raM)
Phenylalanine (raM) Fig. 3. Competitive hydroxylation of salicylate and phenylalanine. 0.1 mM CuSO4, 5 mM ascorbate, 5 mM salicylate and varying concentrations of phenylalanine were incubated in 1 mM Tris-20 mM Mops (pH 7.4) at 37°C until depletion of ascorbate. Samples were analyzed for the formation of DHBA and tyrosine isomers by reversed phase HPLC.
Fig. 4. Salicylate hydroxylation in competition with phenylalanine. The production of DHBA from salicylate was assayed as described in Fig. 3. The relative DHBA production expected under competitive conditions was calculated as described in the text. The results are presented as percentage of the DHBA produced in the absence of phenylalanine (0.267 mM).
with -OH (k t and k2, respectively)as follows:
expected formation of dihydroxybenzoates was calculated using [28]: k(mannitol+.OH
p
k , . [detector 1]
P
k I • [detector 1] + k 2 • [detector 2]
where P and p are the amount of products formed from detector 1 in the absence and presence, respectively, of detector 2. The competition of salicylate and phenylalanine for • O H may thus be calculated from their respective reaction constants [28]:
=
10 9 M -1
s -I
The amount of D H B A formed was higher than expected from the calculated values (Fig. 5). The inability of the non-binding scavengers to intercept . O H to the
tO
1 . 6 . 1 0 ~° M -~ s -
1.7'
k(deoxyribose++OH) = 2.5 • 10 9 M - ] s-
o a,=
k(salicylate +. OH)
) =
1 O0
Jtated
u
k(phenylalanine+. OH) = 6.5 • 1 0 9 M - t
~
80
~
60
rved
s-
When phenylalanine competed with salicylate, D H B A production was consistently lower than calculated for random competition, Fig. 4. This suggested that other factors may be involved in the competition for O H radicals.
c O ~
E
2
