Determination of selenols, diselenides, and selenenyl sulfides by

Methods are described for the determination of selenols. (RSeH), diselenides (RSeSeR), and selenenyl sulfides. (RSeSR'). The methods are based on the ...
2 downloads 0 Views 602KB Size
2283

Anal. Chem. 1988, 6 0 , 2283-2287

Determination of Selenols, Diselenides, and Selenenyl Sulfides by Reversed-Phase Liquid Chromatography with Electrochemical Detection Hamada M. A. Killal a n d Dallas L. Rabenstein*

Department of Chemistry, University of California, Riverside, California 92521

Methods are descrlbed for the determination of selenols (RSeH), dlselenldes (RSeSeR), and selenenyl sulfides (RSeSR’). The methods are based on the separation of the selenium-containing compounds by reversed-phase llquld chromatography followed by detection wlth an electrochemlcal detector wlth dual Hg/Au amalgam electrodes In series In the eluent stream. Selenols are detected directly by facltnatlon of the oxidatbn of Hg from the downstream electrode when It Is set to +0.150 V vs the Ag/AgCI reference electrode. Dlselenldes and selenenyl sulfldes are determined by first reducing them to the selenol and/or thlol form at the upstream electrode followed by detection of the selenol and/or thlol at the downstream electrode. By approprlate cholce of potentlal for the upstream electrode, dlselenMes and selenenyl sulfides can be determlned selecilvely In mixtures contalnlng dlselenides, selenenyl sulfides, and dlwlldes. The results Indicate that hlgh-performance llquld chromatography wlth electrochemlcal detection Is a sensltlve and selective method for the determination of selected selenlumcontalnlng compounds.

Even though the importance of selenium as both an essential and a toxic element has long been recognized ( I ) , relatively few methods have been reported for the speciation analysis of organic and biological selenium-coataining compounds (2). The majority of the reported methods for the determination of selenium in organic and biological matrices involve destruction of the organic material by a digestion procedure, which at the same time converts the organoselenium to Se(1V). The Se(1V) is then determined by one of a variety of methods, including fluorometry, atomic absorption, mass spectrometry, and gas chromatography, after conversion to a suitable derivative (S11). These methods give the total selenium content rather than the concentrations of specific selenium-containing compounds. Methods that have been reported for the determination of organic and biochemical forms of selenium have generally been compound-specific, rather than general methods for all organic and biochemical forms of selenium (2). With the development of mercury-based electrochemical detectors for high-performance liquid chromatography (HPLC) (12, 13),there have been significant advances made in methods for the determination of organic and biochemical compounds of the related element, sulfur. For example, sensitive and selective methods have been developed for the determination of thiols and disulfides in complex matrices, including biological fluids (13-18). The following detector reaction for thiols involves oxidation of the mercury electrode itself (12): 2RSH Hg Hg(SR)2 2H+ 2e(1) Because of the high stability of Hg(SR)2, oxidation takes place

+

-

+

+

Permanent address: Department of Chemistry, Zagazig University, Zagazig, Egypt. 0003-2700/88/0360-2283$01.50/0

at less positive electrode potentials than required for oxidation of mercury alone (19). With the addition of an upstream electrode for the conversion of disulfides to thiols (eq 2) which are then detected at the downstream electrode (13), HPLC is a sensitive and selective technique for the simultaneous determination of thiols and disulfides.

RSSR

+ 2e- + 2H+

-

2RSH

(2)

Considering the similarities in the electrochemical properties of thiols and selenols and of disulfides and diselenides at the mercury electrode (20-22), it would seem that HPLC with electrochemical detection also has the potential to provide a sensitive analytical technique for selected selenium compounds. In this paper, we demonstrate that selenols (RSeH), diselenides (RSeSeR), and selenenyl sulfides (R’SSeR) can all be determined by liquid chromatography with electrochemical detection. Both selenols (23-25) and diselenides (26, 27) have been identified in biological systems, and selenenyl sulfides, which can be formed by the reaction

R’SH

+ RSeSeR e R’SSeR + RSeH

(3)

are of interest in the metabolism of selenols and diselenides and as models for an intermediate in the proposed mechanism of action of the selenoenzyme glutathione peroxidase (28). EXPERIMENTAL SECTION Apparatus. All chromatography was performed on a Bioanalytical Systems BAS 200 liquid chromatograph equipped with a dual electrode detector. The detector cell consisted of two 3.2 mm diameter gold electrodes, which were converted to a mercury-gold amalgam surface by placing triply distilled mercury onto the highly polished gold surface and then, after 2-3 min, removing the excess mercury. The dual electrodes were positioned in series in the flow channel. A BAS RE-3 Ag/AgCl reference electrode was used. For most measurements, the upstream Au/Hg electrode was set at a potential of -1.10 V vs Ag/AgCl to reduce disulfides, diselenides, and selenenyl sulfides or -0.55 V vs Ag/AgCl to reduce only diselenides and selenenyl sulfides. The downstream electrode was set at +0.15 V vs Ag/AgCl to detect thiols and selenols originally present in the sample and those produced by reduction of disulfides, diselenides, and selenenyl sulfides at the upstream electrode. The electrodes could be operated at these potentials for about 2 weeks before they needed to be reconditioned. All tubing was stainless steel to exclude oxygen. A Rheodyne Model 7125 sample injector was used with a 20-pL sample loop; injections were made by overloading the loop with 70 p L of sample. The mobile phase was sparged with helium gas and then maintained under a helium positive pressure during use. Separations were done on a 3 pm diameter particle size Biophase ODS column (100 X 3.2 mm). Mobile Phase. The mobile phase was prepared from NaH2P04 (Fisher Scientific Co.), H3P04(Mallinckrodt), sodium octyl sulfate (SOS) (Aldrich), spectrophotometric grade acetonitrile (Mallinckrodt), and distilled water. After a study of the effect of the concentrations of the various mobile phase constituents on retention times, a mobile phase consisting of 0.05 M NaH2P04,40 mg of sodium octyl sulfate/L, 5% (v/v) CH3CN,and sufficient H3P04to give a pH of 2.9 was selected. The mobile phase was filtered through a 0.45-pm cellulose nitrate filter membrane 0 1988 American Chemical Society

2284

ANALYTICAL CHEMISTRY, VOL. 60, NO. 20, OCTOBER 15, 1988

(Whatman F 2848-24) and was pumped at a flow rate of 0.5 mL/min. Chemicals. Reduced and oxidized glutathione, GSH and GSSG, respectively, were obtained from Sigma Chemical Co. Selenocystamine dihydrochloride (RSeSeR) was obtained from Trans World Chemicals, Inc. Selenocysteamine (RSeH) was prepared by electrochemical reduction of RSeSeR (29). Deaerated solutions of RSeSeR in D20 containing 0.3 M KNOB,0.25 M NaH2P04and 1 mM Na2H2EDTA, and sufficient H3P04to give pH 2.0 were reduced at a Hg pool cathode (2.3 cm2)held at a potential of -1.10 V vs Ag/AgCl. The reductions were carried out in an H-cell with the Hg pool cathode, the Ag/AgCl reference electrode, and the RSeSeR solution in one compartment and saturated KC1 and the counter electrode in the second compartment. The two compartments were separated by a glass frit. A Princeton Applied Research Model 174 Polarographic Analyzer was used as the voltage source. Reduction was monitored by 'H NMR (29). Typical reduction times of 120 min were used for 20 mL of 0.1 mM RSeSeR. The selenenyl sulfide RSeSG was prepared by reaction of GSH, GSSG, and RSeSeR (2:l:l ratio) in distilled water. GSH reacts with RSeSeR to form RSeSG and RSeH, which in turn reacts with GSSG to form RSeSG and GSH. After reaction for 10 min, the solution was bubbled with oxygen for 20 min to oxidize GSH and RSeH, and then the solution was deoxygenated by bubbling with nitrogen. This procedure resulted in a solution containing GSSG, RSeSG, and RSeSeR. The concentrations of GSSG and RSeSeR were determined by HPLC with electrochemical detection using calibration curves prepared by using solutions of GSSG and RSeSeR, and then the concentration of RSeSG was calculated from the difference between the initial concentrations of GSH, GSSG, and RSeSeR and the final concentrations of GSSG and RSeSeR. Blood Plasma. Venous blood was drawn into vacutainers (Becton, Dickinson and Co.) that contained K3HEDTA. A 2-mL sample of plasma, which was separated from red cells by centrifugation, was pipetted into a centrifuge tube and an equal volume of cold 5% (v/v) HCIOl solution added to precipitate the plasma proteins. After 5 min, the sample was centrifuged, the supematant was filtered through a OM-pm membrane, and a 2-mL aliquot was spiked with a mixture of GSSG, RSeSG, and RSeSeR. RESULTS AND DISCUSSION A detector with dual Hg/Au electrodes placed in series in the flowing stream was used in this work. To characterize the electrochemical properties of selenols, diselenides, and selenenyl sulfides at the Hg/Au electrode, hydrodynamic voltammograms were obtained by measuring the chromatographic peak current as a function of electrode potential. Because our interest is ultimately in determining these selenium-containing compounds in the presence of thiols and disulfides, we also measured hydrodynamic voltammograms for reduced and oxidized glutathione under identical conditions. First, hydrodynamic voltammograms were measured for selenocysteamine and reduced glutathione by varying the potential applied to the downstream electrode with the upstream electrode turned off. The hydrodynamic voltammogram for selenocysteamine is presented in Figure 1;that for reduced glutathione is presented in Figure 2. Each point on the voltammograms represents the average of three determinations. The two voltammograms are similar in shape, with the current reaching a relatively constant plateau over the potential range 0 to +0.2 V. Based on these results, the potential of the downstream electrode was set at +0.15 V in all the following measurements. It is interesting to note that the voltammetric wave for the facilitated oxidation of Hg in the presence of selenol is shifted to an even more negative potential than that in the presence of thiol. This is consistent with the expected higher stability of Hg(I1)-selenol complexes. The formation constants for Hg(II)-selenocysteamine and Hg(II)-cysteamine have not been measured; however, those for the analogous methylmercury(I1) complexes are as follows: CH3Hg(II)selenocysteamine, log Kf = 16.49; CH3Hg(II)-cysteamine, log

300

-

2

GSSeR 200

C

L

a V

100

1

RSeSeR

2 I

10.2 VOLTS

-100

vs. A g / A g C I

-200 I

-300

c-a--J-

Flgure 1. Hydrodynamic voltammograms for 2 X M selenocysteamine (RSeH), 1 X lo-' M selenocystamine (RSeSeR) and 7.6 X M selenenyl sulfide formed from selenocysteamine and glutathione. Procedures used for the measurement of the hydrodynamic voltammograms and the convention used for plotting peak currents are described in the text. The mobile phase consisted of 20 mg of sodium octyl sulfatell, 5% CH,CN, 0.05 M NaH,PO,, and sufficient H,PO, to give a pH of 2.9.

200

1

a 150

'

V

/

100

m

GSSG

50

I

Y

1

-0.4

t0.2

-0.6

VOLTS -50

-150

-0.8 VS.

-1.0

-1.2

-1.4

Ag/AgCI

T

1

Flgure 2. Hydrodynamic voltammograms for 2.5 X lo4 M glutathione M oxidized glutathione (GSSG). The conditions (GSH) and 6 X are described in the legend for Figure 1.

Kf = 16.17 (30). When the effect of the competitive protonation of the thiolate and selenolate ligands (cysteamine, pKsH = 8.37; selenocysteamine, p K s e ~= 5.50 (30))is taken into account, the difference between the conditional stability constants for the CH3Hg(II)-selenocysteamine and CH3Hg(11)-cysteamine complexes at the pH of the mobile phase (2.9) is even larger (log Kfc = 13.9 and 10.7, respectively). Hydrodynamic voltammograms for the reduction of RSeSeR, GSSeR, and GSSG were obtained by measuring the chromatographic peak current from the downstream electrode

ANALYTICAL CHEMISTRY, VOL. 60, NO. 20, OCTOBER 15, 1988 40

2285

Table I. Distribution of RSeH, RSeSeR, GSSeR, GSH, and GSSG among Species of Different Charge at pH 2.9O Dercentage having an overall charge of: GSH GSSG GSSeR RSeH RSeSeR

30

-

+1

0

-1

3.3 12.4

12.4 27.0 69.6 100

69.6 44.9 17.9

17.9 23.1

-2 1.7 100

oDistributions for GSH and GSSG were calculated from pK, values referenced in the text. The distribution for GSSeR was estimated from the pK, values for the carboxylic acid groups of GSH. RSeH is 100% in the form H3N+CH2CH2SeH, and RSeSeR is 100% in the form H3N+CH2CH2SeSeCH2CH2NH3+ at pH 2.9

L

0

O

m

U

+2

20

I-

O

(30).

m

P

m

0

.

50

10

=

RSeSeR

A

GSSeR

o

RSeH

0

GSSG

40

0

20

4 0

60

80

100

120

mg SOSiL

Flgwe 3. Dependence of the capacity factors (k’) for RSeH, RSeSeR, GSSeR, GSSG, and GSH on the concentration of sodium octyl sulfate (SOS) in the mobile phase (5% CH,CN, 0.05 M NaH,PO,, and H3P0, to give pH 2.9). Other chromatographic and detector conditions are described in the text.

as a function of the potential of the upstream electrode. The potential of the upstream electrode was varied over the range -0.1 to -1.2 V for RSeSeR and GSSeR and over the range -0.4 to -1.4 V for GSSG. The hydrodynamic voltammograms for RSeSeR and GSSeR are presented in Figure 1; that for GSSG is presented in Figure 2. Although the peak currents measured at the downstream electrode result from the facilitated oxidation of the Hg electrode and thus are negative, they are plotted as positive currents in Figures 1 and 2 because they result indirectly from the reduction of RSeSeR, GSSeR, and GSSG a t the upstream electrode; i.e. they provide indirectly a measure of a reduction reaction at the upstream electrode. The hydrodynamic voltammograms indicate that RSeSeR and GSSeR are reduced at a somewhat less negative potential than is GSSG; however, the upstream electrode was set at a potential of -1.10 V for most chromatographic experiments so that RSeSeR, GSSeR, and GSSG could all be determined simultaneously. T o characterize the reversed-phase liquid chromatographic behavior of RSeH, RSeSeR, GSSeR, GSH, and GSSG, the effect of various mobile-phase parameters on their retention was studied. The procedure involved changing a specific mobile-phase parameter while holding the others constant. First, the effect of ion pairing reagent on retention time was studied. The capacity factors for the five molecules being chromatographed, RSeH, RSeSeR, GSSeR, GSSG, and GSH, are plotted in Figure 3 as a function of the concentration of SOS. As expected, the capacity factors increase as the concentration of SOS increases, and there is more ion pairing capacity (31). However, the effect of increasing the concentration of SOS on the capacity factor is different for the different compounds. All five compounds are polyprotic acids with different combinations of carboxylic acid, selenol, sulfhydryl, and ammonium groups. The degree of ionization of

&

30

0 Y m

.-cz 0 P

m 0

20

10

0 2

3

4

5

6

7

Acetonitrile (%) Figure 4. Dependence of the capacity factors (k’)for RSeH, RSeSeR, GSSeR, GSSG, and GSH on the acetonitrile content of the mobile phase (40 mg of SOS/L, 0.05 M NaH,PO,, and H,PO, to give pH 2.9). Other chromatographic and detector condffions are described in the text.

the acidic groups, and thus the overall charge, will be different for the various molecular species present. The distribution of the five compounds among their various charged forms at pH 2.9, calculated from literature values for the ionization constants (30, 32, 33), are presented in Table I. Assuming that the effect of increasing the concentration of SOS on retention time increases as the percentage of the molecule present in a positively charged form increases, and that dipositive species will have a greater tendency to ion pair with SOS than singly charged species, the effect of increasing the concentration of SOS on retention time at pH 2.9 is predicted to increase in the order: GSH < GSSG < RSeH < GSSeR < RSeSeR, exactly as observed in Figure 3. The dependence of the capacity factors on the acetonitrile content of the mobile phase is shown in Figure 4. As expected, the capacity factors decrease as the percentage of acetonitrile increases. Furthermore, the effect increases in the same order as the percentage of the molecular species present (Table I) which form ion pairs with SOS increases. The effect of

2286

ANALYTICAL CHEMISTRY, VOL. 60, NO. 20, OCTOBER 15, 1988 20

RSeSeR

A

A

B

L

K

GSSeR

GSSG

0 c 0

m

RSeH

o

GSH

\

L

LL

o

10

z

.-0 L

m

I

t

0 5

4

PH Flgure 5. Dependence of the capacity factors (k') for RSeH, RSeSeR, GSSeR, GSSG, and GSH on the pH of the moblle phase (40 mg of SOS/L, 5 % CH3CN, 0.05 M NaHm,, and sufficient H3PO, to &e the desired pH). Other chromatographlc and detector conditions are described In the text.

NaH2P04concentration on the capacity factors was studied by using a mobile phase that contained NaH2P04 at concentrations ranging from 0.015 to 0.07 M and sufficient H3P04 to give a pH of 2.9. The capacity factors for all five compounds decrease as the NaH2P04concentration is increased, with that for RSeSeR being strongly dependent on the NaH2PO4 concentration. The effect of NaH2P04 concentration on the capacity factor increases in the order GSH GSSG < RSeH < GSSeR