ANALYTICAL
BIOCHEMISTRY
189,18-23
(1990)
Determination of Dehydroascorbic Acid Using HighPerformance Liquid Chromatography with Coulometric Electrochemical Detection Kuldeep
R. Dhariwal,
Philip
W. Washko,
Laboratory of Cell Biology and Genetics, National National Institutes of Health, Bethesda, Maryland
Received
February
and Mark Institute 20892
Levine
of Diabetes, Digestive
and Kidney
Diseases,
23, 1990
A method for the detection of dehydroascorbic acid using high-performance liquid chromatography with coulometric electrochemical detection is described. Samples were first assayed for ascorbic acid, then reduced with 2,3-dimercapto-1-propanol to convert dehydroascorbic acid in the sample to ascorbic acid, and subsequently reassayed for total ascorbic acid. The dehydroascorbic acid content was the difference between the two measurements. The dehydroascorbic acid assay provides complete recovery of dehydroascorbic acid, without affecting the ascorbic acid content present prior to reduction. The assay is highly sensitive and reproducible with both standards and biological samples, and was used for routine detection of G 1 pmol per sample injection of dehydroascorbic acid. Prior to reduction, dehydroascorbic acid standards frozen at -80% were stable for at least 1 month; after reduction, stability was limited to 3 days. Dehydroascorbic acid was added to human neutrophil samples; the samples were reduced and ascorbic acid was measured. Ascorbic acid in these samples was stable for a12 h in a refrigerated autosampler (0-2°C). With a run time for each sample of only 4 min, multiple samples can be prepared and placed in the autosampler for unattended assaying. O1990AcademicPrass,Inc.
Ascorbic acid and dehydroascorbic acid have been reported to be present in a variety of tissues (1). Ascorbic acid acts as a reducing agent in biochemical reactions in situ (2), and dehydroascorbic acid is a product of ascorbic acid oxidation (3). Although its true function is uncertain, dehydroascorbic acid has been proposed to have antiscorbutic activity (4,5), to be present in increased concentration in diabetics (6), to be the form of the vitamin which crosses some cell membranes (7-IO), and to
be reversibly reduced for reutilization as ascorbic acid in some tissues (7,8). However, these properties remain unsubstantiated, in part, because of problems in measuring dehydroascorbic acid. Assays are available for direct measurement of dehydroascorbic acid in biological samples, but these assays are not ideal (11). Calorimetric assays for dehydroascorbic acid lack sensitivity and specificity for biological samples and may not account for instability of dehydroascorbic acid (11,12). High-performance liquid chromatographic assays that rely on uv detection of dehydroascorbic acid are also insensitive and are not very specific, since the relatively short wavelength used is subject to interference by other substances which may be present in the sample (13,14). To date, high-performance liquid chromatographic assays with electrochemical detection have not been useful for direct measurement of dehydroascorbic acid. Because of these problems, many assays measure dehydroascorbic acid indirectly (11,15-20). The colorimetric assay developed by Roe and Kuether is dependent on the reaction of 2,4-dinitrophenylhydrazine with the product of dehydroascorbic acid oxidation, 2,3diketo1-gulonic acid (16). Although this assay has been modified by many others (17,18), the assay nevertheless lacks sensitivity and specificity for biological samples. Thus, HPLC assays have been developed which measure dehydroascorbic acid indirectly, by reduction. The ascorbic acid content of the sample is first measured directly using uv or electrochemical detection. The sample is then reduced and the total ascorbic acid content determined (18). The difference between the two measurements is the dehydroascorbic acid content of the sample (l&19,20). An HPLC assay for dehydroascorbic acid for biological samples based on the reduction principle should sat0003-2697/90$3.00
18 All
Copyright 0 1990 rights of reproduction
by Academic Press, Inc. in any form reserved.
LIQUID
CHROMATOGRAPHY
isfy several criteria. The requisite ascorbic acid assay should be highly sensitive and specific. Dehydroascorbic acid should be completely converted to ascorbic acid in a relatively short time period, so that sample oxidation is minimized, The reducing agent should either not interfere with the chromatography or be easily removed. Recovery of dehydroascorbic acid should be complete, yet without affecting ascorbic acid already present. Samples should be stabilized at all steps of the assay, including initial sample preparation and subsequent storage, conversion to ascorbic acid, and postreduction ascorbic acid analysis. Samples should be stabilized so that multiple samples can be processed unattended using an autosampler. Currently available assays fall short in satisfying these requirements. This paper presents a new reduction assay for dehydroascorbic acid that is compatible with coulometric electrochemical detection (21). The method is highly sensitive and reproducible, allows for complete reduction of dehydroascorbic acid in ~10 min, and does not interfere with ascorbic acid chromatography. In addition, for the first time, samples have been stabilized for 212 h during analysis, permitting unattended assay of at least 48 samples. MATERIALS
AND
METHODS
Reagents
Dehydroascorbic acid was obtained from ICN Biochemicals. 2,3-dimercapto-1-propanol and ethyl ether were purchased from Aldrich. All other chemicals were of highest purity available and obtained as previously described (21). HPLC
Analysis
The high-performance liquid chromatographic system used has been described (21) and was modified by the addition of a refrigeration unit (Waters Associates) to the autosampler. This permitted the temperature inside the autosampler to be maintained between 0 and 2°C. The capacity of the autosampler carousel is 48 samples (a 96-sample carousel is also available). The mobile phase consisted of 0.05 M sodium phosphate, 0.05 M sodium acetate, 189 pM dodecyltrimethylammonium chloride, and 36.6 pM tetraoctylammonium bromide in 30/ 70 methanol/water (v/v), pH 4.8. of Dehydroascorbic Acid and Preparation of Standards Ascorbic acid and dehydroascorbic acid standards and 2,3-dimercapto-1-propanol were prepared in 30/70 methanol/water (v/v) containing 1 mM EDTA. Equal volumes of various concentrations of dehydroascorbic acid and a 10 mM solution of 2,3-dimercapto-1-propanol were mixed and incubated in the dark for 10 min at Reduction
OF
DEHYDROASCORBIC
ACID
19
room temperature. Following incubation, the solutions were extracted with 3 vol of water-saturated ethyl ether to remove the 2,3-dimercapto-1-propanol, which otherwise interfered with the chromatography. The extraction procedure was repeated two more times. The samples were then purged with nitrogen for 2 min (3,21), transferred to amber vials (3), and placed in the autosampler. Biological
Samples
Human neutrophils were isolated from normal volunteers as previously described (22). The ascorbic acid and dehydroascorbic acid contents of these cells were measured following extraction of the neutrophils with 60/40 methanol/water (v/v) containing 1 mM EDTA. For determination of dehydroascorbic acid, 0.2 ml of the cell extract was incubated with 0.2 ml of 10 mM 2,3-dimercapto-1-propanol for 10 min in the dark at room temperature, and then extracted with ethyl ether and purged with nitrogen as described above before sample injection. The ascorbic acid content of the neutrophils was determined by centrifugation of cell extracts and injection of supernatants. Data Analysis
Each experimental result is the mean + SD for at least three data points. When SD is not displayed, it was smaller than the size of the symbol. RESULTS Optimum
Conditions
of Assay
Dehydroascorbic acid was measured by reducing it to ascorbic acid for subsequent analysis. We tested several reducing agents for their ability to reduce dehydroascorbic acid completely, quickly, and without causing chromatographic interference. Homocysteine interfered with the chromatography and could not be removed. Dithiothreitol reduced dehydroascorbic acid satisfactorily, but interfered with the chromatography so that sensitivity was reduced. Dithiothreitol could not be easily removed from the samples. 2,3-Dimercapto-lpropanol allowed complete reduction of dehydroascorbit acid with 100% recovery at final concentrations of dehydroascorbic acid as low as 0.5 pmol per injection volume. In addition, the time required for complete reduction was only 10 min. 2,3-Dimercapto-1-propanol also interfered with the chromatography. However, interference was overcome by extracting 2,3dimercapto1-propanol with ethyl ether following the incubation period. Ascorbic acid in the sample was not lost during ethyl ether extractions (21). Once opened, 2,3-dimercapto-1-propanol was discarded within 2 weeks and
20
DHARIWAL,
WASHKO,
AND
LEVINE
Figure 2Ais a representative chromatogram of 1 pmol of reduced dehydroascorbic acid. The ascorbic acid peak is distinct, without interference from 2,3-dimercapto-lpropanol. As a control, 100 pmol of dehydroascorbic acid was injected prior to reduction (Fig. 2B). If dehydroascorbic acid were detectable, the peak should have been off-scale in Fig. 2B. The insignificant peak in Fig. 2B is less than 0.5% of the predicted signal. Thus, dehydroascorbic acid is not directly detectable by this assay, and the material used in these experiments contained ~0.5% ascorbic acid as impurity.
0, 0
Recovery 2
4 Concentration
6
8
10
(pmoles)
FIG. 1. Standard curves of dehydroascorbic acid and ascorbic acid. Standards of dehydroascorbic acid were reduced to ascorbic acid with 2,3-dimercapto-1-propanol as described under Materials and Methods. Injection volume was 10 ~1 and the concentrations shown (0.5-10 pmol) are the final concentrations after the reduction. The detector gain was 100 X 10. Ascorbic acid (0) or dehydroascorbic acid (0) concentrations are expressed as a function of arbitrary integration units. Inset: lo-100 pmol, detector gain 10 X 10.
ethyl ether was discarded after 1 week, creased sample stability and recovery. Chromatography
to prevent
of Dehydroascorbic
Acid in Biological
Samples
Human neutrophils were used to determine if the dehydroascorbic acid assay could be applied to biological samples. The neutrophil cell extracts were assayed for ascorbic acid, treated with 2,3-dimercapto-1-propanol to reduce any dehydroascorbic acid present, and reassayed for total ascorbic acid content. The results showed that the ascorbic acid content of these cells did not increase after reduction, indicating that there was no dehydroascorbic acid present in the samples. We then determined whether added dehydroascorbic acid
de-
with Standards
Standards of dehydroascorbic acid and ascorbic acid were prepared so that the final concentrations were identical. Standards of both were run simultaneously to determine the extent of conversion and recovery of dehydroascorbic acid. Two typical standard curves in the ranges 0.5-10 pmol/lO-pl injection and lo-100 pmol/ lo-p1 injection are shown in Fig. 1. The standard curves obtained by reducing dehydroascorbic acid were identical to those obtained with ascorbic acid. These data imply that reduction of dehydroascorbic acid by 2,3-dimercapto-1-propanol was complete and that recovery after extraction of samples with ethyl ether was 100%. Standard curves were linear (R > 0.99) and the values were highly reproducible on a daily basis. In the concentration range 100-1000 pmol/lO-CL1 injection, the recovery was also 100% and the standard curve linear with a correlation coefficient of >0.99 (data not shown). Dehydroascorbic acid at 1 pmol per injection could be routinely measured. The sensitivity of the system could be increased by increasing detector gain. A detector gain of 9000 could be used with a stable baseline. By increasing detector gain, approximately 0.1 pmol dehydroascorbic acid can be detected per sample injection. While standard curves in Fig. 1 were obtained with a lo-p1 injection volume, the injection volume is variable over at least a lo-fold range. Thus, the lowest final concentration that can be detected is 0.1 pmol/lOO ~1.
k 2.0
4.0
Elution Tme (min)
FIG. 2. (A) High-performance liquid chromatographic elution profile of reduced dehydroascorbic acid. Dehydroascorbic acid was reduced to ascorbic acid with 2,3-dimercapto-l-propanol and extracted with ether. After a purge with nitrogen for 2 min, 10 ~1 (1 pmol) was injected. The detector gain was 100 X 20. Ascorbic acid peak is designated by the arrow. (B) High-performance liquid chromatographic elution profile of dehydroascorbic acid. Dehydroascorhic acid was prepared in 30/70 methanol/water containing 1 mM EDTA. 10 ~1 (100 pmol) was injected; the detector gain was 10 X 40. The elution time of ascorbic acid under these conditions is designated by the arrow.
LIQUID CHROMATOGRAPHY TABLE
Detection
1
and Recovery of Dehydroascorbic Ascorbic Acid in Human Neutrophils
Acid and
Total AA b (pm01110 ~1) DHAA” added (pmol/lOjd)
Peak
None’ 0.97 1.95 2.89 3.92 4.88
9037k 11682 15208 19566 24215 25721
area
+ + f -+ +
48 100 70 214 391 237
Recovered 2.95 3.73 4.77 6.05 7.42 7.87
Expected
Recovery (%I
2.95 3.92 4.90 5.94 6.87 7.83
100 95 97 103 108 100
Note. Neutrophil samples were spiked with various amounts of dehydroascorbic acid, reduced with 2,3-dimercapto-1-propanol, and assayed for total ascorbic acid. The unspiked (control) samples were also treated with reducing agent. Injection volume was 10 al and detector gain 100 X 10. Peak area is in arbitrary integrator units. a Dehydroascorbic acid. b Ascorbic acid. ’ Control.
could be reduced to ascorbic acid with full recovery, without affecting the initial ascorbic acid content of the sample. Neutrophil cell extracts were assayed for ascorbic acid, spiked with various amounts of dehydroascorbit acid, and reduced with 2,3-dimercapto-1-propanol (Table 1). All neutrophil cell extracts spiked with different concentrations of dehydroascorbic acid showed virtually 100% recovery. In addition, the original ascorbic acid content of the cell extracts was not affected. Similar experiments were performed by mixing standards of ascorbic acid and dehydroascorbic acid in various concentrations. The mixtures were reduced and total ascorbic acid content determined. In all cases, the reduction of dehydroascorbic acid to ascorbic acid and ascorbic acid recovery were approximately 100% (data not shown). Stability One prerequisite for a useful dehydroascorbic acid assay is that a large number of samples can be prepared and analyzed in a short time. This can only be achieved if dehydroascorbic acid is stable at each stage of the assay: before, during, and after preparation of the sample. Since the assay measures ascorbic acid, the reduced form of dehydroascorbic acid, attempts were first made to stabilize ascorbic acid during reduction and extraction with ether. It has been shown that a combination of 30/70 methanol/water (v/v) and 1 mM EDTA stabilizes ascorbic acid (21). Therefore, the dehydroascorbic acid standards as well as 2,3-dimercapto-1-propanol were also prepared in 30170 methanol/water containing 1 mM EDTA. After the incubation of dehydroascorbic acid
OF DEHYDROASCORBIC
ACID
21
with reducing agent, the samples were extracted with ether and purged with nitrogen for 2 min. By following this procedure, there was no loss of dehydroascorbic acid prior to conversion and no loss of ascorbic acid after reduction (see Table 1 and Fig. 1). After stabilization of the conversion step of dehydroascorbic acid to ascorbic acid, it was necessary to achieve sample stability during daily assaying. Since dehydroascorbic acid was reduced and assayed as ascorbic acid, we began by studying stability of ascorbic acid in standards and biological samples without dehydroascorbic acid. Ascorbic acid in low concentrations has previously been shown to be stable at room temperature for approximately 4 h (21). It was possible that decreasing the sample temperature would increase stability (3). Therefore, samples in an autosampler were cooled to 2°C with a refrigeration unit, and stability of different concentrations of ascorbic acid was measured as a function of time. Under these conditions, ascorbic acid at a concentration as low as 2 pmol/injection was stable for at least 12 h in the autosampler (data not shown). Stability of ascorbic acid in human neutrophils was also investigated. Neutrophils were prepared such that 10 X lo6 cells were placed in 1 ml of 60% methanol containing 1 mM EDTA; this yielded approximately 43 pmol ascorbic acid/lo-pl injection (10 ~1 = 1O’cell equivalents). Since this amount of ascorbic acid was higher than the amount we wished to test for stability, samples were diluted to obtain three different final concentrations (2.3, 8.7, and 14.5 pmol per lo-b1 injection) of ascorbic acid. The samples were stored in the refrigerated autosampler and injected every 2 h. Results in Fig. 3A show that no loss of ascorbic acid occurred during a 12-h period. We then determined the stability of samples which contained dehydroascorbic acid and were reduced. Neutrophil cell extracts were diluted to contain 10 pmol ascorbic acid/l0 ~1 and were then spiked with 10,20, and 30 pmol/lO ~1 of dehydroascorbic acid. The samples were treated with 2,3-dimercapto-1-propanol, ether extracted, purged with nitrogen, stored in the refrigerated autosampler, and injected approximately every 2 h. The total ascorbic acid content in these samples after extraction should have been 10,15, and 20 pmol/lO ~1 because of 1:l dilution with the reducing agent. There was 100% conversion and recovery of dehydroascorbic acid to ascorbic acid. Again, there was no loss of ascorbic acid content (reduced dehydroascorbic acid) for at least 13 h in the autosampler (Fig. 3B). We then investigated long-term stability, both before and after reduction of dehydroascorbic acid. Dehydroascorbic acid standards were prepared at different concentrations, nitrogen purged, and stored at -80°C. At different time intervals, samples were thawed, reduced with 2,3-dimercapto-1-propanol, and assayed after ether extraction. Dehydroascorbic acid at a concentra-
22
DHARIWAL,
WASHKO,
AND
LEVINE
50-
x
A
\
B
40 -r-J a
h: a
.
,,
30 --
l
0 x 0 F a Y
.
45"
.
P) 0
I
.
-
30--
x
0
0
0 0
P
zo-2 Y
10 -0
e
07
15.-
P a
07 0
3
6 Time
9
12
(hours)
0
3
6 Time
9
12
(hours)
FIG.
3. (A) Stability of ascorbic acid in a refrigerated autosampler, from samples of human neutrophils. Neutrophil cell extracts were prepared in 60/40 methanol/water containing 1 mM EDTA. Extracts were diluted to contain three different concentrations of ascorbic acid, aliquoted in several vials, and placed in the autosampler. Injections (10 ~1) were made every 2 h, and detector gain was 10 X 90. The amounts of ascorbic acid per 10 ~1 injection were 2.3 pmol (O), 8.7 pmol (O), and 15 pmol (A). (B) Stability of reduced dehydroascorbic acid in human neutrophils during storage in a refrigerated autosampler. Neutrophil cell extracts containing 10 pmol ascorbic acid/l0 ~1 were spiked with 10, 20 and 30 pmol of dehydroascorbic acid. The samples were treated with 2,3-dimercapto-1-propanol to convert dehydroascorbic acid to ascorbic acid, extracted with ether, aliquoted into several vials and placed in the autosampler. Injections (10 ~1) were made approximately every 2 h. The detector gain was 10 X 75 and final ascorbic acid concentrations per 10 pl were 10 pmol (O), 15 pmol (O), and 20 pmol (A).
tion as low as 2 pmol/injection was stable for over 1 month. Similarly, stability was determined after the dehydroascorbic acid was reduced to ascorbic acid and then stored at -80°C. Such standards were stable for up to 1 week at concentrations of 10 pmol or above. Standards of lower concentrations were stable for up to 3 days (data not shown). DISCUSSION We describe here a high-performance liquid chromatographic assay for dehydroascorbic acid. Dehydroascorbic acid is first reduced to ascorbic acid which is then measured by using HPLC with coulometric electrochemical detection (21). In a given sample where both ascorbic acid and dehydroascorbic acid may be present, the sample is first assayed for ascorbic acid, then reduced with 2,3-dimercapto-l-propanol to convert any dehydroascorbic acid present to ascorbic acid, and reassayed to determine the increase in ascorbic acid content. The amount of dehydroascorbic acid present is calculated from the difference in ascorbic acid content before and after the reduction. The dehydroascorbic acid assay is reproducible, specific, and highly sensitive, allowing the detection of less than a picomole of dehydroascorbic acid. There is no interference from the reductant, which is removed by extraction with ethyl ether. The reduction process allows for full recovery of dehydroascorbic acid, yet without affecting ascorbic acid previously present. The reduction of dehydroascorbic acid to ascorbic acid is complete in 10 min, allowing a large number of samples to be assayed in 1 day. Following reduction, the reducing agent is easily extracted with ether, eliminating interfer-
ing peaks in the chromatogram. Since ascorbic acid elutes in approximately 2 min, the run time for each sample can be as short as 4 min, permitting 14 or 15 samples to be analyzed in an hour using an autosampler. Because of its high sensitivity, the assay is especially suited to biological samples, where only small amounts of dehydroascorbic acid may be present and assay specificity is important. If a biological sample is thought to contain more than 1000 pmol/lO ~1 (100 mM) dehydroascorbic acid, an aliquot can be diluted with 60% methanol-l IYIM EDTA and reduced for assay with 2,3dimercapto-1-propanol. Stability of ascorbic acid in samples for both short and long periods of time has been recently reported (21). We have now further increased this stability by use of a refrigeration unit attached to the autosampler. The use of a cooling device allows the temperature of the autosampler to be maintained at approximately 2°C. Ascorbic acid at this temperature is stable for at least 12 h at a concentration as low as 2 pmol/injection. This allows for the unattended analysis of a large number of samples. Thus 48-96 ascorbic acid samples (depending on the full capacity of autosampler carousel) can be prepared at a time and put in the autosampler for analysis. The system can be used overnight which provides great versatility to the system. Kutnink et al. (23) used isoascorbic acid as an internal standard of ascorbic acid oxidation, to permit correction for intraassay ascorbic acid instability. An advantage of the method presented here is that no correction is needed for ascorbic acid instability, since ascorbic acid remains stable under the assay conditions. Other dehydroascorbic acid assays have drawbacks when they are used with biological samples. Sensitivity
LIQUID
CHROMATOGRAPHY
OF
in several systems is limited (10,12,13,16-19). Sensitivity can be increased by post-column reduction of dehydroascorbic acid or by derivatizing dehydroascorbic acid (24,25). However, derivatization can affect the assay specificity (11). Post-column reduction and most other assays do not account either for dehydroascorbic acid recovery or for sample instability, a recurrent problem with ascorbic acid and dehydroascorbic acid (25,26). The dehydroascorbic acid assay described here does not have these limitations, and is suitable for measuring small amounts of dehydroascorbic acid in biological samples. This assay is currently being used in our laboratory to determine whether dehydroascorbic is involved in the mechanism of ascorbic acid transport in human neutrophils (7,8,22).
was supported
by National
Research
Service
REFERENCES 1. Hornig,
D. (1975)
Ann.
N.Y.
Acad.
Sci. 258,
103-118.
2. Englard, S., and Seifter, S. (1986) Annu. Reu. Nutr. 3. Lewin, S. (1976) Vitamin C: Its Molecular Biology Potential, pp. 5-39, Academic Press, New York. 4. Fox,
7. Hornig, D., Weiser, Acta 32,33-39.
F. W., and Levy,
L. P. (1936)
Biochem.
6, 365-406. and Medical
J. 30,211-217.
5. Todhunter, E., McMillan, T., and Ehmke, D. (1950) J. Nutr. 42, 297-308. 6. Som, S., Basu, S., Mukherjee, D., Deb, S., Choudhury, P. R., Mukherjee, S., Chatterjee, S. N., and Chatterjee, I. B. (1981) Metabolism 30, 572-577.
23
ACID H., Weber,
F., and Wiss,
8. Bigley, 1092.
R. H., and Stankova,
L. (1974)
9. Tirrell, 186.
J. G., and Westhead,
E. W. (1979)
0. (1971)
J. Exp.
Clin. Chim.
139, 1084
Med.
Neuroscience
4, 181-
10. Stahl, R. L., Farber, C. M., Liebes, L. F., and Silber, R. (1985) Cancer Res. 45,6507-6512. 11. Cooke, J. R., and Moxon, R. E. D. (1981) in Vitamin C (Counsell, J. N., and Hornig, D., Eds.), pp. 167-198, Applied Science, London. 12. Deutsch, M. J., and Weeks, 48, 1248. 13. Finley,
J. W., and Duang,
14. Rose, 145.
R. C., and Nahrwold,
15. Kim,
H. J. (1989)
19. Doner, 230.
C. E. (1965) E. (1981)
Ofi Anal.
C. A. (1943)
O., and Brassard, M. (1980) Clin. L. W., and Hicks,
J. Assoc.
J. Chromatogr.
D. (1981)
J. Assoc.
16. Roe, J. H., and Keuther, 17. Pelletier, 18. Okamura,
ACKNOWLEDGMENT Philip W. Washko Award DK08060-03.
DEHYDROASCORBIC
Anal. Chem. J. Biol.
Off. And.
Chem.
207,449-453.
Biochem.
114, 140-
72,681-686. Chem.
147,399-407.
R. (1977) J. Food Sci. 42,1471-1477. Chim. Acta 103, 259-268. K. B. (1981)
Anal.
Biochem.
115,225-
20. Farber, C. M., Kanengiser, S., Stahl, R., Liebes, L., and Silber, R. (1983) Anal. Biochem. 134,355-360. 21. Washko, P. W., Hartzell, W. O., and Levine, M. (1989) Anal. Biothem. 181,276-282. 22. Washko, P., Rotrosen, D., and Levine, M. (1989) J. Biol. Chem. 264, 18,996-19,002. 23. Kutnink, M. A., Hawkes, W. C., Schaus, E. E., and Omaye, S. T. (1987) Anal. Biochem. 166,424-430. 24. Iwata, T., Yamaguchi, M., Hara, S., and Nakamura, M. (1985) J. Chromatogr. 344, 351-355. 25. Ziegler, S. J., Meier, B., and Sticher, 0. (1987) J. Chromatogr.
391,419-426. 26. Pachla, 367.
L. A., and Kissinger,
P. T. (1976)
Ad.
Chem.
48,
364-
BIOCHEMISTRY
189,18-23
(1990)
Determination of Dehydroascorbic Acid Using HighPerformance Liquid Chromatography with Coulometric Electrochemical Detection Kuldeep
R. Dhariwal,
Philip
W. Washko,
Laboratory of Cell Biology and Genetics, National National Institutes of Health, Bethesda, Maryland
Received
February
and Mark Institute 20892
Levine
of Diabetes, Digestive
and Kidney
Diseases,
23, 1990
A method for the detection of dehydroascorbic acid using high-performance liquid chromatography with coulometric electrochemical detection is described. Samples were first assayed for ascorbic acid, then reduced with 2,3-dimercapto-1-propanol to convert dehydroascorbic acid in the sample to ascorbic acid, and subsequently reassayed for total ascorbic acid. The dehydroascorbic acid content was the difference between the two measurements. The dehydroascorbic acid assay provides complete recovery of dehydroascorbic acid, without affecting the ascorbic acid content present prior to reduction. The assay is highly sensitive and reproducible with both standards and biological samples, and was used for routine detection of G 1 pmol per sample injection of dehydroascorbic acid. Prior to reduction, dehydroascorbic acid standards frozen at -80% were stable for at least 1 month; after reduction, stability was limited to 3 days. Dehydroascorbic acid was added to human neutrophil samples; the samples were reduced and ascorbic acid was measured. Ascorbic acid in these samples was stable for a12 h in a refrigerated autosampler (0-2°C). With a run time for each sample of only 4 min, multiple samples can be prepared and placed in the autosampler for unattended assaying. O1990AcademicPrass,Inc.
Ascorbic acid and dehydroascorbic acid have been reported to be present in a variety of tissues (1). Ascorbic acid acts as a reducing agent in biochemical reactions in situ (2), and dehydroascorbic acid is a product of ascorbic acid oxidation (3). Although its true function is uncertain, dehydroascorbic acid has been proposed to have antiscorbutic activity (4,5), to be present in increased concentration in diabetics (6), to be the form of the vitamin which crosses some cell membranes (7-IO), and to
be reversibly reduced for reutilization as ascorbic acid in some tissues (7,8). However, these properties remain unsubstantiated, in part, because of problems in measuring dehydroascorbic acid. Assays are available for direct measurement of dehydroascorbic acid in biological samples, but these assays are not ideal (11). Calorimetric assays for dehydroascorbic acid lack sensitivity and specificity for biological samples and may not account for instability of dehydroascorbic acid (11,12). High-performance liquid chromatographic assays that rely on uv detection of dehydroascorbic acid are also insensitive and are not very specific, since the relatively short wavelength used is subject to interference by other substances which may be present in the sample (13,14). To date, high-performance liquid chromatographic assays with electrochemical detection have not been useful for direct measurement of dehydroascorbic acid. Because of these problems, many assays measure dehydroascorbic acid indirectly (11,15-20). The colorimetric assay developed by Roe and Kuether is dependent on the reaction of 2,4-dinitrophenylhydrazine with the product of dehydroascorbic acid oxidation, 2,3diketo1-gulonic acid (16). Although this assay has been modified by many others (17,18), the assay nevertheless lacks sensitivity and specificity for biological samples. Thus, HPLC assays have been developed which measure dehydroascorbic acid indirectly, by reduction. The ascorbic acid content of the sample is first measured directly using uv or electrochemical detection. The sample is then reduced and the total ascorbic acid content determined (18). The difference between the two measurements is the dehydroascorbic acid content of the sample (l&19,20). An HPLC assay for dehydroascorbic acid for biological samples based on the reduction principle should sat0003-2697/90$3.00
18 All
Copyright 0 1990 rights of reproduction
by Academic Press, Inc. in any form reserved.
LIQUID
CHROMATOGRAPHY
isfy several criteria. The requisite ascorbic acid assay should be highly sensitive and specific. Dehydroascorbic acid should be completely converted to ascorbic acid in a relatively short time period, so that sample oxidation is minimized, The reducing agent should either not interfere with the chromatography or be easily removed. Recovery of dehydroascorbic acid should be complete, yet without affecting ascorbic acid already present. Samples should be stabilized at all steps of the assay, including initial sample preparation and subsequent storage, conversion to ascorbic acid, and postreduction ascorbic acid analysis. Samples should be stabilized so that multiple samples can be processed unattended using an autosampler. Currently available assays fall short in satisfying these requirements. This paper presents a new reduction assay for dehydroascorbic acid that is compatible with coulometric electrochemical detection (21). The method is highly sensitive and reproducible, allows for complete reduction of dehydroascorbic acid in ~10 min, and does not interfere with ascorbic acid chromatography. In addition, for the first time, samples have been stabilized for 212 h during analysis, permitting unattended assay of at least 48 samples. MATERIALS
AND
METHODS
Reagents
Dehydroascorbic acid was obtained from ICN Biochemicals. 2,3-dimercapto-1-propanol and ethyl ether were purchased from Aldrich. All other chemicals were of highest purity available and obtained as previously described (21). HPLC
Analysis
The high-performance liquid chromatographic system used has been described (21) and was modified by the addition of a refrigeration unit (Waters Associates) to the autosampler. This permitted the temperature inside the autosampler to be maintained between 0 and 2°C. The capacity of the autosampler carousel is 48 samples (a 96-sample carousel is also available). The mobile phase consisted of 0.05 M sodium phosphate, 0.05 M sodium acetate, 189 pM dodecyltrimethylammonium chloride, and 36.6 pM tetraoctylammonium bromide in 30/ 70 methanol/water (v/v), pH 4.8. of Dehydroascorbic Acid and Preparation of Standards Ascorbic acid and dehydroascorbic acid standards and 2,3-dimercapto-1-propanol were prepared in 30/70 methanol/water (v/v) containing 1 mM EDTA. Equal volumes of various concentrations of dehydroascorbic acid and a 10 mM solution of 2,3-dimercapto-1-propanol were mixed and incubated in the dark for 10 min at Reduction
OF
DEHYDROASCORBIC
ACID
19
room temperature. Following incubation, the solutions were extracted with 3 vol of water-saturated ethyl ether to remove the 2,3-dimercapto-1-propanol, which otherwise interfered with the chromatography. The extraction procedure was repeated two more times. The samples were then purged with nitrogen for 2 min (3,21), transferred to amber vials (3), and placed in the autosampler. Biological
Samples
Human neutrophils were isolated from normal volunteers as previously described (22). The ascorbic acid and dehydroascorbic acid contents of these cells were measured following extraction of the neutrophils with 60/40 methanol/water (v/v) containing 1 mM EDTA. For determination of dehydroascorbic acid, 0.2 ml of the cell extract was incubated with 0.2 ml of 10 mM 2,3-dimercapto-1-propanol for 10 min in the dark at room temperature, and then extracted with ethyl ether and purged with nitrogen as described above before sample injection. The ascorbic acid content of the neutrophils was determined by centrifugation of cell extracts and injection of supernatants. Data Analysis
Each experimental result is the mean + SD for at least three data points. When SD is not displayed, it was smaller than the size of the symbol. RESULTS Optimum
Conditions
of Assay
Dehydroascorbic acid was measured by reducing it to ascorbic acid for subsequent analysis. We tested several reducing agents for their ability to reduce dehydroascorbic acid completely, quickly, and without causing chromatographic interference. Homocysteine interfered with the chromatography and could not be removed. Dithiothreitol reduced dehydroascorbic acid satisfactorily, but interfered with the chromatography so that sensitivity was reduced. Dithiothreitol could not be easily removed from the samples. 2,3-Dimercapto-lpropanol allowed complete reduction of dehydroascorbit acid with 100% recovery at final concentrations of dehydroascorbic acid as low as 0.5 pmol per injection volume. In addition, the time required for complete reduction was only 10 min. 2,3-Dimercapto-1-propanol also interfered with the chromatography. However, interference was overcome by extracting 2,3dimercapto1-propanol with ethyl ether following the incubation period. Ascorbic acid in the sample was not lost during ethyl ether extractions (21). Once opened, 2,3-dimercapto-1-propanol was discarded within 2 weeks and
20
DHARIWAL,
WASHKO,
AND
LEVINE
Figure 2Ais a representative chromatogram of 1 pmol of reduced dehydroascorbic acid. The ascorbic acid peak is distinct, without interference from 2,3-dimercapto-lpropanol. As a control, 100 pmol of dehydroascorbic acid was injected prior to reduction (Fig. 2B). If dehydroascorbic acid were detectable, the peak should have been off-scale in Fig. 2B. The insignificant peak in Fig. 2B is less than 0.5% of the predicted signal. Thus, dehydroascorbic acid is not directly detectable by this assay, and the material used in these experiments contained ~0.5% ascorbic acid as impurity.
0, 0
Recovery 2
4 Concentration
6
8
10
(pmoles)
FIG. 1. Standard curves of dehydroascorbic acid and ascorbic acid. Standards of dehydroascorbic acid were reduced to ascorbic acid with 2,3-dimercapto-1-propanol as described under Materials and Methods. Injection volume was 10 ~1 and the concentrations shown (0.5-10 pmol) are the final concentrations after the reduction. The detector gain was 100 X 10. Ascorbic acid (0) or dehydroascorbic acid (0) concentrations are expressed as a function of arbitrary integration units. Inset: lo-100 pmol, detector gain 10 X 10.
ethyl ether was discarded after 1 week, creased sample stability and recovery. Chromatography
to prevent
of Dehydroascorbic
Acid in Biological
Samples
Human neutrophils were used to determine if the dehydroascorbic acid assay could be applied to biological samples. The neutrophil cell extracts were assayed for ascorbic acid, treated with 2,3-dimercapto-1-propanol to reduce any dehydroascorbic acid present, and reassayed for total ascorbic acid content. The results showed that the ascorbic acid content of these cells did not increase after reduction, indicating that there was no dehydroascorbic acid present in the samples. We then determined whether added dehydroascorbic acid
de-
with Standards
Standards of dehydroascorbic acid and ascorbic acid were prepared so that the final concentrations were identical. Standards of both were run simultaneously to determine the extent of conversion and recovery of dehydroascorbic acid. Two typical standard curves in the ranges 0.5-10 pmol/lO-pl injection and lo-100 pmol/ lo-p1 injection are shown in Fig. 1. The standard curves obtained by reducing dehydroascorbic acid were identical to those obtained with ascorbic acid. These data imply that reduction of dehydroascorbic acid by 2,3-dimercapto-1-propanol was complete and that recovery after extraction of samples with ethyl ether was 100%. Standard curves were linear (R > 0.99) and the values were highly reproducible on a daily basis. In the concentration range 100-1000 pmol/lO-CL1 injection, the recovery was also 100% and the standard curve linear with a correlation coefficient of >0.99 (data not shown). Dehydroascorbic acid at 1 pmol per injection could be routinely measured. The sensitivity of the system could be increased by increasing detector gain. A detector gain of 9000 could be used with a stable baseline. By increasing detector gain, approximately 0.1 pmol dehydroascorbic acid can be detected per sample injection. While standard curves in Fig. 1 were obtained with a lo-p1 injection volume, the injection volume is variable over at least a lo-fold range. Thus, the lowest final concentration that can be detected is 0.1 pmol/lOO ~1.
k 2.0
4.0
Elution Tme (min)
FIG. 2. (A) High-performance liquid chromatographic elution profile of reduced dehydroascorbic acid. Dehydroascorbic acid was reduced to ascorbic acid with 2,3-dimercapto-l-propanol and extracted with ether. After a purge with nitrogen for 2 min, 10 ~1 (1 pmol) was injected. The detector gain was 100 X 20. Ascorbic acid peak is designated by the arrow. (B) High-performance liquid chromatographic elution profile of dehydroascorbic acid. Dehydroascorhic acid was prepared in 30/70 methanol/water containing 1 mM EDTA. 10 ~1 (100 pmol) was injected; the detector gain was 10 X 40. The elution time of ascorbic acid under these conditions is designated by the arrow.
LIQUID CHROMATOGRAPHY TABLE
Detection
1
and Recovery of Dehydroascorbic Ascorbic Acid in Human Neutrophils
Acid and
Total AA b (pm01110 ~1) DHAA” added (pmol/lOjd)
Peak
None’ 0.97 1.95 2.89 3.92 4.88
9037k 11682 15208 19566 24215 25721
area
+ + f -+ +
48 100 70 214 391 237
Recovered 2.95 3.73 4.77 6.05 7.42 7.87
Expected
Recovery (%I
2.95 3.92 4.90 5.94 6.87 7.83
100 95 97 103 108 100
Note. Neutrophil samples were spiked with various amounts of dehydroascorbic acid, reduced with 2,3-dimercapto-1-propanol, and assayed for total ascorbic acid. The unspiked (control) samples were also treated with reducing agent. Injection volume was 10 al and detector gain 100 X 10. Peak area is in arbitrary integrator units. a Dehydroascorbic acid. b Ascorbic acid. ’ Control.
could be reduced to ascorbic acid with full recovery, without affecting the initial ascorbic acid content of the sample. Neutrophil cell extracts were assayed for ascorbic acid, spiked with various amounts of dehydroascorbit acid, and reduced with 2,3-dimercapto-1-propanol (Table 1). All neutrophil cell extracts spiked with different concentrations of dehydroascorbic acid showed virtually 100% recovery. In addition, the original ascorbic acid content of the cell extracts was not affected. Similar experiments were performed by mixing standards of ascorbic acid and dehydroascorbic acid in various concentrations. The mixtures were reduced and total ascorbic acid content determined. In all cases, the reduction of dehydroascorbic acid to ascorbic acid and ascorbic acid recovery were approximately 100% (data not shown). Stability One prerequisite for a useful dehydroascorbic acid assay is that a large number of samples can be prepared and analyzed in a short time. This can only be achieved if dehydroascorbic acid is stable at each stage of the assay: before, during, and after preparation of the sample. Since the assay measures ascorbic acid, the reduced form of dehydroascorbic acid, attempts were first made to stabilize ascorbic acid during reduction and extraction with ether. It has been shown that a combination of 30/70 methanol/water (v/v) and 1 mM EDTA stabilizes ascorbic acid (21). Therefore, the dehydroascorbic acid standards as well as 2,3-dimercapto-1-propanol were also prepared in 30170 methanol/water containing 1 mM EDTA. After the incubation of dehydroascorbic acid
OF DEHYDROASCORBIC
ACID
21
with reducing agent, the samples were extracted with ether and purged with nitrogen for 2 min. By following this procedure, there was no loss of dehydroascorbic acid prior to conversion and no loss of ascorbic acid after reduction (see Table 1 and Fig. 1). After stabilization of the conversion step of dehydroascorbic acid to ascorbic acid, it was necessary to achieve sample stability during daily assaying. Since dehydroascorbic acid was reduced and assayed as ascorbic acid, we began by studying stability of ascorbic acid in standards and biological samples without dehydroascorbic acid. Ascorbic acid in low concentrations has previously been shown to be stable at room temperature for approximately 4 h (21). It was possible that decreasing the sample temperature would increase stability (3). Therefore, samples in an autosampler were cooled to 2°C with a refrigeration unit, and stability of different concentrations of ascorbic acid was measured as a function of time. Under these conditions, ascorbic acid at a concentration as low as 2 pmol/injection was stable for at least 12 h in the autosampler (data not shown). Stability of ascorbic acid in human neutrophils was also investigated. Neutrophils were prepared such that 10 X lo6 cells were placed in 1 ml of 60% methanol containing 1 mM EDTA; this yielded approximately 43 pmol ascorbic acid/lo-pl injection (10 ~1 = 1O’cell equivalents). Since this amount of ascorbic acid was higher than the amount we wished to test for stability, samples were diluted to obtain three different final concentrations (2.3, 8.7, and 14.5 pmol per lo-b1 injection) of ascorbic acid. The samples were stored in the refrigerated autosampler and injected every 2 h. Results in Fig. 3A show that no loss of ascorbic acid occurred during a 12-h period. We then determined the stability of samples which contained dehydroascorbic acid and were reduced. Neutrophil cell extracts were diluted to contain 10 pmol ascorbic acid/l0 ~1 and were then spiked with 10,20, and 30 pmol/lO ~1 of dehydroascorbic acid. The samples were treated with 2,3-dimercapto-1-propanol, ether extracted, purged with nitrogen, stored in the refrigerated autosampler, and injected approximately every 2 h. The total ascorbic acid content in these samples after extraction should have been 10,15, and 20 pmol/lO ~1 because of 1:l dilution with the reducing agent. There was 100% conversion and recovery of dehydroascorbic acid to ascorbic acid. Again, there was no loss of ascorbic acid content (reduced dehydroascorbic acid) for at least 13 h in the autosampler (Fig. 3B). We then investigated long-term stability, both before and after reduction of dehydroascorbic acid. Dehydroascorbic acid standards were prepared at different concentrations, nitrogen purged, and stored at -80°C. At different time intervals, samples were thawed, reduced with 2,3-dimercapto-1-propanol, and assayed after ether extraction. Dehydroascorbic acid at a concentra-
22
DHARIWAL,
WASHKO,
AND
LEVINE
50-
x
A
\
B
40 -r-J a
h: a
.
,,
30 --
l
0 x 0 F a Y
.
45"
.
P) 0
I
.
-
30--
x
0
0
0 0
P
zo-2 Y
10 -0
e
07
15.-
P a
07 0
3
6 Time
9
12
(hours)
0
3
6 Time
9
12
(hours)
FIG.
3. (A) Stability of ascorbic acid in a refrigerated autosampler, from samples of human neutrophils. Neutrophil cell extracts were prepared in 60/40 methanol/water containing 1 mM EDTA. Extracts were diluted to contain three different concentrations of ascorbic acid, aliquoted in several vials, and placed in the autosampler. Injections (10 ~1) were made every 2 h, and detector gain was 10 X 90. The amounts of ascorbic acid per 10 ~1 injection were 2.3 pmol (O), 8.7 pmol (O), and 15 pmol (A). (B) Stability of reduced dehydroascorbic acid in human neutrophils during storage in a refrigerated autosampler. Neutrophil cell extracts containing 10 pmol ascorbic acid/l0 ~1 were spiked with 10, 20 and 30 pmol of dehydroascorbic acid. The samples were treated with 2,3-dimercapto-1-propanol to convert dehydroascorbic acid to ascorbic acid, extracted with ether, aliquoted into several vials and placed in the autosampler. Injections (10 ~1) were made approximately every 2 h. The detector gain was 10 X 75 and final ascorbic acid concentrations per 10 pl were 10 pmol (O), 15 pmol (O), and 20 pmol (A).
tion as low as 2 pmol/injection was stable for over 1 month. Similarly, stability was determined after the dehydroascorbic acid was reduced to ascorbic acid and then stored at -80°C. Such standards were stable for up to 1 week at concentrations of 10 pmol or above. Standards of lower concentrations were stable for up to 3 days (data not shown). DISCUSSION We describe here a high-performance liquid chromatographic assay for dehydroascorbic acid. Dehydroascorbic acid is first reduced to ascorbic acid which is then measured by using HPLC with coulometric electrochemical detection (21). In a given sample where both ascorbic acid and dehydroascorbic acid may be present, the sample is first assayed for ascorbic acid, then reduced with 2,3-dimercapto-l-propanol to convert any dehydroascorbic acid present to ascorbic acid, and reassayed to determine the increase in ascorbic acid content. The amount of dehydroascorbic acid present is calculated from the difference in ascorbic acid content before and after the reduction. The dehydroascorbic acid assay is reproducible, specific, and highly sensitive, allowing the detection of less than a picomole of dehydroascorbic acid. There is no interference from the reductant, which is removed by extraction with ethyl ether. The reduction process allows for full recovery of dehydroascorbic acid, yet without affecting ascorbic acid previously present. The reduction of dehydroascorbic acid to ascorbic acid is complete in 10 min, allowing a large number of samples to be assayed in 1 day. Following reduction, the reducing agent is easily extracted with ether, eliminating interfer-
ing peaks in the chromatogram. Since ascorbic acid elutes in approximately 2 min, the run time for each sample can be as short as 4 min, permitting 14 or 15 samples to be analyzed in an hour using an autosampler. Because of its high sensitivity, the assay is especially suited to biological samples, where only small amounts of dehydroascorbic acid may be present and assay specificity is important. If a biological sample is thought to contain more than 1000 pmol/lO ~1 (100 mM) dehydroascorbic acid, an aliquot can be diluted with 60% methanol-l IYIM EDTA and reduced for assay with 2,3dimercapto-1-propanol. Stability of ascorbic acid in samples for both short and long periods of time has been recently reported (21). We have now further increased this stability by use of a refrigeration unit attached to the autosampler. The use of a cooling device allows the temperature of the autosampler to be maintained at approximately 2°C. Ascorbic acid at this temperature is stable for at least 12 h at a concentration as low as 2 pmol/injection. This allows for the unattended analysis of a large number of samples. Thus 48-96 ascorbic acid samples (depending on the full capacity of autosampler carousel) can be prepared at a time and put in the autosampler for analysis. The system can be used overnight which provides great versatility to the system. Kutnink et al. (23) used isoascorbic acid as an internal standard of ascorbic acid oxidation, to permit correction for intraassay ascorbic acid instability. An advantage of the method presented here is that no correction is needed for ascorbic acid instability, since ascorbic acid remains stable under the assay conditions. Other dehydroascorbic acid assays have drawbacks when they are used with biological samples. Sensitivity
LIQUID
CHROMATOGRAPHY
OF
in several systems is limited (10,12,13,16-19). Sensitivity can be increased by post-column reduction of dehydroascorbic acid or by derivatizing dehydroascorbic acid (24,25). However, derivatization can affect the assay specificity (11). Post-column reduction and most other assays do not account either for dehydroascorbic acid recovery or for sample instability, a recurrent problem with ascorbic acid and dehydroascorbic acid (25,26). The dehydroascorbic acid assay described here does not have these limitations, and is suitable for measuring small amounts of dehydroascorbic acid in biological samples. This assay is currently being used in our laboratory to determine whether dehydroascorbic is involved in the mechanism of ascorbic acid transport in human neutrophils (7,8,22).
was supported
by National
Research
Service
REFERENCES 1. Hornig,
D. (1975)
Ann.
N.Y.
Acad.
Sci. 258,
103-118.
2. Englard, S., and Seifter, S. (1986) Annu. Reu. Nutr. 3. Lewin, S. (1976) Vitamin C: Its Molecular Biology Potential, pp. 5-39, Academic Press, New York. 4. Fox,
7. Hornig, D., Weiser, Acta 32,33-39.
F. W., and Levy,
L. P. (1936)
Biochem.
6, 365-406. and Medical
J. 30,211-217.
5. Todhunter, E., McMillan, T., and Ehmke, D. (1950) J. Nutr. 42, 297-308. 6. Som, S., Basu, S., Mukherjee, D., Deb, S., Choudhury, P. R., Mukherjee, S., Chatterjee, S. N., and Chatterjee, I. B. (1981) Metabolism 30, 572-577.
23
ACID H., Weber,
F., and Wiss,
8. Bigley, 1092.
R. H., and Stankova,
L. (1974)
9. Tirrell, 186.
J. G., and Westhead,
E. W. (1979)
0. (1971)
J. Exp.
Clin. Chim.
139, 1084
Med.
Neuroscience
4, 181-
10. Stahl, R. L., Farber, C. M., Liebes, L. F., and Silber, R. (1985) Cancer Res. 45,6507-6512. 11. Cooke, J. R., and Moxon, R. E. D. (1981) in Vitamin C (Counsell, J. N., and Hornig, D., Eds.), pp. 167-198, Applied Science, London. 12. Deutsch, M. J., and Weeks, 48, 1248. 13. Finley,
J. W., and Duang,
14. Rose, 145.
R. C., and Nahrwold,
15. Kim,
H. J. (1989)
19. Doner, 230.
C. E. (1965) E. (1981)
Ofi Anal.
C. A. (1943)
O., and Brassard, M. (1980) Clin. L. W., and Hicks,
J. Assoc.
J. Chromatogr.
D. (1981)
J. Assoc.
16. Roe, J. H., and Keuther, 17. Pelletier, 18. Okamura,
ACKNOWLEDGMENT Philip W. Washko Award DK08060-03.
DEHYDROASCORBIC
Anal. Chem. J. Biol.
Off. And.
Chem.
207,449-453.
Biochem.
114, 140-
72,681-686. Chem.
147,399-407.
R. (1977) J. Food Sci. 42,1471-1477. Chim. Acta 103, 259-268. K. B. (1981)
Anal.
Biochem.
115,225-
20. Farber, C. M., Kanengiser, S., Stahl, R., Liebes, L., and Silber, R. (1983) Anal. Biochem. 134,355-360. 21. Washko, P. W., Hartzell, W. O., and Levine, M. (1989) Anal. Biothem. 181,276-282. 22. Washko, P., Rotrosen, D., and Levine, M. (1989) J. Biol. Chem. 264, 18,996-19,002. 23. Kutnink, M. A., Hawkes, W. C., Schaus, E. E., and Omaye, S. T. (1987) Anal. Biochem. 166,424-430. 24. Iwata, T., Yamaguchi, M., Hara, S., and Nakamura, M. (1985) J. Chromatogr. 344, 351-355. 25. Ziegler, S. J., Meier, B., and Sticher, 0. (1987) J. Chromatogr.
391,419-426. 26. Pachla, 367.
L. A., and Kissinger,
P. T. (1976)
Ad.
Chem.
48,
364-
