- Email: [email protected]
Analysis of Ascorbate in Plant Tissues by High-Performance Capillary Zone Electrophoresis
ANALYTICAL BIOCHEMISTRY ARTICLE NO.
239, 8–19 (1996)
0284
Analysis of Ascorbate in Plant Tissues by HighPerformance Capillary Zone Electrophoresis1 Mark W. Davey, Guy Bauw, and Marc Van Montagu2 Laboratorium voor Genetica, Department of Genetics, Flanders Interuniversity Institute for Biotechnology, Universiteit Gent, K.L. Ledeganckstraat 35, B-9000 Gent, Belgium
Received December 8, 1995
We describe here a simple and rapid capillary electrophoresis method for the determination of ascorbic acid (L-AA) and isoascorbic acid (D-AA) in vegetative tissues. For optimal yields and stabilization, samples are extracted with cold 3% metaphosphoric acid. Hydrophobic contaminants are then removed by passage through a C18 solid-phase extraction cartridge. The analysis itself is performed on a fused silica capillary with 200 mM borate, pH 9, as the carrier electrolyte, using on-line diode array detection over the range 190–350 nm. Quantitation was performed at 260 nm, the uv-absorption maximum for ascorbate at this pH. This method has a minimum detection limit of 84 fmol/ injection and linearity of detector response was observed up to at least 12 pmol/injection. We also describe the influence of electrolyte concentration, pH, and the presence of detergent on separations of L-AA, D-AA, and L-galacturonic acid-1,4-lactone. The protocol has been demonstrated to be suitable for the analysis of L-AA in Arabidopsis, parsley, and mushroom. The method has superior resolution to comparable HPLC separations, a comparable analysis time, but lower sensitivity because of the concentration limitations of the detection system. q 1996 Academic Press, Inc.
acid (L-AA)3 or vitamin C is synthesized by probably all plant species and is widely distributed L-Ascorbic
1 This work was supported by grants from the Belgian Programme on Interuniversity Poles of Attraction (Prime Minister’s Office, Science Policy Programming, #38) and the ‘‘Vlaams Actieprogramma Biotechnologie’’ (ETC). 2 To whom correspondence should be addressed. Fax: 32-92645349. E-mail: [email protected]. 3 Abbreviations used: CZE, capillary zone electrophoresis; D-AA, D-isoascorbic acid/D-erythorbic acid/D-araboascorbic acid; EOF, electroosmotic flow; HPCE, high-performance capillary electrophoresis; L-AA, L-ascorbic acid; L-DHAA, 1-dehydro-L-ascorbic acid; L-GL, Lgalacturonic acid-1,4-lactone; MPA, metaphosphoric acid; RP-HPLC, reversed-phase HPLC; SPE, solid-phase extraction; TCA, trichloroacetic acid.
throughout both the plant and animal kingdoms. In those animals that lack the enzyme L-gulono-1,4-lactone oxidase, L-AA is an obligate component of the diet, and a deficiency of this vitamin leads to development of the symptoms of scurvy. The physiological activity of L-AA is dependent upon its ready oxidation to 1dehydro-L-ascorbic acid (L-DHAA), and most of the functions of L-AA in both plant and animal metabolism can be explained on the basis of its ability to act as a reducing agent in a variety of biochemical processes. These include the detoxification of peroxide, protection against the oxidizing effect of free radicals, and the normal functioning of the photosynthetic apparatus (1, 2). It is also a cofactor for several enzymes including prolyl hydroxylase and myrosinase. Concentrations of L-AA vary widely, with generally the highest levels being found in tissues undergoing rapid growth and development (0.1–2 mg L-AA/g fresh wt) (3), but in certain fruits it can account for up to 3% of the fresh weight (4). There are also intercellular variations in L-AA levels, with highest concentrations being found associated with the chloroplasts (5). L-DHAA concentrations are generally low in plants at less than 5%, but this balance may be disturbed by, for example, mechanical damage. The L-AA/L-DHAA ratio is thus an important indicator of plant stress or disturbances to the metabolism. Any method for the quantitation of L-AA must take into account the conditions that lead to oxidation and degradation of this component, including temperature, pH, and the presence of oxygen and metals, and of stabilizing agents (6). Here we report a rapid and simple HPCE procedure for L-AA analysis from vegetative tissues, using a fused silica capillary. We also make qualitative and quantitative comparisons with HPLC analyses of the same tissue. This differs from other reported capillary electrophoresis methods for L-AA analysis in that the capillary does not require coating (7) and that we use optimal procedures for the extraction and stabilization of ascorbate from plant tissues.
8
AID
0003-2697/96 $18.00 Copyright q 1996 by Academic Press, Inc. All rights of reproduction in any form reserved.
AB 9633
/
6m1a$$$$$1
07-02-96 19:21:14
aba
AP: Anal Bio
CAPILLARY ZONE ELECTROPHORESIS OF ASCORBATE IN PLANTS
EXPERIMENTAL
Reagents EDTA.2Na, Tris base, and sodium hydroxide were purchased from Janssens Pharmaceutica (Beerse, Belgium); TCA, L-AA, and MPA were from Aldrich Chemical (Milwaukee, WI); and HCl was from Merck (Darmstadt, Germany). Isoascorbic acid was obtained from J. T. Baker (Phillipsburg, NJ) and HPLC-grade methanol from Carlo Erba (Milano, Italy). C18 SPE cartridges and 0.22-mm filters were from Millipore (Bedford, MA) and ascorbic acid oxidase was from Boehringer (Mannheim, Germany). Extraction Samples of plant tissue were collected, weighed, and snap frozen in liquid nitrogen. Aliquots (200 mg) of tissue were then pulverized in liquid nitrogen and extracted twice with a total volume of 2.5 ml of 3% MPA/ 1 mM EDTA. After centrifugation, 2 ml of the extract was passed through a preconditioned C18 SPE cartridge, of which only the last 500 ml was kept for sameday analysis by HPLC or HPCE. HPCE Separations were performed on a Model P/ACE 5500, HPCE system (Beckman Instruments, Fullerton, CA), fitted with a diode array detector. The instrument was controlled and data were collected and analyzed by a computer running the Beckman GOLD software, version 8.10. Data were collected at a sampling rate of 2 Hz. Injections were made under hydrostatic (N2) pressure (0.5 psi) onto a 57-cm (50 cm to detection window) 1 75-mm-i.d. fused silica capillary thermostatted at 257C. After installation, the capillary was conditioned by rinsing under pressure (0.5 psi) for 10 min each with 1 M HCl, HPLC-grade water, 1 M NaOH, HPLC-grade water, and finally carrier electrolyte. Constant-voltage HPCE was carried out at /25 kV (439 V/cm) using 200 mM borate, pH 9, as the carrier electrolyte. This gave a current of 103 mA. Samples in 3% MPA/1 mM EDTA were loaded under hydrostatic pressure for 3–10 s. The capillary was hydrostatically prerinsed with carrier electrolyte for 3 min before each analysis and regenerated for use after each analysis by hydrostatic rinses with 0.2 M NaOH (2 min) and two rinses with HPLCgrade water (separate buffer vials) for 1 min each. All buffers were prepared in HPLC-grade water, filtered through a 0.22-mm filter, and helium degassed. HPLC HPLC was carried out using a 600E pump (Waters, Milford, MA) and a Waters 484 variable wavelength
AID
AB 9633
/
6m1a$$$$$1
07-02-96 19:21:14
9
uv detector. Injections were made manually using a Rheodyne 7120 injector (Cotati, CA) and a 20-ml injection loop onto a C18 , 3-mm particle size, 250 1 4.6-mmi.d., RP-HPLC column (Bio-Rad, Hercules, CA). Separations were carried out isocratically at 800 ml/min with phosphoric acid/0.1 mM EDTA, pH 2.5, as the mobile phase. Data were collected and analyzed using the Maxima 820 data integration software (Waters). Identity of L-AA The identity of the L-AA peak was confirmed by migration times, by its characteristic uv-absorption spectrum (absorption max Å 260 nm), by coelution with authentic L-AA, and by incubation of extracts with ascorbic acid oxidase as follows: to 100 ml of MPA extract was added 2 ml of 8 M NaOH to neutralize the sample and then 5 ml of ascorbate acid stock solution (5 mg/500 ml 100 mM phosphate buffer, pH 5.6). After incubation at room temperature for 15 min, samples were acidified with 2 ml of 2.5 M o-phosphoric acid and analyzed. RESULTS AND DISCUSSION
The extraction conditions used in the analysis of LAA are a vital aspect of the work because, as pointed out by others (8–10), many workers have failed to adequately address the problems of L-AA instability and of extraction efficiencies. This was particularly relevant here as the majority of published methods deal with the L-AA analysis of relatively simple tissue matrices such as blood, plasma, and fruit juice (11–19). Of the commonly used extraction solvents (oxalate, TCA, perchlorate, phosphate, etc.), we found that, as previously (20), only MPA adequately fulfilled our requirements. In addition, we also saw that extraction with 10% TCA, as reported in two recent HPCE publications (7, 19), results in the loss of É40% of the total L-AA content, compared to extraction with 3% MPA. This phenomenon has been observed by others (9, 20–22), and supporting data will be published elsewhere. Because of the well-known ability of borate to complex diol-compounds and carbohydrates (23–26), as well as its wide buffering range when used in conjunction with phosphate, borate and borate–phosphate buffers were our first choice of carrier electrolyte for the separation. In addition, the use of borate buffers has been reported to increase the uv absorbance of underivatized monoand oligosaccharides at 195 nm by 2- to 20-fold (27). The suitability of different buffer-electrolyte systems was assessed on the basis of their ability to resolve standard solutions of L-AA and D-AA and of the biosynthetic precursor of L-AA, L-GL (Fig. 1). Concentration of Carrier Electrolyte In 50 mM borate buffer, pH 8.8, under an applied potential of /30 kV, good separation of a 0.1 mg/ml
aba
AP: Anal Bio
10
DAVEY, BAUW, AND VAN MONTAGU
FIG. 1. Chemical structures of L-AA, D-AA, and L-GL.
solution of L-AA and D-AA in nanopure water was obtained. However, when the standards were dissolved in 3% MPA/1 mM EDTA (the extraction buffer), there was a considerable loss of resolution and peak broadening. This is essentially the reverse of the effect that occurs during sample stacking, i.e., the low relative concentration of the carrier electrolyte leads to dilution of the sample due to differential migration rates of the sample and the sample buffer (28). To counteract the anti-stacking effects described above, we decided to increase the carrier electrolyte concentration rather than dilute the sample and risk difficulties with sample stability. Doubling the buffer concentration to 100 mM borate, pH 9, /30 kV and 67 mA, improved the resolution of L-AA and D-AA in nanopure, with migration times relative to the EOF marker of 1.96 and 2.22 min, respectively, but more importantly, the loss of resolution observed with samples injected in MPA was avoided (Fig. 2). There was good separation of biological extracts under these conditions, but as with the standards, there was still some peak ‘‘fronting,’’ indicating a mismatch between the ionic strengths of the background electrolyte and the sample solvent. The inclusion of 10 mM phosphate in 100 mM borate at /30 kV led to unacceptably high currents of 167 mA at 30 kV, broad peaks, and a loss of resolution in comparison to 100 mM borate alone. This is in contrast to the reported improvement in selectivity offered by phosphate/borate mixes in the CZE of carbohydrates (25). A further increase in carrier electrolyte concentration to 200 mM improved peak symmetry and resolution for both standard solutions and biological extracts, allowing easier quantitation of all components present. Along with the increase in the concentration of the background electrolyte to 200 mM, the applied potential was decreased from 30 to 25 kV to avoid Joule heating effects associated with high through-currents. It was also necessary to regularly replenish the carrier electrolyte to prevent migration time drift and this was done after every 10 analyses. Under these conditions,
AID
AB 9633
/
6m1a$$$$$1
07-02-96 19:21:14
L-GL migrated shortly after the EOF marker, as a broad symmetrical peak. Reproducibility Initially we experienced considerable variation in the between-run migration times of standards, and while expressing migration times relative to the EOF marker (the relative migration time) provided more accurate results, absolute peak identification in biological extracts was still difficult. These problems were avoided by increasing the concentration of the postrun NaOH rinse from 0.1 to 0.2 M. Under these conditions, the variation in the absolute migration times for each batch of buffer was within 0.1–0.2%, e.g., L-AA 5.698 { 0.1% and D-AA 6.148 { 0.15%. To prevent alkalinization of the postrun rinses and of the samples through crosscontamination, we used two separate 1-min postrun rinses with nanopure water (rather than one 2-min wash). Influence of pH The pH of the carrier electrolyte has a dramatic influence on the separation of L-AA and D-AA. In 20 mM phosphate, pH 7, there was no observed migration of L-AA because presumably the EOF was too slow to overcome the migration of anionic L-AA/D-AA toward the anode. At pH 7.5 in 200 mM borate, standard solutions dissolved in 3% MPA/EDTA (extraction buffer) were unresolved, with poor peak shape and considerable fronting. Baseline resolution was recovered at pH 8, and increasing the pH to 8.5 and 9.0 progressively increases the migration times of these components and the resolution of the separation. These results are summarized in Table 1. At higher pH, the ionization of the silanol groups of the capillary wall is increased. This leads to an increased z potential, the accumulation of cations near the capillary wall surface, and thus a higher EOF, as discussed in (29). We would therefore expect migration times to decrease with increasing pH. However, in competition with this effect is the influence of pH on the equilibrium between the free and borate-
aba
AP: Anal Bio
CAPILLARY ZONE ELECTROPHORESIS OF ASCORBATE IN PLANTS
11
FIG. 2. Influence of carrier electrolyte concentration on the resolution of L-AA and D-AA dissolved in extraction buffer. Three-second (17.7 nl) hydrostatic injections of 0.05 mg/ml solution of L-AA and D-AA dissolved in 3% MPA/1 mM EDTA. Ultraviolet detection carried out at 194 nm. Separations performed on a 57-cm (50 cm to detector) 1 75-mm-i.d. fused silica capillary, at /30 kV and 257C. Carrier electrolyte was 50 mM borate, pH 9 (A); 100 mM borate, pH 9 (B); and 200 mM borate, pH 9 (C).
complexed forms of the analyte, which favors complex formation under alkaline conditions (26). Therefore, the increased selectivity observed at high pH is the
AID
AB 9633
/
6m1a$$$$$1
07-02-96 19:21:14
result of at least two competing influences: (i) a higher bulk liquid transfer toward the cathode due to the increased EOF and (ii) the increased proportion of carbo-
aba
AP: Anal Bio
12
DAVEY, BAUW, AND VAN MONTAGU
FIG. 2—Continued
hydrate present migrating toward the anode in the form of anionic borate complexes. The EOF is of course still sufficiently high to ensure that these components are swept past the detector window. The relative migration of L-GL increases slightly with pH, suggesting a certain degree of interaction with borate. This is despite the fact that L-GL does not contain the characteristic cis diol function which is preferred for borate complexation (26, 27) and which is present on C2 and C3 of D-AA and L-AA. It is well known that L-AA is unstable at alkaline pH levels, being rapidly oxidized to L-DHAA which can
TABLE 1
Influence of pH on the Relative Mobilities of L-AA, D-AA, and L-GL Relative migration time pH
L-AA
D-AA
L-GL
7.5 8 8.5 9
1.68 2.20 3.99 5.47
1.68 2.20 4.54 6.62
0.15 0.48 0.89 1.76
Note. Relative migration time calculated as the difference between the absolute migration time and the migration time of the EOF marker (acetone) under these conditions. Each value represents the mean of two analyses.
AID
AB 9633
/
6m1a$$$$$1
07-02-96 19:21:14
in turn irreversibly hydrolyze to 2,3-diketogulonic acid. Despite this we saw no evidence for sample degradation over the time course of the analysis. Conceivably complexation with borate stabilizes L-AA at these relatively high pH levels. Influence of SDS To try and improve the resolution of the neutral components of extracts, we investigated the influence of SDS on separations. Using 100 mM borate, pH 9, as the carrier electrolyte and between 0 and 80 mM SDS, injections of L-AA, D-AA, and L-GL standards dissolved in MPA/EDTA were made for each buffer composition. These results are summarized in Table 2. SDS clearly has a significant influence on the migration of L-AA and D-AA, and the relative migration times and resolutions of both progressively increase with increasing concentrations of detergent. This suggests that L-AA and D-AA are able to interact with SDS micelles either directly or as borate complexes and that their migration is then selectively retarded due to the tendency of the anionic SDS micelles to migrate toward the anode. However, SDS appears to have little influence on the migration of L-GL, and the inclusion of 0–80 mM SDS is actually detrimental to the resolution of biological extracts. As a result of this, separations were standardized as /25 kV with 200 mM borate, pH 9.
aba
AP: Anal Bio
CAPILLARY ZONE ELECTROPHORESIS OF ASCORBATE IN PLANTS
13
Biological Extracts—Comparison between HPLC and HPCE Separations
TABLE 2
Influence of Concentration of SDS on Separations of L-AA, D-AA, and L-GL Relative migration time Concentration SDS (mM)
L-AA
D-AA
L-GL
0 20 40 60 80
2.07 2.18 3.97 3.88 3.98
2.39 2.53 4.64 4.60 4.75
0.61 nd 0.89 0.56 0.50
Note. Relative migration time calculated as the difference between the absolute migration time and the migration time of the EOF marker (acetone) under these conditions. Each value represents the mean of two analyses.
Influence of Ascorbate Oxidase The identity of the L-AA peak in biological extracts was confirmed by co-injection of standards and by incubation with ascorbate oxidase (Fig. 3). The identity of some of the other peaks of the electropherogram was determined by injection of appropriate controls. Note that the slight shift in peak migration times following incubation with ascorbate oxidase presumably reflects a difference in the pH or ionic strength of the samples following acidification.
A direct quantitative comparison between the L-AA tissue values calculated following HPCE and HPLC analysis of the same tissue showed good agreement. For separate extractions of the same parsley sample, we obtained values of 2332 { 30 mg/g fresh wt (HPCE) and 2193 { 59 mg/g fresh wt (HPLC). The percentage relative error between the two methods is É6%. The L-AA values calculated for soil-grown Arabidopsis samples are 32.24 { 6.4 mg/g fresh wt (HPCE) and 32.05 { 4 mg/g fresh wt (HPLC). Qualitatively, however, HPCE offers a far superior resolution to HPLC for analysis of the same extracts (Fig. 5). In general, we found that the resolution offered by HPLC was more than adequate for most applications. However, in certain instances such as the analysis of mushroom (Fig. 4) and yeast extracts (data not shown), where low concentrations of L-AA or D-AA are combined with a more complicated uv profile, accurate LAA analysis was only possible using HPCE. One important point to note is that under acidic conditions (the HPLC mobile phase has a pH of É2.5) ascorbate has an absorption maximum of 248 nm, but that under basic conditions (HPCE) this absorption maximum shifts to 260 nm (30). CONCLUSIONS
Quantitation Tissue concentrations of L-AA were calculated using standard curves constructed at 260 nm using different concentrations of L-AA and D-AA dissolved in 3% MPA/ 1 mM EDTA. Minimum mass detection limits were calculated using injection volumes obtained from the Beckman Applications Sheet W23. Although these injection volumes assume a sample viscosity equal to that of water, the fact that standard curves were prepared in the same solvent as the extracts allows direct comparison of the values obtained, even though the absolute injection volumes will not be correct. This is further supported by the good agreement obtained between HPCE and HPLC determinations of L-AA in the same tissue. Minimum detection limits set at approximately three times baseline noise were in the region of 15 pg injected (5.3 mM), and linearity of detection was observed between at least 15 and 2120 pg injected (84 fmol–12 pmol) (y Å 107.4x 0 1.95, r Å 0.999, n Å 24) (Fig. 4).
The HPCE of MPA extracts of plant tissue represents a simple and rapid alternative to HPLC for the analysis of L-AA and D-AA. Initial difficulties encountered with the acidic extraction solvent could be easily compensated for by increasing the concentration of the carrier electrolyte. This is important because like others (20), we have found that MPA is the optimal solvent for the extraction and stabilization of L-AA from plant tissue. The use of borate as carrier electrolyte permits detection of neutral carbohydrates at low uv wavelengths, and because L-AA, D-AA, and probably L-GL are all able to form anionic complexes with borate, resolutions are enhanced due to the competing influences of the EOF and anodic migration of borate complexes at high pH levels. Generally a disadvantage to the use of a high carrier electrolyte concentration is an increased background absorption, and this can compromise the sensitivity of detection. This did not appear to be a significant problem at 260 nm. However, a high carrier electrolyte concentration also has the advantages of
FIG. 3. Influence of ascorbic acid oxidase incubation on electropherograms of parsley extract. Five-second (29.5 nl) hydrostatic injection of parsley extract before (A) and after (B) incubation with ascorbic acid oxidase as described under Experimental. Ultraviolet absorbance set out at 260 nm. Carrier electrolyte of 200 mM borate, pH 9, under separation potential of /25 kV. The capillary was thermostatted at 257C. Inset shows absorbance spectrum obtained for the L-AA peak migrating at 7.8 min.
AID
AB 9633
/
6m1a$$$$$1
07-02-96 19:21:14
aba
AP: Anal Bio
14
AID
DAVEY, BAUW, AND VAN MONTAGU
AB 9633
/
6m1a$$9633
07-02-96 19:21:14
aba
AP: Anal Bio
CAPILLARY ZONE ELECTROPHORESIS OF ASCORBATE IN PLANTS
15
FIG. 4. Minimum detection limits of L-AA and D-AA. Three-second (17.7 nl) hydrostatic injection of standard solution of 1 mg/ml L-AA and D-AA in MPA/EDTA. Peaks represent injections of 84 fmol of sample. All other conditions as in Fig. 3.
higher sample capacities and it minimizes interactions between the analytes and the capillary wall. Although we obtained resolution of L-GL, and although biological extracts showed the presence of a LGL-like peak when extracts were monitored at 194 nm, the poor relative migration times of L-GL, the lack of a characteristic absorption spectrum, and the fact that it could not be resolved from its stereoisomer D-galacturonic acid-1,4-lactone (data not shown) meant that definitive identification of this component in biological extracts was not possible under these conditions. Direct comparison with ascorbate analyses by HPLC shows that CZE has notable advantages in terms of peak
resolution and that there is more than adequate resolution to allow a decrease in analysis times by the use of a shorter capillary length. The minimum concentration limits in CZE are considerably higher with HPLC, but are still low enough to allow measurement of L-AA levels in all extracts examined. However, it is unlikely that the sensitivity of the CZE separation could be much improved by sample stacking, due to the high concentration of the extraction solvent. The use of D-AA as an internal standard to improve the accuracy of determinations as proposed by Koh et al. (19) will not always be appropriate as we have observed D-AA to be a significant component of certain vegetative extracts.
FIG. 5. Comparison between HPCE and HPLC analysis of biological extracts. All conditions for HPCE as described in the legend to Fig. 2, except that injections of mushroom and Arabidopsis extracts were carried out for 10 s under hydrostatic pressure. Isocratic HPLC analysis with uv detection at 248 nm is as described under Experimental. Peaks labeled L-AA and D-AA indicate elution/migration positions of L-ascorbic acid and D-isoascorbic acid, respectively. HPLC (A) and HPCE (B) of parsley extract; HPLC (C) and HPCE (D) analysis of mushroom extract; HPLC (E) and HPCE (F) analysis of Arabidopsis extract.
AID
AB 9633
/
6m1a$$$$$1
07-02-96 19:21:14
aba
AP: Anal Bio
16
AID
DAVEY, BAUW, AND VAN MONTAGU
AB 9633
/
6m1a$$9633
07-02-96 19:21:14
aba
AP: Anal Bio
CAPILLARY ZONE ELECTROPHORESIS OF ASCORBATE IN PLANTS
FIG. 5—Continued
AID
AB 9633
/
6m1a$$9633
07-02-96 19:21:14
aba
AP: Anal Bio
17
18
DAVEY, BAUW, AND VAN MONTAGU
FIG. 5—Continued
AID
AB 9633
/
6m1a$$9633
07-02-96 19:21:14
aba
AP: Anal Bio
CAPILLARY ZONE ELECTROPHORESIS OF ASCORBATE IN PLANTS
ACKNOWLEDGMENTS The authors thank Beckman Instruments for the generous loan of the capillary electrophoresis equipment, Dr. Jens Oostergaard for critical reading of the manuscript, and Martine De Cock for preparation of the manuscript.
REFERENCES 1. Arrigoni, O. (1994) J. Bioenerg. Biomembr. 26, 407–419. 2. Foyer, C. (1993) in Antioxidants in Higher Plants (Alscher, R. G., and Hess, J. L., Eds.), pp. 31–58, CRC Press, Boca Raton, FL. 3. Mapson, L. W. (1970) in The Biochemistry of Fruits and Their Products, Vol. 1 (Food Science and Technology, a Series of Monographs) (Hulme, A. C., Ed.), pp. 369–384, Academic Press, London. 4. Ishihata, K., and Ito, S. (1994) Jpn. J. Trop. Agric. 38, 113–118. 5. Loewus, F. A., and Loewus, M. W. (1987) CRC Crit. Rev. Plant Sci. 5, 101–119. 6. Fleming, J. F., and Bensch, K. G. (1983) Int. J. Pept. Protein Res. 22, 355–361. 7. Chiari, M., Nesi, M., Carrea, G., and Righetti, P. G. (1993) J. Chromatogr. 645, 197–200. 8. Washko, P. W., Welch, R. W., Dhariwal, K. R., Wang, Y., and Levine, M. (1992) Anal. Biochem. 204, 1–14. 9. Kampfenkel, K., Van Montagu, M., and Inze´, D. (1995) Anal. Biochem. 225, 165–167. 10. Lykkesfeldt, J., Loft, S., and Poulsen, H. E. (1995) Anal. Biochem. 229, 329–335. 11. Kutnink, M. A., Hawkes, W. C., Schaus, E. E., and Omaye, S. T. (1987) Anal. Biochem. 166, 424–430.
AID
AB 9633
/
6m1a$$$$$1
07-02-96 19:21:14
19
12. Ross, M. A. (1994) J. Chromatogr. B 657, 197–200. 13. Okamura, M. (1980) Clin. Chim. Acta 103, 259–268. 14. Liau, L. S., Lee, B. L., New, A. L., and Ong, C. N. (1993) J. Chromatogr. 612, 63–70. 15. Nagy, E., and Degrell, I. (1989) J. Chromatogr. 497, 276–281. 16. Iwase, H., and Ono, I. (1993) J. Chromatogr. A 654, 215–220. 17. Iwase, H. (1992) J. Chromatogr. 606, 277–280. 18. Nisperos-Carriedo, M., Buslig, B. S., and Shaw, P. E. (1992) J. Agric. Food Chem. 40, 1127–1130. 19. Koh, E. V., Bissell, M. G., and Ito, R. K. (1993) J. Chromatogr. 633, 245–250. 20. Roe, J. H. (1954) in Methods of Biochemical Analysis (Glick, D., Ed.), Vol. 1, pp. 115–139, Academic Press, New York. 21. Tanishima, K., and Kita, M. (1993) J. Chromatogr. 613, 275– 280. 22. Brown, M. R., and Miller, K. A. (1992) J. Appl. Phycol. 4, 205– 215. 23. Kean, K. L. (1966) J. Lipid Res. 7, 449–453. 24. Honda, S., Iwase, S., Makino, A., and Fujiwara, S. (1989) Anal. Biochem. 176, 72–77. 25. Oefner, P. J., and Chiesa, C. (1994) Glycobiology 4, 397–412. 26. Landers, J. P., Oda, R. P., and Schuchard, M. D. (1992) Anal. Biochem. 64, 2847–2851. 27. Hoffstetter-Kuhn, S., Paulus, A., Gassmann, E., and Widmer, H. M. (1991) Anal. Chem. 63, 1541–1547. 28. Weinberger, R. (1993) in Practical Capillary Electrophoresis (Weinberger, R., Ed.), pp. 197–222, Academic Press, Boston. 29. Weinberger, R. (1993) in Practical Capillary Electrophoresis (Weinberger, R., Ed.), pp. 17–44, Academic Press, Boston. 30. Jaffe, G. M. (1984) in Encyclopedia of Chemical Technology (R. E. Kirk and D. F. Ohtmer, Eds.), Vol. 24, 3rd ed., pp. 8–40, Wiley, New York.
aba
AP: Anal Bio
239, 8–19 (1996)
0284
Analysis of Ascorbate in Plant Tissues by HighPerformance Capillary Zone Electrophoresis1 Mark W. Davey, Guy Bauw, and Marc Van Montagu2 Laboratorium voor Genetica, Department of Genetics, Flanders Interuniversity Institute for Biotechnology, Universiteit Gent, K.L. Ledeganckstraat 35, B-9000 Gent, Belgium
Received December 8, 1995
We describe here a simple and rapid capillary electrophoresis method for the determination of ascorbic acid (L-AA) and isoascorbic acid (D-AA) in vegetative tissues. For optimal yields and stabilization, samples are extracted with cold 3% metaphosphoric acid. Hydrophobic contaminants are then removed by passage through a C18 solid-phase extraction cartridge. The analysis itself is performed on a fused silica capillary with 200 mM borate, pH 9, as the carrier electrolyte, using on-line diode array detection over the range 190–350 nm. Quantitation was performed at 260 nm, the uv-absorption maximum for ascorbate at this pH. This method has a minimum detection limit of 84 fmol/ injection and linearity of detector response was observed up to at least 12 pmol/injection. We also describe the influence of electrolyte concentration, pH, and the presence of detergent on separations of L-AA, D-AA, and L-galacturonic acid-1,4-lactone. The protocol has been demonstrated to be suitable for the analysis of L-AA in Arabidopsis, parsley, and mushroom. The method has superior resolution to comparable HPLC separations, a comparable analysis time, but lower sensitivity because of the concentration limitations of the detection system. q 1996 Academic Press, Inc.
acid (L-AA)3 or vitamin C is synthesized by probably all plant species and is widely distributed L-Ascorbic
1 This work was supported by grants from the Belgian Programme on Interuniversity Poles of Attraction (Prime Minister’s Office, Science Policy Programming, #38) and the ‘‘Vlaams Actieprogramma Biotechnologie’’ (ETC). 2 To whom correspondence should be addressed. Fax: 32-92645349. E-mail: [email protected]. 3 Abbreviations used: CZE, capillary zone electrophoresis; D-AA, D-isoascorbic acid/D-erythorbic acid/D-araboascorbic acid; EOF, electroosmotic flow; HPCE, high-performance capillary electrophoresis; L-AA, L-ascorbic acid; L-DHAA, 1-dehydro-L-ascorbic acid; L-GL, Lgalacturonic acid-1,4-lactone; MPA, metaphosphoric acid; RP-HPLC, reversed-phase HPLC; SPE, solid-phase extraction; TCA, trichloroacetic acid.
throughout both the plant and animal kingdoms. In those animals that lack the enzyme L-gulono-1,4-lactone oxidase, L-AA is an obligate component of the diet, and a deficiency of this vitamin leads to development of the symptoms of scurvy. The physiological activity of L-AA is dependent upon its ready oxidation to 1dehydro-L-ascorbic acid (L-DHAA), and most of the functions of L-AA in both plant and animal metabolism can be explained on the basis of its ability to act as a reducing agent in a variety of biochemical processes. These include the detoxification of peroxide, protection against the oxidizing effect of free radicals, and the normal functioning of the photosynthetic apparatus (1, 2). It is also a cofactor for several enzymes including prolyl hydroxylase and myrosinase. Concentrations of L-AA vary widely, with generally the highest levels being found in tissues undergoing rapid growth and development (0.1–2 mg L-AA/g fresh wt) (3), but in certain fruits it can account for up to 3% of the fresh weight (4). There are also intercellular variations in L-AA levels, with highest concentrations being found associated with the chloroplasts (5). L-DHAA concentrations are generally low in plants at less than 5%, but this balance may be disturbed by, for example, mechanical damage. The L-AA/L-DHAA ratio is thus an important indicator of plant stress or disturbances to the metabolism. Any method for the quantitation of L-AA must take into account the conditions that lead to oxidation and degradation of this component, including temperature, pH, and the presence of oxygen and metals, and of stabilizing agents (6). Here we report a rapid and simple HPCE procedure for L-AA analysis from vegetative tissues, using a fused silica capillary. We also make qualitative and quantitative comparisons with HPLC analyses of the same tissue. This differs from other reported capillary electrophoresis methods for L-AA analysis in that the capillary does not require coating (7) and that we use optimal procedures for the extraction and stabilization of ascorbate from plant tissues.
8
AID
0003-2697/96 $18.00 Copyright q 1996 by Academic Press, Inc. All rights of reproduction in any form reserved.
AB 9633
/
6m1a$$$$$1
07-02-96 19:21:14
aba
AP: Anal Bio
CAPILLARY ZONE ELECTROPHORESIS OF ASCORBATE IN PLANTS
EXPERIMENTAL
Reagents EDTA.2Na, Tris base, and sodium hydroxide were purchased from Janssens Pharmaceutica (Beerse, Belgium); TCA, L-AA, and MPA were from Aldrich Chemical (Milwaukee, WI); and HCl was from Merck (Darmstadt, Germany). Isoascorbic acid was obtained from J. T. Baker (Phillipsburg, NJ) and HPLC-grade methanol from Carlo Erba (Milano, Italy). C18 SPE cartridges and 0.22-mm filters were from Millipore (Bedford, MA) and ascorbic acid oxidase was from Boehringer (Mannheim, Germany). Extraction Samples of plant tissue were collected, weighed, and snap frozen in liquid nitrogen. Aliquots (200 mg) of tissue were then pulverized in liquid nitrogen and extracted twice with a total volume of 2.5 ml of 3% MPA/ 1 mM EDTA. After centrifugation, 2 ml of the extract was passed through a preconditioned C18 SPE cartridge, of which only the last 500 ml was kept for sameday analysis by HPLC or HPCE. HPCE Separations were performed on a Model P/ACE 5500, HPCE system (Beckman Instruments, Fullerton, CA), fitted with a diode array detector. The instrument was controlled and data were collected and analyzed by a computer running the Beckman GOLD software, version 8.10. Data were collected at a sampling rate of 2 Hz. Injections were made under hydrostatic (N2) pressure (0.5 psi) onto a 57-cm (50 cm to detection window) 1 75-mm-i.d. fused silica capillary thermostatted at 257C. After installation, the capillary was conditioned by rinsing under pressure (0.5 psi) for 10 min each with 1 M HCl, HPLC-grade water, 1 M NaOH, HPLC-grade water, and finally carrier electrolyte. Constant-voltage HPCE was carried out at /25 kV (439 V/cm) using 200 mM borate, pH 9, as the carrier electrolyte. This gave a current of 103 mA. Samples in 3% MPA/1 mM EDTA were loaded under hydrostatic pressure for 3–10 s. The capillary was hydrostatically prerinsed with carrier electrolyte for 3 min before each analysis and regenerated for use after each analysis by hydrostatic rinses with 0.2 M NaOH (2 min) and two rinses with HPLCgrade water (separate buffer vials) for 1 min each. All buffers were prepared in HPLC-grade water, filtered through a 0.22-mm filter, and helium degassed. HPLC HPLC was carried out using a 600E pump (Waters, Milford, MA) and a Waters 484 variable wavelength
AID
AB 9633
/
6m1a$$$$$1
07-02-96 19:21:14
9
uv detector. Injections were made manually using a Rheodyne 7120 injector (Cotati, CA) and a 20-ml injection loop onto a C18 , 3-mm particle size, 250 1 4.6-mmi.d., RP-HPLC column (Bio-Rad, Hercules, CA). Separations were carried out isocratically at 800 ml/min with phosphoric acid/0.1 mM EDTA, pH 2.5, as the mobile phase. Data were collected and analyzed using the Maxima 820 data integration software (Waters). Identity of L-AA The identity of the L-AA peak was confirmed by migration times, by its characteristic uv-absorption spectrum (absorption max Å 260 nm), by coelution with authentic L-AA, and by incubation of extracts with ascorbic acid oxidase as follows: to 100 ml of MPA extract was added 2 ml of 8 M NaOH to neutralize the sample and then 5 ml of ascorbate acid stock solution (5 mg/500 ml 100 mM phosphate buffer, pH 5.6). After incubation at room temperature for 15 min, samples were acidified with 2 ml of 2.5 M o-phosphoric acid and analyzed. RESULTS AND DISCUSSION
The extraction conditions used in the analysis of LAA are a vital aspect of the work because, as pointed out by others (8–10), many workers have failed to adequately address the problems of L-AA instability and of extraction efficiencies. This was particularly relevant here as the majority of published methods deal with the L-AA analysis of relatively simple tissue matrices such as blood, plasma, and fruit juice (11–19). Of the commonly used extraction solvents (oxalate, TCA, perchlorate, phosphate, etc.), we found that, as previously (20), only MPA adequately fulfilled our requirements. In addition, we also saw that extraction with 10% TCA, as reported in two recent HPCE publications (7, 19), results in the loss of É40% of the total L-AA content, compared to extraction with 3% MPA. This phenomenon has been observed by others (9, 20–22), and supporting data will be published elsewhere. Because of the well-known ability of borate to complex diol-compounds and carbohydrates (23–26), as well as its wide buffering range when used in conjunction with phosphate, borate and borate–phosphate buffers were our first choice of carrier electrolyte for the separation. In addition, the use of borate buffers has been reported to increase the uv absorbance of underivatized monoand oligosaccharides at 195 nm by 2- to 20-fold (27). The suitability of different buffer-electrolyte systems was assessed on the basis of their ability to resolve standard solutions of L-AA and D-AA and of the biosynthetic precursor of L-AA, L-GL (Fig. 1). Concentration of Carrier Electrolyte In 50 mM borate buffer, pH 8.8, under an applied potential of /30 kV, good separation of a 0.1 mg/ml
aba
AP: Anal Bio
10
DAVEY, BAUW, AND VAN MONTAGU
FIG. 1. Chemical structures of L-AA, D-AA, and L-GL.
solution of L-AA and D-AA in nanopure water was obtained. However, when the standards were dissolved in 3% MPA/1 mM EDTA (the extraction buffer), there was a considerable loss of resolution and peak broadening. This is essentially the reverse of the effect that occurs during sample stacking, i.e., the low relative concentration of the carrier electrolyte leads to dilution of the sample due to differential migration rates of the sample and the sample buffer (28). To counteract the anti-stacking effects described above, we decided to increase the carrier electrolyte concentration rather than dilute the sample and risk difficulties with sample stability. Doubling the buffer concentration to 100 mM borate, pH 9, /30 kV and 67 mA, improved the resolution of L-AA and D-AA in nanopure, with migration times relative to the EOF marker of 1.96 and 2.22 min, respectively, but more importantly, the loss of resolution observed with samples injected in MPA was avoided (Fig. 2). There was good separation of biological extracts under these conditions, but as with the standards, there was still some peak ‘‘fronting,’’ indicating a mismatch between the ionic strengths of the background electrolyte and the sample solvent. The inclusion of 10 mM phosphate in 100 mM borate at /30 kV led to unacceptably high currents of 167 mA at 30 kV, broad peaks, and a loss of resolution in comparison to 100 mM borate alone. This is in contrast to the reported improvement in selectivity offered by phosphate/borate mixes in the CZE of carbohydrates (25). A further increase in carrier electrolyte concentration to 200 mM improved peak symmetry and resolution for both standard solutions and biological extracts, allowing easier quantitation of all components present. Along with the increase in the concentration of the background electrolyte to 200 mM, the applied potential was decreased from 30 to 25 kV to avoid Joule heating effects associated with high through-currents. It was also necessary to regularly replenish the carrier electrolyte to prevent migration time drift and this was done after every 10 analyses. Under these conditions,
AID
AB 9633
/
6m1a$$$$$1
07-02-96 19:21:14
L-GL migrated shortly after the EOF marker, as a broad symmetrical peak. Reproducibility Initially we experienced considerable variation in the between-run migration times of standards, and while expressing migration times relative to the EOF marker (the relative migration time) provided more accurate results, absolute peak identification in biological extracts was still difficult. These problems were avoided by increasing the concentration of the postrun NaOH rinse from 0.1 to 0.2 M. Under these conditions, the variation in the absolute migration times for each batch of buffer was within 0.1–0.2%, e.g., L-AA 5.698 { 0.1% and D-AA 6.148 { 0.15%. To prevent alkalinization of the postrun rinses and of the samples through crosscontamination, we used two separate 1-min postrun rinses with nanopure water (rather than one 2-min wash). Influence of pH The pH of the carrier electrolyte has a dramatic influence on the separation of L-AA and D-AA. In 20 mM phosphate, pH 7, there was no observed migration of L-AA because presumably the EOF was too slow to overcome the migration of anionic L-AA/D-AA toward the anode. At pH 7.5 in 200 mM borate, standard solutions dissolved in 3% MPA/EDTA (extraction buffer) were unresolved, with poor peak shape and considerable fronting. Baseline resolution was recovered at pH 8, and increasing the pH to 8.5 and 9.0 progressively increases the migration times of these components and the resolution of the separation. These results are summarized in Table 1. At higher pH, the ionization of the silanol groups of the capillary wall is increased. This leads to an increased z potential, the accumulation of cations near the capillary wall surface, and thus a higher EOF, as discussed in (29). We would therefore expect migration times to decrease with increasing pH. However, in competition with this effect is the influence of pH on the equilibrium between the free and borate-
aba
AP: Anal Bio
CAPILLARY ZONE ELECTROPHORESIS OF ASCORBATE IN PLANTS
11
FIG. 2. Influence of carrier electrolyte concentration on the resolution of L-AA and D-AA dissolved in extraction buffer. Three-second (17.7 nl) hydrostatic injections of 0.05 mg/ml solution of L-AA and D-AA dissolved in 3% MPA/1 mM EDTA. Ultraviolet detection carried out at 194 nm. Separations performed on a 57-cm (50 cm to detector) 1 75-mm-i.d. fused silica capillary, at /30 kV and 257C. Carrier electrolyte was 50 mM borate, pH 9 (A); 100 mM borate, pH 9 (B); and 200 mM borate, pH 9 (C).
complexed forms of the analyte, which favors complex formation under alkaline conditions (26). Therefore, the increased selectivity observed at high pH is the
AID
AB 9633
/
6m1a$$$$$1
07-02-96 19:21:14
result of at least two competing influences: (i) a higher bulk liquid transfer toward the cathode due to the increased EOF and (ii) the increased proportion of carbo-
aba
AP: Anal Bio
12
DAVEY, BAUW, AND VAN MONTAGU
FIG. 2—Continued
hydrate present migrating toward the anode in the form of anionic borate complexes. The EOF is of course still sufficiently high to ensure that these components are swept past the detector window. The relative migration of L-GL increases slightly with pH, suggesting a certain degree of interaction with borate. This is despite the fact that L-GL does not contain the characteristic cis diol function which is preferred for borate complexation (26, 27) and which is present on C2 and C3 of D-AA and L-AA. It is well known that L-AA is unstable at alkaline pH levels, being rapidly oxidized to L-DHAA which can
TABLE 1
Influence of pH on the Relative Mobilities of L-AA, D-AA, and L-GL Relative migration time pH
L-AA
D-AA
L-GL
7.5 8 8.5 9
1.68 2.20 3.99 5.47
1.68 2.20 4.54 6.62
0.15 0.48 0.89 1.76
Note. Relative migration time calculated as the difference between the absolute migration time and the migration time of the EOF marker (acetone) under these conditions. Each value represents the mean of two analyses.
AID
AB 9633
/
6m1a$$$$$1
07-02-96 19:21:14
in turn irreversibly hydrolyze to 2,3-diketogulonic acid. Despite this we saw no evidence for sample degradation over the time course of the analysis. Conceivably complexation with borate stabilizes L-AA at these relatively high pH levels. Influence of SDS To try and improve the resolution of the neutral components of extracts, we investigated the influence of SDS on separations. Using 100 mM borate, pH 9, as the carrier electrolyte and between 0 and 80 mM SDS, injections of L-AA, D-AA, and L-GL standards dissolved in MPA/EDTA were made for each buffer composition. These results are summarized in Table 2. SDS clearly has a significant influence on the migration of L-AA and D-AA, and the relative migration times and resolutions of both progressively increase with increasing concentrations of detergent. This suggests that L-AA and D-AA are able to interact with SDS micelles either directly or as borate complexes and that their migration is then selectively retarded due to the tendency of the anionic SDS micelles to migrate toward the anode. However, SDS appears to have little influence on the migration of L-GL, and the inclusion of 0–80 mM SDS is actually detrimental to the resolution of biological extracts. As a result of this, separations were standardized as /25 kV with 200 mM borate, pH 9.
aba
AP: Anal Bio
CAPILLARY ZONE ELECTROPHORESIS OF ASCORBATE IN PLANTS
13
Biological Extracts—Comparison between HPLC and HPCE Separations
TABLE 2
Influence of Concentration of SDS on Separations of L-AA, D-AA, and L-GL Relative migration time Concentration SDS (mM)
L-AA
D-AA
L-GL
0 20 40 60 80
2.07 2.18 3.97 3.88 3.98
2.39 2.53 4.64 4.60 4.75
0.61 nd 0.89 0.56 0.50
Note. Relative migration time calculated as the difference between the absolute migration time and the migration time of the EOF marker (acetone) under these conditions. Each value represents the mean of two analyses.
Influence of Ascorbate Oxidase The identity of the L-AA peak in biological extracts was confirmed by co-injection of standards and by incubation with ascorbate oxidase (Fig. 3). The identity of some of the other peaks of the electropherogram was determined by injection of appropriate controls. Note that the slight shift in peak migration times following incubation with ascorbate oxidase presumably reflects a difference in the pH or ionic strength of the samples following acidification.
A direct quantitative comparison between the L-AA tissue values calculated following HPCE and HPLC analysis of the same tissue showed good agreement. For separate extractions of the same parsley sample, we obtained values of 2332 { 30 mg/g fresh wt (HPCE) and 2193 { 59 mg/g fresh wt (HPLC). The percentage relative error between the two methods is É6%. The L-AA values calculated for soil-grown Arabidopsis samples are 32.24 { 6.4 mg/g fresh wt (HPCE) and 32.05 { 4 mg/g fresh wt (HPLC). Qualitatively, however, HPCE offers a far superior resolution to HPLC for analysis of the same extracts (Fig. 5). In general, we found that the resolution offered by HPLC was more than adequate for most applications. However, in certain instances such as the analysis of mushroom (Fig. 4) and yeast extracts (data not shown), where low concentrations of L-AA or D-AA are combined with a more complicated uv profile, accurate LAA analysis was only possible using HPCE. One important point to note is that under acidic conditions (the HPLC mobile phase has a pH of É2.5) ascorbate has an absorption maximum of 248 nm, but that under basic conditions (HPCE) this absorption maximum shifts to 260 nm (30). CONCLUSIONS
Quantitation Tissue concentrations of L-AA were calculated using standard curves constructed at 260 nm using different concentrations of L-AA and D-AA dissolved in 3% MPA/ 1 mM EDTA. Minimum mass detection limits were calculated using injection volumes obtained from the Beckman Applications Sheet W23. Although these injection volumes assume a sample viscosity equal to that of water, the fact that standard curves were prepared in the same solvent as the extracts allows direct comparison of the values obtained, even though the absolute injection volumes will not be correct. This is further supported by the good agreement obtained between HPCE and HPLC determinations of L-AA in the same tissue. Minimum detection limits set at approximately three times baseline noise were in the region of 15 pg injected (5.3 mM), and linearity of detection was observed between at least 15 and 2120 pg injected (84 fmol–12 pmol) (y Å 107.4x 0 1.95, r Å 0.999, n Å 24) (Fig. 4).
The HPCE of MPA extracts of plant tissue represents a simple and rapid alternative to HPLC for the analysis of L-AA and D-AA. Initial difficulties encountered with the acidic extraction solvent could be easily compensated for by increasing the concentration of the carrier electrolyte. This is important because like others (20), we have found that MPA is the optimal solvent for the extraction and stabilization of L-AA from plant tissue. The use of borate as carrier electrolyte permits detection of neutral carbohydrates at low uv wavelengths, and because L-AA, D-AA, and probably L-GL are all able to form anionic complexes with borate, resolutions are enhanced due to the competing influences of the EOF and anodic migration of borate complexes at high pH levels. Generally a disadvantage to the use of a high carrier electrolyte concentration is an increased background absorption, and this can compromise the sensitivity of detection. This did not appear to be a significant problem at 260 nm. However, a high carrier electrolyte concentration also has the advantages of
FIG. 3. Influence of ascorbic acid oxidase incubation on electropherograms of parsley extract. Five-second (29.5 nl) hydrostatic injection of parsley extract before (A) and after (B) incubation with ascorbic acid oxidase as described under Experimental. Ultraviolet absorbance set out at 260 nm. Carrier electrolyte of 200 mM borate, pH 9, under separation potential of /25 kV. The capillary was thermostatted at 257C. Inset shows absorbance spectrum obtained for the L-AA peak migrating at 7.8 min.
AID
AB 9633
/
6m1a$$$$$1
07-02-96 19:21:14
aba
AP: Anal Bio
14
AID
DAVEY, BAUW, AND VAN MONTAGU
AB 9633
/
6m1a$$9633
07-02-96 19:21:14
aba
AP: Anal Bio
CAPILLARY ZONE ELECTROPHORESIS OF ASCORBATE IN PLANTS
15
FIG. 4. Minimum detection limits of L-AA and D-AA. Three-second (17.7 nl) hydrostatic injection of standard solution of 1 mg/ml L-AA and D-AA in MPA/EDTA. Peaks represent injections of 84 fmol of sample. All other conditions as in Fig. 3.
higher sample capacities and it minimizes interactions between the analytes and the capillary wall. Although we obtained resolution of L-GL, and although biological extracts showed the presence of a LGL-like peak when extracts were monitored at 194 nm, the poor relative migration times of L-GL, the lack of a characteristic absorption spectrum, and the fact that it could not be resolved from its stereoisomer D-galacturonic acid-1,4-lactone (data not shown) meant that definitive identification of this component in biological extracts was not possible under these conditions. Direct comparison with ascorbate analyses by HPLC shows that CZE has notable advantages in terms of peak
resolution and that there is more than adequate resolution to allow a decrease in analysis times by the use of a shorter capillary length. The minimum concentration limits in CZE are considerably higher with HPLC, but are still low enough to allow measurement of L-AA levels in all extracts examined. However, it is unlikely that the sensitivity of the CZE separation could be much improved by sample stacking, due to the high concentration of the extraction solvent. The use of D-AA as an internal standard to improve the accuracy of determinations as proposed by Koh et al. (19) will not always be appropriate as we have observed D-AA to be a significant component of certain vegetative extracts.
FIG. 5. Comparison between HPCE and HPLC analysis of biological extracts. All conditions for HPCE as described in the legend to Fig. 2, except that injections of mushroom and Arabidopsis extracts were carried out for 10 s under hydrostatic pressure. Isocratic HPLC analysis with uv detection at 248 nm is as described under Experimental. Peaks labeled L-AA and D-AA indicate elution/migration positions of L-ascorbic acid and D-isoascorbic acid, respectively. HPLC (A) and HPCE (B) of parsley extract; HPLC (C) and HPCE (D) analysis of mushroom extract; HPLC (E) and HPCE (F) analysis of Arabidopsis extract.
AID
AB 9633
/
6m1a$$$$$1
07-02-96 19:21:14
aba
AP: Anal Bio
16
AID
DAVEY, BAUW, AND VAN MONTAGU
AB 9633
/
6m1a$$9633
07-02-96 19:21:14
aba
AP: Anal Bio
CAPILLARY ZONE ELECTROPHORESIS OF ASCORBATE IN PLANTS
FIG. 5—Continued
AID
AB 9633
/
6m1a$$9633
07-02-96 19:21:14
aba
AP: Anal Bio
17
18
DAVEY, BAUW, AND VAN MONTAGU
FIG. 5—Continued
AID
AB 9633
/
6m1a$$9633
07-02-96 19:21:14
aba
AP: Anal Bio
CAPILLARY ZONE ELECTROPHORESIS OF ASCORBATE IN PLANTS
ACKNOWLEDGMENTS The authors thank Beckman Instruments for the generous loan of the capillary electrophoresis equipment, Dr. Jens Oostergaard for critical reading of the manuscript, and Martine De Cock for preparation of the manuscript.
REFERENCES 1. Arrigoni, O. (1994) J. Bioenerg. Biomembr. 26, 407–419. 2. Foyer, C. (1993) in Antioxidants in Higher Plants (Alscher, R. G., and Hess, J. L., Eds.), pp. 31–58, CRC Press, Boca Raton, FL. 3. Mapson, L. W. (1970) in The Biochemistry of Fruits and Their Products, Vol. 1 (Food Science and Technology, a Series of Monographs) (Hulme, A. C., Ed.), pp. 369–384, Academic Press, London. 4. Ishihata, K., and Ito, S. (1994) Jpn. J. Trop. Agric. 38, 113–118. 5. Loewus, F. A., and Loewus, M. W. (1987) CRC Crit. Rev. Plant Sci. 5, 101–119. 6. Fleming, J. F., and Bensch, K. G. (1983) Int. J. Pept. Protein Res. 22, 355–361. 7. Chiari, M., Nesi, M., Carrea, G., and Righetti, P. G. (1993) J. Chromatogr. 645, 197–200. 8. Washko, P. W., Welch, R. W., Dhariwal, K. R., Wang, Y., and Levine, M. (1992) Anal. Biochem. 204, 1–14. 9. Kampfenkel, K., Van Montagu, M., and Inze´, D. (1995) Anal. Biochem. 225, 165–167. 10. Lykkesfeldt, J., Loft, S., and Poulsen, H. E. (1995) Anal. Biochem. 229, 329–335. 11. Kutnink, M. A., Hawkes, W. C., Schaus, E. E., and Omaye, S. T. (1987) Anal. Biochem. 166, 424–430.
AID
AB 9633
/
6m1a$$$$$1
07-02-96 19:21:14
19
12. Ross, M. A. (1994) J. Chromatogr. B 657, 197–200. 13. Okamura, M. (1980) Clin. Chim. Acta 103, 259–268. 14. Liau, L. S., Lee, B. L., New, A. L., and Ong, C. N. (1993) J. Chromatogr. 612, 63–70. 15. Nagy, E., and Degrell, I. (1989) J. Chromatogr. 497, 276–281. 16. Iwase, H., and Ono, I. (1993) J. Chromatogr. A 654, 215–220. 17. Iwase, H. (1992) J. Chromatogr. 606, 277–280. 18. Nisperos-Carriedo, M., Buslig, B. S., and Shaw, P. E. (1992) J. Agric. Food Chem. 40, 1127–1130. 19. Koh, E. V., Bissell, M. G., and Ito, R. K. (1993) J. Chromatogr. 633, 245–250. 20. Roe, J. H. (1954) in Methods of Biochemical Analysis (Glick, D., Ed.), Vol. 1, pp. 115–139, Academic Press, New York. 21. Tanishima, K., and Kita, M. (1993) J. Chromatogr. 613, 275– 280. 22. Brown, M. R., and Miller, K. A. (1992) J. Appl. Phycol. 4, 205– 215. 23. Kean, K. L. (1966) J. Lipid Res. 7, 449–453. 24. Honda, S., Iwase, S., Makino, A., and Fujiwara, S. (1989) Anal. Biochem. 176, 72–77. 25. Oefner, P. J., and Chiesa, C. (1994) Glycobiology 4, 397–412. 26. Landers, J. P., Oda, R. P., and Schuchard, M. D. (1992) Anal. Biochem. 64, 2847–2851. 27. Hoffstetter-Kuhn, S., Paulus, A., Gassmann, E., and Widmer, H. M. (1991) Anal. Chem. 63, 1541–1547. 28. Weinberger, R. (1993) in Practical Capillary Electrophoresis (Weinberger, R., Ed.), pp. 197–222, Academic Press, Boston. 29. Weinberger, R. (1993) in Practical Capillary Electrophoresis (Weinberger, R., Ed.), pp. 17–44, Academic Press, Boston. 30. Jaffe, G. M. (1984) in Encyclopedia of Chemical Technology (R. E. Kirk and D. F. Ohtmer, Eds.), Vol. 24, 3rd ed., pp. 8–40, Wiley, New York.
aba
AP: Anal Bio