- Email: [email protected]
Spontaneous Hydrolysis and Dehydration of Dehydroascorbic Acid in Aqueous Solution
ANALYTICAL BIOCHEMISTRY ARTICLE NO.
260, 223–229 (1998)
AB982700
Spontaneous Hydrolysis and Dehydration of Dehydroascorbic Acid in Aqueous Solution John C. Deutsch1 Division of Gastroenterology, Department of Medicine, University of Colorado Health Sciences Center and Denver VAH, 4200 East Ninth Avenue, Campus Box B-170, Denver, Colorado 80262
Received February 6, 1998
The interaction of water with dehydroascorbic acid was examined by incubating dehydroascorbic acid and ascorbic acid in 18O-labeled water for various amounts of time and then oxidizing the products with hydrogen peroxide or reducing the products with mercaptoethanol, with analysis by gas chromatography mass spectrometry. Based on mass changes, dehydroascorbic acid readily exchanged three oxygen atoms with H218O. When mercaptoethanol was used to reduce dehydroascorbic acid (which had been incubated in H218O) to ascorbic acid, the newly formed ascorbic acid also contained three labeled oxygen atoms. However, ascorbic acid incubated in H218O for the same amount of time under identical conditions exchanged only two labeled oxygen atoms. Electron impact mass spectrometry of derivatized ascorbic acid created a decarboxylation product which had only two labeled oxygen atoms, regardless if 3-oxygen-labeled or 2-oxygen-labeled ascorbic acid was the parent compound, isolating the extra oxygen addition to carbon 1. These data suggest that dehydroascorbic acid spontaneously hydrolyzes and dehydrates in aqueous solution and that the hydrolytic-hydroxyl oxygen is accepted by carbon 1. Ascorbic acid, on the other hand, does not show this same tendency to hydrolyze. © 1998 Academic Press
The degradation of ascorbic acid (AA)2 (vitamin C) is a complex process, which in part may explain some of the interesting biological properties of this important compound. AA is initially oxidized through a free radical intermediate to dehydroascorbic acid (DHA) (1, 2), 1
Fax: (303) 315-8477. E-mail: [email protected]. Abbreviations used: AA, ascorbic acid; DHA, dehydroascorbic acid; DKG, diketogulonic acid; BME, b-mercaptoethanol; TBDMS, tert-butyldimethylsilyltrifluoroacetamide. 2
0003-2697/98 $25.00 Copyright © 1998 by Academic Press All rights of reproduction in any form reserved.
which is then hydrolyzed and further degraded to over 50 different products (3–7). Both AA and DHA have antiscorbutic effects in vivo (8), and the reaction of AA to DHA and, conversely, DHA to AA, must therefore readily occur in vivo (9, 10). It has been previously shown by others (9, 10) that DHA is quite unstable in water since the lactone ring spontaneously opens in aqueous solution to form diketogulonic acid (DKG), and the hydrolysis of DHA to DKG results in an irreversible loss of antiscorbutic activity. However, purified barium precipitates of 2,3-diketogulonic acid regained their antiscorbutic activity when reduced with HI (10), suggesting that the irreversible in vivo reaction could be reversed using stronger agents in vitro. We have recently reported mass spectral evidence which shows that the hydrolysis reaction can be reversed, at least in part, by incubation of hydrolyzed dehydroascorbic acid with b-mercaptoethanol (BME) (11). However, the data are complicated since dehydroascorbic acid forms a hydrated bicyclic hemiketal (12) (Fig. 1) which when derivatized with tert-butyldimethylsilyltrifluoroacetamide (TBDMS) would have a predicted [M-57] ion of m/z 591. This would be identical to the [M-57]1 ion of dehydrated DKG (11, 13). Although preparation of samples for gas chromatography should eliminate the hydrated species (drying under vacuum centrifugation and derivatization in organic solvents), this can be difficult to prove, as ascorbic-acid-containing solutions are able to form many different species at varying concentrations over time (3–7). The specific components in AA-containing solutions and the reactions which those species undergo have practical importance, as each of the different species has its own chemical and potentially different biological properties (14). It is possible and even likely that an effect attributed to DHA could actually be due to DKG or other products derived from DKG (14). The formation of different species under different 223
224
JOHN C. DEUTSCH
circumstances may therefore have profound effects in biological or experimental settings. Furthermore, AA and DHA may act as prooxidants under certain conditions. DHA clearly behaves as an oxidizing agent when it is reacted with sulfhydryls to produce AA (15, 16). AA appears to be a prooxidant since it will damage DNA in the presence of oxygen (17) and initiate oxidative cell death (18, 19). AA supplementation has been shown to increase bleomycin-induced chromosomal abbreviations in healthy individuals (20). Reactive oxygen species such as the OH2 radical (21) and H2O2 (22) are produced when AA is incubated with metals, possibly through simultaneous formation of the semidehydroascorbic acid free radical during AA oxidation to DHA. Since the prooxidant effects possibly involve free radical formation during the reversible AA:DHA interconversion, the first, truly irreversible degradative reaction which vitamin C undergoes could act as a gate to control flux and formation of these reactive species. This study examines interactions between DHA and the aqueous medium (using H218O) at mildly acidic conditions (pH 3.5) and events associated with the oxidation and reduction of DHA to further clarify the behavior of DHA in aqueous systems. METHODS
AA and b-mercaptoethanol were obtained from Sigma Chemicals (St. Louis, MO). [13C6]AA (98% 13C) and [6,62 H2]AA (98% 2H) were obtained from MSD Isotopes (Montreal, Quebec, Canada). DHA was obtained from ICN Pharmaceuticals (Costa Mesa, CA). Threonic acid (hemicalcium salt) was obtained from Fluka Chemicals (Ronkonkoma, NY). Hydrogen peroxide (H2O2) was obtained from Aldrich Chemicals (Milwaukee, WI). H218O (97% 18O) was obtained from Isotec Inc. (Miamisburg, OH). Solvents and other reagents, including purified distilled water, were obtained from Fisher Scientific (Pittsburgh, PA). TBDMS was obtained from Regis Chemicals (Morton Grove, IL). Cation exchange of threonic acid solution was performed by passing 1 ml of a 6 mM solution of threonate: hemicalcium salt in 0.1 N HCl over 50 mg of AGMP-50, 100 –200 mesh, hydrogen form (Bio-Rad Laboratories, Richmond, CA) contained in a plastic column and collecting the runthrough. The pH of AA and DHA solutions was previously determined using a Fisher Accumet pH meter (23) and ranges from pH 3.2 6 0.1 to 3.5 6 0.1. Reactions were carried out on 5- to 200-ml aliquots of 5–10 mM solutions of AA and DHA in 1.1-ml glass autosampler vials. Temperatures ranged from 20 to 37°C. At the end of incubation, samples were dried using a Savant (Farmingdale, NY) vacuum centrifuge system. For oxidation reactions, 5-ml aliquots of 30%
FIG. 1. The structure of anhydrous DHA monomer and the structure of the bicyclic hemiketal which forms during solubilization.
hydrogen peroxide were added to the dried samples and incubated at 20°C for 5–10 min. Reduction reactions were performed by adding 5 ml b-mercaptoethanol to 10-ml aliquots. Reactions were halted by drying at 20°C for 1 to 4 h using a Savant SpeedVac vacuum centrifuge system. Dried aliquots were derivatized by adding 10 ml of TBDMS and 20 ml of acetonitrile and then incubating the capped samples for 1 h at 60°C. Two-microliter aliquots were applied to a Hewlett–Packard (Avondale, PA) 5890 gas chromatograph. Gas chromatography was carried out splitless through a Supelco (Bellfonte, PA) 12-m dimethylsiloxane, fused-silica capillary column (i.d. 0.25 mm) using a temperature ramp of 30°C/ min from 80 to 300°C with helium as a carrier, and mass spectrometry was performed on a Hewlett–Packard 5971A mass spectrometer. The scan mode was used to obtain full spectra (including the [M-57]1 ion) and retention times using a mass range of 215– 650. The electron multiplier was at 1600 –1800 V. RESULTS
Solubilization of DHA As previously noted (12), DHA forms a hydrated bicyclic hemiketal in aqueous solution (Fig. 1). However, samples prepared for GCMS are dried prior to derivatization and resuspended in acetonitrile, which, it is hoped, will eliminate the hydrated product, pro-
HYDROLYSIS AND DEHYDRATION OF DEHYDROASCORBIC ACID
225
strating exchange of at least three 18O atoms between the aqueous medium and DHA. When DHA is incubated in water, several compounds with [M-57]1 ions of m/z 591 (11, 13) eventually form, consistent with hydrolysis of DHA to DKG. When DHA is incubated in H218O and allowed to hydrolyze to DKG (Fig. 3A), the [M-57]1 ion for DKG suggests gains of at least four oxygen atoms (up to m/z 599). This would be expected if DHA which previously exchanged three oxygen atoms is then hydrolyzed to incorporate a fourth labeled oxygen. Oxidation of DHA by H2O2
FIG. 2. (A) The total ion chromatogram of TBDMS-derivatized DHA. (B) The spectrum of TBDMS-derivatized DHA. (C) The expanded spectrum of the [M-57]1 ion of TBDMS-derivatized DHA. (D) The expanded spectrum of the [M-57]1 ion of DHA after incubation of DHA in H218O for 96 h prior to drying and derivatization.
When DHA is exposed to H2O2, a four-carbon product is formed which previously was tentatively identified as threonic acid by our laboratory (14). To definitively characterize this compound, the spectrum of TBDMS–threonic acid along with spectra of the same species formed from oxidation of [2H2]AA and [13C6]AA was generated and is shown in Fig. 4 with [M-57]1 ions of m/z 535, 537, and 539 respectively. Commercially available threonic acid, hemicalcium salt, following cation exchange, gives a spectrum and retention time identical to those of the species shown in Fig. 4A. The structure of threonic acid is shown in Fig. 5. Threonic acid has four derivatizable sites, one on each carbon, and retains carbons 3– 6 from the original AA/DHA molecule.
viding a TBDMS-derivatized DHA monomer with a predicted [M-57]1 ion of m/z 345 (24). As shown in Fig. 2, aqueous solutions of DHA which are dried and derivatized with TBDMS contain a compound with a retention time (Fig. 2A), spectrum (Fig. 2B), and [M-57]1 ion of m/z 345 (Fig. 2C) characteristic of DHA monomer (24). As seen in Fig. 2C, the abundance of the [M-57]1 ion between m/z 345 and m/z 350 falls rapidly and consistently, due to the decreasing natural abundance of the heavier isotopes in that compound. Solubilization of DHA in H218O When DHA is incubated in H218O the spectrum of the derivatized product becomes more complex. Figure 2D shows the [M-57]1 ion of TBDMS-derivatized DHA monomer following incubation in H218O for 96 h at 20°C. As shown in the spectrum, there is an increase in mass at m/z 347, 349, and 351 relative to the pattern shown in Fig. 2C for DHA incubated in water, demon-
FIG. 3. (A) The expanded spectrum of the [M-57]1 ion of a compound with a spectrum consistent with TBDMS-derivatized DKG, formed after incubation of DHA in H218O for 240 h. This ion normally occurs at m/z 591. (B) The expanded spectrum of the [M-57]1 ion of a compound with a spectrum consistent with TBDMS-derivatized threonic acid, formed after incubation of DHA in H218O for 240 h followed by exposure to concentrated hydrogen peroxide. This ion normally occurs at m/z 535.
226
JOHN C. DEUTSCH
FIG. 6. The [M-57]1 ion of TBDMS-derivatized AA formed after BME exposure of (A) a solution of DHA which had incubated in H218O for 25 days and (B) a solution of AA which had incubated in H218O for 25 days.
Since threonic acid contains carbons 3– 6 from the original DHA molecule, this demonstrates that only one of the exchangeable oxygen atoms is above carbon 2, and the rest of the exchange occurred at carbons 1 and 2 of DHA. FIG. 4. The full spectrum of TBDMS-derivatized threonic acid derived following hydrogen peroxide exposure of (A) AA, (B) [2H2]AA, and (C) [13C6]AA.
Oxidation of DHA Previously Incubated in H218O to Threonic Acid by H2O2 Analysis was performed on DHA-degradative products formed from DHA previously incubated in H218O for 240 h. Figure 3B shows the [M-57]1 ion of TBDMSderivatized threonic acid formed following H2O2 exposure. As shown, threonic acid formed from this solution contains only one exchanged atom of oxygen (m/z 537).
FIG. 5. The structure of threonic acid. When formed from AA or DHA, the carboxylic acid is on carbon 3 of the original compounds.
Comparison of Solubilization of DHA and AA in H218O Since AA and DHA both have similar lactone rings, but AA appears to be more stable to hydrolysis than DHA (7, 8, 11, 14), solutions of AA and DHA were incubated in H218O for 25 days at 20°C. Following this, 1 ml of BME was added to 5 ml of both solutions (final BME concentration of 2.4 M) and incubated for 0.5 h at 20°C. Afterward, the samples were dried by vacuum centrifugation for 4 h. As BME has a higher boiling point (157°C) than water, the concentration of BME will reach 14 M during this process. We have previously published a full spectrum for TBDMS-derivatized AA (24). This compound has an [M-57]1 ion of m/z 575. As shown in Fig. 6A, AA made from DHA incubated in H218O contained three exchanged oxygen atoms as shown by ions m/z 577, 579, and 581, whereas Fig. 6B shows that AA incubated in H218O the same amount of time had only two oxygen atoms exchanged (m/z 577 and 579). Recently, our laboratory has shown that the two sites of 18O exchange on AA occur at carbons 2 and 3 (25). In that same report, we characterized the structures responsible for ions in the spectrum of AA (25). Of interest, the ion at m/z 531 formed during electron impact of TBDMS-derivatized AA arises through decarboxylation (loss of carbon 1 and the attached two oxygen atoms). Another ion at m/z 343 arises
HYDROLYSIS AND DEHYDRATION OF DEHYDROASCORBIC ACID
227
DISCUSSION
FIG. 7. The ions formed by electron impact decarboxylation (normally m/z 531) of TBDMS-derivatized AA formed after BME exposure of (A) a solution of DHA which had incubated in H218O for 25 days and (B) a solution of AA which had incubated in H218O for 25 days. This demonstrates that the extra mass in Fig. 6A is on carbon 1. The ions formed by electron impact consisting of carbons 1– 4 (lactone ring, normally m/z 343) of TBDMS-derivatized AA formed after BME exposure of (C) a solution of DHA which had incubated in H218O for 25 days and (D) a solution of AA which had incubated in H218O for 25 days. Taken with A and B, this demonstrates that oxygen exchange occurred at both carbon 2 and carbon 3.
through loss of carbons 5 and 6. As shown in Figs. 7A and 7B, the decarboxylated ions were similar whether DHA or AA was the original starting material. Therefore, the extra oxygen exchanged on DHA (not exchanged on AA) must have occurred at carbon 1. As shown in Fig. 7C the extra oxygen remained with carbons 1– 4 if the AA arose from DHA. Furthermore, when these solutions were oxidized to form threonic acid, the threonic acid made from either H218O-exchanged AA or DHA (consisting of original carbons 3– 6) had [M-57]1 ions which were identical to those in Fig. 3B. These data show that the oxygen exchange on DHA occurs at carbons 1, 2, and 3.
Dehydroascorbic acid undergoes several reactions and interactions in aqueous solution, making it a unique compound in the ascorbate degradative pathway. DHA can be reduced to AA or hydrolyzed to DKG and oxidized to downstream products. DHA forms a bicyclic hemiketal (12) in water and readily dimerizes (8, 11). Over time, in a pH-sensitive reaction, hydrolysis and opening of the lactone ring occurs between carbons 1 and 4 leading to the formation of DKG (11). This report presents data which show ongoing interactions between DHA and the aqueous media. It appears that the formation of the bicyclic hemiketal in aqueous solutions leads to a reversible equilibrium reaction, resulting in the exchange of the keto-oxygens on carbons 2 and 3 of DHA. Despite involvement in a ring between carbons 3 and 6, the hydroxyl oxygen at carbon 6 does not participate in this exchange. Of more interest, the oxygen atoms attached to carbon 1 appear to exchange spontaneously, whereas the same lactone oxygens of AA do not exchange. Hydrolysis at this locus, such as occurs when DHA forms DKG, would be expected to add hydroxyl from the medium to carbon 1, due to the partial positive charge on the carbon at this location. Figure 8 shows a proposed mechanism for the exchange of 16O for 18O on carbon 1. Since structural evidence in this report shows that oxygen exchange occurs on carbons 1, 2, and 3 of DHA, whereas in AA only the oxygen atoms at carbons 2 and 3 exchange (25), the most plausible explanation is that hydrolysis is continually occurring, forming DKG with the labeled oxygen addition on carbon 1. As the carboxyl oxygens are equivalent, and would be expected to rapidly equilibrate, spontaneous reformation (through dehydration) of the lactone will eject labeled or unlabeled oxygen at an equal frequency, so that a proportion of the DHA molecules will retain the label at carbon 1 (Fig. 8). Based on the [M-57]1 ion of a compound structurally consistent with DKG (Fig. 3A), an additional labeled oxygen atom is added during hydrolysis, which concurs with the above assumption. These structural data strongly suggest that the formation of DKG from DHA is not only reversible, but is a spontaneous, ongoing equilibrium reaction at pH 3.5, in which these studies were performed. The rate of this equilibrium reaction would be expected to be profoundly affected by pH, which may explain why feeding DKG to scorbutic animals does not reverse scurvy (9, 10), whereas feeding DHA does, leading to the assumption that the hydrolysis of DHA was irreversible unless the strong reducing agent, HI, was used (9, 10). It may be that some of the reduction ascribed to HI was due to lowering the pH and allowing the spontaneous equilibrium to reestablish itself. It is also important to note that these DHA/DKG-feeding studies were performed
228
JOHN C. DEUTSCH
FIG. 8. A proposed mechanism for the exchange of 16O with 18O on carbon 1 of DHA. (A) AA which can be oxidized to (B) DHA. Oxygen on water can attack carbon 1 of DHA (B), resulting in hydrolysis to DKG (C). Since both oxygen atoms on carbon 1 are equivalent, free rotation can occur between carbon 1 and 2 as shown in (C). Spontaneous dehydration shown in (D) has an equivalent chance of removing the new 18O or 16O. (D) Removal of 16O. (E) DHA, but differing from (B) in that carbon 1 now has 18O. Reduction of (E) will lead to AA molecules containing 18 O. Dashed arrows in (B and D) show movement of electrons.
before the availability of much of the sophisticated equipment commonly used today (9, 10). The chemical identities of the species used in the feeding experiments were therefore not examined in detail, and it is possible that compounds other than those expected by the investigators were being fed to those animals. In addition, we have reported that BME can drive DKG to AA (11). These experiments, done at acidic pH, may also be due to a shift in reactant pools, since BME reduces DHA to AA, driving the equilibrium between DHA and DKG toward DHA. The use of different isotopes of AA during electron impact fragmentation and oxidation reduction reac-
tions allows some concise structural assignments and provides a means to better assess both the interaction of DHA with the aqueous environment and the early steps in DHA degradation. REFERENCES 1. Bendich, A., Machlin, L. J., Scandurra, O., Burton, G. W., and Wayner, D. D. M. (1986) Adv. Free Radical Biol. Med. 2, 419. 2. Buettner, G. R., and Jurkiewics, B. A. (1993) Free Radical Biol. Med. 14, 49 –55. 3. Shin, D. B., and Feather, M. S. (1990) Carbohydr. Res. 208, 246. 4. Shin, D. B., and Feather, M. S. (1990) J. Carbohydr. Chem. 9, 461. 5. Niemela, K. (1987) J. Chromatogr. 399, 235.
HYDROLYSIS AND DEHYDRATION OF DEHYDROASCORBIC ACID 6. Dunn, J. A., Ahmed, M. U., Murtiashaw, M. W., Richardson, J. M., Walla, M. D., Thorpe, S. R., and Baynes, J. W. (1990) Biochemistry 29, 10964. 7. Kimoto, E., Tanaka, H., Ohmoto, T., and Choami, M. (1993) Anal. Biochem. 214, 38. 8. Jaffe, G. (1990) in Handbook of Vitamins (L. Machlin, Ed.), pp. 199 –244, Dekker, New York. 9. Herbert, R. W., Hirst, E. L., Percival, E. G. V., Reynolds, R. J. W., and Smith, F. (1933) J. Chem. Soc. 2, 1270. 10. Penny, J. R., and Zilva, S. S. (1945) Biochem. J. 39, 1. 11. Deutsch, J. C., and Santhosh-Kumar, C. R. (1996) J. Chromatogr. A 724, 271–278. 12. Donner, L. W., and Hicks, K. B. (1986) in Methods in Enzymology, (F. Chytil and D. B. McCormick, Eds.) Vol. 122, pp. 3–10, Academic Press, New York. 13. Deutsch, J. C., Santhosh-Kumar, C. R., Hassell, K. L., and Kolhouse, J. F. (1994) Anal. Chem. 66, 345. 14. Deutsch, J. (1998) Anal. Biochem. 255, 1–7. 15. Washko, P., Rotrosen, D., and Levine, M. (1989) J. Biol. Chem. 264, 18996.
229
16. Dhariwal, K. R., Hartzell, W. O., and Levine, M. (1991) Am. J. Clin. Nutr. 54, 712. 17. Morgan, A. R., Cone, R. L., and Elgert, T. M. (1976) Nucleic Acids Res. 3, 1139 –1149. 18. Sestili, P., Brandi, G., Brambilla, L., Cattabeni, F., and Catoni, O. (1996) J. Pharmacol. Exp. Ther. 277, 1719 –1725. 19. De Laurenzi, V., Melino, G., Savini, I., Annicchiarico-Petruzzelli, M., Finazzi-Agro, A., and Avigliano, L. (1995) Eur. J. Cancer 31A, 463– 466. 20. Anderson, D., Phillips, B. T., Yu, T. W., Edwards, A. J., Ayesh, R., and Butterworth, K. R. (1997) Environ. Mol. Mutagen. 30, 161–174. 21. Zhao, M. J., and Jung, L. (1995) Free Radical Res. 23, 229 – 243. 22. Jameson, R. F., and Blackburn, N. J. (1982) J. Chem. Soc. Dalton Trans. 9 –13. 23. Deutsch, J. C. (1997) Anal. Biochem. 247, 58 – 62. 24. Deutsch, J. C., and Kolhouse, J. F. (1993) Anal. Chem. 65, 321–326. 25. Deutsch, J. C. (1998) J. Chromatogr. A 802, 385–390.
260, 223–229 (1998)
AB982700
Spontaneous Hydrolysis and Dehydration of Dehydroascorbic Acid in Aqueous Solution John C. Deutsch1 Division of Gastroenterology, Department of Medicine, University of Colorado Health Sciences Center and Denver VAH, 4200 East Ninth Avenue, Campus Box B-170, Denver, Colorado 80262
Received February 6, 1998
The interaction of water with dehydroascorbic acid was examined by incubating dehydroascorbic acid and ascorbic acid in 18O-labeled water for various amounts of time and then oxidizing the products with hydrogen peroxide or reducing the products with mercaptoethanol, with analysis by gas chromatography mass spectrometry. Based on mass changes, dehydroascorbic acid readily exchanged three oxygen atoms with H218O. When mercaptoethanol was used to reduce dehydroascorbic acid (which had been incubated in H218O) to ascorbic acid, the newly formed ascorbic acid also contained three labeled oxygen atoms. However, ascorbic acid incubated in H218O for the same amount of time under identical conditions exchanged only two labeled oxygen atoms. Electron impact mass spectrometry of derivatized ascorbic acid created a decarboxylation product which had only two labeled oxygen atoms, regardless if 3-oxygen-labeled or 2-oxygen-labeled ascorbic acid was the parent compound, isolating the extra oxygen addition to carbon 1. These data suggest that dehydroascorbic acid spontaneously hydrolyzes and dehydrates in aqueous solution and that the hydrolytic-hydroxyl oxygen is accepted by carbon 1. Ascorbic acid, on the other hand, does not show this same tendency to hydrolyze. © 1998 Academic Press
The degradation of ascorbic acid (AA)2 (vitamin C) is a complex process, which in part may explain some of the interesting biological properties of this important compound. AA is initially oxidized through a free radical intermediate to dehydroascorbic acid (DHA) (1, 2), 1
Fax: (303) 315-8477. E-mail: [email protected]. Abbreviations used: AA, ascorbic acid; DHA, dehydroascorbic acid; DKG, diketogulonic acid; BME, b-mercaptoethanol; TBDMS, tert-butyldimethylsilyltrifluoroacetamide. 2
0003-2697/98 $25.00 Copyright © 1998 by Academic Press All rights of reproduction in any form reserved.
which is then hydrolyzed and further degraded to over 50 different products (3–7). Both AA and DHA have antiscorbutic effects in vivo (8), and the reaction of AA to DHA and, conversely, DHA to AA, must therefore readily occur in vivo (9, 10). It has been previously shown by others (9, 10) that DHA is quite unstable in water since the lactone ring spontaneously opens in aqueous solution to form diketogulonic acid (DKG), and the hydrolysis of DHA to DKG results in an irreversible loss of antiscorbutic activity. However, purified barium precipitates of 2,3-diketogulonic acid regained their antiscorbutic activity when reduced with HI (10), suggesting that the irreversible in vivo reaction could be reversed using stronger agents in vitro. We have recently reported mass spectral evidence which shows that the hydrolysis reaction can be reversed, at least in part, by incubation of hydrolyzed dehydroascorbic acid with b-mercaptoethanol (BME) (11). However, the data are complicated since dehydroascorbic acid forms a hydrated bicyclic hemiketal (12) (Fig. 1) which when derivatized with tert-butyldimethylsilyltrifluoroacetamide (TBDMS) would have a predicted [M-57] ion of m/z 591. This would be identical to the [M-57]1 ion of dehydrated DKG (11, 13). Although preparation of samples for gas chromatography should eliminate the hydrated species (drying under vacuum centrifugation and derivatization in organic solvents), this can be difficult to prove, as ascorbic-acid-containing solutions are able to form many different species at varying concentrations over time (3–7). The specific components in AA-containing solutions and the reactions which those species undergo have practical importance, as each of the different species has its own chemical and potentially different biological properties (14). It is possible and even likely that an effect attributed to DHA could actually be due to DKG or other products derived from DKG (14). The formation of different species under different 223
224
JOHN C. DEUTSCH
circumstances may therefore have profound effects in biological or experimental settings. Furthermore, AA and DHA may act as prooxidants under certain conditions. DHA clearly behaves as an oxidizing agent when it is reacted with sulfhydryls to produce AA (15, 16). AA appears to be a prooxidant since it will damage DNA in the presence of oxygen (17) and initiate oxidative cell death (18, 19). AA supplementation has been shown to increase bleomycin-induced chromosomal abbreviations in healthy individuals (20). Reactive oxygen species such as the OH2 radical (21) and H2O2 (22) are produced when AA is incubated with metals, possibly through simultaneous formation of the semidehydroascorbic acid free radical during AA oxidation to DHA. Since the prooxidant effects possibly involve free radical formation during the reversible AA:DHA interconversion, the first, truly irreversible degradative reaction which vitamin C undergoes could act as a gate to control flux and formation of these reactive species. This study examines interactions between DHA and the aqueous medium (using H218O) at mildly acidic conditions (pH 3.5) and events associated with the oxidation and reduction of DHA to further clarify the behavior of DHA in aqueous systems. METHODS
AA and b-mercaptoethanol were obtained from Sigma Chemicals (St. Louis, MO). [13C6]AA (98% 13C) and [6,62 H2]AA (98% 2H) were obtained from MSD Isotopes (Montreal, Quebec, Canada). DHA was obtained from ICN Pharmaceuticals (Costa Mesa, CA). Threonic acid (hemicalcium salt) was obtained from Fluka Chemicals (Ronkonkoma, NY). Hydrogen peroxide (H2O2) was obtained from Aldrich Chemicals (Milwaukee, WI). H218O (97% 18O) was obtained from Isotec Inc. (Miamisburg, OH). Solvents and other reagents, including purified distilled water, were obtained from Fisher Scientific (Pittsburgh, PA). TBDMS was obtained from Regis Chemicals (Morton Grove, IL). Cation exchange of threonic acid solution was performed by passing 1 ml of a 6 mM solution of threonate: hemicalcium salt in 0.1 N HCl over 50 mg of AGMP-50, 100 –200 mesh, hydrogen form (Bio-Rad Laboratories, Richmond, CA) contained in a plastic column and collecting the runthrough. The pH of AA and DHA solutions was previously determined using a Fisher Accumet pH meter (23) and ranges from pH 3.2 6 0.1 to 3.5 6 0.1. Reactions were carried out on 5- to 200-ml aliquots of 5–10 mM solutions of AA and DHA in 1.1-ml glass autosampler vials. Temperatures ranged from 20 to 37°C. At the end of incubation, samples were dried using a Savant (Farmingdale, NY) vacuum centrifuge system. For oxidation reactions, 5-ml aliquots of 30%
FIG. 1. The structure of anhydrous DHA monomer and the structure of the bicyclic hemiketal which forms during solubilization.
hydrogen peroxide were added to the dried samples and incubated at 20°C for 5–10 min. Reduction reactions were performed by adding 5 ml b-mercaptoethanol to 10-ml aliquots. Reactions were halted by drying at 20°C for 1 to 4 h using a Savant SpeedVac vacuum centrifuge system. Dried aliquots were derivatized by adding 10 ml of TBDMS and 20 ml of acetonitrile and then incubating the capped samples for 1 h at 60°C. Two-microliter aliquots were applied to a Hewlett–Packard (Avondale, PA) 5890 gas chromatograph. Gas chromatography was carried out splitless through a Supelco (Bellfonte, PA) 12-m dimethylsiloxane, fused-silica capillary column (i.d. 0.25 mm) using a temperature ramp of 30°C/ min from 80 to 300°C with helium as a carrier, and mass spectrometry was performed on a Hewlett–Packard 5971A mass spectrometer. The scan mode was used to obtain full spectra (including the [M-57]1 ion) and retention times using a mass range of 215– 650. The electron multiplier was at 1600 –1800 V. RESULTS
Solubilization of DHA As previously noted (12), DHA forms a hydrated bicyclic hemiketal in aqueous solution (Fig. 1). However, samples prepared for GCMS are dried prior to derivatization and resuspended in acetonitrile, which, it is hoped, will eliminate the hydrated product, pro-
HYDROLYSIS AND DEHYDRATION OF DEHYDROASCORBIC ACID
225
strating exchange of at least three 18O atoms between the aqueous medium and DHA. When DHA is incubated in water, several compounds with [M-57]1 ions of m/z 591 (11, 13) eventually form, consistent with hydrolysis of DHA to DKG. When DHA is incubated in H218O and allowed to hydrolyze to DKG (Fig. 3A), the [M-57]1 ion for DKG suggests gains of at least four oxygen atoms (up to m/z 599). This would be expected if DHA which previously exchanged three oxygen atoms is then hydrolyzed to incorporate a fourth labeled oxygen. Oxidation of DHA by H2O2
FIG. 2. (A) The total ion chromatogram of TBDMS-derivatized DHA. (B) The spectrum of TBDMS-derivatized DHA. (C) The expanded spectrum of the [M-57]1 ion of TBDMS-derivatized DHA. (D) The expanded spectrum of the [M-57]1 ion of DHA after incubation of DHA in H218O for 96 h prior to drying and derivatization.
When DHA is exposed to H2O2, a four-carbon product is formed which previously was tentatively identified as threonic acid by our laboratory (14). To definitively characterize this compound, the spectrum of TBDMS–threonic acid along with spectra of the same species formed from oxidation of [2H2]AA and [13C6]AA was generated and is shown in Fig. 4 with [M-57]1 ions of m/z 535, 537, and 539 respectively. Commercially available threonic acid, hemicalcium salt, following cation exchange, gives a spectrum and retention time identical to those of the species shown in Fig. 4A. The structure of threonic acid is shown in Fig. 5. Threonic acid has four derivatizable sites, one on each carbon, and retains carbons 3– 6 from the original AA/DHA molecule.
viding a TBDMS-derivatized DHA monomer with a predicted [M-57]1 ion of m/z 345 (24). As shown in Fig. 2, aqueous solutions of DHA which are dried and derivatized with TBDMS contain a compound with a retention time (Fig. 2A), spectrum (Fig. 2B), and [M-57]1 ion of m/z 345 (Fig. 2C) characteristic of DHA monomer (24). As seen in Fig. 2C, the abundance of the [M-57]1 ion between m/z 345 and m/z 350 falls rapidly and consistently, due to the decreasing natural abundance of the heavier isotopes in that compound. Solubilization of DHA in H218O When DHA is incubated in H218O the spectrum of the derivatized product becomes more complex. Figure 2D shows the [M-57]1 ion of TBDMS-derivatized DHA monomer following incubation in H218O for 96 h at 20°C. As shown in the spectrum, there is an increase in mass at m/z 347, 349, and 351 relative to the pattern shown in Fig. 2C for DHA incubated in water, demon-
FIG. 3. (A) The expanded spectrum of the [M-57]1 ion of a compound with a spectrum consistent with TBDMS-derivatized DKG, formed after incubation of DHA in H218O for 240 h. This ion normally occurs at m/z 591. (B) The expanded spectrum of the [M-57]1 ion of a compound with a spectrum consistent with TBDMS-derivatized threonic acid, formed after incubation of DHA in H218O for 240 h followed by exposure to concentrated hydrogen peroxide. This ion normally occurs at m/z 535.
226
JOHN C. DEUTSCH
FIG. 6. The [M-57]1 ion of TBDMS-derivatized AA formed after BME exposure of (A) a solution of DHA which had incubated in H218O for 25 days and (B) a solution of AA which had incubated in H218O for 25 days.
Since threonic acid contains carbons 3– 6 from the original DHA molecule, this demonstrates that only one of the exchangeable oxygen atoms is above carbon 2, and the rest of the exchange occurred at carbons 1 and 2 of DHA. FIG. 4. The full spectrum of TBDMS-derivatized threonic acid derived following hydrogen peroxide exposure of (A) AA, (B) [2H2]AA, and (C) [13C6]AA.
Oxidation of DHA Previously Incubated in H218O to Threonic Acid by H2O2 Analysis was performed on DHA-degradative products formed from DHA previously incubated in H218O for 240 h. Figure 3B shows the [M-57]1 ion of TBDMSderivatized threonic acid formed following H2O2 exposure. As shown, threonic acid formed from this solution contains only one exchanged atom of oxygen (m/z 537).
FIG. 5. The structure of threonic acid. When formed from AA or DHA, the carboxylic acid is on carbon 3 of the original compounds.
Comparison of Solubilization of DHA and AA in H218O Since AA and DHA both have similar lactone rings, but AA appears to be more stable to hydrolysis than DHA (7, 8, 11, 14), solutions of AA and DHA were incubated in H218O for 25 days at 20°C. Following this, 1 ml of BME was added to 5 ml of both solutions (final BME concentration of 2.4 M) and incubated for 0.5 h at 20°C. Afterward, the samples were dried by vacuum centrifugation for 4 h. As BME has a higher boiling point (157°C) than water, the concentration of BME will reach 14 M during this process. We have previously published a full spectrum for TBDMS-derivatized AA (24). This compound has an [M-57]1 ion of m/z 575. As shown in Fig. 6A, AA made from DHA incubated in H218O contained three exchanged oxygen atoms as shown by ions m/z 577, 579, and 581, whereas Fig. 6B shows that AA incubated in H218O the same amount of time had only two oxygen atoms exchanged (m/z 577 and 579). Recently, our laboratory has shown that the two sites of 18O exchange on AA occur at carbons 2 and 3 (25). In that same report, we characterized the structures responsible for ions in the spectrum of AA (25). Of interest, the ion at m/z 531 formed during electron impact of TBDMS-derivatized AA arises through decarboxylation (loss of carbon 1 and the attached two oxygen atoms). Another ion at m/z 343 arises
HYDROLYSIS AND DEHYDRATION OF DEHYDROASCORBIC ACID
227
DISCUSSION
FIG. 7. The ions formed by electron impact decarboxylation (normally m/z 531) of TBDMS-derivatized AA formed after BME exposure of (A) a solution of DHA which had incubated in H218O for 25 days and (B) a solution of AA which had incubated in H218O for 25 days. This demonstrates that the extra mass in Fig. 6A is on carbon 1. The ions formed by electron impact consisting of carbons 1– 4 (lactone ring, normally m/z 343) of TBDMS-derivatized AA formed after BME exposure of (C) a solution of DHA which had incubated in H218O for 25 days and (D) a solution of AA which had incubated in H218O for 25 days. Taken with A and B, this demonstrates that oxygen exchange occurred at both carbon 2 and carbon 3.
through loss of carbons 5 and 6. As shown in Figs. 7A and 7B, the decarboxylated ions were similar whether DHA or AA was the original starting material. Therefore, the extra oxygen exchanged on DHA (not exchanged on AA) must have occurred at carbon 1. As shown in Fig. 7C the extra oxygen remained with carbons 1– 4 if the AA arose from DHA. Furthermore, when these solutions were oxidized to form threonic acid, the threonic acid made from either H218O-exchanged AA or DHA (consisting of original carbons 3– 6) had [M-57]1 ions which were identical to those in Fig. 3B. These data show that the oxygen exchange on DHA occurs at carbons 1, 2, and 3.
Dehydroascorbic acid undergoes several reactions and interactions in aqueous solution, making it a unique compound in the ascorbate degradative pathway. DHA can be reduced to AA or hydrolyzed to DKG and oxidized to downstream products. DHA forms a bicyclic hemiketal (12) in water and readily dimerizes (8, 11). Over time, in a pH-sensitive reaction, hydrolysis and opening of the lactone ring occurs between carbons 1 and 4 leading to the formation of DKG (11). This report presents data which show ongoing interactions between DHA and the aqueous media. It appears that the formation of the bicyclic hemiketal in aqueous solutions leads to a reversible equilibrium reaction, resulting in the exchange of the keto-oxygens on carbons 2 and 3 of DHA. Despite involvement in a ring between carbons 3 and 6, the hydroxyl oxygen at carbon 6 does not participate in this exchange. Of more interest, the oxygen atoms attached to carbon 1 appear to exchange spontaneously, whereas the same lactone oxygens of AA do not exchange. Hydrolysis at this locus, such as occurs when DHA forms DKG, would be expected to add hydroxyl from the medium to carbon 1, due to the partial positive charge on the carbon at this location. Figure 8 shows a proposed mechanism for the exchange of 16O for 18O on carbon 1. Since structural evidence in this report shows that oxygen exchange occurs on carbons 1, 2, and 3 of DHA, whereas in AA only the oxygen atoms at carbons 2 and 3 exchange (25), the most plausible explanation is that hydrolysis is continually occurring, forming DKG with the labeled oxygen addition on carbon 1. As the carboxyl oxygens are equivalent, and would be expected to rapidly equilibrate, spontaneous reformation (through dehydration) of the lactone will eject labeled or unlabeled oxygen at an equal frequency, so that a proportion of the DHA molecules will retain the label at carbon 1 (Fig. 8). Based on the [M-57]1 ion of a compound structurally consistent with DKG (Fig. 3A), an additional labeled oxygen atom is added during hydrolysis, which concurs with the above assumption. These structural data strongly suggest that the formation of DKG from DHA is not only reversible, but is a spontaneous, ongoing equilibrium reaction at pH 3.5, in which these studies were performed. The rate of this equilibrium reaction would be expected to be profoundly affected by pH, which may explain why feeding DKG to scorbutic animals does not reverse scurvy (9, 10), whereas feeding DHA does, leading to the assumption that the hydrolysis of DHA was irreversible unless the strong reducing agent, HI, was used (9, 10). It may be that some of the reduction ascribed to HI was due to lowering the pH and allowing the spontaneous equilibrium to reestablish itself. It is also important to note that these DHA/DKG-feeding studies were performed
228
JOHN C. DEUTSCH
FIG. 8. A proposed mechanism for the exchange of 16O with 18O on carbon 1 of DHA. (A) AA which can be oxidized to (B) DHA. Oxygen on water can attack carbon 1 of DHA (B), resulting in hydrolysis to DKG (C). Since both oxygen atoms on carbon 1 are equivalent, free rotation can occur between carbon 1 and 2 as shown in (C). Spontaneous dehydration shown in (D) has an equivalent chance of removing the new 18O or 16O. (D) Removal of 16O. (E) DHA, but differing from (B) in that carbon 1 now has 18O. Reduction of (E) will lead to AA molecules containing 18 O. Dashed arrows in (B and D) show movement of electrons.
before the availability of much of the sophisticated equipment commonly used today (9, 10). The chemical identities of the species used in the feeding experiments were therefore not examined in detail, and it is possible that compounds other than those expected by the investigators were being fed to those animals. In addition, we have reported that BME can drive DKG to AA (11). These experiments, done at acidic pH, may also be due to a shift in reactant pools, since BME reduces DHA to AA, driving the equilibrium between DHA and DKG toward DHA. The use of different isotopes of AA during electron impact fragmentation and oxidation reduction reac-
tions allows some concise structural assignments and provides a means to better assess both the interaction of DHA with the aqueous environment and the early steps in DHA degradation. REFERENCES 1. Bendich, A., Machlin, L. J., Scandurra, O., Burton, G. W., and Wayner, D. D. M. (1986) Adv. Free Radical Biol. Med. 2, 419. 2. Buettner, G. R., and Jurkiewics, B. A. (1993) Free Radical Biol. Med. 14, 49 –55. 3. Shin, D. B., and Feather, M. S. (1990) Carbohydr. Res. 208, 246. 4. Shin, D. B., and Feather, M. S. (1990) J. Carbohydr. Chem. 9, 461. 5. Niemela, K. (1987) J. Chromatogr. 399, 235.
HYDROLYSIS AND DEHYDRATION OF DEHYDROASCORBIC ACID 6. Dunn, J. A., Ahmed, M. U., Murtiashaw, M. W., Richardson, J. M., Walla, M. D., Thorpe, S. R., and Baynes, J. W. (1990) Biochemistry 29, 10964. 7. Kimoto, E., Tanaka, H., Ohmoto, T., and Choami, M. (1993) Anal. Biochem. 214, 38. 8. Jaffe, G. (1990) in Handbook of Vitamins (L. Machlin, Ed.), pp. 199 –244, Dekker, New York. 9. Herbert, R. W., Hirst, E. L., Percival, E. G. V., Reynolds, R. J. W., and Smith, F. (1933) J. Chem. Soc. 2, 1270. 10. Penny, J. R., and Zilva, S. S. (1945) Biochem. J. 39, 1. 11. Deutsch, J. C., and Santhosh-Kumar, C. R. (1996) J. Chromatogr. A 724, 271–278. 12. Donner, L. W., and Hicks, K. B. (1986) in Methods in Enzymology, (F. Chytil and D. B. McCormick, Eds.) Vol. 122, pp. 3–10, Academic Press, New York. 13. Deutsch, J. C., Santhosh-Kumar, C. R., Hassell, K. L., and Kolhouse, J. F. (1994) Anal. Chem. 66, 345. 14. Deutsch, J. (1998) Anal. Biochem. 255, 1–7. 15. Washko, P., Rotrosen, D., and Levine, M. (1989) J. Biol. Chem. 264, 18996.
229
16. Dhariwal, K. R., Hartzell, W. O., and Levine, M. (1991) Am. J. Clin. Nutr. 54, 712. 17. Morgan, A. R., Cone, R. L., and Elgert, T. M. (1976) Nucleic Acids Res. 3, 1139 –1149. 18. Sestili, P., Brandi, G., Brambilla, L., Cattabeni, F., and Catoni, O. (1996) J. Pharmacol. Exp. Ther. 277, 1719 –1725. 19. De Laurenzi, V., Melino, G., Savini, I., Annicchiarico-Petruzzelli, M., Finazzi-Agro, A., and Avigliano, L. (1995) Eur. J. Cancer 31A, 463– 466. 20. Anderson, D., Phillips, B. T., Yu, T. W., Edwards, A. J., Ayesh, R., and Butterworth, K. R. (1997) Environ. Mol. Mutagen. 30, 161–174. 21. Zhao, M. J., and Jung, L. (1995) Free Radical Res. 23, 229 – 243. 22. Jameson, R. F., and Blackburn, N. J. (1982) J. Chem. Soc. Dalton Trans. 9 –13. 23. Deutsch, J. C. (1997) Anal. Biochem. 247, 58 – 62. 24. Deutsch, J. C., and Kolhouse, J. F. (1993) Anal. Chem. 65, 321–326. 25. Deutsch, J. C. (1998) J. Chromatogr. A 802, 385–390.