ANALYTICAL BIOCHEMISTRY 255, 1–7 (1998)ARTICLE NO. AB972293

Ascorbic Acid Oxidation by Hydrogen Peroxide

John C. Deutsch1

Divisions of Gastroenterology and Hematology, Department of Medicine, University of Colorado Health Sciences Centerand Denver Veterans’ Administration Hospital, 4200 East Ninth Avenue, Campus Box B-170, Denver, Colorado 80220

Received May 1, 1997

Ascorbic acid (AA)2 is a water-soluble vitamin whichThe oxidative degradation of ascorbic acid by hydro- reacts with H2O2 (8). AA is a potent antioxidant in some

gen peroxide was examined to determine routes of deg- in vitro models and prevents oxidation of a variety ofradation and identify the initial products which form biomolecules (9, 10). AA reduces or prevents H2O2-in-when ascorbic acid is oxidized. When reacted with hy- duced lipid peroxidation (11, 12), and the formation ofdrogen peroxide, solutions of ascorbic acid and dehy- OH-deoxyguanisine, acting as a free radical scavengerdroascorbic acid are both ultimately oxidized to the in the CrCl3/H2O2 system (13). Furthermore, AA pro-same species, having a mass spectrum consistent with tects thymocytes against oxidant-induced apoptosisthreonic acid. When the intermediate steps in the oxi- (14) and protects aged dermal fibroblasts from H2O2-dation of ascorbic acid are examined in detail, ascorbic induced cytotoxicity (15).acid, dehydroascorbic acid, and solutions containing In contrast to its role as an antioxidant, prooxidanthydrolyzed dehydroascorbic acid are all oxidized properties have also been ascribed to AA under certainthrough a six-carbon compound previously proposed

conditions. AA can cause strand breakage in DNA into be tetrahydroxydiketohexanoic acid. Both dehy-the presence of oxygen (16) and can initiate cell deathdroascorbic acid and hydrolyzed dehydroascorbicin tissue culture, possibly through the generation ofacid (diketogulonic acid) are more susceptible to hy-H2O2 (17, 18). Reactions of AA with metals such asdrogen peroxide oxidation than ascorbic acid. BasedCu(II) are thought to lead to the production of H2O2on mass spectral analysis, diketogulonic acid serves(19) and AA has been shown to increase OH-radicalas an oxygen sink, implying that it may be a betterlevels in Fenton systems (20).reducing agent for toxic oxygen species than ascorbic

Whether or not, and under what circumstances AAacid. These data indicate that oxidation of ascorbicserves as an in vivo antioxidant or prooxidant, it isacid by hydrogen peroxide primarily proceeds

through three major six-carbon intermediates, each clear that AA is easily oxidized (21, 22), and H2O2 willwith distinctive redox properties. The stable metabo- react with and will oxidize AA.lite diketogulonic may be a critical antioxidant in The first step in the oxidation of AA is reportedly theascorbic-acid-containing systems. q 1998 Academic Press formation of dehydroascorbic acid (DHA) with the loss

of hydrogens from carbons 2 and 3 through a free radi-cal intermediate (23–25). This reversible reaction isthought to be primarily responsible for the antioxidantHydrogen peroxide (H2O2) is one of the principle reac-properties attributed to AA (23, 24). The prooxidanttive products of oxygen metabolism. Accumulation ofproperties of ascorbic acid may be due to the simultane-H2O2 can have profound deleterious effects on cellsous formation of the semi-DHA free radical (25) duringthrough base modifications and strand breakage in ge-AA oxidation to DHA.nomic DNA (1–3), damage to lysosomal membranes

Besides being reduced back to AA, DHA can also be(4), and the induction of apoptosis (5). In vivo, H2O2rapidly hydrolyzed to products such as 2,3-diketogu-forms from the dismutation of superoxide, which inlonic acid (DKG) (26) and then can be further oxidizedturn can be generated by phagocytic cells, mitochon-

dria, or autooxidation of endogenous compounds suchas catecholamines (6, 7). 2 Abbreviations used: AA, ascorbic acid; DHA, dehydroascorbic

acid; DKG, 2,3-diketogulonic acid; TBDMS, N-methyl-N(tert-butyldi-methylsilyl)trifluoroacetamide; TIC, total ion chromatograph; THDH,1 To whom correspondence and reprint requests should be ad-

dressed. Fax: 303/315-8477. E-mail: deutschj@essex.uchsc.edu. 4,5,5,6-tetrahydroxy-2,3-diketohexanoic acid.

10003-2697/98 $25.00Copyright q 1998 by Academic PressAll rights of reproduction in any form reserved.

AID AB 2293 / 6m3fh$$401 12-18-97 16:34:19 abas

JOHN C. DEUTSCH2

to over 50 species containing five or less carbons (27–31). This susceptibility to oxidation suggests that theseintermediaries must also be reducing substances.

Certain experimental data imply there is more to theantioxidant effects of AA than simply the reversibleAA:DHA reaction. For instances, it has been reported(32) that solutions of DHA better prevented cupric ion-induced low-density lipoprotein oxidation than AA. Themechanism of this effect is unclear since AA is readilyoxidized to DHA and both AA and DHA should providea source of DHA. However, the formation of AA to DHAgenerates the semi-DHA-free radical (25) whichwouldn’t be present if DHA were the starting material.The situation is further obscured by rapid hydrolysisof DHA to DKG or other degradative products (26).After a short time, a DHA-containing solution will con-tain other species which could contribute importantlyto the overall antioxidant capacity.

This report examines the oxidation of AA and AA-degradative products by H2O2, to determine immediateproducts which form during AA oxidation and the rela-tive stabilities of the oxidative products during H2O2

exposure.

MATERIALS AND METHODS

AA was obtained from Sigma Chemicals, (St. Louis,MO). DHA was purchased from ICN Biochemicals(Cleveland, OH). A 30% (v/v) solution of H2O2 was ob- FIG. 1. (A–D) Total ion chromatograms (m/z 220 to 650) following

elution from a 10 M dimethylsiloxane column generated after TBDMStained from Kodak Chemicals (Rochester, NY). Sol-derivatization of 57 mM AA incubated at 377C for 1 h in (A) water,vents and other reagents, including purified Burdick–(B) 0.1% H2O2 (30 mM H2O2), (C) 1% H2O2 (300 mM H2O2), or (D)Jackson distilled water, were purchased from Fisher 10% H2O2 (3000 mM H2O2). The asterisks mark where DHA (m/z

Scientific (Pittsburgh, PA). N-Methyl-N(tert-butyldi- 345) elutes. The peaks labeled 1 are isomers of THDH, while thepeak labeled 2 is threonic acid.methylsilyl)trifluoroacetamide (TBDMS) was obtained

from Regis Chemicals (Morton Grove, IL). [13C6]AA(98% 13C) and [6,62H2]AA (98% 2H) were obtained from

methylsiloxane, fused silica capillary column (id, 0.25MSD Isotopes (Montreal, Quebec, Canada).mm) using a temperature ramp of 307C/min from 80 toIncubations and solubilization of volumes greater3007C with helium as a carrier, and mass spectrometrythan 500 ml were performed in plastic conical tubes.was performed on a Hewlett–Packard 5971A massReactions were otherwise carried out on 5- to 20-mlspectrometer. The scan mode was used to obtain fullaliquots in 1.1-ml glass autosampler vials containingspectra (including the [M-57]/ ion) and retention timessolutions of AA and DHA (final concentration, 0.3–60using a mass range of 200-650. The electron multipliermM). H2O2 concentrations varied from 0.1 to 30% (30–was at 1650–2100 V.9000 mM). The temperature ranged from 20 to 377C.

The pH of solutions as determined using short-rangeRESULTSand long-range pH paper (Fisher Scientific) varied be-

tween 3.0 and 4.5. No buffers were added. Reactions To determine the sensitivity of AA to H2O2, 114 mM

solutions of AA in water were mixed 1:1 (v/v) with 0.0,were halted by drying at 207C for 30 min to 4 h using aSavant (Farmingdale, NY) vacuum centrifuge system. 0.2, 2, or 20% (0, 60, 600, and 6000 mM) solutions of

H2O2 to give a final concentration of 57 mM AA in 0.0,The dried aliquots were derivatized by adding 20 mlof TBDMS and 40 ml of acetonitrile then incubating the 0.1, 1, or 10% (0, 30, 300, and 3000 mM) H2O2. These

solutions were incubated for 1 h at 377C in triplicate,capped samples for 2 h at 607C. Two-microliter aliquotswere applied to a Hewlett–Packard (Avondale, PA) dried, derivatized with TBDMS and examined by gas

chromatography/mass spectrometry (GC/MS). As5890 gas chromatograph. Gas chromatography wascarried out through a Supelco (Bellfonte, PA) 12-m di- shown in Fig. 1, the total ion chromatograph (TIC) for

AID AB 2293 / 6m3fh$$402 12-18-97 16:34:19 abas

ASCORBIC ACID OXIDATION BY HYDROGEN PEROXIDE 3

FIG. 2. (A–F) Total ion chromatograms (m/z 220–650) following elution from a 12 M dimethylsiloxane column generated after TBDMSderivatization of 28.5 mM AA (A–C) or 28.5 mM DHA (D–F) in 30% (9000 mM) H2O2 at 207C for 24 h (A, D), 72 h (B, E), and 330 h (C, F).The peaks labeled 1 are isomers of THDH, while the peak labeled 2 is threonic acid.

the lower levels of H2O2 (0.1 and 1%) does not appear not AA (Fig. 2B) or DHA (Fig. 2E) is the starting mate-rial, and there is no appreciable change from this chro-different from those in which no H2O2 is added. At

higher concentrations (10%), there is loss of AA and matogram, even after 330 h of incubation (Figs. 2C and2F), demonstrating that DHA and AA go through theformation of other earlier eluting species.

Since the first oxidation product of AA is reported to same degradative pathways in H2O2.The relative formation (based on [M-57]/ ions) ofbe DHA (23, 24), an ion chromatographic analysis was

performed on the data collected in the TICs shown in DHA and THDH to AA in AA-containing solutions ex-posed to H2O2 was examined to determine if AA oxida-Fig. 1, to locate DHA based on the [M-57]/ ion of m/z

345, as previously described (33). DHA elutes earlier tion in H2O2 leads to DHA as an unstable intermediateon the way to forming THDH. Experiments were car-([∗], peak too small to be seen) than the new peaks

shown in Fig. 1D, and the ion abundance of DHA is ried out at 207C in 1% (300 mM) and 10% (3000 mM)H2O2. A parallel experiment was also performed wherebelow visualization on the TIC’s (Figs. 1A–1D). This

suggests that either the new peaks formed in Fig. 1D fresh solution of DHA was incubated in H2O2 and theformation of THDH was monitored.do not arise from DHA or that DHA is less stable than

the other products and is destroyed nearly as fast as As shown in Fig. 3A, DHA appears to form earlyfrom AA in a time-dependent manner. However, as alsoit is formed.

Based on mass spectra, the new products formed in shown in Fig. 3A, the relative DHA content does notcontinue to increase in the high H2O2 incubation. InFig. 1D are proposed to be 4,5,5,6-tetrahydroxy-2,3-di-

ketohexanoic acid (THDH) (labeled 1) and threonic acid contrast, THDH shows a later but steady increase overtime whether AA (Fig. 3B) or DHA (Fig. 3C) is the(labeled 2). THDH has previously been shown by our

laboratory to be a major species formed when AA is starting material. These data suggest that AA is oxi-dized to DHA during exposure to H2O2, but that DHAexposed to H2O2 (34). The proposed structure for THDH

was presented previously (34) and no new structural is actually less stable in the presence of H2O2 than AA.There is no detectable formation of AA from DHA withstudies were performed at this time.

To determine if DHA and AA go through similar deg- H2O2 incubations. Since THDH forms from either AAor DHA in H2O2, and AA forms DHA in H2O2, thenradative routes in H2O2, and to identify species which

may be stable products of H2O2:AA degradation, AA DHA is likely an intermediate species which is rapidlydegraded.and DHA were made into 28.5 mM (5 mg/ml) solutions

in 9000 mM (30%) H2O2 and allowed to incubate at 207C Since it is known that DHA spontaneously hydro-lyzes into DKG and related products in water (26),for 330 h. Aliquots were removed at 24-h intervals,

derivatized with TBDMS, and examined by GC/MS. studies were performed on solutions of DHA whichwere allowed to decompose in water over a 10-day pe-Neither AA or DHA is detectable after the first 24 h

in either sample, but, as shown in Fig. 2, there is less riod. Integrations of the TIC at this point showed thathydrolyzed species were 10-fold more abundant thanthreonic acid relative to THDH if AA is the starting

material (Fig. 2A) than if DHA is the starting material DHA (Fig. 4A).Based on the relative abundance of DKG to DHA(2D). This pattern is still present at 48 h (not shown).

However, by 72 h, the TICs are identical whether or in solutions incubated in 1% H2O2 (Fig. 4B), DHA is

AID AB 2293 / 6m3fh$$402 12-18-97 16:34:19 abas

JOHN C. DEUTSCH4

FIG. 3. (A–C) Relative abundances of [M-57]/ ions of TBDMS-derivatized AA (m/z 575), DHA (m/z 345), and THDH (m/z 607) duringincubation of 5.7 mM AA (A, B) or DHA (C) in 1 and 10% (300 and 3000 mM) H2O2 versus time. (A) Relative ratio of DHA to AA when AAis the starting material. (B) Relative ratio of DHTH to AA when AA is the starting material. (C) Relative ratio of THDH to DHA whenDHA is the starting material.

more stable than DKG since the ratio of m/z 591 lower concentrations (1%) of H2O2 . Together, thesedata show that DKG is more unstable than DHA and(DKG) to m/z 345 (DHA) falls rapidly with time in

H2O2. The formation of THDH was then followed rel- is more easily oxidized to THDH. More importantly,THDH has a mass exactly 16 greater than DKG. Thisative to the residual DHA in solution of degraded

DHA (Fig. 4C) to compare with the formation of mass spectral data implies that DKG acts like anoxygen sink during the formation of THDH, in con-THDH in fresh DHA (Fig. 3C). As shown in Fig. 4C,

there is a rapid increase in the relative formation of trast to previous degradative steps which involve aloss of hydrogen (AA to DHA), or a nonoxidative hy-THDH on exposure to either 1 or 10% H2O2. How-

ever, in solutions of DHA where hydrolysis has not drolysis (DHA to DKG). Furthermore, as DKG is oxi-dized, it avoids the formation of the semi-DHA freeoccurred to any appreciable extent (Fig. 3C), the rel-

ative formation of THDH is slower, particularly in radical. Figure 5 shows a possible scheme for the

AID AB 2293 / 6m3fh$$402 12-18-97 16:34:19 abas

ASCORBIC ACID OXIDATION BY HYDROGEN PEROXIDE 5

FIG. 4. (A) Total ion chromatograph (m/z 220–650) following elution from a 12 M dimethylsiloxane column generated after TBDMSderivatization of a 10-day-old solution of DHA which has primarily hydrolyzed to DKG. The asterisks mark the residual DHA peaks, whilethe peaks marked DKG(X) represent isomers of DKG. (B) Relative ratio of the [M-57]/ ions of DKG (m/z 591) to DHA (m/z 345) when thesolution in (A) is incubated in 1% (300 mM) H2O2. (C) Relative ratio of THDH (m/z 607) to DHA (m/z 345) when the solution in (A) isincubated in 1% (300 mM) and 10% (3000 mM) H2O2.

oxidation of AA to DHA to DKG to THDH. However, rently appreciated. There is some data which agreeswith this hypothesis. It has been shown that solutionsit is important to note that the structures are shown

in an unhydrated state and that other structures are of ‘‘DHA’’ were better at protecting low-density lipopro-tein from cupric ion-indiced oxidation than solutions ofpossible for the compound referred to as THDH.AA (32), and DHA does spontaneously hydrolyze toDKG. Although chelation may play a part in the differ-DISCUSSIONences (35), the other species that arise during contin-

This report shows that secondary species, particu- ued oxidative stress may also be important.larly DHA and DKG, are more susceptible to H2O2 oxi- DKG is a compound which has the potential to be adation than AA, which suggests that these species are critical factor in preventing oxidant injury from reac-preferred reactants. Although not well defined, these tive oxygen species. In this report, DKG is much lesssecondary reactions may contribute more to the antiox- stable in the presence of H2O2 than DHA, which in turn

was less stable than AA, demonstrating that it reactsidant properties of AA-containing solutions than is cur-

AID AB 2293 / 6m3fh$$402 12-18-97 16:34:19 abas

JOHN C. DEUTSCH6

FIG. 5. Possible scheme for the oxidation of AA through the semi-DHA-free radical to DHA to DKG and a possible reaction of DKG plusH2O2 to form THDH.

5. Kazzaz, J. A., Xu, J., Palaia, T. A., Mantell, L., Fein, A. M., andwith H2O2 before either AA or DHA. Based on massHorowitz, S. (1996) J. Biol. Chem. 271, 15182–15186.spectra, THDH is m/z 16 greater than DKG, suggesting

6. Cohen, G., and Heikkila, R. E. (1974) J. Biol. Chem. 249, 2447–that DKG either binds atomic oxygen or exchanges a2452.hydrogen for an H2O2-hydroxyl as it is oxidized to

7. Turrens, J. F., Freeman, B. A., and Crapo, J. D. (1982) Arch.THDH. Furthermore, although AA oxidation is known Biochem. Biophys. 217, 411–421.to produce a relatively stable free radical intermediate, 8. Lowry, J. P., and O’Neill, R. D. (1992) Anal. Chem. 64, 453–456.this has not been reported during the oxidation of DKG 9. Frei, B., England, L., and Ames, B. N. (1989) Proc. Natl. Acad.to THDH. Many of the antioxidant properties ascribed Sci. USA 86, 6377–6381.to AA may actually be due to DKG. In vitro and in 10. Frei, B., Stocker, R., and Ames, B. N. (1988) Proc. Natl. Acad.

Sci. USA 85, 9748–9752.vivo studies looking at the antioxidant properties ofAA should consider using DKG as a complementary 11. Einsele, H., Clemens, M. R., and Remmer, H. (1985) Free Radical

Res. Comm. 1, 63–67.reactant.12. Retsky, K. L., and Frei, B. (1995) Biochim. Biophys. Acta 1257,

279–287.ACKNOWLEDGMENTS13. Tsou, T. C., Chen, C. L., Liu, T. Y., and Yang, J. L. (1996) Carci-

Experiments were funded by a Veterans’ Administration Hospital nogenesis 17, 103–108.Award RAGS 0001. The author thanks J. Fred Kolhouse for the 14. Maellaro, E., Bello, B. D., and Comporti, M. (1996) Exp. Cell Res.use of his mass spectrometer and Valarie Allen for her secretarial 226, 105–113.assistance in preparing the manuscript. 15. Farriol, M., Mourelle, M., and Schwartz, S. (1994) Rev. Esp.

Fisiol. 50, 253–257.REFERENCES 16. Morgan, A. R., Cone, R. L., and Elgert, T. M. (1976) Nucleic Acids

Res. 3, 1139–1149.1. Jaruga, P., and Dizdaroglu, M. (1996) Nucleic Acid Res. 24,17. Sestili, P., Brandi, G., Brambilla, L., Cattabeni, F., and Catoni,1389–1394.

O. (1996) J. Pharm. Exp. Ther. 277, 1719–1725.2. Spencer, J. P., Jenner, A., Aruma, O. I., Cross, C. E., Wu, R., and18. DeLaurenzi, V., Melino, G., Savini, I., Annicchiarico-Petruzzelli,Haliwell, B. (1996) Biochem. Biophys. Res. Commun. 224, 17–

M., Finazzi-Agro, A., and Avigliano, L. (1995) Eur. J. Cancer22.31A, 463–466.3. Cantoni, O., Sestilli, P., Guidarelli, A., Palomba, L., Brambilla,

19. Jameson, R. F., and Blackburn, N. J. (1982) J. Chem. Soc. DaltonL., and Cattabeni, F. (1996) Arch. Toxicol. Suppl. 18, 223–235.Trans. 9–13.4. Brunk, U. T., Zhang, H., Dalen, H., and Ollinger, K. (1995) Free

Radical Biol. Med. 19, 813–822. 20. Zhao, M. J., and Jung, L. (1995) Free Radical Res. 23, 229–243.

AID AB 2293 / 6m3fh$$403 12-18-97 16:34:19 abas

ASCORBIC ACID OXIDATION BY HYDROGEN PEROXIDE 7

21. Grinstead, R. R. (1960) J. Am. Chem. Soc. 82, 3464–3471. 28. Shin, D. B., and Feather, M. S. (1990) J. Carbohydr. Chem. 9,461–470.22. Buettner, G. R. (1986) Free Radical Res. Comms. 1, 349–

29. Lopez, M. G., and Feather, M. S. (1992) J. Carbohydr. Chem. 11,353.799–806.23. Jaffe, G. in Handbook of Vitamins (Machlin, L., Ed.), pp. 199–

30. Grandhee, S. K., and Monnier, V. M. (1991) J. Biol. Chem. 266,244. Dekker, New York, NY.11649–11653.24. Bendich, A., Machlin, L. J., Scandurra, O., Burton, G. W., and

31. Niemela, K. (1987) J. Chrom. 399, 235–243.Wayner, D. D. M. (1986) Adv. Free Radical Biol. Med. 2, 419–444. 32. Retsky, K. L., Freeman, M. W., and Frei, B. (1993) J. Biol. Chem.

268, 1304–1309.25. Buettner, G. R., and Jurkiewics, B. A. (1993) Free Radical Biol.Med. 14, 49–55. 33. Deutsch, J. C., and Kolhouse, J. F. (1993) Anal. Chem. 65, 321–

326.26. Deutsch, J. C., and Santhosh-Kumar, C. R. (1996) J. Chrom. A724, 271–278. 34. Deutsch, J. C., Santhosh-Kumar, C. R., Hassell, K. L., and Kol-

house, J. F. (1994) Anal. Chem. 66, 345–350.27. Shin, D. B., and Feather, M. S. (1990) Carbohydr. Res. 208, 246–250. 35. Davies, M. B. (1992) Polyhedron 11, 285–321.

AID AB 2293 / 6m3fh$$403 12-18-97 16:34:19 abas