Eur. J. Biochem. 138, 479-480 (1984) (0 FEBS 1984

Structure of ascorbic acid and its biological function Determination of the conformation of ascorbic acid and isoascorbic acid by infrared and ultraviolet investigations

Wolfgang LOHMANN, Detlef PAGEL, and Volker PENKA

Institut fur Biophysik im Strahlenzentrum der Universitat GieBen

(Received September 6/November 13, 1983) - EJB 83 0973

The four 0 - H bands of ascorbic acid could be assigned by means of infrared investigations. It could be shown by electron spin resonance and nuclear magnetic resonance measurements that the radical sodium ascorbate is formed by a cyclic side-chain structure resulting in a loss of C(6)-OH and C(3)-OH. The C(2) = C(3) double bond is still maintained as could be shown by infrared and ultraviolet absorption spectroscopy. In the case of complete oxidation of ascorbic acid to dehydroascorbic acid, C(6)-OH is reestablished (indicating the reopening of the furanoid ring), while C(2)-OH as well as the C(2) = C(3) double bond have disappeared due to the deprotonation of C(2)-OH and C(3)-OH. In the case of isoascorbic acid and its radical potassium isoascorbate similar results are obtained with one distinct difference: in the case of isoascorbic acid, C(2)-OH does not appear while C(3)-OH exhibits a shoulder.

Recent results of electron spin resonance and nuclear magnetic resonance which we have obtained [I, 21 have shown that, at neutral pH, the two epimers of vitamin C exist in their radical configuration with the unpaired electron located near the C(4) region. The differences in the shape of the radicals formed by the two epimers are due to the different positions of H and OH at C(5) and thus allow a quick and accurate determination of the two vitamin C epimers. The formation of the radical is accompanied by a ring closure between C(6)-OH and C(3) and thus involves changes of the OH groups at C(3) and C(6). This effect should be, therefore, detectable by infrared spectroscopy.

There is a dispute over the exact structure of the ascorbyl radical, especially in regard to the position of the unpaired electron [3]. Some models propose for the radical configuration the existence of a C(2)-C(3) single bond [4], while others suggest that the double bond is maintained, even though both of the OH groups at C(2) and C(3) are deprotonated [5] .

Since the ultraviolet absorption spectrum of ascorbic acid is assigned to a z - 71 * excitation of the C = C double bond and the OH vibrations can be easily detected by infrared measure- ments, both of these techniques can be applied in order to obtain some more information about the configuration of the ascorbyl radical.

MATERIALS AND METHODS

Ascorbic acid, sodium ascorbate, NaOH, and KOH were purchased from Merck (Darmstadt), isoascorbic acid from Fluka (Buchs). Dehydroascorbic acid was crystallized and purified according to a method proposed by Staudinger and Weis [6]. In addition, sodium ascorbate and potassium isoas- corbate were prepared by dissolving ascorbic acid and isoascor- bic acid in NaOH or KOH, respectively, resulting in a pH value of 7.2, and then lyophilized.

The infrared spectra were obtained with solid samples, ground in Nujol oil, and measured between NaCl plates. For

This is the third paper of a series. For parts 1 and 2 see [I] and [2].

comparison, other samples were recorded as KBr or NaCl pellets (1-mg sample in 200 mg KBr or NaCl) of 1-cm2 surface area. A Perkin-Elmer model 283 spectrophotometer was employed: slit program, 6.0; response, 1.0; gain, 1.0. All spectra were recorded at room temperature.

The ultraviolet spectra were recorded with a Zeiss DMR 10 spectr ophotometer .

RESULTS AND DISCUSSION

The infrared spectra of ascorbic acid, sodium ascorbate, and dehydroascorbic acid are shown in Fig. 1 (this is only a partial spectrum: bands below about 1000 cm-' are not shown). The assignments proposed earlier for the four 0 - H stretching bands above 3000 cm-' are contradictory [7, 81. Because of the importance and relevance of these 0 - H groups for the present discussion, only the 0-Hs as well as the lactone C = 0 (around 1750 cm-') and C = C stretching (z 1680 cm- ') bands will be considered.

Since there is a ring closure in sodium ascorbate between C(3) and C(6)-OH resulting in a disappearance of the 0-Hs at carbons 3 and 6, the two remaining bands in sodium ascorbate (see Fig. 1, middle spectrum) were assigned to C(5)-OH (3328 cm-') and C(2)-OH (3279 cm-'). The distinction be- tween C(5)-OH and C(2)-OH results from the dehydroascorbic acid spectrum. In dehydroascorbic acid, as is well known, the two 0-Hs located at C(2) and C(3) are deprotonated, that is, only the C(6)-OH and C(5)-OH bands should be visible. One of these two bands is located at 3308 cm-' and should, therefore, be assigned to the C(5)-OH stretching band. The remaining band at 3528 cm-' must be assigned, then, to C(6)-OH.

Based on these results, the four 0 - H bands observed in ascorbic acid (see Fig. 1, upper spectrum) can be assigned as follows: 3540 cm-', C(6)-OH; 3425 cm-', C(3)-OH; 3330 cm-', C(5)-OH; 3232 cm-', C(2)-OH.

With the formation of the radical (sodium ascorbate), the C = 0 (1759 cm-') and the C = C (1675 cm-') bands of ascorbic acid are shifted to smaller wave numbers (C = 0:

480

R A

4000 3000 2000 1800 1600 1400 1200 1000

WAVENUMBER (cm-1)

Fig. 1. Infrared spectra of ascorbic acid (ASC) , sodium ascorbate (Na-ASC), and dehydroascorbic acid ( D H A ) at 25°C. Bands below about 1000 cm-l are not shown. Spectra were obtained with solid samples, ground in Nujol oil, and measured between NaCl plates or were recorded as NaCl pellets (1 mg sample in 200 mg NaC1) of 1-cm2 surface area. Slit program: 6.0; response: 1.0; gain: 1.0

1708 cm-'; C = C: 1600 cm-'). The shift to lower frequency of the C = C stretching band suggests a reduction of the double bond character of the C(2) = C(3) bond. This finding agrees very well with the nuclear magnetic resonance results [2], according to which the electron density at C(3) is reduced in the radical state.

It should be also emphasized that the C = C double bond still exists in the case of sodium ascorbate (radical). In dehydroascorbic acid, as can be seen in Fig. 1, that band no longer exists; its C = 0 stretching band is located at a relatively high wave number (1 790 cm- ').

Of the numerous other bands not yet assigned, only one should be pointed out: in ascorbic acid a rather sharp band can be observed at 1500 cm- ', which is not present in isoascorbic acid (see Fig. 2). It is shifted slightly towards smaller wave numbers in sodium ascorbate and dehydroascorbic acid.

It is interesting to note that isoascorbic acid exhibits the two 0 - H bands associated with the carbons 5 and 6. The 0 - H band associated with carbon 3 appears as a shoulder only, while C(2)-OH cannot be detected at all. The band at 1670 cm-', demonstrates, however, the existence of the C(2) = C(3) double bond. For this reason, it is not quite clear why the C(2)-OH and C(3)-OH bands do not exist. Hydrogen bonding might be involved.

In the case of the radical potassium isoascorbate, as in the case of sodium ascorbate, C(5)-OH and C(2)-OH are present. The C = 0 band located at 1765 cm-' in the case of isoascor- bic acid is shifted to 1715 cm-' while a weak band near

LOO0 3000 2000 1800 1600 1400 1200 1000

WAVENUMBER (cm- l )

Fig. 2. Infrared spectra qf isoascorhic acid (Iso-ASC) and potassium isoascorbate (K-lso-ASC) at 25 "C. Details as in Fig. 1 ; instead of NaCl pellets KBr pellets were used. For comparison, the spectrum of sodium ascorbate (Na-ASC) is also shown

1750cm-' is retained when the radical is formed. Also the C = C band is shifted from 1680 cm-' to 1620 cm-'.

When ultraviolet spectra of ascorbic acid, isoascorbic acid, sodium ascorbate, and potassium isoascorbate have been recorded, all of these compounds exhibit a strong absorption band at about 265 nm which is due to the rc - rc * excitation of the C = C double bond. Again, this indicates the existence of the C(2) = C(3) double bond also in the radical configuration of the two epimers of vitamin C.

The excellent technical assistance of Miss P. Braunlich and Miss P. Ferber is appreciated. This work was supported in part by grants from the Bundesministerium fur Forschung und Technologie and from the Fonds der Chemischen Industrie.

REFERENCES 1. Lohmann, W. & Holz, D. (1984) Biophys. Struct. Mech., in the

2. Lohmann, W., Beinhauer, K. & Sapper, H. (1984) Biophys. Struct.

3. Laroff, G. P., Fessenden, R. W. & Schuler, R. H. (1972) J . Am.

4. Sapper, H., Pleyer-Weber, A. & Lohmann, W. (1982) Z.

5. Swartz, H. M. (1982) in Free Radicals andCancer(Floyd, R. A. ed.)

6. Staudinger, Hj. & Weis, W. (1964) Hoppe Seyler's Z. Physiol. Chem.

7. Weigl, J. W. (1952) Anal. Chem. 24, 1483-1486. 8. Falk, W. (1979) Spectrochim. Acta, 35A, 1117-1123.

press.

Mech., in the press.

Chem. Soc. 94, 9062-9073.

Naturforsch. 37c, 129-131.

pp. 81-115, Marcel Dekker, New York.

337, 284-285.

W. Lohmann, D. Pagel and V. Penka, Institut fur Biophysik, Strahlenzentrum der Justus-Liebig-Universitat GieRen, Leihgesterner Weg 217, D-6300 GieOen, Federal Republic of Germany