Measurement of Ascorbic Acid Kinetics in Man Using Stable Isotopes and Gas Chromatography/Mass Spectrometry

JOURNAL OF MASS SPECTROMETRY, VOL. 31, 741-748 (1996)

Measurement of Ascorbic Acid Kinetics in Man Using Stable Isotopes and Gas Chromatography/Mass Spectrometry

Leslie J. C. Bluck, Adrian P. Izzardt and Christopher J. Bates Dunn Nutritional Laboratory, Downhams Lane, Milton Road, Cambridge CB4 lXJ, UK

The feasibility of studying ascorbic acid kinetics in man using stable isotope-labelled tracers and gas chromatographic (GC) separation followed by mass spectrometric (MS) quantitation was assessed. Preliminary studies with 13C-labelled material showed that although better precision at low levels could be achieved using the GCfcombus- tion/MS technique, consideration of likely enrichments in a human study made the simpler GCIMS method just as suitable. On this basis, a small pilot study of the kinetics in man was carried out. The enrichment of the ascorbic acid in plasma was measured for a 24 h period after oral administration of '3C-labelled material. The results were fitted to a simple three-compartment model and rate constants and pool sizes were deduced. The results obtained are comparable to those obtained in other published studies, from which we conclude that the technique may be useful as a non-invasive method for the assessment of nutritional status in a variety of populations.

KEYWORDS : ascorbic acid; compartmental analysis; gas chromatography/mass spectrometry; stable isotopes

INTRODUCTION

Ascorbic acid is probably the most widely used vitamin supplement, but there is no international consensus on requirements. Although it has been known for many years that a daily intake of around 10 mg is needed to prevent scurvy, there is still no definite view on the extra amount required for antioxidant protection and the proper operation of other metabolic processes.' Nutritional surveys provide information which correl- ates much more closely with recent intake than with long term status, and more revealing information on ascorbic acid kinetics is needed; such information can best be obtained from tracer experiments. Studies using radioactive 14C-labelled material have been carried

but not all of these have been in uivo, and often although the rate of transport of the radioactive carbon can be measured, the chemical nature of the material in which it is incorporated was not necessarily known. Hence one cannot be certain if the radioactivity was present in unchanged ascorbic acid or its metabolites. It is now deemed generally undesirable to expose the subjects to the risk of radioactivity, and so the alternative approach of using stable isotopes should be considered. The only published investigation using this method is now over 30 years old7. This prompted us to investigate the possibility of developing a method of assessing ascorbic acid status using stable isotope-labelled material and a combination of gas chromatography (GC) and mass spectrometry (MS), which is suitable for universal application in the assessment of nutritional status.

t Author to whom correspondence should be addressed.

It is known that the absorption of ascorbic acid is dose dependent.' Unchanged ascorbic acid is found excreted in the urine if large doses are given. Recent work has shown that a dose of 60 mg of ascorbic acid is taken up almost completely within 2 h, but that less than a quarter of a 500 mg dose is absorbed.' This puts a ceiling on the size of the dose used in tracer experiments. For the purposes of this work, we chose to limit the dose to 50 mg of labelled material, which is less than the level at which incomplete absorption is reported to start occurring, but large enough to ensure observable enrichment in isotope ratios. To put the dose in a dietary context, it is equivalent to eating one large orange."

METHODOLOGY DEVELOPMENT

Choice of isotope

It is possible to label the ascorbic acid with either 'H or 13C. Labelling should be carried out at non-labile sites. The current possibilities are the use of 13C-labelled material followed by conventional GC/MS or the recently introduced GC/combustion/isotope ratio MS (GC/C/MS) methods." Theoretically the same possibilities exist for 'H label, but a GC/C/MS method for this isotope is not yet available. The GC/C/MS method has not been widely used in studies of this type, but under some circumstances it can offer increased sensi- tivity and precision over the older technique. In order to keep the options of both methods open, we investi- gated the use of a 13C labelled compound as a tracer in man.

CCC 1076-5174/96/070741-08 0 1996 by John Wiley & Sons, Ltd.

Received 7 February 1996 Accepted 21 March 1996

742 L. J. C. BLUCK, A. P. IZZARD AND C. J. BATES

Derivitization and chromatographic considerations

The determination of ascorbic acid by GC techniques has been reported. To obtain sufficient volatility, it is necessary to esterify the hydroxyl groups of the sugar. Ascorbic acid was converted into the trimethylsilyl (TMS) ester by Allison and Stewart12 for GC studies, while Deutsch and Kolhouse13 studied the tert- butyldimethylsilyl (TBDMS) derivative by GC/MS. In general, TBDMS derivatives are much more stable than their TMS analogues. However, the effects of the derivitization on the isotope ratio measurements must be considered before choosing the derivative to be used. If all the active hydrogens are replaced by silyl groups, then ascorbic acid, which has the empirical formula C6H806, is expected to form the tetrasilyl ester. In the case of the TMS derivative this has the empirical formula Cl8H4,O,Si4, while the TBDMS ester is C30H6406Si4. Two mass spectrometric approaches are possible for the determination of isotopic content: GC/MS and the related combustion technique GC/C/ IRMS. Table 1 shows expected results for these two derivatives from both techniques. In compiling this table we have taken the isotopic compositions of all materials to be at natural abundance, except at the label site. The effects of isotopic dilution by the added groups are clearly apparent. For example, the GC/C/MS

method would give an enrichment of only 150.4% for 5% as labelled material if it is separated as the TBDMS derivative, compared with nearly five times this value if the native compound could be analysed. If GC/C/MS is to be used, it is best to choose the derivative which adds the smallest group to the molecule.

A further potential problem which occurred only with the TBDMS derivative is shown in Fig. 1. Here the chromatogram shows two resolved species from ascotb- ic acid. The mass spectra of TBDMS derivatives show a major fragment 57 mass units less than the molecular ion, corresponding to the loss of a tert-butyl group. The mass spectrum of the earlier eluting compound had a fragment ion at m/z 461, and this peak was ascribed to the incompletely esterified tri-TBDMS derivative, which has a calculated molecular mass of 518. The later eluting species was assigned as the tetra-TBDMS ester, which has a molecular mass of 632, and which showed the predicted ion at m/z 575 in its mass spectrum. These assignments were confirmed by measurement of the isotopomer ratios of the [M - 571' fragments. The tri- TBDMS ester was formed even though the reagents used for the derivatives were at least tenfold in excess. The production of this species did not seem to be related to reaction time; indeed, its formation was inconsistent. As the chromatograms obtained from plasma are expected to be crowded owing to the

Table 1. Calculated isotope spectra for ascorbic acid and its derivatives measured by GCIMS and by GC/C/IRMS

Compound GC/MS isatopomer ratios (%) GC/C/IRMS

M M +1 M +2 M + 3 delta values

(a) Ascorbic Acid 100.000 7.095 1.417 0.087 0.0 (b) [l -'3C]ascorbic acid 100.000 5.973 1.350 18447.4 5% (b) +95% (a) 100.000 12.41 8 1.735 0.1 59 720.7

(c) Tetra-TMS-ascorbate 100.000 41.299 22.549 5.279 0.0 (d) Tetra-TMS-[I - '3C]ascorbate 100.000 40.1 76 22.098 5285.2 5% (d) +95% (c) 100.000 46.621 24.6873 6.455 251 .O

(e) Tetra-TBDMS-ascorbate 100.000 55.1 28 29.1 41 8.795 0.0 (f) Tetra-TBDMS-[I - '3C]ascorbate 100.000 54.006 28.535 31 06.6 5% (f) +95% (e) 100.000 60.540 32.01 6 10.31 4 150.4

chrometogram of tbdm 8Scorbate

9.50 1050 11.50 12.50 13.50 14.50 15.50 18.50

Relention Time (M)

Figure 1. Chromatograms of the TBDMS derivatives af ascorbic acid. From the [M - 57]+ ions shown, the earlier eluting peak is assigned as tri-TBDMS axorbate and the later one as tetra-TBDMS ascorbate.

ASCORBIC ACID KINETICS IN MAN USING GC/MS 143

amounts of various sugars in the blood, it was decided to use the TMS derivative to avoid this further possible complication.

Mass spectrometry

The mass spectrum of tetra-TMS ascorbate was obtained under a variety of conditions (Fig. 2). Under electron impact ionization, the spectrum is dominated by the peak at m/z 73 which corresponds to the trimethylsilyl group. This is added at the derivitization stage, and consequently contains no information about the isotopic composition of the ascorbic acid. Compari- son of Fig. 2(a), the EI mass spectrum of labelled material, and Fig. 2(b), the mass spectrum of the unlabelled derivative, shows that the fragments usable for isotope analysis are those with m/z > 330.

The positive-ion chemical ionization mass spectrum in Fig. 2(c) has as its base peak the adduct [M - HI', while the largest peak in the negative-ion mass spectrum (Fig. 2(d)) is at m/z 391, corresponding to the loss of a TMS group from the intact molecule. Both of these look good candidates for isotope ratio studies.

Comparison of GC/MS and GC/C/IRMS

In order to assess the suitability of the two mass spectrometric techniques for the human studies a series

il 1 D -

100 150 200 250 300 350 400 450 500

Figure 2. Mass spectra of the tetra-TMS derivative of ascorbic acid recorded under a variety of conditions: (a) 1 -'3C-labelled material, electron impact ionization; (b) unlabelled material, electron impact ionization; (c) unlabelled material, CH, chemical ionization, positive-ion spectrum; and (d) onlabelled material, CH, chemical ionization, negative-ion spectrum.

of samples with enrichments in the expected ranges were prepared and analysed. GC/MS data were obtained on an INCOS XL system (Finnigan MAT, Heme1 Hempstead, UK) and GC/C/MS data on an Iso- chrom GC combustion spectrometer (VG Analytical, Manchester, UK). Comparison between the two techniques is clouded by the different conventions used to express results. In general, GC/MS uses the unit 'molecule per cent excess', or mpe, and deals with the isotopomer ratios directly. GC/C/MS uses the delta scale proposed by Craig.14 The ratio of 13C to is expressed as

where Rsamp is the ratio of 13C to "C in the sample and Rref that ratio in a reference material.

For the purposes of comparison of the techniques, it is more convenient to convert to mole fractions of labelled and unlabelled material. This unit is also convenient for the analysis of ascorbate kinetics.

GC/MS. The determination of isotopic enrichment from GC/MS data has been considered by a number of

In general, it is accepted that the mole ratio of tracer to tracee should be the basic system unit. Vogt et d.'* proposed a method of obtaining enrichment data in these terms from isotopomer ratios, but their method requires that all the isotopomers in the ion cluster should be measured. Because of the large number of atoms with significant naturally occurring isotopes in the tetra-TMS ascorbate molecule, this was felt to be impractical, and an alternative approach was used in which only the three most intense signals were measured.

Consider the resulting GC/MS trace obtained as a superposition of two spectra due to two different species. If these spectra are denoted in the following way:

M M + l M + 2 M + 3 Species A A, A, A, A3 Species B B, B, BZ B3

then a mixture of NA mol of species A and N , mol of species B will give rise to a spectrum in which the ratio of the (M + n)th ion to the parent Mth ion, R, , is given by

Rn = [An + XABn - An)I/C& + XdBo - 4 1 where X , = N J ( N , + NB) is the mole fraction of species A and A and B represent concentrations of species A and B. As it stands this equation is useless as there are too many unknowns on the right-hand side, but usually approximations can be made which make it applicable to the experimental data. For the particular case of species A being tetra-TMS ascorbate and species B being the same but labelled with I3C at one specific site, we have

R1 = R,* + (I - &-'XB/(l - X, )

R , = Rzo + [(nc - l)A,-(l - A,)-'

+ nSiASi(1 - -4c)-'(1 - A S i - BSi)-']xB/(l - x,)

I44 L. J. C. BLUCK, A. P. IZZARD AND C. J. BATES

where R,' represents the isotopomer ratio for the unlabelled compound, nc is the number of carbon

natural fractional abundance of 13C, Asi is that of 29Si and B,; is that for 30Si. In deriving these eauations. the

and labelled (RJ material:

atoms, nsi is the number of silicon atoms, A, is the R, = Rc R l = [l + (1 - nc)R,-J(l

small effects due to naturally occurring 'H and "'0 have been neglected. Plots of the experimental GC/MS data are presented in Fig. 3.

From these data, experimental values for the standard deviation for a typical measured isotopomer ratio are estimated as 0.27% for R, and 0.16% for R, (measurements performed in triplicate), from which the expected errors in the estimation of mole fraction may be calculated, as shown in Fig. 4.

GC/C/MS. A similar method of analysis may be applied to the GC/C/MS data. If the natural abundance carbons have a ratio of 13C to "C of R,, then we can write expressions for the isotope ratio in unlabelled (R,)

= (1 + ncRc)/(l - n 3

If these are mixed in the proportions Xu to X , , then the isotope ratio in the mix will be given by

R m = (ncRc + X,)/(nc - XI.) Expressing this in terms of a delta value relative to the unlabelled material gives

6 , = (1 + 1/R,)X,/(nc + XI)

The linearity of the GC/C/MS technique is shown in Fig. 5.

It is possible to derive an expression for the error in X, as a function of the error in the measured delta

GGMS Standard Enrichment Curve

1o.m

9.W

8.m P @ 7.m

2.m

1 .W

0.m 0.009 0.020 0.040 0.060 0.W 0.1m 0.120

Mole mlo d latielled to unWled mmkl

Figure 3. GC/MS isotopomer ratios measured for a series of enriched samples of tetra-TMS-ascorbic acid.

Comparison of errom from GCYS and GCCMS

0 i 0.W 0.a 0.10 0.15 0.20

Mole lractionot labelled maWial

Figure 4. Estimated errors in isotopic composition determination of tetra-TMS-ascorbic acid by GC/MS and GC/C/MS. Estimates from the GC/MS isotopomer ratio, R,, are shown as a dotted line and those from R, as a thin solid line. The GC/C/MS data are shown as the thick line.


Llneerfty of GGGMS data

6m.o , 500.0

400.0 1 m.0

2W.0

100.0

0.0 0.0 103.0 2W.O 300.0 uxI.0 500.0 Bm.0

Calculated delta

Figure 5. Linearity of the GC/C/MS technique for the determination of the isotopic composition of tetra-TMS-ascorbic acid.

value, and hence obtain similar estimates in the errors involved in this technique as was done for GC/MS earlier. This is shown in Fig. 4. It will be seen that GC/C/MS is the more precise technique, but that a significant improvement is only apparent at low enrichments. In fact, above mole fractions of 5% there is a less than threefold advantage in wing GC/C/MS.

A further point to consider when choosing between the two methods is that of chromatogram crowding. If there is some chromatographic overlap, GC/MS can often be used to resolve the species since it measures the intact molecule at a mass number which is specific to that compound. In other words, the mass spectrum may be used to separate two chromatographically overlap- ping species. However, GC/C/MS combusts all the materials eluting from the end of the column, and mass spectral specificity is lost. Real samples of human plasma will contain many sugar species in addition to ascorbic acid, and total resolution on a GC column may be difficult. Paying due regard to this and in the light of the expected enrichments in our human study, it was decided that the GC/MS method should be the technique to be used.

HUMAN sEDY: EXPERIMENTAL PROCEDURE

Materials

[l - 13C]ascorbic acid (99 atom% I3C) was purchased from MSD Isotopes (Montreal, Quebec, Canada). Inosi- to1 and trichloroacetic acid were obtained from BDH (Poole, UK) and hexamethyldisilazane (HMDS), tri- methylchlorosilane (TMCS) and pyridine were obtained from Sigma (Poole, UK).

Subjects

Three healthy male non-smoking volunteers were recruited by advertisement. Written informed consent was obtained prior to the commencement of the study.

The study design and protocol were approved by the Ethical Committee of the Dunn Nutrition Unit.

Study protocol

On the morning of the study on arrival at our volunteer suite the subjects had an intravenous canula inserted and the first (predose) 2 ml blood sample was collected. Approximately 50 mg of [l - 13C]ascorbic acid dis- solved in water were administered orally and 2 ml blood samples were collected in heparinized containers at 15, 30, 45, 60, 90, 120, 150, 180, 240, 300, 360, 480, 600 and 720 minutes post-dose. Breath samples were also collected at various times during this period. The canula was then removed. After an overnight stay in the suite, a final blood sample was taken at approximately 1440 min post-dose by venepuncture into a 10 ml syringe. During the entire 24 h period when sampling was being performed the subjects were maintained on a diet free of ascorbic acid, but otherwise nutritionally adequate.

Sample preparation and analysis

Within 1 h of the blood being collected, plasma was separated by centrifugation, stabilized with 10% (v/v) aqueous trichloroacetic acid (TCA) and immediately frozen to -70°C for storage. The samples were not thawed out until just before analysis, when a 300 pl aliquot was removed, the precipitated protein removed by centrifugation and the supernatant spiked with 30 p1 of 100 pg ml-l aqueous inositol as internal standard. TCA was extracted with 1 ml of light petroleum and the aqueous fraction freeze-dried overnight. The residue was resuspended in 225 11 of pyridine and derivitized with 50 pl of HMDS and 25 pl of TMCS for 2 h at room temperature. The precipitated ammonium chlo- ride was removed by centrifugation and the supernatant analysed directly by GC/MS.

The ascorbic acid concentration was measured in the stabilized plasma by a standard fluorimetric method” using a Cobas-Bio centrifugal analyser.

746 L. J. C. BLUCK, A. P. IZZARD AND C. J. BATES

Chromatogram of sugars in plasma

1 G G

#

7.M 7.50 8.00 8.50 9.M 9.50 lo.w Retentbn Tim (min)

Figure 6. Typical chromatogram of sugars in plasma. The two large peaks marked G are due to enantiomers of glucose. The peak marked I is from the inositol added as an internal standard. The peak due to ascorbic acid (A) is incompletely resolved from one of the glucose peaks.

GC/MS was performed using an INCOS XL system (Finnigan MAT) equipped with a 30 m DB-5MS capil- lary GC column (J&W Scientific, Folsom, CA, USA). The typical chromatogram of plasma shown in Fig. 6 demonstrates the separation of the TMS ascorbic acid derivative from those of plasma sugars. Although there is incomplete chromatographic resolution, selected-ion monitoring allows the ascorbic acid to be studied without interference from the glucose enantiomers which flank it. This mode of mass spectrometer operation was used both for isotopomer analysis and also for quantitation, where the ascorbic acid molecular ion at mass 464 was measured in comparison with that of inositol at mass 612. It is important to note that unlike the Cobas determination, this gives the relative concentrations of only the unIabeZled ascorbic acid at each time point. Isotopomer analysis was performed using both EI and positive-ion CI spectra. In both cases the molecular ion group was monitored, with a dwell time of - 0.5 ms per mass.

Isotopomer data was recorded for the M, M + 1 and M + 2 ions. The mole fraction of labelled material was estimated from the expressions given for the isotopomer

ratios:

R1 = Rlo + (1 - AC)-lXB/(1 - X B ) Rz = R,' + [(Nc - l)AC(l - A,)-'

+ nsiASi(1 - A C ) - ' ( 1 - A,j - &i)-']XB/(l - x,)

Since A , is small (- 0.01) and replacing the expression in square brackets with the constant A, these become

R , = Rlo + x~/(1 - XB) R2 = Rzo AXB/(I - XB)

Substitution of the second into the first and rearranging gives

R2 = AR, + (Rzo - ARlo)

A least-squares fit of the isotopomer ratios allows the parameter R and the expression (Rzo+ AR,') to be obtained. The two basal isotopomer ratios are separated by comparison with the measured zero time values R,(O) and R2(0) using the condition IR,(O) - R,*I = IR,(O) - R2*1. It is now possible to make estimates of the mole fraction XB from both of the ratios R , and R ,

Typical time course of label in plasma

0 300 600 900 1200 1500 1800

time Figure 7. Experimental time course for labelled ascorbic acid in human plasma after oral administration of a 50 mg dose.


for all of the time points One of the time courses is shown in Fig. 7.

The breath samples were analysed on a Sira 10 isotope ratio mass spectrometer (VG Isogas, Middle- wich, UK).

RESULTS AND DISCUSSION

A simple compartmental model for ascorbic acid in the body is shown in Fig. 8. This is similar to that used by Piotrovskij et aL9 except that the non-linear absorption route has been omitted and the pathway for irreversible loss has been shifted from compartment 1 to compartment 2. Compartment 1 represents an ascorbic acid pool which is rapidly equilibrated with the plasma. This is the pool which was sampled. Compartment 2 corresponds to the ascorbic acid in the rest of the body and compartment 3 represents the ascorbic acid after inges- tion but prior to absorption. A delay has been introduced between compartments 3 and 1 to account for the time lag between administration of the dose and the appearance of label in the plasma. In common with

&p+p k12

p 0 2

Figure 8. Compartmental model of the disposition of ascorbic acid in man. Compartment 3 represents ingested ascorbic acid, compartment 1 the pool in close kinetic contact with the plasma and compartment 2 the deep body stores.

other workers, we have assumed that there is no loss of ascorbic acid from the gastrointestinal tract,’ which amounts to assuming 100% uptake of the administered dose. However, in contrast, we have assumed that uptake follows first-order kinetics, whereas it has been proposed that there are two parallel absorption routes, one following Michaelis-Menten kinetics and one which is first-order. Our approach is justified on the basis that the dose of ascorbic acid administered in this experiment was small, and replacement of the first-order absorption route with one obeying zero-order kinetics does not improve the fit to the model.

Pharmacokinetic parameters were obtained for each of the subjects by fitting the observed data to the model using a spreadsheet method in the first instance and performing a final fit using the SAAM I1 modelling package (SAAM Institute, University of Washington, Seattle, WA, USA). The advantage of the latter method is that likely errors in the calculated parameters may be obtained and investigations of the covariance of the parameters made. The results of the fitting are shown in Table 2.

The concentrations of ascorbic acid found in the plasma are in good agreement with the levels reported by Kallner et aL4 who found concentrations in the range 14-70 pmol dm-3. The distribution volume for pool 1 ( V , ) is unphysiologically large. Lumping together plasma and intercellular fluid, one might expect this pool to account for 20% of the body weight, or about 15 dm3. Experimentally we found at least twice this. Clearly this pool does not correspond to a simple physiological entity.

The total amount of ascorbic acid stored in the body is similar to that reported p r e v i o u ~ l y . ~ - ~ ~ ~ The relative dispositions between compartments 1 and 2 is similar to that reported by Kallner et ul.: but different from that deduced by Piotrovskij et ~1. ; ’ however, given the mag- nitude of the reported errors in the latter study, this is not unexpected.

The irreversible losses calculated from this model are higher than might be expected. In common with other workers, we found that no enrichment of breath C 0 2

~

Table 2. Pharmokinetic parameters derived from the time course

Subject

Parameter A D J

Age ( w ) 32 22 24 Weight(kg) 70 67 89

k , , (min-‘) 0.01505 (41) 0.05329 (54) 0.03249 (37) k,, (min-’) 0.00423 (17) 0.02874 (23) 0.02664 (31) k , , (min-’) 0.00161 (23) 0.00733 (13) 0.0041 1 (13) ko2 (min-‘) 0.00027 (1 5) 0.001 37 (4) 0.00076 (6)

Q, (mol) 0.00321 (39) 0.001 18 (28) 0.001 51 (32) Q, (9) 0.565 (69) 0.208 (49) 0.266 (56) Q2 0.0084 (26) 0.0046 (12) 0.0098 (25) Q2 (9) 1.48 (46) 0.81 (21) 1.72 (44) Q, +Q, (mol) 0.01 16 (30) 0.0058 (1 5) 0.01 13 (28) Q, +Q, (9) 2.05 (53) 1.02 (26) 1.98 (49)

mean plasma concentration (pmol dm-’) 60.5 (6.2) 36.8 (7.6) 46.0 (8.3) koiQ2 (mg min-‘) 0.40 (35) 1.11 (32) 1.31 (44)

D (min) 39 (3) 24 (3) 21 (3)

v, (dm9 53 (1) 32 (1) 33 (1)

Body pool per unit body mass(mg kg-’) 29 (8) 15 (4) 22 (6)

748 L. I. C. BLUCK, A. P. IZZARD AND C. J. BATES

occurred, indicating that there is no metabolic pathway involving exhalation.20,2 In the UK the current value for the recommended daily intake of ascorbic acid is 40 mg for normal adults. The indications from this model are that this would be insufficient to maintain the body levels at those found from this study. This may indicate that first-order kinetics do not adequately describe the metabolic clearance of ascorbic acid in man. In future studies we intend to continue collection of plasma samples for at least a further 12 h in order to try to resolve some of the ambiguities encountered.

man using a stable isotope-labelled tracer. Although it is expected that the combustion technique GC/C/IRMS would give more precise results at low levels of enrichment, the pool sizes and dosing levels encountered in practice make GC/MS a feasible alternative. We have conducted a small human study, and found that the data obtained are adequately described by a two- compartment model. The kinetic parameters deduced are similar to those reported from other published studies.

Acknowledgements

We are grateful to VG for the opportunity of using a GC/C/MS instrument for the analysis of ascorbic acid standards. A.P.I. is indebted to the MRC for the provision of a research studentship. L.J.C.B. is indebted to the Bristol Myers Squibb Foundation for a research Fellowship.

CONCLUSION

We have demonstrated that the CCmbination of GC and MS is suitable for measuring ascorbic acid kinetics in

REFERENCES

1. C. J. Schorah, Proc. Nutrition SOC. 40, 147 (1 981 ). 2. A. F. Abt, S. von Schuching and T. Enns, Am. J. Clin. Nutri-

tion, 12, 21 (1963). 3. E. M. Baker, R. E. Hodges, J. Hood, H. E. Sauberlich, S. C.

March and J. E. Canham.Am. J. Clin. Nutr. 24,444 (1 971 ). 4. A. Kallner, D. Hartrnann and 0. Hornig, Am. J. Clin. Nurr. 32,

530 (1979). 5. A. B. Kallner, D. Hartrnann and D. H. Hornig, Am. J. Clin.

Nutr. 34,1347 (1 981 ). 6. A. J. Mellors, D. L. Nahrwold and R. C. Rose,Am. J. Physiol.,

233, E374 (1 977). 7. G. L. Atkins, B. M. Dean, W. J. Griffin and R. W. E. Watts, J.

Biol. Chem. 239,2975 (1 964). 8. J. C. Stewart and C. C. Booth, Clin. Sci. 23,15 (1 964). 9. V. K. Piotrovskij, Z. Kallay, M. Gajdos, M. Gerykova and T.

Trnovec, Biopharm. Drug Dispos. 14.429 (1 993). 10. B. Holland, I. D. Unwin and D. H. Buss, Fruit and Nuts-First

Supplement to ‘McCances and Widdowsons. The Composi- tion of Foods. Royal Society of Chemistry, London (1 992).

11. T. Preston, Plant Cell Environ. 15, 1091 (1 992). 12. J. H. Allison and M. A. Stewart, Anal. Biochem. 43, 401

13. J. C. Deutsch and J. F. Kolhouse, Anal. Chem. 65, 321

14. H. Craig, Geochim. Cosmochim.Acta 12,133 (1957). 15. C. Corbelli, G. Toffolo, D. M. Bier and R. Nosadini, Am. J.

16. J. Rosenblatt and R. R. Wolfe, Am. J. Physiol. 254, €526

17. C. CorbeHi, G. Toffolo and D. M. Foster, Am. J. Physiof. 262,

18. J. A. Vogt, D. A. Chapman, V. R. Young and J. F. Burke, Biol.

19. J. P. Vuilleumier and E. Keck. J. Micronutr. Anal. 5, 25

20. L. Hellman and J. J. Burns, J. Biol. Chem. 230, 923 (1 958). 21. E. M. Baker, H. E. Sanberlich, S. J. Wolfskill, W. 1. Wallace

and E. E. Dean, Proc. SOC. fxp. Biol. Med. 109,737 (1 962).

(1 971).

(1 993).

Physiol. 253, E551 (1987).

( 1 988).

E968 (1 992).

Mass. Spectrom. 22,600 (1 993).

(1989).

Documents

Measurement of Ascorbic Acid Kinetics in Man Using Stable Isotopes and Gas Chromatography/Mass Spectrometry