THE JOURNAL OF VITAMINOLOGY 18, 165-171 (1972)

Determination of Vitamins D by Gas-Liquid

Chromatography

II. Rapid Assay for Vitamin Dz in the Presence

of Vitamins A and E1,2

KIYOSHI TSUKIDA AND KAYOKO SAIKI3

Kobe Women's College of Pharmacy, Higashinada-ku, Kobe

(Post No. 658)

(Received July 5, 1972)

In addition to our previous findings on GLC and mass spectroscopic determination of vitamins D, extensive studies on the effect of interfering substances have

presented a new simple GLC determination method of vitamin D2 in the presence of vitamins A and E. Thus, by employing TLC before applying a sample for

GLC evaluation, vitamin E and its related compounds which interfere the GLC

determination are eliminated completely and the greater part of vitamin A is also removed in the form of anhydrovitamin A simultaneously. Exhaustive conver

sion of vitamin A into its anhydro derivative and removal of unreacted vitamin A from vitamin D are not necessary. The proposed method is a simple, inexpen

sive, accurate, and time-saving quantification method of vitamin D2. Main advantages and limitation of the method are discussed in detail.

In a previous paper we presented the data concerning the basic conditions for GLC determination of vitamins D (1). It was also reported

(2) that coexistence of vitamin A could be allowed for GLC determination of vitamins D,

provided A/D value was 50 or less. Otherwise, vitamin A gave rise to cause a tailing phenomenon over vitamin D peaks (twin peaks, pyro and isopyro peaks) which have primarily well differentiated retention times from those of vitamin A and its related compounds. Furthermore, it should be stressed that the presence of vitamin E (a-tocopherol) gave a serious trouble

for the present purpose, though so-called degradation products of vitamins A and D did not

interfere an analysis except for a particular case. For elimination of vitamin E, several attempts employing a basic alumina (3) or an Amberlyst resin column (4) did not give any improvement on GLC analysis of vitamin D.

We now propose a simple and a rapid GLC determination of vitamin D2 in the presence of vitamins A and E by adopting TLC for eliminating vitamin E etc. and for suppressing A/D value below the tolerance limit.

PROPOSED DETERMINATION METHOD FOR VITAMIN D2

a) Saponification and benzene-washing

(Isolation of the unsaponifiable matter)

1 Determination of Vitamins D. III; Part II, Internat. J. Vit. Nutr. Res., 42 (2), in press.2 Following abbreviations are used: GLC

, gas-liquid chromatography; TLC, thin-layer chromatography; A/D, a content ratio of vitamin A to vitamin D; pyro, pyrocalcif erol; isopyro, isopyrocalcif erol; i, d., internal diameter; BHT, butylated hydroxytoluene; G. R., guaranteed reagent.

3 月田 潔,斎 木加代子

165

166 TSUKIDA AND SAIKI

Reagents:Benzene, G. R.Ethanol 99.5% v/v, G. R.2N KOH in ethanol, fresh0.5N and 1N KOH in water, fresh25% Sodium ascorbate in waterBHT, pure

Procedure:

Proceed according to the Mulder's method

(3) essentially. Weigh a sample containing not

less than 50ƒÊg of vitamin D2 into a saponifica

tion flask (200ml). Add BHT (10mg), dissolve

in ethanol (50ml), and then add 2N KOH in

ethanol (10ml) and 25% sodium ascorbate (1

ml) with swirling. Keep a sample solution at

80•‹(bath temperature) for 30min in a stream

of nitrogen and then cool immediately. Add

100.0ml of benzene pipetted accurately and mix

together. Pour the solution into a separatory

funnel without rinsing. Add 1N aqueous KOH

(40ml) and shake vigorously for 10 sec. After

standing for 5min, separate the layers and

discard the turbidz aqueous layer. Wash the

benzene layer similarly with 0.5N aqueous KOH

(40ml), and then repeatedly with water (40ml

portions) until the aqueous layer remains neutral

to phenolphthalein. Filter the benzene solution

through a phase separating paper (Whatman

1PS, siliconized) and store it in a well-stoppered

flask.

b) Acid treatment and TLC (Complete

elimination of vitamin E including its relating

compounds and simultaneous removal of the

greater part of vitamin A)

Reagents:

Acetone, G. R.0.1N HCl in methanol, fresh0.1N KOH in water, fresh

Reference solution containing vitamin D2 and previtamin D2, prepared by refluxing of vitamin D2 (crystals, Philips-Duphar) solution in isooctane (G. R.) for 30min, followed by evaporation to dryness in vacuo and by dissolving the residue in an appropriate volume of acetone (G. R.)

TLC:

Plate (250ƒÊ-thick, 25•~25cm), Kieselgel

HF254 Merck; use the plate within 24hr of its

preparation (activated at 120•‹for 1hr)

Developing solvent, benzene (G. R.)-ethy

lene dichloride (G. R.)-acetone (G. R.)=70:

30:0.5; when vitamin E is not included con

spicuously, another solvent system, e.g., benzene

(G. R.) acetone (G. R.)=9:1 can be employed

Procedure:

Add approximately equi-volume of 0.1N HCl

in methanol to the benzene solution described

above, and stand for 15min at room tempera

ture. Neutralize the solution immediately with

0.1N KOH aqueous solution, and wash repea

tedly with water until the aqueous layer remains

neutral to phenolphthalein. The acid-treated

benzene solution thus obtained is colored yellow

by conversion of the large portion of vitamin A

into anhydrovitamin A. Filter the benzene

solution through a phase separating paper, pipet

50ml of the filtrate, and evaporate in vacuo.

Dissolve the residue in acetone (1ml) and use

it for subsequent TLC as soon as possible.

Streak the acetone solution (0.2ml) (pref e

rably ca. 20-50ƒÊg vitamin D2) onto a thin

layer plate and also apply a single spot of the

reference solution containing vitamin D2 and

previtamin D2 on an extension of the basic

line. Develop the chromatogram for ca. 1hr

and examine the result quickly under ultraviolet

light (254nm). Encircle the corresponding area

at the same height as the spots of vitamin D2

and previtamin D2 ex the reference solution.

Scrape off this area of Kieselgel powder (vita

min D2+previtamin D2, concomitant vitamin

A is allowed) with a flat spatula. Extract the

substances with acetone, followed by washing

with the same solvent (20ml in total) and filter.

c) GLC determinationReagents:Internal standard solution, 7-dehydrochole

steryl acetate (crystals, pure) in acetone (G. R.)

Standard mixture of vitamin D2 and the internal standard, fresh; details are as described in the procedure

GLC:Apparatus, Gas Chromatograph GC-4AP

(Shimazu Seisakusho) equipped with a hydrogen flame ionization detector; Column, glass, 1.5m

GLC DETERMINATION OF VITAMIN D 167

long and 4mm i.d., packed with 1.5% OV

- 17 on Shimalite W (80-100 mesh, silanized);

Operational parameters, temperatures (column

250•‹, injector port 260•‹, and detector 300•‹),

flow rates (nitrogen [carrier gas] 60ml/min,

hydrogen [detector gas] 0.8kg/cm2, and air 1

kg/cm2)

Procedure:

Add the internal standard solution to the

sample solution described above (vitamin D2:

internal standard=2:1 is preferable). Apply

1-5ƒÊl of the acetone solution concentrated to

an appropriate volume (preferably 0.5-2ƒÊg vita

min D2 per an applied volume of the solution)

into the instrument by on-column injection using

a 10ƒÊl microsyringe. Estimate a peak area

(half-width•~peak height) ratio A (pyro or

isopyro/internal standard) from the recording

chart. From an independent GLC tracing,

determine the standard peak area ratio S bet

ween vitamin D2 (pyro or isopyro) and the

internal standard ex a newly prepared standard

mixture (vitamin D2: internal standard=2:1 by

weight, internal standard is the same concentra

tion solution when A value was estimated).

Calculate a quantity of vitamin D2 as follows:

A: S=x:w

x: Vitamin D2 (weight) in a final sample solution applied for GLC

w: Vitamin D2 (weight) in a final solution of a standard mixture applied for GLC

RESULTS AND DISCUSSION

a) GLC of Vitamin D2 and Interfering Substances

Since chemical methods heretofore proposed for microdetermination of vitamin D commonly suffer primarily from lack of specificity, it has been a serious problem how to separate the vitamin in a pure state free from numerous interfering substances. Though it is simple and one of the most representative separation techniques, TLC can not differentiate vitamin D from tachysterol. On the contrary, TLC does separate previtamin D from vitamin D and unfortunately this phenomenon gives rise to trouble how to deal with a thereto-equilibrium

which exists between these two compounds. It is quite impossible to separate minute amount of vitamin D quantitatively on a limited plate from large excess of interfering substances. It is for this reason that purification of vitamin D still relies upon an employment of multi-columns

(adsorption, partition, or gel filtration) in spite of its great disadvantages, viz., insecure reproducibility of a column condition and time-consuming operation. Among recent techniques of instrumental analysis, infrared and nuclear magnetic resonance spectroscopies are not always adequate for microdetermination at present time. Considering every qualifications for a routine assay involving f utural prospects, GLC technique would be most practical and would have unlimited possibilities (1), though mass spectroscopy is the most direct determination method of all (5).

On GLC determination of vitamin D, several papers have been published successively since the work of Ziff er et al. appeared in 1960 (6). Basic studies on an analytical form (unmodified vitamin D), stationary phase of the column (1.5

% OV-17), internal standard (7-dehydrocholesteryl acetate), and operational parameters were already reported by present authors, excellent calibration curves of vitamins D between peakarea ratio and weight ratio (vitamin D/internal standard) had being obtained (1, 2). Since a GLC separation tracing itself reflects a determination value of a definite ingredient immediately, tailing phenomenon as well as overlap with other peaks caused by concomitants would then be urgent problem for present purpose, even though there exist many problems concerning multiplicity of pre-treatment depending upon a form of a sample for analysis (3, 4, 7).

Vitamin A and its related compounds on GLC Under the operated condition, vitamin A (possibly as degradation products) possessing a smaller retention time was eluted rapidly from the column and up to 50 times the weight of vitamin A could be allowed as a concomitant for the present (Table 1). With increasing

quantities over this tentative tolerance limit, vitamin A did induce tailing phenomenon over vitamin D peaks region (Fig. 1). Similar observations were also obtained on the effects of vitamin A related compounds (anhydrovita-

168 TSUKIDA AND SAIKI

TABLE 1GLC determination of vitamin D2 in the presence of vitamin A*

* Vitamin D2 is determined by GLC using the

pyro -and the isopyro-D2 peaks. The ratio of total vitamin D2 to the internal standard is 2:1.

FIG. 1 Gas chromatographic behavior of a mixture of vitamins A acetate and D2

a: pyro thermal cyclization product. b: isopyro thermal cyclization product. Weight ratio of vitamin D2 to vitamin A acetate: A, 1:1; B, 1:50; C, 1:100

min A etc.) and of vitamin A degradation products prepared artificially by irradiation, oxidation (8), or heating of the original vitamin. However, when a sample was treated according to the procedure b) described above, even 500 times the weight of concomitant vitamin A and its related compounds did not interfere with an analysis at all.

Vitamin D related compounds on GLC Interfering effect of so-called vitamin D degrada

tion products obtained by irradiation (white or ultraviolet light) or by prolonged standing at room temperature of vitamin D in crystals or in various solution was investigated. Among these compounds, only lumisterol and a product

(only end absorption in its ultraviolet spectrum) isolated from ultraviolet irradiated vitamin D in hexane solution were not separated quantitatively from vitamin D2 (pyro peak). In this case it is

possible, however, to estimate vitamin D based on isopyro peak instead of pyro peak unless the vitamin content is extremely small.

Vitamin E and its related compounds on GLC Vitamin E and most of its related compounds give serious obstacles for GLC determination of vitamin D. In most types of vitamin drugs, for example, vitamin E is usually contained in a larger amount compared with other vitamins and the peak of vitamin E or its acetate overlaps completely on the pyro or on the isopyro peak, respectively (Table 2). Thus, elimination of vitamin E and its related compounds should now be projected as one of the most important pre-treatment for GLC determination of vitamin D. Separation of vitamin E by employing an appropriate column treatment such as basic alumina or Amberlyst resin column did not give any improvement at a level of GLC analysis, viz., the former column

produced new minor peaks (isomers or a degradation product ?) ex vitamin D and the latter column could not suppress an elution of vitamin

E degradation product (and vitamin E*l) responsible for overlapping on the vitamin D peaks region (Fig. 2). However, when a sample is treated according to the procedure b) described

TABLE 2

Relative retention data on 1.5 on OV-17 column

* Vitamin D2 and D3 give twin peaks, the larger

pyro peak being eluted earlier.

*1 Over 10 times the weight of vitamin E could not be suppressed .

GLC DETERMINATION OF VITAMIN D 169

FIG. 2 Gas chromatographic behaviors of vita

mins D and E

•c•c vitamin D2, -a-tocopherol (a) or ƒ¿-toco

pheryl acetate (b)

A: genuine samples, B: after saponifying and

passing through a basic alumina column, C: after

saponifying and passing through an Amberlyst A-26

column

above, these interfering substances can be eliminated completely.

b) Acid Treatment and TLC

For complete elimination of vitamin E and its related compounds, TLC separation was investigated. Concerning vitamin A, it was

FIG. 3 TLC chromatograms of vitamins A, D, and E g, genuine; s, after saponification; a, after acid treatment Fluorescence: 1B, light blue; Y, yellow; V, violet; RB, reddish brown* Vitamin A acetate: D:E acetate=140:1:160 by weight

practically impossible to separate vitamin D

from excess amount of vitamin A quantitatively,

and hence present TLC should aim at a restrain

of A/D value below 50 or less. This was done

by converting the greater part of vitamin A

into anhydrovitamin A in advance, followed by

TLC separation. When large excess of vita

mins A and E exist, the system of benzene

ethylene dichloride acetone (70:30:0.5) is pre

ferable. Since this developing agent is not

projected for separation of vitamins A and D,

but for complete elimination of vitamin E and

its related compounds as well as anhydrovitamin

A, it is possible to employ another solvent

system, e,g., benzene acetone (9:1), provided

vitamin E is not contained significantly. On a

TLC plate, a rather complicated chromatogram

ex vitamin E was observed and BHT was

found at a front with the solvent. Preferable

amount of vitamin D on a TLC plate would

be 20-50ƒÊg. Typical TLC pattern of vitamins

D, E, and A (genuine, after saponification, and

after acid treatment) were illustrated in Fig. 3

and GLC recoveries of vitamin D both after

acid treatment and after TLC were given in

Table 3.

c) GLC Determination of Vitamin D on a

Laboratory-blended Sample

Surprisingly, little information have been

available on the determination of vitamin D in

the presence of vitamins A and E except for a

few methods employing multi-columns (3, 4) or TLC (7). A GLC determination of vitamin D2 (as the trimethylsilyl ether) in multiple-vitamin tablets, using a 3% OV-210 column was reported by Edlund et al., but the same authors concluded that this method could not be applicable to those containing vitamin E (9)*2. Recently, Feeter et al. (10) separated vitamins D and tocopherols in the form of the propionate on GLC using 2% SE52 column. Even apart from a few problems such as preparation and conditioning of the column or reaction rate and

yield of esterification etc., difficulty will

*2 It may be applicable with much better result than our present method to those containing vitamin E,

only if vitamin E is eliminated according to our present procedure.

170 TSUKIDA AND SAIKI

TABLE 3GLC recoveries of vitamin D2

*1 A ratio of the pyro or the isopyro peak area to the internal standard peak area .

Total vitamin D2 to the internal standard is 1:1 by weight.

*2 Apply the acetone solution in the instrument.

*3 Vitamin D2 (50ƒÊg) is developed on a plate, eluted with acetone. The internal

standard is added and the solution is concentrated prior to its GLC analysis.

*4 After anhydration, neutralization, and washing, the internal standard is added to

the benzene solution. Evaporate the solvent and the residue is dissolved in

acetone.

*5 Saponificationacid treatment TLCGLC; Vitamin A:D:E in a sample=140:

1:160 (by weight).

be encountered at once when a sample is contaminated with a large excess of vitamins A and E, and qualification of trioctanoin for an internal standard is also questionable.

Our method indicates that overall recoveries of vitamin D in a laboratory-blended sample

(vitamin A:D:E=140:1:160 by weight) are satisfactory and sufficiently accurate (Fig. 4 and Table 3). Four to five hours are required for an overall analysis. Practical applicability of our method for determination of vitamin D in commercial multiple-vitamin drugs was confirmed in some pertinent experiments.

d) Scope and Limitation

Main characteristics (Table 4) and advantages of our present method could be summarized as follows:

1. Unmodified vitamin D2 is adopted as an

FIG. 4 GLC determination of vitamin D2 in a laboratory-blended mixture (vitamin A:D:E=140:1:160 by weight)

a:pyro, b:isopyro, c:internal standard

analytical form for GLC determination.2. By employing TLC, vitamin E and its

related compounds are eliminated completely

and the greater part of vitamin A is also

GLC DETERMINATION OF VITAMIN D 171

TABLE 4Determination of vitamin D2 in the presence o f vitamins A and E

removed in the form of anhydrovitamin A simultaneously. Complete conversion of vitamin

A into its anhydro derivative and separation of vitamin D from unreacted vitamin A are not

necessary. Even a sample possessing a high

A/D value can be analyzed without any trouble.

3. Tedious pre-treatment using multi-colu

mns can be omitted. It is a simple, inexpensive,

accurate, and time-saving quantification method of vitamin D2.

4. Consideration on previtamin D and

tachysterol are not necessary.

5. Involving futural prospects such as ad

vanced automation of analysis, it possesses the

qualification for an excellent routine assay.It should be also pointed out as a minor

disadvantage of this method that TLC procedure has still time-consuming factor in some sense

and scraping of Kieselgel powder may cause

fluctuating data in analysis, though our scraping

procedure itself is not necessary to be so strict as in general case.

REFERENCES

1. Tsukida, K., and Saiki, K., J. Vitaminol., 16, 293 (1970).

2. Tsukida, K., and Saiki, K., presented at the 126th Meeting of Fat-Soluble Vitamins Research Committee (Jul., 1970).

3. Mulder, F. J., de Vries, E. J., and Keuning, K. J., Pharm. Weekblad, 100, 1457 (1965).

4. Ueda, F., Makino, T., Kazama, A., and Wata nabe, K., Vitamins, 44, 31(1971); J. Vitaminol., 17, 142 (1971).

5. Tsukida, K., and Saiki, K., Internat. J. Vit. Nutr. Res., 42, (2) (1972), in press.

6. Ziff er, H., Vanden Heuvel, W. J. A., Haahti, E. O. A., and Horning, E. C., J. Am, Chem. Soc.,

82, 6411 (1960).7. Bolliger, H. R., and Konig, A., Z. Analyt.

Chem., 214, 1 (1965).8. Kobayashi, T., Vitamins, 32, 482 (1965).9. Edlund, D. O., and Anfinsen, J. R., J, Ass. Off. Analyt. Chem., 53, 287 (1970).

10. Feeter, D. K., Jacobs, M. F., and Rawlings, H. W., J. Pharm. Sci., 60, 915 (1971).