October, 19501 COX: OBSERVATIONS ON VITAMIN D 521 Observations on the Spectrophotometric Estimation of Vitamin D BY H. E. COX (Read at the meeting of the Society on Wednesday, May 3rd, 1950) SYNoPsIs-The International Unit of vitamin D has recently been changed by the World Health Organisation, but the new unit has absorption constants similar to those of the old unit and gives a similar reaction with the acetyl chloride - antimony chloride reagent first proposed by Nield, Russell and Zimmerli. A study of this method shows that it works well for control purposes provided that (a) the amount of vitamin D present is sufficient, (b) the amount of vitamin A is not more than about five times that of the vitamin D and (c) suitable correction is made for the absorption due to certain other sterols.Data are given on interfering substances and on commonly occurring fats that are used in the preparation of commercial vitamin products. When there is too high a concentration of vitamin A it may be separated by chromatography. THERE is at present no recognised chemical, physico-chemical or even microbiological method applicable to the estimation of vitamin D in foodstuffs. The biological method is so costly and time-consuming that there is great need for a simpler or cheaper process and every contribution to this end is to be welcomed. To appreciate the problem it is necessary first t o define what has to be determined and then to consider possible methods and probable522 COX: OBSERVATIONS ON THE [vol. 75 interfering substances. Until quite recently the International Unit of vitamin D was defined in terms of a solution of irradiated ergosterol that: was in fact mainly a solution of vitamin D, and was equivalent to 0.025 pg.of calciferol. Calciferol is now available as a synthetic chemical of a high degree of purity, has the empirical formula C,,H,OH, and is the subject of a specification in the British Pharmacopoeia. More recently the World Health Organisation, which is the body charged with the duty of defining units and issuing appropriate international standard preparations of certain of the vitamins, has decided that the solution of irradiated ergosterol, being mainly of vitamin D2, has not proved to be sufficiently representative of the D vitamins. Also a suggestion made at an earlier conference that the standard should be replaced by pure crystalline calciferol is now unacceptable. So the standard has been re-defined in terms of the alcohol C2,HBOH, formerly known as vitamin D,.This compound differs from calciferol in the nature of its side-chain; the methyl group at C, is absent and there is one double bond less. It is available as a pure substance in crystalline form and has a melting- point of 87" to 89" C., [a]: + 110" and E:&. at 265 mp. = 490; 0.025 pg. of this substance now becomes the International Unit of vitamin D. It will be noted that this substance is defined as having a high extinction coefficient; in 1 per cent. solution in alcohol it is as high as 480, and E, the molecular extinction coefficient, is 18,800. So it is evident that it would be quite easy to determine the amount of this sub- stance in a pure solution or in the absence of interfering substances.Besides natural products containing vitamin D, there are many proprietary preparations that contain and claim it, and the Labelling of Food Order requires that certain specified proportions of the normal daily requirement (500 I.U. of vitamin D) shall be present. This brings an insistent demand for its estimation, and it might be said that it is the duty of the Public Analyst to determine it. In general, he cannot do so. The vitamin is always associated with the fatty portion of the food in which, of course, it is soluble. Even if the fat is extracted, and saponified and the unsaponifiable matter dissolved in pure solvent, such as cyclohexane, there is almost always enough impurity to mask the specific absorption curve of the vitamin. Nearly all the sterols associated with the unsaponifiable portion of fatty oils have absorptions in the region of 260 mp., so that a direct method such as is used for vitamin A is inapplicable.Moreover, if vitamin A is also present in quantity substantially greater than vitamin D, its absorption will overlap the 260 mp. region. The well known Can- - Price reaction for vitamin A was, of course, very attractive and as long ago as 1936 it was observed that vitamin D or calciferol gave a pink or orange colour with the same reagent, but it was not sufficiently stable and reproducible for a quantitative method, and many substances interfere. In 1940, however, Nield, Russell and Zimmerlil~2 introduced the use of acetyl chloride into the reagent and since then various improvements have been made or proposed.The reagent is a solution containing 20 per cent. w/v of antimony chloride to which is added 4 per cent. of pure acetyl chloride. The reagent must be fresh and must be quite free from alcohol. When 9 volumes of this reagent are added to 1 volume of a solution of vitamin D in cyclohexane or in chloroform, an orange colour is rapidly developed; it reaches a maximum in about 4 minutes and retains it for about 12 minutes. The optimum conditions of time and temperature have been studied by various authors. The colour is fairly specific and has a characteristic absorption curve with a maximum a t 600 mp. At this wavelength the value of EiZ!.is 1880, a figure that makes the test four times as sensitive as the absorption at 265 mp. Vitamin A with this reagent gives the well known blue colour; this fades quickly and is, indeed, somewhat suppressed by the acetyl chloride. The blue colour has Em,,. at about 620 mp., which is well clear of the 500 mp. line. However, when there is much vitamin A relative to the vitamin D, the blue colour does seriously interfere. In my experience the blue colour begins to cause difficulty when the quantity of vitamin A is more than about five times that of the vitamin D. If calciferol is the only sterol present in the mixture the problem is easy and the determina- tion can be made with a good degree of accuracy down to small quantities. Unfortunately the orange or pink colour is given in greater or less degree by various sterols and some polyenes.The literature is somewhat confusing and, as one does not know just what sterols may be present in an unknown material, it is difficult to apply any general correction. Those sterols other than calciferol that do give a colour with acetyl chloride - antimony chloride all give a less intense one, so there is a practical possibility of establishing blanks applicable toOctober, 19501 SPECTKOPHOTOMETRIC ESTIMATION OF VITAMIN D 523 particular types of material. Measurement of the absorption of some of the sterols gave the following extinctions- E;&. at 500 mp. Calciferol . . .. .. .. .. 1800 Sitosterol . . .. .. .. .. nil Phytosterol (B.D.H.) . . . . .. nil Cholesterol .. .. .. . . .. 80 Ergosterol . . .. .. .. .. 3 00 It will be seen that each is small compared with that of calciferol, and for such a quantity as 10 pg., which is convenient for vitamin D, or D,, in a total volume of 2 ml., the extinction will be- E;& a t 500 mp. Calciferol . . .. .. .. .. 0.90 Cholesterol . . .. .. .. f . 0.04 Ergosterol . . .. .. .. .. 0.15 Phytosterol .. .. .. .. 0.00 As the composition of a sample in terms of its sterols is not always known, it becomes necessary to establish data on the basis of the common fats. A few such results may be given; they refer to the unsaponifiable matter calculated on the original sample, i.e., 1 per cent. of the oil in the mixture after addition of the antimony chloride reagent. Ei& at 500 mp.Arachis oil . . 1 . .. .. .. 0-06 Butter fat . . .. .. .. .. 0-05 Castor oil . . .. .. .. .. 0.06 Cocoa butter . . .. .I .. 0.025 Cod-liver oil .. .. .. .. 0.057 Olive oil . . .. .. .. .. 0.033 Tea-seed oil .. .. .. .. 0.025 It will be noticed that margarine is not included; this is because its composition is variable and generally unknown and is likely to include large proportions of fish oils which themselves vary greatly in their content of unsaponifiable matter and its constituents. It has been shown by Henry and Kon3 that cows' butter maintains a constant level of about 0.07 to 0.1 I.U. per gram in the winter and 0.55 to 0.97 I.U. during the summer. These quantities would account for an extinction of only about 0.002 on 1 gram, which is much too small for satisfactory measurement and would be overshadowed by irrelevant absorption. It appears at present that the spectrophotometric method can only be applied when the composition is known or may be assumed.The method has distinct usefulness as a control on manufacture, as once the necessary blanks have been set up it is only necessary to observe the extinction under standardised conditions and ensure that the total E value does not fall below a pre-determined figure. The Public Analyst is not always in the happy position of knowing what is present or of having a good blank. He can, however, establish a maximum. Clearly the vitamin D cannot be more than is indicated by the value E:,%, = 1800 and is almost certainly less, unless the mixture under examination contains calciferol but no animal oil.If the oils or fats present are known or can be determined, then the extinction due to them can be estimated within a margin of error and allowance made. This method has been applied to the control of certain vitamin products now enjoying a considerable sale and it has been found to work. The amounts found have generally agreed reasonably well with the amounts of vitamin added by the manufacturer. For example, the unsaponifiable matter extracted from 3 g. of a sample of a well-known product containing about 33 per cent. of cocoa butter. The extraction of the unsaponifiable matter was carried out as for a cod-liver oil with all due precautions against atmospheric oxidation. The un- saponifiable matter was dissolved in 10 ml. of cyclohexane. To 0.2 ml.of the solution was added 1.8 ml. of the acetyl chloride - antimony chloride reagent a t a temperature of 20" C. The colour was a t first bright blue, but the blue soon faded and a pale orange tint remained; it is necessary to use a quantity of the sample such that the blue colour disappears within 4 or 5 minutes. The mixture was transferred to the l-cm. cell of a Spekker spectrophoto- From the observed value is deducted the appropriate blank, which is usually about 0.03. It is not practicable to measure less than about 50 or 100 I.U. per gram. meter and the extinction measured after 6 minutes, using the stages 0.02,0-04,0.06 . . . 0.30.524 COX: OBSERVATIONS ON VITAMIN D [Vol. 75 If 3-0 grams were taken, the final concentration is 3 per cent., so the amount of vitamin D is: (observed E - blank) x 600 I.U.per gram. It sometimes happens that the amount of vitamin A is inconveniently large in relation to the vitamin D. This is shown by the blue colour being so intense that it does not fade sufficiently within 5 minutes for a clear reading to be obtained, while if a smaller quantity be taken the colour due to vitamin D is too faint for satisfactory measurement. In these circumstances I have applied the chromatographic method devised for whale liver oils by Gridgeman, Gibson and S a ~ a g e . ~ All the vitamin A can be removed and the cholesterol - calciferol fraction afterwards washed out with petroleum spirit and ether. This enables a satisfactory reading at 500mp. to be obtained if it is known that there is no substantial amount of cholesterol present.It remains to be mentioned that certain other reactions have been proposed, e.g., with glycerol dichlorhydrin containing acetyl chloride (by Sobel, Mayer and Kramers) , which are said to be more specific for vitamin D and to be unaffected by cholesterol. However, the extinction curves have a less well-designated peak and the method has not been extensively studied. REFERENCES 1. 2. -,- , Ibid., 1943, 148, 245. 3. 4. 6. 10 BILLITER SQUARE Nield, C. H., Russell, W. C., and Zimmerli, A, J . Biol. Chem., 1940, 146, 73. Henry, K. M., and Kon, S. K., Biochem. J., 1942, 36, 456. Gridgeman, N. T., Gibson, G. P., and Savage, J. P., Analyst, 1948, 73, 662. Sobel, A. E., Mayer, A. M., and Kramer, B., Ind. Eng. Chem., Anal. Ed., 1945, 17, 160. THE LABORATORY LONDON, E.C.3 DISCUSSION DR. J. GREEN said that the reagent must be meticulously prepared and be quite pure, or very erratic results would be obtained. With oils containing usual amounts of vitamin A or products containing irradiation intermediate sterols or 7-dehydrocholesterol, direct estimation fails and separation techniques must be used. DR. Cox, in reply, agreed with Dr. Green that the method was not applicable to oils containing relatively large proportions of vitamin A, or to whale oils and products containing unknown sterols. The method he was putting forward should only be applied to known products or mixtures when the necessary blanks and controls could be established. He agreed as to the need of preparing pure alcohol-free reagents.