DOCTORAL DISSERTATIO N SERIES To A T hysici/ Chemical Of Vitamins D Ass a TITLE AUTHOR open JjruceYounq DATE ^ UNIVERSITY. DEGREE m i I Michigan State College M PUBLICATION NO. ^ 41 uUNIVERSITY wr IT m ANN ARBOR 66 7 MICROFILMS • MICHIGAN STUDIES RELATING TO A PHYSICAL CHEMICAL METHOD OF VITAMINS D ASSAY by Robert Bruce Young A TRESIS Submitted to the Faculty of Michigan State College In Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy Department of Chemistry East Lansing, Michigan' 1943 INTRODUCTION Numerous attempts have been made to quanti­ tatively estimate vitamins D by chemical means. Much of this work was directed toward selective \ reactions, usually colorimetric, by which vitamins D could be determined in the presence of other substances. This search was complicated by the exceedingly complex mixture of vitamin D like substances and other organic material always present with the natural vitamin D. One of the earlier methods, and probably the most specific for vitamins D, concerns the Tortelli-Jaffe color reaction (11). The test consists of layering an alcohol solution of the vitamin with a solution of bromine in chloroform. V/hile the reaction is fairly specific, it is difficult in application and ergosterol interferes. The color reactions of benzaldehyde with sterols and steroids on underlayering v/ith con* centrated sulphuric acid has been the object of investigations (12, 13). is not specific. I This reaction, however, ^ An application of kinetic colorimetry was proposed by Raoul and Heunier (8) in which the difference in the speed of reaction of vitamins D and other sterols was used to estimate th<3 vitamins. This method is inexact and the results t are confused by mixtures and by the presence of vitamin A ^ ' { A colorimetric method, based on the reaction of vitamins D with antimony trichloride, proposed by Erockmann and Chen (5) seemed to offer the most promise. The reagent used was later modified by several investigators (1, 2, 7 and 8). This method, however, depends on the separation of vitamins D from most of the accompanying material. Thip latter was accomplished by Kingsley (4) and, using the reagent of Brockmann and Chen, modified by Nield, et al, (1), excellent results were ob­ tained for about 80 natural fish oils. The present investigation is, for the most part, concerned with the quantitative estimation of irradiated ergosterol by the method proposed by Kingsley. # y \ Early work indicated ^hat assay of samples of irradiated ergosterol by the physical chemical method resulted in values approximtel/ half those obtained by biological assays Abco^clingly it was decided to investigate thoroughly Some of the factors influencing the physical.chemical determination of vitamin D2 and if possihi%\?tO elim>- inate the above mentioned discrepancy. s 4 A ,^ j— "• (1) \ APPARATUS AND MATERIALS The chromatograph tuhes used were Twsett \ columns made from pyrex test tubes 1.7 cm. in * diameter and 14 cm. in length. All extinction values, unless otherwise specified, were made on a visual type universal photometer equipped with a Martins Polarizing unit. The cells were 1 and 5 cm. matched sets with glass spacers and quartz end plates. The photoelectric photometer used was a grating* type and was equipped with 1 cm. open type cells and 1 and 5 cm. stoppered cells with glass spacers and Corex end plates. The various solvents were purified as follows: Skellysolve: A commercial grade was intermittently shaken, with standing, with separate portions of concen­ trated sulphuric acid until a color was no longer imparted to the acid. It was then washed twice with a 10% sodium carbonate and 5% potassium per­ manganate solution. It was washed fifteen times (2) ✓ * v \ with distilled water, dried over sodium for 24 hours and distilled. The fraction boiling from 66° to 68° C. was collected. Benzene: A C.P. thiophene free grade jaf benzene was dried over sodium for 48 hours and distilled. The fraction distilling from 79.5° to 80.5° was collected. Chloroform: It was found that special precautions were necessary in the purification of chlorofornTand to keep it in the stable form. The small amount % of ethyl alcohol, 0.5# to 2%, ordinarily present in commercial and C.P. grades of chloroform, appears to stabilize it for periods of several months. According to Nield, et al, (1) and Ritsert (2) even a small amount of either ethyl alcohol or water in the chloroform, reduces the sensitivity of the antimony trichloride-chloroform-acetyl chloride reagent, consequently both must be removed. The alcohol was removed by washing the C.P. grade of chloroform seven times \ (3) with equal volumes of distilled water, followed by drying over potassium carbonate for 12 hours and fractionating. The fraction distilling at 61° was retained. This dry puri­ fied chloroform was unstable, however, and in from 4 to 7 days gave evidences of a breakdown, as shown by a positive starch-iodide test, and a white cloudiness with silver nitrate solution. When either of these tests was positive, the * chloroform was discarded and in no case used more than four days after being distilled. Shaking the purified chloroform with activated carbon immediately after fractionating, often resulted in chloroform that was perfectly stable for periods of two weeks or longer. However, if the treatment with activated carbon was omitted, breakdown occured in from 4 to 7 days. When a sample of chloroform gave a positive starchiodide test, shaking with activated carbon resul­ ted in a chloroform giving a negative starchiodide test. However, breakdown began again im­ mediately and, since some time was required to prepare the reagent, this chloroform was not used. Results obtained using chloroform giving eviden­ ces of decomposition are shown in Tables I and II and are discussed there. Ethanol: The ethyl alcohol used was a good grade of commercial absolute alcohol. Ether: \ A C.P. grade of ether was shaken for an hour each with two separate portions of a dilute ferrous sulfata solution. It was washed 8 to 10 times with distilled water, preliminarily dried over sodium sulfate and finally dried over night with an excess of soditim. It was then fraction­ ated and the fraction distilling at 34° was coll­ ected. This ether was stored over solid ferrous sulfate in small glass stoppered bottles with a minimum of air space over the ether. Stored in this manner, the ether was stable (no detectable starch-iodide test or precipitate with silver nitrate solution) for as long as two weeks. (5) were detectable (starch-iodide test) in from 24 to 48 hours and with a considerable air space over the ether, this time was lessened. This property is further discussed under the heading "The Effect of Oxidation Products of Ether". Antimony trichloride-chloroform-acetyl chloride reagent: The purity of each of the chemicals used in the preparation of this reagent is extremely critical. This is particularly true of chloroform and has already been discussed. The antimony tri­ chloride and acetyl chloride used were anhydrous reagent grades and care was taken to exclude water as both are decidedly hygroscopic. The reagent was prepared by dissolving 18 •g. of antimony trichloride in 100 ml. of chloro­ form at a temperature below 40°C., filtering the solution, and adding immediately 2 ml. of acetyl chloride. This reagent is good for at least four days but in these investigations was not used more than two days after preparation. The prin­ cipal factor governing the usable life of the 7 I (6) reagent is the stability of the chloroform used. \ STABILITY OF CALCIFEROL' IN VARIOUS SOLVENTS In order to determine whether calciferol was x > sufficiently stable in the various solvents used in this investigation, the following series of experiments were performed. A stock solution was prepared containing 0.005 g. of calciferol in 100 ml. of purified skellysolve. This solution was divided into ten more or less equal parts, (the equality of di­ vision is unimpqrtant here, since only the ques- ■ tion of how the potency ^changed with time was considered) the skellysolve evaporated off and the residues dissolved in the following solvents: Purified chloroform ■+*0.5$ absolute ethanol, purified chloroform, specially purified ether, commercial C.P. ether, anhydrous ether, commercial hexane, commercial skellysolve, absolute ethanol, purified skellysolve and anhydrous thiophene free benzene. » At each of four different intervals of time one ml. were taken from each of these \ solutions, evaporated to dryness and each of the residues taken up in one ml. of purified chloro­ form. Ten ml. of antimony trichloride-chloroform reagent were added to each sample and log I0/l read at 500 mu., after 3 minutes, on the visual photometer. The above procedure was followed in detail for each of the following reagent solutions: Chloroform saturated with antimony trichloride, d$"g. of antimony trichloride in 100 ml. of chloroform, 18 g. of antimony trichloride in 100 ml. of chloroform ml. of absolute ethanol and a saturated solution of antimony trichloride in chloroform that had started to decompose. Sampling and testing these calciferol sol­ utions was discontinued after six days. In the physical chemical method of vitamins D assay no solution of the vitamins was allowed to stand longer than four hours. The values recorded in Table I are log I0/l values as read on the visual photometer and are as suitable as U.S.P. unit potencies for eval- \ Table I. Reagent CHClj saturated with SbClg 18g. SbCl, In 3 100 ml. ®Clj5 18g. SbCl, In 100 ml. CHC1 + 0.4% ethanol Partially decomposed CBClj saturated with SbCl. hours after prep. (feciCEC1 -1-0.5% pure CgHgOH Stability Teats of Calciferol In Different Solvents Treated with Various Antimony Trichloride Reagents 4 Ether specially purified Ether C. P. Hexane Ether anhydrous Comm. Skelly- Skelly­ Ethanol Benzene solve solve absolute pure, purified anhydrous 0 .44 .46 .36 .46 .48 .42 .34 .42 .43 .40 24 .44 .53 .37 .48 .53 .46 .34 .44 .43 .38 72 .47 .48 .36 .48 .49 .46 .36 .43 .43 .39 144 .46 .49 .38 .46 .48 .43 .33 .42 .44 .38 0 .47 1.45 .40 .45 .43 .41 .32 .42 .42 .38 24 .45 .48 .38 .47 .49 .46 .36 •42 .40 .38 72 .45 .50 .37 .48 .50 .44 .34 .43 .43 .40 144 .47 .48 .38 .46 .48 .43 .34 .43 .41 .39 0 .45 .45 .35 .45 .52 .43 .33 .39 .45 .41 24 .46 .49 .41 .43 .54 .47 .35 .42 .44 .40 72 .47 .48 .39 .44 .49 .45 .35 .42 .43 .41 144 .44 .47 .37 .45 .50 .44 .35 .41 .45 .41 0 .be . r8 .12 .07 .18 .06 .10 .08 .08 .06 24 .11 .33 .09 .08 .07 .07 .11 .12 .09 .06 72 .13 .21 .09 .05 .12 .07 .17 .09 .11 .07 144 .09 .19 .ii .08 .10 .08 .11 .14 .12 .05 lo/l values In this table were read at 500 mu., after 3 minutes, on the visual photometer (8) uating stability* It may be pointed out that the log I0/l values for the same sample are reproducible to + 0.02. Conclusions drawn from the summarized data in Table I. Calciferol is stable for at least six days at room temperature in each of the solvents tested. No difference in the color reaction was observed by the use of a reagent prepared by saturating chloroform with antimony trichloride and one composed of 18 g. of antimony tirchloride in 100 ml. of chloroform. (Two ml. of acetyl chloride for each 100 ml. of chloroform were added to each reagent). The addition of absolute ethanol, up to 0.4#, to the reagent makes no difference in the color reaction for calciferol. This agrees with the. findings of Nield, et al, (1) who states that « the-^ppesence of 0.3# ethanol has no effect on the sensitivity of the reagent but that 0.7# (9) ethanol in the reagent reduces its sensitivity. Reagent prepared from partially decomposed chloroform has a large negative effect on the color reaction, causing errors as large as 75$. V STABILITY OF CHOLESTEROL IN VARIOUS SOLVENTS Sterols are present in all natural vitamins D oils, consequently stability studies were made using cholesterol as a representative sterol. These studies are analogous to those described under "Stability of Calciferol in Various Solvents each of the solvents and reagents being the same. The solutions contained approximately 0.2$ choles­ terol, since the extinction value is much smaller than that for calciferol. These measurements were made over a period of 20 days. Thexresults are grouped in Table II. Discussion of the results tabulated in Table II. Cholesterol is quite stable in each of the solvents tested, no change being observed after 20 days at room temperature. The addition of 0.4$ ethanol to the reagent Table II. Reagent raci­ esturated With SbClg 18g. SbClg in 100 JBl. CHCI3 18g. SbCl, in 100 »1T CHC1, 4-0.4% ethanol Partially decomposed 0HC1_ saturated with SbCl3 Hours after prep. CHC1, CHC1, + 0.5* pure CgffgOH Stability Taata of Choleatarol in Different Solvents Treated with Various Antimony Trichloride Reagents Sther specially purified Sther C. P. Sther anhydrous Skelly­ Hexane Skelly- Sthanol Benzene Cosei. solve absolute solve pure, purified enhydrous 0 .08 .09 .26 .11 .18 .13 .12 .12 .15 .10 144 .07 .11 .28 .12 .17 .12 .12 .12 .14 .11 480 .06 .10 .25 .12 .19 .12 .12 .13 .16 .11 0 .07 .11 .27 .11 .18 .12 .14 .13 .16 .11 144 .07 .09 .13 .16 .1? .12 .13 .14 .09 \ 480 .06 .09 .25 .13 .18 .11 .12 .12 .15 .11 0 .12 .16 .35 .19 .23 .18 .17 .18 .23 .19 144 .11 .15 .36 .20 .24 .20 .20 .19 .25 .18 480 .13 .16 .36 .21 .26 .19 .17 .18 .25 .16 0 .67 .73 1.05 .59 1.16 .82 .71 .75 .85 .66 144 .74 .76 1.26 .83 1.32 .93 .85 .91 .88 .78 480 .71 .82 1.13 .79 1.33 .97 .88 .83 .93 .73 The log l0n values in this table were read at 500 mu., after 3 minutes, on the visuel photomete: (10) makes no noticeable difference in the sensitivity of the reagent to cholesterol. Reagent consisting of chloroform saturated with antimony trichloride produces a deeper color than reagents containing smaller amounts of anti­ mony trichloride. Reagent saturated with anti­ mony trichloride results in an increase in color depth of as much as 40# more than reagent contain­ ing 18 g. of antimony trichloride in 100 ml. of chloroform. This effect might be expected since the sterol color deepens on standing, becoming £ almost black in from 24 to 48 hours. Reagent containing partially decomposed chloroform gives extremely high results, as much as a 500# increase, and obviously should not be used. Both C.P. and anhydrous ether gave higher i results than expected, although neither showed any change over the 20 day test period. A series of experiments were performed to check this effect and are described under the heading "The Effect of Oxidation Products of Ether”. STABILITY OF THE COLOR REACTION OF ANTIMONY TRICHLORIDE-CHLOROFORM REAGENT AND CALCIFEROL Tjie color reaction between the antimony trichloride-chloroform reagent and the vitamins D and other sterols has been investigated in several laboratories, Nield, et al, (1), Brockmann and Chen (5), and others (2), (3), (7)”lind (8), and is quite well standardized. These investigators, however, either used the vitamin in a pure state or mixed with substances giving no color with this reagent. In this investigation sterols are assumed to be present in all vitamins D samples and their effect must be evaluated and the proper correc­ tions made to obtain the actual concentration of calciferol present. The rates of change of the extinctions pro­ duced by adding antimony trichloride-chloroform reagent to chloroform solutions of calciferol and of chlosterol was determined on a photoelectric photometer. The concentrations of thie cholesterol solutions were varied. The extinction readings for these concentrations are plotted against time 8 8 8 / / (12) on Figure I. The rates of change of the same color reactions were also measured on the visual photometer. Figure II furnishes a comparison of the visual and of the photoelectric photometer, since it contains plots of the rates of change of the same color reactions measured on both photometers. Conclusions The red color for calciferol develops quickly to a maximum and is stable for 5 minutes, after which it fades slowly. The cholesterol color deepens with time, changing from an extremely pale yellow to a deep red. This change is quite rapid during the first ten minutes, after which is slows slightly but still shows a definite trend. Therefore, if the color reaction is to be applied to a solution containing both vitamins D and sterols or sterols alone, the time when read­ ings are taken must be accurately controlled. The extinction readings at 3 minutes of the 0.2$, 0*4$ and the 0.8# cholesterol solutions are, (13) respectively, 0.06, 0.13 and 0.25 showing, within experimental error, the linearity of the response of the photoelectric cell in this region. If the response was not linear over the range used, a calibration curve would have to be run for each colored substance measured and corrected readings taken from this curve. \ The visual and the photoelectric photometers are not strictly comparable, since differences of 9% and Z% were observed for the cholesterol and the calciferol curves, respectively. Moreover, the deviation for cholesterol was in the opposite direction from that of calciferol. Effect of Diluting a Solution of Calciferol in Corn Oil With Corn Oil A solution of calciferol in corn oil with a biological assay of 200,000 U. S. P. units per gram was chosen for this investigation. One gram aliquots of this solution were diluted by weight to 2, 4, 8 and 16 grams total weight with pure corn oil, thus reducing the calciferol content to Table III Physical Chemical Assay of Samples of Calciferol in Corn Oil Diluted with Corn Oil Cone, g./lOO ml E(l$. 1cm.) 500 mu • Physical Chemical Method U.S.P. u./g. Biological Method U.S.P. u./g. 0,0909 6.61 127,400 200,000 0.04545 3.25 62,650 100,000 0.02273 1.62 31,330 50,000 0.01136 0.81 15,660 25,000 0.00568. 0.385 7,430 12,500 Table IV Assay of Samples of Calciferol in Corn Oil Diluted with Corn Oil E(l$, 1cm.) Cone. 500 mu. g./lOO ml . Physical Chemical Biolorical Method Method U.S.P. u./g. U.S.P. u./g. 0.0909 9.63 186,000 200,000 0.04545 4.84 93,450 100,000 0.02273 2.50 48,300 50,000 0.01136 1.27 24,420 25,000 0.00568 0.715 13,800 12,500 (14) 0.5, 0.025, 0.0125 and 0.00625 its original value. This series was run in duplicate by the method proposed by Kingsley (4) and the results are tabulated in Table III. Since no vitamin A is present in corn oil and since corn Oil has an E (1%, 1cm.) of less than 0.2, this series may also be run by dissol­ ving the oil solution directly in chloroform. Aliquots of these chloroform solutions are then added to antimony trichloride-chloroform reagent and the extinction values measured at 500 mu, after 3 minutes. The results of such a series is shown in Table IV. Discussion of Tables III and IV In Table III, although there is a marked difference in the calculrted potency and the bio­ logical potency of these solutions, the vitamin D concentretion-potency ratio is the same in every case, within experimental error. In Table IV the calculated potencies and the biological^potencies are practically identical and the vitamin D concentration-potency ratios are (15) the sane. Since unsaponified corn oil has an E (1%, 1cm.) of approximately 0.1-5, each dilution should give a value slightly greater than half the value of the preceding dilution. This is true in every case. The data included in both Tables III and IV are in accord with Beers law, that is, the ex­ tinctions produced by the antimony trichloridechloroform reagent are linear functions of the amount of calciferol present. The Effect of Oxidation Products of Ether To test the effect of the products of ox­ idations of ether on the color reaction for calcif erol, the following experiments were performed. Two samples of calciferol in corn oil #3772 were saponified. One was extracted with stock C. P. ether and the other with specially purified ether (ref. Apparatus and ’•aterials). Both were evaporated to dryness, taken up in chloroform and aliquots reacted with antimony trichloride-chloroform reagent. Log I0/T was read at 500 mu., after 3 minutes, on the visual photometer. (16) Extracted with C.P. stock ether " " purified 150,000 U.S.P. u. " 195,300 U.S.P. u. Two samples of oil #3372 were dissolved, one in purified ether that had been standing for 4 months in a glass stoppered bottle and the other in purified ether prepared in the previous 4 hours. The ether was evaporated off, the samples taken up in chloroform, and the color reaction run. Log IQ/I was read at 500 mu., after 3 minutes, on the visual photometer. Ether purified in last 4 hours— 195,300 U.S.P. u./g. Ether after standing 4 months 131,600 U.S.P. u./g. t Further, 30 ml. of purified ether from an­ other bottle, that had been standing for 3 months, was placed in a clean dry flask and the ether evaporated off. A sample of oil #3772 was weighed into this flask and dissolved in chloroform. The color reaction was run with an aliquot of this solution and log IQ/l read at 500 mu., after 3 minutes, on the visual photometer. An assay of 142,200 U.S.P. units/g. resulted. These data clearly show that the oxidation (17) products of ether have a definite large negative effect on the color produced by the reaction of calciferol and antimony trichloride-chloroform reagent, the apparent potency being reduced by as much as 35%. Moreover, the effect does not seem to be limited to solutions of the ether but also to a residue left by the evaporation of the ether. Since this effect is not evident for natural oils, (the above mentioned tests are not applic­ able because of the presence of vitamin A) it was thought that chromatographing might filter out the interfering substance or substances. Accord­ ingly, the following experiment was performed. Two samples of oil #3772 were saponified, one being extracted with purified ether, and the other with stock C.P. ether that had, in the previous hour been run through a 4 cm. super— e filtrol column. The ether was evaporated off, the residues dissolved in chloroform and a color reaction run vdth antimony trichloride-chloroform reagent on an aliquot of each solution. Purified ether sample— — -— -195,300 U.S.P. u./g. (18) Chromatographed ether sample— 191,100 U.S.P. u./g. This indicates that running the ether through a superfiltrol column eliminates, for the most part, the interfering materials. The latter experiment was twice repeated, using ether that had been put through superfiltrol columns 8 and 12 cm. in length. The results ?/ere the same, that is, a slight lowering from the potency obtained using purified ether. The chromatographed ether, however, was very susceptible to further decomposition and within a few hours gave a definite starch-iodide test. Also, when ether that had been run through a super­ filtrol column and then let stand for several hours was used to extract a sample of oil #3,772, a marked lowering in its potency was found. This indicates that further oxidation began immediately. These results show that chromatographing removes most of the oxidation products already present in the ether but does not inhibit the formation of additional oxidation products. As some ergosterol is present in all irrad- (19) iated ergosterols in oils and since sterols are present in all natural vitamins D oils, the effect of the oxidation products of ether on a common sterol was investigated. The cholesterol solutions listed in Table II were used. " \ A 5 ml. aliquot of the purified skellysolve solution, of Table II, was evaporated to dry\ * ness, the residue taken up in stock C.P. ether, #6 the ether evaporated off, and the residue diss­ olved in 5 ml. of chloroform. this solution was added to 10 A 1 ml. aliquot of ml. of antimony trichloride-chloroform reagent and the extinction read, after 3 minutes, at 500 mu. The value was 0.24, as corpared with an extinction of 0.12 before the above treatment. Another aliquot of purified skellysolve solution #6 was treated similarly, except it was redissolved in purified ether that had been standing for 4 months in a glass stoppered bottle. The extinction was 0.71, as compared to an orig­ inal value of 0 .1 2 . Another aliquot of the same solution was (20) treated similarly, except it was redissolved in purified ether prepared during the preoeeding hour. The extinction was 0.13, compared to an original value of 0 .1 2 . The same experiment was repeated in detail except that the residue was redissolved in stock C.P. ether that had been passed through a 4 om. superfiltrol column during the preceeding hour. The extinction was 0.14. These results clearly indicate that partially oxidized ether has the effect of greatly enhancing the extinction of the color produced by adding antimony trichloride-chloroform reagent to a cholesterol in chloroform solution. The effects of partially oxidized ether on the color reactions of calciferol and of sterols with antimony trichloride-chloroform reagent are in opposite directions and, in a solution in which both were present, would tend to cancel. % However, the concentrations v/ould seldom be such that these errors would exactly counterbalance. The ether^used for extraction and for elution (21) of the chromatograph columns should he either purified by the method outlined under "Apparatus and Materials" or drawn through a 4 to 8 cm. superfiltrol column an hour or two before using. The latter is much the simpler procedure if any shortened method is to be used in the analysis of irradiated ergosterols. An example of such a ^shortened method consists of the following steps: 1Saponification, extraction with ether, evaporation of the 'ether, dissolving the residue in chloroform, and treating an aliquot of this solution with antimony trichloride-chloroform reagent. This method obviously could not be used if any vitamin A was present. If the oil is to be run through the entire analysis procedure, (4) the first chromatograph of the method would act to partially purify the ether. Comparison of a Visual Photometer With a Photoelectric Photometer A comparison was made of a photoelectric photometer and the visual, photometer used in this work by measuring the extinction of the antimony x Table V Comparative Extinctions on a Visual and on a -* Photoelectric Photometer Type of Oil or Sample Reacting Material Extinction Visual Photoelectri Photometer Photometer 0.66 0.70 Calciferol D(a) Calciferol in olive oil D 0.39 0.38 Cholesterol S(b) 0.16 0.165 Irradiated ergosterol D+S(c) 0.31 0.28 Calciferol D 0.75 0.70 Calciferol Irradiated ergosterol D S 0.78 0.02 0.71 Tuna liver oil cone, S 0.13 0.13 Tuna liver oil conc. S 0.14 0.13 Mixed high D oil D+S 0.76 0.74 Mixed high D oil D+S 0.78 0.75 Mixed hifh D oil S 0.32 0.32 High P oil S 0.36 0.37 Tuna liver oil conc. D+S 0.76 0.74 Tuna liver oil conc. D+5 0.62 0.57 Mixed fish liver oil D+S 1.12 1.11 Mixed high D oil D+S 0.75 0.76 D distillate D+S 0.69 0.695 D distillate D+ 6 0.95 0.97 D distillate (a) D, vitamins D D+S (b) S, sterols 0.02 0.92 0.92 (c) D S, vitamins D +- sterols (22) trichloride-chloroform reagent-vitamins D color reaction. mu. All the measurements were made at 500 The samples were chosen so that the extinction values covered a wide range and offered a rep­ resentative comparison at the particular wave length used. Measurements were made on the photoelectric photometer using 1 cm. glass stop­ pered cells with both Corex and quartz end plates. Glass covered open type cells were first tried and found very difficult to use with this reagent due to the extremely rapid hydrolysis of antimony trichloride and the fact that the solution tends to creep over the edges of the cell. To check for possible personal differences in the readings on the visual photometer, all except the first six values of Table V were checked independently by another investigator who had had long experience with this instrument. These values never differed by more than 0,02 and in most cases were the samev Extinction values for both instruments are listed in Table V. Discussion of Table V. (23) <, In the lower range, the extinctions, as measured on each instrument, are the same. 'When the intensity of the color is increased, however, the readings are noticeably different. With a few exceptions, the extinctions read on the visual instrument were higher than those read on the photoelectric one. This is probably due to the fact that photoelectric cells in general do not respond linearily to increasing color intensity over a wide range and require a calibration curve for a particular color reaction. If such a cal­ ibration curve is made, the photoelectric photom­ eter may be used interchangeably with the visual one. The visual instrument was used in these studies because the accuracy is well within the accuracy of the method and the cells used are much easier to handle. Sensitivity of the Antimony TrichlorideChloroform Reagent-Vitamins D Color Reaction The question of the applicability of the chromatographic adsorption method for oils of (24) low vitamins D potency led to a determination of the lowest potency accurately measurable by means of the antimony trichloride-chloroform reagent. The solutions for these measurements were prepared by successively diluting a solution of calciferol in purified skellysolve. One ml. aliquots of these solutions were evaporated to dryness, taken up in one ml. of chloroform, ten ml. of antimony trichloride-chloroform reagent added, and the extinctions determined on the visual photometer at 500 mu., after 3 minutes. These values are recorded in Table VI. Table VI Conc. (g.AOO ml.) log I q / I , 3 min., 500 mu. U.S.P. Units/ml. 0.00502 0.81 2,010 0.00251 0.39 1,005 0.00126 0.20 502 0.00063 v 0.10 251 0.00031 0.05 125 0.00016 0.02 63 0.00008 0.00 32 (25) By thele VII consists of values obtained by taking the diff­ erence between the bioassay and the corresponding physical chemical assay and dividing this diff­ erence by the average percentage deviation. This furnishes a method of comparing the various irradiated ergosterols. Thus, it is noted that six of the values fall within Q% of the average percentage deviation, while six of the remaining seven fall within 21$. Therefore, if the con­ version factor was determined from one of these irradiated ergosterols, rather than from the reference oil #47761 (Table X), fair results would be obtained for the other irradiated ergos­ terols. With the results of Table VII in mind, it was decided to further investigate certain steps of the chromatographic adsorption method as app­ lied to solutions of irradiated ergosterol in oil. Solutions of irradiated ergosterol in corn i i (28) oil were used for these investigations. Since such an oil contains no vitamin A and since corn oil has an E(l/6 , 1cm.) of less than 0.2, it may he dissolved directly in chloroform and an ali­ quot of this solution run with antimony trichloride -chloroform reagent. Thus an apparent loss or gain may be determined before and after each step of the procedure. The results of such a series of investigations appear in Table IX. Reference of Table IX will show that a small loss, of from 4 to 7%, occurs during the time a sample of an irradiated ergosterol in corn oil is saponified, washed, and extracted with ether. If purified ether is used to extract the samples of oil #3772, ''calciferol in corn oil, no loss occurs. However, the loss is still evident for the ir­ radiated ergosterols in corn oil. Approximately the same percentage loss, from 3 to 7%, occurs if the samples are treated as follows: Dissolved directly in the developing solution (50-10-1, skellysolve, purified ether, and absolute ethanol, respectively), run through (29) a 6 cm. superfiltrol column, the column cut im­ mediately below the orange band and eluted with ether, the mixed solvents evaporated off, the residue dissolved in purified chloroform and an aliquot of this solution run with antimony tri­ chloride-chloroform reagent. A combination of saponification and chro­ matographing results in a loss of from 8 to 1 2 $. This is true for the samples of calqiferol in corn oil as well as for those of irradiated ergosterol in corn oil.___________________________ Oil #3772 is a 160 times dilution of pure calciferol in corn oil. This calciferol gave a physical chemical assay value of 32,000,000 U.S.P. units/g., and on that basis oil #3722_should have a potency of 200,000 U.S.p-. units/g. This value is almost reached by running an aliquot of a solution of this oil, dissolved directly in chloro­ form, with antimony trichloride-chloroform reagent. This indicates that the corn oil itself has no inhibiting effect on the color re^tion. Corn oil has an E(l$,lcm.) of 0.17, and after (30) s saponification this value drops to 0,06, thus acting only to increase the E(l$, 1cm.) very slightly. Milas (9) and Clover (10) have shown that ethyl ether, and alkyl ethers in general, undergo spontaneous oxidation to form multiple oxidation products, the identity of which are in doubt. Hydrogen peroxide, however, was found in every partially oxidized ether solution. This oxidation is greatly accelerated by ultraviolet light, Milas finding very large amounts of oxidation products titratable with sodium thiosulphate, after an irradiation period of 80 hours. After a standing period o? some weeks, every ether solution tested gave a strong starch-iodide test. All of the irradiated ergosterol samples used in this investigation were irradiated in an ether solution and no attempt was made to purify the ether either before or after irradiation. In view of this fact and in view of the results found under "Effect of Oxidation Products of Ether", it was thought likely that the peroxides and/or (31) other oxidation products of ether were causing the physical chemical assay values to be lower than the corresponding bioassays. As has been shown, passing ether itself through a 3 cm. superfiltrol column eliminates the interfering oxidation products. However, chromatographing an ether solution of an irrad­ iated ergosterol has no effect other than a slight lowering of its vitamin D potency. Hence it was thought that a stronger treatment to elimW> «w> inate the oxidation products might be necessary. i An ether solution of irradiated ergosterol in corn oil #3772 was successively shaken with 0 .1 N. ferrous sulphate, 0 .1 N. and 0.1N. ammonia solutions. sodium thiosulphate, It was then washed with distilled water, evaporated to dryness, and taken up in chloroform. An aliquot of this chloroform solution was added to antimony tri­ chloride-chloroform reagent and, after 3 minutes, log I0/l read at 500 mu. This treatment had no measurable effect on the log I0/l value of the color reaction. (32) This procedure was repeated in detail for all the samples of irradiated ergosterol in corn oil listed in Table IX. TTo change in the vitamin D potency of any of these samples was observed, except a slight lowering in two cases, probably due to mechanical loss. Also, an ether solution of a sample of each of the oils listed in Table IX was shaken with multiple portions of Fehlings solution, washed -, 1cm.) 500 mu. Calculated U.S.P. u./g. (a) Biological method U.S.P. u./g. (b) Irradiated Ergosterol in Corn Oil 6.93 6.82 7.91 133,000 131,600 152,800 5.83 5.38 123,200 112,600 (av.) 250,000 1.00 200,000 0.92 300,000 1.11 250,000 1.16 200,000 0.81 200,000 1.00 138,000 117,900 78272 65751 0.1 0.1 8.21 142,300 148.600 148.600 158,400 149,500 6.06 6.38 116,800 123,000 1 2 0 ,QQO 7.37 7.70 7.70 (B) 3772 0.1 6.61 G *tjl 127.400 127.400 5.94 5.39 114,700 104,000 r> *1 12722 0.1 (C) 56391 0.1 C*ilciferol in Horn Oil 127,400 109,350 Irradiated Ergosterol in Halibut Liver Oil 7.59 7.04 146,500 136,000 225,000 141,250 0.93 Table VII Oil 66701 89052 89182 86632 Wt. (g.) 0.1 0.1 0.1 0.1 E(1%f 1cm.) 500 mu. (continued) Calculated U.S.P. u./g. (a) Biological method U.S.P. u./g. (b) 7.37 7.37 6.82 7.59 7.65 142,300 - 142,300 131,600 146,500 147,700 7.59 7.65 143,500 147,700 7.37 7.26 142,300 140,100 141,200 7.26 6.82 140,100 131,700 135,900 250,000 0.96 250,000 0.92 250,000 0.97 250,000 1.02 3,800 0.79 12,000 1.46 142,000 147,100 (D) Viosterol 2242 62191 1.0 1.0 0.132 2,340 2,550 2,445 0.253 0.176 4,880 3,400 4,140 0.121 0 \ Table VIII Comparative Vitamin D Assays of Fish Liver Oils by Physical Chemical and Biological Methods Sample 44080 76902 P6846 12,300 12,300 18,500 18,050 V3428 1,133,000 1,119,000 77462 61211 Biological Method U.S.P. u./g Physical Chemical Method U.S.P. u,./g. U.S.P. u./g. Kingsley (4) Author 29,300 (av.) 29,720 (av.) • 29,720 29,720 30,100 29,700 8,900 8,900 13,160 11,470 11,470 31,000 12,300 12,400 12,90Q_ 12,650 12,000 18,275 19,700 20,300 20,000 1,126,000 1,262,000 20,000 1 ,2 2 0 , 0 0 0 1 ,2 0 0 , 0 0 0 8,300 . 8,900 12,030 8,110 8,200 6,300 7,720 6,950 7,335 12,000 Table IX Effect of Saponification and Chromatographing on Irradiated Ergosterol in Corn Oil Sample 45120 61691 Chromatographed Saponified Saponified Dissolved and Chromato­ only ' only directly in graphed U.S.P. u./g. Chloroform U.S.P. u./g. U.S.P. u./g. 172,000 180,500 174,100 161,400 160,000 157,100 152,600 172,000 , 193,000 78272 214,500 200,000 200,000 12722 191,100 183,200 186,900 166,400 3772 195,300 184,800 193,200 160,000