DOCTORAL DISSERTATION SERIES

title A Pflysrco-Chemicd Method
Deiermi'h&tson Of The

___________

V;/amVn

Fat The

_

D Potency

Q f Esh Pits_________ _

AUTHORGerJcl Vavg
&h lkihqrsley
yr/l
ifth
Uhjstey___ date
UNIVERSITY.
DEGREE_J_LLAAi_________ PUBLICATION NO.
\u

UNIVERSUy MICROFILMS

S"

A N N ARBOR

•

MICHIGAN

A PHYSICO-CHEMICAL METHOD FOR THE DETERMINATION
OF THE VITAMIN D POTENCY OF FISH OILS

fey
Gerald Vaughan Kingsley

A THESIS

Presented to the Graduate School of Michigan State
College of Agriculture and Apolied Science
in Partial Fulfillment of Requirements
for the Degree of Doctor of Philosophy

Department of Chemistry
Michigan State College
1942

A

nvjiviwv*

T7'T> r* T7,>

x

The writer wishes to express his appreciation to
Dr. D.T. Ewing for his aid ana guidance throughout this
investigation; and to Parke-Davis and Company for the
assistance which made this work possible.

A Physico-Chemical Method for the Determination
of the Vitamin D Potenoy of Pish Oils
Several methods for the physical or chemical determina­
tion of vitamins D have been reported in recent years.

Some

of these give acceptable results on nearly pure samples of
vitamins Da or D3.

In the case of fish oils, satisfactory

agreement with the bioassay values has been obtained only
over limited concentration ranges or for a small number of
oils.
Reerink and van Wijk (16) and later Topelmann and
Schuhknect (22) were successful in determining the concentra­
tion of vitamin Da in fairly pure solutions by measurement
of the ultraviolet absorption maximum at 265 mu.

01 khin (14)

used this method in determining vitamin Da in irradiation
products.
Marcussen (11) applied the measurement of the 265 mu maxi­
mum to the determination of vitamins Da and Ds in fish oils,
knowing from the work of Brockmann and Busse (2) that the
molecular extinctions of the two vitamin forms are the same.
It was necessary to first saponify the oil, then remove vita­
min A, carotenoids, inactive sterols, and other substances
showing absorption in the neighborhood of 265 mu.

This sep­

aration was effected by passing a heptan solution of the nonsaponifiable fraction through a Tswett column filled with
Hydraffin K4 as adsorbent, Isolating the vitamins D in the
chromatographed solution.

The concentration of vitamins D
i%
in this solution was obtained spectr©graphically. The Eicm#
of calciferol (British Drug House) was used as a standard

-2

for conversion of the Eij?^ of the fish oil to biological
units per gram.

Attempts to reproduce results on a given

oil led to fluctuations of about 7?fe.

Results are given on

two tuna oils, one halibut oil, and one oil solution of an
irradiated product.
The search for a color reaction which is specific for
the vitamins D has been unsuccessful.

Halden and Tzoni

(8,9,24) obtained satisfactory quantitative results on solu­
tions of pure calciferol by heating a solution of the vitamin
and pyrogallol in benzene, petroleum ether, chloroform or
absolute alcohol with a fresh solution of anhydrous aluminum
chloride in absolute alcohol.

The deep-violet colored pro­

duct was dissolved in absolute alcohol to give a lilac-red
solution for colorimetric measurements.

For determinations

on fish oils, it was found necessary to remove vitamin A.
Cholesterol, ergosterol, and lumisterol did not interfere,
but some of the irradiation products of ergosterol gave this
color reaction.
Shear (20) suggested a colorimetric method based on the
red color given by vitamins D and liver oils with a mixture
of aniline and hydrochloric acid, but Levine (10) reported
that this color reaction was not specific for vitamins D.
Stoeltzner (21) proposed a colorimetric method based on
the formation of color when phosphorous pentachloride was added
to an oil solution of vitamins D.

That this reaction is typi­

cal of other compounds found in natural oils has been shown
by Christensen (4).
Rutkovsky (19) has shown that the green color of the

-3-

Tortelli-Jaffee reaction (23), formed when an acetic acid
solution of vitamins D is mixed with a 2 % solution of bro­
mine in chloroform, is also characteristic of the provitamins*
Robinson (18) reported a colorimetric method based on the
yellow color formed when an alcohol solution of the vitamins
is boiled with sodium nitrite and acetic acid, then made
slightly alkaline*

This color reaction was not given by

cholesterol, ergosterol, dehydroergosterol, or lumisterol,
but vitamin A gave a strong orange color and the nonsaponifiable fractions of olive oil and arachis oil gave some color.
The antimony trichloride color reaction used by Brockmann
and Chen (3) has been chosen by a number of investigators as
the basis for quantitative methods for vitamins D*

Brockmann

and Chen used a cold saturated solution of antimony trichlor­
ide in dry, alcohol free chloroform.

Vitamins Da and

gave an orange yellow colored solution with an absorption
maxima at 500 mu.

Tachysterol gave a similar reaction, and

cholesterol, sitosterol, ergosterol, 7-dehydrocholesterol,
lumisterol, Suprasterol I and Suprasterol II gave much weaker
colors and hence did not interfere unless present in concen­
trations over thirty times that of the vitamins D.

Vitamin A,

with an absorption maximum at 620 mu, hindered the analysis
when present only in six times the concentration of the vita­
mins D*

This reagent offered a dependable colorimetric method

for vitamins D in the pure state or when vitamin A and sterols
were present in low concentrations.
Emmerie and van Eekelen (5) found that vitamin A must be
removed from fish oils before using the antimony trichloride
color reaction.

-4-

Wolff (25) used the Brockmann and Chen color reaction
for the determination of vitamin Ds in nine samples of ir­
radiated ergosterol, using pure calciferol as reference*

His

results show deviations of from 8% to 40% from the bioassay
values.

Wolff also investigated the determination of vita­

mins D in fish oils, removing the vitamin A and carotenoids
by chromatographic adsorption on Montana earth from benzene
solution.

He found it necessary to remove most of the sterols

by the use of digitonin when the potency of the oil was low.
Ritsert (17) removed vitamin A by chromatographic ad­
sorption on aluminum oxide, but reported difficulty in the
analysis of fish oils or of the mixed products obtained by
irradiation of the provitamins.
Raoul and Meunier (15) modified the reagent of Brockmann
and Chen by adding a small amount of acetic anhydride and
sulfuric acid to the chloroform solution of antimony tri­
chloride.

They attempted to make use of the fading vitamin

A color and the slow formation of cholesterol color to com­
pute the vitamin D concentration by rate of change in the
absorption maxima, and suggested the use of digitonin to
remove most of the sterols.

^

Nield, Russell, and Zimmerli (13) modified the reagent
of Brockmann and Chen by the addition of acetyl chloride and
a change in the antimony trichloride concentration.

They

reported a reagent of increased sensitivity and stability.
Milas, Heggie, and Raynolds (12) determined the vitamin
A potency of fish oils by measurement of the 620 mu maximum.

-5-

Cholesterol, with an absorption maximum increasing slowly with
time, was estimated by measurement of the 480 mu maximum 30
minutes after adding the antimony trichloride reagent to the
fish oil sample#

Corrections for vitamin A and cholesterol,

based on the above measurements, were then applied to the
measured value of the vitamins D absorption maximum at 500
mu.

Better agreement between calculated potencies and bio­

assay values was obtained by a second method, in which vita­
min A, carotenoids, and possibly 7-dehydrocholesterol were
destroyed by treatment with maleic anhydride.

The vitamins

D concentration was then obtained by measurement of the 500
mu maximum.
Gudlet (7) reported the use of maleic anhydride for re­
moving vitamin A and removal of sterols by a freezing out pro­
cess.

This method of removing sterol3 was unsatisfactory and

the use of digitonin was suggested.
Ewing and Tomkins (6) removed vitamin A from the nonsaponifiable fraction of fish oils by chromatographic adsorp­
tion on Superfiltrol from hexane-ether-alcohol solution, after
removal of sterols with digitonin.

Their data for 16 fish oils

shows a mean deviation of 15$ from the bioassay potencies.

The

modified reagent of Nleld, Russell, and Zimmerli (13) was
used for the color reaction.

Special attention was given to

the purification of the chloroform used, including a final
treatment with activated carbon.
These previous investigations all indicate that certain
interfering substances must be removed from the fish oil be­
fore a quantitative determination of the concentration of the

-6-

vitamlns D can be made.

The Nield, Russell and Zimmerli

modification of the antimony trichloride reagent was chosen
as most suitable for the present investigation.

This made

necessary the removal of vitamin A, carotenoids, pigments,
and cholesterol.

The method chosen for the removal of

carotenoids, vitamin A, and pigments was a modification of
the chromatographic adsorption method as used by Ewing and
Tomkins.

In the earlier work, freezing and digitonin were

relied on for removal of sterols, but due to a small error
due to loss of vitamins D during these processes and to a
supply shortage of digitonin, a method was developed for the
separation of cholesterol from the vitamins D by adsorption.
Apparatus and materials.--The adsorption tubes for the
chromatographic determinations were made by sealing a 6 cm.
length of 7 mm. Pyrex tubing to the bottom of 5/8 by 6 inch
Pyrex test tube.

These tubes were cleaned before filling by

soaking in sulfuric acid-dichromic acid solution, rinsing
with distilled water and with alcohol, and finally drying in
the flame of a Meker burner.

The suction apparatus was de­

signed so that a bank of 8 columns could be developed simul­
taneously, controlling the pressure with the aid of an open
tube mercury mannometer attached to the suction flask.

A

finely divided grade of Superfiltrol was used as the adsorbent
throughout this investigation.

The adsorption columns for the

first chromatographic separation were prepared for use by
placing a small wad of cotton in the bottom of one of the ad­
sorption tubes, pressing this down firmly and adding a measured
volume of the adsorbent of such an amount that when very firmly

-7-

pressed down with a cork on the end of a glass rod, under 6
cm, of suction, the height of the packed column was 3 cm.

A

second and equal portion of adsorbent was then added, and
pressed down as before.
hard, level surface.

The top of the column should have a

Any air pockets, particularly between

the adsorbent and the glass walls, caused Irregularly shaped
adsorption bands.

The cork used in packing the columns was

slightly smaller than the inside diameter of the Pyrex tube,
to avoid loosening of the adsorbent by suction when this
piston was raised.

Columns for the chromatographic separa­

tion of vitamins D and cholesterol were prepared in the same
manner, except that the height of the column was only 1.5 cm.
and filling was done under a suction of 4 cm.
Measurements of the absorption maxima at 500 mu were
made on a Bausch and Lomb visual spectrophotometer equipped
with a Martins polarizing unit.

Bausch and Lomb absorption

cells were used, with quartz end plates, 1 or 2 cm. glass
spacers, and Monel metal fittings.
Skellysolve was purified by shaking with concentrated
sulfuric acid, washing twice with 10% sodium carbonate solu­
tion, shaking with 10% sodium carbonate 5% potassium perman­
ganate solution, washing 15 times with distilled water, de­
canting into a dry flask and drying over sodium.

The dried

solvent was then distilled from sodium, discarding the first
5% and the last 10% of the distillate.

The distillate col­

lected boiled over a 2 degree range.
Purified anhydrous ether for the chromatograph was pre­
pared by washing C.P. ethyl ether with ferrous sulfate solu­

-8-

tion to remove peroxides, washing 10 times with distilled
water to remove alcohol, drying with phosphorous pentoxide,
filtering and storing over sodium.

This ether was distilled

from sodium as needed.
C.P. ethyl ether was used without further purification
for extractions after saponification of oils and for elution
of adsorption columns.
C.P. thiophene free benzene was dried over sodium, dis­
tilled, and shaken with Superfiltrol before use.
The absolute ethyl alcohol was a high grade commercial
product•
Alcoholic potassium hydroxide was prepared by dissolving
14 g. C.P. potassium hydroxide pellets in 95$ ethyl alcohol
and diluting to 500 ml.

This stock solution was protected

from carbon dioxide and was filtered through hardened filter
paper as needed.
C.P. chloroform was washed thoroughly with 7 equal por­
tions of water, dried over anhydrous potassium carbonate, de­
canted, and fractionated, discarding the first and last 10$
of the distillate.

This purified chloroform is relatively

unstable and was prepared in small quantities and protected
from light.

Before use it was tested with silver nitrate

solution, and with potassium iodide and starch solution, and
if chlorides and oxidizing agents were absent, it was shaken
with activated carbon and filtered (6).
The antimony trichloride reagent was prepared by dissolv­
ing 18 g. of C.P. antimony trichloride in purified chloroform,
diluting to 100 ml., filtering and adding 2 ml. redistilled

-9-

acetyl chloride (15)*

Best results were obtained when not

more than one weeks supply of purified chloroform or anti­
mony trichloride reagent were prepared at one time.
Samples of fish oils and vitamins D concentrates were
supplied largely by Parke-Davis and Company.

Most of the

bioassays were performed at the Parke-Davis laboratories.
Experimental.--A fish oil sample is weighed out in a
125 ml. erlenmeyer flask.

Samples containing 4,000 to

100,000 U.S.P. units of vitamins D are convenient, 0.5 g.
to 2 g. being the usual weight used for natural oils, and
0.1 g. samples for concentrates.

If the sample weighs 1 g.

or less, 10 ml. of alcoholic potassium hydroxide is added,
but for samples weighing more than one gram, 10 ml. of alco­
holic potassium hydroxide per gram of sample is used.

A

short stemmed funnel is placed in the neck of the flask to
serve as a condenser.

The sample is placed in a water bath

and kept at 70 to 75 degrees for 1 hour, or longer if saponi­
fication is not complete.

Agitation of the sample at frequent

Intervals aids this reaction.
The sample Is cooled to room temperature, 20 ml. water
per 10 ml. alcoholic potassium hydroxide added, and extracted
with four 25 ml. portions of C.P. ethyl ether, using a separa­
tory funnel.

Gentle shaking in the separatory funnel may be

used without danger of emulsification unless more than the
indicated quantity of water has been added.

If an emulsion

does form, it can be readily broken by the addition of a few
drops of alcohol.

The ether layer must always be clear be­

fore separation of the two layers.

»10*>

The ether extracts are combined in the separatory funnel
and 50 ml* of water added*

When both ether and water layers

are clear, the water layer Is withdrawn and discarded, leaving
the ether layer in the funnel.

This preliminary washing to

remove most of the soaps is repeated twice with fresh 50 ml*
portions of water.

Agitation during these preliminary wash­

ings may result in the formation of a quite stable emulsion*
If this occurs, two ml. of alcohol are added to break the
emulsion.

Twenty-five ml. of water are now added to the

ether extract in the separatory funnel and shaken vigorously,
after which 25 ml. more water is added and the water and ether
layers allowed to separate.
fore it is withdrawn.

The water layer must be clear be­

This is repeated twice.

At the end of

the third washing with agitation, the ether and the water lay­
ers usually separate quickly to a sharp interface and are very
clear.

The water layer at this point should be colorless and

not alkaline to phenolphthalein.

Washing must be continued

until these conditions are satisfied.
The washed ether extract is withdrawn into a 125 ml.
erlenmeyer flask through a filter paper containing anhydrous
sodium sulfate.

The separatory funnel is rinsed with 25 ml.

C.P. ether and the rinsings used to recover vitamins D from
the sodium sulfate and filter paper, which should be colorless.
The ether solution Is then evaporated to dryness under reduced
pressure, using a water bath at 50°C.
The dry residue from this evaporation is taken up In 5
ml. of a mixed solvent prepared from purified materials by
mixing 50 parts Skellysolve, 10 parts anhydrous ether, and

-11-

1 part absolute alcohol by volume.

One drop of Sudan III

solution (25 mg. per liter In the above mixed solvent) Is
then added to the sample.
A 6 cm. adsorption column Is wet with 10 ml. of the
mixed solvent, the sample added, followed by 5 ml. used to
rinse the flask and 35 ml. of the same mixed solvent used to
develop the adsorption bands.

Each addition of solvent is

made just before the top of the adsorption column becomes
dry.

If the adsorbent becomes dry, It shrinks away from

the glass walls and development of the column is not regular.
A pressure differential of 6 cm. of mercury is maintained
until after the addition of the 35 ml. of developing solution,
when the suction is increased to 10 cm.

Further increase of

the suction results in crevices in the adsorption column
and packing of the adsorbent to such a degree that the flow
of developing solution is nearly stopped.
When the last of the developing solution has passed
through the column, drying of the adsorbent is accomplished
by drawing air through the column for 5 to 10 minutes.

The

Pyrex tube is then removed from its suction flask and the
adsorbent removed to a point 2 mm. below the red Sudan III
band.

A spatula with a narrow blade and square tip is con­

venient.

This removes vitamin A, carotenoids, and pigments.

The column is replaced in its suction flask and the remainder
of the adsorbate eluted with 25 ml. C.P. ether.
The combined filtrate and eluate are evaporated to dry­
ness under reduced pressure and taken up in 10 ml. purified
chloroform.

To 1 ml. of this solution is added 10 ml. of

-12-

the antimony trichloride reagent.

The flask Is swirled for

30 seconds, the absorption cell filled, and the optical
density determined on the visual or photoelectric spectro­
photometer at exactly 3 minutes after starting the addition
of the reagent to the vitamin sample.

In using the visual

spectrophotometer, optical density readings in the range 0.60
to 1.20 are most easily made, and if the optical density of a
solution is outside of this range, the chloroform solution
may be concentrated or diluted to an appropriate volume and
the color reaction and absorption measurement repeated.

The

vitamin sample at this point contains vitamins D and sterols.
To correct for the absorption at 500 mu due to the
sterols present, a 1 ml. aliquot of the chloroform solution
of vitamins D plus sterols is evaporated to dryness and taken
up in 5 ml. of a solution of 1 part Skellysolve and two parts
benzene by volume.

A tightly packed 1.5 cm. Superfiltrol

column is wet with 5 ml. of the mixed solvent, the sample
added, followed by 5 ml. of solvent to rinse the flask and
50 ml. of the same mixed solvent to develop the liquid
chromatogram.

Vitamins D are fixed at the upper surface of

the adsorption column.

Cholesterol passes Into the filtrate.

The filtrate is evaporated to dryness under reduced pressure
and the residue taken up in 1 ml. of chloroform.

To this is

added 10 ml. of the antimony trichloride reagent, a 2 cm.
absorption cell filled, and the optical density determined
at 500 mu, 3 minutes after mixing.

Other periods of time

may arbitrarily be used, but the same time should be used
for the sterol correction as is used for the vitamins D plus

-131A
sterols determination, and the factor used in converting Ex
values to U.S.P. units per gram is dependent upon this time.
%

Prom these two optical density values,

for vitamins

D plus sterols and for sterols alone are calculated.

The

%

difference between these two Ei^m ^ values is proportional to
the concentration of vitamins D in the original sample.
Experimental results.--Table I shows the results of 30
consecutive determinations made on oil number 47761.
the average of these

Prom

values, a conversion factor was

calculated for converting Ei£m< values of other oils to
U.S.P. units of vitamins D per gram of oil.

Mixed high D oil

number 47761 was selected as a typical fish oil, and was
carefully bioassayed several times, giving a bioassay value
of 15,000 U.S.P. units per gram.

The determinations by the

proposed physico-chemical method were made over a period of
three and one half months, during which time several new lots
of adsorbent, solvents, and antimony trichloride reagent were
%
used. The Ei£m values show a maximum deviation of +, 5%
from the average value.

This indicates that the adsorbent

may be used as received from the manufacturer without con­
trolling its state of activation and that the antimony tri­
chloride reagent used does not require standardization to
control the color forming properties, as is the case with
the reagent of Brockmann and Chen.

-14-

Table I
Oil No*

Type

Date

47761

Mixed High
D oil

3-21

E*cm
cm.

3-23
3-24
3-25
3-26
3-27
3-31
4-1
4-13
4-14
4-27
5-20
5-22
7-6
7-7
Average

«

.75
.80
.76
•80
.74
.75
.80
.82
.77
.78
.80
.78
.77
.76
.79
.80
.74
.75
.77
.79
.78
.78
.76
•80
.81
.79
.78
.81
.79
.76
0.779

Bloassay
15,000 U.S.P.
unlts/g.

Maximum deviation from average = + 5%
15*000

Factor = 0.779
= 19,300
i%
Eicm
unknown X 19,300 = Calculated potency in
U.S.P. units/g.

-15

Table II shows the potencies of oils as determined bio­
logically and as determined by the proposed physico-chemical
method.

In most cases agreement with the bloassay values

is satisfactory.
In the analysis of some of the oils of Intermediate or
low potencies, it was necessary to use rather small samples
in order to assure complete removal of the vitamin A.

This

Is especially true of halibut liver oils, where the vitamin
A concentration is usually high compared to that of the
vitamins D.

Samples of smaller than usual size 3hould also
*

be used when the oil contains such large quantities of
sterols that aggregates of sterol crystals appear in the oil
on standing.

A few oils required prolonged washing of the

ether extract containing the nonsaponifiable fraction.

These

oils were all marked by a tendency to emulsify very easily
during this washing.

Also, green and yellow adsorption bands

appeared below the Sudan III band on the first chromatograph
when the ether extracts were Insufficiently washed.

The

ether extract from a sample of oil number 47761 was washed
with 26 fifty ml. portions of water instead of the usual 6
portions.

No measurable decrease In the vitamins D concen­

tration was found, indicating that no loss of vitamins D
occurs when the ether extract is washed with more than the
usual volume of water.

-16-

Table II

Oil No. Type

Sample
Wt .(g)

EX%

P.O. Method.
Bioassay
U.S.P. u/g U.S.P. u/g

50261

Mixed F.L.O. 2.000

.33
.31

6,000

6,570
5,980

77462

Mixed F.L.O. 1.000

0.43
0.42

6,300

8.300
8,110

60771

Mixed F.L.O. 2.000

.38
.36

6,600

7,310
6.950

66851

Mixed F.L.O. 1.000

.34
.355
.36
.35

6,600

6,550
6.830
6.950
6,750

1.000
74912

Mixed F.L.O. 2.000

0.14
0.15

3,000

2,700
2,890

62321

Mixed F.L.O. 2.000

.27
.25

4.750

5.210
4.830

56211

Mixed F.L.O. 1.000

.26
.27

4.750

5,020
5.210

75442x

Mixed F.L.O. 0.500

2.33
2.42
2.46
2.37

50,000

45,000
46.700
48,100
45.700

1,200

1,270

52051

Mixed F.L.O. 0.4

.066

1.0

1.07
1.03

H .L .0 .
capsules

1.8

0.054
0.052

80352

Swd. L.O.

1.000

.37
.34

5,000

7,140
6,560

80362

Swd. L.O.

1.000

.48
.43

10,000

9,260
8.300

78992

Swd. +
T.L.O.

1.000

.41
.36

4,500

5,520
4,670

68871

T.L.O.

.500

1.38
1.27

25,000

26,600
24,500

llx
66851

20,600
19,900
670 +

1,040
1,000

-17-

Tab 1© IX
(oont•)
Oil No* Type

Sample
Wt.(g)

68881

1.000

T.L.O.

El %
cm.

P.O. Method
Bioassay
TJ*S*P* u/g U.S.P. u/g
12,000

1.000

•60
•60
.62
.59

11,600
11,600
12,000
11.400

76892

T.L.O.

1.000

.797
.802

16,000

15.400
15.500

76902

T.L.O.

1.000

.64
.67

12,000

12.400
12.900

76882

T.L.O.

1.000

1.19
1.15

21,000

1.05

20,000

20.300
19.700

P6846x

T.L.O.

1.000

23.000
22,200

1.02

58011

Bonita
L.O.

0.250

3.11
3.33

65.000

60.000
64.300

57951

T.L.O.

.500

1.40
1.38

28,000.

27,000
26,600

57971

Albacore
L.O.

0.250

3.03
2.94

55.000

58.500
56.700

57991

Yellow Pin
T.L.O.

1.000

0.153
0.162

3,000

2,950
3,130

65221

Mixed High
D Oil

1.000

.70
.75
.69
.67

16.000

13.500
14.500
13.300
12.900

.500
57481

Mixed High
D Oil

1.000

.75
.78

15.000

14.500
15.100

55951

Mixed High
D Oil

.500

1.49
1.56

30.000

28,800
30.100

40090

Mixed High
D Oil

.500

.97
.98

20.000

18.700
18.900

41860

Mixed High
D Oil

.500

.81
.79

16,000

15,600
15,200

55691

Mixed High
D Oil

1.000

.66

12,000

12.700
12.500

.65

-18-

Table XI
(cent•)
Oil No* Type

Sample
Wt.(g)

d>

Ej.__

Bioassay
P*C* Method
U.S.P. We U.S.P. We

44080

D diet.

.250

1.56
1.52

31,000

30,100
29,300

44090

D dist.

.500

1.12
1.13

22,000

21,600
21,800

55031

D dist.

1.000

.96
.934

18,500

18,500
17,900

44280

D dist.

1.000

.26
.25

5,000

5,020
4,830

43050

D dist.

1.000

3,000

1.000

.27
.30
.28
.28

5,210
5,790
5.400
5.400

44170

D dist.

.250

.84
.76

15,000

16,200
14,700

56961

D dist.

1.000

1.11
1.09

20,000

21,400
21,000

60761

Mixed A
and D Oil

.250

1.80
1.89

35,000

34,700
36 ,500

57381

T.L.O.

conc.

.100

260,000

241,300
237,400

63201

T.L.O.

conc.

.100

4.29
4.39

85,000

82,800
84,700

49831

T.L.O.

conc.

.100

7.14
7.47

160,000

151,000
148,400

61751

T.L.O.

conc.

.100

11.7
12.1

300.000 to
350.000

293.000
288.000

44470

T.L.O.

conc.

.100

11.3
12.0

240,000

237,400
229,700

70202

T.L.O.

conc.

.250

65,000

60,400
62,100

V3428

Mixed F.L.O.
Concen.

1,200,000

1.187.000
1.262.000

32,500

31,100
31,300

2x

0.100
0.500

12.3
11.8

3.13
3.22
61.5
65.4
1.43
1.44

-19

Table II
(cont•)
Oil No, Type

Sample
Wt.(g)

3x

Irradiated
ergosterol
in oil

0.100

8x

Irradiated
ergosterol
in oil

0.050

9x

Irradiated
ergosterol
in oil

B3494

x%

Ex cm.

Bloassay
P.C. Method
U*S,P. u/g U.S.P. u/g
192,900

193.000
187.000

23.3
23.8

440.000 to
520.000

450,000
459,000

0.100

26.6
27.0

440,000

513,400
521,100

Irradiated
ergosterol
in oil

0.100

24.7
24.8

over
440,000

477.000
479.000

H-l

Irradiated
ergosterol
In oil

0.1

19.3
19.5

400,000

372,500
376,400

Blend
1

Irradiated
ergosterol
in oil

0.1

21.8

400,000

421.000
405.000

Lot
16792

Irradiated
0.2447
ergosterol
0.2322
in oil(capsules)

10.0
10.8

193,000

193,000
208,400

10.00

9.67

21.0

Lot 200 Irradiated
ergosterol
in oil

0.050

37.8
37.0

880,000

729,500
714,100

Lot 201 Irradiated
ergosterol
In oil

0.100

41.4
42.7

880,000

799.000
824.000

Lot 202 Irradiated
ergosterol
in oil

0.1

42.7
41.9

880,000

824.000
809.000

65251

1.000

12 ,000

12.900
13,100
12.900
13,700

Mixed High
D Oil

.67
.68

.67
.71

-20-

Discus aion.— In the isolation of vitamin Da from tuna
oil, Brockmann (1) concentrated vitamin Da by chromatographic
adsorption on alumina from a 1 to 4 benzene-petroleum ether
mixture.

To locate the crude vitamin Da on the adsorbent,

indicator red 33 (Sudan III) was added to the sample, and a
high concentration of vitamin D a was found In the red dye band.
This work suggested the use of Sudan III as a marker for
locating the colorless vitamins D band in the separation from
vitamin A by adsorption on Superfiltrol from 5 to 1 hexaneether mixture.

For economy, Skellysolve was substituted for

the hexane in this mixture.

Using highly purified anhydrous

ether, the dye band remained fixed at the upper surface of
the adsorbent.

When „C.P. ether was used, the dye band passed

down the column and tests showed that cholesterol and vita­
mins D preceded the dye band.

A 5 to 1 Skellysolve-anhydrous

ether solution was saturated with distilled water and used as
solvent and developer, but the dye remained fixed at the top
of the column.

Also, the water caused hardening of the ad­

sorbent column, uneven development of bands, and an almost
complete stoppage of percolation of the developing solution.
Knowing that ethyl alcohol was a common impurity in com­
mercial ether, varying amounts of absolute ethyl alcohol were
added to the 5il Skellysolve-anhydrous ether solvent.

Tests

showed that a mixture of 50 parts Skellysolve, 10 parts puri­
fied anhydrous ether, and 1 part absolute ethyl alcohol by
volume gave best results.

The Sudan III migrated down the

column in a very sharp, narrow band, and after adequate de­
velopment the filtrate and the adsorbent below this dye band

-21-

contalned the vitamins D but no vitamin A as determined by
measurements of the absorption at 500 mu and at 620 mu when
antimony trichloride reagent was added to a chloroform solution
of the residue from evaporation of the filtrate and eluate.
With less than this amount of alcohol, development was un­
necessarily slow, while with larger amounts of alcohol the
separation of vitamin A was incomplete.

Although small amounts

of Sudan III do not interfere with the maxima in the absorption
spectra at 500 mu, many fish oils contain interfering substances
which develop adsorption bands overlapping the Sudan III band.
For this reason, it is necessary to remove the adsorbent to
a point slightly below the dye band before elution of vita­
mins D and sterols from the column.

The removal of pigments

as well as vitamin A by this chromatographic procedure is in­
dicated by the colorless nature of the filtrate and eluate.
For the separation of sterols from vitamins D numerous
ratios of hexane-ether, hexane-benzene, and Skellysolvebenzene were tested as solvents, using Superfiltrol as ad­
sorbent.

With pure hexane or pure Skellysolve, or with only

small amounts of ether or of benzene present in the mixtures,
both vitamin Ds and cholesterol were very strongly adsorbed.
With larger amounts of ether present, both vitamin D3 and
cholesterol passed down the column without adequate separation
of their adsorption bands.

However, Increasing the ratio of

benzene to Skellysolve in these mixtures caused the cholesterol
to migrate down the column much faster than did the vitamin
Da . Finally, it was found that with a 1 to 2 ratio by volume
of Skellysolve to benzene, and a 2 cm. Superfiltrol column, a

-22-

0*018 gram sample of cholesterol was passed quantitatively
into the filtrate when 50 ml. of developer were used.

Using

270,000 units of vitamin Da as solute and 350 ml. of the'1:2
Skellysolve-benzene mixture as developer, none of the vitamin
Da appeared in the filtrate.

Less than 30% of the calciferol

could toe eluted with ether, ethyl alcohol, or methyl alcohol.
For this reason it was decided to analyze for vitamins D plus
sterols and for sterols alone and obtain the vitamins D con­
centration toy difference.
As a final test to determine whether cholesterol and
ergosterol would interfere in the actual analysis of a fish
oil by the proposed method, varying amounts of these substances
were added to 1 g. samples of reference oil number 47761, and
the samples analyzed for vitamins D (Table III).

These data

show that the presence of even 0.5 g. cholesterol above that
already in the natural fish oil caused an error of less than
13% in the analysis.

This amount of cholesterol is far higher

than would be found in a fish oil.

For 0.050 g. added cho­

lesterol the error due to the added cholesterol is only 2%,
and this is well within the limit of reproducibility of the
method as shown by Table I.
The presence of a provitamin, ergosterol, in concentrations
of slightly higher order than the vitamin Itself, caused only
a slight Increase in the calculated potency.

This is proba­

bly due to a much lower molecular extinction coefficient for
the ergosterol than for the vitamins D in the antimony tri­
chloride color reaction.

The chief value of the additions

of the ergosterol, however, lies in the determination that no

-23-

measurable amount of the provitamin was activated during the
analytical process*
Table III
1%
Ex

Sample Material added
No. to No. 47761

500 mu
3 min.
S
D

D+S

.778 15.000
.20

2.

.025 g« cholesterol

1.00

.23

.77

15,000

14,900

3.

.050 6* cholesterol

1.01

.25

.76

15,000

14,700

4.

.100 8* cholesterol

1.12

.39

.73

15,000

14,100

5.

.250 g- cholesterol

1.36

.64

.72

15,000

13,900

6.

.500 g. cholesterol

1.74 1.11

.73

15,000

14,100

7.

.001 g* ergosterol

1.02 ,.21

.81

15,000

15,600

8.

.005 g- ergosterol

.80

15,000

15,400

•

CD

Refer­ None
ence
1.
.010 g« cholesterol

Bioassay
F.C* Method
U.S.P. u/g U.S.P. u/g

.92

.12

.78

15.000

15,050

Summary •—
1.

A physico-chemical method for the determination of vitamins
D in fish oils has been presented; in which the oil is
saponified, the nonsaponifiable fraction extracted, and
vitamin A, carotenoids, and pigments removed by chromato­
graphic adsorption.
determined.

The vitamins D plus sterols are then

The vitamins D are removed from an aliquot

of this sample by adsorption, and the sterols determined.
Analysis for vitamins D is obtained by difference.
2.

Experiments are discussed and data presented showing that
interference from vitamin A and cholesterol has been re­
moved, and that small amounts of provitamin do not inter­
fere with the analysis.

-24-

3.

A table Is presented showing the potencies o£T

fish oils

as determined biologically on rats and as dets- «rrained by
the proposed physico-chemical method*

The m & & n deviation

of the physico-chemical potencies from the bi-oassay
potencies Is 7

-25Literature Cited
1

Brockmann, H., Z. physiol. Chem., 241. 104 (1936).

2

Brockmann, H., and Busse, A., Ibid., 126, 573 (1939).

3

Brockmann, H., and Chen, Y., Ibid., 241. 129 (1936).

4

Christensen, E., Munch. Med. Wochschr., 75, 1883 (1928).

5

Emmerie, A., and van Eekelen, M., Acta Brevia Neerland.
Physiol. Pharmacol. Microbiol., 6., 133 (1936).

6

Ewing, D.T., and Tomkins, F., forthcoming publication.

7

Gudlet, I., Proc. Scl. Inst. Vitamin Research U.S.S.R.,
3, No. 1, 35 (1941).

8

Halden, W., Naturwissenschaften, 24. 296 (1936),

9

Halden, W., and Tzoni, H., Nature, 137. 909 (1936).

10

Levine, J., Biochem. J., 27,, 2047 (1933),

11

Marcussen, E., Dansk. Tids. Farm., 13, 141 (1939).

12

Milas, N«, Heggie, R., and Raynolds, J., Ind. Eng. Chem.
Anal. Ed., 13, 227 (1941).

13

Nield, C., Russell, W., and Zimmerli, A., J. Biol. Chem.,
156. 73 (1940).

14

01 khin, B., Proc. Sci. Inst. Vitamin Research U.S.S.R.,
3, No. 1, 28-29 (1941).

15

Raoul, Y., and Meunier, P., Comptes rendus, 209. 546 (1939).

16

Reerink, E., and van Wijk, A., Chem. Weekblad., 1952. 645.

17

Ritsert, K . , E. Merck s Jahresberichte, 52, 27 (1938).

18

Robinson, F., Chem. Ind., 56,, 191 (1937).

19

Rutkovskii, L., Biokhimya, 5, 528 (1940).

20

Shear, J., J. Proc. Soc. Exptl. Biol. Med., 23, 546 (1925).

21

Stoeltzner, W., Munch. Med. Wochschr., 715, 1584 (1928).

22

Topelmann, H., and Schuhknect., W., Z. Vitaminforsch.,
4, 11 (1935).

23

Tortelli, M., and Jaffee, E., Ann. Chim. applicata, 2,
80 (1914).
~

-26-

24)

Tzoni, H., Biochem. Z., 28*7, 18 (1936).

25)

Wolff, L., Z. Vitaminforsch., 7, 277 (1938).