THE ISOLATION OF VITAMIN Kl*

BY S. B. BINKLEY, D. W. MAcCORQUODALE, SIDNEY A. THAYER, AND EDWARD A. DOISY

(Prom the Laboratory of Biological Chemistry, St. Louis University School of Medicine, St. Louis)

(Received for publication, June 26, 1939)

Since the discovery of the hemorrhagic syndrome (Dam, 1929, 1930) many interesting contributions to the chemical study of vitamin K have been made by Dam and his collaborators, Alm- quist and his associates, and other investigators. Viewed from the standpoint of our knowledge of the vitamin, one of the most important communications is the recent paper (Dam, Geiger, Glavind, Karrer, Karrer, Rothschild, and Salomon, 1939) in which the preparation of a highly potent oil possessing vitamin K a.ctivity was reported; however, the procedure used in its prepara- tion was not described. According to the extinction coefficient

(3 :%n. = 280 at 248 mp) given in that report and the value re- ported to us by Professor Ewing1 (385; subsequent value 540,

* This investigation was begun in September, 1936. Since early in 1938 we have been assisted in many phases of our work by the active cooperation of Parke, Davis and Company of Detroit. We are happy to acknowledge our appreciation to this concern and particularly to its Scientific Director, Dr. Oliver Kamm.

Preliminary reports on the isolation and preparation of derivatives of vitamins K1 and KZ were published in the May and June (1939) numbers of the Journal of the American Chemical Society. Since two different pure substances possessing vitamin K activity were isolated, the compound from alfalfa which has the smaller molecular weight was called vitamin Ki and the compound from putrefied fish meal having the larger molecular weight, vitamin KS. It is hoped that these temporary names will prove acceptable to the other investigators in this field until the structure of these com- pounds is fully understood. Then names expressive of the structure and action can be selected.

r Our thanks are due to Professor D. T. Ewing of the Michigan State College for the ultraviolet absorption studies. A separate report on the absorption studies will be made.

219

by guest on May 24, 2018

http://ww

w.jbc.org/

Dow

nloaded from

220 Isolation of Vitamin K,

McKee et al., 1939; Binkley et al., 1939) this oil was approximately 50 per cent vitamin K. In this communication we are reporting details of the actual isolation of vitamin K1 and the preparation of a crystalline derivative.

In order to clarify the situation, we wish to say that we have previously reported (Thayer et al., 1938) the isolation of a crystal- line compound possessing vitamin K activity. It was later stated by us (MacCorquodale et aE., 1939) that this compound (m.p. 69”) is not the vitamin and does not possess vitamin K activity. We are calling attention to that statement in order to correct as far as possible our mistake.

Soon after the appearance of our first report, the isolation of vitamins K1 and Kz (McKee et al., 1939) and the preparation of crystalline derivatives (Binkley et al., 1939) were reported. Since the published methods for concentrating vitamin K extracts are wholly inadequate, it seems desirable to publish a method which is simple, easily carried out, and is practically quantita- tive from start to finish. This paper is concerned only with the procedure for the isolation of vitamin K1, the vitamin occurring in alfalfa.

Purification of Vitamin K,

Crude extracts were prepared from artificially dried alfalfa meal or leaf meal by percolation with petroleum ether. Acetone is a good solvent for removing vitamin K from the meal but the extracts contain at least twice the weight of solids obtained with petroleum ether. Although some purification of the crude ex- tracts could be effected by treating them with MgO (Almquist, 1936), the loss of vitamin was often as great as 50 per cent in this laboratory and the amount of concentration was disappointing. Partition between various solvents and crystallization from a number of solvents were equally unsuccessful. The large amount of impurities in crude preparations made distillation at low pres- sures a difficult and impractical process. The inactivity of the vitamin toward many chemical reagents, such as acetic anhydride, hydroxylamine, etc., as well as its instability toward others, such as alkali, eliminated chemical reactions as a means of isolation.

Adsorbents-Dam et al. (1936, 1937, 1938) reported that the vitamin was not recovered from some adsorbents, e.g. alumina and

by guest on May 24, 2018

http://ww

w.jbc.org/

Dow

nloaded from

Binkley, MacCorquodale, Thayer, Doisy 221

MgO, but could be recovered from sucrose, CaC03, and CaS04. In order to isolate the vitamin it was necessary to find an adsorbent toward which the vitamin was stable, from which the vitamin could be readily eluted, and which possessed adsorptive properties favorable to the separation of the vitamin from inert substances. It was found that the vitamin was partially or wholly inactivated by commercial alumina, magnesium oxide, infusorial earth, mag- nesium oxide plus supercel (Nos. 1 to 3) but was stable toward sucrose, decalso, permutit, various activated charcoal preparations such as norit, darco, nuchar, calcium carbonate, and calcium sul- fate. The efficiency of sucrose, calcium sulfate, and CaC03 was not sufficient to warrant their use. Over forty adsorption experi- ments were run before satisfactory adsorbents were discovered and proper methods of using them.

With a modified chromatographic adsorption method it was found that decalso and permutit were very satisfactory for the concentration of crude preparations, while charcoal was ad- vantageous with preparations containing more than 10 per cent vitamin. Decalso is a synthetic zeolite manufactured by The Permutit Company2 (New York) for water-softening purposes. The permutit used was prepared according to Folin. In each case the adsorptive properties of different shipments varied some- what. This difference may be due partially to the water content and partially to variability in the method of manufacture. In general, the greater the water content the less the adsorption. For example, the permutit used for the adsorption illustrated by Table IV did not adsorb the vitamin as strongly as the permutit used for the preparation described in Table III. For this reason petroleum ether containing less benzene was used for elution of the vitamin. The chief advantage of decalso over Folin permutit is one of economy.

Adsorption Columns-Columns of adsorptive material of the following sizes were used: 3.5 X 40 cm., 5.0 X 60 cm., 10.0 X 70 cm., 5.0 x 580 cm., and a column 5 feet tall containing 750 pounds. With the exception of the last column which was cop- per, the columns were of Pyrex glass and were of a height sufficient to hold a convenient volume of solvent above the adsorbent.

2 We appreciate the assistance in this investigation of The Permutit Company, New York.

by guest on May 24, 2018

http://ww

w.jbc.org/

Dow

nloaded from

222 Isolation of Vitamin K,

The Pyrex columns were constricted at the bottom and a 15 mm. outlet tube sealed on. A Gooch plate and then a layer of cotton were used to cover the outlet. The desired amount of permutit or decalso (Folin grain size) was added and packed down firmly by jarring the column. The adsorbent was covered with cotton and the column was ready for use.

El&ion of Vitamin-The material to be adsorbed was dis- solved in petroleum ether (b.p. 90-105”) to make a 2 to 5 per cent solution. The amount of solid that can be used is 1 part of solid to 25 to 60 parts of adsorbent. Any crystalline material that separated was removed by filtration, and the filtrate poured directly into the column and allowed to flow through by gravity. As soon as this solution had passed into the adsorbent, more petroleum ether was added and this was followed in turn by 10 per cent benzene in petroleum ether, 20 per cent benzene in pe- troleum ether, and acetone. The filtrate from the column was collected in fractions from which the solvent was distilled and the residue assayed for vitamin K activity (Thayer, McKee, Binkley, MacCorquodale, and Doisy, 1939). The quantities of solvents used as well as the size and number of fractions collected depended on the size of the column and the adsorbent used, as is shown in Tables I to VI. The amount of solvents used was such that all the vitamin had passed into the filtrate before acetone was added. Washing with acetone was used to prepare the adsorbent for future use.

In the case of the adsorption of crude preparations containing large amounts of chlorophyll it was often necessary to discard the uppermost layer of the adsorbent or to use it again only to remove a considerable portion of the chlorophyll before the adsorptive process. The remainder of the decalso could be recovered by thoroughly extracting it with ethyl alcohol and air-drying for at least 48 hours. When more concentrated preparations which contained no chlorophyll were used, the adsorbent was recovered by merely air-drying for 48 hours after the acetone washing. Either decalso or permutit may be used as often as desired.

It was found that decalso or permutit of a smaller grain size required more time to carry out the operation without an appreci- able increase in efficiency.

by guest on May 24, 2018

http://ww

w.jbc.org/

Dow

nloaded from

Binkley, MacCorquodale, Thayer, Doisy 223

Isolation of Vitamin K1

Four or five adsorptions were usually necessary to obtain a product which was essentially pure vitamin: (1) one or two ad- sorptions on decalso which increased the concentration of vitamin about 20 to 40 times (Tables I and II) ; (2) adsorption on permutit which gave about a 5- to lo-fold concentration (Table III) ; (3) adsorption on a 5.0 X 580 cm. column of permutit which gave a 2- to 5-fold concentration (Table IV) ; (4) adsorption on darco (Table V).

Preparations containing 10 to 30 per cent vitamin are yellow to red in color and contain about 10 per cent of crystalline material which may be removed by crystallization from absolute ethyl alcohol (5 to 10 cc. per gm. of material). The filtrate was then adsorbed on a column (3.5 X 30 cm.) of darco (20 to 40 mesh) and eluted in a manner similar to that used with permutit.

Preparations such as Fractions Tle, Tlf, Tlg, and Tlh (Table V) are lemon-yellow oils which as nearly as can be discerned by assay are essentially pure vitamin K1. In this condition the preparations are extremely sensitive to both light (MacCorquodale et al., 1939) and alkali. The preparations were handled in a dimly lighted room and stored in a dark closet or ice box. All equip- ment was washed with chromic acid cleaning solution, distilled water, and ethyl alcohol before use. A drop or two of glacial acetic acid was added to each fraction as it was collected from the column.

When Fraction Tlf (0.456 gm.) was distilled in a molecular still of the type described by Almquist (1937) at 115” and 145’, and 2 X 1O-4 mm. pressure, 91 per cent of the product was recovered in the fraction distilling between 115“ and 145”. By assay the potency of this product was the same as before distillation; namely, 1000 units per mg.

The product from this distillation as well as several other prod- ucts obtained in a like manner was crystallized from both absolute ethyl alcohol (10 cc. per 100 mg.) and acetone (10 cc. per 100 mg.) by cooling for 3 or more hours in a dry ice-acetone bath. The filtrates contained 2 to 4 mg. per cc. The crystals were light yellow rosettes which melted at about -20” into solvent and a yellow viscous oil which gradually passed into solution.

by guest on May 24, 2018

http://ww

w.jbc.org/

Dow

nloaded from

224 Isolation of Vitamin K,

Equally pure preparations were obtained by crystallizing frac- tions such as No. Tlf without subjecting them to distillation. In either case it was usually necessary first to remove a small amount of less soluble material by cooling an absolute ethyl alcohol solu- tion (5 cc. per 100 mg.) at -5” and filtering.

Carbon and hydrogen analyses and molecular weight determi- nations indicated that these preparations were a single compound having the empirical formula CalH4602. Upon hydrogenation the vitamin absorbed 8 atoms of hydrogen, giving a colorless product which on exposure to air was oxidized to a yellow product similar to the original vitamin in color. This yellow compound absorbs 1 mole of hydrogen to form a colorless product.

The behavior upon hydrogenation, the absorption spectrum, and the lability toward light (MacCorquodale et at., 1939) and alkali indicated that the vitamin was a quinone. Reductive acetylation gave a white crystalline derivative melting at 59”. This derivative possessed the activity of vitamin K and was converted into vitamin K1 which in turn was converted back to the diacetate.

This compound is not readily hydrolyzed by alkali or acids in an aqueous or alcoholic medium. In alcoholic solution its activity is unstable to 1 per cent potassium hydroxide or to 36 hours ex- posure to sunlight but is stable to at least 20 days of exposure to light from a 100 watt bulb at a distance of 4 feet.

Diacetyldihydro vitamin K1 was converted to vitamin K1 by treating an anhydrous ether solution with an excess (fourteen times) of methyl magnesium iodide, followed by shaking an ether solution of the hydrolyzed product with air. After molecular distillation and crystallization from acetone, the product showed the same analysis, assay, and ultraviolet absorption spectrum as the original vitamin preparations.

EXPERIMENTAL

Extraction and Adsorption-Since a specific example of the processes which we have discussed in general terms only may be of value to those who may wish to use our procedure, we are giving the detailed description of the preparation of pure vitamin K1.

1000 pounds of kiln-dried alfalfa leaf meal were percolated with

by guest on May 24, 2018

http://ww

w.jbc.org/

Dow

nloaded from

Binkley, MacCorquodale, Thayer, Doisy 225

petroleum ether (b.p. SO-100’) for 12 hours.a Approximately 500 gallons of petroleum ether were used; after percolation the solvent contained about 20 pounds of extracted solids. After being con-

TABLE I*

Fractionation of Crude Extract of 1000 Pounds of Alfalfa Leaf Meal

10,000 gm. of solids; 4,000,OOO units; 0.4 unit per mg. The volume for each fraction was 240 gallons.

Sdlvent

Petroleum ether I‘ ‘I “ ‘I ‘I ‘(

10% benzene in petroleum ether

Total solids

m-

1034 456 506 200

72

Units in fraction

Units per mg. solids

<0.3 <2

4 7.5

<9

* Performed in the Parke, Davis and Company laboratories.

TABLE II*

Second Fractionation of Active Fractions Obtained from First Fractionation

Decalso was again used. 18,400 gm. of solids; 26,000,OOO units; 1.4 units per gm. 375 gallons of petroleum ether were used as solvent for each fraction.

Fraction No. Solids obtained

gm.

5374 2498 1167 713 475 581

Assay

Units in fraction

Inactive <1,000,000 <1,000,000 10,000,000

> 13,000,000 8,000,000 2,000,000

.- Units per mg.

<0.2 <0.4

8.5 >18

17 3.5

* Performed in the Parke, Davis and Company laboratories.

centrated to about 120 gallons, the extract was ready for further purification.

* Performed in the Parke, Davis and Company laboratories, Detroit.

by guest on May 24, 2018

http://ww

w.jbc.org/

Dow

nloaded from

226 Isolation of Vitamin K,

The concentrated extract was then percolated through a 5 foot column containing from 500 to 750 pounds of decalso. This procedure gave fractions practically free from chlorophyll. Table I shows the results of an actual adsorption experiment in which a crude petroleum ether extract of 1000 pounds of alfalfa contain-

TABLE III

Third Fractionation of Portion of Fraction 6, Table II*

Permutit was used. 106 gm. of solids; approximately 3,000,OOO units; 28 units per mg.

Jolume Fraction NO. Solvent Weight

Units in fraction ;‘,“iEIP$

__- gm.

T85a Petroleum ether 4.22 Inactive T85b 10% benzene

90% petroleum ether 24.30 ‘I

T85c 10% benzene 90% petroleum ether 38.92 <100,000 <3

T85d 10% benzene 90% petroleum ether 14.47 140,000 10

T85e 10% benzene 90% petroleum ether 8.81 440 ) 000 50

T85f 10% benzene 90% petroleum ether 4.78 480,000 100

T85g 10% benzene 90% petroleum ether 3.77 380,000 100

TS5h 20’% benzene 80% petroleum ether 2.75 >275,000 > 100

T85i 20% benzene 80% petroleum ether 2.48 >250,000 > 100

T85j 20% benzene 80% petroleum ether 1.39 >140,000 >lOO

T85k Acetone 2.216 <100,000 <50

* Assay of this fraction after its receipt from Parke, Davis and Company showed that it contained about 3 million units.

liters 6

3

3

3

3

3

3

3

3

6 3

Assay

ing 4 million units of activity and 10,000 gm. of extractives was passed through 750 pounds of decalso.

The results shown in Table II were obtained by a second decalso adsorption of the active fractions obtained from the first adsorp- tion with decalso. These active fractions obtained from the

by guest on May 24, 2018

http://ww

w.jbc.org/

Dow

nloaded from

Binkley, MacCorquodale, Thayer, Doisy 227

first adsorption and representing 6000 pounds of alfalfa contained 18,400 gm. of solids and about 26 million units.

A portion of Fraction 5 (106 gm., see Table II)4 was then dis- solved in 2 liters of petroleum ether and adsorbed on a column of permutit (10 X 70 cm.) by the method already described. The results are shown in Table III.

Fraction e and Fractions f to j were further concentrated by adsorption on a column of permutit 5.0 X 580 cm. The results of

TABLE IV Fourth Fractionation on Column (0.05 X 5.8 Meters) of Permutit

Columns 4, 5, and 6 show fractionation of Fraction T85e and Columns 7, 8, and 9 the fractionation of a combination of Fractions T85f to j of Table III. The weight of the former was 8.81 gm.; of the latter, 15.17 gm.

FrtX- tion No.

(1)

a

i

-

Solvent

(2)

5yo GHe in 95% petroleum ether

‘I ‘C ‘I “ ‘I ‘I ‘I ‘I “ ‘1 ‘I “

10% CLHB in petro- leum ether

Acetone

7

1

Assay Assay “& sygp Solids

“me tained ;$&;i; Units I ?b- Units r

per tamed Units in mg. fraction ggr

(3) (4) (5) (6) (7) (8) (6,. ~__ ___ -~ _____~ riters gm. gm.

4 0.905 <95,000 <100,3.612<180,090 <50

1 3.022 I

<500,000 <160,2.467 <200,000 <80 1 1.930 <300,000 <150,2.627 >870,000 >330 1 0.999 150,000 1502.240 1 0.630 100,000 16011.169

>740,000 >330 >390,000 >330

1 0.718 >250,000’>350,0.980 300,000 300 1 0.301 30,000 100,1.039 300,000 300 2 80.406 <20,000 <501.199 <50,000 <40

2 i0.337 <50~000~<150,1~067 <50,060 <50 _--

this adsorption are shown in Table IV. The results of an adsorp- tion experiment starting with 5.72 gm. of material containing 1.5 million units and with use of darco as adsorbent are shown in Table V.

Andysis of Vitamin K1-In Table VI, we have given the anal- yses of several preparations of vitamin K1. Carbon and hydrogen were determined by the Pregl method and the molecular weight

4 Assay of this fraction after its receipt from Parke, Davis and Company showed that it contained about 3 million units.

by guest on May 24, 2018

http://ww

w.jbc.org/

Dow

nloaded from

228 Isolation of Vitamin K1

TABLE V

Fifth Fractionation with Column 0s Darco

Weight 5.72 gm. 260 units per mg.

Fraction NO.

Tla

Tlb Tic

Tld Tle Tlf

Tlg Tlh Tli

Tlj Tlk Tll

Tlm

Solvent

Ethyl alcohol Petroleum ether 50yo ethyl alcohol,

50% benzene

Benzene “ “ ‘I

I‘ “ “

Hot benzene I‘ I‘ “ ‘I

Volume

-

1.5 liters

3 “

500 cc.

200 I‘ 300 “ 500 (‘

500 “ 500 “ 500 Ii

1 litei 500 cc. 500 ‘(

1 liter -

Weight Units in fraction

gm. *

1.0760 0.4325

Inactive ‘I

0.3485 170,000 0.3075 307 ) 000 0.4567 457,000

0.2845 285,000 0.1795 180,000 0.0973 97,000

0.1789 <50,000 0.1080 <30,000 0.0810 <30,000

0.0706 <20,000

Assay

Units mr mg.

500

1000 1000

1000 1000 1000

<300 <300 <400

<300

* Contains chiefly impurities from carbon.

TABLE VI

Analyses of Vitamin K1 and qf Diacetate of Dihydro Vitamin K1

Fraction No. Carbon Hydrogen Mol. wt.

per cent per cent Vitamin K1

T3b2

T2Oal

T47a2 From diacetate

82.76 10.44 82.54 10.65 82.65 10.66 82.34 10.53 82.47 10.62

82.66 10.54

450, 436

Calculated for

C32H4@2 CLHteOz

Diacetate of dihydro vitamin K1

Calculated for

C3.&404 Cd35204

82.70 10.41

82.62 10.26 78.21 10.07 78.01 10.03

464 450 531

78.50 9.88 550 78.33 9.77 536

hydrogenation

mdes

3.93

4.00 3.91

2.95

by guest on May 24, 2018

http://ww

w.jbc.org/

Dow

nloaded from

Binkley, MacCorquodale, Thayer, Doisy 229

by the Rast method. Nitrogen, phosphorus, sulfur, and halogens were not detected by qualitative analyses. A 1 per cent solution in absolute ethyl alcohol did not rotate polarized light.

Hydrogenation by the method described by Smith (1932) gave the following results: 18.0 mg. of Fraction T20al in 5 cc. of a mixture of equal volumes of glacial acetic acid and n-butyl ether with 80.7 mg. of PtOz required 3.52 cc. of Hz (s.T.P.), correspond- ing to 3.93 moles for a molecular weight of 450; 20.5 mg. in the same solvent with 90.1 mg. of PtOz required 4.05 cc. of Hz (s.T.P.),

corresponding to 4.00 moles; 18.6 mg. of Fraction T47a2 in the same solvent with 85.0 mg. of PtOz required 3.60 cc. of Hz (S.T.P.),

corresponding to 3.91 moles. A colorless product was formed which on exposure to air was oxidized to a yellow product similar to the original vitamin in color. 27.0 mg. of this yellow product in 5 cc. of the same solvent mixture with 55.0 mg. of PtOz required 1.21 cc. of Hz (s.T.P.) corresponding to 0.90 mole.5

The ultraviolet absorption curves show maxima at 243, 248, 261, 270, and 323 rnp with an extinction coefficient of E :Frn, = 540 at 248 mp. Owing to the instability of the vitamin toward light and storage, it is not certain that 540 is the maximum value attainable. After catalytic reduction the maxima at 243, 248, and 325 rnp disappear. Exposure to light destroys the charac- teristic absorption of the vitamin.but produces general absorption in the ultraviolet region with the more intense absorption from 240 to 270 1nE.r.

Diacetate of Dihydro Vitamin Kl-100 mg. of vitamin K1 and 100 mg. of fused sodium acetate were dissolved in 10 cc. of acetic anhydride and the solution was refluxed with 1 gm. of Zn dust for 30 minutes. An additional 1 gm. of Zn dust was added in portions of a few mg. during the period of heating. The mixture was filtered while hot and the zinc washed with a few cc. of acetone. The acetic anhydride was decomposed by the addition of water and the mixture was extracted with ether, and the ether washed with water and evaporated to dryness. The residue was dissolved in absolute methyl alcohol (5 cc.), treated with a small amount of

5 Note Added to Page Proof---As a result of our degradation experiments and the synthesis of vitamin K,, C31H4602, (Binkley, Cheney, Holcomb,

McKee, Thayer, MacCorquodale, and Doisy, J. Am. Chem. Sot., in press) certain corrections have been made in the page proof.

by guest on May 24, 2018

http://ww

w.jbc.org/

Dow

nloaded from

230 Isolation of Vitamin K1

norit, filtered, and the filtrate was evaporated to 3 cc. and cooled to -5”. The product was filtered off and dried in vacua over CaClz; m.p. 56-57”. Yield 85 per cent. After recrystallization from low boiling petroleum ether (b.p. 30-60’) and methyl alcohol it melted sharply at 59”. The product was very soluble in ace- tone, ethyl alcohol, and benzene. The photomicrograph (Fig. 1) is of our most highly purified specimen.

Analysis of Diacetate-Found, C 78.21, 78.01, H 10.07, 10.03, mol. wt. 531. Calculated for CJeH6404, C 78.50, H 9.88, mol. wt. 550; for C3~Hs204, C 78.33, H 9.77, mol. wt. 536. Microhydro- genation: 24.7 mg. in 5 cc. of a mixture of equal volumes of

FIG. 1. Diacetate of dihydro vitamin Kr. Magnification 62 X. This specimen has been crystallized nine times with methyl alcohol, ethyl alcohol, and petroleum ether as solvents. M.p. 62-63”; I$& = 1600 at

232 rnp; C 78.28, 78.22, H 9.76, 9.93; assay, 600 units per mg.

glacial acetic acid and n-butyl ether with 85.0 mg. of PtOz re- quired 3.04 cc. of Hz (s.T.P.), which corresponds to 2.95 moles (mol. wt. 536). There is general absorption in the region from 220 to beyond 300 rnp with intense absorption at 230 rnp where the extinction coefficient of E :F*,. = 1300. Assay, 500 units per mg.

Hydrolysis of Diacetate-200 mg. of the diacetate of dihydro vita- min K1 were refluxed in 20 cc. of anhydrous ether with 0.02 mole of methyl magnesium iodide for 5 hours.6 The reaction mixture was poured into 100 cc. of water to which 10 drops of dilute HCl had

6 For executing this hydrolysis we gladly express our appreciation to Dr. L. C. Cheney.

by guest on May 24, 2018

http://ww

w.jbc.org/

Dow

nloaded from

Binkley, MacCorquodale, Thayer, Doisy 231

been added. Dilute HCl was added until the solution was acid to litmus. The aqueous phase was extracted with ether, and the ether solution washed with water and evaporated. The weight was 170 mg. The product was transferred to a molecular still and distilled at 2 X lo4 mm. pressure and 115’ and 145” for 3 hours. The fraction distilling between 115’ and 145” weighed 120 mg., the residue 50 mg. With a longer period most of the material would have been in the fraction distilling at 115-145”. The fraction of 120 mg. was crystallized from 10 cc. of acetone at -70”. Yield 82 mg. Analysis, found, C 82.47, 82.66, H 10.62, 10.54. Assay, 1000 units per mg. The mother liquors from the acetone crystallization and the residue from distillation (combined weight 81 mg.) were active at 1000 units per mg. and yielded 95 mg. of crude diacetyldihydro vitamin K1 or 81 mg. of purified product.

Reductive acetylation of the fraction of 82 mg. gave a compound which according to melting point (59”), mixed melting point (59’), and bioassay (500 units per mg.) was identical with the original diacetate of vitamin K1.

DISCUSSION

We believe that the evidence submitted in this paper is adequate to show that the isolation of vitamin Kr has been accomplished. Although the oily character and instability of vitamin K1 and the lack of accuracy of assay compared with chemical and physical criteria of purity have made the problem difficult, the preparation of the stable crystalline diacetate has offered conclusive evidence of real isolation. From a light yellow oil which by analysis and assay seemed to be pure, a colorless crystalline diacetate having a sharp melting point was prepared. Upon assay by our proce- dure this diacetate possessed one-half of the potency of the original oil; by hydrolysis the original yellow oil with double the potency of the diacetate was recovered in good yield. From this oil a diacetate agreeing with the original diacetate was prepared.

A similar reduction of potency due to the formation of the di- acetate of the dihydro compounds has been observed withvitamin K2 and also with 1,4-naphthoquinone and a-methyl-1 ,Cnaphtho- quinone (Thayer, Cheney, Binkley, MacCorquodale, and Doisy, 1939).

Though the analytical data presented in this paper are not

by guest on May 24, 2018

http://ww

w.jbc.org/

Dow

nloaded from

232 Isolation of Vitamin K,

adequate to establish whether C32H4802 or G1H4602 is the right formula for vitamin K1, we believe that C31H4602 is correct. The carbon, hydrogen, and molecular weight determinations indicate either formula equally well. However, the results of degradation experiments which are being published elsewhere afford strong evidence for C31H4602.

Perhaps the nearest approach to the isolation of vitamin K1 has been made by Dam; Karrer (1939), and their coworkers. A comparison of their ultraviolet absorption curve with our data shows that they have failed to find a maximum at 243 rnp. Since this has been found in the examination of all of our preparations of vitamin K1, we believe that its existence is real. It would seem from an examination of their curve that this maximum should have been detected even though their preparation was probably only 50 per cent vitamin.

The value of 50 per cent is arrived at by a comparison of the extinction coefficients at 248 mp. The value of E : &. found for one of our best preparations, 540, is approximately double the value, 280, reported by Dam et al. Moreover, the carbon and hydrogen analyses are rather close to the correct values but since the authors did not calculate a formula, it appears that there was a lack of confidence in the purity of their product.

Almquist and Klose (1939) reported that vitamin K1 formed a choleic acid with desoxycholic acid. We have been unable to obtain a choleic acid from the pure vitamin preparations. The only products obtained were vitamin K1 and desoxycholic acid. However, when crude preparations were used, substances with higher melting points than desoxycholic acid were formed. The activity of these preparations was evidently due to adsorption of the vitamin on choleic acids formed by the impurities present.

In an earlier publication (Thayer, McKee, Binkley, Mac- Corquodale, and Doisy, 1939) it was stated that for accurate assay work ten to twenty chicks should be used at a single dosage level. However, it was found that two to four chicks at a single level gave results sufficiently reliable for ordinary concentration work. In fact, most of our routine purifications have, been con- trolled by assays with the smaller number of chicks. If both of two chicks give either a positive or a negative response the result is usually satisfactory and the quantity of the vitamin prepara-

by guest on May 24, 2018

http://ww

w.jbc.org/

Dow

nloaded from

Binkley, MacCorquodale, Thayer, Doisy 233

tion given on the following day is reckoned on the basis of that response. If one of the chicks responds and the other fails to react, the’ administration of the same dosage is repeated.

Since in an investigation of this character the assays become a laborious task, any procedure that lightens the load is welcome. In addition to our finding that a small number of birds usually gives a reliable assay, the report of Ansbacher (1938) that chicks will respond to administration of vitamin K in 4 to 6 hours has been of real assistance. Since there seems to be no advantage of a 6 over an 18 hour period of assay with respect to the chemical work of a laboratory day, we have used the longer period in the expectation that the absorption and response might be more uniform. In practice, the extracts were administered during the afternoon and the results were available early the following morning.

SUMMARY

By the use of decalso, permutit, and darco as adsorbents, vita- min K1 has been isolated in a practically pure state. Crystalliza- tion, or distillation and crystallization, has given the pure vitamin.

Pure vitamin K1 has been converted by reductive acetylation into the diacetate of dihydro vitamin K1; m.p. 62-63”; potency 500 units per mg. By means of the Grignard reaction the diacetate has been converted back to vitamin K1 with 100 per cent en- hancement of potency, 1000 units per mg.

We wish to acknowledge financial assistance from the Theelin Fund administered by the Committee on Grants for Research of St. Louis University. Moreover, we are indebted to Mr. Joseph Yglesias for efficient assistance in the fractionation experiments and to Mr. Patrick Pomphrey for his cheerful help with the assays and to Mr. H. C. Schaefer of the Purina Mills who has advised us on several points connected with this investigation.

Addendum (July Sl, 19S9)-After this manuscript had been accepted for publication and after our “Communications to the Editor” (Journal of

the American Chemical Society, May and June; McKee et al. (1939); Binkley et al. (1939)) had been published, a short report by Karrer and Geiger (1939) on vitamin K from alfalfa appeared in the July 1 number of Helvetica

chimica acta. Most of the data confirm some of our previously published

by guest on May 24, 2018

http://ww

w.jbc.org/

Dow

nloaded from

Isolation of Vitamin K,

observations, but in addition, a study of the oxidation-reduction potential is reported. Apparently, Professor Karrer’s data did not permit him to reach a decision on the nature of the quinone nucleus in vitamin K1. Our ultraviolet absorption studies and data on the degradation products and the biological activities of 1,4-naphthoquinone and 1,4-naphthoquinones substituted in the 2 or 2 and 3 positions enable us to say with certainty that vitamin K1 is a 2,3-disubstituted naphthoquinone.

Our recent report (MacCorquodale et al.,’ 1939) on the sensitivity of vitamin K to visible light, and our communication (McKee et al., 1939) issued on May 5, in which evidence was cited in support of our belief. that vitamins K1 and Kz are quinones, had apparently not come to the attention of Professor Karrer before his paper was submitted for publication.

BIBLIOGRAPHY

Almquist, H. J., J. BioZ. Chem., 114, 241 (1936); 120,635 (1937). Almquist, H. J., and Klose, A. A., J. Am. Chem. SOL, 81,745 (1939). Ansbacher, S., Science, 88,221 (1938). Binkley, S. B., MacCorquodale, D. W., Cheney, L. C., Thayer, S. A.,

McKee, R. W., and Doisy, E. A., J. Am. Chem. Sot., 61, 1612 (1939). Dam, H., Biochem. Z., 216,475 (1929); 220, 159 (1930). Dam, H., Geiger, A., Glavind, J., Karrer, P., Karrer, W., Rothschild, E.,

and Salomon, H., Helu. chim. acta, 22, 310 (1939). Dam, H., Glavind, J., Lewis, L., and Tage-Hansen, E., Skand. Arch.

Physiol., 79, 121 (1938). Dam, H., and Lewis, L., Biochem. J., 31,17 (1937). Dam, H., and Schanheyder, F., Biochem. J., 30,897 (1936). Karrer, P., and Geiger, A., HeZv. chim. acta, 22, 945 (1939). MacCorquodale, D. W., Binkley, S. B., McKee, R. W., Thayer, S. A., and

Doisy, E. A., Proc. Sot. Exp. BioZ. and Med., 40, 482 (1939). McKee, R. W., Binkley, S. B., MacCorquodale, D. W., Thayer, S. A., and

Doisy, E. A., J. Am. Chem. Sot., 61,1295 (1939). Smith, J. H. C., J. BioZ. Chem., 96, 35 (1932). Thayer, S. A., Cheney, L. C., Binkley, S. B., MacCorquodale, D. W., and

Doisy, E. A., .I. Am. Chem. Sot., 61, 1932 (1939). Thayer, S. A., MacCorquodale, D. W., Binkley, S. B., and Doisy, E. A.,

Science, 88, 243 (1938). Thayer, S. A., McKee, R. W., Binkley, S. B., MacCorquodale, D. W., and

Doisy, E. A., Proc. Sot. Exp. BioZ. and Med., 40, 478 (1939).

by guest on May 24, 2018

http://ww

w.jbc.org/

Dow

nloaded from

A. Thayer and Edward A. DoisyS. B. Binkley, D. W. MacCorquodale, Sidney

1THE ISOLATION OF VITAMIN K

1939, 130:219-234.J. Biol. Chem. 

  http://www.jbc.org/content/130/1/219.citation

Access the most updated version of this article at

 Alerts:

  When a correction for this article is posted• 

When this article is cited• 

alerts to choose from all of JBC's e-mailClick here

  tml#ref-list-1

http://www.jbc.org/content/130/1/219.citation.full.haccessed free atThis article cites 0 references, 0 of which can be

by guest on May 24, 2018

http://ww

w.jbc.org/

Dow

nloaded from