TEXAS AGRICULTURAL EXPERIMENT STATION A. B. CONNER, Director College Station, Texas BULLETIN NO. 650 _ JUNE 1944 STUDIES ON TOXIC SUBSTANCES OF LOCOWEEDS, ASTRAGALUS EARLEI A N D O T H E R S G. S. FRAPS and S. H. WENDER Division of Chemistry AGRICULTURAL ANiD MECHANICAL COLLEGE OF TEXAS GIBB GILCHRIST, President D-19-744-1500 [Blank Page in Original Bulletin] The concentrated toxic preparation of the loco weed contains several closely related toxic substances. The compounds precipi- tated by silicotungstic acid are toxic but do not produce all the symptoms of locoism. The spectrum absorption curve of the puri- fied compound so precipitated resembles closely the spectrum ab- sorption curve produced from the compounds not precipitated. When treated with benzoyl chloride, both the benzoylated and non-benzoylated products contained toxic compounds. Adsorption of the picrates on aluminum oxide separated two or more com- pounds. Silver nitrate precipitated a compound having the char- acteristic spectrum absorption curve of the loco compounds. Ex- traction with iso-amyl alcohol and purification with flavianic acid removed impurities. Every active fraction showed the same gen- eral type of spectrum absorption curve, though there were differ- ences in details. Evidence of chemical changes in the toxic sub- stance separated occurred in the solutions after storage even in cold storage, as well as decided losses in the process of separation. The different compounds may be present in the loco weed or some may have been formed during the process of separation. Absorption spectra ................................................................................................. .'...1 Discussion ............................................................................... ....................... ., ......... .. q A Acknowledgment ................ ..................................................................................... .. j CONTENTS " Introduction ............................................................................................................... .- Review of literature ................................................................................................. .- Preparation of concentrated locoine extract with phosphotungstic acid .... Purification of locoine p-hosphotungstate with alcohol .................................. .. a Purification by two precipitations with phosphotungstic acid and fuller’s earth ............................................................................................. .. p: Toxicity tests of fraction 2 PP on cats ............................................................ .. Purification with iso-amyl alcohol ....................................................................... Precipitate with silicotungstic acid .................................................................... Not precipitated with silicotungstic acid .......................................................... ..1 Separation of picrates by chromatographic adsorption ................................ Separation with phenol .......................................................................................... ..1 Benzoylation with benzoyl chloride ................................................................... Portion not benzoylated with benzoyl chloride ..... ......................................... ..1 Silver nitrate separation ...................................................................................... ..1 Flavianic acid separation ...................................................................................... ..1‘ Chromatograph separation of picrates .............................................................. ..1 Stability of locoine to acids and alkali .... ............................................... ....... ..1 Tests on cats for physiological activity ............................................................ ..1 Loco poison in several varieties of loco weeds ................................... ......... ..1 Summary ............................. -.; ...................................... ........................................... .. ‘X References .................................................................................................. .; .............. .. w STUDIES ON TOXIC SUBSTANCES OF LOCOWEEDS, ASTRAGALUS EARLEI AND OTHERS By G. S. Fraps, Chief, Division of Chemistry, and S. H. Wender,* Associate Chemist The ingestion of locoweeds by grazing livestock has continued to re- sult in losses among the animals. Many investigations have been under- taken both in the field and in the laboratory to learn more about the disease, about the weed, and about the toxic substances present in the weed. In a previous bulletin of this Station, No. 537, Fraps and Carlyle (4) described the isolation of a highly toxic concentrate, called ‘locoine” from a locoweed, Astragalus Earlei. The work was continued by A. W. Walde, Joseph Semb, and S. H. Wender. The present bulletin reports further progress made in the isolation of the active constituent or constituents of locoweed. It also reports the results of additional studies made in frac- tionating the locoine concentrate and in investigating the new fractions thus obtained. Three of these new fractions have been found to be toxic to cats; two of these three produced symptoms of locoism. The most concentrated extract previously reported as capable of pro- ducing locoism in cats has now been separated into at least four main fractions, and further separations and investigations have been made on these new substances. The fractions described in this bulletin are more nearly pure and are more potent than those previously reported. The pres- ence in locoweed extracts of several physiologically active substances, in- stead of only one, -has again been experimentally demonstrated to be probable. Spectroscopic studies in the ultra-violet have been undertaken on the various substances found, and of especial value have been the absorption spectrum curves obtained for these new fractions. Since every toxic frac- tion thus far studied has a characteristic absorption in the ultra-violet range, the spectrograph method of studying these substances has proved to be a great aid in following the desired compound after experimental separations. Chromatographic adsorption methods and the use of solvent extraction and selective precipitation methods also have been of much value in making the new fractionations.. REVIEW OF LITERATURE Much of the literature on locoweed investigations has been reviewed in the publications of Fraps and Carlyle (4), of Pease and Elderfield (11), and in the earlier work by Crawford (3). The results of other investiga- tions, as in the field of veterinary science, are found in government publi- cations (17), and in a thorough study of locoism in animals by Mathews (19). The first printed records of locoweed disease apparently were in the report of the Commissioner of Agriculture, U. S. Dept. of Agriculture, for October, 1873. A description of the symptoms of locoed animals was *Now Assistant Professor of Chemistry, University of Kentucky, Lexington, Kentucky. ¢ 6 BULLETIN NO. 650, TEXAS AGRICULTURAL EXPERIMENT STATION given in detail in this and subsequent reports. The very early work was. done by Prescott in 1878 (13) and by Power and Cambier (12) in 1891. Crawford (3) found barium in locoweed plants. He reported that barium compounds gave symptoms of poisoning similar to loco poisoning but this was not confirmed by Marsh (7) or Alsberg and Black (1). Attempts to isolate the toxic principle of the locoweed have encoun- tered many difficulties. Couch (12) isolated a toxic fraction from the locoweed, Oxytropis Lambertii, but not in crystalline form. The solution prepared from the precipitate with silicotungstic acid was non-toxic to cats. The filtrate after treatment with basic lead acetate and charcoal yielded a light yellow, amorphous mass, of honeycomb structure, very soluble in water, extremely hygroscopic, and insoluble in ether, in chloro- form, and in the hydrocarbon solvents. Couch believed his compound not to be of ester, ether, or of glucosidal nature. It was not basic, but nitro- genous and highly hydroxylated. It did not affect a ray of polarized light. Couch was unable to find any precipitant for the substance. Fraps and_ Carlyle (4) isolated a toxic principle from the locoweed, Astragalus earlei, in highly concentrated and toxic form, but not in a defi- nitely pure state. This toxic principle, called “locoine” was shown to be a strong base, very soluble in water and in alcohol, but only slightly soluble in ether, petroleum ether, and in chloroform. This “locoine” was found to be precipitated by phosphotungstic acid and recovered from the precipitate by treatment with barium hydroxide. The base was acetylated, forming an acetylacetate. This “locoine” was found to be stable towards reagents and apparently not affected by boiling with dilute acids or alkalis. The most concentrated preparation required a feeding of only one gram over a ‘ period of fifty days to loco a cat. The toxic solution was found to give reactions with many of the reagents used to test for alkaloids. Pease and Elderfield (11) have isolated two non-toxic substances from the locoweed, Astragalus earlei, which they call alpha-earleine and beta- earleine. They were obtained after a separation involving chromatograph- - ing a solution of their picrates on aluminum oxide. Stempel and Elder- field (14) report that alpha-earleine is identical with betaine, and beta- earleine with choline. Pease, Reider, and Elderfield (10) report the iso- lation of the non-toxic sugar, d-pinite (monomethyl ether of inosite) from Astragalus earlei and from Oxytropis lamberiti. Stempel and Elderfield (15) isolated from Astragalus earlei a non-toxic dihydroxyvalerolactone, or an isomer thereof, along with glycerine. They found that Reinecke salt p precipitates a highly active fraction, from which a crystalline substance was isolated. The solution obtained on decomposition of the Reineckate produced a typical locoism in cats, but they were unable to demonstrate whether this compound was responsible for such symptoms, or whether a still un-isolated substance carried down in the precipitate was the active material. Knowles and Elderfield (6) from Astragalus wootoni have isolated- pinite, betaine, choline, and trigonelline, but have not yet isolated the active constituent or constituents. In tests on cats, no appreciable difference in _. symptoms was found in locoism caused by Astragalus earlei and in loco- - ism caused by Astragalus wootoni. This conclusion is in agreement with the earlier work on these weeds by Fraps and Carlyle (4). STUDIES ON TOXIC SUBSTANCES OF LOCOWEEDS, AND OTHERS _ 7 Although there are many seleniferous “locoWeeds” (16), it has been pointed out by Fraps and Carlyle (4) and by Pease and Elderfield (11) that in the locoweeds they studied, the concentrations of selenium are so low that selenium is definitely not the toxic material. PREPARATION OF CONCENTRATED LOCOINE EXTRACTS WITH PHOSPHOTUNGSTIC ACID (METHOD A) The method of preparation of the concentrated locoine extract is based on the one used by Fraps and Carlyle (1) with some additions and modifi- cations. One-half kilogram of the dry, finely-ground locoweed, Astragalus earlei, was placed in a large evaporating dish, 1500 ml. of 95% ethyl al- cohol added, and the alcohol and weed mixed thoroughly. After standing for at least 30 minutes, with thorough mixing in the meantime, the liquid was pressed out by means of a screw tincture press. The residue was returned to the dish, and this time extracted with 750 ml. alcohol. The residue was discarded, but the combined alcohol extracts were filtered, and then concentrated under reduced pressure to about 150 ml. This concen- trated liquid was set aside, until the alcohol extracts from at least five kilograms of weed were on hand. Extractions of the weed itself were carried out in units of one-half kilogram. After combination and concentration of these extracts, how- ever, solutions representing five kilograms of locoweed were used as units. To the combined alcohol concentrates from 5 kg. of locoweed, an equal amount of water was added, and the solution concentrated under reduced pressure. After cooling, the remaining solution, about 1.5 liters, was de- canted from the water-insoluble part, and treated with a slight excess of basic lead acetate. The resulting precipitate was filtered off and discard- ed. Excess lead was then removed from the filtrate with hydrogen sul- fide, and the lead sulfide discarded. The lead-free filtrate was concen- trated in vacuo to a thick syrup, and the residue was extracted with 500 ml. of commercial absolute alcohol, while thoroughly stirring and warming the solution. The solution was allowed to cool slowly at first, then to re- main over-night in the refrigerator. The supernatant alcohol liquid extract was decanted, and the residue was again extracted, this time with 250 ml. of absolute alcohol. After cooling, the two absolute alcohol supernatant liquid extracts were combined, and concentrated in vacuo to about 250 ml. This concentrated solution was allowed to remain for several days in the refrigerator. Water was added to the filtrate, and the alcohol distilled off under reduced pressure. More water was added, and the solution concen- trated to about 150 ml. The concentrated, aqueous solution was cooled, and in an ice bath, a 24% solution of phosphotungstic acid (P2O5.24WO3.42H2O) in 4% sulfuric acid was added until precipitation was complete. The phosphotungstic acid precipitate was transferred toa centrifuge bottle and extracted with 200 ml. of absolute alcohol. After centrifuging, the supernatant liquid, which had been shown to contain the toxic material, was decanted, and an equal amount of water added. Then the alcohol-soluble precipitate was decom- posed in the cold with saturated barium hydroxide solution. A slight ex- 8 BULLETIN NO. 650, TEXAS AGRICULTURAL EXPERIMENT STATION cess of barium hydroxide solution was used. The insoluble material was’ centrifuged off, and the excess of barium in the residual solution was re- moved by precipitation with dilute sulfuric acid. After removing the pre-e cipitated barium sulfate, the filtrate was concentrated in vacuo to about? 150 ml. The spectrum absorption curves of one of these preparations be-=- fore and after purification with iso-amyl alcohol are shown in Figure 1. - _ DENSITY __L l 1 J I l L I '~ 245 2Z5 28s 30s s55’ k rzcuaa 1, ‘ ’ WAVE LENGTH l Figure 1. Spectrum curve of locoine concentrate pre- pared by precipitation with phosphotung- stic acid and solution of the phosphotung- state in alcohol, Curve A before, and after, B, purification with iso-amyl alcohol. PURIFICATION OF LOCOINE PHOSPHOTUNGSTATE A WITH ALCOHOL Fraps and Semb (5) made many experiments to obtain crystalline phosphotungstates instead of the gummy and resinous precipitate usually produced. No definite method was found to produce clean, crystalline pre- cipitates. Fairly good results seemed to be obtained after the crude _ “locoine” solutions had been first considerably purified. Attempts at re- crystallizations of “locoine phosphotungstate” from its alcohol or acetone» solutions by addition of a number of solvents did not give favorable re- j; sults. The use of barium acetate to decompose the phosphotungstates in- ' stead of barium hydroxide was found to be unsatisfactory. Carrying out 1 _ the precipitation with phosphotungstic acid solution in the cold was afl i‘ preferable way to a precipitation at room temperature. In every method I used for the preparation of the “phosphotungstate,” losses were found to be. i fairly large, g » Stempel and Elderfield (15) have also recently had difficulties with the phosphotungstic acid precipitation, and believe that the precipitation 7i ‘ of the toxic constituent with phosphotungstic acid may possibly be ex- plained by adsorption of the poison on the rather bulky precipitate. STUDIES ON TOXIC SUBSTANCES OF LOCOWEEDS, AND OTHERS Absolute alcohol was found to be a suitable solvent for extracting the “locoine phosphotungstate” from the other phosphotungstates formed on precipitation. Part of the remaining alcohol-insoluble phosphotungstates was soluble in acetone. When the alcohol-soluble “phosphotungstate” was dried thoroughly, it became difficultly soluble in absolute alcohol, but dis- solved easily on addition of acetone. The solution obtained by Fraps and Semb after decomposition of the alcohol-soluble phosphotungstate when fed to cats produced definite signs of locoism in six cats after one or two months of feeding. Two cats were fed preparations from the phosphotungstate which was insoluble in alcohol but soluble in acetone for more than 3 months but failed to show definite signs of being locoed. PURIFICATION BY TWO PRECIPITANTS WITH PH-OSPHOTUNGSTIC ACID AND FULLER’S EARTH A crude locoweed fraction was prepared by two precipitations, as fol- lows: After the first precipitation of the “locoine” solution with phos- photungstic acid, the absolute alcohol soluble part of the precipitate was decomposed with barium hydroxide, treated with basic lead acetate, fil- tered, and the excess lead removed, as in the method of Fraps and Carlyle (4). Then the filtrate was reprecipitated with phosphotungstic acid solu- tion. This second precipitate was extracted with absolute alcohol, water added, and the alcohol-soluble portion of the phosphotungstic acid precipi- tate decomposed with barium hydroxide solution. The neutralized, concen- trated, aqueous solution resulting from the filtrate was called 2PP (twice precipitated by phosphotungstic acid). Trichloracetic acid was added and the aqueous solution extracted with iso-amyl alcohol, then with ether, which removed most of the color. The purified aqueous portion had in its absorption spectrum in the ultra-violet an extremely sharp maximum at 265 mu. and a minimum at 245 mu, see Figure 4. This is one of the sharpest maxima obtained on any locoweed extract. , Measured portions of the fraction 2PP, after having been adjusted with dilute sulfuric acid or barium hydroxide to pH values of 1.5, 3.8, 7.5, and 9.8, were each treated separately with weighed portions of fuller’s earth previously washed with distilled water. More of the greenish-yellow color remained in the solution which had been adjusted to pH 9.8 than in the ones which had been adjusted to the other pH values. The supernatant liquids were neutralized, and the absorption spectra determined. The curves of the solutions at pH 1.5 and pH 3.8 gave a maximum at 265 mu and a minimum at 245 mu. The curve of the solu- tion at pH 7.5 gradually rose to a slight maximum at 255 mu and fell to a minimum at 240-245 mu. The one at pH 9.8 failed to exhibit a maximum or minimum. Elution of fuller’s earth with dilute barium hydroxide after its treat- ment with a solution of 2PP at pH 1.5 produced a solution which had a maximum at 365 mu and a minimum at 245 mu. ‘Toxicity Tests of Fraction 2PP on Cats Five cats fed on various solutions prepared from material adsorbed on fuller’s earth ‘were locoed after about two to two and one-‘half months of 10 BULLETIN NO. e50, TEXAS AGRICULTURAL EXPERIMENT STATION feeding. Four cats fed on solutions prepared from fractions not adsorbed 0n fuller’s earth gave varying results. One showed definite symptoms of locoism; another appeared to be locoed; but two failed to exhibit definite signs of locoism. Diagram 1—Separati0n of locoine concentrate preparations. P-1 Alcohol-soluble phosphotungstate, locoed cats——Fig. 1 P-2 Alcohol-insoluble phosphotungstate, soluble in acetone, did not loco cats P Precipitated with silicotungstic acid, from alcohol-soluble phospho- tungstate—pFig. 2, SP Slightly physiologically active but not locoine—Fig. 3 - SF. Not precipitated by silicotungstic acid——Curve A A. Extracted by phenol B. Not extracted by phenol Chromatographed on aluminum oxide with picric acid A. Colorless percolate B. Yellow percolate C. Adsorbed, soluble in Water D. Adsorbed, not soluble in water NB. Not benzoylated with benzoyl chloride, see diagram 2. Produced symptoms of locoism. Fig. 2, NB- A E. Benzoylated, ether soluble BP. Separated from solution on partial evaporation (BP), and purified (BPS), not locoine. BF. Slightly soluble in ether. BFRS. Ethlgg soluble saponified and base extracted BF . BFRS. Chromatographed in alcohol solution BFRS-W-Adsorbed—-Fig. 3 BFRS-A-Not adsorbed——Fig. 4 BFRS. Chromatographed with picric acid. A. Adsorbed B. Not adsorbed Diagram 2——Separation of unbenzoylated Fraction NB. NB. Chromatographed through aluminum oxide NBW. Adsorbed-eluated with water NBA. Not adsorbed—Symptoms of locoine when fed to cats. Fig. 2 Treatment of NBA with silver nitrate. Filtrate—no maximum or minimum in absorption spectrum curve. Precipitate—locoine spectrum curve. Treatment of NBA with flavianic acid. Precipitate—no maxima or minima in absorption spectrum curve. Filtrate—-chromatographed through aluminum oxide. ' Adsorbed—purified——absorption spectrum maximum 625- minimum 243 Not adsorbed—purified—absorption spectrum maximum 270, minimum 243 Treatment of NBA with picric acid and chromatographed Not adsorbed Adsorbed, yellow zone Purification With Iso-Amyl Alcohol When the locoine solution was made slightly acid and extracted with iso-amyl alcohol and then with ether, substances were removed which were not locoine. The purified solution had given a sharper spectrum curve (See Fig. 1). This method of purification was used on a number of preparations. STUDIES ON TOXIC SUBSTANCES OF LOCOWEEDS, AND OTHERS i’ _ PRECIPITATE WITH SILICOTUNGSTIC ACID When the concentrated locoine solution secured in Method A was treated with silicotungstic acid, a portion was precipitated but most of the organic matter remained in solution. The fraction precipitated by this reagent was toxic but did not produce all the symptoms of loco disease. The fraction not precipitated was toxic and produced the usual symptoms. The concentrated, aqueous solution was separated into two fractions by precipitation with a 25% solution of silicotungstic acid (Merck SiO2.12WO3.26H2O) in 4% sulfuric acid. A grayish, curdy precipitate was formed, rapidly becoming resinous-like and dark. This precipitate was centrifuged off and washed with cold water. Additional water was added to the washed precipitate, then saturated barium hydroxide solution was added in slight excess and the mixture shaken thoroughly. Mechanical breaking up of the precipitate was sometimes necessary. The excess of barium was removed with dilute sulfuric acid and the neutralized aqueous solution was concentrated in vacuo. This solution was treated again with silicotungstic acid solution. This time, a much smaller amount of sticky material accompanied the precipitate. After decomposition with barium hydroxide in the usual manner, after neutralization and concentration, a third precipitation was made with silicotungstic acid solution. This pre- cipitate, when kept cold, did not become resinous-like, nor much darker. This precipitate was washed with cold water, then decomposed with barium hydroxide. The resulting solution was neutralized and concentrated. This solution was called SP (silicotungstic acid precipitate). Some was used for toxicity tests on cats, and proved to be slightly physiologically active, 1.00 - 0.80 _ SP 0.60 .- >< F1 _ 1A ‘S 0.40 - 0.20 _ FIGURE 2 L l I \ 1 Jil- | \- 245 265 285 305 WAVE LENGTH Figure 2. Spectrum curve, SP, of locoine preparation which was precipitated with silicotungstic acid, purified, toxic to cats, did not produce all symptoms of locoism. NBA—Not pre- cipitated with silicotungstic acid, not ben- zoylated, not adsorbed with aluminum ox- ide, locoed cats. 12 BULLETIN NO. 650, TEXAS AGRICULTURAL EXPERIMENT STATION though not producing many of the typical symptoms of locoism. The ab-fi sorption spectrum. is shown in Figure 2 and is similar to that for the pre-i vious preparation (Figure 1). ‘ Not Precipitated With Silicotungstic Acid The filtrate containing the material not precipitated by the silicotung-l; stic acid solution, and still containing an active principle, was treated with p; a saturated barium hydroxide solution until no more precipitation occurred, _ even on long standing. The precipitate was filtered off and discarded, and the excess of barium removed with dilute sulfuric acid. The aqueous so- lution was concentrated to about 100 ml. under reduced pressure. A drop’ of this solution was tested to see if any precipitate was still formed with the silicotungstic acid solution; If so, the silicotungstic acid precipitation I was repeated. When the concentrated aqueous solution no longer gave any precipitation with silicotungstic acid solution, it was used for further sepa- rations. This solution was called SF (silicotungstic acid filtrate). A Alkaloid Tests on the Fraction SF. Alkaloid tests on the preparation _. SF were made according to the methods of the A. O. A. C. Silicotungstic acid solution, ferric chloride solution, Wagner’s reagent, Mayer’s reagent, - Kraut’s reagent, Marme’s reagent, mercuric chloride solution, potassium‘ iodide solution, picric acid solution, and potassium ferrocyanide solution all failed to give any precipitate. Phosphoric acid solution produced a‘ heavy amorphous precipitate. Reinecke’s solution was decolorized, as was . also a 2% potassium permanganate solution. ' An alcoholic solution of SF yielded a trace of precipitate with alcoholic g mercuric chloride solution, and a precipitate with alcoholic tannic acid ‘ solution, but no precipitate with alcoholic picrolonic acid solution. SEPARATION ‘OF PICRATES BY CHROMATOGRAPHIC ADSORPTION An absolute alcohol solution of SF was treated with picric acid in ab- ' solute alcohol, and chromatographed through a column of aluminum oxide. i The upper half of the column was bright yellow in color, and the lower} half, pale yellow. Four fractions were obtained. (A) A colorless alcohol . filtrate. (AA) A yellow alcohol filtrate and alcohol washings. (B) A pale yellow substance adsorbed and removed from the column by elution with j Water. The aqueous solution had a brownish-red tingle. (C) A brightl yellow fraction remaining after washing with alcohol and elution with water = was removed from the column with dilute barium hyroxide solution. Each A solution was then treated in aqueous solution with sulfuric acid and ben- . zene to remove picric acid, then with 150 ml. portions of iso-amyl alcohol, then ether, and finally neutralized and concentrated. The absorption spec- trum curves are given in Figure 4. The curve of the alcohol filtrate had a sharp maximum at 264 mu and no minimum. The curve of the aqueous filtrate had a sharp maximum at 262 mu and a sharp minimum at 240 mu; while that of the barium hydroxide eluate had an extremely sharp maxi-i mum at 256 mu. Another alcoholic solution of SF without picric acid was first passed through aluminum oxide and divided into the alcohol filtrate A and the STUDIES ON TOXIC SUBSTANCES OF LOCOWEEDS, AND OTHERS 1f} 1.00 k A DENS ITY FIGURE 3 | 1 | 1 I l l l %l__ 245 2b5 255 295 WAVE LENGTH Figure 3. Spectrum curves of compounds not precipi- tated by silicotungstic acid, benzoylated, not adsorbed by aluminum oxide, A. The same but adsorbed by aluminum oxide, W. water eluate W. Each of these solutions, transferred to absolute alcohol, was separately treated with alcoholic picric acid solution and separately chromatographéd. Both fractions from the chromatographed “picrates” of A showed a maximum at 268-270 mu. The alcohol filtrate A had a maxi- mum of 264 mu, and the water eluate W had a maximum at 268-270 mu and a minimum at 237 mu. Evidently more than one compound was present. Separation With Phenol Most of the solution SF was removed on extraction with liquified phenol. The remaining aqueous phase, after removal of traces of phenol and purification with iso-amyl alcohol, showed on its absorption spectrum curve in the ultra-violet a broad maximum at 269-283 mu and a major minimum at 245 mu. Also present were a very weak maximum at 297- 306 mu and a weak minimum at 286 mu. Of all the newer fractions studied and tested, this solution was the only one to show the extra maximum and minimum. This extra maximum and minimum, however, has been found in some crude locoweed extracts before they had been separated into the newer fractions. ' The phenol extract, after transfer to aqueous solution, gave a sharp maximum at 268-270 mu. Thus at least one substance, maximum 268-270 mu, was soluble in the phenol and had been separated from the solution by means of the extraction. BENZOYLATION WITH BENZOYL CHLORIDE Previous work showed (4) that the locoine base is acetylated with acetyl chloride. The losses in the acetylation and subsequent saponifica- tion were high, and for this reason benzoyl chloride was tried. Several g odor during the treatment. 14 BULLETIN NO. 650, TEXAS AGRICULTURAL EXPERIMENT STATION fractions were secured by the benzoyl treatment of fraction SF. One e fraction was not benzoylated and not soluble in ether but contained a toxic‘; compound. The benzoylated ether-soluble fraction was divided into two more fractions, one highly soluble in ether, and the other only slightly i‘ The latter fraction did not appear to be poisonous. See soluble in ether. Diagram 1. An excess of 5% sodium hydroxide solution was added to the silico- tungstic acid filtrate, SF; then benzoyl chloride was added. The contents were thoroughly shaken until there was no benzoyl chloride present. A greenish-yellow, foamy precipitate was formed. The precipitate was dis-- The solution was then thoroughly extracted with ether i solved in ether. and with chloroform. The ether extracts were combined and thoroughly washed with dis- The first washings were slightly colored; the later ones were, The ether solution, now deep yellow in color, was set aside to. evaporate. A white substance, difficult to redissolve in ether, was formed when the solution was almost entirely evaporated. The precipitate was This whites Crystals of BP, be-i tilled water. colorless. washed free of oily and resinous-like material with cold ether. precipitate was called BP (“Benzoate” precipitate). fore recrystallization, gave a melting point of 151°-4°C. On recrystalliza- tion from alcohol, the melting point was 178—80°C.' The crystals were long, white, interlacing needles. They were insoluble in water, difficulty; ....i.\a.w...-..~..-1... p‘ .~ soluble in ether and in cold absolute alcohol, but soluble in hot absolute‘ ‘l alcohol. The ether solution from which crystals of BP had separated, was al- e lowed to evaporate still further. petroleum ether extracts yielded color and soluble in ether. due). The crystals of, BP and the ether-soluble residue BFR were each sepa- BP was found to be extremely rately saponified with sodium hydroxide. difficult to saponify completely. BFR produced an extremely penetrating acidified with dilute sulfuric acid. Benzoic acid was precipitated in the case of BFR, but none was observed in the case of- BP. The latter failed to produce any noticeable precipitate on acidification after its treatment with sodium hydroxide. After thorough extraction of each solution with ether, the aqueous phases were each separately neutralized with bariuml hydroxide, and the barium sulfate discarded. The neutralized solutions‘ An oily and sticky residue was left. It" was extracted with cold and then with slightly warm petroleum ether. The another white precipitate, which later‘ studies indicated to be impure benzoic acid. After the petroleum ether extractions a resinous-like material was left which was dark brown in" This was called BFR (benzoate filtrate resi- Each saponified solution was cooled, and ‘then; i were then evaporated to dryness, and the residues thoroughly extracted with absolute alcohol. Absorption spectrum studies in the ultra-violet on a neutral aqueou A solution obtained after the saponification of BP failed to show the maxi mum or minimum which indicates locoine poison. An aqueous, neutral so lution called BFRS obtained after the saponification of BFR showed - STUDIES ON TOXIC SUBSTANCES OF LOCOWEEDS, AND OTHERS 15 maximum at 264 mu and a minimum at 243 mu, before purification. So- lution BFRS was used for toxicity tests on cats and found to be definitely toxic. Chromatographic Adsorption Analysis of BFRS. The absolute alcohol solution of BFRS (saponified solution of BFR) was chromatographed through a column of aluminum oxide (c. p. Baker’s Analyzed, ignited powder). The absolute alcohol filtrate was pale greenish-yellow in color and was called BFRS-A. The spectrum curve is shown in Figure 3. A 1.20 - 1.00 "’ 0.80 - F IGURE 4 l l l 4l_ 2440 ' zéo l aim 3w 4N wm: mnsru l Figure 4. Spectrum curve of fractions separated from preparation SF not treated with benzoyl chloride, by chromatograph of the picrate on aluminum oxide. A, first colorless fil- trate. B, eluted with water. C, eluted by barium hydroxide. brownish ring of material was adsorbed at the top of the column. After thorough washing of the column with absolute alcohol, which failed to remove the adsorbed matter, water was used for the elution. The brown ring now passed into the aqueous filtrate, which was called BFRS-W. The spectrum curve is shown in Figure 3. The two curves are decidedly dif- ferent and show the presence of two different compounds. Chromatograph of the Picrate. A solution of BFRS, which had not been chromatographed previously, showed on its absorption spectrum curves in the ultra-violet a maximum at 264 mu and a minimum at 243 mu. The absolute alcohol solution of this was treated with alcoholic picric acid solution and chromatographed through a column of aluminum oxide. A small ring faintly pink was adsorbed at the top of the column. The remainder of the column was yellow in color. The water eluate of the column, after the usual purifications, had a maximum at 268-270 mu. The l _and boiled with activated charcoal. 16 BULLETIN N0. 650, TEXAS AGRICULTURAL EXPERIMENT STATION column was next eluted with a dilute sulfuric acid solution. The acid re- moved all noticeable color from the column. The acid eluate, after the usual purifications, showed a maximum at 268-270 mu, and a minimum at about 243 mu. Portion Not Benzoylated With Benzoyl Chloride The solution SF, after its treatment with benzoyl chloride, was puri- fied from benzoylated material by extraction with ether, and with chloro- form. The alkaline aqueous solution was then acidified with dilute sul- furic acid. The resulting aqueous solution, as well as the benzoic acid precipitate, were thoroughly extracted with ether. On evaporation of the ether, impure benzoic acid was obtained. The aqueous solution, now acid, was" also extracted with chloroform, but practically no material was re- moved. The remaining acid solution was next neutralized with barium hydroxide solution, and the precipitate of barium sulfate discarded. The neutralized aqueous solution, called NB (not a “benzoate”) was evapo- g rated to dryness, and the residue thrice extracted by commercial absolute alcohol, and filtered. An absolute alcohol solution of NB was chromatographed through a column of aluminum oxide. Most of the color and solid matter of the solution passed through the column and into the filtrate, now to be called NB-A. A small brownish layer was adsorbed near the top of the column. After the column had been thoroughly washed with absolute alcohol, the brownish matter was eluted with water and was called NB-W. On long standing, the alcohol filtrate NB-A deposited a crystalline substance NB-AP, which was not identified or its toxicity tested. The supernatant liquid N B-A was decanted, transferred to an aqueous solution, and the alcohol removed. This aqueous solution was used for toxicity tests on cats, and proved to be highly physiologically active on cats. Its spectrum curve is given in Figure 2. Silver Nitrate Separation On treatment of a solution of N B-A in absolute alcohol with an abso- lute alcohol solution of silver nitrate, in subdued light, a white precipitate was formed. The precipitate was washed with absolute alcohol, suspend- ed in water, and decomposed with hydrogen sulfide. After filtering 01f the lead sulfide, the filtrate was treated with activated charcoal, and the resulting, almost colorless solution showed a flattened maximum at 255- 265 mu. The filtrate from NB-A, after precipitation with silver nitrate, was treated with hydrogen sulfide, the filtrate transferred to aqueous solution, No definite maximum or minimum_ appeared in its absorption spectrum in the ultra-violet, so it did not belong in the locoine group. Thus alcoholic silver nitrate is one of the very few precipitants found for a toxic fraction of locoweed extract, when the latter is highly purified. Flavianic Acid Separation An alcoholic solution of NB-A produced a reddish-orange precipitate on addition of an absolute alcohol solution of flavianic acid (2, 4 di-nitro- S'I‘UDII