2054 TUTIN : THE CONSTITUTION OF ERIODICTYOL,CCXVI.--The Constitution of Eriodictyol, of Homoerio-dictyol, and of Hesperitin.By FRANK TUTIN.IN a paper communicated to the meeting of the AmericanPharmaceutical Association, held at Indianapolis, Ind., in Septem-ber, 1906, Dr. F. B. Power and the present author described theisolation of two crystalline substances of phenolic nature from theleaves of Eriodictyon Calif ornicum (Hooker and Amott), Greene(Proc. Amer. Pharm. Assoc., 1906, 54,352). These two compoundswere designated eriodictyol and homoeriodictyol respectively, theformer having been proved to possess the formula C15H1206, whilstthe composition of the latter was shown to be Cl6H,,O,.I n it subsequent communication (Trans., 1907, 91, 887) it witsnoted that there are certain similarities in the properties ofhomoeriodictyol and its isomeride, hesperitin, which suggested thatthese two compounds were structurally related.Experiments sup-ported this view, for, whilst hesperitin yields isoferulic acid(3-hydroxy-4-methoxycinnclmic acid) aad phloroglucinol on hydrolysis(Tiemann and Will, Ber., 1881, 14, 970), homoeriodictyol, wheOF HOMOERIODICTYOL, AND OF HESPERITIN. 2055similarly treated, gave the same phenol together with ferulic acid(4-hydroxy-3-methoxycinnamic acid).Tiemann and Will (Zoc. cit.) assigned to hesperitin the con-stitutional formula (I). This formula received support through theOHMeO(-)CH:CH*CO*O/-\ \-/HO- OH(1.1work of A. G. Perkin '(Trans., 1898, 73, 1037), who prepared acetyl-hesperitin, and recorded results which indicated that this compoundwas a triacetyl derivative.It appeared 'therefore from the results of the hydrolysis experi-ments mentioned above that homoeriodictyol differed fromhesperitin on5 in the relative positions of the hydroxyl and methoxylgroups in the catechol part of the molecule.When, however,acetylhomoeriodictyol was prepared, it was found to contain fouracetyl groups. It was therefore concluded that homoeriodictyolmust be represented by formula (11).HOHO/-\CH:CH -CO/-\OH \-/ \-/Me0 HO(11.)From this it would appear that homoeriodictyol was not 80similar to hesperitin in structure as had at first been concluded.Nevertheless, Dr. Power and the present author were so convincedof the near relationship of these two compounds that they venturedto suggest that the formula hitherto assigned to hesperitin isincorrect, notwithstanding the statement of Perkin (Zoc.cit.) thatthe latter yields only a triacetyl derivative. Formula (111) wastherefore put forward for hesperitin.HC)M~O/-\CH:CH .c.<-)oH.HO-\-/HO(111.)With regard to the constitution of eriodictyol, the amount ofmaterial which was at first available did not permit of many experi-ments being conducted with this substance, but the view wasexpressed that it was the parent compound of which hesperitin andhomoeriodictyol are monomethyl ethers (formula IV).HOHO<-)CH:CH~CO<-)OHHO- H 2056 TUTIN : THE CONSTl'IUTION OF BRIODIC'I'YOL,Shortly after the appearance of the first paper by Power andTutin on eriodictyon leaves (Zoc. cit.), a communication on the samesubject was published by G.Mossler (Annalen, 1907, 351, 233).This author recorded the isolation of a substance possessing theformula C,,H,,O,, designated " eriodictyonon," which was evidentlyidentical with homoeriodictyol. Mossler, however, did not succeedin isolating any eriodictyol.After having published the account of their work on the con-stitution of homoeriodictyol, Dr. Power and the present authorreceived from Dr. Mossler a reprint of a paper communicated byhim t>o the Academy of Sciences in Vienna (Sitzungsb er. K . A kad.Wiss. Wien, 1907, 116, ii, June, 1907). In this communicationMossler, who was unaware of the more recent work of the above-mentioned authors, admits that his '' eriodictyonon " is identicalwith homoeriodictyol, and sets forth the conclusion that this sub-stance is represented by one of the following formulae:orThis last publication by Mossler was replied t o by Dr.Powerand the present author (Proc., 1907, 23, 243), when it was pointedout that neither of the formulae proposed by Mossler could becorrect, since compounds possessing such a structure could not yieldphloroglucinol.One statement made by Mossler, however, was in direct conflictwith the views which the present author, in conjunction withDr. Power, had expressed regarding the constitution of homo-eriodictyol, namely, that the substance in question waa opticallyactive.The last-mentioned authors were unable to confirm this,and, since the correctness of their conclusions regarding homo-eriodictyol have now been fully proved, it is evident that the abovestatement of Mossler must have been based on an incorrectobservation.It appeared to the present author that there was one possiblealternative to the formula which had been suggested by him inconjunction with Dr. Power for homoeriodictyol, but which was notat all probable, namely, a structure related t o the second formulaproposed by Mossler, as follows OF HOMOERTODICTYOL, AND OF HESPERITIN. 2057A substance possessing such a formula might conceivably yield, onhydrolysis, phloroglucinol and ferulic acid by the addition of twomolecules of water, followed by the elimination of one such molecule,although such a change appeared highly improbable.It was con-sidered advisable, therefore, in order conclusively to prove theconstitution of homoeriodictyol and related compounds, to haverecourse to synthetical experiments.I f the formula suggested by Dr. Power and the present authorfor eriodictyol, homoeriodictyol, and hesperitin be correct, thenthese substances are 2 : 4 : 6-trihydroxyphenyl 3 : 4-dihydroxystyry.Zketone, 2 : 4 : 6-trihydroxyphenyl 4-hydroxy-3-methoxystyryl ketone,and 2 : 4 : 6-trihydroxyphenyl 3-hydroxy-4-met7toxystyryl ketonerespectively, It waa decided therefore to methylate the firstrmentioned three substances, and to compare the fully methylatedproducts with synthetically prepared 2 : 4 : 6-trimethoxyphenyl3 : 4-dimethoxystyryl ketone.The results of the methylation oferiodictyol, homoeriodictyol, and hesperitin are recorded in thepresent paper, and it is shown that each of them yields 2 : 4 : 6-tri-methoxyphenyl 3 : 4-dimethoxystyryl ketone and 2-hydroxy-4 : 6-di-methoxyphenyl 3 : 4-dimethoxystyryl ketone identical in all respectswith these substances as prepared synthetically (see the followingThe correctness of the formulx suggested by Power and Tutinfor eriodictyol, homoeriodictyol, and hesperitin is therefore provedbeyond question.Naringenin, a hydrolytic product of the glucoside, naringin, wasshown by Will (Ber., 1885, 18, 1311) to be related to hesperitin.On heating with aqueous potassium hydroxide, it undergoeshydrolysis in a manner similar to the latter compound, and yieldsphloroglucinol and p-hydroxycinnamic acid.Will therefore con-cluded (Bey., 1887,20, 297) that naringenin was the phloroglucinylester of the above-mentioned acid. I n view of t'he results recordedin the present paper concerning hesperitin, there can be no doubtthat naringenin is also a ketone, namely, 2 : 4 : 6-trihydroxyphenyl4-h ydroxys t yry 1 ketone.I n a previous communication (Power and Tutin, Trans., loc. cit.)a monomethyl ether of homoeriodictyol was described. A largerquantity of this substance has now been prepared, and it has beenproved to be 2 : 6-dihydroxy-4-methoxyphenyl 4-hydroxy-3-methoxy-styryl ketone. Similarly, when one methyl group is introduced intoeriodictyol, it takes up the 4-position in the phenyl radicle, theproduct being a new isomeride of homoeriodictyol and hesperitin,namely, 2 : 6-dihydroxy - 4 - methoxyphenyl 3 : 4 - dihydroxystyrylketone.paper).VOL. XCVII.6 2058 TUTIN : THE CONSTITUTION OF ERIODICTY OL,The observation of Perkin (Zoc. cit.) regarding the anomalouscharacter of the sodium derivative of hesperitin has been confirmed,this substance appearing to have the formula CI6H&Na,C,6H&.On the other hand, the statement made by Perkin that the productof the action of acetic anhydride on hesperitin is a triacetylderivative cannot be confirmed, it having been proved that thesubstance thus formed is tetra-acetylhesperitin.EXPERIMENTAL.Eriodictyol (2 : 4 ; 6-Trihydroxyphenyl 3 ; 4DihydroxystyrytKetone).A quantity (5 grams) of eriodictyol * was dissolved in alcohol, andan excess of methyl sulphate added, after which a concentratedalcoholic solution of potassium hydroxide was allowed to flow intothe hot liquid at such a rate that the mixture was kept gentlyboiling.The liquid at first showed a tendency to darken, owingto the absorption of oxygen, but this soon ceased as methylationproceeded. Finally, the mixture assumed a dark red colour onthe addition of the alkali, which only slowly disappeared. A furtherquantity of methyl sulphate was added, followed by more alkali,after which the mixture was kept for twenty minutes and thenpoured into water. The yellow product precipitated by this treat-ment was extracted by means of chloroform, the solution beingwashed, dried, and the solvent removed.The residue thus obtainedwas boiled with successive portions of dilute, aqueous potassiumhydroxide so long as the decanted alkaline liquid was yellow incolour, after which the material insoluble in the alkali was washedand dissolved in alcohol. On inoculating the solution thus obtainedwith synthetic 2 : 4 : 6-trimethoxyphenyl 3 : 4-dimethoxystyrylketone (see the following paper), crystallisation rapidly ensued.The product so obtained was identical in all respects with thesynthetical compound just mentioned. It crystallised in stout, paleyellow prisms, and, when dried in the air, melted at 85O, but inthe anhydrous state at 117'5O :0.1747 t gave 0.4279 CO, and 0.1046 H20.The potassium hydroxide extracts which had been decanted fromthe crude 2 : 4 : 6-trimethoxyphenyl 3 : 4-dimethoxystyryl ketonewere acidified, and the precipitated yellow product was crystallisedfrom alcohol, in which it was rather sparingly soluble.Deep yellow* For an inipioved method of isolating eriodictyol, homoeriodictyol, and otherphenolic substances fmm Eriodictyoa leaves, compare Tutin and Clewer, Trans.,1909, 95, 81.C = 66.8; H= 6.4.CmH2,06 requires C = 67.0 ; H = 6.1 per cent.t Anhydrous substanceOF HOMOERIODICTYOL, AND OF HESPERITIN. 2059leaflets were thus obtained, which melted a t 154O, and were identicalwith the 2-hydroxy-4 : 6-dimethoxyphenyl 3 : 4-dimethoxystyrylketone described in the following paper :0.1279 gave 0.3099 CO, and 0.0680 HiO.CISH,,O, requires 66.3; H =5*8 per cent.~OnornethyteTiodictyol (2 : 6-Dihydroxy-4-rnet?~oxyp?~enyl 3 : 4-di-hydroxystyryl ketone).-Four grams of eriodictyol were dissolvedin absolute alcohol, and to this solution was added one and a-halfmolecular proportions of methyl sulphate, which had just previouslybeen washed with aqueous sodium carbonate and dried.Slightlymore than the equivalent amount of sodium, dissolved in absolutealcohol, was then gradually introduced into the hot mixture. Afterremoving t'he al-cohol, the residue was dissolved in ether, washedwith water, and then fractionally extracted by shaking with suc-cessive portions of an aqueous solution of sodium carbonate. Thefirst few extractions removed only unchanged eriodictyol, whichformed the greater part of the product., but on acidifying thealkaline liquids subsequently obtained, a yellow product separated,which partly crystallised on keeping.This wits collected, wellwashed with alcohol, and then recrystallised from this solvent, inwhich it was but sparingly soluble. Almost colourless needles werethus obtained, which melted at 2 1 5 O :C = 66.1 ; H * 5.9.0.1284 gave 0.2995 CO, and 0'0583 H,O. C = 63.6 ; H = 5.0.Cl6HI4O6 requires C = 63'5 ; H = 4.6 per cent.This substance was therefore a, monomet?&yZeriodictyoE, and sinceit is not identical with either homoeriodictyol (2 : 4 : 6-trihydroxy-phenyl 4-hydroxy-3-methoxystyryl ketone) or hesperitin (2 : 4 : 6-tri-hydroxyphenyl 3-hydroxy-4-methoxystyryl ketone), and the hydroxylgroup in the 2(0r 6)-position is known to be difficult of methylation,it must be 2 : 6-dih~d~oxy-4-methoxyp?~enyl 3 : 4-dihydroxystyrylketone.Monomethyleriodictyol dissolves in aqueous alkali hydroxides,giving at first a practically colourless solution, but after aboutthirty seconds the liquid suddenly becomes totally bIack. Onscetylation, monomethyleriodictyol yields a tetra-acetyl derivative,which forms colourless needles, melting at 159O.Homoeriodictyol (2 : 4 : 6-T~ihydroxyphenyZ 4-HgcZroxg-3-rnet hoxy-.stpry$ Ifetone).The methylation of homoeriodictyo1 by means of methyl suIphateand potassium hydroxide proceeded analogously t o that oferiodictyol, with the exception that there was no tendency to absorboxygen, and consequently a cleaner product was obtained.The6 T 2060 TUTIN : THE CONSTITUTION OF ERIODICTYOLmethylated material wits examined as above-described, whea itreadily yielded 2 : 4 : 6-trimethoxyphenyl 3 : 4-dimethoxystyrylketone (m. p. 85O when air-dried; 117'5O when anhydrous), and2-hydroxy-4 : 6-dimethoxyphenyl 3 : 4-dimethoxystyryl ketone (m. p.154O).ik4onomethylhornoeriodictyol (2 : 6-DihycEroxy-4-methoxyphenyl4-hydroxy-3-met hoxystyryl ketone). - Monomethylhomoeriodictyol,prepared by the action of methyl iodide on the crystalline sodiumderivative of hoqmoeriodictyol, was previously described by Powerand Tutin (Trans., 1907,91,895). A larger amount of this producthas now been prepared by heating the above-mentioned sodiumderivative with methyl iodide and methyl alcohol.The productthus obtained was dissolved in ether and freed from unchangedhomoeriodictyol by extraction with dilute, aqueous sodiumcarbonate, after which the monomethylhomoeriodictyol was removedby shaking with a concentrated solution of this alkali. The productso obtained crystallised readily from alcohol in hard, yellow, wart-like masses, which melted at 142O.On boiling monomethylhomoeriodictyol for several hours with30 per cent. aqueous potassium hydroxide, hydrolysis occurred atthe double linking, after which vanillin was readily isolated fromthe reaction mixture. It is evident from this, and considerationapreviously given, that the ONa group in the sodium derivative ofhomoeriodictyol, which is converted into methoxyl on treatmentwith methyl iodide, must occupy the $-position in the phenyl groupof the molecule.Monomethylhomoeriodictyol is therefore 2 : 6-cti-hydroxy-4-methoxyphenyl 4-hydroxy-3-methoxystyryl ketone.Hesperitin ( 2 : 4 : 6-Trihydroxyphenyl 3-Hydroxy-4-methoxystyrylKetone.)Hesperitin, which is obtained by the hydrolysis of the glucoside,hesperidin, a constituent of the peel of the orange, lemon, and otherrelated fruits,, has been stated by Tiemann and Will (Ber., 1881,14, 970) to be the phloroglucinyl ester of isoferulic acid. As statedin the introductory portion of this paper, however, the accuracy ofthis conclusion was doubted by Power and Tutin, and the presentauthor has therefore further investigated the question.Hesperitin, as obtained from Schuchardt, was recrystallised fromethyl acetate, when it melted at 224O, but when mixed with homo-eriodictyol, fusion occurred a t 200O.The material so obtained,however, did not agree in its characters with t,he description ofhesperitin as given by A. G. Perkin (Trans., 1898, 73, 1037).Thus it formed pale yellow plates, which could not be distinguisheOF HOMOERIODICTYOL, AND OF HESPERITIX. 2061by inspection from crystals of homoeriodictyol, it was practicallytasteless,* and it dissolved in alkalis with a bright yellow colour.Perkin (Zoc. cit.), on the other hand, has described hesperitin ascrystallising in almost colourless needles, possessing an intenselysweet taste, and dissolving in alkalis with, a t the most, a faintlyyellow colour.Nevertheless, the identity of the material employedby the present author with hesperitin cannot be doubted, inasmuchas the melting points of the compound itself and its acetyl derivativeare in agreement with the corresponding constants given by Perkinfor hesperitin and its acetyl derivative. Moreover, the substanceyielded the abnormal sodium derivative, C,6H,306Na,Cl,Hl,0,,characteristic of hesperitin. (Found, Na= 3.8. Calc., Na= 3.8 percent.)Methylation of Hesperitbz.A quantity (1.5 grams) of hesperitin was methylated by means ofpotassium hydroxide and methyl sulphate in the manner previouslydescribed, when the reaction appeared to proceed precisely as inthe case of homoeriodictyol.A good yield of product was obtained,which was readily separated into the two compounds similarlyprepared from eriodictyol and its homologue, namely, 2-hydroxy-4 : 6-dimethoxyphenyl 3 : 4-dimethoxystyryl ketone (m. p. 154) and2 : 4 : 6-trimethoxyphenyl 3 : 4-dimethoxystyryl ketone (m. p. 8 5 Owhen air-dried.; 117’5O when anhydrous).Tetra-acetyZ~esperiti~.-Half a gram of hesperitin was boiled forthree hours with a considerable excess of acetic anhyaridc, afterwhich the greater part of the solvent was removed and the mixturediluted with ether. After several hours, a crystalline substanceseparated in tufts of colourless prisms, which melted at 120O. Afterrecrystallisation from alcohol, this substance melted at 1 2 7 O , andwas evidently identical with the compound similarly prepared byPerkin (Zoc. cit.), which he regarded as a triacetyl derivative. Thenumber of acetyl groups in the compound were estimated asfollows. A quantity of the substance was hydrolysed with dilutepotassium hydroxide, the mixture then acidified with sulphuric acid,and the acetic acid removed by a current of steam and titrated :0.2616 gave acetic acid equivalent to 0*0900 NaOH. CO-CH,= 36.9.Cl,Hl,0,(CO*CH3)1 requires CO*CH3 = 36.6 per cent.It is evident therefore t<hat this compound was tetra-acetyl-hesperitin, and not a triacetyl derivative. This conclusion is inharmony with the properties of the substance, for it was insoluble* Although hesperitin, when in the solid state, possesses no appreciable taste, itsalcoholic solution is distinctly sweet2062 TUTIN AND CATON : SYNTHESIS OF 2 : 4 : 6-TRIMETHOXY-in cold dilute sodium hydroxide, which would not have been the casehad it contained a hydroxyl group, all the groups of this naturepresent in hesperit,in having phenolic properties.THE WELLCOME CHEMICAL RESFARCH LABORATOR~ES,LONDON, E.C