Anthocyanins and Anthoxanthins

ANTHOCYANINS AND ANTHOXANTHINS. The term anthocyanin has been employed by botanists to denote all the water-soluble pigments of flowers and blossoms, but with new knowledge of the chemical nature of these colouring matters a more exact classification has become possible and the expression is now applied to particular substances which are responsible for the red, blue, mauve, purple and violet colours of flowers. The anthoxanthins are a chemically related class of water-soluble yellow or orange pigments possessing relatively feeble tinctorial power. The bright yellow, orange and green colours are usually due to the presence of one or other of the plastid pigments (xanthophyll, carotin, chlorophyll etc.), insoluble in watery media but soluble in fats and oils. These do not occur, as do the antho cyanins, in the cell-sap, and there is no visible chemical link con necting the plastid and anthocyanin pigments.

One of the most interesting and obvious properties of the anthocyanins is the change of colour which they exhibit when treated with acids or alkalis. It is said that Yorkshire children turn bluebells red by placing them in ant-hills; the formic acid produced by the insects is responsible for this curious result. In 1664 Robert Boyle wrote "Take good Syrrup of Violets, Impraeg nated with the Tincture of the flowers, drop a little of it upon a White paper . . . and on this Liquor let fall two or three drops of Spirit either of Salt or Vinegar, or almost any other eminently Acid Liquor and upon the Mixture of these you shall find the Syrrup immediately turn'd Red.... But to improve the Experi ment, let me add what has not been hitherto observ'd, namely, that if instead of Spirit of Salt, or that of Vinegar, you drop upon the Syrrup of Violets a little Oyl of Tartar per Deliquiurn or the like quantity of Solution of Potashes, and rubb them to gether with your finger, you shall find the Blew Colour of the Syrrup turn'd in a moment into a perfect green." It is probable that the green colour resulted from the produc tion of a bright yellow colour by the action of the alkali on an anthoxanthin in solution, the anthocyanin retaining its blue colour. Many colourless flowers, for example jasmine and Antirrhinum, develop bright yellow colorations in ammonia. A typical antho cyanin is bluish-red in acid solution, violet in neutral solution and blue in alkaline solution. blue cornflower, the bordeaux red cornflower, the deep red dahlia, and the red rose contain one and the same anthocyanin, the variation in colour being simply due to the different degrees of acidity and alkalinity of the cell sap. More than one anthocyanin may be present in a flower or blossom, and the colours of many flowers are due to the presence of both anthocyanins and plastid pigments in the tissues. Yellow wallflowers contain a plastid pigment and an anthoxanthin that contributes very little to the total tinctorial effect ; the different shades of red wallflowers are due to varying proportions of antho cyanin and plastid colouring matters. Moreover, very small changes, botanically, in varieties or species may be associated with the development of different anthocyanins.

In 1905, Molisch demonstrated the existence of anthocyanin crystals in the living plant, and showed that crystals of these pig ments could be readily prepared on a small scale by simple meth ods. For example, petals of the scarlet pelargonium are flattened and bruised on a glass surface, covered with a few drops of 75% acetic acid and then with a cover-glass and the whole placed under a clock-glass to ensure slow evaporation. At the edges of the cover-glass or round the petals, crystals gradually maketheir appearance. Grafe, a few years later, carried cut this experiment on a larger scale and for the first tame isolated an anthocyanin pigment in quantity and in a tolerably pure condition. Our knowl edge of the chemistry of the anthocyanins is, however, chiefly due to Willstatter who, in a series of masterly researches, opened and all but completed a new chapter of organic chemistry. In 1913, Willstatter and Everest published an account of an investigation of the pigment of the blue cornflower. They found that the colour ing matter "cyanin" exists in the plant as its blue potassium salt, but that the substance can also combine with acids to red salts and that advantage may be taken of this property in the isola tion of the pure substance. In this case the work involved was difficult, the dried blue cornflower petals contained only 0.75% of their weight of colouring matter. It was later found that dried, deep red dahlia petals contain 2o% of their weight of the same colouring matter and about 5o% of this may be readily isolated in the following manner. The fresh flowers are extracted with acetic acid, methyl-alcoholic hydrochloric acid and 1 volumes of ether being then added. This precipitates the chloride, which is insoluble in ether, and the salt may then be separated and re crystallized from a solution in 7% hydrochloric acid. It should be added that this process is unusually simple and straightforward owing to the high percentage of colouring matter in the flowers. The work of the florists in developing garden flowers rich in pig ment content has been extraordinarily successful.

Cyanin chloride has the composition C2;H31010C1 and, like all the other anthocyanins yet examined, contains sugar in a combined form. This is readily detached by the action of boiling 20% hydro chloric acid, leaving a salt, termed cyanidin chloride, which tinc torially and in many other properties closely resembles cyanin chloride. The sugar so detached is glucose, C,H,208, each molecule of cyanin chloride giving rise to one of cyanidin chloride and two of glucose : Cyanin chloride is thus a diglucoside of cyanidin chloride. Meco cyanin chloride from the poppy (Papaver rhoeas, purple scarlet variety) has the same composition as cyanin chloride and, like it, is degraded to cyanidin chloride and glucose (2 molecules). The solution of cyanin in aqueous sodium carbonate is blue, whereas that of mecocyanin is violet, and there are other diver gencies. The explanation must be sought in the different mode of the molecular attachment of the cyanidin and glucose com plexes in the two substances. Chrysanthemin, the pigment of the deep red garden chrysanthemum, resembles mecocyanin, but its molecule gives rise to only one molecule of glucose to each mole cule of a cyanidin salt ; it is monoglucosidic. One of the .pig ments of the aster very closely resembles chrysanthemin, whilst the colouring matters of the sweet cherry and of the sloe are apparently allied to mecocyanin, but contain rhamnose as one of the sugar components. Very careful treatment of mecocyanin with hydrochloric acid causes the loss of only one glucose molecule, and the result is chrysanthemin chloride. It will be observed that sev eral anthocyanins may be regarded as derived from cyanidin and that their differences may be traced to the varying nature of the sugars, to the number of sugar molecules attached to one of c} anidin, and to the position of such attachment.

The further work of Willstatter and his colleagues brought to light the curious fact that, despite the existence in nature of a range of colour unrivalled by art, the number of fundamental sugar-free pigments of the type of cyanidin chloride, termed anthocyanidins, is very limited.

Glucosides derived from pelargonidin chloride,

have been found in many plants, for example in Pelargonium zonale, the purple-red summer aster, scarlet-red dahlia, the scarlet Salvia, and in the skins of radishes. Glucosides and complex glucosides derived from delphinidin chloride, occur in the wild purple larkspur (Delphinium consolida) and in the viola. The anthocyanin of the peony, peonin, is a diglucoside of peonidin, which is simply related to cyanidin and may be converted into it. Peonidin chloride has the composition that is, CH2 more than cyanidin chloride, and the action of hydriodic acid on peonidin, furnishes and cyanidin iodide. Similar derivatives of delphinidin chloride are obtained by decomposition of the an thocyanins from the bilberry, wild mallow, grape-skins, the petunia and Primula hirsuta. These all contain one or more groups instead of the —OH groups of delphinidin chloride. Apart from these cases, in which the molecules are slightly modified, there are but three known fundamental anthocyanidins, and these in turn have been found to be built upon one and the same molecular plan. Before considering this, however, it is necessary to pay some attention to the anthoxanthins, the main features of the chemistry of which were well established before any real in sight into the nature of the anthocyanins had been obtained. The anthoxanthins are also glucosides, diglucosides or rhamnogluco sides, although owing to the processes of extraction and isolation adopted, they are frequently decomposed and obtained in the sugar-free condition. These substances are more widely distributed in different plants and in different parts of plants than are the anthocyanins. They have a limited but definite utility as mordant dyestuffs, giving in many cases bright, fast dyeings owing to their capacity for the formation of lakes with certain metallic oxides. A typical anthoxanthin is quercitrin, which is found in quercitron bark, an extract of which was introduced by Bancroft as a colouring matter in 1775. This bark is derived from a species of oak, Quercus velutina, a native of the middle and southern United States. The colouring matter is deposited in the inner bark and may be obtained by extraction with dilute ammonia. On boil ing the solution with hydrochloric acid the quercitrin is hydrated and the products are quercetin, Cir>111007, and the sugar, rhamnose, Ce11.05. Quercetin is a bright yellow, feebly acid substance ; it combines with strong acids to form orange salts, but these are not very stable and are readily dissociated by water. Nevertheless, this property of salt formation in a non-nitrogenous substance is a weaker manifestation of the basic character exhibited by the anthocyanins, and it is due to a structural peculiarity common to both types. Quercetin is a strong polygenetic mordant dyestuff. On wool, with a chromium mordant it gives a reddish-brown shade, with an aluminium mordant a brownish-orange shade, with a tin mordant a bright orange shade and with an iron mordant an olive-black shade. Many representatives of this group have been isolated and closely investigated by A. G. Perkin, St. von Kos tanecki and others. As with the anthocyanins, different glucosides yield the same fundamental colouring matter, and moreover the —OH groups are sometimes replaced by Derivatives of quercetin have been obtained from rue, capers, buckwheat, clover flowers, Eucalyptus macrorhyncha, cotton flowers, onion skins, Persian berries, wallflowers, tea leaves, may blossom, horse chest nut, bark of the apple tree, and also from wings of certain insects and many other sources. Although significance must be attached to this remarkably wide distribution of quercetin, it is to be noted that nature provides more variety among the anthoxanthins than among the anthocyanidins. They are already a numerous clan and fresh representatives are constantly being discovered.

Among the better known of these colouring matters the follow ing may be mentioned, and it must be understood that in all cases the substances occur in the plant in combination with sugars. Chrysin, C15H1004, is contained in the leaf buds of the poplar ; apigenin, in the leaves, stem and seeds of parsley, and also in camomile flowers; galangin, C15H1005, in galanga root; luteolin, C15H1008, in weld. This is the dried herbaceous plant, Reseda luteola, an extract of which formed the oldest known European dyestuff, said to have been used by nations north of the Alps in the time of Julius Caesar. Weld gives a beautiful and fast yellow on silk mordanted with alumina and still finds a limited application in the dyeing of certain materials used in military uniforms. Kaempferol, also C15H1008, occurs in Delphinium con solida and a third isomeride, fisetin, is the colouring matter of the wood of Rhus cotinus which comes into commerce as "young fustic." Old fustic is the wood of the tree Chlorophora tinctoria and contains morin, C15H1007, which is the most important of the natural dyestuffs of this group. Myricetin, is found in the box-myrtle of China and in numerous other plants; its iso merides, gossypetin and quercetagetin, are the colouring matters of cotton flowers and the African marigold respectively.

Constitutional Relationships.

The results of the investiga tions made with the object of unravelling the molecular structure of these substances have demonstrated that they are all based on a single type, and actually the molecules of each of these colour ing matters can be constructed from those of a substance called flavone, C15H1002, by replacing one or more hydrogen atoms by hydroxyl (—OH) groups. This fact was recognized in some of the above cases at a time when flavone itself was unknown, and on account of the interest attaching to the simplest representa tive of an important group of naturally occurring compounds, Kostanecki prepared it artificially in 1898. Muller in 1915 made the remarkable discovery that the "meal" or "farina" of Primula pulverulenta and P. japonica consisted largely of flavone identical with the synthetic preparation of Kostanecki. The atoms in the molecule of flavone are connected in the manner shown in the annexed figure and the various anthoxanthins are substances the molecules of which may be imagined to be derived from those of flavone by replacing a number of the hydrogen atoms (H) by —OH groups, sometimes by groups and then attaching sugar molecules through the —OH groups.

Two of the rings in the flavone molecule are benzene rings and the third, containing an oxygen atom, is a pyrone ring. (The pre cise mode of linkage between carbon atoms is irrelevant to the present discussion.) The various positions in the flavone nucleus are denoted by numerals, and the position of hydroxyl groups in some of the above-mentioned anthoxanthins is as follows :—Chry sin, 5 :7; apigenin, 4':5 :7; galangin, 3 :5 :7; luteolin, 3' :4':5 : 7 kaempferol, 4':3 :5:7; fisetin, 3':4':3 :7; quercetin, 3':4':3 :5:7; morin, 2':4':3 :5 : 7 ; myricetin, 3' : 4' : 5' :3 :5 :7. The nine representatives just cited and several others have been synthesized by methods such as to leave no doubt in regard to their molecular constitution, and in all cases the natural and synthetic products have been found by careful comparison to be identical. The occur rence of a hydroxyl group in position 3 has a considerable influ ence on the properties of the colouring matters ; those that contain it are called "flavonols" and are usually characterized by more powerful tinctorial properties and greater strength as bases than isomerides not hydroxylated in this position. It has been found that the sugar molecules are usually attached through oxygen atoms in positions 3 or 7. Thus quercetin and quercitrin (see above) have the respective constitutional formulae if the con ventional hexagonal symbols are used for rings: These were determined by modifying the free —OH groups in quercitrin, subsequently detaching the sugar group and detecting the situation of the free —OH groups thus produced. A com parison of the composition of the flavones and flavonols with that of the anthocyanidins discloses a simple relation. Cyanidin chloride, Ci5H1,O6C1, is just HCl more than luteolin, C15H1008. Cyanidin chloride is not, however, luteolin hydrochloride, although the relation between the substances is very close, and on treat ment with hot aqueous alkalis, cyanidin and luteolin give identical, fission products. These have the benzene rings intact so that cyanidin and luteolin, if the former is based on the flavone ring system, can differ only in the pyrone ring portion of the molecular structure. In accordance with modern conceptions, the formation of ammonium chloride from ammonia and hydrogen chloride involves the passage of the positively charged hydrogen atom (proton) to the ammonia forming a positively charged complex. This leaves a negatively charged chlorine atom and the process +— -I may be represented by the scheme : and HCI--qNH, 1 C1. Similarly, the pyrone salts are formed by the proton of the acid passing to an oxygen atom, and we can formulate luteolin hydro chloride on the probable assumption that the proton combines with oxygen in position 4.

anthocyanin production in passing from one garden variety to another are echoed by changes in the anthoxanthins.

Furthermore the range of anthocyanidins is so much more re stricted than that of anthoxanthins. In favour of the hypothesis may be counted the circumstance that, so far as we know at pres ent, the groups such as and sugar groups that are attached to the fundamental nuclei occupy corresponding positions in the anthoxanthins and anthocyanins. We may compare, for example, the formulae of peonidin chloride, the anthocyanidin from the peony and isorhamnetin from yellow wallflowers. Both have been synthesized.

The constitution figured for cyanidin chloride was suggested by, rather than deduced from, these facts, but it has been con clusively proved to be correct by synthesis in several different ways. Pelargonidin chloride is similarly constructed, but lacks the —OH group in position 3', whilst delphinidin chloride is cyan idin chloride with an additional —OH group in position 5'. Pelar gonidin chloride and delphinidin chloride have also been artificially prepared. The formulae of the three anthocyanidins (chlorides) are given below.

The sugar is frequently attached to position 3 in both antho xanthins and anthocyanins, but we have still much to learn about this matter, and agreement has not yet been reached in regard to the point of attachment of the sugar residues in cyanin chloride and in pelargonin chloride. It is generally recognized that the diglucosides in both series contain disaccharide units. That is, the attachment involves only one hydroxyl group of the anthocyanidin. One anthocyanin has been artificially prepared, namely, callistephin from the aster, and in this substance the sugar is beyond question in position 3.

Evidently if we were able to remove the oxygen in position 4 in presence of HC1 it would be possible to convert the flavonols into the anthocyanidins as shown in the equation and thus the close relationship of the individuals in the two groups would be established. Actually the transformation is most difficult to effect, and although numerous workers claim to have been successful, the experimental control has not always been sat isfactory. The authentic example is furnished by the work of Willstatter and Mallison, who converted quercetin into cyanidin chloride by means of magnesium in methyl-alcoholic aqueous hydrochloric acid solution. The yield was very poor, since most of the quercetin underwent a different transformation. Nature may have at command a more effective method of reduction, anthocyanins being produced in the plant by deoxygenation (re duction) of the anthoxanthins, considered to be the primary prod ucts. But against this hypothesis, it seems equally probable that the anthocyanins and anthoxanthins represent end-products ob tained by divergent processes from a common parent. It is the exception rather than the rule to find the most closely related anthocyanins and anthoxanthins occurring together in a plant. Thus Delphiniums consolida gives kaempferol and a delphinidin derivative, and there is no evidence that the facile changes in Some characteristic properties of the anthocyanins are the fol lowing: Chemical Reactions. —Amyl Alcohol—Dilute Hydrochloric Acid Distribution Ratio.—The anthocyanidins, the monogluco sidic anthocyanins, and the diglucosidic anthocyanins can be roughly distinguished by their behaviour in presence of a mixture of amyl alcohol and very dilute hydrochloric acid. The antho cyanidins pass completely to the amyl alcohol layer; the digluco sidic anthocyanins remain largely in the aqueous layer, unless one of the sugar groups is rhamnose, when the behaviour tends to be monoglucosidic ; the monoglucosides distribute themselves more evenly between the two layers.

Ferric Chloride Reaction.—Anthocyanins or anthocyanidins with free OH groups in positions 3' and 4' (e.g., cyanidin, delphin idin) give an intense blue colouration in alcoholic solution on the addition of ferric chloride. The colour becomes violet on the addition of water. The fact that all the anthocyanins derived from cyanidin exhibit this reaction proves that the sugar is not attached to positions 3' and 4' in any of them.

Colour Bases and Pseudo-bases.

In most cases the addition of sodium acetate to an aqueous solution of an anthocyanin (idin) produces a violet or purple precipitate of a colour-base. Again, the addition of water to an alcoholic solution of cyanidin chloride causes the slow separation of the colour base in a form exhibiting a characteristic green reflex. Addition of acid repro duces the cyanidin salt. These colour-bases have been more closely investigated in other related series and have a quinonoid nature. They absorb water, giving colourless compounds also convertible by acids to the anthocyanin salts, and these so-called pseudo-bases are often formed on the great dilution of the aque ous or alcoholic solutions of the anthocyanin salts. Decolorization of the solutions of natural pigments and subsequent reappearance of the colour was a puzzling phenomenon encountered by some of the earlier workers in the group.

It remains to be added that anthocyanins and anthoxanthins have a close relation to certain colourless plant constituents, and among these catechin occupies an important position. Freuden berg has shown that both quercetin and cyanidin on treatment with hydrogen and a catalyst yield one of the naturally occurring catechins.

In the course of his experiments on catechin, Freudenberg has had occasion to observe the wandering of the right-hand benzene ring from position 2 to position 3. It is therefore interesting to notice that genistein, an anthoxanthin of dyers broom, has been definitely proved to have the annexed constitution.

Finally, it may be remarked that carajurin, a crystalline con stituent of "carajura," is related to the anthocyanin group. Cara jura, a bright red pigment, is prepared by the Indians of the Rio Meta and Orinoco for use as a flesh paint, since it is a very effec tive colour.