Nierenstein [Footnote: Ber., 1910, 43, 628.] also succeeded in converting tannin into carboethoxytannin, the latter on saponification yielding crystalline, inactive digallic acid. On acetylating pentacetyl leucotannin with acetyl chloride a hexacetyl derivative (M.P. 159° C.) is obtained, the strychnine salt of which is resolved into both of the active components. This proves the presence of digallic acid and leucotannin in tannin lev. pur. Schering investigated by Nierenstein. The latter author [Footnote: Liebig's Ann., 1912, 386, 318; 388, 223.] later considered tannin to be polydigalloylleucodigallic acid anhydride and the simplest tannin to be a digalloylleucodigallic acid anhydride. This view, however, would not stand subsequent criticisms, being in disagreement with the earlier observations of molecular weight and acidic properties of tannin. Manning [Footnote: Ibid., 1912, 34, 918.] believed to have isolated a pentethylester of the pentagalloyl glucoside from tannin, but this was shown to be the ethyl ester of gallic acid.
Feist [Footnote: Ber., 1912, 45, 1493.] had arrived at the conclusion that tannin was a glucose compound, and maintained that tannin from Turkish galls was a compound of glucogallic acid combined as an ester with 2 molecules gallic acid. But Fischer and Strauss [Footnote: Ibid., 1912, 45, 3773.] synthetically prepared a glucoside of gallic acid exhibiting differences from Feist's preparation which were so great that the latter no longer could be considered a single glucoside of gallic acid.
Fischer and Freudenberg [Footnote: Ibid., 1912, 45, 2717; 1913, 46, 1127.] subsequently elaborated a method of purifying tannin, and on investigating the purified substance, arrived at the conclusion that no other hydroxybenzoic acid than gallic acid was present in tannin. On repeating Strecker's hydrolysis they obtained 7-8 per cent, sugar, and hence concluded that 1 molecule of glucose was combined with about 10 molecules of gallic acid. Owing to the difficulty of isolating the intermediary hydrolysis products, and the subsequent impossibility of drawing any conclusions as to the constitution of tannin, the latter investigators decided to adopt the methods offered by synthesis. Their basic idea was the absence of carboxylic groups in tannin, and that hence the total gallic acid must be present in ester form. These conditions are fulfilled if one views tannin as being an ester compound of 1 molecule of glucose and 5 molecules of digallic acid, of similar construction as, for example, pentacetyl glucose. Fischer and Freudenberg succeeded in preparing the former by shaking a mixture of finely powdered glucose, chloroform, and quinoline with an excess of tricarbomethoxygalloyl chloride for twenty-four hours and precipitating the resulting product with methyl alcohol; suitably purified, a light amorphous colourless substance was obtained which proved to be penta-(tricarbomethoxygalloyl) glucose. Careful saponification with excess alkali in acetone-aqueous solution at room temperature yielded a tannin very closely resembling tannin, identified as pentagalloyl glucose. It is doubtful, however, whether this substance is homogeneous, and it is probably a mixture of two stereoisomers.
Fischer and Freudenberg, therefore, further concluded that tannin is mainly an ester compound of glucose and 5 molecules m-digallic acid. Elucidation on this point offered itself advantageously in Herzwig's methylotannin, [Footnote: Ber., 1905, 38, 989.] which is obtained by the interaction of diazomethane and tannin. The first step was then to prepare pentamethyl-m-digallic acid
CH_3.O_______ ______COOH CH_3.O{_______}—CO.O—{______} CH_3.O CH_3.O O.CH_3
from trimethylgalloyl chloride and the m-p-dimethyl ether of gallic acid; the chloride of this substance, coupled with [Greek: a]- and [Greek: b]-glucose, yields—
_CH.OR
| |
| CH.OR H_______O.CH_3
| | R=CO{_______}O.CH_3
O{ CH.OR H O
| | H_____O.CH_3
| CH CO{_____}O.CH_3
| | H O.CH_3
|_CH.OR
CH_2.OR
[Illustration: Penta-(pentamethyl-m-digalloyl)-glucose.]
The [Greek: a]- and [Greek: b]-derivatives thus obtained differ in their behaviour towards polarised light, and are, again, probably mixtures of two stereoisomers, i.e., mixtures of derivatives of [Greek: a]- and [Greek: b]-glucose. Compared to methylotannin, these preparations exhibit very close resemblance to the former, from which it may be concluded that they are closely related to this substance, and probably possess the same or a very similar structure; the result of the above experiments has, therefore, brought us at least in close proximity to the structure of tannin. It must, however, be borne in mind that the analysis and hydrolysis of tannin does not afford an explanation of the question as to whether tannin is a compound of glucose and 10, 9, or 11 molecules of gallic acid; it is also possible, though not probable, that tannin would contain a polysaccharide instead of glucose itself. Similarly to sugar, the true glucosides can be coupled with hydroxybenzoic acids, which is proved by the preparation of tetra-galloyl-[Greek: a]-methyl glucoside; this substance, also, exhibits tannoid character.