CHAPTER VII.

THE BY-PRODUCTS OF ALCOHOLIC FERMENTATION.

When pure yeast is allowed to develop in a solution of sugar containing a suitable nitrogenous diet and the proper mineral salts, the liquid at the close of the fermentation contains not only alcohol and some carbon dioxide but also a considerable number of other substances, some arising from the carbonaceous and others from the nitrogenous metabolism of the cell. Prominent among the non-nitrogenous substances which are thus found in fermented sugar solutions are fusel oil, succinic acid, glycerol, acetic acid, aldehyde, formic acid, esters, and traces of many other aldehydes and acids. In addition to these substances which are found in the liquid, there are also the carbonaceous constituents of the newly formed cells of the organism, comprising the material of the cell walls, yeast gum, glycogen, complex organic phosphates, as well as other substances.

The attention of chemists has been directed to these compounds since Pasteur first emphasised their importance as essential products of the alcoholic fermentation of sugar, and his example was generally followed in attributing their origin to the sugar.

The study of cell-free fermentation by means of yeast-juice or zymin has, however, revealed the facts that certain of these substances are not formed in the absence of living cells, and that their origin is to be sought in the metabolic processes which accompany the life of the cell. Their source, moreover, has been traced not to the sugar but to the amino-acids, formed by the hydrolysis of the proteins, which occur in all such liquids as beer wort, grape juice, etc., which are usually submitted to alcoholic fermentation. This has so far been proved with certainty for the fusel oil and succinic acid, and rendered highly probable for all the various aldehydes and acids of which traces have been detected.

FUSEL OIL.

All forms of alcohol prepared by fermentation contain a fraction of high boiling-point, which is termed fusel oil, and amounts to about [p086] 0·1 to 0·7 per cent. of the crude spirit obtained by distillation. This material is not an individual substance, but consists of a mixture of very varied compounds, all occurring in small amount relatively to the ethyl alcohol from which they have been separated. The chief constituents of the mixture are the two amyl alcohols, isoamyl alcohol,

(CH{3}){2}·CH·CH{2}·CH{2}·OH,

and /d/-amyl alcohol,

CH{3}·CH(C{2}H{5})·CH{2}·OH,

which contains an asymmetric carbon atom and is optically active. In addition to these, much smaller amounts of propyl alcohol and isobutyl alcohol are present, together with traces of fatty acids, aldehydes, and other substances.

The origin of these purely non-nitrogenous compounds was usually sought in the sugar of the liquid fermented, from which they were thought to be formed by the yeast itself or by the agency of bacteria [Emmerling, 1904, 1905; Pringsheim, 1905, 1907, 1908, 1909], whilst others traced their formation to the direct reduction of fatty acids. Felix Ehrlich has, however, conclusively shown in a series of masterly researches that the alcohols, and probably also the aldehydes, contained in fusel oil are in reality derived from the amino-acids which are formed by the hydrolysis of the proteins.

The close relationship between the composition of leucine,

(CH{3}){2}·CH·CH{2}·CH(NH{2})·COOH,

and isoamyl alcohol,

(CH{3}){2}·CH·CH{2}·CH{2}·OH,

had previously led to the surmise that a genetic relation might exist between these substances, but the idea had not been experimentally confirmed. In 1903 Ehrlich discovered [1903; 1904, 1, 2; 1907, 2; 1908; Ehrlich and Wendel, 1908, 2] that proteins also yield on hydrolysis an isomeride of leucine known as isoleucine, which has the constitution

CH{3}·CH(C{2}H{5})·CH(NH{2})·COOH,

and therefore stands to /d/-amyl alcohol,

CH{3}·CH(C{2}H{5})·CH{2}·OH,

in precisely the same relation as leucine to isoamyl alcohol. This suggestive fact at once directed his attention to the problem of the origin of the amyl alcohols in alcoholic fermentation. Using a pure culture of yeast, and thus excluding the participation of bacteria in the change, he found that leucine readily yielded isoamyl alcohol, and isoleucine /d/-amyl alcohol when these amino-acids were added in the pure state [p087] to a solution of sugar and treated with a considerable proportion of yeast [1905; 1906, 2, 3; 1907, 1, 3]. The chemical reactions involved are simple ones and are represented by the following equations:--

(1) (CH{3}){2}·CH·CH{2}·CH(NH{2})·COOH + H{2}O Leucine = (CH{3}){2}CH·CH{2}CH{2}·OH + CO{2} + NH{3} Isoamyl alcohol

(2) CH{3}·CH(C{2}H{5})·CH(NH{2})·COOH + H{2}O Isoleucine = CH{3}·CH(C{2}H{5})·CH{2}·OH + CO{2} + NH{3} /d/-Amyl alcohol

The experiments by which these important changes were demonstrated were of a very simple and convincing character [Ehrlich, 1907, 1]. Two hundred grams of sugar and 3 to 10 grams of the nitrogenous substance to be examined were dissolved in 2 to 2·5 litres of tap water in a 3 to 4 litre flask, the liquid was sterilised by being boiled for several hours, and after cooling 40 to 60 grams of fresh yeast were added and the flask allowed to stand at room temperature until the whole of the sugar had been decomposed by fermentation. In the earlier experiments the amyl alcohols were isolated and identified by conversion into the corresponding valerianic acids, but as a rule the fusel oil as a whole was quantitatively estimated in the filtrate by the Röse-Herzfeld method [Lunge, 1905, p. 571].

The following are typical results. (1) An experiment carried out as above without any addition of leucine gave 97·32 grams of alcohol containing 0·40 per cent. of fusel oil. (2) When 6 grams of synthetic, optically inactive leucine were added, 97·26 grams of alcohol were obtained, containing 2·11 per cent. of fusel oil, which was also optically inactive; 2·5 grams of leucine were recovered, so that 87 per cent. of the theoretical yield of isoamyl alcohol was obtained from the 3·5 grams of leucine decomposed. (3) In the presence of 2·5 grams of /d/-isoleucine (prepared from molasses residues), 200 grams of sugar gave 93·99 grams of alcohol, containing 1·44 per cent. of fusel oil, which was lævo-rotatory. This corresponds with 80 per cent. of the theoretical yield of /d/-amyl alcohol from the isoleucine added.

This change, which Ehrlich has termed the alcoholic fermentation of the amino-acids, although brought about by living yeast, does not appear to occur at all when zymin [Ehrlich, 1906, 4; Pringsheim, 1906] or yeast-juice [Buchner and Meisenheimer, 1906] is substituted for the intact organism, nor is it effected even by living yeast in the absence of a fermentable sugar [Ehrlich, 1907, 1]. The reaction appears indeed to be intimately connected with the nitrogenous metabolism of the cell, and the whole of the ammonia produced is at once assimilated and does not appear in the fermented liquid. Other amino-acids [p088] undergo a corresponding change, and the reaction appears to be a general one. Thus tyrosine, OH·C{6}H{4}·CH{2}·CH(NH{2})·COOH, yields p-hydroxyphenylethyl alcohol, or tyrosol [Ehrlich, 1911, 1; Ehrlich and Pistschimucka, 1912, 2], OH·C{6}H{4}·CH{2}·CH{2}OH, a substance of intensely bitter taste, which was first prepared in this way and is probably one of the most important factors in determining the flavour of beers, etc. Phenylalanine, C{6}H{5}·CH{2}·CH(NH{2})·COOH, in a similar way yields phenylethyl alcohol, C{6}H{5}·CH{2}·CH{2}OH, one of the constituents of oil of roses, whilst tryptophane,

C{6}H{4} ╱ ╲ ╱ ╲ HN C·CH{2}·CH(NH{2})·COOH, ╲ ╱╱ ╲ ╱╱ CH

yields tryptophol,

C{6}H{4} ╱ ╲ ╱ ╲ HN C·CH{2}·CH{2}OH, ╲ ╱╱ ╲ ╱╱ CH

which was also first prepared in this way [Ehrlich, 1912] and has a very faintly bitter, somewhat biting taste.

The extent to which the amino-acids of a medium in which yeast is producing fermentation are decomposed in this sense depends on the amount of the available nitrogen and on the form in which it is present. Thus the addition of ammonium carbonate to a mixture of yeast and sugar was found to lower the production of fusel oil from 0·7 to 0·33 per cent. of the alcohol produced. The addition of leucine alone raised the percentage from 0·7 to 2·78, but the addition of both leucine and ammonium carbonate resulted in the formation of only 0·78 per cent. of fusel oil, The production of fusel oil therefore and the character of the constituents of the fusel oil alike depend on the composition of the medium in which fermentation occurs. This affords a ready explanation of the fact that molasses, which contains almost equal amounts of leucine and isoleucine, yields a fusel oil also containing approximately equal amounts of isoamyl alcohol and /d/-amyl alcohol [Marckwald, 1902], whilst corn and potatoes, in which leucine preponderates over isoleucine, yield fusel oils containing a relatively large amount of the inactive alcohol. The subject is, in fact, one of great interest to the technologist, for as Ehrlich points out "the great variety of the bouquets of wine and aromas of brandy, cognac, arrak, rum, etc., may be very simply referred to the manifold variety of the proteins of the raw materials (grapes, corn, rice, sugar cane, etc.) from which they are derived".

Yeast can also form fusel oil at the expense of its own protein, but this only occurs to any considerable extent when the external [p089] supply of nitrogen is insufficient. Under these circumstances the amino-acids formed by autolysis may be decomposed and their nitrogen employed over again for the construction of the protein of the cell.

The yield is also influenced by the condition of the yeast employed with regard to nitrogen, a yeast poor in nitrogen being more efficacious in decomposing amino-acids than one which is already well supplied with nitrogenous materials. The nature of the carbonaceous nutriment and finally the species of yeast are also of great importance [see Ehrlich, 1911, 2; Ehrlich and Jacobsen, 1911].

A very important characteristic of the action of yeast on the amino-acids is that the two stereo-isomerides of these optically active compounds are fermented at different rates. When inactive, racemic leucine is treated with yeast and sugar, the naturally occurring component, the /l/-leucine, is more rapidly attacked, so that if the experiment be interrupted at the proper moment the other component, the /d/-leucine, alone is present and may be isolated in the pure state. In an actual experiment 3·8 grams of this component were obtained in the pure state from 10 grams of /dl/-leucine [Ehrlich, 1906, 1], so that the whole of the /l/-leucine (5 grams) had been decomposed but only 1·2 grams of the /d/-leucine. This mode of action has been found to be characteristic of the alcoholic fermentation of the amino-acids by yeast. In all the instances so far observed, both components of the inactive amino-acid are attacked, but usually the naturally occurring isomeride is the more rapidly decomposed, although in the case of β-aminobutyric acid both components disappear at the same rate [Ehrlich and Wendel, 1908, 1]. This reaction therefore must be classed along with the action of moulds on hydroxy-acids [McKenzie and Harden, 1903], and the action of lipase on inactive esters [Dakin, 1903, 1905], in which both isomerides are attacked but at unequal rates, and differs sharply from the action of yeast itself on sugars [Fischer and Thierfelder, 1894], and of emulsin, maltase, etc., which only act on one isomeride and leave the other entirely untouched [see Bayliss, 1914, pp. 55, 77, 117].

SUCCINIC ACID.

The origin of the succinic acid formed in fermentation has also been traced by Ehrlich [1909] to the alcoholic fermentation of the amino-acids. It was shown by Buchner and by Kunz [1906] that succinic acid like fusel oil is not formed during fermentation by yeast-juice or zymin, and, in the light of Ehrlich's work on fusel oil, several [p090] modes of formation appeared possible for this substance [Ehrlich, 1906, 3]. The dibasic amino-acids might, for example, undergo simple reduction, the NH{2} group being removed as ammonia and replaced by hydrogen. Aspartic acid would thus pass into succinic acid:--

COOH·CH{2}·CH(NH{2})·COOH + 2H = COOH·CH{2}·CH{2}·COOH + NH{3}.

This change can be effected in the laboratory only by heating with hydriodic acid. Biologically it has been observed [E. and H. Salkowski, 1879] when aspartic acid is submitted to the action of putrefactive bacteria, and almost quantitatively when /Bacillus coli communis/ is cultivated in a mixture of aspartic acid and glucose [Harden, 1901]. In this case a well-defined source of hydrogen exists in the glucose, which when acted on by this bacillus yields a large volume of gaseous hydrogen, which is not evolved in the presence of aspartic acid. Some such source is also available in the case of yeast, although it cannot be chemically defined, for this organism is known to produce many reducing actions, which are usually ascribed to the presence of reducing ferments or reductases in the cell.

A similar action would convert glutamic acid,

COOH·CH{2}·CH{2}·CH(NH){2}·COOH,

into glutaric acid,

COOH·CH{2}·CH{2}·CH{2}·COOH,

which also is found among the products of fermentation, whilst the monamino-acids would pass into the simple fatty acids.

On submitting these ideas to the test of experiment, however, Erhlich found that the addition of aspartic acid did not in any way increase the yield of succinic acid, and that of all the amino-acids which were tried only glutamic acid, COOH·CH{2}·CH{2}·CH(NH{2})·COOH, produced a definite increase in the amount of this substance. Further experiments showed that glutamic acid was actually the source of the succinic acid, the relations being quite similar to those which exist for the production of fusel oil.

Succinic acid is formed whenever sugar is fermented by yeast, even in the absence of added nitrogenous matter, and amounts to 0·2 to 0·6 per cent. of the weight of the sugar decomposed, its origin in this case being the glutamic acid formed by the autolysis of the yeast protein. When some other source of nitrogen is present, such as asparagine or an ammonium salt, the amount falls to 0·05 to 0·1. If glutamic acid be added it rises to about 1 to 1·5 per cent. but falls again to about 0·05 to 0·1 when other sources of nitrogen, such as asparagine or ammonium salts, are simultaneously available, either in the presence or [p091] absence of added glutamic acid. As in the case of fusel oil, the production does not occur in the absence of sugar, and is not effected by yeast-juice or zymin.

The chemical reaction involved in the production of succinic acid differs to some extent from that by which fusel oil is formed, inasmuch as an oxidation is involved:--

COOH·CH{2}·CH·CH(NH{2})·COOH + 2O = COOH·CH{2}·CH{2}·COOH + NH{3} + CO{2}.

From analogy with the production of amyl alcohol from leucine, glutamic acid would be expected to yield γ-hydroxybutyric acid:--

COOH·CH{2}·CH{2}·CH(NH{2})·COOH + H{2}O = NH{3} + CO{2} + COOH·CH{2}·CH{2}·CH{2}·OH.

As a matter of fact this substance cannot be detected among the products of fermentation, but succinic acid as already explained is formed. This acid might, however, possibly be formed by the oxidation of the γ-hydroxybutyric acid:--

COOH·CH{2}·CH{2}·CH{2}·OH + 2O = COOH·CH{2}·CH{2}·COOH + H{2}O,

although this change is on biological grounds improbable.

The conversion of the group ─CH(NH{2})─ into the terminal CH{2}·OH in fusel oil, or COOH in succinic acid, may possibly be effected in several different ways, the most probable of which are the following:--

I. Direct elimination of carbon dioxide, followed by hydrolysis of the resulting amine:--

(1) R·CH(NH{2})·COOH = R·CH{2}·NH{2}+ CO{2}.

(2) R·CH{2}·NH{2} + H{2}O = R·CH{2}·OH + NH{3}.

The reaction (1) is actually effected by many bacteria and has been employed for the preparation of bases from amino-acids [cf. Barger, 1914, p. 7], although there is no direct evidence that it can be brought about by yeast. On the other hand reaction (2) has actually been observed with some yeasts. Thus it has been found [Ehrlich and Pistschimuka, 1912, 1] that many "wild" yeasts produce this change with great readiness in presence of sugar, glycerol or ethyl alcohol as sources of carbon and grow well in media in which amines, such as p-hydroxyphenylethylamine or iso-amylamine, form the only source of nitrogen. /Willia anomala/ (Hansen), a yeast which forms surface growths, succeeds admirably under these conditions, whereas culture yeasts are much less active in this way, although they produce a certain amount of change. It is therefore possible that this mode of decomposition plays some part in the production of fusel oil, but in the case of culture yeasts it is entirely subordinated to the mode next to be discussed. [p092]

II. Oxidative removal of the ─NH{2} group with formation of an α-ketonic acid:--

(1) R·CH(NH{2})·COOH + O = R·CO·COOH + NH{3}

followed by the decomposition of the ketonic acid into carbon dioxide and an aldehyde and the subsequent reduction or oxidation of the aldehyde:--

(2) R·CO·COOH = R·CHO + CO{2}.

(3) (/a/) R·CHO + 2H = R·CH{2}OH. (/b/) R·CHO + O = R·COOH.

The evidence for the occurrence of reaction (1) is supplied by the experiments of Neubauer and Fromherz [1911]. Having previously found that amino-acids undergo a change of this kind in the animal body, Neubauer investigated their behaviour towards yeast. Taking /dl/-phenylaminoacetic acid, C{6}H{5}·CH(NH{2})·COOH, it was found that the changes produced were essentially the same as in the animal body. The /l/-component of the acid was partly acetylated and partly unchanged, whereas the /d/-component of the acid yielded benzyl alcohol, C{6}H{5}·CH{2}·OH, phenylglyoxylic acid, C{6}H{5}·CO·COOH, and the hydroxy-acid C{6}H{5}·CH(OH)·COOH. Since however this hydroxy-acid was produced in the /l/-form it probably arose by the asymmetric reduction of phenylglyoxylic acid, a reaction which can be effected by yeast as was also found to be the case in the animal body [see Dakin, 1912, pp. 52, 78]. Moreover it was shown that when the effects of yeast on a ketonic acid and the corresponding hydroxy-acid were compared, the alcohol was formed in much better yield from the ketonic acid (70 per cent.) than from the hydroxy-acid (3-4 per cent.), the actual example being the production of tyrosol (p-hydroxyphenylethyl alcohol), OH·C{6}H{4}·CH{2}·CH{2}OH, from p-hydroxyphenylpyruvic acid, OH·C{6}H{4}·CH{2}·CO·COOH, and p-hydroxyphenyl-lactic acid, OH·C{6}H{4}·CH{2}·CH(OH)·COOH respectively.

Neubauer by these experiments established two extremely important points. 1. That the amino-acids actually yield the corresponding α-ketonic acids when treated with yeast and sugar solution. 2. That the a-ketonic acids under similar conditions give the alcohol containing one carbon atom less in good yield, whereas the corresponding hydroxy-acids only give an extremely small amount of these alcohols.

It is therefore probable that at an early stage in the decomposition of the amino-acids by yeast a ketonic acid is produced, which then undergoes further change.

The source of the oxygen required for this reaction and the mechanism of oxidation have not yet been definitely ascertained. It is possible [p093]

OH ╱ that hydrated imino-acids of the type R·C─COOH are first formed ╲ NH{2}

[Knoop, 1910], but these have not as yet been isolated.

The spontaneous production of ketonic aldehydes from amino-acids and from hydroxy-acids in aqueous solution, which has been demonstrated by Dakin and Dudley [1913], points however to the possibility that the ketonic acid may be a secondary product derived from the corresponding ketonic aldehyde [see also Dakin, 1908; Neuberg, 1908, 1909]. This itself may either arise directly from the amino-acid or from a previously formed hydroxy-acid, the latter alternative being, however, improbable in view of the small yield of alcohol obtained from hydroxy-acids by the action of yeast in the experiments of Neubauer and Fromherz.

R·CH(NH{2})·COOH → R·CH(OH)·COOH

R·CH(NH{2})·COOH ⇌ R·CO·CHO

R·CO·CHO ⇌ R·CH(OH)·COOH

2R·CO·CHO + O{2} → 2R·CO·COOH

(2) Whatever be the exact mode by which the ketonic acid is formed, it appears most probable that a compound of this nature forms the starting-point for the next stage in the production of the alcohols. The researches of Neuberg, which have already been discussed on p. 81, have revealed a mechanism in yeast--the enzyme carboxylase--by which these α-ketonic acids are rapidly broken up into an aldehyde and carbon dioxide:

R·CO·COOH = R·CHO + CO{2}

and it can scarcely be doubted that this is the actual course of the reaction.

(3) The final conversion of the aldehyde into the corresponding alcohol is also a change which it has been proved can be effected by yeast [Neuberg and Rosenthal, 1913] probably by the aid of the reductase which is one of the weapons in its armoury of enzymes.

Yeast is capable of producing many vigorous reducing actions and rapidly reduces methylene blue and sodium selenite. It is in all probability due to a reaction of this kind that the iso-amylaldehyde and isovaleraldehyde were reduced to the alcohols in Neuberg and Steenbock's experiments [1913, 1914], and that considerable quantities of ethyl alcohol are formed in the sugar free fermentation of pyruvic acid [Neuberg and Kerb, 1913, 1] (see later p. 110 for a discussion of this question).

A further possibility exists that in some cases the aldehyde may [p094] be simultaneously oxidised and reduced or the molecule of one aldehyde reduced and that of another oxidised with production of the corresponding acid and alcohol by an "aldehydo-mutase," similar to that which has been observed by Parnas [1910] in many animal tissues. Finally the aldehyde may simply be converted into the corresponding acid by oxidation as appears to take place in the formation of succinic acid.

The intermediate production of an aldehyde would thus be consistent both with the production of alcohols and acids from amino-acids.

Fusel oil would be formed by the reduction of the aldehydes arising from the simple monobasic amino-acids, succinic acid would be produced by oxidation of the aldehyde derived from the dibasic glutamic acid.

In favour of this view is to be adduced the fact that aldehydes such as isobutyraldehyde and valeraldehyde have been found in crude spirit, whilst acetaldehyde is a regular product of alcoholic fermentation [see Ashdown and Hewitt, 1910]. Benzaldehyde, moreover, has been actually detected as a product of the alcoholic fermentation of phenylaminoacetic acid, C{6}H{5}·CH(NH{2})·COOH [Ehrlich, 1907, 1]. Further, the aldehydes so produced would readily pass by oxidation into the corresponding fatty acids, small quantities of which are invariably produced in fermentation.

This view of the nature of the alcoholic fermentation of the amino-acids is undoubtedly to be preferred to that previously suggested by Ehrlich [1906, 3] according to which a hydroxy-acid is first formed and then either directly decomposed into an alcohol and carbon dioxide or into an aldehyde and formic acid, the aldehyde being reduced and the formic acid destroyed (see p. 115).

R·CH(NH{2})·COOH → R·CH(OH)·COOH

R·CH(OH)·COOH → R·CH{2}OH + CO{2} or R·CHO + H·CO{2}H

R·CHO → R·CH{2}OH

The most probable course of the decomposition by which isoamyl alcohol and succinic acid are produced from leucine and glutamic acid respectively is therefore the following:--

(/a/) /Isoamyl Alcohol./ (1) (CH{3}){2}·CH·CH{2}·CH(NH{2})·COOH (3) (CH{3}){2}CHCH{2}·CHO+CO{2} Leucine Isovaleraldehyde

(2) (CH{3}){2}·CH·CH{2}·CO·COOH (4) (CH{3}){2}·CH·CH{2}·CH{2}OH α-Ketoisovalerianic acid Isoamyl alcohol

(/b/) /Succinic Acid./ (1) COOH·CH{2}·CH{2}·CH(NH{2})·COOH (3) COOH·CH{2}CH{2}·CHO + CO{2} Glutamic acid Succinic semialdehyde

(2) COOH·CH{2}·CH{2}·CO·COOH (4) COOH·CH{2}·CH{2}·COOH α-Keto-glutaric acid Succinic acid

[p095]

GLYCEROL.

Of the three chief by-products of alcoholic fermentation, only glycerol remains at present referable directly to the sugar. This substance, as shown by the careful experiments of Buchner and Meisenheimer [1906], is formed by the action both of yeast-juice and zymin to the extent of 3·8 per cent. of the sugar decomposed, and no other source for its production has so far been experimentally demonstrated. If it be true that during the decomposition of sugar into alcohol and carbon dioxide, substances containing three carbon atoms are formed as intermediate compounds (see p. 100), it is obvious that these might by reduction be converted into glycerol which would thus be a true by-product of the alcoholic fermentation of sugar. [See Oppenheimer, 1914, 2.] It has, however, been suggested that it may in reality be a product of decomposition of lipoid substances or of the nuclein of the cell (Ehrlich).

The effect of Ehrlich's work has been clearly to distinguish the chemical changes involved in the production of fusel oil and succinic acid from those concerned in the decomposition of sugar into alcohol and carbon dioxide, and to bring to light a most important series of reactions by means of which the yeast-cell is able to supply itself with nitrogen, one of the indispensable conditions of life. [p096]