492 Dr. A. Harden and Mrs. D. Norris .
Production [ Nov. 22 Conclusions .
1 .
Antelope may remain in apparently perfect health for a year after having been infected with a human strain of T. gambiense .
2 .
One antelope was still capable of infecting clean laboratory-bred G. palpalis 315 days after it had been infected .
3 .
A small quantity of blood taken from one antelope 327 days after ita infection was proved by inoculation into a white rat to be infective .
4 .
As the interval after the infection of the antelope increases their infectivity , as tested by " cycle " transmission experiments , dissection of flies which have fed upon them , and by the injection of the buck 's blood into susceptible animals , appears to diminish .
5 .
A duiker was infected with a human strain of gambiense by feeding infected G. palpalis upon it .
The Bacterial Production of Acetylmethylcarbinol and 2.3-Butylene Glycol from Various Substances .
By Arthur Harden , F.RS .
, and Dorothy Norris .
( Received November 22 , 1911 , \#151 ; Read February 1 , 1912 .
) ( From the Biochemical Department , Lister Institute .
) In working out the action of B. lactis on glucose quantitatively .
Harden and Walpole ( 1 ) found that , in addition to the products already noted in the action of B. coli communis on glucose ( 2 ) , a small quantity of acetylmethylcarbinol , CH3CH(OH)COCH3 , and a- considerable proportion of 2.3-butylene glycol , CH3CH(OH)CH(OH)-CH3 , were formed , the latter corresponding to about 33 per cent , of the carbon of the sugar fermented .
The production of acetylmethylcarbinol by the action of tenuis , B. subtilis and B. mesentericus vulgatus , and similar organisms on glucose , had been previously noted by Grimbert ( 3 ) and by Desmots ( 4 ) .
The presence of acetylmethylcarbinol is of especial interest , as it has been shown to be the substance responsible for the Yoges and Proskauer reaction ( 5 ) .
In view of this fact , and of the interest attaching to this mode of decomposition of glucose , it becomes a matter of some importance to discover what substances will give rise to acetylmethylcarbinol and butylene glycol during fermentation , and also which bacteria are capable of producing these two compounds .
B. lactis aerogenes and B. cloacce have been shown to 1911 .
] Acetylmethylcarbinol and Glycol .
produce both substances from glucose , mannitol , and fructose ( 6 , 7 , 8 ) ; hence , , in the first place , the action of these organisms was studied , and similar experiments were later carried out with B. coli communis , but with negative results .
It was of especial interest to discover whether these substances could be produced from carbon compounds less complex than the hexoses , and a variety of simpler substances containing fewer carbon atoms were therefore tried , e.g. , glycerol , ethylene glycol , malic acid , etc. The formation of acetylmethylcarbinol or butylene glycol from substances containing three or fewer carbon atoms would necessarily involve a carbon synthesis which would be of considerable theoretical interest .
An instance of this kind , the production of butyric acid and butyl alcohol from lactic acid and from glycerol in the butyric fermentation , has long been known ( 9,10 ) .
To explain this , it has been supposed that acetaldehyde is first formed and then serves as the source of the butyl alcohol and butyric acid .
The aldehyde may be supposed to undergo an aldol condensation followed by reduction , with or without subsequent oxidation : 2CH3-CHO = CH3-CH(OH\gt ; CH2-CHO .
CH3-CH(OH\gt ; CH2*CHO + 4H = ch3-ch2-ch2-ch2-oh+h2o .
It seems possible that the production of acetylmethylcarbinol and butylene glycol may be due to a somewhat similar course of events , which may be represented as follows , although these reactions have not , hitherto , been observed in the laboratory :\#151 ; 2CH3*CHO + 2H = CH3-CH(OH)-CH(OH)-CH3 .
2CH3'CHO = CH3-CO-CH(OH)-CH3 .
Experiments to test this hypothesis were made , with the result that the production of the glycol by bacterial action from acetaldehyde was conclusively established , although the mechanism of this production has not yet been ascertained .
It by no means necessarily follows , however , that in the fermentation of glucose the butylene glycol is actually produced from preformed acetaldehyde .
1 .
Experimental Methods .
As a general rule the culture medium consisted of 1 per cent. Witte 's peptone water containing 5 per cent , of the substance under investigation .
Sufficient chalk was added to neutralise any acid formed during fermentation , and after inoculation the culture was grown for three weeks at 37 ' C. under anaerobic conditions , the flask being fitted with a mercury trap to allow the escape of any gases .
494 Dr. A. Harden and Mrs. D. Norris .
Production of [ Nov. 22 , Among the products of the reaction , acetylmethylcarbinol and butylene glycol only have been estimated quantitatively .
A more complete investigation has , however , been made in the case of glycerol , and the results obtained will form the subject of another communication .
To ascertain whether the carbinol and glycol were derived from the acetaldehyde , sugar or other substances in question , and not simply from a constituent of the peptone , control experiments were carried out with peptone water alone as culture medium .
In no case could the slightest trace of either substance be detected .
2 .
Detection and Estimation of Acetylmethylcarbinol and Butylene Glycol .
As some of the substances employed could only be used in small quantities , it was necessary to elaborate a method for the detection of both acetylmethylcarbinol and butylene glycol in small amounts .
Attempts were first made to separate these two compounds by cautious evaporation and estimate them individually , but it was found impossible to arrive at any quantitative values in this way .
The method ultimately adopted was as follows :\#151 ; The medium in which the organism had been grown was carefully distilled to as small a bulk as possible over a free flame and then to dryness under reduced pressure at 37 ' C. The distillates were then united and made up to a definite volume and portions tested for reducing power and for Voges and Proskauer 's reaction , which , as stated above , is due to the presence of acetylmethylcarbinol .
To perform this reaction 3 c.c. of 1 per cent. Witte 's peptone water are mixed in a test-tube with an equal quantity of 10 per cent , caustic soda , 2 c.c. of the solution to be tested are then carefully poured on to the surface of the liquid , and the tube is allowed to stand at room temperature .
If acetylmethylcarbinol be present , a pink ring forms at the juncture of the two liquids .
With very small quantities of carbinol , this may take some hours to develop , but with larger amounts the colour soon appears and quickly spreads through the whole of the solution , a green fluorescence being also produced .
It was found that more delicate results could be obtained by the above method than by simply mixing the solutions .
Experiments in connection with other work , shortly to be published , have indicated that the carbinol is alone responsible for the reducing power of the distillates , and hence an estimation of this by Bang 's ( 11 ) method at once gives the amount of acetylmethylcarbinol present , the reducing power cf this substance being known ( 12 ) .
The estimation of the butylene glycol is not such a simple matter .
1911 .
] Acetylmethylcarbinoland 2.3- Glycol .
495 A small amount is held in the dry residue after distillation , and must be extracted with dry ether , and the glycol thus obtained , after evaporation of the ether , added to the distillate .
The estimation depends on the fact that the glycol is readily oxidised by bromine under the influence of light to diacetyl , which can be estimated by its reducing power .
Any acetylmethylcarbinol present is quantitatively oxidised to diacetyl in this way , but as the quantity of this substance will already have been determined , the amount of diacetyl formed from it is known , and the difference between this and the total diacetyl represents that due to the glycol .
The details of the process are as follows:\#151 ; An aliquot portion of the distillate is treated with a small quantity of bromine ( 0*1 c.c. for the distillate from the fermentation of 5 grm. of substance ) , and left exposed for 12 to 15 hours to the light of a 50-candle-power electric lamp .
A further addition of bromine is made if the solution becomes completely decolorised , and the exposure continued .
Any free bromine is then removed by the cautious addition of sulphurous acid , excess of this being carefully avoided .
The solution is next saturated with sodium chloride , and the diacetyl distilled off and estimated by determining the reducing power of the distillate .
The results obtained by the above method are low , as the oxidation of the glycol to diacetyl is not quantitative .
Control experiments with known amounts of glycol show that the actual results obtained are two-thirds the correct value , and a correction for this has been made in the tables given below .
In most cases the diacetyl produced by the oxidation was further characterised by the preparation and analysis of the phenylosazone .
Two typical analyses gave the following results:\#151 ; 1 .
Substance fermented\#151 ; galactose .
Organism\#151 ; doacce .
Osazone\#151 ; imp .
243 ' C. 0*2109 grm. gave 39*4 c.c. N at 23*5 ' C. and 765 mm. 1ST = 21*14 per cent. 2 .
Substance fermented\#151 ; arabinose .
Organism\#151 ; B. lactis aerogenes .
Osazone\#151 ; m.p. 244 ' C. 0*1349 grm. gave 25 c.c. N at 22 ' C. and 765 mm. N = 21*13 per cent. Diacetylphenylosazone , Ci6Hi8N4 , requires N = 21*05 per cent. In the sugar experiments the amount unfermented was estimated in the residue left after the extraction of the glycol by dissolving it in water , making up to a known volume , and determining the reducing power after treatment with mercuric nitrate ( Patein ) .
496 Dr. A. Harden and Mrs. D. Norris .
Production of [ Nov. 22 3 .
Results Obtained .
The results obtained with the organisms employed are indicated in the tables given below .
The figures given in Columns 8 and 9 are the values calculated from the actual results found for 10 grm. of sugar fermented .
As previously mentioned , the results obtained with B. coli communis were in every case negative .
Table I. Action of R. lactis aerogenes ( Escherich ) and B. ( Jordan ) on various Sugars .
Experi- meant .
Organism .
Sugar .
Weight fermented .
Time of Carbinol per 10 grm. sugar Glycol per 10 grm. sugar growth .
fermented .
fermented .
i B. lactis aerogenes G-lucose grm. 4-66 3 weeks grm. o-ii grm. 1 *42 2 99 99 99 4-85 99 0-12 1 '39 3 99 99 99 5-00 99 o-ii 1-36 4 99 99 99 2 -47 99 o-ii 1 *44 5 B. cloacae Fructose 4-67 99 o-ii 1 *55 6 99 n 4 15 99 0-11 1*51 7 99 Mannose 2-32 99 Oil 1*36 8 B. lactis aerogenes 99 2 45 99 0*10 1 *42 9 99 99 Galactose 4-03 99 0-06 0*86 10 99 99 2 -40 99 0-07 0*85 11 B. cloacae 99 4-74 99 0-08 1 *41 12 99 B. lactis aerogenes ; j 4-38 0-08 1 *36 13 Arabinose 3-92 99 0*07 1 *18 14 9 9 99 " 2-39 99 0-08 1 *19 15 B. cloacae 99 5-00 99 0 -06 1 *19 16 99 99 1-83 99 0-08 1 *18 17 B. lactis aerogenes Isodulcite 2-03 32 days 0-43 1 *45 18 99 99 99 1 77 3 weeks 0-67 1 *54 19 B. cloacae " 2 -03 99 0-04 i 0*81 It will be seen from the above table that glucose , fructose , and mannose have given practically the same quantities of carbinol and glycol respectively , whilst with galactose the carbinol is slightly less in amount with both organisms used , and the yield of glycol varies with the organism employed .
The results obtained in the case of arabinose show close agreement for both organisms , although in all four cases lower figures are obtained than with the hexoses .
The question of isodulcite is interesting , as the amount of carbinol obtained by means of B. lactis is abnormally high com- pared with that obtained from the other sugars or by using B. cloacce .
The galactose and arabinose used in the above experiments were previously purified from traces of glucose by fermentation with yeast .
The amount of glycol produced from glucose is decidedly lower than that obtained by Harden and Walpole , a result which is possibly due to the employment of a different strain of the organism .
Acetylmethylcarbinol and 2.3 Glycol .
Table II.\#151 ; Action of B. lactis aerogenes ( Escherich ) and B. ( Jordan ) on various Alcohols .
Experi- meant .
Organism .
Alcohol .
Weight taken .
Carbinol per 10 grm. taken .
Glycol per 10 grm. taken .
Glycerol grm. grm. grm. 20 B. lactis aerogenes 5 Ml 0-05 21 n n 99 50 99 0-05 22 n ii 99 10 99 0-05 23 11 99 99 10 99 0-04 24 B. cloacae 99 10 }9 Nil 25 B. lactis aerogenes Ethylene glycol 5 91 0-08 26 99 99 99 5 91 0-08 27 99 99 99 5 91 0-09 28 91 99 Adonitol 2-50 0-06 0-22 29 19 99 99 2-50 0-06 0-24 30 99 99 Mannitol 5 0-03 0-23 31 99 99 99 5 0-03 0-04 The growth was continued for three weeks , except in the case of Experiment 31 , which was for two weeks .
The residual alcohols were not estimated , so that the results stated above are not corrected for amount of .substance unfermented .
Only in the case of mannitol and adonitol was any .carbinol detected .
All these alcohols , however , yielded glycol .
Citric and malic acids gave negative results with both organisms .
The action of B. lactis aerogenes on dihydroxyacetone , CH^OHyCOCH^OH , was also tried , but here again without positive result .
4 .
Experiments on the Synthesis of 2.3-Bu Glycol from Acetaldehyde by means of B. lactis aerogenes .
Eor each experiment a litre of Witte 's peptone water was made up and sufficient chalk added to neutralise any acids which might be formed during fermentation .
The medium was then inoculated with B. lactis aerogenes and incubated at 37 ' C. for 24 hours before any addition of acetaldehyde was made , in order to establish a good growth of the organism .
In some experiments calcium formate ( 10 grin , formic acid per 100 c.c. water neutralised with CaC03 ) was added with the acetaldehyde ( 10 grm. per 100 c.c. ) ; in others acetaldehyde was used alone .
Two cubic centimetres of the above solutions were added to the culture medium each day under sterile conditions , so that each flask received a daily addition of 0'2 grm. of acetaldehyde and the same quantity of formic acid ( as formate ) in the cases where this was used .
This treatment was continued until in two experiments 60 c.c. of the above solutions had been added , and in two others until 80 c.c. had been 498 Dr. A. Harden and Mrs. D. Norris .
Production of [ Nov. 22 added .
At the end of this time and occasionally during the progress of the experiment sub-cultures were made to show that the organism in question was still alive and uncontaminated .
In two experiments the whole of the acetaldehyde had not been used up , so that the liquid obtained by distillation of the culture medium was strongly reducing to Fehling 's solution .
In two other cases sufficient time was allowed to elapse between the last addition of aldehyde and the examination of the products to ensure the complete removal of this factor .
In these two cases the distillates obtained had no reducing power .
In every experiment the distillate was tested by means of the Yoges and Proskauer reaction for acetylmethylcarbinol , and in no case could any trace of this be detected .
The liquid was further examined for butylene glycol , which was in every case found to be present , though in extremely small quantities .
It was detected as described above by the formation of diacetyl , this substance being proved to be present in every case by the positive results given by the Yoges and Proskauer reaction after oxidation of the culture distillate .
In one case sufficient osazone was prepared from the diacetyl for a determination of the melting point , which was found to be 244 ' C. Among the products of the reaction were also found ethyl alcohol , acetic acid , and some succinic acid .
Lactic acid was not present .
These products were estimated in the manner previously described by Harden ( 13 ) with a few slight modifications .
In the estimation of the alcohol any unchanged acetaldehyde present was removed by oxidation with moist silver oxide .
The acetic acid was determined by the method of Macnair(14 ) .
The table below shows the results obtained in three typical experiments :\#151 ; Table III .
Total acetaldehyde added Total formic acid added ( as Ca formate ) ... 1 .
6 *07 grm. 6*07 2 .
8 -0 grm. Nil 3 .
7 -6 grm. 7'6 " Alcohol grm. grm. I grm. 6*5 4*02 4*5 Acetic acid 2*5 1 *38 1 *3 Succinic acid Not estimated 1 *25 0 *12 Butylene glycol + 0*677 0*109 From the point of view of the production of butylene glycol the presence of calcium formate appears to be detrimental , and it is also interesting to note the somewhat large quantity of succinic acid produced in Experiment 2 , which also gave the largest yield of glycol .
1911.J Acetylmethylcarbinol and 2.3-Butylene Glycol .
Summary of Results .
1 .
B. lactis aerogenes and B. cloacae , when grown in a peptone solution containing either glucose , fructose , mannose , galactose , arabinose , isodulcite , mannitol or adonitol , produce both acetylmethylcarbinol and 2.3-butylene glycol .
2 .
Glycerol , ethylene glycol and acetaldehyde , under similar conditions , also give rise to 2.3-butylene glycol in presence of lactis aerogenes , but no acetylmethylcarbinol is produced .
In these three cases a carbon synthesis is involved , analogous to that which occurs in the butyric fermentation of glycerol and lactic acid .
3 .
The fermentation of citric and malic acids , of dihydroxyacetone , and of peptone water , gives rise to neither carbinol nor glycol .
REFERENCES .
1 .
Harden and Walpole , ' Roy .
Soc. Proc. , ' 1906 , B , vol. 77 , p. 399 .
2 .
Harden , 'Chem .
Soc. Journ. , ' 1901 , p. 610 .
3 .
Grimbert , 'Cornpt .
Rend .
, ' 1901 , vol. 132 , p. 706 .
4 .
Desmots , 'Compt .
Rend .
, ' 1904 , vol. 138 , p. 581 .
5 .
Harden , ' Roy .
Soc. Proc. , ' 1906 , B , vol. 77 .
6 .
Harden and Walpole , ' Roy .
Soc. Proc. , ' 1906 , B , vol. 77 , p. 399 .
7 .
Walpole , 'Roy .
Soc. Proc. , ' 1911 , B , vol. 83 , p. 272 .
8 .
Thompson , 'Roy .
Soc. Proc. , ' this vol. , p. 500 .
9 .
Fitz , ' Ber .
, ' 1880 , vol. 13 , p. 1309 .
10 .
Buchner and Meisenheimer , 'Ber .
, ' 1908 , vol. 41 , p. 1410 .
11 .
Bang , 'Biochem .
Zeitschi- .
, ' 1907 , vol. 2 , p. 271 .
12 .
Harden and Walpole , 'Roy .
Soc. Proc. , ' 1906 , B , vol. 77 , p. 399 .
13 .
Harden , 'Chem .
Soc. Journ. , ' 1901 , p. 610 .
14 .
Macnair , ' Chem. News , ' 1887 , vol. 55 , p. 229 .
2 o VOL. LXXXIV.\#151 ; B.

