Vol. 68, 1962] 271 YEAST REPRODUCTION DURING FERMENTATION By R. B. Gilliland, B.A., B.Sc, F.R.I.C. (Arthur Guinness Son & Co. (Dublin), Ltd., Si. James's Gate, Dublin) Received 23rd December, 1962 Numerous yeast strains were tested in skimming-type laboratory fers under controlled conditions. The total amount of yeast recovered varied from 5 g. of pressed yeast per litre of wort to g. per litre, depending on the strain used. As excessive yeast production is undesirable in a brewery, this character, which is stable in subculture, should be considered when selecting yeasts for brewery use. Introduction In normal brewery fers, the yeast first undergoes a lag phase during which little growth takes place; then there is a growth phase, during which the yeast re produces fairly rapidly; and finally comes the fer phase, during which yeast growth gradually slows down and the wort is fermented. A vigorous growth phase is essential in order to obtain an adequate speed of fer and a low final gravity; but brewers do not want more yeast growth than is necessary to produce these results, as excess yeast cannot be sold at an economic price when it has been grown with duty-paid worts as a source of carbohydrate. In addition, increased yeast production involves extra cost in separation of the yeast from its associated barm beer and extra wastage of beer. The effect of different factors on the amount of yeast produced during fermenta tion has, however, received little attention in the literature. The influence of the amount of aeration on yeast crops was studied by White & Munns,8 but they were concerned with yeast production and did not consider the effect of strain difference nor the more anaerobic conditions under which brewery yeast is produced. Fowell1 noted that a series of haploid yeasts gave yields of l-b-yl'2%. per litre (moist) when grown under partially anaerobic conditions. Hough & Hudson,2 in an investigation of the influence of yeast strain on loss of bittering material during fer, reported yeast yields of 3-5-5'6 mg. dry weight per ml. with different strains in wort of O.G. 40, and a range of from 2-8-6-7 mg. dry weight per ml. from a single strain in worts of O.G. varying from -60. It is known that the amount of yeast produced per brl. of beer varies considerably from one brewery to another depending on temperature, aera tion, original gravity of the wort, and other conditions of fer. It is, however, not generally realized that yeast production also depends on the strain of yeast used, and it is the purpose of this paper to show that different strains of brewer's yeast, used for fer under standard conditions, may give a wide range of weights of product yeast. Experimental Most of the results were obtained in standard laboratory fers carried out in the following manner. Fer vessels. These were flatbottomed Pyrex glass tubes, 91 cm. long, with an internal diameter of 6-8 cm. and total capacity of 3 litres. Each was fitted with a water trap to prevent evaporation (Fig. 1). Skimming vessels. These were glass tubes with rounded bottoms, each being 114 cm. long with an internal diameter of 4-6 cm. They were each fitted with an adjustable extension piece by means of which thenlength could be increased to about 1 cm. and their capacity adjusted to 2 litres (Fig. 1). A swan-neck could be fitted to the end of the extension piece. Yeast collecting vessels. These were 400- ml. beakers which were immersed in a water bath at 4 C, so that the yeast did not autolyse during the period of collection (Fig. 1). Growth medium. Mixed brewery worts of O.G. 46 were used. Aeration. The wort was aerated for 5 min. at pitching by passing sterile air into it through a ceramic filter of pore size -22 p.
272 gilliland: yeast reproduction [J. Inst. Brew. The fers were in some cases further aerated after 22 hr. by tumbling to remove excess carbon dioxide, filling the head space with air, and tumbling the beer times to allow further absorption of oxygen. D Fig. 1. Vessels used for laboratory fers. A: Fer tube fitted with a water trap. B: Skimming tube with adjustable extension. C: Yeast collecting vessel which is kept chilled. Temperature. The fers were kept at 18 C. for 22 hr. and were then brought to 21 C, at which temperature the fer was completed and skimming was carried out. Seeding. The seeding rate was 0-5 g. of pressed yeast (approx. 75% moisture) per litre (0-18 lb. per brl.). The weighed seed yeast was well dispersed in a small amount of wort and added to 2 litres of fully aerated wort in a fermenting vessel, which was then closed with a rubber bung and tumbled 6 times to ensure adequate dispersion of the yeast. Fer. The fermenting vessel was fitted with a water-trap and incubated at 18 C. in an upright position. After 22 hr. the fermenting wort was further aerated (if this was necessary) and the vessel was incubated at 21 C. in a sloped position, to the vertical. When the wort had fer mented to a fixed specific gravity (usually 1-018), the fermenting beer was all trans ferred to a skimming vessel. When the fob had settled, the swan-neck was fitted and the extension piece was adjusted so that the beer came to the level of the bottom of the swan-neck. The beer was left for 24 hr. in skimming vessel, which was incubated in an upright position at 21 C. During this time the top yeast worked its way through the swan-neck and into the collecting vessel. At the end of 24 hr. the yeast had ceased coming to the top and the rate of fall of specific gravity of the beer had become very slow. Yeast estimation. The total yeast was estimated in three fractions: (1) The top yeast collected in the beaker and adhering to the swan-neck. (2) The suspended yeast in the beer after it had been poured off. (3) The bottom yeast which had sedimented out in the skimming vessel. The top and bottom yeasts were made up to 250 ml. with water and the yeast contents were determined by capillary centrifuge. The yeast in suspension in the beer was also estimated by capillary centrifuge. In some experiments the yeast products were also estimated by filtering off the yeasts in a buchner funnel and weighing, in order to check that the rapid centrifuge method was giving correct results. This rather elaborate method of estimating the yeast in three fractions was necessary only because the results reported in this communication were part of a larger investigation into the brewing properties of yeast strains, and for
Vol. 68, 1982] gilliland: yeast reproduction 273 this purpose it was essential to know the distribution of the yeast after skimming. In all cases the results are expressed as total number of grams of pressed yeast, i.e. with approximately 75% moisture, recovered per litre of wort. Yeasts. The yeast strains were all either Saccharotnyces cerevisiae or Saccharomyces carlsbergensis. These strains included about 40 strains received from the National Collec tion of Yeast Cultures, which had been isolated from brewer's yeasts by the Brewing Industry Research Foundation, and about 0 strains isolated in this laboratory from English and Irish brewer's yeasts. These strains were kept in pure culture and grown in wort in cotton-wool-plugged flasks at 25 C. for two days before use. Results and Discussion Strain variation in yeast yield. The results have been expressed in the form of histo grams in Figs. 2-6. The crops obtained, in single fers, on 50 different occasions over a period of a year with a standard yeast culture, are plotted in Fig. 2. The -tor.0 6 K 18 22 Fig. 3. Total yeast crops recovered from duplicate fers with 40 different strains of brewer's yeasts received from N.C.Y.C. A miscellaneous collection of 40 brewing yeasts received from the National Collection of Yeast Cultures was tested in duplicate fers under the standard conditions. The total yeast crops in this series varied from 7 to 22 g. per litre, and although evidence of segregation into different classes was not very strong, it seemed probable that three or more classes were present with approximate ranges of 7-, 11-15, and 15- g. per litre. A different type of distribution was observed in duplicate fers with 70 strains isolated from the yeast from one brewery, as is shown in Fig. 4. Here the 50-40 0 4 8 12 16 Fig. 2. Total yeast crops recovered from 60 single fers with one strain of yeast over a period of one year. average yield was 9-5 g. per litre with a range of 7-0-12-0 g. per litre, and the results followed a Poisson-type distribution. As no result differed from the mean by more than 2-5 g. per litre (in spite of the fact that many different worts were used during this series), it must be concluded that no large errors were introduced by unintentional variations in experimental conditions, such as condition of seed yeast, aeration or temperature, and that the experimental errors of the method were reasonably small. Fig. 0 4 8 12 16 4. Total yeast crops recovered from duplicate fers with 70 strains of yeast isolated from the brewing yeast of Brewery "A." average was 12-7 and the range -16 g. per litre. The distribution was more nearly that of a Poisson curve, and the range was comparable to that observed in a single
274 gilliland: yeast reproduction [J. Inst. Brew. culture in Fig. 1, so that it appeared that these strains were homogeneous in this respect. In Fig. 5 duplicate fers with the strains isolated from the yeast from another brewery showed a bi-modal distri bution. The non-flocculent strains gave an average crop of 9-2 g. per litre, while the flocculent strains gave an average crop of 16*2 g. per litre. Wr consistent, the order of increasing yield always being A, B, C, D; the difference between the mean values for any two yeasts is statistically highly significant. As the first and last of these comparisons were done with an interval of about 9 months between them, it is quite clear that the yeast yield in laboratory fers is a strain characteristic. 40 0 * 8 12 16 Total yeast crop (g. pressed /east per litre) Fig. 5. Total yeast crops recovered from duplicate fers with strains of yeast isolated from the brewing yeast of Brewery "B." I: non-flocculent strains; II: flocculent strains. In Fig. 6 the same type of distribution was observed in duplicate fers with strains isolated from another brewery yeast. The bi-modal distribution was quite clear, but in this case it was the non-flocculent strains which gave the higher average yeast crop (17'5 g. per litre), the flocculent strains giving the lower (12-5 g. per litre). In Table I the individual results of 5 separate comparisons between the crops obtained from 4 pure yeast strains are given. It will be seen that the results are quite 0 4 8 12 16 Fig. 6. Total yeast crops recovered from duplicate fers with strains of yeast isolated from the brewing yeast of Brewery "C." I: Non-flocculent strains; II: flocculent strains. Effect of aeration. Yeast reproduction is considerably affected by the amount of oxygen reaching the yeast during its period of growth. The results reported in Table I were obtained in standard fers aerated at starting only. In order to vary the amount of oxygen received during fer some comparisons were carried out in which one set of fers was aerated at starting only, one was aerated at starting and at 22 hr., one was aerated by bubbling air through the wort slowly and continuously during fer, and a TABLE I Yield of Yeast given by 4 Strains or Brbwkr's Yeasts in Period of 0 Months 5 Separate Fers over a Total yield of yeast in g. per litre as pressed yeast in 2-litre laboratory fers; worts aerated at starting only Yeast strain 1st Fer 2nd Fer 3rd Fer 4th Fer 5th Fer 95% Confidence limits for mean A 21 B 14 C C69A D 28 5-2 11-3 151 0 5-6 -8 13-2 19-4 6-4 -8 13-8 18-0 6-0 9-4 13-3 171 5-5 8-6 11-7 17-2 C-34 ±0- -18 ±1-40 13-42 ±1-6 18-44 ±1-83
Vol. 68, 1962] gilliland: yeast reproduction 275 TABLE II Yield of Yeast under Different Conditions of Aeration (Each figure is the average of 4 fers) Yeast strain Yield of yeast in g. per litre as pressed yeast in 2-litre laboratory fers at 18-22 C. Worts aerated at starting only Worts aerated at starting and at 24 hr. Worts aerated by bubbling oxygen through slowly and continuously Yield in g. per litre as pressed yeast in a plugged flask on shaker at 25 C. A 21 B 1164 C S69A D 28 S-S 00 12-5 17-6 0-8 13-7 140 22-3 12-6 18-6 25-1 26-6 25 27 26 28 fourth set was carried out in lightly plugged flasks on a shaker at 25 C. which should give nearly optimum aeration. The yields obtained are given in Table II. Under optimum conditions of aeration these four brewing yeasts give about the same yield. Under conditions approxi mating to a brewery fer, with restricted oxygen, their yields were in the ratios 2:3:4:6; thus, one strain yielded about three times as much yeast as another strain. Intermediate oxygenation intensities gave intermediate yields for each strain, but the high-yielding strains reached their maxi mum yield at a lower level of oxygenation than the low-yielding strains. Conclusions. Since yeast production is uneconomic under brewery conditions it should be reduced to the minimum consistent with good fer behaviour. The results reported above indicate that, in selecting a brewing yeast, the amount of reproduction under brewing conditions should be measured and taken into consideration and that, other things being equal, a strain which does not give high yeast production should be selected. Acknowledgements. The author wishes to express his thanks to Mr. J. P. Lacey for his technical assistance, to Dr. A. K. Mills for his interest and encouragement, and to the Directors of Arthur Guinness Son & Co. (Dublin) Ltd., for their permission to publish this work. References 1. Fowell, R. R., this Journal 1957, 180. 2. Hough, J. S., & Hudson, J. R., this Journal, 61. 241. 3. White, J., & Munns, D. J., this Journal, 51, 170.