Studies on Growth Acceleration in Protozoa and Yeast

In direct contrast, Woodruff (1911) showed that the larger the volume of culture medium, the better the growth of Paramecium. He concluded that Paramecium liberated a toxic substance into the medium, and that in the smaller volumes its concentration soon became sufficient to retard division. We have here a direct apposition of ideas. Following the publications of Robertson (1921, 1923), many workers tried to substantiate his results with varied success. Cutler and Crump (1923–1925) in a series of papers reported their failure to obtain any evidence of the allelocatalyst described by Robertson. In fact, these workers’ results confirmed Woodruff’s observations on Paramecium. Greenleaf (1924, 1926) showed that Paramecium caudatum, P. aurelia and Pleurotricha lanceolata grew better in larger volumes (forty drops) than in small ones (two or five drops). Calkins (1926) obtained similar results on Uroleptus mobilis. Myers (1927) also drew the same conclusion from work on P. caudatum. From this last paper, however, no serious deductions can be made, since the author obtained no divisions over periods of 22 to 57 hours. The works of Calkins, Woodruff, and Metalnikow have shown that P. caudatum can divide almost twice in a day. Therefore when a worker fails to get one division in 36 hours one is forced to suspect that his culture conditions were unfavourable.

On the other side, Robertson has defended his own case by attempting to point out flaws in the technique of Cutler and Crump and of Greenleaf, his chief claim being that the organisms were not washed before isolation. This feature was quickly met by Cutler and Crump (1925) and by Peskett (1925), and shown to be of no importance; washed or unwashed the organisms showed no allelocatalytic phenomena. One paper has recently appeared that purports to substantiate Robertson. Yocom (1928), working on Oxytricha, concluded that an organism divided faster in four drops than in ten. A review of the figures, however, reveals the fact that there are wide variations in the numbers of organisms obtained. It cannot be too strongly stressed that the arithmetic mean of several numbers does not justify deductions from these means without reference to their origin. The average of a series of numbers of totally different magnitude, as was obtained by Yocom, increases the probable error to such an extent that small variations between averages mean nothing. Until a stricter control is attained over the division rate, data of this type cannot be regarded as critical.

The aim of this investigation is not to place on record another case pro or con auto-catalysis, but to explain the origin of these divergent theories.

The main features of the culture medium have already been described (Darby, 1928 b), and only the special precautions for this investigation will be given here. The organisms used were Paramecium caudatum, P. aurelia, and Stylonichia pustulata. The organisms were washed in glass-distilled water, which was buffered at pH 7·0 with KH₂PO₄ and NaOH, and had a final molarity of 1/600. Washing was accomplished by moving a single protozoan through six changes of water by means of a micro-pipette. The efficacy of this method was demonstrated by Hargitt and Fray in 1917. The stocks had been kept under isolation conditions for 6 months, and were of a homozygous nature. The division rate had been carefully followed before these experiments were undertaken.

A series of sister Paramecium (caudatum) in a medium at pH 6·95 ± 0·05 were taken from a single dish, washed, and isolated into a range of volumes. The organisms had just completed their fourth division from a single isolation, and thus gave sixteen cells to choose from for the experiment. The tables below give the results.

Before discussing the above results another series of experiments will be presented. In these the protozoa were isolated daily into new culture medium of the same volume, and the number of divisions rather than the number of organisms recorded.

The foregoing data support the observations of Woodruff (1911) and show no allelocatalytic effect. The division rate of P. caudatum under daily isolation conditions (Table II) conformed to that of the cells in the higher volumes up to 48 hours. After about 72 hours the retardation that appeared earlier in the smaller volumes now started to affect the entire experiment (Table I).

There is one objection that might be raised against the above experiments; namely, that the volume was not large enough to produce an effective dilution of the catalyst. Robertson worked with a very small ciliate, Enchelys, and has claimed that the ratio between organism and volume is the important factor. Cutler and Crump (1925) complied with.this requirement, but still failed to obtain any allelo-catalysis. Paramecium caudatum is much larger than Enchelys, and 2·2 c.c. might not fulfil.the ratio requirement. A much larger volume was therefore tried. Fifteen hay infusions of 1000 c.c. each were made up and carefully sterilised in their containers several times. In ten of these, single Paramecia were isolated by a micropipette, while the other five were left sterile. To avoid wasting time, it was calculated that, at the rate of two divisions per day, the organisms would be numerous in 8 days. On the eighth day an abundant fauna was found in all the inoculated cultures; the controls were still sterile.

It seems unnecessary to add further evidence; but since these studies were commenced before the literature of Cutler and Crump and others had appeared, other experiments were performed with Stylonichia. pustulata and P. aurelia. The results substantiate those of the previous experiments and confirm the results of Woodruff (1913). Two examples are given in Tables III and IV.

Lack of food might possibly account for the poor growth in the smaller volumes, especially as Cutler and Crump have shown that the growth of Colpidium colpoda is dependent on the number of bacteria present. This could not have been the limiting factor in my experiments, however, as smears of the culture fluids showed an abundant bacterial flora. The evidence from the 0·1 c.c. 10-day isolation count is also against such a hypothesis, since the figures show better growth under these conditions, in spite of the bacterial dilution which took place at each successive daily isolation.

Previous workers have gone no further than to disprove Robertson’s thesis. But his findings on Enchelys still have to be explained, especially as some of Greenleaf’s (1926) cultures tended to show a similar effect for P. aurelia. On the other side Woodruff (1911, 1913) attributed the retardation of growth in small volumes to the excretory products of the organism, but the nature of these products was left to conjecture. Weatherby (1927) showed that P. caudatum excreted urea, which was hydrolysed by the bacteria to give ammonia. Similar observations were made independently by myself (Darby, 1928 a) in connection with other studies. The liberation of urea into the medium brings about two separate results; first, the H-ion concentration of the medium is changed, and second, owing to hydrolysis by bacteria free ammonia is produced, and this compound is toxic to the organism.

The first of these results will be discussed briefly with reference to its effect on division rate. Still other aspects of the subject are treated in a previous paper (Darby, 1928 b). The division rate is considerably altered by changing the pH of the medium, and when other factors are kept constant, each species exhibits a maximum rate at a particular H-ion concentration. It must not be assumed, however, that the optimum is the same for all forms, for it varies considerably. Thus a rise (or fall) of pH might bring a given culture medium towards or away from the optimum of a given organism, depending on whether the pH of the medium was originally below (or above) the optimum or not. Although Robertson recognised that the pH of his cultures varied from one to another, and that the organisms were affected by pH changes, he failed to control this factor.

It is obviously impossible to expect equal division rates at different H-ion concentrations any more than one would expect equal division rates at different temperatures. For present purposes it is important to know the significance of such changes in pH as may occur in a 48-hour culture. If cultures are started at pH 6·6 with single Paramecia (caudatum), at the end of 24 hours they contain four organisms each. The second 24-hour period ought therefore to give 16 or 17, but instead gives 21·23. The medium on being tested gives a reaction of pH 7 ·0. The division rate at pH 6·6 is two per day, while at pH 7·0 it is 2·3. A decrease can likewise occur by starting the culture at pH 7·0. If it is poorly buffered, it will change to 7·2; so that the first day will average 2·3 divisions, while the second averages 1·9. In the case of P. caudatum this variation is noticeable, but slight; but in the case of a rapidly dividing form such as Stylonichia pustulata, it is more striking. A single Stylonichia at pH 7·2 will produce eight organisms in 24 hours. If the pH of the medium shifts to 8·05, then the subsequent period will give 179 organisms, as against 64 at a constant pH and division rate.

In this connection, Woodruff’s (1913) observation is of interest; namely, that while the excretion products of Paramecia were toxic to themselves, they were not toxic to Stylonichia. This is entirely explicable, owing to the fact that a pH that is toxic to P. caudatum (7·4 for example) can be raised considerably and yet come nearer to the optimum for Stylonichia.

With reference to the toxicity of ammonia per se, it must be borne in mind that its effect is separate from that of pH. As explained above, by changing the pH of a weakly buffered medium it may or may not cause an unfavourable environment. But even in a well-buffered medium in which no pH changes occur, the ammonia is still free to penetrate the cells with great ease (Jacobs, 1922). The hydrolysis of ammonium salts ensures the presence of small amounts of free NH₃, even in acid solutions.

To the above changes in the medium, must be added the buffering effect of the organisms themselves, quite irrespective of their excretory products. Organisms in an unbuffered solution tend to change the pH to a definite value. In the case of P. caudatum there was a noticeable tendency to bring the solution to pH 7·0 and keep it there. This point coincided with the optimum of the organism, but it is not known whether this would always be the case. Maclnnes (1922) found a similar feature in the growth of moulds, and I have observed the same tendency with yeast.

From the experimental evidence presented here, a possible explanation of Robertson’s results appears to be that his medium was away from the optimum pH and very weakly buffered. The organisms tended to change it, and the more he put in at a time, the quicker was the improvement. In the smaller.volumes it was easier to bring this about than in the large ones. In too small a volume, however, the accumulation of ammonia would soon retard growth ; while in the very large volumes the buffering power of the organisms would be ineffective

The general view on the topic of auto-catalysis is that while it may not hold in the case of protozoa, it certainly has been demonstrated in the case of yeast. This investigation was therefore carried back to the original source. In 1901 Wildiers sought to explain the famous controversy between Pasteur and Liebig by the discovery of a substance unknown to the earlier workers. He claimed that there was another essential which had escaped Pasteur’s notice. This new substance Wildiers called “bios.” In 1902 Amand, working at the same laboratory, and using the same technique, substantiated Wildiers’ findings.

Wildiers’ work has been seriously challenged. In 1921 MacDonald and McCollum, using very pure chemicals, were unable to find any evidence for the necessity of “bios.” To this publication Prof. Ide (1921) of the Institut de Carnoy replied on behalf of his students, Wildiers and Amand, that the yeast cultures of Mac Donald and McCollum must have been contaminated with bacteria (a condition these authors had shown not to be the Case); and secondly, that there were two types of growth, one without “bios” which was slow, and one with bios, which was rapid. This recalls the fact that Wildiers reported a slow growth in cultures which had at first shown no signs of development. Throughout the literature there has been a confusion of two totally different concepts—one of a substance necessary for growth, the other of a substance which accelerates growth. If Prof. Ide’s suggestion is correct, then Wildiers was only dealing with a growth accelerator at the time in question, and not with a factor intrinsically essential for growth.

Since the experiments were performed, the sugar used by Wildiers has been shown to contain an impurity which is deemed necessary for the growth of yeast. This is an entirely different thing from the substance suggested by Wildiers himself. His experiment showed that on adding two drops of inoculum to 125 gm. of medium no growth was obtained, while on adding five drops of the self-same mother culture good growth was obtained. As the same sugar (commercial white cane sugar) was used in all the experiments, it cannot have been the source of the difference in growth. The only variable was the amount of inoculum, as stated before. Recently Peskett (1924, 1925) has attempted to reproduce this volume effect with single cells. He has failed to obtain any difference in growth in various volumes of medium. He has, however, obtained better growth with cultures containing a substance prepared by Eddy, Kerr, and Williams (1924), and named by them “bios.” Peskett concluded that there was no allelocatalytic phenomenon in yeast growth, and that the differences between Pasteur’s and Liebig’s results might have depended on the amount of “bios” added. I shall attempt to show that Wildiers’ and Amand’s results can be explained without reference to “bios”

The first experiment was designed to determine the pH optimum for yeast. The only information available was that of Pearsall and Ewing (1925), that the pH optimum lies between 4·6 and 5·0. A series of experiments was therefore carried out with Pasteur’s medium. A large volume of medium was made up, using saccharose once recrystallised. From this common solution, smaller portions were prepared, differing only in pH. The pH was adjusted with small amounts of dilute HC1. Growth was measured by counts on a Bausch and Lomb haemocytometer. The results are presented in Table V.

In the above experiments, the optimum is obviously at pH 4·4. This is shown in two distinct ways. (I) There is much more growth in the cultures that started at pH 4·4. (II) The other cultures tended to shift toward 4·4 as growth progressed and the number of cells increased. This is the result of the buffering action exerted by the organisms themselves referred to previously. For the purpose of comparison, the media of both Wildiers and Amand were made up and tested. The pH of both lay between 7·2 and 7·4. Amand’s medium has the following composition :

A survey of the above substances will show that the medium could not possibly be acid, or give a reaction anywhere near pH 4·4. The discrepancies in pH between the optimum (4·4), Pasteur’s medium (5·2), and Wildiers’ medium (7·2), led me to investigate the amounts of growth resulting from varied quantities of inoculum in neutralised Pasteur’s medium. The neutralisation (to pH 7·0 by means of dilute NaOH) made Pasteur’s medium more comparable to Wildiers’. Into 100 c.c. portions of this modified Pasteur medium were placed inocula of various sizes. Table VI gives the results of this experiment.

From the above experiments it is clear that the result of a large inoculum is to decrease the initial period of adverse H-ion concentration. This allows the cells to grow better in the medium, and the subsequent 24 hours shows still more growth and a further lowering of the pH. On the other hand, the small inoculum does not change the medium to any appreciable extent. Cell division, therefore, occurs very slowly under these adverse conditions. Growth is, however, obtained from the smaller inocula, but the cells are somewhat abnormal, as Wildiers himself stated.

The entire concept of allelocatalysis seems to me to have been based on comparisons of the growth of organisms in media very different from one another, but whose differences escaped the observers. The significance of the H-ion concentration was unsuspected by the earlier workers, with the notable exception of Pasteur, whose comprehension of the subject may be gathered from the following : “It is easy to show that these differences in temperature which are required to secure organic liquids from ultimate change depend exclusively upon the state of the liquids, their nature and above all upon the conditions which affect their neutrality whether towards acids or bases.”

1. Paramecium if grown in satisfactory media exhibit no allelocatalytic phenomena when the volume of the medium is decreased.

2. The observations of Wildiers on the growth of yeast are explicable by the buffering action of the organisms on a hyperalkaline medium. The original pH of Wildiers’ medium was approximately 7·2 ; the optimum for growth is 4·4.

I wish to express my appreciation to Dr E. J. Allen, F.R.S., and to Mr C. F. A., Pantin M.A., for the courtesies extended to me at the Plymouth Laboratory.