The Nutritional Requirements of Drosophila Melanogaster

Drosophila melanogaster lives naturally on various live yeasts that grow in the presence of bacteria, moulds and other micro-organisms, on decaying organic matter. In the laboratory live or autoclaved yeast is perfectly adequate for development, but media on which live yeasts are growing are very heterogeneous; further, the quantity and quality of the available food changes at successive stages in the developing culture. Change in the nature of the substrate probably favours some yeast strains more than others, and thereby alters the food value of the medium (Robertson & Sang, 1944). Frequently such alterations in the medium are unimportant in genetic experiments and may pass unnoticed, although probably much of the environmentally determined variability in exhibition of genes is attributable to this cause. In the case of some genes there is evidence of control of exhibition by specific chemical compounds. Study of such mutants, sensitive to changes in the diet, can be studied best under aseptic conditions, while a completely chemically defined medium might permit detection of these components of the diet which influence exhibition and perhaps throw considerable light on the biochemistry of gene action. Such a medium would also be useful for comparing the effects of different members of allelomorphic series.

Various authors, notably Tatum (1939, 1941) and Gordon & Sang (1941), have attacked aspects of this problem. In particular, Tatum claimed that riboflavine, nicotinic acid and pyridoxin, in addition to an amino-acid source (casein hydrolysate), tryptophane and cholesterol were essential for the development of Drosophila, but found it necessary to add a yeast residue consisting of the insoluble fraction remaining after the treatment of autolysed yeast with water and hot alcohol. A somewhat similar ‘insoluble’ fraction was described by Goldberg, de Meillon & Lavoipierre (1945) for Aëdes aegypti. These workers also found that folic acid, and probably biotin, were necessary too.

This paper describes attempts to develop a chemically defined medium, together with methods of sterile culture. A preliminary report has already been published (Begg & Robertson, 1948; Begg, 1949).

Most of the tests have been carried out with an inbred wild type Oregon strain of Drosophila melanogaster. Eggs are collected by allowing flies to feed for 3–4 days on a rich yeast diet, after which they are transferred to bottles fitted with watch-glasses filled with agar (Robertson & Sang, 1944). The eggs are removed by means of platinum spoons fixed into glass holders.

Great difficulty was encountered in procuring sterile cultures. Many variants have been tried of which the two most successful are as follows :

• (1) Eggs are collected and transferred to a test-tube containing sterile water which is rotated about its long axis for several minutes after which the eggs are allowed to settle. The water is then replaced by 50% ethyl alcohol. The tube is again rotated slowly for 3–4 min. and the eggs are allowed to collect at the bottom of the tube, when they are removed with a sterile pipette and transferred to a tube containing 5 % antiformin in 10% formaldehyde solution (Glaser, 1943). After 10 min. they are removed to another tube of 50% alcohol, and after settling are transferred in batches of about twenty to small round-bottomed tubes (⁠ × 2 in.), containing 50% alcohol. Two further changes are made into similar tubes, so that the total immersion time at this stage is about 25 min. The eggs are then transferred, with the usual precautions, to the sterile culture media. This is best carried out by using pipettes made by drawing out glass tube, squaring the ends off very carefully and plugging with cotton-wool. With the aid of a length of rubber tubing attached to the top of the pipette the operator can suck the eggs into the end of the pipette and gently blow out excess alcohol, while holding the pipette against the glass bottom to prevent the eggs escaping. The entire operation can be carried out by one person after practice, but at first it is safer for two persons to work together.

• (2) The treatment here is identical with that already described until the immersion in antiformin and formaldehyde. After this step, eggs are transferred to a short sterile glass tube closed at one end by a platinum grid (100 mesh/in.); they can then be moved from one solution to another in this container. After 10 min. in a mercuric chloride solution, ^* the eggs are given four 5 min. washes in alcohol. To ensure constant agitation of the sterilizing liquid we have used an electromagnet in conjunction with a small steel washer. Finally, the eggs are removed in batches to the small tubes and thence into the cultures without further transfers.

All cultures are incubated at 25° C. Any culture showing signs of infection is examined by smears and subcultured on to agar slopes of peptone-malt-beef-extract. The commonest infective agency is B. subtilis.

Since live or dead yeast is quite adequate for normal development, it was decided to proceed by adding to a basal medium, containing known or likely constituents, extracts of yeast which could be progressively fractionated and replaced by known compounds. To begin with, our primary concern was to obtain basal media which, with the addition of preferably water-soluble yeast extracts, would permit complete development. This involved largely empirical modifications of both the basal medium and the yeast extracts ; in consequence, many of the early experiments cannot be strictly compared one with another, and are therefore not dealt with. Throughout the work a very large number of yeast autolysate fractions have been tried. Naturally, most of these proved unsatisfactory, and we shall refer only to the more interesting cases.

The primary extract is prepared by autolysing live brewers’ yeast for 4 days at 35° C. under reflux with ethyl ether, and filtering and centrifuging the resulting mass until a cell-free solution is obtained. This is then dried under reduced pressure at 40–50 ° C. Apparently pH conditions alter the composition of the autolysate. Thus when HC1 is added to the autolysing mass, little or no growth occurs when larvae are reared on the water-soluble autolysate, while addition of 1 % NH₄OH at the start of autolysis yields an extract which is perfectly adequate for normal growth either alone or in conjunction with basal medium. The latter water-soluble extract, prepared in bulk, provided the starting material for the various fractionations, and is referred to throughout as yeast autolysate. A highly homogeneous aseptic medium can be prepared by making up an agar gel from this water-soluble extract (0·3 g. autolysate to 5 ml. gel).

When the culture tubes are removed from the autoclave, they are agitated in cold running water to secure even distribution of the components and then sloped or rotated to ensure that a wide area of gel is presented to the larvae ; consistency of the gel is an important factor in influencing early growth rate.

The basal medium used in the first experiments included the following constituents: casein, gelatin, aneurin hydrochloride, riboflavine, nicotinic acid, calcium pantothenate, agar, water and salts (or Tatum’s salt mixture ^*). Addition of 0·3 mg. of yeast autolysate was adequate for development within 11 days, compared with 8–9 days on live yeast, and 9–10 days on autoclaved yeast. Reduction of this amount lengthened the larval period. Exhaustive extraction of the dried autolyaste with ether removed no essential constituents.

In the next experiments, we used a fuller basal medium containing, in addition to the components already described, ergosterol, cholesterol, dextrin and pyridoxin hydrochloride. By itself, this medium permitted little larval growth. Addition of biotin, however, did improve growth somewhat, although no pupation enused. When tryptophane and inositol were added, a few small malformed pupae developed after a larval period of 22 days, compared with 6–7 days between egg and pupa on the water-soluble autolysate. The detailed compositions of this and later basal media are quoted in the appendix. This more complete medium (A) was used for experiments on the fractionation of the autolysate. This was started by adding ethyl alcohol to the filtrate from an aqueous solution of the extract to bring the concentration to 50% by volume. The precipitate was filtered off and the filtrate dried under reduced pressure at 40–50° C.

Table 1 shows that 0·3 g. of this extract is adequate for normal development. Since similar treatment with acetone gave similar results, these are omitted here. Treatment of the 50% alcohol-soluble constituents in aqueous solution with activated charcoal removed essential constituents, both at low and high pH. We were unable to recover these by elution.

Higher concentrations of ethyl alcohol left behind essential factors (Table 2). Pupation was found generally to be incomplete when the high alcohol-soluble fractions were supplied ; no adults emerged, and larvae usually died in the course of unsuccessful pupation, although larval growth rate was almost normal. Earlier experiments indicated that autolysate could be replaced to some extent by liver extract or liver concentrate.

In view of the work of Goldberg et al. (1944) and of our own results, it seemed worth while testing whether failure to pupate might not be due to lack of folic acid. Since folic acid may be removed from solution by precipitation with basic lead acetate (Laland & Klem, 1936), we treated the 50% alcohol-soluble fraction with acetate until no further precipitation occurred. Lead was removed in the usual manner with H₂S; when the gallocyanine test for minute traces of lead proved negative, the solution was dried under reduced pressure and 0·2 g. added to a slightly altered basal medium (B). This medium contained nucleic acid, which was included in view of its high concentration in yeast. The larvae became fully grown at almost the normal rate but failed to pupate. Many crawled round the cultures for days before they finally died. Later, when pure folic acid became available, we found that addition of 1 /zg. per culture to the basal medium plus autolysate treated with lead acetate allowed emergence in 11 days. Likewise, addition of folic acid to the 90% alcohol-soluble fraction leads to similar successful pupation (Table 3).

Further proof of the need for folic acid is shown by the effect of reducing the concentration below 1 μg. per culture. Although the larvae grow quickly, pupation is greatly delayed. It is possible, in view of the low concentration, that the larvae burrow in the medium until they have accumulated a minimal quantity of folic acid, whereafter pupation ensues. From this point onwards, 2 μg. folic acid/culture were added to the basal medium.

Shortly after this, Schultz, St Laurence & Newmeyer (1946) reported the need for choline ; this was confirmed, since addition of 1 mg. choline chloride to the basal medium containing folic acid allowed complete though slow development, taking on the average about 15 days. However, it was a considerable advance to have a synthetic medium which would permit complete, although slow, development. Growth-accelerating factor(s) remained to be identified.

Since we possessed a chemically defined medium which would permit complete, though slow development, we now turned to those factors concerned with the acceleration of larval growth. Tatum (1939), Goldberg et al. (1944) and Schultz et al. (1946) claim that an insoluble yeast fraction is necessary for rapid larval development. This was not consistent with our findings that, in the presence of folic acid, even the 90% alcohol soluble fraction is adequate for rapid development. It seemed likely that differences in the method of preparing the autolysate (unfortunately these authors do not describe their methods of preparation) accounted for the discrepancy, since we have seen already that pH conditions greatly affect the nutritional quality of the preparation. In particular, autolysis in the presence of HC1 yields an inadequate extract, while when 1 % NH₄OH is added the extract is completely adequate. Apparently some essential component(s) was either left behind or destroyed in the presence of HC1. To test the former possibility, 400 g. live yeast were boiled for ⁠. with 500 ml. 0·2 N-HCI. 170 ml. of absolute ethyl alcohol were added to the resulting mass to ease filtration, and the residue was washed several times with 50% alcohol and then dried over a water bath. When 50 mg. were added to basal medium B which contained, in addition, folic acid, growth was completed in 11 days. Of course, such a residue contains the cell-wall constituents of yeast and is an extremely heterogeneous mixture. However, the tests proved the presence of a semi-essential insoluble factor after this treatment. The factor proved resistant to prolonged oxidation and was not removed by Soxhlet extraction with various organic solvents, including pyridine. Yeast ash did not replace this fraction. Further, when boiled for 1 hr. with concentrations of HC1 above 10% inactivation took place. However, when the pyridine extracted HC1 residue was boiled with 2% NaHCO₃ and filtered, we secured a filtrate which, on neutralization with dilute HC1 and addition of alcohol, yielded an active precipitate. When this was added at the rate of 40 mg./culture to the most complete medium development took place in 11 days.

A similarly active precipitate is secured from the treatment of the 2% NaHCO₃ extract of whole yeast with HC1 and alcohol. Such an extract is best prepared by boiling about 800 g. live brewers’ yeast with 1000 ml. 2% NaHCO₃ for about 15 min., or until the mass begins to smell distinctly of ammonia. Several filtrations are generally necessary. After treatment of the cell-free filtrate with acid and alcohol, the resulting bulky precipitate is collected on a Buchner filter, washed repeatedly with 50% alcohol, then absolute alcohol and finally ether; it is then dried. The time of boiling apparently influences the potency of the extract.

Thymonucleic acid appeared to have a deleterious effect on development. But addition of the base thymine alone in a concentration of 25 μg. per culture accelerated development (Begg, 1949). In the presence of 22 mg. of the NaHCO₃ extract development is completed in about 10 days. (See also Ju-Haw Chu, 1945.)

The solubility properties of this fraction suggested at first that our medium might be deficient in nucleic acid, although this seemed unlikely from earlier experiments. However, increase in the content of this component leads to no change in development time. The fraction shows a positive HCl-aniline reaction for pentoses and also the Sakaguchi reaction for arginine, suggesting that the protamine adjunct of yeast nucleoprotein might be important, although addition of 20 mg. arginine hydrochloride per culture lead to no improvement. Although these experiments are not far enough advanced to suggest a definite conclusion, it is perhaps significant that addition of globin, prepared from haemoglobin by hydrolysis and salting out, produces pupae of greater size; and that in some nutritional studies it has been suggested that whole casein is sometimes more effective in allowing rapid development, than are its hydrolysis products.

These observations suggest that Drosophila may require for rapid growth either some amino-acid supplied in a particular form or possibly a particular short polypeptide molecule. Woolley’s (1945) work with strepogenin is suggestive in this connexion. Work is continuing in the hope of identifying this missing component.

The best synthetic medium we have to date has the following composition

It is possible that this does not constitute an optimum relative concentration of components, and some of the vitamins may even be redundant. Further, for completely normal development we still have to add an unknown fraction which is alkali-soluble. Even in the absence of this fraction, however, there are many purposes for which this medium should prove useful. But naturally its use is limited in studies on the relation of gene exhibition to the composition of the diet, since growth still takes a little longer than normally. This may appear unimportant, but since the delay is primarily in the larval stage, we may be dealing with about 25% greater length of larval life, and disregard of this may lead to erroneous conclusions; for the developmental period is recognized as a factor of prime importance in all such work.

Part of this work was carried out in 1945, when one of us (M.B.) held a Carnegie Teaching Fellowship. We wish to thank the Trust for their financial help.