Staging the metamorphosis of Drosophila melanogaster

The imaginal discs of Drosophila have been widely used in the study of animal development because they exhibit a wide difference between the time of cell determination and that of the expression of the determined state which results in the formation of the adult cuticle. The embryo is the site of the formation of the discs; the larval instars are the sites of cellular proliferation and the source of the discs for experimentation; and the adult bears the results of such experiments in the form of cuticle markers, either on its surface or on metamorphosed implants in the abdomen; so these phases of development are well known. But the events of metamorphosis have generally received less attention, possibly because of the technical difficulties involved. In the cyclorrhaphous Diptera, including Drosophila, the last larval cuticle is tanned and retained throughout metamorphosis as a hard case inside which the developing imaginal hypoderm is at first extremely delicate. In addition, much of the larval structure disintegrates as the various organ systems are histolysed and reconstituted. Nevertheless, ‘pupation’ has recently attracted attention because during this period the realization of the determined state is achieved.

Methods for staging a continuous process have been developed for other systems, e.g. Bownes (1975), for Drosophila embryos; Nieuwkoop & Faber (1967), for Xenopus tadpoles. No method exists for Drosophila pupae, however, although timed animals were employed by Robertson (1936) and Bodenstein (1950) to construct a timecourse of the anatomical changes seen in sectioned material. This takes no account of differences in rate of development between individuals and between the sexes. Rate studies on timed animals were reviewed by Ashburner & Thompson (1978). Bliss (1926) noted four staging criteria but found their timing to be erratic. Several authors recorded the effect of a diurnal rhythm upon eclosion and the greater time required by males to reach eclosion (e.g. Powsner, 1935; Kalmus, 1940; Bakker & Nelissen, 1963).

This paper presents a practical system for the staging of living prepupae and pupae and a method for collecting some batches of animals all in the same stage from timed plates without laborious hand sorting. However, it must be emphasized that staging is useful precisely because it eliminates variation due to stock and culture condition differences, and therefore any timings adopted from our data must be checked first and, if necessary, modified accordingly. Where possible our teminology follows that recommended by Fraenkel & Bhaskaran (1973) for cyclorrhaphous dipterans in general.

It is hoped that this staging method will facilitate a precise approach to the study of metamorphosis together with its underlying molecular mechanisms regulating gene expression.

All observations were made using an Ore-R stock of D. melanogaster maintained on yeasted Lewis’s medium at 25 ±0·1 °C and exposed to a 12:12 h day/night cycle until collected for use in this study unless otherwise stated in the text. The colour of the ‘yellow body’, identified according to Bodenstein (1950), was checked on Instant Drosophila Food (Carolina Biological Supply Co.). The Malpighian tubules are identified according to Wessing & Eichelberg (1978).

White prepupae were picked from the sides of culture bottles at random times throughout the day/night cycle and placed in pupation plates consisting of sealed Petri dishes lined with moist filter paper. Collections continued for about 20 min, after which time some individuals had become pale brown and were discarded. Any larvae still able to crawl were also discarded. The plates were stored in the dark and withdrawn for observation at various times throughout metamorphosis. Time zero was therefore puparium formation (P.F) + 20 min. max. Records of observations were made at 30 min. intervals and the data were pooled. The flies remained fairly synchronous during puparation, becoming less so as pupation-proceeded. 256 observations were made of 44 plates, each of which contained about 25 flies. In a separate experiment we determined the viability under these conditions to be 97 % (total 946).

The typical sequence of events (i.e. visible changes) was assembled, together with a range of times at which each event had been seen to occur. The animals were examined on a white background using a Wild dissecting microscope with lateral illumination close to the stage, interposing the tip of a pair of watchmaker’s forceps between the lamp and the puparium to cast a shadow over any feature which may be obscured by surface reflexion but which would show up with light scattered inside the animal – this was particularly true of the Malpighian tubules and the meconium. The puparia were wetted for observation.

Five mature males and five mature females were placed in each of four fresh culture bottles and kept in constant darkness to encourage continuous egg laying. On days 11–14 mixed populations of prepupae and pupae could be seen on the sides of the bottles. Adult flies were etherized, sexed, staged and counted. Prepupae and pupae were then removed, washed, staged and counted. Staging was performed on a random sample from a bottle’s total population and lasted no more than 1 h, because development continued during this time. Living animals were employed since freezing and fixation altered their coloration and other staging characteristics.

Animals were collected, washed and staged as above (b). The characteristics employed here as staging criteria may be observed through the puparium, but to facilitate the taking of photographs this cuticle is best removed from pupae. The dissections and subsequent photography were performed on animals submerged in 1 % agar in a tissue culture dish. A Wild dissecting microscope was used with a × 50 objective and fitted with a Photoautomat MPS 55 and camera attachment. Illumination was from above using Volpi Intralux Fibre Optics 150H against a dark field.

The following list of 52 numbered events is presented in the order most typical of animals observed under the conditions defined above. Each stage is given both a numerical and a verbal designation, and each is preceded by the change(s) diagnostic of its beginning. Extreme times at which events have been observed to occur in individual animals are given where available (Ext.). Note that these are not extreme stage durations from the pooled observations, (c.f. Table 3). Numbered changes 1–47 are indicated on Figs. 1–6 for which the following list constitutes a running legend. A developmental series of photographs of whole animals is shown (Figs. 7–12) and photographic details of staging criteria are given (Figs. 22–31).

1. Stops feeding, climbs bottle wall:

   Stage LI (Post-feeding) (Fig. 1)

2. Stops crawling (reversible until 5);

3. Everts anterior spiracles :

   Stage L2 (Spiracles everted) (Fig. 1)

4. Withdraws three apparent anterior segments;

5. Body shortens;

6. Sticks to wall of culture bottle:

   Prepupa

   Stage Pl (Whitepuparium) (Figs. 1, 7)

7. Posterior spiracles and ridge between anterior spiracles tan orange (Ext. h);

8. Stops wriggling completely (Ext. 1 h) ;

9. Puparium becomes brown to the unaided eye (Ext. 20 min−1 h):

   Stage P2 (Brown puparium) (Fig. 1)

10. Male gonads become less distinct; (may still be faintly distinguished in stage P3) (Ext. 0–3 h);

11. Oral armature stops moving permanently (Ext. 1–5 h);

12. Dorsal medial abdominal contractions stop, (i.e. heart stops pumping) (Ext. h);

13. Gas bubble becomes visible within the abdomen (Ext. 3 h). (It is not known how this bubble forms; but it appears to have tracheal associations when dissected out, and it fails to develop in submerged prepupae) :

    Stage P3 (Bubble prepupa) (Figs. 2, 22)

14. Ridge of operculum becomes distinct (Ext. 3–6 h):

    Fraenkel & Bhaskaran (1973) place the larval-pupal apolysis at this point (4–6 h); i.e. the puparium becomes separated from the underlying epidermis, starting anteriorly. The resulting stage is termed the Cryptocephalic pupa. (However, see Figs. 24,25. The movements of gas associated with pupal morphogenesis seem to be involved in the final shedding by the prepupa of the larval endocuticle, and so the larval/pupal apolysis is not strictly completed until a few seconds before head evertion (P4 (ii)). The cryptocephalic pupa is then to be regarded as only a transient stage if physical separation of successive cuticles is taken into account, i.e. stage P4 (ii) represents pupation.)

15. Becomes positively buoyant (Ext. h):

    Stage P4(i) (Buoyant) (Figs. 2, 23)

16. Lateral trunk tracheae become obscured (Ext. h);

17. Dorsal medial abdominal contractions observed, (also ventral; mid-intestinal peristalsis (Robertson, 1936)) (Ext. h);

18. Everted leg and wing discs become visible (translucent);

19. Bubble appears in posterior portion of puparium, displacing the pupa anteriorly, and the bubble within the abdomen disappears (Ext. h):

    Stage P4(ii) (Moving bubble) (Figs. 3, 24–26)

20. Bubble is gradually displaced to the anterior portion of the puparium and the pupa withdraws to the posterior end (Ext. h);

21. Imaginal head sac is everted and the oral armature of the larva is expelled (Ext. h):

    Phanerocephalic pupa

    Stage P5(i) (Malpighian tubules migrating) (Fig. 4)

22. Legs and wings achieve their full extension along the abdomen (Ext. h);

23. Enlarged initial segments of anterior pair of Malpighian tubules move from thorax into abdomen;

24. Translucent patch lacking adhering fat body cells becomes discernible in the middle of the eye region;

25. Pair of white Malpighian tubules becomes visible dorsally in the abdomen when viewed in the shadow of forceps (Ext. 15–48 h) :

    Stage P5(ii) (White Malpigian tubules) (Figs. 5–7; Figs. 4, 8–10)

26. Malpigian tubules become prominent and green (Ext. 13–48 h):

    Stage P6 (Green Malpigian tubules) (Fig. 4)

27. Dark green ‘yellow body’ appears between the anterior ends of the two Malpighian tubule segments, mid-dorsally at the anterior of the abdomen (Ext. 25–46 h):

    Stage P7 (‘Yellow body’) (Fig. 4)

28. The ‘yellow body’ moves back between the Malpighian tubules:

    Fraenkel & Bhaskaran (1973) place the pupal-adult apolysis at this point (34–50 h), i.e. the transparent pupal cuticle separates from the underlying epidermis, starting posteriorly. The resulting stage is termed the Pharate adult;

29. Eye cup becomes yellow at its perimeter (Ext. 43–47 h) :

    Stage P8 (Yellow-eyed) (Figs. 5, 11, 27)

30. Pale yellow pigmentation spreads inwards across the eye;

31. Eye becomes bright yellow;

32. Eye colour changes to amber (Ext. 49–57 h) :

    Stage P9 (Amber) (Fig. 5,12)

33. Eye colour darkens to deep amber;

34. Eyes become very pale pink (Ext. 71–78 h):

    Stage PIO (Red-eye Bald) (Fig. 5)

35. Eyes become bright red ;

36. Orbital and ocellar bristles and vibrissae darken:

    Stage P11(i) (Head bristles) (Fig. 5)

37. Dorsal thoracic micro- and macrochaetes become visible (Ext. h):

    Stage P11(i) (Thoracic bristles) (Figs. 6, 28, 29)

38. ‘Tips’ of folded wings become grey:

    Stage P12(i) (Wing tips grey) (Figs. 6, 13–15)

39. Bristles on abdominal tergites become visible, (Ext. 73–97 h):

40. Wings become grey (Ext. 73–78 h) :

    Stage Pl2(ii) (Wings grey) (Fig. 6)

41. Sex comb darkens;

42. Wings darken to black (Ext. 75–86 h):

    Stage Pl3 (Wings black) (Figs. 6, 16–18, 30, 31)

43. Tarsal bristles darken and claws become black :

    Stage P14 (Mature bristles) (Figs. 6, 19, 20)

44. Green patch (the meconium) appears dorsally at the posterior tip of the abdomen (Ext. 87–103 h):

    Stage P15(i) (Meconium) (Figs. 6, 21)

45. Malpighian tubules and ‘yellow body’ become obscured by tanning of the tergites (Ext. 90–103 h);

46. Legs are seen to twitch: flies able to walk if the puparium is removed prematurely (Ext. 90 h) ;

47. Ptilinum expands by hydrostatic pressure, opening the operculum (Ext. 93–c. 105 h):

    Stage Pl5(ii) (Eclosion)

48. Eclosion is completed :

    Exarate adult

    Stage A1 (Newly eclosed) (Fig. 30)

49. Wings unfold, flatten and harden;

50. Abdomen becomes broader :

    Stage A2 (Wings extended)

51. Abdominal tergites tan to shiny brown:

    Stage A3 (Tergites tanned)

Two other species of Drosophila were checked for the presence of the stages listed above. D. simulons (sibling species) is very similar to D. melanogaster except for the site of metamorphosis (on the food) which makes it impossible to collect LI (post-feeding) as wandering larvae. D. pseudo-obscura (obscura group) also shows this character to a lesser extent, but some individuals do move away from the food. However, the diagnostic feature of Stage P7 is lacking in this species since the ‘yellow body’ is not observed until P8 and even at P9 has still to move into position between the M.t.’s. In both spp. the larval male gonads and the sex comb(s) (two in D. pseudo-obscura) are available as criteria for sexing individuals. In general then, the stages of metamorphosis in related species are similar to D. melanogaster though their proportional durations may differ.

Flies were scored from cultures of four ages, namely 11, 12, 13 and 14 days. When the culture contained both Fl adults in stage A3 and post-feeding larvae (stage LI) it was deemed acceptable. Since laying commenced on day 0 and eclosion on day 10, this composite population approximated to a continuous age distribution, and mean percentage frequencies were then obtained for each stage within it. Table 1 shows the sampling procedure and the proportions L:P: A involved.

Table 2 shows the observed frequencies of each stage in the various cultures and their calculated means. However, in order to relate the frequencies to a known time-span (P.F. to eclosion), the scores for all stages (L1–A2) are expressed as percentages of the total Pl-Pl5 taken from Table 1 (*). The time taken from P.F. to eclosion was estimated from the observation of pupation plates. The mid-point of the range of eclosion times was about 100 h, that for females being 98 ± 5 h, and that for males 102 ± 4 h, giving a difference of 4 h which is in agreement with that reported by Powsner (1935), i.e. 3·7 h at 25 °C. So for a mixed-sex population the approximate duration of the period Pl-Pl5 may be taken to be 100 ± 7 h. If we now assume that the size of any given stage in a random sample is directly related to the time spent in that stage, then a 1 % stage size indicates a duration for that stage of 1 h. This principle is employed in Table 3 to describe a timécourse for metamorphosis.

The timecourse obtained from stage frequencies may be compared with timed extremes of stage durations which have been derived from pupation plate data and are shown in Table 3. For the most part these extremes will lie on either side of the frequency-ranges. Timecourses may also be calculated for males and females separately and these, when compared, reveal periods during which we can expect animals of both sexes to be in the same stage. This comparison is made in Table 3. A dash indicates that no such overlap occurs. We now consider the general timecourse and limit the durations of its component stages using both the extreme timings and the predicted male/female overlaps. In this way we define periods of development during which a high proportion of the animals collected together at P.F. may be expected to be in the same stage, i.e. symphasic. Such ‘safe’ collection periods are indicated at Fig. 32.

Note that some stages cannot be obtained homogeneously by timing alone, (e.g. P6) ; and that in any case no sub-stages (i or ii) may be reliably obtained without there first being a visual check performed. So large symphasic batches of animals may be collected by plating prepupae at the beginning of Pl, maintaining them at 25 °C in continuous darkness, and harvesting stages P2–P14 at the times shown. It must be emphasized, however, that these absolute times may vary between stocks, so, at least initially, batches must be checked and times modified if necessary. It is often convenient to segregate males and females as larvae or prepupae and time them separately, but still hand-sorting is required for the short stages Pl l(i) –P13. Note also that after P2 no sexing of flies may be done until P12(ii), though the females are generally larger than the males, so we recommend prior sexing of prepupae in stage Pl for experimental analysis of these stages. This is best done with wet animals on black filter paper. Some difficulty may be encountered at first in drawing the line between two stages, e.g. P8 (Yellow-eyed) and P9 (Amber). To avoid ambiguity in the resulting batches it is best to sort the plate into three sub-groups and discard the animals of intermediate colour. Subdivision of stages may also be performed given a large enough number of flies by employing an internal standard within a given plate: e.g. ‘early P9* may be collected when most of the members of the plate are still in P8, and similarly ‘late P8’ when most are already in P9. With practice, colour sorting may become sufficiently reproducible, though subjective, to allow of the collection of’mid-stage x’ in addition to ‘early’ and ‘late’ by sub-sorting of a plate. It is not convenient, however, to sub-time within a given stage, due to the large number of Pls required to yield a small group of synchronous animals later on.

The narrowness of the ‘safe’ periods (Fig. 32) where all flies are in the same stage for a period of time confirms that staging rather than timing is essential to studies of Drosophila metamorphosis since the members of a timed batch are unlikely to fall unequivocally into one developmental stage. It is possible that the extreme timings are partly due to the effect of the diurnal rhythm, as the prepupae were withdrawn from a day/night cycle and had therefore pupariated under these conditions. It seems likely, however, that the diurnal rhythm induces a block in the latter part of pupation: thus some flies eclose before the meconium becomes visible, while others retain this character but do not eclose for some time. But certainly the difference between males and females is established early, to the extent that, if pupae are segregated on the basis of early or late accession to P5 (i.e. time of head eversion), then they are also segregated sexually with an accuracy of about 70 %. The effects of the day/night cycle are decreased in the estimation of stage durations by frequency, as both oviposition and P.F. took place in the dark to avoid gating.

It is interesting to note that, whilst the timing and frequency data generally correspond, there is one short stage, that of bubble displacement and head eversion (P4(ii)) which is exceptional (see *Table 3). On the basis of frequency the duration of this stage is found to be 0·2 h, i.e. 12 min, which agrees with several observations of the process in action. It is predicted to begin at 8 h post-P.F., and repeated timings have shown P4(ii) to begin no earlier than 12 h at 25 °C. However, this is of no practical significance since P4(ii) can be most conveniently collected at 18 °C, 24–25 h after P.F. Here we assume that morphogenetic and other events employed in staging are not unrelated to the underlying biochemistry and so will not be uncoupled from the metabolism of the organism when metamorphosis occurs at other temperatures. Should the assumption prove to be justified, we may exploit other temperatures to obtain all stages at convenient times of day. Table 4 shows factors by which durations Pl–15 at 18, 22 and 29 °C are shown in the literature to differ from that at 25 °C, (Ludwig & Cable, 1933; Powsner, 1935; Davidson, 1944). If we further assume that the proportions of time devoted to the various stages will remain constant, then these factors may be used, together with Table 3, to predict collection times for stages at other temperatures if these are more convenient; but now hand sorting becomes more than ever essential.

The abdominal bristles are found to tan unreliably in stage P12(i), and they have been observed to appear asynchronously across the abdomen, e.g. they appeared (tanned) first across the second tergite, then along the flanks of the abdomen, then dorsally. This may be related to Trepte’s observations (1980) on Sarcophaga barbata, where the chromosome puffing pattern of bristle cells and the associated process of bristle differentiation were found to be autonomous in different regions of the body surface. In Drosophila the tanning of these bristles cannot be employed as a strict diagnostic feature for staging. The tanning of the tarsi is less variable but not unique to stage Pl 3, although the tanning in the leg does appear to follow a sequence in which the tarsi are tanned last. This is in contrast to the order of acquisition of differentiative competence in the leg disc, (distoproximal to medial; Schubiger, 1974) and the sequence of bristle differentiation in the pupal leg, (distal to proximal; authors’ observations).

The staging system presented in this paper constitutes a more precise approach to the problems of insect metamorphosis than has previously been available to Drosophila workers, who have used timed batches of animals, often of mixed sexes. The visible changes described here presumably do reflect the underlying metabolic state of the animal, so symphasic individuals may be assumed to be reasonably synchronous in metabolic terms, though they may well be of widely differing chronological ages. The hormonal basis of morphogenesis and differentiation could now be readily investigated by assaying hormone titres of the various stages to produce a profile for each sex (in preparation). Protein synthesis may be examined, e.g. in the tanning of cuticle and in those tissues – brain, fat body and Malpighian tubules -which survive metamorphosis. Finally, the involvement of hormones and other factors in the control of gene expression may be studied in relation to the secretion of both pupal and adult cuticles, the latter having well-characterized surface markers whose sequential appearance may be followed in successive stages. The clustering of stages at P10–P12 is fortunate since by this time asynchrony is most significant.

We should like to thank Jo Rennie for her excellent assistance with the photography. The work was supported by the Science Research Council.