ABSTRACT
Results are reported from the culturing in vitro of cells from individual early gastrulae of the following four groups of X-linked embryonic lethal mutants of Drosophila melanogaster. (1) Notch lethals. Five Notch mutants were studied which have been reported to give similar abnormalities in whole embryos: the nervous system displays a three-fold hypertrophy as part of a shift in the pattern of differentiation within ectodermal derivatives, and mesodermal derivatives do not differentiate. An hypertrophy of nerve was found in cell cultures prepared from embryos of all five mutants. In addition, four of the five alleles consistently gave abnormalities of muscle differentiation: when compared to controls, Notch cultures had a reduced frequency of myotubes, and displayed unusual clusters of myocytes which had either failed to fuse or had fused incompletely. Results from mixed cultures prepared from two embryos were consistent with the autonomous expression of nerve and muscle abnormalities by Notch-8 cells in the presence of wild-type cells. It is argued that the Notch locus has a direct role in the differentiation of both nerve and muscle. (2) white deficiencies. Cells carrying either of two deficiencies gave a clear-cut pattern of abnormalities: initial cellular differentiations were normal, but nerve, muscle and fat-body cells progressively deteriorated during the culture period. Mixed cultures showed that wild-type cells could not ‘rescue’ mutant muscle and fat-body cells; however, the status of the autonomy of mutant nerve abnormalities in these cultures was unclear. Both white deficiencies remove cytological band 3C1, and this permits a comparison of results with those from cultures of cells from Notch-S embryos (also deficient for 3C1). Abnormalities displayed in cultures of the two types of mutant show no overlap. Therefore no consistent cellular abnormality can be attributed to absence of band 3C1. (3) lethal(l)myospheroid. In contrast to earlier observations on in vitro cell cultures (Donady & Seecof, 1972) muscle was seen to differentiate, though its morphology was extremely abnormal. Observations indicated that all cell types within the cultures had poor properties of adhesion to a glass substrate. It is argued that the observed abnormalities are not consistent with a mutant lesion which is restricted to the basement membrane (contraWright, 1960), and that all cell types carry a basic defect which may reside in the cell membrane. (4) shibirets alleles. Cultures of two temperature-sensitive lethal shibire alleles (shils1, shits3) were normal at the permissive temperature of 22 °C. At the restrictive temperature (29° C) early cell differentiation was normal but subsequent development was blocked. This blockage could be partially reversed by shifting cultures to the permissive temperature after as much as 10 days exposure to the high temperature. It is suggested that shits cells are mutant in a process which is basic to several cell types.
INTRODUCTION
The intention of this report is to demonstrate the utility of in vitro culture for the characterization and analysis of cellular abnormalities in embryonic lethal mutants of Drosophila melanogaster. The rationale behind this work has been set out fully elsewhere (Cross & Sang, 1978). Results are described from the culturing in vitro of cells from embryos bearing various X-linked lethal mutations. These mutations fall into four groups which are dealt with in turn below.
(1) Notch lethals
The chief abnormalities of embryos bearing extreme Notch alleles are dramatic, and involve a three-fold hypertrophy of the nervous system at the expense of several other ectodermal derivatives (chiefly the hypoderm) and a complete failure of the differentiation of mesodermal derivatives (Poulson, 1940, 1945).
The results presented here show that for each of five Notch mutants which were studied, an hypertrophy of neural tissue occurred within cell cultures of single embryos. Four of the five mutants which were tested also gave muscle abnormalities: at the most extreme (the cytologically deficient Notch-8) functional myoblasts were formed, yet normal myotubes were not observed. Mixed cultures of Notch-R and wild-type cells gave results which were consistent with the autonomous expression of nerve and muscle abnormalities. Based on these observations it is suggested that the Notch locus has a direct role in derivatives of both ectodermal and mesodermal rudiments.
(2) White deficiencies
There have been differing opinions concerning the ontogeny and the cytology of the embryonic lethal ‘white deficiencies’ (reviewed by Kaufman, Shannon, Shen & Judd, 1975).
The results demonstrate that nerve, muscle and fat-body cells show abnormalities in single embryo cell cultures of both of two tested white deficiencies. These abnormalities do not overlap those displayed by Notch-8 cells which, in common with white deficiency cells, lack cytological band 3C1. Therefore, no consistent cellular phenotype has been shown to be associated with deficiency for band 3C1, in contrast to the conclusions of Poulson (1940, 1945) derived from studies of whole embryos. Abnormalities of muscle and fat-body cells are autonomously expressed in the presence of wild-type cells, and, therefore, these are probably direct effects of the mutations.
(3) lethal(l)myospheroid
Wright (1960) undertook the conventional embryological analysis of lethal(1)-myospheroid (1(1) mys) and suggested that the notable pleiotropic effects of the mutation were due to a delay in the formation of the basement membrane such that normal muscular contractions in its absence led to these multiple effects, essentially for mechanical reasons. The results reported here show that, when cultured in vitro l(1)mys cells show abnormalities which are consistent with those observed in both sectioned material and in vivo cultures. However, the observation that l(1)mys cells adhere poorly to a glass substrate provides further information that suggests a new interpretation of the l(1)mys syndrome which emphasizes a general cell membrane defect rather than one specific to the basement membrane. It should be stressed that contractile muscle was observed in our l(1)mys cultures since Donady & Seecof (1972) prepared cultures from l(1)mys cells and found that muscle did not differentiate; this latter result was, however at variance with observations made on whole embryos, which show that functional muscle is clearly formed.
(4) shibiretslethals
Grigliatti, Hall, Rosebluth & Suzuki (1973) isolated six alleles of the shibire (shi) locus (general designation shits). Homozygous stocks of all the mutant alleles are essentially normal at 22 °C, whereas at 29 °C the most severe alleles give reversible larval and adult paralysis; in addition exposure to 29 °C during critical periods of development causes lethality (including embryonic lethality). The present studies were intended to reveal the long-term effects of exposure to the restrictive temperature on individual shits cell types. The results show that cultures of shits1 and shits3 cells display similar and dramatic abnormalities when exposed to the restrictive temperature. The early differentiation of shits cells at 29 °C is qualitatively normal, but subsequent development of all identifiable cell types is blocked. For example, the continued proliferation of nerve axons and the full maturation of fat-body cells is impeded at the restrictive temperature. This blockage may be reversed, at least partially, by shifting cultures to the permissive temperature after as much as 10 days exposure to the high temperature. It is suggested that shits cells are mutant in a process which is basic to several, and possibly all, cell types.
MATERIALS AND METHODS
Flies were raised on a standard cornmeal, yeast, sugar and agar medium.
Unless otherwise noted, full details of all mutations and balancer chromo-somes are given in Lindsley & Grell (1968). The following embryonic lethals and special X chromosomes were used :
(1) Notch mutants: Notch-8 (N8), an 18-band deficiency; N264–39, of doubtful cytology, it may be a one-band Notch locus deficiency (Slizynska, 1938), or else may not be visibly deficient (Welshons, 1965); N264–40, N264–47, cytology indicates no visible deficiency.
(2) white deficiencies: w258–42, a 13- to 15-band deficiency; w258–45, probably associated with two deficiencies which together remove seven bands (Kaufman et al. 1975).
(3) lethal(l) myospheroid (l(l)mys).
(4) shibirets mutants: shits1 and shits3, extreme shits alleles with similar properties (Grigliatti et al. 1973; Poodry, Hall & Suzuki, 1973; Kelly, 1974; Kelly & Suzuki, 1974); shits alleles were kept as homozygous stocks.
(5) Various X chromosome balancers: FM1, FM4, FM7b (Merriam & Duffy, 1972), dl–49, suppress crossing over in heterozygous females, male fertile and homozygous female sterile.
In balanced stocks females were heterozygous for a lethal mutation and a balancer, and males were hemizygous for a balancer; only one quarter of the eggs produced by such stocks are expected to be lethal (hemizygous lethal males). The hatchability of eggs produced by each lethal stock was determined (Wright, 1973). In no case was a significant deviation from the expected frequency of lethality observed: 25 % in the case of balanced lethals, and 100 % for shits eggs at 29 °C.
Eggs generally were collected by standard procedures from stocks kept at 22° C. However, the normal methods of egg collection produced very poor results with shits females. In this case eggs were collected at 29 °C from paralysed females (Cross, 1975); this procedure did not affect embryonic viability. All eggs were incubated at 22 °C, after collection and prior to further handling.
Whole lethal embryos of all balanced stocks were allowed to develop for 26–30 h after laying (at 25 °C), and were then examined under phase-contrast microscopy using standard procedures. Hemizygous l(1)mys embryos displayed a pattern of abnormalities which was as described previously (Wright, 1960). Notch lethal embryos were as expected (Poulson, 1940, 1945), except that individual embryos bearing each of the five alleles occasionally showed muscular contraction in the region of both the dorsal hypoderm and the rudimentary fore- and hindgut structures. The white deficiency embryos were as described by Kaufman et al. (1975): the great majority of embryos died as apparently complete larvae which failed to hatch.
Standard single embryo cell cultures were prepared from early gastrulae as described by Cross & Sang (1978). All cultures, with the exception of those of shits (see Results), were incubated at 25 °C and examined at daily intervals for the first 5 or 6 days, and thereafter every 2 days.
Cell counts were undertaken in cultures of Notch embryos: nerve and muscle cells were counted at 24 h of culture, fat-body cells at 48 h. All distinct clusters of neuron cell bodies were regarded as nerve units and, with the aid of an ocular grid, were counted at 100 × magnification over an area of 0·31 mm2 (chosen blind). This method gave equal weight to isolated cells and very large clusters, but was quick and easy to perform. The number of large clusters of neuron cell bodies (arbitrarily defined as having some dimension greater than 35 μm) also was noted separately within the same field. All distinct muscle figures were scored as muscle units. Again this method gave equal weight to isolated cells and large clusters. Muscle counts were made at 200 × magnification and an area of 0·16 mm2 was scored for each culture. The same area of the culture was also scored for distinct myotubes or groups of myotubes. Fat-body cells were scored at 200 × magnification over an area of 0·16 mm2.
For cell cultures prepared from two embryos, the culture volume was doubled. Such cultures were given an aeration at 5 days by briefly removing the coverslip and then replacing it. Counts were made of nerve and muscle cells and the procedures used for single embryos were followed. Neurons were scored in an area of 0·62 mm2, and muscles in an area of 0·31 mm2.
In one experiment, shits1 cells were cultured according to a bulk embryo procedure (Shields, Dubendorfer & Sang, 1975), employing column-drops prepared from an homogenate of approximately 500 embryos in medium M3(BF) (Cross & Sang, 1978). Embryos were dissociated at 5–6 h after laying and all cultures incubated at 22 or 29 °C from 6 to 7 h.
RESULTS
Notch lethals
Detailed observations were made on cultures prepared from stocks of all five blotch lethals. In order to limit the data presented here, only the abnormalities of the identified cell types which were consistently observed will be mentioned. Full details may be found elsewhere (Cross, 1975). Results from four of the lethals are grouped together, N60g11 being dealt with separately.
A total of 341 single embryo cultures were prepared from stocks of the main group of four lethals and this number was made up as follows: N8/FM1, 120; N8/dl-49, 20; N8/FM7b, 39; N264−39/FM4, 68; N264−40/FM7b,50; N264− 47/FM7b, 44. In the case of each stock a class of abnormal cultures was observed and the frequency of such cultures never significantly differed from 25 %. Notch cultures at 24 h displayed a two- to three-fold increase in the number of large clusters of neuron cell bodies (Fig. 1) : example data for N8 and N264−39 is presented in Tables 1 and 2. In no case was there any indication of an increase in the overall number of neuron clusters. The development of nerve in Notch cultures showed no qualitative variation from that of normal cultures.
A large cluster of neuron cell bodies in a N8 culture at 24 h. The bar in this and other Figures represents 20 μm.
Muscular abnormalities provided the most reliable criterion for the identification of Notch cultures at 24 h. N8 cultures were the most abnormal in this respect and did not display any muscle figures which could be positively identified as myotubes, whereas they did have a number of apparent myocyte ‘bundles’ which were not observed in normal cultures (Fig. 2). These myocyte ‘bundles’ were, in fact, used as the identifying criterion for Notch cultures. Numbers of muscle figures were higher in N8 than in normal cultures (Table 1) and this was probably a consequence of the failure of individual muscle cells to form myotubes. Although organized myotubes were not seen in N8 cultures, it was impossible to rule out myoblast fusion at 24 h. In fact, at 2 days when the muscle cells tended to spread out, groups which showed signs of fusion were often noted. N8 muscle contracted normally and showed no other departure from the normal pattern of development.
(A) A myocyte ‘bundle’ in a N8 culture at 24 h. Three nuclei are arrowed. (B) Three myotubes in a control culture at 24 h. Three nuclei are arrowed.
Cultures of N264−39, N264−40 and N264−47 showed muscle development which was between that of N8 and normal cultures. At 24 h small numbers of myotubes were present, but the unusual ‘bundles’ of myotubes were consistently noted. When compared to controls, myotubes were about half as frequent in Notch cultures (example data for N264−39 in Table 2).
Fat-body cells were present in normal numbers in Notch cultures at 48 h (Tables 1 and 2). A delay in fat-body maturation was displayed by N8 cultures but this was not consistently true in cultures of the other three lethals.
In experiments with the stock N60011/FM7b it was not possible to unequivocally identify an abnormal class of cultures, since no myocyte ‘bundles’ were observed. Some cultures were similar to those of the other Notch lethals in the development of nerve, but no other abnormalities were detected. Pooling all the data on neuron development, 76 % of N8, N264−39, N264−40 and N264−47 cultures gave a count of 10 or more large neuron clusters, whereas only 8-5 % of normal cultures gave a score of 10 or more. Of the 65 cultures from N60g11/ FM7b, 14 (21-5%) gave a count of 10 or more large neuron clusters. It therefore seems likely that about 25 % of the cultures were derived from N60g11 hemizygotes, and that these cultures did show abnormally large quantities of nervous tissue.
In order to investigate the autonomy of N8 abnormalities in the presence of wild-type cells, 26 cultures of two embryos were made from the stock N8/FM7b. One-sixteenth of the cultures were expected to involve two hemizygous N8 embryos, three-eighths to involve one TV5 hemizygote and one non-lethal embryo, and the remainder two non-lethal embryos.
All of the cultures which were set up had clear examples of myotubes, and it seems likely that no culture involved two N8 embryos. Each of 11 cultures displayed ‘bundles’ of myocytes and these were the mixtures of N8 and wild-type embryos. In these cultures large neuron clusters were increased and myotubes reduced in number relative to a sample of apparently normal cultures (Table 3); and increase in overall muscle counts was again observed.
white deficiency lethals
A total of 41 cultures of the stock w258−45/ FM4 gave 8 (19·5 %), showing a clear pattern of abnormalities (i.e. derived from w258−45 hemizygotes); similar abnormalities were displayed by 6 out of 24 (25 %) cultures of the stock w258−42/ dl-49.
white deficiency cultures appeared to be relatively normal at 24 h, with the exception that muscular contractions were weak and very infrequent. By 3 days the three major cell types, nerve, muscle and fat-body, showed clear signs of deterioration. Over the next 5 days this deterioration progressed steadily until all neuron cell bodies had a swollen appearance and muscle and fat-body cells had completely degenerated. Indeed the lethal cultures showed no positive developments after the second day: for instance, there was no sign of either new axon production by the nerve cells or accumulation of fat by the fat-body cells.
Twenty-two cultures were made from the mixed cells of two embryos of the stock w258−45/FM4. In one of them, all nerves, muscle and fat-body cells developed abnormally, and this was presumably a culture of two mutant embryos. Nine cultures showed classes of both normal and abnormal muscle and fat-body cells, and were presumed to consist of mixed wild-type and mutant cells. The remaining cultures were completely normal. In the mixed mutant/wild-type cultures the abnormal muscle and fat-body cells displayed a development which exactly paralleled that observed in cultures of single mutant embryos. Nerve cells, on the other hand, showed little variation within a particular culture, but a good deal of variation between the nine mixed mutant/wild-type cultures: development ranged from a completely normal to a completely mutant pattern.
lethal(1)myospheroid
Seventy-two cultures were prepared from the stock l(1)mys/FM7b and 15 (21 %) were abnormal. The most striking abnormality was in the muscle cells, which could be identified at 24 h by their weak pulsations: they had an unusual form with a rather granular cytoplasm (Fig. 3) and were at first confused with the earliest form of haemocytes. Rather stronger pulsations were noted in a few cell clumps at 48 h, but the muscle cells were hidden and could not be visualized; contractions in these clumps persisted for 6–7 days. Another consistent abnormality in l(l)mys cultures was the apparently poor attachment of cells to the coverslip substrate: in the latter half of the 10-day culture period the coverslip cell density reduced as all types of cell fell away to the bottom of the culture drop. Aside from this poor attachment, nerve, fat-body and chitin-secreting cells were capable of extensive development which in the optimal cases was indistinguishable from that observed in wild-type controls.
shibiretslethals
In the experiments with shits cells control cultures were prepared from embryos of the wild-type Oregon-K stock. Development of Oregon-K cultures at 29 °C was similar to that at 22 °C, with the exception that it proceeded more rapidly at the higher temperature. Cultures of shits cells kept at 22 °C were indistinguishable from controls. When kept at 29 °C, cultures of both shits and shits3, produced either by the single embryo or bulk embryo methods, displayed a consistent pattern of abnormalities. A proportion of cells degenerated in the first 24 h of culture, but nevertheless, nerve, muscle, fat-body and haemocyte cells did appear in their normal forms. However, beyond 24 h no positive developments were observed in the cultures: no new nerve axons were produced, fat-body cells did not mature and haemocytes did not enter cell division. In fact, the cultures progressively deteriorated : muscle cells, for instance, took on a highly condensed form and by 7 days had beome unrecognizable; the cell density of the cultures progressively reduced, through a steady falling away from the coverslip.
A group of 20 cultures of shits, prepared by the bulk embryo method, were initially put at 29 °C and shifted down in pairs to 22 °C every day from day 1 to day 10. All these shift-down cultures showed some degree of recovery from the effects of exposure to the higher temperature. The earliest downshifts (at 1, 2 and 3 days) showed the most dramatic recovery. New nerve axons were clearly developing within 1 day of the shift down and continued to appear. Muscle cells reversed the tendency to condense and showed considerable activity. The later downshifts (4 days and on) showed reduced recovery, which almost certainly reflected the progressive deterioration of the cultures as length of exposure to the restrictive temperature increased. Cultures shifted down at 10 days did, however, show some improvement, with a few fresh axons, active muscles and mature fat-body cells eventually appearing.
shits cultures which were shifted up to the restrictive temperature after 1 day at 22 °C at first showed an arrested development and then a deterioration which was similar to that described for cultures kept continuously at 29 °C.
DISCUSSION
The specific conclusions to be drawn from experiments on each group of mutants will be set out first, leaving generalizations about the work until the end of this section.
(1) Notch lethals
Notch mutations have been the most thoroughly studied embryonic lethals of Drosophila -, their embryology and genetics have been comprehensively reviewed by Wright (1970). The question as to which are the primary abnormalities in the lethal Notch syndrome has interested a number of authors (Poulson, 1945; Counce, 1961; Wright, 1970). The differing views that have been taken both provoked the present work and amply illustrate the difficulties faced in inter-preting the abnormalities of lethal Drosophila embryos (see also Cross & Sang, 1978).
The most straightforward hypothesis to account for the data from culturing experiments is that the Notch locus has a direct role in derivatives of both ectodermal and mesodermal rudiments. There can be little doubt that the hypertrophy of nerve, reported for whole embryos, is part of the ‘Notch syndrome’ in vitro. Though our measurements of the quantity of nerve were crude, they gave consistent results for four of the mutants: in the case of the fifth mutant, N6g11, it was not possible to positively identify Notch cultures within a segregating group, but neuron cluster counts strongly suggest that about 25 % of the cultures showed neural hypertrophy. Furthermore, the consistent observation of reduced myotube frequency and presence of myocyte bundles in cultures of four of the five mutations suggests that these defects, too, are part of this Notch syndrome. The last assertion raises two obvious questions : (1) why did N60g11 cultures give only nerve abnormalities?, and (2) why did the other four alleles vary with respect to the degree of the muscle abnormalities? N60g11 must be set apart from the other alleles because it has temperature-sensitive properties (Foster & Suzuki, 1970) and hemizygotes are likely, therefore, to retain some sort of functional gene product. The available embryological (Poulson, 1939, 1968) and genetic (Welshons, 1965) data suggest that N264–39, N204−40 and N264−61 behave as do Notch deficiencies and they are presumed to be amorphs. However, in culture these three alleles did not behave as did Notch-8, which is obviously a Notch amorph. This particular difference could be inter-preted to mean either that the three alleles are not amorphs, or that the Notch-8 abnormalities are rendered extreme by lesions of genes other than Notch (Notch-8 is an 18-band deficiency). Further experiments with deficiencies restricted to the Notch locus would be fruitful in this regard.
The data from mixed cultures of Notch-R and wild-type cells are consistent with the autonomous expression of muscle and nerve abnormalities and this suggests that these abnormalities are both primary. The mixed cultures showed about 60 % more large neuron clusters and about 35 % fewer myotubes when compared to control cultures from two wild-type embryos (Table 3). However, a mixed culture experiment with reliable genotypic markers would be highly desirable and interesting in this connexion. Indeed, one may speculate that Notch nerve acts to cause the muscle defects since there is evidence that the nervous system has an important role in the development of muscle in insects (see review of Nuesch, 1968).
(2) white deficiencies
Abnormalities of nerve, muscle and fat-body cells were observed in cultures prepared from white deficiency embryos. Initial cellular differentiations were normal but defects became readily detectable after 3 days of culture; this ‘late’ appearance of abnormalities reflects the fact that the mutants die as late, apparently complete, embryos. A central issue has been the possibility of localization of lethal defects to a particular cytological region. Poulson (1940, 1945) found that a series of six white deficiencies gave similar embryonic defects which, on the basis of existing cytology, could be localized to band 3C1. However, Kaufman et al. (1975) revised the cytological data and considered that an embryo lethal effect could not be localized to 3C1. Similar abnormalities were observed for two white deficiencies in culture and this strongly suggests that defects can be localized to the regions common to the two deficiencies: tentatively 3A9–3B1 (3 bands) and 3B3-3C2 (4 bands). Further, abnormalities of white deficiency cells cannot be caused by the lack of bands 3C1-2 since they showed no overlap with those observed in Notch-8 cultures.
Mixed cultures of wild-type and mutant cells gave sharply distinct normal and abnormal classes of both muscle and fat-body cells, thus abnormalities of these two cell types seem likely to be autonomous by this test. The pattern of development of nerve tissue varied among the individual mixed cultures, but was homogeneous within each one, sometimes appearing normal and sometimes typically mutant. Neural defects did not show a clear autonomous expression and this could lead to the conclusion that they were secondary effects in culture. However, since it is not clear to what extent neurons interact with one another in vitro, the homogeneous appearance of nerve in the mixed cultures could be due to some influence of normal on mutant nerve, or vice versa.
(3) lethal(l)myospheroid
Muscle of abnormal morphology was identified in l(1)mys cultures (contraDonady & Seecof, 1972), consistent with the whole embryo data of Wright (1960). We should expect, for instance, that within the embryo these abnormal cells would give rise, as in the case of the somatic muscles, to spheroidal structures which would not attach to the hypoderm. The results from the in vitro cultures do, however, suggest (contraWright, 1960) that at least some of the abnormalities of muscle which occur in whole embryos are intrinsic to the muscle cells themselves, and not related to failures in the basement membrane. It should be noted that electron micrographs show neurons and myocytes cultured in vitro to be free of extracellular material, and therefore suggest that basement membranes are not necessary for the differentiation and survival of wild-type cells in culture (Donady & Seecof, 1972). The observation that all cell types adhered poorly to the glass substrate in the l(1)mys cultures suggests to us that the mutant factor acts directly to alter the picture of the mutant embryo which Wright has suggested; it still remains one of a rather unstable organization which is disrupted by the first muscular contractions.
Amongst the more positive findings is that nerve, fat-body and chitin-secreting cells were capable of considerable growth and development in the in vitro cultures, and this agrees well with Wright’s observation that mutant tissues continue to grow, within the vitelline membrane, for at least 10 h (at 25 °C) beyond the normal hatching time, and for several days if transplanted into larval hosts.
(4) shibiretslethals
Several studies point to a specific neural defect as the cause of the shits paralysis phenotype (Kelly, 1974; Kelly & Suzuki, 1974; Ikeda, Ozawa & Hagiwara, 1976); however, it is difficult to reconcile several of the shits pheno-types with a purely neural defect (Poodry et al. 1973; Suzuki, 1974; Swanson & Poodry, 1976).
The shits cultures showed abnormalities of all cell types at 29 °C, but were normal at 22 °C. The results are consistent with the idea that normal shi locus function is important in all tissues. With continuous culture at 29 °C, the initial differentiation of nerve, muscle, fat-body and haemocyte cells was not blocked, but their continuing development and maintenance was. Taken in isolation, the poor condition of shits cultures at 29 °C could make for difficulties of interpretation, since it could be argued that cell degeneration would have the effect of generally depressing the quality of the cultures, making the identification of specific genetic effects difficult. However, the reversal of many of the observed abnormalities upon downshift to 22 °C even after several days at the restrictive temperature, argues both that any adverse conditioning of the medium at 29 °C was not too serious, and that exposure to the restrictive temperature was not disastrous to cellular integrity.
The ‘bulk embryo’ procedure will be convenient for the culturing of cells from temperature-sensitive mutants; and it is clear that results will be comparable to those from the single embryo culture method: the two procedures gave similar results with both wild-type cells (see also Cross & Sang, 1978) and shit8 alleles.
(5) Overall conclusions
From the outset of this work it was hoped that the culture system would allow cells from mutant embryos to realize more developmental potential than would have been possible insitu : firstly, by limiting the possibility for deleterious inter-actions between cell types; and secondly, by allowing long-term survival of cells. The success of the approach has been demonstrated in the case of each mutant which was studied, and is well-illustrated by the example of the problem of comparing abnormalities of Notch-8 and white deficiency cells. Whole Notch-8 embryos show very limited differentiation of muscle and fat-body cells in situ, and abnormalities of these cell types are developmentally epistatic to those reported in white deficiency embryos ; whereas in vitro Notch-8 cells give extensive differentiation and long-term survival of these two cell types. Furthermore, defects revealed in the white deficiency cell cultures are of the sort that might be expected to be observed well after the normal hatching time in whole mutant embryos. In situ such cellular defects would probably be difficult to distinguish from the degenerative events which must eventually take place in any embryo which fails to hatch. In view of these facts it is apparent that a direct comparison of the abnormalities shown by Notch-8 and white deficiency cells is not possible in situ.
ACKNOWLEDGEMENTS
This research was supported by the Science Research Council, which is gratefully acknowledged. D.P.C. was in receipt of an S.R.C. Research Studentship.