A general cell marker for clonal analysis of Drosophila development

In Drosophila, methods of labelling cells with indelible genetic markers (Sturtevant, 1929) have improved considerably over recent years (Garcia-Bellido & Dapena, 1974; Morata & Ripoll, 1975). Typically, developing larvae heterozygous for a marker mutation are irradiated with X-rays and, occasionally and at random, clones of homozygous cells are produced following somatic recombination (Stern, 1936; Becker, 1957). These experiments have led to a description of the cell lineage of several organs of the adult fly (Garcia-Bellido, Ripoll & Morata, 1973, 1976; Steiner, 1976; Morata & Lawrence, 1979; Struhl, 1977). Some new principles have emerged from these studies, the most important being the compartment hypothesis (Garcia-Bellido et al. 1973; Crick & Lawrence, 1975).

Because the best genetic markers affect the derivatives of the adult epidermis, such as cuticle, bristles, and pigment cells of the eye, our understanding of cell lineage is mainly limited to parts of the adult integument. Attempts to study the cell lineage of the internal organs of both adults and larvae have depended on a less satisfactory set of cell markers (reviewed in Hall, Gelbart & Kankel, 1976). The most useful have been mutations affecting the activity of the enzyme aldehyde oxidase (Janning, 1972; Szabad, Schüpbach & Wieschaus, 1979), but even this enzyme cannot be detected in all tissues.

There is therefore a need for a general-purpose cell marker which would clearly mark each cell in every tissue without affecting development. Such a mutation would allow the study of cell lineage of internal organs and could lead to a description of the role of homoeotic genes in the nervous system, mesoderm and endoderm. Here I describe the isolation of alleles at a locus affecting the activity of succinate dehydrogenase. The mutations are recessive, viable in cells when homozygous and give a clear cell-autonomous phenotype in those tissues so far examined.

Succinate dehydrogenase was chosen for two main reasons: first, specific histochemical methods are available for material to be studied with both the light and electron microscopes (see Pearse, 1972; Lewis & Knight, 1977). Second, the enzyme is located in the mitochondria, and the stain precipitates mainly in those organelles. This is important because, even in heavily stained preparations, it is possible to see nuclei and cell boundaries in sections or whole mounts.

The strategy involved seven steps: (1) identification of a heat-labile variant of the enzyme succinate dehydrogenase; (2) mapping of the locus responsible for the variant; (3) mutagenesis of wild-type chromosomes to produce new mutations at the same locus; (4) ‘cleaning’ of the mutagenised chromosomes;(5) testing for cell viability of the new mutations; (6) developing staining methods for clones of cells homozygous for the new mutations; (7) testing the new mutations for temperature sensitivity.

The indirect flight muscles of Drosophila contain large quantities of succinate dehydrogenase and a single cut with fine scissors removes a piece of cuticle from the notum containing a chunk of flight muscle. This can then be floated on hot Ringer solution (Robb, 1969) for a given period and tested for enzyme activity by transfer to a staining solution containing the substrate sodium succinate and the dye nitroblue tetrazolium. For details, please see section 6 of these methods.

The activity of succinate dehydrogenase is seen as a blue precipitate. In wild-type muscles of Canton S, Oregon R and many other strains the enzyme becomes inactivated after 15 min at 55°C, but is still active after heating for 15 min at 54°C. A collection of 85 strains, all independently isolated from the wild, were therefore screened at 53 and 55°C to look for variants in heat lability. Only one variant was found; the succinate dehydrogenase activity disappeared after heating the piece of thorax at 53°Cfor 15 min, but survived heating at 52°C. It was named sdh¹. This stock, C154, was originally isolated at Bonaké in the Ivory Coast by G. de Jong and was kindly provided by the Department of Genetics, University of Cambridge.

The heat-sensitive phenotype segregated with the second chromosome, and was first located on the right arm between curved (2–75) and brown (2–104.5). It was positioned more accurately at 2-89± 1 (95% confidence limits) by staining 199 recombinants between M(2)S7 (2–77.5) and brown. (For genetic nomenclature see Lindsley & Grell, 1968.)

b pr adult males were fed with ethylmethane sulphonate (Lewis & Bacher, 1968) and crossed to cn sdh¹bw females. Some of these males were crossed to attached-X females to make an independent measure of the induced mutation rate. The sex ratio of their offspring, when compared to controls, indicated a lethal hit rate of about 0·7 per X-chromosome. We took about 3000 Fl males, each was then given the company of two cn sdh¹bw females. Of these individual matings, 2345 produced larvae and subsequently the Fl males were taken and their flight muscles heated at 53°C for 12 min and stained for succinate dehydrogenase. These were processed in batches of ten with It stw sdh¹ and yf^36a, sdh+ muscles as controls. If all ten experimental and the wild-type muscle chunks stained blue and the sdh¹ controls were unstained, all ten tubes were discarded. If one or more failed to stain the batch of tubes was kept and eight F2 cn sdh¹bw/b pr flies heated and stained from each tube (with appropriate controls). If all eight flies from one tube failed to stain, that tube was tested again on another day. If once again the muscle chunks failed to stain males were outcrossed to make a bpr sdh^*/CyO,pr cn² stock (where sdh+ indicates a possible mutant allele of sdh). From the 2345 chromosomes screened in this way eleven putative mutations were isolated. In addition, six tubes generated indirect flight muscles that stained pale blue which indicated partial enzyme activity; these have not been analysed further. The eleven putative alleles were crossed inter se and six were found to be lethal in trans with each other, implying that sdh is a locus in which mutations can be lethal; these six mutations were numbered sdh³–sdh⁶. One other putative allele (sdh²) was viable in trans with all six lethals giving flies having heat-labile succinate dehydrogenase. The muscles of these flies stained positively after heating at 50°C for 15 min but not after heating at 51°C. sdh² is located on a chromosome associated with a dominant female sterile phenotype. The other four putative alleles were discarded as they were viable in most combinations, and when in trans with the lethal alleles of sdh yielded flies with heat-stable succinate dehydrogenase. Dr Geoff Richaids kindly examined salivary chromosomes of sdh², sdh², sdh⁶ and sdh⁸, and detected no aberrations.

In case the mutagenised chromosomes carried lesions at other loci than sdh, the left arm of the second chromosome was completely replaced and the right arm was partially substituted by making two separate recombinations between cinnabar and brown. This was done for sdh³, sdh⁴, sdh⁵, sdh⁶ and sdh⁸.

While chromosome cleaning was under way, the alleles sdh², sdh³, sdh⁴, sdh⁵, sdh⁶ and sdh⁸ were tested to see if they were viable in homozygous clones of cells. The Minute technique (Morata & Ripoll, 1975) was used and the Minute itself scored as a bristle marker. Usually, y; bpr sdh/CyO, pr cn² females were crossed to DfsC^S2, y⁻; M(2)S7 sdh^*Dpsc^S2, y⁺/CyO, pr cn² males; the progeny were X-irradiated as larvae with 1000R (for conditions, see Lawrence & Morata, 1977). Fl females y/DfsC^S2y⁻; bpr sdh/M(2)S7 sdh⁺Dpsc^S2, y⁺were collected and thoraces and legs mounted for scoring under the compound microscope (Dfsc^S2 refers to the first chromosome, and Dpsc^S2 to the second chromosome of Til; 2)sc^S2). Because M(2)S7 is proximal and Dpsc^S2 distal to sdh, yellow Minute clones are either sdh or sdh⁺/sdh, but yellow Minute⁺clones are all homozygous for sdh. The Minute phenotype can be reliably scored in the bristles of the leg, sternopleura and notum. In all four mutants tested, the yield of yellow Minute+ clones was high, showing that all these six sdh alleles are cell viable. The clones are large, normal in phenotype and not noticeably different from control sdh+ clones made by irradiation at the same time.

Staining methods were developed initially in tissues which could be independently marked - the adult epidermis and the eye. For the adult epidermis, thoraces containing large yellow Minute⁺sdh clones are dissected so that the epidermis is exposed to the staining solution and left for an hour or so. The staining solution, adjusted to pH 7·0–7·2, consists of five parts each of 4 mg/ml nitroblue tétrazolium, 0·2 M Tris at pH 7·4,0·02 M magnesium chloride, 1 M sodium succinate and two parts each of 0·1 M sodium azide and 0·1 M sodium cyanide (Pearse, 1972). We do not usually make separate solutions of the cyanide, azide and succinate as recommended by Pearse. Small aliquots of the complete solution are kept frozen and continue to give satisfactory staining after several weeks at −20°C. A control solution, in which the succinate is replaced by malonate, is also useful; if there is blue precipitate after staining in this control solution it must be due to enzymes other than succinate dehydrogenase (Pearse, 1972). Staining is usually carried out in the dark in a few millilitres of solution at room temperature, the pieces of fly frequently floating on the surface. Following staining the preparations are dehydrated in alcohol, fixed in Carnoy’s solution (5–10 min), cleared in oil of cedarwood and mounted in Euparal.

In the adult epidermis the clones are visible as a pale-blue area in a dark-blue surround, and there is a sharp interface between the two areas so that individual cells can be classified as pale blue or dark blue. This interface coincides closely with the boundary between yellow and yellow⁺ bristles. To assess clones in the eye, cn sdh⁸bw/M(2)S7 or cn sdh⁸bw/M(2)c^33alarvae were irradiated. Eyes bearing brown clones were found down the dissecting microscope and the heads cut in half, these bisected heads were stained for 1–4 days and then fixed in half-strength Karnovsky’s solution, dehydrated in acetone, embedded in Araldite, and sectioned at 1–4 μm (Lawrence & Green, 1979). There was a striking difference between the staining of bw sdh and bw⁺sdh⁺territories in the eye, but there was still a pale blue background in the mutant clone (Fig. 3). With help from Michael Wilcox we found the simplest way to remove this staining within the clone is to gently heat-fix the material. The tissue is heated in Drosophila Ringer solution. The conditions must be carefully controlled; we used a waterbath and measured the temperature with an accurately calibrated thermometer - variation in most normal laboratory thermometers may be too great. The optimal temperature has to be worked out by trial and error as some reduction in the succinate dehydrogenase activity outside the clone also occurs. For example the chosen treatment for the imaginai discs is to heat the larvae at 47°C for 15 min (Brower, Lawrence & Wilcox, 1981) while for the flight muscles it is best to heat at 52°C for 15 min (Lawrence, unpublished). The background may be due to other dehydrogenases and their substrates - after heat fixation these substrates are probably free to diffuse out. An additional advantage of heating is that staining times are reduced, probably because the succinate can diffuse in more easily. For example, unheated eyes are best kept in the staining solution for several days; but after heating to 52°C for 10 min, the eyes stain well in about 12 h. Heat fixation does not seriously damage the tissues. Freezing also removes the background but causes more damage.

After chromosome ‘cleaning’, the chromosomes carrying sdh³, sdh⁴, sdh⁵, sdh⁶ and sdh⁸, which are lethal at 25°C, were tested to see if they were viable as homozygotes at 18°C. All trans combinations were also made at 18°C and although most combinations were lethal it was found that sdh⁸ flies were fully viable at 18°C, as were sdh⁸/sdh⁴. A few sdh⁸/sdh⁵ flies also survive at 18°C. Thus, sdh⁸ is a temperature-sensitive lethal; when transferred to 25°C the adults slowly sicken and die after about a week. At 25°C sdh⁸/sdh⁶ first instar larvae hatch normally but die before pupation, sdh⁸ flies grown and kept at 17-18°C are virtually male sterile but female fertile.

The temperature-sensitive lethal, sdh⁸, could be a useful tool: maternal perdurance of succinate dehydrogenase or some precursor RNA might make clonal analysis of the embryo, or even the larva, difficult or impossible. This might be overcome by breeding from sdh⁸females reared at 18°C and then transferring the progeny to 25°C. This tactic has not yet been tried.

An ideal cell marker should be autonomous, so that all and only those cells carrying the mutant genotype express the phenotype. The marker should be gratuitous, so that even large areas of mutant tissue should develop and differentiate normally. The marker should have a short perdurance (Garcia-Bellido & Merriam, 1971) so that it is visible even in small clones. Ideally the marker should be scorable in the cells of embryos, larvae and adults and be expressed in all cell types.

In the Material and Methods I describe the isolation of mutant alleles at a locus affecting the activity of succinate dehydrogenase. These alleles are homozygous lethal but viable in cells, where they reduce or eliminate succinate dehydrogenase activity. In the following section some preliminary results with this cell marker are reported. The tests are designed to see how far the sdh phenotype approaches the ideal described above.

Markers are usually tested for autonomy by linking them genetically to another marker and looking for cell-by-cell coexpression of the two phenotypes (e.g. Janning, 1972). The eye was chosen for these tests, since the pigment and retinula cells contain pigment which is largely removed by the brown mutation. Larvae that were heterozygous for both brown and the marker (sdh³, sdh⁴, sdh⁶ and sdh⁸ have been tried so far) were irradiated and large Minute+ clones produced. Flies bearing brown clones were found under the dissecting microscope, processed for succinate dehydrogenase staining and sectioned.

For each of the four alleles tested most clones show a clear phenotype both in the pigment cells due to bw, and in the retinula cells due to sdh (Figs. 1-4). A few clones show only the brown phenotype, presumably because the mitotic recombination was distal to sdh but proximal to bw. Analysis of the bw sdh clones show first, that cells in the ommatidia are normally arranged and second, that the expression of the phenotype is autonomous, or nearly so. When examined for succinate dehydrogenase staining using the light microscope, the border of the clone is sharp, but in spite of this it is difficult to score unequivocally each retinula cell. This is so even in those sections where individual ommatidia are clearly a mosaic of sdh and sdh⁺ cells (Fig. 3). It is clear that there is some spread of blue precipitate in the retinula cells; this is most conspicuous in the rhabdomeres, but can be seen in the cytoplasm as well (Fig. 3). Nevertheless the boundary of the clone is seen as a jagged interface between dark-blue and pale-blue retinula cells. This boundary coincides very well with the boundary between red (bw⁺) and pale-brown (bw) pigment cells proving that the autonomy is good but not excluding the possibility of some spread of stain across a cell or two. No more precise conclusion is possible, because the retinula cells lose their red pigment during the long staining step, and therefore cannot be independently scored for both the sdh and bw phenotypes.

We have found that heat fixation reduces non-specific precipitation of the dye; therefore eyes were heated to 49, 52 or 53°C before staining. This procedure almost eliminates the background of pale-blue stain within the clone and sharpens the clone border so that it becomes possible to score each retinula cell unambiguously for sdh phenotype (Figs. 1, 2, 4). This result strongly argues for cell-by-cell autonomy of the marker.

One eye bearing a large bw sdh clone was processed for the electron microscope (Ready et al. 1976). It was not heated or stained. The effect of bw can be seen as a reduction in number of pigment granules (Fig. 6). The sdh cells are distinct, their cytoplasm being condensed and darker than the sdh⁺ cells. The endoplasmic reticulum is also more closely packed. This distinctive phenotype is clear in retinula cells 1–6 but not in number 7 and is found in the bw area of the eye. Along the zone where bw and bw+ territories meet, mosaic ommatidia can be found (Fig. 6).

Additional evidence for cell autonomy comes from the appearance of clones in the imaginai discs which have a sharp boundary (Fig. 7, and Brower et al. 1981). Moreover in clones within the adult epidermis, salivary glands (Fig. 8), and gut (Fig. 9), individual cells can be scored. Further, following a suggestion by Antonio Garcia-Bellido, some stw pwn sdh⁸M(2)c⁺clones were examined in newly emerged abdomens and thoraces (see Garcia-Bellido & Dapena (1974) for a description of pawn). The overlying pattern of stw pwn trichomes in the cuticle can be compared with the patch of sdh phenotype in the underlying epidermal cells. I could not determine exactly which trichome was secreted by which cell but the correlation between the mutant cells and mutant cuticle appeared to be perfect. This confirms that the sdh marker is cell autonomous.

The overall appearance, size and shape of the eye clones is normal in young adults, suggesting that they develop as is usual for Minute⁺ clones. Clones in the head, leg and wing are of normal size and shape and respect compartment boundaries. However, in older adults (from 2–3 days after emergence) sdh cells in the eye slowly degenerate. Clones have so far been observed also in the adult muscles, the gut, the salivary glands and the oenocytes and I have seen no other phenotype than the lack of succinate dehydrogenase.

Perdurance is defined as the persistence in cells of wild-type phenotype after loss of the wild-type gene (Garcia-Bellido & Merriam, 1971). A long perdurance would reduce the usefulness of a marker mutation as small clones could not be identified. It was therefore important to measure perdurance for a mutant allele of sdh. Mutant cells were searched for in sections of eyes that had been irradiated late in larval development and heated to 52°C before staining. Small clones of bw pigment cells are found and these are associated with retinula cells that do not stain blue. In eyes from larvae irradiated at 96 h after egg laying, very small clones could be detected, these are sometimes just one or two pigment cells. In these eyes occasional unstained retinula cells are also seen. Unstained cells are not found in eyes irradiated at earlier times and it is probable that these unstained cells are indeed small sdh clones (Fig. 5). Unfortunately, owing to the loss of the red pigment in the retinula cells during staining, it was not possible to decide if some sdh cells were overlooked or if some sdh⁺ cells were incorrectly classified. Further, there is always the possibility that a marked retinula cell may have been part of a slightly larger clone which included unscored cells such as cone or bristle cells. Nevertheless the successful identification of a single sdh retinula cell does suggest that perdurance is not a serious problem in the eye.

One would not expect perdurance of enzyme activity to be the same in all tissues. It might depend on the number of cell divisions and on time elapsed after removal of the sdh+ allele from the cell as well as other factors. In the case of the eye, irradiation was at 96 h after egg laying but the eye was not fixed until after adult emergence, some 5 days later.

A proper answer to this question must await more study, but sdh clones have so far been found in several different organs in the adult; the epidermis, the eye, the flight and abdominal muscles, the heart, the oenocytes, the gut, and the salivary glands. In sdh⁺ flies all cells contain mitochondria which precipitate the stain, so it is conceivable that the marker may be useable in most cell types.

A new cell marker has been isolated and partially tested. The alleles affect the activity of the enzyme succinate dehydrogenase. Clones of cells homozygous for sdh⁸ are cell viable and usually normal, are autonomously expressed, and are scorable in several different tissues. The perdurance, at least in the eye, is such that clones of one or two cells can be scored. The marker thus looks promising and would seem to be an improvement on existing enzyme markers (see Hall et al. 1976).

I thank Brigid Hogan for a key conversation at Titisee in 1978. I am grateful to Paul Johnston and Sheila Green who have done much of the work. The project would have foundered without the good advice and persistent optimism of Gary Struhl who also helped with the mutagenesis. I thank my colleagues at the MRC LMB and Michael Ashburner for encouragement and criticism, and Philip Anderson for improving the manuscript.