Nine X-ray-induced mutations of the bithorax complex (BX-C) have been isolated and characterized. They all show the typical features of the Ultrabithorax mutations. They are homozygous lethal, produce a slight enlargement of the haltere in heterozygous condition and fail to complement the mutations at the bx, bxd and pbx loci. Some of them are associated with chromosomal aberrations in the regions 89E 1-4, where the BX-C lies, while others appear normal cytologically.

The effect of six of these mutants in the adult cuticle has been studied, producing mutant marked clones in heterozygous individuals. The clones were generated by X-radiation at two points in development: the blastoderm stage and the second larval period. In all cases mutant clones showed the same phenotype: clones appearing in the dorsal structures transform metathorax and first abdominal segment towards mesothorax. That is the additive effect of bx, bxd and pbx mutations. Clones in the legs, if induced during the larval period, show an effect homologous to that seen in the dorsal structures. However, when produced at blastoderm they show in addition a transformation of the posterior second (mesothoracic) and third (metathoracic) legs into the posterior first (prothoracic) leg. This transformation, named postprothorax (ppx) has been recently described for the alleles Ubx130 and Ubx1 (Morata & Kerridge, 1981) and appears to be general for the Ubx mutations.

It is concluded that the realm of action of the Ubx gene is defined by part of the meso-thoracic segment (posterior second leg compartment) and the entire metathoracic and first abdominal segments.

The genes of the bithorax complex (BX-C) of Drosophila are necessary for the morphological diversity of thoracic and abdominal segments. Mutations in the complex may transform one or more segments into another but leave the number of segments unaltered (Lewis, 1963, 1978). Segment diversity is achieved by the activation of a specific combination of BX-C genes in each segmental position (Lewis, 1978). In the current view (Lewis, 1978), the lack of BX-C activity results in mesothoracic development; the normal development of the next segment, metathorax, requires the function of the genes bithorax (bx+) and postbithorax (pbx+) for the anterior and posterior compartments respectively. The following segment, first abdominal, requires in addition the function of bithoraxoid (bxd+), for the second abdominal there is the function of infraabdominal-2 (iab-2+) (Kuhn, Woods & Cook, 1981), the third abdominal would need the hypothetical iab-3+ etc….In summary, each ‘level’ of segment development is achieved by the addition of a specific gene function to the pre-existing ones. The last abdominal segment, the eighth, would need the activity of all the BX-C genes.

Although Lewis’s model explains most of the genetic and developmental data about the BX-C, we have found recently (Morata & Kerridge, 1981) that two lethal mutants of the BX-C, Vbx130 and Ubx1 (which behave as deficiencies) produce an unexpected transformation of both posterior meso- and metathoracic leg compartments into prothoracic ones. This indicates the existence of a BX-C function which is necessary for the normal development of the meso-thoracic segment. We named this function postprothorax (ppx) in accordance with Lewis’s nomenclature and suggested the existence of a BX-C gene responsible for this function. Unlike the other BX-C genes, bithorax, postbithorax and bithoraxoid, which are required throughout larval development (Lewis, 1963: Morata & García-Bellido, 1976), postprothorax is only needed until approximately 7 h of development.

The discovery of ppx suggested the existence of a class of early acting BX-C genes which would be necessary for the development of mesothorax and perhaps other segments. In particular, as the postprothorax transformation is exclusively restricted to the posterior leg compartment of mesothoracic (second) and metathoracic (third) legs, we reasoned that there should exist an ‘anterior’ prothoracic gene which in mutant form would produce a homologous transformation in the anterior compartments. This argument was based on the existence of two other BX-C genes, bx and pbx, which perform homologous roles in anterior and posterior compartments respectively (see Morata, 1981 for review).

In an attempt to uncover all the functions of the BX-C, we have begun new experiments with mutagens and X-rays. Using X-rays we have isolated and characterized nine new Ubx alleles. The phenotype of six of them in clones indicates that they are all deficient for the bx, pbx, bxd and ppx activities. None of them shows anterior prothoracic transformation or an effect on any segment but the mesothorax, metathorax and first abdominal segment.

(1) Mutagenesis experiment

New BX-C mutants were induced by X-radiation of a multiply marked chromosome mwh jv st red sbd2e11ro ca(see Lindsley & Grell, 1968, description of mutants). The dose varied between 3000 and 4000 R. The mutations were detected by a quick method suggested by Gary Struhl. We crossed irradiated males to females carrying a weak allele in the bx locus, bx34e(the actual genotype was In(3LR)TM3, ri, pp, sep, bx34e, es/Sb63b). The Fl generation was screened for bx phenotypes. All the possible BX-C mutants, even complete deletions for the complex, will survive over bx34e -although some will be sterile because of the haplo-insufficiency of the Intersex (Ix) gene, which is located to the right and very close to BX-C (Lewis, 1978). Mutants detected in this way were then crossed to Df(3R)bxd100/TM1 flies to test their viability over Df(3R)bxd100(a deletion of part of BX-C) and to isolate stocks over the balancer TM1. This method, while quick and effective, has the obvious limitation that it can only detect those BX-C mutants which do not complement bx34e, such as bx or Ubx pseudoalleles or deficiencies. Point mutations, at the bxd or the pbx loci for example, would not be noticed.

(2) Cytological characterization

Mutant chromosomes were analysed cyto-logically to detect possible chromosomal aberrations. Males of each stock were crossed to females homozygous for the larval marker red. Fl red larvae were selected and their salivary gland chromosomes examined using standard procedures (Lefevre, 1976).

(3) Genetic characterization

The newly induced mutants were tested over recessive mutants of the BX-C. For the locus bithorax (bx) two different alleles were used; bx3(genotype of tester chromosome: sbd2bx3e), and bx34e(a pure homozygous stock was used). For the loci postbithorax (pbx) and bithoraxoid (bxd), one allele was used in each case (tester chromosomes: pbx1e11 and bxd51j). All the crosses were performed at 25 °C and the flies were mounted on slides for microscopical examination.

The expressivity of the homeotic transformations in the different combinations was evaluated and quantified. For transformations of the bithorax type, the number of notum and medial triple row bristles in the metathorax was counted and expressed as a percentage of the number of the same type of bristles in the normal mesothorax (Morata & Kerridge, 1980). combinations showing pbx phenotype were evaluated by scoring the number of bristles in the posterior double row which appear in the posterior haltere. This value is expressed as a percentage of the normal number of double row bristles in the wing. Combinations showing bxd phenotype were evaluated as the percentage of flies showing abdominal cuticle bearing leg bristles.

(4) Clonal analysis

The method of generating mutant clones is very similar to that described for Ubx130 and Ubx1(Morata & Kerridge, 1981). The following cross was made:
formula

In the fathers, the lethal M(l)oSp in the first chromosome is covered by Dp(1;3)-A59 which carries the wild-type M(l)o+ gene. The Dp(3; 1)P115 is a duplication of the entire BX-C in the proximal part of the first chromosome, while Df(3R)P115 is the corresponding deletion of the same genes. Dp(3;3)P5 is a tandem duplication of BX-C. The mutations y, w and f36a are cell-marker mutants used to distinguish mutant clones from surrounding cuticle. Ubxx represents any one of the new Ubx mutants.

The Fl of this cross was irradiated and two types of females were collected and screened for clones:
formula

in females (1) a mitotic recombination event in the first chromosome (see Fig. 1 in Morata & Kerridge, 1981) produces a clone marked with y w f36a which at the same time loses the Minute mutation and the Dp(3; 1)P115 thus uncovering the mutant combination Ubxx/Df(3R)P115(these marked Minute+ clones grow faster than surrounding cells, Morata & Ripoil, 1975). There is always a fraction of ywf36a clones arising from mitotic recombination between Dp(3; 1)P115, which is in the heterochromatin, and f36a which is located in the 15 F band. These clones, which represent about 20% of the total, remain wild type for BX-C. In flies of genotype (2) ywf36a clones also grow excessively, but they are not mutant for BX-C because Dp(3;3)P5 carries two doses of all BX-C genes. These were control clones.

Fig. 1.

Haltere transformation towards wing in (a) bx34e/D,(3R)P9 and (b) bx34e/Ubx130. Note that the amount of wing tissue and the transformation of trichomes is greater in (b) than in (a). Photographs taken at the same magnification.

Fig. 1.

Haltere transformation towards wing in (a) bx34e/D,(3R)P9 and (b) bx34e/Ubx130. Note that the amount of wing tissue and the transformation of trichomes is greater in (b) than in (a). Photographs taken at the same magnification.

Mitotic recombination was induced by X-irradiation at 4 ± 2 h after egg laying (A.E.L.) using a dose of 500-700 R (blastoderm clones) and at 55 ± 12 AEL using a dose of 1000 R (larval clones).

(1) Isolation of new BX-C mutants

A total of 28223 chromosomes were screened and 28 mutant phenotypes were detected. Of these, 10 were fertile, bred true and were isolated as stocks. One of them, that by genetic tests was inferred to be a translocation involving the third and the Y chromosomes, had low fertility and was subsequently lost.

All the mutants are homozygous lethal and lethal in trans with known Ubx alleles and with Df(3R)bxd100. They all exhibit a slight enlargement of the haltere when in trans with chromosomes not carrying any bithorax mutation. Thus all the mutations isolated by this method are typical Ubx alleles. No bx alleles Were found.

Four of the Ubx alleles are associated with gross rearrangements (Table 1). The other five were apparently normal cytologically. In the case of the rearrangements, all showed a breakpoint in the region 89E 1 – 4 where the BX-C lies (Lewis, 1963).

Table 1.

Cytological analysis of new Ubx alleles

Cytological analysis of new Ubx alleles
Cytological analysis of new Ubx alleles

(2) Interactions with recessive alleles

Tables 2 and 3 summarize the expressivity of the phenotype of different combinations with bx alleles. All show a bx phenotype; that is, the anterior region of the metathorax is replaced with pattern elements typical of the mesothorax. The degree of transformation produced by Ubx alleles is in general comparable to that produced by the same recessive mutants over Df(3R)P9 and stronger than that shown by the homozygous recessive mutants. Thus most of the new Ubx alleles behave as the complete lack of bx+ function. Indeed the transformation produced by two of them, Ubx130 and Ubx6-26 over bx34e is stronger than that shown by the Df(3R)P9. This is not clear from Table 3, where the transformations are evaluated exclusively in terms of bristle number, but a close examination of haltere appendages of genotype bx34e/Df(3R)P9 and bx34e/ Ubx130 clearly shows a more complete wing transformation of the haltere trichomes and appearance of wing veins (Fig. 1) in the latter genotype. The difference if probably due to a ‘transvection’ effect (Lewis, 1955) of rearrangements with breakpoints in the bithorax complex on the function of the bx function. That is, the hypomorphic function of bx34e may be reduced still further by gross Ubx rearrangements; an effect which may be due to abnormal somatic pairing of homologous chromosomes in the bithorax complex region.

Table 2.

Degree of mesothoracic transformation produced in the anterior meta-thorax by different mutants in trans with bx3

Degree of mesothoracic transformation produced in the anterior meta-thorax by different mutants in trans with bx3
Degree of mesothoracic transformation produced in the anterior meta-thorax by different mutants in trans with bx3
Table 3.

Degree of mesothoracic transformation produced in the anterior meta thorax by different mutants in trans with bx34e

Degree of mesothoracic transformation produced in the anterior meta thorax by different mutants in trans with bx34e
Degree of mesothoracic transformation produced in the anterior meta thorax by different mutants in trans with bx34e

Two of the point alleles Ubx9-22 and Ubx6-23 showed a transformation weaker than the recessive alleles. Another point mutant, Ubx1, has also been found to behave similarly (Morata & Kerridge, 1980). It is of interest to note that in these three cases the amount of notum structures is greatly reduced compared to the other alleles, while the amount of wing (MTR) is affected to a lesser degree (Tables 2 and 3).

All Ubx mutations, in trans with pbx, transform the posterior metathoracic segment into a posterior mesothoracic one. Estimations of the expressivity of these transformations were made by counting the number of posterior row hairs in the homoeotic wing of each pbx/Ubx combination (Table 4). No attempt was made to quantify the pbx transformations in the proximal metathorax since there are no bristle elements, either in the posterior metanotum or in the posterior mesonotum.

Table 4.

Degree of mesothoracic transformation produced in the posterior meta-thorax by different mutants in trans with pbx

Degree of mesothoracic transformation produced in the posterior meta-thorax by different mutants in trans with pbx
Degree of mesothoracic transformation produced in the posterior meta-thorax by different mutants in trans with pbx

Only two of the rearrangements (Ubx130 and Ubx12-5) and one of the apparent point mutations (Ubx5-2326) gave expressivities comparable to Df(3R)P9 in combination with pbx. The remaining Ubx/pbx trans combinations are significantly different from the expressivity of Df(3R)P9/pbx flies. The mutations Ubx8-8, Ubx6-28 and Ubx9-22 are not significantly different frompbx homozygotes with respect to the numbers of posterior row bristles; indicating that these Ubx alleles have comparable amounts of pbx+ function with the pbx chromosome itself.

The expressivity of the different Ubx mutations in trans with bxd51j is shown in Table 5. In all cases the main transformation is the removal of the first abdominal tergite and occasionally the replacement with metathoracic structures. In addition, in both the metathoracic and homeotic first abdominal segments, elements of the posterior metathorax are replaced by posterior mesothoracic ones (Lewis, 1963; Kerridge & Sang, 1981).

Table 5.

Degree of thoracic transformation produced in the first abdominal seg-ment by different mutants in trans with bxd51j

Degree of thoracic transformation produced in the first abdominal seg-ment by different mutants in trans with bxd51j
Degree of thoracic transformation produced in the first abdominal seg-ment by different mutants in trans with bxd51j

In no case were Ubx/bxd combinations as extreme as in flies of the genotype Df(3R)P9/bxd51j, (Table 5). Furthermore, the leg structures seen in the first abdominal position in Ubx/bxd flies are always rudimentary, with fused leg segments and generally less than 40 leg bristles. On the other hand Df(3R)P9/ bxd51i flies can have one or two superficially normal-looking ectopic legs. In no case were ectopic halteres found.

(3) Clonal analysis

Since all the Ubx alleles used here are lethal when homo-zygous, we have used the technique of X-ray-induced mitotic recombination to generate mutant clones in the adult cuticle of heterozygous flies. Clones were examined in the different body segments after induction at 4 ± 2 h A.E.L, (blastoderm clones) and at 55 ±12 h A.E.L. (larval clones).

(a) Blastoderm clones

The number and distribution of blastoderm clones generated for the different Ubx alleles are shown in Table 6, (a) and (b). In this table the number of clones is underlined in the segments where the clones have a homeotic effect. In the cases of clones in the posterior mesothoracic and anterior and posterior metathoracic legs there are two sets of numbers. The first, not underlined, represents the total number of clones in that compartment and the second, underlined, represents the number of clones clearly showing a homeotic transformation into prothorax or mesothorax. The two numbers cannot coincide because there is always a fraction of ywf33a clones arising from mitotic recombination between the locus of f33a and the Dp(3;1)P115 located in the heterochromatin of the X chromosome. These clones, estimated around 20% of the total (see Fig. 1 in Morata & Kerridge, 1981), remain wild type for the bithorax complex. In the case of the dorsal metathorax (haltere and metanotum) the markers y and f36a cannot be detected unless the clone produces a mesothoracic transformation. For this reason no control clones are given for this structure and all the experimental clones here appear transformed.

Table 6.

Number and homeotic phenotype of y wf36aclones generated at blastoderm after irradiation of the F1 of crosses

Number and homeotic phenotype of y wf36aclones generated at blastoderm after irradiation of the F1 of crosses
Number and homeotic phenotype of y wf36aclones generated at blastoderm after irradiation of the F1 of crosses

Before entering the description of the homeotic transformations, it is worth noting the similarity of clone frequency in the mesothoracic and metathoracic segments in spite of the fact that final structures are of very different size. Summarizing all the experiments, except control, there are 108 dorsal mesothoracic clones and 106 dorsal metathoracic ones (plus 20% of undetected clones). For the legs, where all the clones can be seen, there are 185 clones in the mesothorax and 211 in the metathorax. This confirms a previous observation (Lawrence &Morata, 1979) and strongly suggests that the number of primordial cells for each segment is the same.

With regard to their phenotype, mutant clones for all Ubx alleles had effects only in mesothoracic, metathoracic and first abdominal segments. In all other parts of the body, clones differentiated normally and appeared with frequencies comparable to controls (Table 6). We conclude that the Ubx+ gene provides no essential function in those segments. Note the lack of effect of Ubx alleles in most of the abdomen despite the fact that BX-C deletions (Lewis, 1978) transform all the abdominal segments into mesothorax (see Discussion).

Dorsally, clones in the mesothorax (wing and mesonotum) were untransformed. Three clones occupying the posterior wing and posterior mesonotum caused the development of a patch of bristles in the postnotum(Fig. 2), a feature absent in wild-type flies. The origin and significance of these patches is unknown, but they have been noted previously in gynandromorphs mosaic for Ubx, and wild-type cells (Lewis, 1963).

Fig. 2.

Structure (arrow) formed by bristles and trichomes of thoracic type produced by several of the posterior mesothoracic Ubx-/Df(3R)P115 clones. This type of structure does not normally appear in the posterior mesopostnotum.

Fig. 2.

Structure (arrow) formed by bristles and trichomes of thoracic type produced by several of the posterior mesothoracic Ubx-/Df(3R)P115 clones. This type of structure does not normally appear in the posterior mesopostnotum.

In the haltere and metanotum, clones differentiated mesothoracic structures. Clones were restricted to the anterior or posterior compartments. Anterior clones (Fig. 3) were transformed to anterior wing and mesonotum (bx transformations) and posterior clones were transformed to posterior wing structures (pbx transformation). In no case did a clone cross the anteroposterior compart-ment boundary either in the metathoracic or mesothoracic segments. This agrees with previous results (Steiner, 1976; Lawrence & Morata, 1977) and indicates that establishment of the anterior and posterior compartments in thoracic segments is early and independent of BX-C activity.

Fig. 3.

Large blastoderm clone of genotype Ubx-/Df(3R)P115. The clone forms almost an entire wing compartment in the place of the anterior haltere but does not transgress the anteroposterior compartment boundary in the haltere.

Fig. 3.

Large blastoderm clone of genotype Ubx-/Df(3R)P115. The clone forms almost an entire wing compartment in the place of the anterior haltere but does not transgress the anteroposterior compartment boundary in the haltere.

In the first abdominal segment clones mutant for the Ubx alleles appear not to develop. Clone frequencies in all cases were very low compared with controls (Table 6a). The few clones which appear are normal and are probably due to mitotic recombination distal to Dp(3: 1)P115 and proximal to the marker f36a, that is, cells possessing Ubx+ activity. This apparent lethality of Ubx-clones in the first abdominal segment is probably due to the inability of thoracic cells to develop in an abdominal segment. A similar observation was made for clones homozygous for the viable mutation bxd1, which transform first abdominal segment into metathorax (Morata & García-Bellido, 1976). We concluded that the lack of clones is due to a transformation of first abdominal histoblasts to thoracic cells which fail to differentiate in the abdomen.

In summary, the phenotype of dorsal clones of the genotype Ubx-/ Df(3R)P115 is that expected for the triple deficiency bx,pbx and bxd, with, as well, indication of a new phenotype in the posterior mesonotum.

Special mention is necessary of the effect of the clones on the ventral thoracic appendages; that is the legs and pleurae associated with them. Since the bristle patterns of the three pairs of legs are clearly different (Fig. 4) we can detect homeotic transformations present in clones. The transverse rows of bristles are typical of the anterior compartment of the first leg and the sternopleural bristles, edge bristle in the trochanter and apical bristle in the tibia are typical of the anterior compartment of the second leg. For the posterior compartments the differences are very clear in the femur, where the first leg contains more and larger bristles than the second and third legs (Fig. 4), and the second leg contains trichomes in the ventral side which are missing in the third leg. Also the bristles in the femur are located more posteriorly in the second than in the third leg.

Fig. 4.

Diagram of the bristle and trichome arrangements in the proximal segments of the prothoracic (1), mesothoracic (2) and metathoracic (3) legs (modified after Steiner, 1976). The anterior compartments (A) are located to the left and the posterior compartments (P) to the right of the broken lines which run through the length of the three legs. The dotted areas are those covered with trichomes. The pattern of open circles represents the arrangement of bristles, and their diameter roughly correlates with the size of the bristles. Note that the bristle arrangement in the femur is very different from those of 2P and 3P. Also the 3P femur is almost devoid of trichomes while they are very abundant in 2P and IP. CO, Coxa; TR, trochanter; FE, femur; TI, tibia; bTR, transversal rows bristles; Eb, edge bristle; AB, apical bristle; PAB, preapical bristle.

Fig. 4.

Diagram of the bristle and trichome arrangements in the proximal segments of the prothoracic (1), mesothoracic (2) and metathoracic (3) legs (modified after Steiner, 1976). The anterior compartments (A) are located to the left and the posterior compartments (P) to the right of the broken lines which run through the length of the three legs. The dotted areas are those covered with trichomes. The pattern of open circles represents the arrangement of bristles, and their diameter roughly correlates with the size of the bristles. Note that the bristle arrangement in the femur is very different from those of 2P and 3P. Also the 3P femur is almost devoid of trichomes while they are very abundant in 2P and IP. CO, Coxa; TR, trochanter; FE, femur; TI, tibia; bTR, transversal rows bristles; Eb, edge bristle; AB, apical bristle; PAB, preapical bristle.

The leg clones of all the Ubx alleles were scored and their effects noted (Table 6b). Clones in the anterior compartment of first and second legs were normal in all the cases and found with normal frequency compared to controls.

As expected most of the Ubx clones in the anterior third leg showed a transformation into anterior second, in correspondence with the phenotype of the dorsal clones. This is the phenotype of the bx mutants. The few clones that did not show a transformation are probably the result of mitotic recombination proximal to f36a and distal to Dp(3:l)P115.

In the posterior leg compartments the clones in the second and third legs were transformed into first (Fig. 5) while those in the first remained normal. Thus, all the Ubx mutants showed the ppx transformation which we have described recently (Morata & Kerridge, 1981).

Fig. 5.

Prothoracic transformation shown by clones Ubx-/ Dj(3R)P115(below the dotted line) generated at blastoderm in the posterior compartment of mesothoracic (a) and metathoracic (b) legs. Note in (a) that the large marginal bristles, typical of the posterior prothorax are differentiated by the clone.

Fig. 5.

Prothoracic transformation shown by clones Ubx-/ Dj(3R)P115(below the dotted line) generated at blastoderm in the posterior compartment of mesothoracic (a) and metathoracic (b) legs. Note in (a) that the large marginal bristles, typical of the posterior prothorax are differentiated by the clone.

However, not all the Ubx alleles showed prothoracic transformation with the same strength, Ubx5-2326, Ubx9-22 and Ubx12-5 in addition to Ubx120, produced a strong transformation where the posterior prothoracic pattern was reproduced almost perfectly. By contrast Ubx8-8Ubx5-12 and Ubx6-26 presented imperfect transformations (Fig. 6) with a mixture of prothoracic and meso-or metathoracic patterns. In the case of Ubx8-8 for example, although the majority of clones of the third leg showed prothoracic elements, those on the second leg did not show any clear cases of prothoracic transformation. Similarly, a high frequency of Ubx5-12 clones in the third leg was transformed into first, but only one clone in the second was clearly transformed into first leg.

Fig. 6.

Blastoderm clone of genotype Ubx8-8/Dj(3R)P115 showing a partial transformation into posterior prothoracic leg. The two large bristles (large arrows) located mediodistally are exclusive of the prothorax, but those located more proximally (small arrows) remain largely of mesothoracic type.

Fig. 6.

Blastoderm clone of genotype Ubx8-8/Dj(3R)P115 showing a partial transformation into posterior prothoracic leg. The two large bristles (large arrows) located mediodistally are exclusive of the prothorax, but those located more proximally (small arrows) remain largely of mesothoracic type.

These observations suggested that the alleles Ubx8-8Ubx5-12 and Ubx6-26 have low expressivity (that is were hypomorphic) with regard to that aspect of the phenotype. However, the fact that none of the mesothoracic leg clones in Ubx8-8 showed a prothoracic transformation raised the possibility that the second and third leg might be under independent genetic control. We looked for ways to increase the expressivity of the mutation. As we have described in the preceding section, the phenotype of some bx mutants in trans with Ubx130 is stronger than that of some bx allele over a deficiency of the entire complex (Fig. 1). We suppose that Ubx130 has a greater effect on somatic pairing (transvection effect of Lewis) than Df(3R)P115 and consequently a greater reduction of the hypomorphic function of bx34e.

We then decided to test whether Ubx130 would enhance the expressivity of Ubx8-8 and produce prothoracic transformation in the mesothorax. We generated blastoderm clones of genotype ywf36a; Ubx8-8/ Ubx130 using the same system described in the Material and Methods, but substituting Ubx130 for Df(3R)P115. We now find clones in both meso- and metathoracic legs which are transformed into prothorax (Table 7b). We conclude therefore that Ubx8-8 produces prothoracic transformation in both second and third legs, but with differing expressivity. We believe that this is also the case for Ubx5-12 and Ubx6-28 although these alleles have not been tested over Ubx130.

(b) Larval clones

We generated ywf38a clones mutant for the different Ubx alleles by irradiation during the second larval period. The phenotype of these clones in the different segments was like that of blastoderm clones but there were two differences; the first was that there were no clones with patches of bristles in the postnotum and the second was the absence of the prothoracic transformation of the meso- and metathoracic posterior leg compartments. The effect of the clones was restricted to the metathoracic and first abdominal segments. Clones in the haltere showed a transformation into wing, and those in the metathoracic leg differentiated mesothoracic pattern both in the anterior and in the posterior compartments. Data from the results obtained in the legs are shown in Table 7. In this table we have considered only those clones which appear in regions of the legs where the type of bristle pattern whether pro-, meso-or metathoracic could be ascertained. Of the metathoracic clones, we have considered only those which unequivocally presented a homeotic transformation. We took special care in studying the phenotype of the clones in the posterior compartments of meso- and metathoracic legs because of the prothoracic transformation shown by the blastoderm clones in these compartments. As shown in Table 7, all the 40 mutant clones (for the different Ubx alleles) in the meso-thoracic posterior leg differentiated a typical mesothoracic pattern with no sign of prothoracic transformation. In the metathoracic posterior leg 52 clones were found to produce a homeotic transformation and in every case it was into mesothorax. One such case is shown in Fig. 7.

Table 7.

Number of homeotic phenotype of ywf36aclones generated during the larval period in the Fl of the crosses described in Table 6

Number of homeotic phenotype of ywf36aclones generated during the larval period in the Fl of the crosses described in Table 6
Number of homeotic phenotype of ywf36aclones generated during the larval period in the Fl of the crosses described in Table 6
Fig. 7.

Transformation towards mesothorax shown by a yf36a; Ubx130/Df(3R)P115 clone generated during the larval period in the femur of a metathoracic leg. Normally this area is almost completely blank (see Fig. 4) but the clone differentiates bristles and trichomes of the mesothoracic type. Note that the clone is surrounded by a blank area devoid of bristles and trichomes.

Fig. 7.

Transformation towards mesothorax shown by a yf36a; Ubx130/Df(3R)P115 clone generated during the larval period in the femur of a metathoracic leg. Normally this area is almost completely blank (see Fig. 4) but the clone differentiates bristles and trichomes of the mesothoracic type. Note that the clone is surrounded by a blank area devoid of bristles and trichomes.

As is the case for the blastoderm clones, the larval ones fail to appear in the first abdominal segment, while they appear and differentiate normally in the rest of the abdominal segments and analia and genitalia.

In conclusion, Ubx- clones induced during the larval period show bx, pbx and bxd but not ppx transformation. The six new Ubx alleles analysed show the same temporal sequence of phenotypes described for Ubx130 and Ubx1(Morata & Kerridge, 1981).

Characteristics of the new BX-C mutants

The newly isolated mutants of the BX-C reported here are all Ubx alleles. All of them in trans with bx, pbx and bxd pseudoalleles give the phenotype of the recessive allele in question, have a dominant phenotype in that the haltere is slightly enlarged and are lethal with all other Ubx alleles and a deficiency for the bithorax complex. In addition, all of them show in clones a ppx phenotype, indicating that the prothoracic transformation is a general feature of the Ubx mutations.

Since a weak bx allele, bx34e, was used to detect some of the new mutants it was expected that some new bx pseudoalleles would have been isolated. None was found. Current literature also suggests that there are far more Ubx alleles than bx pseudoalleles (Lewis, 1963; Lindsley & Grell, 1968). However, none of the known bx pseudoalleles is associated with gross rearrangements, whereas from this analysis approximately half of the Ubx mutants are gross rearrangements (Table 2). Of the Ubx mutations isolated over bx34e, five are putative point mutations. The isolation of Ubx mutants only suggests that the DNA responsible for Ubx alleles is either larger or more sensitive to X-ray mutagenesis than that giving rise to bx alleles.

Although all the Ubx alleles, old and new, fail to complement bx, pbx and bxd mutants and also showppx phenotype, the data on the expressivity of the phenotypes indicate that some of them are not completely amorphic for the function of these genes. This is the case of Ubx9-22 and Ubx6-28 which are both point mutants. Another point allele, Ubx1(Morata & Kerridge, 1980), behaves in the same way. However, other point mutants show an expressivity comparable to that of the deficiency Df(3R)P9. All the alleles associated with detectable breakpoints are as strong as the deficiency. This suggests that some of the point mutants might have small chromosomal rearrangements undetectable by cytological examination.

The strong effect of chromosomal rearrangements on the bithorax function is probably related to the phenomenon called ‘transvection’ (Lewis, 1955), where some bithorax mutant phenotypes are enhanced if somatic pairing in the BX-C region is disrupted. In fact the combination of bx34c over Ubx130(Fig. 2) is stronger than over the Df(3R)P9 at least with regard to the transformation of the haltere trichomes. The same effect is seen for the prothoracic transformation of Ubx8-8, which is stronger over Ubx130 than over Df(3R)P115. Our interpretation is that in Ubx130 the somatic pairing is more disrupted than in the deficiency and consequently some of the remaining wild-type activity in Ubx8-8 is eliminated. It is worth noting that in the case of Ubx8-8 it produced bx and pbx transformations as strong as those of Ubx130 or Df(3R)P9, whereas the ppx transformation is weaker. This indicates that the effect of a given Ubx on each of the basic BX-C functions such as bx,pbx, bxd or ppx can be independent.

The interactions of the Ubx mutants with bxd and pbx(Table 4 & 5) show a simpler pattern. In these cases the degree of the transformations is less extreme than with Df(3R)P9, although sometimes the difference is small. Both point mutants and rearrangements behave similarly, which may suggest that pbx and bxd transformations do not show transvection effects.

Early and late functions of the bithorax complex

The results presented in this paper confirm and extend a previous result (Morata & Kerridge, 1981) that some of the BX-C genes act on a temporal sequence or that their products act differently at different times of development. Because all the Ubx alleles tested show the early ppx transformation, we conclude that this is a general feature of the Ubx phenotype. A point which is not yet completely established is the separate existence of ppx as another BX-C gene of the same category but different from bx, pbx and bxd. Indeed to date no independent mutation has been found for ppx. However, we think its existence is strongly suggested by the following observations. (1) There are independent isolates of bx, bxd and pbx mutations. They all show a specific transformation and none of them shows, even in their more extreme combinations, any sign of ppx transformation. (2) ppx affects a segment, the mesothorax, which is not affected by bx, pbx or bxd. (3) The requirement for the normal function of ppx+ is early, until only approximately 7 h of development (Morata & Kerridge, 1981). By contrast bx+, pbx+ and bxd+ are needed until the third larval period (Morata & Garcia-Bellido, 1976). (4) Partial duplications of BX-C like Dp(3;3)bxd100 coverppx but notpbx or bxd, thus indicating that their wild-type functions are in different positions within the complex (Morata & Kerridge, 1981).

It still could be argued that ppx might be the result of the additive effects of the triple mutant combination bx, bxd and pbx. Not having analysed this combination in clones this possibility cannot be completely ruled out. However, we believe this is unlikely because (a) extreme double combinations such as bx3pbx/Ubx130(for bx andpbx) or bxd313/Df(3R)P9(for bxd and pbx) do not show any sign of ppx transformation; (b) it would be difficult to explain why, if ppx is the summation of bx, pbx and bxd defects, it is shown only by blastoderm clones, while bx+, pbx+ and bxd+ are required until much later.

We have proposed (Morata & Kerridge, 1981) that the wild-type function of postprothorax (ppx) is necessary for the normal development of posterior meso- and metathoracic legs. If this function is eliminated, then both compartments develop into posterior prothorax. The period of function of ppx is early and probably preceeds that of bx or pbx.

In theory this difference in the period of function could be explained by: (a) differential gene activity at BX-C, (b) differential response of other genes due to lack of same/or different BX-C activity. Also, unlike the other genes that affect only one segment or compartment, ppx+ is required in two non-adjacent compartments separated by an intervening one (the anterior metathorax) which does not require ppx+ activity.

We believe that the phenotype of ppx may suggest a temporal mechanism of leg determination. The type of leg development, whether prothoracic, mesothoracic or metathoracic, depends on the switching state of BX-C genes in each leg primordium. For the posterior compartments, the formulappx-,pbx- results in prothorax, ppx+, pbx- would be mesothorax and ppx+, pbx+ metathorax. Assuming that anterior and posterior compartments develop in the same manner, we expected the existence of an ‘anterior’ gene as we state in the Introduction. However, we find that neither the Ubx mutants in clones of the adult cuticle nor BC-X deletions in larvae (Lewis, 1978; Struhl, personal communication) show the anterior prothoracic transformation. Therefore we are forced to conclude that if the ‘anterior’ gene exists, it is not in the bithorax complex.

The realm of action of the Ubx gene

All the Ubx mutations tested show a similar phenotype equivalent to the additive effects of bx, bxd, pbx and ppx transformations. On the idea of the existence of a separate ppx gene in addition to bx, bxd and pbx, the Ubx mutants produce the inactivation of these four genes. The mutant clones have an effect on the adult cuticle of three segments, mesothorax (in part), metathorax and first abdominal segment. The clones differentiate normally in the rest of the cephalic or abdominal segments. Since all the Ubx alleles behave in the same way with respect to the extent of the transformation this result strongly suggests that, for the adult cuticle, the realm of action of the Ubx gene is defined by these three segments. The fact that Ubx mutations always inactivate bx, bxd, pbx and ppx genes while there are independent mutations of the latter suggests that these four genes have a common overall control for which the normal function of Ubx is indispensable.

As we have not yet studied the larval phenotype of our Ubx mutants, it is not possible to compare directly the larval and adult phenotypes. However, we can use to this effect the published larval phenotypes of BX-C deficiencies which lacking all the genes defective in the Ubx mutations, for example Df(3R)bxd™Q(see Lewis, 1978 for description of the different deficiencies and larval phenotypes). What is observed in these larvae is that the metathoracic and first abdominal segments are transformed towards mesothorax, but there is no report of a prothoracic transformation homologous to that seen in early clones. Indeed even the biggest deletions where the entire complex is missing, as in Df(3R)P9 or Df(3R)P115, fail to show prothoracic transformations of meso- and metathorax (Lewis, 1978; G. Struhl, personal communication). A likely explanation for this apparent discrepancy is that all landmarks used to define the segment type in the larvae belong to the anterior compartment, which is not transformed into prothorax. In the larval segments there are not clear differences between posterior pro- and mesothorax, and such a transformation would pass unnoticed. In the adult cuticle, on the contrary, the differences between posterior prothorax and mesothorax are very clear. Thus we think that the larval and adult cuticular structures show the same transformation. According to our results, in the larvae homozygous for Df(3R)bxd100, for example, where the genetic and cytogenetic analysis (Lewis, 1978) indicates that the functions bx, bxd, pbx and ppx are deleted, the mesothorax, metathorax and first abdominal segments would develop as mosaics made of anterior mesothorax and posterior prothorax.

The realm of action of the bithorax complex

One important difference emerges when we compare the effect of the Ubx alleles and physical deletions such as Df(3R)P9 on the abdominal segments posterior to the first. While all the Ubx alleles, even the very strong ones like Ubx130, have their effect limited by the first abdominal segment, Df(3R)P9 shows in addition all abdominal segments transformed towards mesothorax (Lewis, 1978, in our view half mesothorax and half prothorax). Df(3R)P9 defines the realm of action of the entire complex; all the segments posterior to mesothorax require BX-C activity. Assuming that the effect of the Ubx mutants on the larval and adult segments is the same, as discussed above, this indicates that the phenotype of Ubx mutants accounts for only part of Df(3R)P9 phenotype. This Suggests that only some of the BX-C genes are inactivated by the Ubx mutations. Since the inactivation always affects the same genes bx, pbx, bxd and ppx, it suggests that these form a separate integrated functional subunit of BX-C. This would require a normal activity of the Ubx gene for its normal expression. The rest of BX-C, the ‘abdominal gene’ region, would be under a different control. It is then possible that the BX-C contains at least two functional subunits, one containing the genes controlling the more anterior segment -mesothorax, metathorax and first abdominal - and the other containing the genes controlling the remaining abdominal segments.

It is of interest to note that the most anterior limit of BX-C activity seems not to be a segment boundary but the anteroposterior compartment boundary in the mesothoracic segment. This result, although unexpected, is not unreasonable, as the thoracic anteroposterior boundaries are laid down very early in development possibly at the same time as the segment boundaries (Steiner, 1976; Lawrence & Morata, 1977).

Bryant
,
P. J.
(
1975
).
Pattern formation in the imaginai wing disc of Drosophila melanogaster: fate map, regeneration and duplication
.
J. exp. Zool
.
193
,
49
78
.
Kerridge
,
S.
&
Sang
,
J. H.
(
1981
).
Developmental analysis of the homoeotic mutation bithoraxoid of Drosophila melanogaster
.
J. Embryol. exp. Morph
.
61
,
69
86
.
Kuhn
,
D. T.
,
Woods
,
D. E.
&
Cook
,
J. C.
(
1981
).
Analysis of a new homeotic mutation (iab-2) within the Bithorax complex in Drosophila melanogaster
.
Mol. gen. Genet
.
181
,
82
86
.
Lawrence
,
P. A.
&
Morata
,
G.
(
1977
).
The early development of mesothoracic compartments in Drosophila
.
Devi Biol
.
56
,
40
51
.
Lawrence
,
P. A.
&
Morata
,
G.
(
1979
).
Early development of wing and haltere discs in Drosophila
.
Wilhelm Roux’ Archiv. devl Biol
.
187
,
375
379
.
Lefevre
,
G.
Jr
(
1976
).
A photographic representation and interpretation of the polytene chromosomes of Drosophila melanogaster salivary glands
.
In The Genetics and Biology of Drosophila
, vol.
la
(ed.
M.
Ashburner
&
E.
Novitski
), pp.
32
66
. London,
New York
:
Academic Press
.
Lewis
,
E. B.
(
1955
).
Some aspects of position pseudoallelism
.
Am. Nat
.
88
,
225
239
.
Lewis
,
E. B.
(
1963
).
Genes and developmental pathways
.
Am. Zool
.
3
,
33
56
.
Lewis
,
E. B.
(
1978
).
A gene complex controlling segmentation in Drosophila
.
Nature
276
,
565
570
.
Lindsley
,
D. L.
&
Grell
,
E. H.
(
1968
).
Genetic variations of Drosophila melanogaster
.
Carnegie Inst. Wash. Publ. no
.
627
.
Morata
,
G.
(
1981
).
The mode of action of the bithorax genes of Drosophila melanogaster
.
Am. Zool. (in the press)
.
Morata
,
G.
&
García-Bellido
,
A.
(
1976
).
Developmental analysis of some mutants of the bithorax system of Drosophila
.
Wilhelm Roux’ Arch, devl Biol
.
179
,
125
143
.
Morata
,
G.
&
Kerridge
,
S.
(
1980
).
An analysis of the expressivity of some bithorax transformations
.
In Development and Neurobiology of Drosophila
(ed.
O.
Sidiqqi
,
P.
Babu
,
C.
Hall
,
J.
Hall
), pp.
141
154
.
London
:
Plenum Press
.
Morata
,
G.
&
Kerridge
,
S.
(
1981
).
Sequential functions of the bithorax complex
.
Nature
290
,
778
781
.
Morata
,
G.
&
Ripoll
,
P.
(
1975
).
Minutes: mutants of Drosophila autonomously affecting cell division rate
.
Devl Biol
.
42
,
211
221
.
Steiner
,
E.
(
1976
).
Establishment of compartment in the developing leg imaginai discs of Drosophila
.
Wilhelm Roux’ Archiv devl Biol
.
180
,
31
46
.