Arthropod appendages are thought to have evolved as out-growths from the body wall of a limbless ancestor. Snodgrass, in his Principles of Insect Morphology (1935), proposed that, during evolution, expansion of the body wall would originate the base of the appendages, or coxopodite, upon which the most distal elements that represent the true outer limb, or telopodite, would develop. The homeobox gene Distal-less (Dll), which is required in the Drosophila appendages for development of distal regions, has been proposed to promote formation of telopodite structures above the evolutionary ground-state of non-limb or body wall. Here, we present evidence that another homeobox gene, extradenticle (exd), which is required for appropriate development of the trunk and the proximal parts of the appendages, represents a coxopodite gene. We show that exd function is eliminated from the distal precursors in the developing limb and remains restricted to proximal precursors throughout development. This elimination is important because, when ectopically expressed, exd prevents distal development and gives rise to truncated appendages lacking distal elements. Moreover, the maintenance of exd expression during larval stages, contrary to Dll, does not require the hedgehog (hh) signaling pathway, suggesting that the proximal regions of the appendages develop independently of hh function. Finally, we show that in the crustacean Artemia, exd and Dll are expressed in comparable patterns as in Drosophila, suggesting a conserved genetic mechanism subdividing the arthropod limb.

In Drosophila, the homeotic genes of the Antennapedia-Complex (ANT-C) and the Bithorax-Complex (BX-C) (hereafter called Hox genes) are responsible for the generation of morphological diversity along the anteroposterior body axis (reviewed in McGinnis and Krumlauf, 1992; Botas, 1993; Lawrence and Morata, 1994). Each body segment is specified by a Hox gene or a particular combination of Hox genes. As they encode homeodomain proteins, Hox genes are supposed to regulate transcription of specific target genes, which are ultimately responsible for the Hox function. However, Hox proteins by themselves show little DNA-binding specificity and their association with cofactors has been proposed as a mechanism to modulate their DNA-binding properties (Wilson and Desplan, 1995).

One of these Hox cofactors is the product of the gene extradenticle (exd) (Peifer and Wieschaus, 1990; Rauskolb et al., 1993). exd is a member of the highly conserved PBC family of homeobox genes, which also includes the mammalian pbx and the C. elegans ceh-20 genes (Bürglin, 1995). PBC proteins have been shown to act as cofactors of the homeotic function by modulating the DNA-binding affinity and specificity of Hox gene products (reviewed in Mann and Chan, 1996).

However, exd is also involved in controlling specific adult subpatterns (González-Crespo and Morata, 1995; Rauskolb et al., 1995), suggesting that some of its functions may be independent of interactions with Hox proteins. This is the case for the legs, where exd does not seem to participate in the control of leg identity mediated by differential Hox gene activity in each of the three thoracic segments; instead, exd is differentially required along the proximodistal (P-D) axis of all six legs (González-Crespo and Morata, 1995; Rauskolb et al., 1995). Other appendages like wings and halteres also show localized requirements for exd. These observations suggest an involvement of exd with discrete patterning events in imaginal discs which would be independent of Hox gene activity, as the latter is concerned with overall segment identity.

The exd requirement in the legs is restricted to the proximal regions and this localized requirement is paralleled by a localized distribution of the exd product in the corresponding precursor regions of the leg imaginal discs (González-Crespo and Morata, 1995). Interestingly, another homeobox gene, Distal-less (Dll) (Cohen et al., 1989), is required and expressed in the distal regions of the legs (Cohen and Jürgens, 1989; Díaz-Benjumea et al., 1994) in a pattern that appears to be nearly complementary to that of exd, suggesting a subdivision of the leg into two distinct regions along the P-D axis.

Such a subdivision of the arthropod leg into proximal and distal components was proposed by Snodgrass (1935) on the basis of comparative morphology in a model of how arthropod appendages arose in evolution. According to this hypothesis, arthropod appendages would have originated from a primitive limb just formed by a basis, called coxopodite, that would be an expansion of the trunk, and a distal arm, or telopodite, that would represent the limb proper. This primitive appendage would have evolved by acquisition of joints both in the coxopodite and the telopodite, thus giving rise to the present morphology.

In this paper, we present a detailed study of the expression patterns of exd and Dll throughout development of the limb primordia in the insect Drosophila and in the crustacean Artemia. We also show that the domains defined by exd and Dll have distinct functional properties. Altogether, these results provide genetic evidence for the Snodgrass hypothesis and suggest that the coxopodite and the telopodite are genetically represented by the exd and Dll domains.

Clonal analysis

exd− clones were generated by X-rays-induced mitotic recombination using the null allele exdXP11 and the markers yellow (y) and forked36a(f36a) as described (González-Crespo and Morata, 1995).

Drosophila and Artemia immunostaining

Drosophila embryos, second and third instar larvae, and pupal leg discs were processed for double immunofluorescence staining as described (Patel, 1994) and analyzed by confocal laser-scanning microscopy. exd and Dll expression were detected with a rat polyclonal anti-exd antibody (González-Crespo and Morata, 1995) and a mouse monoclonal anti-Dll antibody (Díaz-Benjumea et al., 1994). teashirt (tsh) and optomotor-blind (omb) expression were detected using corresponding lacZ lines (Rauskolb et al., 1995; Grim and Pflugfelder, 1996) and a rabbit polyclonal anti-β-gal antibody (Cappel). Secondary antibodies were an anti-rat biotinylated (Amersham) coupled to lissamine-rhodamine-conjugated streptavidin (Jackson), a FITC-conjugated anti-mouse (Jackson) and a FITC-conjugated anti-rabbit (Jackson).

Artemia cysts (San Francisco Bay Brand) were cultured in reconstituted sea water at 30°C and nauplius larvae were collected, fixed in 4% paraformaldehyde in PBS during 3 hours at 4°C, dehydrated through a methanol series and stored at −20°C. For staining, nauplius were rehydrated to PBS, 0.3% Triton X-100, sonicated four times during 5 seconds at an amplitude of 6 μm with an immersion tip sonicator, blocked in 1% BSA, 0.3% Triton X-100 in PBS during 2 hours and incubated in primary antibodies (a rat polyclonal anti-exd antibody (González-Crespo and Morata, 1995) and a rabbit polyclonal anti-Dll antibody (Panganiban et al., 1995)) overnight at 4°C. After washing and second blocking, nauplius were incubated in corresponding biotinylated secondary antibodies during 2 hours and processed for DAB staining using the Vectastain Elite ABC Kit (Vector). In double staining experiments, both antigens were distinguished by the presence or absence of Ni in the developing reaction. Stained nauplius were dehydrated through an ethanol series, treated with xylenes, mounted in Permount (Fisher) and photographed under Nomarski optics.

Ectopic expression

The UAS/GAL4 system (Brand and Perrimon, 1993) was used to target ectopic exd expression. The full-length exd cDNA (Rauskolb et al., 1993) was cloned into the pUAST vector (Brand and Perrimon, 1993) to generate transgenic flies by P-element transformation. Six independent UAS-exd lines were obtained and assayed for their ability to ectopically deliver immunodetectable exd protein, giving similar results. We used as drivers the E132 line (Halder et al., 1995), a Dll-GAL4 (Calleja et al., 1996) and a decapentaplegic (dpp)-GAL4 (Wilder and Perrimon, 1995).

TUNEL analysis

The method of TdT-mediated dUTP-biotin nick end labeling (TUNEL) (Gavrieli et al., 1992) was applied to Drosophila third instar larvae imaginal discs. Discs were dissected in PBS, fixed in 4% paraformaldehyde in PBS for 20 minutes and in 4% paraformaldehyde, 0.1% Triton X-100 and 0.1% sodium deoxycholate in PBS for 20 minutes, washed in PBS, 0.3% Triton X-100 and incubated in 125 mM sodium cacodylate (pH 7.2), 2.5 mM CoCl2, 0.25 mM 2-mercaptoethanol, 0.3% Triton X-100, 6 μM dATP-biotin and 0.5 u/μl of terminal deoxynucleotidyl transferase (TdT) at 37°C for 3 hours. After washing, discs were processed for DAB staining using the Vectastain Elite ABC Kit (Vector), mounted in 80% glycerol and photographed under Nomarski optics.

hh temperature shift

The temperature-sensitive allele hhts2 (Ma et al., 1993) was used to inactivate hh function. Embryos and first instar larvae were kept at a permissive temperature (17°C), transferred to the restrictive temperature (29°C) during the second instar and discs from third instar larvae were processed for immunostaining. Mutant larvae were identified by the absence of the dominant marker Tubby carried on the TM6B balancer.

exd and Dll in the developing Drosophila leg

The functional requirements for exd in the adult Drosophila legs have been described previously (González-Crespo and Morata, 1995; Rauskolb et al., 1995). exd− clones differentiate abnormally in proximal regions but, in distal regions, the loss of exd has no effect. These results suggested that the leg is subdivided into two regions according to their exd functional requirement (González-Crespo and Morata, 1995).

However, Rauskolb et al. (1995) reported that a fraction of exd− clones in the intermediate segments femur and tibia exhibit defective patterns. As this may suggest that exd is required in a graded fashion along the leg, we have re-examined the question of exd requirements in intermediate leg regions by studying a large number of small clones, which were confined within each of these segments. A total of 77 clones were analyzed: 9 in the trochanter, 45 in the femur and 23 in the tibia. In the trochanter all 9 clones exhibited transformation towards a more distal, tibialike pattern (Fig. 1B). The 45 clones found in the femur were subdivided according their position in proximal (13), medial (13) and distal (19). The 13 proximal clones differentiated inappropriate structures that tended to sort out from the surrounding normal tissue. In most cases, the clone bristles showed an associated bract, a feature corresponding to more distal leg elements (Fig. 1A). Of the 13 medial clones, 3 developed abnormally like those in the proximal femur and the remaining 10 differentiated normally. All 19 distal clones were also normal (Fig. 1C). In the tibia, 22 out of the 23 clones differentiated normally. These results establish a precise boundary for exd requirement along the P-D axis of the Drosophila leg at the proximal femur. It is of interest that, within the proximal leg regions where the exd− clones exhibit a phenotype, those in the more proximal areas like coxa and pleura differentiate aberrant patterns unlike any other leg pattern (Fig. 1A), whereas those in the trochanter and proximal femur show transformation to patterns characteristic of more distal regions of the leg (Fig. 1A,B).

Fig. 1.

(A-C) exd is required in proximal, but not in distal, leg. (A) exd− clone in proximal leg (arrow). The proximal femur is reduced and fused to coxa and ventral body wall or sternopleura. The clone extends to distal femur where it does not produce any abnormality. (B) Small exd− clone (arrow) in trochanter forming an outgrowth with bracted bristles which are characteristic of distal leg parts. Only clones in trochanter and proximal femur produce distal patterns, possibly because there is Dll product (see M). (C) exd− clone in distal femur (arrow). Removal of exd in this and any other more distal region does not affect normal leg development. (D-N) exd and Dll expression during Drosophila leg primordia development. In all cases exd is in red and Dll in green. (D) Ventral view of a stage 14 embryo stained with the exd antibody. Thoracic segments are indicated by T1-T3 in the ventral midline. Six ‘holes’ of exd staining appear in the ventrolateral epidermis. (E) Same view as in D showing that exd expression abuts the Dll-expressing imaginal disc primordia. (F) Sagittal view of an embryo in a similar stage as the one in D and E. Cells forming the imaginal disc primordia are stained only with the Dll antibody and invaginate from the outer exd-expressing epidermal cells. (G) Frontal view of a second instar leg imaginal disc stained for exd and Dll. The exd-expressing cells form a peripheral ring surrounding the central Dll-expressing cells. (H) Side view of a second instar disc. Note that there is no overlapping between the exd and the Dll expression domains. The split channels are shown in I (Dll) and J (exd). (K) Frontal view of a third instar leg disc. (L) Middle optical section of the same disc. (M) Basal optical section showing the only region where exd and Dll overlap (yellow). (N) Sagittal optical section of an everting leg disc 2 hours after pupariation. Note the yellow overlapping region. Abbreviations: st, sternopleura; c, coxa; tr, trochanter; f, femur; ti, tibia.

Fig. 1.

(A-C) exd is required in proximal, but not in distal, leg. (A) exd− clone in proximal leg (arrow). The proximal femur is reduced and fused to coxa and ventral body wall or sternopleura. The clone extends to distal femur where it does not produce any abnormality. (B) Small exd− clone (arrow) in trochanter forming an outgrowth with bracted bristles which are characteristic of distal leg parts. Only clones in trochanter and proximal femur produce distal patterns, possibly because there is Dll product (see M). (C) exd− clone in distal femur (arrow). Removal of exd in this and any other more distal region does not affect normal leg development. (D-N) exd and Dll expression during Drosophila leg primordia development. In all cases exd is in red and Dll in green. (D) Ventral view of a stage 14 embryo stained with the exd antibody. Thoracic segments are indicated by T1-T3 in the ventral midline. Six ‘holes’ of exd staining appear in the ventrolateral epidermis. (E) Same view as in D showing that exd expression abuts the Dll-expressing imaginal disc primordia. (F) Sagittal view of an embryo in a similar stage as the one in D and E. Cells forming the imaginal disc primordia are stained only with the Dll antibody and invaginate from the outer exd-expressing epidermal cells. (G) Frontal view of a second instar leg imaginal disc stained for exd and Dll. The exd-expressing cells form a peripheral ring surrounding the central Dll-expressing cells. (H) Side view of a second instar disc. Note that there is no overlapping between the exd and the Dll expression domains. The split channels are shown in I (Dll) and J (exd). (K) Frontal view of a third instar leg disc. (L) Middle optical section of the same disc. (M) Basal optical section showing the only region where exd and Dll overlap (yellow). (N) Sagittal optical section of an everting leg disc 2 hours after pupariation. Note the yellow overlapping region. Abbreviations: st, sternopleura; c, coxa; tr, trochanter; f, femur; ti, tibia.

The functional requirements for Dll have already been described (Cohen and Jürgens, 1989). Cells homozygous for null Dll mutations fail to develop in distal regions, from the trochanter to the tarsus, while they develop normally in the body wall and coxa. Thus it appears that exd and Dll are required in near complementary domains along the leg, with the exception of the trochanter and proximal femur where both gene functions are needed.

To determine if these complementary domains of function correlate with a localized protein distribution, we analyzed exd and Dll expression patterns during leg development by immunofluorescence and confocal laser-scanning microscopy. In early embryogenesis, exd is broadly distributed throughout the embryo and colocalizes with Dll in the limb primordia, where Dll expression is initially turned on (data not shown). However, by stage 14, exd is excluded from the 20-25 Dll-expressing cells that form the thoracic imaginal disc primordia (Cohen et al., 1993) and remains restricted to the nuclei of the surrounding cells in the ventrolateral epidermis of the embryo (Fig. 1D-F). During the second larval instar, exd localizes to a peripheral ring of the leg disc (Fig. 1G-J) surrounding the central Dll-expressing cells (Díaz-Benjumea et al., 1994), with no detectable overlap. In the third instar (Fig. 1K-M) and early pupal discs (Fig. 1N), Dll and exd also show reciprocal expression patterns, except for a proximal ring of Dll expression (Fig. 1M,N) (Díaz-Benjumea et al., 1994). According to the fate map of the third instar leg imaginal disc (Fristrom and Fristrom, 1993), the peripheral exd-expressing cells will give rise to the proximal parts of the adult leg, while central Dll-expressing regions are precursors of the distal elements. Therefore, the reciprocity between exd and Dll genetic requirements correlates with their product distribution during leg development.

This restriction of exd protein to the peripheral parts of the disc is in contrast with its reported uniform mRNA distribution (Rauskolb et al., 1995). We have found that the exd mRNA accumulates preferentially in the periphery of the leg disc, although lower levels are also detected in the central regions (N. Azpiazu and G. Morata, unpublished data). This exd mRNA in central regions may be responsible for the low levels of exd protein detected in the cytoplasm (Mann and Abu-Shaar, 1996), suggesting that the restriction of exd function to proximal leg parts may be controlled not only transcriptionally but also at the level of nuclear transport.

Altogether, these functional and expression data suggest an early subdivision of the leg primordium into two different regions that correspond with the exd and Dll domains. This suggestion is reinforced by the observation that other developmental genes are expressed within these two domains. For example, the product of the gene teashirt (tsh), involved in the specification of thorax and abdomen (Röder et al., 1992), colocalizes with exd in the mature leg imaginal discs (Fig. 2A-C) and, like exd, is absent from the embryonic thoracic disc primordia (data not shown). Conversely, the limb patterning gene optomotor-blind (omb), a dpp target (Grim and Pflugfelder, 1996), is expressed only within the Dll domain and abuts the exd domain (Fig. 2D).

Fig. 2.

tsh and omb expression patterns accommodate to the domains defined by exd and Dll. (A-C) exd (red) and tsh (green) colocalize in third instar leg imaginal discs as shown in the merged image (yellow in A) and in split channels for exd (B) and tsh (C). (D) omb expression (green) is located in the Dll domain and abuts the exd domain (red). No omb expression is detected in the exd domain, although its activator dpp extends all along the P-D axis.

Fig. 2.

tsh and omb expression patterns accommodate to the domains defined by exd and Dll. (A-C) exd (red) and tsh (green) colocalize in third instar leg imaginal discs as shown in the merged image (yellow in A) and in split channels for exd (B) and tsh (C). (D) omb expression (green) is located in the Dll domain and abuts the exd domain (red). No omb expression is detected in the exd domain, although its activator dpp extends all along the P-D axis.

exd expression is confined to proximal precursors in Artemia appendages

We also studied whether the reciprocity between exd and Dll domains is evolutionary conserved by looking at exd expression in the crustacean Artemia. The expression of Dll is restricted to the appendages (Panganiban et al., 1995), in close parallelism to Drosophila. We have found that our anti-exd antibody, made against the Drosophila exd homeodomain, is able to immunoreact in Artemia embryos, which is not surprising since the sequence similarity between members of the PBC family is very high (up to 95 % identity between the exd and the pbx homeodomains).

In early nauplius larvae, exd immunoreactivity is detected along the trunk and also in the limb primordia (Fig. 3A), but as development proceeds the exd product disappears from the most distal tip of the limb bud (Fig. 3B). In older nauplius limbs, exd staining becomes restricted to the gnathobase and the endites in the proximal regions and no staining is detected in the growing exopodite and endopodite that will form the most distal elements of the crustacean natatory biramous appendage (Fig. 3C) (Brusca and Brusca, 1990). Double staining for exd and Dll shows that Dll, but not exd, is present in these most distal precursors (Fig. 3D,E) (Panganiban et al., 1995). In conclusion, the exd product in Artemia, just like in Drosophila, is confined to the proximal limb precursors.

Fig. 3.

exd and Dll expression in a crustacean biramous appendage. In all cases, anterior is to the top and thoracic segments are shown beginning with the first one. Note that the more anterior limb primordia are more developmentally advanced.(A)Young Artemia nauplius larvae showing exd staining along the trunk and the limb primordia (black). (B) Slightly older nauplius. exd staining disappears from the distal tip of the most advanced limb primordia (arrowheads). (C) Ventral view of an older nauplius. exd staining is restricted to proximal parts such as the gnathobase (g) and the endites (e) shown in the second thoracic segment. No exd staining is detected in the distal endopod (‘endo’ in the second thoracic segment) and exopod (‘exo’ in the first thoracic segment in a lower focal plane). (D,E) Double staining for exd and Dll. Dll (brown), but not exd (black), is present in the distal endopod (‘endo’ in D) and exopod (‘exo’ in E; lower focal plane than in D).

Fig. 3.

exd and Dll expression in a crustacean biramous appendage. In all cases, anterior is to the top and thoracic segments are shown beginning with the first one. Note that the more anterior limb primordia are more developmentally advanced.(A)Young Artemia nauplius larvae showing exd staining along the trunk and the limb primordia (black). (B) Slightly older nauplius. exd staining disappears from the distal tip of the most advanced limb primordia (arrowheads). (C) Ventral view of an older nauplius. exd staining is restricted to proximal parts such as the gnathobase (g) and the endites (e) shown in the second thoracic segment. No exd staining is detected in the distal endopod (‘endo’ in the second thoracic segment) and exopod (‘exo’ in the first thoracic segment in a lower focal plane). (D,E) Double staining for exd and Dll. Dll (brown), but not exd (black), is present in the distal endopod (‘endo’ in D) and exopod (‘exo’ in E; lower focal plane than in D).

Ectopic exd expression prevents distal leg development

This parallelism between Drosophila and Artemia concerning the restriction of the exd product to proximal regions suggests that its elimination from distal precursors is a requisite for arthropod limb development. To test this idea, we ectopically expressed exd in distal regions of the Drosophila leg using the GAL4/UAS system (Brand and Perrimon, 1993).

One driver that we used is the GAL4 line E132 (Halder et al., 1995), which directs gene expression in the central region of the leg disc (Fig. 4A). Ectopic exd expression in this region (Fig. 4B) causes deformations in the central folds mainly due, to ring fusions, as visualized by Dll staining (Fig. 4C). The phenotypic effects in the adult leg are shown in Fig. 4D, where the basitarsus is fused to the proximal femur due to disappearance of distal femur and tibia.

Fig. 4.

exd prevents distal development in the Drosophila leg. In all cases, exd is in red and Dll in green. (A) Gene expression driven by the GAL4 line E132 (Halder et al., 1995) as revealed by X-gal staining using an UAS-lacZ strain. (B) exd expression in a pair of leg discs from a cross E132 β UAS-exd. Endogenous exd expression is localized to the periphery, while ectopic exd is in the center. (C) Merged channels of the discs in B. Cells expressing both genes appear yellow. The internal rings are not concentric and most of them are fused to the contiguous (white arrowheads). (D) Adult leg resulting after expressing exd with the E132 driver. (E) Gene expression driven by a GAL4 insertion in the Dll locus which is identical to the endogenous Dll expression pattern (Calleja et al., 1996). (F) exd expression in a third instar larval leg disc from a cross Dll-GAL4 β UAS-exd. exd is present all over the disc, which is shown at 1.5 times the magnification for the ones in E, thus reflecting its great reduction in size after ectopic exd expression. (G) Merged channels of the disc in F. The center of the disc (yellow) is dramatically reduced. No internal rings are observed and the remaining Dll-expressing cells (which also contain exd) form a round flat structure in the center. (H) Adult leg resulting after expressing exd in the Dll domain. (I) Lower magnification of the specimen in H showing the complete set of legs. (J) Adult leg after expressing exd with a dpp-GAL4 driver all along the P-D axis (Wilder and Perrimon, 1995). Coxa, trochanter and proximal femur are not affected by exd overexpression, but all distal structures are reduced. (K) TUNEL staining of a wild-type leg disc. Very few randomly distributed apoptotic nuclei are detected. (L) TUNEL staining of a leg disc from a cross Dll-GAL4 β UAS-exd. Note the large reduction in disc size (same magnification as in K) and the presence of multiple apoptotic nuclei located in the center of the disc where exd is ectopically expressed. Abbreviations: st, sternopleura; c, coxa; tr, trochanter; f, femur; ti, tibia; ta, tarsi.

Fig. 4.

exd prevents distal development in the Drosophila leg. In all cases, exd is in red and Dll in green. (A) Gene expression driven by the GAL4 line E132 (Halder et al., 1995) as revealed by X-gal staining using an UAS-lacZ strain. (B) exd expression in a pair of leg discs from a cross E132 β UAS-exd. Endogenous exd expression is localized to the periphery, while ectopic exd is in the center. (C) Merged channels of the discs in B. Cells expressing both genes appear yellow. The internal rings are not concentric and most of them are fused to the contiguous (white arrowheads). (D) Adult leg resulting after expressing exd with the E132 driver. (E) Gene expression driven by a GAL4 insertion in the Dll locus which is identical to the endogenous Dll expression pattern (Calleja et al., 1996). (F) exd expression in a third instar larval leg disc from a cross Dll-GAL4 β UAS-exd. exd is present all over the disc, which is shown at 1.5 times the magnification for the ones in E, thus reflecting its great reduction in size after ectopic exd expression. (G) Merged channels of the disc in F. The center of the disc (yellow) is dramatically reduced. No internal rings are observed and the remaining Dll-expressing cells (which also contain exd) form a round flat structure in the center. (H) Adult leg resulting after expressing exd in the Dll domain. (I) Lower magnification of the specimen in H showing the complete set of legs. (J) Adult leg after expressing exd with a dpp-GAL4 driver all along the P-D axis (Wilder and Perrimon, 1995). Coxa, trochanter and proximal femur are not affected by exd overexpression, but all distal structures are reduced. (K) TUNEL staining of a wild-type leg disc. Very few randomly distributed apoptotic nuclei are detected. (L) TUNEL staining of a leg disc from a cross Dll-GAL4 β UAS-exd. Note the large reduction in disc size (same magnification as in K) and the presence of multiple apoptotic nuclei located in the center of the disc where exd is ectopically expressed. Abbreviations: st, sternopleura; c, coxa; tr, trochanter; f, femur; ti, tibia; ta, tarsi.

Using a Dll-GAL4 driver (Fig. 4E) (Calleja et al., 1996), we expressed exd all over the Dll domain (Fig. 4F). This results in a complete disappearance of the central rings and a severe reduction in the disc size (Fig. 4G). The adult legs appear truncated, only containing coxa, trochanter and proximal femur (Fig. 4H,I), a phenotype very similar to that observed in strong Dll hypomorphic alleles (Cohen at al., 1989).

We also targeted exd expression all along the P-D axis of the leg in the dorsal anterior regions of both exd and Dll domains using a dpp-GAL4 driver (Wilder and Perrimon, 1995). In this case, exd overexpression produces similar morphological disturbances as described above, but the effects are precisely restricted to the distal leg (Fig. 4J). The excess of exd product in the proximal regions does not have any consequence (Fig. 4J).

Altogether, these results indicate that distal structures cannot develop in the presence of exd. Therefore, the restriction of exd expression to proximal precursors is a requisite during limb development.

Although ectopic exd expression in the Dll domain gives rise to a phenotype similar to that of Dll mutants, exd does not repress Dll expression. In E132/UAS-exd and Dll-GAL4/UAS-exd leg discs both products are present in the corresponding central cells (Fig. 4C,G). Instead, ectopic exd seems to interfere with cell proliferation in the central region of the disc which, in the wild-type situation, grows exponentially during larval stages (Bryant and Schneiderman, 1969). We applied the TUNEL technique (Gavrieli et al., 1992), which stains apoptotic nuclei, to detect programmed cell death in imaginal discs. In wild-type discs, very few cells undergo apoptosis (Fig. 4K) but, when exd is ectopically expressed, a large number of nuclei appear stained in the central region (Fig. 4L), indicating that ectopic exd results in cell death induction in the distal domain of the leg.

Normal exd expression in the absence of hh function

The hh gene plays a major role in controlling pattern formation along the P-D axis of the leg (Campbell et al., 1993; Basler and Struhl, 1994; Tabata and Kornberg, 1994; Díaz-Benjumea et al., 1994). In theabsence of hh function, Dll expression is totally abolished in the central regions of the leg disc, although the proximal ring is unaffected (Díaz-Benjumea et al., 1994). We analyzed whether exd expression is dependent on hh function using a temperature-sensitive hh allele (Ma et al., 1993). As expected, Dll expression in the hh− discs was confined to the femur ring (Fig. 5A) (Díaz-Benjumea et al., 1994), thus confirming the loss of hh function. However, exd expression remained completely normal (Fig. 5A). This is particularly clear in a sagittal view (Fig. 5B) as compared to a wild-type disc (Fig. 5C). This result, together with the fact that hh null clones in the leg only disturb distal regions (Mohler, 1988), indicates that the proximal leg is specified independently of hh function.

Fig. 5.

Normal exd expression in the absence of hh function. (A)Frontal view of a hh− leg disc stained for exd (red) and Dll (green). The yellow ring corresponds to the Dll expression in the trochanter and proximal femur which is the only one that remains and overlaps with the exd expression domain which, instead, appears normal. (B) Sagittal view of another hh− leg disc stained as in A. Note the complete absence of the central folds that correspond to the distal parts and are clearly viewed in the wild-type disc shown in C. The exd expression pattern remains unaffected. The yellow ring is indicated by arrowheads. (C) Sagittal view of a wild-type disc stained as the two previous ones. Also, the yellow ring is indicated by arrowheads. The hh− discs in A, B have been magnified 1.5 times more than this wild-type disc, illustrating the reduction of disc size in the absence of hh function.

Fig. 5.

Normal exd expression in the absence of hh function. (A)Frontal view of a hh− leg disc stained for exd (red) and Dll (green). The yellow ring corresponds to the Dll expression in the trochanter and proximal femur which is the only one that remains and overlaps with the exd expression domain which, instead, appears normal. (B) Sagittal view of another hh− leg disc stained as in A. Note the complete absence of the central folds that correspond to the distal parts and are clearly viewed in the wild-type disc shown in C. The exd expression pattern remains unaffected. The yellow ring is indicated by arrowheads. (C) Sagittal view of a wild-type disc stained as the two previous ones. Also, the yellow ring is indicated by arrowheads. The hh− discs in A, B have been magnified 1.5 times more than this wild-type disc, illustrating the reduction of disc size in the absence of hh function.

The exd and Dll domains in the arthropod leg

The results presented here and those in Cohen and Jürgens (1989) indicate the existence in the Drosophila leg of two domains along the P-D axis with distinct genetic regulation.

The proximal domain is defined by the identical expression of exd and tsh in body wall, coxa, trochanter and proximal femur precursors. Although the functional requirements for tsh have not yet been described, exd is clearly required for normal patterning in these proximal leg parts (González-Crespo and Morata, 1995; Rauskolb et al., 1995). Another common feature of exd and tsh is that they are also expressed in the embryonic trunk, where they play an important role in specifying segment identity (Peifer and Wieschaus, 1990; Röder et al., 1992).

The distal domain is defined by the expression and requirement of Dll, which extends from trochanter to tarsus (Cohen and Jürgens, 1989). Since removal of Dll function results in elimination of distal leg structures, it has been proposed that Dll activates the formation of limb structures above a groundstate of non-limb or body wall (Cohen and Jürgens, 1989). The distinction between ‘limb’ and ‘body wall’ fates is observed already during embryogenesis when Dll is expressed in the invaginating thoracic imaginal disc primordia while exd and tsh are confined to the surrounding epidermal cells. It is worth mentioning that Dll also has an embryonic function, which is the formation of the rudimentary larval appendages called Keilin’s organs (Cohen et al., 1989), in which exd and tsh expression is eliminated.

Additional evidence supporting this two-domain model is the finding that exd and Dll are regulated differently. The expression of Dll in the leg requires hh function, which through dpp and wingless (wg) activates Dll (Díaz-Benjumea et al., 1994), whereas exd is unaffected by the loss of hh function. This distinct gene regulation is also reflected in the expression of omb, a target gene of dpp in the leg (Grim and Pflugfelder,1996). Although dpp is expressed all along the P-D axis, omb expression is only activated in the Dll domain, indicating a control mechanism that discriminates between the two domains.

The fact that, in the trochanter and proximal femur, the expression of exd and Dll overlap argues in principle against a two-domain model. This region is represented in the mature imaginal disc by a proximal Dll ring separated from the main body of Dll expression. However, this Dll ring, which is independent of hh, seems to be a late expression, because we do not observe any overlap between the exd and Dll domains during the second larval instar. Therefore, we propose an original non-overlapping distribution of both products (as suggested by the expression in the younger discs), which would have been secondarily modified by the late expression of Dll in the proximal ring. This idea is also supported by transplantation experiments (Schubiger, 1974) indicating that proximal and distal leg regions are determined first and the middle regions (trochanter, femur and tibia) are determined later, perhaps through interac-tions between proximal and distal cells.

The two-domain model is also supported by the fact that exd and Dll show comparable expression patterns in Drosophila and Artemia. Dll homologs have been isolated in vertebrates and Artemia (Bürglin, 1995; Panganiban et al., 1995), and exd shows a high level of sequence conservation in Drosophila, humans and nematodes (Rauskolb et al., 1993; Bürglin, 1995). Our results indicate that, at least in the arthropods, their respective expression patterns in the limb are also conserved.

The negative effects of ectopic exd expression in the leg distal region strongly suggests that the parallelism between Drosophila and Artemia concerning the elimination of the exd product from the distal precursors is a general phenomenon during appendage formation in arthropods. This process probably involved the singling out of specific groups of cells recruited from the body wall that have to undergo additional proliferation and change of identity. In both Drosophila and Artemia, the gene Dll is specifically activated in those cells and, later, the exd product disappears from the Dll-expressing cells. We do not know why exd function has to be eliminated from those cells, but the Dll-like mutant phenotype caused by ectopic exd expression in the telopodite suggests that exd may be interfering with Dll function. The elimination of distal components associated with localized cell death in central regions of imaginal discs observed after ectopic exd expression in the Dll domain has also been observed in mutations reducing dpp activity (Bryant, 1988), which acts downstream of hh in controlling Dll expression and pattern formation in imaginal discs. In fact, the Dll-GAL4/UAS-exd leg discs are morphologically similar to the ones in which hh function was eliminated in early second larval instar. Even in the proximal ring corresponding to the trochanter and proximal femur where the exd and Dll products coincide, the presence of exd appears to interfere with Dll function since exd− clones in this area develop with a distal identity, as if, in the absence of exd function, the Dll product is able to perform its role in determining distal development. We might speculate that exd suppresses Dll functionally, but not transcriptionally, as in other cases of interactions between Hox gene products (González-Reyes and Morata, 1990).

Genetic evidence in support of the Snodgrass hypothesis

Based on comparative morphology, Snodgrass in 1935 proposed that primitive arthropod appendages were originally formed by two segments, a proximal coxopodite derived from the trunk, and a distal telopodite or outer limb. This primitive appendage would have evolved by acquisition of joints both in the coxopodite and the telopodite (Snodgrass, 1935).

This view was reinforced by recent analysis of Lower Permian Diaphanopterodea fossils, which represent the most primitive known pterygote order. These primitive insects had abdominal appendages in which the free limb was formed by distal segments from prefemur to tarsus, while the proximal segments were included in the ventral trunk (reviewed in Kukalová-Peck, 1991).

Our results provide genetic evidence for the Snodgrass hypothesis (Fig. 6). On the one hand, the fact that trunk-related genes such as exd and tsh are associated with the development of proximal leg parts strongly supports that these originated as an expansion of the body wall and provides a genetic definition of the coxopodite. On the other hand, the true outer limb or telopodite would correspond to the Dll domain (Cohen and Jürgens, 1989) which development relies on the hh signalling pathway and requires the elimination of trunk genes like exd to avoid their inhibitory effects on distal growth.

Fig. 6.

The two major genetic domains along the P-D axis of the arthropod limb. The base of the appendages, or coxopodite (red), would be controlled by genes like exd. The distal part of the appendages, or telopodite (green), would correspond to the Dll domain and represents the true limb.This distal region requires the function of the hh pathway as well as the elimination of trun genes like exd to avoid their inhibitory effects on distal growth.

Fig. 6.

The two major genetic domains along the P-D axis of the arthropod limb. The base of the appendages, or coxopodite (red), would be controlled by genes like exd. The distal part of the appendages, or telopodite (green), would correspond to the Dll domain and represents the true limb.This distal region requires the function of the hh pathway as well as the elimination of trun genes like exd to avoid their inhibitory effects on distal growth.

We thank Stephen Cohen for helpful discussions and a Dll antibody, Eric Wieschaus for the exd cDNA, Sean Carroll and Grace Panganiban for a Dll antibody, Walter Gehring for the E132 and dpp-GAL4 lines, Gary Struhl for the omb-lacZ line, Leandro Sastre for help with the Artemia experiments, Eduardo Moreno for the Dll-GAL4 line, Rosa González for technical assistance, Sonsoles Campuzano for help with the TUNEL technique, Ruth Diez del Corral and Ana Carmena for advice on confocal microscopy, and Ernesto Sánchez-Herrero and Isabel Guerrero for critically reading the manuscript. S. G.-C. is a postdoctoral fellow from the Spanish Ministerio de Educación y Ciencia. This work was supported by grants from the Dirección General de Investigación Científica y Técnica and the Human Frontier Science Program Organization (RG-372/1994) to G. M. and by an institutional grant from the Fundación Ramón Areces.

Basler
,
K.
and
Struhl
,
G.
(
1994
).
Compartment boundaries and the control of Drosophila limb pattern by hedgehog protein
.
Nature
368
,
208
214
.
Botas
,
J.
(
1993
).
Control of morphogenesis and differentiation by HOM/Hox genes
.
Curr. Opin. Cell Biol.
5
,
1015
1022
.
Brand
,
A. H.
and
Perrimon
,
N.
(
1993
).
Targeted gene expression as a means of altering cell fates and generating dominant phenotypes
.
Development
118
,
401
415
.
Brusca
,
R. C.
and
Brusca
,
G. J.
(
1990
).
Invertebrates.
Sunderland
:
Sinauer Associates
.
Bryant
,
P. J.
(
1988
).
Localized cell death caused by mutations in a Drosophila gene coding for a transforming growth factor-β homolog
.
Dev. Biol.
128
,
386
395
.
Bryant
,
P. J.
and
Schneiderman
,
H. A.
(
1969
).
Cell lineage, growth and determination in the imaginal leg discs in Drosophila melanogaster. Dev. Biol.
20
,
263
290
.
Bürglin
,
T. R.
(
1995
).
The evolution of homeobox genes
. In
Biodiversity and Evolution (eds. R. Arai, M. Kato and Y. Doi)
, pp.
291
-
336
. Tokyo: The National Science Museum Foundation.
Calleja
,
M.
,
Moreno
,
E.
,
Pelaz
,
S.
and
Morata
,
G.
(
1996
).
Visualization of gene expression in living adult Drosophila
.
Science (in press)
.
Campbell
,
G.
,
Weaver
,
T.
and
Tomlinson
,
A.
(
1993
).
Axis specification in the developing Drosophila appendage: the role of wingless, decapentaplegic, and the homeobox gene aristaless
.
Cell
74
,
1113
1123
.
Cohen
,
S. M.
,
Brönner
,
G.
,
Küttner
,
F.
,
Jürgens
,
G.
and
Jäckle
,
H.
(
1989
).
Distal-less encodes a homeodomain protein required for limb development in Drosophila
.
Nature
338
,
432
434
.
Cohen
,
S. M.
and
Jürgens
,
G.
(
1989
).
Proximal-distal pattern formation in Drosophila: cell autonomous requirement for Distal-less gene activity in limb development
.
EMBO J.
8
,
2045
2055
.
Cohen
,
B.
,
Simcox
,
A. A.
and
Cohen
,
S. M.
(
1993
).
Allocation of the thoracic imaginal primordia in the Drosophila embryo
.
Development
117
,
597
608
.
Díaz-Benjumea
,
F. J.
,
Cohen
,
B.
and
Cohen
,
S. M.
(
1994
).
Cell interaction between compartments establishes the proximal-distal axis of Drosophila legs
.
Nature
372
,
175
179
.
Fristrom
,
D.
and
Fristrom
,
J. W.
(
1993
).
The metamorphic development of the adult epidermis
. In
The Development of Drosophila melanogaster (eds. M. Bate and A. Martínez-Arias)
, pp.
843
897
.
New York
:
Cold Spring Harbor Laboratory Press
.
Gavrieli
,
Y.
,
Sherma
,
Y.
and
Ben-Sasson
,
S.A.
(
1992
).
Identification of programmed cell death in situ via specific labeling of nuclear DNA fragmentation
.
J. Cell Biol.
119
,
493
501
.
González-Crespo
,
S.
and
Morata
,
G.
(
1995
).
Control of Drosophila adult pattern by extradenticle
.
Development
121
,
2117
2125
.
González-Reyes
,
A.
and
Morata
,
G.
(
1990
).
The developmental effect of overexpressing a Ubx product in Drosophila embryos is dependent on its interactions with other homeotic products
.
Cell
61
,
515
522
.
Grim
,
S.
and
Pflugfelder
,
G. O.
(
1996
).
Control of the gene optomotor-blind in Drosophila wing development by decapentaplegic and wingless
.
Science
271
,
1601
1604
.
Halder
,
G.
,
Callaerts
,
P.
and
Gehring
,
W.J.
(
1995
).
Induction of ectopic eyes by targeted expression of the eyeless gene in Drosophila
.
Science
267
,
1788
1792
.
Kukalová-Peck
,
J.
(
1991
).
Fossil history and the evolution of hexapod structures
. In
The Insects of Australia, 2nd edn
, Vol.
1
, pp.
141
179
.
Ithaca, New York
:
Cornell University Press
.
Lawrence
,
P. A.
and
Morata
,
G.
(
1994
).
Homeobox genes: Their function in Drosophila segmentation and pattern formation
.
Cell,
78
,
181
189
.
Ma
,
C.
,
Zhou
,
Y.
,
Beachy
,
P. A.
and
Moses
,
K.
(
1993
).
The segment polarity gene hedgehog is required for progression of the morphogenetic furrow in the developing Drosophila eye
.
Cell
75
,
927
938
.
Mann
,
R. S.
and
Abu-Shaar
,
M.
(
1996
).
Nuclear import of the homeobox protein Extradenticle induced by Wg and Dpp signalling
.
Nature (in press)
.
Mann
,
R. S.
and
Chan
,
S.
(
1996
).
Extra specificity from extradenticle: the partnership between HOX and PBX/EXD homeodomain proteins.Trends in Genetics
12
,
258
262
.
McGinnis
,
W.
and
Krumlauf
,
R.
(
1992
).
Homeobox genes and axial patterning
.
Cell
68
,
283
302
.
Mohler
,
J.
(
1988
).
Requirements for hedgehog, a segmental polarity gene, in patterning larval and adult cuticle of Drosophila
.
Genetics
120
,
1061
1072
.
Panganiban
,
G.
,
Sebring
,
A.
,
Nagy
,
L.
and
Carroll
,
S.
(
1995
).
The development of crustacean limbs and the evolution of arthropods
.
Science
270
,
1363
1366
.
Patel
,
N. H.
(
1994
).
Imaging neuronal subsets and other cell types in whole-mount Drosophila embryos and larvae using antibody probes
.
Meth. in Cell Biol.
44
,
445
487
.
Peifer
,
M.
and
Wieschaus
,
E.
(
1990
).
Mutations in the Drosophila gene extradenticle affect the way specific homeo domain proteins regulate segmental identity
.
Genes Dev.
4
,
1209
1223
.
Rauskolb
,
C.
,
Peifer
,
M.
and
Wieschaus
,
E.
(
1993
).
extradenticle, a regulator of homeotic gene activity, is a homolog of the homeobox-containing human proto- oncogene pbx1
.
Cell
74
,
1101
1112
.
Rauskolb
,
C.
,
Smith
,
K. M.
,
Peifer
,
M.
and
Wieschaus
,
E.
(
1995
).
extradenticle determines segmental identities throughout Drosophila development
.
Development
121
,
3663
3673
.
Röder
,
L.
,
Vola
,
C.
and
Kerridge
,
S.
(
1992
).
The role of the teashirt gene in trunk segmental identity in Drosophila
.
Development
115
,
1017
1033
.
Schubiger
,
G.
(
1974
).
Acquisition of differentiative competence in the imaginal leg discs of Drosophila. Wilhelm Roux’ Archiv. EntwMech. Org.
174
,
303
311
.
Snodgrass
,
R. E.
(
1935
).
Principles of Insect Morphology,
pp.
83
99
.
New York
:
McGraw-Hill
.
Tabata
,
T.
and
Kornberg
,
T.
(
1994
).
Hedgehog is a signalling protein with a key role in patterning Drosophila imaginal discs
.
Cell
76
,
89
102
.
Wilder
E. L.
and
Perrimon
,
N.
(
1995
).
Dual functions of wingless in the Drosophila leg imaginal disc
.
Development
121
,
477
488
.
Wilson
,
D. S.
and
Desplan
,
C.
(
1995
).
Cooperating to be different
.
Curr. Biol.
5
,
32
34
.

In agreement with recent results of Mann and Abu-Shaar (Nature, in press), we have also detected low levels of exd staining in the cytoplasm of distal cells of third instar larval leg discs.