A Screen for Modifiers of Deformed Function in Drosophila

Proteins produced by the homeotic genes of the Hox family assign different identities to cells on the anterior/posterior axis. Relatively little is known about the signalling pathways that modulate their activities or the factors with which they interact to assign specific segmental identities. To identify genes that might encode such functions, we performed a screen for second site mutations that reduce the viability of animals carrying hypomorphic mutant alleles of the Drosophila homeotic locus, Deformed. Genes mapping to six complementation groups on the third chromosome were isolated as modifiers of Deformed function. Products of two of these genes, sallimus and moira, have been previously proposed as homeotic activators since they suppress the dominant adult phenotype of Polycomb mutants. Mutations in hedgehog, which encodes secreted signalling proteins, were also isolated as Deformed loss-of-function enhancers. Hedgehog mutant alleles also suppress the Polycomb phenotype. Mutations were also isolated in a few genes that interact with Deformed but not with Polycomb, indicating that the screen identified genes that are not general homeotic activators. Two of these genes, cap `n' collar and defaced, have defects in embryonic head development that are similar to defects seen in loss of function Deformed mutants.     

Hox proteins coordinate peripodial decapentaplegic expression to direct adult head morphogenesis in Drosophila

The Drosophila BMP, decapentaplegic (dpp), controls morphogenesis of the ventral adult head through expression limited to the lateral peripodial epithelium of the eye-antennal disc by a 3.5kb enhancer in the 5' end of the gene. We recovered a 15bp deletion mutation within this enhancer that identified a homeotic (Hox) response element that is a direct target of labial and the homeotic cofactors homothorax and extradenticle. Expression of labial and homothorax are required for dpp expression in the peripodial epithelium, while the Hox gene Deformed represses labial in this location, thus limiting its expression and indirectly that of dpp to the lateral side of the disc. The expression of these homeodomain genes is in turn regulated by the dpp pathway, as dpp signalling is required for labial expression but represses homothorax. This Hox-BMP regulatory network is limited to the peripodial epithelium of the eye-antennal disc, yet is crucial to the morphogenesis of the head, which fate maps suggest arises primarily from the disc proper, not the peripodial epithelium. Thus Hox/BMP interactions in the peripodial epithelium of the eye-antennal disc contribute inductively to the shape of the external form of the adult Drosophila head.     

Expression of a homologue of the fushi tarazu (ftz) gene in a cirripede crustacean

In Metazoa, Hox genes control the identity of the body parts along the anteroposterior axis. In addition to this homeotic function, these genes are characterized by two conserved features: They are clustered in the genome, and they contain a particular sequence, the homeobox, encoding a DNA-binding domain. Analysis of Hox homeobox sequences suggests that the Hox cluster emerged early in Metazoa and then underwent gene duplication events. In arthropods, the Hox cluster contains eight genes with a homeotic function and two other Hox-like genes, zerknullt (zen)/Hox3 and fushi tarazu (ftz). In insects, these two genes have lost their homeotic function but have acquired new functions in embryogenesis. In contrast, in chelicerates, these genes are expressed in a Hox-like pattern, which suggests that they have conserved their ancestral homeotic function. We describe here the characterization of Diva, the homologue of ftz in the cirripede crustacean Sacculina carcini. Diva is located in the Hox cluster, in the same position as the ftz genes of insects, and is not expressed in a Hox-like pattern. Instead, it is expressed exclusively in the central nervous system. Such a neurogenic expression of ftz has been also described in insects. This study, which provides the first information about the Hoxcluster in Crustacea, reveals that it may not be much smaller than the insect cluster. Study of the Diva expression pattern suggests that the arthropod ftz gene has lost its ancestral homeotic function after the divergence of the Crustacea/Hexapoda clade from other arthropod clades. In contrast, the function of ftz during neurogenesis is well conserved in insects and crustaceans.     

Establishment of the Deformed expression stripe requires the combinatorial action of coordinate, gap and pair-rule proteins.

In Drosophila embryos, anterior-posterior positional identities are set and maintained by the expression boundaries of homeotic selector genes. The establishment of the initial expression boundaries of the homeotic genes are in turn dependent on earlier acting patterning genes of Drosophila. To define the combinations of early genes that are required to establish a unique blastoderm stripe of expression of the homeotic gene Deformed, we have analysed single and double patterning mutants and heat shock promoter fusion constructs that ectopically express early acting regulators. We find that the activation of Deformed is dependent on combinatorial input from at least three levels of the early hierarchy. The simplest activation code sufficient to establish Deformed expression, given the absence of negative regulators such as fushi-tarazu, consists of a moderate level of expression from the coordinate gene bicoid, in combination with expression from both the gap gene hunchback, and the pair-rule gene even-skipped. In addition, the activation code for Deformed is redundant; other pair-rule genes in addition to even-skipped can apparently act in combination with bicoid and hunchback to activate Deformed.     

The design and analysis of a homeotic response element

We have shown that the 26 bp bx1 element from the regulatory region of Distal-less is capable of imposing control by the homeotic genes Ultrabithorax and abdominal-A on a general epidermal activator in Drosophila. This provides us with an assay to analyze the sequence requirements for specific repression by these Hox genes. Both the core Hox binding site, 5'-TAAT, and the adjacent EXD 5'-TGAT core site are required for repression by Ultrabithorax and abdominal-A. The Distal-less bx1 site thus fits with the model of Hox protein binding specificity based on the consensus PBX/HOX-family site TGATNNAT[g/t][g/a], where the key elements of binding specificity are proposed to lie in the two base pairs following the TGAT. A single base pair deletion in the bx1 sequence generates a site, bx1:A(-)mut, that on the consensus PBX/HOX model would be expected to be regulated by the Deformed Hox gene. We observed, however, that the bx1:A(-)mut site was regulated predominantly by Sex combs reduced, Ultrabithorax and abdominal-A. The analysis of this site indicates that the specificity of action of Hox proteins may depend not only on selective DNA binding but also on specific post-binding interactions.     

Chelicerate Hox genes and the homology of arthropod segments

Genes of the homeotic complex (HOM-C) in insects and vertebrates are required for the specification of segments along the antero-posterior axis. Multiple paralogues of the Hox genes in the horseshoe crab Limulus poliphemus have been used as evidence for HOM-C duplications in the Chelicerata. We addressed this possibility through a limited PCR survey to sample the homeoboxes of two spider species, Steatoda triangulosa and Achaearanea tepidariorum. The survey did not provide evidence for multiple Hox clusters although we have found apparent duplicate copies of proboscipedia ( pb ) and Deformed ( Dfd   ). In addition, we have cloned larger cDNA fragments of pb, zerknullt ( zen / Hox3 ) and Dfd. These fragments allowed the determination of mRNA distribution by in situ hybridization. Our results are similar to the previously published expression patterns of Hox genes from another spider and an oribatid mite. Previous studies compared spider/mite Hox gene expression patterns with those of insects and argued for a pattern of segmental homology based on the assumption that the co-linear anterior boundaries of the Hox domains can be used as markers. To test this assumption we performed a comparative analysis of the expression patterns for UBX/ABD-A in chelicerates, myriapods, crustaceans, and insects. We conclude that the anterior boundary can be and is changed considerably during arthropod evolution and, therefore, Hox expression patterns should not be used as the sole criterion for identifying homology in different classes of arthropods.