The secreted protein Jelly belly (Jeb) is required for an essential signalling event in Drosophila muscle development. In the absence of functional Jeb, visceral muscle precursors are normally specified but fail to migrate and differentiate. The structure and distribution of Jeb protein implies that Jeb functions as a signal to organize the development of visceral muscles. The Jeb receptor is the Drosophila homologue of anaplastic lymphoma kinase (Alk), a receptor tyrosine kinase of the insulin receptor superfamily. Human ALK was originally identified as a proto-oncogene, but its normal function in mammals is not known. Drosophila Alk was identified using a degenerate PCR approach (Lorén, 2001). Like its mammalian counterpart, DAlk appears to be expressed in the developing CNS by in situ analysis. However, in addition to expression of DAlk in the Drosophila brain, careful analysis reveals an additional early role for DAlk in the developing visceral mesoderm where its expression is coincident with activated ERK (Lorén, 2001). In Drosophila, localized Jeb activates Alk and the downstream Ras/mitogen-activated protein kinase cascade to specify a select group of visceral muscle precursors as muscle-patterning pioneers. Jeb/Alk signalling induces the myoblast fusion gene dumbfounded (duf; also known as kirre) as well as optomotor-blind-related-gene-1 (org-1), a Drosophila homologue of mammalian TBX1, in these cells (Lee, 2003).

Signalling molecules and their receptors orchestrate cell fate decisions essential to organogenesis. Studies of mesoderm development in Drosophila have highlighted the role of evolutionarily conserved signalling systems, and the transcription factors they regulate, in the elaboration of the mesoderm into its derivative tissues. The earliest cell fate assignments in the mesoderm are coordinated by inductive signals from the ectoderm. Decapentaplegic (Dpp), a Drosophila BMP signal, induces subjacent dorsal mesoderm to express Tinman (Tin), a homeodomain protein essential for heart, visceral and dorsal somatic mesoderm development. Dpp and Tin, together with Hedgehog, induce visceral mesoderm by activating the expression of two transcription factors, Bagpipe (Bap) and Biniou (Bin). A third signal, Wingless, antagonizes these visceral mesoderm-inducing activities. The combined actions of ectodermally derived Dpp, Hedgehog and Wingless generate segmental clusters of visceral mesoderm precursors in the dorsal mesoderm (Lee, 2003 and references therein).

The secreted protein Jeb is necessary for the subsequent rearrangement of these segmental clusters of visceral mesoderm precursors into bilateral longitudinal bands and for visceral muscle differentiation. Jeb is produced in ventral somatic mesoderm, locally secreted, and is specifically taken up by the visceral mesoderm cells. Its detailed developmental role, however, has not been defined. One critical function of Jeb signalling is to subdivide the pool of visceral mesoderm precursors into two distinct subtypes: muscle founders and fusion-competent cells. This subdivision is key to the muscle specification and fusion pathway, a hierarchical system for patterning muscles. As first shown for somatic muscle development in Drosophila, founder myoblasts are patterning pioneers. They establish specific muscles and recruit fusion-competent myoblasts to fuse with them into mature syncytial muscle fibers. Founder myoblasts and fusion-competent myoblasts are identified by the expression of functional components of the myoblast fusion pathway. Founder cells express Duf, a transmembrane protein necessary for recruitment of fusion-competent cells. Fusion-competent cells express Sticks and stones (Sns), a transmembrane protein also required for fusion (Lee, 2003 and references therein).

Positive regulation of duf and negative regulation of sns implies that Jeb signalling specifies visceral mesoderm founders. As assayed by the markers duf, org-1 and sns, no visceral muscle founders are specified in jeb mutant embryos. Instead all visceral mesoderm precursors become fusion-competent myoblasts. The consequence of absent visceral mesoderm founders, as shown by cell-lineage experiments, is fusion of visceral fusion-competent myoblasts with somatic muscle founders and loss of visceral musculature. Somatic muscle patterning, however, is unaffected (Lee, 2003).

Localized activation of the Ras/mitogen-activated protein kinase (MAPK) cascade in the visceral mesoderm has been noted previously. In the somatic muscle lineage this pathway is required for founder cell specification. It was therefore hypothesized that Jeb signals through the Ras/MAPK cascade in the visceral mesoderm. Activated MAPK is indeed detected in the visceral mesoderm precursors that take up Jeb. The observed overlapping signals for diphospho-MAPK and org-1, as well as the exclusive staining patterns for diphospho-MAPK and sns, confirm that the MAPK pathway is activated in presumptive visceral muscle founders. Moreover, Jeb signalling is necessary and sufficient to activate the Ras/MAPK cascade in visceral mesoderm precursors. Immunostaining of jeb mutant embryos demonstrates absent diphospho-MAPK in the ventral visceral mesoderm cells that normally accumulate Jeb and become founders. As with founder cell markers, ectopic Jeb produces ectopic diphospho-MAPK, but only in the visceral mesoderm (Lee, 2003).

The expanded expression of org-1 upon mesodermal expression of activated versions of Drosophila Ras and human Raf implicates the Ras pathway in MAPK activation and founder cell specification in the visceral mesoderm. If Jeb signals through the Ras/MAPK pathway, then activation of this pathway should rescue jeb mutations. This prediction is true. As judged by expression of Fasciclin III, a marker of visceral mesoderm differentiation, expression of activated Ras can substantially rescue jeb mutant embryos (Lee, 2003).

The observed effects of ectopic Jeb are limited to the visceral mesoderm. Together with the observation that uptake of Jeb into visceral mesoderm cells requires Shibire-mediated endocytosis, these data imply that Jeb acts through a tissue-specific receptor, which is coupled to the Ras/MAPK pathway. The receptor tyrosine kinase Drosophila Alk, a homologue of the human proto-oncogene anaplastic lymphoma kinase (ALK), is expressed in the early visceral mesoderm. It was therefore hypothesized that Drosophila Alk is the Jeb receptor. Alk messenger RNA is expressed in all cells of the trunk visceral mesoderm directly adjacent to the Jeb-expressing cells. In visceral mesoderm cells that both express Alk and take up Jeb1, diphospho-MAPK is detected (Lee, 2003).

Tested was the assumption that Alk activity, similar to Jeb, would be required for the specification of visceral mesoderm founder cells. Embryos homozygous for a deficiency uncovering the Alk locus lack org-1 expression in presumptive visceral mesoderm founders, a phenotype that can be rescued by expressing an Alk minigene in visceral mesoderm precursors. Mesodermal expression of a kinase-deficient, dominant interfering form of Alk produces an identical phenotype. RNA-mediated interference (RNAi) injection experiments further confirm that Alk is specifically required for visceral mesoderm founder specification. Gal staining of bap3-lacZ embryos injected with double-stranded (ds)Alk RNA demonstrates transformation of visceral into somatic muscle fates. Furthermore, injection of dsAlk RNA into duf-lacZ embryos results in strongly reduced or absent expression of this founder cell marker in the visceral mesoderm. These RNAi phenotypes resemble the phenotypes of jeb mutant embryos, although they are less severe (Lee, 2003).

The loss of duf expression and expansion of sns expression in the visceral mesoderm on expression of dominant-negative Alk is identical to a jeb null mutant phenotype as well. Conversely, the expansion of org-1 expression in the visceral mesoderm on expression of activated Alk (a fusion protein analogous to the human oncogenic version, NPM-ALK22) is indistinguishable from the effects of expression of ectopic Jeb, activated Ras and activated Raf. Finally, forced expression of activated Alk in homozygous jeb mutant backgrounds is able to rescue (and compared with wild type expand) org-1 expression in the visceral mesoderm and to restore midgut morphogenesis (Lee, 2003).

To confirm that Jeb signals through Alk, it was determined that Jeb binds Alk with high affinity, and that Jeb binding to Alk activates the Ras/MAP kinase cascade. In these experiments Jeb-alkaline phosphatase fusion proteins (Jeb-AP) was used. To establish qualitatively the binding of Jeb to Alk, the specific association of Jeb-AP with Alk-transfected mammalian tissue culture cells was visualized. Alk-transfected cells bind Jeb-AP. By contrast, Alk-transfected cells do not bind either an equivalent concentration of alkaline phosphatase alone or a Jeb-AP fusion protein that lacks the type-A LDL receptor repeat in Jeb. This truncated version of Jeb resembles a mutant protein encoded by a null allele of jeb. The truncated protein does not accumulate in visceral mesoderm cells. Binding of Jeb depends on Alk, as demonstrated with non-transfected cells that were incubated with full-length Jeb-AP (Lee, 2003).

A similar assay was used to demonstrate that the Jeb-Alk interaction is specific and has high affinity. Jeb binding to Alk-transfected cells is saturable at nanomolar concentrations. Scatchard analysis demonstrates a single class of high-affinity Jeb-binding site with a dissociation constant (Kd) of 2.2 nM. No binding was observed with either alkaline phosphatase alone or Jeb-AP that lacks the type-A LDL receptor repeat. Jeb-dependent activation of the Ras/MAP kinase cascade in this system was confirmed. The concentration dependence of Ras/MAP kinase activation by Jeb correlates well with binding data. Approximately half-maximal activation occurs in the range of 2-3 nM. As in vivo, removing the type-A LDL receptor repeat from Jeb abrogates Ras/MAP kinase activation (Lee, 2003).

This study has shown that Jeb activates the Ras/MAPK cascade both in vivo and in Alk-transfected tissue culture cells. Jeb binds Alk with high affinity. In vivo Jeb accumulates in visceral muscle founder cells and, in late-stage embryos, in axons of the central nervous system. These patterns of Jeb accumulation are absent from Alk-deficient embryos and in jeb mutants that produce an Alk-binding-deficient version of Jeb. Biochemical and genetic interference with Alk function produces phenotypes identical to jeb mutations. A critical function of Jeb signalling is to specify visceral muscle founder cells-patterning pioneers essential to midgut morphogenesis. Structurally Jeb belongs to a class of signalling molecules with type-A LDL receptor repeats as one of their functional domains. Others include Caenorhabditis elegans HEN-1 and MIG-13, and the mammalian proteins 8D6 and sco-spondin. Jeb is the first among these to have an identified signalling receptor and a defined biological pathway. It is anticipated that this discovery will lead to the identification of receptors and modes of action for other members of this class of signalling molecule (Lee, 2003).

The extracellular portions of mammalian and Drosophila Alk have common domain architectures. Their respective ligands are therefore also likely to share structural features. However, two closely related cytokines that are structurally unrelated to Jeb, pleiotrophin and midkine, have been identified by phage display as potential high-affinity ligands for human ALK. In Drosophila two clustered genes, miple1 and miple2, encode polypeptides related to midkine/pleiotrophin. Similar to the mammalian genes, Drosophila miple1 and miple2 are expressed widely during embryogenesis. So, unlike Jeb, Miple1 and Miple2 cannot control the spatially restricted activation of Alk in the visceral mesoderm, although they may have an auxiliary function in Alk activation. The potential functions of Jeb-related molecules in mammalian Alk activation and the possible contribution of midkine/pleiotrophin-related factors to Alk signalling in Drosophila can now be tested by genetic and molecular approaches. The characterization of the Jeb/Alk signalling pathway in Drosophila is also likely to enhance understanding of vertebrate Alk signalling in development and cancer. As most studies of mammalian Alk have focused on the role of oncogenic versions in cellular transformation, current understanding of Alk's normal function in mammals is rudimentary. In light of the known conservation of genetic pathways in the cardiac and splanchnic mesoderm, these insights into the regulation of org-1 expression in Drosophila are potentially relevant for the understanding of the regulation of human TBX1 and its roles in congenital cardiovascular and craniofacial disease. In addition, the specific expression of Drosophila and mouse Alk in the central nervous system suggests a conserved role of Alk signals in the development or function of neuronal tissues (Lee, 2003).

GENE STRUCTURE

In situ hybridization to polytene chromosomes isolated from third instar larva localized DAlk to region 53 on the second chromosome; therefore this novel Drosophila PTK locus was named DAlk53. Using the in situ mapping information, the genomic localization of Dalk was further defined to region 53C10-C11 in the Drosophila genome. In addition to mapping DAlk53, a careful characterization of this region was carried out. The data are derived from P1 genomic sequence data from the BDGP, genomic sequence data from four overlapping cosmids covering the DAlk53 locus, multiple DAlk cDNAs, P element data and the BDGP as well as STS and EST data from the BDGP. DAlk maps to an approximately 15 kb genomic fragment between sts3464 and sts0182 within P1 DS02309 and appears to comprises eight coding exons. Of the multiple DAlk cDNAs obtained, no alternative splicing events within the ORF were observed. However, an analysis of the 5' UTR of six independent cDNA clones revealed alternate splicing events 5' of the ORF. Both of these 5' UTR exons were further characterized and found to map upstream of the ORF encoding exons in the genome. An approximately 8.5 kb DAlk major RNA species was detected by Northern blot analysis, in good agreement with cDNA data, which predict a cDNA of 8466 bp, inclusive of both 5' and 3' UTR regions. The region surrounding DAlk53 contains few P element insertions, none of which are within the DAlk gene; the closest that has been characterized is some 40 kb 3' [l(2)k06503] and within the CDK4/6 gene. (Lorén, 2001).

cDNA clone length - 8435 bps

Bases in 5' UTR - 377

Exons - 9

Bases in 3' UTR - 2952

PROTEIN STRUCTURE

Amino Acids - 1701

Structural Domains

Full length cDNAs were obtained by screening Drosophila melanogaster adult cDNA libraries. Multiple cDNAs were obtained, falling into two classes, based on alternate splicing within the 5' UTR. No alternate splicing was observed within the ORF of these novel cDNA species. This locus was named DAlk53. The DAlk open reading frame predicts a 1701 amino acid, 180 kDa novel protein. Analysis of the predicted amino acid sequence reveals an amino terminal signal sequence, as well as a hydrophobic transmembrane domain. BLAST homology searching of the NCBI database revealed that DAlk does indeed appear to encode a novel RTK in the insulin receptor superfamily. A Drosphila insulin receptor has already been identified, but no Drosophila counterpart for the LTK/ALK single pass RTK branch of the INR superfamily has been described. This novel Drosophila RTK shows the most homology with a previously described mammalian RTK, ALK with 34% identity to ALK (52% in the cytoplasmic region) as well as a conserved overall structure. DAlk, like mammalian ALK, encodes for several putative domains, an amino-terminal signal sequence, an extracellular domain, a hydrophobic transmembrane region and a cytoplasmic PTK domain. The kinase domain of DAlk is most similar (58% identity; 85% homology with hALK) to those of the Insulin Receptor superfamily and contains several sequence motifs conserved among PTKs, including the tripeptide motif DFG that is found in most kinases, and a consensus ATP-binding motif GxGxxG followed by an AxK sequence downstream. The cytoplasmic domain of DAlk contains a ^NPNY putative IRS/Shc-binding consensus sequence at amino acid 1170, homologous to the ^NPXY motif in p80-NPM/ALK, which has been shown to bind to mammalian IRS1 when tyrosine phosphorylated. Within the amino-terminal extracellular domain of DAlk several features are found: (1) an LDLa domain, (2) a MAM domain (named after meprins, A-5 protein and receptor protein tyrosine phosphatase mu) and (3) a glycine-rich region. Careful analysis of the mammalian ALK sequence reveals that both the LDLa and MAM domains can also be seen in mammalian ALK, although they have not previously been noted. The MAM domain is not found in LTK. MAM domains comprise about 160 amino acids that are present in transmembrane proteins such as the meprins and receptor protein-tyrosine phosphatases, where they appear to function in cell/cell interactions. Thus it is proposed that the ALK family of RTKs is novel within the RTK family by virtue of a MAM domain within their extracellular domains. Currently the functions of this domain in the regulation of ALK family RTKs are unknown (Lorén, 2001 and references therein).

date revised: 30 November 2004

Home page: The Interactive Fly © 1995, 1996 Thomas B. Brody, Ph.D.

The Interactive Fly resides on the
Society for Developmental Biology's Web server.