BIOLOGICAL OVERVIEW

Beadex is one of the classic Drosophila mutations, discovered by Bridges in 1923 and described by Morgan, Bridges and Sturtevant in 1925. The gene is X-linked; therefore, only mutant males or homozygous females exhibit beaded-like wings, described as long, narrow and excised along both margins. The mutation is non-lethal. Lifschytz and Green (1979) proposed, with great insight, that Beadex is a dominant mutation in a negatively acting cis-regulatory region. These authors suggested that an overproducing mutation, one producing too much protein product, is expected if a dominant mutation is readily reverted by gene deletion, or, if gene deletion suppresses the expression of the original dominant mutation in flies heterozygous for the deletion. Beadex was shown to meet these criteria. Before Beadex was identified as coding for the LIM-only protein (Lmo), LIM-only was cloned on the basis of homology to genes coding for vertebrate LIM-only proteins. Only in 1998 did three papers appear, almost simultaneously, identifying Lmo and Beadex as being identical.

Beadex has been cloned by three different procedures: homology based cloning (Boehm, 1990, Zhu, 1995 and Shoresh, 1998), using polymerase chain reaction (PCR); P-element insertion (Zeng, 1998), and by a misexpression system (Milan, 1998). The latter method will be described because of its novelty. The modular-misexpression system developed by Rorth (1996) was used to carry out a large-scale screen for genes that perturb wing development (Rorth, 1998). A special P element carrying an EP (enhancer and a basal promoter), oriented to direct expression of adjacent genomic sequences, is randomly inserted into the genome. The EP is regulatable by manipulation of the transcription factor GAL4. When combined with a source of GAL4, the EP element will direct expression of any gene that happens to lie next to its insertion site. Thus the EP element directs a conditional misexpression of genes into which it inserts. The screen identified EP insertions at hh, patched (ptc), and dpp, genes with known roles in limb patterning (Rorth, 1998), as well as a number of new loci implicated in wing patterning by virtue of their overexpression phenotypes. One of the genes identified by the EP screen, Beadex, is involved in dorsal-ventral (DV) patterning of the wing (Milan, 1998).

The LIM-homeodomain protein Apterous (Ap) serves as the selector gene for the dorsal compartment wing imaginal disc. The juxtaposition of Ap-expressing dorsal cells and nonexpressing ventral cells establishes a D-V organizing center at the boundary between these cell types. In the late third instar disc, wingless and cut expression is induced in a three- to six-cell-wide stripe that straddles the D-V boundary; this is the location of the future wing margin. It is known that ap expression needs to be maintained continuously in the dorsal compartment, because loss of Ap function in mutant clones as late as third instar causes cell-autonomous fate transformation from dorsal to ventral. apterous mutant clones in the interior of the dorsal compartment create a new boundary for dorsal versus ventral cells, thereby initiating the genetic program normally reserved for the endogenous D-V boundary. wingless and cut are expressed in stripes of cells flanking the ectopic boundary, signifying the formation of an ectopic margin. Outgrowth organized by this ectopic margin eventually leads to the outgrowth of an ectopic wing from the dorsal surface of the wing blade. An ectopic wing outgrowth phenotype similar to that produced by ap mutant clones is obtained by overexpression of Beadex in the wing disc. When Beadex expression is driven by a patched promoter, an ectopic margin is formed that, as with ap mutant clones, expresses Wingless and Cut in the dorsal compartment of third instar wing discs. These similarities, along with the common LIM domain structure between Beadex and Ap, make it likely that misexpressed Beadex exerts its effect through interference with Ap function (Zeng, 1998 and references).

Beadex has been shown to regulate Apterous activity levels. The classic dominant gain-of-function Beadex mutants reduce Apterous activity. Conversely, loss-of-function Bx mutants appear to increase Ap activity. A LIM-binding protein called Chip has been identified as a possible cofactor for Ap. Bx gain of function mutants overexpress Beadex, which is shown to bind Chip and thereby interfere with formation of a functional complex between Ap and Chip. Ap is shown to induce expression of its antagonist, Beadex, which suggests that Ap and Beadex constitute a feedback mechanism and that the relative levels of Beadex, Chip, and Ap determine Ap activity levels in vivo. Beadex normally is expressed in the wing pouch of the third instar wing imaginal disc during patterning. A mammalian homolog of Beadex is expressed in the developing limb bud of the mouse. This indicates that Lim-only proteins might function in an evolutionarily conserved mechanism involved in patterning the appendages. The dominant interference mechanism of LMO action may serve as a model for the mechanism by which LMO oncogenes cause cancer when misexpressed in T cells (Milan, 1998; Shoresh, 1998; Zeng, 1998).

Selector genes such as ap are often thought of as simple binary switches. The interaction between Chip and Beadex suggest that the situation is more complex and that Ap activity levels are modulated during wing development. The balance in the levels of Ap, Chip, and Beadex proteins is important for determining the level of Ap activity in vivo. This regulation occurs at the level of protein activity, not at the level of gene expression, and it has been suggested that this may reflect a requirement to fine-tune activity levels regionally as the wing develops. Given that Beadex is expressed in regions where Ap is not known to function, it is likely that Beadex may have other functions as well. Chip and Beadex may regulate the activity of LIM-homeodomain proteins in other developmental contexts and in postembryonic homeostatic processes (Milan, 1998).

Drosophila dLMO-PA isoform acts as an early activator of achaete/scute proneural expression

The Drosophila bHLH proneural factors Achaete (Ac) and Scute (Sc) are expressed in clusters of cells (proneural clusters), providing the cells with the potential to develop a neural fate. Mediodorsal proneural patterning is mediated through the GATA transcription factor Pannier (Pnr) that activates ac/sc directly through binding to the dorsocentral (DC) enhancer of ac/sc. Besides, the Gfi transcription factor Senseless (Sens), a target of Ac/Sc, synergizes with ac/sc in the presumptive sensory organ precursors (SOPs). This study investigates, through new genetic tools, the function of dLMO (Beadex), the Drosophila LIM only transcription factor that was already known to control wing development. dLMO gene encodes two isoforms, dLMO-RA and dLMO-RB. dLMO null and dLMO-RA⁻ deletions have similar phenotypes, lacking thoracic and wing margin sensory organs (SO), while dLMO-RB⁻ deletion has normal SOs. At early stages, dLMO-RA is expressed in proneural clusters, however later it is excluded from the SOPs. dLMO functions as a Pnr coactivator to promote ac/sc expression. In the late SOPs, where dLMO-PA is not expressed, Pnr participates to the Sens-dependent regulation of ac/sc. Taken together these results suggest that dLMO-PA is the major isoform that is required for early activation of ac/sc expression (Asmar, 2008).

The lack of dLMO protein leads to very distinctive phenotypes. The mutant animals are not able to fly, they have a short life span and show an abnormal gait behaviour. In addition, they show a discreet bristle phenotype. In Drosophila, there are two paralogous LMO factors, dLMO and CG5708. These genes are expressed in the CNS where redundancy is not excluded. However CG5708 is not expressed in the wing discs and presumptive SOPs. Therefore it is concluded that the mild phenotype observed for the adult PNS in dLMO mutants, is not attributed to gene redundancy. dLMO encodes two distinct isoforms, dLMO-PA and -PB, which only differ from their N-terminus. Only dLMO-RA is broadly expressed in the notum, and contributes to the PNS phenotype. dLMO function is also critical in the developing central nervous system for the activity of the ventral lateral neurons, LN_vs . It is highly probable that dLMO-RB has some subtle biological activities in the brain, where it has a specific pattern (Asmar, 2008).

In vertebrate, multiproteic complexes composed by GATA-1, LMO2, Ldb-1 and the bHLHs E47 and SCL, are required for normal differentiation of haematopoietic cells. The results of this paper highlight several evidences in favour of dLMO as a GATA coactivator in Drosophila. (1) A genetic synergism exists between pnr⁻ and dLMO⁻ null alleles. (2) dLMO modulates the activity of a DC:ac-lacZ reporter, the model target of Pnr, in vivo. Loss of function dLMO mutants show reduced level of the DC:ac-lacZ expression, whereas in gain-of-function dLMO mutants the DC:ac-lacZ expression is increased. (3) dLMO-PA isoform directly interact with Pnr in GST pull down assay. Therefore it is concluded that dLMO might enhance the proneural activity of Pnr through direct interaction with the GATA factor. Consistently, dLMO expression overlaps with the dorsal-most domain of Pnr during third instar larval stages. Though Pnr controls the development of both DC and SC bristles, dLMO null alleles affect only DC bristles. dLMO expression, that overlaps both SC and DC proneural clusters in the notum, is significantly weaker in the SC region, suggesting that regulation of proneural ac/sc expression is differentially sensitive and responds to local combinations of transcription factors. These data support studies demonstrating that the proneural activity of Pnr is prominently repressed in the SC region by the LIM-HD transcription factor Isl (Asmar, 2008).

At later stages, dLMO expression is excluded from the corresponding SOP and its derivative cells. In contrast, the proneural factor Sens, that plays an important role for sensory organ specification, is first broadly expressed in proneural clusters at low levels where it functions as a repressor of ac/sc, and then later, is expressed at high levels in the presumptive SOPs, where it acts as a transcriptional activator that directly interacts and synergizes with the proneural proteins, Ac and Sc. It has been shown recently that both Gfi-1 and GATA-1, the mammalian ortholog of Sens and Pnr respectively, are essential for development of the closed related erythroid and megakaryocytic lineages. This study demonstrated that the Sens/Pnr interaction is evolutionary conserved in Drosophila neurogenesis. It is suggested that Pnr could participate to the Sens-dependent positive autoregulation of Ac/Sc in late SOPs where dLMO is not expressed. The synergism between Pnr and Sens would need more detailed investigations. Taken together, the present studies have shown dLMO-PA as a co-activator for Pnr during the establishment of proneural fields and revealed another level of proneural ac/sc regulation during late neurogenesis in the Drosophila PNS (Asmar, 2008).

GENE STRUCTURE

The structure of the 776-nucleotide-long 3'UTR of Beadex has particular bearing on the analysis of mutants in this gene. It contains four AT_3-5A boxes, seven AT₃ motifs, and several stretches of Ts, the longest of which is T₁₂; this run is collectively referred to as ARE, which in various eukaryotic genes increases destabilization of the transcript. In addition, the 3'UTR of Beadex contains five AGT_3-4A sequences that are closely related to the Drosophila Bearded box (AGCTTTA). Brd boxes also mediate negative post-transcriptional regulation, probably by conferring instability on the transcript (Lai, 1997). Thus, the 3'UTR of Beadex may be involved in negative regulation of Beadex (Shoresh, 1998).

Exons - 5

PROTEIN STRUCTURE

Amino Acids - 266

Structural Domains

The first LIM domain of the conceptual Beadex protein is similar to the first LIM domain of the human LMO1 and LMO2 proteins (79% and 69%) respectively, whereas the second LIM domain of Beadex is similar to the second LIM domain of these human counterparts (94% and 60%) respectively (Boehm, 1990 and Zhu, 1995).

The derived protein sequence of the presumptive oncogene rhombotin is virtually identical when a comparison is made between human and mouse, rendering it difficult to identify functionally important regions or motifs. A sequence that is highly homologous to that of human and mouse rhombotin exists in Drosophila DNA. Comparison of the sequences shows the main conserved feature is a cysteine-rich region (CRR). The mammalian rhombotin gene has tandemly duplicated CRR's (CRR-1 and CRR-2); comparison of CRR-1 and -2 with other known proteins shows close homology to the proposed LIM domains of the nematode cell lineage proteins lin-11 and mec-3, and to a vertebrate transcription factor (Isl-1). The latter three proteins share a homeodomain, in addition to the LIM domains. These observations suggest that the LIM domain might facilitate protein-protein interactions in a manner analogous to the leucine zipper or the helix-loop-helix motifs. Thus, since rhombotin lacks a DNA-binding homeodomain, this protein might belong to a new class of transcriptional regulators that modulate transcription via intermolecular competitive binding to the LIM domains of certain DNA-binding transcription factors (Boehm, 1990).

Members of the human TTG/RBTN family, now renamed ëLMOí for LIM-only proteins, encode proteins with two tandem copies of a LIM motif. There are three members of this family: two have been isolated at the sites of chromosomal translocations in T-cell leukemia. The function of the LIM motifs is at present unknown. The LMO-2 gene is highly conserved between mammals, Drosophila and yeast. As a first step to obtain a model system for studying the function of the LIM motifs the Drosophila homolog Dlmo has been isolated. In contrast to mammals, Drosophila appears to have only one LMO gene. A 2087 bp cDNA clone was isolated from a larval cDNA library, encoding a protein of 266 amino acids. A second transcript with an alternative 5' end was identified in RNA from embryos. The Drosophila LMO protein consists of two tandem copies of the conserved LIM domain characteristic of the human LMO family and an extended amino and carboxy terminus, which is not present in the human proteins. The amino acid sequence similarity with human LMO-1 and LMO-2 in LIM 1 is 79% and 69%, respectively, and in LIM-2, 90% and 60%, respectively. A short stretch of 25 nucleotides with a homology of 83% between LMO-2 and Dlmo is found in the 3' UTR. Dlmo, like LMO-1, has an intron after the second LIM encoding region, which is not present in LMO-2. It is expressed maternally and at a high level in early embryogenesis as well as in adults. The Dlmo protein is found to be immunologically related to LMO-2 and can be detected by immunohistochemistry in early cellular blastoderm embryos. The gene was localized to a genetically well characterized region (17C on the X chromosome) opening the way for identification of mutations (Zhu, 1998).

date revised: 30 September 98

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.