The development of the mature insect trachea requires a complex series of cellular events, including tracheal cell specification, cell migration, tubule branching, and tubule fusion. The Drosophila dysfusion gene encodes a basic helix-loop-helix (bHLH)-PAS protein conserved between Caenorhabditis elegans, insects, and humans; dysfusion controls tracheal fusion events. The Dysfusion protein functions as a heterodimer with the Tango bHLH-PAS protein in vivo to form a putative DNA-binding complex. The dysfusion gene is expressed in a variety of embryonic cell types, including tracheal-fusion, leading-edge, foregut atrium cells, nervous system, hindgut, and anal pad cells. RNAi experiments indicate that dysfusion is required for dorsal branch, lateral trunk, and ganglionic branch fusion but not for fusion of the dorsal trunk. The escargot gene, which is also expressed in fusion cells and is required for tracheal fusion, precedes dysfusion expression. Analysis of escargot mutants indicates a complex pattern of dysfusion regulation, such that dysfusion expression is dependent on escargot in the dorsal and ganglionic branches but not the dorsal trunk. Early in tracheal development, the Trachealess bHLH-PAS protein is present at uniformly high levels in all tracheal cells, but when the levels of Dysfusion rise in wild-type fusion cells, the levels of Trachealess in fusion cells decline. The downregulation of Trachealess is dependent on dysfusion function. These results suggest the possibility that competitive interactions between basic helix-loop-helix-PAS proteins (Dysfusion, Trachealess, and possibly Similar) may be important for the proper development of the trachea (Jiang, 2003).

The insect tracheal system consists of an intricately branched system of tubules that provide oxygen throughout the animal. The formation of the trachea consists of a series of developmental events, and its analysis provides an excellent model system for studying the morphogenesis of other branched structures, such as the vertebrate lung airways, circulatory system, kidney ducts, and excretory epithelia. The trachea is derived from an array of segmentally repeated clusters of precursor cells. After the tracheal precursor cells divide and invaginate, they extend branches, and the branches from neighboring segments fuse to form the mature tracheal tree. The fusion process is mediated by a distinct fusion cell residing on each branch. Branching and fusion are complex cellular processes and pose a number of developmental questions. How are fusion cells specified during tracheal development? What are the short-range and long-range factors that guide tracheal branches to their fusion partners? What is the nature of the adhesive and contact-guidance interactions that mediate fusion and allow the formation of adherens junctions that seal intercellular junctions? How is the cytoskeleton rearranged to allow the tracheal lumen to extend throughout the branch (Jiang, 2003 and references therein)?

The tracheal primordia extend branches in six directions under the guidance of the branchless gene. All of these branches except the visceral branch will fuse with tracheal branches derived from other primordia. The dorsal trunk is formed by fusion of anterior and posterior branches from adjacent segments, as is the lateral trunk, which is formed by fusion of lateral trunk anterior and posterior branches. The dorsal branches travel over the dorsal side of the embryo and fuse along the dorsal midline to their partner from the identical hemisegment. The ganglionic branches migrate ventrally and join at the ventral midline, although only the three anteriormost branches fuse. There is a single fusion cell for each branch, and the fusion cells are characterized by patterns of gene expression distinct from other tracheal cells. The Escargot (Esg) zinc finger transcription factor is prominently expressed during tracheal development in fusion cells and no other tracheal cell. esg mutants show fusion cell defects in the lateral trunk, dorsal branch, and ganglionic branch, but dorsal trunk fusion is relatively normal. Examination of esg mutants indicates that expression of several, but not all, fusion cell-specific genes and markers is absent in the dorsal branch and the ganglionic branch, and there is an excess of branching and gene expression associated with branching. In addition, esg is required for DE-cadherin expression and the ability of fusion cells to form adhesive contacts and adherens junctions. The loss of esg in lateral trunk is the most extreme, resulting in death of the fusion cells. Important issues regarding the tracheal function of esg deal with the identity of genes regulated by esg and why esg is required for tracheal fusion in some branches but not others (Jiang, 2003 and referencest therein).

Three proteins that function prominently in tracheal development are the Trachealess (Trh), Similar (Sima), and Tango (Tgo) basic helix-loop-helix (bHLH)-PAS proteins. Trh and Tgo form a heterodimer that controls transcription and initial formation of the tracheal primordia along with the Drifter (Dfr) POU-homeobox coactivator. Sima and Tgo form a protein dimer that controls the transcriptional response to hypoxia. Since low cellular oxygen conditions induce additional tracheal branching, Sima may autonomously or nonautonomously be required for terminal tracheal branching. Tgo is found in all embryonic cells. In the absence of a bHLH-PAS partner protein, Tgo is found in the cytoplasm, but in the presence of a partner protein, they dimerize, translocate into the nucleus, bind DNA, and activate transcription. Since there may exist multiple bHLH-PAS partners of Tgo in the same cell, it has been proposed that the function and levels of bHLH-PAS proteins may be regulated by competitive interactions. In part for this reason, it is important to identify all Drosophila bHLH-PAS proteins and determine where they are expressed (Jiang, 2003 and references therein).

A novel bHLH-PAS gene, dysfusion, is expressed in all tracheal fusion cells, as well as the epidermal leading edge cells and several other cell types. Tgo accumulates in nuclei of dys-expressing cells, suggesting that it is a partner of Dys in vivo. dys-RNAi experiments reveal tracheal fusion defects in the lateral trunk, dorsal branch, and ganglionic branch but not in the dorsal trunk. The esg gene is expressed in all fusion cells before dys, and esg expression is normal in dys RNAi-injected embryos. However, esg mutant embryos show an absence of dys expression in tracheal fusion cells in most branches, but not the dorsal trunk, further indicating branch-specific function of esg. The appearance of Dys in tracheal fusion cells coincides with a steep drop in Trh levels in fusion cells, and this reduction is dys dependent. This indicates that one function of dys in tracheal fusion cell development is to downregulate Trh protein levels. This provides the first in vivo evidence that bHLH-PAS proteins regulate levels of other bHLH-PAS proteins during development and possibly influence cell fate and morphogenetic decisions (Jiang, 2003).

Drosophila Dys belongs to a novel conserved subfamily of bHLH-PAS proteins that includes C. elegans C15C8.2 and H-NXF. Dys also likely belongs to an extended family of Drosophila bHLH-PAS proteins, including Sim, Sima, Ss, and Trh, that dimerize with Tgo. The protein structure of Dys is conventional for bHLH-PAS proteins. The bHLH domain is near the N terminus of the protein and is followed by the PAS-1 and PAS-2 domains. One unusual feature compared to other bHLH-PAS proteins is the relatively long (152 aa) region N-terminal to the bHLH domain. This region has a large number of glutamine residues and may act as a transcriptional activation domain. The C-terminal residues after the PAS-2 domain are unconserved with C15C8.2, H-NXF, or any other protein but have histidine-rich, proline-rich, and glutamine-rich regions. These residues may also be transcriptional activation domains. The structure of the Dys protein suggests a DNA-binding transcriptional activator, but this needs to be tested biochemically (Jiang, 2003).

Drosophila dys is expressed in a variety of embryonic cell types, including tracheal fusion, leading edge, foregut atrium, brain or stomagastric nervous system, hindgut, and anal pad cells. Initial expression is observed during mid embryogenesis at stage 12. The function of dys in these cell types is unknown, with the exception of the tracheal fusion cells, in which dys plays a developmental role. The dys-RNAi results show that dys is an essential gene. However, dys-dsRNA-injected embryos do not die as embryos, but as second- and third-instar larvae. Thus, the role of dys, as assayed by RNAi, in the various embryonic cell types is not dramatic enough to cause embryonic lethality. Mutations in esg show fusion defects in the same tracheal branches as dys and are also lethal to larvae. The tracheal fusion defects and resulting putative respiratory deficiencies may be the cause of the larval lethality, since other observations have shown that animals with defective tracheae survive until late larval periods. However, both genes are expressed elsewhere, and defects in the other cell types may contribute to lethality (Jiang, 2003).

dys is expressed in tracheal fusion cells, and no other tracheal cells. This was shown by coexpression of dys with esg, a gene that is expressed in fusion cells and regulates tracheal fusion. dys-RNAi experiments were carried out to examine whether dys is involved in tracheal fusion. The results demonstrate that dys is required for fusion of the dorsal branch, lateral trunk, and ganglionic branch but not of the dorsal trunk. This phenotype is similar to the esg mutant phenotype that also affects the dorsal branch, lateral trunk, and ganglionic branch but not the dorsal trunk. Although branches differ in the details of the fusion process, tracheal fusion generally requires migration, recognition, and adhesion of fusion cells. dys-RNAi embryos show relatively normal tracheal branches and migration. The occurrence of a single esg-lacZ cell in each dys-RNAi branch indicates that esg-positive tracheal fusion cells are present, and thus survival and gross cell fate is not controlled by dys. It is possible that dys controls aspects of fusion cell recognition, cell adhesion, or inhibition of nonfusion tracheal functions, such as branching (Jiang, 2003)

Since dys, as well as esg, is expressed in dorsal trunk fusion cells, why is dorsal trunk fusion apparently unaffected in dys-RNAi-injected embryos? It is not likely due to incomplete expressivity of the dys-RNAi, since Dys protein was not detected in dys-RNAi-injected embryos, including dorsal trunk fusion cells, and ~100% of lateral trunk, dorsal branch, and ganglionic branch branches in dys-dsRNA-injected embryos failed to fuse. There are a number of differences between the dorsal trunk and the other branches that could contribute to differences in fusion behavior. The larger diameter dorsal trunk has multiple cells comprising its circumference, unlike most of the other branches, which are thinner and have a single cell comprising the circumference. Dorsal trunk branches are in close proximity to their fusion partner and lack the filopodial extensions that help guide the other branches to their targets. The dorsal trunk also utilizes a mesodermal guidepost cell that mediates fusion. Similar guidepost cells have not been described for the other branches. Finally, breathless RNA levels begin to decline by stage 12 in the dorsal trunk due to spalt repression, which may eliminate the potential need to reduce breathless levels by decreasing Trh levels. These and other possible differences suggest why dys and esg may have different functions in different branches (Jiang, 2003)

The esg gene is required in dorsal branch and ganglionic branch tracheal fusion cells for expression of several genes, including shotgun (DE-cadherin) and three late-expressing fusion cell genes (fusion-4 to fusion-6), as well as repression of terminal branching genes (DSRF and terminal-1). Expression of two early-expressing fusion cell genes (fusion-2 and fusion-3) are not dependent on esg. dys expression is also dependent on esg, in keeping with the role of esg in regulating early-expressing fusion cell gene expression. As with other genes expressed in fusion cells, dys expression is not dependent on esg in dorsal trunk cells. This implies that the ability of esg to activate transcription is fusion cell dependent and is due to the presence of different coregulatory proteins or modifier proteins in the different branches (Jiang, 2003)

The trh gene is required for initiation of tracheal formation. trh expression is maintained throughout embryonic development in most tracheal cells, and this continued expression is due to autoregulation. However, the role of trh beyond its role in initiating tracheal formation is not well understood. Trh protein levels fall specifically in all classes of tracheal fusion cells coincident with the rise in Dys levels. The nuclear levels of Tgo, the partner for both Dys and Trh, remain constant in fusion cells. The biological significance of the reduction in Trh remains to be investigated, but it is possible that fusion requires a reduction in Trh. One possibility is that Trh:Tgo is required for the expression or function of the breathless (btl) tyrosine kinase receptor that guides growing tracheal branches and that btl function must be inhibited in fusion cells. Potentially, the only function of dys is to reduce Trh levels (Jiang, 2003)

There are multiple mechanisms in which Dys could regulate Trh levels. These mechanisms include (1) competition between Dys and Trh for dimerization with Tgo, (2) competitive Dys:Tgo binding to Trh:Tgo autoregulatory binding sites in the trh gene, (3) activation of genes by Dys that encode proteins influencing trh RNA or protein stability, and (4) inhibition of protein kinase B that is required for Trh nuclear transport. Conceptually, the first model is the simplest and most attractive. Trh autoregulates its own expression, and reduction in Trh:Tgo complexes by competition for Tgo by Dys would lead to a reduction in trh RNA and protein. In the second model, Dys:Tgo would function as a transcriptional repressor and extinguish trh RNA synthesis by binding Trh:Tgo autoregulatory sequences within the trh gene (Jiang, 2003)

Evidence for the possible roles of these mechanisms has emerged from studies on vertebrate and Drosophila bHLH-PAS proteins. In one study, it was demonstrated that HIF-1alpha outcompetes the Aryl hydrocarbon receptor (Ahr) bHLH-PAS protein for their common dimerization partner, Arnt, which is the vertebrate Tgo ortholog. In another study, Sim2 was shown to compete with HIF-1alpha for Arnt and partially block expression of a HIF-1alpha:Arnt responsive reporter gene. Sim2 can repress transcription and, by binding to HIF-1alpha:Arnt recognition sites on the reporter gene, Sim2 reduced reporter expression. The third model in which the presence of Dys reduces protein levels by activating the transcription of repressive or inhibitory factors is analogous to how Sim:Tgo represses the expression of genes in the central nervous system midline cells by activating transcription of a repressor. One additional issue is whether Dys reduces the levels of other Drosophila bHLH-PAS proteins in addition to Trh. One possibility is Sim. Both Sim and Dys are expressed in anal pad cells, and sim mutants have anal pad defects (Jiang, 2003 and references therein)

Drosophila has four bHLH-PAS proteins that dimerize with Tgo: Sim, Sima, Ss, and Trh. Dys is also likely to dimerize with Tgo. One particularly interesting observation is that Drosophila Dys, Trh, and Sima are all involved in aspects of tracheal development. Mammals have closely related members of all of these proteins. The mammalian proteins dimerize with either Arnt or the closely related Arnt2. C. elegans has four bHLH-PAS partners for AHA-1, the worm Tgo/Arnt ortholog. These partners include (1) C15C8.2, which is related to Dys and mammalian NXF; (2) AHR-1, which is related to Drosophila Ss and vertebrate Ahr; (3) HIF-1, which is related to Drosophila Sima and vertebrate HIF-1alpha, and (4) T01D3.2, which is related to Sim and Trh. Since orthologs of Dys, Ss, Sima, and Sim/Trh are found in vertebrates, insects, and nematodes, these proteins had already diverged in the common ancestor of these species. The C. elegans C15C8.2 gene is expressed in the pharynx, a feeding organ. The mammalian dys gene, NXF, has been detected only in the brain. The evolutionary conservation of the Dys subfamily of proteins suggests a functional relationship, although the sites of expression in Drosophila and the other organisms, as studied to date, are diverse. One possibility based on the tracheal fusion phenotypes is that Dys regulates aspects of cell recognition or cell adhesion events. It will be interesting to determine in future studies what biochemical, developmental, or physiological features of the Dys proteins are conserved, as well as the evolutionary origins regarding the tracheal functions of Dys, Sima, and Trh (Jiang, 2003)

GENE STRUCTURE

cDNA clone length - 3323

Bases in 5' UTR - 630

Exons - 9

Bases in 3' UTR - 47

PROTEIN STRUCTURE

Amino Acids - 681

Structural Domains

Bioinformatic searches for additional Drosophila bHLH-PAS genes utilized bHLH and PAS sequences screened against the genomic sequence generated by the Celera/Berkeley Drosophila Genome Project. When a human HIF-1alpha PAS-1 region was screened against Drosophila genomic DNA by using the tBLASTN algorithm, a novel bHLH-PAS gene was identified in the BAC clones BACR13F13 and BACR08G22, which map to a cytological position, 96F-97A, on chromosome 3. Later, the gene was also identified by the BDGP-Celera Consortium by using the Genie algorithm and in global searches for bHLH genes. Genie initially predicted four distinct genes in the region, three of which contain multiple predicted exons: CG14554 (one exon), CG12561 (three exons), CG14553 (three exons), and CG14552 (four exons). Using RT-PCR and DNA sequence analysis, these four predicted genes were all shown to constitute a single gene, dys, now referred to as CG32474. dys maps to 96F9-10 according to FlyBase (Jiang, 2003)

Preliminary experiments have indicated that dys is a rare embryonic mRNA, an observation consistent with its absence from the BDGP embryonic expressed sequence tag (EST) collection of 8,809 distinct mRNAs. RNA was isolated from Drosophila embryos, and RT-PCR was carried out with primers designed from genomic sequences. The 5' and 3' ends of the mRNA were identified by RACE. These experiments provide the sequence of the dys mRNA. The genomic sequence comprises 10 exons and spans 21,697 bp. The 5'-untranslated region (UTR) determined by RACE is predicted to be 521 nucleotides (nt), and the 3'-UTR is 500 nt. Of particular interest, the 3'-UTR of dys, which resides in exon 10, overlaps by 475 bp the sequence of EST LP05454, which is transcribed in the other direction (Jiang, 2003)

The coding sequence of dys shows a clear relationship to other bHLH-PAS proteins. It has a bHLH domain near the N terminus, followed by PAS-1 and PAS-2 domains. The C-terminal regions of bHLH-PAS proteins generally contain transcriptional activation domains, although the protein sequences are poorly conserved. Consistent with that observation, the C-terminal region of Dys shows little homology with other proteins but contains histidine-rich and proline-rich regions similar to the C terminus of other bHLH-PAS proteins. The region N terminal to the bHLH domain encoded by exons 1 and 2 is 152 aa, which is uncharacteristically long for bHLH-PAS proteins. This region contains a large number of glutamine residues, suggesting that it may act as an N-terminal transcriptional activation domain. There is also an unusually long spacer region in the PAS-1 domain that contains a large number of serines and repeats of Gly-Gly-Ala. The length of the predicted protein is 918 aa with a predicted molecular mass of 102 kDa (Jiang, 2003)

Sequence comparisons of all existing bHLH-PAS proteins to Dys reveal that it is a member of a discrete subfamily of bHLH-PAS proteins. The complete sequences of representative bHLH-PAS proteins were compared by CLUSTALX and displayed by a phylogenetic tree generated by the neighbor-joining method algorithm. C. elegans C15C8.2, Drosophila Dys (D-Dys), A. gambiae Dys (An-Dys), and H-NXF cluster as a discrete branch of the phylogenetic tree consisting of representative bHLH-PAS proteins. This result indicates that a dys ancestral gene existed before the divergence of nematodes, insects, and vertebrates. Quantitatively, the bHLH region of D-Dys shows the following percent identities to the other Dys subfamily proteins: An-Dys (96%), H-NXF (58%), and C15C8.2 (57%). The percent identity was only 30% within the bHLH region to the next closest bHLH-PAS protein, Drosophila Spineless (Ss). Similar relationships were observed by comparing PAS-1 domains and PAS-2 domains, although the percent identity was less than for the bHLH region, as generally observed with bHLH-PAS protein comparisons. The PAS-1 domain of Dys is characterized by a long insertion containing a number of Gly-Gly-Ala repeats. This insertion is absent in C15C8.2 and H-NXF, but An-Dys has an even larger insertion with a different sequence content consisting of stretches of His, Ser, and Gly. In contrast, the PAS-2 domain is similar in length and colinear between the subfamily members. Sequences N-terminal to the bHLH domain, the spacer between bHLH and PAS-1 domains, and the regions C-terminal to the PAS-2 domain are poorly conserved among subfamily members, whereas the spacer between PAS-1 and PAS-2 shows significant sequence identity (Jiang, 2003)

date revised: 25 October 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.