In mammalian systems, the heterodimeric basic helix-loop-helix (bHLH)-PAS transcription hypoxia-inducible factor (HIF) has emerged as the key regulator of responses to decreased oxygen concentrations (reviewed by Bruick, 2001: Semenza, 2001; see also Kaelin, 2002; Bruick, 2003; Bracken, 2003). A homologous system is present in Drosophila, and its activity has been characterized in vivo during development. By using transcriptional reporters in developing transgenic flies, it has been shown that hypoxia-inducible activity rises to a peak in late embryogenesis and is most pronounced in tracheal cells. The bHLH-PAS proteins Similar (Sima) and Tango (Tgo) function as HIF-alpha and HIF-ß homologs, respectively; a conserved mode of regulation for Sima by oxygen has been demonstrated. Sima protein, but not its mRNA, is upregulated in hypoxia. Time course experiments following pulsed ectopic expression demonstrate that Sima is stabilized in hypoxia and that degradation relies on a central domain encompassing amino acids 692 to 863. Continuous ectopic expression overrode Sima degradation, which remains cytoplasmic in normoxia, and translocates to the nucleus only in hypoxia, revealing a second oxygen-regulated activation step. Abrogation of the Drosophila Egl-9 prolyl hydroxylase homolog, CG1114, causes both stabilization and nuclear localization of Sima, indicating a central involvement in both processes. Tight conservation of the HIF/prolyl hydroxylase system in Drosophila provides a new focus for understanding oxygen homeostasis in intact multicellular organisms (Lavista-Llanos, 2002).

In multicellular organisms, oxygen homeostasis requires precise developmental coordination between the growth of metabolizing tissues and that of systems that supply oxygen. Recent advances have provided new insights into how this complex task is achieved. For instance, in mammals it has long been recognized that local hypoxia is a major stimulus for angiogenesis. More recently, the recognition that specific angiogenic growth factors (such as the vascular endothelial growth factor) are powerfully induced by hypoxia through the action of a DNA binding complex termed hypoxia-inducible factor (HIF) has provided mechanistic insights into the process (Lavista-Llanos, 2002 and references therein).

The HIF DNA-binding complex consists of a heterodimer of basic-helix-loop-helix-PAS (bHLH-PAS) proteins that binds a core element A/(G)CGTG within hypoxia response elements (HREs) (Wang, 1995a and b). Regulation by oxygen involves stabilization of the alpha-subunit in hypoxia, whereas the ß-subunit, a common partner for several other bHLH-PAS proteins, is constitutively expressed regardless of oxygen tension. Normoxic degradation of HIF-alpha is mediated via ubiquitination and subsequent proteolysis, which requires oxygen-dependent interaction with the Von Hippel-Lindau (VHL) tumor suppressor protein (Maxwell, 1999). This interaction is regulated by hydroxylation of specific prolyl residues within the HIF-alpha polypeptides (Ivan, 2001; Jaakkola, 2001, Masson, 2001; reviewed by Masson, 2003), and recent work has identified a series of {alpha}-ketoglutarate-dependent non-heme iron-dependent dioxygenases that catalyze HIF-alpha prolyl hydroxylation and thus regulate stability of the polypeptide in accordance with oxygen availability (Bruick, 2001; Epstein, 2001). Other studies of HIF induction in mammalian tissue culture systems have defined additional regulatory steps that involve oxygen-dependent subcellular localization (Berra, 2001; Kallio, 1998) and coactivator recruitment (Carrero, 2000; Ema, 1999), although these processes are so far less well understood (Lavista-Llanos, 2002 and references therein).

In addition to angiogenic growth factors, HIF-1 drives the expression of genes involved in a broad array of systemic and cellular adaptive responses to hypoxia, suggesting a central role for HIF-1 as a regulator of oxygen homeostasis. However, although targeted inactivation of different HIF-alpha and HIF-ß subunits in the mouse is associated with several severe or lethal phenotypes involving defective vascular development (Carmeliet, 1998; Iyer, 1998; Kozak, 1997; Maltepe, 1997; Ryan, 1998), few mechanistic studies of the HIF-1 system have been conducted in vivo, and the critical interfaces between developmental processes and HIF activation remain largely undefined (Lavista-Llanos, 2002 and references therein).

To better understand HIF regulation in vivo and the role of hypoxia in developmental processes, the system has been characterized in Drosophila. In insects, air reaches the tissues by passive diffusion through a specialized tubular network termed the tracheal system. Moreover, development of tracheal terminal branches is oxygen dependent and shares many features with mammalian oxygen-dependent angiogenesis. For instance, oxygen-regulated expression of Drosophila Branchless/FGF guides tracheal migration during development and also drives extension of plastic terminal branches in a manner similar to the function of vascular endothelial growth factor in mammalian angiogenesis. The existence of a Drosophila HIF homolog has been inferred from DNA-binding assays with nuclear extracts from normoxic or hypoxic SL2 cells (Nagao, 1996), and transfection studies in mammalian cells have suggested that the Drosophila bHLH-PAS protein Similar (Sima) (Nambu, 1996) might function as an HIF-alpha homolog (Bacon, 1998). However, the system has not yet been fully defined or characterized in vivo in the fly (Lavista-Llanos, 2002).

Transgenic flies have been used to demonstrate and characterize in vivo the operation of a hypoxia inducible transcription response homologous to mammalian HIF. The work confirms the candidacy of Sima and Tgo as the Drosophila homologs of mammalian HIF-alpha and HIF-ß, respectively, defines a conserved multistep mode of regulation for Sima and provides new insights into the mechanisms regulating HIF proteins, as well as into the spatial and temporal operation of the hypoxia-responsive system during Drosophila development (Lavista-Llanos, 2002).

By tracking reporter gene activation in developing flies, the oxygen concentration dependence, developmental regulation, and spatial distribution of the transcriptional response were studied. Serial studies of the hypoxia response during development indicate that induction by hypoxia is modest in early embryogenesis and mid-embryogenesis and then rises sharply to peak levels at the end of embryogenesis, thereafter remaining relatively high throughout the larval stages. This developmentally restricted capacity fits well with the adaptive requirements of Drosophila larvae. After eclosion larvae usually dig into the substrate, while feeding actively, and are probably subjected to major variations in environmental oxygen tension so that enhanced activity of the HIF system is likely to be of critical importance at this stage (Lavista-Llanos, 2002).

Interestingly and somewhat unexpectedly, analysis of reporter expression patterns in developing flies shows enhanced hypoxia-inducible activity in the cells of the tracheal system. Although experiments using severe hypoxia and genetic inactivation of Sima proteolysis demonstrate a widespread potential for transcriptional activation by this system, exposure to more moderate hypoxia clearly demonstrates enhanced activity in tracheal cells. This was reflected both in higher expression levels of Sima and in higher activity of different HRE-linked reporter genes and, moreover, was shown to be a cell autonomous function that was preserved in cells of tracheal fate even in the face of mutations that disrupt tracheal architecture (Lavista-Llanos, 2002).

The existence of enhanced responses to hypoxia in cells composing the organ of oxygen delivery is clearly of interest and raises questions as to its function, particularly since current models indicate that the regulation of tracheal development by oxygen is guided by signals arising in the metabolizing tissues outside the tracheae. Interestingly, some of the branches of the tracheal system run alongside the Drosophila nervous system, and one possibility is that the tracheae function as sensory organs for hypoxia, as does the carotid body in mammals. A hypoxia pathway affecting behavioral responses has been described in flies, and it will be interesting to determine whether hypoxia-induced behavioral responses share a regulatory mechanism with the HIF system (Lavista-Llanos, 2002 and references therein).

In the current work, the hypoxia response element (HRE) transgenic reporter system (binding of the Sima/Tango heterodimer to a cognate DNA binding site that triggers a reporter system) was used to define upstream control mechanisms operating on the Drosophila HIF system. These studies identify Sima as the regulatory Drosophila HIF subunit and demonstrate a major mode of regulation through oxygen-dependent proteolysis that involves a central oxygen-dependent degradation domain (ODDD). Interestingly, both of the sites of prolyl hydroxylation that operate in mammalian HIF-alpha subunit ODDD (Masson, 2001) appear to be conserved in Sima. Furthermore, genetic ablation of the Drosophila HIF prolyl hydroxylase homolog CG1114 results in striking upregulation of both Sima and reporter gene activity in vivo. This strongly supports a conserved mode of proteolytic regulation of Sima following prolyl hydroxylation at one or both of these sites (Lavista-Llanos, 2002).

In contrast with the mammalian system, where HIF prolyl hydroxylase activity is represented by the three PHD isoforms (Bruick, 2001, Epstein, 2001), survey of the Drosophila genome revealed only one homolog (Taylor, 2001), raising questions about the potential of this activity to regulate precisely tuned physiological responses. Interestingly, however, the CG1114 gene is itself a Sima target, demonstrating the operation of a conserved feedback control with the potential to contribute to the complex demands of physiological oxygen homeostasis (Lavista-Llanos, 2002).

Studies of Sima regulation also demonstrate an additional regulatory step. Transgenic overexpression of Sima in normoxic embryos resulted in cytoplasmic accumulation of the protein and little transcriptional activity. In contrast, similar levels of overexpression in hypoxia resulted in nuclear accumulation and a strong transcriptional response, demonstrating the presence of a second oxygen-regulated mechanism controlling Sima subcellular localization. An oxygen-regulated nuclear localization step has previously been demonstrated for mammalian HIF-alpha (Berra, 2001, Kallio, 1998, Luo, 2001), although not in every study. However, demonstration of conservation of this mode of regulation in Drosophila Sima provides strong support for the physiological relevance of this process. These findings suggest that Sima subcellular localization is controlled by an active mechanism that maintains the protein in the cytoplasm in normoxia as opposed to an hypoxia-dependent machinery that mediates nuclear import. Although the strong transcriptional activity of mammalian HIF-alpha that is observed after deletion of the ODDD (Elson, 2001, Huang, 1998), mutation of the VHL binding sites (Masson, 2001), or inactivation of VHL (Maxwell, 1999) is consistent with a role for this domain in cytoplasmic localization in normoxia, this has not been tested in studies of mammalian HIF-alpha that have examined subcellular localization directly. Moreover, although induction of nuclear localization by iron chelators and cobaltous ions (Kallio, 1998) suggests a similar mode of regulation to proteolytic regulation, neither the source of the oxygen-sensitive signal nor the mechanism of transduction have been defined. In vivo studies in flies show induction of nuclear Sima after inactivation of CG1114 either by RNAi or by mutation, thus clearly implicating this gene product in the process of cytoplasmic localization in normoxia. Moreover, Sima nuclear localization was also observed in flies bearing the Delta692-863 transgene, indicating that this sequence is absolutely required for cytoplasmic localization (Lavista-Llanos, 2002).

Very recently nonproteolytic regulation of mammalian HIF-alpha subunits involving the C-terminal transactivation domains has been shown to be regulated by hydroxylation of a specific asparaginyl residue by an enzymatic activity that, like the prolyl hydroxylases involved in HIF proteolysis, demonstrates the properties of an alpha-ketoglutarate-dependent dioxygenase (Lando, 2002b). Thus, regulatory hydroxylation of HIF-alpha residues by this class of enzyme appears to extend to both specific asparaginyl and prolyl residues. Currently, the precise substrate requirements of the CG1114 gene product are not defined, and it is not clear whether effects on nuclear localization are mediated through the conserved prolyl residues, possibly reflecting additional functions of the VHL ubiquitylation complex, or whether other sequences within the Sima ODDD mediate this process. Further biochemical and genetic studies should clarify these new insights into the HIF system (Lavista-Llanos, 2002).

Overall, the high degree of conservation in the Drosophila system indicates that genetic studies in this organism should be highly informative in analyses of both the upstream pathways regulating the HIF system, and the downstream physiological effects in an intact organism (Lavista-Llanos, 2002).

GENE STRUCTURE

cDNA clone length - 5763

Bases in 5' UTR - 586

Exons - 9

Bases in 3' UTR - 633

PROTEIN STRUCTURE

Amino Acids - 1505

Structural Domains

The Drosophila similar (sima) gene was isolated using a low-stringency hybridization screen employing a single-minded gene basic helix-loop-helix (bHLH) DNA probe. sima is a member of the bHLH-PAS gene family and the conceptual protein shares a number of structural features, including a bHLH domain, PAS domain, and homopolymeric amino acid stretches. Sima is most closely related to the human hypoxia-inducible factor 1 alpha bHLH-PAS protein. In situ hybridization experiments reveal that sima is transcribed in most or all cells throughout embryogenesis. It has been cytologically mapped to position 99D on the third chromosome, and is not closely linked to other known bHLH-PAS genes (Nambu, 1996).

The Sima bHLH domain has 55% aa identity to Sim. It has even higher identity (63%) to HIF-lalpha, and is closely related to Trh (59%). Within the bHLH region, the highest homology is within the basic region and the loop-helix 2 junction. The presence of identical basic region residues DNA binding residues suggests that all four bHLH-PAS proteins, Trh, Sim, Sima and HIF-lalpha bind the same core half-site sequence. The loop is the least conserved region in sequence although its length is the same in all four proteins. The Sima bHLH sequence clearly places Sima within the Sim-related gene family since it shows significantly lower identity to other bHLH proteins: Ahr (28%), Arnt (28%), MyoD (20%) and Myc (15%) (Nambu, 1996).

The complete coding sequence of sima was determined by analyzing corresponding cDNA clones. The inserts of seven embryonic cDNA clones ranged in size from 3 to 5.8 kb, and sequence and restriction map analysis shows they completely overlap. These results do not provide evidence for multiple embryonic mRNA species. This is consistent with Northern blot analysis of embryonic RNA with a 32P-labeled sima cDNA probe that detected a single 6.1 kb mRNA species. The longest clone, 5763 nt in length, contains a predicted ORF of 1505 aa, considerably longer than Sim (673 aa), HIF-lalpha (826 aa), and Trh (929 aa). In vitro transcription and translation of a full length sima cDNA clone resulted in synthesis of a polypeptide of approximately 150 kDa, close to the expected size of Sima (Nambu, 1996).

The Sima coding sequence shares structural features with bHLH-PAS proteins including a bHLH domain, a PAS domain with the 41-44 aa PAS repeats, and C-terminal Glu-rich regions. Arrangement and spacing of the domains is also conserved between Sima and other members of the family. Both of the PAS repeats have the invariant PAS repeat residues in the 44-aa consensus repeat sequence: Phe¹, His⁴¹, and ASP⁴⁴. Several PAS domain residues have been shown by mutational analysis to affect Per PAS domain dimerization: (1) Val²⁰⁵ of Per (the per^l mutation), (2) Gly³⁷⁵-Tyr³⁷⁶, and (3) Pro3⁷⁸-Leu³⁸¹. These residues are also conserved in the Sima PAS domain (aa 185, 343-344, 346-349, respectively). The C-terminal half of Sima contains numerous long stretches of poly[Glu], His-rich, Pro-rich, Ser-rich, and acidic regions that may function as transcriptional activation domains (Nambu, 1996).

PAS proteins can be partitioned into five groups, all with a single homologous member except the Sim-related group, which is comprised of Sim, Sima, HIF-la, and Trh. Statistical considerations indicate that Sima and HIF-lalpha are significantly more related to each other than to Sim or Trh. However, analysis of the incomplete human Sim and Drosophila Sim sequences indicate they have considerably higher conservation than human HIF-lalpha and Drosophila Sima. Thus, sequence analysis is unclear regarding whether Sima and HIF-alpha are homologous genes (Nambu, 1996).

date revised: 12 June 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.