Ubiquitin-mediated proteolysis regulates the steady-state abundance of proteins and controls cellular homoeostasis by abrupt elimination of key effector proteins. A multienzyme system targets proteins for destruction through the covalent attachment of a multiubiquitin chain. The specificity and timing of protein ubiquitination is controlled by ubiquitin ligases, such as the Skp1-Cullin-F box protein complex. Cullins are major components of SCF complexes, and have been implicated in degradation of key regulatory molecules including Cyclin E, beta-catenin and Cubitus interruptus. The Drosophila Cullin-3 homologue, Guftagu, has been genetically identified and molecularly characterized. Perturbation of Cullin-3 function has pleiotropic effects during development, including defects in external sensory organ development, pattern formation and cell growth and survival. Loss or overexpression of Cullin-3 causes an increase or decrease, respectively, in external sensory organ formation, implicating Cullin-3 function in regulating the commitment of cells to the neural fate. Cullin-3 function modulates Hedgehog signalling by regulating the stability of full-length Cubitus interruptus (Ci155). Loss of Cullin-3 function in eye discs but not other imaginal discs promotes cell-autonomous accumulation of Ci155. Conversely, overexpression of Cullin-3 results in a cell-autonomous stabilisation of Ci155 in wing, haltere and leg (but not eye), imaginal discs suggesting tissue-specific regulation of Cullin-3 function. The diverse nature of Cullin-3 phenotypes highlights the importance of targeted proteolysis during Drosophila development (Mistry, 2004).

Targeted degradation of short-lived proteins is a universal process that regulates diverse cellular functions. A key example of the importance of targeted protein degradation comes from the observation that the rapid and timely destruction of Cyclin proteins regulates cell cycle progression. Cyclins and other proteins are marked for destruction by the action of a ubiquitin ligase whose specific targeting activity mediates the covalent attachment of a ubiquitin polymer to select residues of the target protein. Four types of ubiquitin ligases have been identified: (1) the N-end rule/Ubr1 ligase, (2) the HECT-domain family, (3) the Cyclosome/Anaphase Promoting Complex and the (4) Skp1-Cullin-F box/Elongin C-Cullin-SOCS box (SCF/ECS) complex (Deshaies, 1999; Hershko, 1998; see Drosophila Slimb). Of these, the SCF/ECS complex was the first identified and is the best understood (Mistry, 2004).

SCF and ECS complexes (Deshaies, 1999 and Tyers, 1999) target distinct groups of proteins for degradation. Cullins are conserved proteins of ~800 residues that comprise the scaffold of both SCF and ECS ubiquitin ligases. Cullins interact with Skp1 or Elongin C homologues through their N terminus and with Hrt1/Roc1/Rbx1, a RING-finger-containing protein through their C-terminus. Skp1-like or Elongin C-like proteins interact with F-box or SOCS-box containing proteins, which target specific proteins for ubiquitination by their respective ubiquitin ligase complex. Recently, BTB domain-containing proteins have been shown to interact with Cullin-3 and subsume the role of the Skp1 and F-box proteins in substrate recognition (Geyer, 2003;; Pintard, 2003; Xu, 2003), suggesting the existence of yet another ubiquitin ligase complex, in addition to the four listed above, with a distinct repertoire of protein targets (Geyer, 2003; Pintard, 2003; Xu, 2003). The modular nature of multisubunit ubiquitin ligases endows different complexes with distinct substrate specificity. Targeted protein ubiquitination plays a critical role in mediating the response of multiple developmental signalling pathways. In Drosophila, SCF complexes have been implicated in ubiquitination of protein targets in three signalling pathways; IκBα in the NF-κB pathway; β-catenin/Armadillo in the Wnt pathway and Ci in the Hedgehog pathway. Signal activation in the Hedgehog (Hh) pathway, for example, leads to the tissue-dependent expression of the Wingless (Wg) and Decapentaplegic (Dpp) morphogens (Mistry, 2004).

The Drosophila Cullin-3 homologue (dCul-3) has been identifed and characterized. dCul-3 plays a broad role to regulate the development of many different adult structures. For example, loss of function and over-expression studies indicate that dCul-3 inhibits sensory organ development during adult development consistent with the idea that dCul-3 modulates Notch pathway activity. In addition, genetic studies confirm and extend the relationship between dCul-3 function and the regulation of Ci155 levels and thus, Hedgehog signalling. These results suggest that dCul-3 function impinges on the activity of many different signalling pathways and developmental events via the targeted destruction or modification of specific proteins (Mistry, 2004).

guftagu was identified in a screen for dominant modifiers of an adult viable wing and notal phenotype caused by GAL4::UAS-mediated misexpression of constitutively activated Gα_s (Gα_s*). A collection of overlapping autosomal deficiencies corresponding to ~70% of the autosomal genome was screened to identify genomic regions capable of dominant modification of the Gα_s* phenotype. This screen uncovered 28 genomic regions likely to contain dominant modifiers of the Gα_s* phenotype. Heterozygosity for each of these 28 regions suppresses the Gα_s* phenotype although the extent of suppression varies between regions. To identify the modifying loci, smaller deletions and mutations in individual loci were screened within each genomic region for the ability to modify the Gα_s* phenotype. Seventeen single loci in 13 genomic regions capable of dominant modification of the Gα_s* phenotype were identified. Single modifying loci in the remaining 15 genomic regions were not identified (Mistry, 2004).

Two of the 28 deficiencies, Df(2L)osp29 (35B1; 35E6), and Df(3L)vin7 (69A1-69A5), suppress the Gα_s* phenotype to near wild-type levels. A systematic search of the complementation groups uncovered by Df(2L)osp29 identified l(2)35Cd at polytene position 35C4 in this region as a suppressor of the Gα_s* phenotype. The gene was designated guftagu (gft), an Urdu word that means ‘private conversation’, because it is believed that the gene product has a private conversation with the activated Gα_s signalling pathway in order to modify the Gα_s* phenotype (Mistry, 2004).

Thus dCul-3 was initially detected as a dominant suppressor of Gα_s signal transduction, indicating the ability of dCul-3 to modulate the activity of at least one signalling pathway. Subsequent studies determined that dCul-3 plays a broad role to regulate the development of many different structures during adult development consistent with the idea that dCul-3 activity modulates the strength of multiple signalling pathways during Drosophila development (Mistry, 2004).

Phosphorylation of Ci triggers its subsequent proteolysis. One mechanism that might couple phosphorylation with proteolysis is the ubiquitin-mediated degradation pathway regulated by ubiquitin ligases such as the SCF complex. The F-box-containing factor Slimb is required for the generation of Ci75, the repressor form of Ci. Using the developing eye disc as a model, Ou (2002) has shown that Ci155 stability is controlled differentially by dCul-1 and dCul-3. Slimb and dCul-1 function anterior to the morphogenetic furrow to target Ci155 for proteolysis, while dCul-3 functions posterior to the furrow to mediate the same event (Mistry, 2004).

The current study supports the link between Cullin/SCF function and Ci155 stability during imaginal disc development. Loss of dCul-3 function in posterior compartment cells of the wing disc immediately adjacent to the AP boundary results in a non-autonomous reduction in Ci155 accumulation in anterior compartment cells that abut dCul-3 mutant cells. Furthermore, overexpression of dCul-3 in the anterior but not posterior compartment of wing, haltere and leg imaginal discs leads to a cell-autonomous increase in Ci155 stability. Thus, dCul-1 and dCul-3 are required in distinct developmental contexts to regulate Ci155 stability and Hh signal transduction (Mistry, 2004).

Together with the results of Ou (2002) , these data support the model that dCul-3 functions autonomously to regulate Ci155 stability in a region-specific manner. In the eye, dCul-3 likely acts in a complex to promote the cleavage of Ci155 into the Ci75 repressor form. dCul-3 could mediate this activity directly, by associating with a specific F-box protein that tethers Ci155 to an SCF complex containing dCul-3. Alternatively, dCul-3 could mediate this effect indirectly, by targeted degradation or modification of a protein involved in the regulation of Ci155 stability. In the wing, dCul-3 overexpression could lead to an autonomous accumulation of Ci155 either by titrating other SCF complex components that promote the limited proteolysis of Ci155 to Ci75 or by targeting a protein for degradation that is normally required to promote limited proteolysis of Ci155. At present these data do not distinguish clearly between these models, although the reciprocal phenotypes observed in dCul-3 mutant clones relative to tissues that overexpress dCul-3 suggest dCul-3 does not act solely in a dominant negative manner (Mistry, 2004).

The non-autonomous effect of dCul-3 loss on Ci155 stability suggests that dCul-3 can modulate Ci155 accumulation through multiple mechanisms. In this context, the simplest model is that dCul-3 function is required for the proper expression or transmission of the Hh signal. The apparent ability of dCul-3 to regulate Ci155 stability through at least two different mechanisms and the diversity of dCul-3 phenotypes, suggest that the composition of dCul-3-containing SCF complexes varies in a region- and stage-specific manner. Given this, a clear understanding of the molecular basis through which dCul-3 regulates Ci155 stability as well as the activity/levels of other proteins will require the identification of the direct targets of dCul-3/SCF complexes through biochemical and molecular genetic means (Mistry, 2004).

In addition to its effects on the Hh and Gα_s pathways, loss of dCul3 function results in embryonic and adult phenotypes reminiscent of defects in EGF and Notch signalling. For example, loss of maternal dCul-3 function produces small embryos with fused dorsal appendages, a phenotype similar to that generated by reduction or loss of function in members of the EGF receptor pathway. Loss of dCul-3 function also causes ectopic sensory organ formation and shaft duplication similar to phenotypes that arise as a result of compromised Notch activity. Likewise, the venation phenotypes dCul3 clones are similar to those that arise due to perturbations in either the Notch or EGF-receptor signalling pathways. These results raise the possibility that dCul-3 function also modulates the activity of the Notch and/or EGF-receptor signalling pathways (Mistry, 2004).

Structure function analyses of Cullins identify three distinct domains: an N-terminal domain that binds Skp- or Elongin C-like proteins; a C-terminal domain that binds Hrt1/Roc1/Rbx1 and a ~60 amino acid domain at the extreme C-terminus of unknown function. The latter domain exhibits the highest degree of sequence identity (~50%) between different Cullins and contains the invariant tyrosine-rich motif, Y-X₂-R-X_6-7-Y/F-X-Y-X-A/S, known as the Cullin motif. However, the functional significance of this domain remains unclear. Although dispensable for Cullin function in yeast, the C-terminal domain is modified post-translationally by the covalent attachment of Nedd8, a small, ubiquitin-like protein. An intact neddylation pathway is required for Cullin-dependent ubiquitination of HIFα, p27 and IκBα in mammalian cells and mutations in Cul-1 and nedd8 both lead to heightened accumulation of Ci155 in Drosophila eye discs (Ou, 2002). Therefore, post-translational modification of the extreme C-terminus of Cullins by Nedd8 might be required for SCF-mediated ubiquitination of target proteins (Mistry, 2004 and references therein).

Genetic studies of dCul-3 provide evidence for the in vivo importance of the C-terminal domain. The gft^HG39 lesion encodes a protein with a truncated C terminal domain that is phenotypically one of the most severe dCul-3 alleles, demonstrating the in vivo relevance of the extreme C-terminal domain. Consistent with this, overexpression of a full-length form of dCul-3 is found to be sufficient to induce phenotypes reciprocal to those observed in dCul-3 mutant clones whereas overexpression of a truncated form of dCul-3 that lacks the C-terminal domain yields no overt phenotypes. These data suggest that the dCul-3 C-terminal domain is necessary for dCul-3 activity during Drosophila development. This is the first clear demonstration of the in vivo importance of the C-terminal domain. Future experiments are required to elucidate the precise molecular mechanisms through which the C-terminal domain enables Cullin protein function (Mistry, 2004).

These results demonstrate that the mechanisms by which multisubunit ubiquitin ligases target specific proteins for degradation are complex. For example, dCul-1 and dCul-3 both promote Ci155 proteolysis -- but they do so in distinct spatial domains (Ou, 2002). In addition, complexes containing dCul-1 and dCul-3 can discriminate between distinct forms of the same protein; SCF complexes containing Cul-1 target the phosphorylated form of Cdk2-bound Cyclin E while SCF complexes containing Cul-3 target free unphosphorylated Cyclin E for degradation (Singer, 1999). The specificity of Cul-1 and Cul-3 to target the same protein in different domains and in different conditions, hints at the precision and complexity of targeted protein degradation. Recent work in C. elegans adds another layer of complexity to this equation, as BTB domain-containing proteins, such as MEL-26, appear to identify new adaptor molecules that link Cul-3 with specific target proteins, such as MEI-1, for ubiquitination (Pintard, 2003; Xu, 2003). These data together with the large number of SCF-like components in higher eukaryotes, underline the enormous combinatorial potential for the formation of distinct multisubunit ubiquitin ligases and their ability to target distinct, but potentially overlapping, sets of proteins for degradation (Mistry, 2004).

GENE STRUCTURE

cDNA clone length - 3012 (isoform A)

Bases in 5' UTR - 343

Exons - 11 (isoform A)

Bases in 3' UTR - 347 (isoform A)

PROTEIN STRUCTURE

Amino Acids - 773

Structural Domains

Sequencing and conceptual translation of the longest guftagu cDNA, LD03316, predicts a 773 amino-acid protein that displays at least 25% sequence identity over its entire length and ~50% identity over the C-terminal 60 amino acids to all Cullin family members. Within the Cullin family, gft is most similar to the Cullin-3 class of proteins. Gft shares 69% amino acid identity over its entire length and 80% amino acid identity over the C-terminal domain with human Cul-3. Thus, gft appears to identify the Drosophila homologue of Cullin-3 (dCul-3) (Mistry, 2004).

date revised: 20 February 2005

Home page: The Interactive Fly © 1995, 1996 Thomas B. Brody, Ph.D (Mistry, 2004).

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