BIOLOGICAL OVERVIEW

Dodo is an enzyme that facilitates the degradation of the transcription factor CF2, which regulates expression of the rhomboid gene in follicle cells. This chain of events is required to establish the dorsal/ventral polarity of the developing oocyte. The epidermal growth factor receptor (Egfr) signal transduction system that functions in follicle cells, specifies the dorsal follicle cell fates in the Drosophila egg chamber (Hsu, 2001).

Viewed from a wider perspective, the developing egg chamber of Drosophila is seen to be composed of germline nurse cells and oocyte surrounded by a layer of somatic follicle cells. Inductive interactions between these cells establish anterior-posterior and dorsal-ventral polarities in the egg and the developing embryo. The dorsal-ventral patterning process is initiated by the transmission of a transforming growth factor-like ligand, Gurken (Grk), from the oocyte to the anterodorsal follicle cells, thus activating the Egfr signaling pathway. The immediate outcome of activated Egfr signaling is the degradation of transcription factor CF2 in the dorsal follicle cells, which is where the function of dodo comes in. Degradation of CF2 is required for the expression of rhomboid gene, which in turn initiates a second wave of signaling by activating a second ligand, Spitz, ultimately resulting in the subdivision of different dorsal follicle cell fates (Hsu, 2001 and references therein).

A single MAPK phosphorylation site in CF2 is required for the degradation of CF2. It is speculated that the conformational change induced by proline isomerization after MAPK-mediated phosphorylation might be the structural signature required for CF2 to be recognized by the proteolytic machinery. With this hypothesis in mind, the function of Drosophila Pin1, the dodo gene product, was examined in the development of the egg chamber. Dodo is a peptidylprolyl isomerase involved in facilitating the degradation of CF2, probably by isomerizing a prolyl peptide bond after MAPK-catalysed phosphorylation of the protein (Hsu, 2001).

Dodo serves as a model system for the function of prolyl bond isomerization in regulating the structure of important biological components. One such isomerase is the mammalian Pin1 prolyl cis-trans isomerase. Pin1 is an evolutionarily conserved enzyme that promotes the cis-trans isomerization of the peptide bond at the amino side of the proline residue. Uniquely among various prolyl isomerases, the Pin1 enzyme is specific towards proline residues immediately carboxy-terminal to the phosphorylated threonine or serine. The phosphorylation-dependent specificity suggests a potential role in the kinase-mediated signal transduction pathways. Indeed, the mammalian prolyl isomerase Pin1 and the yeast ortholog Ess1/Ptf1 have been implicated in cell cycle control. For example, decreased Pin1 activity in HeLa cells causes phenotypes such as cell rounding, chromosome condensation and nuclear lamin disassembly, whereas the overexpression of Pin1 induces G2/M arrest. The yeast mutant ESS1 is not viable, exhibiting mitotic arrest and nuclear fragmentation. Human Pin1 interacts in vitro with certain cell-cycle regulators including Cdc25 (Drosophila homolog: String). The molecular consequences of prolyl isomerization are marked conformational changes in the target proteins. How such molecular events are correlated in vivo with cellular functions is therefore a central question about the function of Pin1. The specificity of Pin1, p-T/SP, matches part of the phosphorylated mitogen-activated protein kinase (MAPK) or the cyclin target sites. It is reasonable to consider that isomerization of the prolyl peptide bond at these sites serves to convert a biochemical signal (phosphorylation) into a tangible structural change in the cellular components. These structural changes would probably lead to modifications of enzymatic activity, protein stability or cellular architecture (Hsu, 2001 and references therein).

The dodo mutant resembles Egfr^- and hs-CF2. Deletion mutants in the dodo region are lethal but viability can be restored by replacing the genomic copy of the adjacent fli-I gene with a transgene (Maleszka, 1996). The resulting trans-heterozygotes lacking both the dodo and pen transcription units are viable and the surviving adult females permit the examination of possible oogenic phenotypes, particularly those associated with the MAPK signaling defects in the follicle cells. Initial examination has shown a ventralized eggshell phenotype that is consistent but of low penetrance (35% at 25°C), ranging in expressivity from fused dorsal appendages to no appendages at all. This phenotype is also consistent with impaired CF2 degradation (Mantrova, 1998). Why is the null phenotype only moderately penetrant? It is speculated that, because the phenotype in question is presumed to be related to CF2 protein metabolism, its severity might be dependent on temperature. This is indeed so: the overlapping deficiency displays nearly 100% penetrance at 30°C and only ~3% at 18°C. It is possible that at elevated temperatures CF2 protein is further protected by the presence of heat-induced chaperones (Hsu, 2001).

The ventralized phenotype is reminiscent of that of the hypomorphic Egfr^- mutant or of mutants ectopically overexpressing CF2 (via a heat-shock promoter). As with the Egfr^- and hs-CF2 mutants, dodo mutation also results in maternal-effect embryonic lethality; that is, the progenies of dodo females show early patterning defects even when mated with wild-type males. The embryonic defects are typically expressed in the mutant cuticles as expanded and fused ventral denticles. CF2 degradation in the dorsal follicle cells no longer occurs in the dodo deletion mutant. The repression target of CF2, the rhomboid gene, consequently also exhibits a ventralized expression pattern. Consistent with the variable severity of the eggshell defects, rhomboid expression either is completely abolished or shows a fused two-stripe pattern, the latter reflecting perfectly the fused-appendage phenotype. The variability of rhomboid expression also indicates that the penetrance of the dodo mutant is not 100% (Hsu, 2001).

Because CF2 degradation normally occurs at the beginning of stage 10 of oogenesis and the degradation is sustained into stage 12 (Hsu, 1996), whether or not Dodo protein is present in the follicle cells at these stages was examined. High levels of expression can be detected in all oocyte-associated follicle cells. There is no apparent spatial asymmetry in the expression pattern, suggesting that Dodo protein level is not regulated by Egfr signaling. When compared with similarly treated dodo^- mutant eggs, it seems that Dodo protein is more concentrated in the nuclei, but appreciable levels of cytoplasmic staining are also detected (Hsu, 2001).

In summary, these observations make the compelling argument that Dodo is involved in facilitating the degradation of CF2, probably by isomerizing the prolyl peptide bond after MAPK-catalysed phosphorylation of the protein. To strengthen this hypothesis, three questions were addressed: (1) does Dodo bind preferentially to phosphorylated CF2 protein; (2) does Dodo facilitate the ubiquitination of CF2, and (3) is Dodo function in vivo dependent on the Egfr-MAPK signaling (Hsu, 2001)?

CF2 degradation in the anterodorsal follicle cells requires phosphorylation at the single MAPK site (PAT⁴⁰P). Replacement of Thr 40 with alanine by site-directed mutagenesis renders the mutant protein (CF2^A40) resistant to proteolysis (Mantrova, 1998). This MAPK recognition site also constitutes the potential Dodo target. If Dodo is a responder to the MAPK signal, as is proposed, the phosphorylation status at this site should directly affect the binding affinity of Dodo for CF2. Binding assays in vitro comparing phosphorylated and unphosphorylated CF2 as well as the wild type and the A40 mutant demonstrate that Dodo binding affinity for the CF2 protein is indeed markedly decreased when the MAPK/Dodo recognition site is unphosphorylated, either owing to the absence of MAPK or as a result of the mutation (Hsu, 2001).

Interestingly, although the hs-dodo transgene can rescue the ventralized dodo-null phenotype, expression of the transgene by itself does not cause a significant level of reciprocal (dorsalized) phenotype. This finding is consistent with the hypothesis that the function of Dodo requires, a priori, an activated Egfr-MAPK signaling pathway. In other words, Dodo acts as a responder to the Egfr-MAPK signaling although it is not part of the signaling cascade itself. However, this hypothesis also predicts that increased levels of Dodo should exacerbate the deleterious effect of an ectopically expressed and constitutively activated MAPK signaling pathway. This prediction was tested with a transgenic fly strain containing a constitutively active D-raf mutant complementary DNA (cDNA) under the control of heat-shock promoter. D-Raf acts upstream of MAPK in this pathway and the hs-D-raf gain-of-function (hs-D-raf^gof) transgene has been shown to induce the dorsalized phenotype such as expanded dorsel appendages. One copy of this transgene is weakly penetrant, inducing 16.6% phenocopies. This low penetrance can be enhanced more than threefold by adding only one copy of the hs-dodo transgene, to a degree comparable to two copies of D-raf^gof. The enhancing effect of dodo on D-raf^gof is also reflected in the CF2 expression pattern. In most stage-10 eggs derived from hs-dodo/+;hs-D-raf^gof/+ females, CF2 is degraded throughout the follicle cell population, similar to that observed in homozygous hs-D-raf^gof egg chambers (Hsu, 2001).

On activation by the ligand Grk, Egfr initiates a kinase cascade consisting of the conserved Ras-Raf-MEK-MAPK cassette. It is concluded that CF2 is first phosphorylated by MAPK and CF2 is then recognized by Dodo. It is suggested that Dodo alters the conformation of CF2, resulting in its recognition and degradation by the proteolytic machinery, possibly the ubiquitin- proteosome system. Without CF2, the rhomboid gene is expressed; it promotes a second wave of signaling by activating additional ligands such as Spitz. It should be noted that the Dodo function proposed here is not involved in cell-cycle control because the follicle cells at the oogenic stages described here are not proliferative, with or without the activated Egfr signaling (Hsu, 2001).

It has been a conundrum that the Pin1 enzymes, which are purportedly critical players in signal-transduction-mediated cellular functions, are not essential for viability in Drosophila and mouse. Careful examination of the dodo^- adults did not uncover any defects in the skeletal structures and organs that are specified by various signal transduction pathways, such as wings, eyes and bristles. This is surprising because at least one other essential Drosophila protein, Yan, is also regulated by MAPK-induced protein degradation. This lack of phenotype might be due to genetic compensation by redundant gene functions. A homology BLAST search of the Drosophila genome has revealed at least four uncharacterized prolyl isomerase sequences. Alternatively, a lack of zygotic dodo function might be complemented by maternal contribution. Finally, it is possible that the phosphorylation-dependent prolyl isomerization serves mainly to amplify the molecular signal generated by kinases. Such amplification might not be essential under laboratory conditions but might prove crucial in the variable natural environment. The temperature-dependent phenotype of the dodo mutant supports this speculation. The elucidation of possible embryonic functions of dodo requires the generation of single dodo mutations for genetic manipulations, such as interaction with other mutants and germline clonal analysis. In any case, the dodo gene serves a critical and specific function during oogenesis, much like other maternal-effect (and not embryonic lethal) genes in this pathway such as grk(Hsu, 2001).

Thus one of the functions of the Pin1/Dodo enzyme is to facilitate protein degradation. This might be relevant for the roles of the mammalian Pin1 in cell cycle control: it could conceivably be involved in the proteolysis of cyclins. In addition, human Pin1 has been shown to induce the dephosphorylation of Cdc25C and tau proteins. It is likely that the Pin1/Dodo proteins can have a range of targets involving diverse cellular functions. For example, the yeast mutant PTF1 defective in mRNA 3' end formation is allelic to ESS1. It is not yet clear whether or not there is a causal relationship between defects in mRNA processing and cell cycle arrest in the ESS1/PTF1 mutant. More recently, Ess1 has also been linked to chromatin remodeling and histone deacetylation. The conformational changes induced by Pin1/Dodo might result in many different physiological consequences such as changes in enzymatic activities and cofactor selection, in addition to rendering targets accessible to proteolytic degradation described here. In this regard, Pin1/Dodo can be considered to be a signal amplifier for at least a subset of cellular processes mediated by phosphorylated protein targets (Hsu, 2001 and references therein).

GENE STRUCTURE

Drosophila dodo is located at the chromosomal position 19F2-3, proximal to the flightless-I gene (fli-I) in a genomic region that has been subject to saturation mutagenesis for lethal and flightless phenotypes. However, no dodo point mutations were isolated that exhibited these phenotypes (Maleszka, 1996).

PROTEIN STRUCTURE

Amino Acids - 166

Structural Domains

Dodo shares a high degree of sequence homology with the evolutionarily conserved Pin1 enzymes. For example, it shares 61% similarity to human Pin1 (54% identity) and 60% to yeast Ess1 (45% identity). All Pin1-related proteins are characterized by a highly conserved N-terminal WW-domain involved in protein interaction and the C-terminal isomerase domain. Significantly, the two splice junctions in dodo are conserved in the mammalian gene encoding Pin1, underscoring the notion that they share a common ancestry. dodo is considered a true ortholog of the yeast ESS1 gene because it can rescue the ESS1 lethal phenotype (Maleszka, 1996).

The sequences of three full-length Dodo cDNAs were determined together with the 5' and 3' ends of a number of smaller cDNAs. The underlying genomic DNA was also totally sequenced. Two classes of Dodo transcripts were uncovered that differ in the lengths of their 3' untranslated regions, and the gene has two introns, the first of which is not always removed. The putative 5' end of the dodo transcription unit is 279 bp from the 3' end of the flightless transcript, whereas the 3' ends of dodo and penguin overlap by 353 bp. In other respects, such as intron size, the dodo transcription unit per se is unremarkable (Maleszka, 1996).

The predicted 166 amino acid Dodo protein shows 44% identity at the amino acid level with the ESS1 and PTF1 proteins of S. cerevisiae. It became clear after minor sequence corrections that ESS1 and PTF1 are the same gene, which is present in a single copy in the yeast genome. Furthermore, the Dodo protein has excellent sequence similarity to the human-expressed sequence tags deposited in data bases and it has a peptidylprolyl cis-trans isomerase (PPIase) domain in common with a number of prokaryotic and eukaryotic PPIases. The ESS1, Dodo, and human predicted proteins consist of two domains. The first is the WW domain of ~40 amino acids, which is found in a number of unrelated proteins involved in cell signaling or regulation -- e.g., dystrophin and utrophin, involved in Duchenne and Becker muscular dystrophies; and mouse Nedd4, implicated in embryonic development and nervous system function. The WW domain consists of ß-strands arranged around four conserved aromatic residues, which together with a hydrophobic core and various charged residues is indicative of well-characterized domains involved in protein-protein interactions. Indeed, the WW domain of the Yes-associated protein, YAP, has been shown to bind to identical proline-rich motifs found in two proteins (Maleszka, 1996 and references therein).

The bulk of the Dodo protein sequence has excellent sequence similarity to a new family of PPIases, enzymes involved in protein folding and unfolding that catalyze the cis-trans isomerization of Xaa-Pro peptide linkages. The prototypical member of this family is the parvulin protein of Escherichia coli, which is the smallest of the known PPIases. Its close relatives are lipoprotein PrsA from Bacillus subtilis, SurA from E. coli, protease maturation protein PrtM from Lactococcus lactus, and the NifM family from nitrogen-assimilating bacteria Azotobacter and Klebsiella pneumoniae; all are now thought to function in protein folding (Maleszka, 1996 and references therein).

date revised: 30 June 2001

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.