Since the genetic data indicates that Atro functions closely with eve and hkb, tests were performed for possible physical interactions using a GST pull-down assay. It was found that the full-length radiolabeled Atro can bind to a full-length Eve (GST-Eve) or Hkb (GST-Hkb), but not to GST alone (Zhang, 2001).
To map the domains within Atro responsible for these interactions, a series of Atro deletions were generated and the pull-down assay was performed. The results suggest that the C-terminal region of Atro (aa 1324-1985) is responsible for its binding to Eve, and further deletion in this region diminishes its binding ability. Interestingly, this C terminus Eve binding domain is highly conserved in the Atrophin family proteins, suggesting that this interaction might be evolutionarily conserved (Zhang, 2001).
To define the region in Eve responsible for its interaction with Atro, a series of GST-Eve deletions were generated. Previous work has divided the Eve protein into six regions (regions A-F). It was found that neither Eve's homeodomain (region B) nor the EF region shows significant interaction with Atro. Instead, only the CD region of Eve binds to the Atro protein. Further deletion of either region C or D significantly reduces its binding ability to Atro. This CD region has been previously defined as Eve's minimal repressor domain (Han, 1993), suggesting that the minimal repression domain in Eve functions to bind Atro (Zhang, 2001).
The binding data, along with the genetic data, suggest a possible mechanism in which Atro functions as a corepressor for site-specific repressors like Eve and Hkb. One function of Eve and Hkb might be to recruit Atro to the promoter site where Atro can exert its repressive activities. This hypothesis would predict that Atro can directly repress transcription when it is tethered to DNA via a heterologous DNA binding domain. To test this, an Atro-GAL4 fusion was generated and its function was examined in the Kreggy/NEE-LacZ system. Briefly, full-length Atro was fused to the C terminus of the Gal4 DNA binding domain (Gal4DB::Atro), and the chimeric gene was placed under the control of the Kruppel promoter (Kr-Gal4DB::Atro), which drives gene expression in a broad band in the central region of the blastoderm-stage embryo. The LacZ reporter gene (UAS-NEE)-LacZ, which is driven by a modified rhomboid NEE enhancer that contains three UAS sites for Gal4 binding, is normally expressed in the ventral side of the same stage embryos. However, when the (UAS-NEE)-LacZ flies are crossed with the Kr-Gal4DB::Atro transgenic animals, their progeny show a repressed LacZ transcription in the central region where the Gal4DB::Atro fusion protein is expressed. This result suggests that the full-length Atro protein can behave as a transcriptional corepressor in vivo (Zhang, 2001).
Given the sequence similarity between Atro and human Atrophin-1, it is possible that human Atrophin-1 also functions as a transcriptional corepressor in vivo. Thus, this possibility was tested using the Kreggy/NEE-LacZ system in fly embryos. Interestingly, when tethered to Gal4 DNA binding domain, human Atrophin-1 can repress LacZ transcription in (UAS-NEE)-LacZ reporter embryos, since more than 75% of the examined embryos exhibited a repressed LacZ transcription, suggesting that the function of Atrophin family proteins are evolutionarily conserved. Human Atrophin-1 was further tested with a poly-Q expansion in the same system and it was found that only about 18% of the examined embryos displayed a repressed LacZ transcription, suggesting that poly-Q expansion alters Atrophin's transcriptional repressive activity (Experimental Procedures) (Zhang, 2001).
RNA in situ and Northern analysis have revealed that a single Atro transcript is expressed throughout embryonic development, including 0- to 2-hr-old embryos, indicating a maternal contribution. Immunostainings with anti-Atro antibodies reveal that in early-stage embryos, Atro protein was ubiquitously expressed in every cell, while in later-stage embryos the protein was expressed at a higher level in the nervous system. Atro is also expressed in all larval tissues examined, including brain and imaginal discs. Interestingly, the Atro protein was mainly located within the nucleus in a punctuate form (Zhang, 2001). .
Animals homozygous for the P element insertions or e46-2 alleles died at a late embryonic stage with no obvious morphological defects. Since defects caused by Atro could be masked by its maternal contribution, mutant embryos lacking maternal Atro products (Atromat-) were generated using the FLP-DFS technique. Among the embryos produced from Atromat- mothers and Atro-/+ fathers, those lacking both maternal and zygotic Atro gene product did not develop, whereas those maternally mutant but zygotically rescued embryos displayed a range of patterning defects. To understand the function of Atro in a greater detail, the role of Atro was further investigated in early embryonic patterning where many molecular markers are available (Zhang, 2001).
Cuticle preparations showed that some Atromat- embryos have widened ventral denticle belts in the dorsal/ventral direction. Using an antibody against Twist, a protein expressed in the ventral-most 20 cells of stage 5 wild-type embryos, it was found that Twist expression becomes broader and covers most of the ventral cells in Atromat- embryos. These data indicate that Atromat- embryos are ventralized and suggest a requirement for Atro in dorsoventral patterning (Zhang, 2001). .
In addition, cuticle preparations also showed that many of these embryos have holes or patches of naked cuticle. Such phenotypes have been observed in animals mutant for neurogenic genes in which the hyperplasia of neuronal tissues at the expense of epidermis leaves an insufficient amount of epidermal cells to cover the whole embryo. To test whether Atromat- embryos exhibit neurogenic defects, the mutant embryos were stained with antibodies against Elav, which marks both the central and peripheral nervous system. The number of Elav-positive cells is dramatically increased in the Atromat- embryos, suggesting a role for Atro in the neurogenic process as well (Zhang, 2001).
Anti-Elav staining also revealed that the number of metameric neuronal units in the Atromat- embryos is reduced, suggestive of a simultaneous segmentation defect. Indeed, even in mutant embryos with less severe neurogenic defect, the missing and misshaped ventral denticle belt phenotype is evident. To further study the segmentation phenotype, the expression of engrailed (en), a segment-polarity gene, which in wild-type embryo is expressed as 14 stripes at the anterior boundary of each parasegment, was examined. In Atromat- embryos, the number of en stripes is reduced, and several of those remaining en stripes are incomplete or fused together. Such a segmentation defect was also confirmed by examining the expression of another segment-polarity gene, wg. Taken together, these results indicate that maternal Atro is essential for proper embryonic pattern formation (Zhang, 2001).
Because the expression of segment-polarity genes is defined by the pair-rule genes, it was possible that the abnormal en and wg expression patterns are a result of aberrant pair-rule gene regulation caused by the Atro mutations. Therefore, the expression patterns of pair-rule genes were examined in Atro mutant embryos. In wild-type embryos, the pair-rule genes (such as eve, hairy, runt, and ftz) are expressed as seven precisely defined stripes. However, in Atromat- embryos, these stripes are expanded and their boundaries become less defined. Of special note is the shift in expression of the pair-rule genes into the posterior region in the Atromat- embryos, resembling the phenotype observed in embryos with mutations in the terminal gap gene hkb (Zhang, 2001).
Since the stripe boundaries of the pair-rule genes are restricted by the repressive activities of gap genes, gap gene expression was further examined in the Atromat- embryos. The expression patterns of the four gap genes examined, including hb, kni, kr, and giant, are nearly normal as compared to the wild-type embryos. This result suggests that the repressive activities of the gap genes, but not their expression, might be compromised in the Atromat- embryos. Taken together, these findings suggest that Atro might be required for the repressive activities of multiple transcription factors during embryonic segmentation (Zhang, 2001).
The above results suggest that the maternal Atro is essential for proper embryo pattern formation. To further understand the function of Atro in embryonic development, possible genetic interactions were sought between Atro and genes involved in early embryogenesis. Heterozygous Atro-/+ flies were crossed to flies carrying heterozygous mutations affecting early embryogenesis, and the number of transheterozygous progeny from these crosses was scored. While most mutants tested did not display obvious genetic interactions, mutations affecting the eve and hkb transcriptional repressors showed a dosage-sensitive interaction with maternal Atro. From crosses of heterozygous Atro-/+ females and eve-/+ males, almost all of the eve-/+ progeny (both eve-/+; Atro-/+ and eve-/+; Atro+/+) were absent. In contrast, in the reciprocal crosses where Atro-/+ males were mated with eve-/+ females, a normal percentage of eve-/+ progeny was observed. These data showed that the lethality is caused by a reduction of maternal Atro dosage. Similar interactions were also observed with hkb. These findings suggested that Atro might mediate the functions of eve and hkb repressors in vivo (Zhang, 2001).
The eve-/+; Atro-/+ double heterozygous embryos produced by Atro-/+ mothers die at a late embryonic stage. These embryos lose some or all of the even-numbered ventral denticle belts, mimicking the phenotype of hypomorphic eve mutant embryos. However, staining of these embryos with an Eve antibody reveal that the expression of Eve is normal, indicating that it is the activity of eve that is affected by the reduced Atro dosage. To confirm this, the expression patterns of two eve target genes, wg and en were analyzed. In wild-type embryos, the wg expression stripe is mostly one-cell wide, with only the even-numbered wg expression restricted by eve. In the eve-/+; Atro-/+ double heterozygous embryos, while the odd-numbered wg stripes maintain their one-cell width, there is considerable anterior expansion of the even-numbered wg stripes, indicating that the repressive activity of eve is compromised. Similarly, in wild-type embryos, en is expressed as 14 evenly spaced stripes in the embryo trunk region. Although both the odd- and the even-numbered en expression stripes are regulated by eve, a higher level of eve's repressive activity is required to define the odd-numbered en stripes, while a low level of eve's activity is sufficient for the even-numbered ones. In the eve-/+; Atro-/+ double heterozygous embryos, the odd-numbered en stripes are expanded posteriorly, resembling phenotypes observed in homozygous hypomorphic eve mutant embryos. Taken together, these results demonstrate that Atro is important for the repressive activity of eve and suggest that these two genes might function closely in the segmentation pathway (Zhang, 2001).
Grunge is required for patterning of the ventral proximal leg; in mutant cells proximal patterns are replaced with more distal ones. Gug is also required in all ventrally located cells along the entire proximal-distal axis of the leg for specific ventral identities or processes (Erkner, 1997).
Ellerby, L. M., et al. (1999). Cleavage of atrophin-1 at caspase site aspartic acid 109 modulates cytotoxicity. J. Biol. Chem. 274(13): 8730-6. 10085113
Erkner, A., Vola, C., Fasano, L. and Kerridge, S. (1997). Grunge is a postive regulator of teashirt expression in the proximal parts of the imaginal discs.A. Dros. Res. Conf. 38 1997 :157A
Han K. and Manley J. L. (1993). Transcriptional repression by the Drosophila even-skipped protein: definition of a minimal repression domain. Genes Dev. 7: 491-503. 8095483
Knight, S. P., et al. (1997). Expression and distribution of the dentatorubral-pallidoluysian atrophy gene product (atrophin-1/drplap) in neuronal and non-neuronal tissues. J. Neurol. Sci. 146(1): 19-26. 9077491
Li, C. and Manley, J. L. (1998). Even-skipped represses transcription by binding TATA binding protein and blocking the TFIID-TATA box interaction. Mol. Cell. Biol. 18: 3771-3781. 9632760
Mannervik, M. and Levine, M. (1999). The Rpd3 histone deacetylase is required for segmentation of the Drosophila embryo. Proc. Natl. Acad. Sci. 96: 6797-6801. 10359792
Nagafuchi, S., et al. (1994). Structure and expression of the gene responsible for the triplet repeat disorder, dentatorubral and pallidoluysian atrophy (DRPLA). Nat. Genet. 8(2): 177-82. 7842016
Schilling, G., et al. (1999). Nuclear accumulation of truncated atrophin-1 fragments in a transgenic mouse model of DRPLA. Neuron 24(1): 275-86. 10677044
Takahashi. H., et al. (2001). Neuronal nuclear alterations in dentatorubral-pallidoluysian atrophy: ultrastructural and morphometric studies of the cerebellar granule cells. Brain Res. 919(1): 12-19. 11689158
Wood, J. D., et al. (2000). Atrophin-1, the dentato-rubral and pallido-luysian atrophy gene product, interacts with ETO/MTG8 in the nuclear matrix and represses transcription. L. Cell Biol. 150: 939-948. 10973986
Yamada, M., et al. (2001). Widespread occurrence of intranuclear atrophin-1 accumulation in the central nervous system neurons of patients with dentatorubral-pallidoluysian atrophy. Ann. Neurol. 49(1): 14-23. 11198291
Yazawa, I., et al. (1995). Abnormal gene product identified in hereditary dentatorubral-pallidoluysian atrophy (DRPLA) brain. Nat. Genet. 10(1): 99-103. 7647802
Zhang, S., Xu, L., Lee, J. and Xu, T. (2002). Drosophila atrophin homolog functions as a transcriptional corepressor in multiple developmental processes. Cell 108(1): 45-56. 11792320
date revised: 2 February 2002
Home page: The Interactive Fly © 1997 Thomas B. Brody, Ph.D.
The Interactive Fly resides on the
Society for Developmental Biology's Web server.