Research activity

Expansion of polyglutamine (polyQ) tracts has been identified as the cause for a growing number of neurodegenerative diseases. To date, eight different dominantly inherited diseases have been shown to be associated with such expansions in their respective proteins, including Huntington disease, dentatorubropallidoluysian atrophy (DRPLA), and several spinocerebellar ataxias. Expansion of the polyQ repeat in these disease proteins results in selective death of neurons in different regions of the brain. DRPLA is caused by the expansion of a CAG repeat in the atrophin-1 gene, resulting in the expansion of a polyQ tract within the Atrophin-1 protein. Atrophin-1 mRNA and protein are ubiquitously expressed in the mammalian central nervous system. In humans there is also Atrophin-2, a related protein but void of any polyQ tract. At the cellular level DRPLA results in ubiquitinated inclusions and diffuse accumulation of mutant atrophin-1 in the neuronal nuclei in granule cells in the cerebellar cortex.
The cytopathological mechanism of DRPLA and all other polyQ diseases remain far from being determined. For a long time the dominant model in the field predicted that proteolytic processing by caspases and nuclear accumulation of truncated expanded polyQ fragments may exert toxic effects within cells (Fig. 1). By sequestering other proteins containing short polyQ stretches via a direct interaction involving noncovalent “polar zippers” between the polyQ repeats, the nuclear aggregates would result in transcriptional abnormalities. Although a good deal of experimental evidence supports this model, several experiments have failed to demonstrate any causal role for nuclear inclusions in neurodegeneration and recently it is more commonly assumed that the large polyQ aggregates have a protective role and that the toxicicity of polyQ is exerted by mono or small oligomeres that are still soluble.
Also despite many studies that demonstrate the presence of large-scale alteration of transcription in polyQ cellular and animal models, it has not been determined whether these alterations have a causal role in determining the neuronal death observed in patients.

Figure 1. Model for PolyQ toxic activity inside the cell. From C. A. Ross Neuron, 35, 819–822, (2002) zoom

Atrophins are relatively poorly studied, multifaceted proteins that take part in several cellular and developmental processes. They have been reported to act as transcriptional co-repressors possibly by recruiting Histon-deacetylases (HDAC). There is also some suggestion that they may have a more general influence on chromatin remodelling, being equivalent to the TritoraxG genes. However they are also present in the cytoplasm where they may interact with other proteins. Yeast two hybrid studies have revealed that Atrophin-1 can interact with Membrane Associated GUanilate Kinases (MAGUK) multi PDZ scaffolding proteins as well as with Nedd4/like Ubiquitin E3 ligases with WW domains. Atro, the only Drosophila Atrophin, has been shown to interact with an atypical cadherin (Fat). Atro plays an important role in crucial signalling pathways and is required for diverse processes such as planar cell polarity, some forms of cell adhesion/cell affinities as well as embryonic segmentation, leg and eye development. Our working model about Atrophins views them as functional equivalents of b-catenin (Armadillo, Arm in Drosophila).

Objective 1. Genetic and molecular analysis of Atro functions.

Any modelling the polyQ diseases in flies has been limited by the absence of information about the normal functions of the corresponding endogenous proteins. Atro is the first of these genes to be studied in flies. The possibility of combining classic loss of function genetic analysis with the standard approaches used for polyQ modelling is an unprecedented advantage that can only be fully exploited if we obtain a better picture of the multiple activities of Atro and of the molecular networks in which it is involved. We are currently analysing the relationship between Atro and N signalling. Atro is a negative regulator of EGF signalling, given this pathway’s intense cross-regulation with N in fly development, Atro effect on N signalling could be indirectly mediated by its action on EGF. However, we have identified at least one process, the cell fate choice between photoreceptor R3 and R4 in the eye (clearly affected by N and not by EGF) where Atro’s influence on N may be uncoupled by that on EGF. In this system, which we have further characterised for the components of the N pathway required, Atro acts as a negative regulator of the N signalling. Atro and Notch interact genetically and that Atro is able to repress Notch-dependent transcription, consistent with its function as transcriptional co-repressor. A manuscript is in preparation.
Our biggest effort is however placed in undertaking a series of unbiased screens to identify new interactors of Atrophin, both genetically and physically. The combination of Yeast 2 Hybrid or other physical interaction techniques and Drosophila genetics are particularly powerful because the key point of the functional significance of a given interaction can be readily and thoroughly tested. We are performing a Y2H screen using the C-ter of Atro (containing the most conserved sequences between Atrophins) as bait and a Drosophila adult cDNA library as pray. One of the first interactors recovered is Groucho, a transcriptional co-repressor, known to be required for the repression of Notch targets. We are currently testing the meaning of the Groucho-Atro interaction for the differentiation of the R3/R4 photoreceptors. Interestingly both atro and groucho mutants display hypertrophy of the embryonic nervous system, indicating that they may act together also at that stage of fly development.
The Y2H approach has however a limitation in that we failed to express full length Atro in this system. We have thus resolved to use in the future a Tap-tag strategy to purify biochemically interactors of full length Atro and identify them by mass spectrometry.
Besides looking for binding partners of Atro we plan to exploit the exceptional capability of Drosophila to identify molecular networks through genetic screens that identify interactions on the basis of their functional relevance. Overexpression of Atro in the eye leads to a roughening of the eye surface (Fig. 2), which allows a classic modifier screen for randomly generated mutations that would suppress or enhance this phenotype.

Figure 2. Eye roughening caused by the overexpression of Atrophin

These studies will be complemented by, the generation of a Drosophila model for
We have just started to generate such mutants by classic EMS mutagenesis and are performing the initial test screen to check the mutagenesis rate and the percentage of interacting mutations. We have resolved ourselves for the use of EMS rather than insertional mutagenesis because in such screens the interacting mutation is tested in heterozygosis and EMS delivers a higher percentage of null mutations with complete loss of a gene function (whereas P-elements and other transposons mostly generate hypomorphic mutations). This may be crucial to recover an interaction and we favoured this aspect over the undoubtly more laborious gene mapping procedure necessary after EMS mutagenesis compared to insertional mutations.
Atro and the Atrophins have been described as transcriptional co-repressors that do not bind DNA directly, and Atro and Atrophin-2 recruit Histon Deacetylases at the sites of repression. It is however unknown how specific their repressive action is as they have been implicated in a wide variety of events.
To understand Atro’s role on transcription we have deviced a series of experiment set-ups that aim at identifying Atro targets. First of all we have established a cell culture system of Drosophila neuronal cell lines that will be transfected with Atro under a modular inducible system allowing us manipulations in the levels of Atro expression. This cell culture system will be used in a gene profiling experiment to identify through microarray hybridisation those genes whose transcription varies with the levels of Atro. This in-vitro cell culture set up was chosen strategically as the simplest and most homogeneous cell population, which we consider an essential trait for our purposes (identification of a protein targets rather than genes involved in a given process), in comparison with an in-vivo but more complex system such as embryos. Identified gene candidates will be further verified by Real Time PCR.
Atro targets should not only satisfy the condition that their expression levels vary with Atro content, but their promoter regions should bind complexes in which Atro is present. We will address this question partially by Chromatin Immuno Precipitation starting from the cells culture, but we are planning a more ambitious set up which will add to this information the value of being entirely in-vivo. To achieve this we have fused the bacterial Dam methylase at the N-ter of Atro and will generate transgenic flies with this construct. The Dam-Atro protein should be then recruited at sites that normally bind Atro-containing complexes and this should result in methylation of a region of 2kb around these sites. This newly developed technique called DamID has the advantage over ChIP of not using chemical cross linking and allowing the marking of DNA to happen entirely in-vivo. Methylation can then be easily detected at specific sites by PCR or all hyper-methylated regions can be isolated and hybridised on microarrays slides. In-vivoDrosophila studies will further assess the functional interaction between Atro and the putative targets identified.

Objective 2. Atro and neurodegeneration, with and without polyQ.

The combination of the above-mentioned screens in yeast and flies and the microarray profiling and DamID should give us a thorough knowledge of Atro functions which will be the natural background against which our attempt to model the DRPLA disease in flies will be developed with the aim of addressing key issues of the biology of this disease and of other polyQ disorders.
Our first attempt is running in collaboration with Bernard Charroux and Steve Kerridge at the University of Marseille and is based on the expression of different versions of vertebrate Atrophin-1 and 2. We have established that full length Atrophin-1 both with normal Q stretch and bearing a 65Q expansion are barely detected in flies and have also limited capability of inducing neuro-degeneration. However a C-ter truncated fragment that is found specifically in the disease and is retained in the nucleus is detected at reasonable levels of expression and the 65Q version is specifically able to induce neuronal degeneration when expressed in the eye.
A more ambitious project concerns the expansion of the endogenous polyQ stretches in Drosophila Atro. Atro has two short stretches of 11Q and 14Q which we have expanded to more than 65Q. These constructs in comparison to the use of human Atrophins will allow a more direct comparison of the functionality of polyQ expanded Atro vs. wt Atro in all the above-mentioned projects and experimental set ups and thus this should be viewed as a general leit-motiv which will be present in all our projects. Besides addressing the ability of polyQ Atro to induce neuronal degeneration we will specifically test its ability to repress Notch target genes, to interact with Groucho and the other interactors isolated in flies and yeast, for its ability to regulate Atro target genes in our cell culture system and in the DamID assays.
Such a comprehensive parallelism between wt Atro and polyQ Atro is of great importance for addressing a key issue in the field of polyQ diseases, i.e. the alteration of the wt functions and how these are related to neurodegeneration.

Fig. 3 Degeneration inside a fat homozygous mutant cellular clone generated in a atro heterozygous mutant background.

We have also accumulated evidences that Atro normal functions are connected to degeneration of neuronal photoreceptors in flies and that it co-operates with Fat to maintain the natural homeostasis of the retina. Loss of fat determines mild adult degeneration of the photoreceptors, which is dramatically enhanced by loss of atro only in heterozygosis (Fig 3). We are currently characterising from a cell biological point of view this degeneration, establishing whether it’s retina-specific or is found in other neurones and are planning a deficiency-based genetic screen to identify all the chromosomal regions that in heterozygosis act similarly to atro by enhancing the degenerative eye phenotype.

Fanto Home  | Curriculum | Lab Members | Research | Publications | Links