Evolution of protein interactions involved in epithelial building and patterning (EPIE)


The emergence of epithelial layers is one of the key innovations characterizing the Metazoan lineage. The criteria defining an animal epithelium (basement membrane, junctions, cell polarity) were based on the observation of bilaterian models only, pushing aside the structural diversity of epithelia found in non-bilaterian lineages (Porifera, Ctenophora, Placozoa and Cnidaria). To decipher the minimal and ancestral toolkit of animal epithelia and trace back how and when the bilaterian epithelial features emerged, computational and experimental comparative analyzes on non-bilaterian animals are needed. This project aims at determining the proteins involved in the junctions of homoscleromorph sponges, the only sponge class where adherent-like actin-rich junctions were described as adherence junctions (AJ): this work will enable to test the homology of these structures with bilaterian AJs.


Epithelium, animal evolution, protein domain interactions, in silico prediction, immunolocalization


The project has 3 main objectives: 1) To determine the expression signatures of the two epithelial layers of Oscarella lobularis, pinacoderm and choanoderm, and compare them with recently published data on another class of sponges devoid of adherens-like junctions, 2) To predict protein domains and interactions between newly identified protein candidates in silico and test predicted interactions in vitro; 3) To perform (co)localization assays of proteins/junctions with the help of specific antibodies.

Proposed approach

Aim 1: to identify potential candidate genes for early metazoan AJ in sponges, an unbiased whole organism single cell transcriptomic approach will be performed in Oscarella (at EMBL with chromium 10x technology). This high throughput sequencing data will enable us 1) to separate epithelial (choanocyte + pinacocyte populations) vs mesenchymal (mesohyl cell populations) specific genes in this species and 2) to identify epithelial genes specific to a sponge with AJs by comparison with data published in a sponge devoid of AJs. Candidates will be verified by in situ hybridization.

Aim 2: we will identify protein domains (Hidden Markov Model comparisons) and predict protein-protein interactions (by using protein interaction databases and domain-based interaction prediction algorithms). We will test predicted interactions by biochemical approaches (pull down + mass spectrometry).

Aim 3: co-localization of candidate proteins and junctions will be done by immunolocalization (raised specific antibodies) using SEM/STED.


The success of all three aims fully relies on the identification of a candidate with interdisciplinary training in experimental cell and developmental biology, as well as computational biology, as he/she will have to carry out computational, as well as biochemical and cell biological techniques to address this question. The postdoc will be supervised by 3 PIs with complementary expertise: André Le Bivic (epithelium and biochemistry), Emmanuelle Renard (sponge and evolution) and Bianca Habermann (computational biology).

Expected profile

The candidate must have a strong background in evo-devo approaches and thereby must be used to deal with both cell, developmental and evolutionary concepts and methods. He/she should be proficient in bioinformatics tools related to evolutionary analysis as well as molecular, cellular, imagery and biochemical methods. Team spirit and communication skills are highly necessary for this interdisciplinary project. A previous experience with non-model organisms would be a plus.