Section: New Results

Linking Homology and Function for Algal Desaturases

Polyunsaturated fatty acids (PUFA) such as Omega-3 that are essential for human health cannot be synthesized by the human body and must be acquired through the consumption of certains foods, such as oily fish, that are becoming increasing difficult to produce sustainably. However, fish do not produce them either: their role is to concentrate PUFA through the food chain. There is consequently considerable interest in producing these essential nutriments directly, through the cultivation of domesticated strains of naturally occurring or of engineered strains of green algae.

Ultimately, polyunsaturated fatty acids are produced by molecular machines called desaturases. While desaturases are abundant in all branches of life, the link between gene sequence and the precise activity of the corresponding enzyme is poorly understood. The particular challenge is that, while the catalytic active site is well conserved, the features that recognize the substrate and that determine the regiospecificity of the enzyme are not. In order to produce specific PUFA at industrial scale, it is necessary to develop efficient tools for high-throughput identification of candidate genes in algal species, and precise models for designing desaturases through synthetic biology.

In collaboration with the LBM (UMR 5200 CNRS) and with the support of the Inria Project Lab *In silico algae*, we used a core collection of thirteen desaturases from Osteococcus tauri to explore the link between homology and function in 23 species ([9]). The study reinforced our understanding of the evolutionary conservation of desaturases and confirmed the identification of substrate and regio- specificity through graph neighborhoods. We were further able to extend the identification of PPR motifs correlated with specificity. This work is ongoing. Since most of the pertinent desaturases are membrane bound, the prediction of protein structure has proved perilous, but we are hopeful that future work will allow us to use structure-inspired prediction to narrow in on the sites responsible for specificity despite their poor sequence conservation.