Prokaryotic microbiota outperform eukaryotic microbiota in differentiating between infection states of iconic diseases of two commercial oyster species

K. Mathias Wegner (1) , Benjamin Morga (2) , Laure Guillou (3) , Martina Strittmatter (3) , Cyrielle Lecadet (2) , Marie-Agnes Travers (4, 2) , Delphine Tourbiez (2) , Ophélie Gervais (2) , Isabelle Arzul (2)

1 AWI - Alfred-Wegener-Institut, Helmholtz-Zentrum für Polar- und Meeresforschung = Alfred Wegener Institute for Polar and Marine Research = Institut Alfred-Wegener pour la recherche polaire et marine
2 SGMM - Unité Santé, Génétique et Microbiologie des Mollusques
3 ECOMAP - ECOlogy of MArine Plankton
4 IHPE - Interactions Hôtes-Pathogènes-Environnements

The role of microbiota in health and disease is most often expressed by structural shifts of the taxonomic composition of prokaryote communities in infected and healthy individuals. In cultured aquatic animals with open circulatory systems, such as mollusks, microbiota also harbor a wide range of protists, which are unicellular eukaryotes that could also play an important role during infections. To evaluate the effectiveness of eukaryotic vs. prokaryotic microbiota in characterizing infection states, we examined both microbial compartments under natural conditions in two commercially important oyster species, the flat oyster Ostrea edulis and the Pacific oyster Magallana (Crassostrea) gigas. With O. edulis being infected by two protist parasites, Marteilia refringens and Bonamia ostreae, and M. gigas being infected by the ostreid herpes virus OsHV-1, we chose iconic diseases responsible for substantial mortalities and economic damage within the two species. We analyzed and compared the structural and compositional differences between healthy and infected oysters and used random forest machine learning to classify infection states and identify indicator taxa that distinguish healthy from infected individuals. Both at the structural and compositional levels, bacterial microbiota proved to be better predictors of infection states. By eliminating noisy taxa through variable selection in the random forest models, we enhanced the compositional differences between infection states. In all host-pathogen combinations, only a few taxa (<31) were required to achieve optimal separation. While the identity of indicator taxa will partly reflect the specific environmental conditions at the time of sampling, we recovered several previously described indicator taxa, such as Mycoplasma, Vibrio, Photobacterium, and Arcobacter. Next to these we also discovered new taxa like Motiliproteus that exhibited the potential to differentiate between infection states of the investigated O. edulis specimen. The simultaneous characterization of prokaryotic and eukaryotic microbiota suggests that only few prokaryotic indicator species might be needed to reliably differentiate between infected from healthy individuals and monitor infection risks.

Mots clés: Microbiome; 16s rRNA 18s rRNA amplicon sequencing; mollusk; bivalve; probiotic

Domaines: Médecine vétérinaire et santé animale; Science des productions animales

K. Mathias Wegner, Benjamin Morga, Laure Guillou, Martina Strittmatter, Cyrielle Lecadet, et al.. Prokaryotic microbiota outperform eukaryotic microbiota in differentiating between infection states of iconic diseases of two commercial oyster species. Aquaculture, 2025, 594, pp.741363. ⟨10.1016/j.aquaculture.2024.741363⟩. ⟨hal-04665295⟩