Identification and characterization of thousands of bacteriophage satellites across bacteria

Abstract

Bacteriophage-bacteria interactions are affected by phage satellites, elements that exploit phages for transfer between bacterial cells. Satellites can encode defense systems, antibiotic resistance genes, and virulence factors, but their number and diversity are unknown for lack of a tool to identify them. We developed a flexible and updateable program to identify satellites in bacterial genomes - SatelliteFinder - and use it to identify the best described families: P4-like, phage inducible chromosomal islands (PICI), capsid-forming PICI, and phage-inducible chromosomal island-like elements (PLE). We vastly expanded the number of described elements to [~]5000, finding hundreds of bacterial genomes with two different families of satellites, and dozens of Escherichia coli genomes with three. Most satellites were found in Proteobacteria and Firmicutes, but some are in novel taxa such as Actinobacteria. We characterized the gene repertoires of satellites, which are variable in size and composition, and their genomic organization, which is very conserved. With the partial exception of PICI and cfPICI, there are few homologous core genes between families of satellites, and even fewer homologous to phages. Hence, phage satellites are ancient, diverse, and probably evolved multiple times independently. Occasionally, core genes of a given family of satellites are found in another, suggesting gene flow between different satellites. Given the many elements found in spite of our conservative approach, the many bacteria infected by phages that still lack known satellites, and the recent proposals for novel families, we speculate that we are at the beginning of the discovery of massive numbers and types of satellites. SatelliteFinder is accessible for the community as a Galaxy service at https://galaxy.pasteur.fr/root?tool_id=toolshed.pasteur.fr/repos/fmareuil/satellitefinder/SatelliteFinder/0.9