Plasmid prediction in Micrococcus bacterial strains

Abstract

Plasmids are circular or linear extrachromosomal DNA molecules that replicate autonomously and occasionally provide their guests with bacterial extra genetic material important for their survival and adaptation. The sequencing of bacterial genomes has generated a vast wealth of data that can be processed by different computational tools to identify plasmid sequences. This would allow expanding the knowledge about plasmids and their diversity in most prokaryotic taxa. We tested this idea in a barely studied bacterial genus such as Micrococcus. These are environmental bacteria, and the best-known species is M. luteus, sometimes associated with skin and opportunistic infections. Other species show potential for biotechnological applications, as they can produce antibiotics, biofuels, enzymes and could be applied as biofertilizer or in bioremediation processes. Draftgenomes were obtained from sequencing reads of 20 strains of Micrococcus. The combination of different methods on these genomes allowed us to detect the presence of sequences associated with plasmids in 17 of the selected strains. The predictionsare not complete plasmids, but rather a set of fragments. In these sequences, genes directly associated with plasmid functions (replication and segregation) were detected, as well as accessory genes related to resistance to toxic compounds, oxidative stress, and antibiotics. To test the novelty of these predictions, they were analyzed with the software Copla to identify plasmid taxonomic units (PTUs). Only one set was classified in a PTU containing a diverse set of plasmids that could be involved in horizontal gene transfer between different phyla. Thus, most of the predictions might represent novel plasmids. In addition, a bipartite bacterial network was constructed with the plasmid predictions and known as actinobacterial plasmids. These networks include two types of nodes: genomic nodes representing each plasmid or genetic unit, and protein nodes representing clusters of protein sequences encoded by the different plasmids. Our network included 833 actinobacterial plasmids, 17 predictions, and 112878 proteins. The network had poor connectivity, with most of the nodes consisting of single elements related to isolated plasmids. 80% of the nodes were hypothetical proteins and 69% included only one protein sequence. From the non-hypothetical proteins, 1438 were annotated as transposases, an abundant element in plasmids, and they formed the largest clusters. This suggests that most actinobacterial plasmids are unique and highlights the lack of knowledge on the biology and roles of these mobile genetic elements in Actinobacteria. Still, this represents a significant addition to the Micrococcus plasmid sequences pool and the first step in a study over the whole phylum.