Plasmids can be transferred between cells by conjugation, thereby driving bacterial evolution by horizontal gene transfer. Yet, we ignore the molecular mechanisms of transfer for many plasmids because they lack all protein-coding genes required for conjugation. We solved this conundrum by identifying hundreds of plasmids and chromosomes with conjugative origins of transfer in Escherichia coli and Staphylococcus aureus. These plasmids (pOriT) hijack the relaxases of conjugative or mobilizable elements, but not both. The functional dependencies between pOriT and other plasmids explain their co-occurrence: pOriT are abundant in cells with many plasmids, whereas conjugative plasmids are the most common in the others. We systematically characterized plasmid mobility in relation to conjugation and alternative mechanisms of transfer and can now propose a putative mechanism of transfer for ∼90% of them. In most cases, plasmid mobility seems to involve conjugation. Interestingly, the mechanisms of mobility are important determinants of plasmid-encoded accessory traits, since pOriTs have the highest densities of antimicrobial resistance genes, whereas plasmids lacking putative mechanisms of transfer have the lowest. We illuminate the evolutionary relationships between plasmids and suggest that many pOriT may have arisen by gene deletions in other types of plasmids. These results suggest that most plasmids can be transferred by conjugation. Some plasmids encode complex cellular structures to transfer between bacteria by conjugation, which facilitates the spread of adaptive trains between cells. Yet, half of all plasmids lack the protein-coding genes required for conjugation. Their mobility has remained a mystery that we now contribute to solve by proposing a mechanism of transfer for 9 out of 10 plasmids. We identified the only non-coding sequence required for conjugation, the origin of transfer (oriT), in hundreds of plasmids of Escherichia coli and Staphylococcus aureus. These plasmids might have originated from larger ones and have the highest density of antibiotic resistance genes. Their horizontal transfer depends on complex functional dependencies with other plasmids, which explains their co-existence in bacterial cells.

Antimicrobial resistance is one of the major threats to Public Health worldwide. Understanding the transfer and maintenance of antimicrobial resistance genes mediated by mobile genetic elements is thus urgent. In this work, we focus on the ColE1-like plasmid family, whose distinctive replication and multicopy nature has given rise to key discoveries and tools in molecular biology. Despite being massively used, the hosts, functions, and evolutionary history of these plasmids remain poorly known. Here, we built specific Hidden Markov Model (HMM) profiles to search ColE1 replicons within genomes. We identified 1,035 ColE1 plasmids in five Orders of γ-Proteobacteria, several of which are described here for the first time. The phylogenetic analysis of these replicons and their characteristic MOB_P5/HEN relaxases suggest that ColE1 plasmids have diverged apart, with little transfer across orders, but frequent transfer across families. Additionally, ColE1 plasmids show a functional shift over the last decades, losing their characteristic bacteriocin production while gaining several antimicrobial resistance genes, mainly enzymatic determinants and including several extended-spectrum betalactamases and carbapenemases. Furthermore, ColE1 plasmids facilitate the intragenomic mobilization of these determinants, as various replicons were identified co-integrated with large non-ColE1 plasmids, mostly via transposases. These results illustrate how families of plasmids evolve and adapt their gene repertoires to bacterial adaptive requirements.

The ability of bacteria to horizontally transfer genetic material has turned antimicrobial resistance into one of the major sanitary crises of the 21st century. Plasmid conjugation is considered the main mechanism responsible for the mobilization of resistance genes, and its understanding is crucial to tackle this crisis.

Conjugative plasmids are important drivers of bacterial evolution, but most plasmids lack genes for conjugation. It is currently not known if the latter can transfer because origins of transfer by conjugation ( oriT ), which would allow their mobilization by conjugative plasmids, are poorly known. Here, we identify and characterize occurrences of known oriT families across thousands of plasmids confirming that most conjugative and mobilizable plasmids still lack identifiable families of oriTs . They reveal clear patterns in terms of intergenic position, distance to the relaxases, and MOB-type association. This allowed to develop a computational method to discover novel oriT s. As a proof of concept, we identify 21 novel oriTs from the nosocomial pathogens Escherichia coli , Klebsiella pneumoniae , and Acinetobacter baumannii , some of them responsible for the mobility of critical antimicrobial resistance genes. These 21 oriT families share key characteristics of the others and fill most of the missing diversity of oriTs in relaxase-encoding plasmids both in terms of frequency and phylogeny. We confirmed experimentally the function of six of them. The ability to identify novel oriT s paves the way to explore conjugation across bacterial plasmids, notably among the majority lacking conjugation-related genes.