Evolution of satellite plasmids can prolong the maintenance of newly acquired accessory genes in bacteria

Abstract

Transmissible plasmids spread genes encoding antibiotic resistance and other traits to new bacterial species. Here we report that laboratory populations of Escherichia coli with a newly acquired IncQ plasmid often evolve 'satellite plasmids' with deletions of accessory genes and genes required for plasmid replication. Satellite plasmids are molecular parasites: their presence reduces the copy number of the full-length plasmid on which they rely for their continued replication. Cells with satellite plasmids gain an immediate fitness advantage from reducing burdensome expression of accessory genes. Yet, they maintain copies of these genes and the complete plasmid, which potentially enables them to benefit from and transmit the traits they encode in the future. Evolution of satellite plasmids is transient. Cells that entirely lose accessory gene function or plasmid mobility dominate in the long run. Satellite plasmids also evolve in Snodgrassella alvi colonizing the honey bee gut, suggesting that this mechanism may broadly contribute to the importance of IncQ plasmids as agents of bacterial gene transfer in nature.

Fig. 1. Satellite plasmids evolve from an IncQ plasmid in E. coli.

a Map of the pQGS plasmid used in evolution experiments. The aadA gene for Spec^R was used to select for plasmid maintenance. Constitutively expressed gfp serves as a trackable proxy for a costly accessory gene. Primers used to generate PCR amplicons from pQGS for analysis, AvaI restriction sites (blue), and the fragments generated by digesting the linearized plasmid with AvaI (F1a, F2a, F2, and F3) are shown above the map. Plasmid regions preserved in SPs in evolved strains B253, B353, and B653 are pictured below. b Flow cytometry comparing the distributions of fluorescence in cell populations in all six experimental populations of E. coli after 1 and 5 days of evolution. c Copy number of the replication origin (oriV) and GFP gene (gfp) on the pQGS plasmid relative to a single-copy chromosomal reference gene (dapA) determined using quantitative PCR. The ancestor strain (Anc) and evolved strains from day 5 of the evolution experiment were tested. Error bars are standard deviations for three biological replicates except for ancestor and B454, which have six replicate measurements each, and B554, which has two. P values are for two-tailed t tests for rejecting the null hypothesis that oriV and gfp copy number are equal in a strain. d Linearized plasmid PCR amplicons from the ancestor strain (Anc) and three evolved strains that had significantly different oriV and gfp copy numbers. e DNA fragments produced by an AvaI restriction digest of the samples shown in d. The expected fragments are labeled in a. The smallest fragment (F1a) does not appear on the gel. For each SP fragment, F3 is missing and F2 is truncated to an extent that is consistent with identification of the precise boundaries of the deleted regions in each SP by Sanger sequencing. Source data are provided as a Source Data file.

Fig. 2. Time courses of plasmid evolution in E. coli.

a, b Appearance and persistence in populations B2 and B6 of satellite plasmids and other evolved plasmids with deletions that also change the size of the linearized plasmid PCR amplicon. Compared to the methods shown in other panels, the PCR assay is biased toward detecting satellite plasmids versus other plasmid fates even when they are very rare in a population because they produce smaller amplicons. c, d Distributions of GFP expression in cells in each population determined using flow cytometry. e, f Estimates of the copy numbers of different plasmid regions and the relative frequencies of different evolved plasmids’ fates in each population from Illumina sequencing data. Coverage of the lacI gene was used to monitor the copy number of intact gfp gene copies because lacI is completely deleted in all evolved plasmids, but gfp is only partially deleted in some plasmids. The x-axes are categorical rather than linear in c–f, with even spacing of all samples that were analyzed from different days. g Schematic showing different evolved fates of the pQGS plasmid. Bold lines in blue, red, and dark green show the locations of sequencing reads unique to integrated, deletion, and satellite plasmids, respectively, which were used to estimate their relative frequencies. h Relative fitness of strains reconstructed with different evolved plasmid variants determined using co-culture competition assays versus a control strain with no plasmid. Error bars are 95% confidence intervals based on 15–18 replicate measurements for each strain. Source data are provided as a Source Data file.

Fig. 3. Satellite plasmids evolve in S. alvi in the honey bee gut.

a SPs detected as linearized PCR amplicons during the in vitro propagation of S. alvi at days 8 and 12 in five experimental populations S1–S5. b Fluorescence micrograph showing colonization of the bee gut ileum 24 h after inoculation with S. alvi carrying plasmid pQGS. c Satellite plasmids detected as linearized PCR amplicons from DNA extracts of the guts of individual bees reared in different enclosures (C1–C10) at 72 or 96 h after inoculation. Full results for all populations evolved in vivo are shown in Supplementary Fig. 4. Blue arrows in all panels are bands that were Sanger sequenced to validate that they represent SPs (Supplementary Fig. 3 and Supplementary Table 1). Source data are provided as a Source Data file.

Fig. 4. Satellite plasmid evolution is favored over deletion plasmid evolution.

a pQGS plasmid map showing regions containing the endpoints of deletions leading to satellite plasmids (blue thin line) and accessory gene deletion plasmids (red thin line). The regions required to be contained in each deletion are shown with bold lines in the respective colors. b Microhomology regions like those observed flanking satellite plasmid deletions are more common in the satellite plasmid deletion endpoint regions than in the accessory gene deletion plasmid endpoint regions, but the difference is no greater than the fourfold difference expected if these sites are randomly distributed. c Multilevel stochastic simulations predict an early advantage for cells with newly evolved satellite plasmids due to a reduced phenotypic lag. Their progeny realize more of their full fitness increase earlier, compared to cells that evolve deletion plasmids. Shaded areas are 95% confidence intervals from 200 simulation trajectories. Source data are provided as a Source Data file.

Fig. 5. Evolutionary pathways after plasmid acquisition.

This model summarizes various evolutionary fates of IncQ plasmids observed in this study and how they may relate to the maintenance of accessory genes on these plasmids in nature. After a new cell acquires an IncQ plasmid, the evolution of nonautonomous satellite plasmids that are dependent on a full-length plasmid in the same cell for replication can alleviate the fitness burden of accessory genes encoded on the plasmid by reducing their copy number (Sat). These cells can still potentially benefit from the accessory genes if the environment changes (e.g., if they provide antibiotic resistance and there is treatment), and full-length plasmids in these cells remain capable of being transferred by conjugation to other cells to spread these accessory genes. Formation of satellite plasmids occurs at a high rate, but this state is likely to be a transient evolutionary intermediate. In experimental populations, cells with satellite plasmids are later displaced by cells that have integrated plasmid sequences into the chromosome via a mechanism involving an insertion sequence (Int) or the evolution of autonomous plasmids that delete just the accessory genes (Del). In the current study, cells had to maintain a gene from the plasmid that is encoded adjacent to the replication genes for survival (which is why no arrows are shown to the “no plasmid” state). Even without this constraint, the ability to evolve satellite plasmids is expected to give more opportunities for preserving the functions of accessory genes encoded on IncQ plasmids as they colonize new bacterial cells and hosts. The fitness and accessory gene copy number scales are not meant to be quantitative; they show only the approximate relative values of these parameters. The sizes of the black arrows for transitions are roughly weighted by the relative rates inferred for these processes from the evolution experiments.