Translational demand is not a major source of plasmid-associated fitness costs

Abstract

Plasmids are key drivers of bacterial evolution because they are crucial agents for the horizontal transfer of adaptive traits, such as antibiotic resistance. Most plasmids entail a metabolic burden that reduces the fitness of their host if there is no selection for plasmid-encoded genes. It has been hypothesized that the translational demand imposed by plasmid-encoded genes is a major mechanism driving the fitness cost of plasmids. Plasmid-encoded genes typically present a different codon usage from host chromosomal genes. As a consequence, the translation of plasmid-encoded genes might sequestrate ribosomes on plasmid transcripts, overwhelming the translation machinery of the cell. However, the pervasiveness and origins of the translation-derived costs of plasmids are yet to be assessed. Here, we systematically altered translation efficiency in the host cell to disentangle the fitness effects produced by six natural antibiotic resistance plasmids. We show that limiting translation efficiency either by reducing the number of available ribosomes or their processivity does not increase plasmid costs. Overall, our results suggest that ribosomal paucity is not a major contributor to plasmid fitness costs. This article is part of the theme issue 'The secret lives of microbial mobile genetic elements'.

Keywords: codon usage; fitness cost; horizontal gene transfer; plasmid; ribosome; translational demand.

Figure 1.

Plasmids and their fitness effects. (a) Diagram depicting the six plasmids used in this study, showing the open reading frames (ORFs) as arrows (outer track), with the arrowhead indicating the direction of transcription, and coloured according to their function (see legend). The inner track represents the codon adaptation index (CAI [23]) for each ORF (see legend for colour reference). pOXA-48_K8 is a pOXA-48-like plasmid, and for simplicity, we refer to this plasmid as pOXA-48 throughout the study. (b) Fitness effects associated with plasmid acquisition in E. coli MG1655. Each bar corresponds to the median value of six replicates (36 for MG1655) with each point representing an independent biological replicate. Numbers within each bar show the cost of each plasmid (% reduction in fitness relative to plasmid-free MG1655).

Figure 2.

Reducing ribosomal elongation rates does not increase plasmid costs. (a) Slow ribosomal elongation rates caused by rpsL mutations lead to an increased number of transcript-bound ribosomes, thereby depleting the pool of available ribosomes and consequently reducing translation rate. This is illustrated by comparing translation in wild-type cells (above) and rpsL mutants (below). Note that the total number of ribosomes is conserved between both panels. (b) After 2 h of induction with 0.1% l-arabinose, strains with rpsL mutations show lower fluorescence levels (in arbitrary units; arb. units) than MG1655 (MG) as measured by flow cytometry, indicating a reduced translation capability. For reference, an uninduced MG1655 control (C−) is included. (c) Fitness of strains carrying rpsL mutations relative to MG1655. Each bar corresponds to the median value of 36 replicates, with each point representing an independent biological replicate. Error bars depict standard deviation. Numbers within each bar show the cost of each mutation (% reduction in relative fitness relative to plasmid-free MG1655). (d) Fitness of the plasmid–rpsL mutant strain combinations relative to the fitness of plasmid-carrying MG1655. The height of the bar represents the median fitness of each plasmid–rpsL mutant strain combination relative to MG1655 carrying each of the plasmids. Red diamonds represent median fitness values of rpsL mutants relative to MG1655 (as in figure 2c). Therefore, bars below their respective diamond indicate greater plasmid costs (i.e. lower fitness), whereas those above the diamond show lower plasmid-associated costs than those found in MG1655. Error bars represent standard deviation and each point represents an independent biological replicate (n = 6). (e) Relationship between expected fitness calculated as the product of the fitness effects of each plasmid (Wp) and mutation (WrpsL) separately and the fitness measured for each plasmid–mutant combination in competition experiments. Data points above the grey dotted line indicate positive epistasis, whereas those points below show negative epistasis (i.e. increased plasmid-associated costs). Genotypes are depicted using different symbols (see legend). The blue line shows linear regression of the data with 95% confidence intervals (grey shading).

Figure 3.

Reducing ribosomal availability does not increase plasmid costs. (a) Reduced ribosomal availability depletes the pool of available ribosomes, consequently reducing translation rate. This is illustrated by comparing translation in wild-type cells (above) and rrn deletion strains (below). (b) After 2 h of induction with 0.1% l-arabinose, strains with reduced ribosomal availability show lower fluorescence levels (in arbitrary units; arb. units) than MG1655 (MG) as measured by flow cytometry, indicating a reduced translation capability. For reference, an uninduced MG1655 control (C−) is included. (c) Fitness of strains carrying rrn deletions relative to MG1655. Each bar corresponds to the median value of six replicates, with each point representing an independent biological replicate. Error bars depict standard deviation. Numbers within each bar show the cost of each deletion (% reduction in relative fitness relative to plasmid-free MG1655). (d) Fitness of the plasmid–rrn deletion strain combinations relative to the fitness of plasmid-carrying MG1655. The height of the bar represents the median fitness of each plasmid–rrn deletion strain combination relative to MG1655 carrying each of the plasmids. Red diamonds represent median fitness values of rrn deletion strains relative to MG1655 (as in figure 3c). Therefore, bars below their respective diamond indicate greater plasmid costs (i.e. lower fitness), whereas those above the diamond show lower plasmid-associated costs than those found in MG1655. Error bars represent standard deviation and each point represents an independent biological replicate (n = 6). (e) Relationship between expected fitness calculated as the product of the fitness effects of each plasmid (Wp) and mutation (Wrrn) separately and the fitness measured for each plasmid–mutant combination in competition experiments. Data points above the grey dotted lines indicate positive epistasis, whereas those below show negative epistasis (i.e. increased plasmid-associated costs). Genotypes are depicted using different symbols (see legend). The blue lines show linear regression of the data with 95% confidence intervals (grey shading).