Metabolic redundancy conferred by extrachromosomal replicons

diCenzo et al. 2018. Robustness encoded across essential and accessory replicons of the ecologically versatile bacterium Sinorhizobium meliloti. PLoS Genetics.

This is a summary of our DalMUG journal club discussion of the above paper (when it was online as an uncorrected proof) written by Gavin Douglas.


Several papers have recently demonstrated that knocking out the same gene within different strains of a species (and different species in a genus) can have drastically different fitness consequences. These observations have highlighted that the genomic background in which a gene is expressed partially determines whether a gene is essential or not. However, it’s still unclear how large-scale shifts in the genomic background, such as through gain of a plasmid, affects gene essentiality. Enter: di Cenzo et al. who investigated how the ancient acquirement of replicons (large plasmids) in the common soil bacteria Sinorhizobium meliloti affected gene essentiality. These two replicons, pSymA and pSymA, encode ~45% of the species’ genes. There are only 2 known essential genes on pSymB and otherwise the replicon-encoded genes are thought to be non-essential. The authors used Tn-seq to identify essential and growth-promoting genes while changing two variables: (1) whether the replicons are present or not and (2) whether the bacteria were grown in rich or defined media. They found that the replicons confer a substantial amount of redundancy in essential genes found in the genome. They also highlight insights that would not have been possible without also performing in silico metabolic reconstruction.

Points of Interest

  • Interesting to know background on the relatively low percentages of essential genes that overlap across species/strains
  • They created a transposon with constitutive promoters on both ends of the construct to avoid issues where non-essential genes downstream of an essential gene in an operon are falsely identified as essential
  • Some bias of the transposon to insert in GC-rich regions
  • Identified 307 genes as essential independent of growth medium / strain
  • Tested 4 genes predicted to produce no growth and they all agreed with the Tn-seq data, which was one example of a number of reassuring sanity checks
  • Interesting that some fitness benefits of losing replicons could be due to having less DNA to synthesize
  • Compared their results to Rhizobium leguminosarum, for which there is also a published Tn-seq dataset and found little overlap in their results overall
  • Based on in silico double KOs of a gene from chromosome and a gene from replicons they found 14% of chromosomal genes had predicted negative phenotype, which indicates high metabolic redundancy between chromosome and replicons
  • 1/3 of reactions in reconstructed core metabolism were identified as not having an effect on growth based on Tn-seq
  • They used the uncharacterized essential genes identified through Tn-seq to propose which genes were responsible for the predicted reactions (of unknown enzymes) in their reconstruction of core metabolism

Points of Confusion

  • The authors acknowledge that there could be problems with the constitutive promoters on the transposons, but it would be good to know what kind of off-target effects this would be expected to have
  • Gene essentiality index looks like it is just (# insertions / gene length) - maybe including the possible number of insertion sites per gene would improve this measure?
  • Of course in different environments (besides the two studied here) the replicons could become more or less important (which is really hard to assess)
  • Would have been nice to compare the individual KOs of each replicons, especially since they have such different evolutionary histories
Written on April 26, 2018