An improved estimation of tRNA expression to better elucidate the coevolution between tRNA abundance and codon usage in bacteria

Abstract

The degree to which codon usage can be explained by tRNA abundance in bacterial species is often inadequate, partly because differential tRNA abundance is often approximated by tRNA copy numbers. To better understand the coevolution between tRNA abundance and codon usage, we provide a better estimate of tRNA abundance by profiling tRNA mapped reads (tRNA tpm) using publicly available RNA Sequencing data. To emphasize the feasibility of our approach, we demonstrate that tRNA tpm is consistent with tRNA abundances derived from RNA fingerprinting experiments in Escherichia coli, Bacillus subtilis, and Salmonella enterica. Furthermore, we do not observe an appreciable reduction in tRNA sequencing efficiency due to post-transcriptional methylations in the seven bacteria studied. To determine optimal codons, we calculate codon usage in highly and lowly expressed genes determined by protein per transcript. We found that tRNA tpm is sensitive to identify more translationally optimal codons than gene copy number and early tRNA fingerprinting abundances. Additionally, tRNA tpm improves the predictive power of tRNA adaptation index over codon preference. Our results suggest that dependence of codon usage on tRNA availability is not always associated with species growth-rate. Conversely, tRNA availability is better optimized to codon usage in fast-growing than slow-growing species.

Figure 1

Comparison between tRNA tpm and fingerprinting abundance in E. coli, S. enterica, and B. subtilis. In panels (a) the averaged tRNA abundances across five growth phases retrieved from Dong, et al., (b,c) are tRNA abundances retrieved from Ikemura, and in (c) the tRNA abundances retrieved from Kanaya, et al..

Figure 2

RNA-Seq read map for all 50 unique tRNA sequences in E. coli, split in three sets (a–c). Each line represents the read depth of entire sequence region of one unique tRNA sequence, with sites susceptible to methylation (m²G18, m²C32 or m²U32, m⁵U34 or m²C34 or cmo⁵U34, m⁶A37, m⁷G46 and m⁵U54) highlighted red. In (d) the distribution of mapped reads across the entire length of each unique tRNA sequence. Red indicates tRNAs that are potentially methylated at >4 sites (heavily methylated) and all others are highlighted blue. In (e) the distribution of total tRNA tpm in sets of heavily methylated and other tRNAs.

Figure 3

Relationship between (a) tRNA tpm and averaged tRNA abundance from RNA fingerprinting (tRNA abundance) in E. coli, and (b) RTUs and RSCU in E. coli HEGs.