The birth of a bacterial tRNA gene by large-scale, tandem duplication events

A list of 62 theoretically possible tRNA types (61 elongator tRNAs and one initiator tRNA) is provided. Of these, 23 are absent from the SBW25 tRNA gene set. Codon-tRNA matching patterns are based on the G-U wobble rule (Crick, 1966), and post-transcriptional modifications predicted using a combination of MODOMICS (Boccaletto et al., 2018), tRNAmod (Panwar and Raghava, 2014), and tRNAmodpred (Machnicka et al., 2016). Following current wobble rules (described in Agris et al., 2018) and the predicted post-transcriptional modifications, the 39 tRNA types present in SBW25 are split into 33 theoretically essential types (pink highlighting) and six theoretically non-essential types (green highlighting). Gene copy numbers and genome-wide proportions of codon use (from GtRNAdb; Chan and Lowe, 2016) are listed for each tRNA type present. Further notes are provided where appropriate.

Whole genome sequence data (Illumina NextSeq, 150 bp, paired-end reads; SRA accession number: PRJNA558233) were obtained for seven strains from day 13 of the evolution experiment: W1-L, W3-L (each derived from two independent SBW25 control lineages), M1-L, M2-L, M2-Lop, M3-L, and M4-L (each derived from four independent serCGA deletion lineages). A minimum of 4.5 million raw reads was obtained for each sample, and these were aligned to the P. fluorescens SBW25 genome sequence (Silby et al., 2009) and subsequently analysed using a combination of breseq (Barrick et al., 2014) and Geneious. This file provides a summary and a full report of the mutations predicted in each isolate. In addition to the large tandem duplications detailed in Table 1, two unique point mutations were identified: W3-L carries a non-synonymous point mutation in ytfH (encoding a probable transcriptional regulator), and M2-L carries a synonymous point mutation in edd (encoding phosphogluconate dehydratase). There is evidence for some non-ubiquitous deletions in repetitive, intergenic, DNA elements in M1-L, M2-L, M2-Lop, and M4-L. These deletions affect areas outside of the large duplication regions. Some putative mutations were identified in all (or most) isolates, many of which have also been identified in unrelated experiments (i.e., are likely to be present in the starting strain; Gallie et al., 2015, Gallie et al., 2019). As such, these putative mutations are not expected to be relevant for the fitness effects described in this manuscript.

The spreadsheet lists the SBW25 gene annotations from NCBI (6176 genes; left of the spreadsheet), followed by the details of which genes are duplicated in M1-L, M2-L, M2-Lop, M3-L, and M4-L. Note that for comparison purposes, the duplication details for each strain are provided on the same line numbers as the SBW25 list (scroll down until the first duplicated genes is visible in each duplication isolate). The core set of 45 genes that is duplicated in each of the five isolates is highlighted in pink.

This file contains details of the raw read numbers of proportions of the dominant base called at SBW25 chromosome positions 4,163,616–4,163,861 in strains W1-L, W3-L (each carrying one serTGA copy), M1-L, M2-L, M2-Lop, M3-L, and M4-L (each carrying two serTGA copies). This 245 bp segment encompasses the serTGA gene and ~155 bp of the upstream region, which is expected to contain the serTGA promoter. Any point mutation in either serTGA copy in the duplication strains (M1-L, M2-L, M2-Lop, M3-L, and M4-L) is expected to be reflected by a drop in the proportion of the dominant base (to around 0.5). For example, if in M1-L one serTGA copy had gained a C→T point mutation in tRNA position 34 (i.e., an anticodon switch event; see Discussion), one would expect approximately half the 214 reads covering base 4,163,669 to contain a T, and the other half to carry a C. Therefore, the dominant base proportion would be expected to drop to ~0.5. No evidence was found of any mutations in any copy of serTGA or its promoter; all 245 bp were covered by a minimum of 127 reads in each duplication strain, with at least 94% of reads at each position carrying the dominant base.

These include 42 unique primary tRNA sequences, each of which is listed in this file. These 42 sequences are used as references to align the YAMAT-seq data in this work (see Supplementary file 7). Note that the list of 42 sequences includes one likely pseudo tRNA (11_Cys-GCA-2–1). This sequence is not predicted to form a tRNA with conserved cloverleaf secondary structure (Chan and Lowe, 2019). Further, no YAMAT-seq sequences were aligned to this reference sequence in any sample. We conclude that Cys-GCA-2–1 does not form part of the functional mature tRNA pool in SBW25. In addition, the serCGA sequence (30_Ser-CGA-1–1) is expected to be absent from 21 of 27 samples (samples 2–8, 11–17, and 20–26); this tRNA is encoded by serCGA (the gene that was deleted by genetic engineering in this work and remains absent in all derived strains). As expected, almost no reads were obtained for this reference sequence in these 21 samples. The very low numbers of Ser-CGA reads obtained in some of these samples (e.g., two reads in sample 20, the third replicate of ΔserCGA-1) are likely to be barcode misallocations from one of the six SBW25 or W1-L samples.

The first tab contains index details and a summary of the raw YAMAT-seq reads (GEO accession number GSE144791; Edgar et al., 2002) for each of the 27 samples (three replicates of nine strains): a minimum of 636,995 reads of the expected size (80–151 bp) was obtained per sample. In each case, between 86.4% and 93.0% of these aligned to the list of 42x reference SBW25 tRNA sequences (provided in Supplementary file 6). The subsequent nine tabs contain the YAMAT-seq data for each of the nine strains tested (SBW25, ΔserCGA-1, ΔserCGA-2, W1-L, M1-L, M2-L, M2-Lop, M3-L, and M4-L). Each tab contains (i) the numbers of reads for 42 reference tRNAs, for three replicates (left), (ii) numbers of reference reads for 39 tRNA types in SBW25 (e.g., tRNA-Asn-GTT is the sum of reference sequences 7_Asn-GTT-1–1 and 8_Asn-GTT-2–1; middle), (iii) the proportion of each tRNA type in the mature tRNA pool for each of the three samples (right), and (iv) a scatter plot of the YAMAT-seq proportions versus the proportion of the tRNA gene set encoding the tRNA type. Blue = tRNA-Ser(CGA), green = tRNA-Ser(UGA). The final tab contains information regarding the unused reads for each sample.