Fig 1.
Visualization of the location of transposon insertion sites.
An image of the pst locus of S. meliloti generated using the Integrative Genomics Viewer [78]. Chromosomal nucleotide positions are indicated along the top of the image, and the location and relative abundance of transposon insertions are indicated by the red bars. Non-essential genes contain a high density of transposon insertions, whereas essential genes have few to no transposon insertions. Genes are color coded based on their fitness classification. The pstS, pstC, pstA, pstB, phoU, and phoB genes are co-transcribed as a single operon [92], and previous work demonstrated that non-polar phoU mutations are lethal in S. meliloti, whereas polar mutations are not lethal [93]. The lack of insertions within the phoU coding region is therefore consistent with the non-polar nature of the transposon.
Fig 2.
Characteristics of the core genetic components of S. meliloti.
(A) A circular plot of the S. meliloti chromosome is shown. From the outside to inside: positive strand coding regions, negative strand coding regions (for both positive and negative strands, red indicates the position of the core 489 growth promoting genes), total insertion density, and GC skew. The insertion density displays the total transposon insertions across all experiments over a 10,000-bp window. The GC skew was calculated over a 10,000-bp window, with green showing a positive skew and blue showing a negative skew. Tick marks are every 50,000 bp. (B) A comparison of the overlap between the growth promoting genome (Group I and II genes, shown first) and the essential genome (Group I genes, values in parentheses) of each Tn-seq dataset. Each dataset is labeled with the strain (wild-type or ΔpSymAB) and the growth medium (defined medium or rich medium). (C) Functional enrichment plots for the indicated gene sets. Name abbreviations: Fit–fitness; Dec–decrease; WT–wild-type; ΔAB—ΔpSymAB; Def–defined medium; Rich–rich medium. For example, ‘Fit. dec. WT def > rich' means the genes with a greater fitness decrease in wild-type grown in defined medium compared to rich medium. Legend abbreviations: AA–amino acid; Attach–attachment; Carb–carbohydrate; Cofact–cofactor; e-–electron; Met–metabolism; Misc–miscellaneous; Mot–motility; Nucl–nucleotide; Oxidoreduct–oxidoreductase activity; Prot–protein; Trans–transduction. (D-F) Scatter plots comparing the fitness phenotypes, shown as the log10 of the GEI scores (Gene Essentiality Index scores; i.e., number of insertions within the gene divided by gene length in nucleotides) of (D) wild-type grown in rich medium versus wild-type grown in defined medium, (E) wild-type grown in rich medium versus ΔpSymAB grown in rich medium, and (F) wild-type grown in defined medium versus ΔpSymAB grown in defined medium.
Table 1.
Fitness classification of chromosomal genes.
Genes were ranked from lowest to highest GEI, with the lowest GEI being at the 0 percentile and the highest GEI being at the 100th percentile. The approximate break points for the groupings, determined as described in the Materials and Methods, are shown for each condition. The ranges did not play a role in the classification of the genes, but serve to summarize the number of genes in each category.
Table 2.
Sample genes showing strain specific phenotypes.
The top ten genes from each of the indicated groupings, as determined based on the ratio of GEI (Gene Essentiality Index) scores of the two strains, are shown. GEI scores are shown first for the wild-type (WT) followed by the scores for the ΔpSymAB (dAB) strain.
Fig 3.
Comparison of S. meliloti and R. leguminosarum Tn-seq data.
(A) The fitness phenotypes of essential S. meliloti genes, as determined in this study, are compared to the fitness phenotypes of the orthologous R. leguminosarum genes, as determined by Perry et al. [36]. The data for the S. meliloti copy of the genes are shown in black, while the data for the R. leguminosarum copy of the genes are colored according to their classification by Perry et al. [36]. (B) The fitness phenotypes of essential R. leguminosarum genes is compared to the fitness phenotypes of the orthologous S. meliloti genes. The data for the R. leguminosarum copy of the genes are shown in black, while the data for the S. meliloti copy of the genes are colored according to their classification in this study. (A,B) Normalized fitness values are used to facilitate direct comparison between the studies as different output statistics were calculated. For S. meliloti, the GEI score of each gene for wild-type cells grown in minimal medium broth was divided by the median GEI for all genes under the same conditions. For R. leguminosarum, the insertion density of each gene during growth on minimal medium plates was divided by the median insertion density of all genes. The data underlying this figure and the corresponding gene names are provided in S3 Dataset.
Fig 4.
In silico analysis of genetic redundancy in S. meliloti.
The effects of single or double gene deletion mutants were predicted in silico with the genome-scale S. meliloti metabolic model. All data in this figure comes from in silico metabolic modeling. (A-C) The color of each hexagon is representative of the number of genes, gene pairs, or reaction pairs plotted at that location according to the density bar below each panel. The diagonal line serves as a reference line of a perfect correlation (i.e., where all data would fall if no effects were observed). The data underlying this figure and the corresponding gene/reaction names are provided in S7 Dataset. (A) A scatter plot comparing the grRatio (growth rate of mutant / growth rate of non-mutant) for gene deletion mutations in the presence (wild-type) versus absence (ΔpSymAB) of the pSymA/pSymB model genes. Genes whose deletion had either no effect or were lethal in both cases are not included in the plot. (B) A scatter plot comparing the grRatio for each double gene deletion pair (where one gene was on the chromosome and the other on pSymA or pSymB) observed in silico versus the predicted grRatio based on the grRatio of the single deletions (grRatio1 * grRatio2). Only gene pairs with an observed grRatio at least 10% less than the expected are shown. (C) A scatter plot comparing the grRatio for each double gene deletion pair (both genes on the chromosome) observed in silico versus the predicted grRatio. Only gene pairs with an observed grRatio at least 10% less than the expected are shown.
Fig 5.
Summary schematic of core S. meliloti metabolism.
The iGD726 core metabolic model was visualized using the iPath v2.0 webserver [83], which maps the reactions of the metabolic model to KEGG metabolic pathways; it therefore does not capture metabolism not present in the KEGG pathways included in iPath. Reactions and metabolites are color coded according to their biological role, as indicated. Reactions whose associated genes were not identified as growth promoting in this study are in dashed lines.
Table 3.
The last column indicates reactions whose gene associations are supported by the Tn-seq data of this study. Percentage of all reactions in that category are indicated in brackets.
Fig 6.
Gene essentiality index (GEI) changes for genes of selected biological pathways.
Each data point represents an individual gene, and shows the log10 of the ratio of the GEI for that gene in the ΔpSymAB background compared to the wild-type background. Lines indicate the median value of all genes included from the biological process. The underlying data is given in S9 Table. Genes included in each process are as follows: Cytochrome C oxidase related genes–ctaB, ctaC, ctaD, ctaE, ctaG, ccsA, cycH, cycJ, cycK, cycL, ccmA, ccmB, ccmC, ccmD, ccmG; Proline biosynthesis–proA, proB1, proC; Histidine biosynthesis–hisB, hisD, smc04042; Glycolysis and related genes–glk, frk, pgi, zwf, pgl, edd, eda2, gap, pgk, gpmA, eno, pykA, pyc; Periplamic cyclic β-glucan biosynthesis–feuN, feuP, feuQ, ndvA, ndvB; Arginine biosynthesis–argB, argC, argD, argF1, argG, argH1, argJ; AICAR biosynthesis–purB, purC, purD, purE, purF, purH, purK, purL, purM, purN, purQ, smc00494; UMP biosynthesis–carA, carB, pyrB, pyrC, pyrD, pyrE, pyrF, smc01361; LPS core oligosaccharide biosynthesis–lpsC, lpsD, lpsE.