Population genomics of Australian indigenous Mesorhizobium reveals diverse nonsymbiotic genospecies capable of nitrogen-fixing symbioses following horizontal gene transfer

Abstract

Mesorhizobia are soil bacteria that establish nitrogen-fixing symbioses with various legumes. Novel symbiotic mesorhizobia frequently evolve following horizontal transfer of symbiosis-gene-carrying integrative and conjugative elements (ICESyms) to indigenous mesorhizobia in soils. Evolved symbionts exhibit a wide range in symbiotic effectiveness, with some fixing nitrogen poorly or not at all. Little is known about the genetic diversity and symbiotic potential of indigenous soil mesorhizobia prior to ICESym acquisition. Here we sequenced genomes of 144 Mesorhizobium spp. strains cultured directly from cultivated and uncultivated Australian soils. Of these, 126 lacked symbiosis genes. The only isolated symbiotic strains were either exotic strains used previously as legume inoculants, or indigenous mesorhizobia that had acquired exotic ICESyms. No native symbiotic strains were identified. Indigenous nonsymbiotic strains formed 22 genospecies with phylogenomic diversity overlapping the diversity of internationally isolated symbiotic Mesorhizobium spp. The genomes of indigenous mesorhizobia exhibited no evidence of prior involvement in nitrogen-fixing symbiosis, yet their core genomes were similar to symbiotic strains and they generally lacked genes for synthesis of biotin, nicotinate and thiamine. Genomes of nonsymbiotic mesorhizobia harboured similar mobile elements to those of symbiotic mesorhizobia, including ICESym-like elements carrying aforementioned vitamin-synthesis genes but lacking symbiosis genes. Diverse indigenous isolates receiving ICESyms through horizontal gene transfer formed effective symbioses with Lotus and Biserrula legumes, indicating most nonsymbiotic mesorhizobia have an innate capacity for nitrogen-fixing symbiosis following ICESym acquisition. Non-fixing ICESym-harbouring strains were isolated sporadically within species alongside effective symbionts, indicating chromosomal lineage does not predict symbiotic potential. Our observations suggest previously observed genomic diversity amongst symbiotic Mesorhizobium spp. represents a fraction of the extant diversity of nonsymbiotic strains. The overlapping phylogeny of symbiotic and nonsymbiotic clades suggests major clades of Mesorhizobium diverged prior to introduction of symbiosis genes and therefore chromosomal genes involved in symbiosis have evolved largely independent of nitrogen-fixing symbiosis.

Keywords: ICE; Mesorhizobium; conjugation; evolution; horizontal gene transfer; integrative and conjugative elements; nitrogen fixation; plant-microbe interactions; rhizosphere; soil bacteria; symbiosis; symbiosis island.

Fig. 1.

Map of soil sites. (a) Sampling sites across western Australia. (b),(c) Location of the sampling sites in Badgingarra. Soils were analysed for pH, and content of NH₄ ⁺ and organic C, results are indicated by squares, diamonds or triangles coloured as indicated in the key on the right. The chickpea seed and the ear of wheat cartoons in (c) indicate that the paddocks were cultivated with these crops at the time of sampling.

Fig. 2.

Overview of the Mesorhizobium genus. Maximum-likelihood tree constructed on 436 single-copy core genes with RAxML. The tree was rooted with the type-strain Rhizobium leguminosarum biovar viciae USDA2370 (MRDL00000000). Strains were predicted to be symbiotic if they harboured NodABC and NifHDK genes and/or if they had published symbiotic capacity, and were assumed nonsymbiotic if they lacked these genes or demonstrated symbiotic capacity. Unfilled circles at tip nodes indicate NCBI-downloaded reference genomes (type strains are indicated by ^T), and black-filled circles at tip nodes indicate Mesorhizobium isolated in this study. Scale bar indicates substitutions per site. The histogram in the upper-left displays the distribution of core-proteome amino-acid identity (cpAAI) (bin width 0.5 %) calculated from the same 436 core genes used in tree generation, and the dotted line indicates species within the cpAAI>86 %.

Fig. 3.

Core-gene phylogeny of Mesorhizobium isolated from soil during this study. In total, 1,067 single-copy core-genes were used to build a maximum-likelihood tree using RAxML. The tree was rooted with the M. sp. HJP. Coloured tip nodes indicate the strains were isolated in this study, with colour of the tip nodes representing the species the strain belongs to. The genospecies name attributed by the Genome Taxonomy Database is reported in brackets. Coloured tip nodes with a black border indicate that the strain carries an ICESym of a commercial inoculum, the name of the ICESym is indicated adjacent to the tip node. Other genomes were downloaded from NCBI, grey triangles indicate a collapsed clade and black tip nodes highlight selected type strains. A complete tree with all strain labels is presented in Fig. S1. Scale bar indicates substitutions per site.

Fig. 4.

Legume symbioses with NS-meso exconjugants carrying ICESyms. (a) B. pelecinus cv. Casbah dry foliage weight at 69 days post-inoculation. Uninoculated (N-) and N-fed (N+) (supplied as KNO₃) plants were included as negative and positive controls, respectively. Treatments are shown with standard errors of the means, and the treatments that share a letter are not significantly different according to the least significant difference test (P ≤ 0.05). (b) Number of nodules developed by the L. japonicus Gifu plants grown in test tubes. (c) L. japonicus Gifu foliage weight at 30 days post-inoculation. The experiment was performed with 15 replicates and repeated three times. Treatments are shown with standard errors of the means, and treatments that share a letter are not significantly different according to the least significant difference test (P ≤ 0.01).

Fig. 5.

ICEMsp^B4-1-4, a tripartite ICE lacking symbiosis genes. Gene annotations are colour coded as follows: black, ICE transfer genes; purple, quorum-sensing genes; brown, msi287; green, vitamin biosynthesis; blue, mobile genes such as transposases, integrases, and recombinases; azure, genes likely involved in the broad-specificity catabolism of phosphonates. An asterisk (*) indicates frameshifting required for gene to be translated.

Fig. 6.

ICEMsp^B4-1-4 shares a common ancestor with the archetypal ICESym. (a) Maximum-likelihood tree of ICESym clade ICEs. The tree was constructed with RAxML and it is based on the concatenation of alignments of 16 single-copy backbone genes, the tree was rooted with ICEMc ^WSM1497, scale bar indicates substitutions per site. Light green indicates the ICESym specifies symbiosis with Lotus, blue with Biserrula, pink with Cicer, grey with an unknown plant. The tripartite ICEMsp^B4-1-4 is highlighted in azure. Triangles indicate tripartite ICEs. (b) Maximum-likelihood tree of integrases specifying integration in guaA (intG), met-tRNA (intM) and phe-tRNA (intS). The tree was constructed with RAxML and rooted at midpoint, the scale bar indicates substitutions per site. Dark green highlight indicates integrases of ICESyms, and triangles indicate the integrase belongs to a tripartite ICE. The integrases of the tripartite ICEMsp^B4-1-4 are highlighted in azure.