Subspecies-level genome comparison of Lactobacillus delbrueckii

Abstract

Lactobacillus delbrueckii comprises six subspecies, L. delbrueckii subsp. bulgaricus, L. delbrueckii subsp. lactis, L. delbrueckii subsp. jakobsenii, L. delbrueckii subsp. delbrueckii, L. delbrueckii subsp. sunkii, and L. delbrueckii subsp. indicus. We investigated the evolution of the six subspecies of L. delbrueckii using comparative genomics. While the defining feature of the species was the gene number increment driven by mobile elements and gene fragmentation, the repertoire of subspecies-specific gene gains and losses differed among the six subspecies. The horizontal gene transfer analyses indicated that frequent gene transfers between different subspecies had occurred when the six subspecies first diverged from the common ancestor, but recent gene exchange was confined to a subspecies implying independent evolution of the six subspecies. The subspecies bulgaricus is a homogeneous group that diverged from the other subspecies a long time ago and underwent convergent evolution. The subspecies lactis, jakobsenii, delbrueckii, and sunkii were more closely related to each other than to other subspecies. The four subspecies commonly show increasing genetic variability with increasing genome size. However, the four subspecies were distinguished by specific gene contents. The subspecies indicus forms a branch distant from the other subspecies and shows an independent evolutionary trend. These results could explain the differences in the habitat and nutritional requirements of the subspecies of L. delbrueckii.

Figure 1

Genomic distance and relationship among L. delbrueckii subspecies. (a) Maximum-likelihood tree constructed using amino acid sequences of 689 core genes. The tree was rooted using the outgroups Lactobacillus amylolyticus L6 and Lactobacillus acetotolerans NBRC 13,120. The scale represents the number of substitutions per site. The bootstrap values were 100% in all nodes. (b) Average nucleotide identity (ANI), demonstrating the genomic distance among L. delbrueckii subspecies. The ANI distance was plotted as a heatmap.

Figure 2

Analysis of ancestral genes and minimal gene gain/loss among the complete genome sequences. Black numbers adjacent to the internal nodes indicate the number of estimated ancestral genes. Blue numbers on the branches denote the minimum number of gains and losses under the best fit model. ‘Lac + ’ and ‘Lac−’ indicate the presence and absence of lactose fermentation capability.

Figure 3

Network tree showing the HGT events among L. delbrueckii subspecies. The network tree of complete genomes was generated using the neighbor-net algorithm based on a binary matrix of the presence or absence of gene families generated by the OrthoMCL program. Splits in the tree show the possibility of non-vertical evolution between branches.