Comparative genomic and functional analysis of 100 Lactobacillus rhamnosus strains and their comparison with strain GG

Abstract

Lactobacillus rhamnosus is a lactic acid bacterium that is found in a large variety of ecological habitats, including artisanal and industrial dairy products, the oral cavity, intestinal tract or vagina. To gain insights into the genetic complexity and ecological versatility of the species L. rhamnosus, we examined the genomes and phenotypes of 100 L. rhamnosus strains isolated from diverse sources. The genomes of 100 L. rhamnosus strains were mapped onto the L. rhamnosus GG reference genome. These strains were phenotypically characterized for a wide range of metabolic, antagonistic, signalling and functional properties. Phylogenomic analysis showed multiple groupings of the species that could partly be associated with their ecological niches. We identified 17 highly variable regions that encode functions related to lifestyle, i.e. carbohydrate transport and metabolism, production of mucus-binding pili, bile salt resistance, prophages and CRISPR adaptive immunity. Integration of the phenotypic and genomic data revealed that some L. rhamnosus strains possibly resided in multiple niches, illustrating the dynamics of bacterial habitats. The present study showed two distinctive geno-phenotypes in the L. rhamnosus species. The geno-phenotype A suggests an adaptation to stable nutrient-rich niches, i.e. milk-derivative products, reflected by the alteration or loss of biological functions associated with antimicrobial activity spectrum, stress resistance, adaptability and fitness to a distinctive range of habitats. In contrast, the geno-phenotype B displays adequate traits to a variable environment, such as the intestinal tract, in terms of nutrient resources, bacterial population density and host effects.

Figure 1. Analysis of genome diversity in L. rhamnosus by mapped SOLiD sequencing.

The 100 L. rhamnosus strains were clustered using hierarchical clustering based on their relative shared gene content with L. rhamnosus GG. Strain names were colour-coded as follows: green for dairy isolates, purple for intestinal isolates, orange for oral isolates, magenta for vaginal isolates and blue for clinical/other isolates. Four main groups or clusters were highlighted and numbered. The Figure 1 also shows the 17 variable chromosomal regions identified in GG, as further detailed in Table 1. Each row corresponds to one strain, and each column shows the genes in these variable regions, colour-coded as follows: blue for present and yellow for absent.

Figure 2. Comparison of hierarchical clustering and phylogenetic tree of a selected set of L. rhamnosus strains.

Both hierarchical clustering (panel A) and phylogenetic tree (panel B) were performed on all L. rhamnosus strains, excluding isolates from unspecified or clinical origins. Coloured strings connecting the same strains of both trees aims at highlighting the degree of similarities between both tree methods , .

Figure 3. API 50CH fermentative profile of L. rhamnosus strains.

Fermentation ability is indicated in black for positive, grey for partially positive and white for negative. Strains are organized according to their genetic relatedness as defined in the hierarchical clustering and coloured according to their respective niche/origin (Figure 1). Carbohydrates of interest are marked by a red asterisk. Black arrows show fermentative profile shifts among L. rhamnosus strains.

Figure 4. CRISPR spacer oligotyping and CRISPR-associated protein diversity in L. rhamnosus species.

Panel (A) illustrates the genetic organization of the CRISPR system and its associated genes in L. rhamnosus GG. Panel (B) shows the conservation (blue), the partial conservation (grey) or the absence (yellow) of L. rhamnosus GG spacers. The presence (green) or the absence (red) of the cas genes is also indicated in Panel (C). Strains are organized according to their genetic relatedness defined in Figure 1.

Figure 5. Bile resistance distribution among the different niches or groups.

Strains were classified as resistant, moderately resistant, poorly resistant or sensitive to bile salts. The table below the histogram details the bile resistance distribution of strains in each niche or group.

Figure 6. Mucus adhesion and SpaCBA pili gene diversity among L. rhamnosus.

Panel (A) shows the genotype and phenotype of all strains. Based on our genomic analysis, pilin and sortase genes were assigned as present (green) or divergent (red). Sequences of corresponding genes were further analyzed using blastx. The sequence identity was shown by an upper triangle superposed to the SOLiD genomic data, where the colour gradient corresponds to the identity percentage to GG pili genes. We also indicated if the strains were tested by immunoblotting analysis (DB), electron microscopy (EM) or in vitro competitive binding assay (AB). Green is for pili positive and red for pili negative. Panel (B) shows the human mucus binding ability (%) of all L. rhamnosus isolates ranked from the lowest to the highest mucus binder.

Figure 7. Anthropocentric view of the L. rhamnosus species.

The interactions between L. rhamnosus and the human cavities are frequent and occur in various contexts, i.e. consumption of food products (common scenario) or development of bacteremia (rare event). For each niche or isolation source, the strains were grouped according to their geno-phenotype (radar plot). The geno-phenotype is based on the scoring of distinctive genetic and phenotypic traits measured in this study, i.e. gene-content, CRISPR oligotype, bile resistance, pilosotype, sugar group I (dulcitol, D-arabinose and L-fucose), sugar group II (D-saccharose, D-maltose, methyl-α-D-glucopyranoside and D-turanose) and sugar group III(L-rhamnose, L-sorbose, D-ribose and D-lactose). The distinction between the two main geno-phenotypes mostly relies on gene acquisition and loss, point mutations, genetic reorganization that possibly reflect strain adaptation to an ecological niche.