I am better than I look: genome based safety assessment of the probiotic Lactiplantibacillus plantarum IS-10506

Abstract

Background: Safety of probiotic strains that are used in human and animal trials is a prerequisite. Genome based safety assessment of probiotics has gained popularity due its cost efficiency and speed, and even became a part of national regulation on foods containing probiotics in Indonesia. However, reliability of the safety assessment based only on a full genome sequence is not clear. Here, for the first time, we sequenced, assembled, and analysed the genome of the probiotic strain Lactiplantibacillus plantarum IS-10506, that was isolated from dadih, a traditional fermented buffalo milk. The strain has already been used as a probiotic for more than a decade, and in several clinical trials proven to be completely safe.

Methods: The genome of the probiotic strain L. plantarum IS-10506 was sequenced using Nanopore sequencing technology, assembled, annotated and screened for potential harmful (PH) and beneficial genomic features. The presence of the PH features was assessed from general annotation, as well as with the use of specialised tools. In addition, PH regions in the genome were compared to all other probiotic and non-probiotic L. plantarum strains available in the NCBI RefSeq database.

Results: For the first time, a high-quality complete genome of L. plantarum IS-10506 was obtained, and an extensive search for PH and a beneficial signature was performed. We discovered a number of PH features within the genome of L. plantarum IS-10506 based on the general annotation, including various antibiotic resistant genes (AMR); however, with a few exceptions, bioinformatics tools specifically developed for AMR detection did not confirm their presence. We further demonstrated the presence of the detected PH genes across multiple L. plantarum strains, including probiotics, and overall high genetic similarities between strains.

Conclusion: The genome of L. plantarum IS-10506 is predicted to have several PH features. However, the strain has been utilized as a probiotic for over a decade in several clinical trials without any adverse effects, even in immunocompromised children with HIV infection and undernourished children. This implies the presence of PH feature signatures within the probiotic genome does not necessarily indicate their manifestation during administration. Importantly, specialized tools for the search of PH features were found more robust and should be preferred over manual searches in a general annotation.

Keywords: Comparative genomics; Genome; Lactiplantibacillus plantarum; Safety assessment.

Fig. 1

Circular representation of L. plantarum IS-10506 genome generated by BRIG. Figure A depicts the chromosome and figures B, C, and D the three plasmids

Fig. 2

Summary of L. plantarum IS-10506 gene assignment by KAAS to KEGG pathways and BRITE hierarchies. The number of assigned genes is shown on the x-axis and the pathway or hierarchy group (levels B and C, respectively, in the KAAS annotation output) on the y-axis. Bar colours show the highest informative grouping of pathways or hierarchies corresponding to general functions

Fig. 3

Summary of L. plantarum IS-10506 genes annotated as potentially harmful. The y-axis shows the metabolic annotation of a gene by KAAS; if a gene was not assigned to any potentially harmful pathway, it was assigned as not annotated (NA). Colours indicate the tool used for identification

Fig. 4

Comparison of L. plantarum IS-10506 with L. plantarum genomes from the NCBI RefSeq collection. (A) a heat-map of the average nucleotide identity (ANI) between all analyzed L. plantarum genomes. (B) a heat map showing only a cluster of genomes with the smallest differences in ANI from L. plantarum IS-10506; colored top annotation indicates strain isolation source and location. (C) a circular representation of L. plantarum IS-10506 genome and the six closest strains in terms of ANI genomes from NCBI RefSeq in comparison with L. plantarum SK151 (reference strain). (D) a histogram plotting the number of pairwise comparisons on the y-axis and the corresponding ANI on the x-axis

Fig. 5

Overview of results of the L. plantarum pangenome analysis. Figure A shows the number of genes belonging to different gene categories within the L. plantarum IS -10506 chromosome. Figure B shows the first two axes of the PCoA ordination based on binary dissimilarity distances as a scatter plot

Fig. 6

Summary of genes in L. plantarum genomes from the NCBI RefSeq collection homologous to potentially harmful (PH) genes identified in L. plantarum IS-10506. (A) number of genomes with a homologous gene on the x-axis, with the level of identity encoded as color intensity. (B) the identity of homologous genes to L. plantarum IS-10506 PH genes in the form of boxplots. Colours indicate PH gene detection method for both figures

Fig. 7

Overview of insertion sequences (IS) identified in L. plantarum strains (chromosomes). Figure A and B are heat maps showing the frequency of IS clusters (columns) in the investigated L. plantarum strains (rows); the black frame in Figure A highlights the clusters of IS shown in Figure B; rows and columns are clustered hierarchically. Figure C is a histogram showing the IS frequency (x-axis) in the L. plantarum strains (y-axis); the black dot on the x-axis shows the number of IS in L. plantarum IS -10,506. Figure D is a bar chart showing the total number of IS clusters in all L. plantarum strains; the bars highlighted in blue show IS clusters identified in L. plantarum IS -10,506