Integrative growth physiology and transcriptome profiling of probiotic Limosilactobacillus reuteri KUB-AC5

Culture conditions of L. reuteri KUB-AC5

L. reuteri KUB-AC5 isolated from chicken intestine samples was obtained from the stock culture of Department of Biotechnology, Faculty of Agro-industry, Kasetsart University, Thailand (Nitisinprasert et al., 2000). This strain was preserved at –80 °C in MRS medium (Difco) containing 20% (v/v) glycerol. Culture was propagated twice in MRS medium at initial pH 6.5 and 37 °C for 18–24 h. All cultivations started with the approximate concentration of 10⁹ CFU/ml at late logarithmic phase of L. reuteri as the inoculum for further investigating inhibition effects (Sobanbua et al., 2020). It was cultured with two independent replicates at 37 °C for 24 h in a modified MRS medium (Hamsupo et al., 2005) with either 2% (w/v) sucrose or 2% (w/v) glucose. Samples were taken at different time points (i.e., 0, 3, 6, 9, 12, 15, 18, 21 and 24 h) for further viable cell count and pH measurement. The viability of L. reuteri KUB-AC5 was determined by the standard plate count method. Aliquots of one mL of each sample were diluted by nine mL of sterile 0.85% NaCl solution in serial decimal dilutions. It was afterwards plated in MRS agar (Himedia) and incubated under anaerobic conditions at 37 °C for 48 h. The cell count was performed as log of colony forming units per milliliter (log CFU/ml). For pH measurement, it was performed by a digital pH meter (Mettler Toledo S220).

Determination of inhibition effects of L. reuteri KUB-AC5

Inhibition effects of the L. reuteri KUB-AC5 was determined by 96-well microtiter plate assay according to the method of Sobanbua et al. (2020). In brief, Salmonella Enteritidis S003 grown in nutrient broth (NB) for 12 h used as an indicator strain. The cell pellet obtained from 100 μl of the diluted culture was suspended in two-fold concentrated NB at pH five with approximate concentration of 10⁵ CFU/ml. Then, 100 μl of supernatant of samples was transferred to each of microtiter plate and mixed well by pipetting. Growth of the indicator strain at 37 °C for 24 h was determined by optical density (OD) at 600 nm using a Bio-Rad microtiter reader model 450 (Bio-Rad Laboratories, Hercules, CA, USA). Notably, growth of S. Enteritidis S003 in NB at pH five was also used as a reference for investigating inhibition effects (Tangthong & Nitisinprasert, 2012).

RNA extraction towards sequencing of L. reuteri KUB-AC5

For RNA extraction, cells were initially harvested from two independent cultures grown at two different carbon sources and growth phases i.e., glucose or sucrose and L-phase or S-phase. It was noted that harvested cells were taken at 6 h (mid L-phase) for investigating expression of growth-related genes. Besides, harvested cells were also taken at 12 h (early S-phase) as initial transition of growth ceases for further capturing expression of genes associated with inhibition effects against S. Enteritidis S003 growth. After that the harvested cells were quickly frozen in liquid nitrogen, and then stored at −80 °C. Total RNA samples were then extracted using the TRIzol reagent (Invitrogen, Santa Clara, CA, USA) according to the manufacturer’s recommendations. For purity determination, both contaminants and degradation components of RNA samples were evaluated using ND-1,000 spectrophotometer (NanoDrop Technologies Inc., Waltham, MA, USA) and 1% agarose gel electrophoresis, respectively. RNA integrity number (RIN) (Schroeder et al., 2006) was assessed using Agilent 2,100 Bioanalyzer (Agilent Technologies, Santa Clara, CA, USA). In regard to RNA concentration, the Qubit® RNA assay kit was used in a Qubit® 2.0 Fluorometer (Life Technologies, Carlsbad, CA, USA). The purified RNA samples were then used for cDNA library construction (Novogene Bioinformatics Technology Co. Ltd., Beijing, China) and RNA sequencing was finally performed by Illumina NovaSeq 6,000.

Transcriptome data analysis and functional annotation of L. reuteri KUB-AC5

To analyze transcriptome data, initially raw reads in FASTQ format were first processed by removing all possible low quality reads e.g., adaptor, unknown nucleotides (N) > 10%, and Q20 < 50%. The clean reads from all studied conditions (glucose or sucrose at L-phase or S-phase) were then mapped to the L. reuteri KUB-AC5 genome (Jatuponwiphat et al., 2019) using the alignment package Burrows–Wheeler Alignment tool (BWA) (Li & Durbin, 2009). The resulting mapping files in SAM format were afterwards converted to BAM format using SAMtools version 1.6 (Li et al., 2009) for estimating mapped reads to genes reference. Gene expression levels were estimated using FPKM (Fragments Per Kilobase per Million mapped reads) values based on RNA-Seq data (Trapnell et al., 2012). Once considering FPKM value ≥ 1, gene functions were consequently annotated based on the following integrative databases: the non-redundant (NR) protein database, the SWISS-PROT protein database (Bairoch & Apweiler, 2000), the database of Clusters of Orthologous Groups of proteins (COGs) (Tatusov et al., 2000), and the Kyoto Encyclopedia of Genes and Genomes (KEGG) database (Kanehisa et al., 2010).

Further, DEGs analysis was performed between possible pairwise comparisons between glucose and sucrose at L-phase or S-phase of L. reuteri KUB-AC5 cultures using DESeq2 package (Anders & Huber, 2010; Love, Huber & Anders, 2014). DEGs between two possible conditions were calculated based on the relative ratios of the FPKM values under threshold of FPKM value ≥ 1 and |log₂ fold change| ≥ 1.5. In addition, the false discovery rate (FDR) was used to adjust p-value for the significance of the differences. In this study, the FDR < 0.05 was set for gaining significant genes (Marioni et al., 2008).

Identification of reporter proteins/metabolites in relation to growth and inhibition effects of L. reuteri KUB-AC5

To identify the reporter proteins/metabolites, the metabolic network of L. reuteri KUB-AC5 at a genome-scale was constructed by annotated data (Jatuponwiphat et al., 2019) and KEGG pathway mapping (Kanehisa et al., 2016). Then, the constructed metabolic network of L. reuteri KUB-AC5 as a scaffold was integrated with the list of DEGs in response to preferable carbon source and growth phase using R package Piano (Platform for Integrated Analysis of Omics data) (Väremo, Nielsen & Nookaew, 2013). Notably, the protein or metabolite, which had a distinct up-directional P-value < 0.05, was identified as a reporter designating protein or metabolite around which the most significant transcriptional changes occurred (Patil & Nielsen, 2005; Väremo, Nielsen & Nookaew, 2013). Overlaying reporter protein/metabolite on phenotype data e.g., cellular growth and inhibition effects of L. reuteri KUB-AC5, the global metabolic responses of L. reuteri KUB-AC5 would be uncovered.

Growth characteristics and inhibition effects of L. reuteri KUB-AC5 on different carbon sources and growth phases

The growth characteristics of L. reuteri KUB-AC5 on different carbon sources and growth phases are shown in Fig. 1A. Observably, L. reuteri KUB-AC5 was rapidly growth (>7 log CFU/ml) during L-phase (3–6 h) of both sucrose and glucose cultures (Table S1). The maximum specific growth rate (μ_max) of 0.05 h⁻¹ was obtained in the sucrose culture, which was significantly higher than that of the glucose culture (μ_max of 0.04 ± 0.001 h⁻¹), as well as the bacterial biomass productivity was significantly higher in the sucrose culture (0.36 log CFU/ml/h) than in the glucose culture (0.28 log CFU/ml/h) under P-value < 0.05 (Table S2). As observed in Table S3, the similar pH values were noted in both sucrose and glucose cultures across different time points. As a result, this suggests that the sucrose was preferable source of carbon for L. reuteri KUB-AC5 growth.

To further explore inhibition effects of L. reuteri KUB-AC5, cell-free culture supernatants (CFS) of L. reuteri KUB-AC5 from sucrose or glucose cultures at L-and S-phases were grown in NB at pH five for investigating the growth of S. Enteritidis S003. Here, S. Enteritidis S003 growth was determined at different time points (i.e., 0, 5, 7, 10, 12, 16, 18 and 24 h). As the result in Fig. 1B, this indicates that the CFS of L. reuteri KUB-AC5 from sucrose culture at S-phase showed the highest inhibition effects against S. Enteritidis S003 growth, followed by sucrose at L-phase. It is noted that CFS of L. reuteri KUB-AC5 from glucose culture in such S-phase of growth significantly showed less inhibition effects against S. Enteritidis S003 growth (Fig. 1B and Table S4) when compared to sucrose culture (P-value < 0.05).

Assessment and analysis of DEGs data of L. reuteri KUB-AC5

As different culture conditions by various carbon sources and growth phases resulted in altered phenotypes, it was of interest to investigate transcriptional changes of L. reuteri KUB-AC5 on growth and inhibition effects. L. reuteri KUB-AC5 mRNA pools isolated from glucose or sucrose culture at L-phase or S-phase were sequenced using an Illumina NovaSeq 6,000. As a result, raw reads of 25.14 Megabase pairs (Mb) were gained. After removing adaptor and low-quality sequences and read pollution, clean reads were finally retrieved with an average of 24.92 Mb and an average sequencing quality of 98.37% (Table 1, Table S5). Further, a total of clean reads was then mapped to the L. reuteri KUB-AC5 genome (Jatuponwiphat et al., 2019) which resulted in the total mapped reads average for 86.29% (Table 1). Accordingly, this resulted in 1,945 expressed genes (Table S6). Regarding the annotation of 1,945 expressed genes, 1,357 genes (FPKM value ≥ 1) were considered and annotated functions by COGs database. The results show in Fig. 2. Once considering three selected functional categories, the metabolism (671 genes), the information storage and processing (397 genes) and cellular processes and signaling (289 genes) were identified. Observably, the majority of expressed genes of L. reuteri KUB-AC5 were associated with metabolic functions (Fig. 2A). Interestingly, we found the highest number of genes involved in the amino acid transport and metabolism (142 genes), followed by carbohydrate transport and metabolism (117 genes), coenzyme transport and metabolism (95 genes), nucleotide transport and metabolism (88 genes), energy production and conversion (77 genes), inorganic ion transport and metabolism (67 genes), lipid transport and metabolism (60 genes), as well as secondary metabolites biosynthesis, transport and catabolism (25 genes) (Fig. 2B, Table S7).

To further analyze DEGs, transcriptome data were organized into two pairwise comparisons: sucrose versus glucose at L-phase or S-phase. Under the thresholds of |log₂ (fold change)| ≥ 1.5 with a false discovery rate (FDR) ≤ 0.05 between all possible comparisons (Fig. 3A). Accordingly, the numbers of significant genes from DEGs analysis were identified. Observably, these genes were distributed into sucrose versus glucose conditions at L-phase (214 significant genes) or S-phase (151 significant genes) were identified. Noticeably, we found 52 up-regulated genes and 162 down-regulated genes in sucrose at L-phase as well as 39 up-regulated genes and 112 down-regulated genes in sucrose at S-phase (Fig. 3B). Figure S1 shows an overall list of significant genes from transcriptome profiling.

Reporter proteins/metabolites uncovering global metabolic responses from the integrative transcriptome and metabolic network of L. reuteri KUB-AC5

To identify reporter proteins/metabolites, the constructed metabolic network of L. reuteri KUB-AC5 containing 609 genes, 622 biochemical reactions, and 698 metabolites (Table S8) was used for integrative transcriptome analysis (see “Materials and Methods”). Promisingly, we found 11 reporter proteins under distinct up-directional P-value < 0.05 (Table 2). Of considering these reporter proteins in context of expression patterns, we clearly observed the higher transcriptional changes in a form of log₂ FPKM values occurred in sucrose than glucose in both L-and S-phases as illustrated in Figs. 4 and 5. Once considering at each phase i.e., L-phase, we observed unique reporter proteins i.e. L-histidine transport via ABC system, L-arginine transport via ABC system, L-glutamine transport via ABC system, L-cysteine transport via ABC system, and D-cysteine transport via ABC system. Regarding at S-phase, we found unique reporter proteins, namely levansucrase, sucrose phosphorylase, and sucrose transport via permease, and dihydroorotase. Clearly, the results showed that amino acid transport via ABC systems and sucrose metabolism and transport are key metabolic response-related insight from the transcriptome profiling of L. reuteri KUB-AC5.

Considering on the reporter metabolites as identified in Fig. 6, once shifting from glucose to sucrose at L-phase or S-phase, accordingly we found 14 reporter metabolites in common e.g., sucrose, D-fructose, ATP, ADP, phosphate, diphosphate, L-glutamate, L-glutamine, carbamoylphosphate, N-carbamoyl-L-aspartate, dihydroorotate, orotidine 5′-phosphate, orotate, and bicarbonate. These revealed that L. reuteri KUB-AC5 had a metabolic control in acclimatization to carbon sources (e.g., sucrose and fructose) and energy pools (e.g., ATP) through transcriptional co-regulation, which supported the cell growth at L-or S-phases cultivation.

Regarding on unique metabolites identified in each phase, interestingly we majorly found different amino acids and their intermediates involving in the metabolism of the amino acids (e.g., cysteine, histidine, aspartate, arginine, L-phenylalanine, L-tyrosine, 4-aminobutanoate, phenylpyruvate, N(omega)-(L-arginino)succinate, 3-(4-hydroxyphenyl)pyruvate), 2-oxoglutarate as reporters at L-phase (Fig. 6A, Table S9). These suggest that amino acid metabolism is a key role for L. reuteri KUB-AC5 primary growth. Considering on S-phase, it was the depletion of an essential nutrient and/or the formation of an inhibitory product. Promisingly, we found a potential unique reporter metabolite e.g., levan (Fig. 6B, Table S10) as an inhibitory product. For the other remaining reporter metabolites in S-phase, we also found 5-phospho-alpha-D-ribose1-diphosphate, D-glucose-1-phosphate, D-ribulose 5-phosphate, formate, 2,5-diamino-6-hydroxy-4-(5′-phosphoribosylamino)-pyrimidine, 3,4-dihydroxy-2-butanone-4-phosphate. Taken together, the reporter proteins/metabolites allow identifying hot spots around which significant regulation occurs in amino acid metabolism and transport as well as sucrose metabolism and transport of L. reuteri KUB-AC5.

Identified key gene and metabolic function in relation to inhibition potentials of L. reuteri KUB-AC5

Focusing reporter proteins/metabolites in relation to phenotype data e.g., the highest inhibition effects of L. reuteri KUB-AC5 against S. Enteritidis S003 growth in sucrose culture at S-phase (Fig. 1B), interestingly we found that significant regulation occurred around levan as listed in Table 3. Indeed, we observed sacB genes in a form of cluster (e.g., AC5u0009GL000236, AC5u0009GL000237, AC5u0009GL000238 and AC5u0009GL000239) with overall up-regulation in sucrose. According to S-phase, observably we found higher log₂ fold changes than L-phase (Table 3). Altogether, these clearly indicated that the expression of sacB gene cluster was highly induced by the sucrose at S-phase. Mapping of phenotypic data of inhibition effects and transcriptome profiling therefore suggests that levan biosynthesis from sucrose utilization via levansucrase might be occurred, so that levan can be a target of interest for inhibition potentials of probiotic L. reuteri KUB-AC5.

Availability of data and materials

All sequences have been deposited in the Sequence Read Archive (https://www.ncbi.nlm.nih.gov/sra) under the accession number PRJNA563573 (BioSamples: SAMN18437211, SAMN18437246, SAMN18437399, SAMN18437752, SAMN18438585, SAMN18438691, SAMN18439203 and SAMN18439842).