A comparative genomic investigation of 31 strains of lactic acid bacteria has mapped specific gene clusters to the fermentation of complex carbohydrates, linking genetic presence directly to the utilisation of fructans and ribose. By pairing phenotypic growth assays across 23 different carbohydrates with high-throughput genomic sequencing, the study demonstrated that while some metabolic capacities follow clear operon structures, overall carbohydrate fermentation cannot be inferred strictly from the presence or absence of individual metabolic genes alone.
Lactic acid bacteria have long served as essential starter cultures in traditional food fermentation, playing key roles in extending shelf life through organic acid acidification, improving nutritional and sensory qualities, and supporting the formulation of functional foods. However, the evolutionary history of these organisms reflects adaptation to fluctuating, nutrient-rich environments, including plant tissues, dairy fermentations, and animal gastrointestinal tracts. Strains frequently undergo gene inactivation, gene loss, and horizontal gene transfer, leading to substantial phenotypic divergence within the same species. Traditional phenotypic screening across numerous sugar substrates is labor-intensive, while computational phenotype prediction based solely on draft annotations often falls short without verified genotype-phenotype correlation frameworks.
To investigate these evolutionary and metabolic dynamics, the research team assembled draft genomes for four L. paracasei strains, 12 L. plantarum strains, and 15 P. pentosaceus strains obtained from diverse isolation sources such as fermented plant matter, dairy substrates, brewery settings, and infant stool. Genomic DNA was extracted, prepared with the Nextera XT library kit, and sequenced using Illumina 250-base-pair paired-end technology at a minimum sequencing depth of 30-fold coverage. Read processing and quality trimming through Trimmomatic were followed by de novo assembly using SPAdes. Assembly evaluations via QUAST and CheckM confirmed high genome completeness exceeding 98.9 percent across all taxa, with zero contamination detected in L. paracasei, minimal contamination below 0.15 percent in P. pentosaceus, and between 2.16 and 3.40 percent in L. plantarum, while average nucleotide identity exceeded 98.9 to 99 percent against respective type strains.
Phenotypic characterisation evaluated bacterial growth kinetics under anaerobic conditions across 23 individual carbohydrates comprising monosaccharides, polyols, disaccharides, and long-chain polysaccharides. Cultures were grown in modified MRS broth formulated with specific buffering salts, organic acids, and vitamins, with carbohydrate solutions filter-sterilised and provided at a final concentration of 0.5 percent weight-to-volume. Optical density at 600 nanometres was monitored continuously every hour over a 72-hour incubation period at 30 degrees Celsius across three biological replicates per strain. Growth curves were subsequently modelled using a standard logistic growth equation to extract carrying capacity and growth rates, correcting final optical density measurements against unsupplemented media controls to quantify true carbohydrate-dependent growth.
Pangenomic analysis conducted with ROARY and bidirectional protein alignments revealed distinct genomic architectures reflecting the open pangenomes characteristic of nomadic bacterial lifestyles. The four L. paracasei genomes, averaging 2.96 megabases with a GC content of 46.3 percent, yielded a pangenome of 4,327 orthologous groups and a core genome of 1,940 groups. The 12 L. plantarum genomes, averaging 3.28 megabases with a GC content of 44.0 percent, encompassed 5,035 orthologous groups and a core genome of 2,362 groups. The 15 P. pentosaceus genomes, exhibiting a more compact average length of 1.82 megabases and a GC content of 37.0 percent, comprised 3,318 orthologous groups and a conserved core of 1,379 groups. In all three taxa, strains isolated from plant sources or dairy environments maintained hundreds of unique accessory genes, whereas isolates from infant stool typically displayed reduced genomes.
Phenotypic screening demonstrated wide substrate utilisation alongside notable species-specific constraints. Monosaccharides such as galactose and fructose, as well as disaccharides like maltose, trehalose, and cellobiose, supported growth across nearly all strains, although P. pentosaceus strain YP2.7.1.4.A showed minimal growth on glucose and glucan disaccharides despite harbouring essential genes for glucose-6-phosphate isomerase and phosphotransferase system subunits. Large variations were documented among sugar alcohols: mannitol was fermented by almost all L. plantarum and L. paracasei strains but supported only two P. pentosaceus isolates, while maltitol supported three L. paracasei strains and weak growth in one L. plantarum isolate. Strains also exhibited divergence on pentoses; ribose was utilised by all P. pentosaceus strains but showed strain-dependent fermentation in the remaining species, and xylose was consumed exclusively by six P. pentosaceus strains.
The study also revealed divergent profiles on prebiotic oligosaccharides and complex polysaccharides. All four L. paracasei strains failed to ferment raffinose and melibiose, even though α-galactosidase was annotated within their genomes. Conversely, most L. plantarum and P. pentosaceus strains fermented both compounds, frequently displaying diauxic growth kinetics indicative of catabolite repression during sequential metabolite consumption. Lactulose supported robust growth across L. plantarum and was more readily utilised than lactose among P. pentosaceus strains. When exposed to fructans, L. plantarum strains grew weakly or not at all on short-chain fructooligosaccharides and could not ferment long-chain inulin or levan. In contrast, specific strains of L. paracasei and P. pentosaceus exhibited strong fermentation profiles across short-chain and long-chain fructan polymers.
Through gene-trait matching, the researchers identified the precise genetic architecture responsible for long-chain fructan degradation in L. paracasei and P. pentosaceus. Strains capable of fermenting inulin and levan harboured the fosABCDXE operon, which encodes an integrated transport and degradation apparatus. The fosE enzymes shared approximately 90 percent sequence similarity between L. paracasei and P. pentosaceus, accompanied by adjacent transposase elements indicating acquisition via horizontal gene transfer.
A second major correlation established through gene-trait matching linked ribose fermentation in L. paracasei to a distinct six-gene metabolic cluster. Ribose-positive L. paracasei strains encoded a LacI-family repressor (rbsR), a D-ribose pyranase (rbsD), an ATP-binding cassette transport complex comprising an ATP-binding protein (rbsA), a permease (rbsC), and a periplasmic substrate-binding protein (rbsB), followed by ribokinase (rbsK) to direct entry into central carbon metabolism. The singular L. paracasei strain incapable of ribose fermentation, SOD21, lacked the ribokinase gene within this operon, functionally corroborating the necessity of the complete cluster. Unlike the prevailing phosphotransferase systems that drive carbohydrate uptake in the L. casei group, ribose is imported via this dedicated ATP-binding cassette system.
The findings illustrate both the diagnostic utility and the current boundaries of pure genomic prediction for industrial microbial selection. The authors conclude that systematic gene-trait matching provides an essential bridge between sequence repositories and bioprocess engineering, establishing a foundation for targeted strain screening in food fermentation, dietary FODMAP reduction, and probiotic development.
Beyond the extracellular fos operon, the researchers examined the distribution and functional limits of the intracellular beta-fructofuranosidase gene sacA. In Pediococcus pentosaceus, sacA was identified in multiple strains, including DSM20280, DSM20336, OAT107, and OAT119. Interestingly, isolates carrying both fosE and sacA, such as YP1.7.1.11A and YP3.7.1.9.A, displayed fructan utilisation profiles comparable to those possessing fosE alone. Conversely, several sacA-bearing Pediococcus strains failed to ferment fructooligosaccharides entirely. The authors attributed this phenotypic discrepancy either to low sequence similarity around 32 percent relative to active fermenters or to impaired transport protein expression. Similarly, all twelve Lactiplantibacillus plantarum genomes encoded intracellular sacA while lacking the extracellular fosE operon, yet they exhibited minimal or no growth on fructooligosaccharides, suggesting possible gene silencing via nonsense mutations, frameshifts, or deletions.
Similar incongruities emerged during the evaluation of pentose sugar metabolism. While all Pediococcus pentosaceus strains fermented ribose, xylose breakdown was restricted to six strains and directly correlated with the presence of xylose isomerase, which was absent from the genomes of the other two species. Both L-arabinose and D-arabinose supported the growth of most Pediococcus pentosaceus and Lactiplantibacillus plantarum isolates, whereas none of the Lacticaseibacillus paracasei strains could utilise either isomer. Intriguingly, genome annotations predicted L-arabinose isomerase and L-ribulose-5-phosphate 4-epimerase in both arabinose-positive and arabinose-negative Lactiplantibacillus plantarum strains without a conserved operon structure. Despite lacking predicted ribulokinase or dedicated D-arabinose isomerase genes, arabinose-positive strains successfully metabolised D-arabinose, pointing to alternative regulatory rewiring analogous to mutations documented in other bacterial models. Furthermore, complex arabinoxylans remained unfermented across virtually all tested organisms, showing only negligible absorbance increases in two Pediococcus strains.
These phenotypic specialisations carry direct industrial relevance across varied manufacturing contexts. Strains equipped with dedicated ribose catabolism offer distinct functional advantages in meat processing matrices where fermentable simple sugars are scarce, while also conferring competitive fitness in carbohydrate-rich mammalian intestinal niches. Concurrently, comparative analysis of strain origins highlighted environmental forces shaping these genetic repertoires. Dairy-associated Lacticaseibacillus paracasei isolates maintained over 400 unique orthologous groups, significantly exceeding the 218 unique groups identified in the infant stool isolate R18. Because accessory carbohydrate utilisation determinants often reside on mobile plasmids with elevated mutation rates, the authors emphasise that functional genomic analyses under varying environmental conditions will remain critical to refine phenotypic forecasts for tailored bioprocessing.
Subject of Research: Comparative genomics and carbohydrate fermentation pathways in lactic acid bacteria
Article Title: Comparative genomic analysis of selected lactic acid bacteria and phenotypic association of the key genes involved in fructan and ribose utilisation
Article References: Borowska, M., Buttimer, C., Bottacini, F., Arendt, E. K., & Coffey, A. (2025). Comparative genomic analysis of selected lactic acid bacteria and phenotypic association of the key genes involved in fructan and ribose utilisation. Discover Bacteria, 2(1), Article 5. https://doi.org/10.1007/s44351-025-00016-8
Image Credits: AI Generated
DOI: 10.1007/s44351-025-00016-8
Keywords: Lactic acid bacteria, Comparative genomics, Carbohydrate metabolism, Lacticaseibacillus paracasei, Lactiplantibacillus plantarum, Pediococcus pentosaceus, Gene-trait matching, Fructans, Ribose utilisation, Glycoside hydrolases, Pangenome analysis, Horizontal gene transfer, Food bioprocessing, Bacterial genetics, FOS fermentation, Inulin degradation, Levan hydrolysis
News Source: Juliet Wilcox. (November 13, 2026). Genomic Screening Links Specific Bacterial Gene Clusters to Complex Carbohydrate Fermentation. Scienmag.
