Genomic Characterization of Lactobacillus delbrueckii Strains with Probiotics Properties

Probiotics are health-beneficial microorganisms with mainly immunomodulatory and anti-inflammatory properties. Lactobacillus delbrueckii species is a common bacteria used in the dairy industry, and their benefits to hosting health have been reported. This study analyzed the core genome of nine strains of L. delbrueckii species with documented probiotic properties, focusing on genes related to their host health benefits. For this, a combined methodology including several software and databases (BPGA, SPAAN, BAGEL4, BioCyc, KEEG, and InterSPPI) was used to predict the most important characteristics related to L. delbrueckii strains probiose. Comparative genomics analyses revealed that L. delbrueckii probiotic strains shared essential genes related to acid and bile stress response and antimicrobial activity. Other standard features shared by these strains are surface layer proteins and extracellular proteins-encoding genes, with high adhesion profiles that interacted with human proteins of the inflammatory signaling pathways (TLR2/4-MAPK, TLR2/4-NF-κB, and NOD-like receptors). Among these, the PrtB serine protease appears to be a strong candidate responsible for the anti-inflammatory properties reported for these strains. Furthermore, genes with high proteolytic and metabolic activity able to produce beneficial metabolites, such as acetate, bioactive peptides, and B-complex vitamins were also identified. These findings suggest that these proteins can be essential in biological mechanisms related to probiotics’ beneficial effects of these strains in the host.


INTRODUCTION
Lactobacillus delbrueckii is the type species of Lactobacillus genus after a new proposed taxonomic reclassification that divided this genus into 25 new, based on genetic and phylogenetic analysis associated with ecological and metabolic properties (Zheng et al., 2020). This Lactic Acid Bacteria (LAB) member comprises gram-positive, rod-shaped, facultatively anaerobic, and acid-resistant microorganisms, which occupy diverse carbohydrate-rich environments with final fermentative metabolism-derived lactic acid production (Salvetti et al., 2012;Duar et al., 2017). This species includes mainly two subspecies: bulgaricus and lactis, both with high importance in industrial fermented dairy products (primarily yogurt and cheeses production) and biotherapeutics approaches (Hao et al., 2011;El Kafsi et al., 2014;Santos Rocha et al., 2014).
Some studies have been characterizing the L. delbrueckii strains as probiotics based on their ability to resist gastrointestinal tract (GIT) stressors (Ferreira et al., 2013), pathogens inhibition (De Jesus L. C. L. et al., 2021), and antiinflammatory effects mainly focused on GIT disease treatment, such as colorectal cancer (Wan et al., 2014), ulcerative colitis (Santos Rocha et al., 2014), and intestinal mucositis (De Jesus et al., 2019). In addition, pre-clinical therapeutical applications of these microorganisms to other pathological conditions, such as arthritis (Kano et al., 2013), depression (Qiu et al., 2021), and diabetes (Hallajzadeh et al., 2021), have also been reported. Among this species, L. delbrueckii subsp. lactis CIDCA 133 is the best-characterized probiotics strain whose beneficial characteristics and safety aspects have been widely evaluated by in vitro and in vivo, as well as in silico analysis, for example, its ability to inhibit Escherichia coli, Bacillus cereus, Citrobacter rodentium, and Salmonella Typhimurium pathogens; immunomodulation by inhibition of NF-κB signaling pathway; tolerance to high concentrations of bile salts; no hemolytic or mucin degradation activity, and no adverse effects to clinical and histopathological mice parameters (Rolny et al., 2016;Hugo et al., 2017;De Jesus et al., 2019;De Jesus L. C. L. et al., 2021;De Jesus LCL. et al., 2021;Barroso et al., 2022).
Although most studies focus on the effect and action mechanism of viable probiotic strains, there is a growing interest in applying probiotics as microbiologically non-viable but immunologically active products. This would be more viable and safer for probiotic applications in clinical practice due to safety concerns regarding this active metabolic form favoring the risk of bacterial translocation (Moradi et al., 2020;Teame et al., 2020). Some studies have evaluated the inactivation of these microorganisms or products derived from them in different inflammation models and obtained similar results to their metabolically active form (Sang et al., 2013;Nakai et al., 2021;Trindade et al., 2021).
According to Hill et al. (2014), probiotics are defined as "live microorganisms that confer a health benefit on the host when administered in adequate amounts." However, it should also be highlighted that most of these beneficial effects attributed to probiotics are strain-dependent, revealing that individual characteristics of the strains provide relevant data for the development of effective probiotic products and facilitate individualized or personalized use for clinical applications (Bubnov et al., 2018;McFarland et al., 2018). This strainspecific property was more substantially related by Rocha et al. (2012) that, when screening 57 dairy L. delbrueckii strains, observed that the immunomodulation levels of these bacteria varied depending on the strain. Among the 37 L. delbrueckii subsp. bulgaricus and 20 L. delbrueckii subsp. lactis strains tested, the most effective immunomodulators strains belong to the subsp. lactis (Rocha et al., 2012), including CNRZ327 and CNRZ333 strains.
Individual biological properties of probiotic strains may be related to a high degree of variation in their genomic content. Thus, studies at the genomic level can provide insights into the main genetic factors and molecular mechanisms associated with the probiotic features of these microorganisms, such as GIT survival, pathogens inhibition, and immunoregulation (Ventura et al., 2012;Salvetti and O,Toole, 2018;Castro-López et al., 2021). Probiotics studies using the genome approach have been performed to identify genetic factors involved with features of different potential probiotics strains, such as Lactobacillus helveticus (Fontana et al., 2019), Lactiplantibacillus plantarum (Zhang et al., 2018), Pediococcus sp. (Wanna et al., 2021), Bifidobacterium sp. (Duar et al., 2020), Enterococcus sp. (Hussein et al., 2020), Lactococcus lactis (Oliveira et al., 2017), among others. In this context, the comparative analysis proves to be an essential tool in probiogenomics, contributing to further exploring the diversity and evolutionary relationship of species , and identifying and comparing the gene repertoire in different strains (Fontana et al., 2019) and the relationship of these molecules with reported probiotics effects of these bacteria on the host (Papadimitriou et al., 2015;Sun et al., 2015).
Although the importance of L. delbrueckii strains in the food industry, few studies have focused on genomic studies of L. delbrueckii probiotics strains regarding their host health benefits (El Kafsi et al., 2014;Sun et al., 2015;Kanmani et al., 2018;De Jesus L. C. L. et al., 2021). Thus, this study carried out a comprehensive functional gene characterization of L. delbrueckii species with reported probiotics effects, which may be associated with the specific host health benefits of these strains reported phenotypically, and provide a better comprehension of their probiotics features.

Pan-Genome Analysis
This study used the BPGA (Bacterial Pan Genome Analysis) pipeline for performance pan-genome (Chaudhari et al., 2016). The genome sequences were submitted in FASTA format to Orthofinder software to predict orthologs genes (Emms and Kelly, 2019), using default parameters with a p-value cut-off of 1E −5 . This software bases its inference method on OrthoMCL (Li et al., 2003) through the hybrid Markov Clustering algorithm (Enright, 2002), which computes sequence similarities with BLAST and then uses the MCL clustering algorithm to identify clusters of highly connected sequences. After this process of predicting orthologous genes, through in-house scripts, these genes were classified according to the subsets of the pangenome, being divided into the core genome, shared, and singletons. For the development of the pangenome, after the classification process in its subsets, in-house scripts were used to estimate what would be the fixed parameters of the Heap Law (Soares et al., 2013;Guimaraes et al., 2015) and the Utterance of the Least Squares Principle (for core genome subsets and singletons). For genetic contexts, we can represent Heap's Law according to the formula n = k *N γ, which (n) would be the value for the number of genes for a given number of genomes (N); and then k and γ can be considered as free parameters. By this law, γ can be calculated as α = 1-γ, so when α > 1 (γ < 0), the pangenome is called closed, which means that there is no increase, or there is no significant increase, of genes when more genomes of the studied organism are sequenced. If α < 1 (0 < γ < 1), suggests the pan-genome of the probiotic strains is open, which indicates that there is an increase in the number of genes when more genomes are sequenced. The Least Squares Principle Statement can be represented by the formula n = k * exp [−x/t] + tgθ, where (n) is also the number of genes, and k, t, and tgθ are considered as parameters free. With the result of this law, we were able to estimate, based on the number of singletons added to each new sequencing, how many genomes are still needed for the core genome of the studied group to reach stability.

Prediction of Genes Related to Antibacterial Activity
Proteins involved in antibacterial activity were also evaluated across the probiotics L. delbrueckii genomes. For this purpose, genes coding bacteriocins were predicted through BAGEL4 (http://bagel4.molgenrug.nl/) (van Heel et al., 2018). The bacteriocins-encoding genes' distribution among the genomes was visualized through a heatmap of presence and absence. Furthermore, core proteins producing other antimicrobial compounds, such as hydrogen peroxide and organic acids, were investigated using the KEGG Mapper/BLASTKOALA tool (Almeida et al., 2021).

Identification of Gastrointestinal Tract Stress Response Genes and Proteolytic Enzymes in Core Genome
Identification of core proteins of probiotics L. delbrueckii strains related to GIT stress response (acid and bile) and proteolytic activity were manually predicted through Prokka-derived annotation, based on previous studies (Liu et al., 2010;Papadimitriou et al., 2016;De Jesus LCL. et al., 2021).

Interaction of Core Proteins of Lactobacillus delbrueckii Strains With Human Immune Proteins
To evaluate the potential biological interaction between core proteins of L. delbrueckii probiotic strains and human immune proteins, first, the subcellular localization of proteins identified in the core genome was predicted using SurfG+ software (Barinov et al., 2009). Second, the core proteins were predicted for their ability to be an adhesin calculated by SPAAN software (score >0.7) (Sachdeva et al., 2005). After, immune protein sequences related to the inflammation pathways (TLR2/4-MAPK, TLR2/4-NF-κB, and NOD-like receptor signaling pathways) were mapped and obtained from KEGG pathways and UniProt (UP000005640), respectively (Supplementary Table S1). Finally, the proteinprotein interaction was conducted in the InterSPPI v2 web server (http://zzdlab.com/InterSPPI/) (Lian et al., 2019). Graphical analysis of resulting interaction networks (minimum score: 0.9765; specificity: 0.99) was performed by Cytoscape v3.9.0 software (Shannon, 2003).

Pan-Genome Analysis
The pan-genome (total gene repertoire) obtained through BPGA with nine probiotics L. delbrueckii strains is composed of a total of 2,609 genes ( Figure 1A), of which 1,268 (48.60%) belong to the core genome (number of genes shared by all strains), 892 genes (34.18%) to the accessory genome (genes shared by two or more strains), and 449 (17.20%) are strain-specific (uniques) (genes present in a single strain) ( Figure 1B). Furthermore, the CIDCA 133, DSM20074, and CNRZ327 strains presented the highest exclusive genes, with 102, 76, and 69 genes, respectively, followed by TUA4408L (53 genes), 2038 (47 genes), ATCC11842 (39 genes), KLDS10207 (32 genes), LJJ (16 genes), and DSM20080 (15 genes). According to the curve generated for these nine genomes based on Heap's Law and leats-square fit of the exponential regression decay, the number of genes families in the pan-genome increased with the addition of each other genome (n = 1,848.134*n^0.156), suggesting that the pan-genome of probiotics L. delbrueckii strains remains open. For the subsets of the core genome and singletons developed by the Utterance of the Least Squares Principle, it can be observed a value of tgθ of approximately 1,182 genes (n = 465.995 * exp [−x/4.839] + 1182.675) for the core genome, and a value of approximately 24 (n = 219.676 * exp [−x/4.356] + 24.813) for the strain-specific. This result shows that at each new sequencing, 24 new genes are added to this pangenome, and it is expected that the core genome will stabilize when it reaches around 1,182 genes ( Figure 1A).
Genes encoding proteins such as glucokinase, glucose-6phosphate isomerase, glyceraldehyde 3-phosphate dehydrogenase, phosphoglycerate kinase, ribulose-phosphate 3-epimerase, pyruvate kinase, phosphoketolase, lactate dehydrogenase, and acetate kinase were also identified in the core genome of L. delbrueckii probiotics strains. These essential proteins are involved in the homofermentative or heterofermentative pathways, producing lactate or acetate. Genes involved in the biosynthesis of complex B vitamins were also predicted, including thiamine pyrophosphokinase (thiamine or vitamin B1), a riboflavin kinase (riboflavin or vitamin B2), dihydrofolate reductase (folate or vitamin B9), and cob(I)alamin adenosyltransferase (cobalamin or vitamin B12). No propionate or butyrate-related gene was identified in the core genome (Supplementary Table S4).

Probiotics L. delbrueckii Strains Harbors Genes Related to Antibacterial Profile
The L. delbrueckii strains showed different profiles in terms of bacteriocins. Among all strains, subspecies lactis showed a greater diversity of bacteriocins in their genome, including enterolysin A, helveticin J, and bovicin_255. Few bacteriocins were found for the subspecies bulgaricus. The bacteriocins enterolysin A appears to be conserved in the species (Figure 3). Furthermore, it was identified in the core genome D-lactate dehydrogenase, L-lactate dehydrogenase, acetate kinase, L-lactate oxidase, glycolate oxidase, and pyruvate oxidase genes, which acts like crucial enzymes in the biosynthesis of organic acids (lactate and acetate), and hydrogen peroxide, respectively (Supplementary Table S5).

DISCUSSION
Comparative genomics revealed a high variation level in the genome of nine L. delbrueckii probiotics strains, with the subspecies lactis presenting a larger genome size (Mb) than subspecies bulgaricus, corroborating the findings of El Kafsi et al. (2014). This genomic variation can be related to the differences in the number of unique genes observed across the strains, in which the subspecies lactis had the highest number. The pangenome analysis of Lactobacillus delbrueckii species has already been carried out by Inglin et al. (2018) and Kim et al. (2021). However, the above authors did not perform a functional analysis related to the probiosis of these strains. Thus, in our work, the performance of a combined analysis of pan-genome data of nine potential L. delbrueckii probiotics strains allowed us to obtain more robust data related to the most relevant characteristics of the probiose of these strains, mainly related to their ability to survive the TGI, adhesion, antibacterial activity, and immunomodulation. Functional analysis of the core genome revealed that the proteins of nine L. delbrueckii probiotic strains are mainly involved in genetic and environmental information processing and metabolic activities, which suggests the importance of these genes in conserved cellular processes of these microorganisms to survive and adapt to specific environments or host.
One of the first adaptation steps of probiotics to the host involved molecular/cellular mechanisms related to their response to GIT stressors (stomach acidity and bile salt) (Papadimitriou et al., 2016). The core genome of L. delbrueckii probiotics strains harbors genes related to these stress response mechanisms, mainly including transcriptional regulators expression (e.g., two-component sensor histidine kinase), proton extrusions, and bile efflux (e.g., Na+/H+ antiporter, F0F1 ATPase genes, glycine/betaine ABC transporter permease), metabolic response (e.g., acetate kinase, pyruvate oxidase, ornithine decarboxylase), and heat shock/chaperones proteins production (e.g., GroEL, GroES, DnaK, DnaJ, ClpX). The expression of these genetic factors can be essential to the survival strategy of these bacteria on the GIT, allowing them to arrive in viable amounts sufficient to promote their interactions and beneficial effects with the specific-host sites of action. Genome and phenotype-scale studies demonstrated that these survival and adaptation mechanisms were observed in L. delbrueckii LJJ  strain and are also shared with others, such as UFV H2b20 (Ferreira et al., 2013), 2038 (Hao et al., 2011), ATCC 11842, and CNRZ327 (El Kafsi et al., 2014), and CIDCA 133 (De Jesus LCL. et al., 2021. This study´s probiotics L. delbrueckii strains were mainly isolated from dairy products, supporting the prediction of core genome enzymes related to a conserved proteolytic and metabolic sugar system. This high metabolic property enhances the fermentation ability of these strains with the production of essential metabolites (e.g., bioactive peptides, lactate, SCFA, and vitamins). These compound's synthesis requires specific enzymes (e.g., proteinases and peptidases, glyceraldehyde-3phosphate dehydrogenase, pyruvate kinase, phosphoketolase, acetate kinase, lactate dehydrogenase, riboflavin kinase, thiamine pyrophosphokinase, among others) of these strains, involved in proteolysis, and both phosphoketolase or Embden-Meyerhof (EMP) metabolic pathways (Kandler, 1983;Ye et al., 2021).
A vital feature derived from the fermentation process by probiotic strains is their antimicrobial activity due to organic acids, hydrogen peroxide, and bacteriocins production. The genome of L. delbrueckii probiotics strains has genes coding for these antimicrobial compounds (e.g., bacteriocin enterolysin A, D-lactate dehydrogenase, L-lactate dehydrogenase, acetate kinase, L-lactate oxidase, glycolate oxidase, and pyruvate oxidase), which makes them highly relevant in the food industry, since when used in the fermentation of food dairy products, it can control and preserve these products against the food spoilage of pathogens. The antibacterial effect associated with these L. delbrueckiiproducing compounds against some pathogens, such as Salmonella sp., Enterococcus faecalis, Escherichia coli, Gardnerella vaginalis, Listeria monocytogens, Pseudomonas aeruginosa, has been previously reported (Evivie et al., 2020;De Jesus LCL. et al., 2021;Qiu et al., 2021).
It is essential to identify genes/metabolic pathways and characterize bioproducts produced by probiotic bacteria with high fermentative capacity since studies have demonstrated the beneficial effects of fermented products derived from these microorganisms in GIT inflammatory diseases. For example, milk fermented by L. delbrueckii CNRZ327 (2 × 10 9 CFU/mL) attenuated TNBS-induced colitis in a murine model, improving the epithelial architecture, and reducing inflammatory parameters (IL6, TNFα, MPO) and oxidative markers (COX2 and Hmox) (Plé et al., 2016). Similar effects were reported in L. delbrueckii CIDCA 133, whose milk fermented by the strain preserved the intestinal epithelium from the inflammatory damage caused by the chemotherapy drug 5-FU (300 mg/kg) (De Jesus et al., 2019). Another study demonstrated that intake of yogurt fermented with L. delbrueckii 2038 improves aging by metabolites production and microbiota and intestinal epithelial regulation (Usui et al., 2018). It is suggested that these effects can be associated with the production of organic acids (lactate, SCFA) and bioactive metabolites (vitamins) produced by these bacteria, although these studies did not assess their concentration. However, it is important to highlight that the ability of L. delbrueckii species to produce SCFA or vitamins with host health benefits has been previously reported (Laiño et al., 2012;Levit et al., 2018;Dan et al., 2019), makes them promise to be used as an adjuvant for the treatment of inflammatory GIT diseases and other pathological conditions due to their reported antioxidants, anti-inflammatory, and immunomodulatory properties.
Immunomodulatory and anti-inflammatory properties or probiotics bacterial can also be related to the surface layer proteins or extracellular proteins (Hidalgo-Cantabrana et al., 2020;Chandhni et al., 2021) due to the ability of these proteins to interact with the host cells via pattern recognition receptors (e.g., Toll-like receptors-TLR, NOD-like receptors-NLR) inducing specific signalization pathways responses, as nuclear factor kappa B (NF-κB) and mitogen-activated protein kinase (MAPK) (Delgado et al., 2020). This hypothesis is corroborated by Rocha et al. (2012) when they showed that surface-exposed proteins of the L. delbrueckii CNRZ333 strain played a role in NF-κB immune modulation (Rocha et al., 2012).
L. delbrueckii strains characterized as probiotics shared genes with potential interaction with inflammatory pathways-related human immune proteins, including proteinase B (PrtB), penicillin-binding protein (Pbp1A), and lipoteichoic acid synthase (LtsA), among others, with PrtB being the most interacting protein. PrtB is a cell envelope-associated serine protease essential to milk casein degradation (Gilbert et al., 1996). The expression of this protein and its analogs producing bioactive health-beneficial peptides has been suggested to be crucial to the immunomodulatory properties of L. delbrueckii strains (De Jesus LCL. et al., 2021). For example, De Jesus LCL. et al. (2021) showed that predicted proteins of L. delbrueckii CIDCA 133 with a high adhesin profile, including PrtB protein, interacted with human immune proteins involved with NF-κB signaling pathway activation. These findings can be related to their in vivo results. It was demonstrated that consumption of this probiotic strain presented an antiinflammatory profile by activating TLRs receptors (Tlr2, Tlr4), decreasing Nfkb1 and enhancing immunoregulatory markers Il10 and Tgfb gene expression (De Jesus LCL. et al., 2021). These results are also supported by Espeche Turbay et al. (2012). They demonstrated that milk β-casein degradation by L. delbrueckii CRL581 ameliorates TNBS-induced acute intestinal inflammation by increasing immunoregulatory IL10 and decreasing leukocytes infiltrate and the IFNγ proinflammatory marker (Espeche Turbay et al., 2012). It is believed that these effects can be attributed to its cell envelope-associated proteinase PrtL activity (Villegas et al., 2015).
Anti-inflammatory properties of other surface layer components of L. delbrueckii strains have also been reported as extracellular polysaccharides of L. delbrueckii TUA4408L, which presented antiviral activity against rotavirus infection in porcine cells by modulating TLR2/4, interferon regulatory factor (IRF)-3, and the antiviral factors IFN-β, MxA, and RNase L expression (Kanmani et al., 2018). Altogether, these findings reveal that these bacteria factors are essential to leading the biological process of the host, mainly immune regulation. Therefore, based on these findings, the knowledge at the genomic level of the individual characteristics of probiotics, as well as the genetic factors associated with their immunoregulatory capacity, can facilitate individualized or personalized use of them for clinical applications, thus being an alternative approach to the problems arising from the use of live beneficial microorganisms in clinical practice. Furthermore, the exploration of genetic factors can contribute to validating the role of these probiotics-derived bioactive molecules in different pathological conditions, including their beneficial effects on those that affect distant sites and organs (e.g., skin, respiratory and urogenital tracts, brain, bones, among others) (Reid et al., 2017;Bubnov et al., 2018). Thus, we reinforce that further studies, including knockout genes or heterologous production of these proteins, must be performed to validate these genotypic findings with the phenotypic reported results described for these strains and elucidate the underlying mechanisms involved in their immunomodulatory activities.
In summary, this first probiotic genomic characterization study for potential L. delbrueckii probiotics species shows that these bacteria share a broad gene repertoire that functionally may be responsible for phenotypic features attributed to these strains on the host. The data presented support other studies that aim to identify genetic factors and mechanisms related to the beneficial effects of new probiotic targets from the Lactobacillus species with high commercial and biotechnological relevance. Furthermore, these data open perspectives for new studies to be carried out to evaluate the predicted interacting bacteria proteins with human immune proteins as possible anti-inflammatory molecules to be tested in therapeutic approaches to different inflammatory conditions.

AUTHOR CONTRIBUTIONS
Conceptualization: LJ and FA; Methodology: LJ, TS, AF, and FA; Formal analysis and investigation: LJ and FA; Writing-original draft preparation: LJ; Writing-review and editing: FA, TS, SS, LA, and VA. All authors read and approved the final manuscript.

FUNDING
This work was supported by the National Council for Scientific and Technological Development (CNPq) (Grant number 312045/ 2020-4).