Probiogenomics as a tool to obtain genetic insights into adaptation of probiotic bacteria to the human gut

Abstract

Bifidobacteria and lactobacilli are widely exploited as health-promoting bacteria in many functional foods. However, the molecular mechanisms as to how these bacteria positively impact on host health are far from completely understood. For this reason these microorganisms represent a growing area of interest with respect to their genomics, molecular biology and genetics. Recent genome sequencing of a large number of strains of bifidobacteria and lactobacilli has allowed access to the complete genetic makeup of representative members of these bacteria. Here, we will discuss how the analysis of genomic data has helped us to understand the mechanisms by which these bacteria adapt to the specific environment of the gastrointestinal tract, while also revealing genetic functions that mediate specific host-microbe interactions.

General Features

The bacterial community living in the human gastrointestinal tract (GIT), also known as GIT microbiota, are composed by a vast collection of microorganisms whose composition differs depending on the different regions of the gut. Bifidobacteria and lactobacilli are common inhabitants of the distal regions of the GIT, i.e., the large and the small intestine, respectively.¹ Interestingly, the intestinal microbiota not only includes naturally resident lactobacilli, also known as autochthonous lactobacilli, but also various lactobacilli that have been acquired by food ingestion.

The genera Bifidobacterium and Lactobacillus belong to the phyla Actinobacteria and Firmicutes, respectively, both representatives of Gram positive microorganisms that ferment carbohydrates to mainly organic acids. Bifidobacteria predominantly produce acetate and lactate as fermentation end products, whereas lactobacilli will produce a variety of organic acids, although all produce a significant amount of lactate. Bifidobacteria and lactobacilli are often grouped together based on the fact that these microorganisms share certain metabolic features (i.e., lactic acid production), while both are also extensively exploited by the food industry as health-promoting or probiotic bacteria in functional foods. Nevertheless, one should keep in mind that lactobacilli and bifidobacteria from a phylogenetic perspective occupy distinctly different positions.

The interplay between the GIT microbiota and the human host can be classified as a continuum involving symbiosis and commensalism to pathogenesis. In the human GIT, co-evolution of such host-microbe interactions is the consequence of commensal relationships in which neither partner is disadvantaged, and symbiotic relationships in which both partners benefit, be it from unique metabolic activities or other advantageous properties.

Probiotics and Health

The probiotic concept dates back to 1908 when Metchnikoff noticed that the consumption of certain fermented foods elicited positive effects on human health.² The generally accepted definition of probiotics was proposed by the Food and Agriculture Organization (FAO) World Health Organization (WHO) as follows: “Probiotics are live microorganisms which when administrated in adequate amounts confer a health benefit on the host” (FAO/WHO 2001). This definition implies that safety and efficacy must be demonstrated for each probiotic strain. No specific criteria for selecting new probiotics have been so far proposed, however general criteria that must be satisfied involve the capacity to adhere to the intestinal mucosa and the ability to tolerate acid and bile stress.^3,4 There is accumulating evidence underpinning the capacity of probiotic strains to exert one or more of the following positive activities: anti-inflammatory immune-modulation, reduction of atopic disease symptoms, beneficially influencing the composition and activity of intestinal microbiota, alleviation of acute gastro-enteritis, prevention or suppression of bacterial infections, reduction of lactose intolerance, reduction of intestinal inflammation, production of specific short chain fatty acids, conjugated linoleic acids and vitamins and alleviation of constipation.^5-7

Although there is suggestive evidence for each of these functional claims, the molecular mechanisms behind such probiotic activities remain largely unknown. The decoding of microbial genome sequences, i.e., microbial genomics, offers the possibility of accelerating research into the mechanisms of action of probiotic bacteria.^8-10

Genomics of Probiotic Bacteria

Research in microbiology has remarkably changed during the last decade, largely due to the availability of novel whole-genome sequencing approaches. In fact, the decoding of the genome sequences of more than 1,000 bacteria, as currently present in the NCBI database (www.ncbi.nlm.nih.gov) has greatly advanced our understanding of bacterial biology. The initial microbial genomics efforts were mainly directed toward decoding the genomes of pathogenic bacteria because of their impact on human well-being. The obtained genomic data have opened new avenues of research and even sparked the origin of a new genomics-based discipline, called pathogenomics, which aims to understand the genetic basis of bacterial pathogenesis.¹¹ Recently, genome sequencing has also directed its interest toward food-related bacteria, intestinal commensals and probiotic bacteria. In 2009, a correspondingly novel discipline designated as probiogenomics was coined, which aims to provide insights into the diversity and evolution of commensal/probiotic bacteria and to reveal the molecular basis for their health-promoting activities.⁹ The public availability of full genome sequence data has significantly expanded our understanding of the biology of these microorganisms and has generated an enormous amount of information on metabolic capabilities, genetics and phylogeny of these bacteria.

Currently, the genus Bifidobacterium includes 37 species (for a review see Ref. 10). However, at the time of writing only 11 completely sequenced bifidobacterial genomes were publicly available^12-17 (Table 1), with genome sequences of another 13 strains still unfinished (NCBI source). Notably, for a small number of cases such as B. bifidum, B. longum subsp longum,^12,13 B. animalis subsp lactis^15,16 two or more genome sequences are publicly available.

Table 1. General features of sequenced Bifidobacterium and Lactobacillus genomes.

Species	Genome size (bp)	%GC	Gene numbers
Bifidobacterium strains
B. longum subsp longum NCC2705	2,256,640	60%	1798
B. longum subsp longum DJ010A	2,375,792	59%	2061
B. adolescentis ATCC15703	2,089,645	59%	1701
B. animalis subsp lactis AD011	1,933,695	60%	1603
B. animalis subsp lactis Bl-04	1,938,709	60%	1631
B. animalis subsp lactis DSM 10140	1,938,483	60%	1629
B. bifidum PRL2010	2,214,656	62%	1791
B. bifidum S17	2,186,882	62%	1845
B. dentium Bd1	2,636,367	58%	2197
B. longum subsp infantis 157F	2,400,312	60%	2062
B. longum subsp infantis ATCC 15697	2,832,748	59%	2588
B. longum subsp longum BBMN68	2,265,943	59%	1878
B. longum subsp longum JCM 1217	2,385,164	60%	2009
B. longum subsp longum JDM 301	2,477,838	59%	2035
B. breve UCC2003	2,422,684	59%	1642
Lactobacillus strains
L. acidophilus NCFM	1,993,560	34%	1938
L. casei ATCC334	2,895,264	46%	2909
L. gasseri ATCC 33323	1,894,360	35%	1898
L. johnsonii NCC533	1,992,676	34%	1918
L. plantarum WCFS1	3,308,274	44%	3135
L. fermentum IFO 3956	2,098,685	51%	1912
L. salivarius UCC118	1,827,111	32%	1864
L. amylovorus GRL 1112	2,067,702	38%	2126
L. brevis ATCC 367	2,291,220	46%	2314
L. casei BL23	3,079,196	46%	3090
L. casei Zhang	2,861,848	46%	2906
L. crispatus ST1	2,043,161	36%	2100
L.delbrueckii subsp bulgaricus ATCC11842	1,864,998	49%	2184
L.delbrueckii subsp bulgaricus ATCC BAA-365	185,695,1	49%	2033
L.delbrueckii subsp bulgaricus ND02	2,125,753	49%	2177
L. helveticus DPC 4571	2,080,931	37%	1838
L. johnsonii FI9785	1,755,993	34%	1780
L. plantarum JDM1	3,197,759	44%	3029
L. plantarum ST-III	3,254,376	44%	3137
L. reuteri DSM 20016	1,999,618	38%	2027
L. reuteri JCM 1112	2,039,414	38%	1901
L. rhamnosus GG	3,010,111	46%	2985
L. rhamnosus Lc 705	2,968,598	46%	2954
L. sakei 23K	1,884,661	41%	1963

In contrast, the emphasis on genomics efforts have been firmly placed on the genus Lactobacillus with more than 26 genomes completely decoded (Table 1). This larger number of sequenced genomes of lactobacilli (as compared with bifidobacteria) may be a reflection of a larger number of lactobacilli being included as active ingredients in functional foods. Specific probiotic strains have been sequenced, such as those that belong to the species Lactobacillus acidophilus, Lactobacillus casei, Lactobacillus rhamnosus, Lactobacillus fermentum, Lactobacillus gasseri, Lactobacillus johnsonii, L. plantarum, L. salivarius and L. reuteri (for review see refs. 9 and 18). Genomics data has significantly enhanced and will continue to improve our knowledge on the functionality of various Bifidobacterium and Lactobacillus species. Furthermore, the acquired genomic information has also provided clues as to how bifidobacteria and lactobacilli have adapted to the GIT environment and how they interact with their host (see below).

Comparative genome investigations involving the bifidobacterial strains for which the genomes had been completely decoded revealed that the deduced bifidobacterial pan-genome consists of more than 5,000 genes.¹⁹ The function of many bifidobacterial genes is still unknown but one would imagine that some of these have to be crucial for colonization of and survival in the GIT. Moreover, a set of genes shared by all sequenced bifidobacterial genomes was identified and this represents a presumed core genome of 967 genes, mostly corresponding to housekeeping functions (i.e., cell envelope biogenesis, replication, transcription, translation and signal transduction).¹⁹ This extensive comparative analysis also allowed the identification of genes that were present in just a single genome, while absent in any other currently available bifidobacterial genome [truly unique genes (TUG)]. TUG numbers ranged from 21 to 230 in each of the nine genomes analyzed. The majority of TUG have unknown functions and such genes are thought to perform a specialized function for these bacteria. Thus, the Bifidobacterium genomes display a relative high level of conservation, which is also coupled to a reasonable degree of genome synteny.⁹ Conversely, the Lactobacillus chromosomes reflect the high heterogeneity at phylogenetic, phenotypic and ecological levels among the various members of this genus.²⁰ Sequencing of LAB genomes has indicated that loss and decay of ancestral genes has played a key role in the evolution of Lactobacillales, a taxon that diverged from its Bacillus ancestor with an estimated loss of 600–1,200 genes of its total gene repertoire.²¹ A large proportion of these lost genes encoded biosynthetic enzymes and functions involved in sporulation.²¹ Nevertheless, in addition to major gene decay occurrences, gene acquisition events also occurred which might be a consequence of the nutrient-rich niches occupied by the LAB, such as milk and the GIT. Gene duplication events in lactobacilli have occurred involving genes encoding peptidases and amino acid transport proteins, as well as involving genes implicated in the transport and metabolism of carbohydrates.²¹ Furthermore, comparative analysis between GIT-associated species, such as L. acidophilus, L. gasseri, and L. johnsonii, on the one hand, and dairy species, such as L. helveticus and L. bulgaricus, on the other, has revealed genetic changes that seem to have occurred as a result of niche-specific selective pressure, which appears to have driven genome evolution of these individual species.^22-24

In addition to gene duplication, lactobacilli are also believed to have acquired new genetic information through HGT (horizontal gene tranfer), which is obvious from the analysis of such genomes. Genes encoding cell surface factors in L. johnsonii and the exopolysaccharide cluster in the L. acidophilus complex are clear examples of presumed HGT in probiotic lactobacilli.^23,25

The differences between the genomes of bifidobacteria and lactobacilli highlight their distinct phylogeny, but also reflect the different niches they occupy and the correspondingly niche-adjusted metabolic activities. In this context, the bifidobacterial genome content highlights its relatively broad prototrophy with respect to amino acids, nucleotides and vitamins.²⁶ In contrast, the genome content of lactobacilli reflects a high level of auxotrophy for such compounds.⁹

Adaptation of Probiotic Bacteria to the Human Gut

Probiogenomic investigations have highlighted a plethora of genetic features that may explain how bifidobacteria and lactobacilli have so well adapted to the human GIT. A key example of such an adaptation is represented by the carbohydrate-degrading capabilities of bifidobacteria, which consist of a large arsenal of enzymes involved in the metabolism of complex carbohydrates that are not digested by human enzymes and thus are expected to arrive in the lower regions of the GIT in an intact form. Dissection of bifidobacterial genomes suggests that a relatively large proportion of this genetic arsenal is involved in the metabolism and transport of carbohydrates, with several carbohydrolases predicted to be required for the utilization of various plant-derived dietary fibers or complex sugars.¹² Moreover, genetic and biochemical studies have been directed to analyze the capabilities of various bifidobacterial species to utilize diet-related carbohydrates, such as amylopectin, galactan, starch and pullulan.^27-29

Another example of how bifidobacterial genome data allows us to link the presence of particular genes to a specific ecological niche adaptation has been provided by publications focusing on bifidobacteria isolated from different environments, such as the infant gut (i.e., the case of B. longum subsp infantis ATCC15697¹⁴) and the oral cavity (i.e., the case of B. dentium Bd1¹⁷) or to a bifidobacterial strain that can utilize human mucin (i.e., the case of B. bifidum PRL2010³⁰).

Probiogenomics of Bifidobacteria

As mentioned above, various probiogenomic efforts have been undertaken in order to underpin the genetic and metabolic characteristics of selected members of the genus Bifidobacterium.^12,14,30,31

The genome sequence of B. longum subsp infantis ATCC 15697 contains features that explain the ability of this strain to consume specific human milk carbohydrates known as human milk oligosaccharides (HMO). In particular, the B. longum subsp infantis ATCC 15697 genome harbours a gene cluster that encodes various glycosyl hydrolases and carbohydrate transporters necessary for importing and metabolizing HMOs.¹⁴ This 43 Kb large gene cluster specifies a variety of catabolic enzymes such as fucosidase, sialidase, β-hexosaminidase and β-galactosidase activities, as well as extracellular solute binding proteins and permeases predicted to be active on HMOs.¹⁴ Furthermore, the genome of this microorganism contains additional genetic loci specifying fucosidases and sialidases, as well as a complete urease operon, predicted to be involved in the utilization of urea, which represents an important nitrogen source of milk.¹⁴

Another important member of the bifidobacterial population frequently encountered in the infant gut microbiota is represented by the B. bifidum species.³² Members of this species are, among bifidobacteria, the most capable representatives to metabolise host-derived glycans, such as mucin.³³ Other human gut microbiota members including Bacteroides spp, Ruminococcus spp, Clostridium spp and Akkermansia muciniphila, have been identified as a major bacterial players in mucin degradation,^34-37 although relatively little is known with respect to the genetic elements required for this property. Mucin is the principal component of mucus gel that covers the GIT epithelium and it represents the first barrier between host and intestinal bacteria, as well as host and nutrients present in the gut.

Recently, the genome sequence of B. bifidum PRL2010 was fully decoded,³⁰ revealing novel insights into the metabolic strategies followed by this strain to metabolize mucin-derived carbohydrates. These investigations suggested the existence of specific B. bifidum enzymatic pathways involved in the utilization of host-derived glycans, for example by the activity of enzymes that remove sialic acid and fucose moieties from galacto-N-biose (GNB) and its extended derivatives present in various mucin O-glycans.^38-41 In addition, the action of an endo-α-N-acetylgalactosaminidase is predicted to release such galacto-N-biose-containing glycans from the mucin glycoproteins and once released it may undergo further degradation by the extracellular β-galactosidase and β-N-acetylhexosaminidase, before GNB and other degradation products are translocated across the cell membrane to the cell cytoplasm where, depending on their chemical conformation, they are subjected to further hydrolysis, phosphorylation, epimerization, desulphation and/or deacetylation.

Another clear example of how analysis of genomic data underpins specific adaptations of bifidobacteria to the human GIT is represented by the genome sequencing of another key component of the infant gut microbiota, Bifidobacterium breve UCC2003.³¹ Genome mining of this strain revealed information regarding its genetic adaptation to the colonization and persistence in the human gut through the production of fimbria-like structures belonging to the type IVb (or Tad) pili-family. Mutational analysis demonstrated that the UCC2003 tad gene cluster is crucial for efficient in vivo gut colonization in murine models, while immunogold transmission electron microscopy confirmed the presence of Tad pili at the poles of B. breve UCC2003 cells.³¹ Notably, the Tad pilus-encoding locus was shown to be highly conserved among sequenced Bifidobacterium genomes, thus suggesting the notion of a ubiquitous pili-mediated host colonization and persistence mechanism for bifidobacteria.³¹

Probiogenomics of Lactobacilli

In silico analyses of the genomes between classical intestinal lactobacilli (e.g., L. rhamnosus) and plant or milk isolates (e.g., L. bulgaricus and L. helveticus) have demonstrated functional groups representing their niche adaptation. In this context, the typical milk-adapted L. bulgaricus and L. helveticus genomes^24,42 contain an arsenal of genes that encode enzymes dedicated to the metabolism of typical milk-derived sugars and other carbohydrates.⁹ A clear sign of adaptation of the human GIT is represented by the enrichment of mucus-binding proteins and enzymes that are predicted to be involved in breakdown of complex carbohydrates.^43,44 In addition, specific adaptation to the human intestine is also evident from the existence of a bile salt hydrolase (BSH) encoded by all sequenced intestinal lactobacilli.⁴⁵ Gut-adaptation functions are not only encoded by chromosomal DNA but also by large extrachromosomal replicons such as megaplasmids. The first megaplasmid described in lactic acid bacteria was that of L. salivarius UCC118, representing almost 11% of the overall coding capacity of the L. salivarius genome.⁴⁶ This megaplasmid was shown to encode biologically important characteristics including a locus for bacteriocin production, a bile salt hydrolase-encoding gene, and two genes that complete the phosphoketolase pathway.⁴⁶

Comparative genome analyses within the L. plantarum species revealed the existence of a DNA region, named life-style cassette, encompassing genes predicted to be involved in sugar metabolism (represented by PEP-PTS systems as well as glycosyl hydrolases).⁴⁷

Interaction of Bifidobacteria and Lactobacilli with Their Host

So far, little is known about the genetic basis of interactions between probiotic bacteria and the intestinal host mucosa. Human gut commensals are known to synthesize cell envelope-associated structures, which are claimed to sustain an important role in determining microbe-host interactions (for a review see ref. 48). All sequenced genomes of bifidobacteria and lactobacilli are predicted to encode an extracellular polysaccharide (EPS) or capsular polysaccharide, and such an extracellular structure may be important in bacterial colonization or adherence to host cells, while it could also contribute to resistance to stomach acids and bile salts.^49,50 Moreover, other predicted cell surface-encoding proteins are the sortase-dependent fimbriae-like structures, which are encoded by the genome of enteric,¹² as well as oral bifidobacteria.^17,31,51 The precise role played by these structures in bifidobacteria has not yet been determined, with the exception of the Tad pili as discussed above. However, in other human GIT commensals, such as L. rhamnosus GG, the sortase-dependent pili have clearly been shown to mediate microbial adhesion to and colonization of the epithelial mucus layer.⁵²

Other important mediators contributing to the host interaction in the GIT are represented by serpin-like protease inhibitors, which are encoded by B. longum subsp longum NCC2705 and B. breve UCC2003.^31,53 The serpin encoded by B. longum subsp longum NCC2705 is an efficient inhibitor of human neutrophil and pancreatic elastases, whose release by activated neutrophils at the sites of intestinal inflammation represents an interesting control mechanism of innate immunity.⁵³ A recent survey on the distribution of the serpin-encoding gene in bifidobacteria indicates the presence of this gene in seven different bifidobacterial species (B. longum subsp longum, B. longum subsp infantis, B. longum subsp suis, B. breve, B. dentium, B. scardovii and B. cuniculi), three of which, i.e., B. longum subsp longum, B. longum subsp infantis and B. breve, are commonly encountered within the human gut microbiota.⁵⁴ The presence of such a protease inhibitor may provide an ecological advantage to bifidobacteria since serpin activity may protect them against host proteases.⁵³ The observation of transcriptional activation of the serpin-encoding gene represents a molecular mechanism for immune-modulation, triggered by particular members of intestinal bifidobacteria.⁵⁴

The diversity of cell envelope composition and extracellular structures provides species- and strain-specific features that are most likely driving microbe-host responses. For example, genome analysis of L. plantarum WCFS1 revealed several secreted proteins that are predicted to be involved in adherence to host components including mucins and collagen.⁵⁵ In a similar manner, genome analysis of L. acidophilus NCFM suggests the existence of adhesins that may be involved in binding to host glycans such as mucins.⁵⁶

Genome Evolution of Bifidobacteria and Lactobacilli

In silico analyses of currently available genome sequences of probiotic bacteria has revealed some generally conserved genetic traits (for reviews, see refs. 9 and 10) that may reflect adaptation of these bacteria to the human intestinal niche. Nevertheless, since probiotic bacteria such as bifidobacteria and lactobacilli represent diverse and taxonomically heterogeneous groups of microorganisms, the analysis of gene presence/absence patterns in a particular set of genomes, may be dramatically influenced by the evolutionary distance between these two distant taxa. However, common evolutionary pathways that have been followed by bifidobacterial and Lactobacillus genomes may be identified. These include the loss of genes encoding biosynthetic enzymes, gene duplication and horizontal gene transfer (HGT). From an evolutionary perspective, it must have been crucial for various bifidobacteria and lactobacilli, some of which being exploited as probiotics, to enlarge their genetic arsenal (gene duplication and HGT) in order to successfully colonize the human intestine and to compete with other members of the autochthonous microbiota. Many genes involved in sugar metabolism and transport appear to be duplicated or acquired early in the evolution of bifidobacteria and lactobacilli, including those encoding enolase, β-galactosidase and many other glycosyl hydrolases.⁹ Furthermore, the increase of the number of genes encoding peptidases and amino acid transporters has occurred in several species of bifidobacteria and lactobacilli. Another protein family, frequently found in the genomes of lactobacilli, is presented by the gene products that sustain antibiotic resistance vs other bacteria, i.e., β-lactamases.⁵⁷

With the availability of a growing number of whole genome sequences from bifidobacteria and lactobacilli that have probiotic properties, an important future challenge will be to identify the hypothetical core probiogenome, representing core genome functions of probiotic bacteria. Nevertheless, only seven genes present in the bifidobacteria, but absent in the genomes of other members of the Actinobacteria phylum, are shared with lactobacilli. Only one of these genes, which encodes a functionally uncharacterized membrane protein, is present in all the lactobacilli genomes so far sequenced.²¹

Bioengineering of Probiotic Bacteria and Probiogenomics

Recently, the field of probiotics has embraced the application of bioengineering, which aims to develop “designer probiotics,” for instance by expressing receptor-mimic structures to circumvent pathogens by blocking crucial ligand-receptor interactions.⁵⁸ A good example of such a novel type of bioengineered probiotic is represented by an Escherichia coli strain producing a lipopolysaccharide coupled to a shiga toxin receptor that is able to bind and neutralize toxins in the lumen of the intestine, thus avoiding/limiting adhesion of pathogens to the gut mucosa.⁵⁹ Engineered probiotics include also genetically manipulated Streptococcus gordonii strains and Lactobacillus jensenii strains that have been applied to combat HIV.⁶⁰ Furthermore, engineered probiotics provide a useful way to solve technical problems such as low rate of cell survival of probiotics during the shelf life of the product due to environmental stresses such as cold stress or oxidative stress.⁶¹

However, despite a higher functionality of engineered probiotics as compared with their natural counterparts, the use of such recombinant bacteria in the food chain will meet with significant reluctance expressed by consumers, while they will also have to overcome regulatory requirements as imposed by governmental authorities. As outlined by Steidler et al.⁶² more rigorous scientific studies are required, which should include a careful evaluation of the genetic contents of engineered bacteria and a thorough functional genomics examination. In this context, probiogenomics should represent a mandatory step in the procedure to achieve development and regulatory approval of novel engineered probiotic bacteria.

Conclusions

Almost all probiotic lactobacilli and bifidobacteria that are currently on the market were originally selected based on technological stability, such as resistance and stability during food processing and storing, or on some easily measurable phenotype like the ability to tolerate bile salts or survive GIT passage, but not necessarily for their ability to impart health benefits on the human host. At this point in time, the regulatory requirements regarding probiotic products have shifted toward the need for understanding the precise molecular mechanisms by which probiotic bacteria beneficially influence human health. Characterization through so-called “omics” approaches involving genomics and functional analyses may be a route to satisfy such a regulatory requirement. Moreover, the in-depth knowledge on the composition and functionality of the human gut microbiota will provide molecular criteria that predict susceptibility of individual subjects to specific probiotic supplementation and may be utilized as an a priori criterion for successful probiotic therapy.