While lactic acid bacteria and bifidobacteria have been scientifically important for over a century, many of these are marketed today as probiotics and have become a valuable and rapidly expanding sector of the food market that is leading functional foods in many countries. The human gastro-intestinal tract with its various compartments and complex microbiota is the primary target of most of these functional foods containing lactic acid bacteria and bifidobacteria (LAB&B). In addition, their use as vectors for delivery of molecules with therapeutic value to the host via the intestinal tract is being studied. This review focuses on molecular approaches for the investigation of the diversity of lactic acid bacteria and bifidobacteria in the human intestine, as well as tracking of probiotic bacteria within this complex ecosystem. Moreover, methodologies to determine the viability of the lactic acid bacteria and bifidobacteria and molecular approaches to study the mechanisms by which they adapt, establish and interact with the human host via the digestive tract, are described.
Scientific interest in the lactic acid bacteria and bifidobacteria can be traced back over a century to the pioneering activities of Louis Pasteur, Ilya Mechnikov and Henri Tissier. Pasteur's work at the end of the 19th century illustrated that lactic acid fermentation was due to microorganisms while solving a failed wine production in which lactic acid bacteria replaced the alcohol fermentation of yeast. Mechnikov in fact was most famous for describing phagocytosis, but he proposed the ingestion of lactic acid bacteria in order to promote human health, and essentially founded the probiotic concept in the early days of 1900. At the same time, Tissier discovered the bifidobacteria and speculated about their use as infant probiotics. Probiotics have now been defined as live microorganisms that, when administered in adequate amounts, confer a health benefit on the host . Today, fermented dairy foods supplemented with probiotics have grown into a multi-million Euro business . In most cases, these functional foods include lactic acid bacteria and bifidobacteria (collectively termed LAB&B for this review) that are marketed as probiotics. This explains the large interest in the survival and activity of LAB&B in the human intestine. The food industry and academia continue research in order to understand the physiology, interactions and mechanisms by which probiotic LAB&B provide benefits to the host via the human digestive tract [1,3,4].
The human digestive tract harbours a wealth of niches with many microbial ecosystems that vary according to location of the intestinal tract. Many members of the lactic acid bacteria as well as bifidobacteria naturally form part of this dynamic ecosystem. These include the so-called autochthonous or indigenous ones, as well as the allochthonous LAB&B that are acquired from the environment of which ingested food is presumably the main source. The main lactic acid bacteria found in the human intestine comprise Lactobacillus and Leuconostoc spp., while Bifidobacterium spp. are dominant among the first colonizers of newborns and continue to persist at a low level in adults (see below).
Apart from being indigenous members of the human gut, lactic acid bacteria are found in a plethora of niches, including plant material, fermented dairy, vegetable and meat products and sour dough breads. Foods fermented by lactic acid bacteria are rendered safe by preservation and have improved textures, flavours and tastes. Hence, a variety of lactic acid bacteria, notably Lactococcus and Lactobacillus spp., are used as starter cultures for the production of fermented foods and persist in these in high numbers. However, many of the lactic acid bacteria that are ingested via consumption of these fermented foods do not survive passage of the human intestinal tract. Lactococcus lactis cells used as a starter for industrial cheese production, provide the largest load of living consumed lactic acid bacteria. Studies with a genetically marked Lactococcus lactis strain revealed approximately 1% survival in human volunteers followed by a rapid decline after consumption cessation that resembled a first-order kinetics as expected for a stressed population . Hence, specific strains of lactic acid bacteria that are used to enrich probiotic foods are chosen for their resistance to passage through the human gastro-intestinal tract. This also holds for bifidobacteria that are used as probiotics but as their primary niche is the digestive tracts of humans and other animals, survival is less of an issue than culturing them in industrial environments.
All lactic acid bacteria that include among others the genera Lactobacillus, Lactococcus, Leuconostoc, Pediococcus and Weisella belong to the phylum of the Firmicutes, and include a great number of species that is most numerous for the lactobacilli, for which over 80 species are known . The genus Bifidobacterium belongs to the phylum of the Actinobacteria and, as such, is not closely related to the lactic acid bacteria . However, all LAB&B are gram-positive bacteria that differ in the G + C content of their genomes: low G + C for the lactic acid bacteria and over 50% G + C content for bifidobacteria. Similar to the lactic acid bacteria, bifidobacteria are fermentative and produce several acids including lactate, and are predominantly catalase negative. Due to their phylogenetic relations, shared metabolic properties, and common food-grade status leading to their incorporation in functional foods, lactic acid bacteria are aligned with the bifidobacteria for the purpose of this review and the term LAB&B is used to collectively indicate both groups of bacteria.
During the last few years novel benefits are being discovered of adding specific LAB&B probiotic strains to our food [1,9–11]. This is not so surprising considering we have co-evolved together with our intestinal microbes over millions of years, and these have been programmed to manipulate networks of genes [12–14]. The intestinal microorganisms, collectively called microbiota, consist of at least 1013 microbes and are dominated by anaerobic bacteria, comprising over a 1000 species (see [14–16] for reviews and  for a recent inventory). The digestive tract offers a relatively non-hostile environment and supply of nutrients that is produced and consumed by the host. Studies with germ-free and conventional animals that are colonized by intestinal bacteria have shown that the microbiota contributes to diverse processes including roles in host nutrition, intestinal epithelial development and activity, and fat storage, educates the immune system, and maintains the integrity of the mucosal barrier amongst other functions [18–20]. Lactic acid bacteria and bifidobacteria, both the commensal or ingested members, play a role in the intestinal niche. The rising number of genome sequences of LAB&B and the application of molecular technologies, such as high throughput functional genomics, is already providing novel options to explore the environmental flexibility and adaptability of the LAB&B [21,22]. This paper reviews advances in molecular strategies to determine the diversity, activity and functionality of the LAB&B that are targeted to the microbial intestinal ecosystem for the purposes of improving human health.
2Culture-independent qualitative analysis of intestinal LAB&B
The proposed dietary strategies to influence the health and well-being of the host include the consumption of probiotics, but also prebiotics and synbiotics. Prebiotics are non-digestible food ingredients that are believed to be selectively metabolized by the indigenous LAB&B in the gut [1,23]. The combination of a probiotic with a prebiotic to support probiotic viability and activity has been termed a synbiotic. The effectiveness of these strategies to influence the composition and activity of lactic acid bacteria and bifidobacteria as well as their effect on the rest of the intestinal microbiota is of significant interest for scientific and industrial reasons. Some of the more recent developments in the tracking and identifying members of the LAB&B in complex intestinal samples using the 16S rRNA gene are summarized (Table 1). Several recent reviews are focusing on detecting and identifying lactic acid bacteria and bifidobacteria in the intestine [24–26]. Here, a brief overview of the current techniques for diversity analysis of the dominant microbiota is provided together with a concise picture of its microbial composition (for recent reviews see [15,17,27]).
Table 1. Potentials and limitations of various methods for investigating the diversity of LAB&B in the human intestine
Isolation of pure cultures, enumeration
Not representative for most LAB&B species; insufficiently selective media; time consuming selective
16S rRNA gene libraries and sequencing
Identification and phylogeny
Large scale cloning is laborious; primer bias can be an issue
16S rRNA/rDNA dot-blot hybridisation
Detection, quantification and activity
Gives information about LAB&B activity via ratio rRNA/rDNA; comprehensive set of LAB&B rRNA probes
FISH using 16S rRNA probes
Single cell detection and enumeration
High throughput with image analysis software and flow cytometry; comprehensive group LAB&B rRNA probes available
DGGE/TGGE of 16S rRNA amplicons
Rapid profiling of LAB&B
Relies on specific amplification; pairs semi-quantitative; identification by band extraction and sequencing validated LAB&B rRNA primer
Quantitative real-time PCR of unique genomic DNA or 16S rRNA
Detection and quantification
Requires probe/primers design and validation; various LAB&B genomes; high throughput
A broad range of molecular techniques has become available for the identification, composition and enumeration of the total bacterial community of the intestinal tract (Table 1). Most of these are based on the 16S ribosomal RNA (rRNA) gene. During the last decade, the 16S rRNA gene has revolutionised the manner by which taxonomists classify and identify bacteria. The 16S rRNA gene comprises highly variable to highly conserved regions, and the differences in sequence are used to determine phylogenetic relationships and distinguish bacteria at different levels from species to domain. Databases of over 200,000 16S rRNA genes are available, for example the Ribosomal database project – RDP (http://rdp.cme.msu.edu/html), and EMBL (http://www.embl-heidelberg.de/). These databases enable new 16S rRNA sequences to be compared with existing sequences. The sequencing of 16S rRNA gene libraries of human intestinal microbiota, generated by PCR amplification of the 16S rRNA gene of DNA from human faeces and mucosa-associated bacteria, has revealed that the diversity of the intestinal microbiota had been vastly underestimated [17,28,29]. In the 1960s and 1970s very thorough composition studies were performed on human faeces . Nevertheless, the challenges of obtaining pure cultures, due to the largely anaerobic nature of this community and designing suitable enrichment strategies to simulate intestinal conditions, needed circumvention. Recent database analysis has shown that more than 80% of the 16S rRNA sequences retrieved from the human intestinal tract represent uncultured bacteria [12,17]. A combination of sequence analysis of 16S rRNA gene libraries [17,28,29] and fluorescent in situ hybridization (FISH) approaches targeting the 16S rRNA [31,32] has shown that the most abundant bacterial groups in the human intestine belong to, in order of numerical importance, the phyla of the Firmicutes (including the large class of Clostridia and the lactic acid bacteria), Bacteroidetes, Actinobacteria (including Colinsella and Bifidobacterium spp.) and Proteobacteria. Remarkably, this analysis revealed that the adult intestinal microbiota constitutes of a majority of low and high G + C content gram-positive bacteria. This has been indirectly confirmed by recent analysis of the metagenome of bacterial viruses (bacteriophages) recovered from faecal samples that revealed an abundance of viral sequences with sequence similarity to genomes of bacteriophages specific for gram-positive bacteria . Culture-independent methods have aided in identification and culturing of many novel new intestinal members, some of which are quite numerous in the intestine. This includes the recently discovered phylum of Verrucobacteria that appears to be represented in the human intestine by a single phylotype of Akkermansia mucinophila, isolated for its capacity to utililize mucin as carbon, energy and nitrogen source [16,34].
The vast majority of diversity studies on intestinal lactobacilli has been performed on the accessible faecal samples from healthy individuals. The 16S rRNA libraries of adult faecal samples demonstrated L. ruminis to be a predominant Lactobacillus intestinal species, but many other species were also detected including L. crispatus, L. gasseri, L. plantarum, L. acidophilus, L. delbrueckii, L. casei, L. paracasei and Leuconostoc argentinum[35,36]. Furthermore, characteristically food-associated bacteria such as L. sakei, L. curvatus, Leuconostoc mesenteroides and Pediococcus pentosaceus were also detected. Adult caecal samples contained L. ruminis, L. gasseri, L. vaginalis and Leuconostoc mesenteroides. Infant faecal samples showed various species but especially L. acidophilus, L. casei/paracasei and L. salivarius[36,37]. These data essentially conform to previous culturing studies with the exception of L. reuteri, which was quite frequently cultured from human faeces , but so far has not been detected by molecular methods. While lactobacilli are considered to be culturable (after all, they grow and ferment many foodstuffs), culture-independent 16S rRNA clone libraries generated by Lactobacillus group-specific PCR, indicate the presence of novel, not yet cultured, Lactobacillus species especially within the human intestinal tract (; see below). Recently described alternative incubation conditions to culture some intestinal lactic acid bacteria, hitherto uncultured, resulted in more easy selection of potentially food-associated ones, such as L. sakei and Leuconostoc mesenteroides on agar plates .
Diversity studies of lactic acid bacteria in intestinal samples other than faeces have been obtained from biopsies following endoscopy of patients, samples from sudden death victims, or capsules that upon swallowing open and sample the intestinal contents. The predominance of culturable lactobacilli (and streptococci) in the stomach and especially the small intestine has been reported in many reviews [30,38,40,41]. However, close inspection of the original data reveals that this abundance of lactobacilli is highly variable between individuals while verification of the isolated lactobacilli has not, or only to a very limited extent, been performed [38,42,43]. Hence, the vertical distribution of lactobacilli and other microbial communities in the human intestine needs to be readdressed using modern and molecular approaches. A recent molecular study on three subjects indicated the relative predominance of lactobacilli for the stomach and the upper segment of the duodenum in two of the subjects (Zoetendal, E.G. and von Wright, A., unpublished data). Presumably the subject, precise region of the small intestine, and manner of sampling (biopsy versus ingesta) will influence the numbers of lactobacilli in the small intestine relative to other groups such as Bacteroides and Bifidobacterium spp. In contrast to lactobacilli, data on bifidobacteria are lacking from the classical studies since the appropriate selective media had not yet been developed. However, bifidobacteria can be an abundant group in the caecum, colon and faecal samples as will be discussed below.
3Culture-independent enumeration analysis
Enumeration of LAB&B and their effects on the inherent microbiota in intestinal samples is being used to answer many questions such as LAB&B probiotic survival and colonisation, effects of diet including prebiotics on the native LAB&B as well as those ingested, and furthermore effects of medication, disease and lifestyle. Presently, approximately 20 dominant phylogenetic groups in the human microbiota can be enumerated by a comprehensive set of probes using FISH of the 16S rRNA gene [31,32]. FISH involves whole cell hybridisation with fluorescent oligonucleotide probes targeted against specific bacterial groups and species (see Table 1). Studies with these probes are beginning to reveal some trends in typical microbiota composition of infants, adults and elderly. The first 16S rRNA oligonucleotide probe designed already ten years ago in order to perform FISH, was Bif164 targeting the human faecal bifidobacteria . Application of this probe to faecal samples and comparison to culturing results indicated that bifidobacteria in faeces were for the most part culturable, thus supporting the predominant status of the bifidobacteria within the human intestine . However, since the total culturable counts were only a fraction of the total microscopic counts, the contribution of bifidobacteria to the total intestinal microflora had been over-estimated by almost 10-fold. Today, following extensive studies in faeces of the north European adult population, bifidobacteria are estimated to comprise 4.4 ± 4.3% of faecal microbes . In the majority of infants, bifidobacteria become dominant during the first weeks of life. Both human and formula milks are effective in selecting for high levels of bifidobacteria in infant faeces, typically 40–60% of the total microbiota, with the lower proportion found in formula-fed infants while some breast-fed infants harbour up to 90% bifidobacteria . Some infants have no detectable bifidobacteria, as determined by both molecular and culturing studies, and this does not appear to impact on their health [37,46]. Molecular identification of tentative bifidobacterial colonies on various selective media showed that the traditional media were insufficiently selective and unsuitable for quantitative analyses. The diversity studies have stimulated the development of infant formulas enriched with specific prebiotic oligosaccharides that are bifidogenic, i.e., promote the abundance of bifidobacteria. Use of the present infant formulas results in a similar abundance of bifidobacteria in faecal samples as is found in the faeces of human milk-fed babies . Likewise, synbiotic formulations are being designed that increase bifidobacteria in the elderly, who show a marked reduction in bifidobacteria [48,49].
The design of a Lactobacillus-specific FISH probe appeared to be a challenge due to the non-monophyletic nature of this group. The developed LAB158 probe was found to hybridize to lactobacilli and enterococci, and effective detection often requires permeabilization of the cells prior to hybridisation . The northern European mean for adults is 1.8 ± 1.4% of total human faecal microbiota, which is considerably less than for the bifidobacteria .
In initial FISH studies, fluorescence microscopy and image analysis software was used to enumerate the LAB&B cells in complex samples . Many laboratories still use this approach since it is reliable, has been automated, and can be outsourced. However, FISH has also been adapted for use with the flow cytometry (FCM), which allows for more high-throughput analysis [32,52] (Table 1). More recent studies focus on optimisation of the procedures. For example, bifidobacteria are sensitive to the preparation procedure for FISH, and it is recommended to prepare and store faecal samples using 4% paraformaldehyde fixation and storage at −70 °C after which the samples are stable for up to 8 months . The FISH technique is currently the most advanced technique for enumeration of faecal microbiota, in terms of the numbers of probes available and speed of analysis. It is of course dependent on the availability of the 16S rRNA gene sequence in the databases, and effective probe design and validation (Table 1). Enumeration of the major faecal groups as well as bifidobacteria and lactobacilli during a clinical trial to study the efficiency of consumed probiotics to maintain ulcerative colitis patients in remission was recently achieved with FISH-FCM . This technique will certainly enable further studies on the relationship of diet and disease on the general microbiota and their LAB&B members.
4Molecular profiling techniques for diversity analysis
Analysis of 16S rRNA gene PCR products by denaturing gradient gel electrophoresis (DGGE), a semi-quantitative fingerprinting technique adapted for faecal samples has been widely used to rapidly monitor the microbiota community shifts and compare the communities between different persons, different intestinal locations and due to diets (Table 1). The DGGE technique and variations on it, such as temperature gradient gel electrophoresis (TGGE) or combinations of DGGE and TGGE, are based on 16S rRNA sequence-specific melting behaviour of PCR products, generated with primers one of which contains a 40-bp GC clamp, of complex mixtures of microbes (see review ). Statistical software enables the calculation of similarity indices and cluster analysis to compare the samples. DGGE has been used in numerous applications to study temporal variation of LAB&B, due to effects of diet or probiotic consumption (for reviews see [4,15]). Bands originating from bifidobacteria may often be visualised on DGGE gels since bifidobacteria are usually dominant in humans. The appearance of bifidobacteria in infant faeces within the first days of life, and their reduction in numbers due to withdrawal of milk following weaning, has been visualised clearly using DGGE profiles of the total microbial community . Bands originating from lactobacilli in faecal samples could not be detected on the DGGE profiles since they represent less than 1% of the community, which is approximately the detection limit of this method.
In order to overcome this limitation and focus on the LAB&B, specific PCR primers have been developed and applied for the lactobacilli and bifidobacteria groups [35,36,57]. These allowed tracking of these groups as well as the specific ingested probiotic LAB&B in faeces over time, due to diet or antibiotic treatment [35–37,56–58]. Comparison of DGGE profiles of specifically bifidobacterial strains has suggested their vertical transmission from parents to offspring , confirming earlier studies that were based on a culturing approach and diagnostic PCR on colony DNA . Lactobacillus-specific PCR-DGGE has also been applied to fermented foods . Similarly, this approach been used to characterise the mucosa-associated lactobacilli populations along the human colon . These were found to be similar to the faecal populations, and not significantly different between healthy persons and those with ulcerative colitis . Remarkably, in the latter study bands originating from L. gasseri were found in colonic biopsies of nine out of ten individuals. The possibility to identify the bands observed in the DGGE gels is a strong advantage of this method. Further studies with specific DGGE profiles showed a rather simple population for bifidobacteria along the colonic mucosa, while the lactobacilli profiles were complex and varied with host and sampling location .
Another community fingerprinting technique, terminal restriction fragment length polymorphism (T-RFLP), has been adapted for characterising the human faecal bifidobacteria, as well as the tracking of probiotic Lactobacillus strains in intestinal samples [62–64] (Table 1). A novel phylogenetic assignment database for the T-RFLP analysis of human faecal microbiota (PAD-HCM) has been designed, which enables the prediction of the terminal-restriction fragments at the species level . This research will facilitate and enhance the use of this technique in studies of dietary and probiotic effects on the microbiota. Besides the fingerprinting techniques, a whole range of specific 16S rRNA primers has been designed and applied for identification and distribution of bifidobacterial species in human intestinal samples, which are described in recent reviews [27,65].
5Diversity analysis techniques for the future
More recently, real-time quantitative PCR of the 16S rRNA gene is being developed for the detection and quantification of human intestinal bifidobacteria which has the advantages of being high throughput and measuring very low levels of bifidobacteria [48,65,66] (Table 1). Numerous real-time PCR based assays are being developed for the major groups within the faecal microbiota of humans, as well as lactobacilli and Bifidobacterium species [67–69]. These are been used for various applications such as comparison of healthy persons versus patients suffering irritable bowel syndrome . Besides real-time PCR of the 16S rRNA gene, the option to use the transaldolase gene of Bifidobacterium species has also been investigated and appeared to be superior to the former in quantifying bifidobacterial populations in infants . However, the transaldolase gene is also conserved in many other high GC gram-positive bacteria and hence is of less diagnostic value than the 16S rRNA gene.
DNA microarray technology advances rapidly and offers a fast, high throughput option for detection and estimation of the diversity of microbes in a complex ecosystem . These microarrays may be termed phylochips, microbial diagnostic microarrays or identification arrays. Their principle is based on the dot blot hybridization technique. Typically microarrays contain hundreds of oligonucleotide probes, usually based on the 16S rRNA gene, specific for different strains or species or genera of microorganisms that are detected in a single assay. There are many different forms of arrays to which the probes can be attached including macroarrays, and glass microarrays that are low to medium density, and very high density Affymetrix microarrays (>104 probes typically 25 mer per chip) . Studies are underway to apply microarray technology to the human intestinal microbiota . The application of this technology will undoubtedly impact on studies involving probiotics and the indigenous LAB&B in the future.
6Vitality of LAB&B in the intestine
While great progress has been made in methodology to determine the composition of LAB&B in the complex intestinal ecosystem, the interest in the actual activity as opposed to the presence of the LAB&B is increasing. It is generally believed that probiotics or LAB&B present in fermented foods must endure a harsh transit through the intestinal tract with different conditions depending on the location, which influences their viability. Hence, the LAB&B must adapt to the environment in order to survive. Moreover, the use of lactic acid bacteria and bifidobacteria as vectors for therapeutic delivery of molecules with targeted activity in the host is being investigated [74–76]. These bacteria appear capable of surviving and of being physiologically active at the mucosal surfaces in animal models. Biological containment systems are being developed for these genetically modified LAB&B to limit their activity to the host and allow their use in human healthcare .
Quantitative hybridisation with fluorescent rRNA probes is a useful indicator of activity as there is a correlation between the growth rate, which is coupled to efficient protein synthesis and the number of ribosomes. The FISH technique has been used to estimate growth rates of Escherichia coli cells in faeces . In situ activity of pure cultures of the human commensal L. plantarum WCFS1 strain has been measured by correlating the rRNA, as determined by fluorescence intensity, with the cell growth rate . However, at very high cell densities, a typical property of L. plantarum at late stages of growth, presumably changes in the cell envelope prevented effective entry of the probe into the cells. Permeabilisation issues may confound application of this technique to certain LAB&B in complex environments. Recently, the regulation of the five L. plantarum rRNA promoters were studied using a promoter-probe plasmid, which indicated despite significant differences between the five genes and differential expression over time, their activity was very similar under the circumstances tested . More research into the activity of the different rRNA operons within an LAB&B species under diverse conditions is required to signify a potential role in viability within the intestinal tract.
Recently for the first time the viability of bifidobacteria in faecal samples was assessed by combining a viability assay with flow sorting and subsequent identification by 16S rRNA analysis [70,80,81]. The faecal cells of four adults were initially discriminated with propidium iodide (PI) and SYTO BC into viable, injured and dead cells. This revealed that only approximately half of the microbial population in faecal samples is viable, while the rest suffered from injuries or were dead (each about 1/4 of the total population) . Subsequent bifidobacterial specific PCR-DGGE and 16S rRNA sequencing revealed that, while B. longum and B. infantis were present in all fractions, B. adolescentis was mainly retrieved from the sorted dead portion. In addition, the identification of 16S rRNA sequences with low similarity to the characterised species suggested the potential of as yet uncultured novel bifidobacteria species in humans. This interesting combination of technologies provides ecological information on the in situ diversity and activity of bifidobacteria. It can have more applications such as studying the vitality of LAB&B and probiotics in different subjects, and as a function of diet and lifetime.
The same study as described above for bifidobacteria has also been performed for the lactobacilli in faecal samples and is presented here. Following sorting by a flow cytometer, cells that were active, dead or injured were collected, lysed and prepared for molecular diagnostic analysis as described by Ben-Amor et al. . The results of the separation of Lactobacillus-specific PCR products or amplicons by DGGE (Fig. 1) illustrates that in many cases the amplicons with identical position were obtained from all fractions, indicating that this specific lactic acid bacteria constituted a mix of live, dead or injured cells. However, in some cases the amplicons were specifically found in the fraction of either the active or in contrast the dead fraction. The most abundant amplicons were cloned, sequenced and compared to the 16S rRNA gene sequences in the databases (Table 2). The results show that the nearest cultured representative in all cases belonged to the lactic acid bacteria confirming the specificity of the amplification . However, in about half the cases there was only limited similarity (i.e., <97%) to sequences of cultured lactic acid bacteria. This confirms the earlier observation that the human intestine harbours numerous and diverse Lactobacillus spp. that have not yet been cultured . It is possible that these constitute the species from which the cells were recovered predominantly in the dead or injured state from the faecal samples and hence have escaped culturing. However, this analysis also shows that several of these new Lactobacillus species are highly active as is indicated by the abundance of L. salivarius-, L. citreum- and L. ruminis-like species in the viable fraction. Remarkably, the 16S rRNA gene sequences of L. crispatus and L. plantarum show high similarity to amplicons that are found abundantly in the live fraction, indicating that these cultured species or their close relatives are clearly viable in the intestine.
Table 2. Phylogenetic analysis of live, dead and injured intestinal lactobacilli
Lactobacillus-like 16S rDNA amplicons (Fig. 1) were cloned in E. coli and transformants were selected for subsequent sequence analysis based on the migration position of the cloned PCR fragment in the DGGE in comparison to the fragments in the original DGGE profile as previously described . Insert PCR amplicons of selected transformants were purified and subjected to DNA sequence analysis (Greenomics, Wageningen, The Netherlands). All retrieved sequences were analysed using BLAST tool  at the NCBI database and similarities to cultured species are given. The numbering of the different PCR amplicons corresponds to that in Fig. 1.
7Reporter systems to identify and assess LAB&B gene expression in vivo
The survival and behaviour of probiotic LAB&B from food products once ingested into the intestinal tract is presently receiving much attention. The LAB&B adapt their metabolism to the environment by synthesising proteins whose activities are required for survival and colonisation. A genetic approach based on the fusion of bacterial promoters with genes of the reporter protein luciferase (luxA-luxB genes of Vibrio harveyi) was developed to investigate Lactococcus lactis gene expression in the mouse digestive tract . The latter is applicable to the study of promoter strength and physiology of bacteria in the digestive tract. Subsequently, Lactococcus lactis strains marked with reporter genes for luciferase and the green fluorescent protein were studied for their metabolic activity and survival by assessment of lysis, respectively, which revealed differential expression depending on the intestinal conditions and mode of administration . Following consumption by rats and analysis of the strains in the different regions of the intestinal tract, the lactococci were demonstrated to survive gastric transit quite well but the majority lost activity and underwent lysis in the duodenum. A similar study has been performed using luciferase fused to the promoter of the β-galactosidase (lacZ) gene of Streptococcus thermophilus to monitor expression in the digestive tract . Interestingly, the lacZ gene was expressed in the digestive tract, and this production was enhanced when lactose (the inducer) was added to the diet, although the bacteria did not multiply during transit.
The luciferase gene reporter system has also been applied to L. casei, which is present in fermented dairy products and is reported to have beneficial properties for human health. The live L. casei derivative was grown in milk and consumed by mice harbouring human microbiota (human-microbiota-associated mice), and the ability of L. casei to reinitiate synthesis in the different digestive tract compartments was demonstrated [85,86]. Luciferase activity was undetectable in the stomach to jejunum, but detected when the cells reached the ileum, and the activity remained at a maximum level in the caecum, confirming reinitiation of protein synthesis in the ileal and caecal compartments.
The in vivo expression technology (IVET) is a very elegant technique for identifying promoters of food-grade or commensal bacteria that are induced in vivo as in the intestinal tract of animals, and overcome many limitations of in vitro studies (reviewed by ). An IVET strategy based on in vivo selection of an antibiotic resistant phenotype for the rat intestinal L. reuteri strain led to the identification of 3 genes that are induced by this organism during colonisation of the digestive tract of Lactobacillus-free mice . Two of the gene sequences were similar to xylose isomerase and methionine sulfoxide reductase (MsrB) while the third was unknown. The msrB gene and that for the Lsp protein, implicated in adherence to epithelial cells, were disrupted, and competition experiments with the mutants and wild-type showed reduced ecological performance in the murine intestinal tract .
An alternative system termed recombination or resolvase-based IVET (R-IVET) has the advantage that even transiently expressed genes in the host can be identified, and also avoids the addition of antibiotics to the diet of the animal, which can have dramatic effects on the microbiota. The latter approach was applied to identify 72 genes induced in L. plantarum during passage through the mouse intestinal tract . The genes comprised several sugar-related functions, and acquisition and synthesis of amino acids, nucleotides, cofactors and vitamins, indicating their limited availability in the intestine. Also genes involved in stress-related functions were identified, and extracellular proteins that may mediate interactions with the intestinal epithelial cells. It is noteworthy that the one hypothetical protein found to be induced in L. plantarum was one of the 3 genes (the unknown) identified by the IVET in L. reuteri. Interestingly, many of the functions identified in the non-pathogenic commensal L. plantarum were previously identified in pathogens as being important in vivo during infection suggesting that survival rather than virulence is the explanation for the importance of these genes during host residence [22,89]. A selection of 12 of the 72 genes with a focus on the stress and cell envelope functionality were chosen to construct gene replacement mutants, and subjected to competition experiments by passage through the mouse intestine . Quantitative PCR analysis of the relative population dynamics of the mutants showed that the abundance of many mutants was significantly reduced (even 100- to 1000-fold) compared to the wild-type L. plantarum strain, highlighting their role in persistence.
8Genome sequences of LAB&B provide access to further technologies
The genome sequences that are becoming available for the LAB&B are considerably expanding our insight into the adaptation and functionality of these microbes in the digestive tract. Access to genome sequence information facilitates the above molecular strategies for investigating behaviour of LAB&B in the intestine but also provides opportunities for high throughput functional genomics techniques. There are already many genome sequences available for food-grade LAB&B on the Internet, and a few high-fidelity complete genome sequences have been published (see ). The first bifidobacteria genome sequence was that of B. longum and the first of the lactobacilli was that of a human isolate of L. plantarum followed closely by the genomes of the probiotics L. johnsonii and L. acidophilus. The first comprehensive comparative analysis of two Lactobacillus species, L. plantarum and L. johnsonii has been achieved  and more will be described (see review ). Here, the focus is on the impact of genomics for functionality analysis in the intestine. Especially the complete genome sequences provide opportunities for high throughput functional genomics techniques such as global transcript profiling, proteomics and metabolomics approaches. When applied to LAB&B, these approaches complement the diversity analysis discussed above (Fig. 2).
Diversity arrays based on the complete genome sequences of L. plantarum (, D. Molenaar et al., personal communication) and B. longum are being developed. These will prove useful for genetic diversity analysis and learning more about evolutionary relationships, as well as diagnostic purposes in the future. Besides the genomes of LAB&B, the complete genome of the gram-negative Bacteroides thetaiotaomicron, a member of the dominant colonic microbiota is shedding light on the molecular mechanisms underlying the symbiotic host-bacterial relationships in the human intestine . Complete genome sequences of commensal intestinal microbes will also provide greater insight into the survival strategies and growth in the intestinal tract. Whole genome transcript profiling has been performed on germ-free mice that were maintained on polysaccharide-rich or simple-sugar diets, and subsequently colonized with B. thetaiotaomicron. The microbes selectively induced enzymes for metabolism of the sugars, and upon paucity of the polysaccharides, the commensal microbe shifted its glycan foraging behavior to the host mucous as a nutritive source, a property which aids its stability in the intestine.
Recently the complete genome sequence of L. plantarum WSFS1 has allowed the design of specific DNA microarrays for use in transcript profiling . The L. plantarum species was found to have a high survival rate upon passage of the human stomach and intestine  This property was exploited in a study addressing the activity of L. plantarum in the human digestive tract by analysing RNA extracted from mucosa-associated cells and hybridising to WCFS1 microarrays or gene-specific quantitative real-time PCR (Fig. 2). Appropriate controls were performed and included the absence of specific hybridisation in biopsies from a subject that had not consumed the L. plantarum cells, the absence of interference by human nucleic acids on the microarray, and confirmation of the specificity by sequence analysis of the gene specific real-time PCR (M.C. de Vries et al., unpublished data). The ingested L. plantarum cells were found to be metabolically active in all subjects, and differences between gene expression between the individuals and intestinal location were apparent.
Besides commensal microbes, the lifestyle of the food-borne pathogen Campylobacter jejuni in the gut has recently been investigated using microarray technology . Inoculation of the wild-type cells and disruption mutants into the rabbit ileal loop model apparently resulted in remodeling of the cell envelope, and regulation of the heat shock and stringent response were found to be necessary for efficient colonisation of the tract. Furthermore, the expression of many genes appeared to be dependent on the individual rabbit. The latter phenomenon was also observed for L. plantarum in the human subjects study described above. The specificity of these arrays for the targeted microbial genes (as was confirmed by quantitative RT-PCR and sequence analysis of the PCR products; MC de Vries et al., unpublished data) despite the complexity of the background comprising epithelial cells and mucosa-associated microbiota, suggests a promising outcome for further similar studies.
Transcript profiling of L. plantarum WCFS1 has also been used to investigate the functionality of a locus, lam (Lactobacillusagr-like module), with homology to the staphylococcal agr quorum sensing system . Comparison of gene expression between a disruption mutant and the wild-type indicated that lam was involved in regulating genes for surface polysaccharides, cell membrane proteins and sugar-utilisation proteins, and moreover the mutant showed decreased adherence to glass. Interesting, the lam locus was demonstrated to encode a cyclic thiolactone pentapeptide, which is the first example of such a system in non-pathogenic bacteria.
An important factor determining the applicability of transcriptomics approaches is the speed by which human samples can be obtained and processed. In view of the short half-time of bacterial messengers, a proteomics approach may be more feasible, notably when only faecal samples are available such as in the case of babies and young children. A metaproteomics approach has been applied to the bifidobacteria-dominated human infant faecal microbiota to reveal the temporal microbial activity (E. Klaassens, personal communication). Among the proteins identified using MALDI-Tof-Tof from the two-dimensional gels was a match to a bifidobacterial transaldolase. The ongoing construction of metagenomic libraries and sequence analysis of human intestinal microbiota will enable a meaningful identification of more proteins in the future . A further focus on lactobacilli and bifidobacteria may be realised by applying first a specific sorting step based on LAB&B-specific 16S rRNA hybridisation followed by flow cytometric fractionation (Fig. 2).
The molecular tools described above have many applications for LAB&B in mixed communities or complex ecosystems, such as that of the intestinal tract. The review is an overview of some of the molecular strategies currently being used to investigate LAB&B in the intestinal ecosystem, and focuses on those approaches for in vivo experiments in complex environments. Understanding the relationship between the variable complex intestinal microbiota with their huge microbiome and the host is a phenomenal task, and requires the input of many disciplines. Further application and development of molecular technologies will enhance this and especially impact on functional foods or nutraceuticals that are designed to influence the microbiota. For example the composition and stability of LAB&B in non-European individuals still remains to be established and further approaches such as high throughput DNA microarrays or real-time PCR will aid this. While techniques for diversity analysis have been largely established, the emphasis is shifting towards approaches to study the vitality and the physiological adaptation of lactic acid bacteria and bifidobacteria, whether starter culture, commensal or probiotic, to the digestive environment. Perhaps contrary to initial assumptions, the intestinal environment may not be so antagonistic to at least some LAB&B, as deduced from their ability to survive, adapt and adjust their activity as assessed by reporters, gene expression profiles and R-IVET approaches.
Many of the molecular approaches, for example the construction of reporter systems or defined knock-outs, will be accelerated if the strain is genetically accessible, the challenge of which depends on the specific species and bacteria. Other technology will help to bypass or even solve this constraint. Complete genome sequences facilitate opportunities for high throughput functional genomics techniques such as global transcript profiling, proteomics, metabolomics and systems biology approaches. Great advances are also being made in bioinformatics, which will impact on the potential of the lactobacilli and bifidobacteria . In addition, analysis of the response of LAB&B isolates to intestinal conditions is greatly facilitated by genomics, as can be illustrated by the identification of bile salt responsive promoters in L. plantarums. Comparison of the behaviour of specific members of the commensal microbiota, specific LAB&B and intestinal pathogens, will provide novel insights on their individual roles within the intestinal ecosystem. Microarrays of DNA from metagenomic libraries of the microbiota will become feasible in the future for studying gene expression of the microbiota including the LAB&B members due to external influences such as diet or medication.
The greatest challenge concerns the functionality of the LAB&B, and the mechanisms by which they impact on host health, which are also necessary to strengthen health claims for probiotic LAB&B that are incorporated into functional foods. There has been much research on intestinal pathogens and the manner in which they communicate with the host, some of which involve signal transduction pathways in epithelial cells . The lines of communication between commensals and the host are also being unravelled and it has become more apparent that the human genome does not encode all the information to achieve full health. Thus, while this review has focussed on the effect of the host on the microbe, there are major developments taking place in the commensal microbes effect on the host as outlined in the introduction (see recent reviews [12–14,109]). Individual human microbiota activities are considered to be an essential part of future personalised health-care models, and the mammalian-microbe co-metabolomic studies will play a major role in defining this care . Collectively these studies are solving some of the puzzles of intestinal LAB&B life-style and influences on host health.
This study was partly supported by the EU Quality of life and Management of Living Resources, PROGID project (QLK1-2000-00563) and Microbe Diagnostics (QLK1-2000-00108). It does not necessary reflect its views and in no way anticipates the commission's future policy in this area. We are very grateful to MSc Maaike de Vries, MSc Eline Klaassens, Dr. Erwin Zoetendal and MSc Mark Sturme for sharing data prior to publication.