Commercially available media, such as MRS (Oxoid, Milano, Italy) and Bifidobacterial broth (Hi-Media Laboratories, A-516, Swastik Disha Business Park, via Vadhani Industrial Estate, L.B.S. Marg, Mumbai-400 086, India) and agar medium, were used to grow the different Bifidobacterial species. The first generation of the Bifidobacteria was cultured in MRS broth (and also in Bifidobacterial broth when appropriate) at 37°C (without shaking) in an anaerobic chamber for 36 h. They were sub-cultured and grown in aerobic condition in the second generation.

Bifidobacterial strains, B. animalis, B. longum, and B. adolescentis, were grown at different temperatures: room temperature, 30, 37, and 42°C. The bifidobacterial strains were streaked onto the MRS agar plate to produce single isolated colonies. The streaked plates were incubated at 37°C (for 36 h) under anaerobic conditions. A single colony was isolated, used to inoculate 10 ml of fresh MRS medium, and grown in an anaerobic growth chamber. An aliquot (1%) of each inoculum was subculture in 15 ml of fresh MRS medium to construct growth curves. The bacterial samples were collected every 6 h and the OD₆₀₀ measured. All the experiments were performed in triplicates.

Wild-type C. elegans were obtained from CGC and maintained on NGM at 20°C. The standard E. coli OP50 strain was used as a food source. Age synchronized young adult worms were used in all experiments. Synchronization was obtained through bleaching of gravid hermaphrodites to obtain eggs that were plated onto seeded plates. The young adult nematodes were harvested from NGM plates by using M9 buffer and used for experiments. Was grown on NGM agar plates and fed internationally accepted feed (the standard E. coli OP50 strain), as described previously (Ward, 1973). Standard techniques (Hope, 1999; Stiernagle, 1999) for the maintenance and growth of nematodes on NGM were used.

Escherichia coli OP50 was grown overnight at 37°C at 200 rpm in Luria-Bertani medium. The bacterial lawns used for the C. elegans studies were prepared by spreading 10 μl of the bacterial strain onto nematode growth media (NGM) agar (0.25% peptone) in 3.5-cm plates. The plates were incubated overnight in an incubator at 37°C. They were allowed to equilibrate to room temperature before they were seeded with C. elegans and incubated at 20°C throughout the experiment.

All the bifidobacterial strains cited in the Table 1 (except E. coli) were grown as shown above. Bacterial lawns were made by spreading 10 μl of the bifidobacterial strains onto MRS agar in 3.5-cm plates. A similar procedure to that described above was followed, but C. elegans was fed different Bifidobacterial (B. longum and B. adolescentis) strains rather than E. coli OP50. B. longum and B. adolescentis were considered further for these studies as it is preferred by the C. elegans. The second generation of this bacteria is microaerophilic therefore they can grow in aerobic condition, where C. elegans also grows.

To determine whether the Bifidobacterium and their products (ocins) are non-toxic and safe for C. elegans, several experiments were performed (Brenner, 1974). The procedure follows as below. A Petri plate was divided into four parts, two opposite and one adjacent part labeled “Test”(A, B, and D) and the remaining as “Control” (C). All the sectors (A, B, C, and D) were equidistant from one another. Part A contained B. animalis, part B B. longum, part C the control E. coli OP50, and part D B. adolescentis. Approximately, 25 wild-type C. elegans that had been grown on E. coli OP50 were washed thoroughly and introduced into the center of the plate and the origin was marked with a circle to allow non-motile worms to be eliminated. The plates were incubated under optimal conditions, and analyzed the following day. All the bacteria were equidistant from the center where the worms started their journeys. The numbers of worms in each bacterial sector were scored after 12 and 24 h and the chemotaxis index (CI) was calculated by using the following formula.

Caenorhabditis elegans was used as a model organism to determine the ability of the nematode to consume Bifidobacterium and to understand the intrinsic morphology of B. adolescentis. Different Bifidobacterium (B. animalis, B. longum, and B. adolescentis) were grown in the appropriate medium, as described above. They were harvested, washed, and resuspended in the appropriate medium. C. elegans were fed with different ratios (1:10, 1:20, and so on) of Bifidobacterium and left for 72–96 h. They were harvested at different intervals (after 12, 24, 32 h, and so on) without rupture, and then smashed so that the internal microbes could be observed. The worms were washed, fixed with glutaraldehyde, after a series of wash animals were crushed on the slide and analyzed with SEM. The phenotypic changes in the nematodes fed Bifidobacterium were observed microscopically. The cells of B. adolescentis-fed worms were also lysed, and the supernatants observed with SEM and the bacteria were quantified (CFU).

The method of Garsin et al. (2001), with minor modifications, was used to estimate the approximate numbers of bacterial cells in the worms. Ten worms were selected, and the surface bacteria were removed by washing the worms 4–5 times with 5 μl of M9 buffer on agar plates. Each worm was transferred to a 0.5 ml centrifuge tube containing 20 μl of M9 buffer and mechanically disrupted with a Microtube pestle (Scientific Specialties, Inc., Lodi, CA, USA). The volume was then adjusted to 100 μl with M9 buffer (serially dilute and spread) and the numbers of bacteria [colony-forming units (cfu)] were estimated by spreading the culture on MRS agar plates. The plates were incubated at the appropriate temperature until the colonies were observed. The same colonies were washed well with M9 buffer and immediately fixed and observed with SEM. Genomic DNA was isolated from some bacteria and subjected to 16S rRNA amplification and alignment.

Scanning electron microscopy studies were conducted to understand the diverse morphology of Bifidobacterium. Various Bifidobacterium species were subjected to different conditions, including different media (MRS or Bifidobacterium broth) and different temperatures (in MRS at room temperature, 30, 37, or 42°C for SEM and Growth curve studies). They were also grown in MRS medium (for three generations also), with or without amino acids (667 mg/lt) (alanine + asparagine + glutamine + serine, or individually with alanine, asparagine, tryptophan, etc.) in the presence and absence of CaCl₂ (Husain et al., 1972), and on MRS with and without salts (0.2 M Na₂SO₄, 0.2 M NaCl, or 0.2 M CH₃COONa) (Kojima et al., 1968, the effect of salt concentration studies) and at different pH (5–7.5). The pleomorphism of Bifidobacteria was understood by growing them in vivo (in C. elegans).

The various bifidobacterial strains were harvested, and washed twice with phosphate buffer. The pelleted bacteria were fixed in 2% glutaraldehyde and incubated at 4°C 1 h/overnight. The cells were then harvested and washed with a 10–100% gradient of ethanol. The cells were resuspended in 50–100 μl of absolute alcohol. Approximately 2 μl of the sample was placed on a cover slip, dessicated and examined with SEM. The images were collected at 20,000 magnifications. Only appropriate and convincing images were considered in the analysis.

The commercial GeneJET Genomic DNA Purification Kit (Fermentas, Inc., 830 Harrington Court, Burlington, ON L7N 3N4, Canada) was used for genomic DNA isolation. In brief, a single isolated colony was used to inoculate 10 ml of MRS broth and grown overnight at 37°C without shaking under anaerobic conditions. The following day, the cells were harvested and heat treated at 37°C in the presence of 20 μg of lysozyme in Tris- HCl and EDTA buffer (pH 8.0). After treatment with proteinase K, RNAse A was added and the cells were incubated for 10 min at room temperature. The resulting mix was centrifuged at 12,000 rpm for 20 min, and the supernatant was transferred to a GeneJET column. After two washes with wash buffer, the DNA was finally eluted with Tris buffer (pH 8.0).

The genomic DNA extracted as described above was used as the template for the PCR amplification of 16S rRNA using the bifido16S universal primers 357F/926R (357F 5′-CCTACGGGAGGCAGCAG-3′; 926R 5′-CCGTCAATTCMTTTRAGT-3′) and DyNAzyme II DNA polymerase (Thermo Scientific). The cycling parameters were: initial denaturation at 95°C for 5 min, followed by 35 cycles at 95°C for 30 s, 56°C for 45 s, and 72°C for 1 min, and a final extension at 72°C for 5 min. Each reaction was analyzed on 1% agarose gel. The resulting PCR products were subjected to DNA sequencing with an ABI 310 DNA sequencer (institutional facility). Sequence similarity was determined with the National Center for Biotechnology Information BLAST tool (NCBI).

Another five species, B. thermacidophilum (DSM 15837), B. infantis (DSM 20088), B. asteroides (DSM 20089), B. animalis (DSM 20105), and B. breve (DSM 20213) was also subjected to the different conditions described above, and the results are shown in the Supplementary Figure S1. We noted a marked change in the morphology of B. thermacidophilum from rod- to bifid-shaped when grown on medium containing tryptophan. When B. asteroides was grown at 37°C, a mixed population of rod- and bifid-shaped bacteria were observed. Morphological changes were also observed when B. breve was grown with all four amino acids or with alanine alone, when it metamorphosed from rod- to bifid-shaped cells. Bifidobacterium infantis changed from rods to coccoids at 42°C. Almost all the other strains of Bifidobacterium maintained the rod shape under all the conditions tested, except as noted above. These findings confirm that B. animalis maintains its rod form irrespective of the conditions, implying that this morphology is an intrinsic characteristic. Therefore, B. animalis may be used as a reliable control when investigating the bifid nature of B. adolescentis (in MRS, at 37°C, and under anaerobic conditions). B. adolescentis and B. animalis (a control for rod shape) were exposed to other variable conditions, including various pH (Figure 3), for different growth periods (Figure 4), and in different concentrations of salt (Figure 5). The bifid shape was maintained by B. adolescentis under all the conditions tested, whereas the control, B. animalis, retained the rod form. The bifid shape of B. adolescentis were also retained when it was grown for three successive generations under normal conditions (Figure 6), in all of which the bifid shape was retained. This further affirms that the intrinsic bifid nature in the case of B. adolescentis, and rod shape in the case of B. animalis.

We also found that most Bifidobacterium species stably maintained their morphological (rod) shape, irrespective of the conditions. As shown in the Figure 2, all the Bifidobacterium species were intrinsically rod-shaped, except B. adolescentis. Few of them adopted the bifid structure, and only under stress conditions (Figures 2C,D,F,G B. longum; where C, D conditions corresponds to B. breve [shown in the Supplementary Information], E condition corresponds to B. thermacidophilum [shown in the Supplementary Information], and D, F, G conditions corresponds to B. adolescentis) [where C refers to the combination of four amino acids; D refer to alanine; E refers to tryptophan; F refers to asparagine; G refers to 0.2 M sodium acetate). However, B. adolescentis is intrinsically bifid-shaped and maintains its bifid morphology irrespective of the ambient conditions.

To understand the bifid form of B. adolescentis and the forms of other Bifidobacterium, C. elegans was fed with different Bifidobacterium species and observed microscopically and by SEM. Based on chemotaxis study B. adolescentis and B. longum was considered for the study. Only B. adolescentis maintained the bifid shape in the gut of C. elegans. The B. adolescentis-fed C. elegans were dissected and centrifuged, and the supernatant was spread on an MRS agar plate. The numbers of bacteria were counted. The bacterial genomic DNA was isolated, and the 16S rRNA was amplified and analyzed, confirming that the isolated bacteria were B. adolescentis. The same culture was also subjected to an F6PPK assay, which confirmed the bacterial identity. The same bacteria were examined with confocal microscopy (Figure 7), which confirmed their bifid morphology. Bifidobacteria species other than B. adolescentis maintained their intrinsic rod shape in the digestive tracts of the worms, confirming that this is their intrinsic shape, and changes only under stress conditions.

Experiments were performed to understand the ability of C. elegans to consume Bifidobacterium. The food chosen by C. elegans from among probiotic organisms and a control was estimated both qualitatively (Figures 8A–D) and quantitatively (Figure 8E), and represented as the CI. The CI for B. longum was 8.5 and 9.8 at 12 and 24 h, respectively, and was 0.7 for B. adolescentis at 12 h. C. elegans contained no B. animalis or B. adolescentis at 24 h, indicating that C. elegans had not selected them. To validate the bacterial morphology in vivo, nematodes that had consumed B. longum or B. adolescentis were fixed and examined with SEM to determine the shapes of the bacteria at different magnifications (Figure 9).

The results in the Figure 8E clearly show that the nematodes greatly preferred Bifidobacterium to E. coli OP50 as its staple food. Caenorhabditis also expressed a preference among the three different Bifidobacterium species tested, so C. elegans may be the best model organism in which to investigate this polymorphism of Bifidobacterium. We assayed the preference of C. elegans for bacterial species B. longum, B. adolescentis, and B. animalis. The nematodes showed a preference for B. longum over B. adolescentis and B. animalis (Figures 8A–D). The histogram shows the CIs of wild-type C. elegans for bacterial strains B. longum, B. adolescentis, and B. animalis at 12 and 24 h (Figure 8E). The data are the means ± SD of three independent biological replicates.

The health benefits of probiotic microbes are only observed when they reach the colon, after passing through the small intestine (low pH), and large intestine (alkaline conditions). During their progress, they are exposed to very harsh conditions. To understand the effects on their morphology, these Bifidobacteria were subjected to different pH. When exposed to increasing pH (Figure 3), B. adolescentis slowly became coccoid, with a dramatic reduction in bifid bacteria. At pH 5.5, they maintained the bifid structure, but as the pH increased to 6, the bifid structure started to diminish, and at pH 7 and 7.5, the number of coccoid structures predominated, although no pointed elongated rods were observed (Figure 3). This phenomenon indicates that before it reaches its destination, the morphology of B. adolescentis changes frequently to accommodate the ambient conditions. Bacterial movement may be faster in the coccoid state, and this state may be favored when no adhesion is required. For adhesion, a larger and rougher bacterial surface area is required, and therefore a bifid shape is favored. Thus, B. adolescentis always maintained its bifid shape, unless it senses adverse conditions, when its morphology changes.

When Bifidobacterial species were grown in the presence of different salts (0.2 M sodium acetate or 0.25 M sodium chloride) (Kojima et al., 1968), SEM clearly showed that there were no dramatic changes in their morphologies (Figures 2G,H). A slight elongation of the rods was the only visible difference between salt-exposed B. longum and control B. longum. Importantly, the presence of lower concentration (0.001–0.01 M) NaCl did not affect the morphology of B. adolescentis, and its bifid structure did not change to the rod form. The bifid structure was observed at lower concentrations of salt, and as the concentration of salt increased, the bacteria assumed the rod shape. The rod structures started to diminish at a salt concentration of 0.005–0.01 M, whereas the bifid form started to shorten, until at 0.25 M salt, no bifid structures were seen (Figure 2H, condition of B. adolescentis).

Scanning electron microscopy also showed that Bifidobacterium other than B. adolescentis adopted the bifid form, but only very rarely, and only under stress conditions (Figures 2B–H). Therefore, it is clear that the bifid form is an intrinsic to B. adolescentis, but not to the other Bifidobacterium species, although a few Bifidobacterial species adopt the bifid form under stress conditions. Kojima et al. (1968) reported that changes to the bifid form occurred in different generations of Bifidobacterium, but our study clearly shows that the first, second and third generations of B. adolescentis maintained the bifid morphology (Figure 6). Dramatic changes in B. adolescentis were only observed when it was exposed to different pH (pH 5.5–7.5), when it adopted a coccoid shape (Figure 3). Therefore, the number of coccus-shaped bacteria was much higher depicted in the initiation of coccus formation, but reverted back to the bifid form. CaCl₂ plays a significant role in the conversion of Bifidobacterium from the bifid shape to the rod form, suggesting the involvement of calmodulin or a similar molecule. This phenomenon was observed when B. adolescentis were subjected to different concentrations of CaCl₂ (0.001–0.01 M). As the calcium concentration increased, the cells became rod-shaped, short, and bulging, with terminal buttons, and they also aggregated (Figure 5).

We examined B. longum in experiments similar to those described above because it bears an intrinsic elongated rod structure. As recorded in Figure 2, no dramatic changes were observed under different conditions, except when the cells were exposed to 42°C, their rod structures changed to extend elongated rods. At higher oxygen concentrations (10%), they aggregated and became elongated, with button/hyphal structures at their ends. When the concentration of oxygen was reduced, the hypha-like structures and cell aggregation, decreased (data not shown). A low concentration of calcium chloride (0.001 M) caused some defects, including cell shrinkage, the formation of wrinkles or holes, etc. As the levels increased from 0.005 to 0.01 M, the cells started out to change from rod-shaped to extensively elongated rods, as well as the bifid form. At all concentrations of CaCl₂, the ratio of bifid-shaped B. adolescentis were drastically reduced in number (Figure 5, DSM 20083). Structural alterations of different Bifidobacterial species subjected to different conditions are summarized in Figure 10.