PCR-based screening, isolation, and partial characterization of motile lactobacilli from various animal feces

Most lactobacilli found in animal intestines are generally non-motile, but there are few exceptions. Our previous work showed that Lactobacillus agilis BKN88, which is a highly motile strain originating from a chicken, takes advantage of motility in gut colonization in murine models, and thus motile lactobacilli likely have unique ecological characteristics conferred by motility. However, the ecology and habitat of gut-derived motile lactobacilli are still rarely understood. In addition, the limited availability of motile Lactobacillus isolates is one of the major obstacles for further studies. To gain insight into the ecology and habitat of the motile lactobacilli, we established a routinely applicable detection method for motile lactobacilli using PCR and subsequent selective isolation in semi-solid MRS medium for the collection of additional motile lactobacilli from animal feces. We applied the PCR detection using motile lactobacilli-specific primers, based on the motor switch protein gene (fliG) of flagella, to 120 animal feces, followed by selective isolation performed using 45 animal feces. As a result, motile lactobacilli were detected in 44 animal feces. In the selective isolation, 29 isolates of L. agilis and 2 isolates of L. ruminis were obtained from 8 animal species. These results indicated that motile lactobacilli are distributed in different animal species. Moreover, phylogenetic analysis of the L. agilis isolates suggests co-evolution with the host, and adaptation to a particular environmental niche.


Background
Some lactobacilli are found in the gastrointestinal tract (GIT) of humans and animals and are considered as the gut symbionts [1]. The lactic acid bacteria are known to have specific ecological niches depending on the species/ strains. For example, Lactobacillus johnsonii and Lactobacillus reuteri inhabit the GIT of various animals, but Lactobacillus gorillae and Lactobacillus equigenerosi exhibit host specificity [2][3][4][5]. Additionally the composition of lactobacilli in the gut may be affected by the diets of host animals. Past studies reported that Lactobacillus brevis, Lactobacillus casei and Lactobacillus plantarum were detected in omnivorous animals but not in carnivores, while Lactobacillus ingluviei, Lactobacillus salivarius and Lactobacillus vaginalis are dominant in carnivores, but not in most herbivores and omnivores [6]. This suggests that diet has major impacts on the composition of lactobacilli in animal guts.
Motility is a minor characteristic in the genus Lactobacillus, and only the small part of the species belonging to the phylogenetic group of L. salivarius possesses motility with an exception of the specific species Lactobacillus curvatus [7]. Bacterial motility is mostly mediated by flagella. A flagellum is generally composed of approximately 30 different proteins, which form three substructures: filament, hook, and basal body [8]. In motile lactobacilli, the motilityrelated genes are usually located in single operon, which appears to be conserved [7].
The natural habitats of the motile lactobacilli are variable, such as sewage, freshwater pond, stinky tofu brine, cocoa, cider, wine, shochu, oak tree, grape must, kaburazushi, chicken, horse and human [7][8][9][10][11][12][13][14][15][16][17]. Of the motile lactobacilli, L. agilis and L. ruminis are the only known species originating from animal and human guts [18,19]. The motility related characteristics of the microbes are likely to be involved in their nutrient acquisition and niche colonization. Actually, our previous study showed that L. agilis BKN88, which is a highly motile strain isolated from chicken feces, takes advantage of motility and chemotactic ability in gut colonization in murine models [20]. Thus, the motile lactobacilli might have a unique niche. However, gut-derived motile lactobacilli are not well characterized due to lack of studies on motile Lactobacillus isolates as well as the limited number of isolated strains in public culture collections [19,21]. It is thus important to collect additional bacterial strains of motile lactobacilli. However, the selective isolation method of motile lactobacilli has not been developed yet.
In the present study, motile lactobacilli were detected by PCR using motile lactobacilli-specific primers as a culture-independent technique, and then the selective isolation of motile lactobacilli using a soft-agar medium was performed for clarification and confirmation. In addition, we described the general phylogenetic/genomic features of motile Lactobacillus isolates.

Construction and validation of motile lactobacilli-specific primers
Two PCR primer pairs, the first specific to only L. agilis and L. ruminis (named Lag/Lru primers) and the second specific to all motile-lactobacilli (named universal primers), targeting the fliG gene were designed to detect motile lactobacilli from animal fecal samples (Table 1). To validate the specificity of each primer pair, PCR amplification was performed with genomic DNA from eleven strains of motile lactobacilli including two strains of L. agilis and two strains of L. ruminis, and a non-motile Lactobacillus and E. coli (Table 2). In a PCR amplification with the Lag/Lru primers, fliG-specific PCR products were detected from the two strains of L. agilis and two strains of L. ruminis but not in the non-motile control strains. By using the fliG universal primers, PCR products were observed in all motile lactobacilli but not in a non-motile Lactobacillus and E. coli (Fig. 1a). Furthermore, we tested the primer pairs against chicken and Siberian tiger feces. DNA extracted from chicken feces from which motile lactobacilli could be isolated produced PCR amplicons with each primer pair, but amplicons were not obtained from feces of a Siberian tiger from which motile lactobacilli could not be isolated (Fig. 1b). The sensitivity of each primer pair was tested by PCR amplification with DNA extracted from murine fecal pellets supplemented with serial dilutions of L. agilis BKN88 bacterial cells. Since we could not find any motile lactobacilli in the feces from Balb/c mice, the murine fecal pellets were used as motile lactobacilli-free feces [20]. The detection limits of the Lag/Lru primers and the universal primers were 1 × 10 5 CFU/g feces and 1 × 10 4 CFU/g feces, respectively (Fig. 1c).

PCR detection of motile lactobacilli in animal feces
One hundred twenty fecal samples from various animals were used for screening of motile lactobacilli. As shown in Table 3, the Lag/Lru primers amplified fliG genes from 36 fecal samples, and the universal primers amplified fliG genes from 24 fecal samples. By the use of the two primer pairs, fliG-specific PCR products were detected in 44 of the 120 fecal samples, which included 25 of the 61 omnivores (41%), 16 of the 52 herbivores (31%) and 3 of the 7 carnivores (43%). In terms of animal species, fliG-specific PCR products were detected in 37 of the 93 animal species, which included 20 of the 45 omnivores (44%), 14 of the 41 herbivores (34%) and 3 of the 7 carnivores (43%).

Isolation of motile lactobacilli from animal feces
The selective isolation of motile lactobacilli using semisolid MRS medium was applied to only fresh 45 animal feces samples including 23 omnivores, 20 herbivores and 2 carnivores. As shown in Table 4, 29 isolates of L. agilis and 2 isolates of L. ruminis were obtained from 8 animal species including 5 omnivores and 3 herbivores. No motile lactobacilli were isolated from fecal samples which were negative for PCR detection of fliG. In some fecal samples, the motile bacteria were not isolated although fliG-specific PCR products were detected. Cloning and sequencing analysis of the amplified fliG gene from feces of a Patagonian mara (Sample ID: 063) and Bennett's wallaby (Sample ID: 070), from which motile bacteria could not be isolated but were positive for fliG by PCR, showed that the sequences of the detected fliG gene had a 99% sequence similarity with the fliG gene of L. agilis (GenBank, Accession number KM886859).

Phylogenetic analysis based on flgD gene sequences
Multiple sequence alignment showed that the flgD gene sequence similarities between all pairs of L. agilis isolates were at least 98.5%. Subsequent phylogenetic analysis using the neighbor-joining method revealed that L. agilis isolates could be differentiated into five groups, designated A-E (Fig. 2). Isolates from the same animal species clustered in the same groups. Only the group of A included the isolates from some animal species, and the isolates from the horse, brown lemur and black lemur shared the group. In the isolates from the brazilian tapirs, KZ strains and SY strains, were obtained from animals in different zoos, however, both isolates belonged to the same group. Chicken isolates in our study were bred in different locations, but were located in the same group as L. agilis PTL465. Thus, the representative strains, L. agilis NB11, SN4111, SN811, SN10121, SY111 and SY212, from each animal species were used for further analysis.
Genomic features of L. agilis isolates To obtain more information about the genomic characteristics of L. agilis isolates from different animal species, draft genome sequences of seven strains were determined by the Illumina Genome Analyzer II system. Genome sequences of an additional two strains of L. agilis, DSM 20509 T and La3, were obtained from the NCBI databases for use as references. The GenBank accession numbers of L. agilis, DSM 20509 T and La3 are AYYP00000000 and CP016766, respectively. The genomic features of all strains used in the present study are summarized in Table 5. All genomes possessed a low contamination level (≤ 2% contamination value), and the completeness values for the all strains of L. agilis were 98.23% (Table 5), meaning that all genomes satisfied the criteria to be considered as near-complete genomes with low contamination (≥ 90% completeness value and ≤ 5% contamination value) [23]. The genome size of the type strain (DSM 20509 T ) is 2.06 Mbp and the mol% G + C content of DNA is 41.7 [24]. The genome sizes of the 9 strains of L. agilis ranged from 2.03 to 2.52 Mbp and the number of CDSs ranged from 1963 to 2427. The GC contents of the 9 strains of L. agilis were 40.5 to 41.7%. L. agilis SY212 (Brazilian tapir isolate) had the largest genome and number of CDSs and the lowest GC contents. In a preliminary analysis, we found that the genome of L. agilis SY212 contained more putative mobile genetic elements (MGEs) than the other strains of L. agilis. These MGEs might be related to the unique genomic features of L. agilis SY212. The average nucleotide identity (ANI) values for the tested strains are shown in Additional file 1. ANI values among the 9 strains of L. agilis ranged from 0.973 to 0.982.
Carbohydrate utilization of L. agilis isolates Seven strains of L. agilis obtained from different animal species were further characterized by sugar  fermentation assays using the API 50 CH kit. The carbohydrate fermentation patterns of seven strains are shown in Additional file 2. Different patterns of the acid production from D-mannose, α-methyl-Dglucoside, amygdalin, salicin, cellobiose, lactose, trehalose, melezitose, β-gentiobiose and D-turanose were shown among the 7 strains of L. agilis, but have no obvious correlation to their origins or dietary needs. All L. agilis isolates were able to metabolize Dgalactose, D-glucose, D-fructose, D-mannitol, N-acetylglucosamine, arbutin, esculin, D-maltose, D-lactose, D-melibiose, D-sucrose and D-raffinose.

Discussion
Some lactobacilli reside in the gut of humans and animals as commensal bacteria [1]. Most of the lactobacilli are non-motile, but a few members of the lactobacilli such as L. agilis and L. ruminis are motile. Our previous work showed that L. agilis takes advantage of motility to survive and colonize in the murine gut [20]. This result implies that motile lactobacilli have unique ecological niches utilizing their motility and chemotactic ability. However, their ecological niches remain unknown. In the present study, we focused on the ecology and habitat of motile lactobacilli. At least 15 Lactobacillus species have been reported as motile lactobacilli to date [7]. Most motile lactobacilli originated from alcoholic fermentation samples, e.g. wine and shochu, whereas L. agilis and L. ruminis were found in animal and human feces [9,12,18]. Thus, two kinds of motile lactobacilli-specific primer pairs were designed to detect fecal-motile lactobacilli and all motile lactobacilli. The motility operon of motile lactobacilli appears to be relatively conserved [7,21], and the multiple alignment analysis of the motility operons from fourteen publicly available motile lactobacilli showed that the fliG, fliI, flhA, cheW, cheR and fliY genes are highly conserved among the motility operons in the motile lactobacilli. In this study, the most conserved fliG gene was used as the target gene for detecting potentially motile lactobacilli. The results of PCR using the designed primer pairs against DNA from various samples demonstrated that these primer pairs may be useful to detect motile lactobacilli.
PCR detection using the currently designed primer pairs were applied to 120 animal feces samples for screening of motile lactobacilli. Consequently, PCR using the Lag/Lru primers and the universal primers detected motile lactobacilli in 36 samples and 24 samples, respectively.
By the use of the two primer pairs, motile lactobacilli were detected in 44 animal feces samples (37 animal species) including those from animals which have never been reported as origins of motile lactobacilli, and indicated that motile lactobacilli are distributed in various animals. As shown in Table 3, there were differences in detection between the two primer pairs, and the detection of motile lactobacilli could be improved by the use of the two primer pairs. Although the sensitivity of PCR detection of L. agilis cells in murine feces using the universal primers was shown to be higher than that of the Lag/Lru primers, PCR with the Lag/Lru primers produced amplicons in more animal feces samples than PCR with the universal primers. The reason for this inconsistency might be due to strain-specific nucleotide substitutions at the annealing regions of the universal primers. Further study is essential to improve the inconsistent detection.
In selective isolation of motile lactobacilli using semi-solid MRS medium, 29 isolates of L. agilis and 2 isolates of L. ruminis were obtained from 8 animal species. Previously, L. agilis have been isolated from pigeon, chicken, human, horse, laying hen, piglet and fermented food products such as masau fruits, Nigerian ogi, and cheese [25][26][27][28][29][30][31][32][33][34]. L. ruminis have been isolated from human, cow, horse and pig [35][36][37][38][39][40][41]. Black lemur, brown lemur, turkey, squirrel monkey and brazilian tapir were newly identified as a host of L. agilis in this study. All isolates were obtained only from samples positive by PCR for fliG. This indicated that the PCR detection using designed primer pairs is a useful/powerful tool to exclude fecal samples negative for motile lactobacilli prior to isolation. In some fecal samples, the motile bacteria were not isolated despite fliG specific PCR products having been detected. By sequence analysis, such PCR products were likely to be identified as L. agilis. This discrepancy may be because L. agilis cells were dead/ unculturable or the L. agilis strains were not motile at least in the semi-solid MRS medium. In fact, some strains of L. agilis and L. ruminis do not show motility in MRS medium [38,42,43].
In the phylogenetic relationships based on motilityrelated genes among the 30 L. agilis isolates originating from various animal hosts, L. agilis isolated from the same host species tended to be clustered together, which suggests the co-evolution with each host. Similar phylogenetic relationships have been observed in other species, such as L. casei, L. reuteri, L. rhamnosus, L. johnsoni and L. ruminis [38,[44][45][46][47]. Draft genome sequences of seven L. agilis isolates were then obtained and analyzed. The ANI values are higher than 95% among the seven isolates, which indicate that the isolated strains may belong to L. agilis species. The carbohydrate fermentation patterns of the seven isolates also showed that these strains are highly similar to the type strain of Lactobacillus agilis (https://bacdive.dsmz.de/ strain/6407). These L. agilis isolates and the genome sequences would provide further insight into the ecology of gut-derived motile lactobacilli.

Conclusions
A combination of PCR detection using fliG-specific primer pairs and subsequent isolation with semi-solid MRS medium could successfully isolate a large number of motile lactobacilli from various animal feces, and indicates that motile lactobacilli are distributed in various animals. Phylogenetic analysis on the motility-related gene of L. agilis isolates suggests co-evolution with each host, and the adaptation to a particular environmental niche. Additional genomic studies with the obtained sequences need to be done to provide further insights on the ecological adaption of motile lactobacilli in animals and humans.

Bacterial strains and growth conditions
The bacterial strains, listed in Table 2, were obtained from the NODAI Research Institute Culture Collection (NRIC, Tokyo, Japan), RIKEN BioResource Center (JCM, Ibaraki, Japan) and DSMZ German Collection of Microorganisms and Cell Cultures (DSM, Braunschweig, Germany). Thirty-one motile lactobacilli isolated from animal feces are described in Table 4. All Lactobacillus strains were propagated anaerobically in MRS broth (Difco/BD) at 37°C. Motilities of Lactobacillus strains were determined by visual examination after inoculation into semi-solid MRS medium with 0.15% agar. E. coli NRIC 1023 and E. coli JM109 were grown aerobically in LB medium with or without 100 μg/ml of ampicillin and 25 μg/ml of kanamycin at 37°C.

Design of motile lactobacilli-specific primers
In order to design the motile lactobacilli-specific primers, DNA sequences of the motility operons from fourteen motile lactobacilli were obtained from the NCBI database and analyzed by multiple alignments using ClustalW [48]. The highly conserved fliG gene, which encodes the flagellar motor switch protein of the motility operon, was targeted for the detection of motile lactobacilli in animal feces. Two PCR primer pairs targeting the fliG gene were designed. One primer pair was designed to amplify the fliG gene specifically in L. agilis and L. ruminis, while another was designed to amplify the fliG gene of most motile lactobacilli. These primer pairs are referred to as Lag/Lru primers and universal primers, respectively. The reason we prepared the Lag/ Lru primers is because some strains of L. agilis and L. ruminis have been previously isolated from animal or human feces [18,19]. Motile lactobacilli-specific primers are listed in Table 1. The primer pairs Lag/Lru primers and universal primers produce an approximately 300 bp and 280 bp PCR amplicon, respectively.

Validation of primer specificity and sensitivity
The specificity of each primer pair was determined by PCR amplification with genomic DNA from ten strains of motile lactobacilli including two strains of L. agilis and two strains of L. ruminis, Lactobacillus paracasei IGM393 and Escherichia coli NRIC 1023 as templates. DNA from bacterial cultures was prepared as follows. Bacteria cells from an overnight culture were harvested and washed with TE buffer, followed by bead beating with a FastPrep Instrument (MP Biomedicals) in TE buffer. Then, DNA was extracted using phenol-chloroform and ethanol precipitation according to standard protocols. Further determination of the specificity of each primer pair was performed with PCR amplification using DNA extracted from feces of a chicken or a Siberian tiger. In past studies, motile lactobacilli were found in a chicken but not in a Siberian tiger [32]. In this study we also were able to isolate motile lactobacilli from a chicken but not from a Siberian tiger. Therefore the feces from a chicken and Siberian tiger were used as motile lactobacilli-positive or -negative fecal samples, respectively. To validate the sensitivity of each primer pair, PCR amplification was carried out with DNA extracted from murine fecal pellets supplemented with different concentrations of L. agilis BKN88 cells (1 × 10 1 to 10 8 CFU/g feces). Murine fecal pellets were collected from Balb/c mice, which do not have motile lactobacilli in their feces. Balb/c mice were obtained from Crea Japan, Inc. DNA extraction and PCR amplification from fecal samples are described below.

Fecal samples and DNA extraction
Fecal samples were collected from 120 animals from various zoos in Japan. Based on their dietary needs, they were classified as omnivores, herbivores or carnivores, listed in Table 3. The samples were collected in sterile plastic tubes and immediately taken to the laboratory under refrigerated conditions. The fecal samples were stored at − 20°C until DNA extraction. DNA was extracted from 200 mg of fecal pellet with QIAamp DNA Stool Mini Kit (Qiagen), according to the manufacturer's instructions.

PCR amplification
PCR amplification was performed using Ex Taq DNA polymerase (Takara). PCR was carried out in a total volume of 25 μl, containing 2.5 μL of 10× Ex Taq buffer, 2 μL of dNTP mixture (2.5 mM each), 2.5 μl of each of primers (10 μM), 0.25 μl of TaKaRa Ex Taq (5 U/μl) and 0.5 μl of extracted DNA. PCR conditions were as follow: 5 min at 94°C; followed by 30 cycles of 30 s at 94°C, 30 s at 60 or 50°C, 10 s at 72°C and a final extension step of 5 min at 72°C. The PCR with Lag/Lru primers and universal primers was performed at an annealing temperature of 60 and 50 degrees, respectively. The PCR products were separated in 1.5% agarose gel by electrophoresis. DNA bands were visualized by staining with ethidium bromide.

Isolation of motile lactobacilli
Fecal samples which were used for isolation of motile lactobacilli are highlighted in Table 3. Fecal samples for isolation of motile lactobacilli were suspended in sterile PBS, and homogenized using a vortex mixer. The fecal suspensions were inoculated into semi-solid MRS medium with 0.15% agar, and incubated at 37°C for 1 day. The migrated cells from the outermost ring were collected and inoculated onto MRS agar plates, and the plates were incubated under anaerobic conditions at 37°C for 1 to 2 days. Isolated colonies were randomly selected from the surface of MRS agar plates and restreaked onto MRS agar plates, and incubated under anaerobic condition at 37°C for 1 day as before. Isolates were purified by several streakings. Identification of the bacterial isolates was carried out by Sanger sequencing of the 16S rRNA gene. PCR amplification with 16S primers [49], 27F (5′-AGA GTT TGA TCC TGG C-3′) and 1492R (5′-CGG TTA CCT TGT TAC G-3′), was performed with Ex Taq DNA polymerase (Takara). The PCR products were purified using the NucleoSpin Gel and PCR Clean-up kit (Macherey-Nagel) according to the manufacturer's instructions and sequenced by Macrogen Japan (Tokyo, Japan) with the primers 27F and 1492R. The 16S rRNA gene was sequenced for each isolate and used for BLAST analysis using GenBank.
Cloning and DNA sequencing of the fliG gene The fliG genes were amplified using Ex Taq DNA polymerase (Takara) and the Lag/Lru primers with DNA extracted from feces of a Patagonian mara (Sample ID: 063) and Bennett's wallaby (Sample ID: 070), from which the motile bacteria were not isolated, although fliG specific PCR products were detected. Purification of PCR products from 1.5% agarose gels was performed with the NucleoSpin Gel and PCR Clean-up kits (Macherey-Nagel) according to the manufacturer's instructions. These PCR products were inserted into the pGEM-T Easy Vector (Promega) using T4 DNA Ligase (Takara). E. coli JM109 was used as a cloning host. The constructed plasmid was purified with the NucleoSpin Plasmid EasyPure kit (Macherey-Nagel). The sequences of the insert DNA were determined by Sanger sequencing with the M13 primers M13F (5′-GTA AAA CGA CGG CCA GT-3′) and M13R (5′-CAG GAA ACA GCT ATG AC-3′). The result of the sequencing was analyzed as described above.

Sequencing of flgD gene and phylogenetic analysis
The motility operons from four strains of L. agilis publicly available in databases were analyzed by multiple alignments using ClustalW. The highly conserved flgD, motA, flgE, flgG and fliM genes were selected for the phylogenetic analysis of fecal isolates; however, the motA, flgE, flgG and fliM genes could not be amplified in some of the isolates. Thus, the flgD genes were used for further analysis. The flgD genes were amplified using primers DOKJ507 (5′-AAT TTA AGT GAT GCG GTA GC-3′) and DOKJ508 (5′-ATT TGG CAT CGC CTA CTT GG-3′) with DNA extracted from L. agilis isolates. The sequences of the flgD genes were determined with DOKJ507 and DOKJ508 as explained above. Approximately 355 bp of flgD gene sequences of the isolates and related strains or species were used in the phylogenetic analysis. The flgD gene of L. ruminis ATCC 27782 was used as an outgroup. Multiple sequence alignment was carried out with ClustalW [48], and the phylogenetic tree was generated by the neighbor-joining method with 1000 bootstrap replications using MEGA7 [50].

Fermentation profiles of L. agilis isolates
To determine carbohydrate utilization profiles, API 50 CH (bioMérieux) tests were used according to the manufacturer's instructions. Data for the reference type strain (DSM 20509 T ) was taken from the BacDive database at https://bacdive.dsmz.de/strain/6407.