Simultaneous detection of 37 Lactobacillus species using a real-time PCR assay based on whole-genome sequence analysis

Background Lactobacillus species are used as probiotics and play an important role in food However, use of 16S rRNA gene sequences as standard markers for the differentiation of Lactobacillus species offers a very limited scope, as several species of Lactobacillus share similar 16S rRNA gene sequences. In this study, we developed a rapid and accurate method based on comparative genomic analysis for the simultaneous identification of 37 Lactobacillus species that are commonly used in probiotics and fermented foods. To select species-specific sequences or genes, a total of 143 Lactobacillus complete genome sequences were compared using Python scripts. In 14 out of 37 species, species-specific sequences could not be found due to the similarity of the 16S–23S rRNA gene. Selected unique genes were obtained using comparative genomic analysis and all genes were confirmed to be specific for 52,478,804 genomes via in silico analysis; they were found not to be strain-specific, but to exist in all strains of the same species. Species-specific primer pairs were designed from the selected 16S–23S rRNA gene sequences or unique genes of species. The specificity of the species-specific primer pairs was confirmed using reference strains, and the accuracy and efficiency of the real-time polymerase chain reaction (PCR) with the standard curve were confirmed. The real-time PCR method developed in this study is able to accurately differentiate species that were not distinguishable using the 16S rRNA gene alone. This Real-time PCR method was designed to detect 37 Lactobacillus species in a single reaction. The developed method was then applied in the monitoring of 19 probiotics and 12 dairy products. The applied tests confirmed that the species detected in 17 products matched those indicated on their labels, whereas the remaining products contained species other than those appearing on the label. , that successfully applied to commercial probiotic and dairy products, and showed that some did not accurately match the Lactobacillus listed on their assay be helpful for monitoring the reliability of probiotic and dairy product labels. In to its application in probiotic

casei group and that of the L. sakei group have more than 98.7% similarity between species [19,20].
In this study, we designed species-specific primer pairs targeting the 16S-23S rRNA gene and speciesunique genes, and developed a detection method for 37 Lactobacillus species by a single reaction of real-time polymerase chain reaction (PCR). The developed real-time PCR assay was successfully applied to commercial probiotics and dairy products. We have also confirmed that this assay has the ability to determine the composition of Lactobacillus species present in a product, as well as the presence of species not stated on the label.

Selection of species-specific sequences and primer designs
The species-specific primer pairs of 37 Lactobacillus were designed from unique gene or 16S-23S rRNA region (Table 1). The similarities of the 16S-23S rRNA regions among Lactobacillus species were verified in silico and 23 Lactobacillus species were distinguished with each primer pair designed in the 16S-23S region. Some Lactobacillus species are difficult to distinguish using the 16S-23S rRNA region alone due to the small number of single-nucleotide polymorphisms. Therefore, unique genes of 14 Lactobacillus species were obtained using comparative genomics (Table 2). A membrane protein was found in four L. acidipiscis genomes, but was not present in other species of Lactobacillus.
Adenylosuccinate lyase and leucine-rich repeat protein were detected as the specific genes in L.
amylovorus and L. parabuchneri, respectively. In L. paraplantarum, L. plantarum, L. pentosus, and L. helveticus, MFS (Major Facilitator Superfamily)-type transporter YcnB, LPXTG-motif cell wall anchor domain protein, GHKL domain-containing protein, and decarboxylate/amino acid:cation Na + /H + symporter family protein were detected as the specific genes to each respective species. We also confirmed the specificity of unique genes using BLAST. The unique genes did not match any of the 52,478,804 sequences found in the NCBI database outside of the target species (Table 3). However, some genomes of L. casei contained unique genes of L. paracasei. The presence of unique genes in some, but not all, L. casei strains suggests that the genome information given for the strains is incorrect. These L. casei strains were found to be more similar in the 16S rRNA gene to L. paracasei than to the L. casei described in a previous study [21]. Also, one genome of L. gallinarum contained a unique gene of L. helveticus. To clarify the problem of L. gallinarum strain, we further performed a genomic analysis of L. helveticus and L. gallinarum. The result showed that a L. gallinarum strain containing a unique gene of L. helveticus was more similar to other strains of L. helveticus (Fig. 1).

Specificity of designed primer pairs
To confirm whether primer pairs were species-specific for the identification of each Lactobacillus species, conventional PCR assays were performed with 37 Lactobacillus reference strains. For each of the primer pairs, the amplification product was exclusive to each target strain with a high specificity.
The results of the conventional PCR assays confirmed 100% specificity for all Lactobacillus species.

Specificity and accuracy of the developed real-time PCR assay
The accuracy and efficiency of the real-time PCR assay were validated using the template DNA of the Lactobacillus reference species. All primer pairs exhibited a linear relationship over the range of 0.005 to 50 ng. The slopes for the specific primer pairs of L. acetotolerans, L. casei, L. parabuchneri, and L. lindneri were −3.209, −3.284, −3.207, and −3.595, respectively, and the R 2 values were 1, 0.999, 1, and 0.985, respectively (Fig. 2). The R 2 and slope values of the remaining primer pairs are shown in Table 4.
The specificities of all 37 Lactobacillus reference strains were evaluated for each species-specific primer pair. A non-template was used as a negative control, and the template DNA of 37 Lactobacillus reference stains was used as a positive control for each primer pair. All genomic DNA from Lactobacillus species yielded detectable amplicon signals for each primer pair, whereas none of the non-target Lactobacillus species generated any signals at all (Fig. 3). The C t ranges were 9.0 to 15.0 for each Lactobacillus species (Table 5). Thus, all primer pairs were considered specific for the detection of an individual Lactobacillus species. To verify the accuracy of the assay, a primer pair targeting the 16S rRNA gene was used as an IPC; the amplification of the target region was observed within the C t value range of 5.7 to 9.1 for all tested Lactobacillus species.

Application of the developed real-time PCR assay in probiotics and dairy products
The real-time PCR assay was applied to identify Lactobacillus species from commercial probiotics and dairy products. A total of 31 products were evaluated using the real-time PCR assay we have developed, and the assay results were compared with the probiotic label claims. Probiotic products were tagged as P1 to P19, whereas dairy products were designated as D1 to D12. As a result of the validation process, 17 products were confirmed to match their label claims (Table 6). However, the label claims of four products (P14, P15, P17, and P18) identified L. helveticus but contained L. acidophilus, and three products (P14, P15, and P17) contained L. paracasei instead of the L. casei indicated on the label. In one product (P16), we detected additional Lactobacillus species that were not listed on the label. We were also able to identify the Lactobacillus species from products labeled with the compound LAB. Our real-time PCR results confirmed that these products contained either L. acidophilus and L. delbrueckii or L. paracasei and L. helveticus.

Discussion
A variety of methods have been used to identify LAB in foods or in the environment. The most representative method is a conventional method consisting of phenotypic and biochemical tests, which have limitations in accuracy among isolates possessing similar physiological specificities and fermentation profiles at the species level [22,23]. To overcome these difficulties, several genotypebased methods such as DGGE and metagenome sequencing have been developed [23]. In addition, metagenome sequencing based on the 16S rRNA gene is a common approach in investigating microbial communities but is limited to distinguishing similar species [24]. Because metagenome sequencing remains a time-consuming process and requires specialized equipment and techniques, it is unsuitable for analyzing a large number of samples. To combat this, we have developed a real-time PCR assay that can rapidly and easily analyze Lactobacillus communities in fermented foods and potentially environmental samples.
PCR is generally considered to be a rapid, sensitive, and time-saving method for the detection of bacterial species [25][26][27]. The accuracy of PCR is determined by the specificity of the primer pairs used. The 16S rRNA gene is considered a marker gene for bacterial genotypic analysis and is useful for the accurate identification of bacteria [12,28]. Studies focusing on the identification of To overcome the limitations of the 16S rRNA gene, we developed 37 Lactobacillus species-specific primer pairs based on 16S-23S rRNA gene analysis and comparative genome analysis. Speciesspecific primer pairs were designed to have a small amplicon size (~260 bp) to increase amplification efficiency and detect Lactobacillus species present in processed foods. The specificities of the species-specific primer pairs were confirmed using the 37 Lactobacillus species, and amplification was observed only in the target species DNA without any cross-reactivity. Also, it was confirmed that species such as the L. casei group, L. acidophilus group, and L. plantarum group, which are not distinguished by the conventional identification method, were differentiable using the species-specific primer pairs. According to the CODEX guidelines, the slope values of −3.1 to −3.6 are considered to indicate a high PCR efficiency. The coefficient value of determination should be at least 0.98 to be considered viable data [33]. Therefore, these results demonstrate that the developed real-time PCR assay provides high accuracy and efficiency.
The developed real-time PCR assay was used to assess probiotics and dairy products. Using this assay, 17 products were determined to contain the Lactobacillus species advertised on the label. In the remaining products, the species indicated on the labels were either replaced with or contaminated by another species. For example, L. acidophilus was replaced by L. helveticus and L. casei was replaced by L. paracasei in four probiotic products. Though these products were produced by different companies, the same strains were identified. As described above, L. acidophilus belongs to the same group as L. helveticus, and L. casei belongs to the same group as L. paracasei. The likely reason a label names species other than the one detected is misidentification [20,34]. In one product, additional Lactobacillus species that were not indicated on the label were detected by real-time PCR.
These were detected at much higher C t values than the Lactobacillus species indicated on the label, suggesting that such strains were only present in low concentrations [35]. We were also able to accurately identify the species contained in products labeled compound LAB. In all of these products, Bifidobacterium species claimed on the label [5]. In addition, some products are inconsistent from one lot to another. These results indicate inadequate that quality control for these products.

Conclusion
In this study, we developed specific primer pairs using comparative genomics to identify Lactobacillus accurately and rapidly at the species level, then applied this technology in a real-time PCR assay that can detect 37 Lactobacillus species in a single reaction. The developed real-time PCR method was able to accurately discriminate species that were not distinguishable by the conventional identification method. To verify the developed real-time PCR assay, we compared the label claims of probiotics and dairy products with the Lactobacillus species detected using the real-time PCR method.
The real-time PCR assay that we have developed was successfully applied to commercial probiotic and dairy products, and showed that some products did not accurately match the Lactobacillus species listed on their labels. Thus, this assay will be helpful for monitoring the reliability of commercial probiotic and dairy product labels. In addition to its application in probiotic products, the assay can be applied to identify Lactobacillus communities in various food or environmental samples.

Bacterial strains and probiotic and dairy products
The Lactobacillus reference strains were obtained from the Korean Collection for Type Cultures (KCTC; Daejeon, South Korea; https://kctc.kribb.re.kr/) and the Korean Agricultural Culture Collection (KACC; Jeonju, South Korea; http://genebank.rda.go.kr/) ( Table 7). All reference strains were cultured in Lactobacilli MRS Broth (Difco, Becton & Dickinson, Sparks, MD, USA) at 30°C for 48 h under anaerobic conditions. The probiotic and dairy products tested in this study were obtained from various markets around the world (South Korea, United States, and Canada). The samples used in this study included 19 probiotic products (10 capsule-form pharmaceuticals and 9 powder-form food supplements) and 12 dairy products manufactured by 19 different companies. All products were labeled with bacterial species or LAB compounds.

DNA extraction
All Lactobacillus reference strains were grown in MRS broth at 30°C for 48 h under anaerobic conditions. The cultured cells were harvested by centrifugation at 13,600 × g for 5 min, after which the supernatant was removed. Genomic DNA was extracted using a bacterial genomic DNA extraction kit (Intron Biotechnology, Seongnam, South Korea) according to the manufacturer's instructions. Total genomic DNA from the probiotic and dairy products was extracted using a DNeasy® Blood & Tissue Kit (Qiagen, Hilden, Germany) according to the method described in a previous study [36]. DNA concentration and purity were determined by absorbance using a MaestroNano® spectrophotometer (Maestrogen, Las Vegas, NV, USA).

Identification of Lactobacillus species-specific regions and primer designs
In total, 143 complete genome sequences of 37 different Lactobacillus species were obtained from the National Center for Biotechnology Information (NCBI; ftp://ftp.ncbi.nlm.nih.gov/genomes/) database (Additional file 1: Table S1). The 16S-23S rRNA regions, including the intergenic spacer regions, of 143 strains were extracted from the Lactobacillus complete genomes using a script written in the Python language, and the extracted regions were aligned using the Geneious program ver. 11.1.2 (Biomatters Limited, Auckland, New Zealand). According to the alignment results, primer pairs were designed on the basis of species-specific sequences in the 16S-23S rRNA gene. Some Lactobacillus species are difficult to distinguish at the species level because of the high degree of similarity in their 16S-23S rRNA gene sequences. For these species, we have developed speciesspecific primer pairs from unique genes that exist only in the target species obtained through comparative genomic analysis.
The genome sequences of target species were blasted against the complete genome of target species using the UBLAST function of USEARCH program ver. 9.0 [37], with 80% cutoff identity. The genes that showed a significant match with the genomes of all target species were considered as core genes of target species. Those genes were then blasted against all of the Lactobacillus complete genomes except the target species using the UBLAST function of USEARCH program (ver. 9.0) with default parameter settings of 50% cutoff identity [37]. Genes that found no match to all genomes of the nontarget species were identified as potential unique genes. The identified potential unique genes were verified using the Basic Local Alignment Search Tool (BLAST) for 52,478,804 sequences including Lactobacillus genomes. Genes found only in the target species were determined to be unique genes, and species-specific primer pairs were designed from these genes. To verify the presence of genomic DNA from Lactobacillus species, primer pairs were designed from the conserved regions of 37 Lactobacillus species in the 16S rRNA gene sequence and used as an internal positive control (IPC). All primer pairs were designed using Primer Designer (Scientific and Educational Software, Durham, NC, USA) and synthesized by Bionics Co. Ltd. (Seoul, South Korea).

Specificity of species-specific primer pairs
PCR assays were performed to confirm the specificity of the designed species-specific primer pairs. The specificity was evaluated using 37 Lactobacillus reference strains. PCR products were amplified using the following conditions in a thermocycler (Astec, Fukuoka, Japan): 94°C for 10 min, followed by 30 cycles of 94°C for 30 s, 60°C for 30 s, 72°C for 30 s, and 72°C for 5 min. The 25 µL reaction mixtures contained 20 ng of template DNA of a Lactobacillus reference strain, 0.5 unit of Taq DNA polymerase (TaKaRa BIO Inc., Tokyo, Japan), and species-specific primer pairs. The optimal concentration of each species-specific primer pair obtained from the experiments is shown in Table 1.
The amplification products were confirmed by electrophoresis on a 2% agarose gel, and the product bands were visualized under a UV transilluminator (Vilber Lourmat, Marne La Vallee, France).

Development of real-time PCR assay
The real-time PCR assays were performed on the 7500 Real-Time PCR System (Applied Biosystems, Foster City, CA, USA) using the following conditions: 95°C for 2 min, followed by 30 cycles of 95°C for 5 s and 60°C for 30 s. The melting curve data were generated using one cycle of 95°C for 15 s, 60°C  Table 1. To evaluate the analytical accuracy of the real-time PCR assay, a standard curve was constructed using serial dilutions (50 to 0.005 ng) of genomic DNA from Lactobacillus reference strains in triplicate. The specificities of the species-specific primer pairs were tested using 20 ng of DNA extracted from 37 Lactobacillus reference strains. Real-time PCR amplifications of IPC were also confirmed with 37 Lactobacillus reference strains. The results of the real-time PCR were confirmed using 7500 Software V2.3 (Applied Biosystems).

Application of the developed real-time PCR assay in probiotic and dairy products
We designed a validation test to detect 37 Lactobacillus species with real-time PCR in a single reaction using primer pairs. Each well of a reaction plate contained each primer pair and IPC for the

Competing interests
The authors declare that they have no competing interests.

Availability of data and materials
The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.

Consent for publication
Not applicable.

Ethics approval and consent to participate
Not applicable. approaches for identification and characterization of lactic acid bacteria. J Dig Dis.   Table 6. Results of application test of the developed real-time PCR assay to commercial probiotic and Table 7. Lactobacillus reference strains used in this study