OUP user menu

Intraspecies discrimination of Lactobacillus paraplantarum by PCR

Shinichi Saito, Miho Kobayashi, Hiromi Kimoto-Nira, Reiji Aoki, Koko Mizumachi, Satoshi Miyata, Katsue Yamamoto, Yasushi Kitagawa, Chise Suzuki
DOI: http://dx.doi.org/10.1111/j.1574-6968.2010.02193.x 70-76 First published online: 1 March 2011


Lactobacillus paraplantarum is a species phenotypically close to Lactobacillus plantarum. Several PCR methods were evaluated to discriminate L. paraplantarum strains and among them, a PCR using an enterobacterial repetitive intergenic consensus (ERIC) sequence differentiated L. paraplantarum from other Lactobacillus species. In addition, a combination of ERIC and random amplified polymorphic DNA (RAPD) analysis distinguished among seven strains of L. paraplantarum tested. ERIC-PCR profiles showed several strain-specific DNA fragments in L. paraplantarum, among them, a 2.2-kb ERIC marker, termed LpF1, found to be specific to strain FBA1, which improved the skin integrity in an animal model. The LpF1 encodes three proteins similar to Lactobacillus fermentum AroA, TyrA, and AroK, which are involved in the shikimate pathway. A primer pair specific to FBA1 based on the internal sequence of LpF1 amplified a 950-bp FBA1-specific fragment LpF2. Southern blot analysis of Dra I-digested genomic DNA of L. paraplantarum strains using LpF2 as a probe showed that LpF2 is distinctive of strain FBA1 among 16 L. paraplantarum strains. Because both ERIC- and RAPD-PCR are fast and technically simple methods, they are useful for the rapid discrimination of L. paraplantarum strains and for the development of new strain-specific DNA markers for identifying industrially important strains.

  • Lactobacillus paraplantarum
  • RAPD
  • enterobacterial repetitive intergenic consensus PCR
  • species specificity
  • genetic marker


Lactobacillus paraplantarum, a species phenotypically close to Lactobacillus plantarum, was characterized in 1996 (Curk et al., 1996). Few phylogenetic studies of the species have been reported (Torriani et al., 2001a, b), and methods for discrimination between strains have yet to be developed. On the other hand, some L. paraplantarum strains have received attention owing to their potential uses in food production or preservation (Lee et al., 2007; Chun et al., 2008). We evaluated the effects of 200 heat-killed lactic acid bacteria (LAB) strains on the production of hyaluronate and type I collagen when applied to normal human dermal fibroblast cells in vitro and found five strains with high efficacy (S. Miyata, K. Yamamoto, S. Sakata, C. Suzuki, H. Kimoto-Nira, K. Mizumachi & Y. Kitagawa, unpublished data). These strains (including one called FBA1) improved the skin integrity of HR-1 hairless mice fed a reduced-protein diet. These effects are strain dependent; hence, it is important to develop reliable methods to identify and discriminate strains of L. paraplantarum.

Enterobacterial repetitive intergenic consensus (ERIC) sequences are highly conserved DNA sequences that occur as multiple copies in the genomes of enteric bacteria and Vibrio species (Sharples & Lloyd, 1990; Mercier et al., 1996; Tcherneva et al., 1996; Wilson & Sharp, 2006). Methods using ERIC-PCR have been used to classify closely related strains of enterococci (Wei et al., 2004). The random amplified polymorphic DNA (RAPD) method has been used to classify various organisms from bacteria to plants (Van Reenen & Dicks, 1996; Torriani et al., 2001a; Venkatachalam et al., 2004; Nomura et al., 2006; Walczak et al., 2007). RAPD entails PCR amplification with a single, short oligonucleotide primer that does not strongly match particular sites in target genomes, under low-stringency conditions, for annealing. In most cases, both ERIC- and RAPD-PCR generate several DNA bands that enable species-level or sometimes strain-level differentiation of bacteria.

The aim of this study was to develop a fast and simple method to discriminate strains of L. paraplantarum using PCR and to develop a DNA marker to identify specifically the particular strain. We focused on an L. paraplantarum FBA1 strain, which improved the skin integrity of HR-1 hairless mice fed a reduced-protein diet, and developed a pair of FBA1-specific PCR primers and an FBA1-specific DNA fragment based on ERIC-PCR.

Materials and methods

Bacterial strains and growth condition

The strains used in this study and their sources are listed in Table 1. Several strains were purchased from JCM (RIKEN BioResource Center, Saitama, Japan), NBRC (NITE Biological Resource Center, Chiba, Japan) and the NODAI Culture Collection Center (Tokyo University of Agriculture, Tokyo, Japan), and others were in our culture collection. All strains were grown in MRS broth (Becton & Dickinson) overnight at 37 °C and held as culture stocks in 15% w/v glycerol at −90 °C. Each strain was cultured at least five times on different days for the assessment of the reproducibility of the PCRs.

View this table:
Table 1

The strains used in this study

Species and strainsIsolation source
Lactobacillus curvatus
JCM 1096T
FBA2Radish and carrot pickled with rice bran and salt
LAB10Pickled small eggplant
LAB11Cucumber pickled with rice bran and salt
LAB12Pickled cucumber
LAB13Pickled radish
Lactobacillus paraplantarum
JCM 12533T
A22Fermented rice bran paste
A34Pickled radish
A74Pickled cucumber in plum vinegar
C73Cucumis melo pickled with sake lees
C75Cucumis melo pickled with sake lees
D75Pickled celery
FBA1Chinese cabbage pickled with salt
I71Pickled leaf mustard
ABRD-LbPl-9 (AB9)Asahi breweries collection
ABRD-LbPl-10 (AB10)Asahi breweries collection
ABRD-LbPl-11 (AB11)Asahi breweries collection
ABRD-LbPl-12 (AB12)Asahi breweries collection
ABRD-LbPl-13 (AB13)Asahi breweries collection
ABRD-LbPl-14 (AB14)Asahi breweries collection
ABRD-LbPl-15 (AB15)Asahi breweries collection
ABRD-LbPl-16 (AB16)Asahi breweries collection
Lactobacillus pentosus
JCM 1558T
NRIC 390
NRIC 391
NRIC 1925
NBRC 12011
Lactobacillus plantarum ssp. plantarum
JCM 1149T
LAB1Fermented rice bran paste
LAB2Pickled radish
LAB3Pickled radish
LAB7Salted vegetables
LAB8Salted vegetables
LAB9Radish fermented in rice bran
Lactobacillus sakei
JCM 1157T
FBA3Pickled leaf mustard
  • * After ERIC analysis, the strain was identified as Lactobacillus paraplantarum.


Bacterial cells were collected from 1 mL of an overnight culture containing approximately 1 × 109 cells by centrifugation at 10 000 g for 1 min from which genomic DNA was purified using a DNeasy Blood and Tissue Kit (Qiagen, Tokyo, Japan) according to the manufacturer's instructions. All PCR runs were performed in the same thermal cycler by a single investigator, but each extract was run separately.

ERIC-PCR was performed using ERIC-1R (5′-ATGTAAGCTCCTGGGGATTCAC-3′) and ERIC-2 (5′-AAGTAAGTGACTGGGGTGAGCG-3′) primers (Versalovic et al., 1991). PCR amplifications were carried out in a 50-μL reaction volume containing 1 × PCR buffer [120 mM Tris-HCl, 10 mM KCl, 6 mM (NH4)2SO4, 1 mM MgSO4, 0.1% Triton X-100, 0.001% bovine serum albumin, pH 8.0], 200 μM dNTPs, 1 U KOD Plus DNA polymerase (Toyobo, Japan), 35 ng template DNA, and 0.3 μM ERIC-1R and ERIC-2 primers. Amplifications were performed in a DNA thermal cycler (2400, Perkin-Elmer) under the following cycling conditions: an initial 95 °C for 5 min; 30 cycles at 90 °C for 30 s, 50 °C for 30 s, 52 °C for 1 min, and 72 °C for 1 min; and a final 72 °C for 8 min, with ramping speed 1 °C s−1.

For RAPD-PCR, OPL-01 (5′-GGCATGACCT-3′), OPL-02 (5′-TGGGCGTCAA-3′), OPL-04 (5′-GACTGCACAC-3′), or OPL-05 (5′-ACGCAGGCAC-3′) were used (Van Reenen & Dicks, 1996). PCR amplifications were carried out in a 20-μL reaction volume containing 1 × Ex Taq buffer, 200 μM dNTPs, 0.5 U Ex Taq DNA polymerase, 32 ng template DNA, and 1 μM of primer. Amplifications were performed in a PCR thermal cycler (Dice, Takara, Japan) under the following cycling conditions: an initial 94 °C for 5 min; 45 cycles at 94 °C for 1 min, 36 °C for 1 min, and 72 °C for 2 min; and a final 72 °C for 5 min, with ramping speed 2 °C s−1.

The ERIC- and RAPD-PCR products were separated by electrophoresis in 1.5% agarose gels and photographed. High-resolution images were obtained using a Fluor Chem 8900 fluorescence chemiluminescence and imaging system with alpha ease fc software (Alpha Innotech, San Leandro, CA), and the images were stored as TIFF files.

Cluster analysis

The TIFF images were analyzed using bionumerics v. 5.1 software (Applied Maths, Belgium). The band profiles were entered by a single investigator and saved into a single database. The gels were all normalized against size markers. All bands ranging in size from 200 to 5000 bp and observed three or more times among the five experiments were used in the phylogenetic analyses. Double bands were selected only when two distinct bands could be seen on the gel image and in the bionumerics densitometric curve window.

Phylogenetic analyses were performed using the Dice similarity coefficient (Dice, 1945) and the unweighted pair group method with arithmetic mean (UPGMA) cluster analysis based on numbers and positions of bands by bionumerics (Sneath & Sokal, 1973).

Cloning and sequencing of LpF1

Gel-purified LpF1 was cloned into the pCR-Blunt II-TOPO vector (Invitrogen) and sequenced using the M13 forward (−20) (5′-GTAAAACGACGGCCAG-3′), M13 reverse (5′-CAGGAAACAGCTATGAC-3′), P1-FBA1 (5′-CAGATGGTCAATCAACGATC-3′), and P2-FBA1 (5′-CCGGGTGGTGGATTTAAACC-3′) primers using a BigDye Terminator Cycle Sequencing Kit v. 3.1 (Applied Biosystems) in a 3730 Genetic Analyzer (Applied Biosystems). LpF1 was subsequently characterized by sequence similarity searches against the GenBank database using the blast algorithm (http://blast.ncbi.nlm.nih.gov/Blast.cgi) (Altschul et al., 1997).

Southern blot analysis

The FBA1-specific fragment (LpF2) was amplified using 35 ng of template DNA, P3-FBA1 (5′-TCTATAATTTGTGATACAGGGGTTGCC-3′), and P4-FBA1 (5′-CTCGTAATCACACAGAAATTATGCTGC-3′) under the following cycling conditions: an initial 94 °C for 3 min; 35 cycles at 94 °C for 15 s, 59 °C for 35 s, and 68 °C for 2 min; and a final 68 °C for 7 min. Genomic DNA (1 μg) from L. paraplantarum strains digested by Dra I were separated by a 1% agarose gel and transferred to nylon membranes (Roche Diagnostics GmbH, Mannheim, Germany). The LpF2 fragment (946 bp) was purified using a PCR purification kit (Qiagen) and labeled using a Digoxigenin (DIG) High Prime Kit (Roche Diagnostics GmbH) according to the manufacturers' instructions. Hybridization was carried out at 42 °C. Membrane was washed under conditions of high stringency at 68 °C. Detection was performed using an anti-DIG antibody alkaline phosphatase conjugate and CSPD. Membrane was activated at 37 °C for 10 min and developed to an X-ray film (Roche Diagnostics GmbH).

Results and discussion

Discrimination of Lactobacillus species

Strains were preliminarily classified by sequence analyses of pheS, rpoA (Naser et al., 2005), and 16S rRNA genes (Table 1) and further confirmed using PCR-based methods (Berthier & Ehrlich, 1999; Torriani et al., 2001a, b). To discriminate these strains, we evaluated repetitive element sequence-based (REP-) (Jersek et al., 1999), triplicate arbitrarily primed (TAP-) (Cusick & O'Sullivan, 2000), RAPD-, and ERIC-PCRs, but those except ERIC did not yield a band that was specific to L. paraplantarum strains (data not shown).

In ERIC-PCR, the L. paraplantarum strains tested had similar band profiles (Fig. 1a, lanes 7–13); the shared bands agreed with the type strain of L. paraplantarum (JCM 12533T, lane 7). The DNA bands of approximately 2.8, 1.1, 0.9, and 0.55 kb generated with the primer set ERIC-1R and ERIC-2 were common to strains of the species L. paraplantarum (Fig. 1a, horizontal arrows). Reproducibility can be a problem for fingerprints generated by PCR-based methods as was well documented previously (Cusick & O'Sullivan, 2000). In fact, the thermocycler model affected the fingerprints due to the difference in the thermal ramp. DNA preparation of each strain and enzyme lots affected the fingerprints considerably. Especially, amplification of the 2.8-kb band appeared in strains of L. paraplantarum depended on the activity of the polymerase. Therefore, we used a single thermocycler model with a single program and the bands that appeared at least three or more times among the five experiments were considered. Cluster analysis of the band profiles divided the strains into three clusters: the main cluster, AE, consisting exclusively of the L. paraplantarum strains; cluster BE, consisting of Lactobacillus curvatus and Lactobacillus sakei; and cluster CE, consisting of phenotypically hard-to-distinguish Lactobacillus pentosus and L. plantarum (Fig. 1b). The phylogenetic tree showed a similarity coefficient of 57.0% among the L. paraplantarum strains (Fig. 1b, cluster AE), but only 8.1% between these and BE.

Figure 1

(a) ERIC-PCR amplification profile and (b) dendrogram representing the genetic relationships among strains of Lactobacillus paraplantarum, Lactobacillus curvatus, Lactobacillus sakei, Lactobacillus pentosus, and Lactobacillus plantarum based on (a). (a) Lane 1, L. curvatus JCM 1096T; lane 2, L. curvatus FBA2; lane 3, L. curvatus LAB10; lane 4, L. curvatus LAB11; lane 5, L. pentosus JCM 1558T; lane 6, L. plantarum JCM 1149T; lane 7, L. paraplantarum JCM 12533T; lane 8, L. paraplantarum FBA1; lane 9, L. paraplantarum C75; lane 10, L. paraplantarum I71; lane 11, L. paraplantarum 2-51; lane 12, L. paraplantarum 6-01; lane 13, L. paraplantarum 5-67; lane 14, L. sakei FBA3; lane 15, L. sakei JCM 1157T. The typical band profile is presented. The diagonal arrows indicate strain-specific bands. The horizontal arrows indicate the bands commonly observed in L. paraplantarum species. (b) The dendrogram was constructed using the UPGMA, with correlation levels expressed as percentage values of Dice's coefficient.

In order to confirm the discriminatory effectiveness of the ERIC-PCR-based techniques, we performed ERIC analysis of 141 strains of LAB including 74 identified and 67 unidentified strains in our collection. The phylogenetic tree based on ERIC-PCR showed a cluster consisting of L. paraplantarum strains, in which five unidentified strains were included. After sequencing analysis and multiplex PCR (Torriani et al., 2001a, b), these strains were identified to the species L. paraplantarum (Table 1). This result showed that ERIC analysis is useful for the preliminary discrimination of L. paraplantarum from other Lactobacillus species.

Together with nine additional strains of L. paraplantarum, we performed ERIC analysis of 43 strains of Lactobacillus (Supporting Information, Fig. S1). The phylogenetic tree based on ERIC-PCR showed three clusters: a cluster consisting of L. paraplantarum strains, a cluster consisting of L. plantarum strains, and a cluster consisting of strains of L. pentosus, L. curvatus, and L. sakei. In the third cluster, a subcluster consisting strains of L. pentosus was distinguished from others consisting of strains of L. curvatus and L. sakei. Although L. paraplantarum, L. plantarum, and L. pentosus are considered to be phenotypically close (Curk et al., 1996), ERIC-PCR produced considerable DNA polymorphisms among these species; five bands of 3, 1.25, 1.05, 0.82, and 0.35 kb were typically observed in strains of the species L. plantarum, whereas the band of 0.82 kb was common to strains of the species L. pentosus. Further, three intensive bands of 1.15, 0.95, and 0.45 kb were common to most strains of the species L. curvatus. These data suggest that the ERIC-1R and ERIC-2 primers are useful for generating discriminatory polymorphisms from different species of Lactobacillus.

In RAPD-PCR, none of the four primers yielded a band that was specific to L. paraplantarum strains (Fig. 2a–d). In addition, L. paraplantarum I71 (lane 10) showed a band profile similar to that of L. curvatus LAB10 (lane 3) and L. sakei JCM 1157T (lane 15) and grouped into the same cluster with them (Fig. 2e, cluster BR). These results suggest that the OPL primers are unsuitable for discriminating L. paraplantarum.

Figure 2

RAPD-PCR amplification profiles using OPL-01 (a), OPL-02 (b), OPL-04 (c), and OPL-05 (d) and dendrogram representing the genetic relationships among strains of Lactobacillus paraplantarum, Lactobacillus curvatus, Lactobacillus sakei, L. pentosus, and L. plantarum based on the PCR profiles (a–d) (e). The lane assignments and the analytical conditions are the same as those in Fig. 1.

Discrimination of L. paraplantarum strains

ERIC analysis divided the L. paraplantarum strains into two major groups (Fig. 1b): group AE1 (JCM 12533T, FBA1, C75, I71, and 2-51; Fig. 1a, lanes 7–11) and group AE2 (5-67 and 6-01; lanes 12, 13). Although C75 and I71 were obtained from different sources, they showed highly similar band profiles (lanes 9, 10). Besides ERIC, REP-PCR, and TAP-PCR yielded indistinguishable band profiles (data not shown). In contrast, in RAPD analysis, they showed entirely different band profiles (Fig. 2, lanes 9, 10), suggesting that RAPD-PCR aids discrimination of L. paraplantarum strains. The phylogenetic tree based on RAPD-PCR showed a main cluster, AR, consisting exclusively of L. paraplantarum strains with a similarity level of 43.3% (Fig. 2e), while cluster AE of ERIC-PCR had a similarity level of 57.0% (Fig. 1b), illustrating the discriminatory ability of RAPD-PCR. Thus, the combination of ERIC and RAPD is effective for the molecular identification of L. paraplantarum strains.

Development of the FBA1-specific marker

Besides discriminating a species, it is very important to distinguish a particular industrial or probiotic strain from others to investigate the dynamics of the strain in certain products or in the gastrointestinal tract when ingested. Several strain-specific PCR products were obtained by ERIC or RAPD analysis (Figs 1 and 2, diagonal arrows). The band patterns themselves could be strain-specific DNA markers, but strain-specific PCR primers to amplify a specific product would be more useful. We applied the ERIC-PCR profile to develop an L. paraplantarum FBA1 strain-specific marker to provide a more powerful tool for the discrimination of individual L. paraplantarum strains; we focused on strain FBA1 and a 2.2-kb FBA1-specific product, LpF1, by ERIC-PCR (Fig. 3).

Figure 3

(a) Schematic drawing of the FBA1-specific ERIC fragment LpF1 and LpF2 and the homologous region of Lactobacillus fermentum ATCC 14931. The primer regions are indicated by small black arrows. The nucleotide sequence of LpF1 (2265 kb) was registered under accession number AB520823 in the DDBJ database (http://www.ddbj.nig.ac.jp/). (b) Amplification of LpF2 in Lactobacillus paraplantarum strains by PCR analysis. Lane 1, AB9; lane 2, AB10; lane 3, AB11; lane 4, AB12; lane 5, AB13; lane 6, AB14; lane 7, AB15; lane 8, AB16; lane 9, FBA1; lane 10, 2-51; lane 11, 5-67; lane 12, 6-01; lane 13, C75; lane 14, I71; lane 15, JCM 12533T. (c) Southern blot analysis of Dra I-digested genomic DNA from L. paraplantarum strains using DIG-labeled LpF2. Lane 1, AB9; lane 2, AB10; lane 3, AB11; lane 4, AB12; lane 5, AB13; lane 6, AB14; lane 7, AB15; lane 8, AB16; lane 9, C75; lane 10, D75; lane 11, I71; lane 12, 2-51; lane 13, FBA1; lane 14, 5-67; lane 15, 6-01; lane 16, JCM 12533T.

We cloned and sequenced LpF1. The fragment was 2265 bp long and, contrary to our expectation, had the ERIC-2 primer sequence at both ends (Fig. 3a, arrows), suggesting that LpF1 was amplified by the ERIC-2 primer. In fact, PCR reaction with a single ERIC-2 primer generated four DNA bands, including LpF1 (data not shown). The DNA sequence of LpF1 had no significant similarity to any sequences in the EMBL/GenBank/DDBJ database. The sequence contained three ORFs 831, 864, and 453 bp long, encoding putative proteins of 277, 288, and 151 amino acid residues, respectively; the third ORF is truncated and does not end with a stop codon. The amino acid sequences of ORFs showed 61%, 48%, and 50% identities with the AroA (3-phosphoshikimate 1-carboxyvinyltransferase; 3e−86)(C0WY56), TyrA (prephenate dehydrogenase; 1e−65)(C0WY57), and AroK (shikimate kinase; 9e−35)(C0WY58), respectively, encoded in the aroAtyrAaroK gene cluster of Lactobacillus fermentum ATCC 14931 (Fig. 3a). The same aroAtyrAaroK (I) cluster exists in L. fermentum IFO 3956, L. plantarum JDM1, and Lactobacillus brevis ssp. gravesensis ATCC 27305, but tyrA is missing in the cluster in Lactobacillus antri DSM 16041. These proteins are components of the shikimate pathway, which biosynthesizes aromatic compounds, such as phenylalanine (Herrmann & Weaver, 1999).

Although the nucleotide sequence of LpF1 showed very low similarity to other genes, the ERIC-2 primer region at both ends had similarity to other genes. Therefore, a set of specific primers (P3-FBA1 and P4-FBA1) was designed from the internal sequence of LpF1 to amplify a 950-bp product. PCR analysis showed that the 950-bp product (LpF2) was specifically amplified from genomic DNA of FBA1 among 16 L. paraplantarum strains (Fig. 3b). Further, Southern analysis of Dra I-digested genomic DNA was carried out using LpF2 as a probe. The probe only hybridized with the 4-kb Dra I fragment of FBA1 among 16 L. paraplantarum strains, suggesting that LpF2 is a unique marker of the L. paraplantarum FBA1 strain (Fig. 3c). LpF2 can be applied to assess the survival of FBA1 through the gastrointestinal tract.

In summary, the combination of ERIC- and RAPD-PCR was sufficient for the discrimination of L. paraplantarum strains. Further, when no gene sequence data of a particular strain are available, ERIC-PCR can be an efficient tool to provide the strain-specific information. Genomic Southern blot analysis using the LpF2 probe uniquely identified L. paraplantarum FBA1. Because both ERIC- and RAPD-PCR are fast and technically simple methods, they are useful for the rapid discrimination of L. paraplantarum strains and for the development of new strain-specific DNA markers for identifying industrially important strains.

Supporting Information

Additional Supporting Information may be found in the online version of this article:

Fig. S1. Dendrogram of ERIC-PCR analysis of 43 strains of LAB including five strains of Lactobacillus pentosus, 10 strains of Lactobacillus plantarum, 10 strains of Lactobacillus curvatus, two strains of Lactobacillus sakei, and 16 strains of Lactobacillus paraplantarum.

Please note: Wiley-Blackwell is not responsible for the content or functionality of any supporting materials supplied by the authors. Any queries (other than missing material) should be directed to the corresponding author for the article.


  • Editor: Ezio Ricca


View Abstract