Development of a 16S–23S rRNA intergenic spacer-based quantitative PCR assay for improved detection and enumeration of Lactococcus garvieae


Correspondence: Hyoung-Shik Shin, DDS, PhD, Department of Periodontology, Wonkwang University College of Dentistry, Iksan 570-749, Korea. Tel.: +82 63 850 1965; fax: +82 63 857 6364; e-mail:


Lactococcus garvieae is an important foodborne pathogen causing lactococcosis associated with hemorrhagic septicemia in fish worldwide. A real-time quantitative polymerase chain reaction (qPCR) protocol targeting the 16S–23S rRNA intergenic spacer (ITS) region was developed for the detection and enum-eration of L. garvieae. The specificity was evaluated using genomic DNAs extracted from 66 cocci strains. Fourteen L. garvieae strains tested were positive, whereas 52 other strains including Lactococcus lactis ssp. lactis, Lactococcus lactis ssp. hordniae and Lactococcus lactis ssp. cremoris did not show a specific signal. The minimal limit of detection was 2.63 fg of purified genomic DNA, equivalent to 1 genome of L. garvieae. The optimized protocol was applied for the survey of L. garvieae in naturally contaminated fish samples. Our results suggest that the qPCR protocol using ITS is a sensitive and efficient tool for the rapid detection and enumeration of L. garvieae in fish and fish-containing foods.


Members of the genus Lactococcus are often isolated from food-related sources and are therefore generally regarded as safe. However, among seven Lactococcus species, Lactococcus garvieae is known as an important pathogen that causes disease in fish and mammals (Wang et al., 2007; Li et al., 2008). Lactococcosis outbreaks have been reported in ready-to-eat food products that are consumed without cooking (Novotny et al., 2004). Recently, gastrointestinal disorders, endocarditis, bacteremia, peritonitis, liver abscess, and osteomyelitis have been associated with consumption of raw fish contaminated with L. garvieae; thus, the importance of controlling contamination of foods is increasing (James et al., 2000; Wang et al., 2007; Li et al., 2008; Chan et al., 2011). In South Korea, this microorganism has caused outbreaks in aquaculture, such as in black rockfish (Kang et al., 2004), flounder (Baeck et al., 2006; Jeong et al., 2006), and has been isolated from fermented fish (Jung et al., 2010).

Bacterial identification methods based on biochemi-cal tests, such as the miniaturized API system, or conventional culture have traditionally been used for the identification of L. garvieae in fish (Zlotkin et al., 1998). However, distinguishing this microorganism from other lactic acid bacteria commonly found in fermented foods, such as Streptococcus thermophilus, Lactococcus lactis, or Enterococcus-like strains, remains difficult and unreliable (Casalta & Montel, 2008; Ogier & Serror, 2010).

Real-time quantitative polymerase chain reaction (qPCR) is a popular method for identifying and monitoring contaminant bacterial populations in foods because of its sensitivity and time-efficiency (Casalta & Montel, 2008; Ogier & Serror, 2010). At present, qPCR with 16S rRNA gene-targeted primers is one of the most popular methods for identifying predominant bacterial populations (Matsuki et al., 2004; Haarman & Knol, 2006). However, it has been demonstrated that the sensitivity of qPCR is not sufficient for the accurate quantification of subdominant populations in Gram-positive cocci (Matsuda et al., 2009). Indeed, the ribosomal RNA gene primer set of L. garvieae also amplifies DNA fragments from other bacterial species. By contrast, the 16S–23S RNA intergenic spacer (ITS) is considered a good monitoring tool for bacterial identification and strain typing (Hoffmann et al., 2010).

A real-time PCR for the detection of L. garvieae with primers targeting the 16S RNA gene has been applied directly to foods (Jung et al., 2010). However, this qPCR method showed a false-positive amplification with Lactococcus lactis ssp. lactis strains (Dang et al., 2012), and quantification of L. garvieae after an enrichment step has not been applied in foods. Despite the potential for foodborne transmission to humans, monitoring protocols that target L. garvieae in fish and fish-containing foods have not been developed. In this study, we reported an improved qPCR procedure targeting the ITS for detection and quantification of L. garvieae in fish and fermented fish products.

Materials and methods

Bacterial strains

The bacterial strains used in this study are listed in Table 1. The strains were obtained from the American Type Culture Collection (ATCC, Manassas, VA), the Belgian Co-ordinated Collections of Micro-organisms (BCCM/LMG, Gent, Belgium), the Culture Collection of the University of Gothenburg (CCUG, Gothenburg, Sweden), the Deutsche Samm-lung von Mikroorganismen und Zellkulturen GmbH (DSMZ, Braunschweig, Germany), the Korean Collection for Oral Microbiology (KCOM, Gwangju, Korea), and the Korean Collection for Type Cultures (KCTC, Daejeon, Korea). Lactococci, enterococci, and vagococci were grown aerobically at 30 or 37 °C for 18 h in trypticase soy yeast extract medium (Difco, Detroit, MI), and streptococci were grown under microaerophilic conditions on sheep blood agar (Asan Pharm Co., Seoul, Korea) at 37 °C for 20 h.

Table 1. Bacterial strains (n = 66) and PCR results
No.SpeciesDesignationIsolation sourcePCR
  1. +, PCR product was amplified with the ITSLg-F and ITSLg-R primer set; −, PCR product was not amplified with the ITSLg-F and ITSLg-R primer set.

1 Lactococcus garvieae KCTC 3772TBovine mastitis+
2 Lactococcus garvieae LMG 8162Bovine mastitis+
3 Lactococcus garvieae LMG 9472Raw milk+
4 Lactococcus garvieae LMG 8501Bovine mastitis+
5 Lactococcus garvieae KCTC 5620Diseased yellowtail+
6 Lactococcus garvieae KCTC 5621Turtle eye+
7 Lactococcus garvieae CAU 1101Flounder+
8 Lactococcus garvieae CAU 1102Pollack+
9 Lactococcus garvieae CAU 1103Flounder+
10 Lactococcus garvieae CAU 1104Pollack+
11 Lactococcus garvieae CAU 1105Pollack+
12 Lactococcus garvieae CAU 1106Flounder+
13 Lactococcus garvieae CAU 1107Flounder+
14 Lactococcus garvieae CAU 1108Flounder+
15Lactococcus lactis ssp. lactisKCTC 2013
16Lactococcus lactis ssp. lactisKCTC 3115
17Lactococcus lactis ssp. lactisKCTC 3191
18Lactococcus lactis ssp. lactisKCTC 3899Earthworm intestine
19Lactococcus lactis ssp. lactisKCTC 3769T
20Lactococcus lactis ssp. lactisKCTC 3926Dairy products
21Lactococcus lactis ssp. hordniaeKCTC 3768TLeaf hopper
22Lactococcus lactis ssp. cremorisDSM 20069T
23 Lactococcus raffinolactis KCTC 3982TRaw milk
24 Lactococcus plantarum DSM 20686TFrozen peas
25 Lactococcus chungangensis KCTC 13185TActivated sludge foam
26 Streptococcus anginosus ATCC 33397THuman oral cavity
27 Streptococcus anginosus KCOM 1063Human oral cavity
28 Streptococcus australis KCOM 1439Human oral cavity
29 Streptococcus australis KCOM 1441Human oral cavity
30 Streptococcus gordonii KCTC 3286TBacterial endocarditis
31 Streptococcus infantis KCOM 1375Human oral cavity
32 Streptococcus intermedius KCTC 3268THuman oral cavity
33 Streptococcus mitis KCTC 13047THuman oral cavity
34 Streptococcus mitis KCTC 3556THuman oral cavity
35 Streptococcus oralis KCTC 13048THuman plaque
36 Streptococcus oralis DSM 20066Human throat
37 Streptococcus oralis DSM 20395Human plaque
38 Streptococcus oralis DSM 20379Human plaque
39 Streptococcus oralis ATCC 9811THuman mouth
40 Streptococcus parasanguinis KCTC 13046THuman throat
41 Streptococcus parasanguinis KCOM 1352Human oral cavity
42 Streptococcus pyogenes KCTC 3984TScarlet fever
43 Streptococcus pyogenes KCTC 3208Pharynx of child
44 Streptococcus pneumoniae KCTC 5080TLower respiratory tract
45 Streptococcus pseudopneumoniae CCUG 49455TLower respiratory tract
46 Streptococcus sanguinis KCTC 3284TBacterial endocarditis
47 Streptococcus sanguinis KCOM 1428
48 Streptococcus sinensis KCOM 1017Human sinus
49 Streptococcus sinensis KCOM 1018Human sinus
50 Enterococcus hirae KCTC 3616T
51 Enterococcus mundtii KCTC 3630TSoil
52 Enterococcus casseliflavus KCTC 3638TPlant material
53 Enterococcus malodoratus KCTC 3641TGouda cheese
54 Enterococcus cecorum KCTC 3642TChicken cecum
55 Enterococcus saccharolyticus KCTC 3643TStraw bedding
56 Enterococcus villorum KCTC 13904TPig intestines
57 Enterococcus haemoperoxidus KCTC 13910TService water
58 Enterococcus moraviensis KCTC 13911TService water
59 Enterococcus phoeniculicola KCTC 3818TUropygial gland
60 Enterococcus solitarius KCTC 3923TEar exudates
61 Enterococcus raffinosus KCTC 5189TBlood culture
62 Enterococcus avium KCTC 5190THuman feces
63 Enterococcus faecalis KCTC 3206T
64 Vagococcus salmoninarum LMG 11491TRainbow trout
65 Vagococcus lutrae LMG 19537TOtter blood
66 Vagococcus fluvialis LMG 9464TChicken feces

Genomic DNA preparation

Bacterial genomic DNA used for qPCR was extracted from cultivated bacteria, fresh and fermented fish-containing foods using the bead-beat method, as described previously (Jung et al., 2010). Total bacterial DNA was quantified using the Infinite 200 NanoQuant (Tecan, Männedorf, Switzerland) at a wavelength of 260 nm.

Specificity of real-time PCR amplification

Real-time PCR assay was performed and monitored using SYBR Green chemistry with a 7500 Fast Real-time PCR System (Applied Biosystems, Foster City, CA). To amplify the 16S–23S rRNA ITS region of L. garvieae, we used two specific primers: ITSLg30F (5′-ACTTTATTCAGTTTTGAGGGGTCT-3′) positions 30–53 in L. garvieae KCTC 3772T and ITSLg319R (5′- TTTAAAAGAATTCGCAGCTTTACA-3′) positions 296–319 in L. garvieae KCTC 3772T (Dang et al., 2012). The qPCR amplification was performed in a total volume of 20 μL containing 1 μL of each template DNA, 1 μL of 10 pmole primers, 7 μL nuclease-free water, and 10 μL SYBR Green I master mix (Roche Diagnostics, Indianapolis, IN). qPCR assays were carried out using a standard program: PCR conditions consisted of an initial denaturation step of 5 min at 95 °C, followed by 45 cycles of 95 °C for 30 s, 60 °C for 30 s, and 72 °C for 30 s. A melting curve analysis was then performed from 95 to 65 °C at a rate of 0.1 °C s−1 with the continuous acquisition of fluorescence data. All samples were analyzed in triplicate in MicroAmp 96-well reaction plates (Applied Biosystems). Real-time PCR assay specificity was optimized and tested in real-time PCRs using total DNA extracted from 14 L. garvieae strains and 17 non-L. garvieae bacterial strains belonging to ten different species.

Sensitivity of qPCR

The sensitivity of the qPCR assay was evaluated using Lgarvieae KCTC 3772T. qPCR detection was performed on a 10-fold dilution series of purified DNAs from L. garvieae cells (26.3 × 10−7 to 1 ng μL−1). Lactococcus garvieae KCTC 3772T concentrations were calculated using the viable cell plate count method. Serial 10-fold dilutions of the cultures were plated onto Lactobacillus MRS agar (Difco laboratories), which were subsequently incubated at 37 °C for 24 h, and CFUs were determined in triplicate.

Inhibition control

For confirmation of the qPCR-negative food samples from L. garvieae cell, 8 μL of the purified DNA preparation from the foods was spiked shrimp with 2 μL of the extracted DNA from L. garvieae KCTC 3772T at a concentration corresponding to approximately 101–102 CFU per reaction. The qPCR was follow above same method, along with a positive and a negative control, to identify inhibition.

Standard curves

A calibration curve was constructed using fresh salmon and shrimp matrices that were negative for L. garvieae. For the construction of calibration curves in food samples, the strain of L. garvieae KCTC 3772T was used. The strain was streaked on Lactobacillus MRS agar incubated at 37 °C for 24 h, and then the cells were taken from the agar surface and cultured in 3 mL of Lactobacillus MRS broth (Difco) for 24 h. The cell suspension was serially diluted in Ringer's solution and counted on Lactobacillus MRS agar incubated at 37 °C for 24 h. Each dilution was inoculated into a 10 g of fresh packed salmon and shrimp purchased from super markets in Seoul. Then, 40 mL of Lactobacillus MRS broth was added. The solid sample was homogenized using a Stomacher 80 laboratory blender (Seward Medical, London, UK) for 2 min at maximum speed. Subsequently, 1 mL was recovered and mixed with 9 mL of Ringer's solution, and 1 mL of the diluted sample was used for DNA extraction as described above (DNA at t = 0). The DNA extraction was also performed on the samples after 24 h (t = 24) of incubation at 37 °C in BHI broth. At t = 24, 1 mL of a 10-fold dilution in Ringer's solution was processed. One microliter of the t = 0 DNA (in triplicate) was used in qPCR amplifications, and calibration curves were constructed plotting the threshold cycle against the colony forming units (CFU g−1). Similarly, the t = 24 DNA was used in amplification to determine the detection limit after overnight enrichment. The efficiency of the reactions was calculated according to the study by Rutledge & Cote (2003). Standard curves were constructed at least three times from three independent experiments. Contamination of background flora was checked using the L. garvieae-specific PCR analysis (Rutledge & Cote, 2003) and the API 20 Strep kit on Lactobacillus MRS medium. Suspected L. garvieae colonies from the Lactobacillus MRS agar plates were identified according to standard microbiological methods including the API 20 Strep method (BioMe′rieux sa, Marcy-l'Etoile, France) and L. garvieae-specific PCR analysis (Dang et al., 2012), and then L. garvieae colonies were subjected to qPCR amplification as described above.

Lactococcus garvieae detection in potentially naturally contaminated samples

Forty-three samples of fresh fish and packaged foods containing fermented fish (Korean name: Jeotgal) were obtained from super markets in Seoul and Sokcho in South Korea. The samples used in this study were as follows: yellow corvenia, codfish, Alaska pollock, hairtail, anchovy, oyster, shrimp and fermented sample of yellow corvenia, codfish, pollack, anchovy, squid, octopus, prawn, nautilus, oyster, scallop, and clam. The packaging was removed under a laminar flow hood. From each package, 10 g of the fish or other seafood was aseptically cut and mixed with 40 mL of Lactobacillus MRS broth and homogenized as above. One milliliter of the homogenate was mixed with 9 mL of Ringer's solution, and 1 mL of this mix was used for DNA extraction (t = 0). At the same time, the 1–10 and 1–100 dilutions of the homogenate were plated on Lactobacillus MRS agar. Plates were incubated at 37 °C for 24 h in aerobic conditions. The homogenate was also incubated for 24 h at 37 °C, DNA extraction was carried out from a 10-fold dilution in Ringer's solution, and a loopful of the enriched homogenate was streaked on Lactobacillus MRS agar and incubated at 37 °C for 24 h (t = 24).

Results and discussion

The method for detection and quantification of L. garvieae presented was based on the amplification of the partial 16S–23S rRNA ITS gene. The specificity of the amplification was optimized and tested in qPCRs using total DNA extracted from 14 L. garvieae strains and 17 non-L. garvieae bacterial strains belonging to different species of Lactococcus, Streptococcus, Enterococcus, and Vagococcus. Amplification signals occurred only with the 14 L. garvieae DNA. In contrast, non-L. garvieae strains, including closely related Lactococcus spp., did not show any signal results in the real-time PCR assay.

DNA was obtained from an L. garvieae culture at a concentration of 107 CFU mL−1. Serial 10-fold dilutions were carried out to determine the sensitivity of our qPCR protocol. Each DNA dilution (26.3 ag to 26.3 ng) was used to construct a standard curve and a minimal limit of detection. The minimum limit of detection of L. garvieae genomic DNA using the new ITS gene-based qPCR assay was 2.63 fg (about 1 CFU), with a mean CT value of 39.11 ± 1.54 (Table 2). The sensitivity of PCR amplification is dependent on the copy number, the age of the culture, and the method of cell lysis (Way et al., 1993). One of the advantages of primers based on the ITS region is the multiple copies of rRNA operons. This minimal value is 12 times higher than that of the qPCR method targeting the 16S rRNA gene molecule (Jung et al., 2010), indicating that our method is more sensitive for L. garvieae quantitation than the qPCR method targeting the corresponding 16S rRNA gene. The melting temperature for the amplicon from the L. garvieae type strain with the SYBR Green method was 87.5 °C, and the R2 value was 0.99.

Table 2. CT values for a dilution series of Lactococcus garvieae with 107 cells
DNA concentration (ng µL−1)Cell numberCT (mean ± SE)
  1. ND, not detected.

26.3 × 10−0(5.96 ± 4.51) × 10710.5 ± 0.01
26.3 × 10−1(5.96 ± 4.51) × 10614.5 ± 0.02
26.3 × 10−2(5.96 ± 4.51) × 10518.4 ± 0.10
26.3 × 10−3(5.96 ± 4.51) × 10422.6 ± 0.02
26.3 × 10−4(5.96 ± 4.51) × 10326.5 ± 0.04
26.3 × 10−5(5.96 ± 4.51) × 10230.8 ± 0.07
26.3 × 10−6(5.96 ± 4.51) × 10135.7 ± 0.48
26.3 × 10−7(5.96 ± 4.51) × 10039.1 ± 1.54

Real-time qPCR is regarded as the gold standard for accurate, sensitive, and rapid detection and enumeration of nucleic acid sequences. For this technique, SYBR Green I provides the simplest and most economical format for detecting and quantifying PCR products using the melting curve analysis (De Medici et al., 2003; Audemard et al., 2004). In our protocol, the primer sets with SYBR green fluorescence showed high amplification only to L. garvieae 101–102 CFU extracted DNA sample and have no other fluorescence signals from the spiked fish samples, such as shrimp. No amplification occurred in food samples indicating that this protocol is not interfering with SYBR green fluorescence and early contamination of L. garvieae.

A standard curve was created using the food matrices where there is a high incidence of L. garvieae. In particular, a standard curve was created starting from serially diluted L. garvieae cells in the salmon and shrimp samples. When inoculated salmon sample was used as a matrix, the mathematical expression of the log CFU mL−1 vs. the CT values obtained was as follows: y = −4.0942x + 38.074 with an R2 value of 0.99 (Fig. 1a). The number of bacteria in the inoculated shrimp sample ranged from log10 1.32 ± 0.22 (at CT values 34.83 ± 1.01) to log10 2.98 ± 0.12 (at CT values 23.47 ± 0.07) cells per gram according to the equation y = −4.1285x + 39.163 with an R2 value of 0.99 (Fig. 1b). This genome equivalent was calculated assuming that one molecule of L. garvieae DNA corresponds to 2.6 fg of DNA, using a genome size of 2.5 Mb, and determined according to the following equation: DNA amount in fg = bp × 660 Da bp−1 × 1.6 × 10−27 kg Da−1 × 1 × 10−18 fg kg−1 (Park et al., 2010). As the genome information of L. garvieae is not available from the current public databases, the genome size of 2.5 Mb for the L. lactis ssp. lactis KF147 (GenBank accession no. NC_013657) published on the NCBI GenBank database was utilized.

Figure 1.

Standard curves obtained via qPCR using ITSLg30F and ITSLg319R primers on 10-fold serial dilutions of Lactococcus garvieae in different food matrices (t = 0). (a) Inoculated L. garvieae in salmon, (b) inoculated L. garvieae in shrimp.

The qPCR-based protocol was applied in parallel to traditional microbiological analysis to detect and quantify L. garvieae in food samples both at t = 0 and at t = 24. The results obtained by the two approaches are summarized in Table 3. At t = 0, L. garvieae was detected and enumerated using the ITS-based primer set, ITSLg30F and ITSLg319R, in six of 43 fresh and fermented fish samples (15%). At t = 24, L. garvieae was detected in nine of 43 tested samples (22.5%). The melting analyses of amplicons obtained from qPCR assays were 87.5 °C ± 0.24. These results were confirmed via sequence analysis of the amplicons obtained through PCR and qPCR assays, which were 100% identical to the ITS fragment in the L. garvieae type strain KCTC 3772T (GenBank accession no. HM241913). By traditional microbiological analysis, two of 43 samples (4.7%) at t = 0 and two of 43 samples (4.7%) at t = 24 were positive for L. garvieae. These results are from suspected L. garvieae colonies that were randomly selected from Lactobacillus MRS agar and subjected to qPCR analysis and identification using API 20 Strep kit. Melting curve analysis at the end of each run resulted in melting temperatures for the positive samples that were similar to those of the L. garvieae used as a control in the qPCR.

Table 3. Detection and enumeration of Lactococcus garvieae in fish by qPCR
Sample (no. of positive samples/no. of samples tested)Results t = 0aResults t = 24b
Traditional analysis (CFU g−1)qPCR analysis (cell no. g−1)Traditional analysis (CFU g−1)qPCR analysis (cell no. g−1)
  1. a

    Results at t = 0 are reported as CFU g−1 as determined by the direct counts on Lactobacillus MRS and qPCR or based on the C(t) values obtained for the sample and the appropriate calibration curve.

  2. b

    Results at t = 24 are reported as positive or negative. In the traditional analysis, positive refers to the presence of suspected colonies on Lactobacillus MRS plate and confirmed by qPCR, while in the qPCR analysis it refers to a fluorescence signal obtained during amplification.

Yellow corvenia (0/3)0
Codfish (0/3)0
Alaska pollock (0/3)0
Hairtail (0/3)0
Anchovy (0/3)0
Oyster (0/3)0
Shrimp (0/3)0
Fermented yellow corvenia (0/2)0
Fermented codfish (0/2)0
Fermented pollack (1/2)481.72 ± 0.03++
Fermented anchovy (1/2)1.40 ± 0.16+
Fermented squid (1/2)1.52 ± 0.07+
Fermented octopus (1/2)5922.98 ± 0.12++
Fermented prawn (1/2)1.32 ± 0.22+
Fermented nautilus (1/2)1.64 ± 0.08+
Fermented oyster (1/2)0+
Fermented scallop (1/2)0+
Fermented clam (1/2)0+

In our ITS-based procedures, L. garvieae was successfully detected in six samples (t = 0) ranging from log10 1.32 ± 0.22 to log10 2.98 ± 0.12 cells g−1 and nine fermented fish (t = 24) from a set of 43 tested samples. It is important to note that there is potential for transmission of L. garvieae to humans via the food chain. Rates of contamination of food products vary among samples. There are many documented cases of L. garvieae infection due to eating contaminated food (Wang et al., 2007). The benefit of our new protocol is that it provides real-time results rather than next-day results, which may facilitate more appropriate responses.


This paper was sponsored by Wonkwang University in 2011. The authors declare that there is no conflict of interest related to the present manuscript.