Molecular Fingerprinting of Fish-Pathogenic Lactococcus garvieae Strains by Random Amplified Polymorphic DNA Analysis

Abstract

In this work, we used the random amplified polymorphic DNA (RAPD) technique to evaluate the genetic diversity in Lactococcus garvieae, an important pathogen for fish. Fifty-seven strains with different hosts and geographical origins, including Japan and several countries of the Mediterranean area such as Spain, Portugal, France, Italy, England, and Turkey, were analyzed. Two primers, oligonucleotides 5 and 6 (Pharmacia Biotech) were utilized; primer 5 was the most discriminative, since allowed us to differentiate 10 RAPD -types related to the origin of the strains. Regardless of the oligonucleotide primer employed, the 57 isolates of L. garvieae studied were separated into three genetic groups, composed of the Spanish, Portuguese, English, and Turkish strains (group A), the Italian and French strains (group B), and the Japanese strains (group C). The similarity of isolates within each group, estimated on the basis of the Dice coefficient, ranged from 75 to 100%. Our findings also indicate that RAPD profiling constitutes a useful tool for epidemiological studies of this fish pathogen.

Lactococcus garvieae, one of the major gram-positive cocci pathogenic for fish, is considered a serious problem in cultured marine and freshwater fish species such as yellowtail (Seriola quinqueradiata) in Japan and rainbow trout (Onchorynchus mykiss) in Europe and Australia (4, 8, 10-12, 19-21). In Spain, this disease has appeared in rainbow trout farms since 1991 (9, 28) and at present is considered one of the most important risk factors in the trout industry during the summer months. Besides fish, L. garvieae has been isolated from cows and buffalo (6, 31); it has also been recovered from humans sources (13, 14). In addition, it has been isolated recently from diseased freshwater prawns (Macrobrachium rosembergii) in Taiwan (5). All these facts indicate the expanding importance of L. garvieae.

Despite the role of this microorganism as an infectious agent, few studies of the epidemiology of L. garvieae recovered from fish have been published until now. Epidemiological investigations depend on the availability and reliability of highly discriminatory typing systems which may differentiate between strains from different sources (23). Ribotyping (RT) and pulsed-field gel electrophoresis (PFGE) have been used for epidemiological characterization of L. garvieae and have shown high genetic variability within this species (12, 35). The typing schemes proposed by these authors are complicated and have limited epidemiological value. Moreover, the RT and PFGE techniques usually involve time-consuming steps and specific equipment (16, 25). Randomly amplified polymorphic DNA (RAPD) is an accesible and sensitive method based on the use of arbitrary primers to amplify polymorphic segments of DNA (38). Although this technique has been widely used in recent years for the study of genetic diversity among isolates of a number of bacterial fish pathogens (3, 18, 24, 27, 30, 33, 34, 37) to our knowledge no studies were reported for L. garvieae.

In the present work, RAPD analysis was employed to establish DNA fingerprints for L. garvieae strains from diverse hosts and geographical areas of the world, with the aim of evaluating the applicability of this technique in epidemiological studies.

MATERIALS AND METHODS

Bacterial strains and DNA extraction.

Fifty-seven L. garvieae strains were included in this study. The host and geographic origin of these strains are shown in Table 1. The reference strain, L. garvieae NCDO 2155, was included for comparative purposes. Strains were routinely grown on Columbia sheep blood agar (Oxoid Ltd., Madrid, Spain) plates at 25°C for 24 to 48 h. Stock cultures was maintained frozen at −80°C in tryptone soy broth (Difco, Madrid, Spain) with 15% glycerol.

TABLE 1.

Origin of the L. garvieae isolates and their distribution in the different genomic groups and RAPD profiles

Genomic group	RAPD type with primer:		Strain	Origin (species, country)	Yr of isolation
	P5	P6
A	I	1	TW I	Rainbow trout, Spain	1991
A	I	1	TW II	Rainbow trout, Spain	1991
A	I	1	TW 446.B3	Rainbow trout, Spain	1997
A	I	1	PP 60.1	Rainbow trout, Spain	1997
A	I	1	PP 61.1	Rainbow trout, Spain	1997
A	I	1	EW 175-97	Rainbow trout, Spain	1997
A	I	1	AR1	Rainbow trout, Spain	1999
A	I	1	AR2a	Rainbow trout, Spain	1999
A	I	1	AR2b	Rainbow trout, Spain	1999
A	I	1	301-99	Rainbow trout, Spain	1999
A	I	1	6-99	Rainbow trout, Spain	1999
A	I	1	2725	Rainbow trout, Spain	1999
A	I	1	3712	Rainbow trout, Spain	2000
A	I	1	3682	Rainbow trout, Spain	2000
A	I	1	3756	Rainbow trout, Spain	2000
A	I	1	279-00	Rainbow trout, Spain	2000
A	I	1	321	Rainbow trout, Spain	2000
A	I	1	322	Rainbow trout, Spain	2001
A	I	1	326	Rainbow trout, Spain	2001
A	I	1	350	Rainbow trout, Spain	2001
A	I	1	356	Rainbow trout, Spain	2001
A	I	1	PT 2.1	Rainbow trout, Portugal	2002
A	I	1	PT 4.3	Rainbow trout, Portugal	2002
A	I	1	PT 15.1	Rainbow trout, Portugal	2002
A	I	1	PT 18.3	Rainbow trout, Portugal	2002
A	I	1	00-21	Rainbow trout, England	2000
A	II	2	E1	Rainbow trout, Turkey	2001
A	II	2	K1	Rainbow trout, Turkey	2001
A	II	2	B6	Rainbow trout, Turkey	2001
A	II	2	B28	Rainbow trout, Turkey	2001
A	II	2	Ç1	Rainbow trout, Turkey	2002
A	II	2	G2	Rainbow trout, Turkey	2002
A	III	2	297-8	Catfish, Italy	1999
A	IV	3	2138	Rainbow trout, France	2000
B	V	4	8338	Rainbow trout, France	1998
B	V	4	7898	Rainbow trout, France	2000
B	V	4	1958	Rainbow trout, France	2000
B	V	4	2468	Rainbow trout, France	2001
B	V	4	2753	Rainbow trout, France	2001
B	V	4	2754	Rainbow trout, France	2001
B	V	4	2887	Rainbow trout, France	2001
B	V	4	3031	Rainbow trout, France	2001
B	V	4	3032	Rainbow trout, France	2001
B	VI	4	657-1	Rainbow trout, Italy	1999
B	VI	4	657-2	Rainbow trout, Italy	1999
B	VI	4	2913	Rainbow trout, Italy	1999
B	VI	4	309-C1	Rainbow trout, Italy	2000
B	VI	4	309-C2	Rainbow trout, Italy	2000
B	VI	4	372	Rainbow trout, Italy	2000
B	VI	4	541	Rainbow trout, Italy	2000
B	VII	4	344	Rainbow trout, Spain	2001
C	VIII	5	1753	Rainbow trout, France	2000
C	IX	6	YT-3	Yellowtail, Japan	1974
C	IX	6	NG8206 KG−	Yellowtail, Japan	1982
C	IX	6	NG8206 KG+	Yellowtail, Japan	1982
C	IX	6	SS 91014	Yellowtail, Japan	1991
C	X	7	NCDO 2155	Cow, United Kingdom	1973

DNA isolation was performed using InstaGene Matrix (Bio-Rad, Madrid, Spain). Briefly, bacterial colonies were suspended in 1 ml of autoclaved water and centrifuged at 13,200 × g for 1 min. After the supernatants were removed, the pellets were resuspended in 200 μl of InstaGene Matrix and incubated at 56°C for 30 min. They were then vortexed at high speed for 10 s and boiled in a water bath for 8 min. The lysates were vortexed again at high speed and centrifuged at 13,200 × g for 3 min. The InstaGene DNA preparations were stored at −20°C until used for PCR amplifications.

Identification by PCR.

Confirmatory identification of the strains was performed by PCR. The sequences of the specific PCR primers for identification of L. garvieae isolates were obtained from Zlotkin et al. (39) (PLG-1, 5′ - CAT AAC AAT GAG AAT CGC-3′, PLG-2, 5′-GCA CCC TCG CGG GTT G-3′). These primers targeted a region of the 16S rRNA gene of L. garvieae (EMBL accession number X54262) and produced a 1,100-bp amplicon. The PCR mixture consisted of 10 mM Tris-HCl (pH 9.0), 50 mM KCl, 1.5 mM MgCl₂, 200 μM each deoxynucleoside triphosphate (Pharmacia Biotech, Barcelona, Spain), primers PLG-1 and PLG-2 (8 ng/μl each), 1.5 U of Taq polymerase, and 1 μl of DNA solution in a total volume of 25 μl. Amplification conditions were those described by Zlotkin et al. (39) with minor modifications and included a denaturation step of 3 min at 94°C followed by 35 cycles of denaturation (1 min at 94°C), annealing (1 min at 56°C), and extension (1.5 min at 72°C), with a final extension step at 72°C for 10 min. Amplified products were separated by electrophoresis on a 1.5% agarose gel and stained with ethidium bromide (Bio-Rad). A negative control, consisting of the same reaction mixture but with distilled water instead of template DNA, was included in each run. In addition, DNA from Streptococcus iniae (ATCC 29178) and Enterococcus faecalis (ATCC 19433) was included as negative controls.

RAPD analysis.

The PCR amplifications were performed using Ready-to-Go RAPD analysis beads (Pharmacia Biotech). These commercial beads have been optimized for PCR amplifications and contain buffer, nucleotides, and Taq DNA polymerase. The only reagents which must be added to the reaction are template DNA and primers, which are also supplied in the kit. Six distinct random 10-mer primers (Pharmacia Biotech) were used: primers P1 (GGTGCGGGAA), P2 (GTTTCGCTCC), P3 (GTAGACCCGT), P4 (AAGAGCCCGT), P5 (AACGCGCAAC), and P6 (CCCGTCAGCA). Each 25.0-μl RAPD reaction mixture contained 1.5 U of Taq polymerase, 10 mM Tris-HCl (pH 9.0), 50 mM KCl, 1.5 mM MgCl₂, 200 μM each deoxynucleoside triphosphate, 25 pmol of the respective primer, and 1 μl of DNA solution. Amplification was performed in a Master Cycler Personal (Eppendorf) programmed as follows: an initial denaturation step at 95°C for 5 min followed by 30 cycles of denaturation (95°C for 1 min), annealing (35°C for 1 min), and extension (72°C for 2 min), with a final extension step at 72°C for 5 min. The RAPD products were electrophoresed as described above for the PCR products. The gels were photographed under UV light. A 50- to 2000-bp ladder (Sigma, St. Louis, Mo.) was used as a molecular mass marker.

To determine significant differences in the patterns, the reproductibility of results was assessed by repetition of at least three independent RAPD assays.

Computer data analysis.

All the gels were also scanned and the images were captured by a Gel Doc-2000 gel documentation system (Bio-Rad). The data analysis was performed by using Diversity Database software (Bio-Rad), and the computed similarities among isolates were estimated by means of the Dice coefficient (S_d) (7). Dendrograms were produced on the basis of the unweighted average pair group method (UPGMA).

RESULTS

The isolates used in this study, regardless of their origin, were confirmed as belonging to L. garvieae, since in the PCR assay used for identification, all of them gave the specific amplification product of 1,100 bp (data not shown). No amplification was observed for the E. faecalis and S. iniae reference strains included in the assays. Initially a RAPD analysis of one strain of L. garvieae (TW 446.B3) was performed using each of the six primers provided in the commercial kit. Only two of them, oligonucleotides P5 and P6, generated reproducible patterns with an appropriate number of amplified products (between 7 and 14 bands depending on the primer used) suitable for an accurate analysis (see Fig. 1 and 2). The other primers (P1 through P4) gave a small number of PCR products or very poor amplification, they were therefore not suitable for the genetic study of this pathogen (data not shown). Primers P5 and P6 were then used to analyze the complete collection of L. garvieae strains. A set of reproducible bands produced for a particular primer was defined as a pattern or profile. The RAPD assays were repeated at least three times for each primer tested. Apart from some variations in the band intensity, no significant differences were observed between the profiles obtained, which demonstrated the reproducibility of the RAPD method.

FIG. 1.

RAPD fingerprints obtained for the L. garvieae isolates with primer 5. Lanes: MW, AmpliSize Molecular Ruler (50- to 2000-bp ladder); 1, TW446.B3; 2, E1; 3, 297-8; 4, 2138; 5, 2754; 6, 2753; 7, 372; 8, 344; 9, YT-3; 10, NCDO 2155. The molecular sizes (in base pairs) are indicated on the left.

FIG. 2.

RAPD fingerprints obtained for the L. garvieae isolates with primer 6. Lanes: MW, AmpliSize Molecular Ruler (50- to 2000-bp ladder); 1, TW446.B3; 2, E1; 3, 297-8; 4, 2138; 5, 2754; 6, 372; 7, 344; 8, 1753; 9, YT-3; 10, NCDO 2155. The molecular sizes (in base pairs) are indicated on the left.

The patterns obtained with primer 5, designated types I to X (Table 1), showed between 7 and 14 bands ranging from 150 to 3,000 bp in size (Fig. 1). All but one of the Spanish isolates, the English isolate, and the Portuguese isolates joined in a unique profile I (Fig. 1, lane 1), which showed minimal differences from profiles II, obtained from all the isolates from Turkey (lane 2), and III, obtained from the Italian catfish strain (lane 3). On the other hand, the French strains were distributed into three different profiles (IV, V, and VIII) (lanes 4, 5, and 6, respectively), while the Italian isolates from trout were grouped in profile VI (lane 7). Moreover, all the isolates from yellowtail in Japan belonged to type IX (lane 9). Two profiles contained a unique strain, profile VII, which included Spanish strain 344, isolated from vaccinated fish (lane 8), and profile X, which included reference strain NCDO 2155 of bovine origin (lane 10).

When primer 6 was used, seven distinct patterns were observed and were designated types 1 to 7, comprising four to seven bands with sizes ranging between 150 and 2,200 bp (Table 1; Fig. 2). The majority of the Spanish isolates, together with the English and Portuguese isolates, were included in profile 1 (Fig. 2; lane 1) and showed a close relationship to the isolates from Turkey and the Italian catfish strain, which yielded the same pattern profile, (lanes 2 and 3), and French strain 2138, which exhibited a different profile (type 3) (lane 4). On the other hand, the most the French isolates, all the Italian isolates, and Spanish strain 344 exhibited the same profile (type 4) (lanes 5, 6, and 7). In addition, all the Japanese isolates from yellowtail belonged to type 6 (lane 9), yielding a great similarity to reference strain NCDO 2155 (type 7) (lane 10). Finally, profile 5 contained a unique strain, French isolate 1753 (lane 8).

The results of the analysis of similarity among the different profiles with the Diversity database software employing the S_d and the UPGMA allowed us to identify three genetic groups (A, B, and C) among the L. garvieae isolates studied in this work. Interestingly, although the numbers of RAPD patterns were different when primer 5 and primer 6 were employed, the clustering of the isolates was exactly the same regardless of the primer used (Table 1; see Fig. 3 and 4). In both cases, genetic group A included most of the Spanish isolates, the English, Portuguese, and Turkish isolates, the Italian isolate from catfish, and French strain 2138. Group B comprised the majority of the French isolates, the Italian strains isolated from trout, and Spanish strain 344. Finally, all the yellowtail isolates from Japan, French strain 1753, and reference strain NCDO 2155 were found in group C.

FIG. 3.

Dendrogram established by the Diversity Database software package (Bio-Rad) using the Dice similarity coefficient and UPGMA on the basis of the RAPD profiles of L. garvieae strains obtained with primer 5.

FIG. 4.

Dendrogram established by the Diversity Database software package (Bio-Rad) using the Dice similarity coefficient and UPGMA on the basis of the RAPD profiles of L. garvieae strains obtained with primer 6.

The similarity values obtained in the two computerized analyses were also very similar. Thus, with primer 5, the three genetic groups were defined at S_d values of 90% for cluster A (profiles I to IV), 80% for group B (profiles V to VII), and 75% for group C (patterns VIII to X) (Fig. 3). Genetic groups B and C showed a closer relationship, with a similarity of 45%, while group A exhibited an S_d of 34% with the two other groups (Fig. 3). On the other hand, when primer 6 was used, group A (profiles 1 to 3) yielded an S_d of 93% while groups B (Profile 4) and C (profiles 5 to 7) were defined at similarities of 100 and 83% respectively (Fig. 4). However, and in contrast to the results obtained with primer 5, groups A and C were more closely related, with an S_d of 55%. The three genetic groups joined at a similarity of 28% (Fig. 4).

DISCUSSION

L. garvieae is an important emerging pathogen in intensive aquaculture (2). The identification schemes for L. garvieae, based on biochemical and antigenic characterists (13, 17), can barely differentiate this microorganism from other gram-positive cocci such as the human pathogen L. lactis subsp. lactis (13, 14) or the Enterococcus-like strains isolated from diseased fish (12, 32, 36). Studies involving the phenotypic characterization of L. garvieae strains, collected from different species and countries, have been conducted using conventional methods and miniaturized systems and have given variable results (12, 13, 28, 31, 35). This fact could make the definitive identification of the isolates on the basis of phenotypic tests alone even more difficult, requiring the application of molecular techniques, like specific PCR, to confirm the taxonomic position of the isolates. Two PCR protocols have been reported for L. garvieae, targeting the 16S rRNA (39) and the dihydropteroate synthase genes (1). Due to the advantages of the the 16S rRNA gene-based PCR (see reference 29 for a review), in the present work, this method was used to confirm the identification of all the isolates of L. garvieae. It is important to point out that the PCR conditions had to be slightly changed (the annealing temperature was increased by 1°C) to obtain the specificity reported by Zlotkin et al. (39). The reason for this change in the PCR temperatures can probably be attributed to the use of a different thermal cycling apparatus.

Despite the large number of studies of the phenotypic characterization of this pathogen (12, 13, 28, 31, 35), few works have been done, to our knowledge, on epidemiological aspects of L. garvieae isolated from fish. Eldar et al. (12) used RT to analyze 15 strains of this pathogen isolated from diseased fish in Europe, Asia, and Australia. The use of endonucleases EcoRI and HindIII resulted in two and seven ribotypes, respectively; a close relationship was detected between the Japanese and Italian isolates in both cases. In another study, Vela et al. (35), using the PFGE method, described 19 different pulsotypes in a collection of 84 L. garvieae strains, including trout, bovine, human and water isolates, most of which were isolated from European countries. These results indicated the existence of three genetically unrelated clones, one comprising Spanish and Portuguese isolates and the other two composed of the Italian and French strains, respectively. From these studies, it seems clear that there is high genetic heterogeneity within L. garvieae. However, the usefulness of these techniques as epidemiological tools can be controversial. Ideally, a typing method should be inexpensive, rapid, and simple to perform with commonly available equipment (16, 23, 25). RT and especially PFGE are laborious and time-consuming procedures, are difficult to optimize, and require specific equipment not available in many laboratories. A possible alternative is RAPD, a simpler and faster method, which has proved to be appropriate for epidemiological analysis of a variety of bacteria, including fish pathogens (3, 18, 24, 27, 30, 33, 34, 37).

In the present work, the RAPD procedure was used to analyze a collection of European and Japanese strains of L. garvieae isolated from rainbow trout, catfish, and yellowtail. Of the two primers employed, primer 5, yielding 10 different RAPD types, proved to be more discriminating than primer 6, which distinguished only 7 RAPD profiles. The different patterns obtained with primer 5 could be related to both the geographical origin and the primary host of the isolates, which indicate its potential use in epidemiological studies of this fish pathogen.

The similarity analysis of the patterns obtained RAPD allowed the differentiation of three genogroups for the 52 isolates of L. garvieae examined, regardless of the primer used. Thus, the Spanish, English, Portuguese, and Turkish trout isolates, together with the strain isolated from catfish in Italy, displayed a close genetic relationship. On the other hand, French and Italian strains isolated from trout were genetically related. It is interesting that Spanish strain 344, isolated from an outbreak in vaccinated fish, showed a close relationship to the Italian isolates. Finally, the Japanese isolates from yellowtail are members of a different clone, showing a great similarity to the reference strain, L. garvieae NCDO 2155. The movement of infected fish, asymptomatic carriers, or even contamineted eggs into susceptible populations is one of the main factors implicated in the dissemination of fish diseases (15, 22). This could explain the appearance of the same RAPD type in different countries. On the other hand, it was found that as a consequence of the selective pressure induced by vaccination, a new RAPD type of S. iniae was isolated from trout in Israel (3). Something similar may be occurring with L. garvieae, since in countries with a long history of vaccination against lactcoccosis or in which distinct vaccine formulations are used, such as Italy and Spain, the RAPD types found are completely different. Thus, protection against a particular variant of the pathogen may induce infection with other variants present in the environment. The finding that Spanish isolate 344, responsible for an outbreak in vaccinated fish during the protective period, showed a profile different from the other Spanish strains may support this hypothesis.

In conclusion, we think that RAPD fingerprinting provides a discriminatory and rapid means of comparing L. garvieae isolates and may be an useful tool for epidemiological studies of this genetically heterogeneous fish pathogen. In addition, knowledge of the geographical distribution of different genetic groups of this fish pathogen may be very helpful in designing preventive measures for effective control of lactococcosis, such as vaccine formulations.