OUP user menu

Restriction fragment length polymorphisms of 16S rDNA and of whole rRNA genes (ribotyping) of Streptococcus iniae strains from the United States and Israel

Avi Eldar, Sara Lawhon, Paul F. Frelier, Liliana Assenta, Bruce R Simpson, Patricia W Varner, Herve Bercovier
DOI: http://dx.doi.org/10.1111/j.1574-6968.1997.tb12564.x 155-162 First published online: 1 June 1997

Abstract

Streptococcus iniae (junior synonym S. shiloi) isolated from tilapia and trout in Israel and in the United States were subtyped by restriction length polymorphism (RFLP) based on PCR amplified 16S rDNA and by ribotyping. 16S rDNA RFLP discriminated between S. iniae and other fish pathogens but not between S. iniae strains. HindIII and EcoRI ribotypes of S. iniae discriminated American from Israeli strains rejecting the possibility of an epidemiological link between S. iniae infections in the two countries. Israeli strains isolated from tilapia and trout could not be completely differentiated. The S. iniae ATCC 29178T (T=Type strain) strain, isolated from a freshwater dolphin belonged to a ribotype different from those of all the fish isolates.

Keywords
  • Restriction length polymorphism
  • Streptococcus iniae
  • Epidemiology
  • 16S rDNA
  • Ribotype

1 Introduction

Streptococcus iniae was isolated in 1976 from a diseased freshwater dolphin exhibiting disseminated abscesses of the skin [1] and a few years later from intensively farmed tilapines and trout in Israel [2, 3]. In fish, the disease is primarily a meningoencephalitis accompanied by panophtalmitis without skin manifestations [4]. Losses due to S. iniae infections in Israel ranged between 30% (tilapia) and 50% (trout) of the total predicted crops [2]. Until recently, the economic losses due to streptococcosis in the United States seemed relatively mild. The recent diagnosis of S. iniae in American farmed tilapines and hybrid striped bass [5, 6] indicates that streptococcosis is a real threat to aquaculture, especially when tilapia production is being increased in the United States to respond to the high tilapia demand (47 million pounds of which 32 were imported in 1994).

The taxonomic status of S. iniae was recently clarified and it was shown that S. shiloi which is genetically similar to S. iniae but was described and validated more recently than S. iniae should be considered as a junior synonym of S. iniae[3]. Israeli and American S. iniae isolates from trout and tilapines as well as the type strain are indistinguishable by DNA–DNA hybridization and by phenotypic characterization [3]. The development of fish farming is accompanied by increased movement (export–import) of fish between countries and consequently results in transfer of fish pathogens. It is crucial to have diagnostic tools that allow one to trace the origin of strains isolated in each country in order to establish control measures and import regulations which will protect countries from import of fish pathogens. As many farms in the United States import tilapia fry from Israel, it was necessary to determine if imported fry were a possible source of infection. Isolates of Streptococcus spp. from cattle have been differentiated on the basis of RFLP analysis of the 16S rDNA [7]. Ribotyping of bacterial chromosomal DNA has been widely used to study the molecular epidemiology of various infections [810]. In the present study, a retrospective study of S. iniae strains isolated from infected fish in the United States and in Israel was performed to trace the source of contamination and to prevent the spreading of the condition. Ribotyping and RFLP analysis of the 16S rDNA were employed to investigate the genetic differences within and between (I) S. iniae strains isolated in Israel and in the United States and between (II) S. iniae strains isolated from trout and tilapia.

2 Materials and methods

2.1 Bacterial isolates

For the 16S rDNA RFLP study, 14 strains of Streptococcus iniae from tilapia (Oreochromis nilotica) (12 of the isolates originated from one tilapia farm in Texas and were isolated during various outbreaks in a 2-year period: one from a tilapia farm in Idaho, and one from a hybrid striped bass (Morone saxatilis×M. chrysops) farm in Maryland. The reference strains of S. iniae (ATCC 29178T) (T=type strain), S. shiloi (ATCC 51499T) a junior synonym of S. iniae[3] and S. difficile (ATCC 51487T) were compared. Also compared were the reference strains of S. porcinus (ATCC 43138T), S. uberis (ATCC 19436T), S. parauberis (ATCC 13387T) which are either phylogenetically close to S. iniae or were described as fish pathogen [11, 12], Aeromonas hydrophila (ATCC 7966T), Vibrio parahemolyticus (ATCC 17802T), V. alginolyticus (ATCC 17749T) and Lactococcus garvieae (syn. Enterococcus serolicida ATCC 49156T). For the ribotyping study, nine American S. iniae strains chosen among the 12 Texan strains described above isolated from tilapia during 1993–1994 and nine Israeli S. iniae strains, collected between 1989–1995 from diseased tilapines (three isolates) and trout (six isolates) were included in this study. Each Israeli strains were isolated during distinct outbreaks in time or in location. S. iniae (ATCC 29178T), isolated from a freshwater dolphin in 1976 and S. difficile Dan-31, isolated from a tilapia in Israel were also included in this study. The complete description of the various strains is given in Table 1. All the strains were characterized by biochemical and genetic tests [2, 3], and maintained as stocks in brain–heart infusion broth (BHI, Difco Laboratories, Detroit, MI) with 20% glycerol at −70°C until used.

View this table:
1

Origin of the strains used in the ribotyping study

Strain No.aDate of isolationOrigin and source
1. S. iniae 4B10M3221/9/93Tilapia, brain, Texas
2. S. iniae 4B10M3313/10/93Tilapia, brain, Texas
3. S. iniae 4B10M3313/10/93Tilapia, brain, Texas
4. S. iniae 4B10M3421/10/93Tilapia, brain, Texas
5. S. iniae 4B10M3812/11/93Tilapia, brain, Texas
6. S. iniae 4B10M4115/12/93Tilapia, brain, Texas
7. S. iniae 4B10M4115/12/93Tilapia, brain, Texas
8. S. iniae 4B10M5201/03/94Tilapia, brain, Texas
9. S. iniae 4B10M6026/04/94Tilapia, brain, Texas
10. S. iniae ATCC 29178T1976Dolphin, skin, United States
11. S. iniae Dan 104/89Trout, brain, Dan, Israel
12. S. iniae Dan 302/91Trout, brain, Dan, Israel
13. S. iniae Dan 412/91Trout, brain, Dan, Israel
14. S. iniae Dan 1202/92Trout, brain, Dan, Israel
15. S. difficile Dan 3111/95Tilapia, brain, Dan, Israel
16. S. iniae Dan 3311/95Trout, brain, Daphna, Israel
17. S. iniae Dan 3511/95Trout, brain, Dan, Israel
18. S. iniae ND 5C08/89Tilapia, brain, Afikim, Isreal
19. S. iniae ND 2–1608/89Tilapia, brain, Afikim, Israel
20. S. iniae ND 2405/89Tilapia, E. Hamifraz, brain, Israel
  • aThe serial numbers (1–20) correspond to the well numbers of Figs. 3 and 4.

2.2 Isolation of chromosomal DNA

The methods to lyse the streptococci and to extract DNA have been described previously [2]. Spooled DNAs were suspended in TE (10 mM Tris-HCl; 1 mM EDTA) and stored at 4°C. Alternatively, the bacterial cultures were centrifuged at 7000×g for 10 min and the bacterial pellet was resuspended in digestion buffer (50 mM Tris-HCl; 100 mM NaCl; 20 mM EDTA; 0.05% SDS, at pH 8.5) with 10 mg ml−1 proteinase K and incubated at 37°C overnight. The proteinase K was denatured by heating the suspension to 96°C for 10 min. The DNAs were then purified using Chromaspin tubes (Clontech) and stored at −20°C.

2.3 16S rDNA RFLP

16S ribosomal DNA (rDNA) was amplified by the polymerase chain reaction in a 50 μl reaction. The reaction consisted of PCR buffer, 4 mM MgCl, 4 mM of each dNTP and 0.25 U of Taq DNA polymerase (Promega, Madison, WI). Universal 16S rDNA primers were used (3′-GCGAATTCACGGCTACCTTGTTACCACTT and 5′-GCAAGCTTAGAGTTTGATCCTGGCTCA). Amplification consisted of 35 cycles of denaturation for 30 s at 94°C, annealing for 30 s at 45°C, and extension for 30 s at 65°C. The final cycle included a 5 min extension period at 65°C. The amplified 16S rDNA was digested by RsaI, AvaII, HinfI, and NlaIII restriction endonucleases (New England Biolabs, Beverly, MA). The fragments and a 1 kb molecular mass marker were electrophoresed on a 2% agarose gel stained with ethidium bromide (70 V for 1.5 h). The gels were then examined and photographed under ultraviolet light.

2.4 Ribotyping

Five micrograms of chromosomal DNA was digested with restriction endonucleases EcoRI, HindIII and EcoRV (IBI, New Haven, CT) according to the manufacturer's instructions. The restriction fragments were separated in 0.8% agarose gel and transferred to positively charged nylon membranes (Boehringer-Mannheim), as described by Southern [13]. Lambda DNA digested with HindIII was used as molecular mass marker. The plasmid pKK3535 carrying the E. coli rrnB operon [14](a gift of G. Glaser, Hebrew University Medical School) was digested by EcoRI–HindIII. The resulting 2.5 kb restriction fragment encoding the 16S rRNA and part of the 23S rRNA was purified from low-melting agarose and was used as a probe after labeling by random priming with the DIG DNA labeling kit (Boehringer-Mannheim). Prehybridization, hybridization, washings and chemiluminescence detection procedures (exposure to X-Omat AR 5 films, Kodak, Rochester, NY, for 1–3 h at room temperature) were done as specified by the manufacturer.

3 Results

In RsaI digestions of the 16S rDNA products obtained by PCR amplification, S. iniae and S. shiloi produced patterns identical to S. parauberis whereas S. difficile, S. porcinus and S. uberis each produced unique patterns as seen in Fig. 1. S. difficile had a unique band of approximately 260 bp but lacked a 200 bp band unique to S. uberis. S. uberis is differentiated by the absence of a band common to S. difficile and S. parauberis at 370 bp and the presence of a band at 200 bp. S. porcinus is differentiated by the absence of a 750 bp band common to S. parauberis, S. uberis and S. iniae and the absence of a 200 bp band. S. porcinus also contains a unique band of approximately 900 bp. The S. iniae strains from Texas, Idaho, and Maryland were found to be identical to S. iniae and S. shiloi reference strains but differed from S. difficile reference strain (Figs. 1 and 2). RsaI digestion also differentiated the streptococcal species from L. garvieae, A. hydrophila, E. tarda, V. alginolyticus and V. parahemolyticus (data not shown). Digestion with AvaII (data not shown) demonstrated S. iniae and S. shiloi to be similar and different from S. difficile but could not differentiate between S. difficile and S. porcinus. Digestion with NlaIII did not differentiate between any of the streptococcal species tested or between Streptococcus and L. garvieae but did differentiate Streptococcus from A. hydrophila, E. tarda and the two Vibrio spp. (data not shown). Digestion with HinfI did not differentiate between the streptococcal species tested but did differentiate Streptococcus from A. hydrophila, E. tarda, L. garvieae, V. alginolyticus, and V. parahemolyticus (data not shown).

Figure 1

RsaI restriction endonuclease digestion of PCR amplified 16S rDNA of ATCC Streptococcus spp. followed by electrophoresis on a 2% agarose gel. Lane 2, S. difficile (ATCC 51487T); lane 3, S. porcinus (ATCC 43138T); lane 4, S. uberis (ATCC 19436T); lane 5, S. parauberis (ATCC 13387T); lane 6, S. iniae (ATCC 29178T); lane 7, S. shiloi (syn. S. iniae) (ATCC 51499T); Lanes 1 and 8, 100 bp DNA ladder (Gibco BRL).

Figure 2

RsaI restriction endonuclease digestion of PCR amplified 16S rDNA of 11 different S. iniae strains from the United States. Lane 2 is the ATCC reference strain isolated from a freshwater dolphin. Lane 4 was isolated from a hybrid striped bass from Maryland and lane 5 was isolated from a tilapia from Idaho. Lanes 3, 6, 7, 9–13 were isolated from tilapia from Texas. Lane 14 was the negative control. Lanes 1, 8 and 15 are a 100 bp DNA ladder (Gibco BRL).

The 16S rDNA RFLP could not reveal a pleomorphism among S. iniae strains of various origins, therefore ribotyping was applied to the S. iniae strains studied above and to additional strains (Table 1). A single RFLP pattern (ribotype) was obtained after EcoRV digestion of the DNAs of nineteen S. iniae strains and of the S. iniae strain ATCC 29178T, epidemiologically unrelated to fish disease (data not shown). Two and four different reproducible patterns, which differed by one or more (in the case of the ATCC strain) fragments were obtained after cleavage of genomic DNAs with EcoRI or HindIII, respectively. EcoRI digests (Fig. 3) of DNAs revealed two different ribotypes, allowing the discrimination of the American strains, including the S. iniae ATCC 29178T strain, from the Israeli strains except for the Israeli strain Dan-1 which resembled American strains. American strains and the Israeli strain Dan-1, were found to be identical, presenting five restriction fragments of 5–14 kb (Fig. 3, lanes 1–11), whereas all the other Israeli isolates clustered in a second homogenous ribotype (Fig. 3, lanes 12–14 and 16–20) characterized by six bands. The ribotype of S. difficile (Fig. 3, line 15) was completely different from those of S. iniae. As seen in Fig. 4, HindIII ribotyping revealed a higher degree of polymorphism. All the American tilapia strains and the Israeli strain Dan-1 (Fig. 4, lanes 1–9 and lane 11) presented the same ribotype (ribotype a), different from that of all the other Israeli isolates. Seven out of the nine Israeli strains clustered in a ribotype (ribotype b in Fig. 4, lanes 12–14 and 16–20), distinguishable from the former by a unique 6 kb restriction fragment. The second exception to the homogeneity of the Israeli strains was strain Dan-35 (Fig. 4, lane 17). This isolate was similar to the ribotype b pattern but had an additional band of 5 kb (ribotype c). HindIII ribotyping was found useful in discriminating the S. iniae ATCC 29178T strain (Fig. 4, lane 10) from all other strains. The ATCC strain lacks the high molecular bands present in fish isolates (ribotypes a–c), but possesses a unique 6.5 kb restriction fragment. S. iniae (ATCC 29178T) was classified as ribotype d. The ribotype of S. difficile (Fig. 4, lane 15) was again different from all the ribotypes of S. iniae.

Figure 3

EcoRI ribotyping of American and Israeli S. inaie strains. The numbers of the lanes correspond to the serial numbers of the strains described in Table 1. Lanes 1–9: S. iniae American strains (see Table 1); lane 10: S. iniae ATCC 29178T; lanes 11–14 and 16–20: S. iniae Israeli strains (see Table 1); lane 15: S. difficile strain Dan-31.

Figure 4

HindIII ribotyping of American and Israeli S. inaie strains The numbers of the lanes correspond to the serial numbers of the strains described in Table 1. Lanes 1–9: S. iniae American strains (see Table 1); lane 10: S. iniae ATCC 29178T; lanes 11–14 and 16–20: S. iniae Israeli strains (see Table 1); lane 15: S. difficile strain Dan-31.

4 Discussion

The sudden temporal and geographical emergence of S. iniae infections in fish has been a cause of alarm to Israeli and American fish farmers [3, 5, 6]. Taxonomic and epidemiological studies of S. iniae strains isolated from infected fish have been based on the study of phenotypic characteristics and on the genetic relatedness of S. iniae to other fish Gram-positive cocci pathogens, including streptococci, lactococci, vagococci and enterococci [2, 3, 12, 15, 16]. These methods could not differentiate the strains used in this study; therefore it was necessary to use molecular typing to discriminate between strains of various origins.

The results of RsaI digestion of the 16S rDNA presented in this work support the phylogenetic tree by Bentley et al. [11] who demonstrated 98.4% homology between S. iniae and S. parauberis. The RFLP analysis of the 16S rDNA also supported the findings of Jayarao et al. [7] who were able to differentiate between bovine isolates of S. uberis and S. parauberis. However, RFLP analysis of the 16S rDNA with the endonucleases tested did not allow subtyping of the many S. iniae strains isolated from either Israel or the United States.

Ribotyping was proved useful to subtype several Gram-negative and Gram-positive bacteria [8, 10]. S. iniae Israeli strains except one strain (Dan-1) could be differentiated by ribotyping based on EcoRI digests of DNAs from American strains including the S. iniae ATCC 29178T strain isolated in the United States from a freshwater dolphin in 1970 (Fig. 3). The only Israeli strain similar to the American strains was isolated in 1989 from a trout and not from a tilapia. Since tilapia fry are exported from Israel to the United States on a regular basis, one would expect if S. iniae was contaminating these fry that the ribotypes found in both countries would be similar all the time. The Israeli tilapia strains have a ribotype different from that of the American tilapia strains. The possibility that import of Israeli tilapia fry to the United States concomitantly brought S. iniae is therefore unlikely.

Ribotyping of HindIII digests generated a greater polymorphism than that obtained with EcoRI, enabling a refined study of the epidemiology of S. iniae infections (Fig. 4). All the American fish strains clustered in a single ribotype (ribotype a) indicating an epidemiological relationship among them, whereas the American S. iniae ATCC 29178T strain belonged to a different ribotype (ribotype d). On the other hand, the Israeli strains fell into three distinct groups. Seven out of the nine Israeli strains included in this study belonged to ribotype b, demonstrating the genetic stability of the various strains isolated throughout the period 1989–1995. The first exception to the homogeneity of the Israeli isolates is strain Dan-35 which had the Israeli EcoRI ribotype but the HindIII ribotype c characterized by an additional fragment when compared to the Israeli HindIII ribotype b. The second exception, strain Dan-1, with the American EcoRI ribotype, was found to belong to the HindIII ribotype a, characteristic of the American fish strains. The epidemiological position of this strain is uncertain, but may be a remnant of the first introduction in Israel of S. iniae.

Ribotyping also showed that Israeli isolates from tilapia and from trout do not form two distinct clusters, suggesting that the two species are contaminated by the same strains which is not surprising considering the water distribution system in Israel. When analyzing this finding, one should consider that the number of Israeli trout and tilapia strains used in this study did not reflect their host prevalence. In fact, S. iniae infections in Israel involve mainly trout [3], whereas tilapia are mainly infected with S. difficile[2]. It might be hypothesized that the HindII ribotype a strains are more virulent to tilapines whereas the HindIII ribotype b strains are more virulent to trout. Correlation between ribotype and pathogenicity varies from case to case. For Streptococcus pneumoniae, where capsular typing is an important criteria of pathogenicity, no parallel was found between ribotyping and capsular typing, as isolates sharing the same epidemiological pattern exhibited different capsular types and vice versa [17]. On the other hand, Streptococcus suis ribotyping showed a specific pattern for the virulent strains, although not discriminative for the capsular type of the strain [10]. In the United States mainly tilapia are infected by S. iniae[6] and although the ribotyping data presented here are only on tilapia strains isolated in Texas, the ribotype of a tilapia strain isolated in Idaho was found identical to that of the Texan strains (data not shown). The data on ribotypes of American tilapia strains establish tools to determine the origin of future outbreaks in the United States.

The strains included in this study represent a random sampling of field strains collected from different outbreaks during a 6-year study, starting with early outbreaks in tilapia from each country. The consistent difference in ribotypes between the Israeli and the American strain except for one strain is a unique testimony of the fact that although the etiologic agent is identical, the two epizooties evolved separately. Since most of the Israeli isolates differ from the American ones, the possibility that the disease is actually spreading from one country to the other is unlikely.

Acknowledgements

This work was funded by Grant No. IS 2307-93 from the US-Israeli BARD foundation.

References

View Abstract