Molecular Differentiation of Renibacterium salmoninarum Isolates from Worldwide Locations

Abstract

Renibacterium salmoninarum is a genospecies that is an obligate pathogen of salmonid fish and is capable of intracellular survival. Conventional typing systems have failed to differentiate isolates of R. salmoninarum. We used two methods to assess the extent of molecular variation which was present in isolates from different geographic locations. In one analysis we investigated possible polymorphisms in a specific region of the genome, the intergenic spacer (ITS) region between the 16S and 23S rRNA genes. In the other analysis we analyzed differences throughout the genome by using randomly amplified polymorphic DNA (RAPD). We amplified the spacer region of 74 isolates by using PCR and performed a DNA sequence analysis with 14 geographically distinct samples. The results showed that the 16S-23S ribosomal DNA spacer region of R. salmoninarum is highly conserved and suggested that only a single copy of the rRNA operon is present in this slowly growing pathogen. DNA sequencing of the spacer region showed that it was the same length in all 14 isolates examined, and the same nucleotide sequence, sequevar 1, was obtained for 11 of these isolates. Two other sequevars were found. No tRNA genes were found. We found that RAPD analysis allows reproducible differentiation between isolates of R. salmoninarum obtained from different hosts and different geographic regions. By using RAPD analysis it was possible to differentiate between isolates with identical ITS sequences.

Phylogenetically, Renibacterium salmoninarum is a member of the Micrococcus-Arthrobacter subdivision of the actinomycetes, a heterogeneous group of bacteria typified by high G+C contents (5, 22, 27, 38). R. salmoninarum is a slowly growing, fastidious organism with a narrow temperature range for optimal growth (10 to 20°C) and is an obligate pathogen of salmonid fish. This organism is distributed in much of the Northern Hemisphere and Chile and usually causes a chronic, systemic, granulomatous infection, bacterial kidney disease (BKD), which can be fatal under the appropriate conditions (14). The pathogen survives intracellularly and can be transmitted vertically within an ovum, as well as horizontally between cohabiting fish. There is no effective vaccine or chemotherapy. Furthermore, the presence of subclinical infections complicates attempts to control the disease through eradication programs.

The epidemiology of BKD, particularly the interactions which occur between wild and farmed salmonids, is unclear. This is mainly because attempts to differentiate between isolates of R. salmoninarum so far have been unsuccessful. This bacterium appears to possess remarkable biochemical uniformity, and no reliable serological means of distinguishing between isolates has been found (8, 19). A recent study of 40 isolates of R. salmoninarum from North America in which multilocus enzyme electrophoresis was used indicated that the level of genetic diversity was low (39). The lengthy periods required for growth of the bacterium (often 6 weeks or more) and the consequent degradation of antigenic or enzymically active components cause problems for studies which rely on the use of such components.

There are a variety of DNA-based methods available for differentiating between isolates, strains, and species of bacteria. The 16S-23S rRNA intergenic spacer (ITS) has proven to be useful for such differentiation in many cases (6, 17, 25, 28). The ITS appears to have a higher evolutionary rate than either 16S ribosomal DNA (rDNA) or 23S rDNA (28, 30) has, and there are variations in the ITS length and nucleotide sequence which make it possible to distinguish between closely related bacterial species and, sometimes, between strains and isolates (21). Incomplete 16S rRNA gene sequences of two isolates of R. salmoninarum have been determined (22, 29), but there have been no previous studies of either the 23S rRNA gene or the ITS of R. salmoninarum.

An alternative to using species-specific DNA sequences for isolate or strain differentiation involves a PCR-based method, randomly amplified polymorphic DNA (RAPD) analysis. Usually with this method, short random primers are used to rapidly detect genomic polymorphisms under low-stringency conditions (43, 45). RAPD analysis is widely used for differentiating between bacterial isolates (26, 42) and relies upon small quantities of genomic DNA, which makes it ideally suited to the study of slowly growing and fastidious organisms. We investigated the ITS and also performed a RAPD analysis of the R. salmoninarum genome in order to assess the potential of these methods for examining the molecular variability between isolates.

MATERIALS AND METHODS

Bacterial isolates.

Seventy-four isolates of R. salmoninarum were used in this study. Table 1 shows the isolate designations, countries of origin, and sources of isolation and the GenBank accession numbers for the 16S-23S rRNA ITS sequences which were determined. R. salmoninarum was cultured in SKDM broth supplemented with 5% spent culture broth at 15°C (4, 15). The specificity control species (Table 2) were cultured on nutrient agar at 25°C.

TABLE 1.

R. salmoninarum isolates used in this study

Isolate	Geographic origin	Sourcea	Isolate	Geographic origin	Sourcea
970083-88	Southern England	Rainbow trout (f)	F-120-87(P-2)	Iceland	Atlantic salmon (f)
970083-102	Southern England	Rainbow trout (f)	F-130-87(P-4)	Iceland	Rainbow trout (f)
980106 #1.1.5b	Southern England	Rainbow trout (f)	F-138-87(0-78)	Iceland	Atlantic salmon (f)
980036-150	Wales	Rainbow trout (f)	F-260-87(P-16)	Iceland	Atlantic salmon (f)
980036-87	Wales	Rainbow trout (f)	F-273-87(P-19)	Iceland	Atlantic salmon (f)
970419-1.2.3b	Southern England	Atlantic salmon (w)	F-283-87(P-10)	Iceland	Atlantic salmon (f)
970153-19b	Southern England	Grayling(w)	F-358-87(P-13)	Iceland	Atlantic salmon (w)
A6	Southern England	Rainbow trout (f)	S-182-90(P-7)c	Iceland	Atlantic salmon (f)
A80	Southern England	Rainbow trout (f)	Iwatec	Japan	Coho salmon
W2	Northern England	Rainbow trout (f)	Siletz	Oregon	Coho salmon (f)
W6b	Northern England	Rainbow trout (f)	Marion Forksb	Oregon	Chinook salmon (f)
WMV1	Southern England	Rainbow trout (f)	Little Goose	Washington	Chinook salmon (f)
MT409	Scotland	Unknown	CCM6205	Washington	Coho salmon (f)
MT417b	Scotland	Atlantic salmon (f)	84-019-OC	Washington	Chinook salmon (w)
MT239	Scotland	Atlantic salmon (f)	SS-ChS-94-1	Oregon	Chinook salmon
MT426	Scotland	Unknown	Cow ChS94 P22	Washington	Chinook salmon (f)
NCIMB1111	Scotland	Atlantic salmon (w)	Idaho 91-126	Idaho	Sockeye salmon (f)
NCIMB1112	Scotland	Atlantic salmon (w)	RFL-643.94 #1	Washington	Sockeye salmon (f)
NCIMB1113	Scotland	Atlantic salmon (w)	CCM6206	Oregon	Chinook salmon (f)
NCIMB1114	Scotland	Atlantic salmon (w)	Round Butte	Oregon	Chinook salmon (f)
NCIMB1115	Scotland	Atlantic salmon (w)	NCIMB2196	Wyoming	Brook trout (f)
NCIMB1116	Scotland	Atlantic salmon (w)	ATCC 33209b	Oregon	Chinook salmon (f)
MT420	Scotland	Atlantic salmon (f)	D-6	Oregon	Coho salmon
MT452	Scotland	Rainbow trout (f)	Cole River	Oregon	Unknown
MT1363	Scotland	Rainbow trout (f)	Looking Glass	Oregon	Unknown
MT410	Scotland	Unknown	AcF6-1d	Northwest Territories, Canada	Arctic charr (w)
FT-10	Scotland	Atlantic salmon	Rs 9b	Sweden	Rainbow trout
BA99	Scotland	Atlantic salmon	Rs 19	Sweden	Atlantic salmon
DR143b	Alberta, Canada	Brook trout (f)	Rs 61	Sweden	Arctic char
DR384	British Columbia, Canada	Coho salmon (f)	Rs 116	Sweden	Grayling
3784	British Columbia, Canada	Sockeye salmon (f)	Rs 122	Sweden	Rainbow trout (f)
980002	British Columbia, Canada	Chinook salmon (f)	Rs 125	Sweden	Rainbow trout
960023	British Columbia, Canada	Coho salmon (f)	Rs 126	Sweden	Rainbow trout
960046	British Columbia, Canada	Coho salmon (f)	3015-86b	Norway	Atlantic salmon
GC96-1	British Columbia, Canada	Sockeye salmon (w)	4451-86	Norway	Atlantic salmon
DR-128	British Columbia, Canada	Rainbow trout(w)	DK98951714	Denmark	Rainbow trout (f)
RS-TSA	Nova Scotia, Canada	Atlantic salmon (f)	BY1996b	Alaska	Chinook salmon (f)

^aIsolates were obtained from wild fish (w) or from farmed (captive reared) fish (f). The complete histories of some isolates are not known. 

^bSV1 isolates whose 16S-23S rRNA ITS region was sequenced. The sequences have been deposited in the GenBank database under accession no. AF093461, AF093462, AF093465, AF093466, AF093467, AF093468, AF093469, AF093470, AF093471, AF093472, and AF093473. 

^cSV2 isolates whose 16S-23S rRNA ITS region was sequenced. The sequences have been deposited in the GenBank database under accession no. AF093463 and AF093474. 

^dSV3 isolate whose 16S-23S rRNA ITS region was sequenced. The sequence has been deposited in the GenBank database under accession no. AF093464.  

TABLE 2.

Reference organisms used

Species or group	Strain or sourcea
Arthrobacter aurescens	NCIMB 8912
Arthrobacter globiformis	NCIMB 8907
Arthrobacter nicotianae	NCIMB 9458
Bacillus cereus	NCIMB 11925
Brevibacterium linens	NCIMB 8546
Enterobacter aerogenes	UPCC
Enterococcus faecalis	UPCC
Listeria innocua	UPCC
Listeria ivanovi	UPCC
Listeria seeligeri	UPCC
Micrococcus luteus	NCTC 7563
Micrococcus roseus	NCIMB 9149
Rhodococcus equi	UPCC
Staphylococcus aureus	UPCC
Aeromonas hydrophila	UPCC
Aeromonas salmonicida	UPCC
Escherichia coli	UPCC
Escherichia coli B	UPCC
Escherichia coli enteropathogenic	UPCC
Klebsiella edwardii	NCTC 10896
Klebsiella pneumoniae	NCIMB 418
Pseudomonas aeruginosa	ATCC 9027
Pseudomonas fluorescens	NCIMB 1953
Serratia marcescens	NCIMB 11783
Vibrio anguillarum	UPCC
Vibrio parahaemolyticus	UPCC
Yersinia ruckeri	UPCC

^aUPCC, University of Plymouth Culture Collection. 

DNA preparation and amplification of the ITS and specific R. salmoninarum genes.

Genomic DNA was isolated by using a Puregene D-6000 DNA isolation kit according to the instructions of the manufacturer (Gentra Systems Inc.). DNA extracted by this method was electrophoresed on 1.2% agarose gels. Images of each gel were captured with a Kodak model DC40 digital camera, and the DNA concentration was determined for each isolate by using Kodak Digital Science 1D Image Analysis software.

PCR amplification was performed with a DNA thermal cycler (Perkin-Elmer). The primers used to amplify the ITS sequence were selected from region 2 of the 16S rRNA gene sequence for R. salmoninarum and from two highly conserved regions corresponding to regions 5 and 7 of the 23S rRNA gene sequence (21) for Micrococcus luteus obtained from the GenBank database (29, 36). All of the primers (Table 3), including those used for amplification of the R. salmoninarum msa, hly, and rsh genes (9, 16, 20), were chosen by using Amplify software (13). Each 50-μl reaction mixture contained 1 U of Taq polymerase (Boehringer Mannheim), reaction buffer (Boehringer Mannheim), 1.5 mM MgCl₂, 24 pmol of each primer, each deoxynucleoside triphosphate at a concentration of 0.2 mM, and 10 ng of bacterial DNA. The reaction mixtures were overlaid with mineral oil (Sigma), incubated at 96°C for 2 min, and then subjected to 25 cycles consisting of 96°C for 30 s, 65°C for 30 s, and 72°C for 90 s. Amplification products were analyzed on 1.5 and 2% agarose gels.

TABLE 3.

Primers used to amplify the R. salmoninarum msa, hly, and rsh genes and the 16S-23S rDNA spacer region

Gene or region	Size of PCR prod-uct (bp)	Primer
		Designation	Sequence (5′-3′)
msa	487	Rs57+127	TCGCAAGGTGAAGGGAATTCTTCC
		Rs57−611	GGTTTGTCTCCAAAGGAGACTTGC
	297	Rs57+155	CAACAGTACAAGGCTTCAGCAGTG
		Rs57−449	CCGAAACCTACGTTTAGAGTCGTC
hly	542	RsMP+338	ATCGGCTCAGACTAGCGCCATAAT
		RsMP−877	GCTTCAAGATCGATGACCTTCGAG
	336	RsMP+487	TTACTGAGGTCCTTGATGGTCAGG
		RsMP−820	CGATCGGTGCGGTCATTCAAGATA
rsh	572	RsH+231	TCCGGTCATCATGCTTTCTTCGCT
		RsH−800	ATTGCCACCAAGCTGAAGTACCTG
	374	RsH+401	TGCCCAATCTGAAGACAGCGACTA
		RsH−772	GGTCGATAATGCTCGTCATGCCTA
16S-23S spacer	751	RS+1002	CCGTCCAAGTCACGAAAGTTGGTA
		ML−1329	ATCGCAGATTCCCACGTCCTTCTT
	895	RS+1002	CCGTCCAAGTCACGAAAGTTGGTA
		ML−1469	GTGGGTACTGAGATGTTTCAGTTC

RAPD PCR.

The RAPD analysis was performed with 19 isolates, and two separate methods were employed. First, a Ready-To-Go RAPD Analysis Beads kit (Pharmacia Biotech) containing six distinct random 10-mer primers, including primer P1 (GGTGCGGGAA), primer P2 (GTTTCGCTCC), primer P3 (GTAGACCCGT), primer P4 (AAGAGCCCGT), primer P5 (AACGCGCAAC), and primer P6 (CCCGTCAGCA), was used according to the manufacturer’s instructions. Each 25-μl reaction mixture contained 25 pmol of primer and 2.5 or 10 ng of template DNA. The reactions were performed in a Perkin-Elmer thermal cycler by using one cycle consisting of 95°C for 5 min and then 45 cycles consisting of 95°C for 1 min, 36°C for 1 min, and 72°C for 2 min.

Second, the method described by Atienzar et al. (3) was used. Briefly, the following two primers were selected from the 10 primers in a kit obtained from Operon Technologies Inc.: primer OPA9 (GGGTAACGCC) and primer OPB1 (GTTTCGCTCC). Each 25-μl reaction mixture contained 10 mM Tris-HCl (pH 8.3), 50 mM KCl, 5.11 mM MgCl₂, 0.1% Triton X-100, 0.1% gelatin, each deoxynucleoside triphosphate at a concentration of 0.33 mM, 2 μM primer, 2.5 μg of bovine serum albumin, 2.8 U of Taq DNA polymerase (Immunogen International), and 2.5 or 10 ng of template DNA. The reactions were performed in a Perkin-Elmer thermal cycler by using one cycle consisting of 95°C for 5 min, 39 cycles consisting of 95°C for 1 min, 50°C for 1 min, and 74°C for 1 min, and one cycle consisting of 95°C for 1 min, 50°C for 1 min, and 74°C for 10 min. The PCR products were analyzed on 1.2% agarose gels in Tris-borate-EDTA buffer. Images of each gel were captured with a Kodak model DC40 digital camera, and the DNA profile was analyzed by using Kodak Digital Science 1D Image Analysis software.

Sequence analysis.

PCR products spanning the ITS were sequenced directly by a cycle sequencing method and were aligned by workers at MWG-Biotech Ltd., Milton Keynes, United Kingdom. The R. salmoninarum sequences were compared with the sequences of other organisms obtained from the GenBank database by using the gapped BLAST program (1) and the GeneStream align program (IGH, Montpellier, France) (33).

Nucleotide sequence accession numbers.

The GenBank accession numbers for the nucleotide sequences determined in this study are shown in Table 1.

RESULTS

Amplification of specific R. salmoninarum genes.

In order to confirm the identity of the DNA extracted from R. salmoninarum cultures, six sets of primers were designed to amplify known regions of three R. salmoninarum genes. For each of the 74 isolates of R. salmoninarum tested in the six PCR a single band of the appropriate size was amplified (Table 3). No amplification products were obtained from PCR mixtures containing template DNA derived from the specificity control species or from any of the interspersed negative controls.

Amplification of the 16S-23S rDNA spacer region.

The ITS of 74 isolates of R. salmoninarum were amplified by using primers for highly conserved sequences near the 3′ end of the 16S rRNA gene and the 5′ end of the 23S rRNA gene. Primers RS+1002 and ML−1329 amplified a 751-bp fragment, while primers RS+1002 and ML−1469 amplified a 895-bp fragment. In every case only a single band was detected with the primers that were used. In addition, for each primer set no size differences were detected on 1.5 or 2% agarose gels.

Sequencing of the ITS from total PCR products and sequence analysis.

The complete 16S-23S rDNA spacer region sequences of 14 R. salmoninarum isolates were determined by directly sequencing PCR-amplified products. PCR products were amplified with primers RS+1002 and ML−1329, which bind to highly conserved regions 2 and 5 of the 16S-23S rRNA operon (21). Only a single unambiguous sequence was obtained for each PCR product generated. We found that all of the isolates possessed ITS sequences that were the same length, 534 bp. Furthermore, 11 isolates had the same nucleotide sequence, which was designated sequevar 1 (SV1) (Fig. 1). These 11 isolates were obtained from a broad geographic area, which included the mainland United States, Alaska, Canada, Sweden, England, Scotland, and Norway, and from a variety of host salmonid fish species, including chinook salmon, Atlantic salmon, rainbow trout, brook trout, and grayling. Only three isolates possessed spacer regions whose sequences differed from this sequence. The sequences of isolates S-182-90 and Iwate, obtained from Atlantic salmon from Iceland and coho salmon from Japan, respectively, exhibited three identical single-base differences, and the ITS sequence of these organisms was designated sequevar 2 (SV2). Sequevar 3 (SV3), the ITS sequence of isolate AcF6-1 obtained from Arctic char from the Northwest Territories of Canada, also exhibited three single-base differences, one of which was also found in the ITS sequences of S-182-90 and Iwate (Fig. 1). In order to confirm that the ITS sequences obtained by PCR amplification with primers RS+1002 and ML−1329 each represented a single homogeneous copy of the 16S-23S rRNA ITS region, we sequenced the PCR product amplified with primers RS+1002 and ML−1469 from the genome of type strain ATCC 33209. Primer ML−1469 binds deeper in the 23S rRNA gene than primer ML–1329 in highly conserved region 7 (21). The single unambiguous sequence obtained in this way exactly matched the sequences obtained for ATCC 33209 and the 10 other SV1 isolates by using primers RS+1002 and ML−1329.

FIG. 1.

SV1, SV2, and SV3 of the 16S-23S rRNA ITS of R. salmoninarum. The isolates with each sequevar are identified in Table 1. The sequence of the region from nucleotide 1 to nucleotide 750 was determined for 14 isolates by using PCR-amplified products obtained with primers RS+1002 and ML−1329. The sequence for nucleotides 1 to 895 was confirmed for type strain ATCC 33209 by using PCR-amplified products obtained with primers RS+1002 and ML−1469. The uppercase letters represent the 534-bp ITS sequence. The lowercase letters for nucleotides 1 to 145 represent the 3′ end of the R. salmoninarum 16S rRNA gene (22, 29), while the final 216 bp represents the 5′ end of the R. salmoninarum 23S rRNA gene. The three regions that are substantially the same in members of the actinomycetes are underlined.

The R. salmoninarum ITS exhibited 34 to 47% identity with the 16S-23S rDNA spacer region sequences of actinomycetes in the GenBank database. Three regions that were approximately 20, 27, and 35 bp long (Fig. 1) were found to be highly conserved in a number of other members of the actinomycetes, including Bifidobacterium sp., Brevibacterium sp., Kitasatosporia sp., Rhodococcus erythropolis, Streptomyces sp., Microtetraspora sp., and Streptosporangium sp. Sequences for members of the genera Arthrobacter and Micrococcus, two genera which are closely related to R. salmoninarum, were not available in the database and hence were not included in the comparison.

RAPD analysis as a means of differentiating isolates.

We observed that with all of the primers the geographic origins of 19 isolates were reflected in the RAPD band patterns. Using eight random primers and two RAPD methods, we discerned three arbitrary groups of isolates visually (Fig. 2). Group 1 contained isolates from Canada (Fig. 2, lanes a, b, l, and r), Scotland (lanes h, i, m, q, and s), and England (lanes e and f), as well as two isolates from the United States isolates (lanes g and o); group 2 contained isolates from Iceland (lanes j, k, and n); and group 3 contained the other isolates from the United States (lanes c, d, and p). None of the isolates produced identical RAPD patterns with the eight primers, and in most cases, using two or three primers revealed differences between isolates. We chose primers which consistently gave a distinct and reproducible band pattern for each isolate tested. However, primers P2, P3, P4, P5, and P6 gave the clearest and most discriminatory patterns for each isolate regardless of origin. When these primers were used, it was possible to identify differences between isolates from the same country; e.g., primers P2 and P6 discriminated between Icelandic isolates, while primers P2, P3, and P4 revealed differences between English isolates. Differences in RAPD fingerprints could not be attributed to the presence of plasmid DNA. We previously examined DNA extracts of more than 70 R. salmoninarum isolates and found no evidence of plasmid DNA (unpublished data). In order to assess the reproducibility and variation of RAPD fingerprinting, we performed PCR reamplification analyses by using all of the primers, DNA extracted from R. salmoninarum cultures on different occasions, and two DNA template concentrations, as recommended by Welsh et al. (44). Small differences in the quality and concentrations of two templates can lead to spurious differences in the RAPD pattern; therefore, every experiment should include at least two concentrations of genomic DNA for each individual. The results obtained with two primers, primers OPA9 and P1, are presented in Fig. 3. DNA fingerprints were very reproducible; the only discrepancies were confined to the presence or absence of faint bands. The intensities of these faint bands would render them below the limit for inclusion in any analysis of DNA fingerprints.

FIG. 2.

RAPD fingerprints of 19 isolates of R. salmoninarum from a variety of geographic areas and a variety of host species. The DNA fingerprints were obtained by PCR amplification with primers OPA9 (A), OPB1 (B), P1 (C), P2 (D), P3 (E), P4 (F), P5 (G), and P6 (H). Lane a, isolate DR384; lane b, isolate DR128; lane c, isolate ATCC 33209; lane d, isolate Round Butte; lane e, isolate W2; lane f, isolate W6; lane g, isolate Little Goose; lane h, isolate MT1363; lane i, isolate BA99; lane j, isolate S-182-90; lane k, isolate F-273-87; lane l, isolate RS-TSA; lane m, isolate MT410; lane n, isolate F-138-87; lane o, isolate NCIMB2196; lane p, isolate Marion Forks; lane q, isolate MT417; lane r, isolate DR143; lane s, isolate MT420; lane −, water control. Lanes M contained markers (1-kb DNA ladder [Gibco BRL] in panels A and B and 100-bp DNA ladder [Gibco BRL] in panels C to H). The molecular sizes (in kilobases) are indicated on the left.

FIG. 3.

Reproducibility of RAPD fingerprinting. The DNA fingerprints for DNA extracted on separate occasions with two different concentrations of template were obtained after PCR amplification with primers as described in the Legend to Fig. 2. For the contents of lanes a to s see the legend to Fig. 2. Lanes 1, 2.5 ng of DNA template; lanes 2, 10 ng of DNA template. Only the results obtained with the following two primers are shown: primer OPA9 (A and B) and primer P1 (C and D). Lane M contained markers (see the legend to Fig. 2). The molecular sizes (in kilobases) are indicated on the left.

DISCUSSION

The widespread distribution of R. salmoninarum in the United Kingdom, many European countries, Japan, North America, and Chile and the variety of salmonid host species in these regions suggested the possibility that the genetic diversity of isolates in these areas may be reflected by the number, length, and sequence of the 16S-23S rRNA ITS region. While inter- and intrageneric relationships may be elucidated by examining 16S and 23S rDNA sequences, the ITS has provided information on intraspecific relationships in other bacteria (17, 28, 30, 40). Three distinct ITS sequences (sequevars) were obtained from 14 R. salmoninarum isolates. Isolates from Iceland, Japan, and the Canadian Northwest Territories which had three single-base substitutions in the ITS exhibited some divergence from the highly conserved SV1 which was present in isolates from the United States, the United Kingdom, mainland Europe, and Alberta, Canada. It may be that in areas of the world which could be regarded as relatively isolated from the mainstream intensive salmonid culture areas of North America and Europe the bacterium has diverged from this pattern. It is interesting that the sole Alaskan isolate was an SV1 isolate. BKD has been reported in wild and farmed fish from a number of Alaskan river systems (11, 32), and it seems likely that Alaskan salmon have been exposed to the sequevar of R. salmoninarum carried by salmon from the Pacific coast of Canada or the United States at some stage during their oceanic migrations.

This study provides no evidence that there are multiple copies of the rRNA operon in R. salmoninarum. A single unambiguous nucleotide sequence was obtained for all of the isolates examined, and 11 of the isolates possessed spacer regions that had the same nucleotide sequence. The presence of a nucleotide sequence generated from a highly conserved region deeper in the 23S rRNA gene confirmed these results. Typical tRNA genes were not found in the ITS region of R. salmoninarum. Furthermore, we found no evidence that there were multiple amplicons in PCR mixtures when we used two sets of primers for highly conserved regions of the 16S and 23S rRNA genes. However, absolute proof that there is a single rRNA operon would require direct sequencing from the genome. We concluded that R. salmoninarum probably has a single copy of the rRNA operon, a finding which is consistent with what has been described for a number of other slowly growing organisms (2, 7, 18, 37, 40) and is a further indication of the conservative genetic composition of this obligate pathogen. Generally, our findings suggest that the 16S-23S rDNA spacer region is of limited use for routine discrimination between R. salmoninarum isolates but may offer some clues as to geographic origins.

The lack of a way to differentiate between isolates of R. salmoninarum has constrained epidemiological studies of BKD. In particular, development of a means of contact tracing would allow BKD outbreaks to be traced back to the source of infection and would help resolve some of the difficulties associated with investigation of the interactions between farmed and wild salmonid fish. We used two methods to do this, examination of ITS variation and RAPD analysis, which have been used successfully in studies of other bacteria. Our work shows that compared with ITS variation, RAPD analysis is a better method for discriminating between isolates of R. salmoninarum. In our study, R. salmoninarum isolates from a variety of sources, some with identical 16S-23S spacer region DNA sequences, could be distinguished on the basis of RAPD patterns generated by two different methods. RAPD analysis has provided a reliable and reproducible method for molecular typing and genetic characterization of a variety of microorganisms (23, 24, 34, 41). This method is particularly useful for examining the genomic diversity among strains of bacteria which are indistinguishable by other molecular methods. For example, RAPD analysis of strains of Bacillus cereus revealed a remarkable diversity which was not revealed by rRNA or tRNA ITS-targeted PCR (10). A number of factors have been identified as influencing the outcome of RAPD fingerprinting (12, 31, 35). In our studies, using eight primers and two different methods for PCR amplification of purified DNA template produced RAPD fingerprints which were reproducible with two different DNA concentrations and with DNA extracted on different occasions. In every case, RAPD fingerprints distinguished the same groups of isolates.

So far, R. salmoninarum has defied attempts to find a reproducible way to differentiate between isolates. This study is the first study which revealed the genetic diversity within the species by using a DNA-based method for differentiating between isolates from a wide variety of sources and therefore represents a substantial advance in our understanding of a fastidious intracellular pathogen which is capable of surviving within its host in very low numbers. We are extending our investigations of R. salmoninarum by using RAPD analysis in conjunction with other molecular typing methods as part of a coordinated program to examine farm and wild R. salmoninarum isolates from the United Kingdom and other sources. This work should result in a wide-ranging analysis of isolate differences.

In conclusion, R. salmoninarum is a highly conserved genospecies. The molecular variation in the sequence of the 16S-23S rDNA spacer region of isolates from widely separated environments is extremely limited. RAPD analysis is a reliable and reproducible technique for discriminating between isolates of R. salmoninarum and should facilitate epidemiological studies of this pathogen.