John L. Fryer

Piscirickettsia salmonis, the etiologic agent of piscirickettsiosis, is a systemic disease of salmonid fish. Variations in virulence and mortality have been observed during epizootics at different geographical regions and in laboratory experiments with isolates from these different locations. This raises the possibility that biogeographical patterns of genetic variation might be a significant factor with this disease. To assess the genetic variability the 16S ribosomal DNA, the internal transcribed spacer (ITS) and the 23S ribosomal DNA of isolates from 3 different hosts and 3 geographic origins were amplified using the polymerase chain reaction (PCR). Results of this analysis confirm that P. salmonis is a member of the gamma subgroup of the Proteobacteria and show that the isolates form a tight monophyletic cluster with 16S rDNA similarities ranging from 99.7 to 98.5%. The ITS regions were 309 base pairs (bp), did not contain tRNA genes, and varied between isolates (95.2 to 99.7% similarity). Two-thirds of the 23S rRNA gene was sequenced from 5 of the isolates, yielding similarities ranging from 97.9 to 99.8%. Phylogenetic trees were constructed based on the 16S rDNA, ITS and 23S rDNA sequence data and compared. The trees were topologically similar, suggesting that the 3 types of molecules provided similar phylogenetic information. Five of the isolates are closely related (> 99.4% 16S rDNA similarity, 99.1% to 99.7% ITS and 99.3 to 99.8% 23S rDNA similarities). The sequence of one Chilean isolate, EM-90, was unique, with 16S rDNA similarities to the other isolates ranging from 98.5 to 98.9%, the ITS from 95.2 to 96.9% and the 23S rDNA from 97.6 to 98.5%.

Piscirickettsia salmonis was first recognized as the cause of mortality among pen-reared coho salmon Oncorhynchus kisutch in Chile. Since the initial isolation of this intracellular Gram-negative bacterium in 1989, similar organisms have been described from several areas of the world, but the associated outbreaks were not reported to be as serious as those that occurred in Chile. To determine if this was due to differences in virulence among isolates of P. salmonis, we conducted an experiment comparing isolates from Chile, British Columbia, Canada, and Norway (LF-89, ATL-4-91 and NOR-92, respectively). For each of the isolates, 3 replicates of 30 coho salmon were injected intraperitoneally with each of 3 concentrations of the bacterium. Negative control fish were injected with MEM-10. Mortalities were collected daily for 41 d post-injection. Piscirickettsiosis was observed in fish injected with each of the 3 isolates, and for each isolate, cumulative mortality was directly related to the concentration of bacterial cells administered. The LF-89 isolate was the most virulent, with losses reaching 97% in the 3 replicates injected with 10(5.0) TCID50, 91% in the replicates injected with 10(4.0) TCID50, and 57% in the fish injected with 10(3.0) TCID50. The ATL-4-91 isolate caused losses of 92% in the 3 replicates injected with 10(5.0) TCID50, 76% in the fish injected with 10(4.0) TCID50, and 32% in those injected with 10(3.0) TCID50. The NOR-92 isolate was the least virulent, causing 41% mortality in the replicates injected with 10(4.6) TCID50. At 41 d post-injection, 6% of the fish injected with 10(3.6) TCID50 NOR-92 had died. Mortality was only 2% in the fish injected with 10(2.6) TCID50 NOR-92, which was the same as the negative control group. Because the group injected with the highest concentration (10(4.6) TCID50) of NOR-92 was still experiencing mortality at 41 d, it was held for an additional 46 d. At 87 d post-injection, the cumulative mortality in this group had reached 70%. These differences in virulence among the isolates were statistically significant (p < 0.0001), and are important for the management of affected stocks of fish.

Volume 22, Number 8—August 2016 - Emerging Infectious Disease journal - CDC