Study of Vibrio anguillarum strains from different sources with emphasis on ecological and pathobiological properties

A total of 317 Vibrio anguillarum strains were isolated from water, sediment, and diseased as well as healthy rainbow trout at a Danish mariculture farm and from feral fish caught close to the farm. All strains were examined serologically. Ten sera permitted determination of the O group in 66.7% of the strains from diseased rainbow trout. Furthermore, the O group could be determined in 45.1 to 65.4% of the strains from mucus, gills, and intestinal contents of healthy rainbow trout, while only 22.2 to 28.8% of the isolates from water, sediment, and gills or mucus of feral fish were groupable. Serogroup O1 and to some extent O2 appeared to be associated with trout. Strains from these serogroups were selected for analyses of hemagglutinating activity and surface hydrophobicity. Serogroup O1 comprised hemagglutinating as well as nonhemagglutinating strains; from cases of vibriosis, all O1 strains were nonhemagglutinating. The strains belonging to serogroup O2 were generally hemagglutinating. Examinations of surface hydrophobicity by salt aggregation and hydrophobic interaction chromatography suggested that the O1 strains were more hydrophobic than the O2 strains. In pathogenicity tests, O1 strains isolated from gills and mucus of healthy rainbow trout killed all trout in the test groups. A strain from the intestinal contents of healthy rainbow trout did not produce significant mortality. This strain could, however, be frequently reisolated from the pronephros of fish in the test group concerned. After challenge with strains from eel mucus and seawater, mortality was not produced, and furthermore, these strains could not be reisolated from the pronephros.     

Evidence for the existence of distinct populations of Vibrio anguillarum serogroup O1 based on plasmid contents and ribotypes.

A total of 103 Vibrio anguillarum serogroup O1 strains displaying 15 different plasmid profiles were characterized with respect to biochemical properties and ribotypes. The results confirmed that V. anguillarum O1 is a biochemically homogeneous group. The 103 strains could be allocated to three main clusters with high similarity coefficients. None of the biochemical properties were connected with the presence of plasmids. In total, 12 different ribotypes were demonstrated, with HindIII being used as the restriction enzyme. Forty of the strains were isolated from the same Danish fish farm, some from the kidneys of diseased fish and some from the environment, and some strains were isolated from the mucus, gills, and feces of healthy fish. Nineteen of these isolates possessed the 67-kb virulence plasmid alone or in combination with other plasmids, while 21 had no plasmids. All strains isolated from the kidneys of diseased fish on this farm had plasmids. Irrespective of their origin (kidneys, gills, or mucus), all 19 strains carrying the 67-kb virulence plasmid had the same ribotype, profile 1, while isolates without plasmids belonged to five different profiles, all different from profile 1. These results suggest that pathogenic V. anguillarum O1 strains possessing a virulence plasmid and nonpathogenic strains without plasmids from a small geographical area and even from the same fish may constitute two essentially distinct populations. Thus, it may be suggested that an exchange of virulence plasmids among strains is unlikely to occur in vivo.     

Occurrence of Flavobacterium psychrophilum in fish-farming environments

Occurrence of Flavobacterium psychrophilum in fish farms and fish-farming environments was studied using agar plate cultivation, the immunoflourescence antibody technique (IFAT) and nested PCR. Characteristics of 64 F. psychrophilum isolates from rainbow trout Oncorhynchus mykiss, fish farm rearing water, ovarian fluid and wild fish were serotyped, ribotyped and compared biochemically. Virulence of F. psychrophilum isolates from different sources was compared by injection into rainbow trout. Additionally, the number of F. psychrophilum cells shed by naturally infected rainbow trout was determined. F. psychrophilum was detected and isolated from skin mucus, skin lesions and internal organs of diseased rainbow trout and from fish without clinical disease. The pathogen was also present in wild perch Perca fluviatilis, roach Rutilus rutilus, and ovarian fluids of farmed rainbow trout brood fish. Isolates were biochemically homogenous, excluding the capability to degrade elastin. Five different agglutination patterns with different antisera against F. psychrophilum were found among the isolates studied. Although several different ribopatterns were found (ClaI: 12 ribopatterns and HaeIII: 9 ribopatterns), ribotype A was the most dominant. Farmed rainbow trout brood fish carried a broad-spectrum of serologically and genetically different F. psychrophilum in ovarian fluids. Virulence of the tested isolates in rainbow trout varied and naturally infected rainbow trout shed 10(4) to 10(8) cells fish(-1) h(-1) of F. psychrophilum into the surrounding water.     

Serotyping of Vibrio anguillarum

A serotyping scheme based on the detection of O antigens by slide agglutination in fish-pathogenic strains of Vibrio anguillarum is presented. Over a period of 5 years 270 Vibrio strains from feral and cultured fish, 189 strains from the environment, and 36 strains from invertebrates were collected. The strains were divided into 10 distinct serotypes (O1 through O10). More than 90% of the fish-pathogenic strains, but only 40% of the environmental strains, were typable; 71% of the strains isolated from cultured rainbow trout were serotype O1, whereas 78% of the strains isolated from feral fish were serotype O2. No dominating environmental serotype was found. A serotyping system for V. anguillarum is proposed. A total of 90 strains received from culture collections and laboratories in different countries were typed according to the present system.     

Serotyping of Vibrio anguillarum.

A serotyping scheme based on the detection of O antigens by slide agglutination in fish-pathogenic strains of Vibrio anguillarum is presented. Over a period of 5 years 270 Vibrio strains from feral and cultured fish, 189 strains from the environment, and 36 strains from invertebrates were collected. The strains were divided into 10 distinct serotypes (O1 through O10). More than 90% of the fish-pathogenic strains, but only 40% of the environmental strains, were typable; 71% of the strains isolated from cultured rainbow trout were serotype O1, whereas 78% of the strains isolated from feral fish were serotype O2. No dominating environmental serotype was found. A serotyping system for V. anguillarum is proposed. A total of 90 strains received from culture collections and laboratories in different countries were typed according to the present system.     

Distribution of the flagellate Hexamita salmonis Moore, 1922 and the microsporidian Loma salmonae Putz, Hoffman and Dunbar, 1965 in brown trout, Salmo trutta L., and rainbow trout, Salmo gairdneri Richardson, in the River Itchen (U.K.) and three of its fish farms

The occurrence of Hexamita salmonis Moore, 1922 and Loma salmonae Putz, Hoffman and Dunbar, 1965 was investigated at 10 sites on the R. Itchen (five for brown trout only, three for rainbow trout only, and two for both brown trout and rainbow trout) and at three of its nine fish farms (two for rainbow trout, one for brown trout). Hexamita salmonis was recorded in brown trout from three river sites and the farm, and in rainbow trout from both farms and four river sites. Prevalence of Hexamita salmonis in farmed rainbow trout was higher than in farmed brown trout and was consistent with the former species being more susceptible to infection. H. salmonis was at significantly higher prevalence in rainbow trout from farm no. 5 than farm no. 2 for three size classes of fish. In wild brown trout and feral rainbow trout, the highest prevalences of H. salmonis were recorded at sites in the vicinity of farm no. 2. This distribution was consistent with an area of naturally high infection levels, and with infected fish unintentionally released from farm no. 2 serving as a source of infection, the infection subsequently becoming established in the river fish. Loma salmonae was recorded in wild brown trout and in rainbow trout from both farms. This appears to be the first recording of this parasite from British salmonids and also the first recording of the parasite from brown trout. The distribution of the parasite (particularly the prevalence being higher at farm no. 2 than farm no. 5) was consistent with it being introduced into the R. Itchen via rainbow trout from farm no. 2 (and probably no. 3) much of whose stock derived from imported Californian 'Shasta' rainbow trout.     

Detection of Flavobacterium psychrophilum from fish tissue and water samples by PCR amplification

Rainbow trout fry syndrome and cold-water disease, caused by Flavobacterium psychrophilum, are important diseases in farmed salmonids. Some of the presently available techniques for the detection of Fl. psychrophilum are either time consuming or lack sufficient sensitivity. In the present investigation, the possible detection of Fl. psychrophilum from fish tissue and water samples was examined using nested PCR with DNA probes against a sequence of the 16S rRNA genes. The DNA was extracted using Chelex(R) 100 chelating resin. The primers, which were tested against strains isolated from diseased fish, healthy fish, fish farm environments and reference strains, proved to be specific for Fl. psychrophilum. The obtained detection limit of Fl. psychrophilum seeded into rainbow trout brain tissue was 0.4 cfu in the PCR tube, corresponding to 17 cfu mg-1 brain tissue. The PCR-assay proved to be more sensitive than agar cultivation of tissue samples from the brain of rainbow trout injected with Fl. psychrophilum. In non-sterile fresh water seeded with Fl. psychrophilum the detection limit of the PCR-assay was 1.7 cfu in the PCR tube, corresponding to 110 cfu ml-1 water. The PCR-assay detected Fl. psychrophilum in water samples taken from a rainbow trout farm, but Fl. psychrophilum could not be isolated using inoculation on selective agar. The method presented here has the potential to detect low levels of Fl. psychrophilum in fish tissue and in water samples, and the technique can be a useful tool for understanding the epidemiology of Fl. psychrophilum.     

Surface phenotypic characteristics and virulence of Spanish isolates of Aeromonas salmonicida after passage through fish.

Eleven strains of Aeromonas salmonicida were passaged twice by intraperitoneal injection through rainbow trout and reisolated from the kidney of moribund fish. The surface characteristics and virulence of the strains changed following passage through fish. None of the in vitro tests used could effectively predict the in vivo virulence.     

Relationships between plasmids and phenotypes of presumptive strains of Vibrio anguillarum isolated from different fish species

On the basis of plasmid composition as well as serological and biochemical properties, 26 strains identified as Vibrio anguillarum isolated from diseased fish could be assigned to two different groups. Except for three reference strains, these++ strains were isolated from Norwegian fish. The four strains isolated from rainbow trout (Salmo gairdneri), the only strain isolated from char (Salvelinus alpinus), and three of six strains isolated from Atlantic salmon (Salmo salar) harbored a plasmid of 47 megadaltons (MDa). Restriction endonuclease analysis showed that this plasmid and the virulence plasmid pJM1, carried by V. anguillarum strain 775, were very similar but not identical. Strains harboring the 47-MDa plasmid had nearly identical biochemical properties and were serotype O1. Strains isolated from reared coastal cod (Gadus morhua), turbot (Scophthalmus maximus), halibut (Hippoglossus hippoglossus), free-living saithe (Pollachius virens), and partly from reared Atlantic salmon differed from strains harboring the 47-MDa virulence plasmid by not containing this plasmid, by having different biochemical traits, and by being serotype O2. Rainbow trout which were experimentally infected with a strain isolated from cod suffering from vibriosis developed clinical symptoms similar to those in cod but quite different from those usually seen in rainbow trout.     

Relationships between plasmids and phenotypes of presumptive strains of Vibrio anguillarum isolated from different fish species.

On the basis of plasmid composition as well as serological and biochemical properties, 26 strains identified as Vibrio anguillarum isolated from diseased fish could be assigned to two different groups. Except for three reference strains, these++ strains were isolated from Norwegian fish. The four strains isolated from rainbow trout (Salmo gairdneri), the only strain isolated from char (Salvelinus alpinus), and three of six strains isolated from Atlantic salmon (Salmo salar) harbored a plasmid of 47 megadaltons (MDa). Restriction endonuclease analysis showed that this plasmid and the virulence plasmid pJM1, carried by V. anguillarum strain 775, were very similar but not identical. Strains harboring the 47-MDa plasmid had nearly identical biochemical properties and were serotype O1. Strains isolated from reared coastal cod (Gadus morhua), turbot (Scophthalmus maximus), halibut (Hippoglossus hippoglossus), free-living saithe (Pollachius virens), and partly from reared Atlantic salmon differed from strains harboring the 47-MDa virulence plasmid by not containing this plasmid, by having different biochemical traits, and by being serotype O2. Rainbow trout which were experimentally infected with a strain isolated from cod suffering from vibriosis developed clinical symptoms similar to those in cod but quite different from those usually seen in rainbow trout.     

Morphological organ alterations and infectious diseases in brown trout Salmo trutta and rainbow trout Oncorhynchus mykiss exposed to polluted river water

Poor water quality is discussed as a major factor causing a decline of brown trout populations in Swiss rivers. For our study we have chosen a river in the Swiss midlands, where the brown trout population has decreased dramatically during the last 10 yr and where feral fish have shown distinctive pathological alterations. The objective of our study was to investigate whether river water may be responsible for impaired fish health leading to an increased mortality in the river. In an active monitoring program, groups of brown and rainbow trout were exposed to polluted river water for 24 mo. Fish held in tap water served as a reference. Mortality, macroscopic and histopathologic changes, and infectious agents were investigated. Compared with the reference group, high mortality rates and severe pathological alterations of the inner organs were observed in fish held in river water. Especially gills, liver and kidney of these fish showed significantly higher changes than fish from tap water. These changes were dominated by degenerative and inflammatory reactions. Additionally, several infectious agents were diagnosed in fish exposed to river water. The most important findings were furunculosis and proliferative kidney disease. Brown trout seemed to be more sensitive than rainbow trout to environmental stress and infectious agents.     

Occurrence of plasmids in Danish isolates of Vibrio anguillarum serovars O1 and O2 and association of plasmids with phenotypic characteristics.

Two hundred and twenty-eight isolates of Vibrio anguillarum serovar O1 (125 isolates) and serovar O2 (103 isolates) have been characterized with regard to plasmid contents, biochemical properties, and in vitro hemagglutination and hydrophobic properties. Among 74 V. anguillarum isolates from diseased fish, 63 carried only a 67-kb plasmid (pJM1), 9 carried an additional 98-kb plasmid, and 1 isolate carried only the 98-kb plasmid. Only one isolate was without plasmids. In V. anguillarum serovar O1 from nondiseased fish (mucus and gills), plasmids of the same sizes were present in 29 isolates (58%), whereas 21 isolates (42%) were plasmid free. Based on hemagglutination and biochemical properties, V. anguillarum serovar O1 isolates were divided into eight biovars. The plasmid-carrying strains (102 isolates) all fell within biovars 1 and 2, whereas the 23 strains of biovars 3 to 8 were without plasmids. It was tentatively concluded there are two populations of V. anguillarum serovar O1. One population contains plasmid(s), is hemagglutination negative and trehalose negative, and does not form pellicles in broth cultures, whereas the other population is plasmid free and has the opposite characteristics. The former group is the one related to disease in fish. All 20 V. anguillarum serovar O2 isolates from the environment were without plasmids, whereas 54 (65%) of the isolates from fish (trout and cod) carried plasmids. The biochemical diversity within serovar O2 was pronounced; 13 different biovars were demonstrated. No correlation between the presence of plasmids and biochemical properties was observed.     

Occurrence of plasmids in Danish isolates of Vibrio anguillarum serovars O1 and O2 and association of plasmids with phenotypic characteristics.

Two hundred and twenty-eight isolates of Vibrio anguillarum serovar O1 (125 isolates) and serovar O2 (103 isolates) have been characterized with regard to plasmid contents, biochemical properties, and in vitro hemagglutination and hydrophobic properties. Among 74 V. anguillarum isolates from diseased fish, 63 carried only a 67-kb plasmid (pJM1), 9 carried an additional 98-kb plasmid, and 1 isolate carried only the 98-kb plasmid. Only one isolate was without plasmids. In V. anguillarum serovar O1 from nondiseased fish (mucus and gills), plasmids of the same sizes were present in 29 isolates (58%), whereas 21 isolates (42%) were plasmid free. Based on hemagglutination and biochemical properties, V. anguillarum serovar O1 isolates were divided into eight biovars. The plasmid-carrying strains (102 isolates) all fell within biovars 1 and 2, whereas the 23 strains of biovars 3 to 8 were without plasmids. It was tentatively concluded there are two populations of V. anguillarum serovar O1. One population contains plasmid(s), is hemagglutination negative and trehalose negative, and does not form pellicles in broth cultures, whereas the other population is plasmid free and has the opposite characteristics. The former group is the one related to disease in fish. All 20 V. anguillarum serovar O2 isolates from the environment were without plasmids, whereas 54 (65%) of the isolates from fish (trout and cod) carried plasmids. The biochemical diversity within serovar O2 was pronounced; 13 different biovars were demonstrated. No correlation between the presence of plasmids and biochemical properties was observed.     

Determining Vaccination Frequency in Farmed Rainbow Trout Using Vibrio anguillarum O1 Specific Serum Antibody Measurements

Background

Despite vaccination with a commercial vaccine with a documented protective effect against Vibrio anguillarum O1 disease outbreaks caused by this bacterium have been registered among rainbow trout at Danish fish farms. The present study examined specific serum antibody levels as a valid marker for assessing vaccination status in a fish population. For this purpose a highly sensitive enzyme-linked immunosorbent assay (ELISA) was developed and used to evaluate sera from farmed rainbow trout vaccinated against V. anguillarum O1.

Study Design

Immune sera from rainbow trout immunised with an experimental vaccine based on inactivated V. anguillarum O1 bacterin in Freund’s incomplete adjuvant were used for ELISA optimisation. Subsequently, sera from farmed rainbow trout vaccinated with a commercial vaccine against V. anguillarum were analysed with the ELISA. The measured serum antibody levels were compared with the vaccine status of the fish (vaccinated/unvaccinated) as evaluated through visual examination.

Results

Repeated immunisation with the experimental vaccine lead to increasing levels of specific serum antibodies in the vaccinated rainbow trout. The farmed rainbow trout responded with high antibody levels to a single injection with the commercial vaccine. However, the diversity in responses was more pronounced in the farmed fish. Primary visual examinations for vaccine status in rainbow trout from the commercial farm revealed a large pool of unvaccinated specimens (vaccination failure rate = 20%) among the otherwise vaccinated fish. Through serum analyses using the ELISA in a blinded set-up it was possible to separate samples collected from the farmed rainbow trout into vaccinated and unvaccinated fish.

Conclusions

Much attention has been devoted to development of new and more effective vaccines. Here we present a case from a Danish rainbow trout farm indicating that attention should also be directed to the vaccination procedure in order to secure high vaccination frequencies necessary for optimal protection with a reported effective vaccine.     

THE TOXICITY OF ZINC SULPHATE TO RAINBOW TROUT

The toxicity of zinc sulphate to rainbow trout ( Salmo gairdnerii Richardson) has been investigated in waters of different chemical and physical properties.
Zinc sulphate was less toxic to rainbow trout in hard water than in soft water; when the log concentration of zinc was plotted against log median period of survival of the fish the dose response curve was linear in a very soft water, and curvilinear in a hard water, approaching an apparent threshold concentration. Solutions of zinc sulphate containing calcium chloride were less toxic than those containing an equivalent concentration of calcium as bicarbonate.
An increase in temperature decreased the survival time of rainbow trout in solutions of zinc sulphate in a hard water, but the threshold concentration was not appreciably affected by changes in temperature.
A reduction in the dissolved oxygen concentration of the water increased the toxicity of zinc sulphate, but the effect was reduced when the fish were previously acclimatized to the lower oxygen concentration of the test.
The cause of death of fish in solutions of zinc sulphate was not by the precipitation of mucus on the gills but probably by damage to the gill epithelium.     

Susceptibility of rainbow trout Oncorhynchus mykiss,Atlantic salmon Salmo salar and coho salmon Oncorhynchus kisutch to experimental infection with sea lice Lepeophtheirus salmonis

Physiological, immunological and biochemical parameters of blood and mucus, as well as skin histology, were compared in 3 salmonid species (rainbow trout Oncorhynchus mykiss, Atlantic salmon Salmo salar and coho salmon O. kisutch) following experimental infection with sea lice Lepeophtheirus salmonis. The 3 salmonid species were cohabited in order to standardize initial infection conditions. Lice density was significantly reduced on coho salmon within 7 to 14 d, while lice persisted in higher numbers on rainbow trout and Atlantic salmon. Lice matured more slowly on coho salmon than on the other 2 species, and maturation was slightly slower on rainbow trout than on Atlantic salmon. Head kidney macrophages from infected Atlantic salmon had diminished respiratory burst and phagocytic capacity at 14 and 21 d post-infection (dpi), while infected rainbow trout macrophages had reduced respiratory burst and phagocytic capacities at 21 dpi, compared to controls. The slower development of lice, coupled with delayed suppression of immune parameters, suggests that rainbow trout are slightly more resistant to lice than Atlantic salmon. Infected rainbow trout and Atlantic salmon showed increases in mucus lysozyme activities at 1 dpi, which decreased over the rest of the study. Mucus lysozyme activities of infected rainbow trout, however, remained higher than controls over the entire period. Coho salmon lysozyme activities did not increase in infected fish until 21 dpi. Mucus alkaline phosphatase levels were also higher in infected Atlantic salmon compared to controls at 3 and 21 dpi. Low molecular weight (LMW) proteases increased in infected rainbow trout and Atlantic salmon between 14 and 21 dpi. Histological analysis of the outer epithelium revealed mucus cell hypertrophy in rainbow trout and Atlantic salmon following infection. Plasma cortisol, glucose, electrolyte and protein concentrations and hematocrit all remained within physiological limits for each species, with no differences occurring between infected and control fish. Our results demonstrate that significant differences in mucus biochemistry and numbers of L. salmonis occur between these species.     

Enzymes released from Lepeophtheirus salmonis in response to mucus from different salmonids

Adult and mobile preadult sea lice Lepophtheirus salmonis were incubated with mucus samples from rainbow trout (Oncorhynchus mykiss), coho salmon (O. kisutch), Atlantic salmon (Salmo salar), and winter flounder (Pseudopleuronectes americanus) to determine the response of L. salmonis to fish skin mucus as assessed by the release of proteases and alkaline phosphatase. There was variation in the release of respective enzymes by sea lice in response to different fish. As well, sealice collected from British Columbia responded differently than New Brunswick sea lice to coho salmon mucus. Fish mucus and seawater samples were also analyzed using protease gel zymography to observe changes in the presence of low molecular weight (LMW) proteases after L. salmonis incubation. Significantly higher proportions of sea lice secreted multiple bands of L. salmonis-derived LMW proteases after incubation with rainbow trout or Atlantic salmon mucus in comparison with seawater, coho salmon, or winter flounder mucus. Susceptibility to L. salmonis infections may be related to the stimulation of LMW proteases from L. salmonis by fish mucus. The resistance of coho salmon to L. salmonis infection may be due to agents in their mucus that block the secretion of these LMW proteases or factors may exist in the mucus of susceptible species that stimulate their release.     

Influence of dissolved organic matter on copper binding, and calcium on cadmium binding, by gills of rainbow trout

Complexation of Cu by 5 mg Cl⁻¹ dissolved organic matter (DOM) from a marsh kept Cu from binding to gills of small rainbow trout Oncorhynchus mykiss in 9-day exposures to 0.5 μM Cu in soft water. The protective effect of DOM occurs because the formation of Cu-DOM complexes reduces the amount of free Cu in the water, so the disruptive effects of Cu on ionoregulation, such as inhibited Na uptake, cannot develop. The Cu-DOM complexes themselves do not bind to the gills. Calcium (1100 μm) reduced the accumulation of Cd by trout gills in short, 2-h exposures through competition for gill binding sites but not over longer, 7-day exposures to 0–14 μM Cd. However, the protective effect of Ca against Cd toxicity persisted throughout the longer experiment, likely due to the decrease in the electrochemical gradient for diffusive loss of Ca from the fish to the water. Rainbow trout and fathead minnows Pimephales promelas accumulated Cu and Cd on their gills in a similar manner; thus, binding constants for metal-gill interactions determined for one species of fish can be generalized to other fish species. When literature binding constants determined for fathead minnows were applied to our studies with rainbow trout, computer modelling of Cu-gill and Cu-DOM interactions simulated our results well. In contrast Cd-gill and Ca-gill modelling predicted the initial competitive effect of Ca against Cd accumulation by trout gills, but did not predict the longer-term accumulation of Cd by trout gills.     

Skin mucus protease from rainbow trout, Salmo gairdneri Richardson, and its biological significance

The skin mucus of rainbow trout, salmon, charr, cod, coalfish, plaice and redfish contain, besides lysozyme and haemagglutinins, protease activity. The mucus protease from rainbow trout has been purified, characterized and shown to be indistinguishable from the trypsin produced in pylorus caecae. This enzyme activity can escape detection in mucus since it is inhibited by serum components which may contaminate the skin mucus as a result of stress and handling of the fish prior to collecting the mucus.     

A comparison of cadmium-induced metallothionein gene expression and Me2+ distribution in the tissues of cadmiumsensitive (rainbow trout; Salmo gairdneri) and tolerant (stone loach; Noemacheilus barbatulus) species of freshwater fish

1. The accumulation of cadmium in the liver, kidney and gills of rainbow trout and stone loach was measured during exposure of the fish to the metal at 3 smg/l in their aquarium water. The pattern of accumulation of the toxic metal in the individual organs was different between the two species.2. The tissue concentrations of metallothionein-specific mRNA and metallothionein protein were also determined in these organs from the same fish. In rainbow trout, the induction of metallothionein gene expression resulted in a gradual increase in metallothionein concentration in gill over the course of the experiment whereas increases in metallothionein in the liver and kidney were detected only at the later time points of analysis (beyond 19 weeks). By contrast, in the same tissues from stone loach, relatively minor changes were quantified in specific mRNA and metallothionein concentrations.3. Throughout the experimental period, tissue concentrations of zinc and copper were determined in the liver, kidney and gills of the rainbow trout and stone loach. Subtle decreases were observed in the zinc concentration of gills in rainbow trout and substantial increases were observed in the hepatic copper concentrations in both species at the later time points of analysis.4. The ability of cadmium to induce metallothionein gene expression and its subsequent ability to compete for the sequestration sites on the newly-synthesized protein is discussed with regard to the relative levels of cadmium, zinc and copper in the organs studied and differing regimes of cadmium administration.