Introduction

From a disease perspective, the detrimental effects of invasive species are typically expressed in terms of pathogen pollution, defined as the movement of pathogens to areas outside their native geographic or host ranges (Cunningham et al. 2003). Typically, this occurs as accidental co-introductions of covert pathogens when mammalian, avian, and amphibian hosts are introduced to novel areas (Rupprecht et al. 1995; Laurenson et al. 1998; Hochachka and Dhondt 2000; Hyatt et al. 1997; Daszak et al. 2003; Wyatt et al. 2008). Pathogen pollution also exists in aquatic and marine systems, where the inadvertent transport and introduction of pathogens occur with the live/frozen bait (Arkush et al. 2006; Picco and Collins 2008) and ornamental fish industries (Stewart 1991). For example, a leading hypothesis accounting for massive epizootics among pilchards (Sardinops sagax) in South Australia during the mid-1990’s involved the inadvertent introduction of a herpesvirus via frozen feed used for the captive tuna industry (Griffin et al. 1997; Hyatt et al. 1997; Jones et al. 1997; Gaughan et al. 2000; Ward et al. 2001); ecological effects of the epizootics occurred throughout different trophic levels of the food web (Bruce and Norman 2000; Dann et al. 2000). Pathogen pollution can also occur in aquatic and marine systems through the inadvertent uptake, transport, and release of pathogens in the ballast water from vessels (Ruiz et al. 2000). Detrimental effects of pathogen pollution to native species are realized through pathogen-mediated competition, whereby long-standing relationships between the introduced host and its pathogen confer levels of innate and adaptive resistance, and therefore a competitive advantage, over native hosts that are often highly susceptible to the introduced pathogen (Hudson and Greenman 1998). Additionally, deliberate introductions of exotic species can result in the “Frankenstein Effect,” referring to a myriad of other unexpected consequences that result from well-intended introductions (Moyle et al. 1986).

A less-studied detrimental effect involves the amplification of endemic pathogens by introduced species, followed by “spillback” of the amplified pathogen to native species (Kelly et al. 2009). This phenomenon is difficult to demonstrate using traditional disease surveillance techniques, but may have serious ecological impacts, especially if the introduced population has attained large size or if it transfers the pathogen into new habitats.

American Shad (Alosa sapidissima), an anadromous clupeid fish that is endemic to rivers and nearshore regions along the Atlantic coast of North America, were initially introduced to the Pacific coast of North America in 1871, when fry were transported from the Hudson River, NY to the Sacramento River, CA (Petersen et al. 2003). The range of introduced shad soon expanded northward, and spawning adult shad were first documented in the Columbia River in 1885 (Smith 1886). Biomass of American shad in the Columbia River system remained low until the mid—late 1980’s, when the numbers of returning adults increased rapidly and surpassed the combined numbers of all native adult Pacific salmonids including coho salmon (Oncorhynchus kisutch), Chinook salmon (O. tschawytscha), sockeye salmon (O. nerka), and steelhead (O. mykiss) (Columbia River Data Access in Real Time; http://www.cbr.washington.edu/dart/). The typical shad life history pattern involves a riverine egg and larval period, rapid out-migration of young-of-the-year (YOY) juveniles to seawater, seawater rearing for 3–5 years, and return migration to natal rivers for iteroparous spawning. However an atypical juvenile life history pattern sometimes occurs in the Columbia River, whereby juveniles remain in the river for one to two years prior to their seawater outmigration. The successful colonization of Columbia River by American shad likely resulted from low fishing pressures, successful passage of adults and juveniles beyond the hydroelectric dams, adequate adult spawning habitat, and beneficial conditions for survival of early life history stages (Petersen et al. 2003). Ecological impacts of this introduction and colonization have received little attention.

Ichthyophonus spp. are Mesomycetozoean parasites, primarily of marine fishes, that have caused recurrent epizootics in wild fish populations throughout the North Atlantic (reviewed in McVicar 1999). Incomplete species descriptions fail to provide sufficient detail to allow unequivocal identification of the parasite to the species level (Alderman 1982); consequently, it will hereafter be referred to generically. Since its first description in the NE Pacific during the early 1980’s (Olson 1986), Ichthyophonus has become recognized as ubiquitous in many marine fish populations throughout the region (Marty et al. 1998; Kent et al. 2001; Hershberger et al. 2002; Jones and Dawe 2002) where it has been associated with population-level effects (Marty et al. 1998; Hershberger et al. 2002; Kocan et al. 2004). Causes for its apparent recent emergence in the NE Pacific (Kent et al. 2001) remain undetermined.

During the course of several pilot-scale fish health surveys in the Pacific Northwest, we detected Ichthyophonus in a few samples of wild American shad. To expand upon these preliminary observations and to investigate the epizootiology of the parasite, we performed targeted Ichthyophonus surveys in American shad and sympatric fishes throughout the Columbia River system in 2007. Further, the possible introduction of Ichthyophonus to the Pacific Ocean by American shad was investigated by comparing the genetic relatedness of Ichthyophonus isolates from shad in the Atlantic and Pacific Oceans.

Methods

Epizootiology of Ichthyophonus was investigated by surveying American shad and sympatric fishes in the Columbia River for prevalence and severity of infection. To determine the life history stage when shad become infected, Ichthyophonus prevalence was compared in outmigrating young-of-the-year (YOY) juveniles (n = 120), atypical freshwater juvenile holdovers (n = 62), typical marine phase juveniles from Puget Sound, WA (n = 16), returning pre-spawn adults (n = 201), and post-spawn adults (n = 145) collected at the Bonneville Dam fish passage facility, river kilometer (RKM) 235. Post-spawn adults consisted of a combination of live, dead, and moribund individuals; no attempt was made to compare the infection prevalence between the three post-spawn stages because nearly all fish demonstrated descaling and other signs of distress. To determine whether infection prevalence changed during the adult freshwater migration, the infection prevalence in pre-spawn adults at RKM 235 was compared to that of adults at RM 308 (n = 60). To assess interannual variabilities in infection prevalence, adult shad were sampled at RKM 235 during 2007 (n = 201), 2008 (n = 57), and 2009 (n = 60). The prevalence of Ichthyphonus in sympatric resident and anadromous species utilizing the Columbia River was assessed by sampling smallmouth bass Micropterus dolomieu (n = 14), northern pikeminnow Ptychocheilus oregonensis (n = 73), white sturgeon Acipenser transmontanus (n = 12), adult spring Chinook salmon Oncorhynchus tshawytscha (n = 90), and adult fall Chinook salmon (n = 101) during 2007. Sampled fish were collected by a variety of methods including fish trap, hook-and-line, dip net, beach seine and trawl (Table 1). Infection prevalences were compared using the Chi Square statistic (χ2) and statistical significance was assigned to comparisons where P ≤ 0.05.

Table 1 Sampling details and Ichthyophonus prevalence
Full size table

Prevalence and severity of Ichthyophonus infections were determined by in vitro culture and histological assessment of heart tissue, respectively. Hearts were aseptically removed from sampled fish and approximately 1 g of heart tissue was cultured in 5 mL of tris-buffered Eagle’s Minimum Essential Medium supplemented with 5% fetal bovine serum, 100 IU mL−1 penicillin, 100 μg mL−1 streptomycin and 100 μg mL−1 gentamycin; the remainder of the heart from each fish was fixed in 10% neutral-buffered formalin for histological assessment. Cultures were incubated at 15°C for 14 day, after which prevalence of the pathogen was determined by microscopic examination (40× total magnification) for the presence of Ichthyophonus pseudo-hyphae and/or schizont stages. To quantify infection severity, fixed heart tissues from a subsample (n = 101) of culture-positive samples were dehydrated, and embedded in paraffin; thin tissue slices were mounted on glass slides and Ichthyophonus was selectively stained with periodic acid-Schiff. Infection severity was scored as an index of parasite load within a randomly selected 100× field of view (Kocan et al. 2006) and was defined as “Light” (culture-positive, but no organisms observed histologically), “Moderate” (1–2 organisms observed histologically), or “Heavy” (3 + organisms observed histologically).

To assess the temporal emergence of Ichthyophonus in the Columbia River and the involvement of American shad, a biomass index of infected shad entering the Columbia River was calculated over 20 year intervals. The biomass index was calculated as the product of:

$$ \left( {\text{Annual adult shad passage}} \right) \, \times \, \left( {\text{Mean weight of adult shad}} \right) \, \times \, \left( {\text{Annual infection prevalence}} \right). $$

Annual adult shad passage data were obtained from counts at Bonneville Dam RKM 235 (Columbia River DART; Fig. 1). The mean weight of adult shad (860 g) was determined from a subsample of pre-spawn adults (n = 146) collected from RKM 235 during the 2007 Ichthyophonus surveys. Annual infection prevalence for 2007 (72%) was based on the results from these surveys. Historical infection prevalence was unavailable; therefore, biomass index values prior to 2007 were calculated using a range of presumed historical infection prevalences from 10 to 100%.

Fig. 1
figure 1

Adult anadromous fish counts at the Bonneville Dam fish ladder (RKM 235). Data from Columbia River Data Access in Real Time (DART; http://www.cbr.washington.edu/cgi-bin/dart/dart?report=adult_annual_rpt&format=standard&year=2001&access_time=Monday%2C+10-Aug-2009+16%3A37%3A58+PDT&proj=BON&run=+&startdate=1%2F1&enddate=12%2F31)

Full size image

Concurrent infections of American shad in the Columbia River with other locally important fish pathogens were assessed by performing standard virological and bacteriological diagnostics on a sub-sample of adults collected during 2007–2008 (n = 117). For virology, kidney and spleen samples were aseptically removed, and diluted tissue homogenates were plated onto monolayers of epithelioma papulosum cyprini, Chinook salmon embryo, and fathead minnow cell lines, following standard isolation techniques for Infectious hematopoietic necrosis virus, Infectious pancreatic necrosis virus, and Spring viremia of carp virus (AFS–FHS 2007). For bacteriology, kidney samples were inoculated onto brain-heart infusion agar and tryptone-yeast extract, following standard isolation techniques for Aeromonas salmonicida, Yersinia ruckeri, Edwardsiella ictaulri, Flavobacterium psychrophilum, and Flavobacterium columnare (AFS–FHS 2007). Additional kidney samples were assayed by enzyme-linked immunosorbent assay (ELISA) for the presence of Renibacterium salmoninarum antigen; suspect samples were confirmed by the polymerase-chain reaction (PCR), using R. salmoninarum-specific primers (AFS–FHS 2007).

Genetic relatedness of Ichthyophonus isolates from American shad on the Atlantic and Pacific coasts of North America was assessed by comparing nucleotide sequences in the 18S small subunit and internal transcribed spacer (ITS1, 5.8 s and ITS2) regions of the ribosomal DNA (rDNA) gene complex; Amoebidium parasiticum (AY388646) was used as a Mesomycetozoean outgroup. Atlantic coast Ichthyophonus isolates (n = 3) were obtained from pre-spawn adult American shad collected from the Merrimack River system, Massachusetts and were designated IA65, IA66 and IA67 (Table 2). Isolation of Ichthyophonus in culture was performed from tissue explants cultures, as described above. Techniques for nucleic acid extraction and amplification were similar to those reported previously (Rasmussen et al. 2010); however more specific primers were developed and utilized for these comparisons. Previously published primers embedded in the 3′ end of the 18S and the 5′ end of the 28S genes and amplified the entire ITS1, 5.8 s and ITS2 region; however, these primers were problematic for IA65-67 because they amplified a yeast contaminant present in the Ichthyophonus cultures. To address this problem, we designed and utilized new Ichthyophonus-specific primers located at the 5′ end of ITS1 (Out-ITS1-F: GCGGAAGGATCATTACCAAATAACG) and the 3′ end of ITS2 (Out-ITS2-R: GCCTGAGTTGAGGTCAAATTT). Techniques for assessing sequence variation within isolates, between isolates from the same region, and between isolates from different regions were identical to those reported in Rasmussen et al (2010).

Table 2 Ichthyophonus isolates from American shad including GenBank numbers and characteristics of the partial ITS rDNA gene region used for phylogenetic analysis
Full size table

Results

An Ichthyophonus epizootic occurred among American shad in the Columbia River during 2007, characterized by 72% (145/201) infection prevalence in pre-spawn adults at RKM 235 (Table 1). Among a subsample of positive fish (n = 101), heavy intensities occurred in 27%, moderate intensities occurred in 30%, and light intensities occurred in 44%. Spawning status had no effect on infection prevalence, with Ichthyophonus prevalence in pre-spawn adults (72%; 145/201) similar to that of post-spawn cohorts (70%; 102/145). Ichthyophonus prevalence was slightly lower (58%; 35/60) among pre-spawn shad collected further upriver (RKM 493), but the decrease was not statistically significant (P = 0.057). Infection prevalence was similar (P = 0.467) among a subsample of males and females (66%, n = 117 and 71%, n = 111; respectively). The epizootic waned in successive years, with infection prevalence declining significantly to 49% (28/57) in 2008 (P = 0.002) and 37% (22/60) in 2009. No viral or bacterial pathogens were confirmed in samples of 117 adult shad collected from the Columbia River during 2008–2009.

The emergence of Ichthyophonus into the Columbia River system and the involvement of adult American shad were indicated by the biomass index of infected shad, which increased from 0 mt in 1947–1,595 mt in 2007 (Fig. 2). Although historical prevalences of Ichthyophonus in shad prior to 2007 are unknown, a temporal increase in the amount of infected shad tissues entering the Columbia River is evident, even when assumed historical infection prevalence was 100%.

Fig. 2
figure 2

Emergence of Ichthyophonus in the Columbia River, indicated by the temporal increase in Biomass Index of infected shad tissues at the Bonneville Dam fish ladder (RKM 235). Because Ichthyophonus prevalence values prior to 2007 are unknown, the historical biomass index (prior to 2007) was calculated based on a range (10–100%) of presumed prevalences (indicated in the legend)

Full size image

Ichthyophonus prevalences in shad early life history stages and in other fishes indicated that the infections were likely established in the marine environment and were not vertically transferred from infected parents to progeny. Ichthyophonus was not detected in any (n = 120) outmigrating YOY shad or in any atypical, age 1+ year juvenile holdover shad (n = 62) that failed to out-migrate. However, Ichthyophonus was detected in 33–60% (2/6 and 6/10) of seawater-phase juvenile shad collected from Puget Sound. Among other fishes from the Columbia River, Ichthyophonus was not detected in smallmouth bass (n = 14), northern pikeminnow (n = 73), adult white sturgeon (n = 12), or adult fall Chinook salmon (n = 101); however, it was detected in 4% of adult spring Chinook salmon (Table 1).

Nucleotide sequences in the Ichthyophonus highly conserved SSU gene region differed between American shad from the east and west coasts of North America. Among Ichthyophonus isolates from American shad, nucleotide sequences in both the A and B regions of the 18S rDNA gene are identical throughout the Pacific region (Rasmussen et al. 2010); however, those from the Atlantic coast (IA65-IA67) possessed a single nucleotide substitution in the B region (GQ370802-GQ370804), with no differences occurring in the A region (A: GQ370781-GQ370783).

Further distinction between Ichthyophonus isolates from shad on the east and west coasts of North America was indicated by sequence differences in the ITS regions. The Ichthyophonus-specific outer ITS PCR primers amplified an approximate 640 bp band from the Atlantic Coast isolates (IA65–IA67). Sequencing of 6–8 clones for each isolate showed that intra-isolate variation was an average of 1.2 nucleotides across the amplified region (Table 2). This intra-isolate variation was lower than the level reported by Rasmussen et al. (2010) (e.g. IA6, IA52, IA14; Table 2). Cloning and sequencing of the IA67 isolate with the Rasmussen et al. (2010) general fungal primers (GenBank #: GQ402900-GQ402906) and the Ichthyophonus-specific primers reported here, indicated that the lower intra-isolate variation observed in the present study was due to the primer change; these differences did not significantly change the majority consensus sequence for IA67 or the overall results (data not shown). The three Atlantic coast Ichthyophonus isolates had no nucleotide substitutions relative to each other in their majority consensus sequences. The majority consensus sequences from the Atlantic and Pacific coast Ichthyophonus isolates differed by 4 of 631 aligned nucleotides (Kimura 2-paramter distance = 0.006). The Atlantic and Pacific coast Ichthyophonus isolates differed from the Idaho rainbow trout by 16 and 18 nucleotides, respectively. Phylogenetic analysis using the majority consensus sequence for each isolate grouped the Atlantic coast and Pacific coast shad isolates separately from each other with high bootstrap support (Fig. 3). Additionally, there were no nucleotide substitutions observed between the ITS sequences of Ichthyophonus isolates from the Pacific coast American shad and the endemic Pacific herring.

Fig. 3
figure 3

Relationship of Ichthyophonus isolates based on the majority consensus sequence of the ITS region (partial ITS1, 5.8 s and partial ITS2) inferred using the neighbor-joining method (distances calculated using Kimura 2 parameter method). Tree topologies were validated by bootstrapping 2000 times. Similar tree topology was obtained using the maximum parsimony method. Isolates are derived from American shad from Merrimack River, MA (IA65, IA66, IA67; see Table 2) and compared to sequences previously reported by Rasmussen et al. (2010)

Full size image

Discussion

The introduction and subsequent population expansion of American shad to the Pacific coast of North America ultimately resulted in the amplification of Ichthyophonus, and anadromous shad life history characteristics resulted in the mass transport of the marine pathogen into the Columbia River. After the translocation of American shad to the Pacific coast of North America during the late 1800’s (Reviewed in Petersen et al. 2003), population growth was characterized by an extended lag phase (Kowarik 1995; Sakai et al. 2001) that lasted for nearly 100 years, logarithmic growth in the late 1980’s, and peak in 2004 (Fig. 1). Although historical prevalence of Ichthyophonus in shad is unknown, recent increases in the amount of the parasite entering the Columbia River are well-supported by the recent shad population expansion. For example, even if the historical prevalence of Ichthyophonus in shad was assumed to be 100%, the historical abundance of shad was low; therefore, the historical amount of infected shad entering the system was low (Fig. 2). However, recently high infection prevalence (72% in 2007) combined with the record-high shad abundance (Fig. 1) indicate a massive influx of infected tissues into the Columbia River in recent years (Fig. 2).

Although Ichthyophonus infects American shad along the east coast of North America (reported here), its recent emergence in marine fishes on the west coast of North America (Kent et al. 2001; Hershberger et al. 2002; Kocan et al. 2004) was not likely a result of pathogen pollution via the translocation of shad from the Atlantic. Ichthyophonus phylogenetics, based on 18S and ITS sequences of the ribosomal gene complex, indicate that genetically distinct parasite types occur in American shad from the Atlantic and Pacific (Fig. 3). The single nucleotide substitution occurring in the B region of the highly conserved 18S region is significant because all Pacific coast Ichthyophonus isolates ranging from Alaska to WA from a diverse range of hosts are 100% identical in 18S sequence (Halos et al. 2005; Rasmussen et al. 2010); although a less common 18S rDNA haplotype has been observed in Pacific coast rockfishes from Oregon and British Columbia (Criscione et al. 2002). Worldwide only 5 18S rDNA haplotypes have been reported (Criscione et al. 2002) and none of these haplotypes matched the variant observed in Atlantic coast American shad isolates reported here. The isolates from the Atlantic demonstrated additional genetic separation in their consensus ITS sequences. Although it is possible that the ancestral Atlantic haplotype was transported by shad to the Pacific where the genetic divergence later occurred, epidemiological characteristics also refute the co-introduction hypothesis. For example, only freshwater life history stages of American shad, including eggs, larvae, and juveniles, were translocated to rivers along the Pacific coast of North America (Haskell et al. 2006). There is no indication that the parasite is vertically transferred from the parents to the progeny, and all available data indicate that natural exposures occur and infections are established in the marine environment (reviewed in McVicar 1999). Therefore, freshwater early life stages of shad that were translocated from the Atlantic were likely never exposed to the parasite, nor were they likely infected when moved.

The emergence of Ichthyophonus in the Columbia River system was driven by amplification and transport of a Pacific pathogen by an introduced host. After their introduction to rivers along the Pacific coast of North America, anadromous American shad outmigrated to marine and coastal feeding areas, where exposure to endemic Ichthyophonus occurred. During the early years after the introduction of shad, when biomass numbers remained low (Fig. 1), establishment of the parasite in the introduced host likely had negligible impact on the ecology of the disease in the Pacific. However, the rapid expansion of the American shad population in the eastern Pacific during the 1980’s dramatically changed the ecology of the disease by providing thousands of tons of host substrate suitable for disease amplification. The amplified pathogen was then transported outside its endemic marine range and into the Columbia River by adult shad that returned to their natal river to spawn. An analogous example of disease amplification occurred after the introduction of brushtail possum (Trichosurus vulpecula) to New Zealand, which amplified bovine tuberculosis and represented a catalyst for transmission of the disease to farmed cattle and deer (reviewed in Cunningham 1996). Similar disease amplification often occurs in intensive aquaculture (Hershberger et al. 1999) and agriculture operations (Lafferty and Gerber 2002), where pathogen ‘spillover’ can negatively impact sympatric wild populations (Power and Mitchell 2004; Krkošek et al. 2007). The extent of Ichthyophonus “spillback” (Kelly et al. 2009) from shad to other fishes in the NE Pacific, and the involvement of spillback with the emergence of Ichthyophonus in other NE Pacific fishes were outside the scope of this study. However, the results clearly indicate a strong role for American shad in transporting the amplified pathogen outside its endemic coastal and marine range and into a freshwater system that was largely Ichthyophonus-free. Regardless of the recent influx of infected shad biomass into the Columbia River, paucity of the parasite in sympatric resident and anadromous fishes (Table 1) indicates that an efficient freshwater parasite life cycle has not yet established in the system. However, a high potential exists for the future establishment of a freshwater life history based on elevated riverine infection pressures (Fig. 2), low parasite host specificity (reviewed in McVicar 1999), plasticity of parasite life stages and life history strategies (Okamoto et al. 1985; Kocan et al. in press), and precedent for Ichthyophonus establishment in other freshwater systems (Rucker and Gustafson 1953; Erickson 1965; Hershberger et al. 2008; Rasmussen et al. 2010). Therefore, it is recommended that further expansion of Ichthyophonus throughout the Columbia River system be assessed through continued surveillance efforts in other resident and anadromous fishes.

Substantial crashes of introduced species often occur for unknown reasons after populations are firmly established (Simberloff and Gibbons 2004). Since the peak of shad abundance in Columbia River during 2004 (Fig. 1), the population has undergone a dramatic reduction in abundance. Determination of the driving forces responsible for the population dynamics was beyond the scope of this study; however, a posteriori analyses indicated that the 2007 ichthyophoniasis epizootic coincided with the population decline. The 2007 epizootic was characterized by the highest Ichthyophonus infection prevalence (72%) ever reported from any wild fish population in the Pacific region and 57% of the infections were scored as moderate or heavy intensities. Analogous infection prevalences and intensities occurred during prior ichthyophoniasis epizootics in the Atlantic, when massive mortalities resulted in population declines in Atlantic herring, Clupea harengus (reviewed in Sindermann 1990 and McVicar 1999). Population-level impacts of the epizootic were further suggested by a decline in the Columbia River shad population following the 2007 epizootic and demonstrated high pathogenicity of Ichthyophonus to other clupeids (Kocan et al. 1999). Furthermore, Ichthyophonus-infected fish undergo osmoregulatory distress characterized by significant elevations in sodium levels, which can result in nearly complete mortality during attempted transition from freshwater into seawater (Uno 1990). Therefore, it is possible that mortality among infected adult shad occurred after completion of spawning and prior to re-entry into the seawater. However, it should be cautioned that demonstration of any causal relationships between the ichthyophoniasis epizootic and population-level impacts to American shad require further investigation using well-controlled empirical manipulations, similar to those described by Hudson et al (1998).