INTRODUCTION
The Atlantic snow crab, Chionoecetes opilio, is an economically important decapod in the North Atlantic and Pacific regions. In 2011, the total landings for Atlantic Canada were 84 139 mt with a landed value of CAN$459 million (http://www.dfo-mpo.gc.ca/fm-gp/sustainable-durable/fisheries-peches/snow-crab-eng.htm) (DFO, 2011). One of the major threats to the snow crab fishing industry is the parasitic dinoflagellate Hematodinium sp. that causes bitter crab disease (BCD). The parasite was first reported in C. opilio in 1990 off the coast of Newfoundland at a relatively low prevalence <0·11–3·7% (Taylor and Khan, Reference Taylor and Khan1995). Subsequently peaks in Hematodinium sp. prevalence have approached 35% in some areas and may be contributing to underappreciated impacts on the lucrative snow crab fishery (Shields et al. Reference Shields, Taylor, Sutton, O'Keefe, Ings and Pardy2005, Reference Shields, Taylor, O'Keefe, Colbourne and Hynick2007; Mullowney et al. Reference Mullowney, Dawe, Morado and Cawthorn2011, Reference Mullowney, Dawe, Colbourne and Rose2014; Morado et al. Reference Morado, Siddeek, Mullowney and Dawe2012). Developing a thorough knowledge of the parasite's life cycle is therefore needed in order to understand the disease and predict its effect on crab health and fisheries sustainability (Stentiford and Shields, Reference Stentiford and Shields2005; Stentiford et al. Reference Stentiford, Neil, Peeler, Shields, Small, Flegel, Vlak, Jones, Morado, Moss, Lotz, Bartholomay, Behringer, Hauton and Lightner2012).
During early stages of Hematodinium spp. infections, the parasite proliferates within the haemolymph and haemal spaces of the hepatopancreas eventually spreading systemically to occupy all haemal spaces (Stentiford and Shields, Reference Stentiford and Shields2005). Parasites may be found attached or intimately associated with many tissues including; the outer basal lamina of the hepatopancreatic tubule epithelial cells, the wall of haemolymph vessels, and the myocardial lumina, but tissue invasion beyond the level of haemal spaces is typically not observed (Field and Appleton, Reference Field and Appleton1995; Stentiford and Shields, Reference Stentiford and Shields2005; Small et al. Reference Small, Shields, Reece, Bateman and Stentiford2012). Late stage infections produce gross pathological changes characteristic of BCD including cuticle discolouration and milky white haemolymph associated with large numbers of circulating parasites (Meyers et al. Reference Meyers, Koeneman, Bothelho and Short1987). Throughout late infections, hosts experience haemocytopenia concurrent with profound levels of parasitemia that causes pressure necrosis with respiratory and associated organ malfunctions which may ultimately contribute to the host's death (Shields, Reference Shields1994; Field and Appleton, Reference Field and Appleton1995; Shields and Squyars, Reference Shields and Squyars2000; Wheeler et al. Reference Wheeler, Shields and Taylor2007). Importantly, Hematodinium spp. infections are more commonly observed in juvenile crustaceans which moult more frequently than mature adults and since evidence suggests that all infected crabs succumb to death, infection likely impacts recruitment (Meyers et al. Reference Meyers, Koeneman, Bothelho and Short1987; Eaton et al. Reference Eaton, Love, Botelho, Meyers, Imamura and Koeneman1991; Messick, Reference Messick1994; Stentiford et al. Reference Stentiford, Neil and Atkinson2001; Shields et al. Reference Shields, Taylor, Sutton, O'Keefe, Ings and Pardy2005, Reference Shields, Taylor, O'Keefe, Colbourne and Hynick2007).
Observations from natural infections of Hematodinium spp. have revealed at least three life stages: trophonts, sporonts and dinospores (Meyers et al. Reference Meyers, Koeneman, Bothelho and Short1987; Eaton et al. Reference Eaton, Love, Botelho, Meyers, Imamura and Koeneman1991; Shields and Squyars, Reference Shields and Squyars2000; Li et al. Reference Li, Miller, Small and Shields2011). Trophonts appear vacuolated or ‘foamy’ when viewed by light microscopy and can be uninucleated or multinucleated (Eaton et al. Reference Eaton, Love, Botelho, Meyers, Imamura and Koeneman1991; Appleton and Vickerman, Reference Appleton and Vickerman1998; Morado, Reference Morado2007; Ryazanova, Reference Ryazanova2008; Li et al. Reference Li, Miller, Small and Shields2011). They have been described in vivo as forming ‘sheets’ near various organs, especially the hepatopancreas (Field and Appleton, Reference Field and Appleton1995; Wheeler et al. Reference Wheeler, Shields and Taylor2007; Chualáin and Robinson, Reference Chualáin and Robinson2011). Sporonts arise from trophonts, which have also been described as sporoblasts and pre-spores (Appleton and Vickerman, Reference Appleton and Vickerman1998; Stentiford and Shields, Reference Stentiford and Shields2005; Ryazanova, Reference Ryazanova2008; Li et al. Reference Li, Miller, Small and Shields2011). The final in vivo life stage is the dinospore, the motile dispersal form of the parasite (Appleton and Vickerman, Reference Appleton and Vickerman1998; Stentiford and Shields, Reference Stentiford and Shields2005; Li et al. Reference Li, Miller, Small and Shields2011). Two types of Hematodinium dinospores have been reported in several crustacean hosts: macrodinospores and microdinospores. Hematodinium spp. infections from a single crab have been reported to result in the development of only one type of dinospore (Meyers et al. Reference Meyers, Koeneman, Bothelho and Short1987; Appleton and Vickerman, Reference Appleton and Vickerman1998; Li et al. Reference Li, Miller, Small and Shields2011). Dinospores have been observed in some crustacean hosts to exit the host through the mouth and gills in ‘plumes’ at very high densities (1·6×108 cell mL−1), and may be the infective life stage of this parasite (Stentiford and Shields, Reference Stentiford and Shields2005).
Two previous studies have investigated Hematodinium life stages in vitro; Appleton and Vickerman (Reference Appleton and Vickerman1998) first isolated Hematodinium sp. from the Norway lobster, Nephrops norvegicus, while Li et al. (Reference Li, Miller, Small and Shields2011) isolated Hematodinium perezi from the blue crab, Callinectes sapidus. Both studies identified life stages and presented a life history using light microscopy, but Appleton and Vickerman (Reference Appleton and Vickerman1998) used electron microscopy during their analysis. Chionoecetes opilio can be infected with Hematodinium sp. similar to N. norvegicus (Small et al. Reference Small, Shields, Reece, Bateman and Stentiford2012), but snow crabs inhabit temperatures between −1·5 and 4°C (Dawe and Colbourne, Reference Dawe, Colbourne, Paul, Dawe, Elner, Jamieson, Kruse, Otto, Sainte-Marie, Shirley and Woodby2002; Dawe et al. Reference Dawe, Mullowney, Moriyasu and Wade2012), a temperature range which is below normal for N. norvegicus (4–16°C) and considerably lower than that for H. perezi from C. sapidus (15–30°C) (Cadman and Weinstein, Reference Cadman and Weinstein1988; Baden et al. Reference Baden, Pihl and Rosenberg1990). In order to begin an exploration of the adaptations of this generalist dinoflagellate parasite to cold water hosts, the present study objective was to isolate the Hematodinium sp. from C. opilio and culture the parasite in vitro at 0°C while characterizing developmental life stages using both light and electron microscopy.
MATERIALS AND METHODS
Collection of C. opilio
Atlantic snow crabs (C. opilio) with gross morphological changes (opaque cuticle and milky white haemolymph) consistent with BCD were collected in September 2011. Seven crabs were collected by pot trap in White Bay, Newfoundland at stratum 614 (300–401 m in depth, bottom temperatures ~0–1°C). An eighth crab was obtained off the coast of Nova Scotia at 44°23·23N, 63°28·11W (~100 m in depth, bottom temperatures ~0–3°C). Crabs were kept in coolers layered between seawater-soaked burlap, placed above saltwater ice, and sent via air cargo (Newfoundland) or land transport (Nova Scotia) within 24–48 h to the Atlantic Veterinary College (AVC).
Haemolymph collection and in vitro cultures
On arrival, crabs were stored in an aerated holding tank at 0°C with 33–34 ppt salinity for 1–4 h during which haemolymph was collected aseptically from the base of the first periopod using a 3 or 5 mL syringe with 22 gauge needle. Crabs were exsanguinated and subsequently humanely euthanized by removal of the carapace and disruption of the ventral nerve cord. Infected haemolymph (40 μL) was added to 1·94 mL of modified Nephrops medium (Appleton and Vickerman, Reference Appleton and Vickerman1998) that included gentamicin (12 μg mL−1) (Sigma-G1264) in each of 24 multiwell plates (Falcon) and incubated in the dark at 0±0·2°C (Binder-APT.line™ KB). Manipulations of cultures for microscopy and maintenance required brief exposure to typical room temperature (~18–20°C) and lighting. Cultures were monitored on an inverted phase contrast microscope (Nikon) with transmitted light every 2–3 days. Cultures were diluted 1:1 with fresh medium when cell density covered ~2/3 of culture well. Hematodinium spp. cultures were transferred to fresh culture medium every 7–10 days by pipetting a 1 mL aliquot of non-adherent cells (care was taken to minimize disruption of cells that were attached to plate surface) into a microfuge tube that was centrifuged at 900 g for 5 min at 4°C. After centrifugation, supernatant was discarded, and the cell pellet was resuspended in 1 mL of fresh medium and transferred to a new well on the culture plate. Fresh media was transferred into the original culture well plate containing adherent cells, which had been kept in the dark at 0±0·2°C. If cultures developed contaminating organisms, a 1 mL aliquot of sample was submitted to AVC Diagnostic Services, to classify the bacterial organism and elucidate antibiotic sensitivity and concentrations for treatment options. If contamination could not be resolved, the culture was discarded. Samples were periodically assessed for viability by staining with 0·2% Neutral Red (Sigma-Aldrich) or 0·05% Acridine Orange (Sigma-Aldrich). Samples were stained 24 h after incubation, 3 days following, and after any notable changes in cell morphology. All images were captured using an Axiocam ACC camera. All cultures were terminated after 16 weeks in vitro.
Preparation and imaging of samples for transmission electron microscopy
Haemolymph from BCD suspect crabs was prepared following Gaudet et al. Reference Gaudet, Cawthorn, Morado, Wadowska, Wright and Greenwood2014. Briefly, Haemolymph was fixed in 3% glutaraldehyde (Canemco) in ASW (Instant Ocean, 33 psu) and stored at 4°C for 24 h. Samples were then centrifuged, washed twice in ASW and post-fixed in 1% OsO4 in ASW for 30 min, and concentrated by centrifugation. After post-fixation, osmium tetroxide was removed from samples, and the pellets were overlaid with 4% aqueous molten agar (Fisher) and incubated at 4°C to solidify the agar. Samples were cut and placed in glass vials containing 50% ethanol, dehydrated in an ethanol series, cleared in propylene oxide (PO), infiltrated with Epon resin and placed in BEEM capsules (Canemco) and polymerized.
Cultures were fixed 1:1 in 6% glutaraldehyde in Nephrops saline at pH 7·6 prior to being washed, post-fixed, overlaid in agar, dehydrated, infiltrated with Epon and polymerized as above. Thick sections (0·5 μ m) and ultrathin sections (90 nm) were prepared and viewed with a Hitachi H7500 transmission electron microscope operated at 80 kV and digital images were taken using an AMP XR40 side mounted camera (Gaudet et al. Reference Gaudet, Cawthorn, Morado, Wadowska, Wright and Greenwood2014).
RESULTS
A chronology of the Hematodinium sp. life stages observed by light microscopy and confirmed by transmission electron microscopy (TEM) is detailed in Table 1. All life stages had condensed chromosomes, amphiesmal vesicle membranes, lipid droplets, membrane-bound electron dense granules, mitochondria with tubular cristae and Golgi that often had dilated cisternae. All stages were confirmed as Hematodinium sp. based on ITS PCR and DNA sequencing (data not shown).
Chronology of Hematodinium life stages as determined by light and TEM. Identification numbers correspond to isolates from individual crabs. Amoeboid trophonts were observed in most initial haemolymph collections with the exception of samples #65 and NS-2011, which contained sporonts. Amoeboid trophonts cultures remained at that stage until the completion of the trial or until cultures were terminated. In some cases (#61, #62, #64) early sporonts were observed, but only at one time point. Sporont initiated cultures gave rise to dinospore stages. Schizont-like stages and dense clump aggregates were observed, but were deemed senescent based on ultrastructural analyses. Five cultures were terminated due to bacterial and yeast contamination. All cultures were terminated by 16 weeks
AT, amoeboid trophont; CC, clump colony; S, sporont; D, dinospore; AS, arachnoid sporont; Sb, sporoblast; Sch, Schizont-like; DC, dense clump aggregates; X, cultures terminated.
Six of the initial eight haemolymph samples contained amoeboid trophonts that were uninucleated or multinucleated with condensed chromosomes and possessed numerous intracytoplasmic lipid droplets and membrane-bound electron dense granules but lacked trichocysts. Trophonts (~15–30 μ m in diameter) varied significantly in their morphology; a spherical or irregular (amoeboid) form predominated, while others were elongated (filamentous) in shape (Fig. 1). Occasionally, a few scattered cells within the sample had developing or mature trichocysts when examined ultrastructurally. Trichocysts however were not observed from any time points thereafter, indicating that transition only occurred in a small subset of the sample, and did not represent progression of the life cycle to the sporont stage within the culture as a whole. In vitro cultures seeded with trophonts proliferated but did not progress.
Amoeboid trophonts from haemolymph of C. opilio. (A) Transmission electron micrograph of amoeboid trophonts with ovoid or spherical morphology, nucleus (N) with condensed chromosomes and cytoplasm containing numerous lipid droplets (L) and membrane-bound electron dense granules (ED). Scale bar: 2 μ m; (B) A bi-nucleated amoeboid trophont had filamentous morphology, ultrastructurally similar to A. Scale bar: 2 μ m; (C) Amoeboid trophonts with filamentous (arrow) and ovoid morphology (arrowhead) were found within the same culture. Acridine Orange stain with fluorescence microscopy. Scale bar: 15 μ m.
Sporonts
Two initial haemolymph samples (#65 and NS-2011) contained sporonts (Table 1). In recently isolated haemolymph, sporonts were irregular in shape and similar in morphology and size to trophonts under light microscopy (Fig. 2). Ultrastructurally, sporonts were similar to trophonts except that they possessed trichocysts (Fig. 3). As time elapsed, the outer amphiesmal vesicle membrane and the plasma membrane became tightly apposed, forming spherical cells (Fig. 4). Rapid proliferation of the sporonts was observed with single sporonts giving rise to four daughter cells, each containing trichocysts (Fig. 5). Cultures seeded with sporonts progressed to dinospores in 5–9 weeks.
Light micrograph of aggregated clusters of Hematodinium sp. from in vitro culture. (A) Aggregated amoeboid trophonts with irregular morphology. Scale bar: 30 μm; (B) aggregated sporonts observed by light microscopy. Scale bar: 15 μ m.
Transmission electron micrograph of a typical sporont from the haemolymph of C. opilio. The nucleus (N) contained condensed chromatin, and the inner and outer amphiesmal vesicle membranes (double arrow) were surrounded by the plasma membrane (arrow). Lipid droplets (L), membrane-bound electron dense granules (ED) and trichocysts (T) were evident. Scale bar: 2 μ m.
Light micrograph of uninucleated, spherical, late stage sporonts after 3 weeks in vitro. Scale bar: 10 μ m.
Transmission electron micrograph of a late stage sporont dividing into four uninucleated daughter cells. Note the close association between the amphiesmal membranes and the plasma membrane (arrow), resulting in a spherical morphology of cells. T, trichocysts; N, nucleus; L, lipid droplet; ED, membrane-bound electron dense granules. Scale bar: 2 μ m.
Plasmodia
Multinucleated plasmodial (syncytia) cell masses were observed within in vitro cultures in two forms; a non-adherent spherical or flattened cell mass and an adherent, web-like network of interconnected cellular bodies (Fig. 6). In trophont cultures only the spherical or flattened form was present which was reminiscent of clump colonies (Fig. 6A). In sporont cultures both forms were present, the spherical or flattened cell masses most closely resembled sporoblasts and the web-like network was morphologically similar to arachnoid sporonts (Fig. 6). The arachnoid sporonts adhered to the culture well and were sensitive to mechanical disruption. Changing the medium disrupted these networks and only in a few instances did the networks reform in the new wells. Cytoplasmic extensions from arachnoid sporonts did not contain lipid droplets, membrane-bound electron dense granules, nor did they contain microtubules (Fig. 7). All plasmodial cell masses were associated with uninucleated and multinucleated cells that appeared to arise via nuclear division without subsequent cytokinesis from the plasmodial surface (Fig. 6). Electron microscopy revealed the presence or absence of trichocysts within plasmodia which allowed them to be differentiated as either trophonts or sporonts, respectively.
Light micrograph of plasmodia from in vitro cultures. (A) A spherical or flattened plasmodium with uni- and multinucleated cells budding from the periphery (arrows), stained with Neutral Red vital stain; (B) arachnoid sporont with complex networking (arrowheads) between cellular bodies (unstained). Scale bar: 30 μ m.
Transmission electron micrograph of cytoplasmic extensions from an arachnoid sporont 1 week in vitro. (A) Cytoplasmic extensions (double sided arrow) emerge from the body of the cell. Scale bar: 2 μ m; (B) higher magnification showing a lack of membrane-bound electron dense granules (ED), lipid droplets and microtubules within extensions. Scale bar: 500 nm.
Dinospores
Ellipsoid dinospores measuring ~14 μ m long and 11 μ m wide (range 12–16 μ m long and 9·5–13 μ m wide; n = 50) were observed (Fig. 8). All dinospores were uninucleate, often with nuclear membrane separation revealing small fibrillar material within the intramembranous space (Fig. 9). Dinospores were characterized by motility within culture and by the ultrastructural identification of flagella having typical 9+2 arrangement of microtubules. Each dinospore possessed one longitudinal and one transverse flagellum, both of which emerged from the anterior region of the cell (Fig. 10). Dinospores arose synchronously within cultures; with most cells transforming from sporonts to dinospores 1 week after the first dinospore was observed. Dinospores were observed by TEM within tightly packed monolayers of cells, 1 week prior to visible dinospore motility in cultures (Fig. 10). After ~2 weeks in culture as dinospores, cells lost their flagella becoming non-motile and spherical (Fig. 11). Ultrastructure revealed beaded heterochromatin along the periphery of the nucleus, with euchromatin located centrally (Fig. 12). Trichocysts remained within the cells with some discharged intracellularly within autophagosomes (data not shown).
Light micrograph of uninucleate, ellipsoid macrodinospores ~1 month in vitro. Scale bar: 15 μ m.
Transmission electron micrograph indicating separation of nuclear membranes (arrows) of a mature dinospore. Fibrillar material (F) was observed within the intramembranous space. N, nucleus; L, lipid droplet. Scale bar: 500 nm.
Transmission electron micrograph of a mature dinospore from within a dense cluster of cells 1 week prior to motility in vitro. Two flagella (arrows) were evident at the anterior end of the cell, the nucleus (N) was positioned centrally, and partially degranulated membrane-bound electron dense granules (ED), trichocysts (T) and lipid droplets (L) were found throughout the cytoplasm. The plasma membrane (arrowhead) was seen bulging from the cell. Scale bar: 2 μm.
Non-motile dinospores observed by light microscopy ~2 weeks after dinospore formation in vitro. Hematodinium lost their flagella, became non-motile and rested on the bottom of the culture well. Notably, cells were spherical and uninucleated, with the nucleus (arrow) occupying most of the volume of the cell. Scale bar: 10 μ m.
Transmission electron micrograph of non-motile dinospore stage. The nucleus occupied much of the cell volume with euchromatin central and condensed heterochromatin peripheral (arrows). The cytoplasm contained lipid droplets (L), trichocysts (T), and degranulated membrane-bound electron dense granules (ED). Scale bar: 2 μ m.
Schizont-like stages
Schizont-like cells were observed within in vitro cultures after approximately 3 months and were also found in cultures with concurrent bacterial, yeast or biflagellate contamination. Schizonts were large (>60 μ m in diameter), circular cells that possessed a large vacuole with peripheral cytoplasm and nucleus (Fig. 13A). The vacuole was relatively empty with clumps of electron dense granular material located along the periphery of the cell (Fig. 13B). This material appeared to have a similar consistency to the membrane-bound electron dense granular material. Concurrent with the appearance of schizont-like cells in some cultures, was the appearance of small (4–6 μ m long) ellipsoid biflagellate cells after 3 months in culture (Fig. 13A). TEM revealed that these organisms lacked characteristic dinoflagellate condensed chromosomes and amphiesmal vesicle membranes and possessed flagellar hairs. These were most likely zoosporic fungi belonging to either Labyrinthulale or Thraustochytriale orders (data not shown).
Schizont-like cells from in vitro culture. (A) Several schizont-like cells (S) observed by light microscopy adjacent to numerous small, biflagellated contaminants (arrow). Schizont-like cells appeared as ring shaped due to a large central vacuole. The region where nuclei and other organelles are situated form distinct bulges along the periphery (arrowhead). Scale bar: 10 μm. (B) A binucleated schizont-like cell observed by TEM. The periphery of the large, relatively empty vacuole (V) was lined with granulated material. Notably, the outer amphiesmal vesicle membrane and plasma membrane (arrow) did not bulge but tightly surrounded the cell. N, nucleus; L, lipid droplet. Scale bar: 2 μ m.
Dense Clump-like aggregates
Shortly after initiation of cultures #62 and #63, dense clumps of cells arose within monolayers of amoeboid trophonts (Fig. 14). Ultrastructurally, these clumps consisted of membranous material, lipid droplets, and other degraded organelles resembling cellular debris from culture (Fig. 15). Membrane bound cells were not evident, and nuclear structures were absent within clumps.
A light micrograph of a dense cell aggregate. A large, dense clump (C) measuring ~180 μ m was observed within a monolayer of trophonts (arrow). Scale bar: 30 μ m.
Transmission electron micrograph of a dense clump aggregate. Nuclear and cellular integrity was absent and consisted of membranous material (white arrow), and organellar remnants (double arrow). No distinct membrane boundary was found between the clump and the surrounding environment (black arrow). Scale bar: 2 μ m.
Cytoplasmic autofluorescence and electron dense granules
In all Hematodinium cultures and life stages, cytoplasm-specific autofluorescence was observed in direct smears excited with ultraviolet light (Fig. 16). Additionally, many areas of the cytoplasm stained orange with Acridine Orange and red with Neutral Red (Fig. 17), indicative of acidic organelles such as lysosomes. Similarly, electron dense granules surrounded by a single membrane were observed by TEM within the cytoplasm (Fig. 18). Membrane-bound electron dense granules were 14·0 nm (n = 100) in diameter in parasites from the initial haemolymph samples, and were significantly larger 24·0 nm (n = 100) after 1 month in vitro culture, often with the granular material having a ‘halo’ appearance (Fig. 19). Shortly after cultures were initiated, the electron dense granules appeared to degranulate and this continued throughout in vitro cultivation (Fig. 19). As time in culture progressed, the number of membrane-bound electron dense granules decreased; however the size of vesicles increased. At the end of the 16 week trial, most Hematodinium contained only one large, relatively empty, vesicle containing peripheral clumps of granular electron dense material.
Direct smear of Hematodinium sporonts after 2 weeks in vitro. (A) Hematodinium stained with Wright-Giemsa reveal dense chromatin (purple) within nuclei; (B) when viewed with ultraviolet light passed through an FITC filter, autofluorescent pigments (green) were observed; (C) when images ‘A’ and ‘B’ were merged, autofluorescence was restricted to the cytoplasm. Scale bar: 10 μ m.
Amoeboid trophonts and sporonts vitally stained after 1 month in vitro and viewed by light microscopy. (A) An aggregate of amoeboid trophonts stained with neutral red, revealed many lysosomes (red); (B) sporonts stained with acridine orange fluoresce, revealed orange cytoplasm indicative of acidic organelles such as lysosomes. Permanently condensed chromatin, characteristic of dinoflagellates, were stained green within the nucleus. Scale bar: 10 μ m.
Transmission electron micrograph of an amoeboid trophont ~3 weeks in vitro. Heterogeneous, electron dense granules (ED) were contained in large, single membrane vesicles that were usually non-spherical and varied in shape. These were often seen adjacent to lipid droplets (L). Scale bar: 2 μ m.
Transmission electron micrograph of amoeboid trophont indicating electron dense granule degradation. Electron dense (ED) material degranulation varied from minimal (white arrow), moderate (double white arrow), to considerable (black arrow) degranulation. Granular material became less electron dense as degranulation occurred, giving a ‘halo’ appearance (double white arrow). As degranulation progressed, large electron lucent areas within the vesicle were seen (black arrow). L, lipid body. Scale bar: 500 nm.
DISCUSSION
Hematodinium sp. was isolated from infected C. opilio and in vitro cultures were maintained at 0°C for the first time. Although cultures were initiated from a small number of infected snow crabs, a partial characterization of several distinct life stages was determined including; trophonts, clump colonies, sporonts, arachnoid sporonts, sporoblasts and macrodinospores. Trophonts were observed in all but two of the initial crab haemolymph samples and when cultured in vitro, continued to proliferate. Uninucleate and multinucleate trophont morphologies were observed; a predominant amoeboid form and a less common filamentous form similar to those described previously in Hematodinium sp. from N. norvegicus (Appleton and Vickerman, Reference Appleton and Vickerman1998) and H. perezi from C. sapidus (Li et al. Reference Li, Miller, Small and Shields2011). Both trophont types were ultrastructurally similar based on condensed nuclear chromatin, lipid droplets, membrane-bound electron dense granules and the absence of trichocysts. Motile filamentous forms and gorgon locks life stages as described in Hematodinium sp. from Norway lobsters (Appleton and Vickerman, Reference Appleton and Vickerman1998) and H. perezi from blue crabs (Li et al. Reference Li, Miller, Small and Shields2011) were not observed in this study. This was expected as snow crabs were selected based on the presence of grossly visible clinical signs of BCD (milky haemolymph and opaque shells) which typically only contains Hematodinium sp. late life stages and thus early, motile filamentous forms were unlikely to be observed in the haemolymph. Furthermore, although the trophont stages isolated from C. opilio increased in density, they did not progress in culture. In order therefore to increase the probability of isolating and characterizing the early life stages of Hematodinium sp. to obtain the full life cycle; juvenile C. opilio will need to be collected in the Spring (Shields et al. Reference Shields, Taylor, Sutton, O'Keefe, Ings and Pardy2005).
Trophonts were observed as individual cells, as large aggregations of individual cells (monolayers), or as plasmodial stages resembling clump colonies. Clump colonies appeared to be associated with proliferation of amoeboid trophonts through budding from the plasmodial surface as uninucleate or multinucleate cells, but no progression into further life stages was observed. Along with the gorgonlocks stage from which the clump colonies are thought to be derived, no arachnoid trophonts were observed as found in Hematodinium sp. from N. norvegicus (Appleton and Vickerman, Reference Appleton and Vickerman1998) and H. perezi from C. sapidus (Li et al. Reference Li, Miller, Small and Shields2011). The likely reason for the absence of the gorgon locks and arachnoid trophont stages and the lack of life cycle progression is due to a variety of factors including reduced temperature, salinity, nutrients, contaminants, etc. within the culture conditions. Once again, these observations were only made from two trophont derived in vitro cultures.
Two initial snow crab haemolymph samples contained Hematodinium sp. that morphologically resembled amoeboid trophonts using light microscopy, but were determined to possess trichocysts when viewed ultrastructurally and so were deemed to be sporonts (Gaudet et al. Reference Gaudet, Cawthorn, Morado, Wadowska, Wright and Greenwood2014). The appearance of trichocysts has previously been used in Hematodinium sp. from Chionoecetes bairdi (Meyers et al. Reference Meyers, Koeneman, Bothelho and Short1987) and N. norvegicus (Appleton and Vickerman, Reference Appleton and Vickerman1998) for delineating the transition from trophont to sporont stages. Interestingly, this does not appear to be a consistent feature in Hematodinium species isolated from different hosts (Hudson and Shields, Reference Hudson and Shields1994; Gaudet et al. Reference Gaudet, Cawthorn, Morado, Wadowska, Wright and Greenwood2014). Whether the presence of trichocysts delineates the vegetative trophont from the non-vegetative sporont or whether it defines a commitment of the sporont to dinospore production, will require further descriptions of these organelles using EM in Hematodinium spp. isolated from other hosts.
Sporonts were found as individual cells, as cellular aggregates (monolayers), or as plasmodia within in vitro cultures. The plasmodial stages contained trichocysts and were present in two forms, the adherent arachnoid sporont and the non-adherent sporoblasts. Sporonts were observed to bud from the periphery or surface of the plasmodial cell masses similar to the process suggested for amoeboid trophonts. The predominant plasmodial stage in our cultures was the spherical or flattened sporoblasts and less commonly the arachnoid sporont stage. Both plasmodial forms have been described in Hematodinium sp. from N. norvegicus (Appleton and Vickerman, Reference Appleton and Vickerman1998) and H. perezi from C. sapidus (Li et al. Reference Li, Miller, Small and Shields2011). These plasmodial stages may be comparable to in vivo descriptions of circulating cell aggregations that have been observed within the haemolymph as well as sheet-like stages from histopathological studies of Hematodinium sp. infections of N. norvegicus (Field and Appleton, Reference Field and Appleton1995), C. opilio (Wheeler et al. Reference Wheeler, Shields and Taylor2007) and Cancer pagurus (Chualáin and Robinson, Reference Chualáin and Robinson2011).
Dinospores appeared approximately 1 month after cultures were initiated. Since the two initial sporont cultures gave rise to dinospores, this suggests that Hematodinium sp. may be committed to proceed to dinospore production after it reaches the sporont developmental stage. The biflagellate dinospores contained condensed chromatin, trichocysts, lipid droplets and membrane-bound electron dense granules. Similar to sporoblasts and dinospores of Hematodinium sp. from N. norvegicus and C. bairdi, fibrillar material was present between membranes of the nuclear envelope (Meyers et al. Reference Meyers, Koeneman, Bothelho and Short1987; Appleton and Vickerman, Reference Appleton and Vickerman1998). In the present study the function of the fibrillar material is unknown but since it was only found in mature dinospores possessing flagella, it is unlikely to represent flagellar hairs as originally proposed by Meyers et al. (Reference Meyers, Koeneman, Bothelho and Short1987).
Two morphologically distinct dinospores have been reported within Hematodinium spp., termed macrodinospore and microdinospore (Appleton and Vickerman, Reference Appleton and Vickerman1998; Stentiford and Shields, Reference Stentiford and Shields2005; Li et al. Reference Li, Miller, Small and Shields2011). In H. perezi isolated from C. sapidus macrodinospores were 14·3±1·6 μ m long and microdinospores were 9·2±2·4 μ m long (Li et al. Reference Li, Miller, Small and Shields2011), while Hematodinium sp. dinospores isolated from N. norvegicus showed slightly larger dimensions with macrodinospores and microdinospores measuring 16–20 and 11–14 μ m long, respectively (Appleton and Vickerman, Reference Appleton and Vickerman1998). Macrodinospores have a characteristic ellipsoid shape, and microdinospores have a ‘corkscrew’ appearance (Appleton and Vickerman, Reference Appleton and Vickerman1998; Li et al. Reference Li, Miller, Small and Shields2011). In the present study, dinospores from two cultures were both ellipsoid, ~14 μ m long (ranging 12–16 μ m) and were ~11 μ m wide (ranging 9·5–13 μ m), and were morphologically similar to macrodinospores. Consistent with previous research, only one dinospore type developed within each haemolymph sample arising from an individual crab (Meyers et al. Reference Meyers, Koeneman, Bothelho and Short1987; Appleton and Vickerman, Reference Appleton and Vickerman1998; Li et al. Reference Li, Miller, Small and Shields2011). The lack of microdinospores arising in cultures from this study is likely the result of the small number of haemolymph samples that were cultured. Interestingly, dinospores were found within dense aggregates of cells by TEM 1 week prior to observing dinospore motility in culture. This lag time is most likely a reflection of the development of functional flagella, but may also have been a function of developmental delays due to the relatively cold temperature (0°C) at which the cultures were incubated.
Dinospore germination was initially thought to have occurred as cells became immobile with a concurrent loss of flagella and change from an ellipsoid to a spherical morphology. Ultrastructural examination however revealed that nuclear changes had occurred with the formation of beaded heterochromatin peripherally and granular euchromatin material throughout the nucleus. Additionally, trichocysts were often found in autophagosomes suggesting that dinospores became senescent and therefore could not transition to the trophont stage as in previous in vitro studies (Appleton and Vickerman, Reference Appleton and Vickerman1998; Li et al. Reference Li, Miller, Small and Shields2011). Appleton and Vickerman (Reference Appleton and Vickerman1998) described the need to adjust the cell density to achieve a monolayer of cells to induce further life stage development in Hematodinium sp. from N. norvegicus. In the present study, monolayers were difficult to maintain as the life stages tended to aggregate near the centre of the well even after cultures were diluted and spread uniformly. Thus a lack of monolayer integrity may have impacted life cycle progression in the present study. Additionally, Li et al. (Reference Li, Miller, Small and Shields2011) supplemented their culture media with haemolymph from C. sapidus throughout culturing which likely provided host cues which facilitated successful H. perezi life cycle progression. No additional supplementation was used in the present study and thus insufficient environmental or host cues may have prevented culture progression and viability which may be critical for Hematodinium sp. from C. opilio. It is logical to consider that biotic and abiotic factors including time in culture, cell density, nutrient availability and host factors may play important roles in the successful development and completion of the life cycle of Hematodinium spp.
Dense cell aggregates associated with monolayers in our cultures resembled the ‘clumping cell aggregates’ that occurred within in vitro cultures of Hematodinium sp. from C. bairdi (Meyers et al. Reference Meyers, Koeneman, Bothelho and Short1987). Ultrastructure revealed however that these dense cell aggregates were composed of coalesced cellular debris and were in fact artefacts. Interestingly, Meyers et al. (Reference Meyers, Koeneman, Bothelho and Short1987) also described single cells arising from clump aggregates as hypertrophied and senescent.
Similarly, large schizont-like stages were observed in the present study by light microscopy within both older cultures and those contaminated with other organisms. However, these schizont-like stages did not progress through the gorgonlocks shunt described for the viable schizont life stages of H. perezi (Li et al. Reference Li, Miller, Small and Shields2011). TEM revealed a peripheral nucleus with minimal cytoplasm containing a large vacuole lined with peripheral clumps of coalesced electron dense material. Therefore, the schizont-like stages observed in the present study are an aberrant form indicative of the overall poor health of the culture as they were only seen during unfavourable environmental conditions.
In all Hematodinium sp. stages derived from C. opilio, autofluorescent single membrane-bound electron dense granules were observed in great quantities. These granules resembled lipofuscin as described in vertebrates and invertebrates (Sarna et al. Reference Sarna, Burke, Korytowski, Różanowska, Skumatz, Zaręba and Zaręba2003; Vogt, Reference Vogt2012). Degranulation occurred early during in vitro culture and there was a marked depletion of electron density by the end of the 16 week trial as several Hematodinium sp. cultures became senescent. Although lipofuscin is presumably not completely degradable (Terman and Brunk, Reference Terman and Brunk2004), piecemeal degranulation and ultimate exocytosis of undegradable granule residua has been reported in both mussels and humans (Fonseca et al. Reference Fonseca, Sheehy, Blackman, Shelton and Prior2005; Zorita et al. Reference Zorita, Ortiz-Zarragoitia, Soto and Cajaraville2006). Regardless of the origin, this homogenous granular substance is distinct from lipid material observed in the vegetative and sporoblast stages (trophont and sporont) but not dinospores of Hematodinium sp. from C. bairdi (Meyers et al. Reference Meyers, Koeneman, Bothelho and Short1987), and other granulated vesicles found in many stages of Hematodinium sp. isolated from Chionoecetes tanneri (Morado, unpublished). Specific reference to similar membrane-bound electron dense granules has not been made in Hematodinium sp. isolated from N. norvegicus (Appleton and Vickerman, Reference Appleton and Vickerman1998) or the king crabs Paralithodes camtschaticus and Paralithodes platypus (Ryazanova et al. Reference Ryazanova, Eliseikina, Kukhlevsky and Kharlamenko2010), neither in H. perezi from Liocarcinus depurator (Small et al. Reference Small, Shields, Reece, Bateman and Stentiford2012), nor in Hematodinium australis from Portunus pelagicus (Hudson and Shields, Reference Hudson and Shields1994). The exact nature of the electron dense material will require further analysis; however, it is interesting to speculate that the material is derived from their cold water hosts and plays an important role in the parasites adaptation to this temperature and in life cycle progression.
ACKNOWLEDGEMENTS
The authors wish to thank Dorota Wadowska, Sarah Ramsay-Ogilvie, Sarah Daley and Dr Janet Saunders for technical assistance. Additionally, the study could not have taken place without the commitment of Earl Dawe and Darrell Mullowney along with the crews and technicians on the Fisheries and Oceans trawler survey ships who collected snow crab samples. The authors would like to anonymous reviewers for their constructive comments during their critique of the manuscript.
FINANCIAL SUPPORT
This research was funded by NSERC-Strategic Projects Grant (STPGP 37253-2008) and through the supporting organization at the Canadian Centre for Fisheries Innovation and the Fish, Food & Allied Workers Union.
