Introduction

Only four marine myxozoan life cycles involving polychaete hosts have been elucidated and controlled using DNA analysis. Ellipsomyxa gobii Køie, 2003 and Ellipsomyxa mugilis (Sitjà-Bobadilla and Alvarez-Pellitero, 1993) use Nereis spp. (Nereidae), Gadimyxa atlantica Køie et al., 2007 (Parvicapsulidae) uses Spirorbis spp. (Spirorbidae) and Ceratomyxa auerbachi Kabata, 1962 (Ceratomyxidae) uses Chone infundibuliformis Krøyer, 1856 (Sabellidae), as invertebrate hosts (Køie et al. 2004, 2007, 2008; Rangel et al. 2009). Marine actinosporean stages have in addition been found in the polychaetes Hydroides norvegicus Gunnerus, 1768 (Serpulidae), in unidentified spionids (Spionidae) and in Diopatra neapolitana Delle Chiaje, 1841 (Onuphidae; Køie 2002, 2005; Rangel et al. 2011). In the present study actinospores found in Nereis pelagica (L.) are identified with the garfish (Belone belone L.) myxosporean Myxidium sphaericum on the basis of SSU rDNA sequences. M. sphaericum was originally described from B. belone from the Mediterranean and Atlantic coasts off France (Thélohan 1895). However, the species has later been identified and described from other hosts such as the gadid Merlangius merlangus (L.), widening the species conception. The conspecificity of M. sphaericum from belonid and gadid hosts has been questioned (MacKenzie and Kalavati 1995). Here we redescribe M. sphaericum from B. belone and show that the species does not belong in genus Myxidium and is not closely related to Myxidium spp. from gadids. Based on the morphological and phylogenetic distinctness from genus Myxidium sensu stricto, the novel genus Sigmomyxa is proposed for M. sphaericum.

Material and methods

Specimens of the garfish B. belone (L.) (Teleostei, Belonidae) were obtained from local fishermen in May to June 2008 and in September 2010. These fish were caught in the northern Øresund, Denmark. Also, an M. sphaericum-infected specimen (76 cm) was collected in June 2003, near Misje, western Norway. Fresh smears of parts of the urinary system and bile were examined at high magnification (×1,000), and myxosporean spores and plasmodia were photographed. Measurements were taken from digital images using the software ImageJ 1.43u. Measurements are in micrometres.

About 45 specimens of N. pelagica L. (Annelida, Polychaeta, Nereidae) were dredged in the northern Øresund, Denmark at 4–12 m (12–20‰ salinity) during 2008–2010. Most specimens were examined for actinosporean infections immediately upon capture; a few were kept live in aquaria and examined later. Actinosporean measurements were taken from live specimens.

Two infected gall bladders of B. belone from Denmark in 2008 (isolates SigBel-1 and 2), from Norway in 2003 (isolate Msph) and pieces of infected N. pelagica (isolates Npel-1 and 2) were fixed in absolute ethanol for later DNA extraction. Morphological description is based on material represented by isolate ‘Msph’ and an infection from Denmark in 2010. Additional sequences for comparison were obtained from two Myxidium laticurvum Kabata, 1962-infected Trachinus draco L. caught in northern Øresund, Denmark (sequence isolates Mtra and MyFj-1; GenBank Accession No’s JN033229-30 respectively), and from a Myxidium bergense Auerbach, 1909-infected Pollachius virens (L.) from Bodø, Norway (isolate X1; JN033231).

DNA was extracted from myxosporean spores and developmental stages or pieces of polychaetes using the DNeasy® Tissue Kit protocol for animal tissues (Qiagen, Hilden, Germany). The PCR primers used were the forward primers Erib1/Ur-R1, MarF1/RosR2 and MyxF2/Myxgen4R (see Køie et al. 2008), with annealing temperatures 57°C, 60°C and 58°C, respectively. The novel reverse primer Ur-R1 has sequence 5′-AAG AAT TTC ACC TCT CGC CA. The PCR amplifications were performed in a total volume of 50 μl using 2 μl of template DNA and a reaction mixture consisting of 10 μl 5× PCR buffer, 3 μl 25 mM MgCl2,5 μl 10 mM dNTP, 2 μl (10 mM) of the reverse and forward primer, 2 U of thermostable DNA polymerase (GoTaq) and 26 μl dH2O. The PCR conditions were as previously described (Køie et al. 2008). The PCR products were cleaned with ExoSAP-IT® (Affymetrix Inc.) and then sequenced using the BigDye® Terminator v3.1 Cycle Sequencing Kit. The sequencing was done using the amplification primers. The sequence data were assembled with the Vector NTI 11 software (Invitrogen) and GenBank searches were done with Blast (2.0).

The phylogenetic relationship between M. sphaericum and related members of the ‘marine clade’ (Fiala 2006) was examined using Bayesian inference (MrBayes 3.1.2; 3 × 106 generations), maximum likelihood (ML, Paup 4.0b10; 100 bootstrap replicates) and maximum parsimony (Mega 4.0.2). Sequences were aligned using AlignX (Vector NTI), and manually edited in Genedoc. Hypervariable or ambiguous regions were deleted in order to achieve comparison of homologous positions. The Paup ML analysis used a heuristic search algorithm with 10 random sequence additions and TBR branch swapping. A GTR + Γ + I model of nucleotide substitution was selected (AIC) following the examination of the data set with the program jModeltest (Posada 2008). The model was implemented in Paup (Paup Block) and MrBayes (Invgamma).

Results

Infections in B. belone and N. pelagica

Myxospores identified as M. sphaericum Thelohan were found in the gall bladder of one specimen of B. belone in May 2008 and in one in June 2008. The total number of specimens examined in 2008 was about 100. One of 200 gall bladders examined in September 2010 was infected. All caught in the northern Øresund, Denmark. A single B. belone caught near Bergen, western Norway was also infected with M. sphaericum in the gall bladder. No myxospores were found in the urinary system of the B. belone examined in 2008.

None of the about 40 specimens of N. pelagica caught in 2008–2010 and examined immediately upon capture were infected with actinospores. In August 2010 six specimens (4–6 cm long) were isolated in glass containers (8°C, 20‰ S) without air supply. Accidentally the temperature increased to about 22°C for about 24 h. During this period one specimen died. When it was examined the following day, it contained hundreds of free actinospores. Most of the internal organs had disappeared and bacteria occurred in the body cavity. The remaining five surviving specimens were not infected.

Description of M. sphaericum from B. belone

Plasmodia were spherical to irregular (Fig. 1a, b), attached to gall bladder epithelium or free, often with brushy region representing the zone of epitheliar adhesion. Parts of plasmodia with distinct ectoplasm when unsporulated (Fig. 1a), while ectoplasm often not apparent in fully sporulated ones (Fig. 1a, b). Plasmodia contained refractive granules, scattered or in aggregations, increasing in number during sporogony. Mature plasmodia were usually markedly vacuolate (Fig. 1b). Plasmodia were generally disporic, rarely tetrasporic. Sporogony was disporic. Occasionally, plasmodia releasing spores were seen to contain a second sporoblast at an early stage in sporogony. Plasmodia without spores measured up to 36 μm in average diameter (N = 16), plasmodia with immature spores 20–30 μm (N = 4), and plasmodia with two mature spores 23–37 μm (mean 28 μm, N = 11).

Fig. 1
figure 1

Plasmodia and myxospores of S. sphaerica from the gallbladder of B. belone. a Plasmodium (flattened) without visible indication of sporogony, showing distinction between ecto- and endoplasm. b Sporulated plasmodium (flattened) showing spores in valvular view, vacuolate appearance and refractive granules. Note that any polar capsule lengths taken in valvular view may be erroneously short due to their oblique orientation in the spores. c Spore in sutural view. d, e Spores as seen in the focal plane of one polar capsule, showing polar filament coils and the valvular extensions associated with the protruding part of the capsules. Scale bars a, b 10 μm, c, d 5 μm

Full size image
Fig. 2
figure 2

Line drawings of S. sphaerica myxospores from B. belone. a Valvular view and b sutural view. Scale bar 5 μm

Full size image
Table 1 Measurements (in micrometres) of S. sphaerica from B. belone in Denmark and Norway (this study) compared with those published previously from B. belone
Full size table

Spore main outline was ellipsoid (Fig. 1c), but valvular protrusions were associated with the tip of the polar capsules (PC), giving a sigmoid appearance in sutural view and spindle shape in valvular view (Fig. 2). Immature spores were occasionally crescent shaped. Valves were smooth, thicker along the ellipsoid outline of the main spore body and thin surrounding the protruding part of the PCs (Fig. 1c, d). Suture was weakly sigmoid, faint and symmetrical (Fig. 2). PCs were terminal, equal, elongate pyriform, PC axis in medial plane in valvular view, but with an angle reaching c. 47° to spore axis in sutural view. Polar filament coils are perpendicular to PC axis, with filament coil diameter averaging 70% of PC diameter. Coils are not present in the protruding part of the PCs (Fig. 1c–e). Polar capsule length to spore length was 1:2.1–2.8 (mean 2.3 ± 0.2; N = 21). Fully extruded polar filaments reach 134–153 μm in length. Sporoplasm was binucleate, filling the barrel-shaped spore cavity between the PC cells. Spore measurements are given in Table 1.

Description of actinospores from N. pelagica

All actinospores were of the tetractinomyxon type (Fig. 3). They all occurred free in the decaying polychaete body. No pansporocysts were found. The thick-walled actinospores were spherical to slightly ellipsoidal, length 7.0–8.0 (mean 7.6; n = 10) and diameter 6.0–7.5 (6.7). The actinospores were composed of eight cells most easily identified by the presence of their nuclei; the three nuclei of the shell valve cells appeared as small thickenings internally on the spore wall, the three nuclei of the polar capsules and the two nuclei of the sporoplasm. The diameter of the three identical spherical polar capsules was 2.0–2.3 (2.1). It was not possible to provoke extrusion of the polar filaments.

Fig. 3
figure 3

Actinospores of S. sphaerica in naturally infected N. pelagica from northern Øresund, Denmark. Interference contrast, to same scale. a Apical and lateral views of free actinospores. b Lateral views showing the three nuclei of the shell valve cells (arrows) and the two nuclei of the sporoplasm cells (arrowheads). Scale bar 5 μm

Full size image

SSU rDNA sequences

Partial SSU rDNA sequences were obtained from M. sphaericum-infected gall bladders of two B. belone from Denmark (sequence isolates SigBel-1 and 2, GenBank accession nos. JN033225, JN033226) and the studied infection from Norway (sequence isolate Msph, JN033227). These were identical (1696 nt compared).

Two different parts of a N. pelagica infected with tetractinomyxon actinospores produced two identical partial SSU sequences (sequence isolates Npel 1 and 2, JN033228) (1696 nt compared). These sequences were identical with the M. sphaericum sequences from B. belone, apart for four substitutions (99.8 % identity, 1696 nt compared). On this basis the actinosporean infection in N. pelagica is identified with the myxosporean M. sphaericum infecting B. belone.

Blast searches returned Ellipsomyxa spp. as most similar to the partial M. sphaericum SSU rDNA sequences. Phylogenetic analyses on the basis of the SSU rDNA sequences supported a close relationship between genus Ellipsomyxa and M. sphaericum (Fig. 4). Myxidium queenslandicus Gunter and Adlard, 2008 represent a sister group to Ellipsomyxa spp./M. sphaericum in these analyses (Fig. 4). The congeneric marine clade members M. laticurvum, Myxidium incurvatum Thélohan, 1892 Myxidium gadi Georgévitch, 1916 and M. bergense are not closely related to M. sphaericum (Fig. 4). A schematic illustration of the life cycle of Sigmomyxa sphaerica (syn. M. sphaericum, see below) is shown in Fig. 5.

Fig. 4
figure 4

Phylogenetic affinities of S. sphaerica among related members of the marine clade of Myxosporea. S. sphaerica is closest related to Ellipsomyxa spp., and these two genera represent a sister group to M. queenslandicus incertae sedis in a well-supported clade. Other Myxidium spp. in the marine clade are not closely related to S. sphaerica, including M. laticurvum (JN033229, new sequence) and M. bergense from the type host P. virens in Norway (JN033231, new sequence). All new sequences in bold. Clade support values: upper, MrBayes posterior probabilities (in percent); middle, maximum likelihood bootstrap (N = 100) support values (Paup); lower, maximum parsimony (Mega)

Full size image
Fig. 5
figure 5

Schematic illustration of the life cycle of S. sphaerica. The polychaete N. pelagica acts as the invertebrate hosts and the garfish B. belone acts as the fish hosts. a Actinospore, b myxospore. Not to scale

Full size image

Discussion

Taxonomy

M. sphaericum was originally described from B. belone from the Mediterranean (Banyuls) and Atlantic (Vivier) coasts of France (Thélohan 1895), as producing large (15–20 μm long) spores in disporic plasmodia. Noble (1957) identified the species from the gadid fish M. merlangus in Plymouth, and doubted the validity of M. bergense, a species described from the gadid P. virens by Auerbach (1909) and recorded from other gadids including M. merlangius in Norway (Auerbach 1912). Other authors studying the myxosporea of M. merlangus accepted this, which resulted in a very wide conception of M. sphaericum that has caused much taxonomic confusion. MacKenzie and Kalavati (1995) distinguished between M. sphaericum and M. bergense, and considered it possible that the parasite from B. belone is distinct from Myxidium spp. in the gallbladder of gadid hosts. M. sphaericum as redescribed here is clearly differentiated from all Myxidium spp. recorded from gadids in the northeast Atlantic by characters such as the large and elongated polar capsules with a high number of windings, and show phylogenetic affinity to the Ellipsomyxa spp. rather than to the Myxidium spp. with sigmoid spores recorded from gadids.

Considering only records of M. sphaericum from B. belone, there are two records subsequent to Thélohan (1895), both from the Adriatic Sea (Lubat et al. 1989; Mladineo et al. 2009). Lubat et al. (1989) provided spore measurements agreeing well with the present. However, Mladineo et al. reported spore lengths shorter than the width, which suggests these may be erroneous. The polar capsule lengths of Lubat et al. (1989) and Mladineo et al. (2009) are less than the present, but similar measurements were obtained by us when taken in valvular view. Such measurements tend to represent polar capsule span and not length, due to the oblique angle of the polar capsule axes relative to the spore main axis. However, while such considerations may account for differing length observations, the polar capsule width measurements also differ significantly, being less in the Adriatic samples. Hence verification of conspecificity of Adriatic M. sphaericum isolates with the present through SSU rDNA sequencing would be valuable. The light microscope images of M. sphaericum spores and plasmodia presented by Mladineo et al. (2009) compare well with the present material.

The type species in genus Myxidium, Myxidium lieberkuehni Bütschli, 1882, produce large polysporic plasmodia in the urinary system of pike, Esox lucius L., a freshwater fish. Spores develop in disporic pansporoblasts and show striated valves. Phylogenetic analyses on the basis of SSU rDNA sequences place M. lieberkuehni in a ‘freshwater clade’ of Myxosporea, as a member of a ‘M. lieberkuehni clade’, representing a sister group to a ‘Myxobolus clade’ containing mostly members of the Platysporina (see Fiala 2006; Holzer et al. 2007). M. sphaericum is closest related to Ellipsomyxa spp. in the ‘marine clade’ and belongs to the marine ‘Myxidium clade’ of Fiala (2006). Sequence similarity in the aligned SSU rDNA sequences of M. sphaericum and M. lieberkuehni is only 66% (2,057 sites compared). Hence M. sphaericum is not congeneric with M. lieberkuehni and does not belong in the Myxidiidae typified by that species. We therefore propose a novel genus to encompass M. sphaericum and related species.

Sigmomyxa n. gen.

Coelozoic in gallbladder, sporogony disporic, plasmodia are di- to polysporic, spores smooth, spindle shaped in valvular view and sigmoid in sutural view. Valves are ellipsoid in outline, with thin walled protrusions associated with the PC tips. Polar capsules elongate pyriform, with >7 windings. Intercapsular distance is short. Type species is S. sphaerica (Thélohan, 1895)

Comments

Myxidium elmatboulii Ali et al., 2006 and Myxidium maamouni Abdel-Baki, 2009 are similar to S. sphaerica and likely congeners, but molecular data is lacking (cf. Ali et al. 2006; Abdel-Baki 2009). M. elmatboulii is transferred to Sigmomyxa as Sigmomyxa elmatboulii (Ali et al., 2006) comb. n. on the basis of its morphology, the species is so similar to S. sphaerica that conspecificity is possible. The host, Tylosurus choram (Rüppell, 1837) is also related to B. belone (both belonids). The spores of M. queenslandicus appear morphologically similar to those of S. sphaerica, but these species show only 89% identity in the partial SSU rDNA sequences available. However, expansion segments in the V7 region of M. queenslandicus are responsible for the low identity; exclusion of these gives 94% identity with S. sphaerica. M. queenslandicus is phylogenetically closest related to S. sphaerica and Ellipsomyxa spp., but inclusion in Sigmomyxa appears to render the genus polyphyletic. However plasmodia and sporogony of M. queenslandicus are unknown and the species is therefore considered an incertae sedis. M. laticurvum Kabata 1962 (syn. Myxidium trachinorum Canning et al. 1999, see Karlsbakk 2001) show a protruding polar capsules similar to S. sphaerica but differ in containing very prominent capsulogenic cells in mature spores, a convex spore structure and a different organisation of the polar filaments (Kabata 1962; Canning et al. 1999). Our SSU rDNA sequences of M. laticurvum confirm that this species is not closely related to Sigmomyxa n. gen.

The erection of Sigmomyxa n. gen. removes two species from the polyphyletic genus Myxidium Bütschli, 1882. Several species in the marine group of Myxosporea and currently assigned to Myxidium are not closely related to Myxidium sensu stricto or Sigmomyxa n. gen. on the basis of their SSU rDNA sequences, but show a related morphology and development. Redescriptions and revisions of these taxa are needed.

Life cycle

The actinosporean infection in N. pelagica and the myxosporean S. sphaerica in B. belone is considered different life cycle stages due to the high SSU sequence similarity. Sequence identity has aided the disclosure of all the marine myxosporean life cycles known so far. We observed that five specimens of N. pelagica survived and only the infected specimen died due to stress (high temperature and lack of oxygen). This indicates that an infection with actinosporean stages may affect the survival of the polychaete host. Other observations on the effects of actinosporeans on the annelid hosts are scarce. Shirakashi and El-Matbouli (2009) found feeding and fecundity of actinosporean-infected Tubifex tubifex to be reduced, but did not observe reduced survival.

Apparently only fully developed actinospores of S. sphaerica were found in the examined N. pelagica. However, the wall of the pansporocysts and younger developmental stages may have disintegrated in the decaying polychaete host.

The present actinospores differ from those of E. gobii and E. mugilis (as Zschokkella mugilis), which also use Nereis spp. as polychaete hosts, by being nearly spherical contrary to the elongated actinospores of Ellipsomyxa spp. having nearly twice the length (Køie et al. 2004; Rangel et al. 2009).

Nereis diversicolor and Nereis succinea from less than 1 metre depth may be infected with actinosporean stages (Køie et al. 2004; Rangel et al. 2009). The N. pelagica specimens examined were dredged in among other a shallow sandy bay harbouring N. diversicolor and N. succinea. These two species were only infected with E. gobii, even though specimens of B. belone must have spent some time in this bay. Hence it is possible that these myxosporeans display some degree of host specificity to the invertebrate host; E. gobii uses two species of Nereis as invertebrate hosts (Køie et al. 2004), whereas S. sphaerica apparently use one species, N. pelagica.

Actinospores of the tetractinomyxon type are the actinosporean stages of myxozoans belonging to at least three clades: the Ceratomyxa clade (Køie et al. 2008), the Parvicapsulidae (Bartholomew et al. 2006; Køie et al. 2007) and the Sigmomyxa/Ellipsomyxa clade. A fourth clade is represented by Ceratomyxa shasta, which also show tetractinomyxon actinospores (Bartholomew et al. 1997), but the phylogenetic affinities of C. shasta is unclear (see Fiala and Bartosova 2010).