How to build single‐celled cnidarians with worm‐like motility: Lessons from Myxozoa

2.1. Fish sampling

Bile samples were obtained from fish collected in Brazilian rivers (Table 1; Figure S1). Sampling and access to genetic heritage was authorised by the Brazilian Ministry of the Environment (SISBIO authorisations # 44268‐9 and 67616/1 and SisGen # A656D8E). Fish were transported live to a makeshift field laboratory on river shores or boats and then euthanised prior to dissection. The gall bladder of dissected fish was opened, and bile fluid containing suspended plasmodia was removed using a glass pipette. A drop of the bile fluid was then placed on a microscope slide and examined immediately at 40 or 100× to ascertain infections using a light microscope (Carl Zeiss model Primo Star). In a few cases, cover slips were gently added for examination and image was captured at higher magnification. Videos of ceratomyxids in bile were obtained using a digital camera (Sony CyberShot) coupled to the light microscope objective and edited using the VideoPad Video Editor software v. 8.77 (NCH Software – Canberra, Australia). Ambient conditions occasionally displaced specimens in some videos (e.g. rocking of makeshift laboratory on boat).

2.2. Light, electron and confocal microscopy

Formalin‐fixed plasmodia were analysed and photographed using a Carl Zeiss Axio Imager A2 light microscope with differential interference contrast (DIC) equipped with Axio Cam and AxioVision AxioVs 40V4.8.2 software. For transmission electron microscopy, plasmodia were fixed for at least 12 h in 2.5% glutaraldehyde with 0.1 M cacodylate buffer (pH 7.4), washed in the same buffer and post‐fixed with osmium tetroxide (O_sO₄), all procedures being performed at 4°C. After dehydration in an ascending ethanol series, the samples were embedded in EMbed 812 resin (Electron Microscopy Sciences). Ultrathin sections, double stained with uranyl acetate and lead citrate, were examined using a LEO 906 transmission electron microscope operating at 60 kV. For confocal analyses, infected bile fixed in 10% formalin in 0.1 M PBS was introduced for 30 min on polysine pre‐treated slides. The samples were rinsed three times in PBS and then permeabilised with PBS containing 0.1% Triton X‐100 9for 1 h. Specimens were then stained with Alexa Fluor® 488 Phalloidin (Invitrogen Eugene) at 0.001 mg mL⁻¹ for 4 h and with 4′,6‐diamidino‐2‐phenylindole dihydrochloride (DAPI, Sigma‐Aldrich) at 0.004 mg mL⁻¹ for 10 min. The samples were rinsed in PBS and mounted in 90% glycerol, 10% PBS, 0.5% 1,4‐diazabicyclo [2.2.2] octane mounting medium and examined using a Nikon A1 Confocal Microscope (Nikon).

2.3. DNA extraction, polymerase chain reaction (PCR) and sequencing

DNA was extracted from ethanol‐fixed samples. About 50 µl of bile content was pelleted at 11 200 g for 3 min and the ethanol supernatant removed. Total DNA was then extracted from the pellet using a Qiagen DNeasy Kit, following the animal tissue protocol (QIAGEN). The product was eluted in 50 μl Buffer AE and then quantified in a NanoDrop 2000 spectrophotometer (Thermo Fisher Scientific) at 260 nm.

The partial small‐subunit ribosomal DNA (SSU rDNA) was amplified using a two‐round PCR approach. Initial amplification was performed using the universal primers 18E (CTGGTTGATTCTGCCAGT) (Hillis & Dixon, 1991) and 18R (CTACGCAAACCTTGTTACG) (Whipps et al., 2003), followed by a semi‐nested round, which used a combination of both universal and Ceratomyxa‐specific primers: CerSAZ.1f (GTCCCTTCGATCGTAGTACCAC) paired with CerSAZ.3r (CTATCCCCACAGCCTGAAAACT) (this study); MYXATK3f (CATTTGAGGGCGTTAGTACTTG) (Zatti et al., 2018) paired with 18R and/or CerSAZ.4f (GTTGGTTAGTTTCCACGCGAAA) (this study) with 18R.

PCRs were conducted in 25 μl reaction volumes, which were comprised of 12.5 μl Dream Taq Green PCR Master Mix (Thermo Fisher Scientific), 0.5 μl each primer (10 pmol), 1 μl DNA (5–20 ng/μl) and 10.5 μl ultrapure water. PCR cycling was performed on Applied Biosystems® Veriti® 96‐Well Thermal Cycler (Applied Biosystems), using a block preheated to 103°C. PCR amplification consisted of an initial denaturation step at 95°C for 2 min, followed by 35 denaturation cycles at 94°C for 30 s, annealing at 58°C for 30 s and extension at 72°C for 90 s following by a terminal extension at 72°C for 5 min. Amplicons and a 1 kb Plus DNA Ladder (Thermo Fisher Scientific) were visualised following 1.5% agarose gel electrophoresis using Tris–borate–EDTA buffer (0.045 M Tris–borate, 0.001 M EDTA, pH 8.0), staining with SYBR™ Safe (Thermo Fisher Scientific) and analysing on a Compact Digimage System transilluminator (MajorScience™). Amplicons were then purified using the QIAquick PCR Purification Kit (QIAGEN, California, EUA) according to the manufacturer's instructions and directly sequenced using the same PCR primers at 5 pmol plus MC5 (CCTGAGAAACGGCTACCACATCCA) and MC3 (GATTAGCCTGACAGATCACTCCACGA) (Molnár et al., 2002), using a BigDye 102 Terminator v3.1 Cycle Sequencing kit (Applied Biosystems) in an ABI 3730 DNA 103 Analyser (Applied Biosystems) at the Human Genome and Stem Cell Research Center, University of São Paulo.

Any ambiguous bases were clarified with visual reference to the corresponding ABI chromatograms, then were assembled into a single contig in BioEdit 7.1.3.0 (Hall, 1999). We used BLASTn (Altschul et al., 1997) searches against the NCBI nucleotide database for both individuals and contig sequences to verify the correct amplification of Ceratomyxa DNA.

2.4. Phylogenetic analyses

Two SSU rDNA data sets were subject to phylogenetic analyses. One (n = 65 sequences >1000 bp from NCBI data base) included novel Ceratomyxa sequences obtained in this study (6 sequences; Table 1); congeneric sequences from infected Amazonian fish (4 sequences); other ceratomyxids (Unicapsulocaudum mugilum, Myxodavisia bulani and Palliatus indecorus); representatives of the lato sensu marine Ceratomyxa species (31 sequences), and representatives species of other myxosporean genera (21 sequences). Tetracapsuloides bryosalmonae (KF731712) and Buddenbrockia plumatellae (KF731698) were used as outgroup. The second data set (18 sequences) comprised SSU rDNA sequences of Ceratomyxa species parasitic in South American freshwater fishes (11 sequences) and 6 sequences of marine ceratomyxids closely related to the freshwater Ceratomyxa lineage in the larger molecular phylogeny. Sphaeromyxa zaharoni (AY538662) and Sphaeromyxa kenti (JY443489) were used as outgroup. The sequence of the Amazonian species Ceratomyxa fonsecai (GenBank access number MK796248) [Silva et al., 2020] was excluded due to inconsistent alignment.

Nucleotide sequences were aligned with Geneious 7.1.3. Maximum likelihood (ML) heuristic searches were conducted in the PhyML 3.0 online server (http://www.atgc‐montpellier.fr/phyml/) (Guindon et al., 2010), using automatic model selection by (Smart Model Selection), NNI tree searching and topology assessing by bootstrapping with 1000 replicates. Bayesian inference (BI) analysis was conducted using Plataforma CIPRES (Miller et al., 2010), applying the model of evolution GTR + I + G obtained by JModelTest analysis with corrected Akaike information criterion (AIC) (Posada, 2008). Posterior probabilities were estimated from 10 million generations via two independent runs of four simultaneous Markov Chain Monte Carlo (MCMC) algorithms, with every 10,000th tree saved. Trees were visualised using Figtree 1.3.1 [Rambaut, 2008) and edited using Adobe Photoshop (Adobe Systems Inc.).

3.1. Gall bladder infections

We describe the morphology and motility of plasmodia of 10 Ceratomyxa lineages in gallbladders of fish belonging to six families (Table 1). SSU rDNA sequences were obtained for eight lineages and five represent undescribed species (Ceratomyxa sp. 1, sp. 2, sp. 3, sp. 4 and sp. 5) (Table S1). Sequencing was unsuccessful for Ceratomyxa spp. 6 and 7; however, their plasmodia, myxospore characters and host use strongly suggest they are distinct species. Prevalence of infections varied from 6% to 100%, although some are based on very low sample sizes (Table S1). Plasmodia were vermiform, with the exception of one tadpole‐shaped (pyriform) species, and the poles (tips) of plasmodia were either sharp or blunt (Table 1; Figures 1 and 2). There was no overlap in host use. Co‐infections with other myxosporeans were occasionally encountered. Below we describe the morphology of these vermiform ceratomyxids. We then explore the anatomical basis and functions of motility and the origin and subsequent radiation of vermiform ceratomyxids.

3.2. Organisation of plasmodia

Plasmodial organisation and associated motility as variously revealed by ten taxa (Table 1) collectively provide insights about form and function (Figures 1, 2, 3, 4, 5, 6, 7, 8, 9 and Videos S1–S13). Ultrastructural study revealed an outer cytoplasmic region with organelles (detailed below) and an internal organelle‐free and membrane‐bound region that was observed in C. gracillima, C. vermiformis, C. brasiliensis, Ceratomyxa sp. 1, Ceratomyxa sp. 2 and Ceratomyxa sp. 3 (Figure 3a–d; 4a–d; 5a–d; 6a–d; 7a and b).

These collective observations provide evidence that worm‐like plasmodia are characterised by a large vacuole occupying the internal area – a feature not observed in the only previous ultrastructural investigation (Adriano & Okamura, 2017).

Mitochondria, rough endoplasmic reticulum and ribosomes were consistently observed (Table 1; Figure 3a,b; 4a,b, 5a–d; 6a,d,e and 9a,b). Notable cytoskeletal elements included microtubules and microfilaments, which appear to be common to the worm‐like ceratomyxids. Net‐like microtubules are seen in C. vermiformis (Figure 9a), C. brasiliensis (Figures 5d and 9b,c), C. gracillima (Figure 5a) and Ceratomyxa sp. 3 (Figure 5b). Microfilaments were revealed by phalloidin in C. gracillima (Figure 2a,b), in Ceratomuxa sp. 4 (Figure 2c) and in Ceratomyxa sp. 6 (Figure 2d). Ultrastructural analyses showed microfilaments in C. brasiliensis (Figure 9b,c) and in C. gracillima (Figure 7b). Net‐like arrangements of both microtubules and microfilaments were clearly observed in C. brasiliensis (Figure 9b,c). Long tubular mitochondria were regularly arranged around the circumference of the plasmodia in C. gracillima (Figures 3a,b and 5a), C. brasiliensis (Figures 5c,d and 9b), Ceratomyxa sp. 1 (Figures 4a and 6d), Ceratomyxa sp. 2 (Figure 4b) and Ceratomyxa sp. 3 (Figures 5b and 6a). This mitochondrial arrangement, also observed in C. vermiformis (Adriano & Okamura, 2017), appears be another common feature of worm‐like ceratomyxids. The distribution of mitochondria may be linked to the energetics of motility.

A further notable feature of vermiform ceratomyxids is a secreted glycocalyx‐like layer (Table 1; Figure 3b; 6b,d,e and 7b and Figure 2b in Adriano and Okamura (2017)). Glycocalyx layers perform diverse functions including acting as a barrier between a cell and its surroundings, mediating cell–cell interactions, contributing to differentiation, adhesion and motility, and protection from physical forces and stresses (Butler et al., 2020; Goligorsky, 2020; Monne et al., 2012). Light and scanning electron micrographs of formalin‐fixed material show corrugations of the glycocalyx layer (Figure 7c and Figure 6 in Adriano and Okamura (2017). Corrugations can be observed on both sides of the plasmodia as a result of general contraction during fixation (e.g. Figure 7c). However, alternating waves of contraction on either side of the plasmodia can clearly be seen in videos of live specimens. The corrugations develop where plasmodia undergo bending (Figure 8; Videos S1–S3). We suggest that the corrugations observed in live material represent buckling, or kinking, by the stiffer glycocalyx when axially compressed by forces on the side on which there is microfilament shortening.

Pinocytic vesicles and/or channels were evident in the plasmodia of C. vermiformis and C. brasiliensis (Figure 9a,b). Observations of these transitory structures suggest that vermiform ceratomyxids in bile use the same strategy of obtaining food as histozoic myxosporeans (Current & Janovy, 1978; El‐Mansy & Bashtar, 2002; Hallett & Diamant, 2001).

Plasmodia ranging from small early stages to larger mature spore‐forming stages were generally observed in bile (Adriano & Okamura, 2017; Zatti et al., 2017). Distinct spore‐forming regions within the plasmodia (referred to as sporogonic centres) cause deflection of the vacuolar membrane (Figure 3c,d; 4c,d, 6a,c and 7a). A gradient of spore developmental stages arises with young stages situated near the initiation of sporogony and maturing spores radiating from the sporogonic centre. Details of sporogony will be covered in a separate publication.

3.3. How is movement created?

Hydrostatic skeletons are characterised by a constant volume, fluid‐filled internal cavity surrounded by a muscular body wall. They are used by a diversity of soft‐bodied organisms for support and movement (Kier, 2012). In many worms alternating contractions of circular and longitudinal body wall muscles act in opposition, a process that is mediated by incompressible fluid in the body cavity. Nematodes possess four sets of longitudinal muscles whose alternating contractions on either side of the body produce undulatory propulsion mediated by their stiff cuticle, a reinforcing helically wound fibre system and high internal hydrostatic pressure (Kier, 2012).

Worm‐like Ceratomyxa exhibit undulatory motility along the entire length of the plasmodium (Videos S1–S13). Such motility can result in translocation as observed in plasmodia of Ceratomyxa sp. 1 (Video S4), Ceratomyxa sp. 2 (Video S5), Ceratomyxa sp. 7 (Video S6) and C. vermiformis (Adriano & Okamura, 2017). Undulation originating at a plasmodial pole induces movement in that direction (Ceratomyxa sp. 1 – video S4; Ceratomyxa sp. 2 – Videos S5 and S7 and Ceratomyxa sp. 7 – video S6). We observed undulation generated at both plasmodial poles (whether sharp or blunt) in Ceratomyxa sp. 2 (Video S5) and Ceratomyxa sp. 7 (Video S6) indicating a lack of polarity in movement at least in these taxa. Undulation was observed to be initiated only at the blunt pole of Ceratomyxa sp. 1 (Video S4). Occasionally plasmodia seem to be attached to the slide when introduced with bile fluid (Videos S1 and S2). However, that is just an artefact result of the thin layer of bile formed after the deposition of the drop on the slide, once in their natural environment they are freely undulating in bile fluid and indeed this enables them to be drawn up into a glass pipette.

Ceratomyxid undulatory motility is linked with a plasmodial organisation that includes an outer cytoplasmic region rich in cytoskeletal elements surrounding an extensive internal vacuole as described above. We hypothesise that coordinated activities of peripheral microfilaments result in alternating waves of deformation that proceed along the length of vermiform plasmodia (Videos S1–S3; Figure 8). We propose that the fluid‐filled vacuole transmits forces to the opposing side of the plasmodium thus effecting bending (Videos S1–S13). According to this scenario, the extracellular glycocalyx appears to function like the nematode cuticle, providing stiffness that reinforces the system, preventing the development of aneurisms and maintaining the general vermiform shape. The walls of nematodes (and other animal hydrostats) are typically associated with a collagenous crossed fibre helical array that also contributes to control and limit shape change (Kier, 2012). The net‐like arrangement of microtubules below the glycocalyx may act similarly at the cellular level in vermiform ceratomyxids. According to our scenario, the peripherally organised cytoplasmic elements (especially the net‐like microfilaments), the large internal vacuole, and the relatively rigid glycocalyx collectively enable convergence at the cellular level to a hydrostatic skeleton by worm‐like plasmodia. Undulatory ceratomyxid motility is generated in a strikingly different way from the helical swimming of their distant myxozoan relatives, the vermiform malacosporeans B. plumatellae and T. vermiformis. It also contrasts with the actin‐myosin‐based motility described for the elongated unicellular apicomplexans (parasitic protists) (e.g. tachyzoites of Toxoplasma gondii, trofozoites of gregarines and sporozoites and ookinetes of Plasmodium spp.), which typically glide across surfaces using complex motor machinery (Kováciková et al., 2017; Matuschewski & Schűler, 2008). The outer region of the apicomplexan cell is characterised by an extensive and highly structured inner membrane complex (IMC) below which lies a subpellicular network (SPN) containing net‐like protein filaments, microtubules, actin and myosin (Matuschewski & Schűler, 2008; Morrissette & Sibley, 2002). The glideosome model of apicomplexan movement involves complex interactions of SPN components and the IMC to generate movement (via surficial adhesions and static myosin motors) and to maintain shape as apicomplexans move into and through hosts (Matuschewski & Schűler, 2008). As far as we are aware, the cellular undulatory motility inferred here for vermiform ceratomyxids involves the unique deployment of subcellular components for a novel purpose in unicellular organisms – pressure transmission via an enlarged central vacuole to enable bending. However, both apicomplexans and vermiform ceratomyxids share net‐like arrangements of proteins that likely function to control cell shape.

3.4. Function of motility

The function of motility in vermiform plasmodia is a mystery. Adriano and Okamura (2017) speculated that motility may prevent premature release of immature forms with the bile. This is supported by observations of reduced motility in spore‐filled stages of Ceratomyxa sp. 1 (Video S9) that may promote exit from the gall bladder, excretion from fish and transmission to worm hosts. However, reduced motility was not observed in spore‐filled stages of Ceratomyxa sp. 2, Ceratomyxa sp. 5 and C. gracillima (Videos S10–S12). Videos demonstrate variation in motility amongst lineages that is unrelated to size, with some characterised by relatively slow and others relatively fast undulations (Table 1; Videos S1–S13). An alternative function of motility could be to facilitate acquiring nutrients. By increasing the flow of gall bladder fluid across plasmodial surfaces bending movements may increase the flux of nutrients and subsequent uptake. If this is a primary function, then plasmodia demonstrating reduced motility may develop more slowly, have a lower nutritional demand or may simply be responding to nutrient concentrations. Although motility can result in translocation, most plasmodia tended to remain relatively stationary while undergoing undulations when observed on microscope slides. Whether this is an artefactual wall effect is unknown. Further investigation is required to determine how size, shape and bending rates (undulation frequency) of plasmodia may relate to bile properties such as rate of discharge from the gall bladder and fluid composition or density.

3.5. The radiation of vermiform ceratomyxids

Several key cnidarian traits may have enabled the extensive radiation of endoparasitic myxozoans and further radiations within this clade, including the vermiform ceratomyxids. These traits include exceptional capacities for developmental plasticity, regeneration and transdifferentiation (Bosch, 2008; Leclère & Röttinger, 2017; Piraino et al., 2004), a relatively simple body plan based on epithelial sheets (ectoderm and endoderm), and the principal cnidarian building block (epitheliomuscular cells). Developmental plasticity and the diploblast body plan may have predisposed the evolution of endoparasitism and miniaturisation associated with endoparasitism in the Myxozoa (Okamura et al., 2015b). In the tissue‐forming myxoworms epitheliomuscular cells have been modified by displacement, with the development of four longitudinal smooth muscles that are disassociated from epithelia (Gruhl & Okamura, 2012; Leclère & Röttinger, 2017). We propose that further modification of the fundamental cnidarian building block appears to have promoted a unique radiation of cellular hydrostats in the derived vermiform ceratomyxids. This modification includes the deployment of myofilaments in the peripheral regions of a single‐celled plasmodium. Such modifications of the epitheliomuscular cell in endoparasitic cnidarians demonstrate the variable development and localisation of muscle cells in cnidarians (Jahnel et al., 2014; Leclère & Röttinger, 2017) and support the view that multifunctionality is an ancestral characteristic of epitheliomuscular cells (Arendt, 2008).

All vermiform ceratomyxids known to date are from South American freshwater environments. Our molecular phylogenetic analyses reveal they comprise a monophyletic group of taxa that vary in plasmodial shape and size (Clade 1 – Figure S2; Figure 10; Table 1). This South American freshwater lineage is sister to Unicapsulocaudum mugilum (Figure 10), a parasite of a euryhaline cosmopolitan fish (Mugil cephalus), and together these taxa are sister to a small clade of marine ceratomyxids from Mediterranean Sea. The early‐diverging M. bulani parasitises an amphidromous fish (Megalops cyprinoides) in this freshwater/marine ceratomyxid lineage (Figure 10). Other Ceratomyxa species grouped in an exclusively marine lineage (Clade 2 – Figure S2).

The lack of evidence of vermiform ceratomyxids from other continents suggests that this unique radiation may derive from marine transgressions in the Late Cretaceous (83–67 Ma) to Late Miocene (~10 Ma) in South America (Lundberg et al., 1998; Räsänen et al., 1995) as has been proposed for various other taxa including Mammalia (dolphins and manatees), Chondrichthyes (stingrays), Osteichthyes (sciaenid fishes), Crustacea (shrimp, crabs) and Mollusca (Bloom & Lovejoy, 2017; Lovejoy et al., 2006; Lundberg et al., 1998; Webb, 1995; Wesselingh et al., 2002). This scenario would be challenged if vermiform ceratomyxids were discovered in other biogeographic regions. The hypothesis of a marine origin for this freshwater Ceratomyxa lineage (Zatti et al., 2017, 2018) is consistent with the evidence for sharing a common ancestor with a parasite of a euryhaline fish host (Figure 10; Figure S2). Further research may help to reveal whether marine incursions promoted the invasion of South America by ancestors of vermiform ceratomyxids.