Morphology, ultrastructure, and molecular phylogeny of the ciliate Sonderia voraxwith insights into the systematics of order Plagiopylida

General remarks

The neotype population of Sonderia vorax was discovered in three brackish water samples with a 4–8% salinity range; together with oxygen level (see below), salinity was measured using an OX 22 oxygen meter (Aqualytic, Langen, Germany). All the samples came from the same brackish water pond referred to as “Stagno 1” placed on the coastline of Ligurian Sea, close to Serchio River mouth, Pisa district (Tuscany, Italy) (43°47’39” N, 10°16’4” E), and were collected during October 2005, with a water temperature ranging from 18 to 22°C. The samples also contained ciliates such as Sonderia pharyngea, Plagiopyla sp., Copemetopus sp., and Metopus sp. in moderate or low abundance. In the sediment layer, where S. vorax was mainly discovered, the oxygen level in water was 1–7%; close to the water surface it was 35–66%. Attempts to cultivate in laboratory S. vorax were unsuccessful under full oxygen conditions. The ciliates survived in closed tubes within the original samples for a week, and, sometimes, even longer; thus, all investigations were performed on the specimens of the non-clonal neotype population of the original pond, taken from all of the three collected samples.

Live observations

Live ciliates were observed for morphological details using differential interference contrast (DIC) microscopy with a Leitz (Weitzlar, Germany) microscope at a magnification of 300–1250 × with the help of a compression device [26]. For examination of the swimming behavior, ciliates were observed in a glass depression slide (3 ml) under a dissection microscope (Wild M3, Switzerland) at a magnification of 12.5–50 ×.

Fixation and staining

Ciliates were fixed with Champy’s solution [27] and then silver nitrate-stained according to Corliss, 1953 [28]. Feulgen staining procedure after fixation in Bouin’s fluid [27] was used to reveal the nuclear apparatus.

Cell image capturing and measurements

Computer images were captured from appropriate preparations with a digital camera (Canon PowerShot S45), automatically saved as files during optical observation at a magnification of 500–1250 ×, and used to obtain measurements of living and fixed ciliates.

Schematic line drawings were based on micrographs of typical living and impregnated cells.

Electron microscopy

Scanning electron microscope (SEM) and transmission electron microscope (TEM) preparations were obtained as described in Modeo et al.[29] except for: 1. cell preservation in 2% (w/v) OsO₄ in distilled water for SEM procedure; 2. use of 2.5% (v/v) glutaraldehyde in 0.1 M cacodylate buffer, pH 7.4, for TEM fixation.

Fluorescence microscopy

Fluorescence microscopy to check the possible autofluorescence of cells due to the presence of methanogenic symbionts was used [30, 31]. Specimens were fixed either in 4% (v/v) formaldehyde in PBS or in 2% (w/v) OsO₄ in distilled water, and then observed at the following wavelengths: ~ 495 nm, ~ 550 nm, and UV, with both a Zeiss AxioPlan fluorescence microscope (Carl Zeiss, Oberkochen, Germany) equipped with a HBO 100W/2 mercuric vapor lamp, and a Leica DMR microscope (Leica, Switzerland) equipped with a Osram 50 W/AC L2 mercuric vapor lamp. With the latter microscope, computer images were captured from appropriate preparations by means of a dedicated software called IM1000, version 1.0.

To roughly classify ectosymbionts and possible endosymbionts harbored by S. vorax double fluorescence in situ hybridization (FISH) experiments were performed according to Ferrantini et al.[32]; the oligonucleotidic probes EUB338 5^′-GCTGCCTCCCGTAGGAGT-3^′[33], targeting most of Eubacteria, and Arc915R 5^′-GTGCTCCCCCGCCAATTCCT-3^′[34], specific for Archaea were used.

18S rRNA gene sequence obtainment

Approximately 50 organisms were individually harvested from the original sample and carefully washed three times in sterilized distilled water in order to minimize contaminations from the original medium. The washed cells were fixed in ethanol 70%. Total genomic DNA was isolated with the NucleoSpin™ Plant II DNA extraction kit (Macherey-Nagel) and stored at −20°C in aqueous solution.

A polymerase chain reaction (PCR) was performed with a Primus 96 plus thermal cycler (MWG-Biotech AG) employing the TaKaRa Ex Taq (TaKaRa Bio Inc.) (forward primer: 18S F9 Euk [35]; reverse primer: 18S R1513 Hypo [36]; annealing temperature: 50°C). The PCR products were sequenced in both directions using three internal primers as in Rosati et al.[37]. The three partially overlapping sequences were compared to each other and assembled.

Sequence availability and phylogenetic analyses

The characterized sequence is available under the accession number [EMBL: HF547270].

The 18S rRNA gene sequence of the ciliate was aligned against those available in the SILVA 108 database [38] using the Fast Aligner algorithm of the ARB software package [39]. The alignment was then manually edited in order to optimize base-paring in the predicted rRNA stem regions. For phylogenetic analyses, gaps were coded as a fifth character, while missing data were discarded. Columns containing only one non-gap character were also discarded. The final character matrix contained 40 sequences (29 from the class Plagiopylea and 11 from the class Prostomatea as outgroup) and 1345 columns. The evolutionary model that fits best the data was selected according to the AIC parameter as calculated by jModelTest [40, 41]. The TREE-PUZZLE [42] Likelihood Mapping function was employed in order to check the informational content of the data.

Phylogenetic analyses were performed with Maximum Likelihood (ML) and Bayesian Inference (BI) methods. The software PHYML [40] as provided by ARB was employed for ML, producing 1000 pseudoreplicates for bootstrapping. MrBayes 3.1.2 [43] was employed for BI, using three different runs with one cold and three heated chains each, running for 1,000,000 generations.

The tree topology was also compared against those obtained from modified character matrices. These were generated: (1) retaining only columns with at least one non-gap character conserved in at least 30% of the sequences (modified matrix 1); (2) additionally deleting all columns containing gaps (modified matrix 2); (3) removing the sequences of uncultured organisms (modified matrix 3).

General morphology

Cells are ovoid-ellipsoid in shape with anterior and posterior ends almost equally curved (Figures 1, 2B, 3A, 3B, 4, 5B, 5D). The body is dorsoventrally flattened. In vivo dimensions are ~ 100–150 × 50–75 μm (130 × 69 μm on average); dimensions after fixation in Champy’s solution are ~ 80–130 × 45–70 μm (~ 113 × 65 μm on average) (Table 1). After SEM treatment, cell dimensions are ~ 89 × 42 μm on average (Figures 5A, 5B). The cell surface is uniformly ciliated with 45–62 somatic ciliary rows (~ 54 on average). On the dorsal side 25–31 rows (~ 28 on average) run parallel to each other extending to the posterior end of cell (Figures 3B, 4B, 5B). The dorso-lateral striated band at the right cell margin is visible with DIC microscope as well as on impregnated specimens, arising near the right side of the oral cavity cleft and terminating near the posterior end of the cell (Figures 3B, 4). This structure, that can be considered as a border between dorsal and ventral sides (Figures 1A, 1D, 3A, 3B, 3D, 4), is ~ 2 μm high at SEM, with ~ 0.5 μm-spaced out, ridge-like lamellae ~ 0.15 μm in thickness each (Figure 5F). A single contractile vacuole, apparently without collecting canals, is located in the posterior part of cell and opens on the dorsal side (Figure 1E); its pore was not clearly impregnated with silver staining procedure.

On the ventral side, the oral cavity opening is subapically located as a cleft orientated perpendicularly to the main body axis (Figures 1A, 1B, 3A, 5A). The oral ciliature arises from and is in continuity with the somatic ciliature (Figures 5A, 5C). It runs on the dorsal side (upper oral lip) at first perpendicularly to the front of oral cleft; then, it deviates under some angle to the left reaching the oral cavity’s deepest point (Figures 1A, 1C). The oral ciliature of the lower oral lip consists of kineties perpendicularly inserted with respect to the upper oral lip kineties and forms single ciliary rows with a membranelle-like appearance at SEM (distance between two ciliary rows: ~ 0.3 μm; length of cilia: ~ 5 μm) (Figure 5C). The oral ciliature is represented by 25–30 prebuccal (on the upper oral lip) and 18–20 postbuccal (on the lower oral lip) densely packed kineties. The depth of oral cavity is always not more than 1/3 of body length. On the ventral side, the ciliate kinetom (20–31 ciliary rows, ~ 28 on average) consists of two distinct parts: the ventro-lateral kineties (8–13 ciliary rows, ~ 11 on average), which are continuous along the cell body, and the ventro-frontal kineties (12–18 ciliary rows, ~ 17 on average), which are not continuous (Figure 4A); they start after the membranelle-like ciliary rows and end posteriorly where they meet the left ventro-lateral kineties, forming the so-called ventral secant system [9]; 4–5 rows of the latter group start from the left margin of oral cleft (Figures 3A, 4A).

Many ~ 20 μm long, slightly curved, needle-shaped extrusomes are present in the cortex (Figures 2C-E). They are mainly distributed around the oral cavity opening, but can be found in any part of the cortex and in the cytoplasm. During ejection they appear as long filaments (length at SEM: ~ 17 μm) (Figure 5D). A single quite large micronucleus (diam: 5.4 μm on average) of the compact type is situated nearby or inside the depression of the almost spherical macronucleus (27 × 32.5 μm on average) (Figure 2F). The cell surface is covered by a layer of slightly curved, rod-shaped ectosymbiotic bacteria (size at SEM: ~ 1.5-3.0 × 0.3-0.5 μm), arranged in parallel rows along interkinetal spaces (interkinetal space thickness at SEM: ~ 2.3 μm) (Figures 1E, 3C, 3D, 5C, 5E) except for striated band (Figures 3D, 5F). At SEM observation no gelatinous or mucous coating between the layer of bacteria and ciliate surface was detected (Figure 5).

Notes of behavior

Specimens of Sonderia vorax rotate on the main body axis always anticlockwise (i. e. cells are left spiral swimmers). This species inhabits brackish water sites with oxygen deficiency (oxygen level 1–7%) and is mainly a consumer of diatoms and other algae.

TEM observation

Fluorescence microscopy observation

Cells observed after treatments do not autofluoresce (data not shown). Neither endo- nor ectosymbiotic bacteria are labeled by archeal specific probe Arc915R (data not shown). Ectosymbionts are marked by the universal eubacterial probe EUB338 (Figures 3C, 3D): since they cover most of the cell surface, it is not possible to discriminate signals arising from endosymbionts possibly labelled by the same probe.

Phylogenetic analysis

The ML tree is shown in Figure 10. With one significant exception discussed below, the topology of the ingroup is almost identical in all trees calculated.

Inside class Plagiopylea, the clade containing Epalxella antiquorum (Odontostomatida, Epalxellidae) and four related environmental sequences is the sister group of a major cluster containing all the taxa of the order Plagiopylida. These are distributed in the Trimyemidae clade with three morphospecies of the genus Trimyema, the Plagiopylidae clade including the genera Plagiopyla (non monophyletic) and Lechriopyla, and the Sonderiidae clade with the sequences of S. vorax and Parasonderia vestita. All the aforementioned clades also include environmental sequences from freshwater and marine environments, either suboxic or anoxic. The five sequences most closely related to that of S. vorax were obtained from the supersulfidic and anoxic Framvaren Fjord (Norway).

The monophyly of the families Plagiopylidae and Trimyemidae is well supported. The status of family Sonderiidae is more dubious. It appears monophyletic when trees are calculated either on the unmodified character matrix or the modified matrix 1, although with low statistical support (62/0.75 and 60/0.75 respectively); the support is higher (97/1.00) when sequences from uncultured organisms are discarded (modified matrix 3). In trees calculated on the modified matrix 2, the P. vestita sequence and the closely related environmental sequences AB505461 cluster with Plagiopylidae instead, again with low support (72/0.76).

Identification of our population as Sonderia voraxKahl, 1928 and comparison with related species

Up to now 10 species of Sonderia have been described, most of them by Kahl (8 species) [1, 11]. Four of the originally described Sonderia species were transferred to three different genera [15]: Oncosonderia (Sonderia tubigula); Parasonderia (Sonderia cyclostoma and Sonderia kahli); Kahlisonderia (Sonderia mura). Currently the genus consists of seven species, but the species composition is definitely different with respect to that originally proposed by Kahl. An additional species, Sonderia vestita (Parasonderia vestita according to Xu et al.[16]), still has an uncertain position. The main features of that ciliate do not resemble those of Parasonderia kahli, designated as the type species for its genus [14, 15], so that Jankowski [15] proposed to keep it as S. vestita following Kahl [11]. Finally, the new species Sonderia paralabiata[44] has never been properly described; the only distinctive feature provided by the authors designates the arrangement of kinetids in groups of 3–5 kinetosomes. This is in contradiction with the later statement of Lynn [12] that “the somatic kinetids are monokinetids in the sonderiids, plagiopylids, and trimyemids”. Nevertheless, 3 other sonderiids – S. vestita, Sonderia labiata, S. paralabiata, and maybe Sonderia sinuata, with di- and even up to pentakinetids in the somatic ciliature have been reported [4, 5, 16, 44].

S. vorax can be easily separated from some of the similar-sized species, S. pharyngea and S. labiata, because these ciliates do not show a differentiation of ventral kinetom, the so-called secant system [9] (Table 2). On the contrary, this feature is probably shared between S. vorax and S. sinuata[9], although it has not been indicated neither by Dragesco & Dragesco-Kernéis [4] nor by Al-Rasheid [45]. Anyway, S. vorax differs from S. sinuata for: cell size (180–240 μm vs. 100–150 μm); buccal cavity size (1/2 of cell body vs. 1/3 of cell body); number of dorsal kineties (45 vs. 25–31); size (and probably structure) of the micronucleus [4, 9] (Table 2). The oral ciliature consists of monokinetids in S. vorax while some contradictions do exist in the literature for S. sinuata[4, 9].

General remarks on SEM and TEM analyses

Papers dealing with SEM and TEM observation on the complete cell structure of order Plagiopylida are respectively lacking and scarce. Fine structure descriptions of representatives of family Sonderiidae are limited to the data on S. vorax and Sonderia sp. reported by Fenchel et al.[46], in a study on the interaction between those ciliates and other marine organisms (the so-called “sulphide fauna”) and their prokaryote symbionts. As general information about the determination of their species are lacking, on the basis of the few TEM pictures and data supplied, we could just state the congenerity of our species with S. vorax studied by those authors, but nothing could be argued concerning their conspecificity. Beside the comparison with TEM data reported by Fenchel et al.[46], we also took the opportunity to shed light on general ultrastructure of Plagiopylida; thus, we made a larger comparison with available data on plagiopylid genera such as Lechriopyla and Plagiopyla[30, 47–53], as well as on the single trimyemid genus Trimyema[19, 23, 51, 54, 55].

SEM observation and ectosymbionts

Ectosymbiotic bacteria somewhat covering the cell surface of ciliates have been widely reported (e.g. [56]). Those borne by the hypotrich Euplotidium spp., referred to as epixenosomes [57, 58], are peculiar extrusive symbionts nearly identical at the ultrastructural level to the spherical episymbiotic bacteria of the euglenozoan Bihospites bacati[59], which lives in oxygen-poor habitats. The latter species also bears rod-shaped ectosymbiotic bacteria; these appear to be widespread in protists living in oxygen-poor habitats, as they have been also described in both flagellates such as Calkinsia aureus[60] and Postgaardi mariagerensis[61] and ciliates such as Parablepharisma spp., Metopus spp., and Sonderia spp. [6, 8, 11, 46, 62, 63].

In previous papers on Sonderia spp. the presence of a gelatinous coat between the ectosymbiotic bacteria and the ciliate plasma membrane was either reported as clearly visible under light microscope [5, 9] or at least supposed [46]. Although ectosymbiotic bacteria covering our species appeared partly aggregated on the slide when detached from the ciliate (Figure 5D), a gelatinous coat between ectosymbiotic bacteria and plasma membrane was not evidenced by SEM, while by TEM only a slightly dense layer of material in a few occasions was observed (see below). This result, an apparent discrepancy between SEM and TEM observation, is also evident in pictures of other papers dealing with protists living in oxygen-poor habitats that bear ectosymbionts underlined by a glycocalyx on their cell surface [59, 60].

Curiously, the presence of ectosymbionts as a common feature of the Sonderiidae was not even mentioned in two of the most solid recent ciliate reviews [12, 13]. However, this could be in our opinion a good morpho-biological feature to discriminate members of Sonderiidae family. The ectosymbionts of S. vorax were marked by the universal eubacterial probe in FISH experiments; this indicates their affiliation to Eubacteria.

The height of the striated band on the right surface measured at SEM in S. vorax fits that reported by Lynn [12] in the general description of somatic structures of the families Plagiopylidae and Sonderiidae; thus, we confirmed the size of this structure in sonderiids, but the meaning of this peculiar cortex feature still remains unknown.

TEM observations and endosymbionts

Phylogeny

Our inference partially differs from that of Xu et al.[16] about the position of Parasonderia vestita. It was considered very closely related to Plagiopylidae in their article; in most of our trees it is instead more closely related to S. vorax and its associated environmental sequences. Both their result and ours are poorly supported, and we have demonstrated that the position of P. vestita can change according to how the character matrix is built. Thus, we prefer to leave the question of the Sonderiidae monophyly open, until more data will be obtained and a reliable topology could be inferred. The monophyly of genus Plagiopyla is not recovered, because the sequence of Lechriopyla mystax is nested inside those of the former genus. In none of the papers providing sequences of Plagiopylidae representatives a detailed morphological analysis of the studied organisms is supplied as well. Thus, the possibility that some of them were simply misidentified cannot be excluded. However, a more likely explanation is that the classical morphological characters employed to distinguish between these taxa are not good indicators of evolutionary relatedness, because either plesiomorphic or too vague. It is important to notice, however, that the morphologically similar Plagiopylidae and Sonderiidae do form a robust clade in molecular phylogeny: this clade is clearly separated from Trimyemidae, which members possess a distinctively different morphology.

There are many environmental sequences available in the databases that clearly cluster with characterized plagiopylean species. It is remarkable that all these sequences come from anoxic or suboxic, and often sulfidic, environments. S. vorax-like sequences were apparently obtained only from the Framvaren Fjord (North Sea, Norway) in two different studies [74, 75]. Interestingly, the North Sea is also one of the places where Kahl originally collected ciliates of the genus Sonderia.

Sonderia voraxKahl, 1928

1928 Sonderia vorax – Kahl, Archiv Hydrobiol 19:93–95, Figs 20a-c. [1]

1931 Sonderia vorax – Kahl, Tierwelt Dtl 21:269, Fig. 14. [11]

1934 Sonderia vorax – Kirby, Archiv Protistenknd 82:116. [7]

1969 Sonderia vorax – Fenchel, Ophelia 6:1–182. [6]

1972 Sonderia vorax – Borror, Acta Protozool 10:29–71. [2]

1992 Sonderia vorax – Carey, Marine Interstitial Ciliates 103–104, Fig. 359. [76]

Diagnosis of neotype material

Body dimensions in vivo: ~ 130 × 69 μm (on average). Outline body shape ovoid-ellipsoid with rounded ends, flattened up. On average 56 ciliary rows with ventral kinetom differentiated into two parts (ventral secant system): the ventro-frontal part (~ 17 rows on average) which breaks off posteriorly where it meets from the left the continuous ventro-lateral part (~ 11 rows on average at the ventral surface left). Kineties composed of monokinetids placed at the top of cortical ridges. Depth of oral cavity never more than 1/3 of body length. Oral ciliature arising in continuity with somatic ciliature; oral ciliature of the lower oral lip (18–20 postbuccal rows) consists of kineties perpendicularly inserted with respect to the upper oral lip kineties (25–30 prebuccal rows) and forms single ciliary membranelle-like rows. Dorso-laterally striated band arising near the right side of oral cavity and terminating near cell posterior end. A number of long needle-shaped extrusomes in body cortex visible under light microscope; a second smaller type visible only by TEM. A single contractile vacuole in the posterior part of the dorsal side of the cell. One quite large micronucleus of the “compact” type situated in the depression of the nearly spherical macronucleus. Different hydrogenosomes-endosymbiotic bacteria assemblages distributed throughout the cytoplasm, often in the neighbourhood of endoplasmic reticulum elements. Cell surface completely covered except for striated band by a layer of slightly curved, rod-shaped ectosymbiotic bacteria, arranged in parallel rows along interkinetal spaces. The ciliate rotates about main body axis always anticlockwise (left spiral swimming) and inhabits brackish water sites with oxygen deficiency (level 1-7%).

Neotypification and neotype material

No useable type material (type or voucher slides) is available so far from any of Sonderia vorax populations [15, 17]. The original description [1] is incomplete and apparently based on living observations only. Thus, it seems wise to define S. vorax by the designation of a neotype [77, 78]. Validation of the neotype according to Article 75.3 of the ICZN [77] is justified by the following particulars: (i) the systematic status of S. vorax (it was considered as valid species after Kahl [1], but the description has never been improved according to a modern set of morphological methods); (ii) the differences between S. vorax and related taxa (see Discussion and Table 2); (iii) the neotype specimens (Figures 3A, 3B) representing neotype population from the Ligurian coastline pond (Pisa district, Tuscany, Italy) are described in details (see above); thus, recognition of the neotype designated is ensured; (iv) it is generally known that no type material is available from species described by Kahl; (v) there is strong evidence that the neotype is consistent with S. vorax as originally described by Kahl [1]; (vi) however, the neotype does not come from a site very near to the original type locality (Oldesloe salt marshes, Hamburg region, Germany). Neotype population of the ciliate was found in the middle part of Ligurian Sea (coastline pond nearby Serchio river mouth, Tuscany, Italy), roughly distance: ~ 1000 km; however, both sites are brackish water. Most ciliates, especially marine ones, are cosmopolitans [79], hence this point should not be over-interpreted. A detailed description of the new type locality, that is the sample site of the neotype population, is given in Material and Methods; (vii). One neotype slide of silver nitrate-impregnated specimens (slide № S-11), collected from the pond in Ligurian coastline, Pisa district, Tuscany, Italy, (sampling date 05 October 2005; collector S. I. Fokin), one permanent Feulgen staining preparation (slide № S-17), and Epon-embedded material for TEM investigation have been deposited in the collection of the Museo di Storia Naturale e del Territorio dell’Università di Pisa, Calci (PI), Italy. Two further neotype slides of silver nitrate-impregnated specimens (slides № S-12 and S-14) have been deposited in the slide collection of the Laboratory of Invertebrate Zoology, Biological Research Institute, St. Petersburg State University, St. Petersburg, Russia.

Neotype locality

Owing to the neotypification, the sampling site of the neotype population is the new (valid) type locality of Sonderia vorax: brackish water pond along the coastline of Ligurian Sea close to Serchio River mouth (Pisa district, Tuscany, Italy; 43°47^′16″N, 10°16^′02″E).

Etymology

The derivation of the species-group name is in the original description by Kahl [1].

Gene sequence

The 18S rRNA gene sequence of S. vorax is available under the accession number [EMBL: HF547270].