|Home | About | Journals | Submit | Contact Us | Français|
Neokeronopsis (Afrokeronopsis) aurea nov. subgen., nov. spec. was discovered in soil from the floodplain of a small river in the Krueger National Park, Republic of South Africa. Its morphology, ontogenesis, and 18S rDNA were studied with standard methods. Furthermore, we supplemented the data on N. (N.) spectabilis by reinvestigating the preparations deposited in the British Museum of Natural History. Neokeronopsis (Afrokeronopsis) aurea is a very conspicuous ciliate because it has an average size of 330 × 120 μm and is golden yellow due to the orange-coloured cytoplasm and citrine cortical granules. Further main characteristics include the semirigid body; the urostylid cirral pattern with a distinct corona of frontal and pseudobuccal cirri both originating from the midventral rows; multiple anterior fragmentation of dorsal kineties 1–3; multiple posterior fragmentation of kinety 3, commencing with an unique whirl of kinetofragments; three caudal cirri; an oxytrichid/cyrtohymenid oral apparatus with polystichad paroral membrane and buccal depression; a single oral primordium developing along the transverse cirral row; and an oxytrichid 18S rDNA. These peculiarities are used to establish the new oxytrichid family Neokeronopsidae, the new subgenus Afrokeronopsis, and the new species N. (A.) aurea. Further, these features confirm the CEUU hypothesis, i.e., convergent evolution of a midventral cirral pattern in urostylid and oxytrichid hypotrichs; additionally, N. (A.) aurea is the first (semi)rigid hypotrich with cortical granules and the second one with midventral rows, breaking the granule and flexibility dogmas. These and other observations show that the phylogeny of the hypotrichs is full of convergences. Thus, only a combined effort of classical and molecular phylogeneticists will provide the data needed for a natural classification. Based on the CEUU hypothesis, the molecular data, and literature evidence, we suggest that midventral oxytrichids should be ranked as distinct families; accordingly, we establish a further new family, the Uroleptidae, which forms a distinct clade within the oxytrichid molecular trees. Neokeronopsis is possibly related to Pattersoniella because it has the same special mode of forming the buccal cirri and possesses a buccal depression found also in Steinia, a close relative of Pattersoniella. The large size and conspicuous colour make N. (A.) aurea a biogeographic flagship possibly confined to Africa or Gondwana, while Neokeronopsis (N.) spectabilis (Kahl, 1932) is an Eurasian flagship.
Hypotrichs have fascinated already the ancient protistologists, such as Ehrenberg (1838), Stein (1859), Bütschli (1889), and Wallengren (1900), because the distinctness of the cirri makes it possible to recognize species-specific patterns and to follow pattern ontogenesis without complicated staining procedures. Indeed, hypotrichs were among the first ciliates for which reliable investigations on ontogenesis have been reported (Stein 1859, Wallengren 1900). When silver methods revolutionized ciliate research, the hypotrichs were again among the first for which detailed data became available because they were more easily to impregnate than many other ciliates (Tuffrau 1960, Tuffrau et al. 1968).
Recently, Berger (1999, 2006) monographed part of the hypotrichs, showing the great knowledge that accumulated on their morphology, ontogenesis, and biology. Furthermore, molecular data became available for most main groups (~ families) and some species complexes (Berger 2006, Schmidt et al. 2006). In spite of this, evolution and classification of the hypotrichs remained highly controversial, both at morphologic and molecular level (for reviews, see Berger 1999, 2006; Foissner et al. 2004, Schmidt et al. 2007). One of the most disappointing discrepancies between classical and molecular phylogenies concerned the occurrence of typical midventral hypotrichs (urostylids with two rows of zigzagging cirri in midline, e.g., Urostyla grandis) among the oxytrichids, a large assemblage of species with a highly characteristic pattern usually consisting of 18 fronto-ventral-transverse cirri (18 FVT cirral pattern, e.g., in Stylonychia mytilus).
This fundamental disagreement in morphologic and molecular phylogenies stimulated the CEUU hypothesis (Convergent Evolution of Urostylids and Uroleptids) which proposed that the urostylid midventral pattern evolved from an oxytrichine ancestor, developing a second time within the Oxytrichidae (Foissner et al. 2004). Foissner et al. (2004) could not provide a definite morphologic proof for the CEUU hypothesis because this would have required a hypotrich having the following attributes: a midventral cirral pattern; dorsomarginal kineties; fragmentizing dorsal kineties, preferable posterior fragmentation of kinety 3; and an oxytrichid small subunit (18S) ribosomal RNA sequence. Fortunately, such ciliate has been discovered now, viz., Neokeronopsis (Afrokeronopsis) aurea, which will be described in great detail in the present paper.
Our study shows the validity of the CEUU hypothesis, i.e., that even “very strong” morphologic features evolved convergently. Indeed, evolution of the hypotrichs appears full of homoplasies. For instance, a recent molecular study suggests that even the 18 FVT cirral pattern, which is considered a very stable evolutionary feature (Berger 1999), evolved several times (Schmidt et al. 2007). The same applies to another “strong” morphologic character, viz., body flexibility/rigidity which is not as stable as it has been assumed (Berger 1999, Foissner and Stoeck 2006).
The hypotrichs with their high variety of cirral and ontogenetic patterns could play a major role in understanding evolution and classification of ciliates, especially below the ordinal rank. However, testable evolutionary hypotheses are rare, possibly because researchers underestimated the extent of homoplasy and diversity, thus becoming unable to abstract hypotheses from the data available. Taking into account these problems, we shall propose some hypotheses and translate them into classification units. Presently, about 600 valid hypotrich species have been described (H. Berger, pers. inform.). However, this is only the tip of the iceberg because new taxa are described at a high rate (Berger 2006), and the senior author has about 250 undescribed hypotrichs, mainly from soils globally, in his notes. These many new taxa will make “small” genera and families larger and more reliable.
Neokeronopsis (Afrokeronopsis) aurea was discovered in the Republic of South Africa, i.e., in a soil sample from the bank (active floodplain) of the Matjula River in the surroundings of the Berg-en-dal Lodge, E31°28′ S25°20′. The Matjula River is a small tributary to the large Crocodile River at the southern border of the Krueger National Park. The wet soil was collected from the upper 0–10 cm together with some plant litter and grass roots. The sample was taken in February 1995, air-dried for one month, and stored in a plastic bag.
A second population of N. (A.) aurea was discovered in soil (pH 6.5) from the dry bed of the Mlambane River, about 10 km north of the Berg-en-dal Lodge. This population, which is highly similar to the type, will be described later. Here, we use it mainly for discussion of biogeographic aspects.
Further, we studied the protargol slides of N. (N.) spectabilis deposited by Warren et al. (2002) in the British Museum of Natural History.
In 2006, that is, 11 years after collection, the sample was rewetted with distilled water to obtain a non-flooded Petri dish culture, as described in Foissner et al. (2002). To increase organismal activity and filter capacity of the very fine-grained, alluvial soil, we added some sterilized, chopped wheat straw. This sample yielded an extraordinarily diverse ciliate community with about 150 species, of which at least 30 were undescribed. Neokeronopsis (A.) aurea was sparse. Thus, we isolated some specimens to set up cultures in Eau de Volvic (French mineral water) enriched with some squashed wheat grains to stimulate growth of bacteria and flagellates (Polytomella, Rhodomonas). Further, the ciliate Colpidium kleini was added as a food source, but was rarely ingested. Most specimens fed on starch grains from the squashed wheat kernels and/or on flagellates.
Living cells were studied using a high-power oil immersion objective and interference contrast. Protargol impregnation and scanning electron microscopy (SEM) were performed as described by Foissner (1991). However, Neokeronopsis (A.) aurea was extremely difficult to preserve. Thus, for protargol impregnation we added 1 ml of 2% aqueous osmium tetroxide to each 10 ml of Stieve’s fluid and fixed cultures in toto because cells usually burst when added dropwise. All manipulations on fixed cells were done with fine pipettes because cells tended to break when centrifuged. In spite of much effort, good SEM preparations were not obtained.
Counts and measurements on silvered specimens were performed at a magnification of × 1.000. In vivo measurements were conducted at magnifications of × 40–1.000. Drawings of live specimens were based on free-hand sketches and micrographs; those of impregnated cells were made with a drawing device. In the ontogenetic stages, parental structures are shown by contour, while newly formed structures are shaded black.
For analysis of the 18S rDNA sequence, 30 specimens were isolated with a micropipette from a pure culture, washed in Eau de Volvic, and transferred into 180 μl ATL buffer (Qiagen) and 20 μl Proteinase K (20 mg/ml). Subsequently, the genomic DNA was extracted using the protocol for cultivated animal cells of the DNEasy Tissue Kit (Qiagen), according to the manufacturer’s instructions. We used standard isopropanol precipitation to concentrate the extracted nucleic acids. Amplification of the 18S rDNA fragment was performed via PCR using the universal eukaryotic primers EukA and EukB (Medlin et al. 1988), while cloning was performed as described by Stoeck and Epstein (2003). Three positively screened (M13 reamplification) plasmids were sequenced bidirectionally with MWG (Ebersheim, Germany).
We aligned the 18S rDNA sequence to available Oxytrichinae, Stylonychinae and Urostylidae sequences using Clustal X (Thompson et al. 1997). The alignments were manually refined in MacClade (Maddison and Maddison 2003), according to conserved regions. We applied the program Modeltest (Posada and Crandall 1998) to choose the model of DNA substitution that best fits our data sets from among 56 possible models. Maximum parsimony, evolutionary distance, and maximum likelihood trees we calculated using the PAUP software package 4.0b10 PAUP (Swofford 2002), while a Bayesian inference tree was obtained by using Mr. Bayes (Ronquist and Huelsenbeck 2003). The DNA substitution model as well as parameter settings for the trees constructed are described in the legend to Figure 38. We assessed the relative stability of tree topologies using 1,000 bootstrap replicates and posterior probabilities of 751 Bayesian trees. Heuristic searches for bootstrap analyses employed stepwise addition, starting trees with simple addition of sequences and TBR branch-swapping. Bootstrap analyses settings were chosen according to the Modeltest output. For the Bayesian tree we ran two simultaneous, completely independent analyses starting from different random trees. The analysis also employed GTR+I+G as the DNA substitution model with the gamma distribution shape parameter, the proportion of invariable sites, base frequencies, and a rate matrix for the substitution model as assessed by Mr. Bayes. Metropolis coupling with 3 heated chains and one “cold” chain was employed to improve the Markov Chain Monte Carlo sampling of the target distribution. We ran 1,000,000 generations and sampled every 1,000th generation, resulting in 1,001 samples from the posterior probability distribution. Then, rooted and unrooted trees were calculated. All methods resulted in congruent trees (Fig. 38), that is, none assigned N. (A.) aurea to the Urostyloidea, as suggested by the midventral rows. All alignments and trees are available from the authors upon request.
Terminology is according to Corliss (1979) and Foissner and Al-Rasheid (2006). The term “midventral rows” is used as defined by Berger (2006), that is, it designates a longitudinal series of zigzaging cirri in two rows near the ventral midline. A new term is the “buccal depression”, i.e., a special concavity on the dorsal wall (bottom) of the buccal cavity. As yet, this structure has been found in only two genera, viz., Steinia (Kahl 1932, Voss and Foissner 1996) and Neokeronopsis (Afrokeronopsis) aurea. Further, we introduce the term “semirigid” as a category of body flexibility. This term applies to species which cannot be classified unambiguously in the rigid (e.g. Stylonychia) or flexible (e.g. Oxytricha) group. A further new term, pseudobuccal cirri, is explained in the section on “The origin of buccal cirri”. Finally, the vernacular term “midventral oxytrichids” is introduced to designate that taxa which have urostylidlike midventral rows but appear within the oxytrichids in molecular trees, for instance, Uroleptus and Neokeronopsis.
Nomenclature follows the recent revision of Berger (2006), who maintains the time-honoured naming of Stein and Ehrenberg, and thus abandons the more recent stichotrichs. Accordingly, Berger (2006) distinguishes the Hypotricha Stein, 1859 (e.g. Oxytricha, Urostyla; now widely named Stichotrichida) and the Euplota Ehrenberg, 1830 (e.g. Euplotes, Uronychia; now widely named Hypotrichida).
We shall never know whether Shi et al. (1999) made a mistake or intended to establish a new family Pattersoniellidae. Thus, this family is suspect nomenclaturally, suggesting a correct, new start, that is, to establish a new family Neokeronopsidae including Pattersoniella as a second genus.
Cells golden yellow to brown orange in the dissecting microscope and with beautiful golden sheen under oblique dark-field illumination. Colour caused by a rather intense orange colouration of the cytoplasm, especially in oral area, and numerous citrine cortical granules, forming narrow strands ventrally and dorsally. Cortical granules ellipsoid to obovate, attached with broad end (Figs (Figs7,7, ,34),34), about 2 × 1.3 μm in size, become deeply orange-coloured and ~ 1.5 μm across when detached from cortex or touched by coverslip (Figs (Figs7,7, ,32),32), arranged in complex pattern: (i) between adoral membranelles and within cirral rows (Figs 3–5, ,32);32); (ii) left underneath frontal cirri, forming brick-shaped aggregates each composed of four to six granule pairs (Figs (Figs3,3, ,28);28); (iii) between cirral rows, forming narrow strands of scattered granules (Fig. 4); (iv) around dorsal bristles, thus imitating dorsal kinety pattern (Fig. 6); and (v) lacking between transverse cirral row and left marginal row.
Cirral pattern urostylid (Berger 2006), frequently with small irregularities, such as breaks and/or some supernumerary cirri; frontal cirri and adoral membranelles form an impressive, apical corona (Figs (Figs1,1, ,14,14, 18, 19, ,25,25, 38b; Table 1); cirri associated with complex fibre system very similar in all rows, except of transverse cirri lacking laterally extending fibres (Fig. 8). Most cirri of ordinary thickness and length (18–23 μm), except of enlarged frontal, buccal, and transverse cirri; distances of cirri within rows rather constant, except of narrowly spaced cirri in rear region of marginal rows; thickness of cirri gradually decreasing from anterior to posterior, especially in marginal rows. Both marginal rows commence slightly above level of buccal vertex; right row almost straight ending subterminally, left row J-shaped curving around body end almost touching right row; gap between right and left row dorsally occupied by three inconspicuous, obliquely spread caudal cirri. Midventral rows about 7 μm apart in mid-body, cirri of right row visibly thicker than those of left, extend slightly obliquely and sigmoidally from right anterior end of body to near posterior end; with distinct irregularity at level of right end of adoral zone of membranelles, where rows commence to spread in a frontal corona and a hook-shaped buccal row with long portion of hook extending along right margin of buccal cavity (Figs (Figs1,1, ,14,14, ,25,25, 38b); first frontal cirrus produced by the undulating membranes, frontal bow thus contains one cirrus more than buccal bow. Two inconspicuous frontoterminal cirri underneath right end of adoral zone of membranelles and close to right midventral row. Transverse cirri extraordinary because thick and about 30 μm long in vivo; arranged in a long, J-shaped row commencing underneath level of buccal vertex and extending subterminally around last midventral cirri, thus ending right of cell’s midline (Figs (Figs1,1, 14, 18, 19, ,25,25, 38b; Table 1).
Dorsal bristles 4–5 μm long in vivo, densely spaced within rows, except of central body area occupied by loosely ciliated kinetofragments; encaged by long fibres in fusiform pattern, fibres especially distinct in dorsomarginal kineties. Bristle pattern complex and thus appearing fairly disordered at first glance, forms an average of 14 rows originating by different processes described in the ontogeny section (Figs 6, 11, 14, 15, 19, 20, 25, 26): (i) rows 1–3 in left third of cell, ontogenetically active and almost as long as body, i.e., commence subapically and end posteriorly left of midline with an inconspicuous caudal cirrus each; (ii) rows 1–3 fragmented in anterior portion, especially row 1; (iii) row 3 with multiple posterior fragmentation, producing about seven loosely ciliated kineties extending in middle body third and shortened anteriorly and posteriorly, except of ordinarily ciliated left and right row; (iv) in right body third about six dorsomarginal kineties decreasing in length from left to right.
Oral apparatus conspicuous due to the huge adoral zone of membranelles occupying almost 40% of body length and extending tail-like posteriorly on right body margin (Figs 1, 9, 14, 18–20, 25, 26, 38b; Table 1; for terminology, see Foissner and Al-Rasheid 2006). Adoral zone thus inverted U-shaped, respectively, narrowly spoon-shaped in plane projection (Fig. 10a); right half of U (spoon-handle) on right body margin, composed of minute, 3–5 μm long membranelles gradually increasing to 16–23 μm in left half, that is, at level of buccal cavity, and then gradually decreasing again to 3–5 μm in proximal region of zone covered by an inconspicuous buccal vertex. Membranelles of usual structure, i.e., exactly as described for Sterkiella histriomuscorum by Augustin and Foissner (1992), membranellar cilia, however, up to 20 μm long and highly differentiated (Figs (Figs18,18, 30, 31, 33, 35, 36): (i) length of cilia greatly increases from right to left; (ii) cilia of row 1 acicular and distinctly longer than those of rows 2 and 3; (iii) cilia of rows 2 and 3 with obtuse distal end; (iv) row 4 consists of only three minute cilia with rounded distal end; (v) rightmost cilia of ventral membranelles differentiated to up to 10 μm long, “lateral membranellar cilia” in vivo covered by the buccal seal and extending to right wall of buccal cavity, in scanning electron microscopic preparations usually appearing as a highly disordered stripe of cilia (Figs (Figs18,18, 30, 31).
Buccal cavity and undulating membranes basically of ordinary size and structure. However, the general appearance resembles a mixture of the Oxytricha and Cyrtohymena type because the cavity is rather large and deep, the endoral membrane extends obliquely across the cavity, and the buccal depression appears as a strongly curved elongation of the paroral membrane which, however, is not curved anteriorly (Figs 1, 9, 14, 18, 19, 25, 27–30; Table 1). Buccal depression circa 15 μm across in vivo and gradually deepening to about 5 μm, at anterior end of endoral membrane, left third covered by a hyaline plate possibly belonging to the buccal lip or buccal seal; proximal margin slightly tuberculate because touching anterior end of fibre bundle underneath endoral membrane; appears as a rather distinct, bright area in vivo (Figs (Figs1,1, 28, 29); as a wrinkled, almost invisible structure in protargol preparations (Figs 14, 19, 25, 27); and is invisible in the scanning electron microscope because covered by the buccal seal (Figs (Figs18,18, ,30).30). Buccal seal usually broken at left margin of buccal cavity, partially exposing endoral and lateral membranellar cilia (see above). We could not decide whether there is an upper and a lower seal or only the upper one.
Paroral membrane on base of the about 3 μm wide, inconspicuous buccal lip; distinctly curved but not extending to adoral zone of membranelles anteriorly, as in Cyrtohymena; polystichad, i.e., massive because consisting of very narrowly spaced, oblique kineties, each composed of three to four up to 20 μm long cilia decreasing to 10 μm and less in end regions of membrane (Figs 1, 9, 14, 16, 18–19, 25, 27, 30, 37; Table 1). Endoral membrane extends more or less obliquely across buccal cavity, posterior third rather sharply curved and intersecting optically with paroral membrane; composed of very narrowly spaced mono- or dikinetids; underlayed by a thick fibre bundle performing undulating movements under slight coverslip pressure. Pharyngeal fibres inconspicuous, extend obliquely backwards (Figs 1, 9, 14, 16, 19, 25, 27, ,30;30; Table 1).
Resting cysts conspicuous because 116 μm across on average ( 116.4, M 120.0, SD 5.0, SE 1.8, CV 4.3, Min 105, Max 120, n 14), invariably globular, dark at × 40–100, brown at higher magnifications, consist of two distinct layers (Figs 12, 13, 21–24): outer layer 25 μm thick on average ( 24.6, M 25.0, SD 3.9, SE 1.1, CV 15.8, Min 18, Max 30, n 14), hyaline and without any stratification, colourless, stains red with methyl green-pyronin, surface grown with bacteria and polygonally faceted by 1–2 μm high ridges distinct only in the scanning electron microscope; inner layer about 2 μm thick, compact and colourless. Cytoplasm studded with three kinds of inclusions: (i) colourless lipid droplets 2–10 μm across, usually 4–8 μm; (ii) orange-coloured, bright granules 0.5–3 μm across, providing the squashed cyst contents with a reddish sheen in interference contrast; (iii) colourless (crystalline?) granules about 1 × 0.7 μm in size. Unfortunately, we did not note whether the macronucleus nodules remain separate or fuse.
Many well-impregnated dividers were found in the protargol slides. Thus, each of the stages depicted has been seen in at least three specimens. The description is very detailed because the diagnoses contain sophisticated ontogenetic features and further research might show the need to include even more. We include also the ontogenetic comparison with N. (N.) spectabilis, as far as this is possible from the rather incomplete descriptions of Warren et al. (2002) and Wang et al. (2007).
The fully developed oral primordium extends along the left side of the anterior two thirds of the transverse cirral row, and protomembranelles, each composed of two rows of basal bodies, develop at the right margin of the anterior third of the oral primordium. Concomitantly, the loose array of basal bodies described above enlarges greatly and extends posteriorly to meet an anlagen field originating from the midventral rows (Figs 39a, ,42).42). Slightly later, protomembranelles have formed in the anterior half of the oral primordium and basal bodies begin to segregate for the new undulating membranes; further, the parental undulating membranes are dissolving (Fig. 40). During all these processes, the proximal end of the parental adoral zone and the distal end of the oral primordium are clearly separate, while both almost touch in N. (N.) spectabilis (Wang et al. 2007).
In mid-dividers, the oral primordium develops to a long ribbon with the anterior third sharply curved right-wards (Figs (Figs44,44, ,45).45). At the right side of the straight portion extends a thick streak of anarchic basal bodies, later forming the undulating membranes. The first frontal cirrus segregates from the right anterior end of the streak. The disintegration of the parental undulating membranes and pharyngeal fibres has been completed, leaving a narrow, strongly flattened buccal cavity and a thick streak of anarchic basal bodies right of the straight portion of the adoral zone. The streak of anarchic basal bodies develops by multiplication of basal bodies from the parental membranes, i.e., cirri are not involved. The first frontal cirrus is formed as described for the opisthe. The parental adoral zone of membranelles is inherited unchanged. All these processes are highly similar to those described in N. (N.) spectabilis (Warren et al. 2002, Wang et al. 2007), except of the anterior portion of the adoral zone which is much less curved than in N. (A.) aurea.
In late and very late dividers, where the fission furrow becomes recognizable, the adoral membranelles and the membranellar zone obtain their definite structure and shape, i.e., a third and fourth row of basal bodies are added to the individual membranelles and the distances between the membranelles increase in the anterior third of the zone, forming the long membranellar tail typical for this species (Figs 1, 10a, b, ,14,14, ,4646–48). In both, the proter and opisthe, the two undulating membranes have formed from the anarchic streak of basal bodies described above. Interestingly, the developing membranes do not extend side by side, as in many other hypotrichs, but optically intersect from the beginning in the posterior quarter; this is also recognizable in N. (N.) spectabilis (Warren et al. 2002). Later, the intersecting area migrates to the mid of the membranes (Figs 47, 48). This part of the oral ontogenesis appears highly similar in N. (N.) spectabilis and N. (A.) aurea. Unfortunately, neither Warren et al. (2002) nor Wang et al. (2007) described the further development of the oral apparatus, that is, the origin of the buccal cavity and buccal depression (if present at all!).
In early post-dividers, the oral apparatus is quite similar to that of very late dividers (cp. Figs Figs48,48, ,59),59), although the undulating membranes intersect more distinctly. Only in late post-dividers develop the pharyngeal fibres and the buccal depression as well as a typical cyrtohymenid oral apparatus with a deep buccal cavity and a strongly curved paroral membrane which now extends obliquely across the bottom of the buccal cavity (Fig. 61). Interestingly, the curvature of the paroral becomes flatter and thus Oxytricha – like in the fully grown specimens (cp. Figs. Figs.14,14, ,6161).
Wang et al. (2007) emphasize the unique position of the early oral primordium in N. (N.) spectabilis: “Uniquely, the oral primordium originates below the anteriormost transverse cirrus, which is in contrast to almost all other stichotrichous ciliates in which the oral primordium originates anterior to the transverse cirri or even near the ventral cirri”. However, this is not correct. Identical patterns are found in many oxytrichids, for instance, in Onychodromopsis flexilis (Petz and Foissner 1996) and Sterkiella cavicola (Foissner et al. 2002) as well as in some urostylids, for instance, Pseudoamphisiella lacazei and Holosticha bradburyae (for reviews, see Berger 1999, 2006). It appears that the early oral primordium usually extends along some anterior transverse cirri when these are numerous and form a row extending to mid-body. However, the oral primordium is indeed important in this group of species because there are three anlagen fields in N. (N.) spectabilis while only one in N. (A.) aurea (see above and diagnoses of subgenera).
Ontogenesis of the dorsal ciliature commences within kinety 3, where a dikinetidal anlagen streak each develops above and underneath mid-body (Fig. 49). Then follow kineties 1 and 2 and all anlagen develop to long, dikinetidal streaks extending to the ends of the cell left of the parental bristle rows (Figs (Figs50,50, ,55).55). In early mid-dividers, a unique kinety whirl develops in the posterior third of kinety anlage 3, both in proter and opisthe (Figs (Figs51,51, 57, 58). Although this event has not been described in N. (N.) spectabilis, we suppose that it occurs also in this species because the following stage is present in both, N. (N.) spectabilis and N. (A.) aurea: the whirl spreads and forms four to five short, sigmoidal, staggering kineties in mid-dividers (Fig. 52). Next, the staggering kineties increase in number by further fragmentation (Fig. 53) and then extend in the central third of the cell (Figs (Figs54,54, ,56)56) to form a fusiform bristle field in post-dividers (Fig. 60) and morphostatic specimens (Figs (Figs15,15, ,20,20, ,2626).
During the spread of the posterior fragments of kinety 3, another remarkable process begins not known from any other hypotrich: the anterior third of the newly formed kineties 1 and 2 as well as the leftmost fragment of kinety 3 commence to fragment, forming several short kineties in the broad anterior region of the cell (Figs 53, 54). This unique process is most distinct in kinety 1 and less distinct in kinety 3. When anterior and posterior fragmentation of kineties 1–3 have been completed, cell fission commences and the dorsomarginal kineties migrate onto the dorsal surface in late dividers (Fig. 54) and early post-dividers (Fig. 60). Thus, three main ontogenetic regions occur on the dorsal surface: the left third is occupied by kineties 1 and 2 and their anterior fragments; the middle third is occupied by kinety 3 and its posterior fragments; and the right third is filled with dorsomarginal kineties originating close to the right marginal row.
A caudal cirrus each develops at the posterior end of kineties 1–3 in late dividers (Fig. 53). As usual, the caudal cirrus of kinety 3 is produced by the rightmost fragment, while anterior fragmentation occurs in the left-most fragment (Figs 53, 54). This fragmentation pattern suggests that kinety 3 is homologous to kineties 3 and 4 of the oxytrichids s. str. (for a review, see Berger 1999). Neokeronopsis (N.) spectabilis generates several caudal cirri in each kinety, a conspicuous difference used to define the subgenera Neokeronopsis and Afrokeronopsis as well as the species spectabilis and aurea.
The 18S rDNA sequence of N. (A.) aurea is 1,768 bp long and available under GenBank accession number EU124669. Comparing the N. (A.) aurea sequence to sequences from representative flexible and rigid hypotrichs identifies the flexible Crytohymena citrina (accession numbers AF508755, AY498653) as the closest relative of the semiflexible N. (A.) aurea in all phylogenetic analyses. We here only show the evolutionary distance and Bayesian phylogeny (Fig. 38). The MP and the ML trees as well as trees from the calculations mentioned in the method section are available from the authors upon request. Note the often low bootstrap values, showing that the trees are far from being settled.
Warren et al. (2002) based the redescription of N. (N.) spectabilis on specimens prepared with Wilbert’s protargol method. The cells are distributed over four slides, each possibly representing a separate preparation. We re-analysed the slides for several features possibly differing in N. (N.) spectabilis and N. (A.) aurea.
There are one or several “buccal cirri” at the right margin of the buccal cavity (for a discussion of terminological matters, see Berger 1999). Usually, these cirri originate from cirral anlage II, even in genera which have, like Neokeronopsis, a bicorona, for instance, Pseudokeronopsis (Berger 1999, 2006; Fig. 65). Depending on the size of the anlage, one or several buccal cirri are generated. Rarely, the buccal cirrus has been lost, e.g., in Paragastrostyla, or has been incorporated in the frontal bicorona, as in Uroleptopsis (Berger 2006).
Neokeronopsis and Pattersoniella are the great, as yet unrecognized exceptions: they generate the anteriormost buccal cirrus in the ordinary way, i.e., from cirral anlage II, while the following are midventral cirri, i.e., originate from the anterior portion of the left midventral row (Figs 62–64). At first glance, these cirri are hardly recognizable as buccal cirri because the row extends obliquely away from the margin of the buccal cavity (Figs 62, 63). Indeed, Berger (2006) defines Neokeronopsis as having a bicorona (“two arched rows of frontal cirri”), obviously not recognizing the buccal nature of the inner cirral arch. The same applies to Wang et al. (2007). However, the ontogenetic data show that the inner cirral bow is composed of buccal cirri, both in N. (N.) spectabilis (Wang et al. 2007) and N. (A.) aurea, where the various transition stages are especially distinct (Figs (Figs4646–48, 59, 61).
We consider these two modes of buccal cirri production as a rather fundamental difference, showing not only the close relationship of Pattersoniella and Neokeronopsis, but also their distinctness within the hypotrichs, providing further reason for the family classification suggested in the present study and by Foissner and Stoeck (2006).
Structures of different origin should be separated terminologically. Thus, we suggest the term “pseudobuccal cirri” for cirri located like buccal cirri but not originating form cirral anlage II. This distinction does not exclude to use “buccal cirri” as a general term for both, e.g., when their origin is not important in the context used. The pseudobuccal cirri are not identical with the malar and paramalar cirri coined by Borror (for a discussion of these terms, see Berger 1999, 2006).
Neokeronopsis is the fourth oxytrichid clade with midventral cirral rows (Foissner et al. 2004, Foissner and Stoeck 2006), suggesting the cirral pattern as an ambiguous phylogenetic marker. This is sustained by a recent molecular study (Schmidt et al. 2007), which indicates that even the “conservative” (Berger 1999, 2006) 18 frontal-ventral-transverse oxytrichid cirral pattern evolved convergently several times. Further, N. (A.) aurea has cortical granules, although it is (semi)rigid, a feature as yet found only in the flexible species of the Oxytrichinae (Berger 1999, Foissner et al. 2004). Obviously, N. (A.) aurea breaks both, the flexibility – and the granule dogma (Foissner et al. 2004, Foissner and Stoeck 2006). All these features evolved several (perhaps many) times in different lineages of the hypotrichs. Thus, they cannot be used longer as reliable phylogenetic markers, leaving various ontogenetic features, such as the origin of the ventral and dorsal ciliary pattern and the behaviour of the nuclear apparatus (nodules divide individually or fuse to divide as a single mass). But can we trust in the ontogenetic markers? The present state of knowledge suggests, we can although their interpretation is not easy. For instance, is the special mode of forming the buccal cirri in Neokeronopsis and Pattersoniella a reliable ontogenetic marker for a close relationship or evolved it convergently?
The widespread occurrence of convergences might be one of several reasons for the poor match of classic and molecular phylogenies in hypotrichs (Foissner et al. 2004, Schmidt et al. 2007). However, we emphasize that the molecular phylogenies are also controversial, possibly because they are based on a single gene (18S rDNA), of which we know that it is often too conservative for the classification of families and genera. All these problems show the urgent need of a combined classical and molecular approach, specifically, for detailed comparative ontogenetic and multigene analyses.
There are two highly characteristic cirral arrangements within the hypotrichs, viz., the oxytrichid and the urostylid pattern (Berger 1999, 2006; Foissner et al. 2004). The oxytrichid pattern typically consists of 18 cirri, i.e., 3 frontal, 5 frontoventral, 3 postoral, 2 pretransverse, and 5 transverse cirri (Fig. 38d). The urostylid pattern is characterized by two longitudinally extending rows with cirri arranged in a zigzagging “midventral pattern“ (Figs 38a–c). Traditionally, these patterns are assigned to different families, viz., the Oxytrichidae and the Urostylidae (Borror 1972; Berger 1999, 2006; Foissner et al. 2004).
This classification was only partially supported by the molecular studies which showed that species with typical midventral pattern, i.e., Uroleptus spp. cluster within the 18 cirri Oxytrichidae (Hewitt et al. 2003, Foissner et al. 2004; Fig. 38). Foissner et al. (2004) tried to solve this dilemma by proposing the CEUU hypothesis (Convergent Evolution of midventral cirral rows in Urostylids and Uroleptids) which suggests that the urostylid midventral pattern evolved from an oxytrichid ancestor and developed a second time from a different ancestor within the Oxytrichidae. Unfortunately, Foissner et al. (2004) could not provide a definite morphologic, ontogenetic, or molecular proof for the CEUU hypothesis. Such proof would have required a ciliate with the following combination of key features: midventral cirral pattern; fragmentizing dorsal kineties, preferably kinety 3; dorsomarginal kineties; and an oxytrichid 18S rDNA sequence. Obviously, N. (A.) aurea has all these attributes, and thus confirms the CEUU hypothesis. Likely, N. (N.) spectabilis will join as soon as its gene sequence is available.
Our gene tree (Fig. 38) suggests that a midventral cirral pattern evolved within the oxytrichids not only two times, as proposed by Foissner et al. (2004), but several times, viz., in Pattersoniella, Uroleptus, Rigidothrix, and Neokeronopsis. Likely, further such genera await discovery and/or are misplaced at the present state of knowledge, for instance, Territricha, Bicoronella, Afrophrya, and Holosticha stueberi (for a detailed discussion of these taxa, see Foissner and Stoeck 2006).
The present and former investigations (Foissner et al. 2004, Foissner and Stoeck 2006) show an impressive fact (Fig. 38): an urostylid (midventral) cirral pattern evolved at least four times within the flexible oxytrichids, viz., in Uroleptus, Pattersoniella, Rigidothrix and Neokeronopsis. When Foissner et al. (2004) proposed the CEUU hypothesis, the sequences of Rigidothrix and Neokeronopsis were not yet known. Thus, Foissner et al. (2004) did not comment on classification of Uroleptus. Now, however, time is ripe for conclusions from the CEUU hypothesis and the sequence data accumulated. Each of the “midventral oxytrichids” is associated with a distinct molecular position, showing that they evolved independently in different evolutionary lines of the oxytrichids: Pattersoniella appears in the rigid clade, Rigidothrix clusters near Oxytricha, Neokeronopsis clusters with the Paraurostyla-Cyrtohymena group, and Uroleptus forms its own cluster between the two others (Fig. 38). In accordance with Foissner and Stoeck (2006), we thus propose a distinct family for each of the “midventral oxytrichids”. Although this is not supported by monophylies in the molecular trees, possibly due to insufficient taxon sampling and a low resolution of the hypotrich SSU rDNA in general (Foissner et al. 2004, Schmidt et al. 2006), it is warranted by the morphologic and ontogenetic data (for reviews, see Berger 1999, 2006), i.e., hypotrichs with oxytrichid, respectively, urostylid cirral pattern should not remain in the same family because they represent highly distinct evolutionary lines. Ranking this difference only at genus level would create a massive disproportionality to many other genera often separated only by the presence vs. absence of transverse cirri or by the pattern formed by the undulating membranes. Figure 65 presents a Hennigian argumentation scheme which summarizes the present and former hypotheses.
Family Neokeronopsidae: for definition, see Result section. Here we provide the theoretical background (above) and some important details.
Kahl (1932), who discovered N. (N.) spectabilis, classified it close to Holosticha due to the midventral cirral pattern. Today, hypotrichs with midventral pattern are usually assigned to the Urostyloidea (Berger 2006). Based on a detailed redescription and some ontogenetic data, Warren et al. (2002) recognized that N. (N.) spectabilis has features from both, the oxytrichids (e.g., fragmentation of dorsal kinety 3) and the urostylids (e.g., midventral cirral pattern). Thus, they concluded that placement of Neokeronopsis within the Urostylidae remains uncertain. Berger (2006), in contrast, classified Neokeronopsis in the Oxytrichidae, but emphasized the need of molecular data. Very recently, Wang et al. (2007) provided more complete ontogenetic data and concluded that N. (N.) spectabilis, “very likely represents an intermediate form between oxytrichids and urostylids”. The ontogenetic and molecular data from N. (A.) aurea support the classifications of Berger (2006) and Foissner and Stoeck (2006), who considered Neokeronopsis a derived oxytrichid with a secondarily evolved midventral cirral pattern, quite similar to Uroleptus spp. (Foissner et al. 2004) and Rigidothrix goiseri (Foissner and Stoeck 2006). This is in accordance with the CEUU hypothesis which suggests that a midventral pattern evolved at least two times: the first, older event caused the ancestor to split into an oxytrichid and an urostylid lineage, while the second, more recent event caused the development of a midventral pattern in several oxytrichid lineages (Fig. 64).
The molecular analyses indicate a close relationship of Neokeronopsis with Cyrtohymena, as suggested by Berger (2006), based on the presence of cortical granules and the cyrtohymenid oral apparatus (Fig. 38). However, the Paraurostyla – Cyrtohymena – Neokeronopsis cluster has low bootstrap support (59/< 50/< 50), indicating that it could merge with the large cluster containing Pattersoniella, if taxon sampling is increased. Indeed N. (A.) aurea has two highly specific traits in common with Pattersoniella, viz., the production of buccal cirri from the midventral rows (Figs (Figs4646–48) and the buccal depression (Fig. 9) found also in Steinia, a close molecular relative of Pattersoniella (Fig. 38). Thus, we disagree with Berger (2006) who excluded a close relationship of Pattersoniella and Neokeronopsis.
Certainly, the molecular trees indicate that even special and thus “strong” characteristics, such as the buccal depression and the uncommon mode of forming the buccal cirri, could have evolved convergently. On the other hand, the poor bootstrap support of the Neokeronopsis clade (Fig. 38) and the low resolution of the hypotrich SSU rDNA in general (Foissner et al. 2004, Schmidt et al. 2006) warrant to interpret the molecular trees as critically as the morphologic ones. At the present state of knowledge, the inclusion of Pattersoniella in the Neokeronopsidae avoids the creation of a further family, as explained in the first paragraph of this chapter.
Family Uroleptidae: for definition, see Result section. Here, we provide the justification (first paragraph of chapter) and some important notes.
The diagnosis is based on the molecular data (Fig. 38). However, if it is assumed that Rigidothrix lost dorsal kinety fragmentation (Fig. 65), then the uroleptids can be clearly defined also morphologically: very flexible midventral oxytrichids lacking fragmentizing dorsal kineties.
Uroleptus, type genus of the family, is full of problems which should be solved by a detailed revision of the group. However, Kahl (1932), Foissner et al. (1991), and Berger (2006) addressed some issues, especially, they confined Uroleptus to species with midventral cirral pattern. Another major problem concerns the distinction from Holosticha and several other urostylids because Uroleptus differs from these genera only by the slender, more or less pisciform body shape (Kahl 1932, Foissner et al. 1991). However, the molecular data show that Uroleptus, indeed, is very different from the holostichids: the former belongs to the oxytrichids, the latter to the urostylids (Foissner et al. 2004, Berger 2006, Schmidt et al. 2006). Based on this knowledge, Berger (2006) established an unranked taxon, the Dorsomarginalia, which include all ordinary and midventral oxytrichids. We agree, but suggest family rank for the Oxytrichinae and Stylonychinae (Fig. 65).
Neokeronopsis (Neokeronopsis) spectabilis and Neokeronopsis (Afrokeronopsis) aurea differ by several distinct features possibly useful for generic or subgeneric separation. Unfortunately, the descriptions of N. (N.) spectabilis are not as detailed as one would wish, and the reinvestigation of the preparations from the Polish population could not eliminate all uncertainties (see Result section). Further, gene sequences are available only from N. (A.) aurea. Accordingly, we separate the African species only at subgeneric level from the Eurasian counterpart, emphasizing both similarities and differences.
The buccal depression is a highly characteristic feature of N. (A.) aurea (Figs 25, 27–29). So far, a buccal depression was known only from the genus Steinia (Kahl 1932, Foissner 1989, Berger 1999), i.e., has not been mentioned in the four descriptions of N. (N.) spectabilis (Kahl 1932, Warren et al. 2002, Berger 2006, Wang et al. 2007). The reinvestigation of the Polish population indicates that a buccal depression is, indeed, absent from N. (N.) spectabilis. However, live observation is required to be entirely sure. Certainly, this feature is a “strong” generic character.
The paroral membrane is composed of short, oblique kineties in N. (A.) aurea (Figs (Figs16,16, ,27,27, ,37)37) and of a double row of basal bodies (dikinetids) in N. (N.) spectabilis (Warren et al. 2002). However, the reinvestigation of the Polish N. (N.) spectabilis showed the presence of short, oblique kineties in, at least, the curved portion of the paroral membrane. Thus, this feature is useful only at species level.
Neokeronopsis (Afrokeronopsis) aurea has 3 caudal cirri (Table 1), while N. (N.) spectabilis has an average of about 9 (Warren et al. 2002, Wang et al. 2007). Usually, the number of caudal cirri is not used as a generic feature. However, this might be too conservative because most hypotrichs have 3 caudal cirri, and thus a considerably increased number might indicate a distinct evolutionary branch.
Neokeronopsis (Neokeronopsis) spectabilis develops the oral primordium from three anlagen, i.e., from two small fields underneath the buccal vertex and a narrow field anterior and along the upper three transverse cirri (Wang et al. 2007). In contrast, N. (A.) aurea generates the oral primordium along the upper 8–11 cirri of the transverse cirral row (Fig. 17); this has been checked in three specimens. Usually, such difference is not used as a generic character, but in combination with other features it appears a useful discriminator.
The anterior fragmentation of dorsal kineties 1–3 in N. (A.) aurea is an outstanding feature as yet not observed in any other hypotrich (Figs (Figs15,15, ,20,20, ,26).26). Likely, the fragmentation serves to fill the large anterior area with dorsal (sensory?) bristles. Warren et al. (2002) and Wang et al. (2007) do not mention such fragmentation in N. (N.) spectabilis. However, Figure 18 in Warren et al. (2002) and Figure 30 in Wang et al. (2007) indicate some fragmentation of kineties 1 and 2; on the other hand, Figure 15 in Wang et al. (2007) excludes this possibility. Our reinvestigation shows that Warren et al. (2002) overlooked fragmentation in N. (N.) spectabilis. However, details are markedly different, suggesting that this feature is useable for, at least, species distinction. The micrographs in Wang et al. (2007) show a long anterior fragment near dorsal kinety 1, while multiple posterior fragmentation, which is so prominent in the Polish specimens, is apparently absent. Thus, the Chinese population possibly represents a distinct species. To be sure, very late dividers and post-dividers should be restudied.
We did not find any species in the literature that could be identical with N. (A.) aurea. It differs from N. (N.) spectabilis, as described by Kahl (1932) and redescribed by Warren et al. (2002), Berger (2006) and Wang et al. (2007), not only by the subgeneric features outlined in the previous section but also by the following details: (i) although body size is similar or larger in N. (N.) spectabilis, the number of cirri and adoral membranelles is on average higher by 15–30% in N. (A.) aurea; (ii) the cortical granulation is distinct not only dorsally but also ventrally, especially left of the cirri of the frontal corona (Figs 3–5, ,28);28); (iii) the macronucleus nodules are connected by a fine strand (Fig. 15); (iv) the transverse cirral row extends much farther anteriorly in N. (A.) aurea than in N. (N.) spectabilis due to both a slightly higher number and wider spacing of the cirri (Figs (Figs1,1, ,14,14, ,18);18); (v) the right end of the adoral zone of membranelles extends much farther posteriorly in N. (A.) aurea than in N. (N.) spectabilis (Figs (Figs1,1, ,19,19, ,25);25); (vi) the buccal cirral row is distinctly closer to the margin of the buccal cavity in N. (A.) aurea than in N. (N.) spectabilis (Figs (Figs1,1, ,14,14, ,25);25); (vii) the paroral membrane is less distinctly curved in N. (A.) aurea than in N. (N.) spectabilis, where it almost abuts to the anterior end of the endoral membrane (Figs (Figs14,14, ,18,18, ,2525).
Possibly, N. (A.) aurea represents the plesiomorphic state, i.e., is the ancestor of N. (N.) spectabilis. This is indicated by the oral primordium which develops along the anterior half of the transverse cirral row in N. (A.) aurea, while mainly above the transverse cirral row in N. (N.) spectabilis. Obviously, the transverse cirral row has been shortened in N. (N.) spectabilis.
Both N. (N.) spectabilis and N. (A.) aurea are large, highly conspicuous species, representing ideal biogeographic flagships (Foissner 2006). Indeed, these two species provide an almost perfect proof for the restricted distribution of certain protist species because they occur in quite ordinary habitats (ponds, rivers and their floodplains) present all over the world. In spite of this, N. (N.) spectabilis has been recorded only from Eurasia, while N. (A.) aurea is possibly restricted to the Palaeotropis or Gondwana.
Neokeronopsis (Afrokeronopsis) aurea is, at present, known only from two floodplains in the Krueger National Park, Republic of South Africa, although we investigated several samples each from the Danube floodplain in Austria, the Amazon floodplain in Brazil, and the Murray River floodplain in Australia (Foissner 1997, 1998, 2007; Foissner et al. 2002; Chao et al. 2006). In contrast, eight Eurasian records, of which five are substantiated by detailed data, are known from N. (N.) spectabilis (Berger 2006): Austria, Germany, Slovakia, Poland, Ukraine, and China (Wang et al. 2007). If both, N. (N.) spectabilis and N. (A.) aurea were present in Eurasia, it would be highly unlikely that only N. (N.) spectabilis has been found; likewise, if both occur in the Krueger National Park, it would be unlikely that we found only one.
Financial support was provided by the Austrian Science Foundation (FWF project P-19699-B17) and the German Science Foundation (DFG, STO-414/2-3). The technical assistance of Mag. Birgit Weissenbacher, Mag. Gudrun Fuss, Robert Schörghofer, Hans-Werner Breiner and Andreas Zankl is greatly acknowledged. Special thanks to Dr. Alan Warren (British Museum of Natural History) for sending us slides from N. (N.) spectabilis.