Morphological and molecular phylogeny of dileptid and tracheliid ciliates: resolution at the base of the class Litostomatea (Ciliophora, Rhynchostomatia)

Abstract

Dileptid and tracheliid ciliates have been traditionally classified within the subclass Haptoria of the class Litostomatea. However, their phylogenetic position among haptorians has been controversial and indicated that they may play a key role in understanding litostomatean evolution. In order to reconstruct the evolutionary history of dileptids and tracheliids, and to unravel their affinity to other haptorians, we have used a cladistic approach based on morphological evidence and a phylogenetic approach based on 18S rRNA gene sequences, including eight new ones. The molecular trees demonstrate that dileptids and tracheliids represent a separate subclass, Rhynchostomatia, that is sister to the subclasses Haptoria and Trichostomatia. The Rhynchostomatia are characterized by a ventrally located oral opening at the base of a proboscis that carries a complex oral ciliature. We have recognized two orders within Rhynchostomatia. The new order Tracheliida is monotypic, while the order Dileptida comprises two families: the new, typically bimacronucleate family Dimacrocaryonidae and the multimacronucleate family Dileptidae. The Haptoria evolved from the last common ancestor of the Litostomatea by polarization of the body, the oral opening locating more or less apically and the oral ciliature simplifying. The Trichostomatia originated from a microaerophylic haptorian by further simplification of the oral ciliature, possibly due to an endosymbiotic lifestyle.

Fig. 1

(a–g) Main features of rhynchostomatians in vivo (b), after protargol impregnation (d–g), and in the scanning (a) and transmission (c) electron microscope. From Foissner et al., 1995 (d, f, g) and originals (a–c, e). (a) Monomacrocaryon terrenus, overview showing general body organization. (b–e) There are four basic nuclear patterns: a cylindroidal macronucleus with a single micronucleus in Monomacrocaryon (b); two macronuclear nodules with a single micronucleus in between in Dimacrocaryon (c); many macronuclear nodules and micronuclei scattered throughout cytoplasm in Dileptus (d); and a moniliform macronuclear strand with several micronuclei in Pseudomonilicaryon (e). (f, g) Dileptus margaritifer, right and left side view of oral ciliary pattern. CK – circumoral kinety, E – extrusomes, MA – macronucleus (nodules), MI – micronucleus, OB – oral bulge, OO – oral bulge opening, PE – perioral kinety, PR – preoral kineties. Scale bars: 5 μm (c, d), 20 μm (e–g), and 30 μm (a, b).

Fig. 2

Division modes of macronucleus in rhynchostomatians.

Fig. 3

Cladogram of ten rhynchostomatian genera generated by traditional Hennigian argumentation. For character coding, see Table 2 and section on character states. Only apomorhpies are shown.

Fig. 4

Phylogenetic tree of ten rhynchostomatian genera inferred from 15 characters using the genus Spathidium as the outgroup. Two methods (Bayesian inference and maximum parsimony) were used to construct the tree, both resulting in the same topology. Nodal supports are indicated by posterior probabilities for the Bayesian inference (BI) and bootstrap values for the maximum-parsimony (MP) analysis shown above and by Bremer indexes shown below each node. For character coding and distribution of characters among taxa, see Tables 2 and 3. The scale bar indicates the fraction of substitutions per site.

Fig. 5

Supposed evolution of the oral ciliary patterns and body shapes from a Pseudomonilicaryon-like ancestor. CK – circumoral kinety, CV – contractile vacuoles, MA –moniliform macronuclear strand, OO – oral bulge opening, PE (I + II) – perioral kinety (1 and 2), PE* – perioral-like kinety, PR – preoral kineties.

Fig. 6

Small subunit rRNA gene phylogeny based on 1442 nucleotide characters of 37 litostomatean taxa. The three was constructed using four methods (Bayesian inference, maximum likelihood, maximum parsimony, and neighbour-joining) with GTR + I + Γ substitution model and the variable-site gamma distribution shape parameter at 0.4900, the proportion of invariable sites at 0.6460, and a rate matrix for the model as suggested by Modeltest. Posterior probabilities (PP) for the Bayesian inference and bootstrap values for the maximum-likelihood (ML), maximum-parsimony (MP), and neighbour-joining (NJ) analyses are shown at nodes (a dash indicates values below 0.50 or 50%, respectively). Sequences in bold were obtained during this study. The scale bar indicates two substitutions per one hundred nucleotide positions.

Fig. 7

Small subunit rRNA gene phylogeny based on 1635 nucleotide characters of nine rhynchostomatian taxa. Four methods (Bayesian inference, maximum likelihood, maximum parsimony, and neighbour-joining) were used to construct the tree, all resulting in a very similar topology. Posterior probabilities (PP) for the Bayesian inference and bootstrap values for the maximum-likelihood (ML), maximum-parsimony (MP), and neighbour-joining (NJ) analyses are shown at nodes (a dash indicates values below 0.50 or 50%, respectively). The scale bar indicates six substitutions per one thousand nucleotide positions.