The inadequacy of morphology for species and genus delineation in microbial eukaryotes: an example from the parabasalian termite symbiont coronympha

Abstract

Background: For the majority of microbial eukaryotes (protists, algae), there is no clearly superior species concept that is consistently applied. In the absence of a practical biological species concept, most species and genus level delineations have historically been based on morphology, which may lead to an underestimate of the diversity of microbial eukaryotes. Indeed, a growing body of molecular evidence, such as barcoding surveys, is beginning to support the conclusion that significant cryptic species diversity exists. This underestimate of diversity appears to be due to a combination of using morphology as the sole basis for assessing diversity and our inability to culture the vast majority of microbial life. Here we have used molecular markers to assess the species delineations in two related but morphologically distinct genera of uncultivated symbionts found in the hindgut of termites.

Methodology/principal findings: Using single-cell isolation and environmental PCR, we have used a barcoding approach to characterize the diversity of Coronympha and Metacoronympha symbionts in four species of Incisitermes termites, which were also examined using scanning electron microscopy and light microcopy. Despite the fact that these genera are significantly different in morphological complexity and structural organisation, we find they are two life history stages of the same species. At the same time, we show that the symbionts from different termite hosts show an equal or greater level of sequence diversity than do the hosts, despite the fact that the symbionts are all classified as one species.

Conclusions/significance: The morphological information used to describe the diversity of these microbial symbionts is misleading at both the genus and species levels, and led to an underestimate of species level diversity as well as an overestimate of genus level diversity. The genus 'Metacoronympha' is invalid and appears to be a life history stage of Coronympha, while the single recognized species of Coronympha octonaria inhabiting these four termites is better described as four distinct species.

Figure 1. Morphological features of Coronympha and ‘Metacoronympha’ from Incisitermes termites.

(A–D) Light micrographs of the Coronympha-morph from I. banksi, I. milleri, I. snyderi, and I. schwarzi, respectively, showing a relatively small, generally pyriform cell with eight karyomastigonts in a single anterior circle. (E–H) Light micrographs of the Metacoronympha-morph from I. banksi, I. milleri, I. snyderi, and I. schwarzi, respectively, showing a larger rounded cell with dozens of karyomastigonts arranged in several anterior spiraling rows. (I) Metacoronympha-morph from I. banksi viewed from the top down, showing details of the spiral pattern of karyomastigonts and the flagellar bundles. (J) Dividing Metacoronympha-morph from I. milleri, where two polar zones of karyomastigonts have formed. (K) Coronympha-morph from I. snyderi viewed from the top down, showing details of the eight karyomastigonts in a single circle, the pyriform nuclei, and the eight flagellar bundles. (L–M) SEM of Coronympha-morph and Metacoronympha-morph, respectively, from I. snyderi. In both cases the flagella can be seen to emerge in groups of 4, with three slender anterior flagella bundled, and the thicker recurrent flagellum apart and free from the cell surface. The eight-fold symmetry can be seen in the anterior view of Coronympha (L). In neither cell type was a glycocalyx observed. Significant numbers of bacterial symbionts are sometimes associated with the surface of parabasalia in the termite gut environment, but in this case this was only observed in the Metacoronympha-morph from I. milleri (F). Scale bars in A–K are 10 µm and in L and M are 5 µm.

Figure 2. Maximum likelihood phylogeny of Parabasalia, with an emphasis on the Cristomonada subgroup (to which Coronympha and ‘Metacoronympha’ belong).

Branches are labeled with the GenBank Locus, followed by the genus and species where known, or in the case of unidentified termite symbionts, but the genus and species of the termite and the word ‘symbiont’. Locus IDs were not available for new sequences: the accession numbers for these are FJ986219-22 for Coronympha from I. snyderi, I. banksi, I. milleri, and I. schwarzi, respectively. Numbers at nodes correspond to maximum likelihood bootstrap support (top), and Bayesian posterior probabilities (bottom). The Coronympha-morph and Metacoronympha-morph from a given Incisitermes species are shown in shaded boxes with the host indicated to the right. In all four cases, these two sequences were identical or nearly identical. Environmental sequences from the same host were likewise identical (not shown). The previously reported ‘M. senta’ from I. snyderi branches with the other sequences from I. snyderi, as expected. The C. octonaria and unidentified ‘Ii1’ sequences reported previously from I. snyderi and I. immigrans are both virtually identical to sequences from I. schwarzi, and appear to be due to termite misidentification (see text for details).

Figure 3. Matrix showing uncorrected pairwise distances between: (below the diagonal) the hosts (H) and their symbionts (S), and (above the diagonal) the symbiont:host ratio of distances.