Stilbonematinae

Scharhauser, Florian, Zimmermann, Judith, Gruber-Vodicka, Harald R., Ott, Jörg A. & Leisch, Nikolaus, 2020, Morphology of obligate ectosymbionts reveals Paralaxus gen. nov.: A new circumtropical genus of marine stilbonematine nematodes, Zoologica Scripta 49 (3), pp. 379-394 : 386-387

3.5 | Molecular data corroborates the monophyly of the subfamily Stilbonematinae

and all its genera, including Paralaxus

In addition to the nine Paralaxus specimens, we generated and analysed single worm metagenomes from 20 specimens covering six additional stilbonematine genera and extracted both host 18S and COI gene sequences for all specimens (Table S1). To analyse the monophyly of the Stilbonematinae , we reconstructed their phylogenetic relationships based on the 18S gene, as this is the only gene with sufficient taxon sampling. The 18S gene matrix, with a minimum sequence length of 1,300 bp, consisted of sequences from 78 Stilbonematinae as well as from 52 closely related Desmodorida and Microlaimida, and an outgroup of 24 Chromadorida sequences. The Stilbonematinae formed a highly supported clade, with other non-stilbonematine Desmodoridae and Draconematidae as closest relatives (Figure 3a).

For the analysis of the Stilbonematinae genera, we created a mitochondrial COI tree in addition to the 18S tree, using 27 metagenome-derived stilbonematine COI sequences and an ascarid as outgroup ( Baylisascaris procyonis mitochondrion, JF951366 ) (Figure 3b). Based on both datasets, all Stilbonematinae genera including Paralaxus gen. nov. formed well-supported clades (Figure 3a, b), thus corroborating our morphological assignments. In both the 18S and the COI-based analysis, the genus Leptonemella was phylogenetically most closely related to Paralaxus , with high statistical support in the 18S dataset and acceptable support in the COI-based phylogenies (Figure 3a, b). Other than that, the branching patterns between genera were largely inconsistent between the COI and 18S datasets. This could be linked to the remarkably low support values for many internal nodes in the COI dataset compared to the much higher spectrum of support values for the 18S-based tree topologies.

In an extended 18S and COI data matrix, we also included eight short-length, PCR-amplified sequences of five different Stilbonematinae species with a record of problematic phylogenetic placement— Laxus parvum , Robbea porosum , Stilbonema brevicolle , Catanema exile and Leptonemella brevipharynx (Armenteros, Ruiz-Abierno , et al., 2014; Figs. S9 and S10). With our new 18S- and COI-based datasets, the placements of Leptonemella brevipharynx and Catanema exile sequences were resolved and both clustered with sequences of their respective genera. For the sequences of the other three species, we observed inconsistent grouping with Stilbonematinae genera between the datasets (Figs S9 and S10). In the extended 18S tree, two Stilbonema brevicolle sequences appeared in two different clades, one closely related to Leptonemella , while the other one clustered with a yet undescribed genus-level clade designated as genus ‘B’ ( Zimmermann et al., 2016). In contrast to the 18S-based placement, the Stilbonema brevicolle sequences grouped together in the extended COI tree and formed a novel sister clade to several genera including Stilbonema , Robbea , Leptonemella and Paralaxus (Figure S10). In the extended 18S data matrix, the Laxus parvum sequences clustered within the genus Robbea (Figure S9). Similarly, in the COI dataset the L. parvum sequence grouped with the only available sequence of the genus Robbea (Figure S10). The sequences designated as Robbea porosum formed a divergent genus-level clade but the placement of this clade was not consistent for the two marker genes (Figs S9 and S10). In the 18S-based analyses, they were the sister clade to the genus Laxus , while in the COI dataset, they formed the sister clade to all other Stilbonematinae .

3.6 | Symbiont 16S phylogeny shows genus-level specificity

We constructed a matrix from all full-length 16S ‘Thiosymbion’ sequences, available in the databases and 21 metagenome-derived symbiont sequences—using five free-living Chromatiaceae as an outgroup. All symbiont sequences from the same host genus clustered in well-supported clades, irrespective of the number of gutless oligochaete symbionts included within those clades (Figure 4). In congruence to the host data, the symbionts of all newly sequenced Paralaxus individuals as well as the two previously published sequences from the Caribbean and Australia ( Zimmermann et al., 2016) formed a supported clade in the 16S phylogeny (Figure 4).

3.7 | COI primers cannot cover the compositional diversity in Stilbonematinae

Despite a highly sensitive PCR setup, we could not amplify the COI of any Paralaxus specimen using the widely used primer set by Folmer et al. (1994) (data not shown). To elucidate these PCR-based amplification problems, we analysed all our metagenome-derived full-length stilbonematine nematode COI sequences for compositional patterns. The COI sequences of the 19 stilbonematine nematode species that belonged to seven genera had a highly variable guanine and cytosine (GC) content, ranging from 27.7% in the genus Robbea to 53.1% in the genus Leptonemella . We identified substantial priming problems with the Folmer et al. (1994) primer set as well with as a second set (JB3F and JB5R) that is commonly used for COI amplification (Armenteros, Rojas-Corzo, et al., 2014; Armenteros, Ruiz-Abierno, et al., 2014; Bowles, Blair, & McManus, 1992; Derycke, Vanaverbeke, Rigaux, Backeljau, & Moens, 2010). Across the Stilbonematinae diversity, the Folmer et al. (1994) primers had a range of 3–11 mismatches for the 1490F primer and 1–5 mismatches for the 2198R primer. Similarly, the primers JB3F and JB5R used in a recent study by Armenteros, Ruiz-Abierno, et al. (2014) showed up to 11 mismatches for each primer (Figure 5). Of all four primers tested, only the JB5R perfectly matched and only to three sequences from Robbea hypermnestra (Figure 5). We performed an in silico analysis that mimics PCR conditions of low stringency, which would overcome up to 11 primer mismatches and would cover the whole diversity of Stilbonematinae . Our analysis showed that such permissive conditions would lead to unspecific binding at multiple alternate sites of the COI gene. We tried to design alternative primer sets, but no primer pair, neither for the full-length nor for partial regions, would reliably target the COI gene and would cover all genera of Stilbonematinae .

3.8 | The symbiont coat as a new diagnostic feature for Stilbonematinae genera

In addition to several host-based morphological differences, the presence of a multilayered coat distinguishes the new genus Paralaxus from Laxus , which carries a monolayer of symbionts. Extending from this observation, we analysed the structure and morphology of the bacterial coat of all Stilbonematinae genera and found substantial and highly informative differences. We therefore propose to include the description of the bacterial coat given below into the genus trait sets as a new character (see also iconographic overview of the different genera in Figure 6).