Eriosomatinae (von Dohlen and Moran, 2000)

Li, Xing-Yi, Jiang, Li-Yun, Qiao, Ge-Xia & Online, Published, 2014, Is the subfamily Eriosomatinae (Hemiptera: Aphididae) monophyletic?, Turkish Journal of Zoology 38 (2), pp. 285-297 : 289-295

4.1. Eriosomatinae is not monophyletic

All the trees obtained in the analyses with single-gene datasets were totally comb-like, which was similar to previous results based on single-gene datasets ( von Dohlen and Moran, 2000; Martínez-Torres et al., 2001). This would be due to the powerlessness of single-gene datasets, rather than the species tree originally being comb-like, whereby thus the nodal supports would be totally poor. Meanwhile, the support values in the major studies based on multigene datasets were significantly improved, especially within each subtribe. It was interesting that the 14 MPTs obtained were almost identical, which provided valuable references for the actual positions of the related out-groups on the phylogeny. Comparing the tree topologies from the major studies through BI/ML/MP reconstructions, it was concluded that the positions of some out-groups were flexible. Mindarinae and Thelaxinae were constantly inserted into the in-group, which suggested that the monophyly of Eriosomatinae might be in doubt. However, the basal node supports were still not high, which suggested that the topology was paraphyletic. Compared with the previous phylogenetic studies concerning similar paraphyletic topology ( von Dohlen and Moran, 2000; Martínez-Torres et al., 2001; Zhang and Qiao, 2008; Ortiz-Rivas et al., 2010), there were relatively high nodal support values for the clade E-root (0.66/100/56) in our results, revealing a constant clustering identical to the “E+T” clade ( Ortiz-Rivas et al., 2010). We thus conducted the test of the out-group to handle the relationships between the taxa in the “E+T” clade. In all trees obtained from the test of the out-group, only Phloeomyzinae was clustered away from the in-group with high nodal supports. The other 4 out-groups clustered into the in-group with high nodal supports, respectively. We could therefore conclude that there were certain related out-groups, i.e. Hormaphidinae , Anoeciinae , Mindarinae , and Thelaxinae , inserted between the 3 tribes of Eriosomatinae in the phylogeny, allowing us to confirm that Eriosomatinae is not monophyletic but is rather paraphyletic.

There were several possible explanations for this paraphyletic cladogram. Incompetence of datasets is one of the main possible causes, as referred to in previous studies ( von Dohlen and Moran, 2000; Ortiz-Rivas et al., 2010). Molecular phylogenetics is generally based on several premises, including that the gene trees can reveal the species tree and that the molecular phylogeny is resolvable ( Nei and Kumar, 2000). The aim of phylogenetic studies is to reconstruct a reasonable phylogeny, i.e. with perfect nodal support values, that could be treated as the approximate restoration of the species tree ( Rannala and Yang, 2008). However, to date, no results showed even medium nodal support values for the key nodes concerning the relationships among aphid tribes. Thus, more genes are needed in improving the confidence level (high nodal support values) of the reconstructed phylogeny, as suggested in phylogenetic studies using genomic data ( Dunn et al., 2006; Rannala and Yang, 2008). Furthermore, it was argued that the molecular phylogeny was originally comb-like, which might be possible evidence for the fast radiation hypothesis among the aphid tribes ( von Dohlen and Moran, 2000). Obviously, little phylogenetic information could be preserved during a fast radiation process, and so there were not enough phylogenetic informative sites (i.e. nucleotide substitutions) to resolve the serried short branches. In this scenario, the molecular phylogeny of these fast-radiated taxa was finally arranged in parallel. Additionally, the fast radiation hypothesis had been partially proved in Hormaphidinae through the estimation of divergent time ( Huang et al., 2012). Regardless, the classification of these aphid taxa might need adjustment in reference to the results of phylogenetic studies.

4.2. Monophyly of the 3 tribes

The position of Hormaphidinae and Anoeciinae in the major studies was slightly different to that in the test of the out-group. It was implied that Fordini and Pemphigini might be more closely related to Mindarinae and Thelaxinae , and meanwhile Eriosomatini might be more closely related to Hormaphidinae and Anoeciinae . However, these out-groups clustered together in the con-tree of the major studies, and thus the monophyly of the 3 tribes could be revealed. Taking no account of the monophyly of Eriosomatinae , it was notable that our results had several points in common with an earlier cladistic study ( Zhang and Chen, 1999). For example, Eriosomatini branched at a basal node, Pemphigini was closely related to Fordini , and Fordini consisted of Fordina and Melaphidina . However, some issues remaining from the earlier study were still unresolved, such as the monophyly of Eriosomatini , Pemphigini , and Fordini .

According to our results, it was implied that Eriosomatini and Fordini were monophyletic, which was consistent with several previous studies ( Inbar et al., 2004; Zhang and Qiao, 2007 b, 2008; Sano and Akimoto, 2011). The nodal support values of node E and node F were very high in different reconstructing methods, and these high support values were not affected by the changes of the out-groups. Additionally, Fordini consisted of 2 monophyletic subtribes ( Fordina and Melaphidina ), which was also consistent with the previous studies ( Zhang and Qiao, 2007a, 2008). However, the monophyly of Pemphigini was a little problematic, because node P did not exist in a few results of the test of the out-group. This was mainly attributed to the contribution of Formosaphis . Formosaphis is a monotypic genus, and the type species, Formosaphis micheliae Takahashi 1925 , is distributed in India, Japan, and China. It presents a hind wing typical of Pemphigini , but antennal segments III–V of the alatae exhibit many reticulate small secondary rhinaria and irregular sclerotizations identical to the Fordini species ( Zhang et al., 1999). Results of the test of the out-group showed that Formosaphis clustered in parallel with the other Pemphigini taxa and Fordini . However, the results of the major studies through BI/ML/ MP algorithms supported that Formosaphis clustered into Pemphigini with relatively high nodal supports, and the Pemphigini taxa formed a monophyletic clade. Therefore, our results reconfirmed that Formosaphis should be placed in Pemphigini rather than Fordini , which is consistent with a previous study ( Zhang and Qiao, 2007b).

Meanwhile, the monophyly of Eriosomatini , Pemphigini ,andFordiniisalsosupportedbymorphological characters and biological data ( Heie, 1980; Blackman and Eastop, 1994; Zhang et al., 1999). Not only do the apterous viviparous females of the 3 tribes show distinctly different characters (such as the presence and absence of siphunculi, the number of wax gland plates, shape of wax cells, and so on), but so do the alatae viviparous females (such as the shape of secondary rhinaria on antennae, veins of hind wing, tor he number of gonapophyses and wax gland plates) ( Table 4). Furthermore, the primary host plants of the 3 tribes show extreme specificity: Eriosomatini on Ulmaceae (Ulums and Zelkova ), Pemphigini (except Prociphilina ) on Salicaceae ( Populus ), and Fordini on Anacardiaceae ( Pistacia and Rhus ) ( Blackman and Eastop, 1994). The secondary host plants of the tribe/subtribes are also distinct, in that Eriosomatini , Pemphigina , Prociphilina , Fordina , and Melaphidina mostly feed on herbal monocots, herbal dicots, Pinaceae , Graminaceae, and mosses, respectively ( Zhang et al., 1999). Therefore, the monophyly of the 3 tribes is inferred from molecular data, morphological characters, and biological data.

We have concluded that Eriosomatinae is not monophyletic but is rather paraphyletic, because 4 out-group subfamilies were found to be placed between Eriosomatini and the other 2 tribes in the phylogeny. However, the monophyly of Eriosomatini , Pemphigini , and Fordini is supported by the obtained tree topologies, the morphological features, and the biological data. In addition, the tangled phylogenetic relationship between Fordini and Pemphigini was mainly attributed to the contribution of Formosaphis . It was reaffirmed based on more molecular data that the genus should be placed in Pemphigini rather than in Fordini .

Though the support values of the basal nodes might be improved in further studies, the phylogeny of the subfamily was still resolved well in this study. Morphological data could also provide abundant phylogenetic signals, such that it will be valuable to combine molecular and morphological data to reconstruct a total-evidenced phylogeny of the subfamily. In addition, the interrelationships of the “E+T” clade (including Eriosomatinae and 4 related subfamilies) should be the focus of other important work in the future, with more samples and larger datasets.

Acknowledgments

We thank the predecessors and colleagues of our lab who accumulated abundant samples for our study. We are indebted to FD Yang for making all the slide preparations and to Dr Z Wang for the help in the experiments. We also thank L Liu for sharing sequences of the out-groups.