Genomic Mining of Phylogenetically Informative Nuclear Markers in Bark and Ambrosia Beetles

Abstract

Deep level insect relationships are generally difficult to resolve, especially within taxa of the most diverse and species rich holometabolous orders. In beetles, the major diversity occurs in the Phytophaga, including charismatic groups such as leaf beetles, longhorn beetles and weevils. Bark and ambrosia beetles are wood boring weevils that contribute 12 percent of the diversity encountered in Curculionidae, one of the largest families of beetles with more than 50000 described species. Phylogenetic resolution in groups of Cretaceous age has proven particularly difficult and requires large quantity of data. In this study, we investigated 100 nuclear genes in order to select a number of markers with low evolutionary rates and high phylogenetic signal. A PCR screening using degenerate primers was applied to 26 different weevil species. We obtained sequences from 57 of the 100 targeted genes. Sequences from each nuclear marker were aligned and examined for detecting multiple copies, pseudogenes and introns. Phylogenetic informativeness (PI) and the capacity for reconstruction of previously established phylogenetic relationships were used as proxies for selecting a subset of the 57 amplified genes. Finally, we selected 16 markers suitable for large-scale phylogenetics of Scolytinae and related weevil taxa.

Fig 1. Schematic tree showing well supported relationships between tribes within the subfamily Scolytinae and other weevil families and subfamilies considered in this study.

Fig 2. Structure of the PCR amplified gene fragments.

The graphics illustrate intron-exon patterns in 16 markers with coding regions shown as black bars and introns as thin black lines. Length variable coding regions (indels) were colored in light grey (Iap2 and Arr2).

Fig 3. Phylogenetic trees based on Bayesian analyses of 16 selected genes.

Trees were rooted with the most distant outgroup available for each marker. Posterior probabilities are given to the left of the nodes. Sequences of D. ponderosae (ToDen00) were obtained from GenBank.

Fig 4. Phylogentic informativeness profiles.

The K2P model was used to estimate substitution rates in HyPhy as implemented in the software PhyDesign. Different evolutionary models produced similar results (data not shown). The dated phylogenetic tree was obtained using BEAST v1.8.2.

Fig 5. Phylogenetic trees resulting from Bayesian analyses of 8 excluded gene fragments.