Spider mite (Acari: Tetranychidae) mitochondrial COI phylogeny reviewed: host plant relationships, phylogeography, reproductive parasites and barcoding

Abstract

The past 15 years have witnessed a number of molecular studies that aimed to resolve issues of species delineation and phylogeny of mites in the family Tetranychidae. The central part of the mitochondrial COI region has frequently been used for investigating intra- and interspecific variation. All these studies combined yield an extensive database of sequence information of the family Tetranychidae. We assembled this information in a single alignment and performed an overall phylogenetic analysis. The resulting phylogeny shows that important patterns have been overlooked in previous studies, whereas others disappear. It also reveals that mistakes were made in submitting the data to GenBank, which further disturbed interpretation of the data. Our total analysis clearly shows three clades that most likely correspond to the species T. urticae, T. kanzawai and T. truncatus. Intraspecific variation is very high, possibly due to selective sweeps caused by reproductive parasites. We found no evidence for host plant associations and phylogeographic patterns in T. urticae are absent. Finally we evaluate the application of DNA barcoding.

Fig. 1

Overview of the COI fragments sequenced in different studies and their relative position after alignment. The position of the fragment analyzed in this study and of the standard DNA barcoding fragment (Folmer fragment) on the total mitochondrial COI gene (position 1474–3009) are indicated on top. Base pair numbers correspond to the Drosophila melanogaster mitochondrial DNA sequence (GenBank accession nr. U37541). For each study, thick horizontal lines represent the fragment sequenced in all specimens and thin lines indicate the maximum sequence length. Number of sequences (between parentheses) and references are listed at the right. For references of unpublished studies, see Appendix

Fig. 2

Base compositions for each codon position of the 390-bp aligned COI region, averaged over all tetranychid samples. Error bars depict minimum to maximum range. Results of the homogeneity test are given for each codon position

Fig. 3

Saturation plots of transversion and transition rates against uncorrected p-distance at each codon position

Fig. 4

Maximum likelihood tree of the tetranychid dataset based upon COI sequences. GenBank accession numbers and associated species names are given. If a haplotype is found more than once, the accession number is followed by the haplotype number (see Appendix) and the number of times the haplotype is found between parentheses. Numbers on the branches indicate the percentage bootstrap values (>50) based on NJ bootstrapping with ML settings (1,000 replicates). Bar at the lower left corner depicts the branch length corresponding to 10% maximum likelihood distance

Fig. 5

Maximum likelihood tree of the genus Tetranychus based upon COI sequences. GenBank accession numbers and associated species name are given (except for T. cinnabarinus, see text). If a haplotype is found more than once, the accession number is followed by the haplotype number (see Appendix) and the number of times the haplotype is found between parentheses. Accessions deposited on GenBank as T. turkestani are marked in grey. Accessions followed by the letter A or B indicate samples belonging to clade A and B respectively, deduced from Navajas (1998) and Navajas et al. (1998). Numbers on the branches indicate the percentage bootstrap values (>50) based on NJ bootstrapping with ML settings (1,000 replicates). Bar at the lower left corner depicts the branch length corresponding to 10% maximum likelihood distance

Fig. 6

Histogram of pairwise differences (p-distance) between 91 COI sequences within the family Tetranychidae. Pairwise differences are separated into three categories: 1. between individuals in the same species; 2. between individuals in the same genus (excluding intraspecific differences); 3. between individuals in the same family (excluding intraspecific and intrageneric differences). n = number of pairwise comparisons