The authors have declared that no competing interests exist.
In this study, the first complete mitogenome of Andrenidae, namely
Bees (Hymenoptera: Apoidea: Anthophila) are widely distributed and comprise approximately 20,000 described species [
The extant bees are generally classified into seven families (Apidae, Megachilidae, Colletidae, Melittidae, Andrenidae, Halictidae, and Stenotritidae) [
(A) bees were divided into long-tongued bees and short-tongued bees; (B) Melittidae was inferred as the basal lineage of bees or sister to other bee families; (C) and (D) Andrenidae was suggested as sister to all other bees except Melittidae or sister to Melittidae, respectively.
Although most studies based on morphology and nuclear genes suggested that Melittidae was sister to all other bee families, and the remaining bees were classified into two groups: (Apidae + Megachilidae) and (Andrenidae + (Halictidae + (Stenotritidae + Colletidae))) [
To date, only 12 complete mitogenomes have been sequenced for bees, (GenBank, December 1, 2017), and the corresponding phylogenetic analyses are still limited. In this study, we present the first complete mitogenome of Andrenidae (
The specimen of
Ten pairs of primers were used (
The overlapping PCR products were assembled using SeqMan program included in the Lasergene software package (DNAStar Inc., Madison, Wisc.). The transfer RNA genes (tRNAs) were predicted using the Mitos WebServer [
A total of 64 species were analyzed in this study, including 62 bees and two outgroups from the families Vespidae (
Nucleotide sequences for each of the 13 PCGs, two rRNAs, and 22 tRNAs were imported into separate files using BioEdit 7.1.3.0 [
To eliminate poorly aligned positions and divergent regions, Gblocks 0.91b [
In order to test the effects of the third codon positions, gene types, and the Gblocks treatment on phylogeny, 16 datasets were carried out: 1) all codon positions of PCGs, with the Gblocks treatment (P123_G); 2) P123 and rRNAs (P123R_G); 3) P123 and tRNAs (P123T_G); 4) P123, rRNAs, and tRNAs (P123RT_G); 5) first and second codon positions of PCGs (P12_G); 6) P12R_G; 7) P12T_G; 8) P12RT_G; 9) P123 analyzed without Gblocks (P123_UnG); 10) P123R_UnG; 11) P123T_UnG; 12) P123RT_UnG; 13) P12_UnG; 14) P12R_UnG; 15) P12T_UnG; 16) P12RT_UnG. The best partitioning schemes and nucleotide substitution models were simultaneously confirmed with PartitionFinder 2.1.1 [
Two inference methods, maximum likelihood (ML) and Bayesian inference (BI), were performed using RAxML 8.2.9 [
The mitogenome of
Gene identity is obtained by BLAST searches, with the reference genome of
To better visualize the gene identity in mitogenomes of bees, the comparable circular map was generated (
Gene rearrangement could be classified into local inversion (inverted in the local position), remote inversion (translocated and inverted), gene shuffling (local translocation), and translocation [
Compared with the putative ancestral gene arrangement of insects (
PCGs, rRNAs, tRNAs, and the control region are marked with yellow, pink, green, and grey, respectively. Gene with underscore indicates that it is encoded in the N strand.
Substitution saturation index (
Gene regions | NumOTU | Psym | Pasym | ||||
---|---|---|---|---|---|---|---|
32 | 0.517 | 0.809 | 0.0000 | 0.554 | 0.0001 | ||
32 | 0.518 | 0.809 | 0.0000 | 0.554 | 0.0008 | ||
32 | 0.751 | 0.809 | 0.0000 | 0.554 | 0.0000 | ||
32 | 0.533 | 0.818 | 0.0000 | 0.572 | 0.0000 | ||
32 | 0.874 | 0.802 | 0.0000 | 0.539 | 0.0000 | ||
32 | 0.704 | 0.79 | 0.0000 | 0.52 | 0.0000 | ||
32 | 0.484 | 0.808 | 0.0000 | 0.551 | 0.0000 | ||
32 | 0.348 | 0.808 | 0.0000 | 0.551 | 0.0000 | ||
32 | 0.735 | 0.808 | 0.0000 | 0.551 | 0.0000 | ||
32 | 0.496 | 0.817 | 0.0000 | 0.571 | 0.0000 | ||
32 | 0.774 | 0.777 | 0.7069 | 0.496 | 0.0000 | ||
32 | 0.634 | 0.782 | 0.0000 | 0.507 | 0.0000 |
Notes: NumOTU, number of OTUs;
In order to test the effects of gene types, the combined analyses of PCGs + rRNAs, PCGs + tRNAs, and PCGs + rRNAs + tRNAs were compared with PCGs alone (
Clade | P123 | P12 | P123R | P12R | P123T | P12T | P123RT | P12RT | ||||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
G | UnG | G | UnG | G | UnG | G | UnG | G | UnG | G | UnG | G | UnG | G | UnG | |
100 | 100 | 100 | 100 | 100 | 100 | 100 | 100 | 100 | 100 | 100 | 100 | 100 | 100 | 100 | 100 | |
98 | 99 | 98 | 98 | 96 | 98 | 97 | 98 | 99 | 98 | 99 | 97 | 98 | 100 | 99 | 98 | |
100 | 100 | 100 | 100 | 100 | 100 | 100 | 100 | 100 | 100 | 100 | 100 | 100 | 100 | 100 | 100 | |
- | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | |
99 | 99 | 94 | 96 | 99 | 100 | 99 | - | 100 | 100 | 98 | 98 | 100 | 100 | 100 | - | |
- | - | - | - | - | - | - | 99 | - | - | - | - | - | - | - | 100 | |
- | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | |
57 | 47 | 42 | 39 | 50 | 40 | 40 | - | 60 | 54 | 51 | 51 | 56 | 40 | 43 | - | |
- | - | - | - | - | - | - | 36 | - | - | - | - | - | - | - | 42 | |
100 | 100 | 100 | 100 | 100 | 100 | 100 | 100 | 100 | 100 | 100 | 100 | 100 | 100 | 100 | 100 | |
100 | 100 | 100 | 100 | 100 | 100 | 100 | 100 | 100 | 100 | 100 | 100 | 100 | 100 | 100 | 100 | |
84 | 81 | 86 | 83 | 97 | 98 | 99 | 99 | 91 | 88 | 94 | 94 | 99 | 99 | 100 | 99 | |
100 | 100 | 100 | 100 | 100 | 100 | 100 | 100 | 100 | 100 | 100 | 100 | 100 | 100 | 100 | 100 | |
100 | 100 | 100 | 100 | 100 | 100 | 100 | 100 | 100 | 100 | 100 | 100 | 100 | 100 | 100 | 100 | |
1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | |
1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | |
1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | |
- | - | - | - | - | - | 1 | 1 | - | - | - | - | - | - | - | - | |
1 | 1 | 1 | 1 | 1 | 1 | - | - | 1 | 1 | 1 | 1 | 1 | 1 | 1 | - | |
- | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | |
- | - | - | - | - | - | 0.66 | 0.73 | - | - | - | - | - | - | - | - | |
0.98 | 0.99 | 0.68 | 0.53 | 0.74 | 0.88 | - | - | 0.99 | 1 | 0.86 | 0.93 | 0.93 | 0.98 | 0.69 | - | |
- | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | |
1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | |
1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | |
1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | |
1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | |
1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 |
Notes: G, usage of Gblocks; UnG, without Gblocks; -, not recovered; LT-bees, long-tongued bees; ST-bees, short-tongued bees; ME, Melittidae excluded; Mel, Melittidae; Hal, Halictidae; And, Andrenidae; Col, Colletidae; P123, all codon positions of PCGs; P12, first and second codon positions of PCGs; P123R, P123 and rRNAs; P12R, P12 and rRNAs; P123T, P123 and tRNAs; P12T, P12 and tRNAs; P123RT, indicates P123, rRNAs, and tRNA; P12RT, indicates P12, rRNAs, and tRNAs.
Gblocks generally reduced the degree of substitution saturation and presented positive effects on tree topology, but had opposite effects on nodal support between the two inference methods. Under the ML framework, datasets with Gblocks treatment presented the same tree topology and had much higher nodal support. In BI, P12RT that analyzed without Gblocks treatment even failed to present a clear relationship among Andrenidae, Halictidae, and Colletidae. However, the nodal support of Halictidae + Colletidae generated with Gblocks was slightly lower.
The two inference methods produced the same tree topology, except for the analyses of P12R_G, P12R_UnG, and P12RT_UnG. It was indicated that when rRNAs excluded, the tree topology showed little sensitivity to inference methods. In addition, compared with ML method, the BI generally presented higher nodal support, such as the node of Halictidae + Colletidae.
Phylogenetic analyses were performed on 64 complete or nearly complete mitogenomes, representing six bee families. In order to assess the effects of datasets (based on codons and gene types), Gblocks treatment (Gblocks and UnGblocks), and inference methods (BI and ML), a total of 32 independent phylogenetic analyses were carried out. Based on the relationships among bee families, four different tree topologies were recovered. As shown in
Numbers on branches are Bayesian posterior probabilities.
The effect of RNA genes on tree topology and nodal support is a long-standing debate [
Whether to include the third codons is also an ongoing debate. Some studies proposed that exclusion of third codons could produce more consistent topologies [
Based on algorithm, Gblocks was used to increase signal-to-noise ratio by eliminating poorly aligned positions and highly divergent regions. These regions might be nonhomologous or include inaccurately defined gene boundaries, or have been saturated by multiple substitutions [
For the same dataset, especially when the third codons of PCGs included, the tree topology had low sensitivity to inference methods. The nodal support in BI was generally higher than in ML. However, it had also been suggested that posterior probabilities were somewhat liberal [
In this study, we presented the most comprehensive mitochondrial phylogeny of the family-level relationships of bees. Our analyses highly supported the monophyly of each family, except for Melittidae which had only one species analyzed. This was consistent with other morphological and molecular studies [
The monophyly of long-tongued bees was highly supported (BP = 100, PP = 1.0), which had also been supported by both morphological and other molecular studies [
This study firstly presented a comprehensive mitochondrial phylogeny of the family-level relationships of bees. As described above, the tree topology described in
(PNG)
(DOCX)
(DOCX)
(DOCX)