Comparative Mitochondrial Genomics And Phylogeny Of The Main Pollinator Of Camellia Oleifera

The oil-tea camellia(Camellia oleifera)is a unique woody oil crop in China.It has self-incompatibility,with its sexual reproduction dependent on the pollinators.It has been showed that pollination is the main factor affecting the yield.The main pollinators of oil-tea camellia are Colletes gigas,Andrena camellia,Andrena hunanensis,Andrena striata,and Andrena chekiangensis.These wild bees play an important role in improving the pollination efficiency,remov:ing the limitation of pollen,and improving the fruit setting rate and yield.Although our field observations found that A.chekiangensis,A.hunanensis,and A.striata are also widely distributed in oil-tea forest,they are rarely described as the pollinators.The reason may be that these three flower visitors exhibit similar morphologic features with A.camellia.They may be incorrectly identified as A.cfamellia.Therefore,it is very important to elucidate the taxonomic status and phylogenetic relationship of the pollinators of C.oleifera.In this study,we sequenced the mitochondrial genomes of four endemic species of the genus Andrena,which were also the pollinators of C.oleifera,and carried out detailed comparative genomics and phylogenetic analysis in combination with other reported bee species.The main research results are as follows:(1)In this study,the mitochondrial genomes of four genera species were sequenced,among which the sequences of A.camellia(15 065 bp)and A.chekiangensis(15 804 bp)were complete.However,due to high A + T content and complicated secondary structure,the control regions of A.hunanensis(14 780 bp)and A.striata(14 736 bp)were unable to be amplified.The mitochondrial genome contained 13 protein-coding genes,22 tRNA genes,and 2 rRNA genes.The A+T content of these genomes was 75.86%-78.58%,and the deviation degree of base composition was similar to that of other Andrena species,i.e.,positive AT-skews,and negative GC?skews.Compared with putative ancestral gene arrangement of insects,the three mitogenomes presented identical gene rearrangements,including local inversion(trnR)and gene shuffling(trnQ/trnM,trnK/trnD,and trnW/trnC-trnY),which were also consistent with other mitogenomes reported in the same genus of Andrena.Comparative analysis showed that the mitochondrial genes of bees presented a complex rearrangement phenomenon,with most rearrangement events located in the following tRNA clusters:trnI-trnQ-trnM,trn W-trnC-trnY,and trnA-trnR-trnN-trnS1-trnE-trnF.(2)The newly sequenced Andrena mitogenomes exhibited the typical start codon ATN.Most of the protein-coding genes used TAA or TAG as the stop codon,while the incomplete termination codon T was detected in atp6 gene of A.chekiangensis.In addition,the nad4 gene used a single T in all species.Relative synonymous codon usage indicated that A/T was more frequently used than G/C in degenerate codons.The third codon positions of the six most commonly used codons in Andrena,including TTA,TCA,GTT,ACA,CCA,and CGA,were all composed of A or T.In contrast,some GC-rich codons were rarely used in Andrena species,such as CGC,GGC,CCG,and GCG.Among the protein-coding genes,high conservation was observed in cytochrome oxidase genes with coxl exhibited the highest conservation.On the contrary,NADH dehydrogenase and ATPase subunit genes were more variable with atp8 presented the maximal variation(3)In this study,we presented the most comprehensive mitochondrial phylogeny of the family-level relationships of bees and the species-level relationships within the genus Andrena.Phylogenetic analysis showed that each bee family was monogenic.Given that the sister relationship between Melittidae and other bee families was highly supported in all datasets,the short-tongued bees were inferred as a paraphyletic group,in which the family Andrenidae was inferred as the sister to Halictidae +Colletidae.The mitochondrial phylogeny of bees in our analyses exhibited promising congruence with most of the molecular studies,suggesting that mitogenomes were suitable for resolving higher-level relationships within bees.Within the genus Andrena,A.hunanensis and A striata formed a group and were close to A.camellia.Although A chekiangensis was also difficult to be distinguished from A.camellia by traditional morphological methods,A.chekiangensis and A.haemorrhoa formed a clade and were sister to the other taxa of the genus Andrena.In addition,in order to improve the resolution of phylogenetic analysis,various data set and analysis methods were also used in this study.It showed that the data set including the third codon of the protein-coding gene was beneficial to produce the same topology and higher node support.The tRNA genes with typical clover structure also showed similar positive effects.However,rRNA genes presented poor sequence alignment and distinct substitution saturation,which had negative effects on both tree topology and nodal support.In addition,Gblocks treatment could increase the consistency of the topology,but had opposite effects on the nodal support between the two inference methods of maximum likelihood and Bayesian inference.