More than 20,000 species of bees (Hymenoptera, Anthophila) are estimated to occur worldwide [1] and they became a posterchild for conservation biology as they are regarded as the economically most important group of pollinators in both natural and managed ecosystems [2]. Unfortunately, bee diversity and abundance has been reported to decline at different levels across continents [3]. To better understand changes in occurrence of bee species across scales exhaustive faunistic inventories are necessary. However, bee taxonomy can be difficult and different ways to accelerate their biomonitoring have been suggested [4, 5]. The most prominent approach is DNA barcoding, where a specific segment of the mitochondrial cytochrome oxidase 1 gene (cox1) is used for species identification [6]. As such, several geographic region specific barcoding initiatives for bees have been launched or already successfully finished, e.g., Central Europe [7], Ireland [8], Canada [9], providing the necessary background for DNA-based identification.
While in general the barcoding approach has been found working well for the so far analysed regional bee faunae, several practical problems have been identified. Most notoriously, when using standard barcoding primers (eg., LCO1490 and HCO2198 [10], or LepF1 and LepR1 [11]) sequencing failures are reported and misamplification of endosymbiotic bacteria of the genus Wolbachia has been found to interfere with the amplification of the targeted bee DNA [8, 12]. Screening submitted data to the BOLD database revealed that Hymenoptera account for the highest number of cases where it came to un-intended amplification of Wolbachia DNA [13]. Wolbachia are maternally inherited intracellular Alphaproteobacteria that are found to be widespread in (terrestrial) arthropods, but also are also known from selected nematode taxa [14]. By screening the German bee fauna, a prevalence of > 60% of Wolbachia infected species has been reported [15, 16]. There seem to be taxon-specific patterns regarding the frequency of infected species, and especially a high number of species from the species-rich genera Andrena, Halictus, Lasioglossum, Nomada or Sphecodes are infected [16]. In congruence with what was reported by [8] our own routine work in the lab showed that individuals of these genera are difficult to barcode using standard approaches, as sequencing revealed mixed signals or the Wolbachia sequence. This comes to no surprise, as the most common used standard barcoding primers (LCO1490/HCO2198, LepF1/LepR1) actually show a high similarity to the homologous region of the Wolbachia coxA gene (Fig. 1). Alternative primer pairs have been already suggested as a workaround, but these seem to need rather low (and therefore unspecific) annealing temperatures (the primer pair BarBeeF/MtD9, as proposed for corbiculate bees by Francoso & Arias [17]) or involve highly degenerate priming sites (Hymeno-1/HCO-Hymeno-2 as proposed by Vilalta et al. [12]), which often hampers ultimate sequencing success. Here we present a newly designed primer combination and test its suitability for a phylogenetically representative taxon sampling of bees, with a special focus on its suitability in the case of Wolbachia infected species.