Pollen DNA metabarcoding of Australian honey

Supporting information for "Pollen DNA metabarcoding identifies regional provenance and high plant diversity in Australian honey". Abstract: Accurate identification of the botanical components of honey can be used to establish its geographical provenance, while also providing insights into honey bee (Apis mellifera L.) diet and foraging preferences. DNA metabarcoding has been demonstrated as a robust method to identify plant species from pollen and pollen-based products, including honey. We investigated the use of pollen metabarcoding to identify the floral sources and local foraging preferences of honey bees using 15 honey samples from six bioregions from eastern and western Australia. We used two plant metabarcoding markers, ITS2 and the trnL P6 loop. Both markers combined identified a total of 55 plant families, 67 genera and 43 species. The trnL P6 loop marker provided significantly higher detection of taxa, detecting an average of 15.6 taxa per sample, compared to 4.6 with ITS2. Most honeys were dominated by Eucalyptus and other Myrtaceae species, with a few honeys dominated by Macadamia (Proteaceae) and Fabaceae. Metabarcoding detected the nominal primary source provided by beekeepers amongst the top five most abundant taxa for 85% of samples. We found that eastern and western honeys could be clearly differentiated by their floral composition, and clustered into bioregions with the trnL marker. Comparison with previous results obtained from melissopalynology shows that metabarcoding can detect similar numbers of plant families and genera, but provides significantly higher resolution at species level. Our results show that pollen DNA metabarcoding is a powerful and robust method for detecting honey provenance and examining the diet of honey bees. This is particularly relevant for hives foraging on the unique and diverse flora of the Australian continent, with the potential to be used as a novel monitoring tool for honey bee floral resources.\nLineage: Honey selection\nWe selected 15 honey samples from twelve localities located within six Australian IBRA regions. The raw honey samples were collected in January (n = 3), September (n = 1) and October (n = 11) 2015 (Table 1). Each sample was collected in clean food-grade containers filled directly from beekeeper extractions from single apiaries, and sampling equipment was cleaned between each collection to avoid cross-contamination as detailed in Sniderman et al. (2018). Samples were stored at room temperature until 2019, when they were transferred to a -20 ºC freezer. To take a sub-sample, the containers were first placed in a water bath at 40 ºC for 30 minutes to reduce viscosity. Between 6.4 g and 11 g of honey were taken from each container using a sterile large syringe, and transferred to sterile 50 mL tubes, which were stored at -20 ºC until DNA extraction.\n\nDNA extraction and sequencing\nDNA extractions from each sub-sample were conducted in November 2019 and March 2020. The pollen isolation protocol for honey was modified from de Vere et al. (2017). Samples were diluted with sterile water to make up to 15 ml. These mixtures were incubated for 30 minutes in a water bath at 65 C, with occasional vortexing. After incubation, mixtures were centrifuged for 15 minutes at 7000 rpm at room temperature. The supernatant was discarded and the remaining pellet resuspended in 400 L of lysis buffer (CF buffer in the Macherey-Nagel DNA Food extraction kit) and transferred to a 1.5 mL tube, to which 20 L of Proteinase K was added. The samples, including a blank lysis control (consisting of 400 L buffer CF and 20 L Proteinase K), were then incubated in a shaker oven at 65 C for 1 hour. DNA extraction then proceeded as per the manufacturer’s protocol (Macherey-Nagel, Düren, Germany). Final DNA concentrations were quantified using a Quantus (Promega Corporation, Manheinn, Germany) fluorometer.\n\nDNA metabarcoding and sequencing\nAn initial PCR was performed in duplicate reactions on each DNA extraction, to determine the DNA dilution for optimal amplification by adding DNA template either directly to the PCR master mix, or as a 1/10 dilution. PCR master mixes comprised: 2.5 mM MgCl2 (Applied Biosystems, USA), 10x PCR Gold buffer (Applied Biosystems), 0.25 mM dNTPs (Astral Scientific, Australia), 0.4 mg/ml bovine serum albumin (Fisher Biotec, Australia), 0.4 μmol/L forward and reverse primers, 0.6 μl of a 1:10,000 solution of SYBR Green dye (Life Technologies, USA) and 1 U AmpliTaq Gold DNA polymerase (Applied Biosystems). All PCR reactions had a volume of 25µL and were performed using StepOne Plus Instruments (Applied Biosystems). PCR cycling conditions consisted of denaturation at 95C for 5 minutes, followed by 50 cycles of: 95C for 30s, 52C (trnL) or 55C (ITS2) for 30s, 72C for 45s, finishing with a final extension stage at 72C for 10 min. \n\nIndexing of samples was achieved using unique, single use combinations of 8 bp multiplex identifier tagged (MID-tag) primers as described in Koziol et al. (2019) and van der Heyde et al. (2020). MID-tag PCR reactions were prepared using a Qiagility instrument (Qiagen) using the same master mix and PCR conditions as described. Negative and positive PCR controls for both trnL and ITS2 were included to ensure validity of results. Sequencing libraries were pooled equimolarly based on the PCR amplification results. Libraries were size selected using a Pippin instrument (Sage Sciences, USA), quantified on a Qubit (Thermofisher) and diluted to 2nM. Libraries were sequenced on an Illumina MiSeq instrument using 300 cycle V2 kit (trnL) or 500 cycle kit (ITS2) with custom sequencing primers.