Quantitative multi-locus metabarcoding and waggle dance interpretation reveal honey bee spring foraging patterns in Midwest agroecosystems

Abstract

We explored the pollen foraging behavior of honey bee colonies situated in the corn and soybean dominated agroecosystems of central Ohio over a month-long period using both pollen metabarcoding and waggle dance inference of spatial foraging patterns. For molecular pollen analysis we developed simple and cost-effective laboratory and bioinformatics methods. Targeting four plant barcode loci (ITS2, rbcL, trnL and trnH), we implemented metabarcoding library preparation and dual-indexing protocols designed to minimize amplification biases and index mis-tagging events. We constructed comprehensive, curated reference databases for hierarchical taxonomic classification of metabarcoding data and used these databases to train the Metaxa2 DNA sequence classifier. Comparisons between morphological and molecular palynology provide strong support for the quantitative potential of multi-locus metabarcoding. Results revealed consistent foraging habits between locations and show clear trends in the phenological progression of honey bee spring foraging in these agricultural areas. Our data suggest that three key taxa, woody Rosaceae such as pome fruits and hawthorns, Salix, and Trifolium provided the majority of pollen nutrition during the study. Spatially, these foraging patterns were associated with a significant preference for forests and tree lines relative to crop fields and herbaceous land cover.Quantitative pollen metabarcoding

Experimental quantification of pollen with DNA metabarcoding using ITS1 and trnL

“…In this study, we investigated whether it is possible to predict the abundance of DNA sequencing reads from the quantities of pollen grains and if so, how accurately, and how the prediction is affected by the methodology used to estimate pollen abundances in mock solutions, PCR conditions, plant species and the type of markers (nuclear versus plastid). In agreement with previous studies [20][21][22]25,40 but in contrast to others 17,26,30,31 , we found a significant, and often strong positive relationship between the number of DNA sequences and the number of pollen grains in the mock solutions. However, the strength of the relationship was influenced by the pollen counting methodology, the marker, the species and the number of PCR cycles.…”

“…www.nature.com/scientificreports www.nature.com/scientificreports/ Effects of molecular markers and PCR conditions. We found that for trnL, very accurate prediction of DNA read abundance was obtained from pollen grain quantities estimated by light microscopy, with determination coefficients and slopes generally higher than those found in the literature with the same markers 40 or with other markers 17,21,26,30,31 . Moreover, except at PCR25, the good relationship was conserved across species and PCR cycles and trnL sequence abundance increased steadily with increasing PCR cycles.…”

“…which aim to evaluate the potential of metabarcoding for pollen quantification, the estimates of pollen abundance in suspension and the DNA extractions have been performed using different subsamples 17,[20][21][22][25][26][27]30,31,40 , resulting in unavoidable variation in pollen concentrations. Moreover, since counting pollen grains under a light microscope is very time consuming and as mock solutions often have high concentrations of pollen, only small subsamples and a small proportion of microscope slides (and consequently of pollen population) are usually inspected 23 .…”

“…www.nature.com/scientificreports www.nature.com/scientificreports/ Recommendations and methodological considerations when using metabarcoding for pollen quantification. Despite its high potential, we know only three other studies that (successfully) used trnL in pollen quantification 20,22,40 . Therefore, our findings call for further studies to determine whether trnL can be applied to many other plant species and routinely used to take greater account of interaction strength in pollination network studies.…”

“…It is worth noting that, probably due to their relatively low GC content, plastid loci such as trnL 20,22,40 (≈35% GC content), rbcL 26,40 (≈42% GC content), matK (≈35% GC content) or trnT 23,48 (≈25% GC content) may provide a better estimation of pollen number than nuclear markers (ITS1 and ITS2 ≈ 60% GC content). However, the drawbacks of ITS could to some extent be alleviated 49 by using high-fidelity DNA polymerases such Phusion High-Fidelity DNA polymerase 25,40 or Herculase II Fusion DNA polymerase (the present study), 3% (our study) or 5% DMSO 50 and by applying low primer annealing temperatures 45 . The detrimental impact of the accumulation of polymerase inhibitors could be reduced by increasing the number of PCR steps (3 to 5 successive PCRs) and by diluting a subsample of the previously obtained amplicons in fresh reagent mixtures at each step 45 .…”