Identification of potential insect ecological interactions using a metabarcoding approach

Abstract

Species interactions are challenging to quantify, particularly when they happen cryptically. Molecular methods have become a key tool to uncover these interactions when they leave behind a DNA trace from the interacting organism (e.g., pollen on a bee) or when the taxa are still present but morphologically challenging to identify (e.g., microbial or fungal interactions). The decreasing costs of sequencing makes the mass analysis of thousands of target species possible. However, the challenge has shifted to selecting molecular markers which maximize information recovery while analyzing these data at broad biological scales. In this manuscript we use model arthropod groups to compare molecular markers and their analysis across life stages. We develop protocols for two ecologically and economically devastating pests, the spongy moth (Lymantria dispar dispar) and the emerald ash borer (Agrilus planipennis), and a group of pollinators including bees and wasps which regularly deposit eggs in "bee hotels" where the larvae develop. Using Illumina MiSeq and Oxford Nanopore MinION platforms we evaluate seven primer pairs for five molecular markers which target plants, fungi, microbes, insects, and parasitic phyla (e.g., nematodes). Our data reveals hundreds of potential ecological interactions and establishes generalized methods which can be applied across arthropod host taxa with recommendations on the appropriate markers in different systems. However, we also discuss the challenge of differentiating co-occurring DNA signals and true ecological interactions, a problem only starting to be recognized as eDNA from the environment accumulates on living organisms.

Keywords: DNA barcoding; Insects; Metabarcoding; Species interactions.

Figure 1. Co-amplified taxa from host spongy moths.

(A) In addition to the amplification of host DNA, COI primers co-amplified DNA from six orders of insects as minor background signals. (B) ITS primers recovered 19 orders of plant and fungi in the DNA samples of moths with more taxa detected in egg masses than larval samples. (C) 16S primers targeting the microbiome detected a remarkably similar microbial community in egg vs. larval stages. (D) rbcL primers targeting plant DNA detected a similar, but not completely overlapping plant community to ITS primers (B) with the same increased taxonomic richness in egg masses compared to larval samples.

Figure 2. Co-amplified taxa from host emerald ash borer across life stages.

(A) In addition to the amplification of host DNA, COI primers co-amplified four orders of insects as minor background signals. (B) ITS primers targeting primarily fungal DNA recovered 11 orders in the DNA samples of larvae, pupae, and adults. (C) 16S primers targeting the microbiome found an increasingly complex microbial community with life stage. (D) 18S primers amplified the widest range of taxa including a series of nematodes which are likely parasites. Low taxonomic recovery from larva size 3 specimens likely reflects the small sample size.

Figure 3. Co-amplified taxa from pollinator larva in artificial bee homes.

(A) ITS primers targeting primarily plant DNA recovered 21 orders in the DNA samples of larva collected from tubes in bee homes; (B) rcbL primers targeting plant DNA detected 22 orders of plants of which 14 were also detected by ITS; (C) ITS primers targeting primarily fungal DNA detected 10 orders; (D) 16S primers targeting the microbiome recovered nine orders of bacteria; (E) in addition to the amplification of host DNA COI primers co-amplified 16 additional arthropods including known kleptoparasites, prey, and some evidence of cross amplification from other neighbouring larva (e.g., S. bifasciatus is a host but also detected in M. relativa larva); (F) a greater number of genera were recovered from individual samples than pooled samples across all primers tested.

Figure 4. Highly similar co-amplification in eggs and larval spongy moths.

The recovery of various targets by target amplified region (left) shows high variability in taxonomic richness. Interestingly the profile of taxonomic recovery of co-amplified taxa is highly similar between egg masses and larval stage spongy moths (right) suggesting little change in the richness of interacting taxa over development.