Background

Insect–fungus interactions are ubiquitous in the environment and span the full range of symbiotic interactions, from obligate mutualism to obligate parasitism [1]. One end of the spectrum of symbiosis is obligate mutualism, such as fungus-farming insects, including attine ants (Hymenoptera; Formicidae; Attini), some species of termites (Blattodea; Macrotermitinae), and ambrosia beetles (Coleoptera; Scolytinae and Platypodinae) that engage in behaviors akin to agriculture [2]. Further along the spectrum are insects that still rely on fungi as nutritional sources but lack the specialization seen in the fungus-farming insects; examples of this intermediate mutualism include ship timber beetles (Coleoptera; Lymexylidea), woodwasps (Hymenoptera; Siricidae and Xiphydriidae), and gall midges (Diptera; Cecidomyiidea). Obligate parasitism falls on the opposite side of mutualism on the symbiotic spectrum and includes highly specialized fungal parasites that alter host behavior; these include Cordyceps (Ascomycota; Cordycipitaceae and Ophiocordyceps) and Entomophtora and Massospora (Entomophthoromycota) [3]. A myriad of less specialized facultative fungal pathogens infect insect hosts, making use of the fungal kingdom’s diverse enzymatic abilities to function as pathogens in a range of species. Studies of the molecular mechanisms of these symbioses can inform our understanding of ecology, evolution, and developmental biology, but as these research systems are emerging, they still lack well-established protocols that can be easily adopted and ensure reproducible results.

Determining the genetic and molecular underpinnings of emerging systems comes with a set of challenges not seen in systems with robust molecular procedures. Most insect–fungus systems cannot benefit from established protocols for extensive molecular biology–based research approaches, such as those used in studies of Tribolium and Drosophila [4], due to the range of potential community member abundance between systems. Tribolium and Drosophila have contributed greatly to the fields of developmental biology and genetics, but do not include the unique insect–fungus symbioses. Unfortunately, challenges exist in applying protocols from model systems to non-model systems due to differences in collection, sample composition and morphology, and the question of study [5].

Transcriptomic analysis through RNA sequencing, for example, can be used to determine genetic underpinnings of insect–fungus interactions by sampling at different points in the symbiosis on a temporal or morphological basis. However, insect–fungus samples of different systems will vary significantly in their relative composition of insect and fungal cells and subsequent RNA yields. Consider the following two insect–fungus interactions representing opposite ends of the symbiosis spectrum: The ambrosia beetle Euwallacea validus stores its obligate fungal mutualist Fusarium oligoseptatum in tiny pocket-like organs called mycangia within the head [6,7]. Samples are composed primarily of beetle tissue with relatively few fungal cells, which are aggregated in a discrete location. On the opposite end is the 17-year cicada Magicicada septendecim and its obligate fungal parasite, Massospora cicadina, which consumes and replaces the cicada’s abdomen with fungal spores [8,9]. The remaining “fungal plug” is mostly fungal tissue, while relatively little cicada tissue remains inside in a nonhomogeneous distribution [10].

In hopes of reducing the need for protocols individualized to each system and to promote the study of insect–fungus interactions, a standard transcriptomic analysis workflow for insect–fungus samples of varying composition has been created. These two systems described above, with opposing insect–fungus compositions, were chosen to test the applicability of this protocol across varied sample types representing extreme scenarios. We found that the compositional ratio of fungal tissue to insect tissue did not influence the experimental outcome of this protocol, as RNA extraction of both sample types yielded sufficiently high-quality material that could be sequenced, making it widely applicable across a range of insect–fungus interactions. However, we found that the proportion of transcripts recovered for each partner is dependent on its abundance in the sampled tissues (see Validation, section C).

This protocol is designed to be accessible to all scientists. Price and accessibility of materials were considered. A TRIzol extraction was chosen due to its affordability and applicability to a range of sample types over RNA extraction kits. Two means of quantification are produced. Data processing of RNASeq samples utilizes the open-source workflow and pipeline manager Nextflow [11,12]. Its community and extensive documentation allow its use by those with little to no bioinformatics background. With high accessibility and potential wide application, this protocol aims to bridge the gap of established protocols observed when comparing non-model and model systems, allowing a more effective study of insect–fungus interactions.