Parasite presence induces gene expression changes in an ant host related to immunity and longevity
Most species are either parasites or exploited by parasites, making parasite-host interactions a driver of evolution. Parasites with complex life cycles often evolve strategies to facilitate transmission to the definitive host by manipulating their intermediate host. Such manipulations could explain phenotypic changes in the ant Temnothorax nylanderi, the intermediate host of the cestode Anomotaenia brevis. In addition to behavioral and morphological alterations, infected workers exhibit increased survival, comparable to that of queens, which live up to two decades. We used transcriptomic data from cestodes and ants of different castes and infection status to investigate the molecular underpinnings of phenotypic alterations in infected workers, and whether the extended lifespan of queens and infected workers has a common molecular basis. To this end we sequenced the abdominal RNA of infected and uninfected ants of parasitized nests and queens and workers from unparasitized nests and assembled a transcriptome using StringTie. After filtering the transcriptome for those transcripts with an ORF greater or equal than 150 bp and filtering the gene count matrix for only those genes that had at least 5 counts in at least 3 samples, we first conducted a Principal Component Analysis. We tested each PC for its association with infection status and for the significant PCs we extracted the loadings of each gene (dataset S1). Moreover, we analyzed gene expression using DESeq2, comparing infected workers with their uninfected nestmates and on the other hand queens and their workers from unparasitized nests. The significant upregulated genes in each group can be found in the dataset S2. GO enrichment was performed using topGO on both the genes and their loadings on PC2 using a Kolomogorov-Smirnov test and the upregulated genes compared to the whole transcriptome using Fisher's exact tests (dataset S3). Afterward, we analyzed the overlap between genes upregulated in infected workers and in queens. Infected workers and queens commonly upregulated only six genes, one of them with a known anti-aging function. Our text-mining approach revealed that both groups overexpressed immune genes, although not the same ones. Our findings suggest that the lifespan extension of infected workers is not achieved via the expression of queen-specific genes. To be able to filter the abdominal RNA sequences of infected workers from cestode RNA, we additionally sequenced RNA of 320 cysticercoids and de novo assembled a transcriptome of A. brevis. The analysis of the cestodes’ transcriptome revealed dominant expression of genes of the mitochondrial respiratory transport chain indicating an active metabolism and shedding light on the physiology of the parasite in its cysticercoid stage (dataset S4).
Steps to reproduce
The analysis of the cestode transcriptome included the following steps: - filter contaminant RNA using FastQScreen (including ants) - trim data using Trimmomatic - de novo assemble transcriptome using Trinity - assign functions based on homology using blastx against the nr invertebrate database (E-value cut-off 10^-5) - filter transcriptome to remove all transcripts that have a best BLAST hit against a Hymenopteran species - translate transcript sequences into amino acid sequences using TransDecoder - get functional annotation using InterProScan - map reads against transcriptome using Bowtie2 - count reads using RSEM The analysis of the ant transcriptomes included the following steps: - filter for contaminants (including cestodes) using FastQScreen - trim reads using Trimmomatic - map reads against genome using HISAT2 - assemble transcriptome using StringTie - assign functions based on homology using blastx against the nr invertebrate database (E-value cut-off 10^-5) - translate transcript sequences into amino acid sequences using TransDecoder - get functional annotation using InterProScan - analyse gene expression using the gene count matrix (dataset S5) using DESeq2 - additional analysis including colony size as batch effect in the comparison between queens and nurses indicated by "batch" All analytic scripts for analyses done in R are provided.