Trans-regulatory changes underpin the evolution of the Drosophila immune response

Abstract

When an animal is infected, the expression of a large suite of genes is changed, resulting in an immune response that can defend the host. Despite much evidence that the sequence of proteins in the immune system can evolve rapidly, the evolution of gene expression is comparatively poorly understood. We therefore investigated the transcriptional response to parasitoid wasp infection in Drosophila simulans and D. sechellia. Although these species are closely related, there has been a large scale divergence in the expression of immune-responsive genes in their two main immune tissues, the fat body and hemocytes. Many genes, including those encoding molecules that directly kill pathogens, have cis regulatory changes, frequently resulting in large differences in their expression in the two species. However, these changes in cis regulation overwhelmingly affected gene expression in immune-challenged and uninfected animals alike. Divergence in the response to infection was controlled in trans. We argue that altering trans-regulatory factors, such as signalling pathways or immune modulators, may allow natural selection to alter the expression of large numbers of immune-responsive genes in a coordinated fashion.

Fig 1. The immune response of two species of Drosophila and their hybrids.

(A) The proportion of larvae that had encapsulated eggs of the parasitoid wasp L. boulardi 96 hours post infection (96hpi). (B and C). The proportion of encapsulated droplets of mineral oil containing wasp homogenate in (B) larvae 24 hours post injection or (C) in adult flies that developed from injected larva. Error bars are 95% binomial confidence intervals.

Fig 2. Divergence of the transcriptional response to immune challenge in fat body and hemocytes.

Larvae were immune challenged by injecting them with mineral oil containing wasp homogenate while control larvae were not injected. The transcriptional response was then measured by RNA sequencing 24 hours post injection. (A) Volcano plots comparing the magnitude and statistical significance of changes in gene expression after immune challenge. Each point is a gene, and they are coloured above an FDR threshold of 5%. (B) The magnitude of the transcriptional response to immune challenge in D. simulans compared to D. sechellia. Each point is a gene.

Fig 3. Changes in gene expression after immune challenge.

(A) Proportion of lamellocytes in different transcriptional states in D. simulans and D. sechellia larvae under control and immune-challenged conditions, estimated from bulk RNA-seq data with CIBERSORTx. Error bars are 95% confidence intervals. (B) Changes in the expression of immunity genes in the D. simulans and D. sechellia fat body after immune challenge. Bars represent the scaled mean count per million of gene expression in control larvae and 24 hours after injection with mineral oil containing wasp homogenate. Only genes with a greater than two-fold change in expression in at least one species are shown. Significance of the change in expression after infection was defined by false discovery rate (FDR): 0.05*, 0.01**, 0.001***.

Fig 4. Expression divergence of immune responsive genes between D. simulans and D. sechellia in hemocytes.

Plots only show genes that are significantly differentially expressed in response to infection in at least one species. (A) Expression differences between parental species (total regulatory divergence) plotted against allelic expression differences within F1 hybrids (cis-regulatory divergence). Gene expression was measured under unchallenged control conditions (left) and immune-challenged conditions (centre), and this was used to estimate the expression change in response to immune challenge (right). Each point is a gene, the expression difference of each gene is represented as the relative log₂ fold change. (A and B) The genes are colour coded according to whether gene expression divergence is controlled in cis or in trans, with panel (B) depicting the number of genes in each category. The ambiguous category refers to genes with significant divergence in expression between the parental species but no significant divergence in either cis or trans. (C) Histograms depicting the frequencies of cis- and trans-diverged genes across the magnitudes of total expression divergence between D. simulans and D. sechellia. Fly larvae were immune-challenged by injecting them with oil droplets containing wasp homogenate.

Fig 5. Expression divergence of immune responsive genes between D. simulans and D. sechellia in the fat body.

Plots only show genes that are significantly differentially expressed in response to infection in at least one species. (A) Expression differences between parental species plotted against allelic expression differences within F1 hybrids (cis-regulatory divergence). Gene expression was measured under unchallenged control conditions (left) and immune-challenged conditions (centre), and this was used to estimate the expression change in response to immune challenge (right). (A and B) The genes are colour coded according to whether gene expression divergence is controlled in cis or in trans, with panel (B) depicting the number of genes in each category. (C) Histograms depicting the frequencies of cis- and trans-diverged genes with differing expression divergence between D. simulans and D. sechellia. (D) Changes in the expression of immunity genes encoding components of major humoral immune response signalling pathways in the fat body after immune challenge. Bars represent the change in gene expression 24 hours after injection with mineral oil containing wasp homogenate. Significantly differentially expressed genes within each species were indicated by * on top of columns. Divergent gene expression change responding to immune challenge between D. simulans and D. sechellia was indicated by * on top of bars between columns. Significance level was defined by false discovery rate (FDR): 0.05*, 0.01**, 0.001***.

Fig 6. Cis regulatory divergence in humoral immunity genes.

Transcripts were amplified by PCR from F1 hybrids between D. simulans and D. sechellia, and the PCR products Illumina sequenced. RNA was extracted from whole larvae that were uninfected (control) or immune challenged (wasp). Genomic DNA (gDNA) was included to check there were no biases towards one allele. (A) The proportion of reads from the two alleles. (B) The natural log of the odds (logit) that a transcript is from the D. simulans allele. Error bars are 95% confidence intervals on model coefficients. Fly larvae were immune-challenged by injecting them with oil droplets containing wasp homogenate. The horizontal dashed line (y = 0.5) signifies that there is no cis-regulatory divergence between the two species.

Fig 7. Relative expression between D. simulans and D. sechellia alleles and Trans regulatory divergence in humoral immunity genes.

Relative gene expressions were quantified by qPCR in parental and F1 hybrid larvae. The relative allelic expression in hybrids were estimated using proportions inferred from Illumina MiSeq sequencing of hybrid transcriptions for each gene, and the relative allelic expression between parental larvae were calibrated to half accordingly. (A) Relative allelic expressions between parental larvae and F1 hybrids under unchallenged (control), and infected (immune challenge imposed by injections of wasp homogenate) conditions. (B) Trans-regulatory divergence between D. simulans and D. sechellia in unchallenged and infected larvae, measured as the log₂ ratio of relative D. simulans over D. sechellia expression. Error bars are 95% confidence intervals on model coefficient.