Comparison of antiviral responses in two bat species reveals conserved and divergent innate immune pathways

Abstract

Bats host a range of disease-causing viruses without displaying clinical symptoms. The mechanisms behind this are a continuous source of interest. Here, we studied the antiviral response in the Egyptian fruit bat and Kuhl's pipistrelle, representing two subordinal clades. We profiled the antiviral response in fibroblasts using RNA sequencing and compared bat with primate and rodent responses. Both bats upregulate similar genes; however, a subset of these genes is transcriptionally divergent between them. These divergent genes also evolve rapidly in sequence, have specific promoter architectures, and are associated with programs underlying tolerance and resistance. Finally, we characterized antiviral genes that expanded in bats, with duplicates diverging in sequence and expression. Our study reveals a largely conserved antiviral program across bats and points to a set of genes that rapidly evolve through multiple mechanisms. These can contribute to bat adaptation to viral infection and provide directions to understanding the mechanisms behind it.

Keywords: Immunology; Phylogenetics; biological sciences; evolutionary biology; transcriptomics.

Graphical abstract

Figure 1

Characterization of dsRNA-stimulated genes in of cells from Rousettus and Pipistrellus (A) System overview: dermal fibroblasts derived from wing biopsies from two bat species were cultured and stimulated with dsRNA and control, followed by profiling of the response using bulk RNA-seq, gene quantification and differential expression analysis using edgeR. (B and C) Go term enrichment analysis of genes upregulated in response to dsRNA-stimulus in (B) Rousettus and (C) Pipistrellus. Selected non-redundant terms are shown in a decreasing order of significance (FDR-corrected p-values are shown). Dot colors and sizes denote p-values. Detailed analyses, including all enriched terms appear in Tables S4 and S5, respectively. (D) Heatmap of log₂(fold change) of dsRNA-stimulated Rousettus genes across different immune stimuli with different types of IFN and in response to infection with Sendai virus and Marburg virus.

Figure 2

Correlation of response across species (A) A schematic phylogenetic tree of the six studied species: human, macaque, mouse, rat, and the two bat species. The two bat species, belonging to Yinpterochiroptera and Yangochiroptera, are estimated to have a last common ancestor (LCA) ∼60–70 million years ago (MYA), whereas primates and rodents split ∼90MYA. (B–E) Spearman’s rank correlation of gene fold change in response to dsRNA, in 2,865 one-to-one orthologous genes between (B) Rousettus and Pipistrellus, (C) Human and macaque (estimated LCA – 28MYA⁹⁶), (D) Mouse and rat (estimated LCA – 13MYA⁹⁶), and (E) Human and mouse. Genes included are dsRNA-regulated (FDR-corrected<0.01) in at least one of the six species appearing in figures B–E (2,865 genes).

Figure 3

Regulatory and evolutionary characteristics of conserved and divergent dsRNA-stimulated genes between Rousettus and Pipistrellus (A and B) Coding sequence divergence versus transcriptional divergence: Genes were partitioned into three groups based on divergence in response to dsRNA stimulation between the two bat species (termed high, medium, and low transcriptional divergence), and coding sequence divergence was compared between these three groups: (A) Distributions of P-values of test for positive selection across bat orthologs (using the dN/dS values from a previous analysis⁷⁹), and (B) sequence similarity (precent of identity) between the two bat species, are shown for the three gene groups (1,899 and 2,728 genes with available data in A and B, respectively). High- and low-divergence groups are compared using a Mann-Whitney test. (C) Divergence in response to dsRNA between the bat species in genes with and without CpG Island (CGI) in their promoters, and with and without TATA-box in their promoters (896 CGI and 1,969 CGI-less genes; 107 TATA and 2,758 TATA-less genes). Groups are compared using a Mann-Whitney test. (D–E) Using single-cell RNA-seq, the distribution of cell-to-cell variability in gene expression, as measured using the distance to median approach (DM), is shown for each of the three groups mentioned in (A), for (D) dsRNA-stimulated Rousettus cells and (E) stimulated human cells (1,980 and 2,578 genes with computed DM values, respectively). High- and low-divergence groups are compared using a Mann-Whitney test. See Figure S3 for analysis of unstimulated Rousettus and human cells. (F) The distribution of measures for tolerance and resistance are shown for genes that are highly upregulated in response to dsRNA in Rousettus versus Pipistrellus, or the opposite (1,928 and 1,949 genes, respectively). The distributions of Rousettus-specific and Pipistrellus-specific genes are compared using a Mann-Whitney test. (∗ = p < 0.05, ∗∗∗ = p < 0.001).

Figure 4

Duplication of TNFRSF14 in Rousettus and duplicate divergence in sequence and expression (A) Reconstructed tree of TNFRSF14 of human, mouse, and five Rousettus duplicates. (B) Left: TNFRSF14 gene expression across tissues of human, mouse, and five Rousettus duplicates; Middle: log₂(Fold Change) of TNFRSF14 of human, mouse, and five Rousettus duplicates in response to dsRNA or IFNB, as measured in cells from human,mouse and Rousettus (significance of upregulation is shown within the square). Right: log₂(Fold Change) of five Rousettus duplicates in response to infection by Sendai virus or Marburg virus. (C) Left: Structure of human TNFRSF14 with the herpes glycoprotein D, in red and blue respectively (PDB: 1JMA). Human TNFRSF14 is colored by relative evolutionary rate of residues (as measure across all orthologs and paralogs appearing in A). Right: The distributions of evolutionary rates of residues found at the surface of TNFRSF14 (surface) compared with evolutionary rates of residues interacting with the herpes glycoprotein D (interface). Comparison is made using a Mann-Whitney test (∗ = p < 0.05).