Genomic insights into host and parasite interactions during intracellular infection by Toxoplasma gondii

Abstract

To gain insights into the molecular interactions of an intracellular pathogen and its host cell, we studied the gene expression and chromatin states of human fibroblasts infected with the Apicomplexan parasite Toxoplasma gondii. We show a striking activation of host cell genes that regulate a number of cellular processes, some of which are protective of the host cell, others likely to be advantageous to the pathogen. The simultaneous capture of host and parasite genomic information allowed us to gain insights into the regulation of the T. gondii genome. We show how chromatin accessibility and transcriptional profiling together permit novel annotation of the parasite's genome, including more accurate mapping of known genes and the identification of new genes and cis-regulatory elements. Motif analysis reveals not only the known T. gondii AP2 transcription factor-binding site but also a previously-undiscovered candidate TATA box-containing motif at one-quarter of promoters. By inferring the transcription factor and upstream cell signaling responses involved in the host cell, we can use genomic information to gain insights into T. gondii's perturbation of host cell physiology. Our resulting model builds on previously-described human host cell signalling responses to T. gondii infection, linked to induction of specific transcription factors, some of which appear to be solely protective of the host cell, others of which appear to be co-opted by the pathogen to enhance its own survival.

Fig 1. Project overview and host gene transcriptional response to T. gondii infection.

We show in (A) the overview of the experiments performed, depicting the in vitro infection of human fibroblasts with T. gondii, with transcriptional (RNA-seq) and chromatin accessibility (ATAC-seq) studies performed before and 24 hours after infection, and alignment of the reads obtained to both the human and T. gondii genomes. The RNA-seq results in (B) show more human genes upregulated (yellow) than downregulated (blue) with infection, with ADAMTS15 strongly upregulated. Using network analysis, we extracted genes interacting with ADAMTS15 (C) and showed that other metalloproteases were upregulated with infection (D). (E-G) show the major groups of gene ontologies in the remaining upregulated genes.

Fig 2. Inferred transcription factors mediating host cell chromatin accessibility changes.

In (A) we show the response to T. gondii infection in the human genome is marked by increased accessibility of chromatin at numerous loci. The results in (B) describe the motifs enriched in these loci that open their chromatin, corresponding to the binding sites of three known TFs. The results in (C) show the numbers of expressed genes containing one or more of the TF binding site motifs within 5 kb of their transcription start site. For those genes with only one motif, testing their ontological properties in (D) revealed the common property of induction of immune responses, but also other properties that are likely to be more favorable to the pathogen.

Fig 3. Distinct properties of T. gondii genes expressed during tachyzoite stage.

The T. gondii transcriptome in its tachyzoite stage while infecting human fibroblasts can be classified into three groups of genes based on their expression levels (A). In (B) we show the ontological properties of the genes in each of these three categories.

Fig 4. Chromatin accessibility profiles in the T. gondii genome indicate incomplete annotations of genes.

The chromatin accessibility pattern at all of the ATAC-seq peaks in the T. gondii genome at the tachyzoite stage of its life cycle is shown in (A). When the subset within ±5 kb of annotated T. gondii transcription start sites (TSS) is plotted with the associated genes represented to the right of the TSS (B, blue line), the patterns shows a skewing relative to the distributions of all peaks from panel (A) (gray distribution). Examples of genes are show to represent how the annotated TSS can be several kilobases (kb) from the upstream ATAC-seq peak and the start of the RNA-seq reads, with examples of 1.53 kb (C) and 4.03 kb (D) differences illustrated. In (E) we gain the additional insight that the two annotated genes are only associated with a single upstream ATAC-seq peak, indicating that there is only one transcript at this locus and the genes could be combined into a single annotation.

Fig 5. Chromatin accessibility and transcriptional profiling identifies new features in the T. gondii genome.

The gene expression and chromatin accessibility information combine to allow new T. gondii genomic annotations. In (A) we show the RNA expression at ATAC-seq peaks that are not within 5 kb of an annotated TSS. We sorted these peaks by clustering the characteristics of the RNA-seq data nearby, with the top clusters (1–3 especially) showing evidence for adjacent transcripts. Examples of the loci are shown on the right, including what appears to be a new gene (B), an antisense transcript from the 3’ end of an annotated gene (C), the use of an alternative intragenic promoter (D) and a locus of open chromatin with no nearby genes or RNA expression, potentially representing a distal cis-regulatory element.

Fig 6. Inference of transcription factor binding motifs in the T. gondii genome.

Analysis of the loci of open chromatin in the T. gondii genome reveals enrichment for several motifs (A). The purine (AG)-rich motif is the most abundantly represented at T. gondii promoters (B) followed by the known AP2 motif (GCATGCA), with the TATA-containing motif in 25.1% of promoters. Panel (C) shows the properties of the subset of genes with the AP2 motif at their promoters.

Fig 7. A model derived from genomic assay data for the host cellular response to T. gondii infection.

By inferring the TFs mediating the host cell response, we can further predict the cell signaling pathways induced by T. gondii infection in human fibroblasts. How the known GRA24 induction of p38 MAPK signaling influences the transcriptional response remains uncertain, and we include the possibility that T. gondii TFs may contribute to host cell transcriptional dysregulation.