Toxoplasma and Eimeria co-opt the host cFos expression for intracellular development in mammalian cells.

Successful asexual reproduction of intracellular pathogens depends on their potential to exploit host resources and subvert antimicrobial defense. In this work, we deployed two prevalent apicomplexan parasites of mammalian cells, namely Toxoplasma gondii and Eimeria falciformis, to identify potential host determinants of infection. Expression analyses of the young adult mouse colonic (YAMC) epithelial cells upon infection by either parasite showed regulation of several distinct transcripts, indicating that these two pathogens program their intracellular niches in a tailored manner. Conversely, parasitized mouse embryonic fibroblasts (MEFs) displayed a divergent transcriptome compared to corresponding YAMC epithelial cells, suggesting that individual host cells mount a fairly discrete response when encountering a particular pathogen. Among several host transcripts similarly altered by T. gondii and E. falciformis, we identified cFos, a master transcription factor, that was consistently induced throughout the infection. Indeed, asexual growth of both parasites was strongly impaired in MEF host cells lacking cFos expression. Last but not the least, our differential transcriptomics of the infected MEFs (parental and cFos-/- mutant) and YAMC epithelial cells disclosed a cFos-centered network, underlying signal cascades, as well as a repertoire of nucleotides- and ion-binding proteins, which presumably act in consort to acclimatize the mammalian cell and thereby facilitate the parasite development.

Introduction

The protozoan phylum Apicomplexa consists of over 6000 known parasite species, many of which have significant medical, veterinary and ecological relevance because they infect humans, livestock as well as the wildlife [1]. Amongst all, Toxoplasma gondii has become a popular model due to relative ease of its propagation, genome engineering and phenotyping [2]. There is only a single species in the genus Toxoplasma. Its sexual reproduction is confined to a feline family member (primary or definitive host), whereas asexual growth can occur in most homeothermic organisms (secondary or intermediate host). The asexual stages of T. gondii can also transmit between the intermediate hosts and thus bypass the need of a definitive host unlike most other apicomplexans, which has contributed to its widespread success. Besides, the parasite can reproduce in numerous types of nucleated host cells of a given organism. Consequently, there has been substantial interest in decoding how T. gondii reprograms and survives in diverse intracellular environments. Some of these studies have deployed transcriptomic profiling of the parasitized mammalian cells to identify and validate the host determinants of parasite development [3], [4], [5].

In notable contrast to Toxoplasma, another apicomplexan genus Eimeria comprises >1800 extant species, which have primarily evolved to reproduce in the gastrointestinal epithelial cells of distinct host organisms [6]. The lifecycle of Eimeria species is completed in a single host and the inter-host transmission requires gyrating asexual and sexual reproduction. A high natural diversity of Eimeria species in conjunction with its monoxenic lifestyle and fecal-oral transmission make it a prevalent pathogen of livestock and wildlife. One particular species, E. falciformis, infecting rodents to accomplish its entire lifecycle, is an emerging model to study the pathogen-host interactions, sexual development, immune response and microbiome-parasite relationship in conventional mouse models [7], [8], [9], [10] and in the natural environment [11]. Our own previous research on E. falciformis has reported its genome [12] and infection in different rodent species [13], [14], investigated the parasite’s membrane biogenesis [15] and host immune response [16], as well as studied parasite-induced host manipulation by ex vivo expression analyses of the mouse caecum tissue epithelial cells [7], [8]. The latter work identified a retinue of IFNγ-regulated pathways including tryptophan catabolism, chemokine signaling and cell-intrinsic immunity, some of which play opposing roles during in vivo parasite infection.

Toxoplasma and Eimeria, both being tissue-dwelling parasites, belong to the subclass coccidia. Given their differences and similarities, we surmise that T. gondii and E. falciformis together could enable a more inclusive understanding of the coccidian biology. In this regard, the intestinal epithelial cells – the primary site of infection by both pathogens – impart an outstanding prospect to compare their host reprogramming; however, a parallel (head-to-head) examination in a common host cell type remain elusive. Herein, we utilized the young adult mouse colonic (YAMC) epithelial cells to evaluate how these two phylogenetically-related parasites alter the respective host niche. Among other findings, we discovered the cellular FBJ osteosarcoma oncogene (cFos) – a vital component of the AP-1 transcription factor that governs differentiation, proliferation, apoptosis and immune response in mammalian cells [17], [18] – is persistently upregulated upon infection by both parasites. Our subsequent phenotyping and comparative transcriptomics using the cFos-knockout and parental mouse embryonic fibroblasts (MEFs) unveiled a pro-parasite role of cFos alongside its infection-associated network and underlying signaling events.

Results

Eimeria and Toxoplasma infection of YAMC epithelial cells modulate a gamut of host genes

To investigate the gene expression in parasitized YAMC cells, we infected them with tachyzoites of T. gondii or with sporozoites of E. falciformis, both of which invaded with similar efficiency (40%), as deduced by immunostaining 4 h post-infection (Fig. 1A). Unlike Toxoplasma tachyzoites, which divide to form the identical progeny [19], Eimeria sporozoites can only develop into trophozoite and schizont stages, and culture is usually aborted 24 h post-infection. We chose 4 h and 16 h for our transcriptomic analysis to discern the infection-linked rewiring of gene expression in host cells. As illustrated in Fig. 1B, RNA was isolated from infected and uninfected YAMC cells, and subsequently the mRNA samples were labeled with Cy3 and Cy5 dyes for hybridization to the whole-genome (4x44K) mouse microarray chips. Differences in gene expression of the parasitized cells with a cut-off of 1.5-fold and an error-weighted p-value ≤ 0.05 in comparison to uninfected samples were considered significant for further analysis.

Fig. 1

The young adult mouse colonic epithelial cells show a significant transcriptional modulation upon infection by coccidian parasites. (A) Invasion efficiency of Eimeria sporozoites and Toxoplasma tachyzoites in YAMC cells. Intracellular parasites (4 h infection) were quantified after staining with anti-E. tenella serum (E. falciformis), or using anti-TgSag1 and anti-TgGap45 antibodies (T. gondii). (B) Schematics of the mouse microarray analysis using YAMC host cells parasitized with either E. falciformis sporozoites or T. gondii tachyzoites. The RNA samples collected after 4 h and 16 h of infection were labelled with Cy3 or Cy5 fluorescent dye, and then hybridized to the whole-genome microarrays. (C) Scatter plot of the fold-changes between uninfected and infected YAMC cells after 4 h or 16 h infection either with T. gondii or with E. falciformis. Differentially-regulated probes (fold change ≥ +1.5, p ≤ 0.05) are color-coded, whereas others appear in black. Probes altered only by T. gondii or E. falciformis are colored blue and red respectively, while those regulated by both are depicted in violet (same trend) or green (opposite trend). (D) Venn diagram of genes regulated upon infection by T. gondii or E. falciformis. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

A majority of genes did not change in expression irrespective of the parasite infection or time point (black dots, Fig. 1C); although, distinct modulation of YAMC cells by T. gondii and E. falciformis was quite obvious (colored dots, Fig. 1C). We observed 3853 genes regulated in Eimeria-infected cells, of which 2002 transcripts were altered after 4 h, 1851 at 16 h, and 155 at both time periods (Fig. 1D). Equally, T. gondii infection modulated 3464 transcripts; most of them (2920) however were affected 16 h post-infection, a much smaller set of 544 genes responded within 4 h, and 226 transcripts were perturbed at both time points. These results indicated that the host response during Eimeria infection remained rather stable, while a striking difference was recorded between 4 h and 16 h infection by T. gondii. Surprisingly, only 10 exclusive transcripts (5 induced, 5 repressed) were regulated by the two parasites 4 h post-infection, whereas 182 unique genes were commonly upregulated, and 158 were downregulated after 16 h incubation. A major fraction of the host transcriptome was affected by one or the other pathogen, signifying a markedly divergent modulation of YAMC epithelial cells in a parasite-specific manner.

Several distinct and shared host-signaling cascades are regulated by coccidian parasites

Our downstream analysis using KEGG pathway database revealed a significant enrichment of about 90 pathways among all transcripts differentially expressed upon E. falciformis and T. gondii infection. Nearly one third of these pathways are related to signaling events and a quarter each to oncogenesis and infection processes (Table S1). PI3K-Akt and RIG-I-like receptor pathways were most enriched during T. gondii infection, whilst cGMP-PKG, Rap1, oxytocin, NOD-like receptor and NF-κB signaling were modulated mainly by E. falciformis (Appendix A, Appendix A). Illustration of two such cascades (PI3K-Akt and cGMP-PKG) showed transcript regulation of several actuators and mediators after 4 h and/or 16 h infection. Interestingly, growth factors (PDGF family) and extracellular matrix collagen (Col15a1) that are known to activate PI3K-Akt signaling as well as various mediators (PI3Ks, Pten, Nur77) were induced by both parasites (Fig S1B). Most effectors involved in protein synthesis and/or cell cycle (4EBPs, eIF4B, p27, cyclin, p27Kip1) were however affected only by Toxoplasma. In contrast, Eimeria exerted a nearly exclusive induction of PKG and other players of cGMP signaling (RhoA, MLC, CaM) (Fig S1C).

Of multiple commonly-affected host pathways (Fig. 2A), we illustrated cAMP signaling as it converges with other differentially-altered pathways, such as PI3K-Akt, Ca2+, cGMP-PKG, insulin and cFos (see below). Notably, T. gondii and E. falciformis appear to impact different mediators of cAMP signaling (Fig. 2B). Soon after Eimeria infection (4h), the adenylate cyclase (AC) transcript was induced, whereas cAMP-specific phosphodiesterase PDE4a was repressed, which suggested the actuation of cAMP signaling. We did not witness this phenomenon upon T. gondii infection, albeit genes located upstream of AC (Ptger3, GHRL) and downstream of PKA (Creb5, Creb3l3, Gli3, Crebbp) were significantly altered. Several ion transporters (Ca2+, Na+, K+, Cl-) and calcium-responsive calmodulin (CaM) were affected by both coccidians, implying a perturbed ion homeostasis and calcium signaling in infected host cells. One of the most distinguished transcripts among all was cFos – a master transcription factor known to be induced by cAMP signaling [20] – that was upregulated by the two parasites at 4 h as well as 16 h post-infection. Given the multifaceted roles of cFos in mammalian cells, our latter work investigated its physiological importance and interaction network during parasitic infection.

Fig. 2

cAMP signaling is one of several differentially-modulated signaling cascades in coccidian-infected YAMC cells. (A) Signaling pathways regulated upon infection by T. gondii or E. falciformis. Significantly-altered genes were analyzed for KEGG pathway enrichment and selected transduction cascades shared by both parasites were plotted. For an inclusive list of enriched pathways, refer to Table S1. The pathways were enriched using DAVID v6.7 (EASE score ≤ 0.1). (B) Scheme depicting transcriptional reprogramming of cAMP signaling in parasitized host cells. Genes shown with colored symbols are regulated by either or both pathogens. Heatmap blocks next to individual genes illustrate the fold-change according to the actual up- or down-regulation relative to uninfected controls.

cFos is one of the few host transcripts mutually regulated by coccidian parasites

Our further work focused on those transcripts that were inversely or similarly changed in YAMC cells by Toxoplasma and Eimeria (Fig. 3). Of >3000 altered genes, 74 were perturbed in an opposing manner by specified pathogens (Fig. 3A-B), representing the divergent portion of host response. In early cultures (4 h infection), E. falciformis displayed a mostly repressive effect (10 repressed and 4 induced), whereas T. gondii exerted an inductive effect (10 induced and 4 repressed). Surprisingly, this trend was reversed at 16 h, as 52 transcripts were upregulated and 8 genes were downregulated by Eimeria, and the converse was true for Toxoplasma. Quite notably, many interferon-associated genes were repressed in YAMC cells inhabited by T. gondii, but induced upon E. falciformis infection (Fig. 3B). Equally, many other genes involved in epithelial cell growth and inflammatory response (e.g., Muc13, Prkce, Trp53inp1, Pxt1, Dusp1, Dusp10) were downregulated by Toxoplasma, although induced during Eimeria infection.

Fig. 3

Only a limited set of host transcripts are conjointly regulated by Toxoplasma and Eimeria. (A-B) Genes altered in an opposite manner (inversely-regulated) upon parasitism of YAMC epithelial cells with T. gondii or E. falciformis (4 h, 16 h infection). Asterisks indicate those genes that can be modulated by type-I or type-II interferon (Orange, type-I; green, type-II; violet, type-I and type-II) (C) Genes similarly affected by both parasites. Heatmaps in panel A-C show fold-induction or repression according to the color scale. (D) Quantitative PCR expression profile of the Fos/Jun family members in comparison to the microarray dataset. The qPCR results were obtained by the method, and threshold modulation of transcripts (≥1.5 or ≤ 1.5) is marked by dotted lines. (E) Immunostaining of cFos upon infection of YAMC epithelial cells by T. gondii tachyzoites. Host cells (uninfected or 16 h post-infection) were stained with mouse anti-cFos monoclonal (green) and rabbit anti-TgGap45 (red) antibodies. DAPI was used to visualize the parasite and host nuclei. Images shown herein represent one of the three independent assays. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

We identified 13 transcripts that were similarly regulated during both infections (Fig. 3C). Among these, Rassf3, a member of the RASSF family tumor suppressors that is known to induce apoptosis [21], was repressed, indicating inhibition of apoptosis in parasitized cells. Besides, repression of Adrb2, which controls inflammation by rapid induction of IL10 [22], may shift the immune equilibrium in favor of the host cells. Likewise, downregulation of Ikzf2 – a chromatin remodeler maintaining self-renewal in leukemic stem cells [23] – may promote the cellular differentiation upon infection. On the other hand, transcripts of Saa4, Slc8a3, Me3 and cFos were steadily upregulated. Serum amyloid A (SAA) is a family of microbe-inducible retinol-binding proteins expressed in the intestinal epithelium [24]. Retinol is pivotal to develop the innate and adaptive immunity; hence the induction of Saa4 is likely a defensive response of the YAMC epithelial cells. Upregulation of solute carrier 8a3 (Slc8a3) and malic enzyme 3 (Me3) by contrast may foster the parasite growth by adjusting ionic and pyruvate homeostasis, respectively.

A persistent induction of cFos echoed with potential activation of cAMP signaling (Fig. 2), implying a role of this key transcription factor during coccidian infection. We therefore tested the expression of cFos family genes (cFos, FosB, Fra-1, Fra-2, cJun, JunB, JunD) by performing a quantitative PCR analysis of T. gondii-infected YAMC cells (Fig. 3D). Overall, the qPCR data correlated well with our microarray results. We observed that both cFos and JunB were significantly induced, while cJun and JunD were modestly suppressed upon infection. Expression of Fra-1, Fra-2 and FosB was either unaffected or did not match between the qPCR and microarray datasets. Our extended work attempted to validate the induced expression of cFos by immunoblot analysis, which did not however yield a reproducible detection of the protein. Nonetheless, immunofluorescent staining of tachyzoite-infected YAMC epithelial cells (Fig. 3E) indicated apparent upregulation of cFos when compared to the uninfected sample, and prompted us to examine its physiological relevance as described below.

cFos is required for optimal in vitro growth of Eimeria and Toxoplasma

We next studied the importance of cFos for the parasite development in murine embryonic fibroblasts lacking its expression (cFos-/- MEFs). As expected, infection of the parental (cFos+/+) MEF cells with Eimeria sporozoites resulted in the formation of schizont and trophozoite stages (Fig. 4A). In contrast, the development of these two parasite stages in cFos-knockout cells was strikingly reduced to 25%. A lack of cFos expression did not impact the sporozoite invasion into mutant MEFs, which was akin to the parental host cells (20% infection). A similar phenotype was observed when we examined the growth fitness of T. gondii tachyzoites in plaque assays using cFos+/+ and cFos-/- MEFs (Fig. 4B). The wild-type cells supported the recurring lytic cycles of tachyzoites and, as anticipated, eventually led to the formation of plaques by disruption of confluent monolayers. In contrast, we scored a prominent reduction by > 75% in the number and size of plaques formed in the cFos-/- host cells.

Fig. 4

Lack of cFos compromises the asexual growth of coccidian parasites in mouse fibroblasts. (A) Intracellular development of E. falciformis in MEF cells with or without cFos expression. The wild-type (cFos+/+) or mutant (cFos-/-) cells were infected with freshly isolated sporozoites of E. falciformis, and the developmental stages were quantified by immunostaining with cross-reactive anti-E. tenella serum. (B) Plaque assays representing the overall growth fitness of T. gondii tachyzoites in cFos+/+ and cFos-/- cells. The white spots are formed by sequential lytic cycles of tachyzoites in a confluent monolayer of MEF cells (stained with crystal violet). (C) Intracellular replication of tachyzoites in MEF cells, as deduced by staining with anti-TgGap45 antibody. The mean fraction of the parasitophorous vacuoles enclosing variable number of tachyzoites are depicted in a bar graph (n = 3 assays, >400 vacuoles). (D) Plaques formed by the PruΔku80 (parental) and PruΔku80Δgra24 strains of T. gondii in cFos+/+ and cFos-/- host cells. Left panel displays crystal violet-stained plaques in MEF monolayers, whereas the confetti plot (right panel) reveals the area of at least 200 plaques in arbitrary units (a.u.), as quantified by ImageJ program (n = 3 experiments; mean ± S.E.M.; *, p ≤ 0.05; **, p ≤ 0.01; ***, p ≤ 0.001). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

A defect in plaque formation may be caused by impaired host-cell invasion, intracellular proliferation and/or egress of tachyzoites. Hence, we preformed additional phenotyping assays to determine the effect of cFos expression on individual steps of the lytic cycle. Both cell lines were parasitized equally (50%); however, the parasite replication was markedly reduced, as judged by counting tachyzoites within parasitophorous vacuoles during the course of infection (Fig. 4C). The fraction of large vacuoles (8–16 parasites/vacuole) was higher in the parental cells compared to cFos-knockout cells. Conversely, the latter host cells harbored a higher portion of smaller vacuoles (2–4 parasites/vacuole). The phenotype was evident even after 12 h infection, signifying that reproduction of tachyzoites in cFos-/- MEFs was arrested at an early stage. All above results taken together demonstrate a need of host cFos for the intracellular development of Toxoplasma tachyzoites and Eimeria sporozoites in mouse fibroblasts.

It has been previously reported that cFos can also be induced via an effector protein Gra24, secreted by a cyst-forming (type II) strain of T. gondii [25]. We therefore examined the growth fitness of parasites lacking the Gra24 expression (Fig. 4D). As expected, the parental strain (PruΔku80) displayed a significantly impaired development in the cFos-/- host cells. A severe growth defect was also quite evident in the Δgra24 mutant, suggesting that Gra24 does not underlie the specified phenotype in the cFos-knockout cells. Interestingly, we observed that when compared to the Δgra24 mutant, the parental strain reproduced slightly better in the cFos-/- but not in the cFos+/+ MEFs, which reflects a plausible role of Gra24 in promoting the parasite development in the absence of cFos expression.

Transcriptomics of parasitized cFos+/+ and cFos-/- cells reveals a perturbation of cFos network

To decrypt how cFos can foster the coccidian development, we performed gene expression analysis of the wild-type (cFos+/+) and cFos-knockout (cFos-/-) MEFs infected with Eimeria or Toxoplasma, as shown for YAMC epithelial cells (Fig. 1). Infection-induced changes correlated between the parental and mutant MEFs for a large majority of genes notwithstanding the pathogen or infection period (Fig. 5A, see Supplement Text). There were nonetheless clear differences between the two host-cell types when infected by a given parasite. A small number (16 genes during E. falciformis and 20 during T. gondii infection) were inversely correlated, suggesting a divergent response of cFos+/+ and cFos-/- cells (green dots, Fig. 5A). In both infections, we witnessed numerous transcripts that were uniquely induced or repressed either in the parental or cFos-knockout cells (Fig. 5B). In case of Eimeria, the number of altered genes was similar between the two cell types and time periods. Toxoplasma-infected cFos-/- cells by contrast displayed 2–4-fold fewer differentially-expressed genes in early and late cultures when compared to the cFos+/+ host cells. The Venn diagram also showed a considerable number of commonly-regulated genes in the parental and mutant fibroblasts infected by individual parasites at a specified time point. These findings were further endorsed by Pearson correlation matrix (Fig S2A, refer to Supplement Text for additional details).

Fig. 5

Transcriptomic analysis of the wild-type and cFos-knockout mouse embryonic fibroblasts infected with T. gondii or E. falciformis. (A) Scatter plot illustrating the gene expression (fold-change) of cFos+/+ and cFos-/- cells upon infection, as indicated. Differentially-regulated probes (≥or ≤ 1.5, p ≤ 0.05) are color-coded, whereas others appear in black. Transcripts altered only by T. gondii or by E. falciformis are presented in blue and red, respectively, while those regulated by both parasites are depicted in violet (same trend) or green (opposite trend). To see the complete correlation matrix (heatmap) of parasitized MEF and YAMC cells, refer to Fig. S2A. (B) Venn diagram of genes regulated upon infection by T. gondii or E. falciformis. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

Of various plausible means to analyze our results, we first selected transcripts that were shared by YAMC and parental MEF cells upon infection with each parasite (Fig. 6A). We then filtered the chosen transcripts on the condition that they failed to be significantly modulated in cFos-knockout cells. Our approach eventually enriched only those genes that are likely to be cFos-dependent (cFos-related) and associated with infection. Our study yielded 384 and 187 genes differentially regulated by T. gondii and E. falciformis, respectively, in YAMC and cFos+/+ but not in cFos-/- host cells. The GO-term analysis of these transcripts identified a repertoire of nucleotide-binding and ion-binding factors (Fig. 6A, Table S2). Intriguingly, while the former category dominated during T. gondii infection, the latter was more prominent in Eimeria-infected cells. A total of 18 genes were altered by both parasites in YAMC and wild-type MEF cells, which indicated a substantially minimal convergence between them.

Fig. 6

Infection-relevant cFos-network as deduced by comparative transcriptomics of YAMC and MEF cells parasitized with T. gondii or E. falciformis. (A) Genes regulated in infected YAMC epithelial cells and in the parental MEFs but not in cFos-/- cells. The gene-ontology (GO) clusters most enriched upon T. gondii and E. falciformis infection are labeled. For other significant categories, refer to Table S2. The GO terms were enriched using DAVID v6.7 (threshold of EASE score ≤ 0.1). (B) Quantitative PCR illustrating the expression profile of 12 genes randomly selected from panel A. The bar graph shows a comparison of fold-change results, as calculated by the qPCR method and microarray analysis. Values ≥ or ≤ 1.5 were defined as ‘significant’ (dotted lines). (C) cFos-centered network predicted by STRING analysis of genes identified in panel A. The image shows primary genes (directly linked to cFos), and secondary genes (connected to cFos-networked primary genes). The breadth/thickness of connecting lines indicates the level of confidence. The primary genes were networked with STRING v11.0 at medium confidence (score ≥ 0.4), whereas the secondary genes were defined at high confidence (score ≥ 0.7).

We next validated the expression profile of 12 randomly-selected genes (Adar, Ccl4, Cldn1, Hes1, Peg10, Plxna4, Saa4, Sp100, Psmb8, Vcam1, Strbp, Zfp60) that were altered in both parental lines (YAMC and cFos+/+ host cells), but unchanged in cFos-/- MEFs upon infection by tachyzoites of T. gondii (16 h) (Fig. 6B). As seen in microarray analysis, a majority of chosen transcripts (11 out of 12, except Ccl4) exhibited no significant modulation in cFos-/- MEFs. Besides, 7 genes (Adar, Ccl4, Cldn1, Hes1, Peg10, Plxna4, Saa4) displayed the same trend of regulation in both parental lines, as recorded by microarrays. Only 1 transcript (Vcam1) was differently modulated in the two wild-type host cells when compared to corresponding microarray data, and 4 genes (Sp100, Psmb8, Strbp, Zfp60) were inversely correlated either in YAMC epithelial or cFos+/+ cells. In brief, the qPCR results added further confidence to our designated list of infection-relevant cFos-related genes.

cFos network associated with coccidian infection

In continuation of the above work, we constructed a cFos-centered protein–protein interaction network by STRING software (Fig. 6C). Indeed, 16 genes were directly networked with cFos and an additional panel of 39 genes were indirectly associated with cFos through its primary interaction network, which retrospectively endorsed our intersection analysis. Only 2 genes, Serpind1 and Ccl4, in the network were influenced by both parasites, whereas 16 others were either affected by Eimeria or Toxoplasma. The KEGG pathway classification of all genes appearing in the cFos-network enriched Ca2+, MAPK and pattern recognition receptor signaling (Fig. 7). Not least, a direct comparison of KEGG pathways enriched among all infected host-cell types used herein disclosed an exclusive enrichment of insulin signaling in cFos-/- cells (Fig S3, Table S3). In conclusion, our comparative transcriptomics yielded an infection-relevant network of cFos comprising several deferentially-regulated putative determinants of coccidian development.

Fig. 7

A cFos-centered network of host proteins linked to intracellular parasitism by coccidians. The model was constructed based on the comparative gene expression analyses of YAMC and MEF (cFos+/+ and cFos-/-) cells infected by T. gondii or E. falciformis. The cFos-linked pathways influenced by both pathogens in the wild-type host cells include calcium, MAPK and pattern recognition receptor (Toll-/RIG-I like) signaling cascades. Several other genes encoding for nucleotide-binding and ion-binding proteins were affected by individual parasites. Note that according to our model, insulin signaling is enriched only in MEF host cells lacking cFos expression (Fig S3, Table S3), likely due to feedback regulation. Insulin, PI3K-Akt and cAMP signaling – differentially regulated upon coccidian infection – probably operate upstream of cFos. Transcriptomic reprogramming of the host cell may, at least in part, be driven by secretory effector molecules of the two parasites (question-marked).

Discussion

Our transcriptomic analysis of the young adult mouse colonic epithelial cells infected with Eimeria falciformis or Toxoplasma gondii identified a large retinue of pathways, some of which were modulated by both, whereas others were regulated primarily by each pathogen. Infection-mediated reprogramming of YAMC epithelial cells signifies either host-cell defense and/or apparent subversion by the two parasites. A major group of differentially-expressed genes may, at least in part, be influenced by parasite-secreted proteins, as elaborated below. We found cFos as one of the 13 genes that were upregulated throughout the course of coccidian infection. Importantly, a lack of cFos expression retarded the growth of both parasites in murine embryonic fibroblast cells, advocating a pro-coccidian role of this transcription factor. In additional work using parasitized MEF and YAMC host cells, we discovered a cortege of cFos-networked factors related to infection irrespective of the host-cell type. Strikingly, genes differentially regulated in infected cells expressing cFos networked together and even cFos-related host signaling pathways were often shared, all of which plausibly act in consort to benefit T. gondii and E. falciformis (Fig. 7).

cFos belongs to an immediate early gene family of transcription factors, which is barely expressed under normal conditions but rapidly induced by a repertoire of internal and external stimuli [17], such as by growth factors [26], [27], cytokines [28], calcium [29], cAMP [30], oxidants [31], antioxidants [32], LPS [33], IFN-γ [34], as well as by physical [35] and mechanical [36] stress. Once induced, cFos and its family members (FosB, Fra-1, Fra-2) heterodimerize with members of the Jun family (cJun, JunB, JunD) to exert their function as transcription factors [37]. Heterodimerization results in the formation of AP-1 complex, which binds in the regulatory regions of many genes [38] and thereby controls differentiation, proliferation, apoptosis and immune response in a wide range of cell types [17], [18]. The AP-1 complex has also been reported to be modulated by other pathogens e.g., Leishmania [39], [40], Salmonella typhimurium [41], Bacillus anthracis [42], hepatitis C virus [43] and polyomavirus [44]. Its activity correlates with oncogenic transformation and propagation of viruses [43], [45], [46], [47]. On the other hand, AP-1 is known to facilitate the protective immunity against Salmonella [41] and Leishmania [39], [40], and a lack of cFos promotes the growth of S. typhimurium [41]. Our findings on T. gondii and E. falciformis reverberate with the viral but contrast the bacterial pathogens, and thus epitomize opposing roles of cFos during infection by different pathogenic organisms.

We identified several cFos-networked factors involved in calcium, MAPK and RLR/TLR signaling, some of which are potentially subverted by coccidians. In accord, multiple transcripts of nucleotide and ion-binding proteins were modulated in infected YAMC and parental MEF but not in the cFos-/- host cells, suggesting a substantial transcriptional reprogramming associated with cFos expression (Fig. 7). Apicomplexan parasites are known to secrete an entourage of effector proteins from highly specialized organelles namely rhoptries, micronemes and dense granules [48]. Proteins discharged by the latter organelle usually enable adaptive modification of parasite-specific niches [49], [50]. Tachyzoites of T. gondii, for instance, excrete many dense granule proteins (Gra15/16/24/44, IST etc.) that are known to modulate IFN, JAK-STAT, p38α MAPK, NF-κB and p53 signaling in host cells [25], [51], [52], [53], [54], [55], [56]. In context of this work, TgGra24 is reported to trigger prolonged autophosphorylation and nuclear translocation of the p38α MAP kinase [25], [51], which correlates with the induction of immediate early genes. Similarly, TgGra16 and TgGra44 have been shown to induce cMyc in infected cells [52], [57], and TgIST counteracts the IFNβ and IFNγ-mediated defenses by binding to STAT1 and STAT1/STAT2 heterodimers [54]. Based on these studies, we surmise a role of dense granule proteins in the observed transcriptomic rewiring of YAMC and MEF cells. The genome of E. falciformis also encodes a broad range of predicted secretory factors [12]; however, the counterparts of Toxoplasma effector proteins are yet to be characterized.

Toxoplasma and Eimeria infection induced primarily discrete transcriptional response in a given host cell despite their close phylogenetic relationship. Similarly, different host cells of a selected organism displayed broadly distinct expression profile upon infection by individual parasites. Our comparative analysis disclosed regulation of multiple pathways in parasite- and/or host-specific manner (Supplement Text), some of which may underlie in vitro development of E. falciformis. Going further, a comparison of YAMC (this work) and caecum epithelial cells infected by Eimeria (in vivo) [8] might reveal additional determinants of parasite development. For example, tryptophan catabolism was one of the most outstanding (IFNγ-dependent) pathways by ex vivo transcriptomics, which we found to have a pro-parasite role in vivo. Even though numerous IFNγ-linked genes were regulated during in vitro infection of YAMC and MEF cells, tryptophan catabolism was not enriched in any of our datasets (Tables S4–S6). Similar observations were made with many IFNγ-regulated immunity-related GTPases and guanylate-binding proteins. In vitro optimization guided by comparative expression analyses therefore holds promise to develop a sustained culture of E. falciformis.

In conclusion, we demonstrate a physiological requirement of the mammalian cFos for intracellular development of T. gondii and E. falciformis. Additionally, we revealed an infection-affiliated network of cFos and signal cascades that are likely co-opted by coccidian parasites. A string of credible host determinants identified herein shall enable comprehensive dissection of pathogen-host interactions.

Materials and methods

Biochemical resources and bioethics statement

Oocysts of Eimeria falciformis were procured from Bayer (Germany), and tachyzoites of Toxoplasma gondii (type I RH strain) were provided by Carsten Lüder (Georg-August University, Germany). The cyst-forming type II strains of T. gondii (PruΔku80, PruΔku80Δgra24) were acquired from Mohamed Ali-Hakimi (University of Grenoble, France). The young adult mouse colonic epithelial cells were obtained from Robert Whitehead (Vanderbilt University, USA) [58]. The murine embryonic fibroblasts (parental, cFos+/+ 1–98 M; mutant, cFos-/- 7–98 M) were offered by Marcus Christmann (University of Mainz, Germany) [59]. The primary antibodies recognizing TgSag1 (mouse) and TgGap45 (rabbit), and anti-Eimeria tenella serum (rabbit) were donated by Jean-François Dubremetz (University of Montpellier, France), Dominique Soldati-Favre (University of Geneva, Switzerland) and Fiona Tomley (Royal Veterinary College, London, UK), respectively. The primary antibody against the cFos protein (sc-271243) was purchased from Santa Cruz Biotechnology. NMRI mice were acquired from Charles River Laboratories (Sulzfeld, Germany). All animal experiments were executed following the guidelines of Landesamt für Gesundheit und Soziales (LaGeSo), Berlin.

Propagation of E. falciformis and sporozoite purification

The natural life cycle of E. falciformis was maintained by continuous passage in NMRI mice. In brief, oocysts were purified from animal feces by NaOCl floatation method [60], counted using a McMaster chamber, and stored in potassium dichromate at 4 °C for up to a maximum of 3 month. To isolate the free sporozoites, purified oocysts were digested with 0.4% pepsin (pH 3, 37 °C, 1 h), washed with PBS (1800g, 10 min), mixed with glass beads (0.5 mm, 1:1 ratio) and then vortexed briefly to release the sporocysts. Sporozoites were excysted by incubating the sporocyst preparation with 0.25% trypsin and 0.04% sodium tauroglycocholate (MP Biomedicals) in Dulbecco’s modified Eagle’s medium (DMEM) containing 2 mM glutamine, 100 µg/mL streptomycin and 100 U/mL penicillin (37 °C, 2 h), and then purified using DE52 anion-exchange chromatography, as described elsewhere [61].

The parasite and host cell culture

Tachyzoites of T. gondii were maintained in confluent monolayers of human foreskin fibroblast (HFF) cells, as reported previously [62]. HFFs were cultured in DMEM with 4.5 g/L glucose, 10% FCS, 2 mM glutamine, 1x MEM non-essential amino acids, 100 µg/mL streptomycin and 100 U/mL penicillin in a humidified incubator (37 °C, 5% CO2). The YAMC epithelial cells were cultured in RPMI medium containing glucose (4.5 g/L), FCS (5%), insulin (1 µg/mL), α-thioglycerol (10 µM), hydrocortisone (1 µM) and IFNγ (5 U/mL) (33 °C, 5% CO2). These cells originate from mice harboring a temperature-sensitive mutation of the SV40 large tumor antigen expressed under the control of a IFN-γ regulated MHCII promoter [58]. Hence, they require IFNγ and permissive temperature of 33 °C for optimal growth. The murine embryonic fibroblasts (cFos+/+ 1–98 M wild-type and cFos-/- 7–98 M mutant) were cultured in DMEM supplemented with 4.5 g/L glucose, 10% FCS, 2 mM glutamine, 100 µg/mL streptomycin and 100 U/mL penicillin (37 °C, 5% CO2). Host cells were harvested weekly by trypsinization method for routine propagation and seeded for infection, as indicated elsewhere. To perform the microarray analyses, we infected YAMC epithelial or MEF (wild-type and cFos-knockout) cells with tachyzoites of T. gondii (MOI: 2) or sporozoites of E. falciformis (MOI: 6), resulting in a normalized infection rate of about 40% across the host-cell types. We were not able to separate uninfected and infected cells; therefore, an impact of bystander cells on our final gene expression datasets cannot be excluded.

RNA isolation and microarray hybridization

Total RNA was isolated from uninfected and infected cells suspended in TRIzol (PureLink RNA kit, Life Technologies)8, and analyzed for gene expression by dual-color hybridization using the whole-genome mouse microarray chips (4x44K, AMADID 014868, Agilent Technologies, Germany). We performed two biologically independent assays, each with dye-swapped replicates, and all sample-processing steps were executed according to the manufacturers’ protocol. Briefly, 5 µg of purified RNA was reverse-transcribed, amplified and labeled with Cyanine 3-CTP or Cyanine 5-CTP using oligo-dT-T7 promoter primer (QuickAmp kit, Agilent Technologies). After precipitation, purification, quality test and quantification, 1 µg of each cRNA preparation was fragmented and hybridized overnight to microarrays, followed by washing steps, as recommended. Images were recorded using a laser scanner (G2565CA) at 5-µm resolution.

Computational analysis of microarray data

The microarray data were analyzed with the Agilent image analysis and feature extraction software using default settings (G2567AA). Dye ratios were calculated using the most conservative estimate between the universal and propagated error. The extracted MAGE-ML files were evaluated with the Rosetta Resolver Biosoftware. Only anti-correlated genes of dye-reversal hybridizations with ≥1.5-fold and an error-weighted p-value ≤0.05 were considered as differentially regulated. Microarray data have been deposited in the Gene Expression Omnibus (GEO; ), and can be accessed with the GEO accession number (GSE157395). Heatmaps were visualized with the R package “pheatmap” and/or GraphPad Prism (v8.0). The gene enrichment analysis was performed using the Database for Annotation, Visualization and Integrated Discovery (DAVID, v6.7) [63], [64] (threshold of EASE score ≤0.1) and Kyoto Encyclopedia of Genes and Genomes (KEGG). cFos-centered network was predicted using the STRING program (v11.0) [65]. The primary genes directly networked with cFos were selected with a confidence score ≥0.4, while the secondary genes that were indirectly linked to cFos but directly networked with the primary genes were chosen with a confidence score ≥0.7.

Quantitative real-time PCR analysis

The YAMC epithelial cells infected with tachyzoites of the RH strain (MOI: 2; 16 h) were scraped in TRIzol solution for isolating the total RNA, which was immediately subjected to the cDNA synthesis using SuperScript III kit (Thermo Fisher Scientific). 10 ng cDNA of each sample was examined for the expression of designated transcripts by Platinum SYBR kit (20 μL reaction, Thermo Fisher Scientific). Samples from three independent assays were processed in duplicate reactions (Applied Biosystems 7300). The fold-change calculation was performed by the ΔΔCt method [66] using rps18 as a housekeeping gene with respect to corresponding uninfected control groups. Primers for qPCR analysis of cFos family proteins were designed as reported elsewhere [67] (see Table S7 for all primers).

Indirect immunofluorescence assay

The parasitized host cell monolayers grown on coverslips were fixed with 4% paraformaldehyde (10 min), followed by neutralization with 0.1 M glycine/PBS (5 min, RT) and permeabilization by 0.2% Triton X-100/PBS (20 min). Samples were treated with 2% BSA in 0.2% Triton X-100/PBS (20 min) to minimize any unspecific binding of antibodies, and then incubated for 1 h with the primary antibody (anti-TgGap45, 1:10000; anti-cFos, 1:200) or anti-Eimeria tenella serum (1:2000, cross-reactive to E. falciformis). Samples were washed 3x with 0.2% Triton-X100/PBS (5 min), followed by addition of secondary antibodies (Alexa Fluor488 or 594, 1:3000, 45 min). After 3x washing steps with PBS, samples were mounted in Fluoromount G/DAPI and kept in dark at 4 °C. Imaging was performed by a fluorescence microscope (AxioVision, Zeiss, Germany).

The parasite phenotyping

All assays with tachyzoites of T. gondii were set up essentially as reported previously [68]. Parasitized HFF cells (MOI: 2–3; 36–44 h infection) were washed with the culture medium, scraped, and extruded through a 27G syringe (2x) to collect fresh extracellular parasites. For the invasion assay, host cells seeded on glass coverslips were infected with tachyzoites for 1 h (MOI: 6). Cells were subsequently fixed (4% paraformaldehyde, 15 min), neutralized (0.1 M glycine/PBS, 5 min) and then blocked (3% BSA/PBS, 30 min). Uninvaded or extracellular parasites were stained with anti-TgSag1 antibody (1:10000, 1 h) prior to detergent permeabilization. Cultures were washed 3x with PBS (5 min), permeabilized with 0.2% Triton-X 100/PBS (20 min), and then treated with anti-TgGap45 antibody (1:10000, 1 h) to detect invaded or intracellular parasites. Samples were finally washed and incubated with Alexa488 and Alexa594-conjugated secondary antibodies (1:3000, 1 h). Invasion efficiency was determined by counting parasites stained with anti-TgGap45/Alexa594 (in), but not with anti-TgSag1/Alexa488 (out). Invasion by E. falciformis sporozoites (MOI: 6) was assessed 4 h post-infection by counting the host cells per high power field (10 HPFs/sample, 400x magnification) after staining with cross-reactive anti-E. tenella serum, as mentioned above.

To set up the replication assay, host cells grown on coverslips were infected with 3x104 tachyzoites (12–24 h infection) before fixation, permeabilization, neutralization, blocking and immunostaining with anti-TgGap45 and Alexa594 antibodies, as explained elsewhere. The cell division was assessed by enumerating intracellular parasites within their parasitophorous vacuoles. For the plaque assay, we cultured MEF cells to near-confluence in 6-well plates in standard culture medium, and then infected with 150–200 tachyzoites. The culture medium was adjusted to 1% FCS prior to infection to inhibit the overgrowth of uninfected cells, and cultures were kept at 37 °C for 7 days (12 days for Type II strain) without any perturbation. Samples were fixed with ice-cold methanol (-80 °C, 10 min), stained with crystal violet (12.5 g dye in 125 mL ethanol mixed with 500 mL of 1% ammonium oxalate) for 20 min and then washed with PBS to visualize plaques. The plaque size and number were measured by ImageJ program (National Institute of Health, USA).

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.