|Home | About | Journals | Submit | Contact Us | Français|
Conceived and designed the experiments: GMZ JCB. Performed the experiments: GMZ KLN. Analyzed the data: GMZ JCB. Contributed reagents/materials/analysis tools: GMZ JMT SMH JCB. Wrote the paper: GMZ JCB.
The apicomplexan parasite Toxoplasma gondii can infect and replicate in virtually any nucleated cell in many species of warm-blooded animals; thus, it has evolved the ability to exploit well-conserved biological processes common to its diverse hosts. Here we have investigated whether Toxoplasma modulates the levels of host microRNAs (miRNAs) during infection.
Using microarray profiling and a combination of conventional molecular approaches we report that Toxoplasma specifically modulates the expression of important host microRNAs during infection. We show that both the primary transcripts for miR-17~92 and miR-106b~25 and the pivotal miRNAs that are derived from miR-17~92 display increased abundance in Toxoplasma-infected primary human cells; a Toxoplasma-dependent up-regulation of the miR-17~92 promoter is at least partly responsible for this increase. The abundance of mature miR-17 family members, which are derived from these two miRNA clusters, remains unchanged in host cells infected with the closely related apicomplexan Neospora caninum; thus, the Toxoplasma-induced increase in their abundance is a highly directed process rather than a general host response to infection.
Altered levels of miR-17~92 and miR-106b~25 are known to play crucial roles in mammalian cell regulation and have been implicated in numerous hyperproliferative diseases although the mechanisms driving their altered expression are unknown. Hence, in addition to the implications of these findings on the host-pathogen interaction, Toxoplasma may represent a powerful probe for understanding the normal mechanisms that regulate the levels of key host miRNAs.
The intracellular protozoan parasite Toxoplasma gondii is a ubiquitous pathogen of warm-blooded animals with approximately one to two billion humans infected . While infections in immunocompetent adults are typically subclinical, infection persists for the life of the host, and serious disease can occur in fetal infections and in primary or recrudescent infection of immunocompromised persons. During host cell invasion, Toxoplasma uses a specialized set of secretory organelles to inject parasite-derived effector molecules into its host cell . Some of these effectors are known to interfere with host cell signaling pathways and alter host defenses although the means by which they do this are not known –. Toxoplasma appears to block apoptotic responsiveness in its host cells at myriad points  and to co-opt host apoptotic signaling pathways as an environment-sensing mechanism . Toxoplasma-infected cell cultures can also cede control of their cell cycle, progressing through G1/S and halting at G2/M and Toxoplasma attaches to, and invades host cells in S phase up to 4 times more efficiently than host cells in G1 –. In addition to these alterations in host cell signaling and cell cycle control, Toxoplasma specifically modulates host cell gene expression. Changes to the Toxoplasma-infected host transcriptome have been interrogated by expression-profiling and the data demonstrate that 24 hours post-infection, upwards of 15% of host mRNAs display altered abundance relative to uninfected cells ; similarly, quantitative analysis of the host proteome during Toxoplasma infection showed that the abundance of many host proteins is modulated in expression by Toxoplasma .
We have developed the hypothesis that some of these changes in host gene expression involve Toxoplasma co-opting host microRNAs (miRNAs), which are central components of an ancient post-transcriptional gene-regulatory mechanism that is highly conserved among the mammalian and avian hosts of Toxoplasma. MiRNAs are one of the most abundant classes of gene-regulatory molecules in the cell  and in silico target prediction suggests that miRNAs may be involved in the regulation of up to 30% of the human transcriptome . Mature miRNAs are ~23 nucleotide double-stranded RNA molecules that base-pair with target mRNAs. Most reports show that miRNAs act to negatively influence the translation and/or stability of target mRNAs , although under conditions of cell cycle arrest, some miRNAs appear to increase the translation of their target mRNAs . Some closely related miRNAs are grouped into families by virtue of containing identical seed sequences (the first 6-8 nucleotides of their mature 5′ ends); miRNA seeds are instrumental in target mRNA discrimination  and so individual miRNAs grouped into the same family are believed to be functionally equivalent , .
Although miRNAs were first identified due to their regulatory roles in sculpting metazoan development , , miRNA dysregulation also plays a central role in cancer and infection , -. The dysregulation of several miRNAs have been experimentally demonstrated to promote oncogenesis , . Examples relevant to the present work are four families of miRNAs (miR-17, miR-18, miR-19 and miR-25) that are encoded by three paralogous loci; these related loci, which are miR-17~92, miR-106b~25 and miR-106a~363 (see Figure 1B), produce primary transcripts that are post-transcriptionally processed to yield mature miR-17, miR-18, miR-19 and miR-25 family members. MiR-17~92 is also known by the synonym OncomiR-1 due to the observed acceleration of lymphomagenesis when these miRNAs are over-expressed in Eµ-Myc mice .
Functionally, slight perturbations to the levels of mature miR-17~92-derived miRNAs can have profound biological consequences; knock-in mice engineered to produce twice as much miR-17~92-derived miRNAs as wildtype mice suffer lymphoproliferative phenotypes, autoimmunity and a shortened life-span linked to the loss of the PTEN tumor suppressor and the pro-apoptotic factor Bim, which are both validated targets of miR-17~92 . Indeed, the miR-17~92 and miR-106b~25 clusters appear to attenuate apoptotic responsiveness by targeting several mRNAs encoding pro-apoptotic effectors and favor progression from G1 to S-phase by targeting mRNAs that encode negative regulators of the cell cycle , .
To investigate whether Toxoplasma infection drives changes to host miRNA levels, we profiled host miRNA levels during Toxoplasma infection with miRNA microarrays and confirmed these results by independent means. The data reveal numerous, Toxoplasma-specific changes to host miRNA levels and some of these changes are at least partly mediated at the transcriptional level.
To investigate whether Toxoplasma infection alters the steady-state levels of host miRNAs, we profiled the levels of miRNAs extracted from primary human foreskin fibroblasts (HFFs) during a 24-hour Toxoplasma infection timecourse (where host RNA was extracted at 6-hours, 12-hours and 24-hours from mock- and from Toxoplasma-infected cells) using miRNA microarrays spotted with duplicate probes that should hybridize to all human, mouse and rat miRNAs included in miRbase v.8. When we compared hybridization intensities between arrays that were hybridized with RNA derived from uninfected HFFs and Toxoplasma-infected HFFs, the results demonstrated that at 24-hours post-infection, ~14% of the spots displayed dissimilar hybridization intensities (Figure S1). We focused our attention on 18 microarray spots (9 closely related human and mouse probe sequences that were spotted in duplicate) that displayed comparable increases in hybridization intensities on arrays hybridized with RNA derived from infected HFFs (Figure S1, red boxes); these 18 spots contained probes that hybridized to members of the miR-17 family (the miR-17 family is composed of miR-17, miR-106a, miR-106b, miR-20a, miR-20b and miR-93; see Figure 1A for a sequence alignment of the miR-17 family). As miR-17 family members are co-transcribed with members of the miR-18 (Figure S1, blue box), miR-19 and miR-25 families, and are encoded in three separate paralogous loci (miR-17~92, miR-106a~363 and miR-106b~25; Figure 1B), we assembled a heat-map from our microarray data that contains the averaged fold-change values for all probes that hybridized to members of the miR-17, miR-18, miR-19 or miR-25 families (18, 4, 4, and 4 probes, respectively; Figure 1C). When evaluating these data it is important to note that within these individual miRNA families, mature miRNAs have nearly identical nucleotide sequences, and are virtually indistinguishable from the other members of their own miRNA family by microarray, northern blotting, and conventional primer-extension analyses.
The microarray profiling data showed that the miR-17 family and the miRNA families that are co-transcribed with it were generally more abundant in RNA samples from Toxoplasma-infected cells and that the abundance of these miRNAs marginally increased at 6-hours post-infection and showed greater increases at 12-hours and 24-hours post-infection (Figure 1C). MiR-21 is a highly abundant unrelated miRNA in HFFs that is derived from an independent intergenic locus and is included in the microarray heat-map and subsequent analyses for comparison.
To validate the microarray data and to specifically measure the effects of Toxoplasma infection on miRNAs encoded by miR-17~92, miR-106a~363 and miR-106b~25, we took two orthogonal approaches. First, we performed high-resolution northern blot analysis for miR-17. Consistent with the microarray data, the northern blot results showed that, collectively, the level of mature miR-17 family members increased ~2.7-fold in Toxoplasma-infected HFFs relative to uninfected HFFs (Figure S2). The data also showed that the pre-miR-17 signal was a small fraction of the hybridization signal observed for mature miR-17, suggesting that the contribution of pre-miR-17 to the observed miRNA microarray hybridization intensity was negligible, an interpretation that is consistent with previous data .
The second approach used to follow up on the microarray results was primer-extension analysis with primers specific for each miR family. The raw data shown in Figure 2A confirmed that the levels of mature miR-17 family members increased as a function of time in RNA samples derived from Toxoplasma-infected HFFs. Primer-extensions performed with probes specific for 5S rRNA and miR-21 gave very similar patterns to each other with no significant effect of Toxoplasma infection (Figure 2A). MiR-18, miR-19 and miR-25 family members also showed an increase in abundance in RNA samples derived form Toxoplasma-infected HFFs (Figures 2C). A graphical representation of all replicate primer-extensions normalized to 5S rRNA is plotted in Figure 2B. Upon Toxoplasma infection, miR-17 family members collectively increased ~3-fold, and miR-19 and miR-25 family members increased ~1.5-fold. When normalized to 5S rRNA, miR-18 also increased upon Toxoplasma infection, but the measurements for the miR-18 primer-extensions were complicated by low signal and high background (see Figure 2A), making its measurement less reliable; it was therefore not included in Figure 2B. Similar primer-extension analyses were performed with probes for the miR-17 family and 5S rRNA on RNA samples derived from two Toxoplasma-infected mouse embryonic fibroblast lines (C57/B6/129  and NIH 3T3 ) and from a human B-cell line (Raji ); the results of these experiments were consistent with the primary HFF results detailed above (data not shown).
The data from the miRNA microarray profiling and primer-extension analysis were generally in agreement except in one instance. The microarray data (Figure 1B) indicated that the miR-25 family of miRNAs did not appreciably increase in arrays hybridized with RNA derived from Toxoplasma-infected cells at the 24 hour timepoint whereas primer-extension for miR-92 (the miR-25 paralog derived from the miR-17~92 locus) indicated a modest but reproducible increase (Figure 2B). The Cy5 channel intensities of the miR-25 family probe spots were near the upper limits of the dynamic range of the microarrays, which likely obscured the signal increase seen in the primer-extensions. Excluding this discrepancy, the primer-extension data closely approximate the results of the northern blot and miRNA microarray profiling and confirm that Toxoplasma increases the levels of mature miR-17, miR-18, miR-19 and miR-25 family members in infected human cells.
To investigate whether the observed increase in host miR-17 levels is a specific response to Toxoplasma infection or a general response of apicomplexan-infected host cells, we assayed the levels of host miR-17 in HFFs infected with Neospora caninum. Neospora and Toxoplasma are very closely related coccidians, have similar asexual growth cycles, and are morphologically almost indistinguishable . Although Neospora is not a human pathogen and has a far more restricted host range than Toxoplasma, Neospora readily infects HFFs and grows with kinetics comparable to Toxoplasma in these cells . To assess whether Neospora infection changes the levels of host miR-17 family in a manner similar to Toxoplasma infection, we performed primer-extension analysis on RNA samples derived from HFFs comparably infected with Neospora. The results showed that Neospora infection did not significantly alter the levels of host miR-17 family members using either a miR-17 probe (Figure 3A and Figure 3B), or miR-20 probe (data not shown) at any time-point in our assay. We conclude that the increased level of miR-17 family members seen in Toxoplasma-infected cells is Toxoplasma-specific and is not a general host response to apicomplexan infection.
To investigate whether Toxoplasma manipulates the levels of pri-miRNAs encoding miR-17, miR-18, miR-19 and miR-25 family members, and to query which of the paralogous clusters encoding miR-17 family members are manipulated by Toxoplasma, we performed northern blot analysis on pri-miR-17~92, pri-miR-106a~363 and pri-miR-106b~25 (Figure 4). This was done using probes that hybridize to sequences that are unique to each cluster (sequences that encode the pre-miRNA hairpins are not recognized by the probes; sequences provided in Figure S3), thereby allowing us to discriminate between the three paralogous clusters. The hybridization products with different probes were quantified by phosphorimager analysis, normalized to the signal for the constitutively expressed RPS29 to correct for loading variation, and expressed as a fraction of the signal derived from uninfected HFFs for each probe. A parallel analysis was performed using a probe for AldoA as a loading control and yielded comparable numbers to those obtained by normalization with RPS29 (data not shown).
Figure 4 shows that hybridization with the pri-miR-17~92 probe yields a ~3.2kb band which is a size previously published for the pri-miRNA derived from this cluster . The level of pri-miR-17~92 in HFFs infected with Toxoplasma relative to uninfected HFFs was ~3.6-fold greater whereas HFFs infected with Neospora showed no such increase (~0.7-fold relative to uninfected HFFs). When the membrane was stripped and hybridized with a pri-miR-106b~25 probe, infection with either Toxoplasma or Neospora showed a major band at ~850nt, and this band demonstrated an increase of ~2.5-fold and ~2.6-fold, respectively, relative to uninfected HFFs (Figure 4). There are no previous reports of pri-miR-106b~25 northern blots and so, while we cannot be certain that this band corresponds to pri-miR-106b~25 or to a splice variant of MCM7, it was consistently seen in multiple experiments and is within the size range expected for this molecule. No reproducible pri-miR-106a~363 hybridization products were visible (data not shown).
These results suggest that the mature miR-18 and miR-19 family members that increase upon infection with Toxoplasma are derived from miR-17~92; although pri-miR-106b~25 is also increased in Toxoplasma-infected cells, miR-106b~25 does not encode miR-18 and miR-19 family members and, consistent with previous reports , , no pri-miR-106a~363 was detectable.
To investigate whether the Toxoplasma infection-dependent increase in pri-miR-17~92 abundance is the result of increased transcription from this locus, we performed a dual luciferase assay with a reporter plasmid  containing a 1.3kb fragment located upstream of C13ORF25 fused to a promoter-less firefly luciferase (FLUC) cassette (pro1353::FLUC; see Figure 5A); this 1.3kb upstream fragment of C13ORF25 overlaps the 5′ terminus of the C13ORF25 expressed sequence tag (EST) by 10 nucleotides and has been previously shown to drive firefly luciferase expression . The pro1353::FLUC plasmid was co-transfected into HFFs with a plasmid that constitutively expresses Renilla luciferase (RLUC; pRL4.75). The transfected cells were either mock- or Toxoplasma-infected and the ratio of FLUC/RLUC was determined at 18 hours post-infection.
The data demonstrate that Toxoplasma-infected HFFs show a 4.3-fold increase in FLUC/RLUC relative to mock-infected HFFs (p=8.8×10−5; Figure 5B). These data strongly suggest that the increased level of pri-miR-17~92 seen in Toxoplasma-infected HFFs is a result of increased miR-17~92 transcription rather than alterations in the stability of pri-miR-17~92, as only 10 nucleotides from the pri-miR-17~92 transcript are present in the luciferase reporter construction.
The results presented here demonstrate that Toxoplasma infection specifically drives an increase of 2–3 fold in the levels of mature miR-17~92-derived miRNAs in primary human foreskin fibroblasts. The mature miRNAs encoded by miR-17~92 and its paralog miR-106b~25 play important roles in mammalian biology. MiR-17~92 is crucial in development, as mice harboring a homozygous deletion of miR-17~92 die shortly after birth due to pulmonary and cardiac defects and develop abnormal B-cell lymphocytes . In adult animals, miR-17~92 and miR-106b~25 have been shown to influence the functionally intertwined pathways of apoptosis and G1/S cell cycle progression by targeting multiple components of each pathway . The importance of these functions have been demonstrated by showing that retroviral overexpression of a cassette containing miR-17, miR-18 and miR-19 in mice results in c-Myc-induced lymphoma . More subtly, transgenic mice that have been engineered to overexpress miR-17~92 by 2-fold in lymphocytes develop lymphoproliferative disease and autoimmunity and die prematurely, phenotypes that were shown to be partially a consequence of the translational repression of PTEN and Bim . Hence, even relatively modest increases in the levels of miR-17~92, of a magnitude less even than seen here with Toxoplasma infection, have profound biological consequences in vivo.
The ultimate, downstream target of these changes in miRNAs, in terms of benefit to the parasite, is not clear; in the set of HFF mRNAs that decrease in abundance upon Toxoplasma infection , there is no significant enrichment for mRNAs containing predicted miR-17~92 or miR-106b~25 binding sites (data not shown). This lack of apparent enrichment could be due to the very large number of changes induced by Toxoplasma through other means; for example, many transcription factors are up-regulated in Toxoplasma-infected cells, which significantly increases the background for this kind of bioinformatics analysis. It could also signify that the critical, intended effect of modulating these miRNAs is on translation rather than transcription , .
To understand the role of increased miR-106b~25 and/or miR-17~92 in Toxoplasma infection we took several approaches. First, we infected primary mouse fibroblasts (MEFs) harboring a miR-17~92 deletion . The results showed no reproducible difference in Toxoplasma growth in such cells relative to wild type controls (data not shown). Second, concurrent analyses using locked nucleic acid-mediated knockdown of miR-17~92- and miR-106b~25-derived miRNAs was performed using a wide range of conditions in mouse fibroblasts and HFFs. The results showed no reproducible effects of knockdown of these miRNAs on the ability of Toxoplasma to infect, grow within and lyse the host cell in vitro (data not shown). This absence of an effect is consistent with the fact that while many of these miRNAs are extremely important in the animal , , they are largely dispensable for simple, in vitro growth of cell lines (miR-17~92 knockout MEFs have no obvious in vitro growth defect and, in general, most commonly used transformed cell lines dramatically overexpress miR-17~92). Extensive Affymetrix microarray profiling of Toxoplasma-infected wildtype- or miR-17~92-knockout MEFs showed no statistically significant differences in mouse mRNA levels 24h post infection (data not shown).
To address what the important biological target(s) of these miRNAs are in the context of Toxoplasma infection, in vivo studies will be needed. Unfortunately, as homozygous miR-17~92 knockout mice are not viable  dissecting the impact of these genes on Toxoplasma pathogenesis is technically difficult. The situation is further complicated by the fact that both pri-miR-17~92 and pri-miR-106b~25 are increased upon Toxoplasma infection and these two clusters are believed to be partially functionally redundant. Ultimately, a conditional knockout of miR-17~92 in a miR-106b~25-null background will be necessary to score the full biological role(s) of these miRNAs in Toxoplasma-infected animals.
Whatever the biological role of these miRNAs in Toxoplasma infection, the results presented here demonstrate that increases in miR-17~92-derived miRNAs are a specific response to Toxoplasma infection. During such infection only a subset of miRNAs are affected and an increase in miR-17~92-derived miRNAs is not seen when cells are similarly infected with the closely related coccidian Neospora. To our knowledge this represents the first report of a pathogen that specifically increases the levels of miR-17~92 and/or miR-106b~25. Thus, in addition to the importance of understanding the host-pathogen interaction, Toxoplasma and Neospora could prove extremely useful as probes for dissecting the biogenesis and function of miRNAs in normal, uninfected human cells.
The levels of miR-17~92-derived miRNAs reported here are at least partially due to a Toxoplasma-dependent increase in pri-miR-17~92 transcription. The transcription of pri-miR-17~92 has been shown to be positively regulated by E2F3  and c-Myc  transcription factors, and pri-miR-106b~25 is positively regulated by E2F1 . Expression-profiling of Toxoplasma-infected HFFs  demonstrated that there are modest increases in the levels of E2F3 (+1.4-fold), c-Myc (+1.9-fold) and E2F1 (+1.5-fold) in Toxoplasma-infected HFFs relative to uninfected HFFs at 24h post-infection. It is, therefore, probable that the Toxoplasma-dependent increase in the levels of pri-miR-17~92 and pri-miR-106b~25 are at least in part due to such changes in host transcription factors.
Whether the increases in miR-17~92-derived miRNAs are mediated by c-Myc, E2F or another pathway, the identity of the parasite factor(s) that drive these changes are unknown. These changes could involve one or more of the recently reported effectors that the parasite injects into the host cell during infection. These effectors include protein phosphatases  and protein kinases , , one of which has been shown to influence Toxoplasma virulence by dramatically altering host transcription through interacting with STAT3 and STAT6 functions . Whether this or one of the other injected Toxoplasma effectors is responsible for the alterations in host miR-17~92 and miR-106b~25 expression, and what role these changes play on the host-pathogen interaction in vivo are the focus of on-going work.
Parasites were propagated in primary HFFs and Vero cells using standard cell culture techniques as described previously . Manipulations and host cell passage histories are expanded in the Detailed Materials and Methods (Appendix S1).
RNA was extracted with TRIzol (Invitrogen), and fractionations were performed with the miRvana™ miRNA isolation kit (Ambion) or the Oligotex kit (Qiagen) according to the manufacturer's instructions. Expanded RNA methods are provided in the Detailed Materials and Methods (Appendix S1).
Spotted oligonucleotide mirVana™ miRNA Probe Set version 2 (Ambion) miRNA microarrays were purchased from the Stanford Functional Genomics Facility. Extracted miRNAs were 3′ end labeled with pCp-Cyanine dyes (pCp-Cy3 for common reference, which was a mixture of all hybridized samples and pCp-Cy5 for each individual experimental sample). pCp-Cyanine dyes were synthesized at the Stanford PAN facility. Arrays were hybridized, washed and processed according to the manufacturer's instructions (Ambion). Expanded microarray methods are provided in the Detailed Materials and Methods (Appendix S1).
MiRNA and pri-miRNA northern blotting were performed essentially as previously reported . For pri-miRNA northern blotting, 100 µg total RNA samples were fractionated with Oligotex (Qiagen) according to the manufacturer's instructions prior to 1% agarose/formaldehyde separation and capillary transfer. Expanded northern blot methods are provided in the Detailed Materials and Methods (Appendix S1).
miRNA microarray profiling data of Toxoplasma-infected primary human foreskin fibroblasts. Columns are individual microarrays hybridized with labeled, size-fractionated RNA derived from mock infected HFFs or Toxoplasma-infected HFFs at the indicated time-points. Heatmap scale is from −3 to +3, and these values reflect the log2-transformed ratios of the hybridization intensities of Cy5-labeled sample/Cy3-labeled common reference for each array. Common reference RNA was a pooled mixture of RNAs from all samples. Data was filtered in SMD according to the default settings, and only spots that were called ‘present’ on 80% of the arrays were included in the heatmap. Log2-transformed ratios of the hybridization intensities of sample/common reference for each spot on each array were hierarchically clustered (Euclidean) in MEviewer (MeV4.2; TIGR) by arrays and to genes. Red boxes are miR-17 family members; blue boxes are miR-18 family members. has=human, mmu=mouse, rno=rat, ambi=proprietary Ambion probe sequences.
(9.27 MB TIF)
Northern blot analysis demonstrates that Toxoplasma infection results in increased levels of mature miR-17 family members. RNA samples derived from uninfected HFFs (- lanes), and from HFFs infected with Toxoplasma for 24h (+ lanes) were resolved through a 10% acrylamide/8M urea gel, transferred to nylon membrane, and hybridized to miR-21, miR-17 and U6 oligonucleotide probes (corresponding probes are indicated underneath autoradiographs). The U6 snRNA hybridization is shown as a loading control. Putative pre-miRNA bands and mature miRNA bands are indicated with arrows. A co-electrophoresed 10nt ladder is shown for nucleotide size comparison.
(0.42 MB TIF)
Probe sequences used in this study. All oligonucleotides are listed from 5′ to 3′. In ‘notes’ column, target nucleotide positions are provided for all primer-extension probes. The forward and reverse primers used in PCR amplification of the pri-miR-17~92, pri-miR-106a~363, pri-miR-106b~25, AldoA, MCM7 and RPS29 probes are shown.
(0.33 MB DOC)
We thank C.Z. Chen and members of the Boothroyd and Sarnow laboratories for helpful suggestions, discussions and reagents through the course of this work, J. Boyle, J. Coller and A. Brotke for assistance with microarrays and data analysis, and E. Sheagley and S. Curran for critical reading of the manuscript.
Competing Interests: The authors have declared that no competing interests exist.
Funding: G.M.Z. was supported by the NIH (5F32AI066538). K.L.N was supported by the Alberta Heritage Foundation for Medical Research postdoctoral fellowship. J.M.T. was supported by the American Cancer Society (PF-05-152-01). S.M.H. was supported by the NIH (5R01GM070674). J.C.B. was supported by the NIH (AI21423 and A173756). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.