T. gondii encodes three separate mRNA capping enzymes
To examine the presence of 7-methylguanosine (m7G) cap in Toxoplasma, we performed immunofluorescence analysis using an anti-m7G antibody in the asexual stages (tachyzoite and bradyzoite) of the parasite. Analysis of intracellular parasites revealed predominant punctate staining for m7G at the nuclear periphery and cytoplasm of tachyzoite (Tz) and bradyzoite (Bz) (Fig. 1A). To detect whether the m7G mark is present on RNA, we performed an immuno-dot blot using an anti-m7G antibody on total RNA and genomic DNA (gDNA) extracted from filter-purified tachyzoites. The results demonstrated that the m7G mark was exclusively present in RNA (Fig. 1B). Collectively, these results show that m7G cap RNA is a feature of Toxoplasma.
To identify capping enzymes, which could add m7G structure to RNA, we performed BLASTP homology searches of the Toxoplasma genome (http://toxodb.org/toxo/) using amino acid sequences of yeast Cet1, Ceg1, and Abd1 proteins as queries. The search analysis revealed three separate RNA capping enzymes in Toxoplasma that are similar to yeast. The identified candidate Toxoplasma RNA triphosphatase, guanylyltransferase, and guanine-N7 methyltransferase were named TgRT (TGME49_224650), TgGT (TGME49_305320), and TgGMT (TGME49_272720), respectively (Fig. 1C). The MEGA analysis was performed to determine how closely the Toxoplasma capping enzymes (TgCEs) are related to the alveolates' and other eukaryotes' capping proteins. Phylogenetic analysis revealed a distinction between two apicomplexan classes, Conoidasida and Aconoidasida, for all three capping proteins. Members of the Sarcocystidae family were found to encode the largest triphosphatase and methyltransferase proteins (Supplementary Fig. 1A-C).
The putative TgRT gene encodes a 920-aa polypeptide and has characteristic features of fungal triphosphatases9, including two glutamate-containing metal-binding motifs, homologs of β strands that comprise the active site tunnel, and conserved hydrophilic amino acids required for catalysis. (Fig. 1D, Supplementary Fig. 2). The putative TgGT is a 509-aa protein contains two conserved domains, a nucleotidyl transferase (NTase) domain and a C-terminal oligonucleotide-binding domain (OB)41,42. The NTase domain has six conserved motifs (I, III, IIIa, IV, V, and VI)43 (Supplementary Fig. 3) with a lysine-containing KxDG motif I (Fig. 1D), which comprises the active site of GTP-binding and nucleotidyl transfer44,45. The putative TgGMT encodes 1283-aa polypeptide with large N-terminal extension and conserved glycine-rich sequence in the SAM-binding motif (Fig. 1D). Among all three capping enzymes, TgRT and TgGMT are significantly larger than yeast's TPase and guanine-N7 MTase; however, each of the three Toxoplasma capping enzyme orthologue is essential for parasites according to a genome-wide CRISPR screen (Fig. 1C), suggesting that M7G capping of RNA is important and required for parasite fitness.
Owing to the sequence similarity between Toxoplasma capping enzymes (TgCEs) and yeast enzymes, we performed yeast complementation to test whether TgCEs function in the cap-synthetic pathway and sustain the growth of yeast cells that lack one or more capping enzymes. We separately cloned the TgRT, TgGT, and TgGMT genes into a yeast 2µ TRP1 pYES3 plasmid, and the function of each of these genes was tested by plasmid shuffle in S. cerevisiae Δcet1 or Δceg1 or Δabd1 cells that contain respective gene on a CEN URA3 plasmid. The mutant strain cannot survive on a medium containing 5-FOA, a drug that selects against the URA3 plasmid unless it is transformed with a second plasmid containing a functional homolog from another source. We found that 2µ TgRT, TgGT, and TgGMT supported the growth of Δcet1, Δceg1, and Δabd1 cells, respectively (Fig. 1E-G). Similarly, 2µ TgRT + TgGT + TgGMT supported the growth of triple mutant yeast cells (Fig. 1H). These results demonstrate that Toxoplasma encodes biologically active capping enzymes.
Toxoplasma RNA Triphosphatase TgRT shows metal-dependent triphosphatase activity
To gain insight into the expression, localization, and biochemical function of TgRT, full-length His6-TgRT protein of ~ 100 kDa was purified (Fig. 2A) and used to generate specific anti- TgRT antibodies. The expression and localization studies revealed that TgRT is robustly expressed (Fig. 2B) in both the asexual stages and localized in the nucleus of the tachyzoite and bradyzoite stages (Fig. 2C), as demonstrated using anti-TgRT antibodies. Further, we tested the triphosphatase activity of TgRT using a detailed biochemical characterization. The recombinant wild type (WT) TgRT catalyzed the release of 32Pi from [γ-32P]ATP in the presence of manganese (Fig. 2D); however, the ATP hydrolysis was ineffective in the presence of magnesium (Fig. 2D). The extent of ATP hydrolysis was proportional to TgRT concentration (Fig. 2E). No ATP hydrolysis was observed without divalent cation (Fig. 2E). ATP activity was nominal in the presence of calcium or zinc (Fig. 2F). The optimal ATPase activity of TgRT was detected as early as 5 min and remained stable for 30 min (Fig. 2G). The enzyme activity of TgRT was similar from 20oC to 70oC (Fig. 2H). TgRT was found to be catalytically active in a wide pH range (5.5–10) with an optimal activity from pH 6.5 to 8.0 (Fig. 2I). We also tested the specificity for NTP hydrolysis using two different triphosphorylated nucleoside substrates. The rate of release of 32Pi from [γ-32P]ATP was similar to the rate of conversion of [α-32P]ATP to [α-32P]ADP and [α-32P]UTP to [α-32P]UDP in a parallel reaction mixture containing the same concentration of TgRT (Fig. 2J). Two glutamate residues corresponding to the metal binding sites of the tunnel were replaced by alanine (E366A and E726A) (Fig. 1D and 2K), and enzyme activities were compared. ATPase activity of the E366A mutant was 50% of the activity of wild-type TgRT (Fig. 2L), whereas in comparison, E726A showed < 10% activity (Fig. 2L). Less activity in the mutant proteins was due to a change in the secondary structure, as demonstrated by circular dichroism (Supplementary Fig. 3). Together, these results demonstrated that TgRT belongs to the family of triphosphate tunnel metalloenzymes (TTMs).
Characterization of Toxoplasma Guanylyltransferase TgGT
The full-length His6-TgGT protein of ~ 60 kDa was purified (Fig. 3A) and used to generate anti-TgGT antibodies. TgGT is robustly expressed in the asexual stages (Fig. 3B) and localized in the nucleus of the tachyzoite and bradyzoite stages (Fig. 3C). All known guanylyltransferases accomplish nucleotidyl transfer through a covalent enzyme-(lysyl-N)-GMP intermediate that can be detected by label transfer from [α-32P]GTP to the enzyme. To determine the guanylyltransferase activity of TgGT, protein was incubated with [α-32P]GTP and a divalent cation, which resulted in the formation of an SDS-stable ~ 60 kDa enzyme-GMP intermediate (Fig. 3D). TgGT activity requires a divalent cation cofactor, either manganese or magnesium; however, enzyme activity was more effective in the presence of manganese than magnesium (Fig. 3D). Enzyme-guanylate formation was linear with respect to TgGT concentration (Fig. 3E). We mutated conserved Lys (essential for GMP interaction during the guanylyltransferase reaction for GTases)39, Tyr, Glu, and Gly residues to alanine (K133A, T134A, D135A, and G136A) of TgGT and compared enzyme activities (Fig. 3F). None of the mutant proteins showed GTase activity (Fig. 3G); however, after long autoradiographic exposure, a trace of residual activity was observed for T134A (Fig. 3G). No activity in the TgGT mutant proteins was due to a change in the secondary structure, as demonstrated by circular dichroism (Supplementary Fig. 4). We conclude that the observed guanylyltransferase activity is intrinsic of TgGT.
Characterization of Toxoplasma guanine-N7 methyltransferase TgGMT
His6−TgGMT716 − 1283 (subscript denotes amino acid coordinates) protein of ~ 60 kDa was purified (Fig. 4A) and used to generate anti-TgGMT antibodies. The expression and localization studies showed TgGMT expressed in the asexual stages (Fig. 4B) and primarily localized in the nucleus of the tachyzoite and bradyzoite stages (Fig. 4C). To test the guanine-N7 methyltransferase activity of TgGMT, we first generated 32 mer RNA (pppN32RNA) using in vitro transcription and sequentially treated with TgRT to generate ppN32, TgGT to generate GpppN32, and TgGMT in the presence of methyl donor S-adenosyl methionine (SAM) to generate m7GpppN32, as shown in Fig. 4D. The individual reaction was spotted on the membrane and capping of RNA substrate was determined using anti-m7G antibody. The immune-blot analysis revealed that TgGMT could successfully add m7G to the guanylated RNA (Fig. 4E). Vaccinia Capping Enzyme (VCE) was used as a positive control to generate m7GpppN32 (Fig. 4E). The specificity of TgGMT to use SAM was tested using sinefungin, a structural analog of SAM and inhibitor of methyltransferases46,47. Immunoblot analysis using anti-m7G antibody revealed that sinefungin inhibits m7GpppN32 synthesis in a concentration-dependent manner (Fig. 4F). Together, these results demonstrate that TgCEs are biochemically active and function in a cap-synthetic pathway in Toxoplasma.
Cap-dependent translation in vitro
The eukaryotic translation initiation factor, eIF4E, binds to the m7G cap of mRNA and initiates translation48,49. Similarly, we wanted to test whether in vitro-generated capped RNA could be recognized by eIF4E-like protein and productively translated into protein in Toxoplasma. BLASTP search using Plasmodium50 and human51 eIF4E revealed three eIF4E-like homologues (TGME49_223410, TGME49_315150, and TGME49_312560) in T. gondii genome52. Of these three homologue, TGME49_223410 showed highest amino acid similarity, including conserved residues required for m7G interaction with Plasmodium (e score: 1e-69) and human protein (e score: 1e-14) (Supplementary Fig. 6). While the eIF4E protein is highly conserved in eukaryotes, TgeIF4E clusters with fungi and other protozoan parasites, and the metazoans form a separate clan (Supplementary Fig. 7). Hence, we name this putative protein TgeIF4E, a 225-aa protein (Fig. 5A). The full-length TgeIF4E-His6 protein of ~ 26 kDa was purified (Fig. 5B) and used to generate specific antibodies. TgeIF4E is robustly expressed in the asexual stages (Fig. 5B) and, as expected, localized in the cytoplasm of the tachyzoite and bradyzoite stages (Fig. 5C).
To determine the interaction between TgeIF4E with capped or non-capped RNA variants generated using TgCEs, a microscale thermophoresis (MST) assay was performed using fluorescently labeled TgeIF4E. As measured, TgeIF4E showed strong binding affinity (KD=8.01 ± 1 nM) towards m7GpppN32RNA (Fig. 5F) and no binding affinity was measured for pppN32RNA (Fig. 5D) and GpppN32RNA (Fig. 5E). These results confirm the m7G cap specificity of TgeIF4E. Further, to evaluate m7G RNA were utilized by Toxoplasma to promote translation, we generated luciferase transcript variants in vitro with and without 5’-m7G cap (using TgCEs) and 3’-poly(A)tails (Fig. 5G). We found that luciferase activity was detected after transfection of capped and polyadenylated RNA (m7G-FLuc-[A]n), but not after transfected with either only IVT RNA (pppFLuc) or capped RNA (m7G-FLuc) or polyadenylated RNA (pppFLuc-[A]n)14 (Fig. 5H). As a positive control, 5’capped and 3’polyadenyaled FLuc RNA was generated using VCE and PolyA polymerase and transfected. However, after comparing the luciferase activity, we found that TgCEs are less effective than VCE (Fig. 5H). Together, these results demonstrate that in vitro-generated capped RNA using TgCEs can be utilized by Toxoplasma to promote translation in vivo.
Depletion of TgRT impairs overall m7G levels and arrest parasite replication
Toxoplasma RNA triphosphatase is a metalloenzyme unique to the parasite. To understand better the role of this protein in the transcription-associated processes, we endogenously tagged TgRT at the C-terminus with mini auxin-inducible degron (mAID) fused with three copies of HA (3HA) epitope (TgRT-mAID-3HA) in RH strain parasites expressing TIR1, which allows rapid degradation of the TgRT-mAID-HA protein (Fig. 6A) upon the addition of indole 3-acetic acid (IAA). The resulting TgRT-mAID-HA strain was confirmed by diagnostic PCR (Fig. 6B). Western blot analysis using anti-HA antibody revealed complete loss of the TgRT-mAID-HA protein in as little as 1 h following addition of 500 µM IAA in the culture medium (Fig. 6C). Using IF analysis, we also observed no staining for TgRT-mAID-HA in the 1 h IAA treated parasites (Fig. 6D). Consistent with the role in the mRNA capping, depletion of TgRT for 8 h diminished m7G capped RNA in the parasite (Fig. 6E and 6F). Initially, we checked the m7G levels of total parasite RNA of control vs treated parasites at 1 h, 4 h, and 8 h; however, we could observe a significant decline in m7G levels after 8 h of TgRT depletion. Hence, we selected an 8 h IAA treatment period for all the further experiments. The effect of TgRT depletion on productive translation was examined by transfecting TgRT-mAID-HA parasites with Luciferase plasmid. As measured, the luciferase activity was decreased after 4 h, and a drastic reduction in the activity was observed after 8 h of RT depletion (Fig. 6G). Collectively, these results show that depletion of TgRT leads to reduced RNA capping in the parasite.
We next examined the effect of TgRT depletion on parasite-specific processes. The complete arrest of parasite replication was observed in the TgRT-depleted parasites tested using a standard parasite counting assay (Fig. 6H). A significant number of parasites displayed morphological defects upon TgRT depletion (Fig. 6I). The impact of TgRT depletion on parasite growth was tested using plaque assays. Unlike the parental strain, TgRT-depleted parasites produced no visible plaques (Fig. 6J-L). To determine whether parasites can recover from a transient loss of TgRT expression, we conducted a plaque assay that used six different IAA treatment regimens (Fig. 6M). After an initial 24 h of growth, parasites were treated with IAA or vehicle for 1/2/4/8/12/24 h; at this point, the media was replaced with fresh IAA (or vehicle) and incubated for 5 days. The number of plaques formed was similar for 1 h IAA or vehicle-treated parasites (Fig. 6M and 6N). Two hours of IAA-treated TgRT-mAID-HA parasites showed a 50% reduction in the number of plaques formed, whereas no plaques were observed when treated for 4 h or 8 h or 12 h, or 24 h IAA (Fig. 6M and 6N). These results indicate that parasites could not recover after depletion of capping for > 4 h. The depletion of TgRT significantly decreased the invasion efficiency of the parasite (Fig. 6O); however, no difference was observed in the ability of parasites to egress upon inducing egress using calcium ionophore (Fig. 6P). Together, these data provide compelling evidence that TgRT is essential for Toxoplasma viability and proliferation.
Depletion of TgRT perturbs gene expression
Given that M7G RNA capping governs overall RNA metabolism, we reasoned that the loss of parasite viability upon TgRT depletion could be due to a defect in the capped RNA abundance and gene expression. We investigated the consequences of TgRT depletion (the absence of capping) on gene expression using cap sequencing, which was aimed at capturing all the m7G capped RNA population to provide a transcriptome-wide profile that will reveal the identity and relative levels of capped RNA. The m7G-capped RNAs were enriched using sequential enzyme treatments (Fig. 7A) on 8 h IAA- or vehicle-treated TgRT-mAID-HA parasites (in biological duplicates). In the first step for the enrichment capped RNA, total RNAs were treated sequentially with RppH (in the presence of NEBbufer2) and Xrn1 to remove rRNA and 5’-monophosphate-ended RNA (Fig. 7B). Subsequently, RNAs were treated with RppH in the presence of Thermopol buffer to convert M7G-capped RNA (decapping) to monophosphate-ended RNA (Fig. 7C). These decapped RNAs were subjected to library preparation and sequencing, and the obtained transcriptomic data were utilized for differential gene expression analysis (DESeq). The gene expression data from two biological experiments were analyzed using a k-means clustering algorithm. Clustering of 2000 transcripts revealed four clusters (A, B, C, and D) with similar expression profiles in two biological replicates of IAA or vehicle samples (Fig. 7D). The significantly altered capped transcripts were identified by edgeR (FDR ≤ 0.005; fold change ≥ 1), resulting in 185 and 186 genes whose abundance was decreased and increased, respectively, upon depletion of TgRT (Fig. 7E). To understand what processes the differentially expressed genes might be involved in, we performed gene ontology (GO) enrichment analyses (Supplementary Table 2). We could identify clear terms for cellular components (CC) and biological processes (BP) for the significantly dysregulated genes. The multiple GO terms were identified for the downregulated genes but were mostly related to ‘DNA packaging’ and ‘cell membrane’. The most significant downregulated genes (Fig. 7F and 7H) were histones (H1 like protens, H2A1, H2Ba, H2Bb, H2Ax, H3, and H4) and inner membrane complex proteins (IMC16, IMC17, IMC20, IMC22, IMC26, IMC34, IMC35). Transcripts related to protein phosphorylation (Fig. 7F and 7G) were found to be significantly downregulated (Rhoptry kinases-ROPs, aurora kinase, calcium-dependent protein kinase). However, transcripts related to transporter activities (phosphate transporters, MC family transporter) and transferase activity (glucosamine transferases, dehydrogenases, esterases) were found to be significantly upregulated (Fig. 7F and 7H). Further analysis revealed glucose (PYK1, ENO1, GAPDH1, ENR, and ACC1) and glutamate metabolism (glutathione synthase, glutamate cysteine ligase) pathways upregulated upon loss of capping (Fig. 7F and 7H).
Depletion of TgRT protects mice from lethal toxoplasmosis
TgRT protein is essential for parasite fitness in tissue culture; however, to test its essentiality in establishing an infection in the host, we performed mouse infection studies. First, we tested whether TgRT-mAID-HA could be depleted in vivo (Fig. 8A). Mice were infected with 50 tachyzoites of RH TgRT-mAID-HA parasites intraperitoneally and then treated orally with 200 mg/kg/day IAA34 or vehicle from day 2 to 4 post-infection (Fig. 8A). On day 5 and 6 pi, the mice were sacrificed and peritoneal exudate cells (PECs) were collected for IF microscopy. Parasites from IAA-treated mice showed depleted TgRT-mAID-HA levels but normal levels of a control protein, TgIMC1 as compared to parasites from the vehicle-treated mice (Fig. 8B). This experiment confirmed the successful in vivo depletion of TgRT-mAID-HA protein as early as 3 days after IAA treatment. Next, to determine the effect of prolonged depletion of TgRT on parasite survival in vivo, mice were infected with RH TgRT-mAID-HA parasites and then treated with IAA or vehicle orally for 15 days to deplete TgRT (Fig. 8C). The depletion TgRT was confirmed by IF microscopy in one of the sacrificed mice at 5 day of IAA treatment (data not shown). All mice receiving the vehicle control treatment succumbed to lethal toxoplasmosis by day 11 post-infection (Fig. 8D). Conversely, IAA treatment (i.e., TgRT depletion) rescued all mice from lethal toxoplasmosis, indicating that TgRT is necessary for acute infection in T. gondii (Fig. 8D). Severe morbidity and complete mortality were seen in the control treatment group compared to the IAA treatment group (Fig. 8E and 8D). By day 2 pi, the control group showed weight loss, a sign of illness, and continued to lose up to a quarter of their initial weight by the time of death. The IAA treatment group showed no weight loss. Suppressing TgRT expression during the acute phase of infection effectively blocked replication of T. gondii since discontinuation of IAA treatment following day 15 did not result in morbidity (Fig. 8E), mortality (Fig. 8D), or parasite presence (tested by collecting PECs -data not shown) as monitored for an additional 15 days. Overall, these results show the essential role of TgRT for parasite survival and replication in the mouse host.
Structural analysis of T. gondii RNA triphosphatase
The tunnel family RNA triphosphatases are potential antiinfective targets owing to the complete divergence in structures and mechanisms of the RNA triphosphatases of the unicellular pathogen and the mammalian host. We showed that the mechanism of action of TgRT is different from its host counterpart; however, we have yet to know the structural details of this protein, which can help develop a parasite-specific inhibitor. To gain insight into the structure of TgRT, we first attempted to perform comparative modeling of TgRT with known PDB structures of other RNA triphosphatases; however, we failed to obtain the structure owing to poor (< 40%) sequence homology. Next, we retrieved Alphafold predicted structure of TgRT using the AlphaFold database (https://alphafold.ebi.ac.uk/entry/S8F0G9). The obtained structure showed a moderate to high confidence score (the per-residue model confidence score: 40–95) for the region corresponding to the triphosphate tunnel (361-729aa) and a high predicted error (PAE) for the other regions of TgRT (Supplementary Fig. 8A and 8B). Finally, using a ColabFold, the structure of the triphosphate tunnel region (361-729aa) was predicted (Supplementary Fig. 8C). The predicted structure comprised 10 anti-parallel β barrels (green) with 3 α helices (blue) surrounding the active tunnel site (Fig. 9A and 9B). Structural comparison with S. cerevisiae RNA triphosphatase (Cet1) revealed a nearly identical structure (for triphosphatase tunnel) of the two enzymes (Fig. 9C) with 9 of 15 side chain positions important for Cet1 activity conserved in TgRT (Fig. 9D). The triphosphatase tunnel structure of TgRT also showed high structural similarity, including active site residues with the recently deciphered crystal structure of Trypanosoma cruzi RNA triphosphatase. (Fig. 9D and 9E). Next, we checked whether any host protein has similar structure as of TgRT. Using psi BLAST, we found human thiamine triphosphatase (hTTP) is the only TTM with the available structure that resembles TgRT predicted structure (Fig. 9F). Structural comparison with hTTP revealed that the metal binding sites are identical between the two enzymes (Fig. 9D); however, TgRT lacks the C-terminal plug-in helix, which ensures a topologically closed structure and provides substrate specificity only for thiamine triphosphatase (Fig. 9F). Collectively, these results demonstrate the overall similarity in the structure of TgRT with other TTMs. Although the central tunnel structure of TTM is conserved amongst TgRT and hTTP, the substrate recognition and specificity vary between them, and this raises the possibility of development of an inhibitor that specifically targets the entry of RNA substrate into TgRT with minimal to no effect on the hTTP.