Characterization of the dual role of Plasmodium falciparum DNA methyltransferase in regulating transcription and translation

Abstract DNA modifications are critical in fine-tuning the biological processes in model organisms. However, the presence of cytosine methylation (5mC) and the function of the putative DNA methyltransferase, PfDNMT2, in the human malaria pathogen, Plasmodium falciparum, remain controversial. Here, we revisited the 5mC in the parasite genome and the function of PfDNMT2. Low levels of genomic 5mC (0.1–0.2%) during asexual development were identified using a sensitive mass spectrometry procedure. Native PfDNMT2 displayed substantial DNA methylation activities, and disruption or overexpression of PfDNMT2 resulted in reduced or elevated genomic 5mC levels, respectively. PfDNMT2 disruption led to an increased proliferation phenotype, with the parasites having an extended schizont stage and producing a higher number of progenies. Consistent with PfDNMT2’s interaction with an AP2 domain-containing transcription factor, transcriptomic analyses revealed that PfDNMT2 disruption led to a drastic alteration in the expression of many genes, some of which provided the molecular basis of enhanced proliferation after PfDNMT2 disruption. Furthermore, levels of tRNAAsp and its methylation rate at position C38, and the translation of a reporter containing an aspartate repeat were significantly reduced after PfDNMT2 disruption, while the levels of tRNAAsp and its C38 methylation were restored after complementation of PfDNMT2. Our study sheds new light on the dual function of PfDNMT2 during P. falciparum asexual development.


INTRODUCTION
DNA methylation plays a critical role in numerous biological functions via an inheritable epigenetic regulatory mechanism through regulation of transcription, chromatin structure, and chromosome stability by genomic imprinting, Xchromosome inactivation, and suppression of repetitive elements (1)(2)(3)(4). DNA methylation is conserved in most major eukaryotic organisms, including plants, animals, and fungi, with a few exceptions in certain model organisms such as the budding yeast Saccharomyces cerevisiae and the nematode worm Caenorhabditis elegans (5,6). In general, DNA methylation occurs 'globally' in vertebrates, with CG sites being heavily methylated genome-wide except for those in CpG islands, whereas invertebrates, plants and fungi have a 'mosaic' methylation landscape, characterized by sporadically methylated and unmethylated areas (3,5). Methylation at the CG sites is the major type, mostly symmetrical on both DNA strands, whereas methylation in non-CG contexts (CHG and CHH, where H can be any nucleotide but G) mainly occurs in plants and certain human developmental stages such as embryonic stem cells, germ cells and differentiated neuronal cells (3,(7)(8)(9)(10)(11)(12).
DNA methylation at cytosine is catalyzed by DNA methyltransferases (DNMTs), and it is a widely accepted paradigm that methylation is introduced de novo by the DNMT3 DNA methyltransferase. DNMT1 maintains methylation patterns by methylating hemimethylated CG, although accumulated evidence indicates functional overlaps of both DNMTs (4,10). DNMT2 was the third DNMT based on its conserved motifs among DNMTs (13). However, its function is still equivocal, and data from previous studies disputed whether DNMT2 is a primary DNMT or redundant (13). It localizes in the nucleus of Entamoeba histolytica (14), distributes in both the cytoplasm and nucleus of the Drosophila embryo (15), and shuttles from the nucleus to cytoplasm during Dictyostelium discoideum development from S-phase to mitosis (16), suggesting it has a dual function. In agreement with its nuclear and cytoplasmic localization, the recombinant DNMT2 proteins from E. histolytica, Homo sapiens, and Spodoptera frugiperda have low-level cytosine methylation activities on DNA, but relatively highlevel RNA methylation activities on tRNA ASP (14,(17)(18)(19). However, other studies showed that DNMT2s from humans and D. melanogaster had no detectable in vitro DNA methylation activities (19,20). Loss of DNMT2 had only subtle phenotypic effects in Mus musculus, D. melanogaster, S. pombe and D. discoideum (13,21,22). In sharp contrast, DNMT2 is essential for E. histolytica and Danio rerio (23,24). In the genomes of 'DNMT2 only' organisms, such as D. melanogaster, D. discoideum, S. pombe and E. histolytica, which harbor one DNMT2 ortholog as the only known DNMT, DNA is methylated at a low-level without an obvious methylation pattern except that CG methylation is not prevalent in these organisms (13,(25)(26)(27)(28)(29)(30).
The notorious malaria parasite Plasmodium falciparum, responsible for about half a million human deaths annually (31), is another 'DNMT2 only' organism with a single PfDNMT2 (PF3D7 0727300) gene identified (32,33). P. falciparum undergoes a 48-h developmental cycle in human red blood cells (RBCs) through morphologically different stages (ring, trophozoite, and schizont) (34). This intraerythrocytic development cycle (IDC) is tightly controlled by multiple regulatory mechanisms, including the selection and activation of gene-specific transcription factors and epigenetic mechanisms such as histone modifications (34)(35)(36)(37)(38)(39)(40)(41)(42). DNA methylation in P. falciparum remains obscure. Early studies failed to identify methylated cytosine in P. falciparum, whereas partial methylation in the dihydrofolate reductase-thymidylate synthase (dhfr-ts) coding region was detected (43,44). Two attempts to determine the 5-methyl-2'-deoxycytidine (5mC) levels in the P. falciparum genome using the DNA bisulfite sequencing approach led to controversial results, with one study detecting a relatively high level (0.36-1.3%) but the other an extremely low level (0.01-0.05%) (32,45). Recent studies reported that two truncated PfDNMT2 proteins expressed in bacteria showed weak or no DNA methylation activity (32,46), whereas the latter protein also harbored apparent methylation activity on C38 of the tRNA ASP . Conditional knockout of PfD-NMT2 resulted in no noticeable growth phenotype, but methylation of the tRNA ASP C38 was significantly reduced, with a concomitant reduction in the expression of Asp (D)rich proteins (33). These alterations were considered the reasons why PfDNMT knockout (KO) parasites became more sensitive to stress and prone to gametocyte production.
These controversial studies have prompted us to perform a detailed functional analysis of PfDNMT2 in P. falciparum. Here, using an accurate mass spectrometry method, we determined the presence of low levels of 5mC in the parasite genomic DNA during the IDC. We conclusively demonstrated the substantial DNA methylation activity of PfDNMT2 using an in vitro enzyme assay. Interestingly, the PfDNMT2-disrupted parasite had a more rapid growth phenotype partially due to enhanced merozoite production and invasion efficiency, which were further verified by genetic complementation and overexpression of PfDNMT2.
Transcriptomic analysis revealed considerable alterations in gene expression that were consistent with the growth phenotype. In addition, the disruptant line showed substantially reduced methylation of the tRNA ASP C38 and the level of tRNA ASP , correlating with reduced synthesis of a poly-Dcontaining reporter. This study provides solid evidence of the dual role of PfDNMT2 in DNA and tRNA methylation in P. falciparum, which influences gene expression and protein translation.

Parasites culture, leucocyte depletion, and isolation of DNA
The P. falciparum 3D7 strain and transgenic lines were cultured according to a standard method with 0.5% Albu-MAX II added to the buffered medium (RPMI-1640 supplemented with 25 mM HEPES, 50 mg/l hypoxanthine, 40 mg/ml gentamycin sulfate and 25 mM sodium bicarbonate) and O + human RBCs. Cultivation was performed at 37 • C in a gas mixture of 5% CO 2 , 3% O 2 and 92% N 2 (47). To minimize DNA contamination, parasites used for DNA methylation analysis were cultured in leucocyte (white blood cell, WBC)-depleted blood by removing the 'buffy coat' (48) followed by using Plasmodipur filters (EuroProxima) for filtering WBCs (49) and keeping blood at 4 • C for 1 week for the WBC decay (50). The number of WBCs in the blood was measured by flow cytometry after nuclear DNA was stained by Hoechst 33342. Parasite development was synchronized twice with 5% sorbitol treatment, and genomic DNA was purified at the ring, trophozoite and schizont stages according to established methods (51,52).

Mass spectrometry-based quantification of DNA methylation in P. falciparum
Global 5mC levels of the parasite genome at the ring, trophozoite and schizont stages were quantitively analyzed by liquid chromatography coupled with tandem mass spectrometry (LC-MS/MS) (53,54). Briefly, [U-15 N]-labeled nucleoside internal standard ([ 15 N 3 ]-5C, [ 15 N 3 ]-5mC) was added to 1 g DNA/sample, which had been digested completely to nucleosides for 14-16 h following our simplified one-step protocol (54) before the mixture was analyzed by LC-MS/MS. Methyldeoxycytidine and deoxycytidine concentrations were calculated using the analyte/internal standard peak area ratios and standard curves. The percentage of methyldeoxycytidine was calculated by dividing methyldeoxycytidine concentration by total deoxycytidine concentration (methyldeoxycytidine plus deoxycytidine). The same approach was also used to measure the 5hydroxymethyl-cytidine (5hmC) by using [D3]-5hmC as an internal standard.

Tagging, disruption, and episomal expression of PfDNMT2
To tag PfDNMT2 with PTP or GFP, a C-terminus PfD-NMT2 fragment [nucleotides (nt) 948-2118] was amplified using primers Int-F and Int-R (Supplementary Table S1) from P. falciparum genomic DNA and cloned into modified pBluescript SK to fuse with the PTP or GFP and pDT 3' UTR as described earlier (55,56). The PTP tag consists of two protein A-and one C-binding motifs for protein purification using a tandem affinity purification (TAP) procedure (57)(58)(59). This cassette was then subcloned into pHD22Y at the BamHI and NotI sites to produce pHD22Y/DNMT2-PTP/GFP (56). To disrupt PfD-NMT2 by single cross-over recombination, a PfDNMT2 fragment [nt 2-1089] was amplified using primers KO-F and KO-R (Supplementary Table S1) and cloned into pHD22Y at the SpeI and NotI sites to produce the plasmid pHD22Y/ DNMT2. For episomal expression of PfD-NMT2, the entire length of PfDNMT2 was amplified using primers Exp-F and Exp-R (stop codon was omitted) (Supplementary Table S1) and cloned into modified pBluescript SK between HSP86 5' and pDT 3' UTRs to produce pBlue/DNMT. The GFP sequence was cloned at the end of PfDNMT in pBlue/DNMT to form pBlue/DNMT-GFP. The dihydroorotate dehydrogenase (DHODH) drug cassette was released from pUF1-Cas9 (60) and subcloned at SpeI and XbaI into pBlue/DNMT-GFP to produce pBlue/DNMT2-Exp.
Parasite transfection was done by an RBC loading method (61). Briefly, 100 g of plasmid was introduced into fresh RBCs by electroporation. Purified schizonts were used to infect the loaded RBCs and selected with 2.5 nM of WR99210 or 1.5 M of DSM1 for approximately four weeks with weekly replenishment of fresh RBCs until resistant parasites appeared. Resistant parasites were subjected to three cycles of on-off drug selection (62), and single clones of parasites with stable integration of the constructs were obtained by limiting dilution (63). Correct integration of the plasmids into the parasite genome was screened by integration-specific PCR, consisting of a forward primer upstream of the homologous region and a reverse primer located in the PTP or GFP tag (Supplementary Table S1). Correct integration resulting in PfDNMT2 domain deletion was further confirmed by Southern blot with a specific probe located in the homologous region, which was generated by DIG-labeling PCR (Supplementary Table S1).

Expression of PfDNMT2 in P. falciparum
Real-time quantitative PCR was performed to monitor the relative transcription of PfDNMT2 to the endogenous seryl-tRNA synthetase gene (stRNA, PF3D7 0717700) as the control. RNA harvested at an 8 h interval was reverse transcribed using the SuperScript III RNase H reverse transcriptase (Invitrogen). cDNA was PCR amplified with the following primer pairs: PfDNMT2 Fw/Rv and stRNA Fw/Rv (Supplementary Table S1). PCR was done in quadruplicates in 96-well MicroAmp plates in 20 l (Applied Biosystems), 900 nM of each forward and reverse primer and 2 ng of template. PCR (45 cycles of 95 • C for 15 sec and 56 • C for 1 min) was performed in an ABI sequence detector 7300 (Applied Biosystems). The detection threshold was set above the mean baseline value for the first 6-15 cycles. The standard deviation of the quotient was calculated according to User Bulletin 2 (Applied Biosystems, http://www.appliedbiosystems.com). Results were visualized as percentages of PfDNMT2 mRNA relative to the stRNA mRNA.
To assess the PfDNMT2 protein expression during the IDC, synchronized parasite cultures were lysed with 0.06% saponin, and the parasite pellet was washed thrice with PBS. Proteins were extracted by incubating parasite pellets with 2% SDS. For subcellular fractionation, parasites were under a freeze-thaw process three times. After centrifugation at 10 000g for 5 min, the supernatants were obtained as the cytoplasmic fractions, the pellets were suspended in 2% SDS, and the supernatants were used as the nuclear fractions after the same centrifugation procedure. Western blot analysis was performed using transgenic parasites with PTP-tagged PfDNMT2 and the anti-Protein C antibodies (1:1000, Gen-Script). Cytoplasmic and nuclear fractions of rings, trophozoites, and schizonts were equally loaded in SDS-PAGE for detecting PTP-tagged PfDNMT2. To monitor the GFP signal in live PfDNMT2::GFP parasites, images were captured using an epifluorescence microscope (Nikon Eclipse Ni, USA; 100×, 1.4 oil immersion lens) after the nuclei were stained by DAPI at 0.5 g/ml (Invitrogen).

TAP purification and MS
TAP was performed using the PTP-tagged PfDNMT2 parasite line according to the published method (55,57,59). Briefly, 10 9 parasites were lysed in 5 volumes of the hypotonic buffer (10 mM HEPES, pH 7.9, 1.5 mM MgCl 2 , 10 mM KCl, 0.5 mM DTT, 0.5 mM EDTA) at 4 • C for 10 min followed by centrifugation for 20 min at 500 × g. The supernatant was harvested as the cytoplasmic fraction. Five volumes of the PA150 buffer (150 mM KCl, 20 mM Tris-HCl, pH 7.7, 3 mM MgCl 2 , 0.5 mM DTT, and 0.1% Tween 20) containing a protease inhibitor cocktail (Roche) were added to the pellets and mixed thoroughly. The lysate was centrifuged for 10 min at 16 000 × g, and the supernatant was harvested as the nuclear fraction. The supernatants were incubated with 100 l (settled volume) of IgG agarose beads (GE Healthcare) at 4 • C for 2 h. The beads were washed twice with PA150 and equilibrated twice with the TEV buffer (150 mM KCl, 20 mM Tris-HCl, pH 7.7, 3 mM MgCl 2 , 0.5 mM EDTA, 1 mM DTT, and 0.1% Tween 20). The beads were incubated with 2 ml of TEV buffer containing 150 U of TEV protease and rotated overnight at 4 • C to release the PfDNMT2 and its associated proteins from IgG beads. The supernatant was collected, and the beads were rinsed with another 4 ml of the PC-150 buffer (150 mM KCl, 20 mM Tris-HCl, pH 7.7, 3 mM MgCl 2 , 1 mM CaCl 2 , 0.1% Tween 20). These eluted proteins either proceeded directly to protein identification by LC-MS/MS or were subjected to further purification. For the second step of TAP, 7.5 l of 1 M CaCl 2 was added to chelate the EDTA from the TEV buffer, and the combined supernatant was incubated with anti-protein C beads for 2 h at 4 • C. The beads were washed four times with PC150 and eluted with the buffer containing 10 nM EGTA/5 mM EDTA.
The proteins from the first elution step were concentrated by Amicon Ultra centrifugal filters (Millipore Sigma) and separated briefly in a 10% Bis-Tris SDS-PAGE gel for 10 min. Proteins in the gel were excised, and in-gel digestion was done as described (64). The digests were analyzed by nanoLC-MS/MS using a Waters NanoAcquity HPLC system interfaced to a Q Exactive Hybrid Quadrupole-Nucleic Acids Research, 2023, Vol. 51, No. 8 3921 Orbitrap Mass Spectrometer (Thermo Scientific). Peptides were loaded on a trapping column and eluted over a 75 m analytical column at 350 nL/min. MS and MS/MS were performed at 70 000 FWHM and 17 500 FWHM resolutions. The 15 most abundant ions were selected for MS/MS. Parasite proteins were identified by searching the UniProt P. falciparum protein database (v01/2014). Data were filtered at 1% protein and 0.2% peptide false discovery rates (FDR) and at least two unique peptides per protein. Mascot DAT files were parsed into the Scaffold software for validation and filtering to create a non-redundant list per sample. Protein data were analyzed by significance Analysis of INTeractome (SAINT) using a probability threshold above 93% and an FDR below 1% (65). The proteomics data were deposited into the ProteomeXchange Consortium via the PRIDE (66) partner repository with the dataset identifier PXD032860 and 10.6019/PXD032860.

DNA methylation activities of PfDNMT2
The DNMT activity of nuclear extracts and the TAP eluate were measured using the EpiQuik™ DNA Methyltransferase Activity Ultra Kit (Fluorometric) (EpigenTek) following the manufacturer's instructions. Assays were performed in triplicate on 10 g of nuclear protein extracts from ∼2 × 10 7 infected RBCs harvested from ∼5% parasitemia culture (∼4 × 10 8 RBCs from WBC-depleted blood) or ∼200 ng of TAP eluate purified from ∼2.5 × 10 8 Pf-DAMT2::PTP parasite-infected RBCs with 100 ng of purified bacterial C5-methyltransferase as the positive control. Nuclear extract from ∼4 × 10 8 of uninfected RBCs from WBC-depleted blood was used as uninfected RBC control. SG1-1027, a DNMT inhibitor (67), at 10 and 100 M was used to inhibit DNMT activities in the nuclear extracts. A buffer-only blank was used for background subtraction. DNMT activity was expressed in relative fluorescence units per hour and per mg of proteins (RFU/h/mg).

In vitro drug susceptibility assay
The in vitro drug susceptibility assay was performed as previously described (68,69). Briefly, synchronous ring parasites were grown in the presence of different concentrations of SG1-1027 (Cayman chemical company) ranging from 0 to 1500 nM in 96-well plates at 2% hematocrit, 0.5% starting parasitemia and 200 l of culture media. IC 50 was determined after 72 h of parasite growth. 20 l of 10x SYBR green lysis buffer (20 mM Tris-HCl [pH 7.5], 5 mM EDTA, 0.08% Triton X-100, 0.008% saponin in phosphatebuffered saline [PBS; 137 mM NaCl, 2.7 mM KCl, 10 mM Na 2 HPO 4 , 1.8 mM KH 2 PO 4 ], 0.2 l SYBR green I) was added and incubated at room temperature for 30 min in the dark before measuring fluorescence on a TECAN spark multimode plate reader using a 485/535 nm filter. Analysis of the obtained counts plotted against the logarithm of SG1-1027 was performed with GraphPad Prism software 8.0 by fitting a curve with nonlinear regression (sigmoidal dose-response/variable slope equation) to generate IC 50 values.

Growth phenotype analysis
Growth phenotypes of the PfDNMT2 disruptant (ΔPfDNMT2), complementation (episomal expression of PfDNMT2 under HSP86 promoter in ΔPfDNMT2), and overexpression (episomal expression of PfDNMT2 under HSP86 promoter in 3D7) lines during the IDC were compared with the wild-type (WT) as described (55,59). To measure the cell cycle progression, highly synchronous rings obtained by incubating schizonts with RBCs for 3 h were initiated at 2% parasitemia. Parasite development through the IDC was monitored using Giemsa-stained smears every 2 h over 50 h. Cycle time was determined as the duration between the peak ring parasitemias of two consecutive cycles. After two consecutive rounds of synchronization by sorbitol, synchronous cultures were initiated with 0.5% rings, and parasitemia was monitored daily for seven days without replenishing RBCs to measure parasite proliferation. The number of merozoites produced per schizont was determined from mature segmenters. To delineate individual merozoites, highly segmented schizonts were stained by 4',6-diamidino-2-phenylindole (DAPI). The smears were read under a fluorescence microscope. Three independent biological replications were done for each parasite line.
Merozoite invasion assay was performed as described (70,71). To isolate merozoites for RBC invasion, schizonts were pelleted and resuspended in a culture medium supplemented with 10 M of the protease inhibitor transepoxysuccinyl-l-leucylamido (4-guanidino) butane (E64, Sigma) to prevent rupture. The cultures were incubated for 4 h in standard culture conditions with monitoring for segmentation by Giemsa staining. Merozoites were isolated by mechanically passing schizonts (25% hematocrit) twice through a filter unit with a 1.2 m pore size FP 30/1.2 CA-S (Whatman) pre-equilibrated with medium and attached to a 3 ml syringe. The filtrate containing free merozoites was divided to mix with pre-warmed (37 • C) uninfected RBCs, and the number of merozoites contained therein was counted. The same numbers of purified merozoites from the WT, ΔPfDNMT2, complementation, and overexpression lines were mixed with fresh RBCs, and 24 h later, the parasitemia of the culture was determined. The invasion rate was calculated as % RBCs invaded × [(RBCs per l)/ (merozoites per l)].

Parasite transcriptome analysis
RNA-seq was performed to compare the transcriptomes during the IDC among the WT, ΔPfDNMT2, complementation, and overexpression lines by the established methods (59). Three replicates of total RNA from parasites at the ring, trophozoite, and schizont stages were harvested with the ZYMO RNA purification kit and used to generate sequencing libraries using the KAPA Stranded mRNA Seq kit for the Illumina sequencing platform according to the manufacturer's protocol (KAPA biosystems). Pooled libraries were sequenced on the Illumina NextSeq 550 instrument using 150 bp paired-end sequencing and dual indexing. Reads from Illumina sequencing were mapped to the P. falciparum genome sequence (Genedb v3.1) using HISAT2.
The expression levels and differential expression were calculated by FeatureCounts and DESeq2, respectively. The differential expression analysis was performed using a Padjustment value of < 0.01 as the cut-off. RNA-seq data were also normalized by transcripts per million (TPM) to validate the DEseq2 results. Differentially expressed genes between transgenic parasites and WT were selected according to the following criteria: 1) the absolute fold change of TPM higher than 1.5 in biological replicates, and 2) Padjustment from the DEseq lower than 0.1. RNA-Seq data were submitted to NCBI GEO (72) (GSE199368).

Bisulfite sequencing of tRNA ASP and stability of tRNA ASP
Total RNA was isolated from the WT, ΔPfDNMT2, complementation, and overexpression lines in the trophozoite stage. The RNA bisulfite conversion was performed using EpiNext 5-mC RNA Bisulfite-Seq Easy Kit (Epigentek) according to the manufacturer's instructions with minor modifications (73). Briefly, 200 ng RNA was treated with a conversion solution, and samples were incubated in a thermal cycler using the following program: 65 • C for 5 min, 60 • C for 90 min and 4 • C on hold. The bisulfite-converted RNA was column cleaned. For reverse transcription, bisulfiteconverted total RNA was used as a template for cDNA synthesis with a stem-loop primer (55Pf-AspStem-loop) (Supplementary Table S1) (46). The reverse transcription reaction mixtures were heated at 25 • C for 5 min, 50 • C for 60 min in a thermal cycler, followed by RNA digestion at 37 • C for 10 min according to the manufacturer's instructions. cDNA was purified by MQ binding beads. cDNA coding for Aspartic acid tRNA was enriched by PCR using GoTaq qPCR master mix and Bryt Green dye (Promega) with a deaminated specific forward primer (30Pf-Asp-deaminated-F) and a stem-loop specific reverse primer (30Universal stem-loop R) (Supplementary Table S1). The amplicons were cleaned using the NucleoSpin Gel and PCR cleanup kit (Macherey-Nagel) and quantified using the Qubit® 4.0 Fluorometer and Invitrogen™ Qubit® Quantitation Kit (Thermo Fisher Scientific). The PCR products were cloned using pGEM®-T Easy Vector Systems (Promega) and transformed into JM109 Competent Cells (Promega). The positive clones were confirmed by blue-white screening on X-gal LB plates. At least 50 positive clones per parasite line were selected, and the plasmids were prepared and sequenced using the universal M13 forward primer. The sequenced clones were converted to fasta files, and the presence of methyl cytosines was analyzed using BISMA methylation analysis software (74).
To analyze tRNA ASP level in the ΔPfDNMT2, complementation, and overexpression parasite lines compared to the wildtype 3D7 parasite line, total RNA harvested from 30 h trophozoites were reverse transcribed using the Su-perScript III RNase H reverse transcriptase (Invitrogen). cDNA were PCR amplified with the following primer pairs: tRNAasp F/R and stRNA Fw/Rv (Supplementary Table  S1) in triplicates following a 45-cycle program of 95 • C for 15 sec and 42 • C for 30 sec) performed on a QuantStudio 3 platform (Applied Biosystems). Results were visualized as fold change of tRNA ASP mRNA relative to the stRNA mRNA.

Microscopic, flow cytometric, and transcriptional analysis of GFP reporter expression
To determine the translational efficiency of Asp-rich proteins, we designed a GFP reporter, where a 6 × Asp tag was introduced by infusion cloning into the BamHI site at the N-terminal of GFP in the modified pBluescript SK with HSP86 5' and pDT 3' UTRs described earlier. The primers used were D6F and D6R (Supplementary Table  S1). The 6 × D-GFP expression cassette was then subcloned into pCC4 (55,75) at the NotI and SpeI sites to produce pCC4/6D-GFP. The DHODH drug cassette was subcloned from pUF1-Cas9 via AflII and EcoRI into pCC4/6D-GFP to produce pBlue/6D-GFP. Parasite transfection was done in the WT 3D7 and the ΔPfDNMT2 parasites. Drug selection was made with 1.5 M of DSM1 to obtain the transgenic parasites. The synchronized parasites at 36 h trophozoite were stained with 10 M Hoechst 33342 for 10 min at 37 • C and then were observed under an epifluorescence microscope to assess GFP expression (Nikon Eclipse). Totally 100000 cells per sample from the two transgenic lines with 6D-GFP expressed episomally were assayed using the Beckman Coulter MoFlo Astrios flow cytometer. The relative GFP fluorescence intensities in 36 h trophozoites were measured with a Violet laser (395 nm), and bandpass filter of 450/50 nm for detecting Hoechst labeled nucleus, and a blue laser (488 nm) and bandpass filter of 530/30 nm for measuring GFP fluorescence in the cells. The flow cytometry data were analyzed with the FlowJo software. Real-time quantitative PCR with a pair of primers (6 × D-GFP-F and 6 × D-GFP-R) (Supplementary Table S1) was performed to assess the 6 × D-GFP transcripts levels in the WT and disruptant line at 30 h post-invasion (hpi) relative to mRNA of the endogenous control stRNA.

Statistics
Parasite proliferation rates were compared by two-way ANOVA. DNA methylation levels, DNA methyltransferase activities, cycle lengths, merozoite number in mature schizonts, merozoite invasion rates, and mRNA levels of the reporter and tRNA Asp levels were analyzed by paired t-test. Methylation levels at Cytosine 38 of tRNA were compared by paired Wilcoxon test. Spearman correlation was used to compare the transcriptomes among the parasite lines with PfDNMT2 manipulation and the published transcriptome data of WT parasites.

The P. Falciparum genome during IDC contains low-level 5mC
To accurately measure the methyldeoxycytidine in the P. falciparum genome, we cultured the parasites in WBC-free human blood to eliminate contamination from human DNA. Flow-cytometry only detected ∼5-10 WBCs in one million RBCs after WBCs depletion (Supplementary Figure  S1), which only accounts for 0.001-0.002% 5mC of parasite genomic DNA harvested from parasite culture at 5% parasitemia given that the human genome harbors ∼4% 5mC (1). We performed complete digestion of the purified genomic DNA using our improved DNA digestion  (53) to obtain an absolute measure of 5mC. Internal standards allowed the retention times of the analytes to be compared within the same run for accurate identification of the analytes. Using three replicates of genomic DNA prepared from highly synchronous P. falciparum during the IDC, we quantified that 5mC accounted for 0.13%, 0.19% and 0.18% of total cytosines in the ring, trophozoite, and schizont stages, respectively ( Figure 1 and Supplementary Figure S2). The 5mC levels determined in this study are several-fold lower than those (1.16-1.31%) reported by Ponts et al. (32) except for the schizont stage (0.36%). Hammam et al. detected very low levels (0.01-0.02%) of 5mC but relatively high levels (0.2-0.4%) of a 5hmC-like molecule in the P. falciparum genome (45). However, our analysis with an internal standard of [D3]-labeled 5hmC showed that 5hmC was below the limit of detection and there was no trace of a 5hmC-like peak in the DNA samples from parasites at different stages compared to the positive control of the human genome (Supplementary Figure S2).

PfDNMT2 is expressed throughout the IDC and localized in the cytoplasm and nucleus
A time-course study by RT-qPCR revealed that PfDNMT2 was expressed throughout the IDC, with the peak mRNA level detected in the trophozoite stage (Figure 2A), consistent with the expression pattern identified in several transcriptomic studies (36,76,77). To study protein expression and localization, we tagged the endogenous PfDNMT2 in the 3D7 strain with the GFP or PTP tag (Supplementary Figure S3A) and verified the correct integration of the plas-mid into the parasite genome by diagnostic PCR (Supplementary Figure S3B, C). The genetically modified parasites with C-terminal tagging of PfDNMT2 did not show any noticeable change in growth. Western blot analysis of the PTP-tagged parasites with the anti-protein C antibody detected a single protein band at ∼100 kDa, which agreed with the predicted size of the PfDNMT2-PTP fusion protein (Supplementary Figure S3D). The PfDNMT2-PTP protein was detected throughout the IDC, with its relative abundance showing a similar trend as the PfDNMT2 transcript ( Figure 2B). Live imaging of the GFP-tagged PfDNMT2 showed that PfDNMT2-GFP colocalized extensively with the nuclear staining in all the stages of the IDC, while it also had substantial localization in the cytoplasm in the trophozoite stage ( Figure 2C). Consistently, cell fractionation of the PfDNMT2::PTP parasites followed by western blots confirmed this predominantly nuclear localization pattern of PfDNMT2-GFP in rings and schizonts, whereas PfDNMT2-PTP was detected in both the nuclear and cytoplasmic fractions in trophozoites ( Figure 2D).

PfDNMT2 has robust DNA methyltransferase activity
Previous studies showed that DNMT2 orthologs possessed weak or no DNA methyltransferase activities (14,(17)(18)(19)(20). The DNMT activity of the recombinant truncated PfD-NMT2 expressed in bacteria was either weak (32) or undetectable (46) from in vitro assays. We speculated that the poor enzyme activities of the recombinant PfDNMT2 might be due to improper folding, or PfDNMT2 may need to be in its native protein complex for enzyme activity. To address this problem, we purified ∼200 ng of PfDNMT2-PTP from the nuclear extracts of ∼2.5 × 10 8 synchronized trophozoites under native conditions using the TAP procedure (57-59). We measured the DNMT activity using 200 ng of the TAP eluate with a DNA methyltransferase activity kit. Surprisingly, the TAP eluate showed higher DNMT activity (∼4600 RFU/mg) than the positive control (∼2860 RFU/mg), 100 ng of purified bacterial C5methyltransferase M.SssI ( Figure 2E). M.SssI contains all ten conserved motifs of C5-DNMTs and has been used as a positive control because it has the same specificity as the mammalian DNMTs (78,79). When 10 g of the crude nuclear extracts from ∼2 × 10 7 parasites was used in the enzyme assays, the DNMT activity was relatively low at ∼610 RFU/mg, consistent with a previous study (32). Western blotting showed that the abundance of PfDNMT2 in 200 ng TAP eluate was substantially higher than that from 10 g of the crude nuclear extracts, suggesting that a higher abundance of PfDNMT2 in TAP eluate contributed to its higher DNMT activities (Supplementary Figure S3E).

PfDNMT2 associates with distinct proteins in the nucleus and cytoplasm of parasites
We next wanted to determine the proteins associated with PfDNMT2. To capture proteins that may be transiently or weakly associated with PfDNMT2, we performed IPs with only the IgG Sepharose column using lysates from the nuclear and cytoplasmic fractions of trophozoites. Bound proteins were released from the column by TEV proteinase digestion and subjected to protein identification by LC-MS/MS. SAINT identified 16 proteins from the nuclear fraction in all three biological replicates, including an AP2 transcription factor (PF3D7 1239200), the SPT16 subunit of the putative FACT complex, a protein kinase, and two RNA-binding proteins ( Table 1, Supplementary Table S2). In the cytoplasmic fraction, SAINT identified 13 proteins, including the eukaryotic translation initiation factor eIF2A, an RNA binding protein, a Zinc finger protein, two aspartic proteases (plasmepsin I and III), a ubiquitin-like protein, and two exported proteins (Table 1, Supplementary  Table S2). Only three identified proteins were shared in both fractions, indicating that PfDNMT2 associates with distinct protein partners to potentially regulate transcription in the nucleus and protein translation and degradation in the cytoplasm.

PfDNMT2 disruption enhances parasite proliferation
To investigate the function of PfDNMT2, we sought to disrupt this gene by single cross-over recombination (Supplementary Figure S4A, B). PfDNMT2 is dispensable for the IDC (80), and its disruption was confirmed by the Southern blot (Supplementary Figure S4C). Compared to the WT parasite, the ΔPfDNMT2 parasite displayed an increased proliferation rate ( Figure 3A). Starting from 0.5% ring-stage parasitemia, the ΔPfDNMT2 parasite reached 11.3% on day 7 as compared to 7.5% for the WT parasite (P < 0.01). To understand this growth phenotype, we examined the parasite development progression through different IDC stages. We found that the schizont stage of the ΔPfDNMT2 parasite was ∼3 h longer than the WT ( Figure  3B), significantly extending the duration of the IDC to ∼51 h compared to ∼48 h in the WT ( Figure 3C, P < 0.05, paired t-test). We then determined if the increased proliferation rate of the ΔPfDNMT2 parasite resulted from enhanced efficiency of parasite multiplication or merozoite invasion. We found that the ΔPfDNMT2 mature schizonts contained significantly more merozoites than the WT parasite ( Figure  3D, P < 0.0001, paired t-test). Merozoites isolated from the ΔPfDNMT2 parasite also showed a higher RBC invasion rate ( Figure 3E). However, this higher rate did not reach statistical significance (P > 0.05, paired t-test). These data collectively indicate that the higher proliferation rate of the ΔPfDNMT2 parasite was mainly due to the higher number of merozoites produced during the longer schizont development and partially due to higher invasion capacity.
To corroborate that the phenotypic change of the ΔPfDNMT2 was caused by PfDNMT2 gene disruption, we performed genetic complementation and overexpression studies by episomally expressing PfDNMT2-GFP in the ΔPfDNMT2 and 3D7 parasites, respectively (Supplementary Figure S5A). The episomally expressed PfDNMT2-GFP showed a similar localization pattern as observed for the endogenously tagged PfDNMT2-GFP (Supplementary Figure S5B). Episomal expression of PfDNMT2 in the ΔPfDNMT2 parasite fully restored the parasite growth phenotypes, including the proliferation rate, the IDC progression, merozoite production, and merozoite invasion efficiency ( Figure 3A, C-E, Supplementary Figure S5C). Moreover, overexpression of PfDNMT2 in 3D7 further reversed the growth phenotype compared to the WT 3D7, showing a significantly smaller number of merozoites per mature schizont ( Figure 3D, P < 0.0001, paired t-test), Nucleic Acids Research, 2023, Vol. 51, No. 8 3927 as well as a slightly lower proliferation rate, shorter IDC length, and lower merozoite invasion efficiency ( Figure 3A, C, E, Supplementary Figure S5D, P > 0.05, paired ttest). Taken together, these PfDNMT2 genetic manipulation studies demonstrated that PfDNMT2 is involved in parasite development, proliferation, and invasion.

Alteration of DNA methylation upon PfDNMT2 manipulation
To further corroborate the PfDNMT2's DNA methylase activity, we first measured the DNA methylation activity in 10 g of nuclear extract from ∼2 × 10 7 of ΔPfDNMT2 parasites, which exhibited a ∼50% reduction compared to the WT parasite ( Figure 3F, Supplementary Figure S6A). The nuclear extracts from the same numbers of WBC-depleted RBCs for harvesting ∼2 × 10 7 of ΔPfDNMT2 parasites harbored only ∼3% of DNMT activity in nuclear extracts of WT parasites, indicating that the DNMT activity in the nuclear extract of the ΔPfDNMT2 parasites is not from contaminating WBC from the asexual stage blood culture (Supplementary Figure S6A). This finding is in line with the similar levels of reduction in DNA methylase activity observed in the Drosophila embryos and D. discoideum after DNMT2 gene deletion, suggesting that there is probably another DNMT in these 'DNMT2 only' organisms (81,82). The DNMT activity in ΔPfDNMT2 parasites was substantially inhibited by SGI-1027, a DNMT inhibitor (67), corroborating that P. falciparum harbors another DNMT (Supplementary Figure S6B).
We next performed absolute measurement of methyldeoxycytidine by LC-MS using parasite genomic DNA isolated from the ΔPfDNMT2, complementation, and overexpression parasite lines. Compared to the WT, the 5mC levels in the ΔPfDNMT2 parasite decreased by ∼25.6% (0.13% versus 0.09%), 16% (0.19% versus 0.15%) and 26% (0.18% versus 0.13%) in the ring, trophozoite and schizont stages, respectively ( Figure 3G). Notably, the level of 5mC reduction in the schizont stage was similar to the 30% reduction reported earlier in two conditional PfDNMT2 knockout clones (33). The smaller scale of reduction in 5mC (16-26%) compared to the reduction in DNMT activity (50%) in ΔPfDNMT2 parasites could be a compensation effect from an unknown DNMT. Upon PfDNMT2 complementation, DNA methylation levels reached ∼146%, 100% and 103% of the WT values in the ring, trophozoite, and schizont stages, respectively. Compared to the WT, overexpression of PfDNMT2 in 3D7 further increased the methylation levels by ∼115%, ∼221% and ∼51% in the ring, trophozoite, and schizont stages, respectively ( Figure 3G). These results reaffirmed that PfDNMT2 disruption substantially reduced DNA methylation in P. falciparum, which was restored by complementation and increased by overexpression of PfDNMT2. Additionally, in vivo drug susceptibility assays showed that ΔPfDNMT2 parasites became more sensitive to SGI-1027 (IC 50 Figure S6C, D). Collectively, these results further demonstrated that PfDNMT2 is a bona fide DNMT.

PfDNMT2 disruption causes specific transcriptional changes
To assess the effect of altered DNA methylation on gene expression, we compared the transcriptomes of the WT and PfDNMT2 manipulation parasite lines (disruption, complementation, and overexpression) by RNA-seq. RNA was prepared from tightly synchronized parasites at 10, 30, and 40 hpi in three biological replicates except for ΔPfDNMT2 at the schizont stage, which was prepared at 43 hpi to compensate for the delay in development compared to the WT ( Figure 3B Table S3). Gene ontology (GO) enrichment analysis revealed that the processes such as invasion and protein export were downregulated in the trophozoite stage, whereas DNA replication was upregulated at the schizont stage, explaining the ∼3 h extension of the schizont stage and more progeny (merozoite) produced in the ΔPfDNMT2 schizont (Figure 4D, E). Upon complementation of ΔPfDNMT2, transcription was restored to the WT levels with only 1% (55), 0.4% (22) and 0.36% (19) of genes differentially expressed in the ring, trophozoite, and schizont stages, respectively  Table  S3). GO enrichment analysis showed that invasion-related transcripts were downregulated in schizonts of the PfD-NMT2 overexpression line, consistent with its reduced invasion phenotype (Supplementary Figure S9, Figure 3E).

PfDNMT2 disruption reduces tRNA ASP C38 methylation
Previous studies reported that the recombinant truncated PfDNMT2 had in vitro tRNA methylation activity (46), and conditional knockout of PfDNMT2 resulted in a reduction of methylation at C38 of tRNA ASP (33). To confirm that PfDNMT2 is the specific tRNA methyltransferase for C38 of tRNA ASP , we determined the methylation status of tRNA ASP in the WT, ΔPfDNMT2, complementation, and overexpression lines by RNA bisulfite sequencing (73) in trophozoite stage when PfDNMT2 had peak expression and cytoplasmic location. We performed deaminationsensitive PCR to enrich bisulfite-converted tRNA ASP from these parasite lines followed by cloning and sequencing the PCR products (Supplementary Figure S10). Overall, PfD-NMT2 disruption significantly decreased C38 methylation, whereas complementation restored C38 methylation to the WT level ( Figure 5A). PfDNMT2 overexpression maintained a methylation level similar to the WT. The tRNA ASP gene has seven sites in the CpG context, and C38 showed 96.3% methylation in the WT (Supplementary Figure S11). PfDNMT2 disruption resulted in a 72% reduction in C38 methylation. Other sites not known as PfDNMT2 substrates did not show significant alterations in methylation upon PfDNMT2 manipulation (84,85). These results confirmed the direct role of PfDNMT2 in C38 methylation.

Loss of tRNA ASP methylation decreases tRNA ASP and polyaspartate protein expression
tRNA cytosine methylation has been previously highlighted in protein synthesis (13,(84)(85)(86), notably in proteomes of or-ganisms with substantial proportions of Asp residues (46). A recent study showed that disruption of PfDNMT2 resulted in the decrease of Asp-rich proteins in the parasite (33). To confirm that PfDNMT2 affects the translational efficiency of Asp-rich proteins, we examined whether reduced tRNA ASP C38 methylation conferred a defect in the synthesis of Asp-rich proteins using a GFP reporter gene with six aspartate residues inserted in its N-terminus, consistent with the reduction of poly-Asp-tagged reporter proteins after DNMT2 KO in murine embryonic fibroblast cells (84) ( Figure 5B). The GFP reporter construct was episomally expressed in the WT and ΔPfDNMT2 parasites and analyzed by fluorescent microscopy and flow cytometry. Although the GFP transcript levels in the two parasite lines were similar ( Figure 5C), quantitative flow cytometry detected a significant decrease in relative GFP fluorescent in- Total RNA isolated from trophozoites of WT, PfDNMT2, PfDNMT2 complementation and overexpression lines was used for quantitative analyses with seryl-tRNA synthetase gene as a reference. Data from three biological replicates represent the fold change of tRNA ASP relative to the constitutively expressed seryl-tRNA synthetase gene. *P < 0.05 (unpaired two-tailed t-test). tensity in the ΔPfDNMT2 parasite compared to the WT control ( Figure 5D, E). These data provide direct evidence that loss of tRNA ASP methylation reduces the translation efficiency of Asp-rich proteins.
Earlier studies indicated that C38 methylation of tRNA ASP by DNMT2 protected the cleavage of tRNA ASP in Drosophila (87) and double KO of DNMT2 and Nsun2, which generates 5mC at other tRNA positions, caused the reduction of tRNA levels in mice (86). To investigate whether disruption of PfDNMT2 led to any change in tRNA ASP abundance, we measured the level of tRNA ASP in ΔPfDNMT2, PfDNMT2 complementation, and overexpression parasite lines compared to the WT control. As shown in Figure 5F, disruption of PfDNMT2 resulted in a significant reduction in tRNA ASP level whereas complementation and overexpression of PfDNMT2 restored and escalated the tRNA level in the parasite, respectively, indicating that PfDNMT2 regulates tRNA ASP stability probably by protecting it from cleavage via methylation.

DISCUSSION
In P. falciparum, the biological functions of DNMT2, the most widely conserved gene of the DNMT family in eukaryotes, are presently contentious with tRNA and potentially DNA methyltransferase activities. We have comprehensively interrogated the PfDNMT2 enzyme activities on genomic DNA and tRNA throughout the IDC in a panel of transgenic parasites through a series of complementary experiments, including accurate MS, in vitro DNMT activity, subcellular localization, interactome, genetic manipulation, tRNA methylation and levels, and reporter gene translation. We provide evidence that PfDNMT2 possesses both DNA and tRNA methyltransferase activities and plays important roles in the transcriptional regulation of specific bi-ological pathways such as DNA replication and invasion in the nucleus, and protein translation in the cytoplasm.
An earlier study detected relatively high levels (1.16-1.31%) of 5mC in the P. falciparum genome (32), whereas a follow-up study identified extremely low levels of 5mC (0.01-0.02%) in the parasite genome (45), leading to the speculation that the relatively high 5mC contents reported earlier were probably due to the method used without internal control or human DNA contamination during parasite culture. To solve this discrepancy, we developed a protocol to incorporate steps to eliminate potential contamination of human DNA from WBCs in the blood, improve DNA digestion, and precisely identify the target analytes by including stable isotope-labeled internal standards in the same LC-MS/MS runs. Including internal standards in the same MS/MS runs improves the accuracy of the quantitative assay, especially when there are significant concentration differences and variations between the analytes. We found that 0.13-0.19% of the total cytosines in the parasite genome were methylated during the IDC ( Figure 1) and these 5mC levels are similar to the levels in other 'DNMT2 only' organisms (27,82). The study that identified 5mC at extremely low levels also detected relatively high levels of 5mC-like and 5hmC-like peaks at longer retention times compared to 5hmC and 5mC peaks in the parasite genome (45). In contrast, our analyses did not reveal any trace of 5mC-like and 5hmC-like peaks (Figure 1 and Supplementary Figure S2). Previous studies showed that more than one peak at different retention times indicated nucleoside dimers, salt adduct byproducts, or oligomers existed in the incompletely digested DNA (88,89). Therefore, we speculate that the appearance of 5hmC-like and 5mC-like peaks could be derived from the incomplete digestion of genomic DNA or salt adduct byproducts.
Our complementary approaches conclusively attest that PfDNMT2 is the methylase catalyzing the cytosine methylation in the parasite genome. First, the native PfDNMT2 enzyme (complex) purified from the parasite using a TAP procedure displayed high-level DNA methylase activity, rivaling a bacterial DNA methylase included as the positive control. This result is in stark contrast to the insignificant DNA methylase activity of recombinant truncated PfDNMT2 expressed in bacteria, which might have improperly folded, missed essential DNA methylase motifs (32), or lacked the endogenous complex (90). Previous reports showed that DNA methylation activities of DNMT2 from malaria parasites (32,46) and other 'DNMT2 only' organisms (14,(17)(18)(19)(20) were not detectable or at low levels, which led to suspicion that DNMT2 may not have real DNA methylation activity. However, by using TAP, we enriched PfDNMT2 at a higher concentration from parasite nuclei and defined for the first time that PfD-NMT2 has strong DNA methylation activity like the positive control. Second, PfDNMT2 disruption led to a significant decrease in the DNA methylase activity of the nuclear extract and a reduction in the global DNA methylation level in parasite genomic DNA compared to the WT. Third, PfDNMT2 complementation restored the 5mC levels, while PfDNMT2 overexpression further elevated parasite genomic DNA 5mC level up to 221% of that in the WT parasite (Figure 3). The shift in IC 50 values after PfD-NMT2 manipulation further confirms that PfDNMT2 is an authentic DNMT. Finally, the DNMT activity of PfD-NMT2 is also consistent with its predominant nuclear localization in the ring and schizont stages ( Figure 2). These data demonstrated the presence of relatively low levels of 5mC modification in the P. falciparum genome and incriminated PfDNMT2 as the DNA methylase. On the other hand, PfDNMT2 disruption resulted in the reductions of DNA methylation activities and 5mC levels by only 50% and 16-26%, respectively, and SGI-1027 significantly reduced the DNA methylation activity in the ΔPfDNMT2 nuclei, suggesting that there might be an unidentified DNMT in the parasites like other 'DNMT2 only' organisms (81,82).
Does the low-level 5mC in the P. falciparum genome have any functional significance? Our transcriptome analysis detected extensive transcription perturbance from PfDNMT2 disruption, with the most significant effect on the trophozoite stage coincidental with peak PfDNMT2 expression ( Figure 4). While the transcriptomic changes resulting from PfDNMT2 disruption were fully rescued by complementation, PfDNMT2 overexpression led to a more substantial disturbance of the transcriptome. The upregulated genes upon PfDNMT2 disruption are enriched in transcription regulation and DNA replication, which may be responsible for the extended schizont stage and enhanced merozoite production, underlying the increased proliferation of the ΔPfDNMT2. The association of PfDNMT2 with the AP2-TF (PF3D7 1239200) indicates that PfDNMT2 is involved in transcriptional regulation. These data, together with the predominant nuclear localization of PfDNMT2 in ring and schizont stages, demonstrate PfDNMT2's involvement in the transcription regulation of specific cellular pathways. Recent studies showed that DNMT2 has a novel function in RNA-mediated epigenetic heredity in mice (91)(92)(93) and whether PfDNMT2 participates in this mechanism needs more investigation.
We observed a marked dual localization pattern of PfDNMT2 in the nucleus and cytoplasm, particularly in trophozoites, when protein synthesis is most active (94,95). Although we could not speculate on the cytoplasmic functions of PfDNMT2 based only on its interactome, its action on epitranscriptomic tRNA methylation will probably influence protein translation. Similar to the observation from a recent study (33), PfDNMT2 disruption led to a ∼72% reduction in tRNA ASP methylation at position C38, which was restored to the WT levels by genetic complementation. Using a reporter construct incorporating poly-D, we provided a biological implication that the deficit in tRNA ASP methylation is directly linked to the reduced translation of the poly-D-containing reporter protein. C38 methylation of tRNA Asp was fully restored after PfDNMT2 complementation, which further confirmed that PfDNMT2 was responsible for this methylation. tRNA Asp level in ΔPfDNMT2 was significantly reduced whereas complementation and overexpression of PfDNMT2 restored and escalated the tRNA Asp levels, respectively, suggesting that PfDNMT2 regulates tRNA ASP stability probably by protecting it from cleavage via methylation. Further characterization of RNA binding proteins identified from PfDNMT2 interactome opens a window to study the underlying mechanisms.
In summary, we provide solid evidence of the presence of low-level 5mC in the P. falciparum genome during its asexual development in the RBCs. Through a series of experiments, we conclusively showed that PfDNMT2 was responsible for depositing methyl groups on genomic DNA and C38 of tRNA ASP , which have functional significance in transcriptional and translational regulation, respectively. The regulation of general biological processes such as DNA replication and parasite-specific biological pathways such as invasion indicates that P. falciparum, a DNMT2-only organism, has evolved to designate DNMT2 novel regulatory functions.

DATA AVAILABILITY
The proteomics data were deposited into the Pro-teomeXchange Consortium via the PRIDE partner repository with the dataset identifier PXD032860 and 10.6019/PXD032860. RNA-Seq data were submitted to NCBI GEO (GSE199368).