Spermidine is a key polyamine required by intracellular parasites for survival within host erythrocytes

Abstract

Intracellular parasites, including Babesia and Plasmodium, the agents of human babesiosis and malaria, depend on the salvage or de novo synthesis of critical nutrients for survival within human erythrocytes. Among these, polyamines play a pivotal role, but their specific requirements and molecular functions in intraerythrocytic parasites remain poorly understood. We identify spermidine as a key polyamine for Babesia duncani and Plasmodium falciparum for intraerythrocytic development. We demonstrate that spermidine is indispensable for regulating protein translation through hypusination of the eukaryotic translation initiation factor eIF5A, and its depletion leads to increased production of reactive oxygen species. Disruption of spermidine biosynthesis or its conversion from spermine results in parasite death. We also show that B. duncani and other Babesia species use an ancestral spermidine synthase–like enzyme, highlighting a distinct evolutionary adaptation from P. falciparum. Our results reveal the spermidine’s dual role in oxidative stress defense and translation regulation, positioning spermidine biosynthesis as a critical vulnerability and a promising therapeutic target.

Spermidine is crucial for Babesia and Plasmodium survival, making its biosynthesis a promising target for antiparasitic therapy.

INTRODUCTION

Apicomplexan protozoan parasites, including Plasmodium species and Babesia sp., comprise a broad group of obligate intracellular pathogens that cause various diseases in humans and animals with major economic and global health impact worldwide. The majority of human babesiosis clinical cases worldwide have been linked to Babesia microti, which is transmitted by Ixodes ticks and carries the smallest genome among Apicomplexa (1, 2). Other babesiosis cases in Europe and United States have been linked to several Babesia species including Babesia divergens, Babesia MO1, Babesia venatorum, Babesia duncani, and Babesia odocoilei (3–5). A crucial but poorly understood factor in understanding Babesia pathogenesis and virulence is how these parasites survive within host erythrocytes. The recent development of the B. duncani in culture–in mouse propagation model has enabled investigation of the molecular mechanisms underlying survival of Babesia parasites within erythrocytes and their interaction with the host (6). The intraerythrocytic life cycle of B. duncani initiates following invasion of a human erythrocyte by a free merozoite. Upon invasion, the parasite develops into metabolically active forms (ring and filamentous stages) and multiplies to produce four daughter merozoites within the infected cell (6). Following egress from the host cell, the released merozoites initiate a new intraerythrocytic cycle leading to a rapid increase in parasite burden. As in malaria, the repeated rounds of parasite development and multiplication and destruction of host erythrocytes are responsible for all the clinical symptoms associated with human babesiosis.

Survival of Apicomplexan parasites in host erythrocytes is absolutely dependent on the uptake and/or de novo synthesis of several essential nutrients (7–9). Accordingly, gene conservation or loss and, in some instances, gene redundancy have been crucial in determining the reliance of these parasites on either the uptake of exogenous nutrients from host plasma or their synthesis de novo from available precursors. Among these nutrients are the polyamines putrescine (PUT), spermidine (SPD), and spermine (SPM), which have been shown to play critical roles in the viability of intraerythrocytic parasites by regulating cell growth and differentiation (10). While Plasmodium species rely primarily on de novo synthesis of polyamines from l-arginine (11, 12) but can also salvage these nutrients from their environment (13), the possible synthesis or salvage pathways used by Babesia parasites remain completely unknown. Furthermore, the specific polyamines required for Babesia and Plasmodium survival and the cellular processes controlled by such polyamines have also not been elucidated.

Here, we provide evidence that B. duncani lacks a de novo biosynthesis pathway for the synthesis of PUT and instead relies on the uptake of exogenous PUT from the host for survival within human erythrocytes and operation of critical cellular activities, chief among them is protein translation and protection against reactive oxygen species (ROS). Unlike other eukaryotes, B. duncani uses an unusual SPD synthase (SPDS), a putative aminopropyl transferase [B. duncani putative aminopropyl transferase (BdPAPT)], related to an ancestral eukaryotic methyltransferase from which SPDSs have evolved. Furthermore, we demonstrate that SPD synthesized from salvaged PUT or SPM is the primary polyamine required for parasite proliferation and survival in human erythrocytes. Consistent with the modulation of B. duncani protein translation by polyamines, pharmacological studies showed that inhibition of hypusination of the parasite translation factor eIF5A (eukaryotic translation initiation factor 5A) from SPD results in parasite death. We demonstrate that similar to B. duncani, SPD is the key polyamine used by Plasmodium falciparum parasites, and its essential function is through its regulation of eIF5A hypusination.

This study reveals distinct strategies for polyamine acquisition and utilization in the intracellular parasites B. duncani and P. falciparum. B. duncani lacks a de novo biosynthesis pathway for PUT and instead relies exclusively on host-derived PUT for survival within human erythrocytes. In contrast, P. falciparum synthesizes PUT de novo from ornithine (ORN), highlighting divergent evolutionary adaptations in nutrient acquisition. Despite these differences, both parasites rely on SPD as the critical polyamine for proliferation and survival. SPD regulates protein translation via hypusination of eIF5A and provides protection against ROS. Pharmacological inhibition of eIF5A hypusination leads to parasite death in both species, underscoring the essential role of SPD in parasite biology. Notably, B. duncani uses an unusual SPDS, BdPAPT, related to an ancestral eukaryotic methyltransferase, setting it apart from P. falciparum. These findings highlight both shared and species-specific dependencies on polyamine metabolism and suggest that targeting SPD synthesis and utilization could provide a unified therapeutic approach against these parasites.

RESULTS

PUT is salvaged by B. duncani and is essential for its survival and development within host erythrocytes

Recent studies have shown successful continuous propagation of B. duncani in human erythrocytes in Dulbecco’s modified Eagle’s medium (DMEM)/ Nutrient Mixture F12 (DMEM/F12) but not the base DMEM culture medium (Fig. 1A) (14). Parasites maintained in DMEM/F12 over multiple 3-day propagation cycles showed increased parasite load overtime, with parasitemia doubling every 22 to 24 hours, whereas those transferred from DMEM/F12 to the DMEM base medium experienced a steady decline in parasite survival over time with no viable parasites detected by the fifth cycle of growth in this medium (Fig. 1A). The morphology of parasites in DMEM medium showed accumulation of highly stressed and dead parasites displaying condensed chromatin as well as abnormal cellular morphology (Fig. 1B). Examination of the nutritional composition of the two media identified 15 components that are present in DMEM/F12 but not in DMEM media (table S1). These include six amino acids, two vitamins, four salts, two lipids, and one polyamine. To investigate which of the components is vital for the growth of the parasite, B. duncani cultures were grown in DMEM medium either alone or supplemented with the missing amino acids, vitamins, inorganic salts, lipids [linoleic acid (LNA) + lipoic acid (LA)] or PUT or various combinations of these nutrients. Parasite growth was monitored for 15 days with parasitemia diluted to 1% every third day. Parasites cultured in DMEM supplemented with either vitamins or inorganic salts failed to support parasite growth, and the parasitemia levels were similar to those observed in the cultures grown in DMEM alone after 5 cycles of propagation (Fig. 1C). Supplementation of DMEM with missing amino acids was also insufficient to support the optimal growth of B. duncani, although the parasitemia levels at the end of cycle 5 were slightly higher in comparison to the parasitemia levels in DMEM (Fig. 1C). The parasites grown in DMEM supplemented with either PUT (+PUT) or lipids (LNA + LA) showed significantly (P < 0.0001) higher parasitemia levels in comparison to the DMEM control (Fig. 1C). At the end of cycle 5, the parasites grown in DMEM supplemented with either the combination of PUT, lipids (LNA + LA), and six amino acids or the combination of PUT and lipids (LNA + LA) showed parasitemia levels similar to those observed in DMEM/F12, indicating that the combination of PUT and lipids (LNA and LA) is sufficient to support optimal in vitro growth of B. duncani (Fig. 1C). Similar analyses were conducted using drop-out media, which also identified PUT, LNA, and LA as key factors for parasite survival (Fig. 1D). While the roles of lipids like LA and LNA in parasite development have been extensively characterized (15–19), the specific contributions of PUT and other polyamines remain largely unexplored. Recognizing the critical importance of PUT uptake for B. duncani’s intraerythrocytic development, we sought to determine whether PUT itself, or one or both of its downstream derivatives, plays a key role in supporting parasite survival and proliferation. We examined whether the growth defect caused by PUT depletion can be restored by addition of downstream polyamines, including SPD and SPM. The parasites propagated in media lacking PUT but supplemented with either SPD or SPM were viable and replicated at similar rates as those grown in the presence of PUT (Fig. 1E). In contrast, the growth defect of B. duncani in the absence of exogenous PUT could not be rescued by the addition of ORN to the culture medium (Fig. 1E). Furthermore, treatment of B. duncani with DL-ɑ-difluoromethylornithine (DFMO), an ORN decarboxylase (ODC) inhibitor, did not result in inhibition of parasite growth (Fig. 1F). Together, these data indicate that B. duncani lacks a de novo pathway for synthesis of PUT from ORN. In some organisms, an alternate de novo pathway of PUT biosynthesis from arginine has also been reported (20). To assess whether such a pathway is used by B. duncani, we examined the effect of DL-ɑ-difluoromethylarginine (DFMA), an inhibitor of the enzyme arginine decarboxylase (ADC), which catalyzes the first step of this pathway, on parasite growth. DFMA treatment failed to inhibit B. duncani growth, indicating the absence of this pathway in these parasites (Fig. 1G). Moreover, a search for ODC and ADC homologs in the recently annotated B. duncani genome failed to identify these enzymes in this parasite (21). Together, these data demonstrate that B. duncani does not synthesize PUT de novo and therefore relies solely on the salvage of PUT from its milieu for survival within human erythrocytes.

Fig. 1. Essential nutrients for B. duncani in vitro growth in human RBCs.

(A) Comparison of B. duncani growth in DMEM/F12 and DMEM media over five 3-day intraerythrocytic cycles (IEC). Parasitemia was monitored at the end of each cycle, with 3000 to 3500 red blood cells (RBCs) counted. The parasite growth in DMEM/F12 was used for normalization. Data presented as means ± SD from two independent experiments with biological triplicates. Statistical significance was assessed by multiple t tests; *P < 0.05. (B) Representative Giemsa-stained smear images of B. duncani–infected RBCs from IEC 1 and 3, highlighting differences between healthy and stressed rings and tetrads in DMEM versus DMEM/F12. (C) Comparison of B. duncani growth in DMEM/F12, DMEM, and DMEM supplemented with missing components [amino acids (aa), vitamins (Vit), salts, lipids, and PUT (from table S1)] at cycle 5. (D) B. duncani growth in DMEM/F12, DMEM, DMEM + all, and DMEM + all with individual component depletions over 5 cycles. Parasitemia was monitored at cycle 5, with data normalized to DMEM/F12. Statistical significance was determined using Welch’s t test; ****P < 0.0001; ns, no significant difference. (E) Restoration of growth in PUT-depleted cultures supplemented with SPD or SPM, but not ORN, over 6 days. Statistical significance between conditions was assessed by Welch’s t test; ****P <0.0001; ns, no significant difference. (F) In vitro efficacy of DL-α-difluoromethylornithine (DFMO) on B. duncani growth. (G) In vitro efficacy of DFMA on B. duncani growth. Data in (F) and (G) are shown from three independent experiments in biological triplicates. Statistical significance was evaluated by Welch’s t test; ns, no significant difference. Ala, alanine; Asn, asparagine; Asp, asparatic acid; Pro, proline; Cys, cysteine; Glu, glutamine; LNA, linoleic acid; LA, lipoic acid.

SPD is the key polyamine essential for optimal B. duncani growth

Our finding that PUT uptake is essential for B. duncani intraerythrocytic development and survival led us to investigate whether PUT itself or one or both of its derivatives, SPD and SPM, are the key polyamines required for parasite viability. To assess this, we evaluated the effect of inhibition of the parasite SPDS by 4-methylcyclohexylamine (4-MCHA) and SPM oxidase by MDL 72527 on the growth of the B. duncani in DMEM/F12 medium alone or supplemented with excess PUT (DMEM/F12 + PUT) or SPD (DMEM/F12 + SPD) or SPM (DMEM/F12 + SPM) (92 μM, each) (Fig. 2A). While the inhibitory activity of 4-MCHA was not altered by PUT supplementation (Fig. 2B), addition of SPD or SPM to the culture medium led to a statistically significant reduction in the in vitro efficacy of the drug, resulting in ~10× and ~3.5× increases in the inhibitor’s median inhibitory concentration (IC₅₀) in the presence of SPD and SPM, respectively (Fig. 2B). Consistent with MDL 72527’s specific inhibition of SPM synthase activity, its inhibitory activity was similar in media lacking or supplemented with SPM (Fig. 2C). Supplementation of the culture medium with either SPD or PUT significantly reduced MDL 72537 activity, with ~3.5× and 1.4× increases in the compound’s IC₅₀ in the presence of SPD and PUT, respectively (Fig. 2C). While SPD supplementation significantly reversed the effect of MDL 72527, PUT supplementation only achieved a modest 1.4-fold reversal. This suggests that SPD is directly incorporated into essential downstream pathways critical for parasite growth and development within the host. These data demonstrate that SPD is the key polyamine required for parasite growth in vitro. Consistent with this observation, the inhibition of parasite’s growth following treatment with a combination of 4-MCHA and MDL 72527 was significantly mitigated in media supplemented with SPD, whereas supplementation with PUT or SPM failed to restore parasite growth (Fig. 2D). Together, these data indicate that SPD is the primary polyamine that is essential for successful propagation of the parasites within human erythrocytes.

Fig. 2. SPD is the key polyamine required for intraerythrocytic growth of B. duncani in human RBCs.

(A) Schematic representation of the polyamine biosynthesis pathway in B. duncani. (B) In vitro efficacy of the SPDS inhibitor trans-4-MCHA against B. duncani in DMEM/F12 and DMEM/F12 supplemented with 92 μM of either PUT or SPD or SPM. (C) In vitro efficacy of the SPM oxidase inhibitor, MDL 72527, against B. duncani in DMEM/F12 and DMEM/F12 supplemented with 92 μM of either PUT, SPD, or SPM. (D) SPD rescues the growth defect caused by the inhibition of SPDS and SPM/polyamine oxidase. In vitro efficacy of combinations of 4-MCHA [790 μM (IC₉₀)] and MDL 72527 [84 μM (IC₉₀)] against B. duncani in DMEM/F12 and DMEM/F12 supplemented with 92 μM of either PUT, SPD, or SPM. ***P < 0.001; ns, no significant difference; SPS, SPM synthase; SMO, SPM oxidase; RBCM, red blood cell membrane; PPM, parasite plasma membrane.

Evidence of active SPDS driving PUT-to-SPD conversion in B. duncani

Our findings, which demonstrate that SPD produced from salvaged PUT is essential for the survival of B. duncani, suggest the presence of an active SPDS within the parasite. To test this hypothesis, we assessed whether B. duncani cell-free extracts could catalyze the conversion of PUT into SPD. Liquid chromatography–mass spectrometry (LC-MS) analysis revealed efficient conversion of PUT into SPD by B. duncani cell-free extracts (Fig. 3A), whereas no conversion occurred when PUT was incubated with heat-inactivated cell-free extracts (fig. S1A). Consistent with our pharmacological data and bioinformatic analyses, neither intact nor heat-inactivated B. duncani cell-free extracts catalyzed the conversion of ORN into PUT (Fig. 3B and fig. S1B). Further supporting these results, thin-layer chromatography (TLC) demonstrated that B. duncani cell-free extracts catalyze the conversion of ¹⁴C-labeled PUT into SPD (Fig. 3C). Control experiments confirmed similar activity using cell-free extracts from P. falciparum and purified Saccharomyces cerevisiae Spe3 (Fig. 3, D and E). To exclude the potential involvement of host enzymes in this conversion, we tested hemolysate extracts from both B. duncani and P. falciparum. These extracts showed no detectable aminopropyl transferase activity and failed to convert ¹⁴C-labeled PUT into SPD (Fig. 3, C and D).

Fig. 3. B. duncani has aminopropyl transferase activity.

(A) Aminopropyl transferase activity was assessed by incubating the cell-free extract with PUT, followed by LC-MS analysis to detect the levels of PUT, and the downstream polyamines SPD, and SPM. (B) B. duncani lacks ODC activity, as tested by incubating cell-free extract with ORN, followed by LC-MS to detect ORN and PUT. (C and D) TLC of APT reactions in B. duncani and P. falciparum hemolysate and cell-free extracts using ¹⁴C-labeled PUT as a substrate. (E) TLC showing the APT reaction catalyzed by S. cerevisiae SPDS (ScSpe3) using ¹⁴C-labeled PUT. ¹⁴C-PUT, ¹⁴C labeled PUT; ¹⁴C-SPD, ¹⁴C labeled SPD; Ori, origin; * represents extra bands present in the ¹⁴C-PUT purchased from American Radiolabeled Chemicals Inc. (F) Recombinant BdPAPT-Cterm catalyzed the conversion of PUT into SPD when incubated with PUT and dcSAM, followed by LC-MS to detect PUT, SPD, and SPM. (G) The BdPAPT-Cterm^D1337A mutant (Asp¹³³⁷ to Ala¹³³⁷ substitution) failed to convert PUT into SPD. All data (A to D) are presented as means ± SD of three independent experiments performed in biological triplicates. (H) TLC showing reactions with active, heat-inactivated BdPAPT-Cterm and BdPAPT-Cterm^D1337A, as well as with active and heat-inactivated recombinant CD BdPAPT-CD (amino acids 1020 to 1454) at different time intervals. (I) BdPAPT-Cterm–mediated SPDS assays were performed with active and heat-inactivated enzymes, using 100 μM each of PUT and dcSAM at 37°C for 0 to 60 min. Control reactions were done at 4°C. SPDS activity was measured as micromoles of SPD formed per milligram enzyme. (J and K) The rate of SPD formation was measured as a function of dcSAM and PUT concentrations. The assays used 700 ng of enzyme and varying concentrations of substrates at 37°C for 60 min. V_max and K_m were determined using allosteric sigmoidal kinetics in GraphPad Prism. Data are presented as means ± SD from three independent experiments with biological triplicates.

To identify the SPDS enzyme required for SPD biosynthesis in B. duncani, we conducted a bioinformatic search of the recently published B. duncani genome using prokaryotic and eukaryotic SPDSs as reference sequences. This search identified a candidate gene, BdPAPT (accession number: BdWA1_000190), which encodes a protein containing an aminopropyl transferase catalytic domain (CD). This domain shares 21% similarity and 11% identity with Escherichia coli SPDS and 24% similarity and 13% identity with P. falciparum SPDS (fig. S2). Careful analysis of the annotated gene revealed that the previously annotated BdWA1_000191 and BdWA1_000190 genes are part of a single, fused gene encoding a single protein of ~160 kDa (1454 amino acids) with similar enzymes found in other Babesia and Theileria species (figs. S3 and S4A). Phylogenetic analysis of proteins containing the IPR029063 InterPro domain, including SPDS orthologs, confirmed that BdPAPT clusters with aminopropyl transferases from yeast, Theileria orientalis, Babesia bovis, B. divergens, and P. falciparum (fig. S4B). These findings support the classification of this protein as an SPDS ortholog.

To evaluate the enzymatic activity of BdPAPT, we expressed its C-terminal domain (BdPAPT-Cterm) as a recombinant protein. In addition, we generated a mutant version (BdPAPT-Cterm^D1337A) by substituting a conserved aspartate residue (D1337) in the predicted catalytic site with alanine (fig. S5, B and C). Recombinant proteins were purified, and their ability to catalyze the conversion of PUT to SPD was assessed using LC-MS, TLC, and a fluorescent-based assay [1,2-Diacetyl benzene-aminopropyl transferase (DAB-APT) assay] (Fig. 3, F to K, and fig. S5) (22). The BdPAPT-Cterm enzyme efficiently catalyzed the conversion of PUT to SPD in the presence of decarboxylated S-adenosylmethionine (dcSAM), as demonstrated by LC-MS (Fig. 3F). In contrast, no activity was detected in reactions containing heat-inactivated BdPAPT-Cterm (fig. S1C) or the catalytic mutant BdPAPT-Cterm^D1337A (Fig. 3G and fig. S1D). No conversion of PUT to SPM was observed, indicating that, unlike the P. falciparum counterpart, BdPAPT lacks SPM synthase activity (Fig. 3F). Consistent with these results, TLC analysis confirmed the efficient conversion of PUT to SPD by BdPAPT-Cterm and its CD (amino acids 1020 to 1454; BdPAPT-CD) (Fig. 3H). Inactive forms of these enzymes failed to catalyze the reaction (Fig. 3H). These results firmly establish BdPAPT as the SPDS in B. duncani. The enzymatic activity of BdPAPT-Cterm was further assessed using the DAB-APT (fig. S5) (22). A time-dependent increase in fluorescence, corresponding to the formation of SPD, was observed when BdPAPT-Cterm was incubated at 37°C. In contrast, no significant fluorescence increase was detected when the reaction was conducted at 4°C or when the enzyme was heat-inactivated (fig. S5D). Enzymatic activity was also shown to increase proportionally with higher concentrations of BdPAPT-Cterm (fig. S5, E and F). Using the DAB-APT assay, the kinetic parameters of BdPAPT-Cterm were determined. The Michaelis constant (K_m) and maximal velocity (V_max) for dcSAM were 5.976 nmol/μg per min and 33.46 μM, respectively (Fig. 3J). Similarly, the K_m and V_max for PUT were calculated as 5.38 nmol/μg per min and 32.11 μM, respectively (Fig. 3K).

PUT depletion correlates with impaired oxidative damage defense and protein translation in B. duncani

Given the well-established protective role of polyamines against ROS, either through direct scavenging of free radicals, and thereby acting as antioxidants, or indirectly by binding to and stabilizing DNA, RNA, and proteins against oxidative damage (23, 24), we investigated whether parasites grown in the absence of PUT were under oxidative stress. B. duncani–infected red blood cells (RBCs) grown in the absence of exogenous PUT were stained with dihydrorhodamine 123 (DHR123) and Hoechst 33342 (a nuclear stain), and the resulting fluorescence was compared to that of parasites cultured in the presence of PUT, SPD, or SPM. DHR123 is a nonfluorescent molecule that is cleaved inside cells into its fluorescent by-product, rhodamine, in the presence of ROS (25). Rhodamine-positive B. duncani–infected RBCs in cultures maintained in media lacking or supplemented with polyamines were examined by fluorescence microscopy (Fig. 4A) and quantified using flow cytometry (Fig. 4B). Parasites cultured in the absence of PUT exhibited a significant increase in rhodamine-positive RBCs, with 22×, 50×, and 75× more rhodamine-positive B. duncani–infected RBCs compared to those maintained in the presence of PUT, SPD, or SPM, respectively (Fig. 4B). As a control, hydrogen peroxide (H₂O₂) or the antiparasitic drug artemisinin, known to induce ROS production in apicomplexan parasites (26, 27), also showed increased rhodamine-positive cells, even when PUT was present (Fig. 4B). Fluorescence microscopy corroborated these findings. Parasites maintained in media supplemented with PUT, SPD, or SPM were predominantly rhodamine negative, indicating low ROS levels. In contrast, parasites grown in PUT-depleted media or treated with H₂O₂ or artemisinin were mostly rhodamine positive, reflecting elevated oxidative stress (Fig. 4A). These results demonstrate that polyamines plays a critical role in protecting B. duncani from oxidative damage.

Fig. 4. Insights into oxidative stress levels in B. duncani cultured in PUT-depleted medium.

(A) ROS levels were higher in B. duncani grown in PUT-depleted medium. Representative live cell fluorescence images show rhodamine signal (red) after incubation with DHR123 in parasites cultured in different media, including PUT-depleted medium (−PUT), −PUT medium supplemented with 9.2 μM PUT (+PUT), 9.2 μM SPD (−PUT + SPD), 9.2 μM SPM (−PUT + SPM), or treated with 100 μM H₂O₂ (+PUT + H₂O₂) or 1 mM artemisinin [+PUT + artemisinin (ART)]. Hoechst staining (blue) highlights parasite nuclei. Scale bars, 2 μm. Data are presented as means ± SD from three independent experiments in biological triplicates. (B) Percent of rhodamine-positive parasites was compared across different media. Data were normalized to H₂O₂-treated control. Statistical significance was assessed using Welch’s t test. *P < 0.05, ****P < 0.0001; ns, no significant difference. Bd-iRBCs, B. duncani–infected RBCs.

We further investigated the impact of PUT depletion on global gene expression in B. duncani. The parasites were cultured in media lacking or supplemented with PUT for 24 hours, and their transcription profiles were analyzed by Illumina RNA sequencing (RNA-seq). The normalized counts and rlog-transformed data were generated using DESeq2 analysis (28, 29), and a heatmap was constructed (fig. S6, A and B). Of the 4222 genes encoded by the B. duncani genome (21), 66 were found to be differentially expressed (twofold difference) between parasites cultured in the absence or presence of PUT (fig. S6 and data S1). Of these, 37 genes were found to be down-regulated, and 29 were genes up-regulated in parasites propagated in the absence of PUT compared to those maintained in the presence of PUT (fig. S6B). Gene ontology analysis of these differentially expressed genes identified unique biological processes that were affected by PUT supplementation or depletion. Approximately half of the down-regulated genes in parasites cultured in the absence of PUT were found to encode hypothetical proteins (fig. S6C). Of the down-regulated genes with predicted functions, 17% were found to play a role in protein translation, 6% in protein phosphorylation, and 3% in other functions including adenosine triphosphate (ATP) binding, DNA binding, phosphoinositide-mediated signaling pathway, regulation of mRNA stability, nuclear-transcribed mRNA catabolism, DNA regulation, hydrolase activity, and inositol phosphate biosynthesis (fig. S6C). Of the genes up-regulated in media lacking PUT, 32% encode hypothetical proteins, 32% are involved in nuclear-transcribed mRNA catabolism, and ~6% involved in each of the following processes: protein folding, intracellular protein transport, carbohydrate derivate metabolism, mRNA splicing via spliceosomes, negative regulation of translation, and cellular glucose metabolism (fig. S6D).

PUT depletion and inhibition of SPD synthesis disrupt eIF5A hypusination and survival of B. duncani and P. falciparum

The finding that PUT depletion significantly affects protein translation led us to investigate whether the primary role of SPD, synthesized from PUT, is to regulate protein translation. In other organisms, SPD serves as a donor of the aminobutyl group to lysine 51 of eIF5A to form hypusine (N^ε-[4-amino-2-hydroxybutyl]-lysine) (30), a posttranslational modification critical for protein translation. This posttranslational modification is a two-step enzymatic process catalyzed by deoxyhypusine synthase (DHS) and deoxyhypusine hydroxylase (DOHH) (Fig. 5A). To examine the role of B. duncani DHS enzyme (BdWA1_002959-T1) in hypusination, we used the SPD analog N1-guanyl-1,7-diaminoheptane (GC7) (fig. S7A) and 6-bromo-N-(1H-indol-4yl)-1-benzothiophene-2-carboxamide (fig. S7B), both known to inhibit DHS enzyme and prevent eIF5A hypusination in other systems (31–33). While GC7 inhibited the growth of B. duncani in vitro (IC₅₀, ~100 μM) (fig. S7A), no inhibition of the parasite growth was observed with 6-bromo-N-(1H-indol-4yl)-1-benzothiophene-2-carboxamide (IC₅₀, >2.5 mM) (fig. S7B). Supplementing growth media with PUT, SPD, or SPM reduced GC7 efficacy by ~35, ~65, and ~35%, respectively (Fig. 5B). To confirm the effect of GC7 on eIF5A hypusination, we performed immunoblot analyses using anti-eIF5A and anti-hypusine antibodies on B. duncani lysates collected over 48 hours of treatment. While total eIF5A (BdeIF5A) levels remained unchanged (Fig. 5, C and D), hypusinated eIF5A (BdeIF5A^Hyp) levels decreased significantly, with ~60 and ~80% reductions observed at 36 and 48 hours, respectively (Fig. 5, C and E). These effects were specific to GC7, as treatment with pyrimethamine, an unrelated inhibitor of the folate pathway, had no impact on BdeIF5A hypusination (fig. S8).

Fig. 5. PUT depletion results in altered eIF5A hypusination (BdeIF5A^Hyp) in B. duncani.

(A) Schematic representation of the eIF5A hypusination pathway in B. duncani. (B) In vitro efficacy of GC7 (100 μM) against B. duncani in DMEM/F12 and DMEM/F12 supplemented with 9.2 μM of either PUT, SPD, or SPM. (C) Western blot analysis using anti-eIF5A and anti-hypusine antibodies using lysates from parasites grown in the absence or presence of GC7 (100 μM) for 0, 12, 24, 36, and 48 hours. Anti–Bdhsp70-2 antiserum was used as a loading control. (D) Relative levels of BdeIF5A in lysates from untreated or GC7-treated parasites. (E) Relative levels of hypusinated eIF5A (BdeIF5A^Hyp) in parasite lysates from untreated or GC7-treated parasites. (F) Immunoblot analyses using anti-eIF5A and anti-hypusine antibodies using lysates from parasites cultured in the absence of PUT for 0, 12, 24, 36, and 48 hours. Anti–Bdhsp70-2 antiserum was used as a loading control. (G and H) Relative levels of BdeIF5A (G) and BdeIF5A^Hyp (H) in lysates from parasites maintained in PUT-depleted medium at different time points. (I) Western blot analysis using anti-eIF5A and anti-hypusine antibodies using lysates from parasites cultured in PUT-depleted (−PUT) medium, PUT containing medium (+PUT), and −PUT media supplemented with either SPD (+SPD; 9.2 μM) or SPM (+SPM; 9.2 μM) for 48 hours. Anti-Bdhsp70–2 antiserum was used as a loading control. (J and K) Relative levels of eIF5A (J) and BdeIF5A^Hyp (K) in lysates from parasites cultured in the aforementioned media. All data (A to K) presented as means ± SD of three independent experiments. Statistical significance of differences was calculated using Welch’s t test. *significant P < 0.05, **significant P < 0.001, ****P < 0.0001; ns, no significant difference.

To assess whether the growth defect observed upon PUT depletion in B. duncani is due to altered eIF5A hypusination, parasites were shifted to PUT-free media, and samples were collected for immunoblot analyses using anti-eIF5A and anti-hypusine antibodies at 12-hour intervals. While the levels of BdeIF5A were comparable at all time points tested (Fig. 5, F and G), the levels of BdeIF5A^Hyp declined significantly starting at 24 hours postPUT withdrawal (Fig. 5, F and H), with ~80% reduction in BdeIF5A^Hyp levels measured 48 hours postdepletion (Fig. 5, F and H). Supplementation with SPD or SPM restored BdeIF5A hypusination to levels comparable to PUT-containing cultures (Fig. 5, I to K). Together, these pharmacological and cell biological data demonstrate that PUT plays a key role in SPD-mediated hypusination of the translation factor BdeIF5A in B. duncani.

To assess the broader relevance of SPD in apicomplexans, we extended our studies to P. falciparum. Unlike B. duncani, P. falciparum synthesizes polyamines de novo from ORN (Fig. 6A). As reported previously, the intraerythrocytic stages of P. falciparum are sensitive to inhibitors of the polyamine biosynthesis pathway, including DFMO (targeting ODC), 4-MCHA (targeting SPDS), and MDL 72527 (targeting SPM oxidase), as well as to GC7 (targeting DHS) (fig. S9, A to D) (34–37). The inhibitory effects of DFMO on P. falciparum growth were reversed by PUT, SPD, or SPM supplementation (fig. S9E). In contrast, 4-MCHA inhibition was reversed by SPD or SPM but not PUT (fig. S9F), and MDL 72527 inhibition was reversed by PUT or SPD but not SPM (fig. S9G). These results confirm SPD as the primary polyamine required for P. falciparum development.

Fig. 6. eIF5A hypusination in P. falciparum is inhibited by DFMO and GC7.

(A) Schematic of polyamine biosynthesis and hypusination pathways in P. falciparum. (B) Western blot analysis of eIF5A and hypusinated eIF5A using lysates from parasites grown with or without DFMO (IC₅₀ concentration, 0.5 mM) for 0, 12, 24, 36, 48, and 72 hours. Glyceraldehyde phosphate dehydrogenase (GAPDH) was used as a loading control. (C and D) Relative levels of eIF5A and hypusinated eIF5A in DFMO-treated and untreated parasites. (E) Western blot analysis of eIF5A and hypusinated eIF5A in parasites grown with or without DFMO (IC₅₀ concentration, 0.5 mM) or DFMO plus 9.2 μM PUT (+PUT), 9.2 μM SPD (+SPD), or 9.2 μM SPM (+SPM) for 72 hours. GAPDH was used as a loading control. (F and G) Relative levels of eIF5A and hypusinated eIF5A under the aforementioned different media conditions. (H) Western blot analysis of eIF5A and hypusinated eIF5A in parasites grown with or without GC7 (IC₅₀ concentration, 15 μM) for 0, 12, 24, 36, 48, and 72 hours. (I and J) Relative levels of eIF5A and hypusinated eIF5A in untreated and GC7-treated parasites. (K) Western blot analysis of eIF5A and hypusinated eIF5A in parasites grown with or without GC7 (IC₅₀ concentration, 15 μM) or GC7 plus 9.2 μM PUT (+PUT), 9.2 μM SPD (+SPD), or 9.2 μM SPM (+SPM) for 72 hours. (L and M) Relative levels of eIF5A and hypusinated eIF5A under these different conditions. Data are presented as means ± SD of three independent experiments. Statistical significance of differences was calculated using Welch’s t test. *P < 0.05 and **P < 0.01; ns, nonsignificant P value.

Further analysis of P. falciparum PfeIF5A hypusination revealed a similar dependence on polyamine availability. Treatment with DFMO or GC7 reduced hypusinated PfeIF5A (PfeIF5AHyp) levels by ~80 and ~95% at 48 and 72 hours for DFMO and ~40, ~60, and ~90% at 36, 48, and 72 hours for GC7, respectively (Fig. 6, B to D and H to J). Supplementation with PUT, SPD, or SPM restored PfeIF5AHyp levels in DFMO- or GC7-treated cultures (Fig. 6, E to G and K to L). Together, these data indicate that despite differences in polyamine acquisition, both B. duncani and P. falciparum rely on SPD for eIF5A hypusination and intraerythrocytic development, underscoring its critical role in parasite biology.

DISCUSSION

PUT, SPD, and SPM are organic polycationic amines that play essential roles in all eukaryotes including pathogenic protozoa (38–41). While some parasites use de novo synthesis routes for production of these polyamines, others rely on the uptake of preformed polyamines for survival. However, in most parasites, it remains unclear which of the three polyamines are critical for parasite development and by what mechanism. In this study, we demonstrate that B. duncani exclusively relies on a salvage pathway for PUT uptake through a putative polyamine general transporter (PGT) and demonstrated that the key output of this pathway is SPD, which plays a critical role in the operation of the translation machinery through hypusination of eIF5A (Fig. 7). The availability of full genome sequences from several protozoan parasites and, in some cases, further genetic and pharmacological studies have helped map the components of the polyamine biosynthesis pathway in Plasmodium spp., Trypanosoma spp., Toxoplasma spp., and Cryptosporidium spp. (41). In organisms primarily reliant on de novo biosynthesis, the pathway initiates with the conversion of arginine to ORN by arginase 1. ORN is further decarboxylated to form PUT by the action of an ODC, a rate-limiting step in the biosynthesis of polyamines. PUT is further converted first to SPD and then to SPM by the actions of SPDS and SPM synthase, respectively, through the sequential addition of aminopropyl groups (Fig. 7). In Trypanosoma cruzi (the causative agent of chagas disease in humans) and Toxoplasma gondii (the causative agent of toxoplasmosis), a functional ODC is absent, and PUT is taken up from the host (42). The loss of a functional ODC in these parasites is a classic example of reductive genome evolution (43) where the parasites have lost the genes involved in de novo polyamine biosynthesis pathway and rely on their host for polyamine salvage. We provide pharmacological, biochemical, and genomic evidence that B. duncani lacks ODC and ADC enzymes and rely on PUT uptake from the host. In this study, we showed that the intraerythrocytic stages of B. duncani parasites are insensitive to inhibition by the ODC inhibitor DFMO. Furthermore, we found that ORN supplementation failed to rescue parasite growth on media lacking polyamines, indicating that a functional ODC is absent in B. duncani. Accordingly, bioinformatic analyses failed to identify an ODC gene in B. duncani, and biochemical assays using whole parasite extracts showed that no conversion of ORN to PUT occurs in this parasite. Together, these data suggest that B. duncani has undergone reductive evolution, which resulted in the loss of de novo biosynthesis of PUT and relies exclusively on its host to salvage this polyamine.

Fig. 7. Model of the modes of PUT acquisition, polyamine biosynthesis, and eIF5A hypusination used by B. duncani and P. falciparum.

In B. duncani, PUT is salvaged from the host, through a putative PGT, and used for the synthesis of SPD and SPM. In P. falciparum, PUT production relies on the conversion of arginine to ORN by the parasite arginase and subsequent conversion of ORN to PUT by ODC. Alternatively, P. falciparum can also salvage PUT from host via PGT. In both parasites, PUT is processed further to form SPD and SPM, and SPD is used as a substrate in the hypusination pathway and results in the addition of hypusine to eIF5A. The hypusinated eIF5A is critical for protein translation. SPS, SPM synthase; SMO, SPM oxidase; ODC, ORN decarboxylase; SPDS, SPD synthase; DHS, deoxyhypusine synthase; DOHH, deoxyhypusine hydroxylase.

Whereas no ODC activity could be detected using B. duncani cellular extracts, PUT conversion into SPD and SPM using cell-free extracts was detected by mass spectrometry (Fig. 3A) and by TLC analysis of cell-free extract catalyzed reactions using ¹⁴C-PUT as substrate (Fig. 3C), suggesting that SPDS aminopropyl transferase activity exists in this parasite. Guided by this finding, we searched for the encoding gene(s) using available sequences from known bacteria and eukaryotes and identified a single gene encoding a putative SPDS homolog in B. duncani (BdPAPT). A homolog of this enzyme was also identified in B. microti (BMR1_02g00531) and shares ~30% identity with the B. duncani protein (fig. S2). The BdPAPT protein is much larger than their homologs in P. falciparum, E. coli, and Homo sapiens. The regions of homology between BdPAPT and other SPDSs encompass the region around the catalytic residues known to be critical for catalysis including the conserved Asp¹³³⁷ residue, the equivalent of residue Asp¹⁵⁸ in E. coli SPDS. The activity of B. duncani BdPAPT-Cterm enzyme was confirmed using mass spectrometry by the formation of SPD from PUT (Fig. 3A), as well as by TLC analysis of the reaction catalyzed by recombinant BdPAPT-Cterm (Fig. 3H), as well as by recombinant CD (BdPAPT-CD) (Fig. 3H). Mutation of Asp¹³³⁷ in BdPAPT-Cterm to alanine resulted in loss of this activity (Fig. 3H). Moreover, using a fluorescence-based DAB-APT assay, the kinetics of BdPAPT-Cterm was performed, and the kinetic constants were determined (Fig. 3, I to K). Unlike classical single-substrate enzymes that follow Michaelis-Menten kinetics, aminopropyl transferase enzymes, which have two substrates (PUT or SPD and dcSAM), often exhibit a significant lag at the beginning and follow a nonhyperbolic sigmoidal kinetic profile (Fig. 3, I to K) (22).

In P. falciparum, the SPDS enzyme has been demonstrated to have dual activity catalyzing the formation of both SPD and SPM from PUT and SPD, respectively, and no canonical SPM synthase gene is found in the genome (36). While the gene encoding Babesia SPM synthase remains unknown, our analyses identified a SPM oxidase that shares homology with the yeast FMS1 enzyme known to catalyze the conversion of SPM into SPD (22, 44). Further studies are needed to characterize the remaining components of the Babesia polyamine biosynthesis pathway. These will be further facilitated by the development of efficient genetic tools for manipulation of the parasite.

In this study, we demonstrated that depletion of PUT severely diminishes the ability of B. duncani to complete its life cycle within human erythrocytes (Fig. 1D). However, PUT depletion from the growth medium did not result in 100% parasite death. The residual growth seen in media lacking PUT may be due to low amounts of polyamines present either in serum, a critical supplement for continuous propagation of the parasite in vitro, or in human erythrocytes. When the growth of B. duncani in media lacking PUT was compared between RBCs collected from fresh blood (<30 days old) or aged blood (60 and 90 days old) (fig. S10), residual growth was higher with young RBCs than with aged RBCs. This is consistent with the hypothesis that human RBCs contain polyamines that are depleted over time as these cells age. However, the polyamine levels even in freshly isolated human RBCs are insufficient to support optimal in vitro parasite growth since the parasites cannot successfully replicate in media lacking polyamines.

Research aimed to elucidate the physiological roles of polyamines in mammalian cells suggested a plethora of functions associated with the alterations in the polyamine biosynthesis pathway, including defects in protein translation and cell proliferation (45). The importance of PUT salvage in B. duncani indicates that either PUT or its downstream products are essential for parasite growth. Apart from playing an important role in cell growth, the polyamine SPD has been reported to prolong longevity in yeast, Caenorhabditis elegans, Drosophila, and mouse (46). Our pharmacological data indicated that SPD is the main polyamine required for intraerythrocytic proliferation of B. duncani (Fig. 2). SPD is the substrate for DHS, the first enzyme of the two-step hypusination pathway that leads to the formation of a hypusine residue by conjugation of the aminobutyl moiety of SPD to a specific lysine residue of eukaryotic translation factor eIF5A (30) (Fig. 5A). The hypusination pathway has evolved in eukaryotes and is highly conserved, indicating that the cellular function of eIF5A has been maintained throughout evolution (47). The addition of a hypusine residue on Lys⁵⁰ of eIF5A is an important posttranslational modification which is orchestrated by the sequential actions of the enzymes DHS and DOHH. On the basis of sequence homology with DHS and DOHH from other species, B. duncani genome was found to encode genes for DHS (BdWA1_002959-T1) and DOHH (BdWA1_002323-T1).

Our data showed that inhibition of the first step in eIF5a hypusination in B. duncani with GC7 results in a marked decrease in BdeIF5A^Hyp (Fig. 5, C to E). We further showed that depletion of PUT leads to reduced hypusination of eIF5A in B. duncani. The levels of hypusinated eIF5A (BdeIF5A^Hyp) were significantly reduced in the parasites cultured in PUT-depleted growth medium compared to PUT-containing medium (Fig. 5, I to K). This finding is consistent with our transcriptomic data that confirmed altered expression of genes involved in protein translation (fig. S6).

Similar to its importance in B. duncani, SPD is also critical for the survival of P. falciparum in human RBCs. In P. falciparum, the primary route of SPD production is through de novo synthesis from arginine (Fig. 6A). Previous studies have demonstrated that the inhibition of ODC, which catalyzes the rate-limiting step in the polyamine pathway, leads to growth arrest that can be reversed by the addition of polyamines (34). In this study, we demonstrated that inhibition of ODC by DFMO, as well as inhibition of the hypusination machinery with GC7 in P. falciparum, results in significant reduction in the levels of PfeIF5A^Hyp (Fig. 6). Moreover, studies in Plasmodium vivax have demonstrated the importance of the eIF5A hypusination pathway, suggesting that parasite eIF5A to be a promising therapeutic target for human malaria (37).

In summary, our data indicate that unlike P. falciparum that can use both de novo polyamine biosynthesis and uptake by salvage mechanisms, B. duncani uses a salvage pathway to acquire PUT from the host and use it as a precursor for the biosynthesis of downstream polyamines. Our pharmacological data suggest that, in both organisms, the primary function of PUT and SPM is to produce SPD that serves as a precursor for hypusination of eIF5A, a critical posttranslational modification for the operation of the translation machinery. Catalytic steps in polyamine utilization and eIF5A hypusination represent promising therapeutic targets for the development of drugs to treat human babesiosis, malaria, and potentially other diseases caused by intraerythrocytic parasites.

MATERIALS AND METHODS

In vitro culture of B. duncani in human RBCs

B. duncani parasites were cultured in vitro as previously described (48, 49). Parasite growth was monitored and evaluated in the following media: DMEM (Thermo Fisher Scientific, 11965-092) and DMEM/F12 (Lonza, BE04-687/U1). The media were supplemented with 20% heat-inactivated fetal bovine serum, 2% 50× Hypoxanthine-Thymidine (HT) Media Supplement Hybrid-Max (Sigma-Aldrich, H0137), 1% 200 mM l-glutamine (Gibco, 25030-081), 1% 100× penicillin/streptomycin (Gibco, 15240-062), and 1% gentamicin (10 mg/ml; Gibco, 15710-072). The DMEM and DMEM/F12 media supplemented with the above listed components were named as cDMEM and cDMEM/F12 media, respectively. Parasitemia was monitored by light microscopy examination of Giemsa-stained blood smears.

Parasite propagation in various culture media

B. duncani parasites were cultured in vitro in human RBCs in complete DMEM/F12 (cDMEM/F12) medium. The parasitemia was determined following staining of thin blood smears with Giemsa. To compare the growth of parasites in DMEM versus DMEM/F12 medium, parasites were precultured in cDMEM/F12 medium [5% hematocrit (HC); A⁺ human RBCs], washed three times using incomplete DMEM (iDMEM), and diluted to 0.5% parasitemia (using fresh human RBCs) in iDMEM medium. Subsequently, 2 ml of the parasite cultures (0.5% parasitemia and 5% HC) was aliquoted into different tubes and centrifuged for 5 min at 1800 rpm at room temperature (RT). Following centrifugation, the supernatant was discarded, and the growth media to be tested (for example, PUT-depleted medium, PUT-depleted medium supplemented with polyamines of interest, etc.) were added to different tubes in triplicates. The cultures from different tubes were plated in wells of a 12-well plate, and day 0 Giemsa smears were prepared for microscopic analysis. The cultures were maintained for 2 days at 37°C, and parasitemia was determined at day 3 after which all the cultures were diluted to 0.5% parasitemia by addition of fresh human A⁺ RBCs and the respective growth media and incubated at 37°C for 2 days. A similar procedure was conducted on days 6 and 9. From day 9 onward, the parasites were maintained continuously until day 15 without dilution. The respective culture media were replaced on days 12 and 14, and parasitemia was monitored on days 12 and 15.

Nutrient supplementation assays

To test whether the growth defect of parasites maintained in media lacking PUT can be rescued by addition of ORN (the precursor of PUT), B. duncani parasites were grown in PUT-containing medium (cDMEM/F12) to 10% parasitemia. The parasites were washed three times in iDMEM, and the cultures were diluted to 0.5% parasitemia (5% HC) in iDMEM. One-milliliter cultures (0.5% parasitemia and 5% HC in iDMEM medium) were aliquoted in 15 different microcentrifuge tubes (three replicate tubes per experimental condition). The cultures were centrifuged for 5 min at 1800 rpm at RT, and the supernatants were discarded. To the first set of three microcentrifuge tubes (triplicates), 1 ml of PUT-containing medium (cDMEM/F12) was added (positive control). To the remaining 12 tubes, 1 ml each of PUT-depleted medium (cDMEM + all except PUT) was added. To a set of three tubes, l-ornithine was added to 0.92 μM final concentration. To a second set of three tubes, SPD was added to 0.92 μM final concentration. To the third set of three tubes, SPM was added to 0.92 μM final concentration. The cultures from all the tubes were seeded in the wells of a 24-well plate. Day 0 smears were prepared, and the parasitemia was estimated. The cultures were allowed to grow continuously for 6 days, with media change (respective media as listed above) on day 3. The final parasitemia on day 6 was monitored by the examination of Giemsa-stained blood smears.

In vitro drug susceptibility assays

The susceptibility of the intraerythrocytic development cycle of B. duncani WA-1 cultured in cDMEM/F12 to DFMO (Sigma-Aldrich, D193) and DFMA (Cayman Chemical, 16415) was determined by incubating the parasite cultures (0.5% parasitemia and 5% HC in either PUT containing or depleted medium) to different concentrations of these drugs (5, 2.5, 1.3, and 0.6 mM DFMO; 250, 125, 62.5, and 31.3 μM DFMA). The untreated parasites treated with the vehicle (water) were used as no drug controls. Uninfected RBCs (5% HC) were used as the negative control. The assays were performed in triplicates in a 96-well plate in a volume of 200 ml and maintained for 60 hours at 37°C in an incubator with a mixture of 2% O₂, 5% CO₂, and 93% N₂. Subsequently, 100 ml of the culture per well of the 96-well plate was used in a SYBR Green I assay. Briefly, 100 ml of the culture from the 96-well assay plate was transferred to the Costar 96-well black plate, and 100 ml of SYBR Green I lysis buffer [20 mM tris (pH 7.4), 5 mM EDTA, 0.008% saponin, 0.08% Triton X-100, and 1× SYBR Green I (Molecular Probes, 10,000× solution in dimethyl sulfoxide)] was added to the wells containing the parasite culture. The plate was incubated at 37°C in the dark for 1 hour and subsequently read on a BioTek SynergyMX fluorescence plate reader with an excitation of 497 nm and an emission of 520 nm. Three independent experiments were performed in biological triplicates, and the data were analyzed using GraphPad Prism (version 9.4.1).

Activity of 4-MCHA, MDL 72527, and GC7 against B. duncani intraerythrocytic development in vitro

B. duncani WA1 parasites were cultured in vitro in cDMEM/F12 medium or cDMEM/F12 supplemented with either 9.2 μM PUT, SPD, or SPM. The cultures maintained in these media were diluted to 0.5% parasitemia (5% HC) and treated with decreasing concentrations [twofold dilution starting from 10 mM (4-MCHA), 1 mM (MDL 72527), and 0.4 mM (GC7)] of different drugs, including 4-MCHA (Sigma-Aldrich, 177466), MDL 72527 (Sigma-Aldrich, M2949), and GC7 (Santa Cruz, sc-396111) for 60 hours in a 96-well plate. The parasite viability was determined by the SYBR green I method (mentioned above). Three independent experiments were performed in biological triplicates, and the data were analyzed using GraphPad Prism (version 9.4.1).

The combined in vitro efficacy of 4-MCHA and MDL 72527 against B. duncani was evaluated using parasites cultured under different conditions (either depleted of PUT or supplemented with PUT, SPD, or SPM). Parasites (0.5% parasitemia and 5% HC) were exposed to the IC₉₀ concentrations of MDL 72527 (84 μM) and 4-MCHA (790 μM) for 60 hours, after which the parasite survival was estimated using SYBR green I labeling. Three independent experiments were performed in biological triplicate, and the data were analyzed using GraphPad Prism (version 9.4.1).

Immunoblot analysis

Whole-cell lysates from cultures of B. duncani grown in different media (PUT-containing, PUT-depleted, and PUT-depleted media supplemented with SPD or SPM) were collected as follows. B. duncani–infected RBCs were pelleted, treated with 0.1% saponin (Sigma-Aldrich, SAE0073) to lyse the RBCs and subjected to centrifugation at 14,000g. The parasite pellet was washed three times with 1× phosphate-buffered saline (PBS), resuspended in 1× Laemmli buffer (Bio-Rad, 161-0747) containing 5% β-mercaptoethanol (Bio-Rad, 1610710), boiled for 5 min at 95°C, and loaded onto Mini-Protean TGX Stain Free Gels 4-20% (Bio-Rad, P4568096). The gels were transferred onto 0.45-mm nitrocellulose membranes (Bio-Rad, 1620115) and blocked in 5% skimmed milk for 2 hours at RT. Following this, the membranes were exposed to either anti-eIF5A rabbit polyclonal antibodies (1:500) (Abcam, ab137561) or anti-hypusine rabbit polyclonal antibodies (1:500) (Sigma-Aldrich, ABS1064-I) and incubated overnight at 4°C. As a loading control, anti–Bdhsp70-2 rabbit polyclonal antisera were used. Following the primary antibody incubation overnight, the membranes were washed three times in 1× PBS containing 0.1% Tween 20 (PBS-T; Thermo Fisher Scientific, 85114) and two times in 1× PBS. The membranes were then incubated with goat anti-rabbit immunoglobulin G (H + L) horseradish peroxidase–conjugated secondary antibody (1:5000) (Thermo Fisher Scientific, 31460) for 1 hour at RT. Following this, the membranes were washed three times in PBS-T and two times in 1× PBS, and the signals were developed using the Super signal West Pico PLUS chemiluminescent substrate (Thermo Fisher Scientific, 34577). The blot was imaged using the LI-COR Odyssey-Fc imaging system. The signal from the anti–Bdhsp70-2 rabbit serum was used for normalization.

Sample collection for RNA-seq

B. duncani parasites were cultured in vitro in PUT-containing growth medium to 10% parasitemia. Following this, the cultures were collected in microcentrifuge tubes and washed five times with iDMEM medium. The cultures were then diluted to 1% parasitemia using fresh human RBCs and either PUT containing or depleted growth media. A total of 25-ml cultures (1% parasitemia and 5% HC) in duplicate plates in either PUT containing or depleted growth medium were seeded and maintained for 24 hours. After this, the cultures were harvested in different tubes, and the pellets were resuspended in 5 ml of TRIzol (Sigma-Aldrich, T9424) each. Following this, 1 ml of chloroform was added to each tube, mixed vigorously for 5 min, and incubated for 10 min at RT. The tubes were then centrifuged at 13,000g for 20 min. The aqueous layer was collected, and the equal volume of isopropanol was added, mixed, and incubated for 10 min at RT. The tubes were then centrifuged at 13,000 rpm for 10 min, and the supernatants were discarded. The pellets were washed twice in freshly prepared 70% ethanol and air dried after the final wash. The pellets were resuspended in 25 ml of diethyl pyrocarbonate–treated water. The concentration and quality of RNA were estimated using NanoDrop (Synergy HTX multimode reader, Agilent), and Illumina RNA-seq was performed on the samples. Total RNA was then treated with a DNA-free DNA removal kit (Thermo Fisher Scientific, AM1906) followed by mRNA purification using the NEBNext Poly(A) mRNA Magnetic Isolation Module [New England Biolabs (NEB), E7490S]. RNA-seq library was constructed using an NEBNext Ultra II RNA library preparation kit (NEB, E7770S) according to the manufacturer’s instructions. The libraries were amplified for 15 polymerase chain reaction cycles [45 s at 98°C followed by 15 cycles of (15 s at 98°C, 30 s at 55°C, and 30 s at 62°C), 5 min at 62°C]. Libraries were subjected to 150-bp paired-end sequencing on the Illumina NovaSeq platform.

RNA-seq data analysis

Raw sequencing reads in FASTQ format were imported into Partek Flow for preprocessing and analysis. Quality control was performed using FastQC, and low-quality reads as well as adapter sequences were trimmed. Reads were aligned to the reference genome (B. duncani) using the Spliced Transcripts Alignment to a Reference (STAR) aligner with default parameters. Read counts per gene were quantified using the featureCounts tool. Differential gene expression analysis comparing PUT-depleted versus PUT-containing samples was performed using the DESeq2 package within the Partek Flow software. Genes with an adjusted P value (false discovery rate) of <0.05 and a fold change of ≥1.5 were considered significantly differentially expressed. Last, hierarchical clustering was performed, and heatmaps were generated in Partek Flow to visualize differentially expressed (up-regulated or down-regulated) genes between the two conditions.

Flow cytometry and measurement of ROS

B. duncani WA-1 cultures were initiated at 0.5% parasitemia in different growth media, including PUT containing or depleted media or PUT-depleted medium supplemented with 9.2 μM SPD or 9.2 μM SPM in a volume of 1 ml. All the cultures were maintained at 37°C for 2 days. In parallel, another set of parasites grown in PUT-containing media were treated with either 1 mM artemisinin for 2 days or 100 μM H₂O₂ for 12 hours. On day 3, all cultures were washed twice with 1× PBS and resuspended in 1 ml of 1× PBS. DHR123 (0.25 mg/ml; Sigma-Aldrich, D1054) and Hoechst H33342 (1:1000) were added to all cultures and incubated for 30 min at 37°C in the dark. The cultures were then washed twice with 1× PBS to remove the dyes and resuspended in 1 ml of 1× PBS in the dark. Unstained B. duncani–infected RBCs and uninfected human RBCs were used as negative controls. The cells were analyzed using a BD LSRII flow cytometer. The cells positive for rhodamine (formed by breakdown of DHR123 through the action of ROS) and Hoechst were counted. Rhodamine-positive B. duncani–infected RBCs in H₂O₂-treated cultures were counted and used as controls. The data were plotted on GraphPad Prism, and the statistical significance of differences between parasite growth in different growth media was estimated by multiple t tests.

Determination of ODC and SPDS (aminopropyl transferase) activity in B. duncani whole-cell lysate using LC-MS and TLC

ORN, PUT, SPD, and SPM (>99% purity) were purchased from Sigma-Aldrich. Water, 0.1% formic acid, and acetonitrile were purchased from Fisher Chemical (Hampton, NH). The standards were prepared at the concentration range of 25 to 2500 ng/ml. Quality controls were prepared at the concentrations of 100 and 500 ng/ml.

B. duncani cultures were grown to a parasitemia of 20%, collected by centrifugation, and lysed using 0.15% saponin (Sigma-Aldrich, S7900). Following saponin lysis, the supernatant (referred to in this manuscript as “hemolysate” and containing the host cytosolic fraction) was separated from the parasite pellet by centrifugation at 14,000g for 30 min at 4°C. The parasite pellet was then used to prepare cell-free extracts. The parasite pellet was washed with PBS and resuspended in 10 volumes of hypotonic lysis buffer [1 mM ATP, 1 mM E64, 20 mM Hepes (pH 8.0), and 0.03% SDS] and subjected to 5 to 7 quick freeze-thaw cycles in liquid N₂. Following this step, the parasite lysate was centrifuged at 14,000g for 5 min at 4°C. The supernatant (cell-free extract) was collected and used as an enzyme source for the activity assays.

For detection of ODC activity and aminopropyl transferase activity by mass spectrometry analysis, parasite lysates and their heat-inactivated (incubation at 80°C for 45 min) counterparts were collected in buffer A used for ODC activity [10 mM ORN, 25 mM sodium phosphate, 0.2 mM EDTA, and 0.1 mM pyridoxal phosphate (pH 7.5)] or buffer B used for aminopropyl transferase activity assay [0.1 or 0.5 mM dcSAM (BOC Biosciences), 0.1 or 0.5 mM PUT, 1 mM EDTA, 1 mM dithiothreitol (DTT), 1 mg of bovine serum albumin (BSA), and 50 mM potassium phosphate buffer (pH 7.5)] for 1 hour at 37°C. As a negative control, lysates (with and without heat inactivation) from uninfected human RBCs were incubated with 10 mM each of ORN, PUT, SPD, and SPM for 1 hour at 37°C. After completion of the enzyme reactions, heat inactivation was performed by putting the samples at 85°C for 5 min. For analyzing the reaction products formed by cell-free extract or recombinant enyzme, 10 ml of the reaction products was spotted onto Silica 60 plates (500 mm; Merck), and the TLC was run using a solvent system consisting of n-butanol–acetic acid–pyridine water (3:3:2:1, v/v/v/v). Polyamines were visualized with ninhydrin spray, followed by incubation at 110°C for 5 min.

For samples that were analyzed with LC-MS to detect the enzyme products (SPD and SPM), the samples were resuspended in equal volumes of sodium phosphate buffer (pH 7.2) and incubated for 3 hours at 4°C with intermittent mixing. Subsequently, three volumes of cold acetonitrile was added to the reaction tubes, and the samples were centrifuged at 2100g at 4°C for 15 min. The supernatants were collected and subjected to mass spectrometry analysis to detect the peaks of different polyamines.

Five microliters of each sample was injected onto the LC-MS system for analysis. LC-MS analysis was performed on an API 4000 QTrap mass spectrometer (Applied Biosystems Sciex, Toronto, Canada) coupled with the Agilent HP1200 high-performance liquid chromatography system (Agilent Technologies, Santa Clara, CA). Analyst 1.7 software was used for data acquisition and analysis. A targeted multiple reaction monitoring (MRM) method was used to quantify the level of PUT, ORN, SPD, and SPM. An Agilent Eclipse XDB-C18 column (3.5 micros, 2.1 mm by 100 mm) coupled with an Agilent C18 guard column was used for the liquid chromatography separation at 50°C. The gradient started with 98% of 0.1% formic acid in water (A) and 2% of 0.1% formic acid in acetonitrile (B), maintained for 0.2 min, and increased to 30% B in 4.3 min, further increased to 50% B in 1.5 min and 85% in 0.5 min, maintained at 85% B for 1.5 min, went back to 2% B in 0.5 min, and equilibrated for 3.5 min before next injection.

The API 4000 Qtrap mass spectrometer was operated using an electrospray ionization source in positive ion mode. MRM transitions monitored for PUT, ORN, SPD, and SPM were 89.1/72.0, 133.1/116.2, 146.2/72.0, and 203.2/129.1, respectively. The declustering potentials were 48, 45, 35, and 60 eV; entrance potentials were 5, 5, 5, and 9 eV; collision cell exit potentials were 10, 7, 10, and 10 eV; and collision energies were 14, 13, 20, and 17 eV, respectively. The ion spray voltage was 5500 eV, the source temperature was 400°C, ion source gas 1 and 2 pressures were both 50 psi (344.738 kPa), and curtain gas and Collisionally Activated Dissociation (CAD) gas were set at 15 psi (103.421 kPa) and high mode.

Recombinant BdPAPT-Cterm expression and purification

A codon-optimized BdPAPT-Cterm (BdWA1_000190; amino acids 727 to 1454) and CD BdPAPT-CD (amino acids 1020 to 1454) genes were synthesized and cloned into the pMAL-c4x-1-H(RBS) bacterial expression plasmid, and the resulting constructs were introduced into Rosetta (DE3) E. coli cells (Thermo Fisher Scientific, 713973). Two clones were selected and tested for the expression of each construct. Briefly, each clone was inoculated into 5 ml of LB containing ampicillin (50 mg/ml) and allowed to grow overnight at a 37°C incubator shaker (200 rpm). The following day, secondary cultures were initiated using the primary cultures and grown to 0.6 OD₆₀₀ (optical density at 600 nm). Following this, 0.5 mM isopropyl-β-d-thiogalactopyranoside (IPTG) was added, and the cultures were shifted from 37° to 16°C and allowed to grow for 16 hours. The cultures were collected by centrifugation at 5000g for 5 min. The pellets were resuspended in 1× Laemmli sample buffer, boiled at 95°C for 10 min, and centrifuged at 10,000g for 5 min. The supernatants from uninduced and induced samples were run on 4 to 20% SDS–polyacrylamide gel electrophoresis (SDS-PAGE) to the check protein expression using Coomassie staining.

For purification of the recombinant enzymes, 500-ml culture was grown in LB medium containing ampicillin (50 mg/ml) and induced with 0.5 mM IPTG. The harvested bacterial pellet was resuspended in lysis buffer [25 mM tris-HCl (pH 8.0), 500 mM NaCl, 0.5% glycerol, 50 mM l-arginine, deoxyribonuclease (250 U/ml), protease inhibitor cocktail, and 0.002% CHAPS] and disrupted by sonication on ice (Omni Sonic Ruptor 400 Ultrasonic Homogenizer) by 15-s burst at 70% amplitude, five times, with 30-s cooling intervals. The bacterial lysate was centrifuged at 16,000g for 20 min, and the supernatant was collected. A column containing 500 ml of amylose resin (NEB, E8021L) was packed and equilibrated with column buffer (200 mM NaCl, 20 mM tris-HCl, 1 mM EDTA, and 1 mM DTT). The supernatant was incubated with the amylose resin for 1 hour at 4°C with end-to-end shaking. Following this, the column was washed with 10 volumes of the column buffer, and the protein was eluted using 10 mM maltose. The purified recombinant enzymes were run on SDS-PAGE and stained with Coomassie blue to check the purity. The protein concentration was determined using NanoDrop (Biotek SynergyMX with take3 plate; TAKE3-SN).

SPDS activity assay

Recombinant BdPAPT-Cterm and BdPAPT-CD were used for the activity assay. The aminopropyl transferase activity was determined by the formation of SPD from the substrates PUT and dcSAM. The reaction mixture contained 0.1 mM dcSAM (BOC Biosciences), 0.1 mM PUT, 1 mM EDTA, 1 mM DTT, 1 mg of BSA, 2.5 mg of either BdPAPT-Cterm, BdPAPT-CD, or mutant BdPAPT-Cterm^D1337A, and 50 mM potassium phosphate buffer (pH 7.5) in a total volume of 50 μl. The enzyme reaction was incubated at 37°C for 1 hour. Subsequently, three volumes of cold acetonitrile was added to the reaction tubes, and the samples were centrifuged at 2100g at 4°C for 15 min. The supernatant was collected and subjected to mass spectrometry analysis to detect the peaks of different polyamines.

Alternatively, for detecting SPDS activity using 1,2-DAB–based assay, the enzyme reaction was performed as mentioned above for 1 hour, and after that, 35 μl of enzyme reaction was mixed with 85 μl of buffer containing 1,2-DAB (pH 9.6) and allowed to incubate at RT in the dark for 1 hour in a clear-bottom black 96-well plate. Subsequently, the total fluorescence intensity was measured using 364-nm excitation and 425-nm detection wavelengths. The net fluorescence intensity was calculated from total fluorescent intensity by subtracting the fluorescence values of substrates (dcSAM + PUT).

For the aminopropyl transferase activity assay using ¹⁴C-PUT (American Radiolabeled Chemicals, ART 0279), a total of 50 μl of enzyme reaction contained reaction buffer (recipe above), ¹⁴C-PUT (3.5 nmol, 0.2 μCi), 150 μM dcSAM, and 100 μg of cell-free extracts or hemolysates (B. duncani or P. falciparum) or 1 μg of S. cerevisiae Spe3. The protein content of the cell-free extracts and hemolysates was determined using the Bradford assay. The reaction was done at 37°C for 120 min, and the reaction was terminated by heating the tubes at 95°C for 5 min.

Thin-layer chromatography

The formation of the product (SPD) from substrates (PUT and dcSAM) in APT reactions was analyzed by using TLC on Silica 60 plates (500 μm; Merck). The solvent used for TLC consisted of n-butanol–acetic acid–pyridine water (3:3:2:1, v/v/v/v). The APT reaction products were spotted on to TLC plates, allowed to dry at RT for 1 hour, and then run in the aforementioned buffer system in a TLC tank for 4 hours. Following this, the plates were removed from the tank and allowed to air dry overnight. For signal development, 0.2% ninhydrin (Thermo Fisher Scientific, A1040990) solution (in ethanol) was sprayed on the plate, and the plate was incubated at 110°C for 5 min.

In vitro culture of P. falciparum (3D7) in human RBCs

P. falciparum 3D7 blood-stage cultures were maintained at 4% hematocrit in human A⁺ erythrocytes in RPMI 1640/Hepes medium (Gibco, 11875119) supplemented with 0.5% Albumax II (Gibco, 11021037), HT media supplement (Sigma-Aldrich, H0137), and gentamicin (Thermo Fisher Scientific, 15710-072) under mixed gas (5% O₂, 5% CO₂, and 90% N₂). Parasitemia was monitored by light microscopy examination of Giemsa-stained blood smears.

Bioinformatic analysis

To validate our pharmacological and biochemical observations, we performed mining of the B. duncani complete genome sequence to identify ODC and ADC enzymes. Homology searches were performed (BLASTn and BLASTp) using ODC and ADC from different apicomplexan parasites (P. falciparum and Theileria parva), human, mouse, plant, and other lower eukaryotes (Leishmania donovani and Trypanosoma brucei) as queries. We did not identify any ODC and ADC genes in the B. duncani genome.

Supplementary Materials

The PDF file includes:

Table S1

Figs. S1 to S10

Legend for data S1

Other Supplementary Material for this manuscript includes the following:

Data S1