Introduction

Plasmodium parasites are responsible for malaria and cause yearly half a million deaths in sub-Saharan countries. Malaria continues to be a major worldwide health concern despite recent successful attempts to lower its prevalence (1–4). The infection starts when an infected female Anopheles mosquito, during a blood meal, injects salivary gland sporozoites into the skin of the vertebrate host. Some of the motile sporozoites escape the host immune defence and invade blood vessels. They are carried by the blood flow to the liver, where they invade and infect hepatocytes (5). Approximately 50% of the parasites are often killed by host cell responses (6–8). Those that managed to establish a productive invasion form a parasitophorous vacuole (PV) in which the parasite undergoes massive growth and asexual replication known as schizogony. One single parasite gives rise to thousands of daughter parasites called merozoites in 3 to 6 days depending on the Plasmodium species (9–11). Fully developed merozoites exit the liver cells in merosome vesicles through adjacent blood vessels (11–14). The merosomes then rupture and release merozoites into the blood stream. The parasites immediately invade erythrocytes, where they undergo again asexual replication to form the blood stage merozoites. Some merozoites differentiate into sexual stages, termed gametocytes, which are ingested by female mosquitos during a blood meal. In the mosquito midgut, the gametocytes develop into microgametes and macrogametes, which fertilize and form a zygote. The later undergoes differentiation into motile ookinetes, which penetrate the gut wall, forming oocysts (15–18). The oocyst produces the midgut sporozoites from the sporoblasts. The sporozoites migrate to the salivary gland from day 11 post-infection in P. berghei (Frischknecht et al., 2004; Frischknecht & Matuschewski, 2017; jerome P. Vanderberg, 1974; J. P. Vanderberg, 1975). During the liver stage development, the parasite relies on the fast and massive asexual replication which is essential for the transmission and pathogenesis in erythrocyte stage, representing key targets for malaria control strategies. Investigating the parasite microtubule dynamics during the rapid replication could provide important insight on the parasite biology and to develop strategies for halting the massive proliferation during the clinically asymptomatic liver stage.

Microtubules are cytoskeletal filaments of about 25nm outer diameter, made up of usually 13 chains of protofilaments composed of globular α-tubulin and β-tubulin heterodimers. They are crucial components of the cytoskeleton and offer structural support to cells and also function as tracks for the movement of organelles, vesicles, and other cellular components (23,24). Microtubules are involved in important cellular activities such as cell division and the production of cilia and flagella to allow cell motility. Microtubules exhibit inherent dynamics, alternating between polymerisation (rapid growth) and spontaneous depolymerisation (shrinkage) in a process termed dynamic instability (23,25,26). The dynamics of microtubules and their interactions with other cellular structures are regulated by specific microtubule-associated proteins (MAPs), which selectively bind to either the plus or minus ends of microtubules (27–30).

Tubulin is subject to seven identified post-translational modifications (PTMs), among which are acetylation of the lysine residue at position 40 in the luminally exposed part of α-tubulin, glycylation, which entails the incorporation of polyglycine side chains of variable length, phosphorylation, palmitoylation, tyrosination/ detyrosination and polyglutamylation. These modifications are extensively preserved and enabling cells to modulate microtubule characteristics, dynamics, and interactions with microtubule-associated proteins (MAPs) and motor proteins (31,32). Tyrosinated microtubules are often associated with microtubule stability. These stable microtubules represent a specialised subset that exhibits enhanced resistance to depolymerization, even in the presence of destabilizing agents. They play a crucial role in maintaining cellular architecture, facilitating intracellular transport, and supporting specialized cellular functions. In contrast to dynamic microtubules, which continuously undergo cycles of polymerization and depolymerization due to dynamic instability, stable microtubules persist for extended periods. They are also frequently polyglutamylated, a modification that regulates their interactions with motor proteins and other cellular components. These microtubules are predominantly localized in structures requiring sustained structural integrity, such as cilia, flagella, neuronal axons, and motile cell types (33–36). Glutamylation, or polyglutamylation, is a post-translational modification characterized by the addition of multiple glutamate side chains to glutamate residues on both α- and β-tubulin. This modification occurs primarily at the C-terminal tails and is catalyzed by a family of enzymes known as polyglutamylases. Polyglutamylation plays a critical role in regulating the interactions of microtubules with MAPs and motor proteins, thereby influencing microtubule dynamics and cellular functions (37–40). The properties and functions of tubulin are largely encoded within their protein sequences, contributing to molecular diversity that influences key microtubule characteristics. This variation has been conceptualized as the “tubulin code”, a regulatory framework through which cells control microtubule diversity to achieve specific physiological functions (24). Nevertheless, the precise mechanisms by which cells establish and interpret this code remain largely unknown. Although tubulin is highly conserved across evolution, it exhibits structural and functional variations between species, cell types, and even within single cells.

In eukaryotic cells, tubulins are key components of the microtubule cytoskeleton, and their genetic complexity reflects the organism’s level of cellular specialization. In yeast, two distinct genes encode α-tubulin, while a single gene encodes β-tubulin. In contrast, the human genome contains nine genes encoding α-tubulin and an equal number encoding β-tubulin, reflecting a greater level of complexity and functional specialization in higher eukaryotes (24,41,42).

The Plasmodium genome encodes two α-tubulins (α1-tubulin and α2-tubulin) and one β-tubulin (43–47). The α1-tubulin gene encodes a protein of 450 amino acids, while α2-tubulin encodes 453 amino acids (48). The two isoforms share 94% amino acid identity. However, α2-tubulin lacks a terminal tyrosine residue, which is present in α1-tubulin. This difference suggests potential variation in post-translational modifications and functional roles (44,49). In P. berghei, the α1-tubulin gene is expressed across most stages of the Plasmodium life cycle, whereas α2-tubulin expression appears restricted to male gametocytes, gametes, and newly formed zygotes, indicating distinct functional roles for the two isoforms (32,50). Deletion of α1-tubulin blocks sporozoite formation in oocysts and lowering α1-tubulin expression leads to less infective sporozoites (50).

There is a significant degree of amino acid conservation between human and Plasmodium tubulins, with approximately 83.7% identity for α-tubulin and 88.5% for β-tubulin. However, Plasmodium tubulin shows greater similarity to plant tubulin than to mammalian tubulin. This raises the potential for developing compounds that selectively target parasite microtubules while sparing the host cell cytoskeleton, offering a promising avenue for therapeutic intervention (43,45–47,51).

Recent studies on Plasmodium falciparum blood stage schizonts have revealed that hemi-spindles lack polyglutamylation, whereas subpellicular microtubules do exhibit this post-translational modification (52–54). Similarly, SPMTs appear to be stabilized by microtubule inner proteins, while spindle microtubules do not show these luminal proteins (55). These differences may reflect functional specialisation of microtubule subpopulations in different cellular structures and developmental stages.

The liver stage of Plasmodium exhibits one of the fastest rates of replication known among eukaryotic cells, a process likely dependent on microtubule dynamics. However, to date, no studies have investigated microtubule dynamics during the liver stage. This study was therefore designed to explore the dynamics of parasite tubulin during the asexual liver stage, using recently developed expansion microscopy protocols (54,56–58).

Results

Expansion microscopy reveals the presence of the sporozoites subpellicular microtubules during the liver stage of infection

To investigate microtubule organization during the Plasmodium parasite liver stage, HeLa cells were infected with sporozoites and fixed them at different time points using cold methanol. We then performed IFA to stain host and parasite microtubules using an α-tubulin antibody (Figure 1A). The IFA revealed a marked accumulation of microtubules in proximity of the PVM (stained with anti-UIS4) from early infection (24 hours post-infection (hpi)) to the schizogony (30 to 54 hpi) (Figure 1A inlets). To examine these structures in greater detail, we applied the recently developed pre-staining expansion microscopy (PS-ExM) (56). By expanding the cells five-fold, PS-ExM enabled visualization of a large multimeric tubulin structure within the parasite, detectable from 6 hours post-infection (hpi) to 48 hpi (Figure 1B). Based on observations of sporozoites prior to liver infection, we hypothesize that this structure originates from the sporozoite subpellicular microtubules (SSPM) (Figure 1C top panel). We refer to this structure as the liver stage parasite microtubule bundles (LSPMB). Notably, the LSPMB progressively shrank over time and was no longer detectable by the end of schizogony (54 hpi) (Figure 1A, B). Using the 3D and 4D imaging software IMARIS, we quantified the LSPMB volume at various stages of liver infection. The results confirmed a consistent decrease in LSPMB size between 6 and 56 hpi (Figure 1D).

Figure 1:

To verify that the LSPMB is not an artifact of the IFA and PS-ExM, we performed live cell imaging of infected HeLa cells. Cells were incubated with the tubulin tracker SPY650 for one hour prior to imaging. The LSPMB was clearly visible in sporozoites before invasion (Movie 1, ∼10 hpi) and persisted until late liver stage development, gradually shrinking during schizogony (Figure 1E; Movies 2–4, ∼36 hpi to 48 hpi). These live imaging results corroborate our findings from IFA and expansion microscopy, validating the presence and dynamic nature of the LSPMB during liver stage development.

The liver stage parasite subpellicular microtubules are dynamic

The SSPM has previously been described as highly stable structures, enabling sporozoite formation and invasion (50,55,59). These microtubules, similar to the corresponding microtubules in T. gondii, are resistant to conventional depolymerizing agents (59). To investigate whether the LSPMB retain this stability during liver stage development, we treated infected HeLa cells with the microtubule-depolymerizing agents Amiprophos-methyl (AMP) and oryzalin herbicide. These compounds are known to selectively target dynamic Plasmodium microtubules without affecting mammalian cells (43,45,47,51,60) (Figure 2 and Supplementary Figure 1A and B). We first validated drug specificity and cytotoxicity by testing various concentrations on HeLa cells over one week. Neither compound affected host cell viability (Supplementary Figure 1A and B) or disrupted host microtubules (Supplementary Figure 1D and E). However, oryzalin showed a stronger inhibitory effect on parasite growth compared to AMP (Supplementary Figure C) and was therefore selected for further experiments (Figure 2A). Sporozoites were dissected into medium containing 10 µM or 20 µM oryzalin and then used to infect HeLa cells (Figure 2A and B), which were incubated with the same drug concentrations. At 48 hpi, the cells were fixed and analyzed by IFA and PS-ExM. To test reversibility, oryzalin was washed out at 24 hpi prior to the onset of schizogony (Figure 2B). At 10 µM, oryzalin significantly impaired nuclear replication, spindle formation, and LSPMB volume when present the entire cultivation time (Figure 2C, D and E). Partial recovery of hemi-spindles and nuclear replication was observed, suggesting that the effect of 10 µM oryzalin is reversible at this stage (Figure 2C, D, and E). However, at 20 µM, neither the hemi-spindles nor the LSPMB recovered after drug removal, indicating irreversible microtubule disruption at higher concentrations (Figure 2A and B). To investigate whether the microtubule depolymerization by oryzalin affects the parasite growth and survival during the liver stage, we measured parasite size within infected HeLa cells at 6, 24 and 48 hpi. Treatment with 10 µM oryzalin led to a significant reduction in parasite size and survival (Figure 2 F, G, H, and I). We further examined merozoite formation by staining for MSP1 at 56 hpi. Parasites treated with 10 µM oryzalin failed to express MSP1, indicating a block in merozoite development and suggesting that intact microtubules are essential for DNA replication and merozoite formation (Figure 2J). Interestingly, at lower concentrations of 0.5 µM, oryzalin treatment led to an unexpected elongation of the LSPMB and hemi-spindles up to four times longer than controls suggesting that microtubule turnover is tightly regulated (Supplementary Figure 1F, G and H). In conclusion, oryzalin-mediated depolymerization confirms that both the LSPMB and the hemi-spindles are dynamic structures during the liver stage. This dynamic behaviour was further supported by live-cell imaging using the SPY650 tubulin tracker (Movie 5, 6, and 7).

Figure 2:

Given the observed microtubule remodeling in response to oryzalin, we next sought to determine whether known microtubule-associated proteins, specifically those regulating plus-end dynamics, are expressed during the liver stage and might contribute to the observed plasticity. Only a limited number of motor proteins and microtubule associated proteins, such as centrin and the kinesin protein family, have been identified during the Plasmodium liver stage (11,61–63). The recently characterized EB1 protein, known to regulate microtubules plus-end dynamics and length, has so far been reported for being expressed during the sexual blood stage (Mauer et al., 2023; Yang et al., 2023; Zeeshan et al., 2023). However, these studies did not assess EB1 expression during the asexual mosquito and liver stages. Given our observations of dynamic microtubule behavior and excessive polymerization following oryzalin treatment, we explored whether EB1 might be expressed during the liver stage. To this end, we generated a parasite line expressing endogenous EB1 fused to eGFP via double homologous recombination (Supplementary Figure 2C and D). As expected, EB1-eGFP was detected in sexual blood stages (Supplementary Figure 2E and F) consistent with previous reports (64,65). Interestingly, we also observed EB1 expression in both midgut and salivary gland sporozoites (Supplementary Figure 2G). Importantly, EB1 was also detected during liver stage development, where it localized to the LSPMB and hemi-spindles (Supplementary Figure 2H–I).

These findings suggest that EB1 may play a broader role in Plasmodium microtubule regulation than previously recognized, potentially contributing to the dynamic remodeling of microtubules during the liver stage.

The liver stage parasite subpellicular microtubule carries tubulin post-translational modifications

Microtubule functions are often regulated by different post-translational modifications (PTMs), which influence their stability, interactions, and dynamics. To gain insight into the dynamic nature of LSPMB during the liver stage development, we examined α-tubulin PTMs using immunofluorescence (IFA) and expansion microscopy (Figure 3 and Movie 5, 6 and 7). To investigate whether the LSPMB is present in cell types other than HeLa cells, we infected primary hepatocytes with Plasmodium sporozoites and fixed the cells at different timepoints (6, 24, 36, 48, 54 and 56 hpi). We also analyzed sporozoites isolated from mosquito salivary glands. Using antibodies specific for polyglutamylated (anti-IN105) (66) (Figure 3A) and tyrosinated (anti-Ty) α-tubulin (Figure 3B), we observed that sporozoite microtubules, including the LSPMB, are both polyglutamylated and tyrosinated. Quantitative colocalization analysis confirmed a strong overlap between LSPMB and polyglutamylated tubulin in both sporozoites and liver stage parasites (Supplementary Figure 2A). Interestingly, following hepatocyte invasion, the LSPMB lost tyrosination (i.e., became detyrosinated) while retaining polyglutamylation. This detyrosination persisted throughout liver stage development, including early and late schizogony (24–56 hpi). During this period, polyglutamylation was observed on the LSPMB and at the poles of the hemi-spindles, but not along the polymerizing hemi-spindles themselves, which eventually disappeared along with the LSPMB (Figure 3A and C). Given previous findings implicating polyglutamylation in hemi-spindle pole formation in erythrocytic merozoites (31), we investigated whether there is a structural relationship between the LSPMB and the hemi-spindle poles. Using PS-ExM and the tubulin tracker SPY650, we visualized direct connections between LSPMB filaments and hemi-spindle poles. Live cell imaging confirmed this connection between LSPMB and hemi-spindle poles, mainly during the shrinkage of the LSPMB (Fig 3 F; Movie 8, 9 and 10, Supplementary Movie 1). These observations suggest that the LSPMB may be redistributed to form or support the hemi-spindle poles during schizogony. We also stained for γ-tubulin and observed its localization at the hemi-spindle poles in both the parasite and host cell (Supplementary Figure 2B), supporting the idea that Plasmodium may utilize γ-tubulin for nuclear division and microtubule assembly, as reported in other systems (67–70).

Figure 3:

To further define the spatial localization of the LSPMB, we used P. berghei parasites expressing the plasma membrane protein PMP1 fused to superfolder GFP (sfGFP) previously generated (71). ImageJ RGB profiling revealed that the LSPMB lies between the parasite plasma membrane (PM) and the parasitophorous vacuole membrane (PVM), suggesting that the sporozoite subpellicular microtubules undergo subcellular relocalization during liver stage development (Figure 3D and E). At later stages, particularly the cytomere stage, we observed reappearance of tyrosinated tubulin near the parasite nuclei (Figure 3B). During merozoite formation, this tyrosinated tubulin appeared as punctate structures within the forming merozoites (Figure 3B).

Tyrosination has previously been associated with stable microtubules involved in motility, such as those in sporozoites, flagella, and spermatozoa (33,34,50,55). The absence of tyrosination during early intracellular liver stage development when the parasite is non-motile may facilitate its transformation from a banana-shaped sporozoite to a rounded liver stage form (Figure 1 and 3). LSPMB detyrosination may contribute to microtubule dynamic instability, providing both energy and tubulin monomers for rapid growth and spindle formation during schizogony. Finally, the reappearance of tyrosinated microtubules during merozoite formation may serve to stabilize the microtubules required for merozoite motility and erythrocyte invasion.

The Plasmodium α-tubulin C-terminal modification are important for sporozoite formation

To investigate the functional role of parasite α-tubulin PTMs during the liver stage development, we generated P. berghei mutant parasite lines with targeted mutations in the α-tubulin C-terminal region, where most PTMs are predicted to occur (Figure 4A and Supplementary Figure 3A and B). Specifically, we created three mutant lines: PsΔ: a point mutation replacing the polyglutamylation site glutamic acid (E445) with alanine (A), TyΔ: deletion of the tyrosination/detyrosination motif (YEADY) from position 449, CtΔ: deletion of the entire C-terminal tail (EGEDEGYEADY) to prevent all possible C-terminal tail PTMs (Supplementary Figure 3A to C).

Figure 4:

Using AlphaFold predictions, we confirmed that the mutations were unlikely to affect tubulin folding (Supplementary Figure 3C). The mutants were generated via double homologous recombination and verified by integration PCR and sequencing of the mutated α-tubulin gene (Figure 4A and B; Supplementary Figure 3B). The development of mutant parasites was investigated during the mosquito, and liver stages. Significant defects emerged during mosquito stage development. The PsΔ and TyΔ mutants were able to form oocysts (Figure 4 C, Supplementary Fig 4A) and gliding sporozoites, albeit in reduced numbers (Figure 4D, E and G; Supplementary Fig 4B; and Supplementary Movie 2, 3, 4 and 5). Quantification of salivary gland sporozoites at 18 days post-blood feeding showed that PsΔ and TyΔ mutants produced approximately half as many sporozoites as the control line (NoΔ) (Figure 4D). We next examined sporozoite motility, which typically consists of continuous circular gliding (16,72–76). PsΔ and TyΔ salivary gland sporozoites were assessed and compared to controls. While PsΔ have a similar gliding velocity compared to controls, TyΔ sporozoites displayed increased gliding velocity (2.5 µm/s vs. 1.4 µm/s) (Figure 4G), suggesting a regulatory role of the tyrosination motif in fine-tuning motility. To assess whether these motility phenotypes influenced infectivity, we performed in vivo infection assays with PsΔ and TyΔ. All mutant lines sporozoites were able to infect mice similarly to control parasites (Supplementary Figure 4C). However, the CtΔ mutant failed to produce salivary gland sporozoites entirely (Figure 4D). Since CtΔ mutant parasites failed to reach the salivary glands, we analyzed hemolymph sporozoites for this line compared to the controls (Figure 4F). These showed severely impaired motility, lacking the typical gliding pattern and exhibiting significantly reduced speed (Figure 4F and H). Given that PsΔ and TyΔ mutants produced motile salivary gland sporozoites, we infected HeLa cells with PsΔ and TyΔ mutant sporozoites and CtΔ mutant hemolymph sporozoites to assess their liver stage development and identify any downstream phenotypic effects.

The Plasmodium α-tubulin C-terminal mutants undergo liver stage development

Using polyglutamylation-specific antibodies, we found that PsΔ mutant parasites still exhibited polyglutamylated liver stage parasite microtubule bundles (LSPMBs) in both sporozoites and liver stage forms, although the signal intensity was significantly reduced compared to controls (Figure 5A, B and C and G). As tyrosination was previously observed only in sporozoites and late liver stages, we performed PS-ExM on sporozoites and stained them with α-tubulin tyrosination-specific antibodies. TyΔ salivary gland sporozoites and CtΔ hemolymph sporozoites both showed detectable tyrosination signals, although the signal in CtΔ parasites was markedly weaker than in controls (Figure 5D, E and F). Interestingly, as PTM-specific antibodies continued to recognize signals in the PsΔ and TyΔ mutants, it is possible that the parasite compensates by relocating the PTM sites to adjacent residues within the C-terminal sequence. Alternatively, the antibody signal may stem from β-tubulin modifications rather than α-tubulin.

Figure 5:

To assess parasite survival and growth during liver stage development, we performed quantitative microscopy (Figure 5H to O). Both PsΔ and TyΔ mutant parasites successfully infected HeLa cells and progressed through liver stage development (Figure 5B, D, H, I, J and K). This was further confirmed by staining the parasitophorous vacuole membrane (PVM) using anti-UIS4 antibodies (Figure 5A, B and C). Parasite size was measured at 6 hpi and 48 hpi to assess developmental progression. No significant differences were observed at 6 hpi (Figure 5I). However, by 48 hpi, mutant parasites were significantly smaller than controls, indicating impaired growth (Figure 5J). Survival rates of PsΔ and TyΔ mutants were comparable to controls (Figure 5K), with infection rates of 15% for PsΔ (matching the control) and 27% for TyΔ, exceeding the control. As expected, the CtΔ hemolymph sporozoites failed to infect and to develop liver stage forms (Figure 5C, L, M, N and O). The sporozoites failed to establish productive infections with a typical PVM at all examined time points (Figure 5C, M, N and O).

To confirm the phenotypes, we compared our mutants with previously generated P. berghei α1-tubulin mutants described by Spreng et al. (2019), including parasites with intron deletions (intronΔ), C-terminal tyrosination deletions 540ADY (C-termΔ), and α2-tubulin exon swaps. These mutants showed similar phenotypes, including a complete absence of salivary gland sporozoites in α2-tubulin exon swap parasites, mirroring the CtΔ mutant phenotype (Figure 4D and H; Supplementary Figure 3D and E). These results underscore the importance of α-tubulin C-terminal PTMs in sporozoite formation and migration from the midgut to the salivary glands.

The failure of CtΔ parasites to generate viable sporozoites highlights the essential role of α-tubulin PTMs in sporozoite development and migration. Co-mutation of both α- and β-tubulin PTM sites may prove lethal to parasite development in the mosquito stage. Future studies should focus on identifying and characterizing the enzymes responsible for tubulin PTMs in Plasmodium, and genetic knockout strategies should be pursued to further elucidate the function of these modifications.

Discussion

Our study provides the first in-depth characterization of microtubule architecture and dynamics during the Plasmodium liver stage. Using tubulin trackers and α-tubulin antibodies, we identified two major microtubule populations: a large, multimeric subpellicular network that we termed the liver stage parasite microtubule bundles (LSPMB), and the polymerizing hemi-spindles involved in nuclear division. The LSPMB appears to originate from the SSPM and persists until late schizogony. These observations align with earlier electron microscopy studies by Meis et al. (77) who reported the presence of subpellicular microtubules and inner membranes in liver stage parasites up to the point of mitotic spindle formation.

The behavior and function of microtubules are closely regulated by PTMs. We found that α1-tubulin expressed during the liver stage carries residues associated with polyglutamylation and tyrosination. Using PTM-specific antibodies, we confirmed that Plasmodium sporozoites display both modifications on their subpellicular microtubules. Tyrosination, a well-known marker of stable microtubules, has been described in highly motile systems such as cilia, flagella, and spermatozoa (33,34,50,55). This is consistent with the need for high motility in sporozoites, which must traverse multiple tissue barriers during transmission (16,72,73).

Polyglutamylation, in contrast, introduces negative charges on the tubulin surface that enhance interactions with MAPs, and is commonly associated with dynamic microtubule behavior (78–80). This modification has previously been linked to merozoite formation and intracellular trafficking in Plasmodium blood stages (31). The co-occurrence of tyrosination and polyglutamylation in sporozoites likely ensures a microtubule network that is both stable and functionally interactive during mosquito stage migration.

Upon hepatocyte invasion, we observed a loss of tyrosination while polyglutamylation was maintained, suggesting a functional switch from stable to dynamic microtubules (Figure 6). This transition may facilitate morphological transformation from a motile, slender sporozoite into a rounded liver stage form and could also support the parasite’s rapid replication. The liver stage is one of the fastest-replicating phases in the Plasmodium life cycle, with replication rates among the highest observed in eukaryotic cells (81,82). During this phase, the parasite synthesizes proteins de novo and hijacks host cell lipids and proteins for membrane biogenesis (83–86). We propose that the parasite brings its SSPM into hepatocytes to be repurposed as the LSPMB (Figure 6). The observed microtubule disassembly over time may support tubulin turnover and safes energy for intracellular development.

Figure 6:

In addition to their structural roles, LSPMB filaments were occasionally found in contact with hemi-spindle poles, suggesting that LSPMB contribute to spindle formation during schizogony. This connection, although infrequent in fixed cells, was frequently observed during live imaging suggesting a transient connection. We also detected expression of EB1, a microtubule plus-end binding protein previously believed to be restricted to the sexual blood stage (64,65,87), during mosquito and liver stages. Its localization to LSPMB and hemi-spindles suggests a broader role for EB1 in regulating microtubule dynamics throughout the parasite life cycle.

To directly test the function of tubulin PTMs, we generated P. berghei mutants with targeted C-terminal modifications of α1-tubulin. These included a glutamate-to-alanine substitution at a polyglutamylation site, deletion of the tyrosination motif, and complete deletion of the C-terminal tail. Despite the mutations, PTM-specific antibody staining remained detectable in the first two mutants, suggesting compensatory mechanisms such as modification of adjacent residues or of β-tubulin. The tyrosination antibody used (anti-Ty, T9028, Sigma-Aldrich) is a purified monoclonal antibody targeting C-terminal tyrosine residues on tubulins, whereas the monoclonal GT335 antibody (AG-20B-0020-C100, AdipoGen Life Sciences) recognizes the epitope formed at the branching point of polyglutamylation side chains (GEGE*EE, * marking the modification site) and detects both short and long glutamate side chains. The polyclonal antibody (AG-25B-0030-C050, AdipoGen Life Sciences) was raised against at least three glutamate residues and is thereby specific to long polyglutamylation side chains, though both antibodies may cross-react with β-tubulin or other proteins containing genetically encoded C-terminal modifications (66,88). The reduced fluorescence signal observed (Figure 5) in mutant parasite microtubules suggests either antibody cross-reactivity with β-tubulin or compensatory PTM relocation to adjacent α-tubulin residues. In the CtΔ mutant (lacking the entire C-terminal tail), PTM relocation is abolished, and residual signal likely stems from β-tubulin which was insufficient to sustain sporozoite motility and transmission. These possibilities warrant further investigation. Previous studies have also demonstrated partial redundancy between α1- and α2-tubulin modified (49,50,89). However, the complete C-terminal deletion mutant failed to produce salivary gland sporozoites and could not establish liver stage infection. These results are consistent with findings from Spreng et al. (2019), who showed that α1-tubulin mutants lacking the C-terminal region produced defective sporozoites with altered microtubule numbers and impaired motility.

We also observed that the LSPMB relocates during liver stage development. Using expansion microscopy on PMP1::GFP-expressing parasites (71), we found that the LSPMB localizes between the parasite plasma membrane and the PVM. This differs from earlier reports of SSPM positioning (77,90–92) but both notions are not mutually exclusive. It is possible that the LSPMB first localizes within the parasite but then re-localizes at later stages or even re-localizes dynamically. Its localization between the PM and the PVM positions it well for interaction with the host cell indicating an interaction with host cell structures such as ER tubules or host microtubules (71,84,93,94). However, whether this re-localization plays a role in nutrient acquisition or host-pathogen signaling remains to be explored.

Finally, during late schizogony and merozoite formation, we observed the reappearance of tyrosinated tubulin. This may reflect the need for stable microtubules in newly formed merozoites, which must be competent for host cell invasion.

In conclusion, our study reveals a previously unrecognized complexity in the microtubule biology of Plasmodium liver stages. The identification of the LSPMB as a dynamic structure regulated by tubulin PTMs and repurposed from the sporozoite cytoskeleton opens new avenues for understanding how the parasite controls its intracellular architecture. These PTMs appear critical for parasite shape, replication, and stage transition, and may represent valuable targets for therapeutic intervention. Future work should aim to identify the enzymes responsible for these modifications and explore their potential as drug targets. The behavior and function of microtubules are closely tied to their PTMs. We found that α1-tubulin expressed during the liver stage carries residues predicted to undergo polyglutamylation and tyrosination.

Using PTM-specific antibodies, we confirmed that sporozoite microtubules are both tyrosinated and polyglutamylated. Tyrosination is typically associated with stable microtubules and has been described in highly motile structures such as cilia and spermatozoa. This is consistent with the sporozoite’s need for high motility gliding at approximately 1.5 µm/s as it migrates from the mosquito midgut to the salivary glands and eventually traverses skin and vasculature to reach hepatocytes.

Methods

Ethics statement

All experimental procedures were performed in full compliance with the Swiss Animal Protection Act (TSchG) and were approved by the Canton of Bern’s Animal Experimentation Commission (Authorization, BE118/22). Personnel involved in animal studies held the Federation of European Laboratory Animal Science Associations (FELASA) accredited certifications. Female C57BL/6 and BALB/c mice of 6 to 10 weeks old and 20-30 g at infection were sourced from Janvier Labs (France) or bred in the animal facility at the University of Bern’s Institute of Cell Biology. Animals were maintained under standard housing conditions with daily health monitoring by trained staff.

P. berghei lines used in this study and the generation of transgenic parasites

The P. berghei ANKA parasite line was utilized for both murine and cellular infection models. To generate the α1-tubulin C-terminal mutant and eb1::eGFP parasite lines, the C-terminal coding and 3’ untranslated regions (UTRs) were cloned into the double homologous recombination vector pDHR, which contains the 3’UTR PbDHFR, the ef1α promoter region, the selectable markers hdhfr::yfcu, and additional 3’UTR PbDHFR sequences. The pDHR vector backbone originated from the plasmid pG362 (kindly provided by the Waters lab) (95).

First, the pG362 plasmid was digested with SacII and PstI to remove the AID, selectable marker, and GFP sequences. The amplified product (linker-eGFP-3’UTR PbDHFR) from the pOB277 vector was then cloned using T4 DNA ligase (NEB, M0202L). For subsequent cloning, restriction sites were added to the forward (KpnI and ApaI) and reverse (EcoRI and SmaI) primers. The plasmid was then digested with EcoRI and SmaI to clone the amplified product: ef1α promoter region-hdhfr::yfcu marker (a fusion gene of the positive selection marker human dihydrofolate reductase (hdhfr) and the negative selection marker yeast cytosine deaminase-uridyl phosphoribosyl transferase (yfcu) 3’UTR PbDHFR from the genomic DNA of the P. berghei GIMO parasites (RMgm-687) parental line (96).

The complete α1-tubulin gene sequence (PBANKA_0417700), including 5’ and 3’ UTRs, was retrieved from PlasmoDB (https://plasmodb.org). The linker-eGFP region of the pDHR vector was removed using ApaI and NotI restriction enzymes. The 5’UTR of the α1-tubulin gene was then subcloned into the digested pDHR using Gibson assembly. The resulting vector was digested with SmaI, and the PbHsp70 promoter and mCherry regions from RMgm-687, along with the 3’UTR of the α1-tubulin gene, were subcloned into the vector at the ApaI and SmaI restriction sites, respectively, using Gibson assembly.

Site-directed mutagenesis (QuickChange II XL Kit) was used to introduce the C-terminal modifications, with primer sequences listed in Tables 2 and 3. The final plasmid was linearized with ScaI before being transfected into the PbWT line (P. berghei ANKA line cl15cy1) (97). Similarly, the end-binding protein eB1 (PBANKA_0405600) with an eGFP tag was engineered using identical cloning strategies. The 5’ and 3’ UTRs were retrieved from PlasmoDB (https://plasmodb.org) and subcloned into the pDHR vector at the ApaI and SmaI restriction sites, respectively, using Gibson assembly. The final vector was linearized with NdeI before being transfected into the Pb1868 line (constitutively expressing Hsp70-mCherry-ef1α-NLuc) (98). Primer sequences are listed in Table 1.

Table 1:

Table 2:

Table 3:

Vector transfections were performed in schizont cultures as previously described (16,99). Briefly, P. berghei schizonts were cultured and purified (16,99,100). Electroporation of schizonts from the respective parasite lines was performed using the Amaxa® Human T Cell Nucleofector® system with the U33 program, followed by intravenous injection into naïve mice. The construct was integrated by double-crossover homologous recombination at the locus of interest. Pyrimethamine drug selection was administered via drinking water, and parasitemia was monitored daily from day 7 post-infection until mice tested positive (99).

Parasitaemia measurement and cloning of transgenic parasite lines

To measure the parasitemia ratio of infected mice, approximately 2 μL of blood from the tail vein of each infected mouse was added to 500 μL of PBS and analyzed by flow cytometry using the NovoCyte (Agilent, 2023) as described (101). The data were analyzed using NovoExpress Flow Cytometry Analysis Software (version 1.6.2, Build 230717.50551; Agilent Technologies).

To clone the different engineered transgenic parasite lines, mice were infected with the transgenic parasites. Three days post-infection, parasitemia was assessed and when parasitemia reached at least 1%, blood was collected from the infected mouse via heart puncture following euthanasia as described (16). The collected blood was passaged into a new naïve mouse. When parasitemia in the newly infected mouse reached 1-2%, 2 μL of tail vein blood was collected and mixed with RPMI-1640. Maintaining low parasitemia was crucial to prevent both double erythrocyte infection by P. berghei and gametocyte formation. Single P. berghei-infected erythrocytes were then sorted into individual wells of 96-well plates (Cellvis) containing 200 μL of RPMI-1640 supplemented with 10% FCS using a BD FACSDiscover S8 with CellView. Supplementing the medium with 10% FCS allowed the sorted infected erythrocytes to recover from the sorting process. Five mice were then intravenously infected with a single sorted infected erythrocyte each. The mice underwent drug selection via drinking water, and parasitemia was monitored daily from day 7 post-infection until parasites were detected. Clonality of the parasites in positive mice was verified by integration PCR using primers listed in Tables 1, 2, and 3.

Cell culture

Human epithelial HeLa cells (European Cell Culture Collection) were maintained in Minimum Essential Medium (MEM; 1-31F01-I BioConcept) supplemented with Earle’s salts and 10% heat-inactivated fetal bovine serum (FBS; Sigma-Aldrich), hereafter referred to as complete MEM (cMEM). Primary hepatocytes were cultured in William’s Medium E (WME) complete medium. All cells were incubated at 37°C under 5% CO₂. For infection assays, 40,000 HeLa cells were seeded onto coverslips in 24-well plates and allowed to adhere overnight. Cells were then infected with P. berghei sporozoites and maintained in cMEM containing 2.5 µg/mL amphotericin B (PAA Laboratories, P11-001) to prevent fungal contamination. Medium was exchanged every 24 hours.

Primary hepatocyte isolation

The mice primary hepatocytes were isolated according to the method described (8). Briefly, the cells were isolated from 12 to16 week old Balb/c or C57BL/6 mice using a two-step perfusion technique (8,83,102). Mice were euthanized via CO₂ inhalation (1–2 minutes), and after securing the animal in a supine position, a midline abdominal incision was made to expose the portal vein or inferior vena cava (IVC). A peristaltic pump was pre-rinsed with 70% ethanol and HEPES buffer, then set to 50 for initial cannulation of the portal vein (or IVC) using a Teflon-coated needle. Upon successful insertion, the flow was increased to maximum (999), and the opposing vein was severed to allow perfusion with HEPES buffer for 5 minutes, during which periodic clamping (10 seconds per minute) of the cut vein was performed. The tubing was subsequently changed to a collagenase solution, and perfusion was held for 7 minutes with repeated clamping. The liver was excised, transferred to a collagenase-containing well, and mechanically dissociated using filter tips and gentle shaking with forceps. Cell suspensions were sequentially collected, pooled, and washed with William’s Medium E (WME) by centrifugation at 50g for 2 minutes. The pellet was resuspended in WME, optionally passed through a 70 µm nylon strainer, and washed twice more. The final pellet was resuspended in WME complete medium (WME supplemented with L-glutamine, penicillin-streptomycin, amphotericin B, insulin-transferrin, and aprotinin), and cells were counted and seeded at densities appropriate for T25 flasks, GB dishes, or 24-well plates. After initial adhesion (1 hour), culture medium was supplemented or exchanged after 2–4 hours.

Immunofluorescence assay (IFA) of fixed primary hepatocytes and HeLa cells

Immunofluorescence was carried out as previously described (56,86,103). Cells were seeded onto coverslips in culture 24 well-plates and allowed to adhere overnight before infection. Infection was maintained for 24 hours post-infection (hpi), after which cells were fixed using cold methanol at -20°C for 15 minutes and subsequently washed twice with Phosphate buffered saline (PBS). Blocking was performed by incubating cells in 10% FCS in PBS for 20 minutes at room temperature to reduce non-specific binding. Primary antibodies staining was carried out by incubating cells in 10% of FCS diluted in PBS for 1 hour at room temperature, followed by three washes with PBS. Secondary antibodies incubation was performed in 10% of FCS as well for 1 hour at room temperature in the dark. The antibodies used are listed in the enumerated session bellow. After three additional washes with PBS, nuclei were counterstained with DAPI. Coverslips were then mounted onto slides with DAKO mounting medium and fluorescence imaging was performed using either Nikon W1 or LIPSI and Crest V3 confocal microscope equipped with filters for DAPI (4,6-Diamidin-2-phenylindol), AF488, AF594, and Cy5. Single-stain controls were carried out for every staining to ensure that bleed-through was negligible.

Sporozoite isolation from mosquito salivary glands and haemolymph

Anopheles mosquitoes were fed on Plasmodium-infected mice exhibiting a gametocytaemia of 0.2 to 1% and the sporozoite isolations were performed from day 18 post-infection as previously described (16,61). For salivary gland sporozoites, mosquitoes were anesthetized with chloroform, sterilized in 70% ethanol, and rinsed twice in PBS. The mosquitoes were then dissected, and salivary glands were extracted under a binocular microscope in Iscove’s Modified Dulbecco’s Medium (IMDM) pooled in 50 µL of ice-cold IMDM and mechanically homogenized using a pestle (Sigma-Aldrich, Z359947) powered with cordless motor (Sigma-Aldrich, Z359971) as described (16,61). For hemolymph sporozoite extraction, glass capillaries were heat-pulled to fine tips. The abdominal terminal segment of the mosquito was excised, and hemolymph was flushed from the thorax by PBS perfusion via capillary insertion (75,104) . Hemolymphs extracted were centrifuged (11,000 × g, 1 min), and pellets were resuspended in PBS or IMDM. Sporozoite quantification was performed using a Neubauer hemacytometer. Midgut, hemolymph, or salivary gland suspensions were mixed 1:1 with RPMI-1640 supplemented 10% FCS, and sporozoites within four corner grids were counted. Total yields (sporozoites/mL and per mosquito) were calculated based on dilution factors and dissection numbers.

Sporozoites IFA and gliding motility assay

To perform the sporozoite immunofluorescence assay (IFA), 12 mm cover glasses (0111530, Paul Marienfeld, Germany) were coated with 0.1 mg/mL poly-D-lysine (A38904-01, Gibco, USA) (56,105) and incubated overnight at room temperature. The following day, sporozoites were isolated from mosquito midguts, hemolymph, or salivary glands as described above. The extracted sporozoites were suspended in RPMI-1640 medium to a final volume of 200 µL in an Eppendorf tube and centrifuged at 1000 rpm for 3 min. The pellet was fixed by resuspension in 200 µL of cold methanol and incubated for 15 min at −20°C. After fixation, sporozoites were washed twice with PBS (1000 rpm centrifugation) and transferred onto poly-D-lysine-coated cover glasses in a 24-well plate (400024, Bioswisstec, Switzerland). The plate was incubated for 30 min at room temperature to facilitate sporozoite adhesion before proceeding with either the IFA or expansion microscopy protocol.

For the gliding motility assay, 5 µL of the sporozoite suspension in RPMI-1640 (supplemented with 10% FCS) was added to 96-well glass-bottom plates (P96-1.5H-N, Cellvis, USA) (56,72,73). The plate was centrifuged (1000 rpm, 1 min) and immediately live imaged using a Leica DM5500 widefield fluorescence microscope.

Expansion microscopy

Expansion microscopy (PS-ExM) was performed as previously established (56–58). Infected cells (HeLa cells, primary hepatocytes and sporozoites) were seeded on 13 mm coverslips in 24-well plates and fixed with either 4% paraformaldehyde (PFA; F8775, Sigma) or cold methanol. The cells were anchored by incubation in PBS containing 0.7% formaldehyde and 1% acrylamide at 37°C for 2 hours. Gelation was then performed by adding 5 µL each of 10% tetramethylethylenediamine (TEMED; 17919, Thermo Fisher) and 10% ammonium persulfate (APS) to 90 µL of monomer solution containing 19% sodium acrylate (408220, Sigma), 10% acrylamide (A4058, Sigma), and 0.1% N, N’-methylenebisacrylamide (M1533, Sigma). A 35 µL of monomer solution was dispensed onto parafilm in a humidity chamber, and coverslips containing the anchored cells were inverted onto droplets. Following 5 minutes incubation on ice to improve gel penetration, gels were polymerized at 37°C for 30 minutes. Gel-embedded samples were transferred to denaturation buffer (200 mM SDS, 200 mM NaCl, 50 mM Tris-HCl, pH 9) in 6-well plates and agitated (200 rpm) at room temperature for 15 minutes, followed by incubation in fresh buffer at 95°C for 1 hour with gentle shaking. Gel expansion was performed in autoclaved MilliQ water (50 mL) for 1 hour with optional buffer exchange. For immunostaining, gels were incubated with primary antibodies in 2% BSA for 2 hours at 37°C, washed with PBS-Tween 0.1%, then labelled with secondary antibodies plus 1 µg/mL DAPI for nuclear staining in 2% BSA for 2 hours at 37°C. Final gel expansion was conducted in MilliQ water containing 0.2% propyl gallate (02370, Sigma) to prevent photobleaching. Expanded samples were mounted on poly-D-lysine-coated (A38904, Gibco USA) glass-bottom dishes with cellular surfaces facing downward and imaged with confocal microscopy.

Antibodies used for IFA and expansion microscopy

The primary antibodies used were: anti-UIS4 chicken (1:1000 IFA/1:500 expansion; Proteogenix), anti-UIS4 rabbit (1:1000 IFA/1:500 expansion; P. Sinnis, Baltimore), polyglutamylation IN105 rabbit (1:1000 IFA/1:500 expansion; AG-25B-0030-C050, AdipoGen Life Sciences), anti-GFP rabbit (1:1000 IFA/1:500 expansion; SP3005P, Origene), anti-α-tubulin guinea pig (1:500 IFA/1:125 expansion; AA345, ABCD), anti-β-tubulin guinea pig (1:500 IFA/1:125 expansion; AA344, ABCD), anti-tyrosinated tubulin mouse (1:1000 IFA/1:500 expansion; T9028, Sigma Aldrich), anti-polyglutamylation GT335 mouse (AG-20B-0020-C100, AdipoGen Life Sciences), anti-γ-tubulin mouse (1:1000 IFA/1:500 expansion; T6557, Sigma Aldrich), and anti-GFP mouse (1:1000 IFA/1:500 expansion; 814460001, Roche).

Secondary antibodies included anti-chicken IgY Alexa Fluor 594 (1:1000 IFA/1:500 expansion; A11042, Invitrogen Molecular Probes), anti-rabbit IgG Alexa Fluor 594 (1:1000 IFA/1:500 expansion; A21207, Life Technologies) and 488 (1:1000 IFA/1:500 expansion; A11034, Sigma Aldrich), anti-rabbit IgG ATTO 647N (1:1000 expansion only; 40839, Sigma Aldrich), anti-guinea pig IgG Alexa Fluor 647 (1:1000 IFA/1:500 expansion; SAB4600033, Sigma Aldrich) and 594 (1:1000 IFA/1:500 expansion; A11076, Invitrogen Molecular Probes), and anti-mouse IgG ATTO 647N (1:1000 expansion only; 50185, Sigma Aldrich) and Alexa Fluor 488 (1:1000 IFA/1:500 expansion; A11001, Invitrogen Molecular Probes).

In vivo assay

To assess the infectivity of parasite mutant sporozoites to hepatocytes, Anopheles stephensi mosquitoes infected with different parasite mutants were allowed to feed on anesthetized naïve mice as previously described (Atchou et al., 2025). Three mice per parasite line and five mosquitoes per mouse were used. The blood stage infection progression was monitored daily via flow cytometry, starting from the day of infection until mice tested positive.

Oocyst size measurement

To quantify oocyst size, Anopheles stephensi mosquitoes were allowed to feed on anesthetized mice infected with the parasite (Rogerio et al. 2006; Hegge et al., 2009). At 7- and 11-days post-feeding, ten mosquitoes per time point were dissected, and their midguts were extracted and placed in 1× PBS. The midguts were then transferred to a 21-well microscopy slide, covered with a coverslip, and imaged using a Nikon Crest V3 confocal microscope (10× objective). The oocyst area was measured using Fiji software (ImageJ). Each experiment was performed in triplicate for every parasite line.

Liver stage parasite size measurement

To quantify the parasite size during liver stage development, HeLa cells (40,000 cells/well, triplicate) were seeded in 96-well plates and infected 24 hours later with 20,000 sporozoites per well. After a 2-hour incubation at 37°C, infected cells from each well were washed and redistributed into eight wells across two 96 well plates (4 wells/plate) to optimize cell density for imaging (Pino et al., 2017; Atchou et al., 2025; Sturm et al., 2006). Parasite development was analyzed at 6, 24, and 48 hpi using an InCell Analyzer 2000 with 10x objective. The image processing parameters such as kernel size, sensitivity, and size thresholds were optimized in InCell Developer software to distinguish individual parasites from background signals. Parasite growth was assessed based on size measurements, while survival rates were calculated as percentages relative to 6 hpi baseline counts.

Live cell imaging of Plasmodium microtubule dynamism

HeLa cells were seeded in a 96-well plate and infected the following day with P. berghei salivary gland sporozoites. After a 2-hour incubation at 37°C, cells were washed with MEM medium to remove uninfected parasites and residual mosquito salivary gland materials. The infected cells were then reseeded in a dark 24-well glass-bottom plate (Cellvis) and maintained in fresh cMEM at 37°C with 5% CO₂ (16,106). Prior to live imaging, cells were incubated for 1 hour in cMEM containing the SPY650-tubulin probe (1:1000 dilution; SC503, Spirochrome, Switzerland). Images were acquired using a Nikon W1 LIPSI spinning disk microscope with a 100X objective. Marked infected HeLa cells were imaged at defined time points to generate time-lapse movies capturing tubulin dynamics and parasite development.

Image processing and data analysis

All the live cell imaging, expansion microscopy, and IFA images in this study were acquired with Nikon W1 LIPSI and Crest V3, both in spinning disk mode. The microscope images were processed and analyzed using FIJI software, Huygens deconvolution Professional version (Huygens Remote Manager v3.8, Scientific Volume Imaging, The Netherlands), or, when specified, the 3D and 4D IMARIS software (Imaris 10.2.0; June 18, 2024). The 3D/4D image analysis software was used to segment the parasite LSPMB, hemi-spindle, and poles into either isospots or isosurfaces, as previously described (16,56). The counts, volumes, and coefficients of the isospots or isosurfaces were retrieved in .csv files, which were then imported into GraphPad Prism for statistical analysis. All data were analyzed using GraphPad Prism version 10.4.1 for macOS (GraphPad Software, San Diego, California, USA). Statistical comparisons were performed using one-way ANOVA with Dunn’s multiple comparisons test (*p ≤ 0.05, p ≤ 0.01, *p ≤ 0.001). All experiments in the study were carried out in triplicate.

Supplementary figures

S1 Figure

S2 Figure

S3 Figure

S4 Figure

Acknowledgements

The authors gratefully acknowledge the Microscopy Imaging Center (MIC) at the University of Bern, Switzerland, for their technical support and access to microscopy equipment. K.A. acknowledges the Swiss Confederation for the award of a Government Excellence Scholarship. This work was funded by the Swiss National Science Foundation (SNSF) (grant number 310030_212795) and an Multidisciplinary Center for Infectious Diseases (MCID), University of Bern, Switzerland, grant to Volker Heussler (MA-09).

Additional files

Movie 1. Time-lapse live imaging of P. bergei salivary gland sporozoites expressing GFP cytosolically. Sporozoites were isolated and incubated with tubulin tracker SPY650. Microtubules are shown in magenta; sporozoites are shown in green. Scale bar: 5µm.

Movie 2. Time-lapse live imaging of HeLa cells infected with P. berghei liver stage parasites expressing cytosolic GFP. Infected cells were incubated with tubulin tracker SPY650 for one hour prior to imaging. The movies show the development and dynamics of the LSPMB during liver stage infection. Microtubules are shown in magenta; parasites are shown in green. Imaging from 10 hpi to 25 hpi. Scale bar: 5 µm.

Movie 3. Time-lapse live imaging of HeLa cells infected with P. berghei liver stage parasites expressing cytosolic GFP. Infected cells were incubated with tubulin tracker SPY650 for one hour prior to imaging. The movies show the development and dynamics of the LSPMB during liver stage infection. Microtubules are shown in magenta; parasites are shown in green. Imaging from 30 hpi to 35 hpi. Scale bar: 5 µm.

Movie 4. Time-lapse live imaging of HeLa cells infected with P. berghei liver stage parasites expressing cytosolic GFP. Infected cells were incubated with tubulin tracker SPY650 for one hour prior to imaging. The movies show the development and dynamics of the LSPMB during liver stage infection. Microtubules are shown in magenta; parasites are shown in green. Imaging from 37 hpi to 49 hpi. Scale bar: 5 µm.

Movie 5. Time-lapse live imaging of HeLa cells infected with P. berghei liver stage parasites expressing cytosolic GFP. Infected cells were incubated with tubulin tracker SPY650 for one hour prior to imaging. The movies highlight the high dynamicity of the LSPMB during liver stage development, starting from early schizogony (∼30 hpi). Microtubules are shown in magenta; parasites are shown in green. Scale bar: 5 µm.

Movie 6. Time-lapse live imaging of HeLa cells infected with P. berghei liver stage parasites expressing cytosolic GFP. Infected cells were incubated with tubulin tracker SPY650 for one hour prior to imaging. The movies highlight the high dynamicity of the LSPMB during liver stage development, starting from early schizogony (∼30 hpi). Microtubules are shown in magenta; parasites are shown in green. Scale bar: 5 µm.

Movie 7. Time-lapse live imaging of HeLa cells infected with P. berghei liver stage parasites expressing cytosolic GFP. Infected cells were incubated with tubulin tracker SPY650 for one hour prior to imaging. The movies highlight the high dynamicity of the LSPMB during liver stage development, starting from early schizogony (∼30 hpi). Microtubules are shown in magenta; parasites are shown in green. Scale bar: 5 µm.

Movie 8. Time-lapse live imaging of HeLa cells infected with P. berghei liver stage parasites expressing cytosolic GFP. Infected cells were incubated with tubulin tracker SPY650 for one hour prior to imaging. The movies reveal the interaction between the LSPMB and the parasite’s hemi-spindles and spindle poles during liver stage development. Microtubules are shown in magenta; parasites are shown in green. LSPMB (white arrow) localizes at the periphery of the parasite and connects to hemi-spindles. Scale bar: 5 µm.

Movie 9. Time-lapse live imaging of HeLa cells infected with P. berghei liver stage parasites expressing cytosolic GFP. Infected cells were incubated with tubulin tracker SPY650 for one hour prior to imaging. The movies reveal the interaction between the LSPMB and the parasite’s hemi-spindles and spindle poles during liver stage development. Microtubules are shown in magenta; parasites are shown in green. LSPMB (white arrows) connects to the parasite’s hemi-spindle poles. Scale bar: 5 µm.

Movie 10. Time-lapse live imaging of HeLa cells infected with P. berghei liver stage parasites expressing cytosolic GFP. Infected cells were incubated with tubulin tracker SPY650 for one hour prior to imaging. The movies reveal the interaction between the LSPMB and the parasite’s hemi-spindles and spindle poles during liver stage development. Microtubules are shown in magenta; parasites are shown in green. Shrinkage of the LSPMB during the liver stage development (ROI marked by a white arrow). Scale bar: 5 µm.

Supplementary Movie 1. 3D projection of PS-ExM images showing LSPMB connected to the hemi-spindle pole in P. berghei-infected primary hepatocytes. Fixed and expanded primary hepatocytes were stained with anti-IN105 (polyglutamylated tubulin, green), anti-α-tubulin (magenta), and anti-UIS4 (PVM, grey). The movie shows the LSPMB in contact with the parasite’s hemi-spindle pole. Image analysis was performed using IMARIS 3D and 4D visualization software.

Supplementary Movie 2. Plasmodium berghei salivary gland (SG) sporozoite motility of P. berghei NoΔ (control), PsΔ, and TyΔ mutants, as well as hemolymph sporozoites of P. berghei NoΔ and CtΔ mutants. Scale bar: 5 µm