Pre-gelation staining expansion microscopy for visualisation of the Plasmodium liver stage Open Access

Malaria is a debilitating and potentially fatal disease caused by Plasmodium parasites. It is a major global health issue, affecting millions of people in developing countries, particularly in sub-Saharan Africa (Snow et al., 2005; WHO, 2021, https://www.who.int/publications/i/item/9789240040496). Despite extensive research efforts made, malaria remains a major public health challenge, and new treatment options are needed to effectively control and prevent the disease (Tse et al., 2019). A major challenge in the development of new effective anti-malaria drugs is the ability of the parasite to rapidly develop resistance to antimalarial drugs (Plowe, 2022; White, 2004). The Plasmodium life cycle is complex, consisting of both asexual and sexual stages that are present in the vertebrate host and in the mosquito. The transmission of malaria occurs when an infected female Anopheles mosquito bites a human and releases up to several 100 sporozoites into the skin (Amino et al., 2006; Beier, 1998; Venugopal et al., 2020). Sporozoites are motile and a proportion of them enter blood vessels and are transported with the blood flow to the liver where they invade hepatocytes (Tavares et al., 2013). Inside the liver cells, the parasite resides within a parasitophorous vacuole (PV), in which the parasite proliferates from one single sporozoite to thousands of daughter parasites, called merozoites. This process is also known as schizogony. The merozoites exit the liver cells in vesicles known as merosomes that are transported to adjacent blood vessels (Mota et al., 2001). When the merosomes rupture, they release the merozoites, which subsequently invade erythrocytes. The liver stage represents a crucial opportunity for anti-malaria interventions as it is clinically silent. The characteristic symptoms of malaria only occur upon repeated rounds of asexual blood stage replication and the associated rupture of the red blood cell (Derbyshire et al., 2011; Vaughan and Kappe, 2017). In the infected red blood cells, some merozoites develop into gametocytes (Cowman and Crabb, 2006). Gametocytes are the sexual form of the parasite which, when taken up during the blood meal of a mosquito, differentiate into gametes and mate in the midgut of the insect to form a zygote. This motile parasite stage transmigrates the midgut epithelium and forms a so-called oocyst, inside which sporozoites develop. Upon rupture of the oocyst, sporozoites migrate to the salivary gland of the mosquito and can be transmitted to the next human host. An important tool for the discovery of new treatment options is a better understanding of molecular details of the life cycle of the parasite and host–pathogen interactions. Conventional light microscopy allowed us to gain insight into cellular organisation (Rankin et al., 2010). However, due to the diffraction of light, the maximal resolution lies between 200–300 nm (Heintzmann and Ficz, 2006). Subcellular imaging of the parasite therefore remains challenging and requires super-resolution techniques, such as photo-activated localisation microscopy (PALM), stochastic optical reconstruction microscopy (STORM), structured illumination microscopy (SIM), and stimulated emission depletion (STED) microscopy (reviewed by Heintzmann and Ficz, 2006). However, they require expensive and complex microscopical equipment and special sample preparation (Heintzmann and Ficz, 2006). In comparison, the recently developed expansion microscopy (ExM) is relatively straightforward and does not require complicated preparations or specialised equipment, making it accessible to researchers all over the world (Chen et al., 2015; Gambarotto et al., 2019; Gao et al., 2017). In a relatively short period of time, ExM became a new cutting-edge imaging technique that has revolutionised the field of microscopy. The physical expansion of the sample results in a significant improvement in spatial resolution compared to traditional light microscopy techniques. Expansion microscopy is based on the usage of a hydrogel matrix. The hydrogel matrix is composed of a polymer network that is covalently attached to the biological sample and expands upon hydration (Chen et al., 2015). There are several different expansion microscopy protocols that have been developed, each with its own strengths and limitations (Chen et al., 2015; Gambarotto et al., 2021; Tillberg et al., 2016). The original ExM protocol developed by Edward Boyden's group in 2015 has been widely used and has been modified and improved over the years to increase the accuracy and specificity of the technique (Chen et al., 2015; Gao et al., 2017). Nevertheless, the basic principles of the protocol remain the same. One of the currently most widely used expansion microscopy protocols is ultrastructure expansion microscopy (U-ExM), which has been shown to preserve the cellular architecture in a near-native manner, achieving a four-fold expansion (Gambarotto et al., 2019, 2021). In comparison to the original protocol from Edward Boyden's group, the staining is performed after (instead of before) embedding the sample in the swellable gel. The post-expansion staining reduces antibody competition through the decrowding of the epitopes and prevents fluorophore loss (Gambarotto et al., 2019). However, the protocol is time-consuming, requiring roughly 2 days. Given that the antibody staining is performed on the gel, the incubation times are rather long with ∼3 h for each antibody incubation (Gambarotto et al., 2021). Another limitation is the loss of epitopes during the denaturation step. An alternative to the post-expansion labelling protocol is the proExM protocol developed by the group of Joshua Vaughan (Chozinski et al., 2016). Here, biological samples are stained with fluorescent antibodies before the gelation process begins. The fluorescent labels are afterwards covalently linked to the polymer. Subsequent protease treatment is used for exposing protein epitopes while preserving the signal of the fluorescent labels, which are fairly protease resistant. This approach has the advantage that existing stained samples can be entered into the expansion process and non-expanded samples can be performed under the same conditions (Chozinski et al., 2016). Since its first application in 2015, ExM has been applied to a variety of samples, including cells, tissues and even entire organisms (Chen et al., 2015; Gao et al., 2017). Recently, the U-ExM was used to study the molecular architecture of Plasmodium gametocytes (Chozinski et al., 2016; Rashpa and Brochet, 2022). Furthermore, U-ExM allowed insight into the asexual Plasmodium blood stage to be gained (Liffner and Absalon, 2021). The aim of the present study was to optimise the U-ExM protocol developed by Gambarotto et al. to study the Plasmodium liver stage (Gambarotto et al., 2019, 2021). With some major adaptations for pre-stained samples, we used this method to visualise different developmental stages of the Plasmodium liver stage. This fast and easy protocol will now allow researchers to gain a more detailed insight into the Plasmodium liver stage.

The aim of this study was to adapt the U-ExM protocol to investigate intracellular Plasmodium parasites during the liver stage development. The protocol developed by Gambarotto et al. in 2019 results in a near-native expansion of cultured cells (Gambarotto et al., 2019, 2021). In the original protocol, the antibody staining was performed after embedding the cells in a swellable gel, which resulted in a rather long antibody incubation time of ∼3 h. Furthermore, some epitopes were lost during the denaturation prior to the antibody staining. Additionally, the non-expanded controls had to be prepared separately. In this study, the original U-ExM protocol was modified to allow expansion of pre-stained specimens (pre-gelation staining ExM; PS-ExM). Overall, the reagents used for fixation (formaldehyde), anchoring (acrylamide), gelation and denaturation were identical to the ones previously described for U-ExM (Gambarotto et al., 2019, 2021). However, the incubation time for each of the steps was strongly reduced to better preserve the fluorescent signal. A comparison between the original and the adapted PS-ExM protocol version is shown in Table 1.

An overview of the PS-ExM protocol is depicted in Fig. 1A, and the detailed description can be found in the Materials and Methods section. An immunofluorescence assay was performed on infected HeLa cells. Primary and secondary antibody incubations were performed for 1 h. In comparison, in the original U-ExM protocol, each antibody incubation took 3 h (Table 1). The incubation time with formaldehyde and acrylamide was reduced to 2 h compared to the original U-ExM protocol, in which the incubation lasted for 5 h (Table 1). The addition of acrylamide results in functionalisation of the cells and subsequently allows to chemically link the proteins into a polyacrylamide gel matrix. Gelation was performed for 30 min in comparison to the 1.5 h in the original protocol (Table 1). For isotropic expansion, the cells were denatured for 30 min, which is three times shorter compared to the original protocol (Table 1). DNA of host cells and the parasites was stained with DAPI. After gel expansion, the cells were imaged. A major challenge in studying the Plasmodium liver stage in expansion microscopy is to be able to expand the intracellular parasite as well as the host cell under the same conditions. Representative images of non-expanded and expanded cells at 6 hours post-infection (hpi) are depicted in Fig. 1B. The parasitophorous vacuole membrane (PVM) was stained with anti-up-regulated in infective sporozoites gene 4 (UIS4) antibodies to identify infected HeLa cells. When comparing expanded and non-expanded cells, the morphological integrity of the PVM was preserved. The expansion factor measured was 5-fold for HeLa cell nuclei (Fig. 1C) and 5.1-fold for the parasite nuclei measured at 56 hpi (Fig. 1D). The PS-ExM protocol yielded isotropic expansion, as accessed through 3D sphericity measurements of parasite nuclei in both non-expanded and expanded samples (Fig. S1A,B). The isotropic expansion of the PS-ExM protocol was determined using the trophozoite developmental stage, as the parasite has a single nucleus with a spheric appearance at this stage (Roques et al., 2023). Furthermore, microtubule (α-tubulin) staining demonstrated the preservation of the cell morphology in the PS-ExM protocol (Fig. S1D). We compared the preservation of different epitopes between the original protocol and the adapted PS-ExM protocol – the epitopes of UIS4, lysosome-associated membrane protein 1 (LAMP1) and α-tubulin were adversely affected using the original protocol, whereas the PS-ExM protocol resulted in excellent preservation of epitopes (Fig. S1C,D).

During the Plasmodium liver stage, the parasite resides within a PV. The applicability of the PS-ExM was accessed for different phases of parasite liver stage development (Fig. 2A). The earliest time point was chosen as 6 hpi where only one single parasite nucleus is visible in the infected HeLa cell. The sporozoite, which initially has a banana-like shape, started becoming rounded at 24 hpi but still had a single nucleus. At 30 h post-infection, the parasite initiates the schizogony with massive replication, which is visible at 48 and 56 hpi as cells now show many parasite nuclei (Fig. 2A). Subsequently, the parasites undergo merogony and the PVM is ruptured. This time point was not included in this study given that the host cells detach from the coverslip at this phase. Besides being able to expand the whole Plasmodium liver stage, the applicability of the new PS-ExM to study host–pathogen interaction was tested using antibodies that detect proteins in the parasite and host cell. Previous studies have shown the association of lysosomes to the PV (da Silva et al., 2012; Niklaus et al., 2019). The interaction of lysosomes with the PVM was investigated in greater detail with PS-ExM. LAMP1, a lysosome membrane marker, was used to stain host-cell lysosomes and anti-UIS4 antibody staining was performed to outline the PVM. As previously described, lysosomes were found in close proximity to the PVM (Niklaus et al., 2019) (Fig. 2B). The lysosome interaction with the PVM was investigated at different time points of parasite development (24 hpi, 30 hpi and 48 hpi), as shown in Fig. S2A,B. By utilising Imaris software, we measured the fusion of lysosomes with the PVM, revealing that 25% of host lysosomes were attached to the PVM in non-expanded cells, whereas this number increased to 35% in expanded cells (Fig. S2C,D). We hypothesise that the improved staining achieved with the PS-ExM protocol contributes to this observed difference. In the PS-ExM samples, lysosomes appeared as vesicle-like structures, allowing for more accurate identification and quantification of lysosomes on the PVM compared to non-expanded controls, where lysosomes appeared as dots. The vesicle-like morphology of lysosomes provided valuable insights into the interaction between host-cell lysosomes and the PVM (Fig. 3). We observed that the lysosome sizes varied at the early time point of 6 hpi (Fig. 3A). The host cells might attack the pathogen by engulfing parts of the parasite PVM into lysosomes for digestion. Imaging showed that some large lysosomes contain PVM structures (Fig. 3A, ROI1 and ROI2). At the contact sides of lysosomes with the PVM, we observed a weaker UIS4 signal, suggesting fusion of the two membranes and their components (Fig. 3A, ROI3). Once the parasite became established within the host cell at 48 h post-infection, we observed a diffuse LAMP1 signal on the PVM (Fig. 3B, ROI1), again suggesting lysosome fusion with the PVM as previously described (Niklaus et al., 2019). Furthermore, we observed lysosomes that were attached or fused to the PVM (Fig. 3C, ROI1–ROI3). The fusion of the host lysosomes with the PVM resulted in a weaker UIS4 signal at the attachment point (Fig. 3C, ROI3) as observed at earlier time points of infection (Fig. 3A, ROI3).

To further evaluate the utility of the novel PS-ExM protocol in enhancing our understanding of parasite biology, we examined the preservation of mitochondria following expansion (Fig. 4). Plasmodium mitochondria were labelled with antibodies against Toxoplasma gondii heat shock protein 70 (TgHSP70), which also cross-reacts with P. berghei mitochondrial heat shock protein 70 (Eickel et al., 2013). We investigated mitochondrial development during the Plasmodium liver stage at various time points post-infection (6 hpi, 30 hpi, 48 hpi and 56 hpi). Remarkably, the expansion process did not compromise the structural integrity of the mitochondrial network of the parasite (Fig. 4; Movie 1). Utilising the TgHSP70 signal, we generated a 3D model of the expanded mitochondrial network using Imaris software (Fig. 4). The mitochondria of the parasite could be seen to undergo extensive branching following hepatocyte invasion, ultimately distributing within the forming merozoitesm as previously described (Stanway et al., 2011).

In this study, U-ExM was used for the first time to investigate Plasmodium liver stage parasites. The protocol originally developed by Gambarotto et al. was adapted to be used on pre-stained Plasmodium-infected HeLa cells (Gambarotto et al., 2019, 2021). Although we used the same chemical approach as in the U-ExM protocol, we performed a pre-expansion staining ExM (PS-ExM). Importantly, the time required for the U-ExM protocol was approximately halved from 1 to 2 days of work. In the original protocol developed by Gambarotto et al., the antibody incubation times were rather long, with 3 hs per incubation (Gambarotto et al., 2021). Performing the staining on cells prior to gelation allows for reducing the incubation time to 1 hour. Furthermore, non-expanded controls can be performed simultaneously using the same staining protocol as for the expanded samples.

With the new PS-ExM protocol, both the host cell and the developing parasite exhibited an isotropic 5-fold expansion throughout the entire Plasmodium liver stage (Fig. 1). We evaluated the effectiveness of the PS-ExM protocol in preserving epitopes by staining the parasite vacuole membrane (UIS4), host lysosomes (LAMP1), and parasite mitochondria (TgHSP70). In comparison to the original protocol, the PS-ExM protocol demonstrated superior preservation of the UIS4 and LAMP1 epitopes (Fig. S1C,D). Additionally, the structural integrity of the mitochondrial networks of the parasite was maintained throughout the liver stage expansion, as depicted in Fig. 4. In summary, the PS-ExM protocol provides excellent preservation of parasite mitochondria, the PVM, and host lysosomes, making it a valuable tool for in-depth organelle studies. By using α-tubulin to delineate single-cell borders for quantifying lysosome attachment to the PVM, we also observed that the PS-ExM protocol better preserved microtubules compared to the original protocol (Fig. S1D). These results suggest that, as compared to the original protocol, the PS-ExM maintains epitopes more effectively. This might be due to epitope loss during the denaturation time in the U-ExM protocol, which has been previously observed for some epitopes (Chozinski et al., 2016; Gao et al., 2017; Tillberg et al., 2016). By performing the antibody staining at the beginning of the protocol, this problem was solved. Additionally, as previously described, the antibodies themselves are linked to the swellable hydrogel during the acrylamide treatment potentially enhancing the antibody signal (Gao et al., 2017).

Additionally, PS-ExM was used to investigate the host–pathogen interaction of lysosomes on the PVM. The lysosomes were seen as vesicle-like structures, and further details of the interaction between lysosomes and the parasite PVM were observed using the PS-ExM protocol (Figs 2 and 3). Therefore, this novel protocol might be a helpful tool for further research of the molecular processes underlying the fusion of lysosomes to the parasite PVM.

However, expanding and imaging whole HeLa cells for quantitative purposes results in huge data sets (after deconvolution, a single image was ∼10 GB for wide-field images, and for every experiment, at least ten images were acquired). Here, data storage and processing must be considered. Taken together, the PS-ExM protocol is faster compared to the original protocol, and it might be better suited for some epitopes that are otherwise lost during denaturation. The protocol here described can be further used to study host–pathogen interactions and gain further insight into the Plasmodium liver stage.

Mice were obtained from Harlan Laboratories or Charles River. All mice (C57BL/6 and BALB/c) were maintained and bred in the central animal facility of the University of Bern. Studies were strictly performed under the guidelines and laws of the Animal Research Ethics Committee of the Canton of Bern, Switzerland (Permit Number: BE98/19) and the University of Bern Animal Care and Use Committee Switzerland. The mice were between 6 and 26 weeks old.

Mosquitoes (Anopheles stephensi) were maintained inside stock cages in the insectarium at the Institute of Cell Biology of the University of Bern. Briefly, mosquitoes were fed with blood by placing a Petri dish closed with parafilm containing the blood upside-down onto the mosquito cage. Blood was kept warm with a small Erlenmeyer containing hot water that was put on top of the Petri dish. The mosquitoes were allowed to feed for 30 to 40 min before the Petri dish containing the blood was removed. At 2 days after the feed, a beaker with water and filter paper was added to the mosquito cage to collect the mosquito eggs. The next day, the eggs were collected, washed once with 70% ethanol, and twice with water before being added to a metal bowl with a drop of about 10 µl of NobilFluid Artemia (JBL 33801 Germany), where they developed into larvae. After 2 days in the metal bowl, the larvae were transferred to a water-containing plastic bowl and fed with grounded Tetra TabiMin complete food tablets (Olibetta, 400080, Germany). After 7 days, the first pupae were collected and put in a stock cage, where they were kept at 27°C and 80% humidity until hatching. Adult mosquitoes were fed with 8% fructose supplemented with 0.2% PABA (Sigma-Aldrich, Cat. No. 100536). After 2 days, the female A. stephensi mosquitoes were ready to be infected with P. berghei parasites.

A blood stabilate of the P. berghei ANKA strain expressing cytosolic mCherry under the Hsp70 promoter (Burda et al., 2015) was injected intraperitoneally into a naïve Balb/c mouse. When the mouse reached a parasitemia of 3 to 5%, 50 µl of its infected blood was diluted with 150 µl 1× phosphate-buffered saline (PBS) and injected intravenously into a new mouse that was intraperitoneally injected with 200 µl phenylhydrazine (6 mg ml⁻¹ in PBS, Sigma-Aldrich, 114715) 3 days prior to infection. After 3 days, 5 µl of blood was collected from the mouse tail and diluted in 500 µl of 1× PBS to check the parasitemia with fluorescence-activated cell sorting (FACS). When the mouse reached a parasitemia of 5% with at least 0.5% to 1% of gametocytes, the mouse was anaesthetised and used to feed 100–150 female Anopheles stephensi mosquitoes for 1 h. Infected mosquitoes were kept at 20.5°C and 80% humidity and fed with 8% fructose containing 0.2% PABA for 16 to 26 days for their salivary gland dissection and mammalian HeLa cell infection.

Human epithelial HeLa cells (European Cell Culture Collection) were maintained at 37°C with 5% CO₂. They were grown in minimum essential medium (MEM, BioConcept, 1-31F01-I) supplemented with Earle's salts and 10% heat-inactivated fetal calf serum (FCS, GE Healthcare) referred to as complete culture MEM. Medium contained 1% L-glutamine (Bioconcept, 5-10K00-H) and 1% penicillin-streptomycin (Bioconcept, 4-01F00-H. Cells were split every 4 days by treatment with accutase (Sigma-Aldrich, A6964). 40,000 HeLa cells were seeded on a coverslip in a 24-well plate (Greiner Bio-one, 662160) containing complete culture MEM. The next day, when the infected female A.s stephensi mosquitoes were between days 16 and 26 post-infection, the mosquitoes were anaesthetised with chloroform vapour (Merck, 67663 Germany) for ∼20 s. Anaesthetised mosquitoes were briefly dipped into 70% ethanol and then transferred to 1× PBS. The mosquitoes were then put into 200 μl of non-supplemented MEM medium, referred to as infection medium, on a slide, and the head was then meticulously removed using surgical forceps. The two sets of each of the three salivary gland lobes were isolated from the head and transferred into an Eppendorf tube containing 20 μl of infection medium. The salivary glands that had been dissected were kept on ice until infection. The sporozoite parasites were mechanically released from the salivary glands using a pestle (Sigma-Aldrich, Z359947) that was powered by a cordless motor (Sigma-Aldrich, Z359971) with 10 pulses of 1–2 s. To determine the number of sporozoites, 10 μl of the infection medium was introduced to a Neubauer chamber. The seeded HeLa cells were then infected by the addition of 20,000 sporozoites in 200 μl. The 24-well plate was spun at 1000 g for 1 min to allow the sporozoites to settle quickly before being incubated at 37°C with 5% CO₂ for 2 h. To get rid of mosquito salivary gland debris and sporozoites that failed to infect HeLa cells, cells were washed twice with 500 μl of pre-warmed complete culture medium. Following that, cells were then incubated with 1 ml of complete culture medium at 37°C with 5% CO₂ as previously described (Kaiser et al., 2017). The medium was exchanged with the cells at 24 and 48 hpi. At specific time points after infection (6 hpi, 24 hpi, 30 hpi, 48 hpi and 56 hpi), the cells were fixed and kept at 4°C for the immunofluorescence assay.

Infected HeLa cells were fixed with 4% paraformaldehyde (PFA) in PBS for 10 min at room temperature and subsequently washed twice with 1× PBS. Cells were permeabilised with 0.05% Triton X-100 (Fluka Chemie, T8787) for 5 min at room temperature, supplemented with 0.1% saponin (for washing out cytoplasmic background when staining lysosomes). Nonspecific antibody binding was reduced by blocking with 10% FCS in PBS for 20 min at room temperature. Subsequently, the cells were incubated with primary antibodies diluted in 10% FCS for 1 h at room temperature. The primary antibodies used were: anti-UIS4 rabbit (P-Sinnis, Baltimore, 1:1000 in PS-ExM and 1:500 in post-staining ExM), chicken anti-UIS4 (produced by Proteogenix in 2021, 1:1000), mouse anti-LAMP1 (Developmental Hybridoma Bank, clone H4A3, 1:1000 in PS-ExM and 1:500 in post-staining ExM), rabbit anti-TgHSP70 (gift from Dominique Soldati Favre, 1:500) and guinea pig anti-α-tubulin (Geneva Antibody Facility, AA345, 1:200 in PS-ExM and 1:125 in post-staining ExM) antibodies. The cells were then washed three times for 5 min each with 1× PBS and incubated with secondary antibodies at room temperature for 1 h, protected from light. The secondary antibodies used were: anti-rabbit-IgG conjugated to ATTO 647 (Sigma, 40839, 1:1000 in PS-ExM and 1: 500 in post-staining ExM), anti-mouse-IgG conjugated to Alexa Fluor 488 (Invitrogen Molecular Probes, A11001, 1:1000 in PS-ExM and 1:500 in post-staining ExM), anti-guinea pig-IgG conjugated to Alexa Fluor 594 (Invitrogen Molecular Probes, A11076, 1:1000 in PS-ExM and 1:500 in post-staining-ExM) and anti-chicken-IgY conjugated to Alexa Fluor 594 (Invitrogen Molecular Probes, A11042, 1: 500). All antibodies are additionally listed in Table S1. For non-expanded samples, the coverslips were stained with DAPI (Sigma, D9542, 1:100) for 5 min before being mounted with 5 μl ProLong™ Gold Antifade Mountant (Invitrogen, P36930). Images were acquired as described below.

The chemical approach used was based on the U-ExM protocol published by Gambarotto et al. (Gambarotto et al., 2019, 2021). The protocol was optimised to make it applicable for pre-gelation-stained samples. The incubation times were shortened to better preserve the fluorophores of the pre-stained sample. Infected HeLa cells were stained as described above and samples were kept protected from light until imaging. Cells were incubated in PBS containing 0.7% formaldehyde (FA, 36.5–38%, F8775, Sigma) and 1% acrylamide (AA, 40%, A4058, Sigma). The latter will link the cells to the polymer during gelation. Compared to the original U-ExM protocol (5 h incubation), the formaldehyde and acrylamide incubation time was shortened to 1.5 h at 37°C. Subsequently, gelation was performed as described by Gambarotto et al. Gelation solution was prepared freshly: 0.5% ammonium persulfate (APS, 17874, Thermo Fisher Scientific) and 0.5% tetramethylethylenediamine (TEMED, 17919, Thermo Fisher Scientific) were added to the monomer solution containing 19% sodium acrylate (SA, 97–99%, 408220, SIGMA), 10% acrylamide (AA, 40%, A4058, Sigma) and 0.1% N,N′-methylenebisacylamide (BIS, 2%, M1533, Sigma). Per cover slip, 35 µl gelation solution was used. After incubating the sample on ice for 5 min, the gelation was carried out for 30 min at 37°C. Subsequently, gels were incubated in 1 ml denaturation buffer (200 mM SDS, 200 mM NaCl and 50 mM Tris-HCl in deionised water, pH 9) with gentle agitation at room temperature for 15 min to detach the gel from the coverslip. Denaturation was carried out at 95°C in denaturation buffer for 30 min. Gels were briefly incubated in PBS and afterwards, the DNA was stained with 5 µg ml⁻¹ DAPI (D9542-5MG, Sigma) diluted in 2% bovine serum albumin (BSA) at 37°C with gentle agitation for 1 h. Samples were expanded in deionised water for 1 h. If the sample was imaged a day after expansion, the gel was incubated in 0.2% propyl gallate (02370, Sigma) to avoid photobleaching of samples until imaging. The chemicals are additionally listed in Table S2.

After the final expansion in deionised water, the gels were cut and mounted on poly-D-lysine (A38904, Gibco) functionalised 35 mm glass-bottom dishes (D35-20-1.5-N, Cellvis, 35 mm glass bottom dish with 20 mm micro-well #1.5 cover glass). These closed sample holders strongly reduced evaporation and shrinkage of the gels. All images were acquired with a 60× oil objective (NA=1.4). The NIKON Ti 2 CREST V3 equipped with a Hamamatsu Flash 4.0 camera and a celesta light engine was used in widefield or spinning disc. Images were acquired with a Z-step size of 300 nm and pixel size of 108 nm. Images acquired in wide-field mode were deconvolved with Huygens Professional version (Huygens Remote Manager v3.8; Scientific Volume Imaging, The Netherlands, http://svi.nl).

The open-source ImageJ platform was used to measure cell nuclei to determine the expansion factor (Schindelin et al., 2012). To measure parasite expansion, parasite nuclei were measured at 56 hpi. In comparison, a minimal HeLa cell diameter was determined at different time points after infection. The expansion factor was calculated by dividing the average nuclear diameter of expanded cells by the average nuclear diameter of non-expanded samples. The 3D and 4D image analysis software Imaris (Imaris X64, 9.9.0; Mar 11, 2022) was used to segment parasite mitochondria and nuclei (Fig. 3) and the lysosome attachment to the parasite PVM (Fig. S2). Briefly, to measure the number of lysosomes attached to the parasite PVM in infected HeLa cells, the PVM (stained with anti-UIS4) was used to generate the isosurface, and the lysosomal vesicles (stained with LAMP1) were used to generate isospots. The generation of the isosurface from the anti-UIS4 staining was carried out using an isosurface detail value of 0.215 for non-expanded and expanded PVM in conjunction with the absolute intensity. The lysosome isospots were derived in the ‘local contrast’ mode using the estimated point spread function (PSF). For non-expanded lysosomes, an estimated XY-diameter of 0.4 μm and an estimated Z-diameter of 0.5 μm were used, whereas for expanded lysosomes, an estimated XY-diameter of 1.5 μm and 2.4 μm of estimated Z-diameter were used. The isospots attached to the PVM were classified using the ‘Shortest Distance to Surfaces’ parameter, with the threshold set at 0 μm distance to the PVM. The isospots within infected HeLa cells were manually selected based on anti-α-tubulin staining, which was used to delimitate the single-cell border. The isospots of non-infected HeLa cells were excluded from the analysis. Isospot numbers as well as the shortest distance to surfaces and isosurface data were exported into a Microsoft Excel file. The proportion of LAMP1 isospots attached to the PVM was calculated by dividing the number of LAMP1 isospots attached to the PVM by the total number of LAMP1 isospots within the infected cell. The Imaris software was also used to compute the isosurface of parasite nuclei in acquired Z-stack images in a similar manner to the PVM isosurface details to determine isotropic expansion by measuring the sphericity of the nuclei in non-expanded and expanded samples. The data were structured and analysed using GraphPad Prism version 9.0.0 for Mac, GraphPad Software, San Diego, California, USA (www.graphpad.com). Statistical analysis was performed using a two-tailed unpaired Student's t-test and are denoted as ***P<0.001; **P<0.01; *P<0.05.

We are grateful to Dominique Soldati for the anti-TgHSP70 antiserum. The authors acknowledge Ruth Rehmann, Ado Crnovrsanin, Clirim Jetishi and Dr Sandro Käser for discussions. The authors gratefully acknowledge Christin Berger for her assistance in revising the manuscript. Microscopy was performed on equipment supported by the Microscopy Imaging Center (MIC) of the University of Bern, Switzerland. K.A. acknowledges the Swiss Confederation for the awarded Government Excellence Scholarship.