Erythrocyte Membrane Protein 3 (EMAP3) Is Exposed on the Surface of the Plasmodium berghei Infected Red Blood Cell
Sophia Raine C. Hernandez, Ravish Rashpa, Thorey K. Jonsdottir, Martina S. Paoletta, Josy ter Beek, María Rayón Díaz, Jelte M. M. Krol, Severine Chevalley‐Maurel, Takahiro Ishizaki, Ronnie P.‐A. Berntsson, Chris J. Janse, Blandine Franke‐Fayard, Mathieu Brochet
TL;DR
A new protein called EMAP3 is found on the surface of malaria-infected red blood cells, offering a new way to study how the parasite sticks to blood vessels.
Contribution
EMAP3 is a novel Plasmodium berghei protein exposed on the surface of infected red blood cells, providing a new platform for studying cytoadherence.
Findings
EMAP3 is trafficked to the outer membrane surface of Plasmodium berghei infected red blood cells.
EMAP3 is not essential for parasite growth or sequestration but can display Plasmodium falciparum proteins on infected red blood cells.
Abstract
The human malaria parasite Plasmodium falciparum invades red blood cells (RBCs) and exports parasite proteins to transform the host cell for its survival. These exported proteins facilitate cytoadherence of the infected RBC (iRBC) to endothelial cells of small blood vessels, protecting iRBCs from splenic clearance. The parasite protein PfEMP1 and the host protein CD36 play a major role in P. falciparum iRBC cytoadherence. The murine parasite Plasmodium berghei is a widely used experimental model that combines high genetic tractability with access to in vivo studies. The P. berghei iRBC also sequesters by CD36‐binding via an unknown parasite ligand and few parasite proteins, including EMAP1 and EMAP2, have been localised to the iRBC membrane. We have identified a new protein named EMAP3 and demonstrated its export to the iRBC membrane where it likely interacts with EMAP1, with only EMAP3…
Genes, proteins, chemicals, diseases, species, mutations and cell lines named across the full text — each resolved to its canonical identifier and authoritative record.
Click any figure to enlarge with its caption.
FIGURE 1
FIGURE 2
FIGURE 3
FIGURE 4
FIGURE 5- —Vetenskapsrådet10.13039/501100004359
- —Knut och Alice Wallenbergs Stiftelse10.13039/501100004063
- —Swiss National Science Foundation10.13039/501100001711
- —Japan Society for Promotion of Science (JSPS)
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
Taxonomy
TopicsMalaria Research and Control · Hereditary Neurological Disorders · Trypanosoma species research and implications
Introduction
1
The etiological agents of malaria are unicellular Plasmodium parasites that cause disease by invading and replicating within red blood cells (RBCs), which ultimately leads to the destruction of the infected cells (Bannister and Mitchell 2003). Upon invasion of the RBC, the parasite proliferates within the protective parasitophorous vacuole and remodels the RBC to support its rapid growth by exporting a large range of parasite proteins across the parasitophorous vacuole membrane (PVM) into the infected RBC (iRBC) (Marti et al. 2004). A number of malaria parasite species, including the human parasite Plasmodium falciparum, export proteins onto the surface of the iRBC that facilitate cytoadherence to the vascular endothelium of small blood vessels and thereby evasion of clearance of the iRBC by the spleen. This causes parasites to accumulate in blood vessels of inner organs and contributes to disease pathology through blockage of blood capillaries, vascular leakage and endothelial inflammation, which correlates with severe malaria infections by P. falciparum (Lee et al. 2019). In P. falciparum infections, endothelial cytoadherence and the resulting organ sequestration are mediated by the parasite protein PfEMP1 (erythrocyte membrane protein 1), which is encoded by the var multigene family. This protein is the major parasite surface ligand binding the host protein CD36 that is expressed on the vascular endothelium. In addition, it binds to the host proteins iCAM‐1 (intercellular adhesion molecule 1), as well as to CSA (chondroitin sulphate A) in the placenta of pregnant women (Craig and Scherf 2001). The identification of new therapies blocking cytoadherence has the potential of being translated into much needed life‐saving interventions that can be deployed to interrupt the rapid and often fatal disease progression characterising P. falciparum severe malaria. Experimentally studying P. falciparum cytoadherence and sequestration in vivo is challenging due to the lack of appropriate whole organism models.
Plasmodium berghei is a widely used murine experimental model that benefits from high genetic tractability and facilitates in vivo studies. The P. berghei iRBC also sequesters in small blood vessels, mediated by binding to CD36 (Franke‐Fayard et al. 2005). Importantly, P. berghei infections can be monitored by whole body imaging, where both the parasite and host can be manipulated and the effect on sequestration and virulence can be directly assessed (Franke‐Fayard et al. 2005; Fonager et al. 2012). Despite being a commonly used model, the molecular mediators of P. berghei sequestration are not fully understood. There is no orthologue of PfEMP1 in P. berghei (Hall et al. 2005). However, many fundamental aspects of parasite biology and virulence are conserved between P. falciparum and P. berghei. Significantly, P. berghei also exports an array of parasite proteins into the infected RBC, including proteins encoded by multigene families (Pasini et al. 2013; Siau et al. 2023). The molecular machinery for protein export across the PVM is also shared between P. falciparum and P. berghei, and experimental deletion of several of these conserved proteins in P. berghei results in reduced blood‐stage growth and sequestration in vivo (de Koning‐Ward et al. 2009; Matthews et al. 2013; Matz et al. 2013; Elsworth et al. 2014; Chisholm et al. 2016; Batinovic et al. 2017). In addition, P. berghei induces the generation of membranous structures in the cytoplasm of iRBC, termed intra‐erythrocytic P. berghei‐induced structures (IBIS), akin to Maurer's clefts in P. falciparum that mediate trafficking of proteins to the surface of the infected cell (Ingmundson et al. 2012). Importantly, some of the key molecular players of these membrane structures are also shared between P. falciparum and P. berghei (Blisnick et al. 2000; Spycher et al. 2003; Petersen et al. 2015; De Niz et al. 2016).
None of the P. berghei exported proteins that are known to play a role in CD36‐mediated sequestration are presented on the surface of the iRBC. Instead, they are located in the iRBC cytoplasm and therefore likely play a role in trafficking CD36‐binding ligand(s) to the surface (Fonager et al. 2012; De Niz et al. 2016; Gabelich et al. 2022). For example, deletion of the P. berghei gene encoding SMAC (schizont membrane‐associated cytoadherence protein) results in slow growth of blood stage parasites in vivo and impaired CD36‐mediated sequestration of iRBCs. SMAC is exported into the iRBC but does not localise to the iRBC membrane (Fonager et al. 2012). In addition, exported proteins encoded by several different multigene families, including the large pir (Plasmodium interspersed repeat) family, are located in the PV or in the cytoplasm of the iRBC with no clear evidence for an iRBC surface localisation, neither in P. berghei (Fougère et al. 2016) nor in another rodent malaria parasite Plasmodium chabaudi (Fougère et al. 2016; Giorgalli et al. 2022). The P. berghei proteins EMAP1 and EMAP2 (erythrocyte membrane associated protein 1/2) are exported to the iRBC membrane. Although these two proteins have been identified based on their absence in iRBCs of a non‐sequestering laboratory strain of P. berghei (K173), they do not play a critical role in growth of blood stage parasites or iRBC sequestration and thus are not P. berghei surface ligands that mediate binding to CD36 (Pasini et al. 2013).
PfEMP1 binds CD36 via its extracellular cysteine‐rich interdomain region (CIDR) domain. There are no known P. berghei proteins with CIDR domains (Fonager et al. 2012). We therefore set out to identify by bioinformatic analyses P. berghei genes encoding novel exported proteins that are putatively expressed on the surface of the iRBC and could mediate iRBC cytoadherence. These analyses resulted in the identification of a single copy gene coding for a small four‐transmembrane‐domain protein that we named EMAP3. We demonstrate that in the sequestering schizont stage EMAP3 interacts with EMAP1, where EMAP3 is inserted into the iRBC membrane and displays its C‐terminal domain on the outer side of the iRBC membrane. Deletion of the gene encoding EMAP3 does not significantly affect asexual blood stage growth nor iRBC sequestration or disease characteristics. We conclude that EMAP3 is exposed on the outer surface of the iRBC, but it is not the major ligand for CD36‐mediated sequestration of P. berghei iRBCs. Nevertheless, its extracellular exposure opens the possibility to use EMAP3 as a scaffold to display P. falciparum proteins on the surface of the P. berghei iRBC for screening in vivo putative inhibitors of P. falciparum cytoadherence.
Results
2
Bioinformatic Analysis Identifies EMAP3, a Novel Putatively Exported Protein
2.1
To identify Plasmodium exported proteins, we applied previously reported selection criteria for identifying putative P. berghei blood‐stage surface proteins. This selection was based on existing data for transcription and/or protein expression in P. berghei blood‐stages and the presence of protein export motifs as previously described (Pasini et al. 2013; Fougère et al. 2016). By refining these selection criteria for expression in both P. berghei blood‐ and liver‐stage parasites (Caldelari et al. 2019) and annotation as conserved Plasmodium protein, with unknown function we identified emap3 (PBANKA_0825900). The emap3 gene encodes a small protein (352 amino acids) that lacks functional domain prediction and is conserved among non‐laveranian human, simian, avian and rodent malaria parasites. Bioinformatic analysis indicates that it lacks a signal peptide but has four transmembrane domains at its N‐terminus (Figure 1A and Figure S1C) (Tsirigos et al. 2015). The N‐terminal positioning of the four transmembrane domains is preserved in syntenic orthologues of the human malaria parasites Plasmodium vivax and Plasmodium knowlesi (Figure 1A), as well as Plasmodium ovale (https://plasmodb.org/), (Alvarez‐Jarreta et al. 2023). Structural modelling using AlphaFold 3 also confidently predicts a four transmembrane helical bundle connected by short loops (pTM score = 0.79 for amino acids residues 1–130). However, no reliable model could be predicted for the soluble C‐terminal half of the protein (pTM score = 0.11), (Figure S1C) (Abramson et al. 2024).
EMAP3 is a novel exported protein in Plasmodium. (A) Schematic presentation of EMAP3 in P. berghei, P. knowlesi and P. vivax with the amino acid (aa) positions of the 4 transmembrane helices. (B) Schematic presentation of genes encoding EMAP1, EMAP2, EMAP3 and SMAC with a C‐terminal triple‐myc tag (referred to as myc) of four different transgenic P. berghei parasite lines. Plasmodium export element (PEXEL), transmembrane helix (TM) and signal peptide (SP). Molecular weight of proteins given in kilodalton (kDa), excluding myc‐tag. (C) Western blot analysis of 3×‐myc tagged lines to confirm 3×‐myc tag expression and detection of each protein at the expected size. A rabbit α‐myc antibody was used to detect the tagged proteins and a rabbit α‐Pfhsp70 antibody was used as a loading control. Full length blots can be found in Figure S2.
In agreement with earlier work (Pasini et al. 2013), our analysis indicates that emap1 (PBANKA_0836800) and emap2 (PBANKA_0316800) both have a signal peptide at their N‐terminus but lack transmembrane domains. In addition, emap2 also carries a classical Plasmodium export element/host targeting motif (PEXEL/HT) (Hiller et al. 2004; Marti et al. 2004). Out of the three P. berghei emap genes, only emap1 has a known functional domain, a pyst‐A domain that has been implicated in lipid binding and transport (Frech and Chen 2013; Pasini et al. 2013) (Figure 1B). Structural modelling of EMAP1 and EMAP2 individually gave a reliable prediction only for the known emap1 pyst‐A domain (Figure S1A,B). This pyst‐A domain was found to contain an α/β helix‐grip structure, as described for steroidogenic acute regulatory protein‐related lipid transfer (START) domains (Frech and Chen 2013; Clark 2020; Abramson et al. 2024; van Kempen et al. 2024). Furthermore, the hydrophobic pocket within EMAP1 could be confidently modelled bound to both cholesterol and phosphatidylcholine with ipTM scores above 0.8 (Figure S3A,B) (Abramson et al. 2024; van Kempen et al. 2024). One additional gene encoding the known exported P. berghei protein, smac (PBANKA_0100600), was also revisited in this study. The smac gene contains a PEXEL/HT motif, an N‐terminal signal peptide, and two centrally located transmembrane domains predicted by structure and sequence (Figure 1B, Figure S1D) (Tsirigos et al. 2015; Teufel et al. 2022; Abramson et al. 2024). We note that some AlphaFold models of SMAC generated a single TM helix as previously reported (Fonager et al. 2012), however, this single helix is 7 nm long, which exceeds the thickness of the lipid bilayer, which is approximately 5 nm with a 3 nm hydrophobic core.
EMAP3 Is Exported Onto the iRBC Surface During Asexual Blood Stage Development
2.2
To determine the subcellular location of EMAP3, we generated a P. berghei mutant with a C‐terminal triple‐myc (referred to as myc) tag in the endogenous emap3 locus by standard genetic modification technology (Janse, Ramesar, and Waters 2006). In addition, we generated P. berghei mutants with triple‐myc tagged EMAP1 or EMAP2 proteins that are known to be exported to the iRBC membrane or SMAC, which is known to be involved in P. berghei iRBC sequestration (Fonager et al. 2012; Pasini et al. 2013). Correct myc tagging was confirmed by diagnostic genotyping PCR (Figure S4A–D). Expression of the tagged proteins in blood stage parasites (schizonts) was verified by Western blot and immunofluorescence assays (IFA) detecting the myc tagged proteins (Figures 1C and 2A). In schizonts, EMAP3‐myc and EMAP1‐myc show a location largely concentrated to the iRBC membrane, while EMAP2‐myc displayed a more diffuse localisation throughout the iRBC cytoplasm. SMAC‐myc was also concentrated at the iRBC periphery but, in contrast to EMAP3 and EMAP1, it is present in distinct foci seemingly lining the inside of the iRBC (Figure 2A).
The C‐terminal of EMAP3 is exposed on the surface of P. berghei iRBC during the schizont stage. (A) Immunofluorescence assays (IFAs) of schizonts expressing myc tagged EMAP1, EMAP2 (left panel), EMAP3 and SMAC (right panel). Images are captured using a widefield microscope. (B, C) Time course IFA of (B) EMAP3‐myc and (C) EMAP1‐myc expression in different blood stages (ring, early trophozoite, late trophozoite, schizont). Images are captured using a confocal microscope. (D, E) The effect of permeabilisation using Triton X‐100 and trypsin treatment on the detection of (D) EMAP3‐myc and (E) EMAP1‐myc in P. berghei schizonts by IFA. Images captured with a confocal microscope. In all microscopy panels the myc‐tagged protein is visualised by staining with an anti‐myc antibody (cyan) and the red blood cell membrane with an anti‐TER‐119 antibody (red). DNA is labelled by 4′,6‐diamidino‐2‐phenylindole (DAPI), (yellow). Scale bar = 5 μm.
To further investigate the location and timing of expression of EMAP3 we took samples from cultures containing roughly synchronised blood stages (ring, early trophozoite, late trophozoite and schizont) and performed IFAs to detect EMAP3‐myc and EMAP1‐myc. EMAP3 is expressed already in ring forms, where it appears closely associated with the nucleus (likely still residing in or close to the endoplasmic reticulum, ER). Already in early trophozoites EMAP3‐myc is found at the iRBC membrane and remains located at the iRBC membrane in schizonts (Figure 2B). The active export of EMAP3 and the specificity of the EMAP3‐myc signal at the iRBC membrane was verified by treatment with brefeldin A (BFA), which traps parasite proteins during transport between the ER and Golgi apparatus (Crary and Haldar 1992). As expected, BFA treatment inhibited export of EMAP3‐myc and abolished EMAP3‐myc iRBC membrane staining in mature parasite stages (schizonts/late trophozoites). In younger parasite stages (rings), EMAP3‐myc accumulated in close proximity to the parasite ER (visualised by co‐staining of the ER‐resident protein BiP (binding immunoglobulin protein)), in both control and BFA treated parasites (Figure S5A,B). EMAP3‐myc expression and location closely mirrors that of EMAP1‐myc, which is also exported to the iRBC membrane in trophozoites and schizonts. A notable difference is that EMAP1‐myc is not detectable in ring forms, which is in agreement with what has previously been reported for EMAP1 (Pasini et al. 2013). In a subset of mature schizonts, EMAP1‐myc is also concentrated to small foci adjacent to the nuclei of the merozoites developing within (Figure 2C), in a pattern that suggests localisation to apical organelles such as the rhoptries (Baldi et al. 2000).
EMAP3 is predicted to have four transmembrane domains at its N‐terminus, and a soluble C‐terminal domain roughly half the size of the protein (Figure 1A). This C‐terminal part of EMAP3 lacks homology to known domains and its structure cannot be predicted (Figure S1C), which might indicate it will only fold into its functional conformation when interacting with other proteins. We hypothesised that the C‐terminus of EMAP3 contains functional domain(s) that may play a role in interacting with host cell proteins outside the iRBC when exposed on the surface of the iRBC. We therefore set out to identify how the protein is oriented on the iRBC membrane through surface shaving using trypsin and permeabilisation with Triton X‐100. These differential permeabilisation experiments were performed by first treating the cells with or without trypsin, followed by permeabilisation with Triton X‐100 or a no detergent control. EMAP3‐myc is detectable without prior permeabilisation, demonstrating that the C‐terminal end of the protein containing the myc tag is exposed on the iRBC surface. This is supported by the observation that the EMAP3‐myc signal is lost upon treatment with trypsin, irrespective of permeabilisation (Figure 2D). In contrast, EMAP1‐myc could only be detected when the cell is permeabilised with Triton X‐100 and the EMAP1‐myc signal is only lost when treated with Triton X‐100 prior to trypsin treatment (Figure 2E), indicating that the C‐terminus of EMAP1 is not exposed on the surface of the iRBC. The lack of EMAP1‐myc signal in the trypsin treated but non‐permeabilised sample is reflecting an inability of the antibody to access EMAP1‐myc in absence of Triton X‐100, not a loss of EMAP1‐myc signal from the trypsin treatment (Figure 2E). A control experiment wherein the cell is first permeabilised, followed by treatment with trypsin, shows loss of EMAP1‐myc signal upon trypsin treatment (Figure S5C,D). Despite its association with the iRBC membrane, EMAP1 lacks predicted transmembrane domains (Figure 1A) and the orientation of the protein at the iRBC membrane was not further studied.
To examine the location of EMAP1, EMAP3 and SMAC more closely, we performed ultrastructure expansion microscopy (U‐ExM). To analyse the orientation and location at the iRBC membrane we stained the iRBC with BODIPY TR ceramide to visualise membranes and/or N‐hydroxysuccinimide (NHS)‐ester that binds to protein dense regions. U‐ExM analysis confirms that EMAP3‐myc predominantly localises to the iRBC membrane in mature schizonts, where it co‐localised with the BODIPY‐stained RBC membrane (Figure 3A). EMAP1‐myc could also be observed in close proximity to the iRBC membrane. However, the EMAP1‐myc staining in U‐ExM was both more intense and more widespread compared to EMAP3‐myc. EMAP1‐myc was observable more widely across the host cell cytoplasm in the U‐ExM images (Figure 3B) compared to the conventional IFA images (Figure 2A,C). Nevertheless, the weaker EMAP3 signal observed in the U‐ExM analysis, compared to EMAP1, may affect the detection of EMAP3 also present elsewhere in the iRBC. This potentially influences direct comparisons of distribution between the two different proteins. Interestingly, in mature schizonts EMAP1‐myc appears to also localise to the apical organelles, where it co‐localises with the bright foci of NHS‐ester that stains the protein‐rich rhoptry organelles (Figure 3B, upper panel) (Liffner and Absalon 2021). This correlates with an apical, punctuated pattern as observed for schizonts using IFA (Figure 2C, Figure S6A). Co‐staining with the rhoptry marker RAP1 using IFA and U‐ExM shows some evidence of co‐localisation of EMAP1‐myc with RAP1 (Figure S6B). However, the RAP1 signal was not optimal in the conventional IFA and in the U‐ExM bleed‐through of the RAP1 signal could not be excluded. Nevertheless, in absence of co‐staining with RAP1 the apical localisation is still preserved, supporting the rhoptry localisation of the protein (Figure S6B). In addition, in agreement with IFA results, SMAC‐myc was concentrated in distinct patches, localising to the periphery of the iRBC during the schizont stage (Figure 3C). Taken together, these analyses show that EMAP3 is exported into the iRBC where we hypothesise that it is inserted into the iRBC membrane as a multi‐pass transmembrane protein with its C‐terminus exposed on the surface of the iRBC.
Ultrastructure expansion microscopy showing EMAP3 localisation to the iRBC membrane during the schizont stage. Ultrastructure expansion microscopy (U‐ExM) of (A) EMAP3‐myc and (B) EMAP1‐myc in schizonts with the c‐myc tagged proteins in cyan with BODIPY TR ceramide to visualise membranes (magenta) and NHS‐ester marking protein dense regions (yellow). The upper (with NHS‐ester) and lower panels (without NHS‐ester and with a zoomed‐in panel denoted by dashed white square) show two different representative cells for EMAP3‐myc (A) and EMAP1‐myc (B). The protein‐rich foci brightly stained by NHS‐ester at the apical end of the merozoites packed in mature schizonts marks the apical rhoptry organelles (upper panels A, B). (C) U‐ExM of SMAC‐myc schizonts with c‐myc tagged proteins in cyan and NHS‐ester stain in yellow. Left and right panels show two different representative cells for SMAC‐myc (C). Scale bar = 10 μm or 3 μm. The images are maximum intensity Z‐projections of three to five slices and captured with a confocal microscope.
EMAP3 Interacts With EMAP1 but Not With SMAC at the iRBC Membrane
2.3
To determine the putative parasite and host proteins that interact with the EMAPs and SMAC and to identify novel exported proteins we performed cross‐linking co‐immunoprecipitation (Co‐IP) assays using antibodies against the myc tagged proteins. We performed Co‐IPs of iRBCs containing schizonts without and with saponin treatment of the iRBC prior to cross‐linking. Proteins that are soluble in the iRBC will be lost with saponin treatment but those that are bound to the iRBC membrane or cytoskeleton will be retained. The saponin treatment is also expected to remove soluble exported proteins that are not strongly interacting with the myc‐tagged bait proteins. The control samples used were schizont preparations obtained from cultures of PbCas9‐FLAG_Tir1‐myc mutant expressing the unrelated and non‐exported, triple‐myc tagged Tir1 protein (transport inhibitor response 1), referred to as Tir1‐myc. Correct insertion of the Tir1‐myc expression cassette was confirmed by diagnostic PCR‐genotyping (Figure S4F) and expression of the myc tagged Tir1 protein in blood stages was analysed by Western blot analysis (Figure S4G).
Western blot analysis of the Co‐IP samples prior to mass spectrometry analysis confirmed the presence of the myc‐tagged EMAPs and SMAC bait proteins with the expected sizes (Figure 4A, Figure S7). A list of putative interacting proteins was generated for each myc‐tagged bait protein and visualised in heatmaps. Only proteins that were not present in the Tir1‐myc control, or if detected were present with at least four times the peptide abundance compared to the control in at least two Co‐IP replicates (under the same conditions), were included (Tables S1–S4, Figure 4B). EMAP3‐myc robustly co‐immunoprecipitates EMAP1 in all biological replicates, with a stronger interaction observed for the samples without saponin lysis. EMAP3‐myc also co‐immunoprecipitates EMAP2, but this interaction is detected less robustly (only in two out of three replicas for non‐saponin treated samples only). In contrast, there was no evidence of EMAP3‐myc interacting with SMAC (Tables S1–S4, Figure 4B).
EMAP3 interacts with EMAP1 at the iRBC membrane but not with SMAC. (A) Western blot analysis of co‐IP samples (eluate) for bait proteins EMAP1‐myc, EMAP2‐myc, EMAP3‐myc and SMAC‐myc. (B) Heatmap summary of results from co‐immunoprecipitation (Co‐IP) using EMAP1‐myc, EMAP2‐myc, EMAP3‐myc or SMAC‐myc as bait proteins and analysed by nanoscale liquid chromatography coupled to tandem mass spectrometry (nano LC‐MS/MS). Co‐IPs were performed with (+) and without (−) saponin treatment followed by formaldehyde cross‐linking. Columns depict spectral counts for proteins co‐immunoprecipitated in each IP replicate with and without saponin, with the corresponding bait protein labelled above. Proteins shown are those absent in the Tir1‐myc control or if present in the Tir1‐myc control, occur at a four‐fold abundance to the control. Spectral counts in the Tir1‐myc control are shown in the last two columns of the heat map. Proteins annotated include bait proteins, those that are predicted to interact with the EMAPs and SMAC, and proteins uniquely present in the EMAP3 IPs. (C) Schematic summary of selected shared interactions between the EMAPs and SMAC at the iRBC membrane, where only proteins detected in a minimum of two replicas are shown as solid lines indicating their interaction. EMAP3 has a dotted line indicating its absence from mass spectrometry detection. (D) Proposed model of orientation and interactions of EMAP3 at the iRBC membrane.
The EMAP2‐myc Co‐IPs showed evidence that EMAP2 interacts with EMAP1 (but not in all replicas), but EMAP2 was not present in the reciprocal Co‐IP of EMAP1‐myc. This was possibly due to a low abundance of EMAP2, as was evident from the low amount of EMAP2 peptide hits in the EMAP2‐myc Co‐IPs. In addition, peptide abundances in the EMAP1‐myc Co‐IPs appeared low overall, as reflected in that the EMAP1‐myc signal was stronger in the EMAP‐3 Co‐IP than when EMAP1 itself was used as bait. Therefore, the possible interaction between EMAP1 or EMAP3 with EMAP2 remains uncertain. Unexpectedly, EMAP3 was not detected in any of the EMAP Co‐IPs, including the EMAP3‐myc Co‐IP. Likely, EMAP3 is not easily detectable by mass spectrometry, which would explain its absence in the EMAP1‐myc and EMAP2‐myc co‐IPs. Transmembrane proteins are known to be difficult to detect by mass spectrometry and their detection can require extensive optimisation (Savas et al. 2011). The difficulty in detecting EMAP3 by mass spectrometry is highlighted by its absence in several P. berghei blood stage proteomes (Hall et al. 2005; Pasini et al. 2013; Siau et al. 2023). In agreement with the EMAP3‐myc Co‐IPs, SMAC was not identified in the EMAP1‐myc or EMAP2‐myc Co‐IPs, which suggests that SMAC is not interacting with any of the EMAPs (Figure 4B, Tables S1–S4).
Overall, only a few proteins were detected above the set thresholds in the EMAPs and SMAC Co‐IPs. This may indicate low numbers of interacting proteins or may result from the approach used where we combined cross‐linking with RIPA solubilisation. For this reason, we also performed EMAP3‐myc Co‐IPs without crosslinker and lysed the samples with NP‐40 detergent. While the interaction with EMAP1 was confirmed, only a few additional interactions were identified (Figure S8A). To further study the interaction between EMAP1 and EMAP3, we modelled this interaction with AlphaFold 3 and AlphaBridge (Abramson et al. 2024; Álvarez‐Salmoral et al. 2024). When performing modelling using the full‐length models of EMAP1 (without signal peptide) and EMAP3 (with a pTM folding score of 0.46), the predicted interaction is of low confidence (ipTM = 0.46) and with the poorly folded C‐terminal extracellular domain of EMAP3 crossing the membrane to interact with EMAP1 (Figure S9A). In contrast, when modelling an interaction between only the properly folded domains of each protein model (pTM = 0.63), EMAP1 is placed on top of the short intracellular loops of EMAP3, formed in between its transmembrane domains. However, this interaction is predicted with even lower confidence (ipTM = 0.27) (Figure S9B).
All EMAPs and SMAC show evidence of interacting with two shared proteins, a tryptophan‐rich protein (PBANKA_0623300) and a Plasmodium helical interspersed subtelomeric (PHIST) protein (PBANKA_122900). In addition, all bait proteins, but particularly EMAP1‐myc and EMAP3‐myc, also co‐immunoprecipitated BiP (PBANKA_0818900), which is the ER heat shock protein 70 (HSP70) chaperone, also known as HSP70‐2. In addition, EMAP3‐myc co‐immunoprecipitated PBANKA_1437300, another ER resident heat shock protein 90 (HSP90), also named endoplasmin or glucose response protein 94 (GRP94) (Shahinas et al. 2013). These interactions likely take place in the ER and might indicate that EMAP3 is also actively secreted from the ER in schizonts. The shared interactions between the EMAPs and SMAC at the iRBC membrane are summarised in a schematic figure (Figure 4C).
Furthermore, we investigated the mouse host proteins that were present in the EMAP3‐myc and SMAC‐myc cross‐linked Co‐IPs. We found that both EMAP3 and SMAC appear to interact with the mouse RBC membrane skeleton protein ankyrin‐1 as well as spectrin (alpha and beta chain) and Band 3 protein, which links the RBC membrane to the membrane skeleton via ankyrin (Warncke et al. 2016), (Figure S8B, Table S4). These interactions seem specific to SMAC and EMAP3 and most likely do not result from contamination due to their sheer abundance since these proteins are absent or present at very low levels in Co‐IPs with the non‐exported Tir1‐myc control (Figure S8). In addition, EMAP3‐myc but not SMAC‐myc was found to bind mouse IgG immunoglobulin (peptide hits mapping to gamma (heavy) and kappa (light) chain variable and constant regions). This again is consistent with an iRBC surface exposure of EMAP3 where it would be exposed to mouse IgG antibodies. EMAP3‐myc also co‐immunoprecipitated Protein 4.2, which is part of the RBC cytoskeleton. In summary, the analysis of the co‐precipitation of the EMAPs and the co‐precipitation of these proteins with host proteins indicates that likely EMAP3 interacts with EMAP1 at the iRBC membrane where it might also interact with EMAP2 and the RBC membrane skeleton, either directly or as part of the same complex. To confirm such interactions, further validation by heterologous expression of the proteins and using, for example, surface plasmon resonance would be required. Based on the Co‐IP interaction data, the presence of the four transmembrane domains of EMAP3 and the demonstration that the C‐terminus of EMAP3 is exposed on the surface of the iRBC, we propose a model of the interactions and localisation of the EMAP3 at the iRBC surface (Figure 4C,D).
EMAP3 Neither Mediates Organ Sequestration of iRBC nor Influences Virulence
2.4
Based on the timing of expression and its location on the surface of the iRBC, we hypothesised that EMAP3 could mediate cytoadherence and tissue sequestration of the iRBC. To test this hypothesis we generated a gene‐deletion mutant (emap3 ko) lacking expression of EMAP3 using standard genetic modification technology (Janse, Ramesar, and Waters 2006). The emap3 gene was deleted in a dual reporter P. berghei (ANKA) reference line that constitutively expresses mCherry and luciferase, driven by the hsp70 and eef1a promoter, respectively (Prado et al. 2015). This dual hsp70:mcherry‐eef1a:luciferase (mCherry‐Luc) reporter line was used as a control in all experiments. Synchronised blood‐stage infections were established in Balb/c mice by intravenous injection of histodenz‐purified schizonts of the emap3 ko or mCherry‐Luc line. The blood stage development of emap3 ko parasites in the mice was monitored using Giemsa‐stained smears of tail blood. The blood stage growth rate of emap3 ko parasites was comparable with mCherry‐Luc blood stages. Although emap3 ko parasites appear to have a mildly reduced growth rate during the later days of the blood stage infection compared to the mCherry‐Luc control, this reduction was not statistically significant (p‐value = 0.183467) (Figure 5A).
EMAP3 does not play a critical role in iRBC sequestration and virulence (A) In vivo blood stage growth assay of emap3 ko compared to the mCherry‐Luc P. berghei background line. Balb/c mice were injected with 1 × 106 parasites intravenously (IV) on day zero. Parasitaemia was determined by Giemsa‐stained smears where each day in the graph represents the average of eight Balb/c mice from two independent biological experiments. Statistical analysis was performed on day five using a Student's t‐test (p‐value = 0.183467). (B, C) Whole body imaging of emap3 ko and mCherry‐Luc parasites in Balb/c mice using the Spectrum in vivo imaging system (IVIS) to visualise luciferase expressing parasites in (B) adipose tissue and (C) spleen on day one to five after injecting 10–30 × 106 purified schizonts IV. Statistical analysis was performed on day one using a Student's t‐test on adipose tissue signal, p‐value = 0.162349. Bioluminescence signal in each organ is normalised to the signal in lungs and expressed as normalised flux. Representative experiment with five Balb/c mice per parasite line is shown in (B, C) together with (D) representative whole‐body images from day one and four of infection. Corresponding regions of the image used to measure luciferase signal are shown. Area surrounded by the dashed square box was used to acquire the lung or total signal. The left belly of the mouse surrounded by a solid rectangle was used to measure adipose tissue signal and the area outlined by the oval was used to determine spleen signal. Analysed IVIS data for a second independent experiment together with all IVIS images are available in Figure S10. (E) Spleen weight of Balb/c mice infected with emap3 ko and mCherry‐Luc parasites on day five of infection where spleen weight is normalised to mouse body weight. Student's t‐test p‐value = 0.0050. (F) Evaluation of virulence of emap3 ko and mCherry‐Luc parasites in C57BL/6 mice, expressed as probability of mouse survival where mice are culled when they reach moderate symptoms. Eight (emap3 ko) or ten (mCherry‐Luc) C57BL/6 mice were assayed in two independent experiments. (G, H) Malaria symptom scoring of mCherry‐Luc (G) or emap3 ko (H) infected C57BL/6 mice that are culled on the day the mice are maintaining or exceeding a total score of six, which corresponds to moderate symptoms and was used to generate the probability of survival plot (F). Parasitaemia and spleen weight for C57BL/6 infections are available in Figure S10 and Table S5.
In parallel, we investigated whether EMAP3 plays a role in tissue sequestration by performing whole‐body in vivo imaging of the infected mice to measure parasite‐expressed luciferase as a measure of parasite load in organs using the Spectrum in vivo imaging system (IVIS), (Figure 5B–D, Figure S10). As expected, in all emap3 ko and mCherry‐Luc infected mice the lungs showed the highest parasite load, which is a major site of parasite accumulation along with abdominal adipose tissue (Franke‐Fayard et al. 2005). During early stages of infection (day one post‐infection), the emap3 ko infected mice displayed a mildly reduced luciferase signal in adipose tissue, compared to the mCherry‐Luc infected mice; however, this difference was not statistically significant (p‐value = 0.162349), and this small difference disappeared as the infection progressed (Figure 5B). It has been shown previously that P. berghei sequestration deficient blood stages commonly display reduced sequestration in adipose tissue but an increased luciferase signal in the spleen as a result of splenic clearance of the non‐sequestering iRBC (De Niz et al. 2016). We did not observe an increase in luciferase signal in the spleen of emap3 ko infected mice (Figure 5C). However, the weight of isolated spleen from sacrificed emap3 ko infected mice was significantly increased compared to spleens of mCherry‐Luc infected mice (p‐value = 0.0050), (Figure 5E). Consistent with our observations that EMAP3 does not play a significant role in blood stage growth or sequestration in Balb/c mice, we observed no difference in virulence between emap3 ko and mCherry‐Luc infected C57BL/6 mice. The emap3 ko and mCherry‐Luc infected mice developed moderate symptoms (as determined by scoring piloerection, mobility and hunching) at a comparable time (day six to eight post‐infection), at which point they were culled (Figure 5F–H). The growth rate of emap3 ko blood stages in C57BL/6 mice was not significantly lower compared to mCherry‐Luc blood stages, and thereby mirrored results obtained in Balb/c infected mice (Figure S11A). Curiously, no significant difference in spleen weight was observed between emap3 ko and mCherry‐Luc infected C57BL/6 mice, although the trend is the same as in Balb/c mice with emap3 ko mice exhibiting enhanced spleen weight (Figure S11B). The enhanced spleen weight observed for emap3 ko infected mice thus supports a potential but minor role in sequestration of EMAP3. We thereby conclude that EMAP3 does not play a significant role in the growth and multiplication of P. berghei blood stages or iRBC sequestration and does not influence parasite virulence characteristics.
Discussion
3
We here identify EMAP3, a novel P. berghei protein that is exported by blood stage parasites into the iRBC and inserted into the iRBC membrane where its C‐terminal is exposed on the outer side of the iRBC membrane. So far, the location of relatively few putative P. berghei exported proteins has been characterised. This includes proteins encoded by the multigene families pir, fam‐a, fam‐b, etramp (early transcribed membrane protein) and phist (Plasmodium helical interspersed subtelomeric), and the tryptophan‐rich protein family (Currà et al. 2012; Pasini et al. 2013; Fougère et al. 2016; Moreira et al. 2016; Gabelich et al. 2022). It was proposed that proteins encoded by the pir multigene family are located at the iRBC surface (Cunningham et al. 2005; Di Girolamo et al. 2008; Yam et al. 2016), like the PfEMP1 protein of P. falciparum encoded by the var multigene family (Craig and Scherf 2001). However, several recent studies analysing multiple different PIR proteins by tagging with fluorescent markers demonstrate a location in the iRBC cytoplasm and no evidence was found for a location at the membrane surface (Fonager et al. 2012; Fougère et al. 2016; Giorgalli et al. 2022).
Several proteins encoded by single copy genes including IBIS1 (intra‐erythrocytic P. berghei‐induced structures 1), SBP1 (skeleton binding protein 1), MAHRP1a (membrane‐associated histidine‐rich protein 1a), SMAC and exported proteins lacking functional annotation have also experimentally been shown to be exported into the iRBC. However, these proteins have either a diffuse location pattern in the iRBC cytoplasm or have a discrete punctate distribution associated with IBIS membranous structures (Fonager et al. 2012; Pasini et al. 2013; De Niz et al. 2016). For EMAP1 and EMAP2, a location at the iRBC membrane periphery has been shown. Nevertheless, no strong evidence was found for exposure of these proteins at the outer side of the iRBC surface (Pasini et al. 2013). The difference in localisation for EMAP2 previously shown (Pasini et al. 2013), compared to what we observed, may be due to the different epitope tags used in the two studies. The results of the surface shaving and differential permeabilisation experiments presented here further confirm that despite localising to the iRBC membrane, EMAP1 is not exposed on the iRBC surface. To our knowledge, only one other protein, the merozoite surface protein PbTiP (P. berghei T‐cell immunomodulatory protein), has been demonstrated to be exported to the iRBC membrane and exposed on the outer surface of the iRBC during asexual development (Kalia et al. 2021).
We here demonstrate that EMAP3 localises to the outer surface of the iRBC and provide evidence for an interaction with the iRBC membrane protein EMAP1 and EMAP2 (Pasini et al. 2013). We did not find any evidence for interaction with the exported protein SMAC, which we here show is located to distinct foci at the iRBC membrane and is previously known to mediate P. berghei iRBC sequestration (Fonager et al. 2012; Pasini et al. 2013). Deletion of EMAP3 has no significant effect on blood stage development, sequestration of schizonts nor virulence. This is similar to the absence of an effect on blood stage development or iRBC sequestration of mutants lacking either EMAP1 or EMAP2 (Pasini et al. 2013). We therefore conclude that EMAP3 is not a major mediator of iRBC sequestration and is thereby not the unidentified P. berghei surface ligand that binds to CD36. Nevertheless, EMAP3 might play a minor role in sequestration and this role could be masked by a potential functional redundancy where multiple parasite ligands can bind CD36.
The molecular function of EMAP2 and EMAP3 remains unknown due to lack of predicted functional domains, where structural modelling using Alphafold 3 did not infer any new clues to their potential function. The unstructured extracellular C‐terminal domain of EMAP3 suggests that folding into its functional confirmation requires binding of a yet to be identified host or parasite protein. EMAP1 is a member of the Pb‐fam‐1 family, which carries a PYST‐A domain with strong homology to START domains and has therefore implicated Pb‐fam‐1 proteins in lipid binding and transport (Frech and Chen 2013; Pasini et al. 2013; Fougère et al. 2016). Furthermore, at least one of the Pb‐fam‐1 proteins (PBANKA_1327251) has been confirmed to have phosphatidylcholine (phospholipid) transfer activity (Fougère et al. 2016). Structural modelling of EMAP1 using Alphafold 3 confirms its START domain and confidently predicts its ability to bind lipids including cholesterol and phosphatidylcholine. Intriguingly, we observed some evidence that EMAP1‐myc has an apical localisation in mature schizonts, potentially also localising to the rhoptries. Plasmodium proteins important for early stages of iRBC infection and establishment of the PV are known to be packaged into the bulb of the pear‐shaped rhoptry organelles and discharged upon infection (Riglar et al. 2011; Ghosh et al. 2017). This warrants further investigation of the putative EMAP1 rhoptry localisation, but leads us to propose a potential role for EMAP1 in membrane biogenesis during the extensive membrane remodelling that is a hallmark of malaria RBC infection.
EMAP3‐myc reproducibly co‐immunoprecipitated EMAP1 in all Co‐IP conditions tested, and to a lesser extent also EMAP2. We performed reciprocal Co‐IPs but EMAP3 was not detected neither in EMAP1‐myc or EMAP2‐myc Co‐IPs, nor when EMAP3‐myc itself was used as the bait. This is despite the presence of residues for tryptic digestion in the EMAP3‐myc amino acid sequence and confirmation of EMAP3‐myc in the Co‐IP lysates by Western blot. The detection of SMAC, EMAP1 and EMAP2 in multiple P. berghei blood stage proteomic datasets stages (Hall et al. 2005; Pasini et al. 2013; Siau et al. 2023), while simultaneously failing to detect EMAP3, strongly supports that EMAP3 is not readily detectable by mass spectrometry in the blood. Nevertheless, EMAP3 has been detected by mass spectrometry in the ookinete (Tremp et al. 2020) and liver stages of P. berghei (Shears et al. 2019) where it is more highly transcribed (Howick et al. 2019). Two proteins, tryptophan‐rich protein (PBANKA_0623300) and a PHIST protein (PBANKA_122900), showed evidence of interacting with the three EMAPs and SMAC. The PHIST protein PBANKA_122900 has previously been identified by proteomics, found to be dispensable to blood stage growth and to localise to IBIS in the iRBC (Pasini et al. 2013; Moreira et al. 2016). There is no knockout phenotype or localisation data available for the tryptophan‐rich protein PBANKA_0623300; however, it was previously detected in a proteomic study as a putative liver merosome protein (Shears et al. 2019), the same data set that identified EMAP3 as present in the liver stages. Structural modelling of the EMAP1 and EMAP3 interaction failed to confidently predict an interaction between the proteins. Future work aimed at further validating and studying the interaction between EMAP1 and EMAP3 should investigate their reciprocal dependence by studying their location when the other protein is absent.
The role of SMAC in P. berghei iRBC sequestration and the observation that SMAC‐myc localises to dense and large patches associated with the iRBC membrane supports the hypothesis that SMAC might serve a trafficking role to present a CD36‐binding parasite ligand on the surface of the iRBC (Fonager et al. 2012). We found no evidence that SMAC interacts with any of the EMAPs and we hypothesise that it does not play a role in anchoring EMAP3 at the iRBC membrane. SMAC instead possibly plays a role in trafficking of an unidentified parasite ligand binding to host CD36 that was not captured here. Nevertheless, the decoration of the surface of the iRBC with parasite proteins including EMAP3 is a process that likely takes place during the trophozoite stage, so an interaction between EMAP3 and any protein(s) presenting it on the surface could be stage dependent and such interactions missed in our schizont Co‐IP setup. There was overall a paucity in interactions identified from each Co‐IP, and in particular for the SMAC‐myc Co‐IPs. This indicates that conditions we have used in our Co‐IPs might have enabled detection of only the most robust or abundant interactions during the schizont stage such as those between EMAP3 and EMAP1. Further efforts to optimise Co‐IP conditions might identify more SMAC‐interacting proteins, including a potential P. berghei surface ligand that binds to host CD36.
The lack of direct evidence for EMAP3 interacting with SMAC indicates that EMAP3 is trafficked onto the iRBC surface in a mechanism independent of SMAC. Analysis of transport and location of EMAP3 in a mutant lacking expression of SMAC would conclusively prove this. Interestingly, it has been shown that EMAP1 and EMAP2 were still localised to the iRBC membrane in the absence of SMAC (Pasini et al. 2013), which supports the hypothesis that EMAP3 is also transported to the iRBC surface membrane independently of SMAC. These observations suggest the feasibility of generating P. berghei mutants where the CD36‐binding of iRBC is abolished by deletion of the smac gene and where EMAP3 is used as a scaffold presenting P. falciparum proteins on the outside surface of the iRBC by fusing them to the C‐terminus of EMAP3, after the last transmembrane domain. For example, EMAP3 could be fused with cytoadherence proteins or domains, including PfEMP1 variants. The dispensable nature of emap3 for blood stage growth is an advantage as it indicates that it is likely possible to truncate the endogenous emap3 gene after the last transmembrane domain to display the P. falciparum proteins/domains onto the iRBC surface.
EMAP1 has previously been used as the scaffold to display Var2CSA on the surface of P. berghei infected iRBCs as a model to study pregnancy associated malaria experimentally in vivo. However, only a very low proportion (< 6%) of iRBCs were found to express EMAP1‐Var2CSA at the outer side of the iRBC (de Moraes et al. 2016). This is in agreement with previous observations that only a low percentage (around 7%) of iRBCs exposed EMAP1 on their outside surface (Pasini et al. 2013), which concurs with our observations that the C‐terminus of EMAP1‐myc is not exposed on the outer surface of the iRBC. With no predicted transmembrane domain in EMAP1, we propose that the association of EMAP1 with the RBC membrane is by its interaction with the N‐terminus and/or the short intracellular loops of the transmembrane protein EMAP3. An improved model, where EMAP3 is used as an expression scaffold in a mutant lacking SMAC expression, could be combined with humanised mice expressing, for example, human CD36 (Xie et al. 2023) offering a much needed platform to test iRBC sequestration and disease modulating therapies aimed at interrupting P. falciparum cytoadherence and sequestration in vivo.
Materials and Methods
4
All antibodies, key reagents and primer sequences are available in Table S6.
Bioinformatic Analyses
4.1
The selection criteria applied to identify putative novel P. berghei blood‐stage surface proteins included the presence of an N‐terminal signal peptide, PEXEL/HT motif, transmembrane domain(s), experimental evidence of export through protein tagging and localization studies, gene expression profiles, gene ontology (GO) annotations related to ‘host cell remodelling’, ‘cytoplasm’ or ‘membrane trafficking’ and proteomic evidence from stage‐specific datasets and subcellular host compartments (e.g., cytosol and membrane). Gene annotation information as well as transcription and proteomic data was obtained from PlasmoDB (https://plasmodb.org), (Alvarez‐Jarreta et al. 2023). For the analysis and revision of gene models, signal peptides and transmembrane helices were de novo predicted by SignalP 6.0, TopCons and DeepTMHMM (Tsirigos et al. 2015; Hallgren et al. 2022; Teufel et al. 2022). AlphaFold 3 models (Abramson et al. 2024) were predicted for EMAP1 and EMAP2 without their signal peptides (without the first 25 and 20 amino acids, respectively) and for EMAP3. EMAP1 was also modeled bound to cholesterol and phosphatidylcholine using a local installation of AlphaFold 3. Interaction models for EMAP1 together with EMAP3 were made in various stoichiometries, but only the highest confidence model is shown. Possible interaction sites between the two proteins in the AlphaFold models were determined by AlphaBridge (Álvarez‐Salmoral et al. 2024).
Animal Work
4.2
The parasitaemia of infected animals was determined by microscopy of methanol‐fixed and Giemsa‐stained thin blood smears from small‐volume tail blood samples. Infected blood was harvested by heart puncture on terminally anaesthetised mice [90 mg/kg ketamine; 20 mg/kg xylazine in phosphate buffered saline (PBS)] and blood collected into a syringe containing heparin (Sigma‐Aldrich). Cervical disolcation was performed to euthanise mice.
Animal Work at Umeå University
4.2.1
Animal experiments were performed under and according to ethics permits A34‐2018 and A24‐2023 as approved by the Swedish Board of Agriculture (Jordbruksverket). For routine propagation of parasites, BALB/c mice were used and for pathology experiments, C57BL/6 mice. Female‐only mice were purchased from Charles River Europe and used for experiments from 6 weeks of age. Housing was in groups of four mice in individually ventilated cages (IVC) with autoclaved wood chips and nesting material and a temperature of 21°C ± 1°C under a 12:12 h light–dark cycle, with a relative humidity of 55% ± 5%. Specific pathogen‐free conditions were maintained and biannual Exhaust Air Dust (EAD) monitoring and analysis were performed. Fresh water and commercial dry rodent diet were freely available (ad libitum) at all times. Visual health checks were performed daily to monitor animal health.
Animal Work at Leiden University Medical Centre
4.2.2
All animal experiments were approved and performed under a license given by competent authority after ethical evaluation by the Animal Experiments Committee Leiden (AVD1160020171625). All animal experiments were carried out in agreement with the Experiments on Animals Act, which is the applicable legislation in The Netherlands that follows European guidelines (EU directive no. 2010/63/EU). All animal experiments were performed in Leiden University Medical Center; LUMC, a licensed establishment for animal experimentation. Experiments were performed with female 6–7 week old OF1 mice, which were purchased from Charles River Laboratories. Mice were housed in ventilated cages supplied with autoclaved aspen wood chips, a wood chew block, a fun tunnel, and nestlets (at 21°C ± 2°C; 12‐h light–dark cycle; relative humidity of 55% ± 10%). Mice were fed and supplied with water ad libitum. They were fed commercially available autoclaved dry rodent diet pellets.
Animal Work at University of Geneva
4.2.3
Animal experiments were conducted with the authorisation numbers GE102 and GE‐58‐19, according to the guidelines and regulations issued by the Swiss Federal Veterinary Office. P. berghei ANKA strain clone 2.34 together with derived transgenic lines were grown and maintained in CD1 outbred mice. Six to twelve‐week‐old mice were obtained from Charles River laboratories, and females were used for all experiments. Mice were specific pathogen free (including Mycoplasma pulmonis ) and subjected to regular pathogen monitoring by sentinel screening. They were housed in individually ventilated cages furnished with a cardboard mouse house, tunnel and Nestlet, maintained at 21°C ± 2°C under a 12 h light/dark cycle, and given commercially prepared autoclaved dry rodent diet and water ad libitum.
Plasmids and Parasite Lines
4.3
A complete list of parasite lines used in this study can be found in Table S7. The emap3 ko (PBANKA_0825900) plasmid was generated by amplifying 0.5 kb (5′) and (3′) homology arms flanking the emap3 open reading frame (ORF) from P. berghei genomic DNA and subsequently cloned into the pL0001 plasmid (www.mr4.com) carrying the mutated tgdhfr/ts (Toxoplasma gondii dihydrofolate reductase‐thymidylate synthase) selection marker expressed under control of the pbdhfr/ts (P. berghei dihydrofolate reductase‐thymidylate synthase) 5′ and 3′ UTRs (untranslated regions). The resulting emap3 ko pL2296 plasmid was linearised (using KpnI, NotI and ScaI) and transfected into the P. berghei mCherry_Luc background line (hsp70p:mCherry eef1ap:Luciferase, 1868Cl1), (Prado et al. 2015) and cloned by standard limited dilution cloning (Janse, Ramesar, and Waters 2006) to generate the PbEMAP3 ko line (3076cl1).
The emap3‐myc (PBANKA_0825900) tagging vector was generated by amplifying a 0.7 kb 3′ fragment of emap3 that was cloned in‐frame with the coding sequence for myc, which upon integration of the vector into the emap3 target locus inserts a 3′ (C‐terminal) myc tag. To this end the homology arm fragment was amplified from P. berghei genomic DNA and cloned into the pL1672 vector (Rijpma et al. 2016) containing the selection marker tgdhfr/ts using BamHI/EcoRI restriction enzymes to generate the emap3‐myc vector (pL2302), which was linearised with NdeI prior to transfection. The emap1‐myc (PBANKA_0836800, pL1594), emap2‐myc (PBANKA_0316800, pL2358) and smac‐myc (PBANKA_0100600, pL1435) tagging vectors were generated by replacing the mCherry tag with a 3×‐myc tag in existing C‐terminal tagging vectors carrying the tgdhfr/ts selection marker (Fonager et al. 2012; Pasini et al. 2013). The vectors were linearised prior to transfection using NdeI (pL1594), ClaI (pL2358) or NdeI (pL1435). The resulting 3×‐myc tagging vectors were transfected either into PbANKA cl15cy1 reference line (Janse, Franke‐Fayard, and Waters 2006), (emap1‐myc, line 1583 and emap3‐myc, line 3084), mCherry_Luc (emap2‐myc, line 3296) or GFP_Luc (Franke‐Fayard et al. 2005), (smac‐myc, line 1413).
The PbCas9‐FLAG_Tir1‐myc line was generated by cloning the auxin inducible osTir1 (transport inhibitor response 1 gene from Oryza sativa ) (Yesbolatova et al. 2020), fused to a C‐terminal 3×‐myc tag and flanking it by 260 bp (5′) and 288 bp (3′) homology arms targeting insertion into the p230p locus, downstream of the GIMO locus (Lin et al. 2011). The 5′ p230p and 3′ p230p targeting sequences were cloned into the pPbU6‐hdhfr/yfcu plasmid (Addgene #216422), (Jonsdottir et al. 2025), carrying the dual positive and negative selection marker hdhfr‐yfcu (human dihydrofolate reductase/yeast cytosine deaminase and uridyl phosphoribosyl transferase) under control of pbeef1a (P. berghei elongation factor alpha) 5′UTR and pbdhfr/ts 3′UTR. The selection marker was made recyclable by addition of an additional pbdhfr/ts 3′UTR motif upstream of the pbeef1a 5′UTR. Finally, a gRNA targeting p230 was designed using EupaGDT and ligated using T4 ligase (New England Biolabs, Ipswich, MA, USA), resulting in the pYCs‐PbU6‐hDHFR/yFCU‐NS‐p230p plasmid. The eef1a promoter used to drive osTir1 was amplified from the pYCs plasmid (Qian et al. 2018) and the P. falciparum histidine‐rich protein 2 (pfHrp2) terminator was amplified from pDC2‐Cam‐Cas9‐hDHFR (Lim et al. 2016). The eef1a promoter, osTir1‐3×‐myc and hrp2 terminator were seamlessly ligated into the pSKA‐001 vector using NEBuilder HiFi DNA Assembly Master Mix (NEB) and the intermediate fragment was cloned into the pYCs‐PbU6‐hDHFR/yFCU‐NS‐p230p vector. The resulting vector was transfected into the PbCas9 line where spCas9 ( Streptococcus pyogenes CRISPR associated protein 9) is integrated in the GIMO (gene in marker out) locus (Lin et al. 2011; Jonsdottir et al. 2025).
Transfections
4.4
The parasite lines described above were generated by transfection of P. berghei schizonts and transgenic parasites were selected for by pyrimethamine, all as previously described (Janse, Ramesar, and Waters 2006). In brief, P. berghei parasites were cultured ex vivo from infected blood (approximately 1%–5% parasitaemia) at 37°C with 80 rpm shaking for 22 h. The culture was then checked by Giemsa staining for the presence of schizonts. Schizonts were harvested by using a 15.2% w/v Histodenz gradient (Sigma‐Aldrich), (from 27.6% w/v Histodenz buffered stock solution), collecting the brown parasite layer on the interface. Collected schizonts were washed and then re‐suspended in P3 buffer (Lonza) and mixed with DNA. The schizont‐DNA mixture was then loaded onto a Lonza transfection cuvette and electroporated. Fresh media was added to the cuvette and aspirated for direct intravenous injection into a mouse. Mice were treated with pyrimethamine in drinking water a day after transfection to select for transgenic parasites. The transgenic lines obtained were genotyped by diagnostic PCR with all genotyping primers available in Table S8. DNA for genotyping was extracted from whole blood (> 1% parasitemia) typically using the Qiagen DNAeasy or GeneJet Genomic DNA purification kit (Thermofisher).
Immunofluorescence Assays
4.5
Immunofluorescence assays (IFA) were carried out by allowing 20–100 μL of cells fixed with 4% paraformaldehyde (diluted in RPMI or PBS) to settle on a poly‐D‐lysine (GIBCO) coated slide for 15 min. The cell suspension was then removed, and the slide was washed with 200 μL of 1X PBS. To permeabilize cells, 200 μL of 0.2% Triton X‐100 was added and allowed to incubate for 5 min. The cells were then washed 3 times with 200 μL of 1X PBS. Blocking buffer was then added (200 μL of 3% BSA in 1X PBS) and incubated at room temperature (RT) for 30–40 min. The blocking solution was removed, and 100 μL of primary antibody solution (diluted in blocking buffer) was added to the slide that was placed in a humid chamber and allowed to incubate at 4°C overnight. The primary antibody was then removed, and the slide was washed three times with 1X PBS before adding the secondary antibody (diluted in blocking buffer). The slides were incubated in a dark humid chamber for 1–2 h at RT. The secondary antibody was removed, and the slides were washed three times with 200 μL of 1X PBS. The cells were mounted with 5 μL of Vectashield with DAPI and coverslip and sealed using nail polish. Slides were allowed to dry overnight before imaging. Widefield images were taken using a Zeiss Apotome microscope at 63X objective, and confocal images were taken using a Leica SP8 microscope at 63X objective.
Surface Shaving and Permeabilisation
4.6
Surface shaving and permeabilisation was performed by adding 200 μL of ice‐cold 0.003% trypsin in 1X PBS to 50 μL of Percoll purified schizonts. This was allowed to incubate at RT for 50 min. After incubation, the tube was spun at 500 g for 1 min, the supernatant was removed and 200 μL of RPMI was added to stop trypsin activity. In parallel, cells with 200 μL of RPMI were allowed to incubate at RT for 50 min as a negative control for trypsin treatment. The aforementioned IFA protocol was then performed on the cells where the permeabilisation (0.2% Triton X‐100) step was omitted for non‐permeabilised set‐ups.
Brefeldin A Treatment
4.7
Treatment with Brefeldin A (BFA, Sigma) was performed by taking infected mouse blood (mostly rings and early trophozoites) and setting up 3 mL ex vivo cultures in 6‐well plates. BFA was added to a final concentration of 5 μM. As a control, the same volume of DMSO was added to control wells. An aliquot (T = 0) was taken at the start of the experiment and fixed with 4% paraformaldehyde. The plate was placed in an airtight container following the candle jar method (Jensen and Trager 1977) and incubated at 37°C. After 3 h, an aliquot of the sample was taken and fixed with 4% paraformaldehyde. The standard IFA protocol was then followed to determine the localisation of the myc tagged protein with and without BFA treatment.
Ultrastructure Expansion Microscopy
4.8
Ultrastructure expansion microscopy (U‐ExM) was performed on schizonts separated from uninfected erythrocytes on a density gradient (3.6 mL Percoll, Cytiva, 1 mL RPMI and 400 μL 10X PBS). Schizonts were fixed with 4% paraformaldehyde according to protocols adapted to Plasmodium parasites (Bertiaux et al. 2021). Cells were allowed to settle on 12 mm round coverslips coated with poly‐d‐lysine for 10–15 min. The coverslips with cells were then incubated in a 1.4% formaldehyde (FA) and 2% acrylamide (AA) solution to add protein anchors. This was done by taking the coverslips and placing them in individual wells of a 24‐well plate and adding the FA/AA solution; the plate was incubated at 37°C for 5 h, water was added to the empty wells to prevent dehydration of the plate. The gel solution was prepared by adding 5 μL of 10% ammonium persulfate (APS) and 5 μL of 10% TEMED to an aliquot of monomer solution (23% Sodium acrylate; 10% AA; 0.1% BIS‐AA in PBS). 35 μL of the gel solution was then placed on parafilm and the coverslips were inverted onto the gel (cell side down) to perform gelation. Coverslips embedded in gels were allowed to gelate for 1 h at 37°C. Denaturation was performed by taking the gels and placing them in 1.5 mL tubes with denaturation solution. The tubes were then incubated at 95°C for 1 h. Gel expansion was then performed by placing the gels in 200 mL of double distilled (ddH_2_0) at RT for 30 min. This step was repeated two more times, replacing the water each time. The gels were then left to expand overnight at room temperature. The following day, the gels were shrunk by incubating them twice in 1X PBS for 15 min each time. The shrunk gels were transferred to 6‐well plates and incubated in 1 mL of primary antibody solution at 37°C for 2 h with 125–150 rpm shaking. The gels were then washed three times with 1 mL of 0.1% PBS‐Tween. The gels were then incubated in a secondary antibody solution at 37°C for 2 h with 125–150 rpm shaking, followed by three washes with 1 mL of 0.1% PBS‐Tween. Gels were subsequently incubated in stains (NHS ester, Merck and BODIPY, Thermofisher) for 1 h at RT shaking at 125–150 rpm. The gels were then washed three times with 1 mL of 0.1% PBS‐Tween. Final gel expansion was performed by placing the gels in 200 mL of ddH20 and incubating for 30 min; this was repeated three times. The gels are allowed to incubate in ddH_2_0 overnight in the dark. Finally, the gels were cut into small cubes and placed on a poly‐d‐lysine coated 24 mm coverslip in an O ring. The expanded cells were visualised on a confocal Leica SP8 microscope using 63X and 100X objectives.
Growth Assays
4.9
To evaluate parasite blood stage growth, mice were injected with 1 × 10^6^ parasites intravenously. Parasitemia was tracked by smearing daily starting 3 days after initial infection. Thin blood smears were made by small volume tail bleeds collected by lancet and smeared on glass slides. The smears were fixed with methanol and stained with 10% Giemsa for 10 min. The parasitemia was calculated by counting the number of P. berghei infected cells per 1000 RBCs. The parasitemia over 5 days of infection of the emap3 KO was compared to the mCherry_Luc background line.
In Vivo Imaging
4.10
To evaluate parasite sequestration to organs, in vivo imaging (IVIS) was performed following the established protocol for P. berghei (Franke‐Fayard et al. 2005; Fonager et al. 2012). In brief, to establish a roughly synchronous infection, mice were injected with 10–30 × 1 × 10^6^ Percoll‐purified schizonts at 12 noon, intravenously. After 24 h (12 noon the next day), mice were anaesthetised with isoflurane and injected with 30 μL of 80 mg/mL D‐luciferin (Synchem) subcutaneously. After 1–2 min of injecting D‐luciferin, imaging with the IVIS Spectrum imager was performed. Mice were imaged at the same time (between 11 and 12 noon) each day, until day five post‐infection. On the last day of imaging, mice were anaesthetized and dissected to collect spleens for determination of spleen weight. IVIS images were analysed using the Living Image software, taking only images with no saturated pixels. Regions corresponding to the lungs, adipose tissue and spleen were drawn for each mouse using the region of interest (ROI) tool in the software.
Co‐Immunoprecipitation
4.11
A formaldehyde cross‐linking co‐immunoprecipitation (Co‐IP) protocol was performed to fix protein complexes together and co‐immunoprecipitate any proteins interacting with the myc‐tagged bait proteins according to a previously established protocol (Fang et al. 2018). Co‐IPs were performed on Percoll‐purified schizonts from a 100 mL schizont culture with 5 mL of infected mouse blood with a 1%–5% parasitemia. The schizonts were harvested from the Percoll gradient and pelleted at 300 g for 3 min. The schizont pellet was washed two times with 1 mL of fresh RPMI. The pellet was then resuspended in 500 μL of RPMI; this was then split equally between two tubes. To one pellet, 500 μL of 0.1% ice‐cold saponin was added, then incubated on ice for 5 min. The other pellet was kept untreated with saponin; instead, 500 μL of RPMI was added. After saponin treatment, 500 μL of RPMI was added to stop the reaction. To cross‐link the sample, formaldehyde was added to a final concentration of 1% and incubated at room temperature for 10 min with gentle agitation. To stop the crosslinking reaction, 10 mL of 0.125 M glycine was added to the samples and incubated at RT with gentle agitation for 5 min. The sample was spun at 1000 g for 10 min at 4°C and the supernatant was removed. The pellet was then used to perform the next IP steps on the same day or stored at −80°C for future use.
To immunoprecipitate myc tagged proteins, Protein G dynabeads (Thermofisher) were coated with rabbit α‐myc antibody (Cell Signalling Technologies) following the manufacturer's protocol for coating beads for immunoprecipitation. In brief, magnetic beads (60 μL per sample) were placed on a magnet and allowed to separate from storage liquid for 10–30 s. The beads were then washed with 1 mL of 0.01% PBST. After washing, 7 μg of rabbit α‐myc antibody was added. The beads were allowed to bind to the antibody for 15 min at room temperature with gentle rotation. To the cross‐linked pellet, 1 mL of ice‐cold RIPA lysis buffer was added. The tubes were then incubated on ice for 10 min. The pellet‐RIPA solution was then passed through an insulin syringe 20× to solubilize the pellet. The sample was then spun at 14,000 g for 15 min, 4°C. A 30 μL aliquot from this solution was saved as the “total protein” sample. The supernatant was then mixed with 60 μL of the α‐myc antibody coated magnetic beads and allowed to bind overnight at 4°C with gentle rotation. The bead‐sample slurry was transferred to a fresh tube, then placed on the magnetic rack for 30 s to 1 min. The liquid was collected as the ‘flow through (FT)’ sample. The tube was then separated from the magnetic rack and 1 mL of ice‐cold RIPA was added to resuspend the beads. The tubes were placed back on the magnetic rack and the liquid was removed. The RIPA wash was repeated two times. After the last wash with RIPA, the beads were resuspended with 1 mL of ice cold 1X PBS with protease inhibitors and placed in a fresh tube. This was then allowed to settle on the magnet for 30 s to 1 min. The wash with 1X PBS with protease inhibitors was repeated two times. The beads were resuspended in 500 μL of 1X PBS with protease inhibitors after the last wash. An aliquot of the beads was collected as the ‘immunoprecipitation (IP)’ sample and analysed by Western blot together with the total protein and flow‐through samples.
Mass spectrometry data was analysed using the Scaffold software. A 1% protein false discovery rate (FDR) with at least 2 peptides identified per protein and a 0.1% peptide FDR was set as thresholds. To generate a list of putative protein interactions for each myc tagged P. berghei line (Tables S1–S4), the hits for each myc‐tagged bait protein were normalised to that of the TiR1‐myc control. Shown in the heatmap are only those parasite or host protein hits that were either (1) not present (with a spectral count of zero) in the corresponding TiR1‐myc co‐IP control and were represented in at least two co‐IP replicates, or (2) if a protein was present in the TiR1‐myc sample, it was included in the heatmap if the abundance was greater than fourfold of the spectral counts of the TiR1‐myc control in at least two of the co‐IP repeats.
Western Blot
4.12
Samples for western blot were mixed with a 4X Laemmli buffer with β‐mercaptoethanol (BME). These were incubated at 70°C for 10 min. Samples were run on either a 4%–12% polyacrylamide pre‐cast gel (NuPage, Thermofisher) or a 4%–20% tris‐glycine polyacrylamide gel (Mini‐PROTEAN TGS, BioRad) at constant voltage (90–150 V), until the running dye had reached the bottom of the gel. Transfer to a PVDF or nitrocellulose membrane was performed by activating the membrane in methanol (PVDF only) then soaking in cold transfer buffer. The transfer stack was assembled following Thermofisher or Bio‐Rad guidelines for Western blot. Wet transfers were performed at 30 V for 1 h. The membrane was removed and stained with Ponceau S staining solution to visualise total protein amount and confirm successful transfer to membrane. The membrane was blocked in 5% Milk in 0.05% PBS‐Tween for at least 1 h on a shaker. The blocking solution was removed and primary antibody diluted in blocking solution was added and incubated overnight at 4°C. The primary antibody was then removed and membrane washed three times with 0.05% PBS‐Tween, 10–15 min each wash. The secondary antibody diluted in blocking solution was then added and allowed to incubate for 2–3 h whilst shaking. The membrane was subsequently washed three times with 0.05% PBS‐Tween, 10–15 min each wash. Detection solution was prepared (Amersham, ECL) and added to the blot, followed by imaging with the BioRad Chemidoc.
Virulence Assay
4.13
Virulence assays were performed by injecting 1 × 10^6^ parasites intravenously into 6 week old female C57BL/6 mice. Mice were health checked daily and given a symptom score following a scoring matrix. Briefly, a mouse was given a score from 1 (not affected) to 3 (significantly affected) for each category: piloerection (fur condition), movement and hunching (body posture). Infection was terminated when a mouse reached a score above six, which is equivalent to moderate symptoms, or scored six at two subsequent scoring occasions. Moderate symptoms equate to substantial piloerection, reduced mobility even under stimulation, and persistent hunching. Symptoms of cerebral malaria were also checked for by testing reflexes, self‐preservation, balance, motor skills and muscle tonus. Each day, the weight of the mouse was recorded, and thin blood smears were collected to enumerate parasitemia.
Author Contributions
Sophia Raine C. Hernandez: conceptualization, investigation, writing – original draft, writing – review and editing, data analysis and curation. Ravish Rashpa: investigation, writing – review and editing, data analysis and curation. Thorey K. Jonsdottir: investigation, writing – review and editing. Martina S. Paoletta: investigation, writing – review and editing. Josy ter Beek: investigation, writing – review and editing. María Rayón Díaz: investigation, writing – review and editing. Jelte M. M. Krol: investigation, writing – review and editing. Severine Chevalley‐Maurel: investigation, writing – review and editing. Takahiro Ishizaki: investigation, writing – review and editing. Ronnie P.‐A. Berntsson: funding acquisition, supervision, writing – review and editing. Chris J. Janse: conceptualization, writing – review and editing, supervision, resources. Blandine Franke‐Fayard: conceptualization, writing – review and editing, supervision, resources. Mathieu Brochet: funding acquisition, supervision, writing – review and editing. Ellen S. C. Bushell: conceptualization, funding acquisition, writing – original draft, writing – review and editing, supervision, resources, project administration, data curation.
Funding
This work was supported by Vetenskapsrådet (2021‐06602 and 2023‐02423), Knut och Alice Wallenbergs Stiftelse (2019.0178), Swiss National Science Foundation (31003A_179321 and 310030_208151), Japan Society for Promotion of Science (JSPS) (202160312) and Leiden University Medical Center (LUMC) internal funds.
Ethics Statement
All animal work was conducted under ethical permits that were evaluated and approved according to the rules and regulations of respective host countries and awarded to the principal investigators: E.S.C.B (A34‐2018 and A24‐2023, Sweden), B.F.F. and C.J.J (AVD1160020171625, The Netherlands) and M.B. (GE102 and GE‐58‐19, Switzerland).
Conflicts of Interest
The authors declare no conflicts of interest.
Supporting information
Figures S1–S11: mmi70050‐sup‐0001‐FiguresS1‐S11.pdf.
Tables S1–S8: mmi70050‐sup‐0002‐TablesS1‐S8.xlsx.
The reference list from the paper itself. Each links out to its DOI / PubMed record.
- 1Abramson, J. , J. Adler , J. Dunger , et al. 2024. “Accurate Structure Prediction of Biomolecular Interactions With Alpha Fold 3.” Nature 630, no. 8016: 493–500.38718835 10.1038/s 41586-024-07487-w PMC 11168924 · doi ↗ · pubmed ↗
- 2Alvarez‐Jarreta, J. , B. Amos , C. Aurrecoechea , et al. 2023. “V Eu Path DB: The Eukaryotic Pathogen, Vector and Host Bioinformatics Resource Center in 2023.” Nucleic Acids Research 52, no. D 1: D 808–D 816.10.1093/nar/gkad 1003 PMC 1076787937953350 · doi ↗ · pubmed ↗
- 3Álvarez‐Salmoral, D. , R. Borza , C. Maiella , et al. 2024. “Alpha Bridge: Tools for the Analysis of Predicted Macromolecular Complexes. bio Rxiv.” 10.1101/2024.10.23.619601. · doi ↗
- 4Baldi, D. L. , K. T. Andrews , R. F. Waller , et al. 2000. “RAP 1 Controls Rhoptry Targeting of RAP 2 in the Malaria Parasite Plasmodium Falciparum.” EMBO Journal 19, no. 11: 2435–2443.10835342 10.1093/emboj/19.11.2435 PMC 212767 · doi ↗ · pubmed ↗
- 5Bannister, L. , and G. Mitchell . 2003. “The Ins, Outs and Roundabouts of Malaria.” Trends in Parasitology 19, no. 5: 209–213.12763426 10.1016/s 1471-4922(03)00086-2 · doi ↗ · pubmed ↗
- 6Batinovic, S. , E. Mc Hugh , S. A. Chisholm , et al. 2017. “An Exported Protein‐Interacting Complex Involved in the Trafficking of Virulence Determinants in Plasmodium‐Infected Erythrocytes.” Nature Communications 8: 16044.10.1038/ncomms 16044 PMC 550813328691708 · doi ↗ · pubmed ↗
- 7Bertiaux, E. , A. C. Balestra , L. Bournonville , et al. 2021. “Expansion Microscopy Provides New Insights Into the Cytoskeleton of Malaria Parasites Including the Conservation of a Conoid.” P Lo S Biology 19, no. 3: e 3001020.33705377 10.1371/journal.pbio.3001020 PMC 7951857 · doi ↗ · pubmed ↗
- 8Blisnick, T. , M. E. Morales Betoulle , J. C. Barale , et al. 2000. “Pfsbp 1, a Maurer's Cleft Plasmodium Falciparum Protein, Is Associated With the Erythrocyte Skeleton.” Molecular and Biochemical Parasitology 111, no. 1–2: 107–121.11087921 10.1016/s 0166-6851(00)00301-7 · doi ↗ · pubmed ↗