Introduction

Malaria remains a significant global health problem, and the most severe parasite associated disease. Worldwide, 247 million malaria cases and 619,000 deaths were estimated in 2021 due to malaria, mainly caused by Plasmodium falciparum1. In the last decades, many efforts to develop vaccines against malaria have been employed2,3,4, but until a highly efficacious vaccine has been developed, active drug treatment to cure malaria infections remains a central part of malaria control5,6. However, historically several antimalarials have been lost to clinical use by the repeated emergence and spread of drug resistance, and the current frontline antimalarial, artemisinin, is now also under threat1. A better understanding of malaria biology and pathogenesis is needed to guide the search for new drugs and vaccines.

In P. falciparum infected red blood cells (iRBCs), parasite-encoded adhesion ligands are localised to protrusions, so-called knobs, on the iRBC surface, which lead to the formation of rosettes (an aggregation of iRBCs and/or uninfected RBCs)7 as well as the cytoadherence of iRBCs to endothelial cells, allowing the parasite to avoid spleen clearance mechanisms, and causing obstruction in the microvasculature8,9. These ligands are also highly variable, driven by a complex process of antigenic variation which allows iRBCs to evade the adaptive immune response10. The formation of knobs is only one aspect of iRBC remodelling, which involves the export of hundreds of P. falciparum proteins to the iRBC11,12, primarily through tubulovesicular network vesicles termed Maurer’s clefts (MC)8,13.

KAHRP (knob-associated histidine rich protein) is widely thought to be the backbone of the cytoadherence machinery14,15,16. Knobs are organized in a spiral-shape ring structure comprising modular KAHRP units that is assembled as the parasite matures, with up to 60 KAHRP units per knob17. KAHRP is a non-essential gene for parasite growth in vitro, but plays a critical role in pathogenesis in vivo18. How KAHRP catalyses the formation of knobs is still not completely understood. It is presumed to interact with additional P. falciparum exported proteins to do so, but a complete map of KAHRP-interacting proteins has not yet been established.

In recent years, proximity-dependent biotinylation techniques have increasingly used to identify protein–protein interactions. One such approach, BioID, uses a promiscuous biotin ligase fused to a protein of interest expressed in living cells, where it forms an amide bond between reactive biotin and lysine residues (K) of proximal endogenous proteins. Biotinylated proteins are then captured by biotin-affinity, and mass-spectrometry based proteomics is used to identify neighbouring and potentially interacting proteins of the BioID tagged targets19,20. This technology has been showing its usefulness in the study of Apicomplexan parasites, including P. falciparum21,22,23, Plasmodium berghei24 and Toxoplasma gondii25,26. For instance, BioID has been used to define the proteomes of the apicoplast, parasitophorous vacuole and rhoptries in P. falciparum21,22,27.

An alternative proximity-dependent biotinylation technique is based on the use of the engineered ascorbate-peroxidase (APEX2)28,29,30. The main difference between BioID and APEX2 lie in the incubation time for the biotinylation reaction. While BioID has a labelling period of 15–18 h, APEX2 has a shorter labelling period of the biotin phenoxyl radical of 1 min and, a tagging radius smaller than 20 nm20,28. APEX2 is ideal to study short-lived events, however, the labelling appear to be more difficult compared to BioID20.

In this study, we aimed to implement APEX2 proximity tagging followed by label-free mass-spectrometry based proteomics to identify candidate interactors of KAHRP, in order to expand understanding of potential KAHRP interactors. We modified the endogenous kahrp gene using CRISPR/Cas9 genome engineering to insert both a C-terminal APEX2 domain and FLAG tag, to allow us to follow its location by immunofluorescence. Carrying out in vivo biotinylation in this strain followed by mass spectrometry to specifically identify only biotinylated peptides, we classified 30 proteins as potential interacting partners for KAHRP. Some of these proteins have previously been shown to localize at MC by immunofluorescence or proteomic analyses, but its evidence as a KAHRP interactor was never investigated. These interactions may occur either in the knobs, or during KAHRP transport to its final subcellular location. In summary, we demonstrated for the first time the feasibility of using APEX2-proximity labelling in P. falciparum.

Results

Construction and validation of KAHRP-FLAG-APEX2 expressing parasites

The plasmids pCC1-KAHRP-FLAG-APEX2 and pDC2-U6-Cas9-gRNAkahrp were co-transfected into the P. falciparum 3D7 strain to generate a transgenic strain in which FLAG and APEX2 tags were inserted at the 3′ end of the endogenous kahrp gene, creating a chimeric KAHRP-FLAG-APEX2 gene (Fig. 1A). The first transgenic parasites were observed after 21 days under drug pressure. Diagnostic PCR confirmed integration of the cassette to the endogenous KAHRP locus—amplicons from the primers PS110/PS102 and PS103/PS114 of 1.1 kb and 1.6 kb, respectively, confirming correct integration (Fig. 1B).

Figure 1
figure 1

CRISPR/Cas9 guided knock-in scheme to tag the Plasmodium falciparum kahrp gene with APEX2. (A) A schematic figure illustrating integration of the cassette from the pCC1-KAHRP-FLAG-APEX2 vector into parasite gDNA, resulting in integration of the FLAG tag and APEX2 domain at the 3′ end of the gene, in frame with the KAHRP ORF. (B) PCR validation of cassette integration. Legend: K (KAHRP-FLAG-APEX2) and C- (negative control).

To test the ability of the integrated APEX2 enzyme to catalyse biotinylation, enriched late-stage KAHRP-FLAG-APEX2 parasites were incubated for 30 min in the presence or absence of Biotin-Phenol (BP+ /−), followed by 1 min of oxidising in the presence or absence of H2O2. Samples were therefore created from four conditions: BP−/H2O2− (both without BP and H2O2); BP−/H2O2+ (without BP and in the presence of H2O2); BP+/H2O2− (in the presence of BP and without H2O2); and BP+/H2O2+ (both in the presence of BP and H2O2)—biotinylation is only expected in the presence of both BP and H2O2. After quenching, iRBCs were solubilised using saponin treatment, which disrupts the RBC membrane but not the parasite membrane, allowing release of the RBC cytosol with its high levels of haemoglobin. The saponin-treated pellet, which contains both parasites and membrane fragments from the iRBC such as Maurer’s Clefts, was then solubilised using RIPA buffer (RIPA-soluble proteins) and separated by NuPAGE.

Anti-FLAG tag antibodies confirmed the expression of the chimeric protein KAHRP-FLAG-APEX2 in all conditions (Fig. 2A), although the chimeric protein ran higher than predicted molecular mass of 100 kDa, most likely due to charged residues in the C-terminus of KAHRP, or the histidine-sequence repeats, as similar repeats are well-known to produce atypical migration of other proteins in P. falciparum31. Multiple bands of smaller size were also detected, suggesting partial proteolytic cleavage. It is not known whether this occurred within the parasite or post-lysis, although it should be noted that proteolytic processing is a common feature in P. falciparum proteins. Wildtype P. falciparum 3D7 strain parasites and a P. falciparum transfectant strain in which FLAG-APEX2 was expressed from an episome (Fig. S1A), unlinked to any endogenous protein, were used as controls and show clearly different sized products in Western blot (Fig. S1B), confirming the bands in Fig. 2A are specific to the presence of KAHRP-FLAG-APEX2.

Figure 2
figure 2

Confirmation of chimeric protein expression and biotinylation activity. (A) Ponceau S stained nitrocellulose membrane (left panel), Western-blot (middle panel, anti-FLAG-HRP dilution at 1:800) and streptavidin-blot (right panel, streptavidin-HRP dilution at 1:10,000) of KAHRP-FLAG-APEX2 parasites treated and/or not treated with BP and H2O2. In all cases, material was generated by saponin lysis of iRBCs, which releases RBC cytosolic contents but retains parasite and exported membrane material, followed by solubilisation of all retained material using RIPA buffer. (B) NuPAGE of material enriched from the KAHRP-FLAG-APEX2 line following biotinylation, lysis with saponin and then RIPA, and purification of biotinylated proteins using streptavidin-coated beads enrichment. Eluted proteins stained by Coomassie-blue were only detected in the material from the BP+/H2O2+ condition (right panel). Legend for all gels/blots: first lane BP−/H2O2− (not treated with BP or H2O2); second lane BP−/H2O2+ (not treated with BP, treated with H2O2); third lane BP+/H2O2− (treated with BP but not treated with H2O2); fourth lane BP+/H2O2+ (treated with both BP and H2O2).

To test the efficacy of the APEX2 biotinylation assay, RIPA-soluble proteins were probed with streptavidin-HRP to label biotinylated proteins. Only the condition in which both the BP substrate and the oxidizing agent H2O2 were present revealed biotinylated proteins, with multiple bands over a large molecular weight range, suggesting biotinylation of multiple proteins (Fig. 2A). In the episome-based FLAG-APEX2 line, biotinylated proteins were also observed by streptavidin-HRP blot only in the presence of BP and H2O2 (Fig. S1B). No signal was observed in the negative controls, in which either BP or H2O2 were omitted in the KAHRP-FLAG-APEX2 line (Fig. 2A) or in wildtype 3D7 parasites (Fig. S1B). Biotinylated proteins were enriched using streptavidin-coated beads from all KAHRP-FLAG-APEX2 samples from all four conditions—again, protein material were observed only in the lane containing the sample treated with both BP and H2O2 (Fig. 2B). This establishes that APEX2-tagged KAHRP catalyses biotinylation of proteins within the parasite, and that this biotinylation is dependent on incubation with H2O2 and the BP substrate.

Localisation of KAHRP-FLAG-APEX2 and biotinylated proteins by confocal immunofluorescence

Immunofluorescence assays (IFA) using anti-FLAG antibodies showed a punctate distribution of KAHRP-FLAG-APEX2 inside the iRBC cytoplasm (Fig. 3A, grey arrows), consistent with previous reports of KAHRP location32. Staining was also visible inside the parasite near the nucleus, potentially representing KAHRP that was being synthesised, however, similar staining was also present with anti-FLAG antibodies in wild-type parasites, suggesting the labelling is non-specific (Fig. 3A). By contrast the punctate distribution of anti-FLAG staining was absent from 3D7 parasites, suggesting that this localisation inside the iRBC cytoplasm is specific to the expression of KAHRP-FLAG-APEX2.

Figure 3
figure 3

IFA of KAHRP-FLAG-APEX2 and biotinylated candidate interactors of KAHRP. (A) 3D7 wild type and KAHRP-FLAG-APEX2 lines incubated with DAPI and anti-FLAG antibodies, followed by detection with Alexa Fluor 555-conjugated anti-mouse antibody. KAHRP-FLAG-APEX2 chimeric protein displayed punctuate localization throughout the parasitized erythrocyte (grey arrows as examples). (B) BP+/H2O2− and BP+/H2O2+ treated KAHRP-FLAG-APEX2 parasites were incubated with DAPI and streptavidin-AF488. As expected, fluorescence, indicating the presence of biotinylated proteins, was observed only in the BP+/H2O2+ treated parasites. (C) BP+/H2O2+ treated KAHRP-FLAG-APEX2 parasites were stained with DAPI, streptavidin-AF488 and anti-MAHRP antibody followed by detection with Alexa-Fluor 546-conjugated anti-rabbit IgG donkey antibody. A clear co-localization between biotinylated proteins (green) and MAHRP (purple) was observed, suggesting that KAHRP interacting partners and MAHRP share the same vesicular compartment in the Maurer’s clefts (merge, white arrows as examples).

To establish the localisation of proteins biotinylated by KAHRP-FLAG-APEX2, we used IFA and fluorescently-tagged streptavidin (streptavidin AF488) to detect proteins that were labelled after incubation with H2O2 and the BP substrate. Structures within the iRBC cytoplasm were present in schizonts treated with BP+/H2O2+, but not schizonts from which H2O2 was omitted, confirming that the labelling was dependent on APEX2 activity (Fig. 3B). Co-localisation of streptavidin-AF488 with antibodies against membrane-associated histidine-rich protein (MAHRP), a specific marker of MC known to colocalize with KAHRP33,34,35, showed that proteins biotinylated by KAHRP-FLAG-APEX2 were partially co-localised to the MC (Fig. 3C, white arrows). Conversely, IFAs using streptavidin-AF488 with the parasite line expressing episome-based FLAG-APEX2 showed biotinylated proteins were primarily localised to the cytosol, and did not co-localise with anti-MAHRP (Fig. S2). In this parasite line, a partial co-localization between biotinylated proteins and the Golgi apparatus at the perinuclear regions of P. falciparum segmented schizonts was observed using an antibody raised against the ERD236,37,38,39 (Fig. S2).

KAHRP-FLAG-APEX2 parasites produce iRBCs that display knobs and bind to human endothelial cells

To confirm that the addition of the FLAG-APEX2 tags did not affect KAHRP function, the presence of knobs in the trophozoite/schizont-iRBCs of the transfectant line was examined by scanning electron microscopy (SEM) compared to the 3D7 parental strain and uninfected RBCs. SEM showed a typical knobbed surface on both 3D7 and transfectant trophozoite/schizont-iRBCs (Fig. 4). This pattern was consistent with previous studies15,18,31, indicating that KAHRP-FLAG-APEX2 is correctly expressed and can support the production of knob structures in the iRBC membrane. To test for any impact on cytoadherence, we incubated human umbilical vein endothelial cells (hUVECs) with both 3D7 and transfectant trophozoite/schizont-iRBCs, and saw adherence of iRBCs from both strains, whereas no adherence was seen when hUVECs were exposed uninfected RBCs (Fig. S3A–C)40. There was no statistical difference in the binding frequency of iRBCs from 3D7 and KAHRP-FLAG-APEX2 transfectant lines to hUVECs (Fig. S3D). Together with the SEM data, this suggests addition of the FLAG-APEX2 tags to the endogenous KAHRP protein has no discernible impact on knob production or function in cytoadherence.

Figure 4
figure 4

Scanning electron microscopy of P. falciparum-infected erythrocytes displaying the knob morphology. Uninfected RBCs (A, B) and Percoll-purified trophozoite/schizont iRBCs of 3D7 (C, D) and KAHRP-FLAG-APEX2 (E, F) lines. Scale bars: (AF) 1 µm.

Proteomic identification of proteins biotinylated by KAHRP-FLAG-APEX2

Biotinylated proteins were enriched using streptavidin magnetic beads from triplicate samples from KAHRP-FLAG-APEX2 line, with material purified in parallel from both the BP+/H2O2+ condition and the BP−/H2O2+ condition to act as a negative control. Proteins were eluted from the beads, precipitated, alkylated and digested with trypsin as described in the Methods, and subjected to LC–MS/MS using an Orbitrap mass analyser. Proteins were identified by searching peptide sequences against the 3D7 proteome, with H. sapiens erythrocyte databank.

40,327 peptides were identified in the KAHRP-FLAG-APEX2 BP+/H2O2+ condition and 35,845 peptides were identified in the KAHRP-FLAG-APEX2 BP−/H2O2+ condition (Table S1). The high number of peptides identified in the pull-downs from both conditions does not appear to correlate with the Western blotting (Fig. 2A) and IFA results (Fig. 3B), which both indicate that biotinylated proteins are present at much higher abundance in the BP+/H2O2+ condition than the BP−/H2O2+ condition. However, this is likely due to non-specific purification of non-biotinylated proteins during enrichment. Although the affinity between streptavidin and biotin substrate is strong41,42, theoretically favouring highly specific enrichment, nonspecific binding of non-biotinylated proteins to streptavidin beads is a known concern, mostly because of the amount and composition of the streptavidin coated beads used for the protocol43.

Given these potential sources of background caused by non-biotinylated proteins being enriched along with biotinylated proteins, we took a deliberately conservative and stringent approach to minimise the risk of false positives. We only considered proteins as potential KAHRP-FLAG-APEX2 interactors if we could detect that protein from at least 2 different peptides, where at least one must be a unique proteotypic peptide44 that contains a variable biotin-phenol (+ 361.14601) modification on Tyr residues, confirming that it was definitively biotinylated. We considered only unique biotinylated peptides by which we could assess the signature of the biotin-phenol fragmented ions on the mass spectra as already reported elsewhere42. This analysis identified a total of 42 non-redundant biotinylated peptides that were found exclusively in the KAHRP-FLAG-APEX2 BP+/H2O2+ condition. These peptides corresponded to 8 H. sapiens erythrocyte proteins and 22 P. falciparum proteins, including KAHRP-FLAG-APEX2 (Table S2).

Remarkably, many proteins observed localizing at MC by IFA or proteomic analyses in previous studies were found biotinylated in our dataset. Among these proteins are ring exported protein-1 (REX1)45, ring exported protein-2 (REX2)46, skeleton-binding protein 1 (SBP1)47, early transcribed membrane protein 10.2 (ETRAMP10.2)48, small exported membrane protein (SEMP1)49 and parasite-infected erythrocyte surface protein (PIESP2)48. It is also noteworthy to mention that host proteins were also tagged by KAHRP-FLAG-APEX2. Among these proteins are α/β-spectrin, Protein 4.1, glycophorin C and band 3. These proteins are intimately associated with RBC cytoskeleton, which could explain their labelling considering KAHRP as main knob constituent17 (Table 1). Once KAHRP is also transported through MC47, it is not at all surprising that other MC proteins may interact with KAHRP at some time point during their export into iRBC membrane. Thus, this study remarks the first report of such direct interactions during in vitro schizogony in P. falciparum development.

Table 1 List of the candidate neighbouring parasite protein partners of the knob-associated histidine rich protein during schizogony in P. falciparum.

Bionformatic analysis of potential KAHRP-interacting proteins

KAHRP is exported through the Plasmodium Translocon of EXported proteins (PTEX) localized at parasitophorous vacuole membrane (PVM) during its transport into the iRBCs12,50,51. Many P. falciparum proteins exported by this mechanism possess a conserved motif (RxLxE/Q/D), frequently referred as the Plasmodium EXport ELement (PEXEL) motif, at their N-terminal end11,12,52, and 8 of the 22 biotinylated proteins from our dataset were predicted to have a PEXEL motif at their N-terminal end (Table 1). Among these proteins are PTP4 and 5, PfEMP1-trafficking proteins 4 and 5 (PF3D7_0730900 and PF3D7_1002100), GBP130 (PF3D7_1016300), MESA (PF3D7_0500800), PIESP2 (PF3D7_0501200) and KAHRP-FLAG-APEX2 (named as its endogenous gene name, knob-associated histidine rich protein, PF3D7_0202000).

Overall, a total of 6 and 12 of the 22 potential parasite KAHRP-interacting proteins were predicted to be secreted by classical or non-classical pathways by SignalP and SecretomeP, respectively. Moreover, 15 and 16 of the 22 potential parasite KAHRP-interacting proteins were predicted to contain at least one transmembrane protein domain by TMHMM and Phobius, respectively (Table S3). Altogether, 21 proteins are predicted to have an extracellular or cell surface localization corresponding to 95% of the 22 potential parasite KAHRP interactors identified. Only protein histone H4 (PF3D7_1105000) was not predicted to be exported by any algorithm used (Table 1).

These 22 potential parasite KAHRP-interacting proteins were also GO-categorized in PlasmoDB (https://plasmodb.org/plasmo/app) to assess the “Cellular Compartment” ontology terms. A total of 24 GO terms for “Cellular Compartment” were returned from our data (Table S4). A significant enrichment was found in the expected GO terms. The most overrepresented GO terms statistically significant in the WordCloud and Bubble plot were related to the host environment (Fig. 5). Together, these bioinformatic approaches establish that the majority of the parasite proteins identified as candidate KAHRP interactors were secreted, exported and/or associated with membranes.

Figure 5
figure 5

GO term annotation. (A) GO Cloud obtained from PlasmoDB (https://plasmodb.org/plasmo/app/) corresponding to the “Cellular Component” terms annotated. (B) Bubble plot from top 10 GO terms with most significant p-values from panel (A).

Discussion

In this work, we have used APEX2-based proximity biotin-labelling and shotgun proteomics to identify proteins that are candidate interactors of P. falciparum KAHRP, the critical component of the knobs that form in the iRBC membrane and play a key role in cytoadherence. The APEX2 promiscuous biotinylation domain was integrated into the P. falciparum genome in-frame with the endogenous KAHRP protein, and electron microscopy and cytoadherence assays both detected no effect of the FLAG-APEX2 tag on either knob formation or function (Fig. 4 and Fig. S3), suggesting that KAHRP remained fully functional. After labelling assays, Western blotting confirmed that a set of proteins were biotinylated in this line only when the substrate BP was added, followed by H2O2 to allow oxidation (Fig. 2A). IFA confirmed that these biotinylated proteins were predominantly found in subcellular structures within the iRBC (Fig. 3C). By contrast, in the FLAG-APEX2 line with episomally expressed APEX2 that is not fused to any endogenous protein and hence is cytosolic, the APEX2 was active (Fig. S1) but the localisation of biotinylated proteins were intracellular (Fig. S2). No labelling was also detected in a wildtype control line (Fig. S1). Based on these data, APEX2 has significant promise to identify interacting proteins in P. falciparum.

Despite these clear results from Western blotting and IFA, enrichment of biotinylated proteins by streptavidin beads followed by identification using mass spectrometry revealed high levels of non-specific background in both conditions of the KAHRP-FLAG-APEX2 transfectant. It was therefore expected that there would be some level of non-specific background detectable by mass spectrometry and not by Western blotting/IFA, but there is no question that the level of background in the mass spectrometry data was a surprise. There is a consensus that the use of more streptavidin beads than necessary can increase the background by providing more surface area for nonspecific binders. Elution is also a bottleneck, specifically the difficulty of detaching biotinylated proteins from streptavidin, given the interaction is exceedingly strong42,43,53.

Given this significant issue with background, we applied highly stringent filters, assigning proteins as potential KAHRP interactors only if they were identified by two peptides, with at least one being unique containing a biotinylated Tyr, in at least one of the test replicates, and in none of the control replicates. From these unique biotinylated peptides, an additional criteria were employed: the presence of MS/MS spectral data of fragmentation ion assignment of dehydrobiotin (m/z 227.08), immonium of the bpY (m/z 497.22) and/or immonium of the bpY with loss of –NH3 (m/z 480.195), as already reported42.

This conservative and stringent approach identified with high confidence only 42 exclusive non-redundant biotinylated peptides from KAHRP-FLAG-APEX2 BP+/H2O2+ condition, representing a total of 22 parasite and 8 host proteins. It is important to be clear on the limitations and outlines of this dataset. Some interacting proteins may have biotinylation sites that are inaccessible, either because their coiled three-dimensional structure means that the tyrosine residues facing inwards, or they are membrane proteins with little of their sequence exposed to APEX2 activity. Moreover, some proteins that are more transient KAHRP partners but were not in the vicinity of KAHRP when the ‘snapshot’ of the APEX2-based biotinylation was carried out may be missing.

These caveats emphasise that there are undoubtedly false negatives, but there are multiple reasons to be confident that our dataset has captured true KAHRP interactors. To reach the host cell membrane, many parasite proteins exported across the parasitophorous vacuole membrane towards MC need to be processed by PTEX machinery51,54. Proteins exported by this mechanism contain a PEXEL motif at N-terminal end12,55. KAHRP possesses a PEXEL motif and is processed by this mechanism12,52, and transiently carried over into MC47. Among our data, beyond KAHRP, several other proteins identified as potential KAHRP interactors also have a PEXEL motif: PTP4 and 5, GBP130, MESA, PIESP2, PHISTa, and Plasmodium exported protein (PF3D7_1353100) (Table 1)11,12. Although we cannot establish whether these proteins were tagged by KAHRP-FLAG-APEX2 before, during or after translocation through PTEX machinery, the presence of proteins widely known to be exported through this machinery increases confidence in the hits identified in our dataset.

KAHRP-FLAG-APEX2 was confirmed by IFA to be localized in punctate structures (likely MC), as were biotinylated proteins catalysed by it (Fig. 3A). As mentioned already, several of the exported proteins we identified as putative KAHRP interactors are known to be found in the MC, such as REX145, REX246, SBP156, ETRAMP10.248, SEMP149 and PIESP248 (Table 1). Interestingly, all these aforementioned proteins are characterized as PEXEL-Negative Exported Proteins (PNEPs)57. Given KAHRP is a true PEXEL-exported protein12,52, it is possible that labelling of these PNEPs indicates that some export PEXEL proteins and PNEPs are shared after PEXEL proteins have exited from the PTEX machinery.

Some of the proteins identified in our dataset, such as REX145, SBP156 and; PTP4 and 511, are already known to be involved in PfEMP1 trafficking. PfEMP1 is often associated with KAHRP, supporting knob formation and/or cytoadherence to the host endothelium18,47. Some reports have indicated that KAHRP and PfEMP1 could interact with each other58,59, but this statement has been controversial. Structurally, PfEMP1 is a transmembrane protein found at the MC periphery prior redistribution to the iRBC membrane35. Moreover, the intracellular portion ATS (acidic terminal segment) at C-termini from PfEMP1, supposedly responsible to create mechanical connections once in the iRBC membrane, does not bind to KAHRP60. These topological and biochemical features of PfEMP1 would be reasonable to explain its absence in our dataset, once KAHRP-FLAG-APEX2 would not biotinylate it.

Other structures have been implicated in parasite protein trafficking within the infected erythrocyte, such as transient J-dots/Maurer’s clefts (JAM) complex trafficking61. One potential KAHRP interactor identified in our data, the Plasmodium exported protein (PF3D7_0702500), has been identified as resident both in the MC and also in the JAM complex61. In addition, KAHRP has also been observed in proteomic analysis of extracellular vesicles (EVs) released from ring-stage iRBCs, as have other putative KAHRP interactors found in our dataset such as GBP130, histone H4 and MESA62.

Several host proteins were also confirmed to be biotinylated in our dataset possibly concentrated in assembled knobs (Table S1). For instance, KAHRP was already reported to interact with domains 10–14 of β-spectrin, proximal to the spectrin-ankyrin complex16, and N-5 fragment (repeat 4) of α-spectrin63. Both α and β-spectrin were found biotinylated in our dataset (Table S1). It is also known that spectrin interact with Protein 4.164, and Protein 4.1 is also biotinylated by KAHRP-FLAG-APEX2 in our data set, suggesting a proximity tagging that includes the 4.1-spectrin-KAHRP network in knobs and adjoining erythrocyte cytoskeleton. In addition, other P. falciparum proteins that are biotinylated in our findings have previously been shown to interact with Protein 4.1, specifically Plasmodium exported protein (PHISTa, PF3D7_0402000)65 and mature parasite infected erythrocyte surface antigen (MESA)66,67. Two additional erythrocyte proteins, glycophorin C and band 3, which are known to form a direct or indirect bridge with Protein 4.1 and spectrin, respectively68,69, were also biotinylated by KAHRP-FLAG-APEX2. These findings confirm an intimate relation between KAHRP-FLAG-APEX2 and cytoskeleton proteins from host, corroborating KAHRP location and knob architecture in P. falciparum-infected erythrocytes.

In summary, while the biotin-based enrichment used clearly had issues of non-specific binding, our stringent protein identification approach established a list of high confidence list of potential KAHRP interacting proteins that include multiple known exported proteins, as well as host proteins that are known to be enriched in knobs. A schematic representation of how these proteins could interact in spatial and temporal context with KAHRP, either stably (for example by co-assembly into knobs) or transiently (by co-export using different cellular pathways) is proposed in Fig. 6.

Figure 6
figure 6

Diagram of potential cellular locations for interaction between KAHRP and parasite and host cell partners during protein export. Based on our proteomic data, APEX2-tagged KAHRP has biotinylated a set of proteins. Different trafficking routes throughout iRBC, including translocation via the PTEX complex and possibly alternative export pathways (e.g. PNEP proteins) are proposed to be used by KAHRP until its destination at knobs under the RBC membrane. Identified biotinylated proteins are labelled by a red circle with the letter B. PfEMP1, the main partner of KAHRP is described as a periphery protein found in the MC and its ATS domain does not interact with KAHRP. These topological features probably hampered APEX2 biotinylation. Other MC proteins, such as SBP1, PIESP2, REX1, REX2, SEMP1 and ETRAMP10.2 were found biotinylated in this study in agreement with the literature. Moreover, host proteins such as α/β-spectrin, protein 4.1, glycophorin and band 3 that compose cytoskeleton erythrocyte proteins were also biotinylated by KAHRP-APEX2, confirming the involvement of these proteins on the knob architecture. It is known that MC and J-dots comprise the JAM trafficking complex. One protein from our dataset found in this trafficking hub connecting these structures is the Plasmodium exported protein (PF3D7_0702500). Of special note, KAHRP and other proteins, such as GBP130, histone H4 and MESA, which were also found in our dataset as biotinylated, were reported to be in extracellular vesicles from ring-stage parasites, suggesting that protein export mechanisms in P. falciparum is much more complex than presumed.

Concluding remarks

Protein export in malaria parasites is undoubtedly complex, and is still a field of intense exploration across Plasmodium species. Maurer’s Clefts have been widely studied and viewed as a primary structure involved in protein export12,13,33,34,35,70, but recent reports have shown that alternative mechanisms exist61,62,71,72. A better understanding of the protein composition of these pathways is needed, and APEX2 and BioID strategies to establish protein proximity represent technological breakthroughs that could help with this19,28,29,30. Tagging endogenous KAHRP with APEX2 resulted in the biotinylation of parasite proteins already described as exported into the RBC, suggesting that they are neighbouring KAHRP in a given time point. Among our dataset there are likely both direct and transient interactors with KAHRP, which may help to give a broader comprehension in protein export. The study also clearly identifies the methodological challenges concerning biotinylated protein/peptide elution in these proximity-tagging approaches42,43,53,73,74. Overall, our work has showed for the first time that APEX2 technology can be used for tagging a protein target in P. falciparum during the erythrocyte cycle and that while it has broad potential, rigorous controls and stringent analysis are essential to interpretation.

Materials and methods

P. falciparum culture conditions

P. falciparum strain 3D7 parasites were originally obtained from Malaria Research and Reference Reagent Resource Centre (available at: www.mr4.org) and were continuously maintained in O+ human erythrocytes as previously described75. All methods were carried out by relevant guidelines and regulations. The use of erythrocytes from human donors for parasite culture was approved by the Ethics Committee of NHS Cambridgeshire 4 Research (15/EE/0253) and all donors supplied written informed consent.

Generation of transfection vectors

Vectors were designed using SeqBuilder software (DNASTAR), and a two-plasmid CRISPR-Cas9 system was used. The gRNA and Cas9 were expressed using pDC2-cam-Cas9-U6-sgRNA-yDHODH (yeast dihydroorotate dehydrogenase)76,77. To target specifically the kahrp gene, the first ranked gRNA target site localised after the stop codon was identified using CHOPCHOP (https://chopchop.cbu.uib.no) and confirmed by sgRNA Designer (http://www.broadinstitute.org/rnai/public/analysis-tools/sgrna-design). The backbone plasmid pCC1-hDHFR was used to build the targeted vector. FLAG-APEX2 sequence was amplified from pcDNA3APEX2-NES plasmid (Addgene plasmid # 49386)78. The hsp86 3’ UTR sequence was amplified from pINT plasmid79. The final assembled vector pCC1-FLAG-APEX2-HSP86-3’UTR was obtained in a single Gibson assembly reaction The Homology repair regions to target the insert to the 3’-end of kahrp were amplified from P. falciparum 3D7 genomic DNA, which was extracted and purified using a QIAamp DNA blood mini kit (QIAGEN). SacI and AflII digested homology region 1 (HR1) and EcoRI and NcoI digested homology region (HR2) were subcloned into multiple cloning site 1 and 2, respectively, of pCC1-FLAG-APEX2-HSP86-3’UTR. The resulting vector contained the drug-selectable marker hdhfr (human dihydrofolate reductase), which confers resistance to the antifolate drug WR9921080. All primers used for cloning procedures in this study are listed in the Table S5. Plasmids for P. falciparum transfection were obtained from transformed E. coli TOP10 cells containing either pCC1-FLAG-APEX2-HSP86-3’UTR-KAHRP-HR1-HR2 or pDC2-cam-Cas9-U6-sgRNAkahrp-yDHODH and purified using Maxiprep kit (QIAGEN).

Transfection of parasites

For the transfection in P. falciparum, 50 µg of each plasmid was co-precipitated using 3 volumes (v/v) of ethanol and 10% (v/v) of the initial volume of DNA solution of 3 M sodium acetate pH 5.2. The DNA pellet was resuspended and mixed in 100 µL of buffer P3 (Lonza) with 4 µL ATP (625 mM). A total of 100 µL of ring stage-iRBCs with a parasitemia higher than 10% were pelleted and washed in cold sterile cytomix81. The washed iRBCs were resuspended in the cytomix containing both DNA plasmids and electroporated with Lonza Nucleofector 4D using programme P3/CM-15082. After 3 weeks of drug pressure using 1.25 nM of WR99210 (Jacobus Pharmaceuticals Company, Inc.) for pCC1 selection and 1.5 µM of DSM-1 for pDC2 selection82, the first parasites were seen through blood smears. DNA extraction was performed using QIAamp DNA Blood Mini Kit (QIAGEN), and the cassette integration into the specific loci of the genomic DNA were assessed by PCR.

Biotinylation assay

Biotinylation assays were carried out using both wildtype control strains, and the KAHRP-APEX2 tagged line. Parasites were routinely synchronized using sorbitol treatment. Before biotinylation around 2 mL of RBCs at 5% parasitemia (corresponding to 1 × 109 iRBCs) of predominantly late stages (trophozoites/schizonts) were enriched by centrifugation over 63% (v/v) isotonic Percoll cushion (Percoll Plus 17-5445-01, GE Healthcare), with the interface, which contains primarily late-stage parasites, isolated. This parasite fraction was then divided in equal volumes between the different test conditions in order to have the same cell number in all conditions. The enriched parasites were resuspended in complete RPMI medium without phenol red, supplemented with or without 250 µM biotin-phenol (Iris Biotech, CDX-B0270), and incubated for 30 min at 37 ºC, 5% CO2 to allow biotinylation to occur. After the incubation period each condition was split in two, with half having 1 mM H2O2 added (final concentration; GE Healthcare) for 1 min, and half not. Samples from a total of four conditions were therefore generated—in the presence or absence of biotin-phenol (BP), and each one with or without H2O2. Immediately after H2O2 addition the reactions were quenched with an antioxidant solution containing 10 mM sodium azide (Sigma-Aldrich), 10 mM sodium ascorbate (Sigma-Aldrich) and 5 mM of Trolox (Sigma-Aldrich) in PBS. A fraction of the iRBCs were pelleted to make blood smears used for immunofluorescence assays. The remaining iRBCs were lysed with 0.1% (w/v) saponin (Sigma-Aldrich) in PBS containing EDTA-free protease inhibitor cocktail (cOmplete™, Roche), and the pellets (which contain both intact parasite and membrane material from the iRBC including Maurer’s Clefts) were frozen for subsequent use in Western blots and proteomics. Material from all four conditions were used in Western Blots, but only two were then used for subsequent pull-down of biotinylated proteins and proteomic analysis: BP+/H2O2+ and BP−/H2O2+.

Protein extraction and pull-down of biotinylated proteins

For Western blotting and proteomic analyses, frozen saponin parasite pellets were thawed and then lysed by using 4–5 volumes of RIPA buffer containing 50 mM Tris–HCl pH 7.0 (Invitrogen™), 150 mM NaCl (Sigma-Aldrich), 0.1% (w/v) SDS (Invitrogen™), 0.5% (w/v) sodium deoxycholate (Sigma-Aldrich), 1% (v/v) Triton X-100 (Sigma-Aldrich), EDTA-free protease inhibitor cocktail (cOmplete™, Roche), 10 mM sodium azide (Sigma-Aldrich), 10 mM sodium ascorbate (Sigma-Aldrich) and 5 mM of Trolox (Sigma-Aldrich) in PBS28,29,30. The lysates were incubated for 30 min, 4 ºC, followed by a centrifugation at 15,000 × g, for 20 min, 4 ºC to remove insoluble material26. Biotinylated proteins were enriched from the soluble fraction using Pierce™ streptavidin magnetic beads (Thermo Scientific). Fifty microliters of slurry beads were added per sample and incubated for1 hour with mild agitation at room temperature. The beads were separated from the flowthrough by using MagJET separation rack (Thermo Scientific) and washed as follow: (1) 2 × 1.4 mL with RIPA buffer; (2) 1 × 1.4 mL with RIPA buffer with 1 M of urea and (3) 2 × 1.4 mL with RIPA buffer. Biotinylated proteins were eluted from the beads using 100 µL of a solution containing 20 mM DTT (GE Healthcare), 2 mM of biotin (Sigma-Aldrich) and NuPAGE™ LDS Sample Buffer supplemented with NuPAGE Sample Reducing Agent and NuPAGE™ Antioxidant (Invitrogen™), heated at 95 ºC, for 5 min28,29,30.

NuPAGE™, Western-blot and streptavidin-blot

NuPAGE™ Novex™ 4–12% Bis–Tris protein gels in MES buffer (Invitrogen™) were used to separate 15–20 µg of RIPA soluble proteins, before blotting onto nitrocellulose membrane (GE Healthcare). After transfer, the membranes were stained with 0.1% (w/v) Ponceau S (Sigma-Aldrich) in 5% (v/v) of acetic acid (Sigma-Aldrich). After blocking with 5% (w/v) skimmed milk, nitrocellulose membranes were incubated with anti-FLAG coupled to horseradish peroxidase (HRP; Sigma-Aldrich; 1:800 dilution) or with streptavidin-HRP (Invitrogen™; 1:10,000 dilution). Blots were developed with Amersham ECL Prime chemiluminescence substrate (GE Healthcare) and fluorograms were collected.

Immunofluorescence assays

Smears from KAHRP-APEX2 or WT parasites enriched for late stages were prepared on glass slides and dried for at least 30 min before fixing with 4% (v/v) paraformaldehyde (Sigma-Aldrich) for 20 min and washing once with PBS. Cells were permeabilized with 0.1% (v/v) Triton X-100 (Sigma-Aldrich) in PBS and blocked for 3 h at 8 °C using 3% (w/v) BSA in PBS. To identify the location of the KAHRP-APEX2 fusion, the parasites were probed with mouse monoclonal anti-FLAG M2 antibody (Sigma-Aldrich, F3165; 1:800 dilution) for 1 h at room temperature in blocking solution followed by staining with an Alexa Fluor 555-conjugated anti-mouse IgG goat antibody (Invitrogen™, A21424; 1:800 dilution). To identify proteins biotinylated following the biotinylation assay, slides were co-probed with Alexa-Fluor 488-conjugated streptavidin (Invitrogen™, S11223; 1:800 dilution) and IgG rabbit antibodies directly raised against PfMAHRP or PfERD2 (MRA; 1:200 dilution for both) for 1 h at room temperature in blocking solution. After three washes in PBS, cells were incubated with Alexa-Fluor 546-conjugated anti-rabbit IgG donkey antibodies (Invitrogen™, A10040; 1:800 dilution). After a final wash, the slides were mounted with anti-fade mounting medium with 4′,6-diamidino-2-phenylindole dihydrochloride (DAPI, Molecular Probes). Confocal images were acquired with a LSM510 laser scanning confocal microscope (Zeiss).

Scanning electron microscopy (SEM)

Trofozoites and schizonts from both 3D7 and KAHRP-FLAG-APEX2 lines were enriched by using Percoll cushion 63% (v/v) (Percoll Plus 17–5445-01, GE Healthcare). Knob selectivity was not performed as used in previous studies15,83. Cells were then washed with PBS and fixed with 2.5% (v/v) glutaraldehyde in 0.1 M sodium cacodylate buffer for 1 h. After fixation, cells were conditioned in 0.1 M sodium cacodylate. Following the procedure, cells were fixed in contrast solution (0.1% (w/v) OsO4 in 0.1 M sodium cacodylate buffer) at room temperature. After fixation, cells were washed with distilled water. Coverslips were prepared with poly-l-lysine to attach trophozoites/schizonts-iRBCs, followed by acetone gradient dehydration (50%, 70%, 90% and 100%). Critical drying of samples was performed by using CPD 030 critical point dryer (BALZERS, USA). Dried samples were mounted on suitable sample holders using conductive adhesive and taken to the SCD 500 metalizer (Leica, Germany). All analyses were performed in the JSM-7001F (JEOL, Japan) SEM microscope, and the images were collected at an increase of 10.000 × with 15.0 kv.

Cytoadherence assay of P. falciparum-iRBCs to the human endothelial cells

hUVEC cultures were maintained in DMEM-F12 (Gibco™) medium, supplemented with 10% (v/v) fetal bovine serum (FBS, Gibco™) and 1% (v/v) penicillin/streptomycin solution (Sigma-Aldrich), at 37 ºC in 5% CO2 atmosphere. hUVEC cells were cultured until a confluence of about 70–80% were achieved. Only trophozoites/schizonts P. falciparum-iRBCs were used, obtained by Percoll enrichment as already described. A total of 5 × 103 of hUVEC cells and ~ 50% parasitaemia of P. falciparum trophozoites/schizonts (adjusted for 0.5% (v/v) haematocrit) were used for cytoadherence assay. Briefly, hUVEC cells were seeded into 8-well Lab-Tek CC2 chamber slides (Nalge Nunc International) coated with 1% (w/v) gelatin (Sigma-Aldrich) during overnight. hUVEC cells were washed once with binding medium (RPMI medium supplemented with 0.5% (w/v) AlbuMAX II (Gibco™) and cell attachment was observed by PrimoVert invert microscope (Zeiss). Next, P. falciparum trophozoites/schizonts were seeded over hUVEC cells at 37ºC in 5% CO2 atmosphere, during 1 h. After incubation, cells were immediately washed four times with binding medium to remove unattached cells. hUVEC and P. falciparum-iRBCs were fixed with methanol and stained with InstantProv hematological panotic kit (NewProv) and observed in a 1000× oil immersion lens. The same protocol was performed by using uninfected red blood cells (uRBCs) as negative control. All assay was performed in four independent biological replicates. Parasite-like bodies were counted onto direct hUVEC membrane association40. For statistical comparisons, GraphPad Prism v. 5.04 was used for normality test, Mann–Whitney test and graphical representation. A comparison was statically significant when p value was ≤ 0.05.

Proteomic analysis

After confirmation by streptavidin-HRP blot that the KAHRP-FLAG-APEX2 line was producing biotinylated proteins, independent biological triplicate samples from the conditions BP+/H2O2+ and BP−/H2O2+ were obtained, with the same independent triplicates samples from the same conditions also obtained from the wildtype 3D7 parasite strain as an additional control. All samples were enriched using streptavidin beads as described above, and eluted material precipitated with 4 volumes of cold acetone to the sample in a proportion of 80:20. The protein pellet was washed twice with cold ethanol 70% (v/v), followed with the air drying in the fume hood84,85,86. The protein pellet was then resuspended in 25 mM triethylammonium bicarbonate buffer (TEAB, Sigma-Aldrich), reduced with 5 mM dithiothreitol (GE Healthcare) for 30 min, at 55 ºC, followed by alkylation of disulfide bonds with 14 mM iodoacetamide (Sigma-Aldrich) for 40 min in the dark, at room temperature. Next, the sample was digested with trypsin gold (Promega) at 37 ºC, for 18 h, then desalted using a homemade desalting system of packed P-200 C18 columns (3 M™ Empore™) and Poros OligoR3 resin (Applied Biosystems™)87. The eluted peptides were dried in SpeedVac centrifuge (Eppendorf™) using LoBind® tubes (Eppendorf™) and stocked at -20 ºC until the day of usage.

Liquid chromatography (LC) was performed with an RSLCnano system (Ultimate 3000, Thermo Scientific) coupled online to an Orbitrap Exploris 480 mass spectrometer (Thermo Scientific). Peptides were trapped on a C18 column (75 μm inner diameter × 2 cm; nanoViper Acclaim PepMap™ 100, Thermo Scientific) with buffer A (2/98 MeCN/H2O in 0.1% formic acid) at a flow rate of 3.0 µL/min over 4 min. Separation was performed on a 50 cm × 75 μm C18 column (nanoViper Acclaim PepMap™ RSLC, 2 μm, 100 Å, Thermo Scientific) regulated to a temperature of 40 °C with a linear gradient of 3% to 29% buffer B (B: 100% MeCN in 0.1% formic acid, A: 100% H2O in 0.1% formic acid) at a flow rate of 300 nL/min over 91 min. MS full scans were performed in the ultrahigh-field Orbitrap mass analyzer in ranges m/z 375–1500 with a resolution of 120,000 at m/z 200. The top 20 most intense ions were subjected to Orbitrap for further fragmentation via high energy collision dissociation (HCD) activation and a resolution of 15,000 with the auto gain control (AGC) target set to 100%. We selected ions with charge state from 2+ to 6+ for screening. Normalized collision energy (NCE) was set at 30 and the dynamic exclusion of 40 s.

For identification, the data were searched against the P. falciparum strain 3D7 proteome databank (Proteome UP000001450, from August, 2021) with 5,456 entries downloaded from Uniprot (https://www.uniprot.org/). To this file, 438 entries that contained “red blood cells”, “hematopoiesis” and “erythrocytes” associated key words from the Homo sapiens proteome databank (Proteome UP000005640, from August, 2021) were manually added to identify potential erythrocyte proteins, and the KAHRP-FLAG-APEX2 chimeric sequence was also manually added. Enzyme specificity was set to trypsin and a maximum of two miss cleavages sites were allowed. Carbamidomethylation of cysteine (+ 57.02) was set as fixed modification. Oxidized methionine (+ 15.99), biotinyl-tyramide of tyrosine residues (+ 361.14601) and N-terminal acetylation (+ 42.01) were set as variable modifications42. Maximum allowed mass deviation was set to 10 ppm for monoisotopic precursor ions and 0.02 Da for MS/MS peaks. The resulting files were further processed using PEAKS Studio 7.0. FDR was set to 0.01 at the peptide level for the whole study. To determine a protein as “biotinylated”, a manual selection of 2 peptides, with at least one being unique, was used44,88. All identified proteins must have at least one unique peptide with a variable biotin-phenol (+ 361.14601) modification on Tyr residues in at least one of the test replicates. To ensure that all unique biotinylated peptides used to identify KAHRP candidate interactors are indeed tagged by biotinylation, a manual screening was performed to detect through MS/MS spectra data from fragmentation ions signature of dehydrobiotin (m/z 227.08), immonium bpY (m/z 497.22) and/or immonium bpY with loss of –NH3 (m/z 480.195), as described previously42. All intensities of these fragmentation ions were also listed.

Bioinformatic analysis

Proteins predicted to contain the export PEXEL motif (RxLx[EQD]) were obtained from PlasmoDB (https://plasmodb.org/plasmo/app). Overrepresented Gene Ontology (GO) terms were predicted by GO Cloud tool through PlasmoDB (https://plasmodb.org/plasmo/app) databank. The bubble plots from those enriched GO terms were obtained from R programming language by using ggplot2 package89. The potential of identified proteins to be secreted and/or exported to the extracellular environment of all proteins were assessed using the SignalP90,91 and SecretomeP WebServer92,93. For transmembrane protein prediction, TMHMM94 and Phobius95 were additionally employed.