All experiments were performed in accordance with the UK Animals (Scientific Procedures) Act 1986 (PPL 70/8326 and PADD88D48) and were approved by The Francis Crick Institute Ethical Committee. V(D)J recombination activation gene RAG1 knockout (RAG1^-/-) (Mombaerts et al., 1992) on a C57BL/6 background and C57BL/6 WT mice were obtained from the specific-pathogen-free (SPF) unit and subsequently conventionally housed with autoclaved cages, bedding and food at the Biological Research Facility (BRF) of the Francis Crick Institute. Experiments were performed with six- to eight-week-old female mice under reverse light conditions (light 19:00pm-07:00am and dark 07:00am-19:00pm), at 20-22°C.

Cloned lines of P. chabaudi chabaudi AS parasites were used (originally obtained from David Walliker, University of Edinburgh, UK). To initiate infection, mice were injected intraperitoneally (i.p.) with 1 x 10⁵ infected erythrocytes or submitted to the bites of 25 infected mosquitoes, as previously described (Spence et al., 2012). Parasitaemia was monitored by the counting of blood stage parasites in air dried, methanol fixed, and Giemsa (Sigma)-stained thin tail-blood smears.

Protein sequences from all members of the S7 and L1 PIR clades were collated and analysed via the MEME suite of bioinformatics software to identify the top ten most conserved peptide motifs within those clades. From these motifs, those located within the C-terminal hydrophobic domain of the proteins were excluded from further analysis. The remaining motif sequences were aligned to their target and off-target PIR clades, to choose those that match as many proteins as possible within the target clade and vice versa. The original sequences, taken directly from the candidate PIR proteins, were taken forward. L1 motif 1: CKTNYERINALGAYLY; L1 motif 2: ACTLLREVDAYFNNE; S7 motif 1: CEKCSKDAKEFVNKYNELN; S7 motif 2: CHSYDEMISSAVLFF. Both L1 motifs match the N-terminal of the L1 clade PIRs (39% and 76% conserved among all L1-clade members, respectively), whereas one of the S7 motifs aligns at the N-terminal (58% conserved among all S7-clade members) and the other at the C-terminal (45% conserved among all S7-clade members) of the S7 clade PIRs ( Figure 1A ). The proportion of L1 and S7 clade PIRs targeted by each peptide are summarised in Supplementary Table S1 . BlastP hits and scores are listed in Supplementary Table S2 .

The peptides were commercially generated and chemically coupled to a KLH carried protein in order to stabilize the peptide and elicit a stronger and more persistent immunological response. The conjugated to KLH peptides were then used to immunise rabbits and generate the anti-S7 and anti-L1 specific antisera (Cambridge Research Biochemicals, Billingham, UK). Their immunoreactivity to the peptides was verified using standard Enzyme-Linked Immunosorbent Assay (ELISA) procedures. To maximise the proportion of PIRs belonging to the same clade that are recognised by the individual peptide antisera (named anti-S7 motif 1 peptide antiserum, anti-S7 motif 2 peptide antiserum etc), we pooled together the two antisera generated against the two individual motifs that belong to each clade (named anti-S7 and anti-L1 peptide antisera). The individual peptide antisera were mixed proportionally to guarantee similar levels of reactivity with the PIR proteins.

Synthetic genes encoding a panel of P. chabaudi PIR proteins (PCHAS_1200500, PCHAS_1100300, PCHAS_0600600, PCHAS_0400300 and PCHAS_1300101) were cloned into the AscI (NEB) and NotI (NEB) restriction sites of the pTT3 vector (Crosnier et al., 2013) that carries a C-terminal His6 tag. Putative C-terminal hydrophobic domains were removed, and putative N-linked glycosylation sites were removed through mutation of the S or T residue of the NxS/T motif to alanine.

Protein expression and purification were carried out as previously described (Harrison et al., 2020). Briefly, all plasmids were transfected into HEK293F cells (Thermo Fisher Scientific) using polyethylenimine, and after five days (density of about 2.5x10⁶ cells/mL) harvested by centrifugation at 5,000 × g. The supernatant was buffer-exchanged into 20 mM HEPES pH 7.5, 150 mM NaCl, and 20 mM imidazole by tangential flow filtration, and the protein was purified by immobilised metal affinity chromatography using Ni²⁺-NTA resin, followed by size exclusion chromatography using a Superdex 75 10/300 column (GE Healthcare Life Sciences).

PCHAS_1200500, PCHAS_1100300, PCHAS_0600600, PCHAS_0400300 and PCHAS_1300101 were purified as described above and re-solubilised in 1 x PBS. For the preparation of parasite protein lysates, whole blood was depleted of leukocytes by filtration (Plasmodipur, EuroProxima) and red cell debris by saponin lysis (Sigma) and centrifugation. Purified parasites were then lysed in RIPA buffer (Sigma) supplemented with 1% SDS (ThermoFisher) and 10 μg/mL (1X) Protease Inhibitors (PI) (cOmplete EDTA free, Roche). Approximately 5 μg/mL in 1 x Phosphate Buffered Saline (PBS) (pH 7.4) of each recombinant protein and 50 μg/mL in 1 x PBS of parasite protein lysate were propped with the titrated (two-fold serial dilutions) rabbit anti-peptide antisera, using standard ELISA assay procedures (Spence et al., 2013). Starting concentrations of the antisera were anti-S7 motif 1 peptide antiserum (1:500), anti-S7 motif 2 peptide antiserum (1:500), anti-L1 motif 1 peptide antiserum (1:500), anti-L1 motif 2 peptide antiserum (1:500), anti-S7 peptide antiserum (pooled motif 1 and 2) (1:1000), anti-L1 peptide antiserum (pooled motif 1 and 2) (1:1000), anti-S7 pre-immune serum (1:1000) and anti-L1 pre-immune serum (1:1000) in 1% (w/v) BSA, 0.3% (v/v) in 1x PBS, incubated at 37°C for 1.5 h. Following three washes with 0.25% (v/v) Tween-20 (Sigma) in 1 x PBS, wells were incubated with goat anti-rabbit IgG (H+L) alkaline phosphatase conjugated secondary antibody (Southern Biotech) for 1 h at 37°C. Samples were developed with 1 mg/mL 4-Nitrophenyl phosphate disodium salt hexahydrate (PNPP; Sigma) and OD was measured at 405 nm on a Tecan Infinite M1000 plate reader (LifeSciences). Data were analysed with Magenta software (LifeSciences).

For the preparation of P. chabaudi membrane fractionated samples, magnetically enriched infected erythrocytes at the trophozoite-stage were pelleted in aliquots of 2.5 x 10⁶ iRBC and then resuspended into approximately 20 μL (10 volumes of the pellet volume) in 1 x PBS containing 4 haemolytic units of activated Streptolysin-O (SLO) (Sigma), as previously described (Jackson et al., 2007; Ruecker et al., 2012). Following incubation for 15 min at 37°C, the parasite-containing fractions were separated by centrifugation at 12,400 × g for 10 min at 4°C. Parasite fractions were then washed twice in ice cold 1 x PBS containing 10 μg/mL (1X) PI and resuspended in 10 volumes of 0.15% (w/v) saponin (Sigma) in 1 x PBS for 10 min at 4°C. Following centrifugation, pellets were washed twice in ice cold 1 x PBS/PI and resuspended in 50 mM Tris HCl-PBS-PI (pH 7.4) buffer, mechanically homogenised for 1 min and stored at -80°C until frozen. Immediately upon thawing, the samples were centrifuged and washed twice in ice cold 1 x PBS/PI. The remaining membrane soluble fractions were extracted in RIPA whole cell lysis buffer supplemented with 1% SDS and PI for 20 min at 4°C. Following centrifugation, Laemmli buffer (BioRad) was added to all samples used for western blot analysis, or directly stored at -80°C until further proteomic analysis. Whole protein lysates were prepared as described above. After the last centrifugation, the supernatant was mixed with Laemmli buffer until further western blot analysis. Approximately 70 ng of purified recombinant proteins were mixed with Laemmli buffer until further western blot analysis. For reducing conditions, +3-5% (v/v) 2-mercapthoethanol was added to the sample buffer.

Proteins from all samples were separated on a 4–20% gradient polyacrylamide gel electrophoresis (Mini-PROTEAN TGX™ Precast Protein Gels, BioRad) in Tris-Glycine buffer (BioRad), followed by transfer to nitrocellulose membranes (iBlot Transfer Stack, Invitrogen). The blots were then blocked in 3% (w/v) BSA in 1 x PBS (blocking buffer) and probed with the following antibodies: rabbit anti-S7 motif 1 peptide antiserum (1:1000), rabbit anti-S7 motif 2 peptide antiserum (1:1000), rabbit anti-L1 motif 1 peptide antiserum (1:1000), rabbit anti-L1 motif 2 peptide antiserum (1:1000), rabbit anti-S7 peptide antiserum (pooled motif 1 and 2) (1:1000), rabbit anti-L1 peptide antiserum (pooled motif 1 and 2) (1:1000), rabbit anti-S7 pre-immune serum (1:1000), rabbit anti-L1 pre-immune serum (1:1000), mouse anti-his (1:1000; Novagen), mouse anti-9G8 (1:5000), mouse anti-MSP1 (1:2000), rat anti-Ter119 (1:5000), rabbit anti-PyEXP2 (a kind gift of James Burns, Drexel University College of Medicine, Philadelphia, United States; Meibalan et al., 2015) (1:500) overnight at 4°C. After three washes with 0.05% (v/v) Tween-20 in 1 x PBS, anti-rabbit IgG 680LT IRDye, anti-mouse 800CW IRDye and anti-rat 800CW IRDye secondary antibody (Li-Cor), diluted 1:10000 in blocking buffer, was applied for 1 h at room temperature. Odyssey Imaging system (Li-Cor) was used for detection and image analysis.

P. chabaudi has a synchronous erythrocytic life cycle and therefore blood could be collected at several time points in the 24-hour cycle in order to study the different asexual stages. Thin smears at the ring-, early trophozoite-, mature trophozoite- and late trophozoite/schizont-stages were prepared at 01 h, 16 h, 19 h and 23 h, respectively. For IFAs on in vitro maturated schizonts, infected mouse blood was cultured for five hours as previously described (Spence et al., 2011) and cells were then centrifuged, prior to preparing thin smears. Thin smears were fixed in cold acetone:methanol (9:1) for 10 min at -20°C and air-dried. Fixed cells were then re-hydrated in 1 × PBS, washed once in 1 × PBS, blocked in 3% BSA (w/v) in 1 x PBS for 1 h at room temperature and subsequently incubated with primary antibodies for 1 h at room temperature or overnight at 4°C. Rabbit anti-S7 peptide antiserum, rabbit anti-L1 peptide antiserum, mouse anti-9G8 (Hartz et al., 1993), mouse anti-MSP1 (NIMP23; Boyle et al., 1982), rabbit anti-SBP1-C (a kind gift of Tobias Spielmann, Bernhard Nocht Institute for Tropical Medicine, Hamburg, Germany; De Niz et al., 2016) and rat anti-Ter119 antibodies were used at 1:1000, 1:1000, 1:1000, 1:500, 1:100 and 1:1000 dilution in blocking buffer, respectively. After three washes with 1 x PBS, slides were incubated for 1 h at room temperature with secondary antibodies in blocking solution. Secondary antibodies used were goat anti-rabbit Alexa Fluor 568, goat anti-mouse Alexa Fluor 488 and donkey anti-rat Alexa Fluor 647 (Life Technologies) at 1:1000 dilution in blocking solution. Following three washes with 1 x PBS, nuclei were stained with 1 μg/mL−1 4′,6′-diamidine-2′-phenylindole dihydrochloride (DAPI) (Roche) in 1 x PBS for 5 min at room temperature. Slides were mounted with ProLong antifade mounting medium (ThermoFisher Scientific) and viewed under a Leica SP5 MP confocal laser-scanning microscope. Images were processed by deconvolution using the Huygens software and visualised using the ImageJ x 38 (NIH) software. In these experiments, Ter119 antibodies were used to define the red cell membrane and not direct light microscopic images.

For live cell imaging, parasites were incubated with the primary antibodies for 30 min at room temperature, washed three times in 1 x PBS, and incubated with the secondary antibodies for 20 min at room temperature. The cells were then viewed immediately under a Leica SP5 MP confocal laser-scanning microscope between slide and coverslip, as described above. After acquisition, contrast and brightness levels were optimised using the Image J x 38 software.

Magnetically enriched infected erythrocytes at the trophozoite-stage were pelleted and fixed with 4% (v/v) paraformaldehyde/0.01% (v/v) glutaraldehyde in 1 x PBS for 30 mins at room temperature, followed by a second fixation step using 4% formaldehyde (v/v)/2.5% (v/v) glutaraldehyde in 0.1 M Sorenson’s phosphate buffer (0.1 M Na₂HPO₄/ddH2O, pH 7.4) for 60 min at room temperature.

The cell pellet was then processed using a modified version of the NCMIR protocol (Deerinck et al., 2010). Briefly, the cells were washed in 0.1 M phosphate buffer and post-fixed with 1% reduced osmium (1% w/v OsO4/1.5% w/v K₃Fe(CN)₆) for 60 min at 4°C, and then washed in double distilled water (ddH2O). The cells were incubated in 1% thiocarbohydrazide (TCH) for 20 min at room temperature, rinsed in ddH2O and further fixed with 2% (w/v) osmium tetroxide for 30 min at room temperature. The cells were then stained with 1% (w/v) uranyl acetate at 4°C overnight, washed in ddH2O and stained with Walton’s lead aspartate for 30 min at 60°C. The cells were then washed in ddH2O and dehydrated stepwise using serial dilutions of ethanol: 30% (v/v) at room temperature for 5 min, then 50% (v/v), 70% (v/v), 90% (v/v) and twice with 100% for 10 min each. The cells were incubated in propylene oxide (PO) for 10 min. The cells were subsequently infiltrated with 3:1 mixture of PO : Durcupan resin (Sigma) for 45 min at room temperature, followed by 1:1 and 1:3 mixtures for 45 min each at room temperature, then with 100% Durcupan resin for 48 h. The cell pellet was then polymerised in fresh Durcupan resin at 60°C for 48 h. The sample was cut into 70 nm ultrathin sections using an ultramicrotome (UC7, Leica Biosystems UK) and picked up onto copper mesh grids (Agar Scientific). Images were obtained on a 120 kV transmission electron microscope (TEM) (JEM-1400Flash, JEOL) using a sCMOS camera (Matataki Flash, JEOL).

Investigation of iRBC surface antigens was performed using a flow cytometry-based assay modified protocol from Elsworth et al., 2014. Two microliters of whole blood from C57Bl/6 RAG1^{-/- P} . chabaudi AS infected mice (parasitaemia 15-20%) at the trophozoite-stage, were collected in 20 μL of saline-heparin (25 U/mL, Wockhardt) in U-shaped 96-well plates. After two washes with 1 x PBS supplemented with 2% (v/v) foetal bovine serum (FBS) and 3 mM EDTA (ThermoFisher), cells were either stored at 4°C to be stained later or fixed in 500 μL of 1:1 acetone:methanol for 10 mins at -20°C. Fixed cells were then washed twice in 1 x PBS/FBS/EDTA. Both live and fixed cells were then incubated with the anti-S7 and anti-L1 rabbit peptide antisera diluted in 1 x PBS/FBS/EDTA for 45 min at 4°C. Equal amounts of the anti-S7 motifs 1 and 2 and anti-L1 motifs 1 and 2 were mixed together prior to cell staining, and diluted 1:500 in 1 x PBS/FBC/EDTA. Following three washes in 1 x PBS/FBS/EDTA, cells were stained simultaneously with anti-Ter119 PE conjugated antibody (BioLegend) and goat anti-rabbit Alexa Fluor 568 (Life Technologies) for 20 mins at 4°C in the dark. Both the rat anti-Ter119 PE and the goat anti-rabbit Alexa Fluor 568 antibodies were diluted 1/1,000 in 1 x PBS/FBS/EDTA. Cells were then washed twice and stained immediately in 200 μL Hoechst 33342 (10 μg/mL in 1 x PBS/FBS/EDTA) for 15 mins at 37°C to stain for DNA. After washing twice in 1 x PBS/FBS/EDTA, cells were resuspended in 250 μL 1 x PBS/FBS/EDTA and acquired by flow cytometry on an Amnis^® CellStream^® benchtop flow cytometer (Luminex).

A total of 500,000 single RBC (selected based on their FCS/SSC and FSC/AspectRatio FSC) were analysed for each sample. The gating strategy established for the analysis of the results is described in ( Supplementary Figure S1 ). FlowJo (Tree Star) was used for all flow cytometry analyses. Fixation in acetone:methanol also permeabilises the cells, therefore fixed cells were used as a control to validate antibody integrity. “Fluorescence minus one” (FMO) and single staining controls were included as controls, as well as a negative control sample from a non-infected mouse to define the threshold of positivity for the antibody signal.

Statistical analysis was determined by using Graphpad Prism5 software. The one-way analysis of variance (ANOVA) test was used. P-value of <0.05 was considered as statistically significant.

The two largest clades of P. chabaudi PIRs, the S7s and the L1s, together account for 68% of the PIRs in P. chabaudi. There are 70 S7s and 82 L1s PIRs in the P. chabaudi genome. PIR members of each clade share conserved regions (Brugat et al., 2017) and in order to examine their expression and localisation more holistically, we generated peptide antisera that targeted as many members of each PIR clade as possible.

The localisation of PIR proteins within P. chabaudi-parasitised erythrocytes during the acute- and chronic-phases of infection was investigated, using rabbit polyclonal antisera that were generated against 20-conserved residue sequences of the S7 (acute-associated) and of L1 (chronic-associated) PIR clades, respectively (Freitas-Junior et al., 2000; Otto et al., 2014), the two major pir clades expressed in P. chabaudi blood-stage parasites (Brugat et al., 2017; Little et al., 2021). Two peptide motifs were chosen for each clade ( Figure 1A ), whose sequence had the highest number of BlastP hits within the same clade they were designed for and the lowest off-target activity. In summary, the peptides recognised by the anti-S7 peptide antisera designed against motif 1 and motif 2 are present in 37/70 and 12/70 S7 PIRs, respectively. Similarly, the peptide motif 1 and motif 2 identified for the L1 clade can be found in 49/82 and 29/82 of the L1 PIRs, respectively. Some off-target proteins (i.e. PIRs of a different clade) could not be avoided. Peptide sequences are described in Supplementary Table S1 , and materials and methods. The BlastP hits for each motif designed against each clade are summarised in Supplementary Table S2 .

Antisera to the selected peptides, generated in rabbits (Cambridge Research Biochemicals, Billingham, UK) were tested by ELISA against their respective target peptides ( Figure 1B ). Peptide antisera that were generated against S7 motif 1 and L1 motif 1 recognise their target peptide at high titres, whereas peptide antisera against the S7 motif 2 and L1 motif 2 antigens indicated lower titres. Western blot analysis against the his-tagged recombinant PIR proteins PCHAS_0600600 (L4), PCHAS_1100300 (L1), PCHAS_0400300 (S7), PCHAS_1200500 (S7) and PCHAS_1300101 (S7) confirmed specificity of each antiserum to their target clades (S7 at approximately 33-37 kDa, L1 at approximately 75-151 kDa, and L4 at approximately 40 kDa). Pre-immune rabbit antisera did not recognise any of the recombinant proteins ( Figure 1C ).

To maximise the proportion of PIRs belonging to the same clade that are picked up by the individual peptide antisera, we pooled together the two antisera generated against the individual motifs from each clade. The pooled antisera were then checked for reactivity against the PIR proteins in P. chabaudi WT parasites by ELISA and western blot analysis. ELISA assays using a lysate of P. chabaudi blood-stage parasites showed that both rabbit anti-S7 and anti-L1 sera contained antibodies that recognised proteins from lysates of both MT and serially passaged blood-stage (SBP) parasites ( Figure 1D ), with the overall reactivity being higher for both antisera with the parasite lysate derived from MT blood-stage parasites. The increased reactivity of the antisera with MT blood-stage parasites agrees with the increased RNA expression in MT parasites compared with SBP parasites, previously observed (Spence et al., 2013; Brugat et al., 2017). Western blot analysis of the pooled rabbit anti-S7 and anti-L1 sera against whole protein lysates of P. chabaudi WT MT parasites demonstrated that both antisera are able to recognise naturally produced PIR proteins. The two antisera display differing patterns of bands across the lysates, with S7 and L1 antisera having predominant bands of 33-50 kDa, and 75-151 kDa, respectively, reflecting the expected size of PIR proteins of the S7 and L1 clades ( Figure 1E ). The S7 antisera also interact with a protein of approximately 100 kDa, which may be due to formation of multimers, as purified recombinant protein exhibits a similar migration and interaction ( Figure 1C ). Arrows indicate the major interacting proteins.

The proportion of acetone:methanol fixed/permeabilised P. chabaudi iRBC that stained positively for S7 and L1 PIRs is similar between acute and chronic blood-stage infection and between MT and SBP parasites ( Figure 2A ). However, the proportion of L1-positive iRBC (mean value: 95.33 ± 0.53% SEM) is significantly higher (p=0.0286) in all cases, compared to the percentage of positive cells enumerated for the anti-S7 serum (mean value: 78.95 ± 0.60% SEM). Individual P. chabaudi parasites are thought to co-express multiple pir genes, as well as expressing many pir genes at a population level (Bozdech et al., 2008; Cunningham et al., 2009; Yam et al., 2016). Our data suggest that at each stage, individual parasites also express multiple PIR proteins across different clades.

Subcellular localisation of S7 and L1 PIRs was investigated in fixed/permeabilised RBC from P. chabaudi infections, at both acute- and chronic-phases of the infection (at day 10 and day 35 post infection, respectively) (Brugat et al., 2017) in immunofluorescence assays using the anti-S7 and anti-L1 peptide antisera described above. Trophozoite-stage parasites from MT blood-stage P. chabaudi were initially assessed, as it has been previously demonstrated that pir transcripts are increased in blood-stage parasites following mosquito transmission (Spence et al., 2013; Yam et al., 2016; Brugat et al., 2017). In iRBC taken from an acute-stage MT infection, S7 PIRs are mainly exported into the host’s iRBC cytoplasm where they can be seen as multiple discrete foci. They are also seen, less abundantly, close to the PVM. This pattern changes in iRBC collected at the chronic phase of infection, where in approximately 98% of iRBC, PIRs were localised closed to the iRBC surface and to a lesser extent exported into its cytoplasm ( Figure 2B ; top panel, Supplementary Figure S2 ). L1 PIRs in acute-phase MT blood-stage parasites are localised on or close to the PVM, but with progression to the chronic-phase of infection, they also translocate to the iRBC cytoplasm. Specifically, in approximately 80% of the iRBC, L1 PIRs persist in close proximity to the PVM while also being secreted within the iRBC cytoplasm where they accumulate in granular structures that further co-localise with clone 6 following labelling with 9G8, a monoclonal antibody that recognises a tryptophan-rich protein (PCHAS_0624800) known to be found in the cytoplasm of P. chabaudi iRBC (Hartz et al., 1993) ( Figure 2B ; bottom panel, Supplementary Figure S2 ).

Transcription of pir genes in MT P. chabaudi iRBC differs from that observed in the more commonly used SBP parasites, in both amount and diversity of pir genes expressed (Spence et al., 2013; Brugat et al., 2017). This seems to also apply at the expression level where PIR proteins are more highly abundant in MT parasites compared to SBP, as described above ( Figure 1D ). We therefore asked whether PIRs in P. chabaudi SBP infected erythrocytes also have a different localisation pattern compared to the MT parasitised erythrocytes. Immunofluorescence staining of P. chabaudi SBP iRBC at the acute- and chronic-phases of infection with the anti-S7 and anti-L1 peptide antisera, have shown that S7 PIRs are exported into the iRBC cytoplasm during both the acute- and chronic-phases of a SBP infection ( Figure 2C ; top panel), whereas L1 PIRs are localised in close proximity to the PVM in both acute- and chronic-phases parasites ( Figure 2C ; bottom panel).

Neither S7 nor L1 PIR proteins could be detected at the surface membrane of RBC infected with either MT or SBP blood-stage parasites ( Figures 2B, C ). However, as there are predicted transmembrane domains in PIR proteins (Harrison et al., 2020), S7s may be located on vesicle-like structures within the iRBC cytoplasm and L1s on the PVM. In agreement with this, S7 showed co-localisation with 9G8 monoclonal antibodies, and L1 was found to be localised closely to MSP1, known to be concentrated at the parasite periphery which is consistent with PV localisation (Holder et al., 1992). Some co-localisation of the S7 PIRs with MSP1 was observed in acute-phase MT blood-stage parasites, whereas L1 PIRs seem to translocate on a similar vesicle-like pattern as 9G8 when progressing into the chronic-phase of MT blood-stage parasites. For clarification, the localisation data are summarised in a schematic ( Figure 2D ).

PIR expression in iRBC during acute and chronic P. chabaudi infection described above was determined using only the trophozoite-stage of the parasite. We next examined whether the subcellular localisation of the S7 and L1 PIRs changed during the 24-hour developmental cycle in iRBC taken from acute- and chronic-stages of an MT infection using the S7 and L1 peptide-specific antisera. PIR localisation was examined at rings (01 h), early trophozoites (16 h), mature trophozoites (19 h) and late trophozoites/schizonts (23 h) on blood films directly ex vivo ( Figure 3A ; Supplementary Figure S3 ). However, as P. chabaudi schizonts are normally not present or only at very low levels in peripheral blood because of sequestration in organs and tissues (Gilks et al., 1990; Mota et al., 2000; Brugat et al., 2014), PIR expression was also determined in schizont-iRBC that had been matured in vitro ( Figure 3B ).

S7 PIRs were exported to the iRBC cytosol and appeared as dense granules, in both trophozoite- and ring-stages of both acute- and chronic-phase parasites. In agreement with our observations for trophozoites described in Figure 2 , S7 PIRs appeared to be more closely associated with the iRBC membrane in the chronic-phase of infection throughout the 24-hour cycle. Furthermore, co-localisation of the S7 PIRs with the tryptophan-rich protein recognised by the 9G8 antibody indicates vesicle-like distribution of the S7s across the iRBC cytoplasm ( Figure 3A ; top panel). L1 PIRs were found to be associated with the PVM during the ring- and trophozoite-stages; further supported by their co-localisation with the merozoite surface protein MSP1. With progression to the chronic-phase of infection and during ring-to-trophozoite developmental transition, traces of L1 PIRs were seen in the host cell cytoplasm ( Figure 3A ; bottom panel), validating our previous observations.

In in vitro-matured schizonts, S7 PIRs were found to be exclusively exported beyond the parasite plasma membrane and PVM, and distributed evenly throughout the iRBC cytosol. However, in the few cases where free merozoites were detected, upon schizont rupture, S7 PIRs were found to be highly associated with their cytoplasm as a single vesicle-like structure; about 80% of the released merozoites were found to be positive for S7 expression. Progression from acute- to chronic-phases of infection did not alter S7 PIR localisation ( Figure 3B ; top panel). L1 PIRs in in vitro schizonts from both acute- and chronic-phase were found to be localised within the individual merozoites’ cytoplasm as discrete dots. Upon schizont rupture, L1 expression could still be observed in approximately 95% of the released merozoites possibly suggesting a potential function of the L1s in merozoite egress or invasion. Co-localisation of the L1s with MSP1 was not observed, further suggesting that L1 PIRs are not localised on the merozoites surface ( Figure 3B ; bottom panel). Similar distribution pattern of the S7 and L1 PIRs has been also observed in SBP acute-stage in vitro maturated schizonts ( Supplementary Figure S4 ).

These results show that localisation of both S7 and L1 clades does not change across P. chabaudi 24-hour developmental cycle in the blood. The only difference observed between expression in the different developmental stages was the proportion of iRBC that stained positively with the peptide antisera ( Figure 3C ). The proportion of detectable S7-expressing cells peaked at 19 h (mature trophozoites), after which they rapidly declined with minimum expression levels at 01 h (ring-stage). The proportion of L1-expressing cells did not significantly change between early (16 h) and late trophozoites (19 h), at 97-98% abundance. After the trophozoite-stage, the proportion of L1-positive iRBC decreased to about 77-78% abundance in the ring-stage. Importantly, the proportion of S7+/L1+ iRBC in acute and chronic P. chabaudi parasites did not change indicating the lack of further regulation between acute- and chronic-phase.

Previous staining of MT acute and chronic P. chabaudi blood-stage parasites with the anti-S7 peptide antisera displayed a similar punctuate localisation pattern to that of clone 6 following immunostaining with the 9G8 monoclonal antibody ( Figure 2B ). This vesicle-like pattern of both S7 PIRs and clone 6 is typical of Maurer’s clefts. Maurer’s clefts are sack-like structures spread across the iRBC cytosol of most malaria species (Langreth et al., 1978; Atkinson and Aikawa, 1990; Henrich et al., 2009) and they have been shown to be involved in sorting parasite proteins on the surface of iRBC (Mundwiler-Pachlatko and Beck, 2013). The presence of Maurer’s clefts has not yet been examined in P. chabaudi. By conducting electron microscopy ( Figure 4A ) and fluorescence ( Figure 4B ) imaging analysis of P. chabaudi iRBC at the trophozoite-stage, we investigated whether Maurer’s clefts exist in the cytoplasm of P. chabaudi parasitised erythrocytes, in similar vesicle-like structures as in other malaria species.

Transmission electron microscopy (TEM) analysis revealed membrane structures with sizes ranging from 30 to 150 nm, scattered between the PVM and the erythrocyte membrane ( Figure 4A ). They appear as narrow flattened clefts with either a dense or slightly expanded lucent lumen, suggesting that these vesicles are likely to be Maurer’s clefts. However, to confirm that these vesicle-like structures are Maurer’s clefts, we performed fluorescence microscopy-based imaging of MT blood-stage parasites at the acute- and chronic-phases of the infection following staining with the P. berghei anti-SBP1 antiserum (a gift of Tobias Spielmann; De Niz et al., 2016). PbSBP1 is essential for parasite sequestration (De Niz et al., 2016) and it has been shown to co-localise with IBIS1, a marker protein of previously identified Maurer’s clefts-like structures in P. berghei (Ingmundson et al., 2012). Since there are very few available antibodies for P. chabaudi and P. berghei SBP1 (PbSBP1) shares 64% similarity with its P. chabaudi orthologue, anti-PbSBP1 was used as a marker for identifying Maurer’s clefts in P. chabaudi. Moreover, since staining with the 9G8 monoclonal antibody displays the same localisation pattern as that of the S7 PIRs, both the anti-PbSBP1 antiserum and 9G8 monoclonal antibody were used to co-stain P. chabaudi iRBC. Direct co-localisation experiment using the P. berghei anti-SBP1 and the P. chabaudi anti-S7 peptide antisera simultaneously was not conducted, as both were produced in rabbits. Immunofluorescence assays showed that SBP1 and clone 6 are exported into the cytosol of P. chabaudi acute-stage infected erythrocytes where they were found to be co-localised in multiple discrete foci (83% co-staining of clone 6 and SBP1) ( Figure 4B ), resembling the typical labelling of proteins located in P. falciparum Maurer’s clefts. Both clone 6 and SBP1 were found to have similar localisation patterns in chronic-stage infected erythrocytes, where about 76% of the discrete foci detected were positive for both proteins. Some localisation of clone 6 was also observed close to the periphery of the iRBC in 91% of the examined discreet dots ( Figure 4B ), however our flow cytometry data on unfixed iRBC ( Figures 5B, D ) suggest that 9G8 is not externally located on the iRBC membrane at this stage of the infection. We have previously shown that S7 PIRs, at the chronic phase of infection, are found to be localised as granular structures across the iRBC cytosol and near the iRBC membrane ( Figures 2B , 3A ). However, the use of anti-PbSBP1 to identify P. chabaudi Maurer’s clefts indicates that their localisation does not change in iRBC at the acute and chronic infection ( Figure 4B ). These data suggest that S7 PIRs do not co-localise with the Maurer’s clefts in chronic-phase iRBC and are possibly released as infection progresses to the chronic-phase.

With progression to the chronic-phase of infection and during ring-to-trophozoite developmental transition, traces of L1 PIRs were seen in the host cell cytoplasm to co-localise with clone 6 following 9G8 labelling ( Figure 3A ; bottom panel). This suggests that Maurer’s clefts might also be involved in the trafficking of L1 PIRs but to a lesser extent compared to the S7 PIRs.

It has been suggested that PIR proteins, similar to proteins of other Plasmodium multigene families (Niang et al., 2009; Niang et al., 2014; Saito et al., 2017; Kyes et al., 1999; Blythe et al., 2008), are present on or close to the surface of iRBC (Del Portillo et al., 2001; Cunningham et al., 2005; Pasini et al., 2013; Yam et al., 2016; Harrison et al., 2020). The distribution of flexible, sequence-variable loops across their structures is also characteristic of a protein family undergoing diversification to avoid immunoglobulin-mediated clearance (Harrison et al., 2020). As shown above ( Figures 2B , 3A ), immunofluorescence studies of acetone:methanol fixed P. chabaudi RBC using the anti-S7 and anti-L1 peptide antisera showed that S7 appeared to be associated with vesicle-like structures in the iRBC cytoplasm and L1 PIRs with the PV, in both acute- and chronic-stage parasites. However, in P. chabaudi iRBC obtained from chronic infection, S7 PIRs were observed to localise in close proximity to the iRBC membrane suggesting they might be localised on the surface. Fixation can adversely destroy the integrity of the cell membrane and thus it is not possible to conclude that PIR proteins are expressed on the outer surface of the iRBC. Therefore, live imaging and flow cytometry analysis of P. chabaudi iRBC following labelling with the same peptide antisera were performed to investigate whether S7 and L1 PIRs are externally localised on the iRBC membrane ( Figure 5 ).

Live-cell immunofluorescent imaging of approximately 500-600 iRBC, across two individual biological replicates, obtained from acute and chronic P. chabaudi infections, showed complete absence of S7 and L1 PIRs on the iRBC surface, following staining with the anti-S7 or anti-L1 peptide antisera ( Figure 5A ; top panels), confirming that the S7s and L1s recognised by these antisera are not associated with the outer iRBC plasma membrane at the trophozoite-stage. They seem rather to remain within the iRBC, as shown in fixed sample preparations (81.18 ± 0.10% SEM S7+ iRBC and 98.03 ± 0.21% SEM L1+ iRBC). Labelling with the 9G8 and anti-MSP1 monoclonal antibodies, known to specifically target proteins located within the iRBC (Holder et al., 1992; Hartz et al., 1993) ( Figure 4A ; bottom panels), revealed positive staining only of acetone:methanol fixed/permeabilised iRBC.

To increase detection sensitivity of PIRs, we also conducted flow cytometric analysis of P. chabaudi iRBC at the acute- and chronic-phases of infection. Live (unfixed) iRBC were stained with anti-S7 and anti-L1 peptide antisera and compared with the same antibody staining on acetone:methanol fixed/permeabilised iRBC ( Figures 5C, D ; Supplementary Figure S5 ). The gating strategy used for the flow cytometry analysis is described in Supplementary Figure S1 . The contour plots clearly demonstrate that live iRBC (Hoechst+) are not stained positively for the anti-S7 or anti-L1 antisera (mean 99.3 ± 0.3% SEM S7 and 99.5 ± 0.2% SEM L1 acute-phase parasitised RBC; mean 99.4 ± 0.4% SEM S7 and 99.4 ± 0.4% SEM L1 chronic-phase parasitised iRBC), whereas almost the entire population of the fixed/permeabilised iRBC (Hoechst+) was stained positive for either the anti-S7 or anti-L1 peptide antisera (mean 99.7 ± 0.1% SEM S7 and 99.7 ± 0.2% SEM L1 acute-phase parasitised RBC; mean 99.0 ± 0.5% SEM S7 and 98.2 ± 1.5% SEM L1 chronic-phase parasitised RBC). As with IFAs, positive staining with the 9G8 and anti-MSP1 monoclonal antibodies was observed only on fixed/permeabilised iRBC ( Figure 5B ; bottom panels).

Therefore, no significant staining of live iRBC on IFA or by flow cytometry could be seen with either the anti-S7 or anti-L1 peptide antisera. By contrast, fixed/permeabilised iRBC demonstrated clear labelling of the internal S7 and L1 PIR proteins with anti-S7 and anti-L1 peptide antisera. These data confirm that the S7 and L1 PIRs as recognised by these antisera, are not exposed on the surface of live iRBC but rather localised within them.

To further investigate PIRs subcellular localisation, membrane fractionations were carried out using samples of iRBC at the trophozoite-stage to separate proteins translocated to the periphery and the host-cell cytoplasm from those localised in the PV/PVM or those remaining in the internal of the parasite. Figure 6A illustrates schematically the strategy and workflow of the fractionation procedures employed to separate proteins associated with each subcellular compartment. Briefly, to isolate proteins localised on the iRBC membrane and host-cell cytoplasm, P. chabaudi MT acute- and chronic-stage erythrocytic parasites at the trophozoite-stage were magnetically enriched and treated with Streptolysin-O (SLO fraction). The main function of SLO is to cause haemolysis, and therefore to permeabilise the iRBC plasma membrane without affecting the PVM (Jackson et al., 2007; Ruecker et al., 2012). SLO-treated cells were then treated with saponin (SAP fraction), which disrupts the PVM without affecting the plasma membrane of the parasite. This allows extraction of proteins that are localised either on the PVM or within the PV. Subsequently, the membrane soluble fraction was extracted in a buffer containing Tris.HCl (Tris fraction), which causes osmotic lysis. Following this step, all proteins localised in the parasite cytoplasm are released. Lastly, hydrophobic proteins associated with the parasite plasma membrane and proteins that interact with other parasite plasma membrane proteins were extracted in a RIPA lysis buffer supplemented with additional SDS and Triton (RIPA fraction) ( Figure 6A ).

Samples collected as previously described were subjected to western blot analysis using the rabbit anti-S7 and anti-L1 peptide antisera ( Figure 6B ). The results showed that the antisera-specific S7 PIRs are present in the SLO-fraction, represented by bands that migrate at about 40 kDa and 100 kDa, meaning they are localised either on the host-cell membrane or cytoplasm. No differences in the localisation were observed between SBP acute and MT acute- or chronic-phase parasites. The weak signal observed in the SAP fraction of MT parasites might be related to carry over proteins from the SLO fraction, since a Ter119 faint band can be also observed in the SAP fraction. The size of S7 PIRs is predicted to be between 33-37 kDa, therefore the specific band observed at about 100 kDa suggests polymerisation. L1 PIRs have an expected size range between 75-151 kDa. Following western blot analysis of membrane fractionated samples of MT acute- or chronic- and SBP acute-phase parasites ( Supplementary Figure S6 ), L1 antisera-specific PIRs were found to migrate at around 100 kDa and are predominantly present in the SAP fraction, where proteins associated with the PV and PVM are expected to be localised. In the MT chronic-phase parasites, a clear band was also observed in the SLO fraction. This agrees with the fluorescence microscopy data above ( Figures 2B , 3A ), further suggesting secretion of L1 PIRs in the iRBC cytoplasm at this stage. The purity of the different samples was assessed by western blot probed with antibodies against the RBC membrane protein Ter119 and the PVM protein EXP2 (a gift of James Burns; Meibalan et al., 2015). The anti-EXP2 antibody revealed a ~30 kDa polypeptide, exclusively in the SAP fraction. Ter119 can be predominantly found in the SLO fraction, however some cross-contamination can be also observed in the SAP fraction. This is possibly associated with the high abundance of this protein on the RBC membrane, which is maintained in the sample after multiple washes, rather than RBC membrane debris contamination.

The anti-S7 and anti-L1 peptide antisera do not target all members of the PIR clades ( Supplementary Table S2 ), therefore we performed mass spectrometry-based proteome analysis of the same fractionated samples (described above) as a more holistic approach to investigate the localisation of PIRs. Protein samples from three individual experiments were in-gel digested and analysed by LC-MS/MS using data independent acquisition. Acquired MS/MS spectra were then searched against databases of tryptic peptides predicted from all P. chabaudi as well as all host proteins. The results indicated that purity of the samples is likely to be of high quality since proteins whose location is known e.g. MSP1 (Holder et al., 1992), EXP2 (Meibalan et al., 2015), PCHAS_0624800 (Hartz et al., 1993), were identified in the corresponding fraction. Nevertheless, localisation of some of the recovered proteins indicated low-level contamination between serially prepared fractions. We compared the proteome of the acute-stage parasites with the proteome of the chronic-stage parasites to identify (1) PIR proteins that are highly expressed during the acute- and/or chronic-phases of blood-stage infection, and (2) their localisation on the iRBC (iRBC membrane/Host-Cell Cytoplasm (HCC), PV/PVM, Parasite Cytoplasm (PCyt) or Parasite Plasma Membrane (PPM)). PIR proteins identified in the fractionated samples are shown in Table 1 , 2 . The full list of P. chabaudi and Mus musculus proteins identified is given in Supplementary Table S3 .

Proteomic analysis of four fractionated samples prepared from acute- or chronic-stage parasitised erythrocytes recovered on average 2,311 P. chabaudi and 3,634 Mus musculus proteins, across three individual biological replicates ( Supplementary Table S3 ). Altogether, 23 individual PIRs were identified across all fractions of acute- and chronic-stage parasite preparations and all replicates conducted. Among them, 22 acute-stage associated PIRs and 16 chronic-stage associated PIRs were identified ( Table 1 , 2 , respectively). Among the PIRs recovered samples prepared from acute-stage parasites, 6 (27%) belong to the L1 clade, 8 (36%) to the L4 clade and 4 (18%) to the S7 clade. Additionally, 9 (41%) PIRs are located within the AAPL loci and 5 (23%) within the ChAPL loci. In total, 7 PIRs are located on chromosome 3 and 8 PIRs on chromosome 6. With regard to their subcellular localisation, 13 PIRs were picked up in at least 3 fractions suggesting that acute-associated PIRs are not limited to one subcellular compartment. Only 5 (23%) PIRs were exclusively found in one fractionated sample/subcellular compartment (1 in PV/PVM, 1 in PCyt and 3 in PPM). No association between clade or AAPL/ChAPL loci and subcellular compartment was found, however 14 PIRs were found to be localised in the iRBC membrane/HCC fraction, suggesting that they are exported proteins and further confirming our previous IFA findings ( Figure 2B ). Among the PIRs detected in samples prepared from parasites at the chronic phase of infection, only 1 (6%) of them belong to the S7 clade, 5 (31%) to the L4 clade and 7 (44%) to the L1 clade. Among the recovered PIRs, 3 (19%) are associated with the AAPL loci and 5 (31%) with the ChAPL loci. The majority of PIRs detected (9 PIRs; 56%) are located on chromosome 6, previously shown to be largely composed of ChAPL-associated clusters of pir genes (Brugat et al., 2017). Additionally, 7 PIRs (44%) were identified in all four subcellular compartments, 1 (6%) PIR in three subcellular compartments, and 8 PIRs (50%) were recovered exclusively within the parasite (either in PCyt or PPM). All identified PIRs, except PCHAS_0200031, were always being detected in the PPM fraction suggesting parasite membrane localisation within the iRBC. Comparing the PIRs identified at each phase of infection, a larger variation (e.g. AAPL/ChAPL loci association, chromosomal location, localisation pattern) and number of PIRs was recovered in acute-phase compared to chronic-phase parasite preparations ( Figure 6C ) possibly reflecting the complexity at this phase of infection.

Some of the PIRs detected in protein preparations of acute-phase parasites showed a wide distribution across all subcellular compartments, but as infection progresses to the chronic-phase, they appear to concentrate exclusively in one subcellular compartment e.g. PCHAS_0200011, PCHAS_0200021 and PCHAS_0200031 were recovered in all subcellular compartments of acute-phase parasites, but were not found to be secreted at the chronic-phase of infection.

The ancestral PIR, PCHAS_0101200 (Little et al., 2021), was consistently recovered in all samples prepared from both acute- and chronic-phase parasites and across all replicates conducted. Its high abundance in all subcellular compartments of the iRBC suggests that it is likely to be essential during parasite blood development.

In conclusion, there does not seem to be an exclusive subcellular compartment where PIR proteins are localised during both the acute- and chronic-phases of infection. This observation further supports the diverse nature of this family of proteins and consequently, the diversity in their function during Plasmodium blood development.