An N-terminal export domain in PF08_0005 (PF3D7_0830300).

Representative live cell images of P. falciparum 3D7 iRBCs expressing episomal GFP fusion constructs under the crt promoter. Construct names and schematics are depicted above respective micrographs. Nuclei were stained with DAPI. Scale bar 5 µm. (A) The PF08_0005 sequence downstream of the signal peptide (SP) was divided into three parts. Shown are parasites expressing the full-length protein, and the protein lacking part 2 and 3, respectively. (B) PF08_0005 part 1 was further subdivided into three parts, a, b and c. Parasites expressing deletion constructs missing part a, b and c, respectively, are shown. (C) Parasites expressing minimal constructs containing only part a and b, or only part a.

Export domains in MSRP6 and PFB0115w.

Representative live cell images of P. falciparum 3D7 iRBCs expressing episomal GFP fusion constructs under the crt promoter. Construct names and schematics are depicted above the respective micrographs. Nuclei were stained with DAPI. Scale bar 5 µm. (A) MSRP6 downstream of the signal peptide (SP) was divided into three parts (1-3). Shown are parasites expressing the indicated constructs. The SP replacement was done with the SP from ETRAMP10.1. Arrowheads indicate a large focus of GFP fluorescence likely representing an overexpression induced aggregate. (B) Parasites expressing minimal constructs of MSRP6 containing part 1 and a fraction of part 2 or only part 1. (C) Parasites expressing minimal construct of PFB0115w containing the SP and downstream 50 aa.

The MSP-7-related region in MSRP6 is sufficient for export and Maurer’s clefts attachment.

Representative live cell images of P. falciparum 3d7 iRBCs expressing episomal GFP fusion constructs under the crt promoter. Construct names and schematics are depicted above respective micrographs. Yellow bar, PEXEL. Nuclei were stained with DAPI. Scale bar 5 µm. (A) Schematic showing subdivision of MSRP6 part 3 into 4 parts, a, b, c, and d. (B) Parasites expressing MSRP6 constructs with the indicated deletions. (C) Parasites expressing REX3 containing MSRP6 part c, d or c+d at its C-terminus or only the MSRP6 SP fused to parts c and d.

Bio-ID of Maurer’s cleft proteome and MSRP6 MAD interactome.

(A) Live cell images of P. falciparum 3D7 infected RBCs expressing the episomal BirA*-GFP fusion constructs schematically shown above the panels. Yellow bar, PEXEL (top, N-terminal 70 aa of REX3 (REX3trunc) fused to BirA* and GFP (host cell soluble control), middle, REX3trunc fused to MSRP6 parts c and d, BirA* and GFP (MAD-specific), bottom, N-terminal 260 aa of STEVOR fused to GFP and BirA* (general Maurer’s clefts control)). Nuclei were stained with DAPI. Scale bar 5 µm. (B) Western blots with parasite lysates of the 3 cell lines in (A) and 3D7 control probed with α-GFP (left) and Streptavidin (right). Two different exposures of the strepavidin probed blot are shown for STEVOR1-260-GFP-BirA* (low and high) for comparability to the REX3 samples. Asterisks, bands representing the respective fusion proteins. (C-E) Scatterplots of quantitative Bio-ID experiments (see Fig. S1 for all plots of all experiments and Table S1 for full data) identifying proteins significantly enriched or depleted in the MAD interactome compared to the host cell cytosolic proteome (C) or Maurer’s cleft proteome (D) or in the Maurer’s cleft proteome compared to the host cell cytosolic proteome (E). Normalized ratios were calculated for proteins identified by at least two peptides and normalized log2-ratios of replicate experiments were plotted. Intensity-based outlier statistics (two-sided Benjamini-Hochberg test) was applied to calculate FDR values and proteins enriched or depleted with an FDR below 5% in both replicates were labelled with a color-code reflecting the level of significance in the least significant experiment (<5% dark green, <1% light green, <0.5% blue, <0.1% purple, <5e-4 yellow, <1e- 4 orange, <5e-5 pink, <1e-5 red). Significant hits are numbered and gene-IDs or short unique names are given. Proteins encoded by multigene families (STEVORs, RIFINs, PfEMP1) that might differ in expression between the distinct cell lines and hence may show as false-positives, are marked with rhomboids with black frame. (F) A heatmap representation of all replicates each of REX3-MAD-over-STEVOR (MAD over unrelated Maurer’s cleft proteome), REX3-MAD-over-REX3 (MAD over host cell cytosolic proteome) and STEVOR-over-REX3 (Maurer’s cleft proteome over host cell cytosolic proteome) quantitative Bio-ID experiments (C-E and Fig. S3, see Table S1 for full data) for proteins enriched with an FDR <5% in at least 2 out of 4 replicate reactions. Proteins are ranked from high to low on the average normalized log2 ratio of the REX3-MAD-over-STEVOR comparison and color intensity portrays the normalized log2 ratio per experiment (red enriched, yellow neutral, blue depleted based on the 1-99 percentile of values for all identified proteins). Grey blocks are proteins not identified in an experiment or for which no ratio could be calculated due to a missing label. PlasmoDB gene identifiers, short names and protein product descriptions are listed. Yellow labelled proteins were selected for follow-up as putative MAD interactors, orange labelled proteins showed moderate to no enrichment as MAD interactors but were enrichment in the Maurer’s cleft proteome. Proteins encoded by clonally variant multigene families (likely false-positives) are labelled grey. The heatmap was generated using the web-based Morpheus tool from the Broad Institute (Harvard, 2017).

Validation of Bio-ID hits.

(A,B) Representative live cell images of P. falciparum infected RBCs expressing endogenously GFP-tagged (using the standard 2xFKBP-GFP tag (Birnbaum et al., 2017)) proteins enriched in REX3trunc-MAD-over STEVOR1-260-STEVOR-GFP-BirA*in BioID experiments (Fig. 4), representing potential MAD specific interactiors (A) or enriched in MAD or STEVOR over REX3trunc-GFP-BirA* (Fig. 4), representing Maurer’s cleft specific proteins (B). (C) Ring stage parasites showing stage specific localisations of PeMP1, PF08_0003 and PeMP3. (D) Co-IP experiments showing interaction of MSRP6 with the pulled down endogenously GFP-tagged bait protein (indicated above the blot) and probed with the indicated antibodies. IP, total input lysate; U, unbound protein of the lysate after IP; W, fifth wash; E, eluate. Asterisk shows Pf332 degradation product, this protein is too large to detect full length on a standard PAGE. All replicas and full blots shown in Fig. S6. (E) Quantification of IP experiments in (D) and Fig. S6, see Table S2 for exact values. Error bars show SD, green line the mean. Average of enrichment of MSRP6 over SBP1 is indicated above each graph. (F) Representative live cell images of P. falciparum young trophozoite iRBCs expressing endogenously GFP-tagged PeMP3 with (bottom) and without episomally expressed MSRP6-mCherry. If no simple short name was known (Fig. 4F) the protein designations in the imaging panels were 10_0024-GFP for PF3D7_1002000/PF10_0024, 8_0003-GFP for PF3D7_0830500/PF08_0003, C0070c for PF3D7_0301400/PFC0070c and L0055c for PF3D7_1201100/PFL0055c. A, B, and F, Nuclei were stained with DAPI; merge, overlay of red and green channel; Scale bar 5 µm.

Maurer’s clefts numbers and anchoring in MSRP6 complex member mutants.

(A) Quantification of the number of Maurer’s clefts (MCs) per iRBC in the indicated cell lines (Δ indicates disrupted gene of the indicated protein). Each dot represents the Maurer’s clefts number in a single cell. n = 20 cells for each cell line except ΔPeMP2 (23 cells). (B) Live cell images of ΔPIESP2-GFP (top panel), ΔPf332-GFP (panel 2 and 3) or ΔPeMP2-GFP (panel 4 and 5) parasites, respectively, co-expressing episomal SBP1-mScarlet as a Maurer’s clefts marker. Boxes, areas enlarged on the right. Nuclei stained with DAPI. (C) Quantification of the number of Maurer’s clefts per infected RBC in the indicated cell lines as in (A). n = 41 cells for the PeMP2-GFP parasites, 67 cells for the ΔPIESP2 parasites and 20 cells for the ΔPIESP2-complemented parasites, respectively. (D) Short-term time-lapse images of PeMP2-GFP (top panel), ΔPeMP2-GFP (panel 2), ΔPIESP2-GFP (panel 3) or ΔPf332-GFP (panel 4 and 5) parasites, respectively, co-expressing episomal SBP1-mScarlet. Images of the same cell were taken at the time points indicated. Images from the second time point were converted to blue and overlaid with the first time point to visualise MC movement (purple, overlay of the SBP1-signals indicating arrest of MCs, signal with original colour indicative of MCs movement). (B, D), scale bars, 5 μm and 2 µm for enlarged areas. (A, C), mean and error bars (SD) are shown; two-tailed unpaired t-test, P-values indicated. DIC, differential interference contrast; ns, not significant.

Detailed phenotpyes of PIESP2- and Pf332-TGDs.

(A) Quantification of the number of Maurer’s clefts (MCs) per iRBC in the indicated integration cell lines in relation to the parasite age determined by parasite diameter (n = 41 cells for PeMP2-GFP parasites and 67 cells for ΔPIESP2 parasites, data from Fig. 6C). (B) Transmission electron microscopy images of 27 - 36 hours post invasion (hpi) endogenously tagged PeMP2 (PeMP2-GFP) and PIESP2 (PIESP2-GFP) parasites, and ΔPIESP2 or ΔPf332 parasites. Black squares with numbers show enlarged areas. For PIESP2-GFP and ΔPIESP2, MCs are highlighted in red. Scale bars, 0.5 μm. (C, D) Quantification of MC length (C, n = 67 (PeMP2-GFP), 148 (PIESP2-GFP) and 117 (ΔPIESP2) sections), and number of detectable MCs per cell in a section (D, n = 27 (PeMP2-GFP), n = 24 (PIESP2-GFP) and n = 123 (ΔPIESP2) individual sections). Red lines, mean; error bars, SD; two-tailed unpaired t-test, P-values indicated. (E-G) Selected images (full set in Fig. S6) of long-term 3D time-lapse experiments in PeMP2-GFP (E), ΔPIESP2 (F) and ΔPf332 (G) parasites co-expressing episomal SBP1-mScarlet. DIC shows single slice per time point, red signal maximum intensity projection of reconstructed z-stack. Time is indicated as h after start of imaging; re-invasion after completion of the cycle is indicated by white arrows. Onset of MC arrest in PeMP2-GFP indicated by blue line. Scale bars, 5 µm. One representative of 32 (PeMP2-GFPendo), 18 (ΔPIESP2), 12 (ΔPf332) cells. (H,I) Quantification of the number of MCs per iRBC of indicated cell lines in relation to the parasite age determined by parasite diameter (n = 17 and 26 cells for ΔPf332+SBP1-mScarlet and ΔPf332+MSRP6-mCherry parasites, respectively (H) and n = cells including those in Fig. 6A; PeMP2-GFP same data as in (A) shown as a reference in both graphs). (A, H and I), linear regression lines indicated.

Anchoring defects in PIESP2- and Pf332-TGDs have no profound effect on tether-Maurer’s clefts connection.

(A) Schematic of the disruption of PIESP2 by homologous recombination using SLI2a in parasites with SLI-based endogenously GFP-tagged Pf332; mSca (mScarlet); yDHODH, yeast dihydroorotate dehydrogenase; asterisk stop codon; black arrows, native promotors; black boxes, N-terminal signal peptide and transmembrane domains. (B, C) Representative live cell images of Pf332-GFPendo parasites with a disrupted, mScarlet-tagged PIESP2 (ΔPIESP2-mSca) (B) or endogenously mScarlet-tagged MAHRP2 (MAHRP2Scaendo) corresponding to tethers. Top panel in (C), short-term time-lapse as done in Fig. 6D. (D) Live cell images of 3D7 parasites episomally expressing SBP1-GFP (MC) and MAHRP2-mScarlet (tethers) using a single expression cassette (SBP1-GFP-T2A-MAHRP2-mScarlet). (E-H) Representative live cell images of parasites with an endogenously mScarlet-tagged MAHRP2 (MAHRP2Scaendo) with a disrupted Pf332 (ΔPf332) (E) or PIESP2 (ΔPIESP2) (G) or disrupted Pf332 with episomal SBP1-GFP-T2A-MAHRP2-mScarlet (F). Short term time overlays in (E) and (F) done as in Fig. 6D. (I), Time overlay of merge of red and green signal from the top cell shown in (H) for which the signal in the second time point was converted from green to magenta (SBP1, Maurer’s clefts) and from red to turquoise (MAHRP2, tethers). Individual clefts and tether foci are numbered to show relative position between the two time points. (C-F) Nuclei were stained with DAPI, white arrowheads show fainter mobile tether signals. (H, I), arrowheads show connection of tethers with MC. Schematics of the modified loci are shown for (E) and (G). White boxes show enlarged regions. Scale bars, 5 μm. DIC, differential interference contrast.

Analysis of the MSRP6-TGD.

(A) Schematic MSRP6 disruption using SLI2a; asterisk, stop codon; yDHODH, yeast dihydroorotate dehydrogenase; mSca, mScarlet; black arrows, native promotor; black box, N-terminal signal peptide. (B) Short-term time-lapse live cell images of ΔMSRP6-mScarlet parasites, expressing episomal SBP1-GFP; time overlay as done in Fig. 6D. (C-E) Quantification of the number of Maurer’s clefts per iRBC in the indicated cell lines. (D) Shows the relation to parasite age determined by parasite diameter. Linear regression lines are shown. (C, D), n = 60 and 58 cells for 3D7 and ΔMSRP6 parasites, respectively. (E) Shows only late-stage parasites (diameter ≥6 µm) including the parasites from (C) fulfilling this criterion (n = 74 and 82 cells for 3D7 and ΔMSRP6 parasites, respectively). (F) Live cell images of PeMP3GFPendo parasites with a disrupted MSRP6 (ΔMSRP6-Sca). The graph shows a quantification of the Maurer’s clefts-associated PeMP3-GFP signal relative to the distributed signal (soluble population of the protein) in the indicated cell lines (n = 16 and 17 cells for PeMP3-GFP in 3D7 and PeMP3-GFP in ΔMSRP6 parasites, respectively). (G, H) Live cell images of ΔPIESP2-GFP (G) or ΔPf332-GFP (H) parasites with an endogenously mScarlet-tagged MSRP6 (MSRP6-Scaendo). (I) One possible model of MSRP6 complex (’late stage complex’) at the Maurer’s clefts. (B, F-H), nuclei were stained with DAPI; scale bars, 5 μm; DIC, differential interference contrast. (F-H) Show schematics of the modified loci. (C, E, and F), mean and error bars (SD) are shown; two-tailed unpaired t-test, P-values indicated.

Cytoadherence and PfEMP1 transport in MSRP6 complex member mutants.

(A-D) IT4 parasites with a SLI-activated Var01 (IT4var01-HAact parasites) with disruptions of the genes encoding the indicated proteins (schematics above the panels). Images show live cells demonstrating loss of localisation of the disrupted protein (top), or IFAs with acetone fixed cells showing the PfEMP1 (a-HA) together with SBP1 (Maurer’s clefts) or KAHRP. For the SBP1 disruption REX1 was used instead of SBP1 and EXP2 to show the PVM. Nuclei were stained with DAPI; scale bars, 5 μm; DIC, differential interference contrast. (E) Binding of iRBCs of the indicated cell lines to CHO cells expressing CD36, ICAM1 or GFP shown by super plots (Lord et al., 2020) (three independent experiments with technical replicas. Error bars (SD) indicates mean of averages of biological replicates with SD; P-values (unpaired t-test) are shown; small grey dots: bound iE infected erythrocytes) per field of view extrapolated to mm2. Coloured shapes: average of bound iE/mm2/replicate). F) Immunoblots with extracts of saponin-lysed parasites of the indicated the cell lines treated with or without trypsin (n = 2, replica shown in Fig. S11B). α-HA detects the activated Var01 and the protected fragment is indicative of surface exposure. α-SBP1N served as control that the RBC was not breached which would also result in a protected fragment. Asterisks indicate protected fragments. (G) Fluorescence microscopy images of Bodipy stained parasites of the indicated cell lines, showing the same cell after a short interval. Second time point image converted to turquoise and overlaid (time overlay) with the first time point to illustrate Maurer’s clefts movement. A boosted and enlarged image of the overlay is shown at the bottom. Arrow shows a ΔPf332-typical aggregate. Scale bar, 5 μm. DIC, differential interference contrast.