Fig 1.
Visualizing knob structures at the external and internal surfaces.
(A) SEM of an intact CS2-infected RBC (~36 h post-invasion). Zoom shows a representative 1 x 1 μm section. (B) Schematic representation of KAHRP. Protein domains and features are as follows: black–signal sequence; orange–PEXEL; yellow–histidine-rich (His) repeats (amino acids 60–123), blue—5’ repeats (amino acids 329–420), green—3’ repeats (amino acids 506–555), ankyrin-binding domain (amino acids 282–361), spectrin-binding domain (amino acids 370–441) and the binding regions of anti-KAHRP (mAb89: amino acids 424–539 and mAb18.2: amino acids 282–362). Full length CS2 KAHRP has 634 amino acids. (C) dSTORM imaging of a sheared membrane preparation from CS2-infected RBCs (~30 h post-invasion) labelled with anti-KAHRP (mAb89) and anti-mouse Alexa-647 secondary antibody. Zoom shows a representative 1 x 1 μm section. Blue arrows: puncta. White arrows: ring structures. (D) The number of fluorescence events detected by dSTORM within 200 x 200 nm regions drawn around individual anti-KAHRP labelled puncta and ring structures. Counts are presented from 23 (color-coded) sheared membranes. (E) Sheared membranes prepared from CS2-infected RBC membranes at 30 h post-invasion imaged by SEM. Zoom shows a representative 1 x 1 μm section. White arrow: Dimpled disk-structure. (F) STORM-SEM showing overlay of physical dimpled disks and anti-KAHRP labelled ring structures (white arrow). Scale bars: (A,C, E, F) 500 nm.
Fig 2.
The density of KAHRP-labelled ring structures increases during parasite development.
(A) CS2-infected RBC membranes imaged by dSTORM and SEM during knob formation. Representative 1 x 1 μm sections are presented from cells at 16, 20, 24 and 30 h post-invasion (hpi). Scale bar: 500 nm. (B) The numbers of puncta (structures 5–50 pixels in size) and ring structures were quantified for multiple regions from different cells (n = 35, 25, 26 and 15 membranes for 16, 20, 24 and 30 h post-invasion respectively). Data is presented as the mean ± SEM (unpaired t-test, * p < 0.05, ** p < 0.005, *** p < 0.0001).
Fig 3.
Actin remodeling is required for knob formation.
Sheared membranes were prepared from CS2-infected RBCs treated with or without cytochalasin D (CytoD), from 16 h post-invasion. Parasite infected RBCs were harvested at 22 h post-invasion, or cytochalasin D was removed at 22 h and cells harvested 8 h after removal at 30 h post-invasion (CytoD 8 h recovery). (A) dSTORM images showing KAHRP distribution. Scale bar: 500 nm. (B,C) Quantification of ring structures and puncta from untreated control (n = 31) and cytochalasin D-treated (n = 28) membranes at 22 h post-invasion (unpaired t-test, *** p < 0.0001). (D) Infected RBCs were passaged though a bed of microbeads and the parasitemia in the flow-through was assessed. Data represent three separate experiments each in triplicate (unpaired t-test, *** p < 0.0001). (E,F) Quantification of ring structures and puncta from control (n = 15) and cytochalasin D-treated (n = 30) cells following 8 h recovery (unpaired t-test, *** p < 0.0001). (G) Microbead filtration of cells following 8 h recovery showing no significant difference. Data represent three separate experiments each in triplicate. (Unpaired t-test, NS p = 0.53). Data is presented as the mean ± SEM.
Fig 4.
Spectrin binding is required for correct knob formation.
(A,B) Sheared membranes prepared from RBCs infected with CS2, K530, K405 and K362 parasite lines at 40 h post-invasion were imaged by SEM and dSTORM (A; anti-KAHRP mAb18.2) or only SEM (B). Images are displayed as 1 x 1 μm zoomed panels. Scale bar: 500 nm. (C) Projections from electron tomograms of negatively-stained preparations of schizont-infected RBCs from the CS2, K530 and K405 lines. The spiral structures were segmented manually, shown as an overlays in the bottom panel. Scale bar: 50 nm. (D) The proportions of knobs containing complete, fragmented or no spiral were recorded (n = 52, 24 and 21 knobs for CS2, K530 and K405 infected RBCs, respectively). Data is presented as a percentage of the total number of recorded knobs.
Fig 5.
PfEMP1 associates with knobs laterally.
(A) Sheared membranes prepared from RBCs infected with PfEMP1A-GFP transfectants at 20 h post-invasion. Samples were imaged by SEM and dSTORM (anti-GFP). STORM-SEM shows fluorescent foci at knobs (blue arrows), adjacent to knobs (yellow arrows) or independent from knobs (white arrows). Scale bar: 500 nm. (B) Total number of PfEMP1A-GFP foci on sheared membranes at 20, 24 and 30 h post-invasion. (C) Knob counts from sheared membranes viewed by SEM. (D) Quantitation of the number of fluorescent anti-GFP foci and SEM knobs per μm2 from RBC membranes infected with A4 PfEMP1A-GFP parasites at 20, 24 and 30 h post-invasion (n = 11, 12 and 10 membranes, respectively). Data plotted as fluorescent foci at a knob (<50 nm, blue), adjacent to a knob (50–100 nm, yellow) or independent from knobs (>100 nm, white). See S10B Fig for explanatory diagram.
Fig 6.
A model for virulence complex formation.
A schematic diagram of the proposed model for assembly of the virulence complex. (A) The membrane skeleton of an uninfected RBC (uRBC). (B) KAHRP is delivered to the RBC membrane skeleton where it interacts with spectrin. Host cell remodeling is initiated (white arrows) via the dissociation of the actin junction points, allowing spectrin remodeling. PfEMP1 is delivered to the RBC membrane. (C) Host cell remodeling continues with spectrin octamers being formed (white arrow). KAHRP undergoes higher order assembly (with potential involvement of a bridging protein) forming a horseshoe-shaped structure that scaffolds the spiral knob core. PfEMP1 moves laterally to associate with the virulence complex. (D) KAHRP modules form the base of the knob complex. Most of the PfEMP1 is assembled adjacent to or at knob complexes.