Atomic models of the Toxoplasma cell invasion machinery

Abstract

Apicomplexan parasites, responsible for toxoplasmosis, cryptosporidiosis and malaria, invade host cells through a unique gliding motility mechanism powered by actomyosin motors and a dynamic organelle called the conoid. Here, using cryo-electron microscopy, we determined structures of four essential complexes of the Toxoplasma gondii conoid: the preconoidal P2 ring, tubulin-based conoid fibers, and the subpellicular and intraconoidal microtubules. Our analysis identified 40 distinct conoid proteins, several of which are essential for parasite lytic growth, as revealed through genetic disruption studies. Comparative analysis of the tubulin-containing complexes sheds light on their functional specialization by microtubule-associated proteins, while the structure of the preconoidal ring pinpoints the site of actin polymerization and initial translocation, enhancing our mechanistic understanding of gliding motility and, therefore, parasite invasion.

Fig. 1. Cryo-EM structure of the CF.

a, Schematic diagram of T. gondii, adapted from a previous study. b, Representative cryo-EM image showing well-separated CFs following sonication and protease treatment. Overlapping boxes (yellow rectangles) were used to extract CF particles. c, Selected 2D class averages of CF particles, showing different views. d, Cross-sectional view of the cryo-EM structure of a CF segment, with component proteins displayed in distinct colors. Three clusters of proteins bound to the external tubulin surfaces are indicated by black dashed circles. e, Longitudinal view of the CF segment shown in d. f, Interprotofilament angles of the CF. g, Single-particle cryo-EM structure of CF (dark gray) fitted into a previously reported subtomogram average of CF (light orange) (EMD-66190). The red arrow indicates additional densities above the CPH1–DCX–CF6 network that are absent in our single-particle cryo-EM structure. Source data

Fig. 2. Atomic model of the CF.

a, Cross-sectional view of the atomic model of a CF segment, with component proteins shown in distinct colors. Viewing angles for b, d, e and g are indicated. b, Close-up view of TgDCX bound at the intradimer interface on the microtubule lattice. c, Close-up view of human DCX (PDB 6RF2) bound at the interdimer interface on the microtubule lattice. d, Longitudinal view of a CF segment with 24-nm periodicity, showing component proteins in distinct colors. Atomic models of α,β-tubulin are hidden for clarity. e, Close-up view of T. gondii SAS6L linear array in the region indicated by dashed rectangles in d. f, Crystal structure of zebrafish SAS6 homodimer (PDB 2Y3W). g, Luminal view of the CF structure showing that C-terminal short helices of CF2 bind across the seam. h, External surface view of the seam between protofilaments 4 and 5. i, Cutaway view of the CF structure showing CIP3 penetrating through the tubulin wall.

Fig. 3. Structure-guided genetic disruption of identified CF proteins within the bridging complex.

a, Atomic model of two adjacent CFs fitted into a previously reported subtomogram average of the CF (light orange) (EMD-66190). b, Positions of eight CF component genes in the phenotypic ranking of all T. gondii genes (x axis), as determined in a genome-wide KO screen. The phenotype scores (fitness scores) for each gene (y axis) were obtained from ToxoDB, where genes with low scores are predicted to be essential. Inset: the two synthetic lethal pairs (black lines) and one synthetic defective pair (gray line) identified in this study. c, Costaining of the four bridging complex proteins with selected markers using U-ExM. Markers include PCR2 (green), APR1 (blue) and tubulin (gray). Freshly egressed extracellular parasites were labeled with chicken anti-Myc and anti-chicken IgY Alexa Fluor 405 (blue), mouse anti-Ty and anti-mouse IgG Alexa Fluor 488 (green), rat anti-HA and anti-rat IgG Alexa Fluor 555 (red), and rabbit anti-Tubulin and anti-rabbit IgG Alexa Fluor 647 (gray). d, Plaque assay of four bridging complex protein mutants, generated through clean KO or cKD, on HFF monolayers treated with IAA or vehicle control (−IAA) for 8 days with 200 parasites per monolayer. Scale bar, 5 mm. e, Plaque assay of 2 synthetic lethal pairs and 1 synthetic defective pair on HFF monolayers treated with IAA or vehicle control (−IAA) for 8 days with 200 parasites per monolayer. Scale bar, 5 mm. f, Quantification of plaque area and number in parental lines and synthetic mutants (n = 6), from 3 independent experiments, each with 2 technical replicates. Data are shown as the mean ± s.d. Each parasite line was analyzed individually for statistical significance using an unpaired Student’s t-test (IAA versus vehicle). NS, not significant (P ≥ 0.05); *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001 and ****P ≤ 0.0001. Source data

Fig. 4. Cryo-EM structures of ICMTs and apical SPMTs.

a, Representative cryo-EM image showing a pair of ICMTs within the conoid. b, Example of an ICMT pair dislodged from the conoid, revealing fibrous densities on their sides. c, Selected 2D class averages of ICMT particles showing different views. d, Cross-sectional view of the cryo-EM structures of ICMT-1 and ICMT-2, with associated proteins. The light-purple density in the background is our single-particle 3D reconstruction of the ICMT pair, obtained before refining each ICMT individually. e, Tilted view of ICMT-2 showing that TLAP3 spans laterally across 11 of the 13 protofilaments, while TLAP4 binds specifically at the microtubule seam. TLAP3 and TLAP4 in ICMT-1 and apical SPMT exhibit similar appearances. f,g, Costaining of TLAP3 (f) and TLAP4 (g) with tubulin markers using mouse anti-Ty and anti-mouse IgG Alexa Fluor 488 (green) and rabbit anti-Tubulin and anti-rabbit IgG Alexa Fluor 568 (red) for U-ExM. h,i, Longitudinal views of ICMT-1 showing ICMAP1 (h) and ICMAP2 (i), respectively. j, Longitudinal view of ICMT-2 showing MAP densities (green), with some confidently assigned to ICMAP4 (medium orchid).

Fig. 5. Cryo-EM structures of apical SPMTs.

a, Representative cryo-EM image showing an isolated conoid after sonication but before protease treatment. b, Representative cryo-EM image of a conoid sample after sonication and protease treatment, showing SPMTs connected to the APR. c, Cross-sectional view of the cryo-EM structure of apical SPMT, obtained by manually selecting only the SPMTs near the APR, as shown in a and b. d, Luminal view of the seam in the apical SPMT. e, Close-up view of the TLAP2 homodimer bound to the apical SPMT. f, Cross-sectional view of the cryo-EM structure of the central SPMT (EMD-23869). g, Luminal view of the seam in the central SPMT, showing the absence of TLAP3 and TLAP4 molecules. h, TLAP3 inserts a segment into the taxol-binding pocket. In the central SPMT, this pocket is occupied by a small molecule-like density (cyan).

Fig. 6. Cryo-EM structure of PCR-P2.

a, Representative cryo-EM image of a conoid sample after sonication and protease treatment, showing a PCR dislodged from the conoid. Overlapping boxes (yellow rectangles) were used to extract the ring segments. b, Selected 2D class averages of ring particles showing different views. c, A hypothetical structure of a complete ring, generated by extending the single-particle structure (red dashed rectangle) using the measured angle between two consecutive repeating units. A slight misalignment was observed at the rejoining point (red arrow). Viewing angles for d and e are indicated. d,e, Top view (d) and bottom view (e) of the atomic model of the PCR-P2 containing ~3 repeating units, with component proteins shown in distinct colors. f, Atomic model of the two copies of PCR4/5 heterodimers located at the core of the PCR-P2. g, Atomic model of the two SEC23/24 heterodimers located at the inner rim of the PCR-P2. h, Atomic model of MyoL interacting with three potential myosin light chain subunits (CAM1, CAM4 and MLC4).

Fig. 7. PCR-P2 contains several essential genes.

a, The single-particle cryo-EM structure of PCR-P2 (dark green) fitted into a subtomogram average of the PCR P2–P3 rings (gray) (EMD-28126). b, Locations of seven identified P2 essential proteins in this study (FRM1, FLM1, FLM2, AKMT2, PCR10, CAM4 and SEC24). c, Rows 1–2: immunoblot of lysates from parasites containing C-terminal mAID–HA-tagged PCR proteins treated with vehicle (−IAA) or 500 μM auxin (+IAA) for 14 h. Western blots were performed for ALD1 (red) and HA (green). Row 3: localization of mAID–HA-tagged PCR proteins by U-ExM. Parasites were detected with rabbit anti-HA (green) and mouse anti-acetylated tubulin (magenta). Scale bars, 2 μm. Rows 4–5: plaque assay of mAID–HA-tagged lines on HFF monolayers treated with IAA or vehicle control (−IAA) for 8 days. Scale bars, 5 mm. d, Quantification of plaque size (n = 3 independent experiments, each with n = 10 technical replicates; total n = 30). Data are presented as the mean ± s.d. Statistical analysis was conducted using a multiple-comparison unpaired Student’s t-test with false discovery rate correction. ****P < 0.000001. None indicates no formed plaques. e, Immunofluorescence microscopy of intracellular dividing SEC23–mAID–HA and SEC24–mAID–HA parasites expressing GRASP2-Ty (Golgi marker) growing on HFF monolayers for 16 h. Parasites were detected with mouse anti-Ty (green) and rabbit anti-HA (magenta). DAPI (blue) was used to stain the nuclei of both HFFs and parasites. Scale bars, 2 µm. In these images, SEC23 and SEC24 signals at the conoid are not evident because of their notably weaker intensity compared to those at the ER–Golgi interface. Source data

Extended Data Fig. 1. Data processing workflow for conoid fibers (CFs).

Operations in CryoSPARC, Relion and FREALIGN are highlighted in light green, orange and turquoise, respectively.

Extended Data Fig. 2. Protein identification workflow.

Representative examples of both sequence-based and structure-based approaches are shown.

Extended Data Fig. 3. Domain analysis of CF and ICMT/SPMT component proteins.

Protein domain assignments are based on AlphaFold predictions^,, ToxoDB and FoldSeek. Regions resolved in the cryo-EM maps are indicated with green boxes.

Extended Data Fig. 4. Synthetic lethal analysis of bridging complex proteins.

(a) Time course of auxin-mediated degradation dynamics of the four bridging complex proteins. (b) Representative immunoblot from CF1-mAID and SAS6L-mAID lines. Blots were probed with Rabbit anti-ALD1 and anti-rabbit IgG IRDye 680RD (red), Mouse anti-HA and anti-mouse IgG IRDye 800RD (green). (c) Quantification of plaque area and number in parental lines and four bridging complex protein mutants. N = 6, from 3 independent experiments, each with two technical replicates, means ± SD. Each parasite line was analyzed individually for statistical significance using an unpaired Student’s t test (IAA vs. vehicle or knockout vs. TIR1), P values: ns ≥ 0.05, ** ≤ 0.01, *** ≤ 0.001, **** ≤ 0.0001. (d) Immunoblot of lysates from SAS6L-mAID parasites expressing endogenously Myc-tagged CF1, CF5 and CIP3. Blots were probed with Rabbit anti-ALD1 and anti-rabbit IgG IRDye 680RD (blue), Rabbit anti-HA and anti-rabbit IgG IRDye 680RD (red) and Mouse anti-Myc and anti-mouse IgG IRDye 800CW (green). (e) Immunofluorescence (IF) microscopy of extracellular SAS6L-mAID parasites expressing endogenously Myc-tagged CF1, CF5 and CIP3. Intracellular parasites were treated with IAA or vehicle for 24 hr and freshly egressed extracellular parasites labeled with mouse anti-Myc and anti-mouse IgG Alexa Fluor 488 (green), rabbit anti-HA and anti-rabbit IgG Alexa Fluor 555 (red) and DAPI (blue). Scale bar, 1 µm. (f) Schematic diagram of pairwise synthetic lethal analysis used in this study. (g) Plaque assay of two other synthetic mutants with modest defect on HFF monolayers treated with IAA or vehicle control (-IAA) for 8 days with 200 parasites per monolayer. Scale bar, 5 mm. Source data

Extended Data Fig. 5. Synthetic lethal/defective pairs of CF proteins disrupt the conoid structure in both intracellular and extracellular parasites.

(a) Transmission electron microscopy (TEM) images of intracellular parasites grown with or without IAA for 24 h. Conoid fibers are shown in tangential (section-1) and cross-sectional views (section-2). Blue brackets highlight the continuity of the conoid fibers. Scale bar = 100 nm. (b) TEM images of extracellular parasites prepared by negative staining. Parasites were grown with or without IAA for 24 h. Scale bar = 100 nm.

Extended Data Fig. 6. Data processing workflow for ICMTs.

Operations in CryoSPARC, Relion and FREALIGN are highlighted in light green, orange and turquoise, respectively.

Extended Data Fig. 7. Cryo-EM structure of the preconoidal ring P2 (PCR-P2).

(a, b) Orthogonal views showing our single-particle cryo-EM structure of PCR-P2 (dark green) fitted into a subtomogram average of the PCR P2-P3 rings (EMD-28126). (c) Fit of the same structure into a subtomogram averaged density of the PCR P1-P2-P3 rings (EMD-28246). (d, e) Top (d) and bottom (e) view of the cryo-EM density of PCR-P2 containing ~3 repeating units.

Extended Data Fig. 8. Domain analysis of PCR-P2 component proteins.

Protein domain assignments are based on AlphaFold predictions^,, ToxoDB and FoldSeek. Regions resolved in the cryo-EM maps are indicated by green boxes.

Extended Data Fig. 9. Functional analysis of other identified PCR proteins.

(a) Invasion assay of MyoL-mAID-HA parasites. Parental and MyoL-mAID-HA parasites were pre-treated ±500 μM IAA for 24 h, then allowed to invade HFFs for 16 h before SAG1 staining. % of invasion was determined by calculating the average number of parasitophorous vacuoles per host cell and normalizing to the untreated ( − IAA) group. Means ± SD (N = 3 independent experiments, each with n = 10 technical replicates). Multiple t-test with FDR approach, n.s., not significant, ****, P < 0.000001. (b) Expression, localization and growth phenotype of PCR10-mAID-HA parasites. Top left: Immunoblots of parasites treated ±IAA (14 h) were probed with anti-ALD1 (red) and anti-HA (green). Bottom left: U-ExM image showing PCR10 localization (HA, green; acetyl-Tubulin, magenta). Scale bar, 2 μm. Middle: Plaque assays were performed ±IAA for 8 days on HFF monolayers (200 parasites per well). Scale bar, 5 mm. Right: Plaque sizes were shown as means ± SD from three independent experiments with 10 replicates for each (N = 3, n = 30). Two-tailed unpaired Student’s t test, n.s., not significant. (c) Plaque assay showing no growth defect of SEC23-KO parasites. Scale bar, 5 mm. Data were pooled from three independent experiments with 10 replicates for each (N = 3, n = 30), means ± SD. Two-tailed unpaired Student’s t test, n.s., not significant. (d) Expression, localization and growth phenotype of SEC23-mAID-HA parasites. Experimental design and data analysis were performed as described in (b). Data were pooled from three independent experiments with 10 replicates for each (N = 3, n = 30), and represented as means ± SD. Two-tailed unpaired Student’s t test, n.s., not significant. Source data