Structural Basis of Phosphatidic Acid Sensing by APH in Apicomplexan Parasites

Abstract

Plasmodium falciparum and Toxoplasma gondii are obligate intracellular parasites that belong to the phylum of Apicomplexa and cause major human diseases. Their access to an intracellular lifestyle is reliant on the coordinated release of proteins from the specialized apical organelles called micronemes and rhoptries. A specific phosphatidic acid effector, the acylated pleckstrin homology domain-containing protein (APH) plays a central role in microneme exocytosis and thus is essential for motility, cell entry, and egress. TgAPH is acylated on the surface of the micronemes and recruited to phosphatidic acid (PA)-enriched membranes. Here, we dissect the atomic details of APH PA-sensing hub and its functional interaction with phospholipid membranes. We unravel the key determinant of PA recognition for the first time and show that APH inserts into and clusters multiple phosphate head-groups at the bilayer binding surface.

Keywords: NMR; Plasmodium falciparum; Toxoplasma gondii; egress; exocytosis; gliding motility; invasion; microneme; phosphatidic acid; pleckstrin homology domain.

Graphical abstract

Figure 1

Structural Characterization of PfAPH_106-235 and TgAPH_99-229 Reveals the C-Terminal Region of APH (APH) Adopts a Conserved Pleckstrin Homology Domain-like Fold (A) Schematic representation of Toxoplasma parasite actively invading the host cell. (B) Schematic representation of microneme fusion with parasite plasma membrane. (C) Close-up of fusion event. APH embedded into the microneme surface via acylation interacts with PA accumulating on the inner leaflet of the plasma membrane, facilitating microneme exocytosis. PA is represented in purple, APH in light blue, micronemes and their contents in green. (D) Multiple sequence alignment for APH full-length sequence from different apicomplexan species. Residues are colored in a purple spectrum according to the level of sequence identity, secondary structural elements are indicated, and numbering is shown for PfAPH. The consensus sequence is given below, invariant residues are colored red, highly conserved residues are colored blue, semi-conserved sequence identity is indicated by (+), and invariant residues are indicated by (−). The highly conserved 21 N-terminal residues required for targeting to the micronemal surface are highlighted by a red box, myristoylation (G2) and palmitoylation (C7) lipid anchor sites are indicated by green asterisks. A basic region within the linker sequence containing several conserved basic residues is highlighted by a cyan box. (E) Overlay of ¹⁵N-labeled TgAPH_22-229 (black) and TgAPH_99-229 (green) 2D ¹H-¹⁵N HSQC spectra. In comparison to TgAPH_99-229, additional backbone amide peaks belonging to the linker region are visible in the TgAPH_22-229 spectrum. There is expected to be an additional 77 backbone amide peaks in this linker region, but it is estimated only ∼54 peaks are visible. Residues that could be assigned in TgAPH_99-229 are labeled, sc indicates resonances could be assigned to side chains (W161sc and W215sc). (F) Left, aligned cartoon representations of the lowest-energy structures calculated for PfAPH_106-235 (PDB: 6F24, blue) and TgAPH_99-229 (PDB: 6F8E, green), the first 11 and 10 residues are omitted from PfAPH_106-235 and TgAPH_99-229 respectively as these were shown to be disordered. Right, ensembles of the ten lowest-energy structures calculated for PfAPH_106-235 and TgAPH_99-229.

Figure 2

Mapping the APH:PA Interface (A) Overlay of representative 2D ¹H-¹⁵N HSQC spectra of PfAPH_106-235 recorded upon titration with increasing molar ratios of short-chain PA. HSQC spectra are colored according to the molar ratio between ¹⁵N-labeled PfAPH_106-235 and short-chain PA; black 1:0, green 1:1, blue 1:3, orange 1:7, purple 1:15. (B) Plot of CSPs observed in (A) upon titration with 15-fold molar excess of short-chain PA, versus PfAPH_106-235 sequence number. Residues that could not be assigned are indicated by a gray bar. Prominent CSPs are categorized as greater than 2σ from the mean noise (0.041 ppm), which is represented by a dotted line. (C) CSPs mapped onto the structure of PfAPH_106-235, colored in a 20-interval red spectrum. A more intense coloring indicates a greater CSP as each interval represents 0.5σ from the mean noise. Key residues clustered around the β1/2 strands and β3-β4 loop region are labeled, unassigned residues are colored dark gray. (D) Representative ¹H-¹⁵N HSQC spectra and ¹⁵N 1D profiles for PfAPH_106-235 recorded in the presence of PA-enriched bicelles doped with and without a paramagnetic 5% PE-DTPA-Gd3+ lipid. PREs and therefore proximity to the PA-binding sites are indicated by a reduction in peak intensity. (E) Plot of peak intensity reduction observed in (D) relative to the mean noise (61.80%), which is shown as the baseline, versus PfAPH_106-235 sequence number. (F) PREs mapped onto the structure of PfAPH_106-235, residues are colored if greater than 1σ (yellow), 2σ (orange), or 3σ (red) from the mean noise, while unassigned residues are colored dark gray.

Figure 3

APH Specifically Binds PA-Enriched Membranes (A and B) PfAPH_106-235 1D ¹H NMR spectral region corresponding to the upfield-shifted methyl region (0.255 to −0.170 ppm) was monitored upon titration with increasing concentration of LUVs composed of (A) POPC (100%) or (B) POPC and POPA (50%:50%). PfAPH_106-235:LUVs molar ratios: blue, free PfAPH_106-235 in solution; red 1:2; green 1:4; purple 1:7; yellow 1:15; orange 1:20; lime 1:25; black 1:30. (C) This region was monitored upon titration with variable LUV compositions (POPC [100%] green, POPC:POPS [50%:50%] purple, or POPC:POPA [50%:50%] red), integrated, expressed as the fraction of bound protein, and plotted against total lipid concentration to generate binding curves. Data are represented as mean ± 1σ. (D) Apparent dissociation constants (Kd_app) for binding LUVs were calculated from fitting binding curves. Data are shown as mean ± 1σ for fitting curves. (E and F) (E) and (F) are identical to (C) and (D), but for TgAPH_99-229 using the downfield-shifted amide region (9.4–6.4 ppm).

Figure 4

Conserved APH Basic Residues Mediate Binding to PA Within a Membrane Environment (A–D) APH contains only two, highly conserved KxK motifs present on the β1 strand (K138 and K140 in PfAPH, K130 and K132 in TgAPH) and within the β3-β4 loop region (K163 and K165 in PfAPH, K155 and K157 in TgAPH). Coulombic colored surface representation of (A) PfAPH_106-235 and (C) TgAPH_99-229 reveals KxK motifs form patches of solvent-exposed basic charge. 1D NMR LUV titration experiments show that mutation of KxK motifs reduce (B) PfAPH_106-235 and (D) TgAPH_99-229 affinities for PA-enriched LUVs. A glutamate residue (E146 in PfAPH, E138 in TgAPH) present in the β2 strand disrupts the β1 strand KxK surface exposed basic charge (A and C). As measured by 1D NMR LUV titration experiments, mutation of this glutamate residue increases PfAPH_106-235 and TgAPH_99-229 affinity for PA-enriched LUVs (B and D).

Figure 5

In Vivo Functional and Mutagenesis Studies of APH (A) Schematic representation of APH-Ty mutant generation. (B) Western blot analysis of endogenous and second copy TgAPH ± ATc 48 hr. Catalase provides a loading control. (C) Plaque assay on human foreskin fibroblast monolayer 7 days ± ATc. (D) Intracellular growth assays at 24 hr ± ATc treatment, with 24-hr pre-treatment. Data are presented as mean ± 1σ. (E and F) Microneme secretion assay of mutants in the PH domain (E) and linker region (F). Extracellular secreted antigen (ESA) MIC2 was compared with parental strain ± ATc 48 hr. Catalase represents a loading control for parasite number and lysis, GRA1 represents a control for constitutive secretion.

Figure 6

Coarse-Grained MD Simulation of APH Binding to PA-Enriched Membranes (A) Snapshots from an individual binding series of PfAPH_106-235 (gray) to a 10% PA membrane. The hydrophobic residues I143-F144-H145 that become anchored in the membrane are shown as a green surface. POPA residues within 6 Å of the protein surface are shown as spheres colored individually and the lipid head-groups are shown as a transparent red surface. The recruitment of POPA following the initial association is apparent in the final panel. (B) Average occupancy of PA (magenta) and PC (yellow) head-groups averaged over five simulations of 50% PA membranes, PfAPH_106-235 is shown in light blue. POPA is found to be preferentially in the first shell of lipids around the buried anchor residues I143-F144-H145 (green) whereas PC is found in the second annular layer. Rough lipid shell boundaries are indicated by gray-shaded circles. The protein backbone is shown as a gray trace. (C) Relationship between average time (μs) bound to membrane and PA membrane enrichment for PfAPH_106-235 coarse-grained MD simulation (5 μs total simulation time). (D) Binding between PfAPH_106-235 and a fixed concentration of LUVs (500 μM total available lipid) increasingly enriched with PA (Mol% PA). Hill plot analysis indicates PfAPH_106-235 binds to PA in a positively cooperative manner. (E) Comparison between coarse-grained MD simulation and NMR experiments probing binding between PfAPH_106-235 and PA reveal three regions key to interaction with a PA-enriched membrane. Coarse-grained MD simulations indicate residues 99–110 (including β5-β6 loop) are involved in initial contact with a PA-enriched membrane (red). Hydrophobic residues located at the tip of the β1-β2 loop region (green, I143/F144) dip into the membrane. This anchoring is stabilized by electrostatic interaction between conserved charged residues including KxK motifs (blue, K138-K140 and K163-K165), and PA head-groups.