Quantitative in vivo analyses reveal calcium-dependent phosphorylation sites and identifies a novel component of the Toxoplasma invasion motor complex

Abstract

Apicomplexan parasites depend on the invasion of host cells for survival and proliferation. Calcium-dependent signaling pathways appear to be essential for micronemal release and gliding motility, yet the target of activated kinases remains largely unknown. We have characterized calcium-dependent phosphorylation events during Toxoplasma host cell invasion. Stimulation of live tachyzoites with Ca²⁺-mobilizing drugs leads to phosphorylation of numerous parasite proteins, as shown by differential 2-DE display of ³²[P]-labeled protein extracts. Multi-dimensional Protein Identification Technology (MudPIT) identified ∼546 phosphorylation sites on over 300 Toxoplasma proteins, including 10 sites on the actomyosin invasion motor. Using a Stable Isotope of Amino Acids in Culture (SILAC)-based quantitative LC-MS/MS analyses we monitored changes in the abundance and phosphorylation of the invasion motor complex and defined Ca²⁺-dependent phosphorylation patterns on three of its components--GAP45, MLC1 and MyoA. Furthermore, calcium-dependent phosphorylation of six residues across GAP45, MLC1 and MyoA is correlated with invasion motor activity. By analyzing proteins that appear to associate more strongly with the invasion motor upon calcium stimulation we have also identified a novel 15-kDa Calmodulin-like protein that likely represents the MyoA Essential Light Chain of the Toxoplasma invasion motor. This suggests that invasion motor activity could be regulated not only by phosphorylation but also by the direct binding of calcium ions to this new component.

Figure 1. 2-DE display of protein phosphorylation events following in vivo stimulation of Ca²⁺ pathways in Toxoplasma gondii tachyzoites.

Toxoplasma parasites were incubated in media containing ³²[P]-sodium orthophosphate to radioactively label intracellular ATP pools and calcium-dependent signal transduction was stimulated in vivo by treating free tachyzoites with Ca²⁺-mobilizing drugs, as detailed in Materials & Methods. As negative controls parasites were pre-incubated in the presence of membrane-permeable Ca²⁺-chelator BAPTA-AM (A) or treated with DMSO solvent alone (C), and for Ca²⁺ pathway stimulation parasites were incubated with 1% (ie. 172 mM) ethanol (B) or 100 µM ionomycin (D), respectively. Arrows indicate the positions of 50 ³²[P]-labeled 2D protein spots that are exclusively phosphorylated following Ca²⁺ signaling pathway stimulation. E) A false colour image overlay of corresponding unstimulated (C, green) or stimulated (D, red) autoradiographs showing the relative position of multiple confident LC-MS/MS protein identifications (<1% false protein identification rate). A detailed list of proteins IDs to accompany this figure is provided (Supplementary Table S1). F) Radio-immunoprecipitation analyses confirm the phosphorylation of HA-tagged candidate parasite phosphoproteins HA-PKA-R, PP2C-HA, ARM1-HA, MLC1-HA. Free tachyzoites expressing HA-tagged parasite proteins were radioactively labeled with ³²[P]-sodium orthophosphate and treated with 1% ethanol in the presence (+) or absence of BAPTA-AM (-), as indicated. Anti-HA Western blots (αHA, left panels) and corresponding autoradiographs (³²[P], right panels) of radio-immunoprecipitates demonstrate a Ca²⁺-dependent increase in the phosphorylation of HA-reactive tachyzoite protein bands (arrows). Abbreviations: PKA-R, cAMP-dependent protein kinase A regulatory subunit; GAP45, Glideosome-Associated Protein 45; CKIIβ, Casein kinase II beta subunit; ANP32, IPP2A-1/ANP32-like protein; SET, IPP2A-2/SET-like protein; SNAP, alpha-soluble NSF attachment protein; PP2C, putative protein phosphatase 2C; ARM1, armadillo repeat-containing protein; MLC1 myosin light chain 1, RAB5B, GTP-binding protein RAB5B; Hypo, hypothetical phosphoprotein.

Figure 2. Functional annotation analysis of Toxoplasma phosphoproteins identified by MudPIT and comparison of proteomics strategies.

A) Venn diagram summarizing the number of confident tachyzoite phosphoprotein identifications (<1% FDR) detected by 2-DE- or MudPIT-based LC-MS/MS analyses. A comparison of 2-DE and MudPIT datasets enabled us to cross-validate 10 major Ca²⁺-dependent tachyzoite phosphoproteins represented by 21 ³²[P]-labelled 2-DE spots and 86 phosphopeptide spectra (see Supplementary Figure S2, Tables S1-S4 for details). B) Summary of LC-MS/MS evidence for Ca²⁺-dependent tachyzoite phosphoproteins and their associated phosphorylation sites. Listed in the columns (from left to right) are: Top rows (grey background): The ToxoDB entry (ToxoDB release 6.0), protein name, number of 2-DE spot(s) in which each protein was identified (cross-reference to Supplementary Table S1), amino acid sequence coverage, molecular weight (Mr), isoelectric point (pI) and, based on 2-DE gel LC-MS/MS identifications. Bottom row (white background): The database accession (ToxoDB release 4.0 or UniProt entries), predicted phosphorylation site(s), total number of unique spectra (rank 1), highest individual Mascot score (rank 1), delta Mascot score (rank 2), based on LC-MS/MS analyses of MudPIT data. The above information was used in conjunction with independent literature searches to assign keywords associated with Biological Processes or Cellular Components and identify homologs and conserved phosphorylation sites in P. falciparum, as indicated in the last column. ^(a-o) MS/MS evidence spectra shown in Supplementary Figure S2.

Figure 3. Quantification of calcium-dependent regulation of phosphorylation sites of Toxoplasma invasion motor complex components.

A) Work flow to identify individual phosphorylation sites and quantitatively assess their responsiveness to calcium signals using a SILAC-based proteomics approach. A 1∶1 mixture of Triton X-100 lysates from “Heavy” (H; Arg4/Lys8)-labeled ethanol-stimulated tachyzoites or "Light" (L; Arg0/Lys0)-labeled non-stimulated parasites was generated, and a TiO₂-enriched phosphopeptide sample of H/L-labeled Toxoplasma invasion motor complexes was prepared and analysed by LC-MS/MS on an LTQ-Orbitrap instrument. Mascot and MaxQuant search engines facilitated subsequent manual identification, phosphosite localization and quantification of proteins or peptides as detailed in materials and mathods. B) Sypro Ruby-stained SDS-PAGE separation of the relative amounts of light (lane 1) or heavy (lane 2) Triton X-100 whole protein extracts are shown. Intact tachyzoite invasion motor complexes comprising the five major components MyoA, GAP50, GAP45 and MLC1 were precipitated from a 1∶1 H/L mixture by GAP45-specific immuno-affinity chromatography (lane 3).

Figure 4. Identification of a potential Essential Light Chain of the Toxoplasma myosin motor.

(A) Immunofluorescence analyses of Toxoplasma parasite lines expressing HA-tagged ELC1 confirms the co-localization of this proteins with GAP45 at the parasite periphery, and (B) demonstrates a stable association with the inner membrane complex. (C-F) Immunoprecipitation analyses validate that ELC1 is an integral component of the intact MyoA invasion motor complex machinery. C) Western blot analyses of anti-HA immunoprecipitates prepared from detergent-soluble protein extracts of wild-type (WT) or transgenic parasites expressing HA-tagged MLC1 (MLC1-HA) or ELC1 (ELC1-HA) were probed by Western blot using antibodies against the HA epitope tag, as indicated. Arrows show the relative size of MLC1-HA (∼35 kDa) and ELC1-HA (∼18 kDa) protein bands. We also observed smaller immunoreactive MLC1-HA or ELC1-HA bands in HA pull downs (*). Western blot analyses using specific rabbit polyclonal antibodies against MyoA (top), GAP45 (center) or MLC1 (bottom) confirm the co-purification of other invasion motor complex components in anti-HA immunoprecipitates of parasite lines expressing MLC1-HA (lane 2) or ELC1 (lane 3), but not wild-type controls (lane 1). Arrows show the sizes of endogenous Toxoplasma MyoA, GAP45, MLC1, and the size of the immunoreactive band corresponding to MLC1-HA (lane 2, arrowhead). D) Immunoprecipitates of parasites expressing MLC1-HA were purified using magnetic microbeads coated with anti-HA antibodies and eluted proteins stained with Sypro Ruby. Major bands corresponding to MyoA, GAP50, GAP45, GAP40, and MLC1 were confirmed by LC-MS/MS. A ∼15 kDa protein band was precipitated in addition to the other known invasion motor complex components and were excised from a preparative 10% SDS-PAGE gel (arrow). This protein band yielded 44% sequence coverage for ELC1 (Supplementary Figure S8). E, F) Immunoprecipitates of parasites expressing ELC1-HA were prepared using anti-GAP45 antibodies. Eluted proteins were stained with Sypro Ruby (E) or probed by Western blot using anti-HA antibody (F). A ∼18 kDa protein band corresponding to ELC1-HA was detected in addition to the other known invasion motor complex components (E, arrow). Western blot analyses confirm the co-purification of ELC1-HA with motor complexes prepared from parasite lines expressing HA-tagged ELC1 using specific rabbit polyclonal antibodies against GAP45 (lane 2, arrow), but not in pull-downs prepared using non-specific rabbit IgG (lane 1). An asterisk indicates a putative degradation product, proteolytic fragment or posttranslational modification of ELC1-HA. G.) Structural model of Toxoplasma MyoA-tail (green) interacting with MLC1 (Cyan) and the newly identified ELC1 (Magenta) based on P. polycephalum myosin regulatory complex .

Figure 5. Predicted structure and phosphorylation of Toxoplasma GAP45.

A. Domain model of Toxoplasma GAP45 structure. The protein encodes a short lipid anchored N-terminal domain (aa1-15), an extended alpha-helical coiled-coil domain (aa16–151), a region predicted to be intrinsically disordered (aa152–192) and a globular C-terminal domain (aa193–245). The upper limit of the dimension of the coiled-coil, intrinsically disordered and globular domains derived from structural modeling studies are shown. B. Prediction of the GAP45 coiled-coil and disordered domains. Plots of the overall probability of GAP45 amino acid residues to form a coiled coil (black line) or form a dimeric coiled-coil (broken blue line) are based on the Multicoils program . Plots of the disorder tendency of the unphosphorylated GAP45 amino acid sequence (blue line, -phos) or GAP45 sequence with glutamate replacement of all S/T/Y residues in region aa152-192 are based on the IUPred algorithm . C. ClustalW v.2.1 alignment of GAP45 amino acid sequences from (top to bottom) P. falciparum (PFL1090 w), P. knowlesi (PKH_143920), P. vivax (PVX_123765), P. chabaudi (PCAS_143960), P. yoelii (PY03448), Neospora caninum (NCLIV_048570) and T. gondii (TGME49_023940). Highly conserved residues are highlighted in red and potential S/T/Y phosphorylation sites are highlighted in yellow. The positions of PfCDPK1 in vitro phosphorylation sites of recombinant PfGAP45 identified by Winter et al. (2009) (top) and in vivo phosphorylation sites localized on Toxoplasma GAP45 in this study (bottom) are shown. The sequence of a merozoite-derived PfGAP45 phosphopeptide identified by Green et al and the S163 and S167 phosphorylation sites of TgGAP45 are marked (*), and residues that are part of the coiled-coil (C) alpha helices (H) or beta-sheets (E) are annotated. Arrows indicate Ca²⁺-insensitive (↔) or Ca²⁺-sensitive phosphorylation sites (↑) on TgGAP45 identified in this study.

Figure 6. Predicted structure and phosphorylation of Toxoplasma MLC1.

A. ClustalW v.2.1 alignment of the N-terminal sequences of myosin A tail interacting protein (MTIP) from (top to bottom) Plasmodium falciparum (PFL2225 w), P. knowlesi (PKH_146380), P. vivax (PVX_101215), P. chabaudi (PCAS_146180), P. yoelii (PY00409), and MLC1 protein sequences of Neospora caninum (NCLIV_029420) and T. gondii (TGME49_057680). The positions of PfMTIP residues S47 and S51 phosphorylated by PfCDPK1 in vitro (top), and Ca²⁺-independent TgMLC1 phosphorylation site S55 (↔) localized in this study are shown. Residues that are part of the conserved N-terminal region or intrinsically disordered region of MLC1 are highlighted in red or yellow, respectively. The sequence of a post-translationally modified Toxoplasma MLC1 peptide known to inhibit myosin motor activity is marked . B) Domain structure and ClustalW alignment of the C-terminal sequences of MTIP/MLC1 proteins. The positions of PfMTIP residues S85/6 and S107/9 phosphorylated by PfCDPK1 in vitro (top), and Ca²⁺-dependent Toxoplasma MLC1 phosphorylation site T98 and S132 (Ca²⁺↑) localized in this study are shown. C) Structural model of the Toxoplasma MyoA regulatory tail domain bound to MLC1 of GAP45. The positions of the two Ca²⁺-dependent in vivo phosphorylation sites T98 and S132 (Ca²⁺↑) are highlighted.