Article pubs.acs.org/jpr
The Plasmodium falciparum Schizont Phosphoproteome Reveals Extensive Phosphatidylinositol and cAMP-Protein Kinase A Signaling Edwin Lasonder,*,† Judith L. Green,‡ Grazia Camarda,§ Hana Talabani,∥ Anthony A. Holder,‡ Gordon Langsley,∥ and Pietro Alano§ †
Centre for Molecular and Biomolecular Informatics, Nijmegen Centre for Molecular Life Sciences, Radboud University Nijmegen Medical Centre, PO Box 9101, 6500 HB Nijmegen, The Netherlands ‡ Division of Parasitology, MRC National Institute for Medical Research, The Ridgeway, Mill Hill, London, NW7 1AA, United Kingdom § Dipartimento di Malattie Infettive, Parassitarie ed Immunomediate, Istituto Superiore di Sanità, Viale Regina Elena n.299, 00161 Rome, Italy ∥ Laboratoire de Biologie Cellulaire Comparative des Apicomplexes, Institut Cochin, Inserm U1016, CNRS UMR 8104, Faculté de Médecine, Université Paris Descartes Sorbonne Paris Cité, 27, rue du Faubourg-Saint-Jacques, 75014 Paris, France S Supporting Information *
ABSTRACT: The asexual blood stages of Plasmodium falciparum cause the most lethal form of human malaria. During growth within an infected red blood cell, parasite multiplication and formation of invasive merozoites is called schizogony. Here, we present a detailed analysis of the phosphoproteome of P. falciparum schizonts revealing 2541 unique phosphorylation sites, including 871 novel sites. Prominent roles for cAMP-dependent protein kinase A- and phosphatidylinositol-signaling were identified following analysis by functional enrichment, phosphoprotein interaction network clustering and phospho-motif identification tools. We observed that most key enzymes in the inositol pathway are phosphorylated, which strongly suggests additional levels of regulation and crosstalk with other protein kinases that coregulate different biological processes. A distinct pattern of phosphorylation of proteins involved in merozoite egress and red blood cell invasion was noted. The analyses also revealed that cAMP-PKA signaling is implicated in a wide variety of processes including motility. We verified this finding experimentally using an in vitro kinase assay and identified three novel PKA substrates associated with the glideosome motor complex: myosin A, GAP45 and CDPK1. Therefore, in addition to an established role for CDPK1 in the motor complex, this study reveals the coinvolvement of PKA, further implicating cAMP as an important regulator of host cell invasion. KEYWORDS: malaria, phosphoproteome, schizonts, inositol/PKA signalling pathways
■
INTRODUCTION The life cycle of the Apicomplexan malaria parasite Plasmodium falciparum is complex and composed of several developmental stages in the mosquito and human hosts. Asexual blood stage development causes the mortality and morbidity associated with malaria. After red blood cell invasion by a single merozoite, the parasite develops within a parasitophorous vacuole (PV), from ring stage to trophozoite, and then undergoes 4 to 5 rounds of DNA synthesis, mitosis and nuclear division that result in a schizont with 16−22 nuclei, which produces and releases new invasive merozoites to continue the cycle.1 Merozoite release and reinvasion requires signal transduction and secretion of organellar contents at egress, and the formation of a moving junction between parasite and host cell surface at invasion that involves transfer of some parasite proteins to the newly invaded erythrocyte: processes that may be controlled in part by calcium flux and calcium-dependent phosphorylation.2 Fine-tuning of the eukaryotic cellular machinery is regulated by several mechanisms involving transcriptional control, post© 2012 American Chemical Society
translational control, and post-translational modification of proteins (PTM). It has become evident that protein kinases and signal transduction pathways are integral to regulation of the Plasmodium parasite life cycle.3,4 In other organisms interand intracellular signals trigger the release of secondary messenger molecules that activate protein kinases, which transmit signals to the nucleus via a cascade of phosphorylation events to regulate the transcription machinery of the cell. P. falciparum is considered in this respect an “insensitive beast”, because there is a relatively small number of known external stimuli,5 and because it apparently lacks the tyrosine kinase signaling pathways found in most eukaryotes that are activated by signals from outside the cell (e.g., growth factors). Phylogenetic studies6,7 revealed that the kinome of P. falciparum (approx 100 eukaryotic protein kinases (ePKs)) diverges profoundly from those of mammalian species, with a large proportion of kinases classified either as semiorphan ePKs Received: June 20, 2012 Published: October 1, 2012 5323
dx.doi.org/10.1021/pr300557m | J. Proteome Res. 2012, 11, 5323−5337
Journal of Proteome Research
Article
PAGE) loading buffer and fractionated on a 10% polyacrylamide gel. The pellet fraction was loaded onto four 12.7 mm wide wells, and the soluble fraction was loaded onto 10 12.7 mm wells using Biorad Protean III mini gels. Following electrophoresis, gels were stained with colloidal Coomassie blue stain (Severn Biotech), and each track was cut into 5 slices. Proteins in the gel slices were reduced with dithiothreitol (DTT), alkylated with iodoacetamide and digested by trypsin as described.20 Digested samples were acidified with a final concentration of 0.1% TFA, and peptides were purified by large STAGE tips,21 which were made from two 8 mm diameter plugs of 3 M Empore C18 Extraction disk inserted into a 0.5 mL Handee spin column (Pierce).
that do not cluster directly with a mammalian orthologue, or as orphan Plasmodium specific ePKs such as the FIKK kinase family. Thus, unravelling the Plasmodium signal transduction pathways cannot be achieved by bioinformatic approaches alone. Gene disruption studies of several ePKs (reviewed in refs 3, 8) show that many are essential for asexual proliferation at the blood stage and many are important at other stages of the life cycle. Ca2+-dependent phosphorylation has well documented roles in parasite motility and microneme secretion,4 and protein phosphorylation mediated by cAMP-dependent protein kinase A (PKA) has recently been shown to be involved in two very different areas: anion channel regulation and erythrocyte invasion.9,10 Phosphoinositides (PIs) are essential components of membranes in eukaryotes and play key roles in many biological processes. PIs can serve both as docking sites at the membrane for components of signaling pathways and as precursors of lipid second messengers; for a review, see ref 11. Reversible phosphorylation of the myo-inositol headgroup of phosphatidylinositol (PtdIns) at positions 3, 4, and 5 gives rise to the seven known PI isoforms. Individual PI isoforms act as unique lipid signaling entities, where for example, they contribute to cytokinesis12 and vesicle traffic and fusion with membranes.13 Tremendous advances have been made recently in global phosphoproteome analyses of several organisms that reveal with high confidence thousands of sites collected in dedicated databases such as Phospho.ELM14 and Phosida.15 However, at the start of this study protein phosphorylation in P. falciparum had been little studied with no more than 21 known phosphorylation sites.16,17 Thus, unravelling Plasmodium signal transduction pathways and other functions of phosphorylation urgently required high throughput analysis of in vivo phosphorylation by mass spectrometry. The aim of this study was 2-fold: to generate an in-depth phosphoproteome data set for the P. falciparum blood stage schizont and to analyze this phosphoproteome by bioinformatic and biochemical methods to identify prominent signaling pathways that are promising drug targets.
■
Sample Preparation by the Gel-Free Filter Assisted Sample Preparation (FASP) Method
4 × 108 P. falciparum schizont-infected erythrocytes and 1 × 108 uninfected RBCs were lysed by freeze/thaw and divided into soluble and pellet fractions by centrifugation as described above, after the addition of Halt phosphatase inhibitor cocktail (Thermo Fisher) and complete protease inhibitor cocktail (Roche). The Halt phosphatase cocktail is a 100× concentrated solution of sodium fluoride, sodium orthovanadate, sodium pyrophosphate and β-glycerophosphate, and was diluted 100 times, and the complete protease inhibitor cocktail tablet contains AEBSF, aprotinin, bestatin, E-64, leupeptin, pepstatin A and EDTA and was diluted according to the specifications of the manufacturer. Additionally, hemoglobin was depleted from the soluble fraction with HemoVoid beads according to the manufacturer’s instructions (http://www.biotechsupportgroup. com/), to yield a hemoglobin-depleted and a hemoglobinenriched soluble fraction. The soluble cellular fractions were diluted 1 to 1 in 100 mM Tris-HCl pH 8.0, 8 M urea and further processed according to the FASP protocol,22 while the pellet fraction was dissolved in 100 μL of phosphate buffered saline (PBS) containing 4% SDS and then diluted 1 to 1 in 100 mM Tris-HCl pH 8.0, 8 M urea. The fractions were placed in a 0.5 mL Amicon Ultra 30K spin filter unit (30 kDa cutoff, Millipore), reduced with 10 mM DTT at room temperature for 25 min, concentrated by centrifugation at 14000g and alkylated in the dark with 50 mM iodoacetamide at room temperature for 25 min. The samples were concentrated by centrifugation and diluted 10-fold with 100 mM Tris-HCl pH 8.0 containing 8 M urea. This step was repeated 3 times for the soluble fractions and 6 times for the pellet fraction. Following the washes, 5 μg of lys C protease (Wako) per cellular fraction was added in 100 mM Tris-HCl pH 8.0, containing 4 M urea and incubated overnight at room temperature. Samples were diluted with 50 mM ammonium bicarbonate to a final concentration of 2 M urea and then incubated with 5 μg of trypsin (Promega) for 18 h. Digested samples were acidified to a final concentration of 0.1% TFA and purified as described above using large STAGE tips. Purified digests were dissolved in 500 μL of Britton and Robinson buffer, which is composed of 20 mM CH3COOH, 20 mM H3PO4, 20 mM H3BO3, adjusted to pH 11.0 with 1 M NaOH, and loaded on large anion exchange STAGE tips, made from four 3 M Empore anion exchange extraction disks. Peptides were eluted with Britton and Robinson buffer solutions of pH 8.0, 6.0, 5.0, 4.0 (adjusted with 1 M NaOH to the desired pH) followed by a final elution with 1% TFA. Fractionated peptides were acidified to a final concentration of 0.1% TFA and purified by large STAGE tips.
EXPERIMENTAL PROCEDURES
Parasites
P. falciparum clone 3D7 18 was cultured in human 0 + erythrocytes provided by Prof. G. Girelli, Università La Sapienza, Rome in medium containing 10% human serum as described previously.19 Percoll purified schizonts were used to seed a 50 mL culture at 5% hematocrit, which was sorbitol treated to obtain a ring stage culture with a 4 h synchronization window. Parasites were harvested after two reinvasion cycles in a total culture volume of 400 mL at 3.5% parasitaemia, when analysis of Giemsa-stained smears indicated that 90% of parasites were at the multinucleated schizont stage (from 2 nuclei to segmenters). Parasitized red blood cells (RBCs) were harvested after purification on a 60% Percoll cushion, yielding a sample containing 8 × 108 parasitized and 1 × 108 uninfected RBCs, which were washed in PBS and snap frozen in liquid nitrogen. Sample Preparation by Gel-Electrophoresis
A P. falciparum-infected erythrocyte sample (4 × 108 infected RBCs and 1 × 108 uninfected RBCs) was lysed by a freeze/ thaw procedure as described20 and divided into a soluble and a pellet fraction following centrifugation. Each sample was solubilized in SDS-polyacrylamide gel electrophoresis (SDS5324
dx.doi.org/10.1021/pr300557m | J. Proteome Res. 2012, 11, 5323−5337
Journal of Proteome Research
Article
Phosphopeptide Enrichment Using TiO2 Beads
2−3 times with exclusion lists of previous runs in segments of 10 min.
Phosphopeptides obtained from 8 × 10 iRBCS + 1 × 10 RBCs (approximately 2 mg of protein determined by the micro BCA protein assay kit from Thermofisher) were enriched by batch-wise incubation with Titansphere 10 μm TiO2 beads (GL Sciences, Inc., Japan) as described23 with some modifications. Before use, TiO2 beads were washed with 80% acetonitrile, 0.1% TFA and suspended in 30 mg/mL of dihydroxybenzoic acid in 80% acetonitrile, 0.1% TFA and diluted 1:4 with 0.1% TFA before use. Twenty microliters of this slurry (1 mg beads) was added to all peptide digest fractions in 1.5 mL Eppendorf vials and incubated under continuous shaking for 1−2 h. Then, the TiO2 bead slurries were transferred to a microcolumn made from a 200 μL precision pipet tip with an inserted 2 mm fused silica frit (50 μm internal diameter). The unbound fractions were collected and mixed with a freshly prepared portion of 1 mg beads and reincubated for 1−2 h. This step was performed twice resulting in 3 portions of TiO2 beads with bound phosphopeptides per peptide fraction. Then, the beads in the micro columns were washed three times with 100 μL of 30% acetonitrile, 3% TFA, followed by three washes with 100 μL 80% acetonitrile, 0.3% TFA. Phosphopeptides were eluted from the beads sequentially with 100 μL NH4OH in 20% acetonitrile pH 11.0, and 100 μL 5% NH4OH, 5% piperidine, and 5% pyrrolidine.24 Finally, the eluted peptides were acidified with TFA, purified using STAGE tips21 and dissolved in 0.1% TFA and 10 mM EDTA. 8
8
Peptide Identification and Validation by MASCOT and MaxQuant
Raw spectrum files were converted into Mascot generic peak lists by MaxQuant25 version 1.013.13 (http://maxquant.org//) using different peak reduction settings for improved sensitivity.26 Peak lists were generated for the top 3, 6, 12, 18, 24, 30, 35, 40, and 50 peaks per 100 Da window. Peptides and proteins were identified by searching the peak lists with Mascot version 2.2 (Matrix Science) against the P. falciparum database (version 7.0, 5491 sequences downloaded on 13 Nov 2011 from http:// plasmodb.org/plasmo/) supplemented with the human International Protein Index (IPI) database version 3.65 (downloaded on 16 Oct 2009 from ftp://ftp.ebi.ac.uk/pub/databases/ IPI/; 86 379 sequences) and frequently observed contaminants and concatenated with reversed copies of all entries. Mascot search parameters for protein identification specified a mass tolerance of 10 ppm for the parental peptide and 0.5 Da for fragmentation spectra and a trypsin specificity allowing up to 2 miscleaved sites. Carboxyamidomethylation of cysteines was specified as a fixed modification, and phosphorylation at serine, threonine and tyrosine, oxidation of methionine, deamidation of glutamine and asparagine and protein N-terminal acetylation were set as variable modifications. The required minimal peptide length was set at 6 amino acids. Internal mass calibration of measured ions and peptide validation by the target decoy approach was performed by MaxQuant as described.25 We accepted proteins with a false discovery rate (FDR) better than l % and multiple charged peptides with a Mascot score >15, a FDR better than 1% and not more than 1 variable oxidized M and 1 variable deamidation (Q,N). Redundancy in identifications from different peak list settings was removed by requiring unique best localization MS/MS scan numbers per MS file. For redundant identifications, priority was given to the identification with the lowest posterior error probability (PEP) and then to the highest localization probability for all peptides with Mascot scores higher than 15. Next, a PEP threshold of 0.025 was set to obtain a phosphopeptide FDR at 1%. These validated phosphopeptides were classified according to their phosphosite PTM localization probability in peptides with mapped sites with probabilities better than 0.75 (class Ia, Table S1, Supporting Information) and sites with lower probabilities.
Liquid Chromatography Tandem Mass Spectrometry
Tandem mass spectrometry experiments were performed using the nano EASY LC system (Proxeon, Denmark) connected to a 7-T linear ion trap cyclotron resonance Fourier transform (LTQ-Ultra FT) mass spectrometer (Thermo Fisher, Bremen, Germany) with 15 cm 100 μm internal diameter PicoTip columns (New Objective, Woburn, USA) packed with 3 μm Reprosil C18 beads (Dr. Maisch GmbH, Ammerbuch, Germany). Peptides were separated by a 90−120 min gradient of 92% buffer A/8% buffer B to 73% buffer A/27% buffer B (buffer A contains 0.5% acetic acid and buffer B contains 0.5% acetic acid in 100% acetonitrile) with a flow-rate of 300 nL/ min. Peptides eluting from the column tip were electrosprayed directly into the mass spectrometer with a spray voltage of 2.2 kV. Data acquisition with the LTQ-FT Ultra instrument was performed in a data-dependent mode to automatically switch between MS, MS2 and phosphoric acid neutral loss triggered MS3. Full-scan MS spectra of intact peptides (m/z 350−1500) with an automated gain control accumulation target value of 1 000 0000 ions were acquired in the Fourier transform ion cyclotron resonance cell with a resolution of 50 000. The four most abundant ions were sequentially isolated and fragmented in the linear ion trap by applying collision induced dissociation using an accumulation target value of 10 000, a capillary temperature of 100 °C, and a normalized collision energy of 27%. Multistage activation was switched on for neutral loss dependent MS3 fragmentation on the masses of phosphoric acid at charge states 2+, 3+ and 4+. The four most abundant ions were sequentially fragmented again under identical settings as the first selection, with the exception of a normalized collision energy of 40%. A dynamic exclusion of ions previously sequenced within 180 s was applied. All unassigned charge states and singly charged ions were excluded from sequencing. A minimum of 500 counts was required for MS2 selection and 5 counts for MS3 selection. Peptide fractions were measured
Gene Ontology Analysis
All P. falciparum gene ontology analyses were performed with the software package Ontologizer (http://compbio.charite.de/ index.php/ontologizer2.html), with the following Open Biological Ontology and Gene association components from http://www.geneontology.org: gene ontology v1.2.obo, goslim_generic.obo and Gene_association.GeneDB-Pfalciparum_2011-5-31. The OPI GO terms were taken from ref 27 and rearranged to a gene association compatible file. Ontologizer was used to identify overrepresented GO terms for the schizont phosphoproteome relative to the background of the P. falciparum proteome (∼5500 proteins). GO term enrichment was computed by the parent−child union approach and corrected for multiple testing by the Benjamini and Hochberg method and was considered significant for adjusted p-values lower than 0.05.28 GO enrichment analysis for phosphorylated human proteins was carried out using the web tool Database for Annotation, 5325
dx.doi.org/10.1021/pr300557m | J. Proteome Res. 2012, 11, 5323−5337
Journal of Proteome Research
Article
Figure 1. Overview of the phosphoproteome experimental procedures and their output. (A) Schematic workflow of the mass spectrometry procedures to detect phosphopeptides in tryptic digests derived from 4 × 108 infected RBCs and 1 × 108 uninfected RBCs (approximately 2 mg of total protein) P. falciparum schizonts were lysed and separated into cytosolic and pellet fractions and subsequently processed by two approaches to generate 10 and 18 peptide fractions for approach I and II, respectively. Approach I utilized a gel-based approach with a 1D SDS PAGE gel, and approach II was based on a gel-free digestion method by FASP in combination with strong anion exchange (SAX) disks for peptide fractionation. A detailed description of the procedure is provided in the Experimental Procedures. (B) Overview of the number of unique phosphopeptides prepared by approach I and II, which were identified with a confidence level better than 1% FDR at the phosphopeptide level and phospho site PTM localization probabilities better than 75%. (C) Distribution of P. falciparum phosphorylated serine, threonine and tyrosine residues in the phosphoproteome.
Motif Analysis
Visualization and Integrated Discovery (DAVID, version 6.7. http://david.abcc.ncifcrf.gov/) using standard settings and corrected for multiple testing as above.
Phosphorylation sites were categorized by their chemical properties as acidic, basic, proline-directed, tyrosine or other by a decision tree method from ref 29 as follows: (1) Get the 6 neighboring amino acids before and after the phosphorylation site. (2) pY at position 0 then classify as “tyrosine”. (3) P at +1 then classify as “proline-directed”. (4) Positions +1 to +6 contain more than one D and E residues then classify as “acidic”. (5) K or R at position −3 then classify as “basic”. (6) D or E at +1, +2, or +3 then classify as “acidic”. (7) Between −6 and −1 more than 2 K or R residues then classify as “basic”. (8) Remaining peptides are classified as “other”. Phosphorylation motifs were identified using Motif X30 that tested for motif overrepresentation in the phosphorylation data proteins with the following parameters: phosphorylation motif window = 13 amino acids, p-value threshold = 1 × 10−4 for S, T
Domain Enrichment Analysis
P. falciparum domains collected from several databases (PFAM, SUPERFAMILY, TIGRFAMs, SMART, PROFILE, PRINTS, PIR, PRODOM) were downloaded from PlasmoDB (http:// plasmodb.org/plasmo/). Redundancy between domain annotations from different sources was removed by prioritizing domain annotations with the lowest E-value. Overrepresented domains in the schizont phosphoproteome relative to the background of all P. falciparum predicted proteins (∼5500 proteins) was determined by the Fisher exact test adjusted for multiple testing by the Benjamini and Hochberg correction. We accepted domains that were detected more than 5 times in the proteome with adjusted p-values 5, and a background of all P. falciparum PlasmoDB 7.0 proteins. Motif X analysis was performed for a normal and a degenerate amino acid set. The degenerate amino acid set was enabled for conservative amino acid substitutions within the motif window according to the following: A = AG, D = DE, F = FY, K = KR, I = ILVM, Q = QN, S = ST, CC, H H, P = P, W = W. When different motifs were found for a peptide by the analyses with different amino acid residues, priority was given to the motif with the highest Motif X score. The motifs were matched to known motifs using the motif matcher of Phosida (http://www.phosida.de/) and the phosphomotif finder of the HPRD database (http://www. hprd.org/PhosphoMotif_finder).
followed by digestion with trypsin, and phosphopeptide enrichment with titanium dioxide (TiO2) binding, and then analysis by liquid chromatography tandem mass spectrometry (LC−MS/MS); and (2) filter assisted sample preparation (FASP) that includes digestion with trypsin, subsequent strong anion exchange chromatography (SAX) to fractionate peptides according to their isoelectric points, and finally TiO2-based phosphopeptide enrichment and LC−MS/MS measurements. Peptides were identified by searching the mass spectrometry data with Mascot against a combined database of P. falciparum and human proteins and were validated with the target−decoy methods of the MaxQuant software,25 where in addition to the standard validation criterion of 1% false discovery rate (FDR) for peptides and proteins, we required a minimal peptide Mascot score of 15 and a maximal posterior error probability (PEP) of 2.5% to limit the phosphopeptide FDR to 1%. The MaxQuant analysis software includes a PTM scoring algorithm35 that was applied to those phosphopeptides that contain more than one potential phosphorylation site (either Ser, Thr, or Tyr) for finding the most likely phosphorylation site. The validated phosphopeptides were grouped into 2737 class Ia unique P. falciparum and human peptides (Table S1, Supporting Information) with a PTM localization probability score ≥0.75 and 566 class Ib unique peptides, where the phosphorylation site is not precisely known with PTM localization probability scores
