Quantitative Analysis of Kinase-Proximal Signaling in Lipopolysaccharide-Induced Innate Immune Response Kirti Sharma,†,§ Chanchal Kumar,‡,# Gyo ¨ rgy Ke´ri,|,⊥ Susanne B. Breitkopf,† † Felix S. Oppermann, and Henrik Daub*,† Department of Molecular Biology and Department of Proteomics and Signal Transduction, Max Planck Institute of Biochemistry, Am Klopferspitz 18, 82152 Martinsried, Germany, Vichem Chemie Ltd., Herman Otto´ u. 15., Budapest, 1022, Hungary, and Pathobiochemistry Research Group of the Hungarian Academy of Science, Semmelweis University, Puskin u. 9., Budapest, 1088, Hungary Received December 23, 2009
The innate immune system senses invariant microbial components via toll-like receptors (TLRs) to elicit a host defense program against invading pathogens. Lipopolysaccharide (LPS), a constituent of Gramnegative bacteria, is recognized by TLR4 and triggers protein kinase signaling to orchestrate immune responses such as inflammatory cytokine production. To analyze kinase-proximal signaling in murine macrophages, we performed prefractionation experiments with immobilized kinase inhibitors to enrich for protein kinases and their interaction partners. In conjunction with SILAC-based quantitative mass spectrometry and phosphopeptide enrichment, we recorded five time point profiles for more than 850 distinct phosphorylation events on protein kinases and copurifying factors. More than 15% exhibited significant changes and many of those mapped to LPS-regulated kinase networks. We identified many unreported TLR signaling events including LPS-triggered phosphorylations of Akt substrates, which point to previously unknown molecular mechanisms in innate immune response. We further detected extensive phosphoregulation of TANK-binding kinase 1, inhibitor of nuclear factor-κB kinase ε and their associating scaffolding factors, and none of these events were known despite the key roles of these proteins in LPS signaling. Thus, our data expands previous knowledge for functional analyses of innate immune response. Keywords: phosphoproteomics • SILAC • affinity purification • protein kinases • signaling • lipopolysaccharide • innate immune response • bioinformatics
Introduction
genes involved in cell proliferation, migration, and survival as well as inflammatory and immune responses.4
Toll-like receptors (TLRs) play a central role in the priming of immune responses against invading pathogens and represent the first line of defense in humans. TLRs are activated upon extracellular recognition of pathogen-associated molecular patterns.1 One of the most intensively studied receptors is TLR4 which is the exclusive TLR for lipopolysaccharide (LPS), a major component of the outer cell membrane of Gram-negative bacteria.2,3 LPS-induced TLR4 activation mediates intracellular signaling cascades which orchestrate a gene expression program that is fundamental to the macrophage-dependent immune response. These changes include the regulation of a broad spectrum of
At the molecular level, TLR4 activation triggers two intracellular signaling pathways through recruitment of Toll/interleukin-1 receptor-like domain (TIR)-containing adaptor molecules to the TIR domain of the TLR. Binding of the adaptor protein MyD88 activates nuclear factor kappa B (NF-κB) and activator protein 1 (AP1), eventually leading to the transcription of genes encoding pro-inflammatory cytokines and chemokines such as tumor necrosis factor R, interleukin 12 and interferonγ. The TIR-domain-containing adaptor protein-inducing interferon-β (TRIF)-dependent pathway leads to the activation of genes involved in the production of interferon beta. Notably, these transcriptional responses are mediated through LPStriggered phosphorylation events on signaling factors such as inhibitor of NF-κB (IκB), IκB kinases (IKKs), and IKK-related kinases, interferon-regulatory factor 3 (IRF3), and mitogenactivated protein kinases (MAPKs).4,5
* To whom correspondence should be addressed. E-mail: daub@biochem. mpg.de. † Department of Molecular Biology, Max Planck Institute of Biochemistry. § Present address: ITC Research and Development Centre, Peenya Industrial Area, Bangalore, India. ‡ Department of Proteomics and Signal Transduction, Max Planck Institute of Biochemistry. # Present address: Lilly Singapore Centre for Drug Discovery, 8A Biomedical Grove #02-05, Immunos, Biopolis, 138648, Singapore. | Vichem Chemie Ltd. ⊥ Semmelweis University. 10.1021/pr901192p
2010 American Chemical Society
LPS-mediated macrophage activation must be tightly regulated per se, not only to respond appropriately to the bacterial invasion, but also to prevent excessive activation of TLR4mediated signal transduction pathways. Aberrant TLR4 signaling is associated with conditions such as septic shock, inflamJournal of Proteome Research 2010, 9, 2539–2549 2539 Published on Web 03/12/2010
research articles matory diseases, and cancer. Thus, the identification of new components in TLR4-mediated signaling is of high importance as this might define new molecular targets in acute and chronic inflammatory diseases.6 Mass spectrometry (MS)-based proteomics enables unbiased approaches to identify new components in LPS signaling. Recently, a quantitative proteomics analysis by Dhungana et al. revealed that LPS triggers the redistribution of components of the ubiquitin-proteasome system to macrophages lipid rafts where local proteasome activation then triggers the extracellular signal-regulated kinase (ERK) MAPK pathway.7 Generally, members of the protein kinase superfamily fulfill key roles in phosphorylation-mediated signal transmission upon TLR activation, and represent a major class of proteins tractable by small molecule intervention. As the phosphorylation states of protein kinases can provide a read-out for their signaling activities within the cellular system, proteome-wide monitoring of LPS-triggered phosphorylation events on protein kinases can guide the identification of new potential targets for anti-inflammatory therapy. However, protein kinases are often expressed at relatively low levels compared to their numerous cellular substrate proteins. To address the need for comprehensive and sensitive kinase analysis, we have recently developed a phosphoproteomics strategy that combines selective kinase enrichment by affinity chromatography with stable isotope labeling by amino acidsincellculture(SILAC)-basedquantitativemassspectrometry.8,9 Efficient kinase prefractionation was enabled by combinations of immobilized small molecule inhibitors, which collectively captured a major part of the expressed protein kinase complement (kinome). In our present study, we adapted this chemical proteomics approach to enable the enrichment of kinases along with their cellular interaction partners. Moreover, we performed the first time-resolved analysis of kinase-proximal signaling upon LPS challenge, and used bioinformatics analysis to place new components in the inflammatory immune response into the functional cellular context.
Material and Methods Cell Culture. For quantitative kinome analysis based on SILAC, RAW264.7 macrophages were grown in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% dialyzed fetal bovine serum (Invitrogen), 1% (10 mg/mL) streptomycin/(10 000 U/mL) penicillin (PAA Laboratories), 1% (200 mM) L-glutamine (PAA Laboratories), 1% (100 mM) sodium pyruvate (Invitrogen), and either unlabeled L-arginine (Arg0) at 42 mg/L and L-lysine (Lys0) at 71 mg/L or equimolar amounts of the isotopic variants L-[U-13C6,14N4]arginine and L-[2H4]lysine (Arg6, Lys4), or L-[U-13C6,15N4]arginine and L-[U13 C6,15N2]lysine (Arg10, Lys8) (Cambridge Isotope Laboratories or Sigma).10 After five cell doublings on culture dishes, macrophages were treated with 10 ng/mL LPS (Sigma) for 0, 10, and 30 min. A second, identically labeled set of RAW264.7 macrophages was treated with LPS for 3, 10, and 90 min prior to cell lysis. For each labeling and stimulation condition, RAW264.7 macrophages were grown in 20 larges dishes (15 cm diameter) to a final cell number of about 5 × 107 per dish yielding 5 mg of isotope-encoded protein. Immunoblotting of lysate from LPS-stimulated RAW264.7 macrophages was performed with rabbit anti-phospho-Akt (Ser473) antibody (Cell Signaling Technology, Inc.). Affinity Enrichment by Immobilized Kinase Inhibitors. Kinase inhibitor resins were generated according to published procedures.8 Cell lysis was performed essentially as described.11 2540
Journal of Proteome Research • Vol. 9, No. 5, 2010
Sharma et al. Briefly, RAW264.7 cells were lysed with 500 µL of buffer per 15 cm dish containing 50 mM HEPES-NaOH, pH 7.5, 150 mM NaCl, 0.5% Triton X-100, 1 mM EDTA, 1 mM EGTA plus additives (10 mM sodium fluoride, 2.5 mM orthovanadate, 50 ng/mL calyculin, 1% phosphatase inhibitor cocktail 1 (Sigma), 1% phosphatase inhibitor cocktail 2 (Sigma), 10 µg/mL aprotinin, 10 µg/mL leupeptin, 1 mM phenylmethylsulfonyl fluoride) for 1 h at 4 °C. Total cell extracts were then sonicated (3 × 5 s) and cleared by centrifugation for 30 min at 20 000 rpm. Supernatants were again sonicated (2 × 5 s), centrifuged for 10 min at 20 000 rpm, and then passed through a 0.45 µM PVDF filter membrane. In case subsequent affinity purification was performed in high salt conditions, lysate was adjusted to a final concentration of 1 M NaCl. Equal protein amounts of cell extracts (100 mg each) were pooled and then adjusted to a total volume of 40 mL prior to further sample processing. For each triple labeling SILAC experiment, 300 mg of mixed RAW264.7 cell lysate (with a final concentration of either 150 mM or 1 M NaCl) was subjected to in vitro association with the kinase inhibitor resin mix for 2.5 h at 4 °C in the dark. The resin mix comprised 0.5 mL of the VI16832 and purvalanol B, along with 0.33 mL of SU6668, AX14596, and bisindolylmaleimide X matrices.8,12 After three washing steps, the resin-bound material was eluted with buffer containing 20 mM Tris-HCl, pH 7.5, 5 mM DTT, and 0.5% SDS solution by consecutive 10 min incubations at 50 °C. The protein-containing fractions were identified by silver staining, lyophilized, and then resuspended with water in one-fifth of the initial volume followed by a final protein precipitation step according to the method from Wessel and Flu ¨ gge.13,14 Mass Spectrometry Analysis. Sample preparation and MS analysis were performed as described earlier.8,15 Briefly, twothirds of the inhibitor resin-purified proteins were loaded on six lanes of ready-made gels (NuPage 4-12% Bis-Tris, Invitrogen) and resolved by gel electrophoresis. The gels were then cut into slices, followed by in-gel digestion with trypsin and either titansphere enrichment of phosphopeptides or peptide extraction with StageTips.15,16 In parallel, the remaining onethird of the kinase-enriched sample was subjected to insolution digestion with trypsin followed by three consecutive TiO2 enrichments for sequential phosphopeptide extractions.17 All MS analyses were done with a nanoflow HPLC system (Agilent Technologies 1100, Waldbronn, Germany) coupled online to a LTQ-Orbitrap (Thermo Fisher Scientific, Bremen, Germany) mass spectrometer equipped with a nanoelectrospray ion source (Proxeon Biosystems, Odense, Denmark) essentially as described.15 Tryptic peptide and phosphopeptide mixtures were separated in a 15 cm analytical column (75 µm inner diameter) packed with 3 µm C18 beads (Reprosil-AQ Pur, Dr. Maisch) with a 2 h gradient from 5% to 40% acetonitrile in 0.5% acetic acid. The LTQ-Orbitrap was operated in datadependent mode to automatically switch between full scan MS and MS/MS acquisition. Survey MS spectra (from m/z 300-2000) were acquired in the orbitrap mass analyzer with resolution of 60 000 at m/z 400. The five most intense peptide ions with charge states g2 were sequentially isolated to a target value of 5000 and fragmented in the linear ion trap by multistage activation.18 All fragment ion spectra were acquired in the LTQ part of the instrument. Multistage activation (MSA) was enabled to activate phosphopeptide-derived neutral loss species at 97.97, 48.99, or 32.66 m/z below the precursor ion for 30 ms during fragmentation (pseudo-MS3).19 For all measurements with the orbitrap detector, a lock-mass ion from ambient air
Quantitative Analysis of Kinase-Proximal Signaling 18
(m/z 429.08875) was used for internal calibration. Mass spectrometric conditions were spray voltage, 2.4 kV; no sheath and auxiliary gas flow; heated capillary temperature, 150 °C; normalized collision energy 35% for MSA in LTQ. The ion selection threshold was 500 counts for MS2. An activation q ) 0.25 and activation time of 30 ms were used. Assigning Peptide Sequences Using MASCOT and MaxQuant. Raw full-scan MS and ion trap MS2/pseudo-MS3 spectra data were collectively analyzed with the MaxQuant software package (version 1.0.11.1 for all data analyses and version 1.0.13.12 for spectra extraction) as previously described.20,21 Briefly, peak lists were generated, and identified SILAC triplets were quantified and searched against the mouse IPI protein database version 3.37 (51 292 entries). Peptides and proteins were identified using the MASCOT search engine (version 2.2.04, Matrix Science, London, U.K.) from all tandem mass spectra searched against a target/decoy database with concatenated forward and reversed version of the murine IPI protein sequence database supplemented with 175 frequently observed contaminants. Cysteine carbamidomethylation was set as a fixed modification, and N-acetyl protein, N-pyroglutamine, oxidized methionine, and phosphorylation of serine, threonine, and tyrosine as variable modifications. Depending on prior knowledge about the parental ions determined by MaxQuant presearch, Arg10, Arg6, Lys8 and Lys4 were used as additional fixed or variable modifications. For all detected SILAC triplets, MaxQuant integrates repeated mass measurements and further corrects for linear and nonlinear mass offsets as described in detail by Mann & Cox, thereby considerably improving mass accuracy for SILAC precursor ions prior to MASCOT searches.21 The maximum initial mass tolerance for precursor ions in MS scans was 7 ppm and 0.5 Da for fragment ions in MS/MS scans. The minimum peptide length was set to 6 amino acids and up to three missed cleavages and three isotopically labeled amino acids were permitted per peptide. Data Analysis. The resulting Mascot dat output files together with the raw data files were loaded into the MaxQuant software for further processing. The accepted false discovery rate (FDR) was 1% for proteins and peptides. Peptide FDR was applied by successively including best scoring peptide hits until the list contained 1% reverse hits. Posterior error probabilities (PEP) of peptides were multiplied for each protein and proteins were then sorted according to these products until a 1% FDR was reached. In addition to the FDR threshold applied for protein identification, proteins were only included when they were identified with at least one unique peptide for the protein and quantified if at least one quantifiable SILAC pair was associated with them. Outliers were not removed as protein ratios were calculated as the median instead of the average of all peptide ratios. For peptides shared among different identified proteins, SILAC ratios were only considered for the ratio of the protein identified with the highest number of unique peptides. Phosphosites were made nonredundant with regards to their surrounding peptide sequence. Finally, to pinpoint the actual phosphorylated amino acid residue within the identified phosphopeptide sequences in an unbiased manner, the localization probabilities were calculated for all serine, threonine, and tyrosine phosphorylation sites using the PTM score algorithm as described.15 Further, the phosphosites were termed as class I sites when they had a localization probability at least 0.75 and a difference of probability localization score higher or equal to 5 observed in at least one set of SILAC experiments. The phosphopeptide ratios were calculated as median ratios if several quan-
research articles tifications for a phosphopeptide with the same number of phosphorylation events were available. To combine the full time courses of class I phosphosites and phosphopeptides generated from the experiments performed under different salt conditions, we used the phosphorylation ratios that had not been normalized for protein abundance and took average ratios in case quantification was performed in both experiments. Clustering of Regulated Phosphopeptides and Phosphosites. The data from regulated phosphopeptides and phosphosites were combined before clustering. The raw ratios for the time profiles of this collated data set were log10 transformed and then normalized so that, for each profile, the mean was zero and standard deviation was one. The normalization of data ensures that phosphorylation events with similar temporal patterns are close in Euclidean space. The transformed profiles were then clustered using the Mfuzz package in the statistical platform R.22 We used the fuzzy c-means (FCM) clustering algorithm, which is part of the package. For FCM, the number of clusters (c) and fuzzification parameter (m) was set to 5 and 2, respectively. Proteomic Phenotyping by Hierarchical Clustering Based on Gene Ontology (GO) Enrichment Analysis. We employed the approach of ‘proteomic phenotyping’ to analyze the proteins identified in this work.23 Briefly, proteins identified in experiments performed under different salt conditions were divided into three classes as follows: (a) proteins identified in high and low salt condition (HS + LS), (b) proteins identified in LS condition (LS), and (c) proteins identified in high salt condition (HS). The enrichment analysis for gene ontology (GO) molecular function (MF) was done separately for these classes with respect to the whole mouse GO annotations for IPI version 3.37 using “hypergeometric test” with multiple hypothesis correction by “Benjamini & Hochberg False Discovery Rate”.24 For hierarchical clustering, we first collated all the categories obtained after enrichment along with their p-values, and then filtered for those categories which were at least enriched in one of the classes with p-value
