Report CopyRight/DMCA Form For : Proteome Wide Drug And Metabolite Interaction Mapping By
the feasibility of such a methodology for drug discovery However despite its pioneering nature that study also revealed some limita tions as the BCR ABL inhibitor dasatinib surprisingly did not alter the cognate target BCR ABL in K562 cells The report also did not address the possibility of revealing target proteins other than kinases
brief communications,Figure 1 Schematic representation of the N NH2. thermal profiling methodology a Binding of a N,ligand to its target protein top increases the Cl. enthalpy required for unfolding As a result the Seed and. treat cells,melting temperature Tm is shifted bottom. This can be exploited with for example,differential scanning fluorimetry Tm shift. Natural state Stabilized, assay A U arbitrary units b Representative Wash harvest. workflow Cells are treated with either small and heat cell. molecule or vehicle washed and harvested,Fluorescence A U. T1 T2 T10 T1 T2 T10,Equal amounts of cells are then heated to 4. increasing temperatures and lysed and the Tm,supernatant obtained after centrifugation is 2. digested and labeled for quantitative mass Lyse cells and. spectrometry MS analysis Detection of centrifuge, changes in protein stability in treated versus 40 60 80 100 T1 T2 T10 T1 T2 T10. control samples is enabled by bioinformatic Temperature C. processing and data plotting in the form of, differential protein abundance graphs c The Digest. LYTLVLVLQPQR LYTLVLVLQPQR,supernatant, bioinformatics scoring parameters used for data Tm. Relative abundance,LYTLVLVLQPQR LYTLVLVLQPQR, analysis Continuous lines depict raw data and 100 peptides LYTLVLVLQPQR. dashed lines are curve fitting results,2015 Nature America Inc All rights reserved. Relative abundance,for other proteins linked to DNA repair DMSO. 0 Compound,such as POLA Supplementary Fig 2 and MS and. Temperature bioinformatic,Supplementary Table 1 analysis. To validate the thermal stabilization of,Temperature. MTH1 by S crizotinib as detected by, mass spectrometry under these experimental conditions we per Our results prompted us to further investigate the possibility. formed an analogous experiment using SW480 colon carcinoma of using the methodology to elucidate the molecular interaction. cells and analyzed the supernatants by immunoblotting Indeed partners of cellular metabolites which would make it an option. S crizotinib protected MTH1 from unfolding whereas DMSO for applications beyond drug target deconvolution We exam. treated control cells showed a sharp decrease in MTH1 content ined whether the approach could recover the cellular targets of. Fig 2b and Supplementary Fig 3 the recently reported cGAMP synthase metabolite 2 3 cGAMP14. To investigate potential differences between intact cells and Binding of 2 3 cGAMP to its receptor STING stimulator of. lysates we performed an analogous experiment using SW480 cell interferon genes also known as TMEM173 elicits the transcrip. lysates Supplementary Table 2 As reported for other compounds tion of proinflammatory cytokines To ensure sufficient engage. tested with the immunoblotting method8 we applied S crizotinib ment of STING by its endogenous ligand we transfected RAW. at a higher final concentration 100 M to avoid potential dilution macrophages with 2 3 cGAMP and analyzed the samples by. effects in the lysate In contrast to the experiment with intact cells mass spectrometry Bioinformatic filtering showed a reproducible. the analysis revealed not only MTH1 but also several kinases as tar shift for STING as well as for well known downstream media. gets many of which had been highlighted by the previous in vitro tors of innate immune signaling such as NF B and STATs sig. kinase screen Supplementary Figs 4 and 5 and Supplementary nal transducers and activators of transcription 15 Fig 2e and. Table 3 This supports the notion that cellular compartmentaliza Supplementary Table 5 We also confirmed a stabilizing effect. tion may affect a compound s target profile and that thermal profil on STING by immunoblotting Fig 2f and Supplementary. ing experiments using intact cells may be more likely to reflect the Fig 7 demonstrating not only that the approach can recapitu. actual physiological context than experiments with lysates late the activation of the inflammatory response by the cGAMP. Next we evaluated the cellular target profile of the antimetabolite synthase metabolite but also that the procedure is compatible with. methotrexate MTX in K562 cells Our analysis identified the cog transmembrane proteins. nate and well studied target dihydrofolate reductase DHFR as the Our study demonstrates that this type of approach is applicable. top candidate Fig 2c and Supplementary Table 4 We confirmed to enzymes other than kinases as was shown by Savitski et al 9. the stabilization by immunoblotting Fig 2d and Supplementary such as hydrolases and oxidoreductases as well as to the detec. Fig 6 We also detected a highly reproducible shift for thymidylate tion of transmembrane protein targets Moreover our procedure. synthase TYMS an important mediator of MTX action MTX differed from that of Savitski et al 9 in sample processing mass. exhibits inhibitory activity toward TYMS only at concentrations spectrometric analysis and computational and statistical inter. above 10 M whereas it has been shown that MTX polyglutamate pretation of the data Supplementary Note 1 Our approach. metabolites are active at concentrations in the nanomolar range13 may be further tailored to specific target classes subcellular. As our treatment conditions were compatible with the formation of compartments or tissues. MTX polyglutamate species the data indicate that our method in The choice of temperature steps may be critical for captur. contrast to classical target identification approaches could enable ing differences in individual stabilities in particular groups of. tracking of both drug and drug metabolite effects proteins Therefore several experiments with different temperature. ADVANCE ONLINE PUBLICATION nature methods,brief communications. Figure 2 Thermal profiling results for the MTH1 a 150 MTH1 versus NUDT5 intact cells b SW480. Intact cells, inhibitor S crizotinib MTX and 2 3 cGAMP 47 0 C 61 4 C. Relative abundance,a Protein abundance graph derived from mass. spectrometry TMT 10 plex experiments showing 100,data for the cognate S crizotinib target MTH1. S crizotinib 5 M MTH1,and the Nudix family member NUDT5 Data are 50. representative of two independent experiments MTH1 S crizotinib. NUDT5 DMSO, n 2 b Western blot confirming the 0 NUDT5 S crizotinib. stabilization of MTH1 by S crizotinib in intact 48 6 51 8 55 0 58 2 61 4. SW480 cells Images are cropped for clarity full Temperature C. length blots are presented in Supplementary, Figure 3 c Protein abundance graphs c 150 DHFR intact cells d K562. Intact cells,derived from mass spectrometry TMT 10 plex. Relative abundance,38 0 C 62 5 C,experiments profiling DHFR as a target of MTX 100. Data are representative of two independent DMSO DHFR. experiments n 2 d Immunoblot confirming,50 MTX 10 M DHFR. the stabilization of DHFR by MTX in intact K562, cells Images are cropped for clarity full length DHFR DMSO. blots are presented in Supplementary 0, Figure 6 e Results of mass spectrometry 40 7 46 1 51 6 57 0 62 5. Temperature C,analysis of RAW macrophages treated with. 2 3 cGAMP to profile the stabilization of e 150 f,2015 Nature America Inc All rights reserved. STING intact cells RAW 264 7, the transmembrane receptor STING Data are Intact cells. Relative abundance, representative of two independent experiments 37 0 C 73 0 C. n 2 In a c and e individual data points,H 2O STING. are marked by curves represent lines of best, fit f Western blot validating the increased 50 2 3 cGAMP 10 M STING. thermal stability observed for STING after H2O,treatment with 2 3 cGAMP Images are cropped 0. for clarity full length blots are presented in 41 0 49 0 57 0 65 0 73 0. Supplementary Figure 7 Temperature C, ranges may be required to identify the unknown targets of less well Acknowledgments. studied compounds Future systematic work will undoubtedly show We are grateful to W Berger Institute of Cancer Research Vienna Austria for. providing SW480 cells and P Majek CeMM Vienna Austria for assistance with. correlations between ligand pocket architecture and many other data processing This work was supported by the Austrian Academy of Sciences. parameters Moreover thorough charting of the general thermal the European Union FP7 259348 ASSET and the Austrian Science Fund FWF. stability of the proteins in a cell should uncover many interesting F4711 MPN. features regarding the thermal stability of cellular proteins as well. AUTHOR CONTRIBUTIONS, as their subcellular localization abundance multiprotein complex K V M H designed experiments and jointly performed them with K M O. formation and post translational modifications Finally there is an A C M and K L B performed mass spectrometry C S H T and J C performed. increasing appreciation of the fundamental role of cellular metabo bioinformatics analysis and K V M H and G S F conceived the study and. lism in human disease and an exact mapping of the functional wrote the manuscript All authors contributed to the discussion of results and. participated in preparation of the manuscript, relationships between metabolites and their protein counterparts. could not only increase our mechanistic understanding but also COMPETING FINANCIAL INTERESTS. reveal new potential drug targets Importantly the approach should The authors declare no competing financial interests. allow for target engagement in human biopsies Combined with. targeted quantitative proteomics approaches the method could Reprints and permissions information is available online at http www nature. com reprints index html, focus on specific biomarkers and thus act in synergy with other. efforts in this area16 1 Lee J Bogyo M Curr Opin Chem Biol 17 118 126 2013. In summary we have presented a methodology that takes advantage 2 Kambe T Correia B E Niphakis M J Cravatt B F J Am Chem Soc. 136 10777 10782 2014, of alterations in protein thermal stability after ligand binding provid 3 Huber K V M et al Nature 508 222 227 2014. ing a straightforward approach for identifying the cellular proteins 4 Feng Y et al Nat Biotechnol 32 1036 1044 2014. that are engaged by a metabolite or a drug in intact living cells 5 Lomenick B et al Proc Natl Acad Sci USA 106 21984 21989 2009. 6 Niesen F H Berglund H Vedadi M Nat Protoc 2 2212 2221 2007. 7 Fedorov O et al Proc Natl Acad Sci USA 104 20523 20528 2007. Methods 8 Martinez Molina D et al Science 341 84 87 2013. Methods and any associated references are available in the online 9 Savitski M M et al Science 346 1255784 2014. version of the paper 10 Gad H et al Nature 508 215 221 2014. 11 Dayon L et al Anal Chem 80 2921 2931 2008, 12 Parlanti E Locatelli G Maga G Dogliotti E Nucleic Acids Res 35. Accession codes The raw mass spectrometric data used in this study 1569 1577 2007. are available via PeptideAtlas with the identifier PASS00693 13 Allegra C J et al J Biol Chem 260 9720 9726 1985. 14 Ablasser A et al Nature 498 380 384 2013, Note Any Supplementary Information and Source Data files are available in the 15 Cai X Chiu Y H Chen Z J Mol Cell 54 289 296 2014. online version of the paper 16 Lambert J P et al Nat Methods 10 1239 1245 2013. nature methods ADVANCE ONLINE PUBLICATION, ONLINE METHODS described18 using an Agilent 1200 series HPLC system Separation. Cell culture K562 and RAW 264 7 cells were obtained from DSMZ was performed at a flow rate of 100 l min 1 on a Phenomenex col. and ATCC respectively SW480 cells were kindly provided by W umn 150 2 0 mm Gemini NX 3 m C18 110 with a 50 min. Berger Institute of Cancer Research All cells were cultured in the linear gradient of 5 70 vol vol acetonitrile containing 20 mM. recommended media containing 10 FBS and 10 U ml 1 penicil ammonium formate Twenty time based fractions were auto. lin streptomycin Gibco and checked for mycoplasma by PCR or matically collected and peptides were lyophilized in a vacuum. ELISA before experimental use centrifuge and reconstituted in 5 vol vol formic acid for on. line liquid chromatography mass spectrometry LC MS The pep. Immunoblotting The following antibodies were used accord tide abundance per fraction was estimated on the basis of the UV. ing to the manufacturer s instructions mouse anti DHFR trace and samples were diluted accordingly to avoid overloading. A 9 sc 377091 Santa Cruz Biotechnology 8 rabbit anti MTH1 the nano LC MS system Fractions were injected onto a Dionex. NB100 109 Novus Biologicals 3 and rabbit anti STING 3337 Ultimate 3000 system Thermo Fisher Scientific coupled to a. Cell Signaling 17 Q Exactive mass spectrometer Thermo Fisher Scientific. Peptides were loaded onto a trap column Zorbax 300SB C18. Sample preparation SW480 cells were cultured in 12 well cell 5 m 5 0 3 mm Agilent Biotechnologies at 45 l min 1 and. culture plates to 80 confluency and treated with media containing separated on a customized 16 cm 50 m inner diameter reversed. either DMSO or 5 M S crizotinib for 3 h K562 cells were cultured phase column packed with 3 m C18 particles ReproSil Pur 120. in T75 cell culture flasks and upon reaching about 80 confluency C18 AQ Dr Maisch Peptides were separated at a constant flow. aliquoted into 12 well cell culture plates and treated with media rate of 100 nl min 1 with an 85 min gradient of 3 35 vol vol. 2015 Nature America Inc All rights reserved, containing either DMSO or 10 M MTX for 3 h After treatment solvent B that was increased to 70 vol vol solvent B within. cells were detached with trypsin SW480 only collected by centrif 12 min and finally to 100 vol vol solvent B for 10 min before re. ugation and resuspended in PBS Cell suspensions were transferred equilibration to 3 solvent B Solvent A consisted of 0 4 vol vol. into 0 2 ml PCR tubes and heated for 3 min After a subsequent formic acid and 70 vol vol methanol and solvent B consisted of. 3 min incubation at room temperature cells were lysed by the addi 20 vol vol isopropanol and 0 4 vol vol formic acid Analysis. tion of 30 l of lysis buffer and three repeated freeze thaw cycles on the Q Exactive was performed in data dependent mode with. using liquid nitrogen Precipitated proteins were separated from the top ten most intense ions selected for fragmentation MS and. the soluble fraction by centrifugation at 17 000g for 20 min at 4 C higher energy collision induced dissociation HCD MS spectra2. For metabolite experiments RAW cells were cultured in 100 mm were acquired at 70 000 and 35 000 resolution at m z 200 respec. culture plates to 80 confluency and transfected with either water tively Automatic gain control was used to prevent overfilling of the. or 2 3 cGAMP 10 M InvivoGen using Lipofectamine 2000 ion trap and was set to 3 106 and 2 105 ions for MS and HCD. After 4 h cells were harvested and treated as described above For respectively Maximum ion accumulation times were set to 250. cell lysate experiments cells were lysed with lysis buffer 50 mM and 120 ms for MS and HCD respectively The dynamic exclusion. Tris HCl 100 mM NaCl 0 2 NP 40 5 glycerol 1 5 mM MgCl2 for selected ions was 60 s The singly charged siloxane mass at m z. 25 mM NaF 1 mM Na3VO4 1 mM phenylmethylsulfonyl fluoride 445 120024 was used for lock mass correction19 The threshold. 1 mM dithiothreitol DTT 10 g mL TLCK 1 g mL leupeptin for switching from MS to tandem mass spectrometry MS MS. 1 g mL aprotinin and 10 g mL soybean trypsin inhibitor was 1 of the target ion value The normalized collision energy. Sigma pH 7 5 and after centrifugation incubated with DMSO was 29 with a fixed first mass of 100 m z Software versions. or S crizotinib 100 M for 30 min 6 mg total protein per con used for the operation of the Q Exactive were Tune 2 3 SP1 and. dition Six aliquots were prepared for each sample and heated for Xcalibur 3 0 63. 3 min in a PCR machine Precipitated proteins were separated from. the soluble fraction by centrifugation as described above Bioinformatic analysis All MS MS spectra were searched against. UniProtKB Swiss Prot human with Mascot20 and Phenyx21 Results. In solution tryptic digestion TMT derivatization and 2D were merged and false discovery rate imposed a maximum of 1. reversed phase liquid chromatography The protein concentra according to a previously described procedure22 Only spectra pro. tion of cell lysates was determined using the Bradford method Bio viding a signal above the estimated background noise23 in all ten. Rad Thirty SW480 fifty K562 and eighty RAW micrograms TMT channels were considered and we further limited the analysis. of total protein were denatured with 8 M urea and reduced with to proteins identified by at least two such spectra. DTT 10 mM and cysteine residues were alkylated with iodoaceta. mide 55 mM and digested with modified porcine trypsin 1 100 Data normalization Classical normalization approaches could. Worthington Biochemical overnight at 37 C after the dilution of not be applied because the total amount of proteins in samples. urea to less than 2 M After quenching with trifluoroacetic acid at increasing temperatures was reduced on purpose Nonetheless. TFA peptides were purified by solid phase extraction using C18 it could be appreciated that un normalized data especially those. SPE columns The Nest Group Eluted peptides were lyophilized in obtained with DMSO treatment showed a departure from the. a vacuum centrifuge and resuspended in TEAB buffer to a final con expected sigmoid curve at the first three temperatures above the. centration of 100 mM before being labeled with the TMT reagents lowest temperature Supplementary Fig 1 This was observed in. Thermo Fisher Scientific Labeling was performed according to the original western blotting procedure8 and we hypothesize that. the instructions provided by the manufacturer After quenching this phenomenon represents an artifact due to protein crowding. all samples were pooled purified by solid phase extraction and effects at the lower temperatures We thus reasoned that data. separated at pH 10 by reversed phase liquid chromatography as could be normalized by their shape instead of their values. nature methods doi 10 1038 nmeth 3590, To this end we summed 50 000 randomly selected spectra in each the analysis to proteins for which the assay generated meaning. sample separately modeling the average temperature dependent ful interpretable data This procedure also eliminated some. abundance decrease which we fitted by a sigmoid of the form oddly shaped melting profiles that generated tentatively decent. scores by coincidence and would be discarded immediately upon. 1 1 eb a t expert inspection, where t is the temperature and a and b are the curve parameters Target selection Exploiting the two replicates available in each. determined by nonlinear least square optimization At each cell line for each compound and vehicle control we created. temperature the summed spectrum was compared with the sig two pairs of vehicle versus compound treated samples. moid and a relative shift was determined that was subsequently Proteins identified in a pair were scored as described above In. applied to all the proteins of the corresponding sample as a means each pair we retained proteins in the top p of both scores and. of normalization Supplementary Fig 1 intersected such selections across the two biological replicates. We used P 10 for MTX P 15 for S crizotinib and P 20. Target scoring Two scores were computed to capture differ for 2 3 cGAMP We calculated the coefficient of variation CV. ent features of the thermal shift that should be observed in tar values reported in Supplementary Tables 1 4 and 5 by applying. gets a temperature shift at 50 reduction of protein abundance the classical definition s d divided by the mean to the two rep. in the sample and a signed area between the DMSO or vehi licates of a given score temperature shift or signed area Namely. cle treated sample curve and the compound treated curve Fig if s1 and s2 are the replicates of one of the two scores CV 2. 1c and Supplementary Fig 8 The sigmoid curves required to s1 s2 s1 s2. 2015 Nature America Inc All rights reserved, estimate the thermal shift at 50 concentration reduction were. of the form above The two scores were necessary because more. than 3 000 proteins had to be scored in each sample comparison 17 Yoh S M et al Cell 161 1293 1305 2015. and odd curves were able to yield good values according to one 18 Gilar M Olivova P Daly A E Gebler J C J Sep Sci 28 1694 1703. of the two scores by chance In addition it was required that a 19 Olsen J V et al Mol Cell Proteomics 4 2010 2021 2005. potential drug target feature at least a 25 protein abundance 20 Perkins D N Pappin D J C Creasy D M Cottrell J S Electrophoresis. reduction at the highest temperature under compound treatment 20 3551 3567 1999. 21 Colinge J Masselot A Giron M Dessingy T Magnin J Proteomics. and at least a 75 reduction under vehicle for MTX these val 3 1454 1463 2003. ues were respectively 25 and 50 for S crizotinib and 10 22 Burkard T R et al BMC Syst Biol 5 17 2011. and 10 for 2 3 cGAMP These nonstringent conditions limited 23 Breitwieser F P et al J Prot Res 10 2758 2766 2011.