High Field Asymmetric Waveform Ion Mobility Spectrometry (FAIMS) expands the comprehensiveness and precision of mulplex proteomics 1 Ins�tute for Research in Immunology and Cancer (IRIC); 2 Department of Chemistry, Uni ver sité de Montr éal, Montréal, QC; 3 ThermoFi sher Scien�fic, San Jose, CA Sibylle Pfamma�er 1,2 ; Eric Bonneil 1 ; Francis P. McManus 1 ; Satendra Prasad 3 ; Derek J. Bailey 3 ; Michael Belford 3 ; Jean-Jacques Dunyach 3 ; Pierre Thibault 1,2 O VERVIEW AIM:Improved coverage and quantification in proteomic analysis using differential ion mobility. METHODS: A novel high field asymmetric waveform ion mobility (FAIMS) interface was coupled to an Orbitrap Fusion Tribrid (Thermo Fisher Scientific). Isobaric labeling of peptides was used to profile dynamic changes in protein abundance upon heat shock. RESULTS: LC-FAIMS-MS2 provided 2.5-fold higher number of quantifiable peptides compared to the SPS-MS3 strategy for TMT labeling with comparable fold changes. measurements. I NTRODUCTION A KNOWLEDGMENTS 1. Pfammatter et al., J. Proteome Res., 2016, 15 (12), pp 4653–4665 2. Keshishian et al., Nat. Protoc., 2017, 12, 1683-1701 3. Bonneil et al., J. Mass Spectrometry., 2015, 50 (11), pp 1181-1195 METHOD R ESULTS R ESULTS SF3A1 SRRM2 CDC5L SNRPG PRPF6 CRNKL1 SF3B2 PRPF31 LSM2 EIF3J EIF3L TPM1 THOC1 DIS3 EIF3G ACTN1 THOC6 EIF3B DDX27 PTCD3 GNL3 RPS17L WDR33 ZC3H18 PKN2 HNRNPC HNRNPA0 CPSF1 YTHDC2 DDX6 PNN RBMX USP10 C14orf166 DDX1 CHD1 SRP14 PAF1 C22orf28 TIMM10 CTR9 ALDH18A1 SRP9 SDE2 LUC7L3 CYP51A1 THRAP3 SUPT16H PFAS PRPF3 PRPF19 NHP2L1 DHX38 SNRPB2 DDX23 NAA38 SLIRP DDX46 DDX3X TIMM13 TIMM8B U2AF2 PPP1R8 GMPS MFAP1 SMN1 EIF1B COPG1 CLINT1 DNAJA1 LMAN1 CCT2 DNAJB4 HSPE1 HUWE1 SSR1 RANBP2 PHB2 NUCKS1 DNAJB1 NDUFA4 COX4I1 COX5A VDAC3 NDUFA2 MYBBP1A UTP6 DHX37 NOP14 NAP1L1 WDR75 WDR43 H1FX UTP20 NOL6 CDK1 HSPA6 STUB1 SUPT6H CDK4 SUGT1 LMAN2 SDHB RCC1 SCP2 UQCRC1 NDUFA7 GTF2I THOC2 MATR3 NDUFA5 NDUFS3 UNC45A HSPA1A COPA HSPA8 AHSA1 HSP90AA1 RUVBL1 SYAP1 PRKDC XRCC6 NTPCR XRCC4 SDHA TERF2IP XRCC5 ARPC1B ACTR3 CPSF6 PDS5A WAPAL ASUN BCL7A ARHGAP17 ARPC5 SMARCA5 SMARCA1 SMARCC2 BAZ1B SMARCC1 POLE3 SMARCB1 PDCD10 HEATR2 PPP2R1A INTS3 SLC25A3 ANKHD1 PPP4R2 AFG3L2 TMCO1 PPP4C PPP2R5D CAPZB PLS3 CFL1 WDR1 TXN TYMS DCTN4 CAPZA2 PSMB3 UCHL5 UBFD1 PSMB1 PSMA4 G3BP1 PSMA7 ADRM1 PSMB4 NVL NCL RRP12 POLA2 UBA1 TSR1 PDCD11 UBE2N BMS1 RPL19 RPS18 RPL23A RPL6 RPL38 MRPS31 RPS16 MRPL46 RSL1D1 TXLNA MRPL37 MRPL38 MRPL39 RPL7 MRPS7 MRPL12 RPL31 RPL4 RPL13A RPL24 RPS15A RPL35 RPL13 RPL18A DAP3 RPS19 RPL5 EEF2 SF1 RPL7A BOP1 RRP1B RPS28 RPL35A RPL34 RPS13 RPS15 RPL17 RPL14 RPL3 RPS26 SRSF11 TRA2A DBR1 AARS IK MSH2 TSR3 MSH6 YBX1 SRSF7 SRSF3 SLC3A2 GTF3C5 TRA2B GTF3C1 SMU1 MRPL55 RTFDC1 LARS IARS2 EPRS GARS MRPS30 GMNN CCNA2 HSPH1 COX6B1 ATP5C1 RPA3 SRRT MCM3 FASN RFC4 MCM2 Ribosome p-value 1.9e-26 Proteasome complex p-value 4.98e-5 Response to unfolded protein p-value 1.01e-4 Mitochondrial respiratory chain p-value 9.27e-4 Spliceosomal complex p-value 7.64e-14 Chromatin remodeling p-value 2.76e-4 Mitochondrial Ribosomes Cytoplasmic Ribosome Time (h) 9h 1h EARLY Response LATE Response Down Regulated Up Regulated 43°C a b Chromatin Remodelling Mitochondrial Proteins Cell-Cell adhesion Cell-Cell adhesion Mitochondrial Proteins Proteasome Complex HSPs aggregation missfolding Response to unfolded proteins Ribosome Spliceosome Exon 1 Intron Exon 2 Exon 1 Exon 2 Spliceosome Exon 1 Intron Exon 2 Exon 1 Exon 2 Spliceosome/Spliceosome Spliceosome/Spliceosome SPS-MS3 FAIMS-MS2 # MS 20437 22450 # MS2 # MS3 105820 105797 124005 # identified PSM (peptide sequences) 27146 (13642) 49968 (38374) # identfied protein groups (2 uniques peptides) 1260 2673 # quantified PSM (peptide sequences) 25142 (12400) 44490 (30848) # quantified protein groups (2 uniques peptides) 1229 2646 # dynamic proteins 375 902 b a # unique peptides identified # Injections 9650 22.8% 28724 67.8% 3992 9.4% without FAIMS with FAIMS 0 5000 10000 15000 20000 25000 30000 35000 40000 45000 1 2 3 1195 43.6% 1478 54.0% 65 2.4% without FAIMS with FAIMS peptides protein groups * quantified = present in min 7 TMT channels # cluster proteins 149 502 −2 −1 0 1 2 n=54 - - - without FAIMS with FAIMS Protein groups (cluster membreship >0.9) Cluster 4 - - - -1.5 -1.0 -0.5 0.0 0.5 1.0 1.5 - log 2 FC ( HS / CTR ) 0123456789 Time( h) 0 100 200 300 400 500 Cluster 3 Cluster 2 Cluster 1 Cluster 4 Cluster 3 Cluster 2 Cluster 1 0123456789 Time( h) 0 20 40 60 120 140 80 100 Protein groups (cluster membreship >0.9) 0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 HSPA1A Cluster 1 -0.5 -0.4 -0.3 -0.2 -0.1 0 0.1 LRPPRC Cluster 3 log2 FC(HS/CTR) log2 FC(HS/CTR) SPS-MS3 FAIMS-MS2 SPS-MS3 FAIMS-MS2 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 without FAIMS Cluster 1 Late Up Cluster 2 Early up Cluster 3 Late Down Cluster 4 Early down Time HS (h) log 2 Fold Change (order n=2) −2 −1 0 1 2 −2 −1 0 1 2 −2 −1 0 1 2 −2 −1 0 1 2 −2 −1 0 1 2 −2 −1 0 1 2 n=116 n=153 n=56 n=177 n=41 −2 −1 0 1 2 n=24 n=30 relative Fold Change 0.0 0.2 0.4 0.6 0.1 0.2 0.3 −0.4 −0.3 −0.2 −0.1 0.0 −0.5 −0.4 −0.3 −0.2 −0.1 0.0 0.1 0.2 0.3 0.4 0.1 0.2 −0.3 −0.2 −0.1 0.0 −0.4 −0.3 −0.2 −0.1 Time HS (h) 1 2 3 4 5 6 7 8 9 1 2 3 4 5 6 7 8 9 with FAIMS c d log2 FC(SPS-MS3) log2 FC(FAIMS-MS2) −2 −1 0 1 2 3 −2 −1 0 1 2 3 8h HS protein groups e f Fig.1: Experimental workflow. (a) All experiments were performed using LC-MS/MS on a Orbitrap tribrid Fusion mass spectrometer with a two hours LC gradient, 500ng/injection with 3s cycles. For SPS-MS3, 10 notches were selected 2 . Without FAIMS we did 3 replicates, with FAIMS we combined CVs together to cover in 3 injections the whole transmission range. (b) Optimised FAIMS method shows little overlap between the different injections and comparable number of identification. (c) Yeast and HEK293 cell extracts were reduced, alkylated and digested with trypsin prior to labeling with TMT 6-plex. (d) HEK293 cells were incubated at 43°C, and collected at 1h intervals up to 9 h prior to digestion and TMT 10-plex labeling. d TMT labeling TMT 10 126 TMT 10 127N TMT 10 128N TMT 10 129N TMT 10 130N TMT 10 131 Protein extraction Tryptic digestion 0h 1h 2h ... 8h 9h Human Heat Shock 43°C TMT 10 127C TMT 10 128C TMT 10 129C TMT 10 130C 1:1:1:1:1:1:1:1:1:1 LC-MS/MS/MS (SPS) LC-FAIMS-MS/MS 3 x 2h MS Inlet ESI MS Inlet CV ESI IE 100°C - OE 100°C a MS-MS/MS MS-MS/MS MS-MS/MS MS-MS/MS CV 1 CV Transmission range: -37V to -93V MS-MS/MS CV 2 MS-MS/MS CV 3 MS-MS/MS CV 4 MS-MS/MS CV 5 MS-MS/MS CV 6 MS-MS/MS CV 7 MS-MS/MS CV 8 MS-MS/MS CV 9 -37V /-44V /-51V -58V /-65V -73V /-80V /-87V /-93 V Rep1 Rep3 Rep2 Inj #1 Inj #3 Inj #2 without FAIMS with FAIMS -58V/-65V -73V/-80V/-87V/-93V -37V/-44V/-51V 1% 26% 4% 29% 4% 7% 29% b Yeast TMT labeling Yeast 4 : 6 : 4 : 0 : 0 : 0 Human 0 : 2.5 : 4 : 10 : 4 : 2.5 Human Tryptic digestion WITHOUT Interferences WITH Interferences Interferences 1 : 1.5 : 1 1 : 1.6 : 4 : 1.6 : 1 TMT 6 126 TMT 6 131 TMT 6 127 TMT 6 128 TMT 6 129 TMT 6 130 c Isobaric labeling of peptides provides a convenient approach to enhance the throughput of quantitative proteomic measurements, and can be achieved using different reagents including Tandem Mass Tag (TMT) labeling. However, the fragmentation of co-eluting isobaric ions can lead to chimeric MS/MS spectra and distorted reporter ion ratios. Synchronous precursor selection (SPS)-based MS3 method can alleviate this problem, though this approach can result in a reduced number of quantifiable proteins compared to traditional MS2 method. Here, we compared the analytical merits of SPS-MS3 to that of LC-MS/MS that combines a new high field asymmetric waveform ion mobility spectrometry (FAIMS) interface. LC-MS/MS experiments performed using FAIMS enhanced peak capacity and sensitivity while reducing peptide co-fragmentation thus extending the coverage of multiplex proteomic measurements 1 . • The New FAIMS interface provides significant advantages in terms of instrument speed compared to the old generation FAIMS, allowing in three injections to cover the CV transmission range with optimized CV distribution (Fig.1a and Fig.1b). • Using a known two-proteome model of Yeast - Human peptides (Fig.1b), FAIMS MS2 improves peak capacity 3 and reduces the occurence of co-fragmentation of isobaric precursors with precision comparable to the SPS MS3 approach (Fig.2). • The dynamic changes of HEK293 cells exposed to hyperthermia was analyzed with total 1.5ug of peptides. FAIMS increases the number of quantifiable TMT-labeled peptides by 3-fold and proteins by 2-fold compare to the SPS MS3-based strategy for LC-MS/MS analyses (Fig.3b). • Proteins showing changes in abundance upon heat shock were separated in 4 groups: early up- or down-regulation and late up- or down-regulation and showed similar trends for FAIMS and SPS (Fig.3c-d). • Hyperthermia affects several key cellular processes that impact protein homeostasis, such as responses to unfolded proteins (Fig.4). b * quantified = present in min 2 TMT channels 2 instrumental replicates for SPS and FAIMS Human126 Human127 Human128 Human129 Human130 Yeast127 Yeast128 Yeast129 Yeast130 Yeast131 0 2 4 6 0 2 4 6 0 2 4 6 0 2 4 6 0 2 4 6 0 1000 2000 3000 0 1000 2000 3000 0.22 1.22 1.62 4.12 1.54 1 1 1.57 1.13 0.53 0.18 0.10 0 2 4 6 0.08 1.12 1.65 4.32 1.60 1 1 1.55 1.06 0.24 0.06 0.03 TMT 6 126 0 2 4 6 SPS-MS3 FAIMS-MS2 TMT 6 127 TMT 6 128 TMT 6 129 TMT 6 130 TMT 6 131 TMT 6 126 TMT 6 127 TMT 6 128 TMT 6 129 TMT 6 130 TMT 6 131 Interferences Human Yeast Interferences 0 2 4 6 0 2 4 6 0 2 4 6 0 2 4 6 0 2 4 6 Interferences Interferences ratio to TMT 6 131 ratio to TMT 6 126 # unique peptides # unique peptides ratio to TMT 6 131 ratio to TMT 6 126 c d Yeast Yeast 4 : 6 : 4 : 0 : 0 : 0 Human 0 : 2.5: 4 : 10 : 4 : 2.5 Human 1 : 1.5 : 1 1 : 1.6 : 4 : 1.6 : 1 500ng/injection a 3 x 2h RP LC SPS-MS3 FAIMS-MS2 # identified PSM • peptide sequences 28232 • 9354 79302 • 40631 # identified protein (# uniques peptides) 1120 (>2); 301 (=2); 485 (=1) 3987 (>2); 798 (=2); 1226 (=1) # quantified PSM • peptide sequences 26158 • 8685 77373 • 39636 # quantified protein groups 1062 3424 # quantified unique peptide sequences 6074 28088 • Yeast • 2240 • 9444 • Human • 3834 • 16844 Fig.2: Precision of TMT quantification (a) To determine the extent of co-fragmentation using FAIMS, we labeled separate aliquots of yeast and HEK293 tryptic digests with TMT reagents, and mixed those aliquots to obtain final TMT ratios of 1:1.5:1:0:0:0 for yeast and 0:1:1.6:4:1.6:1 for human extracts. Human peptides were not labeled with TMT-126, whereas channels TMT-129 to TMT-131 were not used for yeast peptides. This labeling scheme facilitated the identification of co-fragmentation arising from interfering peptides of each species. 70% of all unique sequences in SPS-MS3 were also quantified in FAIMS-MS2. (b) Summary table comparing MS analysis between SPS-MS3 (middle column) and FAIMS-MS2 (right column) for identified and quantified features. (c) Distortion of TMT ion ratios and extent of ion contamination for the two-proteome model analysed with SPS-MS3 (grey) and FAIMS-MS2 (green). Box plot and (d) frequency distribution of TMT reporter ion ratios normalized using TMT-126 and TMT-131 for for unique yeast and human peptide sequences, respectively. Fig.3: FAIMS improves TMT quantification of the human proteome. HEK293 cells were exposed to a 43˚C heat stress for up to 9 h in 1h increments. (a) Cumulative number of unique peptides identified as a function of repeat injections for SPS-MS3 (black) or CV stepping program with FAIMS-MS2 (green). Overlap in peptide and protein identifications between the two methods are depicted in Venn diagrams to the right of the curves.(b) Summary table comparing MS analysis parameters between SPS-MS3 (middle column) and FAIMS-MS2 (right column). (c) Dynamic clusters for heat shock regulated proteins without FAIMS (left) and with FAIMS (right). The grey lines show the relative fold changes of the individual proteins with high membership (≥0.9), the blue lines the average fold changes of all the proteins in the corresponding cluster. (d) Corresponding heat map for all proteins in the clusters from (c). (e) Representative dynamic profile of HSPA1A (assigned to a late up regulation) and LRPPRC (assigned to a late up downregulation) for SPS-MS3 (grey) and FAIMS-MS2 analysis (green), highlighting virtually identical profiles/quantifications for both acquisition methods. (f) Scatterplot representations for the common dynamic (c) proteins at time point 8h. C ONCLUSION Fig.4: Heat stress affects several key cellular processes that impact protein homeostasis. (a) Interaction network for upregulated proteins (green) and down regulated proteins (red) based on the clustering shown in Figure 3c. Proteins that belong to enriched GO-terms are outlined by coloured shapes. (b) Cellular processes affected in early and late response to heat shock. • A compact FAIMS interface combined with an Orbitrap tribrid mass spectrometer improves accuracy and coverage in multiplex quantification. • TMT-based MS2 quantitation made possible with FAIMS enabled a 2-3 fold gain in identification with high TMT ratio accuracy compared to SPS-MS3 quantitation. R EFERENCES 4260 23828 1814 SPS-MS3 with FAIMS-MS2 unique peptides This work was carried out with financial support from the Natural Sciences and Engineering Research Council and the Genomic Applications Partnership Program (GAPP) of Genome Canada. IRIC receives infrastructure support from IRICoR, the Canadian Foundation for Innovation, and the Fonds de Recherche du Québec - Santé (FRQS). IRIC proteomics facility is a Genomics Technology platform funded in part by the Canadian Government through Genome Canada.