New Developments in LC-IMS-MS Proteomic Measurements and Informatic Analyses

Size: px

Start display at page:

Download "New Developments in LC-IMS-MS Proteomic Measurements and Informatic Analyses"

Ralph Parker
5 years ago
Views:

New Developments in LC-IMS-MS Proteomic Measurements and Informatic Analyses Erin Shammel Baker Kristin E. Burnum-Johnson, Xing Zhang, Cameron P.

1 New Developments in LC-IMS-MS Proteomic Measurements and Informatic Analyses Erin Shammel Baker Kristin E. Burnum-Johnson, Xing Zhang, Cameron P. Casey, Yehia M. Ibrahim, Matthew E. Monroe, Tao Liu, Brendan MacLean, Brian Pratt, Michael J. MacCoss and Richard D. Smith Pacific Northwest National Laboratory

Introduction Why are we interested in using IMS-MS? 1.

feature numbers with fast LC separations or no LC separations at all 2.

IMS adds complementary information to MS measurements which helps lower false discovery rates

2 Introduction Why are we interested in using IMS-MS? 1. We would like to increase throughput of measurements and IMS-MS analyses are able to detect high feature numbers with fast LC separations or no LC separations at all 2. IMS-TOF MS provides greater dynamic range of detection relative to trapping (e.g. Orbitrap) instruments 3. IMS adds complementary information to MS measurements which helps lower false discovery rates and separates isomers 4. Detection of structural changes in peptides/ proteins that can help characterize specific disease states (structural biomarkers)

3 Ion mobility concept E F friction Ion F el p (Drift Gas) Drift Cell velocity is constant v = K E K = ion mobility

4 Ion mobility concept E in out Drift Time Pulse of 2 ions with same m/z but different shape Drift Cell Different conformers are separated in time with peak heights representing the amount of each velocity is constant v = K E K = ion mobility

IMS-MS concept ESI Source High Pressure Ion

Features: NanoESI ion source with 2 inlets for

hourglass ion funnel/accumulation trap before

IMS Segmented quadrupole for CID Electric

5 IMS-MS concept ESI Source High Pressure Ion Funnel Trapping IF Drift Cell Rear IF Q00 MS Features: NanoESI ion source with 2 inlets for on-the-fly calibration DRIFT CELL Off-axis hourglass ion funnel/accumulation trap before IMS Sugars Peptides DNA Rear ion funnel after IMS Segmented quadrupole for CID Electric Field High dynamic range Agilent TOF or Q-TOF MS Drift Time

30 40 50 60 Drift Time (ms) 100 20 30 40 50 60 Drift Time (ms) 100 20

6 Intensity Ion mobility concept LC (minutes) IMS (~60 ms) MS (~100 µs) IMS MS Elution Time Drift Time m/z 1100 m/z 1100 m/z 1100 m/z Drift Time (ms) Drift Time (ms) Drift Time (ms) Elution Time (minutes)

7 8 peptides spiked in human serum Spiking Level Non-Serum Peptide 60-min LC- IMS-TOF MS 60-min LC- TOF MS 100-min LC- Velos-Orbitrap 100 pg/ml Melittin ND ND ND 100 pg/ml Dynorphin A Porcine Fragment 1-13 ND ND 1 ng/ml Des Pro Ala Bradykinin ND ND 1 ng/ml Leucine Enkephalin ND ND 10 ng/ml 3X FLAG Peptide ND 10 ng/ml Substance P 100 ng/ml [Ala92]-Peptide ng/ml Methionine Enkephalin Sample analyzed using Velos-Orbitrap, TOF MS and IMS-TOF MS instruments

8 8 peptides spiked in human serum Scaled abundance and (CV values) Spiking Level Non-Serum Peptide 60-min LC- IMS-TOF MS 60-min LC- TOF MS 100-min LC- Velos-Orbitrap 100 pg/ml Melittin ND ND ND 100 pg/ml Dynorphin A Porcine Fragment (18) ND ND 1 ng/ml Des Pro Ala Bradykinin 21 (12) ND ND 1 ng/ml Leucine Enkephalin 23 (10) ND ND 10 ng/ml 3X FLAG Peptide 115 (8) 125 (20) ND 10 ng/ml Substance P 126 (7) 138 (18) 112 (19) 100 ng/ml [Ala92]-Peptide (4) 848 (11) 841 (12) 100 ng/ml Methionine Enkephalin 1000 (3) 1000 (9) 1000 (10) Sample analyzed using Velos-Orbitrap, TOF MS and IMS-TOF MS instruments

9 Benefits of IMS drift time separation 1. Improved Sensitivity & Increase Feature Detection & Confidence QTOF MS of Bradykinin (100 pm) IMS-QTOF MS of Bradykinin (100 pm)

10 Benefits of IMS drift time separation 1. Improved Sensitivity & Increase Feature Detection & Confidence Only 3 features discerned without drift time dimension (*)

11 Benefits of IMS drift time separation 2. Utilizing IMS in databases reduces FDR due to extra dimension Mass Error LC Error Drift Time Error Human Serum Peptide Analysis 100-min LC-Velos Orbitrap Features with FDR <5% 60-min LC-IMS-MS Features with FDR <5% If you only use LC and MS Features with FDR <5% Extra dimension adds confidence to LC-IMS-MS Features Matches to AMT Tag DB

12 Benefits of IMS drift time separation 3. Distinguish different classes of compounds and sequence isomers P 4+ P 3+ L 2+ P 4+ P 3+ L 2+ P 2+ P 2+ P + P + L + L + L + Peptide : red Lipid : blue Peptides and lipids are easily distinguished

13 Benefits of IMS drift time separation 3. Distinguish different classes of compounds and sequence isomers RRGPFPSPF + RPPGFSPFR + GPFRPRFPS + NW Chem used to model 3-D conformations

14 Simultaneous Targeted and Discovery Proteomic Measurements with LC-IMS-MS

15 Increasing sample throughput and identifications 1. To obtain the highest number of proteins, tissue samples are usually fractionated 2. However, fractionation increases analysis time, introduces errors and causes quantitation difficulties (peptides split between fractions) 3. Our goal is to understand how many proteins can be detected in the unfractionated samples with the Velos Orbitrap, Q-Exactive and IMS-MS platforms using a single 100 minute LC gradient

Patient derived xenograft (PDX) tissue sample 6-9 weeks Mice engrafted bilaterally with WHIM16 breast tumor (Luminal A) Pooled and cryopulverized WHIM16 PDX samples for MS

2013, 4: 1116-30) Has been used for evaluation of the ischemia effect on proteomic and phosphoproteomics analysis (Mertins et al, Mol Cell Proteomics.

16 Patient derived xenograft (PDX) tissue sample 6-9 weeks Mice engrafted bilaterally with WHIM16 breast tumor (Luminal A) Pooled and cryopulverized WHIM16 PDX samples for MS analyses Genomically characterized breast-cancer-derived xenografts (Li et al, Cell Rep. 2013, 4: ) Has been used for evaluation of the ischemia effect on proteomic and phosphoproteomics analysis (Mertins et al, Mol Cell Proteomics. 2014, 13: ) Part of the Comparison Reference (CompRef) sample (has one basal, WHIM2, and one luminal A sample, WHIM16) that has been used as the QC for both identification and quantification throughout the CPTAC proteomic characterization of the TCGA tumors (Zhang et al, Nature. 2014, 513: 382-7)

17 Platform identifications MS/MS Identifications at 1% FDR For Qexact and VOrbi QExact QExact VOrbi 2498 VOrbi MS/MS undersampling (especially for Velos) led us to use an AMT tag database approach

18 Platform identifications MS/MS Identifications at 1% FDR For Qexact and VOrbi AMT Tag Matching at <1% FDR (MS/MS Database 0.01% FDR) QExact VOrbi QExact VOrbi QExact VOrbi 2498 IMS MS/MS undersampling (especially for Velos) led us to use an AMT tag database approach AMT tag database was built from 48 fractions Each unfractionated run was aligned to the database for identifications

22401 IMS 2618 A representative protein was chosen for redundant peptides (peptides that

19 Identifications using AMT tag approach Unique Peptides Non-redundant Human Proteins with at least 2 peptide identification QExact VOrbi QExact 2164 VOrbi 2070 IMS IMS 2618 A representative protein was chosen for redundant peptides (peptides that are shared between multiple proteins) 8,055 peptides and 319 proteins were detected only by IMS-MS

20 Platform sequence coverage difference Non-redundant Human Proteins Of the 1855 proteins detected by all platforms # of Peptides/Protein IMS QE VOrbi > Of the 1855 proteins detected by all platforms, IMS-MS has the least number of proteins with only 1 and 2 peptide identification and the highest number of proteins with more than 4 peptide identifications (best sequence coverage)

21 Platform sensitivity difference Unique Peptides 8055 peptides detected only by IMS-MS 9832 peptides detected by all platforms The 8055 peptides detected only by IMS-MS had lower intensities than the peptides found in all platforms (IMS more sensitive)

22 Simultaneous targeted and discovery IMS-MS approach Protein Peptides Detected/Protein Q-Exactive Orbitrap-Velos IMS-MS Membrane-Associated Progesterone Receptor Component 2 (PR) Kynureninase Isoform B Epidermal Growth Factor Receptor (EGFR) Isoform A Precursor Transmembrane & TPR Repeat-Containing Protein MAP7 Domain-Containing Protein Validated 5 low concentration proteins by spiking in heavy labeled peptides for each (2 peptides/protein except MAP7) Proteins were chosen due to their low abundance in previous studies and potential role in breast cancer

Targeted and discovery IMS-MS approach CPTAC_LabelFree_P6_4_17Sep14_Frodo_14... 9/17/2014 11:03:30 PM RT: 0.00-99.00 100 NL: 2.

23 Targeted and discovery IMS-MS approach CPTAC_LabelFree_P6_4_17Sep14_Frodo_ /17/ :03:30 PM RT: NL: 2.38E TIC F: FTMS + p NSI Full ms [ ] MS CPTAC_LabelFree_P6_ 4_17Sep14_Frodo_ Time (min)

24 Analysis of EGFR peptide (TIQEVAGYVLIALNTVER 3+ ) CPTAC_LabelFree_P6_4_17Sep14_Frodo_ /17/ :03:30 PM RT: NL: 2.38E Target peptide from EGFR Protein 10 nm heavy added TIC F: FTMS + p NSI Full ms [ ] MS CPTAC_LabelFree_P6_ 4_17Sep14_Frodo_ Time (min)

Analysis of EGFR peptide (TIQEVAGYVLIALNTVER 3+ ) CPTAC_LabelFree_P6_4_17Sep14_Frodo_14.

38E9 90 80 70 60 Target peptide from EGFR Protein 10 nm heavy added TIC F: FTMS + p NSI Full ms [400.

25 Analysis of EGFR peptide (TIQEVAGYVLIALNTVER 3+ ) CPTAC_LabelFree_P6_4_17Sep14_Frodo_ /17/ :03:30 PM RT: NL: 2.38E Target peptide from EGFR Protein 10 nm heavy added TIC F: FTMS + p NSI Full ms [ ] MS CPTAC_LabelFree_P6_ 4_17Sep14_Frodo_ nm heavy added Time (min) 0 nm heavy added

Analysis of MAP7 peptide (TPETLLPFAEAEAFLKK 3+ ) CPTAC_LabelFree_P6_4_17Sep14_Frodo_14... 9/17/2014 11:03:30 PM RT: 0.00-99.00 100 NL: 2.

26 Analysis of MAP7 peptide (TPETLLPFAEAEAFLKK 3+ ) CPTAC_LabelFree_P6_4_17Sep14_Frodo_ /17/ :03:30 PM RT: NL: 2.38E Target peptide from MAP7 Protein 10 nm heavy added TIC F: FTMS + p NSI Full ms [ ] MS CPTAC_LabelFree_P6_ 4_17Sep14_Frodo_ nm heavy added Time (min) 0 nm heavy added

27 Biological importance 319 proteins only detected by IMS-MS 297 IMS-MS unique proteins found in Human Protein Atlas Cancer database

Of these 60 proteins: All have functional roles in biological process necessary for cancer progression Hanahan D, Weinberg

28 Biological importance 319 proteins only detected by IMS-MS 297 IMS-MS unique proteins found in Human Protein Atlas Cancer database The Hallmarks of Cancer 60 of these unique IMS-MS proteins ranked as High in at least 25% of breast cancer stains Of these 60 proteins: All have functional roles in biological process necessary for cancer progression Hanahan D, Weinberg Robert A. Cell. 2011;144: > 26% have documented mechanistic roles in breast cancer progression in complementary literature

29 Summary LC-IMS-MS Measurements: Detected lower concentration proteins than the Velos Orbitrap and Q-Exactive platforms Performed discovery and targeted studies simultaneously Identified hallmark of cancer proteins in unfractionated samples, which are usually only observed in fractionated studies

30 IMS-MS Informatic Analyses Using Skyline

31 Why do we want to use Skyline to analyze our IMS-(CID)-MS all-ions fragmentation data? Currently, all of our IMS-(CID)-MS all-ions fragmentation data is manually assigned which is very time consuming and limits our ability to analyze the benefit of IMS We need to understand the effect of the IMS separation on DIA identifications Skyline provides a rapid analysis tool to study MS and MS/MS quantitation with and without the IMS dimension

32 IMS-(CID)-MS all-ions data acquisition 4-12 Torr of Buffer Gas Fragmentation takes place after the drift cell

33 Fragments have the same drift time as precursors Parent Spectrum Fragmentation Spectrum

Due to time consuming manual analyses, until now we have only studied specific fragments No Heme (Heme) + Heme Containing Peptides Heme Containing Peptides Parent Spectrum (3

34 Due to time consuming manual analyses, until now we have only studied specific fragments No Heme (Heme) + Heme Containing Peptides Heme Containing Peptides Parent Spectrum (3 Peptides Observed in 1 sec LC acquisition) 2 hemepeptides identified Zoomed into 617 region Fragmentation Spectrum Find multiple hemepeptides simultaneously by looking for heme + ion at 617

35 Due to time consuming manual analyses, until now we have only studied specific fragments Phosphopeptides Parent Spectrum (11 Peptides Observed in 1 sec LC acquisition) 3 Phosphopeptides Identified Find multiple phosphopeptides by looking for PO 3 or PO 4 mass difference Fragmentation Spectrum

36 Concentration dependent IMS-(CID)-MS experiment Spiked 8 different concentrations of tryptic BSA digest from 100 pm to 1 µm into 0.1 ug/ul yeast digest Analyzed the effect of IMS on the MS and MS/MS quantitation for the DIA data All ions fragmentation IMS-(CID)-MS All ions fragmentation (CID)-MS BSA Concentration in Yeast 100 pm 1 nm 5 nm 10 nm 50 nm 100 nm 500 nm 1 µm

37 IMS-MS and IMS-(CID)-MS were alternated IMS-MS IMS-(CID)-MS (CE=29V) CE of 29 V was chosen since it fragmented 2+ ions well

38 Skyline provides the ability to perform fast DIA analyses Good quantitation from 1 µm to 1 nm for all peptides when drift time filtering was used When drift time filtering wasn t used, the lower concentration peptides quantitation wasn t as linear

39 Drift time removes interference in MS quantitation

40 Another peak interferes when drift time filtering isn t performed 5 nm 10 nm

41 Drift time filtering removes interference in MS/MS quantitation Good quantitation from 1 µm to 1 nm for 7 of 10 peptides when drift time filtering and 10 for 10 from 1 µm to 10 nm When drift time filtering wasn t used, many interferences caused bad quantitation

42 Baseline noise increases when drift time isn t used

43 Interfering peaks cause quantitation problems when drift time isn t used

44 Summary Skyline: Quickly analyzed the IMS-(CID)-MS data Showed that IMS drift time allows better MS and MS/MS quantitation by removing interfering peaks Enables further studies to better understand the effect of IMS on other types of DIA data (such as SWATH)

45 Conclusions IMS-MS: Separates isomeric peptides Distinguished different classes of compounds Allows detection of low concentration components Is easily coupled with LC and other front end separations Data analysis is becoming automated with new programs such as Skyline

Acknowledgements Agilent Technologies National Institute of Environmental Health Sciences of the NIH (R01ES022190) NIH: General Medical Sciences Proteomics Center at PNNL (2 P41 GM103493-11) and

46 Acknowledgements Agilent Technologies National Institute of Environmental Health Sciences of the NIH (R01ES022190) NIH: General Medical Sciences Proteomics Center at PNNL (2 P41 GM ) and National Cancer Institute Leidos Biomedical Research, Inc. PNNL Laboratory Directed Research and Development Program Environmental Molecular Sciences Laboratory Biological Separations & Mass Spectrometry Group

Increasing Molecular Coverage in Complex Biological and Environmental Samples by Using IMS-MS

Increasing Molecular Coverage in Complex Biological and Environmental Samples by Using IMS-MS Erin Shammel Baker Kristin E. Burnum-Johnson, Jon M. Jacobs, Yehia M. Ibrahim, Daniel J. Orton, William F.