Upload
others
View
2
Download
0
Embed Size (px)
Citation preview
Multi‐Omics Profiling of Methionine‐Restricted MCF7 Cells in 24 Hours Using a Prototype UPLC‐
Compatible Microfluidic Device
J. Will Thompson 1, Jay Johnson2, Giuseppe Astarita2, XiaohuTang1, Giuseppe Paglia3, Jim Murphy2, Steven Cohen2, Mark Bennett4, Jen‐Tsan Chi1, Jim Langridge4, Geoff Gerhardt2,
and M. Arthur Moseley1
1Duke University School of Medicine; 2Waters Corporation, Milford, MA and Manchester, UK;
3Center for Systems Biology, University of Iceland, 4Nonlinear Dynamics, Durham, NC
High Throughput MS‐Based ‘Omics Methods
Cox, J., and Mann, M. Nat. Biotechnol. 2008, 26, 1367–1372.
Nagaraj, N and Mann, M. et al. Mol. Cell. Proteomics 2012 11, M111.013722.
ProteomicsLipidomicsShevchenko et al. J Mass Spectrom. 2012 Jan;47(1):96‐104.
Reid et al. Anal Chem. 2012 Nov 6;84(21):8917‐26.
Castro‐Perez et al, J Proteome Res. 2010 May 7;9(5):2377‐89.
MetabolomicsKurwin IJ et. al, Ann. N.Y. Acad. Sci.2013 1287, 1–16 C
Want E et. al Nature Prot. 2013 8(1): 17‐32
A Case for 150 um Separations
Assumptions: UPLC conditions of 0.45 cm/sec linear velocityAiming for fast, efficient separation (10‐15 min per sample or fraction)Signal is proportional to analyte concentrationSwept volume (‘delay’) is 3 ul
• It offers a nice blend of:– Sensitivity of capillary scale– Solvent and sample use of
capillary scale– Throughput of analytical
(UPLC) scale• But, there have been ‘historical’
hangups:– Spray stability and robustness
at 2‐4 ul/min– Columns that are easy to
install, with UPLC performance
390 ul/min
90 ul/min (1 mm)
0.5 ul/min (75 um)
0.22 ul/min (50 um)
8 ul/min (300 um)
2 ul/min (150 um)
1.4 ml/min
Tile Design and Flow Diagram
Incoming flow
Analytical Column
Electrical Connections(EEPROM, Heater)
ESI EmitterAssembly
Flexible Configuration of 150 um Tile DeviceConfiguration for Infusion and 1DLC(Polar and Nonpolar Metabolites)
Configuration for High/Low pH 2DLC(Proteins)
Tile options tested:5, 10, 20 cmBEH C18HSS T3 C18CSH C18
BEH C4BEH Amide HILICInfusion Tile
The Beauty of the Tile Interface… seamless column installation for UPLC
Pressure
Solvent AFlow
Solvent BFlow
‐70.0 ‐60.0 ‐50.0 ‐40.0 ‐30.0 ‐20.0 ‐10.0 0
Injections w 20 cm column Injections with 10 cm columnTILE CHANGE
Target Timeline of a “Comprehensive” MS‐based ‘Omics Profile on a Single System
LIPID INFUSION METHODS: 5 min/sampleMETABOLITE UPLC METHODS: 15 min/samplePROTEIN 2D UPLC METHOD: 3 hr/sample
0 1 2 3 4 5 6 7 8 9 10 24
Tune and CalibrateLIPID (ESI+)LIPID (ESI‐)
Metabolite (RPLC, ESI+)
Metabolite (HILIC, ESI+)Metabolite (HILIC, ESI‐)
Proteome (3‐fraction RP/RPLC, ESI+)
transition (column change)
transition (column and wash solvent change)
transition (change LC and column)
Data Collection Timeline (Real Time Capture)
Tune and Calibrate+ESI Lipids
‐ESI Lipids
Tile Change
Equilibration+ESI HILIC ‐ESI HILIC
Tile Change
Wash Solvent Exchange
+ESI RPLC
7:00 am 9:00 am 11:00 am 1 pm 3 pm
Data Collection Timeline, continued…Switch LC System
2DLC Sample 1 Sample 2 Sample 3 Sample 4 Sample 5 Sample 6
4 pm 7 pm 10pm 1 am 4 am 7 am 10am 1 pm(Will goes to beach)
Bradford Assay(normalize by total lysate)
Cell Disruption(Sonication in AmBic pH8)
Polar Metabolites80/20 MeOH/water
1 hr extraction, N2 dry
Lipids80/20 MTBE/MeOH1 hr extraction, N2 dry
Proteins0.25% w/v RapigestDTT/IAA/trypsin overnight
ResuspendMeOH (HILIC)
OrH20/MeCN (RPLC)(15 min/sample)
Resuspend4/2/1
IPA/MeOH/CHCl3(5 min/sample)
Acidify, Spin, Dry 20 mM Ammon Formate pH 10Inject 8 ug for 2DLC‐MS/MS(3 hr/sample)
Sample Preparation StrategyMCF7
+/‐Methionine
Bradford Assay(normalize by total lysate)
Cell Disruption(Sonication in AmBic pH8)
Polar Metabolites~48%
80/20 MeOH/water1 hr extraction, N2 dry
Lipids~48%
80/20 MTBE/MeOH1 hr extraction, N2 dry
Proteins~4%0.25% w/v RapigestDTT/IAA/trypsin overnight
ResuspendMeOH (HILIC)
OrH20/MeCN (RPLC)(15 min/sample)
Resuspend4/2/1
IPA/MeOH/CHCl3Inject 4% for FIA(5 min/sample)
Acidify, Spin, Dry 20 mM Ammon Formate pH 10Inject 8 ug for 2DLC‐MS/MS(3 hr/sample)
Sample Preparation StrategyMCF7
+/‐Methionine
BPI
MP B Co
mpo
s.
m/z
Time (min)
Development of a Versatile RPLC Metabolomics Method
150 um x 10 cm 1.7 um BEH C18 tile, F = 2.0 ul/min at 45°CMobile phase A: 0.1% Formic acid, 0.02% HFBA, in water Mobile Phase B: 0.1% Formic acid in 10/90 IPA/MeCNMass Spectrometry: Synapt G2 HDMS, Resolution mode (25,000 Rs) @ 5Hz
~16,500 total deisotoped features~2,900 ‘high confidence’ features
Polar amino acids
phospholipids
‘native’ peptides
Moderate polarity metabolites
Software for Quantitative Analysis, “Omic” Tools(5’‐methyl thioadenosine (MTA), from RPLC data)
+Met ‐Met
Progenesis COMET/TransOmics(Nonlinear/Waters)
Rosetta Elucidator(Rosetta Biosoftware)
min2.0 4.0 6.0 8.0 10.0 12.0
%
0
100
F1:TOF MS,ES+298.089
-M1 sample +ESI metab
8.740e+003MTA6.22
401.258685
5.784.901.88 3.82
min2.0 4.0 6.0 8.0 10.0 12.0
%
0
100
F1:TOF MS,ES+298.089
pool of +M and -M samples, 2 ul injection, 4x dilution
4.003e+004MTA6.22
1850.2239909
4.89
+Met ‐Met
Software for Quantitative Analysis, “Targeted” Tools(5’‐methyl thioadenosine (MTA), from RPLC data)
Targetlynx(Waters)
Skyline v1.4.1(MacCoss, MacLean)
‘Direct’ HILIC Method Translation from 2.1mm to 0.15 mm
Parameter 2.1 mm UPLC (Column) 0.15 mm UPLC (Tile)
Column Length 150 mm 100 mm
Mobile phase A/B 0.1% FA in H20/MeCN 0.1% FA in H20/MeCN
Flow Rate (Velocity) 0.4 ml/min (0.45 cm/sec) 0.002 ml/min (0.44 cm/sec)
Loop offline* 0.1 min 2.5 min
Gradient 99% to 30 % B in 5.9 min 99% to 30% B in 7 min
Sample reconstitution* 50/50 MeCN/H2O 99/1 MeCN/H2O
Methionine
Xanthene
Lactose
Arginine
Bradford Assay, 1.8mg/sample(normalize by total lysate)
Cell Disruption(Sonication in AmBic pH8)
Polar Metabolites~48%
80/20 MeOH/water1 hr extraction, N2 dry
Lipids~48%
80/20 MTBE/MeOH1 hr extraction, N2 dry
Proteins~4%0.25% w/v RapigestDTT/IAA/trypsin overnight
Resuspend2/1/0.2 MeCN/
Formic Acid/HFBAInject 1% for LC‐MS/MS
(30 min/sample)
Resuspend4/2/1
IPA/MeOH/CHCl3Inject 4% for FIA(10 min/sample)
Acidify 1/2/97 TFA/MeCN/waterInject 20% for 2DLC‐MS/MS(3 hr/sample)
Summary of Multi‐Omics Sample Preparation Strategy
Lipid Profiling using Flow Injection Analysis and an Infusion Tile
Analysis of the Lipid Isolate from MCF7 cells (prepared using MTBE/MeOH extraction). ‐ Ion‐Mobility Data‐Independent Analysis‐ Synapt G2, 0.6 sec scans (6V then 15‐45V)‐ 3 ul/min flow rate‐ Mobile phase was 10/90 IPA/MeCN with
0.1% formic acid
Rapid Lipid Profiling and Statistical Analysis (AUC)
TG (57:9)
METLIN Search (10 ppm) (http://metlin.scripps.edu/metabo_batch.php)
+ Met ‐Met
m/z
Time (min)
6 samples, ESI + and ‐871 Confident, Charge = 1 Features
71 Features Passing ANOVA p<0.05 (Hochberg FDR correction)
Bradford Assay, 1.8mg/sample(normalize by total lysate)
Cell Disruption(Sonication in AmBic pH8)
Polar Metabolites~48%
80/20 MeOH/water1 hr extraction, N2 dry
Lipids~48%
80/20 MTBE/MeOH1 hr extraction, N2 dry
Proteins~4%0.25% w/v RapigestDTT/IAA/trypsin overnight
Resuspend2/1/0.2 MeCN/
Formic Acid/HFBAInject 1% for LC‐MS/MS
(30 min/sample)
Resuspend4/2/1
IPA/MeOH/CHCl3Inject 4% for FIA(10 min/sample)
Acidify 1/2/97 TFA/MeCN/waterInject 20% for 2DLC‐MS/MS(3 hr/sample)
Summary of Multi‐Omics Sample Preparation Strategy
Proteomics Analysis3‐Fraction 2DLC with IMS‐ToF (DIA) 3 hrs/sample
0
100
200
300
400
500
600
5 15 25 35 45 55 65 75 85 95
More
Protein Co
unt p
er Bin
% CV Bin
% CV (n=2207 proteins)
8 ug MCF7 digestHigh/Low pH RPLC
80% of proteins below 25% CV (biological variation)
Technical SummaryMCF7 Multi‐Omics Profile in 24 Hours
Method Time/sample(min)
Total Features
ConfidentFeatures
Median % C.V.(n=6)
Lipid Infusion, ESI + 5 1750 6621 27
Lipid Infusion, ESI ‐ 5 851 2001 52
Metabolite, HILIC 16.5 6,123 8502 55
Metabolite, RPLC 16.5 16,745 2,9162 25
Protein, 2D RPLC 180 172,569 2,2073 14
1 Filtered based on charge state = 12 Filtered based on charge state = 1, peak width < 0.5 min, peak confidence score > 0.83 Identified and quantified proteins
Cutting Run Time in Half while Maintaining the Separation Efficiency
5.00 10.00 15.00 20.00 25.00 30.00 35.00 40.00
%
0
100
5.00 10.00 15.00 20.00 25.00 30.00 35.00 40.00
%
0
100 BPI1.31e5
22.85995.21
20.31748.48
18.32748.928.94
478.817.62
585.84
13.28622.88
22.91995.21 39.16
795.7930.30706.7528.48
1196.27 32.42964.08
38.46987.5534.83
750.9639.26
1022.2143.99995.54
ID09049_01_UCA195_3252_092712_04 1: TOF MS ES+ BPI
9.02e418.19858.47
12.57549.8410.76;720.95
9.26613.40
7.11617.37
4.43;409.25
15.63724.43
18.34858.47 20.63
983.53 24.96995.54 28.75
774.41 31.62748.42
32.74935.25 39.16
949.15
5.00 10.00 15.00 20.00 25.00 30.00 35.00 40.00
%
0
100
5.00 10.00 15.00 20.00 25.00 30.00 35.00 40.00
%
0
100 BPI1.28e5
13.09995.546.21
730.43
5.50585.84
20.47987.55
16.99706.75
20.56927.15
09156_01_UCA195_3252_092712_04 1: TOF MS ES+ BPI
9.77e410.84858.467.07;720.94
5.90;648.84
4.18550.80
14.09995.53 17.89
935.49
20.51;1118.03
150 um x 100 mm BEH C18 nanoTile7 to 35% MeCN; 3.0 ul/min
Fxn 4
Fxn 5
Fxn 4
Fxn 5
3
912.5
3
9 12.5
150 um Tile Loading Test150 um x 100 mm, 37 min gradient
Loading test performed with E. Coli lysateMethod: 5 to 40% MeCN in 37.1 min, 3 ul/min, 35COnly very minor effects on chromatographic efficiency at 4 ug load
AcknowledgmentsDuke University Proteomics Core Facility
http://www.genome.duke.edu/cores/proteomics/
FundingNIH S10 grantDuke School of MedicineCTSA grant UL1RR024128