Integration of HPLC-FTICR MS and HPLC-QIT MS2 to Achieve Enhanced Proteome Characterization Chongle Pan1,2,3 Nathan VerBerkmoes1,3 Praveen Chandramohan2 Nagiza Samatova2,3 Robert Hettich1,3

1Chemical Sciences Division; 2Computer Science and Mathematics Division, Oak Ridge National Lab; 3Genome Science and Technology Graduate School, ORNL-University of Tennessee

Proteomics Systematic analysis of protein complement in a given cell, tissue or organism for their identity, quantity and function. Major challenges: proteome coverage, low abundance proteins, membrane proteins, post-translationally modified proteins, General procedures: Gel-based or LC-based separation, MS-based characterization, informatics-based identification.

Current approaches to proteomics MudPIT technology (1)

1. Biphasic column separation of tryptic digested proteome, integrating SCX resin and RP resin in one column2. Introduction of eluent directly to QIT through ESI interface; data-dependent MS2 scan of eluding peptides3. Identification of peptides by SEQUEST database searching (2)4. Assemble of peptide identification to protein identification by DTAselect (3)

Switching LC technology (4)1. Autosampler loading of tryptic digested proteome onto a trapping cartridge2. 1D RP-LC separation or 2D switching SCX/RP LC separation3. QIT data-dependent MS2 scan of eluding peptides, possible multiple mass range scan4. SEQUEST/MASCOT database searching and protein identification

Advantages: HPLC separation of peptides are relatively unbiased. MS2 scans are informative of peptide sequenceDisadvantages: The score from MS2 database searching algorithm is the only base for

identification calls The selection of cutoff is a compromise between identification specificity

and sensitivity Common cutoff for SEQUEST Xcorr (+1) > 1.8, (+2) > 2.5, (+3) > 3.5

AMT technology (5)“An AMT tag is a peptide with a sufficiently distinctive mass and LC elution time to act as a biomarker for a given protein.” 1. Construction of potential mass tag (PMT) library

1.1 Capillary HPLC / QIT MS2 characterization of a given organism’s proteome under a variety of conditions1.2 Establishment of PMT for a peptide using a very liberal cutoff ( SEQUEST Xcorr (+1, +2, +3)> 2.0)

2. Construction of accurate mass and time tag (AMT) library2.1 Capillary HPLC / FTICR MS characterization of those proteome samples2.2 Establishment of AMT for a peptide matched with its PMT’s retention time and calculated mass

3. Proteome examination by capillary HPLC / FTICR MS 3.1 Capillary HPLC / FTICR MS characterization of the proteome samples of interest3.2 Identification of protein by searching AMT library with measured mass and retention time

Advantages: High throughput proteome examination once AMT library is constructed High proteome coverage given a confident AMT libraryDisadvantages: Large initial effort investment in constructing AMT library Multiple dimensional separation not straight-forward False positive protein identification due to insufficiently distinctive mass and LC elution time of an AMT tag

Objectives

Proteome characterization generally consists of peptide separation by high performance liquid chromatography (HPLC) and peptide identification by tandem mass spectrometry (MS/MS). We aim to enhance the specificity and sensitivity of peptide identification by in silico integrating nanoLC-FTICR-MS and nanoLC-QIT-MS/MS.

Methods nanoLC-FTICR-MS: 1D reverse phase LC coupled online to IonSpec 9.4T FTICR with MS scans

nanoLC-QIT-MS/MS : 1D reverse phase LC coupled online to LCQ with data dependent MS/MS scans

in silico integration: Retention time normalization; peptide correlation by retention time and mass

Results Development of a robust nanoLC-FTICR-MS system

Development of retention time normalization and peptide correlation

Demonstration of enhanced peptide identification from simple protein mixture digest

Methods:

1D-reverse phase Liquid Chromatography Autosampler injection, LC packings FAMOS 50 ul Preconcentration: 300 um x 5 mm C18 PepMap RP-LC: Vydac 75 um id x 25 cm C18 nanocolumn

Nanospray Tip: 10 um ID New Objective Picotip Vendcap: -1800V Vtip: 0V Ventrance: -1800V Distance: 3mm

FT ICR MS 9.4 Tesla IonSpec 2 sec. hexapole ion accumulation 256K data points @ 1Mhz ADC 2-scan signal averaging, ~ 9 sec. per spectrum

Results

Standard protein mixture tryptic digest Constituents: ALCOHOL DEHYDROGENASE I and II, CARBONIC ANHYDRASE, HEMOGLOBIN A and B chains,

MYOGLOBIN, LYSOZYME C, SERUM ALBUMIN

Rhodopseudomonas palustris proteome cleared fraction (Soluble proteins)

3 consecutive nanoLC- FTICR MS runs with proteome samples

1. Parallel LC/MS experiments

2. Data process and extraction

3. Normalization of retention time

1. Selection of highly confident peptide hits from LCQ data

Xcorr (+1) > 2.3; (+2) > 3.0; (+4) > 3.5 Total 206 I.D.

2. Search against FT ICR data with calculated monoisotopic masses

Mass error < 0.05 Da

3. Data representation and visual inspection to remove spurious hits

Deletion of the obvious outliers

4. Linear regression: RTLCQ = 0.9669*RTFT - 11.5872

correlation coefficient = 0.94

4. Correlation and Integration

Sample preparation and protein digestion The protein sample is denatured with 6-8 M Guanidine or Urea, and reduced with DTT or some other reducing agent at

60oC for 10-60 minutes. The denaturant concentration is lowered by dilution with Tris; sequencing grade trypsin is added to the sample and

incubated overnight The sample is desalted by solid phase extraction and organic solvent is removed by SpeedVac

1D LC / QIT MS2 One-dimensional LC-MS/MS experiments were performed with an Ultimate HPLC

(LC Packings, a division of Dionex, San Francisco, CA) coupled to an LCQ-DECA ion trap mass spectrometer (Thermo Finnigan, San Jose, CA) equipped with an electrospray source. Injections were made with a Famos (LC Packings) autosampler onto a 50ul loop. Flow rate was ~4ul/min with a 240min gradient for each run.

A VYDAC 218MS5.325 (Grace-Vydac, Hesperia, CA) C18 column (300µm id x 15cm, 300Å with 5µm particles) or a VYDAC 238EV5.325 monomeric C18 (300µm id x 15cm, 300Å with 5µm particles) was directly connected to the Finnigan electrospray source with 100µm id fused silica.

For all 1D LC/MS/MS data acquisition, the LCQ was operated in the data dependent mode with dynamic exclusion enabled, where the top four peaks in every full MS scan were subjected to MS/MS analysis.

SEQUEST and DTASelect The resultant MS/MS spectra from the sample were searched with SEQUEST against the six constituent protein sequence

and all predicted ORFs from R. palustris. The ORFs from R. palustris serves as indication of false position rate. The raw output files were filtered and sorted with DTASelect. The filter criteria include the minimal Xcorr, the validation flag

for FT ICR data hits and the FTICR mass measurement error

1800

1900

2000

2100

2200

2300

2400

2500

2600

2700

2800

2900

3000

3100

3200

3300

LCQ

_RT

2000 3000FTICR_RT

Mass tolerance < 0.05 DaRT tolerance < 3 minXcorr (+1) > 1.3, (+2) > 2.0, (+3) > 3.0)

Export of results to DTASelect readable format. Use of manual validation flag to flag the presence of FTICR data hits

1000

2000

3000

4000

5000

6000

Mas

s

1000 2000 3000 4000 5000

RT

Total ion chromatogram

1 hour gradient

MS @ 21min MS @ 34min

Zoom-in of m/z 610 – 640 High dynamic range and sensitivity

Deconvoluted spectrum 4000- 4100 DaCharge state measurable

Total ion chromatogram

1 hour gradient

MS @ 15 min Zoom-in of m/z 800 - 1000

Deconvoluted spectrumHighly complex low abundance peptides resolved

Automatic 3 repetitive sample injectionsHighly reproducible results (mass spectra)A robust LC/MS system

LC-FTICR MSTotal Ion ChromatogramStandard protein mixture digest60 min gradient LC

LC-QIT-MS2

Total Ion ChromatogramStandard protein mixture digest60 min gradient LC

SEQUEST program Peptide identification

thru database searching (8 protein sequences + 4800 distracting protein sequences from R. palustris) No trypsin cleavage specificity used.

DTASelect program Filter and assemble

peptides. Identification (Xcorr (+1) > 1.8, (+2) > 2.5, (+3) > 3.5)

Total 301 MS/MS SEQUEST I.D.s passed cutoff

Ionspec FTdoc program

Determine charge state De-isotope cluster Filter noise Export monoisotopic

masses, retention time and intensity to ACSII files

Total 5496 data points observed

1000

2000

3000

4000

5000

6000

Mas

s

1000 2000 3000 4000

RT

0

1000

2000

3000

4000

5000

6000

7000

8000

9000

10000

11000

12000

13000

14000

15000

16000

17000

Mon

o M

ass

1900 2000 2100 2200 2300 2400 2500 2600 2700 2800 2900 3000 3100 3200 3300 3400 3500 3600

RT

Mo

no

iso

top

ic

ma

ss

Ca

l. M

on

ois

oto

pic

m

ass

Retention time

FTICR RT

LC

Q R

T

Normalization of retention time of LC-FTICRMS to the same scale as LC-LCQMS2

Peptide identification with lowered Xcorr cutoff. (Xcorr (+1) > 1.3, (+2) > 2.0, (+3) > 3.0)

Total 442 MS/MS SEQUEST I.D.s passed cutoff

Retention time

Ca

l. Mo

no

isoto

pic

ma

ss

OVERVIEW

INTRODUCTION

1. Link AJ, Eng J, Schieltz DM, Carmack E, Mize GJ, Morris DR, Garvik BM, Yates JR 3 rd Nat Biotechnol. 1999 17, 676-82.

2. Eng JK, McCormack AL, Yates JR 3rd, J. Am. Soc. Mass Spectrom. 1995 67, 1426-1436

3. Tabb DL, McDonald WH, Yates JR 3rd, J Proteome Res. 2002 1, 21-6.

4. VerBerkmoes NC, Bundy JL, Hauser L, Asano KG, Razumovskaya J, Larimer F, Hettich RL, Stephenson JL Jr. J Proteome Res. 2002 1, 239-52.

5. Smith RD, Anderson GA, Lipton MS, Masselon C, Pasa-Tolic L, Shen Y, Udseth HR. OMICS. 2002 6, 61-90

Acknowledgement :

Dr. David Tabb and Dr. Hayes McDonald are acknowledged for technical input and discussions.

C.P. and N.V. thank Genome Sciences and Technology graduate program for financial support.

Research was sponsored by the U.S. Department of Energy, Office of Biological and Environmental Research. Oak Ridge National Laboratory is operated and managed by the University of Tennessee-Battelle, LLC. for the U.S. Department of Energy under contract DE-AC05-00OR22725

REFERENCES

NANOLC-FTICR-MS METHODS AND RESULTS INTEGRATION METHODS AND RESULTS

CONCLUSIONS

“Y”, if there is an FTICR hit within the mass and RT tolerance

LC QIT MS2 METHODS

LC-MS/MS (base peak chromatogram)

MS (full scan) at 60 .4 minutes

MS/MS spectra of 881 precursor

Algorithm identifies the peptideas originating from a

hypothetical ORF

Database Search Algorithm

Lysate_4mz_vydac_2nd_103001 #1475 RT: 60.44 AV: 1 NL: 4.81E7T:+ c ESI Full ms [ 780.00-1200.00]

800 850 900 950 1000 1050 1100 1150 12000

10

20

30

40

50

60

70

80

90

100

Lysate_4mz_vydac_2nd_103001 #1476 RT:60.46 AV:1 NL:5.23E5T:+ c d Full ms2 881.21@35.00 [ 230.00-2000.00]

400 600 800 1000 1200 1400 1600 1800 20000

10

20

30

40

50

60

70

80

90

100

RT:0.00 - 183.59

0 20 40 60 80 100 120 140 160 1800

10

20

30

40

50

60

70

80

90

100 NL:1.14E9Base Peak m/z= 100.0-2000.0 MS Rpal_Pellet_Full_112402

m/z 881 peptide

Development of a robust nanoLC-FTICR-MS system Stable electrospray and good sensitivity in highly aqueous solution Low sample consumption (~ 1 ug of total peptide loaded) Good mass accuracy, mass resolution, dynamic range, and reproducibility

Development of retention time normalization and peptide correlation Use of linear regression to offset gradient start time and normalize gradient slope High correlation coefficient in retention times between LC-FTICR-MS and LCQ Demonstration of comparability of FTICR data and QIT data

Demonstration of enhanced peptide identification from simple protein mixture digest FTICR data used as a validation method for SEQUEST identification Improved specificity and sensitivity in peptide identification for simple protein mixtures; should be much more enhanced for

proteomes Improvements in the mass measurement accuracy of FTICR and retention time reproducibility of LC system should provide

more narrow windows Improvements in peptide identification scoring should provide a more rigorous method to integrate FTICR accurate mass

measurements and QIT MS/MS measurements.

0

1000

2000

3000

4000

5000

6000

7000

8000

9000

10000

11000

12000

13000

14000

15000

16000

17000

Mon

o M

ass

1900 2000 2100 2200 2300 2400 2500 2600 2700 2800 2900 3000 3100 3200 3300 3400 3500 3600

RT

Correlation

Inte

gra

tio

n

Normalized retention time

Mo

no

isoto

pic

ma

ss