(2525) Structural Characterization Of Glycopeptides: Integration Of ...
Embed Size (px)
Transcript of (2525) Structural Characterization Of Glycopeptides: Integration Of ...
l h f Structural characterization of glycopeptides: Integration of g y p p g
complementary CID and ETD based information.information.
K.F. Medzihradszky1,2, A.J. Lynn1, P. Baker1, R.J. Chalkley1, Z. Darula2 and A.L. Burlingame1.
1Mass Spectrometry Facility, University of California San Franciscop y y y2Proteomics Research Group, Biology Research Center of the Hungarian
Academy of Sciences, Szeged, Hungary.
O-glycosylationg y y
Two functionally different types:1. Long-chain modifications of extracellular proteins2. Single residue regulatory modification of
intracellular proteins O GlcNAcylation (analogous intracellular proteins O-GlcNAcylation (analogous to phosphorylation)
O-GlcNAcylation is increasingly studied for its regulatory role and interaction with phosphorylation.
Secreted O-glycosylation is a somewhat neglected research area, even though aberrant glycosylationhas been linked to cancer and Alzheimers diseasehas been linked to cancer and Alzheimer s disease.
The study of both types of O-glycosylation present similar analytical challenges by mass spectrometry.
All O-linked glycopeptides display characteristic fragmentation in CID:
The glycosydic bond is weaker than the peptide bonds The glycosydic bond is weaker than the peptide bonds, thus the spectrum is usually dominated by ions present due to carbohydrate losses and non-reducing end oxonium i ions.
Gas-phase deglycosylaton occurs - the sugar is eliminated without leaving any telltale sign on the originally modified without leaving any telltale sign on the originally modified amino acid, making site assignments impossible in most instances. (Figure 1).
MS/MS analysis utilizing ETD permits both the identification of the modified sequence and the identification of the modified sequence and the unambiguous determination of the modification site(s) (Figures 2 & 3) [2,3].
P i / 714 8520(4 )
Figure 1: CID of SAGalGalNAc-Modified Peptide
T8110318 #2241 RT: 30.40 AV: 1 NL: 2.36E4T: ITMS + c NSI d Full ms2 [email protected] [185.00-2000.00]
Precursor ion = m/z 714.8520(4+).
300 400 500 600 700 800 900 1000 11000
20 518.40509.48 734.27
1035.39 1100.91828.06500.51 1165.57866.37 975.34415.08375.05 553.74
300 400 500 600 700 800 900 1000 1100m/z
Sugar structure can be identified, but neither peptide sequence nor modification site can be determined.
Figure 2: ETD of Same SAGalGalNAc-Modified Peptide as Figure 1.
T8110318 #2242 RT: 30.40 AV: 1 NL: 1.48E4T: ITMS + c NSI d Full ms2 [email protected] [50.00-2000.00]
(2+) c14(2+) c17
1106.82939.17 1250.311046.44 1149 68 1307 09854 94394 38 646 13
c2 c3 z4 z6 z10 z12
400 600 800 1000 1200 1400 1600 1800m/z
1149.68 1307.09854.94394.38 646.13472.15247.17 1709.66 1934.591521.97
636KTFMLQASQPAPT(GalNAcGalSA)HSSLDIK655 from Inter-alpha-trypsin inhibitor heavy chain H1 precursor was confidently identified.
Sample Preparation & Mass Spectrometry
* GlcNAc-modified glycopeptides were enriched on a wheat germ agglutinin column as published earlier . * Secreted glycopeptides bearing mucin core-1 type Secreted glycopeptides bearing mucin core 1 type structures were isolated by Jacalin-affinity chromatography, and were analyzed intact or after di ti t l l th G lNA  digestion to leave only the GalNAc core .
MS analyses were performed on an LTQ-Orbitrap mass y p pspectrometer in LC/MS mode. The instrument was operated in a data-dependent fashion: MS acquisitions were followed by CID and ETD analyses of 3 software-selected multiply by CID and ETD analyses of 3 software selected multiply charged ions. Precursor masses were measured in the Orbitrap, while MS/MS experiments were performed and th f m nts m s d in th lin t pthe fragments were measured in the linear trap.
Data-processing I Raw data were converted into peaklists using in-house
software, PAVA. Database searching was performed against the UniProt database supplemented with a random sequence pp qfor each entry, and species specified as Bos taurus (31074 entries searched).
Trypsin was specified as the enzyme, 1 missed cleavage and Trypsin was specified as the enzyme, 1 missed cleavage and non-specific cleavage at one of the peptide termini were permitted.
Mass accuracy: 15 ppm for precursor ions and 0 6 Da for Mass accuracy: 15 ppm for precursor ions and 0.6 Da for fragment ions.
Carbamidomethylation of Cys = fixed modification; variable modifications: acetylation of protein N termini; Met modifications: acetylation of protein N-termini; Met oxidation; cyclization of N-terminal Gln residues; and HexHexNAc or SAHexHexNAc (for intact glycopeptides) and HexNAc (for partially deglycosylated or GlcNAc-and HexNAc (for partially deglycosylated or GlcNAcmodified peptides) modification on Thr and Ser residues. 3 modifications per peptide.
Data-processing II The same peaklists were used by 2 versions of the search
engine, Protein Prospector: v.5.3 and v5.4: Scoring in v5.3 was optimized for tryptic peptides Scoring in v5.4 uses different weighting depending on
precursor charge and basic residue locationprecursor charge and basic residue location.
Acceptance criteria: Protein score: 22, E>0.05; peptide Acceptance criteria Protein score 22, E 0.05; peptide score: 15, E>0.1.
P i P ( 5 5) i bli l il bl Protein Prospector (v5.5) is publicly available:
http://prospector ucsf eduhttp://prospector.ucsf.edu
New features in Protein Prospector v5.4 t d T d (P t 2781)presented on Tuesday (Poster 2781)
1) Neutral loss peaks from charge reduced species are d i t d t b hi removed prior to database searching.
2)Multiply-charged fragment ions are now considered in ETD data (for precursors of charge state 3+ or higher). ( p g g )
3)New scoring was introduced that gives different fragment ion type weighting depending on the precursor ion charge state and presence of basic residues at the ion charge state and presence of basic residues at the peptide termini.
Re-interrogation of Published  Datasets
Modified with SA1 0GalGalNAc: 16 LC/MS/MS files
Table 1. Comparison of the performance of different versions of ProteinProspector BatchTag. Version 5 3 3 Version 5 4 2 improvement
Modified with SA1-0GalGalNAc 16 LC/MS/MS files
Comparison of Number of Glycopeptides Identified at Different Acceptance Thresholds Version 5.3.3 Version 5.4.2 improvement
Figure 3: One of the new glycosylation sites identified: Thr-605 of Kininogen-1
T8110314 #1483 RT: 25.77 AV: 1 NL: 1.98E3T: ITMS + c NSI d Full ms2 [email protected] [50.00-2000.00]
717 8379 (4 )
90 717.8379 (4+)
1075.98620.31 1405 83921 74
400 600 800 1000 1200 1400 1600 1800 2000m/z
1405.83921.74 1231.90563.45 1035.21 1381.83505.79362.27 1136.05897.77801.41 1472.70 1760.85 1843.101572.93
c3 z4 c6z13
Filtering for glycopeptides CID data provides useful information to support ETD identification:p pp
1. Characteristic sugar losses indicate the presence of the carbohydrate
2. Limited peptide fragmentation may aid ID confirmation.
Identifying the characteristic neutral losses could be automated: Input: separated CID and ETD peaklists in mgf format.
P l s i t fi d t l l ss l ( si 0 2 1) Perl script find_neutral_loss.pl (version 0.2.1):1. Identifies CID spectra in which peaks corresponding to neutral
losses of sugar moieties (for example, -203.1 or -291.1 Da, corresponding to HexNAc or sialic acid, respectively) from the p g , p y)precursor ion are observed. The script looks for 1 or 2 losses at all the possible charge states (1+ to zn-1, where n is the precursor charge).
2 Identifies corresponding ETD spectra of the same precursor mass 2. Identifies corresponding ETD spectra of the same precursor mass in the ETD MGF file to create a subset MGF file that only contains ETD spectra of potential glycopeptides.
Mass of the neutral loss, the threshold for how many of the most intense peaks are searched for the neutral loss peak and the intense peaks are searched for the neutral loss peak and the precursor retention time window for correlating ETD and CID spectra can all be altered.
Testing the filtering scriptTesting the filtering script
CID data from LC/MS/MS experiments of fractions f p f fenriched in GalNAc- or GlcNAc-modified peptides were screened for glycopeptides
~90% of the unmodified peptides were eliminated ~5% of the glycopeptides may be affected: either
eliminated or identified from a different charge eliminated or identified from a different charge state
m/z 203 07
Figure 4: False-Positive Glycopeptide IDd by the Script
Precursor 203 Da?
784.4 834.0707.21001.6667.1490.3 1083.9452.4 843.3258.2 287.1 344.3191.1 928.6 1174.5
200 300 400 500 600 700 800 900 1000 1100m/z
Unmodified peptides may produce glycopeptide-like CID fragmentation !
778.750% z6* z8*/c9*1556.9
Figure 5: False-Positive Glycopeptide ousted by the scripte
LFSLPYS(GlcNAc)RTRL vs. LFSLPAQPLWNNR
SummarySummary The altered ETD searches with Protein
Prospector yielded superior results from the very same dataI ti CID d t ith th h l f Incorporating CID data with the help of a glycopeptide-identifying script may boost the confidence level of glycopeptide confidence level of glycopeptide identifications
Linking different charge states of the same g gcomponent will be beneficial, since different charge states may produce the best CID and ETD dataETD data
1. O-linked N-acetylglucosamine proteomics of postsynaptic density preparations using lectin weak affinity chromatography and mass spectrometry. Vosseller K, Trinidad JC, Chalkley RJ, Specht CG, Thalhammer A, Lynn AJ, Snedecor JO, Guan S, Medzihradszky KF, M ltb DA S h f R B li AL M l C ll P t i 5 923 34 Maltby DA, Schoepfer R, Burlingame AL. Mol Cell Proteomics. 5, 923-34 (2006).
2 Affinity enrichment and characterization of mucin core 1 type 2. Affinity enrichment and characterization of mucin core-1 type glycopeptides from bovine serum. Darula Z, Medzihradszky KF. Mol Cell Proteomics. 8, 2515-26 (2009).
3. Identification of protein O-GlcNAcylation sites using electron transfer dissociation mass spectrometry on native peptides. Chalkley RJ, Thalhammer A, Schoepfer R, Burlingame AL. Proc Natl Acad Sci U S A., p , g106, 8894-9 (2009).
This work was supported by NIH NCRR grant P41RR001614 (to ALB) and Hungarian Science Foundation grants OTKA T60283 (to KFM)