Limits...
Improved identification of O-linked glycopeptides from ETD data with optimized scoring for different charge states and cleavage specificities.

Darula Z, Chalkley RJ, Lynn A, Baker PR, Medzihradszky KF - Amino Acids (2010)

Bottom Line: Interpretation of the corresponding ETD data using Protein Prospector is also presented.We show that the improved scoring is more than doubled the glycopeptide assignments under very strict acceptance criteria.This study illustrates that "old" proteomic data may yield significant new information when re-interrogated with new, improved tools.

View Article: PubMed Central - PubMed

Affiliation: Proteomics Research Group, Biological Research Center, 62 Temesvari krt, Szeged, 6726, Hungary.

ABSTRACT
This article describes the effect of re-interrogation of electron-transfer dissociation (ETD) data with newly developed analytical tools. MS/MS-based characterization of O-linked glycopeptides is discussed using data acquired from a complex mixture of O-linked glycopeptides, featuring mucin core 1-type carbohydrates with and without sialic acid, as well as after partial deglycosylation to leave only the core GalNAc units (Darula and Medzihradszky in Mol Cell Proteomics 8:2515, 2009). Information content of collision-induced dissociation spectra generated in collision cell (in QqTOF instruments) and in ion traps is compared. Interpretation of the corresponding ETD data using Protein Prospector is also presented. Search results using scoring based on the frequency of different fragment ions occurring in ETD spectra of tryptic peptides are compared with results obtained after ion weightings were adjusted to accommodate differential ion frequencies in spectra of differing charge states or cleavage specificities. We show that the improved scoring is more than doubled the glycopeptide assignments under very strict acceptance criteria. This study illustrates that "old" proteomic data may yield significant new information when re-interrogated with new, improved tools.

Show MeSH

Related in: MedlinePlus

Linear ion trap CID spectrum of the same glycopeptide AVGAQVLESTPPPHVMR modified with SAGalGalNAc as in Fig. 1. Precursor ion was at m/z 815.7307(3+). The abundant peptide fragments in the spectrum were formed by cleavage N-terminal to the first Pro residue. The glycosylated half of the structure (b10) can be observed with the sugar attached as well as partially or completely deglycosylated—the number of asterisks indicates the number of carbohydrate units lost
© Copyright Policy
Related In: Results  -  Collection


getmorefigures.php?uid=PMC3102200&req=5

Fig2: Linear ion trap CID spectrum of the same glycopeptide AVGAQVLESTPPPHVMR modified with SAGalGalNAc as in Fig. 1. Precursor ion was at m/z 815.7307(3+). The abundant peptide fragments in the spectrum were formed by cleavage N-terminal to the first Pro residue. The glycosylated half of the structure (b10) can be observed with the sugar attached as well as partially or completely deglycosylated—the number of asterisks indicates the number of carbohydrate units lost

Mentions: As published earlier (Darula and Medzihradszky 2009), mucin core 1-type glycopeptides were enriched from bovine serum and analyzed by mass spectrometry. LC/MS/MS analysis of such a mixture using a QqTOF instrument (QTOF Premier, Waters) indicated the presence of more than 100 glycopeptides based on the presence of diagnostic carbohydrate fragment ions, but only a handful of these yielded sufficiently informative CID spectra to allow identification by automated database searching or even manual sequencing (e.g. Fig. 1). When the same mixtures were subjected to LC/MS/MS analysis in an LTQ-Orbitrap, CID spectra almost exclusively contained only carbohydrate fragments (Fig. 2). From the intact glycopeptide datasets discussed here only four glycopeptides produced sufficient peptide fragmentation to be identified using CID data for database searching. The rest of the CID data, although generally not good enough for peptide identification, were nevertheless useful for manual confirmation of the glycopeptide identification. For example, when the modified peptide contains a proline residue, which is frequently the case with O-glycosylation, then usually both halves are detected as illustrated in Fig. 2, with the glycosylated part frequently present at different degrees of deglycosylation. (Similar CID fragmentation is presented in Supplemental Figures 5A, 5B, and 10 in online resource, Supplementary material 1). The corresponding ETD spectra contained a lot more information; however, the necessary data interpretation tools were still being developed when we performed the previous analysis. In order to be able to identify modified peptides, database searching had to be restricted to bovine proteins within the Swiss Prot database, and very liberal acceptance criteria were applied, followed by manual analysis of all candidate modified spectra. We also experimented with permitting non-specific cleavages at both termini and multiple missed cleavages. Altogether, we reported the identification of 49 glycopeptides (Darula and Medzihradszky 2009).Fig. 1


Improved identification of O-linked glycopeptides from ETD data with optimized scoring for different charge states and cleavage specificities.

Darula Z, Chalkley RJ, Lynn A, Baker PR, Medzihradszky KF - Amino Acids (2010)

Linear ion trap CID spectrum of the same glycopeptide AVGAQVLESTPPPHVMR modified with SAGalGalNAc as in Fig. 1. Precursor ion was at m/z 815.7307(3+). The abundant peptide fragments in the spectrum were formed by cleavage N-terminal to the first Pro residue. The glycosylated half of the structure (b10) can be observed with the sugar attached as well as partially or completely deglycosylated—the number of asterisks indicates the number of carbohydrate units lost
© Copyright Policy
Related In: Results  -  Collection

Show All Figures
getmorefigures.php?uid=PMC3102200&req=5

Fig2: Linear ion trap CID spectrum of the same glycopeptide AVGAQVLESTPPPHVMR modified with SAGalGalNAc as in Fig. 1. Precursor ion was at m/z 815.7307(3+). The abundant peptide fragments in the spectrum were formed by cleavage N-terminal to the first Pro residue. The glycosylated half of the structure (b10) can be observed with the sugar attached as well as partially or completely deglycosylated—the number of asterisks indicates the number of carbohydrate units lost
Mentions: As published earlier (Darula and Medzihradszky 2009), mucin core 1-type glycopeptides were enriched from bovine serum and analyzed by mass spectrometry. LC/MS/MS analysis of such a mixture using a QqTOF instrument (QTOF Premier, Waters) indicated the presence of more than 100 glycopeptides based on the presence of diagnostic carbohydrate fragment ions, but only a handful of these yielded sufficiently informative CID spectra to allow identification by automated database searching or even manual sequencing (e.g. Fig. 1). When the same mixtures were subjected to LC/MS/MS analysis in an LTQ-Orbitrap, CID spectra almost exclusively contained only carbohydrate fragments (Fig. 2). From the intact glycopeptide datasets discussed here only four glycopeptides produced sufficient peptide fragmentation to be identified using CID data for database searching. The rest of the CID data, although generally not good enough for peptide identification, were nevertheless useful for manual confirmation of the glycopeptide identification. For example, when the modified peptide contains a proline residue, which is frequently the case with O-glycosylation, then usually both halves are detected as illustrated in Fig. 2, with the glycosylated part frequently present at different degrees of deglycosylation. (Similar CID fragmentation is presented in Supplemental Figures 5A, 5B, and 10 in online resource, Supplementary material 1). The corresponding ETD spectra contained a lot more information; however, the necessary data interpretation tools were still being developed when we performed the previous analysis. In order to be able to identify modified peptides, database searching had to be restricted to bovine proteins within the Swiss Prot database, and very liberal acceptance criteria were applied, followed by manual analysis of all candidate modified spectra. We also experimented with permitting non-specific cleavages at both termini and multiple missed cleavages. Altogether, we reported the identification of 49 glycopeptides (Darula and Medzihradszky 2009).Fig. 1

Bottom Line: Interpretation of the corresponding ETD data using Protein Prospector is also presented.We show that the improved scoring is more than doubled the glycopeptide assignments under very strict acceptance criteria.This study illustrates that "old" proteomic data may yield significant new information when re-interrogated with new, improved tools.

View Article: PubMed Central - PubMed

Affiliation: Proteomics Research Group, Biological Research Center, 62 Temesvari krt, Szeged, 6726, Hungary.

ABSTRACT
This article describes the effect of re-interrogation of electron-transfer dissociation (ETD) data with newly developed analytical tools. MS/MS-based characterization of O-linked glycopeptides is discussed using data acquired from a complex mixture of O-linked glycopeptides, featuring mucin core 1-type carbohydrates with and without sialic acid, as well as after partial deglycosylation to leave only the core GalNAc units (Darula and Medzihradszky in Mol Cell Proteomics 8:2515, 2009). Information content of collision-induced dissociation spectra generated in collision cell (in QqTOF instruments) and in ion traps is compared. Interpretation of the corresponding ETD data using Protein Prospector is also presented. Search results using scoring based on the frequency of different fragment ions occurring in ETD spectra of tryptic peptides are compared with results obtained after ion weightings were adjusted to accommodate differential ion frequencies in spectra of differing charge states or cleavage specificities. We show that the improved scoring is more than doubled the glycopeptide assignments under very strict acceptance criteria. This study illustrates that "old" proteomic data may yield significant new information when re-interrogated with new, improved tools.

Show MeSH
Related in: MedlinePlus