Structural Analysis of Glycosylated Peptides in Complex Mixtures with
Ion Trap MSn
Shiaw-Lin Wu, Pavel Bondarenko, Tom Shaler,
Paul Shieh, and William S. Hancock
The data presented here can be acquired using
any Thermo Finnigan LCQTMion trap mass spectrometer with MSn
In humans, glycoproteins on cell surfaces are
important for communication between cells, maintaining cell structure and
self-recognition by the immune system. Of particular clinical concern is
how viruses, bacteria and parasites bind to cell-surface glycoproteins and
use them as portals of entry into cells. Understanding the detailed structure
of glycoproteins at the molecular level may provide insights to aid in combating
glycoprotein- mediated induction of disease.
A variety of mass spectrometry-based approaches have been applied to the
fundamental study of glycoproteins. Typically, a glycoprotein is enzymatically
digested, the resulting peptide fragments are separated using HPLC and the
peptides are identified by on-line MS analysis using electrospray ionization.
Alternatively, fractions may be collected after the HPLC separation and
analyzed off-line by matrix-assisted laser desorption ionization (MALDI)
or nanospray ionization.
Peptides that do not correspond to predicted masses may be present in a
glycosylated form. These putative glycopeptides are treated with a glycosidase
to cleave the bond between the peptide and the oligosaccharide. The difference
in mass following cleavage is used to infer the carbohyd
of the cleaved glycoform. Generally, neither the oligosaccharide structure
nor the exact site of attachment to the peptide can be determined. This
technique is labor-intensive, time-consuming, and it requires a large amount
of sample all of which severely limit its general utility.
In this report, we describe a novel, highthroughput
technique that will allow researchers to determine:
- The amino acid sequence of glycopeptides
- The exact site of attachment of the oligosaccharide
- Accurate, detailed structures of the attached oligosaccharide
Starting materials for this technique are tryptic
fragments of the protein of interest. Treatment with glycosidase is not
required, and very little sample is needed. Analysis can be accomplished
in a single, automated process using any LCQ ion trap mass spectrometer
with Data DependentTM
in this case up to MS4
and Dynamic ExclusionTM
to successively isolate, fragment,
and analyze the peptide and oligosaccharide structures. In this case, as
is the case for most glycoprotein analyses, MS/MS is simply not enough.
Thermo Finnigan LCQ Deca XP mass spectrometer
fitted with orthogonal ESI probe (positive ion mode)
Ion capillary tube temp: 140 C
Needle voltage: +3.8 kV
Sheath gas: 9 units
Normalized Collision Energy: 65%
Dynamic Exclusion duration: 3 min
Autosampler with Surveyor MS Pump: 120 L/min before
splitting, 1 L/minute after splitting
HPLC column: Microte
, 0.15 x 100 mm
HPLC gradient: 2% acetonitrile, 0.1% formic acid for 3 minutes, ramp to
60% acetonitrile in 90 minutes, ramp to 80% acetonitrile in 5 minutes, hold
at 80% for 20 minutes.
Recombinant human tissue plasminogen activator
(rt-PA) was expressed in Chinese hamster ovary cells and purified. This
glycoprotein has an approximate molecular weight of 64 kDa and a heterogeneous
distribution of N-linked glycoforms. The purified protein was reduced and
alkylated at two sites, then digested with trypsin. A 2-g aliquot of the
digest mixture was analyzed. Automated Data Dependent LC-MS/MS was used
to generate the chromatogram shown in Figure 1.
Figure 1. Base peak chromatogram from Data
Dependent LC-MS/MS of a rt-PA tryptic digest.
The product ion spectra from the tryptic digest
mixture were used by TurboSEQUEST software to search the human protein
database. Figure 2 shows that the protein was identified as tPA. Sequence
coverage is shown. The peptides highlighted in red were immediately identified
using TurboSEQUEST, representing 79% coverage of the protein. The peptides
labeled in blue were not initially identified, however each of these unknown
peptides contains an Asn-X-Ser/Thr motif (with the Asn labeled in green),
strongly suggesting that these peptides might be glycosylated.
N-linked oligosaccharides in human proteins are generally predictable but
cell lines may incorporate uncommon structures making direct confirmation
of complete structure with MSn
important. All have the same core structure
with extensions and branches composed prim
arily of sialic acid, galactose
Figure 3 designates the expected mass-to-charge ratios for the possible
forms of the glycosylated tryptic peptide T45 from rt-PA. This glycoform
was found to have a typical structure including one sialic acid, two galactoses
and two N-acetylglucosamines (GlcN) (labeled as structure 211).
Figure 2. Sequence of rt-PA. Observed peptides
labeled in red. Blue highlighted peptides correspond to glycoforms identified
after possible glycoform structures were manually added to the database.
Small peptides in black, representing 10% of the amino acids, were not observed.
Figure 3. Typical theoretical oligosaccharide
sequences for rt-PA peptide T45. Mass-to-charge ratios and charge states
for the heterogeneous structures are listed in parentheses. The unglycosylated
peptide was determined to have a molecular weight of 1131 amu.
Figure 4. Top panel Base peak ion chromatogram
of tryptic peptides from rt-PA.
Bottom panel Full-scan mass spectrum of the 211-T45 glycoform eluting
at 15.57 minutes.
Figure 5. MS/MS on the 211-T45 ion at m/z
1064. The glycopeptide fragmented into many different glycoforms. The 210-G,
GlcN ion m/z 1267.6 (+2) was selected for MS3 analysis.
Theoretical m/z values for all possible cleavage
products of the glycoform were added to the database to enable TurboSEQUEST
to automatically match glycopeptide structures. From these, TurboSEQUEST
able to identify the 211 form of the T45 peptide (211-T45) at a retention
time of 15.57 minutes. As would be expected from such a complex mixture,
a number of co-eluting ions were observed. However, the triply-charged 211-T45
ion at m/z 1064 was easily identified. The base peak ion chromatogram and
full-scan mass spectrum for the targeted glycopeptide are shown in Figure
The signal intensities for the doubly- and triplycharged glycopeptides (m/z
1595 and m/z 1064, respectively) were less intense than those for co-eluting
peptides because glycosylated peptides typically have lower ionization efficiencies
than unmodified peptides. The use of Dynamic Exclusion allowed for automated
analysis of these low-intensity glycopeptides. The LCQ Deca XP was operated
in Data Dependent Ion Tree mode, where the most abundant ion in the MS2
spectrum was selected for MS3
, and the most abundant MS3
product ion was
chosen for MS4
Figure 6. MS3 on the 210-G, GlcN
ion. The dominant ion in the spectrum, the T45 peptide +GlcN at m/z 1333,
was selected for MS4.
Figure 7. MS4 on the T45 +GlcN
ion at m/z 1333.
are presented in Figures 57. The MS2
generated a heterogeneous collection of glycosylated product ions. GlcN
was not easily cleaved from the peptide, so it serves as a marker to identify
the site of glycosylation. The CID energy was sufficient at the MS4
stage to begin fragmenting the T45 peptide and to confirm the peptide sequence.<
When all of the glycosylated peptides were identified and combined with
the unmodified peptides, the sequence coverage obtained in the experiment
was increased to greater than 90%.
Figure 8. MSn on the LCQ ion trap
mass spectrometer confirms predicted sequence, linkage site and glycostructure
of an rt-PA peptide. The precursor ion in glycostructure 211 was determined
using MS/MS. To determine oligosaccharide structure, glycostructure 210
was selected for further fragmentation using MS3. To further
confirm the linked peptide sequence and the site of attachment, ion m/z
= 1333 was trapped and fragmented further (MS4).
The Thermo Finnigan LCQ Deca XP ion trap mass
spectrometer coupled with TurboSEQUEST protein identification software is
an extremely powerful tool for the elucidation of post-translational modifications
of proteins. In this case, the combination of Data Dependent MSn
and Dynamic Exclusion have been used to quickly generate results which are
more comprehensive than those provided by traditional approaches to glycoprotein
analysis. While MS/MS can give limited oligosaccharide structure information,
can unambiguously confirm complete structure and the sites
of attachment. This is yet another example of why the LCQ is an essential
tool for every proteomics laboratory.
Source:Page: All 1 2 3 4 5 6 Related biology technology :1
. Structural Determination of Flavonoids Using MSn2
. Continuous-Elution Electrophoresis for Purification of the Baculovirus-Expressed Coronavirus Structural Proteins, Rev A3
. A Positive Selection Assay for Mutation Analysis in Big Blue
. Epitope-Tagging Vectors for Functional Analysis in Yeast5
. Map and Link Human Genetic Disorders with SSLP Analysis6
. CastAway Precast Gels for Rapid Automated DNA Sequence Analysis7
. Single-cell RT-PCR Analysis of Paramecium Primaurelia with the Eppendorf
cMaster RTplusPCR System8
. Use of Transcriptor Reverse Transcriptase
in Microarray Analysis9
. Increased Sensitivity in Microarray Analysis10
. Gene-Expression Analysis of TP, DPD, and TS Using
the LightCycler mRNA Quantification KitsPLUS11
. Agilent 2100 Bioanalyzer
Automated Analysis System