Glyco

PMCID: PMC4261947

Legend: Gene, Sites, Suger

Section : Qualitative Analyses ofGlycopeptides from PSA and PSA-HighpI Isoform

Content :

Prostate specific antigen in the samples analyzedhere was identified with sequence coverages of 85.6 and 86.5% from PSA and PSAH samples, respectively (data not shown)
Peptides containingthe glycosylation sequon (N69KS) were not detected in theproteomics analysis, thus suggesting that the N69 residueis glycosylated
Protein identifications in proteomics analysis werebased on an ion score cutoff higher than 30 and a minimum number ofpeptides of two
According to spectral count data generated by Scaffoldsoftware, spectral counts of PSA accounted for 86% of the total spectralcount data (15 identified peptides ), whereas the spectral counts ofPSAH accounted for 78% of the total spectral count data (17 peptides )
Other proteins that were detected in the case of the PSA sample were triosephosphate isomerase (6% of total spectral counts, three peptides ), prostaglandin-H2 d-isomerase (3% of the total spectral countdata, two peptides ), and growth/differentiation factor 15 (5% of totalspectral count data, two peptides )
Other proteins detected in thecase of the PSAH sample were triosephosphate isomerase (12% of thetotal spectral count data, eight peptides ), prosaposin (5% of thetotal spectral count data, two peptides ), prostaglandin-H2 d-isomerase (2% of the total spectral count data, two peptides ), andgrowth/differentiation factor 15 (3% of the total spectral count data,two peptides )
Furthermore, GlypID analysis using these proteins didnot identify any valid glycopeptides associated with above-mentioned proteins , yet they all have the glycosylation sequences
Therefore,all glycopeptides reported here are unique to PSA and are not originatingfrom the other proteins observed in the LC–MS/MS analysis
Additionally, the Y1 ions reported here are unique to PSA and arenot from any contaminant proteins observed in the LC–MS/MSanalysis
Since PSA is known to have only one N-linked glycosylationsite at Asn69 , it is adequate to interpret its glycosylations withCID or HCD tandem MS. If the CID or HCD spectra include glycan fragmentationor glycopeptide diagnostic ions in addition to the expected Y1 ion( peptide + GlcNAc) , then the glycopeptides associated with Asn69 wereconfirmed
For example, a Y1 ion associated with Asn69 detected atan m/z value of 602 or 601 correspondedto doubly charged NKSVILLGR + GlcNAc
Hexose rearrangement could occurduring CID tandem analyses
It is observed in lowabundance (<8%) when Y1 ion intensities are incredibly high
However,hexose rearrangement is not the case in this study because Y1 ionintensities associated with the PSA glycosylation site are not thathigh, as can be seen in CID MS/MS scans (SupportingInformation Figure 1)
In cases when the theoretical m/z values of glycopeptides were used,they were confirmed by matchingthe retention times of the glycopeptides that were confirmed by tandemMS
Using a 5 ppm cutoff mass accuracy allowed glycopeptides thathave similar molecular weights to be distinguished
For example, aglycopeptide possessing 2 Fuc residues is only 1 Da different fromthat possessing an NeuAc residue , which can be differentiated usingthis 5 ppm cutoff
Moreover, their retention time differences werealso used to distinguish them, since a glycopeptide containing NeuAcis retained longer by LC than a counterpart glycopeptide containing2 Fuc
Overall, the use of the aforementioned criteria results inan overall averaged mass accuracy of 3.05 ppm for all identified glycopeptides
In total, tryptic digestion of the samples resulted in the formationof three peptide backbones containing the glycosylation site , namely,NKSVILLGR (17–23 min), AVCGGVLVHPQWVLTAAHCIRNK(26–21 min), and AVCGGVLVHPQWVLTAAHCIRNKSVILLGR(31–36 min), as shown in Figure 1A
There are 56 glycopeptides observed for PSA , as summarized in Table 1
Because 55 of the structures were identified onthe NKSVILLGR backbone, the discussion will be mainly focused on thispeptide backbone
One missing glycoform associated with the NKSVILLGRbackbone is HexNAc3Hex3dHex1NeuAc1 , which was detected on theAVCGGVLVHPQWVLTAAHCIRNK backbone
Figure 1B–D illustrates averaged full MS with observedN-glycans detected on the NKSVILLGR (=P) backbone
There were 11 neutralN-glycans observed at 17.5 min (Figure 1B),36 monosialylated and 2 sulfated/phosphorylated N-glycans were observedat 19.5 min (Figure 1C), and 6 disialylatedN-glycans were detected at 22 min (Figure 1D)
The m/z values (charge state= 3) of these glycopeptides were listed with confirmed glycan structuresif they were supported by tandem MS data
Otherwise, the most likelystructures were described as previously reported
There were 32 N-glycansconfirmed by tandem MS data, as shown in Table 1
For PSAH, similar retention times were detected withdifferenttypes of N-glycans, as mentioned above (Figure 2A)
In total, 57 N-glycans were identified from three peptide backbones,as summarized in Table 2
All of the N-glycanswere detected on the NKSVILLGR backbone except HexNAc6Hex4dHex2NeuAc1 ,which was detected only on the AVCGGVLVHPQWVLTAAHCIRNKbackbone
Figure 2B–D depicts an averagedfull MS with identified N-glycans associated with the NKSVILLGR backbone
There were 15 neutral N-glycans observed at 17.5 min (Figure 2B), 29 monosialylated and 2 sulfated/phosphorylatedN-glycans observed at 19.5 min (Figure 2C),and 10 disialylated N-glycans detected at 22 min (Figure 2D)
Manual annotations of the CID MS/MS of 37 glycopeptidesare depicted in Supporting Information Figure1
Assigning glycan structure from CID MS/MS is achievablebecausethe mass difference between the glycan residues of glycopeptides ,such as Hex, HexNAc, Fuc, or NeuAc, is rather large
Also, sequencingN-glycan structures follows some rules because specific glycosyltransferases/exoglycosidasesare involved in trimming or attachment of each glycan residue duringthe biosynthesis of N-glycans
The precursor m/z values of glycopeptides were confirmed using a 5 ppm cutoffmass accuracy because the detection of precursor ions was achievedby an Orbitrap FTMS analyzer
The intriguing features of N-glycansassociated with PSA are thepresence of an GalNAc residue instead of Gal followed by GlcNAc andsulfation/phosphorylation
In this study, many N-glycanswith these characteristics were detected, some of which were confirmedby tandem MS, as depicted in Figure 3
Theglycan fragments of a glycopeptide with m/z 1034.1406 were assigned with diagnostic ions originatingfrom glycan residues (Figure 3A)
The annotationof this tandem MS demonstrates that this glycopeptide is core fucosylatedand monosialylated followed by either a Gal or GlcNAc residue
Thecore fucosylation was confirmed on the basis of the existence of aY1+2 ion (peptide backbone + GlcNAc) with fucose, whichhas an m/z value of 675.3
One ofthe fragment peaks with an m/z valueof 1303.95 represents the loss of a GalNAc residue
Also, we confirmedthat there are two co-eluting isomeric glycopeptides
In the low m/z region , diagnostic ions with m/z values of 495.16 and 698.35 confirmedGalNAc + NeuAc and GlcNAc + GalNAc + NeuAc glycan structures, respectively
In addition, the presence of a 657.19 m/z value affirms the presence of a GlcNAc \+ Gal \+ NeuAc lycan structure
Therefore, the sialylation followed by Gal and GalNAc residues co-eluted,as confirmed in this single tandem MS. Also, on the basis of the assignmentof fragment ions of glycans, we distinguished between the attachmentof GalNAc followed by antenna GlcNAc and bisecting structures
Whena series of HexNAc fragmentation were detected as associated withantenna GlcNAc without assignment of a Glc residue , the presence ofGalNAc on an antenna GlcNAc was confirmed
Regarding bisection structures,the presence of 2GlcNAc + Man + GlcNAc, 2GlcNAc + 2Man + GlcNAc, and2GlcNAc + 3Man + GlcNAc was detected, which correspond to m/z values of 886, 967, and 1048, respectively
Figure 3B illustrates annotated tandemMSof a sulfated/phosphorylated glycopeptide corresponding to m/z 963.7589
The sulfation/phosphorylationon a GlcNAc residue was determined on the basis of the detection of fragment ions at m/z 1303.87 and1344.45, the difference of which corresponds to a doubly charged sulfateor phosphate ion
Because the peak at m/z 1344.45 originated from the loss of a GalNAc residue , this suggeststhat the occurrence of sulfation/phosphorylation is on a GlcNAc residue
Also, the presence of m/z valuesof 1221.05 and 1262.04 supports the occurrence of sulfation/phosphorylationon a GlcNAc residue
The difference between these two assignmentsof sulfation/phosphorylation was 0.5 Da, which fits well within themass accuracy of CID MS/MS
On the other hand, no diagnostic ionsrepresenting a sulfated/phosphorylated glycan residue were observedfrom HCD MS/MS, as shown in Supporting InformationFigure 2
Overall, more of core-fucosylated and/or disialylatedglycoforms were identified in PSAH than in normal PSA
Moreover, highlybranched glycan structures such as tri- or tetra-antennary structureswere more abundant in PSAH than in normal PSA
These observationsare in agreement with previously published results
Glycoform differences between PSA and PSAH are illustrated in Supporting Information Figure 3
Therewas a difference in the number of identified N-glycans fromthe three peptide backbones, as shown in Tables 1 and 2
This might be due to the fact thatthe ionization efficiency of peptides becomes substantially higheras the mobile phase’s organic content increases
As a result,some of the glycopeptides with minor intensities would not be subjectedto tandem MS experiments
This observation is supported by the datashown in Supporting Information Figure 4
Quantitative results of all identified glycopeptides associatedwith the three peptide backbones are illustrated in Supporting Information Figure 4A,B for PSA and PSAH, respectively
Glycopeptides with relatively low intensities associated with theNKSVILLGR backbone were not detected in the other two peptides backbones,which is believed to be mainly due to competitive ionization

*Output_Site_Fusion* (sent_index, protein, sugar, site):

Section : QuantitativeAnalyses of Glycopeptides from PSA and PSA-HighpI Isoform

Content :

The quantitation of all identified glycopeptidesassociated with the three detected peptide backbones is summarizedin Tables 1 and 2
Averagepeak areas and standard deviation values ( STD ) were measured fromtechnical triplicates
Figure 4 showsthe quantitation results of N-glycans on the NKSVILLGR backbone
Thereare 66 common N-glycans with 68 total identified glycoforms that werequantitatively compared between PSA and PSAH samples
The abundancesof two glycopeptides associated with the peptides AVCGGVLVHPQWVLTAAHCIRNKand AVCGGVLVHPQWVLTAAHCIRNKSVILLGR were excludedbecause their ionization efficiencies might be different from thoseof other glycopeptides associated with the peptide NKSVILLGR
Figure 4B depicts a separate comparison of two glycoformsthat demonstrated a higher abundance for PSAH relative to that inall other structures
The most abundant glycoform for PSA was determinedto be HexNAc3Hex4dHex1NeuAc1 (13.3%)
Four glycopeptides possessingdifferent glycan structures, namely, HexNAc3Hex6NeuAc1 (9.3%), HexNAc4Hex4NeuAc1(7 .9%), HexNAc4Hex4dHex1NeuAc1 (6.8%), and HexNAc5Hex5NeuAc1(5 .7%) were detected as the next most abundant ions
On the otherhand, glycopeptides with HexNAc4Hex5dHex1NeuAc1 (28.9%) andHexNAc4Hex5dHex1NeuAc2 (27%) were observed at higher intensitiesin the PSAH sample
Glycopeptides possessing HexNAc5Hex5NeuAc1(6 .8%), HexNAc5Hex4dHex1NeuAc2 (6.6%), HexNAc5Hex5NeuAc2(5 .3%), and HexNAc5Hex4dHex1NeuAc1 (5.2%) glycans were observedas the next most abundant ions
The glycoform of HexNAc5Hex4dHex1s/pwas identified in both PSA and PSAH with different abundances (0.52and 0.28%, respectively)
An interesting observation related to thisstructure is that the summation of the relative abundances of HexNAc5Hex4dHex1and HexNAc5Hex4dHex1s/p from PSAH was determined to be comparablewith the intensity of HexNAc5Hex4dHex1s/p from PSA
On the other hand,HexNAc5Hex4dHex1 was not detected for PSA
The abundances of the 46 commonN-glycans were compared betweenthe PSA and PSAH samples
For example, the glycopeptides possessingHexNAc3Hex6NeuAc1 (>21.7-fold), HexNAc4Hex4NeuAc1 (>13.8-fold),HexNAc4Hex3dHex1NeuAc1(>8-fold), HexNAc3Hex4dHex1NeuAc1 (>6.5-fold), and HexNAc5Hex4NeuAc1(>5-fold) were detected with higher intensities in the PSA samplecompared with that of the PSAH sample
On the other hand, the glycopeptideswith HexNAc5Hex4dHex1NeuAc2 (>9.2-fold), HexNAc4Hex5dHex1NeuAc2(>4 .8-fold), and HexNAc4Hex5dHex1NeuAc1 (>3.2-fold) wereobservedwith higher abundances in PSAH than that in PSA

*Output_Site_Fusion* (sent_index, protein, sugar, site):

Section : Comparisonof the Data with the 2012 ABRF Data and a Data Setfrom a Laboratory Participating in the 2012 ABRF Study

Content :

Asone of the 26 global laboratories that participated in the 2012 ABRFGlycoprotein Research Group (gPRG) study (the 2012 ABRF study), the qualitative and quantitative results werecompared to evaluate our glycoproteomic methods
The goal of the 2012ABRF study was to provide a global overview of glycoproteomics methodsusing a consensus among interlaboratory data
As a result, 57 N-glycanswere reported in total
Twenty six data sets collected from the participatinglaboratories were classified into four clusters from A to D basedon data similarity
The major cluster is C, where the data were summarizedfrom 21 laboratories, including our laboratory
According to abundances,N-glycans were grouped into three categories, including major, intermediate,and minor abundance structures
A separate result has been publishedby Behnken et al
They were also an ABRF-participatinglaboratory that identified 42 glycoforms of intact PSA /PSAH glycoproteinusing HR– ESI /TOF–MS. As shown in Supporting Information Figure 5A, the three studies reported85 total glycoforms associated with PSA and PSAH, with 29 common glycoforms
Seventeen additional glycoforms were commonly identified between the2012 ABRF study and this study
Forty six glycoforms were observedboth in the ABRF study and here
Eleven glycoforms were detected onlyin the ABRF study
Twenty two structures were unique to this study
Three glycoforms were reported by Behnken et al. and were observed in our study but were not described inthe ABRF study
Our study and that by Behnken et al. study had 32 common glycoforms
Ten glycoforms were describedin Behnken et al.’s study but were not observed in our study,of which four were described in the ABRF study
This discrepancy maybe due to the use of different methods: bottom-up gycoproteomics inthis study and top-down glycoproteomics in the Behnken et al. study
As shown in SupportingInformation Figure 5B, the relative abundances of 18 major/intermediateN-glycans appeared to be comparable with that in the 2012 ABRF study
In particular, the N-glycans HexNAc3Hex4dHex1NeuAc1, HexNAc4Hex5dHex1NeuAc2,HexNAc5Hex3dHex1NeuAc1, and HexNAc5Hex4dHex1NeuAc1 wereobserved with a very high overlap in intensities
From minor N-glycans,we noticed that 6 of the 13 missing structures were sulfated/phosphorylatedN-glycans with very low intensities
One sulfated/phosphorylated glycoform,HexNAc5Hex4dHex1s/p , was reported in the 2012 ABRF study as beingless than 1%, but it is present in PSA at approximately twice theamount as that in PSAH
This particular glycoform was observed inour analysis and with a similar intensity profile variation
Behnken et al. reported a single analysisof quantitation; thus, no variation was inserted (Supporting Information Figure 5C)
Among the 32 common glycoformsthat were observed in our study, most of them showed a comparabletrend of intensities between PSA and PSAH
With regard to the 18 major/intermediateglycoforms defined by the 2012 ABRF study, 16 glycoforms were observedby Behnken et al., whereas 2 major/intermediateglycoforms were not reported, including HexNAc2Hex5 and HexNAc4Hex4dHex1.
Most of the major/intermediate glycoforms show comparable abundanceprofiles except for HexNAc4Hex5dHex1NeuAc2 , in which its abundancein PSA appeared to be 4.60-fold lower when comparing this study andthe 2012 ABRF study

*Output_Site_Fusion* (sent_index, protein, sugar, site):

Section : Total ion chromatogram of PSA digests with different retentiontimes of three peptide backbones possessing the Asn69 glycosylationsite (A)

Content :

Assigned glycoforms in MS of PSA are presented with theNKSVILLGR backbone
Different retention times correspond to the detectionof neutral glycoforms (B), sulfated/phosphorylated (s/p) neutral glycoformsand monosialylated glycoforms (C), and disialylated glycoforms (D)

*Output_Site_Fusion* (sent_index, protein, sugar, site):

Section : Total ion chromatogram of PSAH digests with different retentiontimes of three peptide backbones possessing the Asn69 glycosylationsite (A)

Content :

Assigned glycoforms in MS of PSAH are presented with theNKSVILLGR backbone
Different retention times correspond to the detectionof neutral glycoforms (B), sulfated/phosphorylated (s/p) neutral glycoformsand monosialylated glycoforms (C), and disialylated glycoforms (D)

*Output_Site_Fusion* (sent_index, protein, sugar, site):

Section : Two examplesof identified glycoforms whose structures were confirmedby CID MS/MS

Content :

The glycan fragments of HexNAc5Hex4dHex1NeuAc1glycopeptide with m/z 1034.1406(A) support the presence of a GalNAc residue followed by a GlcNAcresidue on its antenna
Diagnostic ions highlighted in red representthe presence of a GalNAc residue
The annotation of a sulfated/phosphorylated(s/p) glycopeptide , determined as HexNAc5Hex4dHex1s/p1 glycopeptidewith m/z 963.7589 (B), suggeststhe occurrence of sulfation/phosphorylation on the GlcNAc residuebefore the GalNAc residue

*Output_Site_Fusion* (sent_index, protein, sugar, site):

Section : Quantitative values (relativeabundance) of N-glycans on the NKSVILLGRbackbone

Content :

Sixty six total N-glycans were quantitatively compared betweenPSA and PSAH
(A) Relative abundances of 64 glycoforms
(B) Relativeabundances of two glycoforms, which shows low-abundance glycoformsfor PSAH by separating the extremely high abundances of HexNAc4Hex5dHex1NeuAc1and HexNAc4Hex5dHex1NeuAc2 in PSAH
An image representing aglycoform was added to the side for which higher quantities were observed

*Output_Site_Fusion* (sent_index, protein, sugar, site):

Protein	NCBI ID	SENTENCE INDEX