Glycobiology Advance Access originally published online on October 26, 2006
Glycobiology 2007 17(2):141-156; doi:10.1093/glycob/cwl063
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
New insights in Rapana venosa hemocyanin N-glycosylation resulting from on-line mass spectrometric analyses
3 Laboratory of Protein Biochemistry and Protein Engineering, Ghent University, K.L. Ledeganckstraat 35, B-9000 Ghent, Belgium
4 Institute of Organic Chemistry, Bulgarian Academy of Sciences, 1113 Sofia, Bulgaria
1 To whom correspondence should be addressed; Tel: +32 9 264 5109; Fax: +32 9 264 5338; e-mail: jozef.vanbeeumen{at}ugent.be
Received on April 27, 2006; revised on October 16, 2006; accepted on October 20, 2006
 |
Abstract |
|---|
 Top Abstract IntroductionResultsDiscussionMaterials and methodsSupplementary dataAcknowledgmentsReferences |
|---|
The N-glycosylation of structural unit 1 of Rapana venosa hemocyanin was studied. Enzymatically liberated N-glycans were analyzed by matrix-assisted laser desorption ionization-time-of-flight-mass spectrometry (MALDI-TOF-MS) and capillary electrophoresis (CE)-MS following 8-aminopyrene-1,3,6-trisulfonate labeling and labeling with 3-aminopyrazole, a new dedicated sugar reagent. Structural information was obtained by exoglycosidase sequencing, on-line MS/MS, permethylation, and amidation. A mixture of high-mannose and complex glycans with so far unknown and unusual acidic terminal structures was revealed. As the hemocyanin protein sequence is currently unknown, de novo sequencing of the glycopeptides had to be carried out. The N-glycans were therefore enzymatically removed with simultaneous partial (50%) 18O-labeling of glycosylated asparagine residues prior to proteolysis. Following nano-liquid chromatography-MALDI-TOF-MS, the originally glycosylated peptides could be revealed and their sequences determined by MS/MS. The site occupancies were subsequently elucidated by precursor ion scanning of the intact glycopeptides using a Q-Trap mass spectrometer.
Key words: capillary electrophoresis-mass spectrometry / glycosylation / hemocyanin / MALDI-TOF / TOF / Rapana venosa
|
Introduction |
|---|
|
TopAbstractIntroductionResultsDiscussionMaterials and methodsSupplementary dataAcknowledgmentsReferences |
|---|
Hemocyanins (Hcs) are high-molecular-mass oxygen-transporting proteins, freely dissolved in the hemolymph of several arthropods and molluscs. Arthropodan Hcs occur as hexamers or multiples of hexamers (up to 8 x 6-mers) of approximately 75 kDa subunits, each containing a Cu(I) pair able to reversible bind dioxygen. Molluscan Hcs, in contrast, are decamers or di-decamers of approximately 350–450 kDa subunits, which each are folded into seven or eight globular functional units (FUs) of approximately 50 kDa. These FUs contain one dioxygen-binding Cu(I) pair and are assigned by the letters a–g (h) starting from the N-terminus. The (di)decameric structure has the shape of a hollow cylinder (Van Holde and Miller 1995
, Van Holde et al. 2001).
There are large differences in the carbohydrate content and monosaccharide composition of Hcs from arthropods and molluscs (Van Kuik et al. 1990). The carbohydrate content of arthropodan Hcs is relatively low (0.1–2%, w/w). Molluscan Hcs usually have a higher carbohydrate content (2–9%, w/w) and may contain unusual monosaccharides. The terrestrial snail Helix pomatia contains, besides the commonly occurring monosaccharides Man, Gal, Fuc, GlcNAc, and GalNAc, also Xyl and 3- and 4-O-methyl-Gal in their N-glycans (Lommerse et al. 1997). The carbohydrate moiety of molluscan Hcs has recently received particular interest for its immunostimulatory properties. The occurrence of Xyl in H. pomatia Hc is considered to be highly immunogenic in mammalian species. Keyhole limpet (Megathura crenulata) hemocyanin (KLH), which is used as hapten carrier and immunostimulant in the treatment of bladder cancer and other cancers, presumably carries, among others, an epitope consisting of an oligosaccharide with a terminal Gal ß-1,3 linked to GalNAc, essential for its efficacy as an immunotherapeutic agent (Harris and Markl 1999, 2000; Kurokawa et al. 2002; Wuhrer et al. 2004).
In the present work, the N-glycosylation of Hc isolated from the marine snail Rapana venosa (RvH), previously referred to as R. thomasiana, has been studied. This Hc consists of two structural subunits, RvH1 and RvH2 (± 400 kDa), each composed of eight FUs (± 50 kDa) and arranged as two independent homo-oligomeric isoforms (Stoeva et al. 1999). It has been shown that FU RvH1-f and RvH1-a, respectively, contain one and two N-glycosylation site(s) with carbohydrate chains of the complex type (Stoeva et al. 1997; Dolashka-Angelova, Beck et al. 2003; Dolashka-Angelova et al. 2004). Two potential N-glycosylation sites were shown for FU RvH2-a, although the occupancy was not studied. FUs RvH2-b and RvH2-d both possess only one N-glycosylation site, carrying glycans of the high-mannose type (Idakieva et al. 2004). FU RvH2-c is devoid of sugars (Idakieva et al. 2001). Stoeva et al. (2002) reported on the occupancy of Asn 127 of RvH2-e by the high-mannose structure Man5GlcNAc2. The attachment of other structures to Asn 127 was shown by Gielens et al. (2005): besides the core structure for N-glycosylation, Man3GlcNAc2, they reported on more complex glycans having internal fucose residues substituted with 3-O-methyl-Gal and GalNAc.
As RvH1 glycosylation is relatively unexplored, and the structures determined thus far are different from those observed on RvH2, we undertook a further study of the RvH1 N-glycan structures and the occupied N-glycosylation sites.
A completely different strategy was applied in comparison to the above-mentioned reports, which described work on tryptic digests of isolated FUs. Glycopeptides present in the FU hydrolysates were enriched by affinity chromatography using concanavalin A (Gielens et al. 2005) or by reversed-phase liquid chromatography with carbohydrate-specific orcinol staining of the collected fractions (Dolashka-Angelova, Beck et al. 2003; Dolashka-Angelova et al. 2004; Idakieva et al. 2004). Structural information on the glycans was obtained by studying the glycopeptides in a mass spectrometry (MS) experiment (either by MS/MS or exoglycosidase sequencing). Following deglycosylation, the peptide part was characterized by Edman N-terminal sequencing (Dolashka-Angelova, Beck et al. 2003; Dolashka-Angelova et al 2004; Idakieva et al. 2004) or by higher order MS (Gielens et al. 2005). The strategy that we followed for the characterization of RvH1 N-glycosylation is presented in Figure 1. The intact RvH1 was subjected to peptide-N-glycosidase F (PNGase F) digestion and/or tryptic digestion. The N-glycans were analyzed by matrix-assisted laser desorption ionization-time-of-flight (MALDI-TOF) and capillary electrophoresis (CE)-MS/MS following 8-aminopyrene-1,3,6-trisulfonate (APTS) as well as 3-aminopyrazole (3-AP) labeling. The former label has recently been proven to be of great value in the CE-MS/MS analysis of glycans (Sandra, Van Beeumen et al. 2004). It permits high-resolution CE, allows the simultaneous detection of uncharged and charged glycans, and provides easily interpretable spectra. The latter label, representing a novel sugar tag, allows CE-MS detection in the positive ion mode and provides complementary information.
|
As the protein sequence of RvH1 is currently unknown, de novo MS sequencing had to be performed on the glycopeptides. This is difficult to perform on the intact glycopeptides because the sugar fragments often dominate the peptide fragments due to glycosidic bonds being more labile than peptide bonds. This phenomenon is typically observed when using collision-induced dissociation (CID) or post-source decay as fragmentation techniques. Some approaches have been developed to overcome this problem, mainly based on multistage MS (Bateman et al. 1998
; Demelbauer et al. 2004), electron-capture dissociation (Håkansson et al. 2001, 2003) or electron-transfer dissociation (Hogan et al. 2005). In the present study, the glyco-moiety was removed from the glycopeptide, greatly facilitating de novo sequencing using CID. An important issue here is how to preserve information at the site of glycosylation for identification by mass spectrometry. Some time ago, it has been shown that N-glycosylation sites can be labeled with 18O during enzymatic deglycosylation using PNGase F (Gonzales et al. 1992). Therefore, prior to proteolysis, we partially (50%) labeled N-glycosylation sites with 18O by performing PNGase F digestions in buffers containing 50% H218O. N-glycosylated peptides can be detected using MS by the 2 Da spacings between the monoisotopic peaks of unlabeled and labeled peptides. Following nano-liquid chromatography (LC) with on-target fraction collection, MALDI-TOF was used to screen for the 18O-labeled peptides. De novo sequencing was performed using either MALDI-TOF/TOF or nano-LC-Q-Trap-MS. The specific Q-Trap scan function of time-delayed fragmentation (TDF) proved to be of great value for the de novo sequencing (Hager 2003). On the basis of the known masses of the 18O-labeled peptides, the glycan occupancy could easily be determined from the precursor ion scanning experiment performed on the intact glycopeptides using the Q-Trap MS, as described before (Sandra, Devreese et al. 2004). The strategy as presented can be applied to unravel the glycosylation puzzle of proteins in general.
|
Results |
|---|
|
TopAbstractIntroductionResultsDiscussionMaterials and methodsSupplementary dataAcknowledgmentsReferences |
|---|
Analysis of the N-glycans
Compositional analysis
The MALDI-TOF mass spectrum of the underivatized RvH1 N-glycans is presented in Figure 2A. Whereas compositional information can be deduced from the m/z values, partial structural information can be obtained based upon exoglycosidase sequencing. The peaks at m/z 1905.61 (peak 20) and 1743.55 (16) disappear upon
-1,2-mannosidase treatment, whereas the intensities of the peaks at the m/z values 1581.51 (12) and 1419.47 (10) are strongly reduced and the intensity of the peak at m/z 1257.41 (7) increases (Figure 2B). The corresponding compounds are thus predominantly of the high-mannose type. Prolonged incubation did not result in the complete removal of the peaks 12 and 10. Although the presence of a Gal residue on those structures is possible, subsequent treatment with bovine testes ß-galactosidase, however, did not result in the disappearance of the peaks. Also, the other glycans were unreactive towards ß-galactosidase. On the contrary, and as presented in Figure 2C, a substantial number of peaks disappear upon -1,2,3,4,6-fucosidase treatment. The peaks at m/z 1661.55 (14) and 1864.62 (19) appear to be unreactive, although, based upon the molecular weight information, these structures are supposed to contain a fucose residue. As demonstrated in Structural analysis of glycans by CE-MS/MS, the fucose in those structures is internally located. Table I presents the proposed compositions for the RvH1 N-glycans, for which more structural information will now be given.
|
|
Structural analysis of glycans by CE-MS/MS
CE-MS/MS was performed to provide additional structural information. As described before, performing MS/MS on APTS-derivatized sugars results in easily interpretable spectra, as Y-fragments predominate (Sandra, Van Beeumen et al. 2004
). The CE-MS electropherogram of the APTS-labeled RvH1 N-glycans is presented in Figure 3. By way of an example, the on-line MS and MS/MS data of the compound migrating at 10.0 min are added as Supplementary data (Figure S1). This particular glycan is detected as a doubly and triply negatively charged ion and the composition corresponds to a fucosylated Man3GlcNAc2 core structure. From the MS/MS data obtained on the [M-2H]2– ion at m/z 747.8, one can deduce that the fucose is attached to the proximal GlcNAc.
|
The CE-MS(/MS) data also revealed the presence of unusual structures for which we now present the logics behind their discovery. Because MS detection is performed at the cathodic side, the electro-osmotic flow (EOF) is mandatory to drive the negatively charged sugars towards the detector. Consequently, larger species migrate faster than smaller ones, and charged oligosaccharides will migrate slower than their neutral counterparts. For example, compound 17 (Table I), detected at m/z 555.7 as a 4-fold negatively charged ion, migrates slower than compound 7, which is smaller and detected at m/z 557.8 as a 3 times negatively charged ion. It is thus logical to conclude that the ion at m/z 555.7 corresponds to a charged N-glycan. A sulfated oligosaccharide was proposed to occur in RvH1 (Dolashka-Angelova, Beck et al. 2003
), and for the ion at m/z 555.7, the following sulfated structures can indeed be considered: SO4FucHex4Man3GlcNAc2 or (SO4)2 FucHexHexNAc2Man3GlcNAc2. In order to distinguish these two possibilities, the sample was reinjected for an on-line MS/MS experiment, keeping Q1 fixed to select the ion at m/z 555.7 (Figure 4). Apparently, CE could separate two isomeric compounds (10.5 and 10.9 min) with similar tandem mass spectra, and hence, the difference does not reside in the sequence. The spectra do not fit any of the sulfated glycans mentioned earlier. Instead, a complex glycan with an internal fucose should be proposed, which itself is decorated with a hexuronic acid (HexA) and an N-acetylhexosamine (HexNAc). This trisaccharide, which is linked to a GlcNAc residue commonly found in N-glycans, represents a novel N-glycan motif. The structural difference between the isomers can then be attributed to the presence of this unusual unit on either the -1,3 or the -1,6-arm of the trimannosyl-chitobiose core. On the basis of on-line MS/MS data, and taking the electrophoretic migration into account, the presence of HexA and an internal fucose can also be proposed for the [M-4H]4– ions at m/z 519.2, 569.8, and 606.5 (Table I). The structure at m/z 606.5 again displayed two clearly resolved isomers with identical spectra.
|
Underivatized acidic glycans are not observed in the EOF peak at time 4.41 min (Figure 3). Instead, they migrate later (5.36 min), a behavior typical for charged glycans, as reasoned earlier. At 6.08 min, another population (low intensity) of unlabeled glycans is observed and their masses correspond to complex structures carrying even two HexA residues (Supplementary Figure S2). No MS/MS data could be obtained for these glycans (only present in trace amounts). Structures with the HexAHexNAcFucGlcNAc oligosaccharide at both the
-1,3 and the -1,6-arms seem possible, but so far this only remains a hypothesis.
Confirmation of the presence of the HexA residue by CE-MS/MS, amidation, and permethylation
As these unprecedented acidic structures focussed our attention, we carried out additional experiments. The glycans were therefore labeled with a novel positively charged tag, 3-AP, and were subjected to CE-MS and -MS/MS. When compared with the APTS-labeled sugars, which are drawn by the EOF towards the mass spectrometer, the 3-AP-labeled glycans move ahead of the EOF peak. The smaller species will migrate faster than the larger ones, and the negatively charged glycans will migrate more slowly than their neutral counterparts. The base peak electropherograms of a 3-AP-labeled oligosaccharide standard mixture and of the RvH1 N-glycans are shown in Figure 5.
|
The acidic sugars migrate as two clearly resolved populations (14.87 and 15.59 min). This is further illustrated by the selected ion electropherograms of the [M + 2H]2+ ions at m/z 853.7 and 1028.5 (Figure 5, inset), corresponding to the derivatives of the structures FucHexAHexNAcGlcNAcMan3GlcNAc2 (compound 14) and Fuc2HexAHexNAcGlcNAc2Man3GlcNAc2 (compound 21), respectively. As was stated before, the structural differences between the isomeric species most likely results from the substitution of either the
-1,3 or the -1,6-arm of the trimannosyl-chitobiose core with the unusual acidic tetrasaccharide. The MS/MS data of the isomers are indeed identical. The tandem mass spectra of two double-charged acidic glycans, namely FucHexAHexNAcGlcNAcMan3GlcNAc2 (compound 14) and Fuc2HexAHexNAcGlcNAcMan3GlcNAc2 (compound 17), at m/z of 853.7 and 926.8, respectively, are presented in Figure 6. In the former spectrum, the loss of Fuc is only observed after the loss of a HexNAc and a HexA, allowing to conclude that it does not have a terminal location. From the MS/MS data of the second species, it can be determined that one Fuc is internally located, whereas the other is positioned at the proximal GlcNAc (see the intense Y1 ion at m/z 435.1 and the Y6ß/Y1
ion at m/z 1503.2). Such informative spectra can never be obtained when performing MS/MS on underivatized glycans.
|
Further evidence on the presence of the HexA was obtained following amidation and permethylation of the glycan mixture. Tanaka and coworkers recently described a derivatization method for stabilizing sialic acids in MALDI-MS (Sekiya et al. 2005
). The method involves the modification of carboxyl groups by converting the carboxylic acid to the amide in the presence of both an amine constituent (NH4Cl) and a condensing agent [4-(4,6-dimethoxy-1,3,5-triazin-2yl)-4-methylmorpholinium chloride (DMT-MM)]. We implemented this strategy for the unambiguous identification of the HexA moiety on the RvH1 N-glycans. The use of ammonia as a source of the amine functionality results in a mass decrease of 0.984 Da per carboxylic acid. The MALDI mass spectrum of the RvH1 N-glycans following amidation is presented in Figure 7 (compare with the inset of Figure 2A). A mass decrease is observed for structures 14, 17, 19, and 21, indicating that these glycans do carry a HexA. As expected, the other glycans were unreactive towards amidation.
|
The RvH1 glycans were further analyzed by MALDI-MS following permethylation and the measured masses are presented in Table I. It is once more demonstrated that glycans 14, 17, 19, and 21 carry a HexA. When permethylated, HexA can be discriminated from methyl-hexose by a 14 Da mass increase. Performing MS/MS on these permethylated glycans, using the Q-Trap mass spectrometer, resulted in the appearance of two very informative ions at m/z 660.3 and 919.4, corresponding to the B3/Y6ß- and B3-ions, respectively, further confirming the existence of the unusual tetrasaccharide (data not shown). The MS3 spectrum of such a permethylated B3-ion is presented in Figure 8. It is clear that the fucose residue is internally located, and branched, because this moiety is lost with a mass of 160 Da. This means that only one hydroxyl group is methylated and that three of them are involved in a linkage. The presence of the 3,5A3-ion (and the absence of a 0,4A3-ion) further led us to conclude that the Fuc is 1,4 linked to the GlcNAc.
|
Characterization of glycosylation sites and determination of the occupancy at a particular site
Characterization through 18O labeling
We carried out the method of labeling the N-glycosylation sites by performing a PNGase F digestion in a buffer containing 50% H218O. The peptides containing the glycosylation sites can then be detected via the 2 Da spacing between the unlabeled and labeled ions using MS, and sequence information can be obtained by performing higher-order MS. The 18O-labeling step can be performed either prior or following trypsinolysis. When choosing the former approach, it is important to remove the remaining H218O to prevent unspecific incorporation of the label into all tryptic peptides. In the latter case, trypsin needs to be removed from the hydrolysate. As H218O can easily be removed by membrane filtration, we chose the method of labeling the protein prior to the tryptic digestion.
The approach applied to RvH1 resulted in a number of glycopeptide candidates of which six could unambiguously be identified as such. The enormous complexity of the RvH1 tryptic digest (hundreds of peptides) required a preceding separation step. Nano-LC was used, and fractions were directly collected onto a MALDI-target. The complexity of the tryptic peptide mixture is reflected in the nano-LC-UV chromatogram which is added as Supplementary data (S3). The MALDI-TOF spectra corresponding to fractions 30 and 54, and an expanded view of the regions containing the 18O-labeled peptide ions at m/z 906.5460 (spot 30) and 2406.3296 (spot 54), are presented in Figure 9. The former peptide could easily be identified as being 18O-labeled. This is more difficult in the latter case because the contribution of the other isotopes increases with m/z. However, upon comparing this peptide with other peptides in the same m/z-region, it could be identified as 18O-labeled.
|
The MALDI-TOF/TOF spectrum of the ion at m/z 906.5460 is presented in Figure 10. The y- and b-ions that contain the glycosylation site occur as doublets with the characteristic 2 Da spacings. If the peptide would contain more than one potential N-glycosylation site (consensus sequence Asn-Xxx-Ser/Thr), the presence of these doublets would make it straightforward to identify the actual glycosylation site (see below). An important aspect to be considered is that, upon deglycosylation, Asn is converted into Asp. This results in a mass increase of 1 Da for the peptide and for the y- and b-ions containing the glycosylation site. The consensus sequence could thus be identified as NI/LT. The N-terminal amino acids could not be unambiguously elucidated, although the possible combinations are limited to LS, IS, AE, or VT.
|
Despite the fact that some strategies have been developed (Samyn et al. 2004
), de novo sequencing using the TOF/TOF instrument remains difficult. In cases where TOF/TOF data were insufficient, the 18O-labeled peptides were subjected to higher-order MS using nano-LC- electrospray ionization (ESI)-Q-Trap. TDF proved to be powerful because it simplified tandem mass spectra, especially in the low m/z region. The TDF mass spectrum of the double-charged 18O-labeled peptide at m/z 966.0 is shown in Figure 11. Almost a complete y-ion series is observed and the consensus sequence for N-glycosylation is NGT. The characterized RvH1 18O-labeled peptides are presented in Table II.
|
|
For the peptides at m/z 2409.34 and 2419.35, two consensus sequences are possible, namely, DDT or DTS. Because the b3-ion (m/z 391.2) appears as a doublet in the TDF spectra, it could be elucidated that the former sequence is the one in which the glycan is attached. This could also be deduced from the tandem mass spectra of the intact glycopeptides (see what follows) where the b3-ion appears at m/z 390.2, i.e. 1 Da lower due to the Asn/Asp deamidation. A drawback of the labeling approach is that it may be of limited value if glycosylation sites are only partially occupied. This could make it hard to ‘fish out’ the labeled peptides, as they would reside in the isotopes of the non-glycosylated peptide.
Occupancy of particular glycosylation sites
To reach this goal, we decided to perform precursor ion scanning using the nano-LC-MS approach described before (Sandra, Devreese et al. 2004). Occupancy can be revealed by combining the known peptide masses (Table II) with the known glycan masses (Table I). Additional information is derived from MS/MS data on the glycopeptides which are generated ‘on-the-fly,’ using information-dependent acquisition.
The precursor ion scan (monitoring m/z 204 for HexNAc+) at time 23 min during chromatography, the enhanced resolution (ER) scan of the most dominant ion (973), and the MS/MS data of that ion are presented in Figure 12. The ER scan was incorporated to determine the charge state and to improve mass accuracy. The glycopeptide appeared to be 2 times charged. By substracting the peptide masses (Table II) from that of the molecular ion [(972.3 x 2)–1 = 1943.6], the glycan mass and structure could be derived. The FucMan3GlcNAc2 glycan apparently occupies the NI/LT glycosylation site. This was confirmed by the tandem mass spectrum. The ion at m/z 905.2 corresponds to the peptide without the glycan. The m/z value differs by 1 Da compared with the peptide detected via PNGase F digestion and 18O-labeling (Table II and Figure 9A). The ions at m/z 1108.3, 1254.2, 1311.1, 1457.3 and 1473.3 correspond to the peptide with GlcNAc, FucGlcNAc, GlcNAcGlcNAc, FucGlcNAcGlcNAc and ManGlcNAcGlcNAc attached, respectively. Some y- and b-ions are observed. Again, the ions that contain the glycosylation site differ by 1 Da in comparison with the fragments measured on the 18O-labeled peptide. From the precursor ion scan spectrum, additional occupancies were revealed (Table III). Some of the glycopeptides were only present in trace amounts. Although precursor ion scanning at m/z 204 is an extremely useful tool, it gave false-positive results in some cases due to the limited resolution of the quadrupole. These signals could easily be excluded by examining the MS/MS data for other oxonium ions (e.g. m/z 163 for Hex+, 366 for HexHexNAc+, etc.) or by performing precursor ion scanning experiments on those oxonium ions (data not shown).
|
|
Additional glycopeptides, for which no corresponding 18O-labeled peptides could be detected, were revealed by precursor ion scanning. Because the oxonium ions dominated the tandem mass spectrum, only minor sequence information could be extracted for one of them.
Especially notable in Table III are the glycopeptides containing acidic glycans. The tandem mass spectra of these structures displayed the typical oxonium ions at m/z 526 (HexAHexNAcFuc) and 729 (HexAHexNAc2Fuc). Both these ions can be used in a precursor ion scanning experiment for the selective detection of glycopeptides containing the HexA. However, it is the latter ion that is preferred, because the ion at m/z 528 (Hex2HexNAc) might interfere with the detection of the oxonium ion at m/z 526.
We draw the attention to the fact that performing a tryptic digest on a glycosylated protein may result in partial proteolysis. Therefore, it might sometimes be difficult to correlate glycopeptides characterized through labeling with those measured during precursor ion scanning. However, in the present work, the detected 18O-labeled peptides could readily be attributed to the glycopeptides measured using precursor ion scanning.
|
Discussion |
|---|
|
TopAbstractIntroductionResultsDiscussionMaterials and methodsSupplementary dataAcknowledgmentsReferences |
|---|
Although an initial analysis of the glycosylation of the FUs a and f in the structural unit RvH1 from RvH has been published (Dolashka-Angelova, Beck et al. 2003
; Dolashka-Angelova et al. 2004), we now carried out a detailed investigation of the glycosylation of intact RvH1, making use of state-of-the-art mass spectrometric techniques coupled to nano-LC and electrophoretic separations.
A global view on the composition of the N-glycans linked to the several different FUs was obtained by digestion of RvH1 with PNGase, followed by MALDI-TOF-MS of the mixture in the absence and presence of -1,2-mannosidase and -1,2,3,4,6-fucosidase. This revealed the existence of typical high-mannose and complex glycans, and it was shown that some expected fucose residues were not cleaved off, most likely because of being internally located. This preliminary conclusion was proved to be correct by subsequent labeling experiments. Labeling with APTS and 3-AP, a novel reagent that we here introduced, followed by CE-MS(/MS) analysis revealed the presence of a novel N-glycan motif composed of an internal fucose residue being linked to a HexA and a HexNAc residue. This trisaccharide itself is linked, via the fucose residue, to a GlcNAc residue commonly found in complex N-glycans. Previously, Gielens et al. (2005) presented a structure in RvH2 where fucose is decorated with a methyl-galactose and a N-acetylgalactosamine. Although the mass difference between methyl-hexoses and HexA s is only 0.036 Da, we exclude the possibility of a methylated structure, as it is not supposed to behave as a charged sugar during electrophoresis. It is noteworthy to mention that CE-MS further revealed that some of the glycans carrying this unusual acidic motif occur as isomers.
Further proof for the presence of the HexA was obtained by amidation and permethylation. MSn of those permethylated structures confirmed that the fucose residue is internally located, branched, and linked in a 1,4 linkage to GlcNAc.
The unusual acidic tetrasaccharide described represents a novel N-glycan motif. Fucose residues are frequently observed on N-glycans and have been detected before on molluscan Hcs (Lommerse et al. 1997; Kurokawa et al. 2002; Gielens et al. 2005). They are usually -1,6-linked to the proximal GlcNAc. Recently, Wuhrer et al. (2004) reported on the 4-substitution of this core fucose with Gal residues on KLH. The Gal(ß-1,4)Gal(ß-1,4)Fuc(-1,6) substitution very likely contributes to the immunostimulatory properties of KLH. The structure presented in the present study is of a completely different origin. The fucose, which is attached to the GlcNAc residue located at the non-reducing end, is substituted at two positions with a HexNAc and a HexA. HexA moieties, which are common constituents of proteoglycans, occur rather rarely in glycoproteins. Nevertheless, they have been observed on mammalian N-glycans as constituent of the HNK-1 carbohydrate epitope (SO4-3GlcAß1-3Galß1-4GlcNAc) (Jacques et al. 1996; Voshol et al. 1996; Geyer et al. 2001; Liedtke et al. 2001).
Determination of which sites in the glycoprotein are glycosylated and determination of the extent of occupation at each site are generally accomplished by performing tryptic or other degradative reactions, generating a peptide/glycopeptide mixture. Some time ago, Carr and coworkers demonstrated that glycopeptides can be selectively detected in such mixtures by precursor ion scanning (Huddleston et al. 1993). The next step is usually MS/MS for the structural characterization of the detected glycopeptides. This approach, however, is only straightforward for proteins with known amino acid sequence, which is not the case for RvH1. This restriction, which originates from the dominance of the sugar fragments, can be overcome by performing MS/MS or MS3 on the Y1-ion (peptide + GlcNAc) generated via in-source fragmentation or MS/MS, respectively. As we experienced, generating such an ion in the source is hard to perform in a controlled manner. This, together with the fact that a low mass cut-off exists when performing MS3, made the method of limited value. In the present work, an elegant strategy using 18O-labeling, nano-LC-MALDI-MS, precursor ion scanning, and higher-order MS was introduced to characterize glycopeptides without the need for prior knowledge of the amino acid sequence of the protein. In a first step, N-glycans were enzymatically removed with simultaneous partial (50%) 18O-labeling of glycosylated asparagine residues prior to proteolysis. Following nano-LC-MALDI-TOF-MS, the originally glycosylated peptides could be revealed by their occurrence as doublets and their sequences determined by MS/MS. By way of example, the peptide with m/z 906.546 could easily be identified as being 18O-labeled (Figure 9). The TOF/TOF spectrum of this peptide provided an unambiguous sequence around the glycosylation site residue (Figure 10). In cases where this method gave no good result, we switched over to the use of the Q-Trap mass spectrometer as shown in Figure 11. After the glycosylation sites had been characterized, precursor ion scanning was used to determine the site occupancies as demonstrated (Figure 12). Using this approach, seven glycosylation sites could unambiguously be determined and their occupancy revealed (Table III). The peptides were searched against the National Center for Biotechnology Information (NCBI) non-redundant protein database, using FASTS (Mackey et al. 2002). All of them show homology with molluscan Hcs.
In terms of future developments, we think that the incorporation of an additional dimension in separation (e.g. ion exchange chromatography) could make the method valuable for the large-scale characterization of glycosylation sites and glycopeptides in a trypsinolate of a complex biological sample, e.g. the proteins of serum. Performing a database search following MS/MS might unambiguously identify the glycoproteins. This approach, which is only limited by the PNGase F specificity, is an alternative to lectin affinity chromatography for the characterization of glycoproteins (Kaji et al. 2003; Yang and Hancock 2004; Madera et al. 2006).
|
Materials and methods |
|---|
|
TopAbstractIntroductionResultsDiscussionMaterials and methodsSupplementary dataAcknowledgmentsReferences |
|---|
Materials
All chemicals, unless otherwise noted, were purchased from Sigma-Aldrich (St. Louis, MO). Native Hc was purified from the hemolymph of the Black Sea marine snail R. venosa as described previously (Dolashka-Angelova, Schwarz et al. 2003
). Dissociation of native Hc was achieved by dialyzing the protein against 0.13 M glycine/NaOH buffer (pH 9.6), and RvH1 and RvH2 were purified by means of ion exchange chromatography.
Deglycosylation and trypsinolysis
For deglycosylation, approximately 2 mg of protein was dissolved in 20 µL of denaturing solution (1% SDS, 0.5 M mercaptoethanol, 0.1 M EDTA). Following incubation at room temperature during 30 min, 160 µL of sodium-phosphate buffer (200 mM at pH 8.6) was added and the solution was placed in a boiling water bath for 5 min. After cooling to room temperature, 20 µL of Triton X-100 and 2 µL of PNGase F (2 U) (Roche Diagnostics GmbH, Mannheim, Germany) were added. This mixture was incubated for 20 h at 37 °C. The liberated N-glycans were purified from the reaction mixture by solid phase extraction on a Carbograph column (Alltech, Lokeren, Belgium). The glycans were eluted with 2 mL of 25% acetonitrile/0.05% trifluoroacetic acid (TFA). The collected fraction was dried and dissolved in 20 µL H2O (stock solution).
For 18O-labeling and proteolysis, approximately 500 µg of protein was dissolved in 600 µL of 50 mM NH4HCO3 and 50 µL dithiothreitol (45 mM) was added. Reduction took place at 60 °C for 30 min. After cooling to room temperature, 50 µL iodoacetamide (100 mM) was added and the mixture was placed in the dark for 30 min. The reduced and alkylated protein was purified from the reaction medium by membrane filtration using an Ultrafree-0.5 centrifugal filter (Millipore, Bedford, MA) with a molecular mass limit of 10 kDa. A volume of 40 µL 100 mM NH4HCO3 (50% H218O) and 2 U of PNGase F was added and the mixture was incubated at 37 °C for 20 h. The mixture was subjected to membrane filtration to remove the excess of H218O, and the purified 50% 18O-labeled protein was subsequently dissolved in 100 µL 50 mM NH4HCO3. An amount of 2 µg of modified trypsin (Promega, Madison, WI) was added and the protein was incubated overnight at 37 °C. A 20 times diluted solution was used for LC-analysis.
Proteolysis of the intact glycoprotein (500 µg) was performed following reduction and alkylation, as described above. The reduced and alkylated protein was purified by membrane filtration. A volume of 100 µL 50 mM NH4HCO3 and 2 µg trypsin was added and the protein was incubated overnight at 37 °C. A 20 times diluted solution was used for LC-analysis.
APTS and 3-AP labeling
For CE-MS analyses, 5 µL of the stock solution was dried and derivatized with APTS (Beckman-Coulter, Fullerton, CA) or 3-AP according to the previously described procedure (Sandra, Van Beeumen et al. 2004). Following derivatization, the labeled sugars were precipitated by addition of ice-cold acetone. The supernatant was removed and 10 µL of H2O was added to the precipitate. Prior to CE-MS analyses, the solution was diluted 5 times.
Exoglycosidase sequencing
For the digestion with -1,2-mannosidase, Trichoderma reesei mannosidase was used (Maras et al. 2000). One unit was added to the glycans (0.5 µL of the stock solution), which were dissolved in a 100 mM sodium acetate buffer, pH 5 (final volume of 10 µL). For digestions with ß-galactosidase from bovine testes (Sigma-Aldrich), 1 µL (1.14 milli-units) was added to glycans dissolved in acetate (100 mM) and phosphate buffers (50 mM) of different pHs (pH 4–8). For digestions with bovine kidney, -1,2,3,4,6-fucosidase (Prozyme, San Leandro, CA), 1 µL (1.65 milli-units) was added to the glycans dissolved in 50 mM NH4OAc (pH 5). Overnight incubations at 37 °C were performed. RNase B, N-pro-opiomelanocortin, and desialylated fetuin N-glycans were treated in parallel by means of control of the enzymatic activity.
Permethylation and amidation
Glycans were permethylated as described before (Sandra, Devreese et al. 2004). Amidation was carried out as described recently (Sekiya et al. 2005). RvH1 N-glycans (1 µL of the stock solution) were dissolved in 25 µL of 1 M NH4Cl and then mixed with 15 µL of 1 M DMT-MM. After incubation at 50 °C for 24 h, the reaction mixture was desalted by hydrophilic affinity isolation of the oligosaccharides. The sample solution was mixed with 100 µL Sepharose CL-4B slurry (Amersham Biosciences, Uppsala, Sweden) in 1 mL of 1-butanol/ethanol/H2O (4:1:1, by vol.). After gentle shaking for 45 min, the mixture was centrifuged and the supernatant was removed. The gel was subsequently washed with the same solution (3 x 500 µL). Ethanol/H2O (1:1, v/v) (250 µL) was added to the gel and the resultant mixture was incubated for 30 min. The solution phase was recovered and dried. The sample was then dissolved in 10 µL of water for MALDI-MS analysis.
Capillary electrophoresis-mass spectrometry
CE experiments of the APTS-labeled N-glycans were performed as described before (Sandra, Van Beeumen et al. 2004). The fused silica capillary dimensions were 50 µm x 60 cm. 3-AP-labeled glycans were separated in a 50 µm x 80 cm untreated fused silica capillary using a 25 mM ammonium-formate buffer (pH 3). Samples were introduced by hydrodynamic injection (50 mbar) for 10 s. A plug of water was introduced prior to sample injection (3 s, 50 mbar). The separation voltage was set at +25 kV, and the resultant current was 17 µA. Sheath liquid was delivered at a flow rate of 2 µL/min and consisted of 75% isopropanol. No sheath gas was applied. The ESI-needle voltage was set at +3 kV.
Nano-LC-MALDI and -ESI-MS
Nano-LC-MALDI and the nano-LC-ESI-Q-Trap analyses of glycopeptides were performed as described before (Sandra, Devreese et al. 2004; Vanrobaeys et al. 2005). For LC-MS/MS studies on the 18O-labeled peptides, Q1 resolution was set to "low" so that all isotopes of a particular isotope were transmitted. In enhanced product ion (EPI) experiments with Q1-to-Q2 ion activation, the collision energies were set between 25 and 50 eV. For TDF experiments with Q2-to-Q3 ion activation, collision energy values were set between 15 and 35 eV, cool time at 5–10 ms, and the fill mass was set at 85–90% of the precursor m/z value. The trap fill time in EPI and TDF was 200 ms and the scan rate 4000 Da/s.
Matrix-assisted laser desorption ionization-time-of-flight/time-of-flight
The 4700 Proteomics Analyser with TOF/TOF optics (Applied Biosystems, Framingham, MA) was used in this study. The mass spectrometer uses a 200 Hz frequency-tripled Nd-YAG laser operating at a wavelength of 355 nm. MS/MS analysis of the LC-fractions was performed in the simple metastable decomposition mode (no collision gas and with the collision energy set at 1 keV, defined as the potential difference between the source and the collision cell) or in the CID mode at 1 keV (air as collision gas). A total of 1500 shots were acquired in the MS mode and 5000 in the MS/MS mode.
Prior to fraction collection, the LC-MALDI target plate was manually spotted with 0.4 µL matrix (8 mg/mL -cyano-4-hydroxycinnamic acid dissolved in 50% acetonitrile/0.1% TFA, containing 2 mM dibasic ammonium citrate). After fractionation was completed, 0.1 µL of a matrix saturated solution (50% acetonitrile/20% ethanol/0.1% TFA) was added on top of each spot and allowed to air-dry.
For glycan analysis, the matrix solution was prepared as a 10 mg/mL dihydroxybenzoic acid solution in 50% acetonitrile. The sugar stock solution and the exoglycosidase digestions were diluted 100- and 10-fold, respectively, and 1 µL of a 1:1 sugar–matrix mixture was applied onto the MALDI target.
Off-line ESI-MS
Off-line ESI-MS measurements of the methylated glycans were performed on the Q-Trap mass spectrometer equipped with a nanospray ion source (Protana, Odense, Denmark) using Protana medium nanospray needles. Typically, 5 µL of sample in 50% methanol was introduced. The needle voltage was set at 1000 V. In the product ion scanning and MS3 mode, the scan speed was set to 1000 Da/s, with Q0-trapping being activated. The trap fill time was 200 ms in the MS/MS- and 300 ms in the MS3-scan modes. For operation in the MS/MS and MS3 modes, the resolution of Q1 was set to "low." For MS3 operation, the excitation coefficient (slightly mass dependent) was set at a value to excite only the first and part of the second isotope for a single-charged precursor. Excitation time was set at 100 ms.
|
Supplementary data |
|---|
|
TopAbstractIntroductionResultsDiscussionMaterials and methodsSupplementary dataAcknowledgmentsReferences |
|---|
Supplementary data are available at Glycobiology online (http://glycob.oxfordjournals.org/).
|
Acknowledgments |
|---|
|
TopAbstractIntroductionResultsDiscussionMaterials and methodsSupplementary dataAcknowledgmentsReferences |
|---|
The authors wish to thank Applera Benelux for instrumental support and Roland Contreras for the exoglycosidase gift. This work was supported by a research grant to K.S. from the Institute for the Promotion of Innovation by Science and Technology in Flanders (IWT-Vlaanderen), by the Fund for Scientific Research-Flanders (FWO-Vlaanderen) through project G.0312.02 to J.V.B. and B.D, and by the Bulgarian National Science Fund through grant X-1202 to P.D.-A.
|
Footnotes |
|---|
2 Present address: Peakadilly nv., Technologiepark 4, VIB Bio-incubator, B-9052 Zwijnaarde/Ghent, Belgium, e-mail: koen.sandra{at}peakadilly.com
|
Abbreviations |
|---|
3-AP, 3-aminopyrazole; APTS, 8-aminopyrene-1,3,6-trisulfonate; CE, capillary electrophoresis; CID, collision-induced dissociation; DMT-MM, 4-(4,6-dimethoxy-1,3,5-triazin-2yl)-4-methylmorpholinium chloride; EOF, electro-osmotic flow; EPI, enhanced production ion; ER, enhanced resolution; ESI, electrospray ionization; FU, functional unit; Hc, hemocyanin; Hex, hexose; HexA, hexuronic acid; HexNAc, N-acetylhexosamine; KLH, keyhole limpet hemocyanin; LC, liquid chromatography; MALDI, matrix-assisted laser desorption ionization; PNGase F, peptide-N-glycosidase F; RvH, Rapana venosa hemocyanin; TDF, time-delayed fragmentation; TFA, trifluoroacetic acid; TOF, time-of-flight
|
References |
|---|
|
TopAbstractIntroductionResultsDiscussionMaterials and methodsSupplementary dataAcknowledgmentsReferences |
|---|
Bateman KP, White RL, Yaguchi M, Thibault P. (1998) Characterization of protein glycoforms by capillary-zone electrophoresis-nanoelectrospray mass spectrometry. J Chromatogr A 794:327–344.[CrossRef]
Demelbauer UM, Zehl M, Plematl A, Allmaier G, Rizzi A. (2004) Determination of glycopeptide structures by multistage mass spectrometry with low-energy collision-induced dissociation: comparison of electrospray ionization quadrupole ion trap and matrix-assisted laser desorption/ionization quadrupole ion trap reflectron time-of-flight approaches. Rapid Commun Mass Spectrom. 18:1575–1582.[CrossRef][Web of Science][Medline]
Dolashka-Angelova P, Beck A, Dolashki A, Beltramini M, Stevanovic S, Salvato B, Voelter W. (2003) Characterization of the carbohydrate moieties of the functional unit RvH1-a of Rapana venosa haemocyanin using HPLC/electrospray ionization MS and glycosidase digestion. Biochem J. 374:185–192.[CrossRef][Web of Science][Medline]
Dolashka-Angelova P, Beck A, Dolashki A, Stevanovic S, Beltramini M, Salvato B, Hristova R, Velkova L, Voelter W. (2004) Carbohydrate moieties of molluscan Rapana venosa hemocyanin. Micron. 35:101–104.[CrossRef][Web of Science][Medline]
Dolashka-Angelova P, Schwarz H, Dolashki A, Stevanovic S, Fecker M, Saeed M, Voelter W. (2003) Oligomeric stability of Rapana venosa hemocyanin (RvH) and its structural subunits. Biochim Biophys Acta. 1646:77–85.[Medline]
Domon B and Costello CE. (1988) A systematic nomenclature for carbohydrate fragmentations in FAB-MS/MS spectra of glycoconjugates. Glycoconj J. 5:397–409.[CrossRef][Web of Science]
Geyer H, Bahr U, Liedtke S, Schachner M, Geyer R. (2001) Core structures of polysialylated glycans present in neural cell adhesion molecule from newborn mouse brain. Eur J Biochem. 268:6587–6599.[Web of Science][Medline]
Gielens C, Idakieva K, Van den Bergh V, Siddiqui NI, Parvanova K, Compernolle F. (2005) Mass spectral evidence for N-glycans with branching on fucose in a molluscan hemocyanin. Biochem Biophys Res Commun. 331:562–570.[CrossRef][Web of Science][Medline]
Gonzales J, Takao T, Hori H, Besada V, Rodriguez R, Padron G, Shimonishi Y. (1992) A method for determination of N-glycosylation sites in glycoproteins by collision-induced dissociation analysis in fast atom bombardment mass spectrometry: identification of the positions of carbohydrate-linked asparagine in recombinant -amylase by treatment with peptide-N-glycosidase F in 18O-labeled water. Anal Biochem. 205:151–158.[CrossRef][Web of Science][Medline]
Hager JW. (2003) Product ion spectral simplification using time-delayed fragment ion capture with tandem linear ion traps. Rapid Commun Mass Spectrom. 17:1389–1398.[CrossRef][Web of Science][Medline]
Håkansson K, Chalmers MJ, Quinn JP, McFarland MA, Hendrickson CL, Marshall AG. (2003) Combined electron capture and infrared multiphoton dissociation for multistage MS/MS in a Fourier transform ion cyclotron resonance mass spectrometer. Anal Chem. 75:3256–3262.[Medline]
Håkansson K, Cooper HJ, Emmett MR, Costello CE, Marschall AG, Nilsson CL. (2001) Electron capture dissociation and infrared multiphoton dissociation MS/MS of an N-glycosylated tryptic peptide to yield complementary sequence information. Anal Chem. 73:4530–4536.[Medline]
Harris JR and Markl J. (1999) Keyhole limpet hemocyanin (KLH): a biomedical review. Micron. 30:597–623.[CrossRef][Web of Science][Medline]
Harris JR and Markl J. (2000) Keyhole limpet hemocyanin: molecular structure of a potent marine immunoactivator. A review. Eur Urol. 37:24–33.[CrossRef]
Hogan JM, Pitteri SJ, Chrisman PA, McLuckey SA. (2005) Complementary structural information from a tryptic N-linked glycopeptide via electron transfer ion/ion reactions and collision-induced dissociation. J Proteome Res. 4:628–632.[CrossRef][Web of Science][Medline]
Huddleston MJ, Bean MF, Carr SA. (1993) Collisional fragmentation of glycopeptides by electrospray ionization LC/MS and LC/MS/MS: methods for selective detection of glycopeptides in protein digests. Anal Chem. 65:877–884.[Medline]
Idakieva K, Parvanova K, Nicholson G, Voelter W, Genov N. (2001) Carbohydrate content and monosaccharide composition of dioxygen-binding functional units from Rapana thomasiana hemocyanin isoform RtH2. Compt Rend Bulg Sci. 54:73–76.
Idakieva K, Stoeva S, Voelter W, Gielens C. (2004) Glycosylation of Rapana thomasiana hemocyanin. Comparison with other prosobranch (gastropod) hemocyanins. Comp Biochem Physiol. 138B:221–228.[CrossRef][Medline]
Jacques AJ, Opdenakker G, Rademacher W, Dwek RA, Zamze SE. (1996) The glycosylation of Bowes melanoma tissue plasminogen activator: lectin mapping, reaction with anti-L2/HNK-1 antibodies and the presence of sulphated/glucuronic acid containing glycans. Biochem J. 316:427–437.
Kaji H, Saito H, Yamauchi Y, Shinkawa T, Taoka M, Hirabayashi J, Kasai K, Takahashi N, Isobe T. (2003) Lectin affinity capture, isotope-coded tagging and mass spectrometry to identify N-linked glycoproteins. Nat Biotechnol. 21:667–672.[CrossRef][Web of Science][Medline]
Kurokawa T, Wuhrer M, Lochnit G, Geyer H, Markl J, Geyer R. (2002) Hemocyanin from the keyhole limpet Megathura crenulata (KLH) carries a novel type of N-glycans with Gal(ß1-6)Man-motifs. Eur J Biochem. 269:5459–5473.[Web of Science][Medline]
Liedtke S, Geyer H, Wuhrer M, Geyer R, Frank G, Gerardy-Schahn R, Zähringer U, Schachner M. (2001) Characterization of N-glycans from mouse brain neural cell adhesion molecule. Glycobiology 11:373–384.
Lommerse JPM, Thomas-Oates JE, Gielens C, Préaux G, Kamerling JP, Vliegenhart JFG. (1997) Primary structure of 21 novel monoantennary and diantennary N-linked carbohydrate chains from d-hemocyanin of Helix pomatia. Eur J Biochem 249:195–222.[Web of Science][Medline]
Mackey AJ, Haystead TAJ, Pearson WR. (2002) Getting more from less: algorithms for rapid protein identification with multiple short peptide sequences. Mol Cell Proteomics. 1:139–147.
Madera M, Mechref Y, Klouckova I, Novotny M. (2006) Semiautomated high-sensitivity profiling of human blood serum glycoproteins through lectin preconcentration and multidimensional chromatography/tandem mass spectrometry. J Proteome Res. 5:2348–2363.[CrossRef][Web of Science][Medline]
Maras M, Callewaert N, Piens K, Claeyssens M, Martinet W, Dewaele S, Contreras H, Dewerte I, Penttila M, Contreras R. (2000) Molecular cloning and enzymatic characterization of a Trichoderma reesei 1,2--d-mannosidase. J Biotechnol 77:255–263.[CrossRef][Web of Science][Medline]
Samyn B, Debyser G, Sergeant K, Devreese B, Van Beeumen J. (2004) A case study of de novo sequence analysis of N-sulfonated peptides by MALDI TOF/TOF mass spectrometry. J Am Soc Mass Spectrom 15:1838–1852.[CrossRef][Web of Science][Medline]
Sandra K, Devreese B, Stals I, Claeyssens M, Van Beeumen J. (2004) The Q-Trap mass spectrometer, a novel tool in the study of protein glycosylation. J Am Soc Mass Spectrom 15:413–423.[CrossRef][Web of Science][Medline]
Sandra K, Van Beeumen J, Stals I, Sandra P, Claeyssens M, Devreese B. (2004) Characterization of cellobiohydrolase I N-glycans and differentiation of their phosphorylated isomers by capillary electrophoresis-Q-Trap mass spectrometry. Anal Chem 76:5878–5886.[Medline]
Sekiya S, Wada Y, Tanaka K. (2005) Derivatization for stabilizing sialic acids in MALDI-MS. Anal Chem 77:4962–4968.[Medline]
Stoeva S, Dolashka P, Genov N, Voelter W. (1999) Domain structure of the Rapana thomasiana (Gastropod) hemocyanin. J Chem Soc Pak 21:229–238.
Stoeva S, Idakieva K, Betzel C, Genov N, Voelter W. (2002) Amino acid sequence and glycosylation of functional unit RtH2-e from Rapana thomasiana (gastropod) hemocyanin. Arch Biochem Biophys 399:2149–158.[CrossRef][Web of Science][Medline]
Stoeva S, Idakieva K, Genov N, Voelter W. (1997) Complete amino acid sequence of dioxygen-binding functional unit of the Rapana thomasiana hemocyanin. Biochem Biophys Res Commun 238:403–410.[CrossRef][Web of Science][Medline]
Van Holde KE and Miller KI. (1995) Hemocyanins. Adv Protein Chem 47:1–81.[Web of Science][Medline]
Van Holde KE, Miller KI, Decker H. (2001) Hemocyanins and invertebrate evolution. J Biol Chem 276:15563–15566.
Van Kuik JA, Kamerling JP, Vliegenthart JFG. (1990) Carbohydrate analysis of hemocyanins. In Préaux G and Lontie R (Eds.). Invertebrate dioxygen carriers(Leuven University Press, Leuven, Belgium) pp. 157–163.
Vanrobaeys F, Van Coster R, Dhondt G, Devreese B, Van Beeumen J. (2005) Profiling of myelin proteins by 2D-gel electrophoresis and multidimensional liquid chromatography coupled to MALDI TOF-TOF mass spectrometry. J Proteome Res 4:2283–2293.[CrossRef][Web of Science][Medline]
Voshol H, van Zuylen CWEM, Orberger G, Vliegenthart JFG, Schachner M. (1996) Structure of the HNK-1 carbohydrate epitope on bovine peripheral myelin glycoprotein P0. J Biol Chem 271:22957–22960.
Wuhrer M, Robijn MLM, Koeleman CAM, Balog CIA, Geyer R, Deelder AM, Hokke CH. (2004) A novel Gal(1-4)Gal(ß1-4)Fuc(1-6)-core modification attached to the proximal N-acetylglucosamine of keyhole limpet haemocyanin (KLH) N-glycans. Biochem J 378:625–632.[CrossRef][Web of Science][Medline]
Yang Z and Hancock WS. (2004) Approach to the comprehensive analysis of glycoproteins isolated from human serum using a multi-lectin affinity column. J Chromatogr A 1053:79–88.
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||