Glycobiology Advance Access originally published online on July 31, 2006
Glycobiology 2006 16(12):1207-1218; doi:10.1093/glycob/cwl034
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Site-specific glycan analysis of human chorionic gonadotropin ß-subunit from malignancies and pregnancy by liquid chromatography—electrospray mass spectrometry
2 Department of Clinical Medicine, Division of Clinical Chemistry, Biomedicum, University of Helsinki, PO Box 63 (Haartmaninkatu 8), Helsinki FIN-00014, Finland; and
3 Department of Obstetrics and Gynecology, Columbia University College of Physicians and Surgeons, New York, NY 10032
1 To whom correspondence should be addressed; e-mail: leena.valmu{at}helsinki.fi
Received on May 12, 2006; revised on July 7, 2006; accepted on July 28, 2006
 |
Abstract |
|---|
 Top Abstract IntroductionResultsDiscussionMaterials and methodsAcknowledgmentsConflict of interest statementReferences |
|---|
Glycosylation is an important posttranslational modification in proteins, and aberrant glycosylation occurs in malignancies. Human chorionic gonadotropin (hCG) is a glycoprotein hormone produced in high concentrations during pregnancy. It is also expressed as particular glycoforms by certain malignancies. These glycoforms, which are called "hyperglycosylated" hCG (hCGh), have been reported to contain more complex glycan moieties. We have analyzed tryptic glycopeptides of the ß-subunit of hCG of various origins by liquid chromatography (LC) connected to an electrospray mass spectrometer. Site-specific glycan structures were visualized by the use of differential expression analysis software. hCGß was purified from urine of two patients with testicular cancer, one with choriocarcinoma, one with an invasive mole, two pregnant women at early and late gestation, from a pharmaceutical preparation and culture medium of a choriocarcinoma cell line. N-glycans at Asn-13 and Asn-30 as well as O-glycans at Ser-121, Ser-127, Ser-132, and Ser-138 were characterized. In all samples, the major type of N-glycan was a biantennary complex-type structure, but triantennary structures linked to Asn-30 as well as fucosylation of the Asn-13-bound glycan are increased in cancer-derived hCGß. There were significant site-specific differences in the O-glycans, with constant core-2 glycans at Ser-121, core-1 glycans at Ser-138, and putative sites unoccupied by any glycan. Core-2 glycans at either Ser-127 or Ser-132 were enriched in cancer. The glycans of free hCGß were larger and had a higher fucose content of Asn-13-linked oligosaccharides than intact hCG. This may facilitate the detection of this malignancy-associated variant by a lectin assay. Analysis of hCGh affinity purified with antibody B152 confirmed that this antibody recognizes a core-2 glycan on Ser-132.
Key words: antibody B152 / cancer / hCG / hyperglycosylated hCG / mass spectrometry / N-glycosylation / O-glycosylation
|
Introduction |
|---|
|
TopAbstractIntroductionResultsDiscussionMaterials and methodsAcknowledgmentsConflict of interest statementReferences |
|---|
Glycosylation is one of the most important and common posttranslational protein modifications. It may affect both protein structure and function, but the impact of glycosylation is protein dependent. Because changes in protein glycosylation are associated with numerous diseases, including cancer (Hakomori, 2002
), protein glycosylation is important in clinical research.
Recent advances in glycoproteomics, a field combining proteomics and glycomics, have been greatly influenced by methodological development. Especially mass spectrometry (MS) provides important structural information on glycoproteins (Harvey, 2001). MS analysis of glycoproteins based on the release of glycans either enzymatically or chemically provides structural glycan information, but the information on site occupancy is lost. Analysis of glycopeptides obtained by proteolytic digestion, typically with trypsin, provides site-specific glycan information, but this approach may be hampered by difficulties in the digestion of resistant glycoproteins. Furthermore, the analysis is often complicated by the heterogeneity of structurally related oligosaccharides occupying a single glycosylation site.
Human chorionic gonadotropin (hCG) is a glycoprotein hormone produced by placental trophoblasts and trophoblastic tumors. hCG is a heterodimer composed of 
- and ß-subunits (hCGß). The -subunit is shared with the other glycoprotein hormones, luteinizing hormone, follicle stimulating hormone, and thyroid stimulating hormone, whereas the ß-subunit is specific for each hormone. hCG is expressed throughout gestation, and it occurs at high concentrations both in serum and in urine of pregnant women.
About one-third of the molecular weight of hCG consists of carbohydrates, and glycosylation is of structural and functional importance, affecting both the half-life in circulation and the signal transduction induced by this hormone (Lustbader et al., 1998). The glycosylation pattern of hCG has been extensively studied. Two asparagine-linked carbohydrate units (N-glycans) on hCGß at Asn-13 and Asn-30 (Carlsen et al., 1973) have been shown to consist of biantennary, complex-type N-glycans with terminal sialic acids and a variable content of fucose (Kessler, Reddy, et al., 1979; Weisshaar et al., 1991). The site-specific N-glycan structures have been further characterized by MS using matrix-assisted laser desorption ionization (MALDI) technique (Laidler et al., 1995; Jacoby et al., 2000) and liquid chromatography (LC)–electrospray ionization (ESI) MS (Liu and Bowers, 1997). Four serine-linked oligosaccharides (O-glycans) attached to serine residues 121, 127, 132, and 138 have been shown to consist of monoantennary, so-called core-1 O-glycan structures with two sialic acids attached to each glycan (Kessler, Mise, et al., 1979). However, core-2 type O-glycan at Ser-121 as well as putative glycosylation sites apparently not glycosylated has also been detected (Liu and Bowers, 1997). All MS studies on the glycan structure of hCG have been performed using pharmaceutical hCG preparations that have been partially purified from pooled urine from pregnant women.
Except in pregnancy, hCG is expressed by certain malignancies and especially by trophoblastic and testicular germ cell tumors, for which hCG is a sensitive marker (Stenman et al., 2004). Furthermore, the concentration of free hCGß in serum is often elevated in patients with nontrophoblastic cancers, and this is a sign of adverse prognosis (Alfthan, Haglund, Dabek, et al., 1992; Alfthan, Haglund, Roberts, et al., 1992; Stenman et al., 2004). Various isoforms of hCG have been shown to be associated with malignancies. These include "nicked" hCG (hCGn), in which several peptide bonds in loop 2 of hCGß are cleaved (Puisieux et al., 1990; Birken et al., 1999) as well as "hyperglycosylated" hCG (hCGh), which is the major form of hCG in trophoblastic cancer (Elliott et al., 1997; Birken et al., 1999) and early pregnancy (Kovalevskaya et al., 2002). hCGh has also been associated with Down’s syndrome (Cole et al., 1999) and early pregnancy loss (Kovalevskaya et al., 2002). Tumor-derived hCGh has been shown to contain increased amounts of triantennary N-glycans (Elliott et al., 1997), abnormal biantennary N-glycans (Kobata and Takeuchi, 1999), and biantennary core-2 type O-glycans (Elliott et al., 1997; Birken et al., 2003).
Most studies on the carbohydrates of hCGh have been performed by chromatographic analysis of glycans released from the polypeptide, and this gives no site-specific structure information (Mizuochi et al., 1983; Elliott et al., 1997). An antibody to hCG, B152 (Birken et al., 1999), recognizes a core-2 type O-glycan attached to Ser-132 and the surrounding peptide structures in hCGß (Kovalevskaya et al., 2002; Birken et al., 2003). All clinical studies on hCGh have been based on the use of this antibody. No studies on site-specific oligosaccharide heterogeneity of cancer-derived forms of hCG, which are often lumped together under the term "hyperglycosylated" (Birken et al., 2003), have been performed.
In this study, we have used novel differential expression analysis software to analyze site-specific glycan structures of hCGß from various sources by LC–MS of tryptic protein fragments.
|
Results |
|---|
|
TopAbstractIntroductionResultsDiscussionMaterials and methodsAcknowledgmentsConflict of interest statementReferences |
|---|
Characterization of hCGß purified from urine
hCG was purified from urine by serial immunoaffinity chromatography and gel filtration. Purified hCG was reduced and alkylated with 4-vinyl pyridine causing the dissociation of
- and ß-subunits. hCGß was isolated by reverse-phase (RP) chromatography, and its purity was checked by sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS–PAGE) and further by LC–MS analysis of tryptic peptides. All the potential tryptic peptides (Table I) without attached glycans were visualized by LC–MS analysis with better than 20 ppm mass accuracy (Table II). The identity of the peptides was verified by LC–tandem mass spectrometry (MSMS) identification (data not shown). In addition to the predicted tryptic cleavage sites, additional cleavage sites were observed in peptides 5 and 9 (data not shown). The cleavage pattern of peptide 5 is in agreement with the proposed cleavage of loop 2 in the ß-subunit in hCGn (Cole et al., 1991), but in contrast to a previous report (Elliott et al., 1997), the major sites cleaved were after amino acids 45 and 49, with some minor cleavage after amino acids 44, 46, and 48. In P9, the major cleavage occurred after Asn-77, but minor cleavage was also seen after amino acids 76, 78, 79, and 80.
|
|
Of the putative N-glycan-containing peptides P3 and P4 as well as the O-glycan-containing peptides P13, P14, and P15, some lacking the glycan were also observed (Table III), but with minor intensities.
|
When the N-glycans of hCGß were enzymatically removed by peptide-N-glycosidase F (PNGase F), a mass shift from 37 to 29 kDa was observed by SDS–PAGE (data not shown). In LC–MS analyses of the tryptic digest, peptides P3 and P4 were detected at high intensity, with a mass adduction of 1 Da due to the conversion of asparagine to aspartic acids during PNGase F treatment (data not shown).
Site-specific N-glycan analysis of hCGß
For glycan analysis, 3 pmol of hCGß tryptic digest was subjected to capillary LC on an Atlantis dC18 column specially designed to retain hydrophilic peptides such as glycosylated ones. The LC–MS data generated were incorporated into DeCyder MS software, where two-dimensional (2D) intensity maps were visualized (Figure 1A), with each spot representing the intensity of a peptide in a certain charge state separated based on its m/z value and retention time, that is, hydrophilicity. The same peptide is represented by several charge states within the map. To identify tryptic peptides containing N-glycans, we compared the 2D LC–MS patterns using the DeCyder MS software PepMatch module to patterns of tryptic digests of hCGß, from which the N-glycans had been enzymatically removed (Figure 1B). Two highly heterogeneous peptide patterns or "clouds" were observed for the N-glycosylated hCGß peptides. In peptide-by-peptide analysis, peptides occurring at equal concentrations in the two digests (Figure 1C) were easily separated from peptides present only in the N-glycosylated molecule (Figure 1D). The identity of the differentially expressed peptides was studied by LC–MSMS fragmentation analysis (Figure 1E), which revealed sugar-specific oxonium ions (m/z 204 for N-acetylhexosamine (HexNAc), 274 for N-acetylneuraminic acid (NeuAc)-H2O, 292 for NeuAc, 366 for Hex-HexNAc, and 657 for NeuAc-Hex-HexNAc) as a result of oligosaccharide dissociation. In addition, a clear fragmentation profile of the attached oligosaccharide portion of the glycopeptide was seen. Minor peptide-derived y-ions could also be detected in some peptide fragmentations facilitating differentiation between the P3 and P4 containing glycans (Figure 1A). The deglycosylation of the Asn-30-attached oligosaccharides from P4 was incomplete in part of the peptides.
|
, HexNAc; 
, Hex; 
, Fuc; 
, NeuAc.In the LC–MS analysis, glycopeptides containing Asn-13-linked oligosaccharides were mainly preset as [M + 3H]3+ and [M + 4H]4+, with minor intensities of smaller peptides as [M + 2H]2+. Ion counts of all different charge states of the same peptide were summarized to calculate the intensity of each glycopeptide. The masses of the oligosaccharides attached to P3 were calculated by subtracting the theoretical polypeptide portion mass of 1418.8 Da from the observed deconvulated masses of glycosylated peptides. The sugar structures were thus derived by the mass-matching approach (Figure 2). The same approach was used to calculate the oligosaccharides attached to Asn-30 in P4. Owing to their larger mass, P4 glycopeptides appeared in the LC–MS analysis mainly as [M + 4H]4+ and [M + 5H]5+, with minor intensities as [M + 3H]3+. Analysis of the Asn-30-attached oligosaccharides was complicated because Met-41 in P4 occurred both in native and in oxidized forms, and the intensities of these needed to be summed together (Figure 3). Interestingly, incomplete deglycosylation of P4 with PNGase F was observed in P4 with oxidized Met-41 (Figure 1B).
|
, Man;
|
The N-linked oligosaccharides attached to Asn-13 and Asn-30 in hCGß purified from the urine of a pregnant woman at 35 weeks of pregnancy, from the urine of a patient suffering from invasive mole, and from the urine of a patient with testicular cancer are shown in Figures 2 and 3. The relative intensities, which roughly represent the relative abundance of different glycan structures present, are shown with a value of 100 for the most intense component. Intensities of smaller glycan structures might be slightly increased relative to sialylated structures because of possible sialic acid release in mass spectrometric analysis. To simplify the picture of the enormous oligosaccharide heterogeneity, glycan structures with a relative intensity <15 in all three samples are omitted from the figures, but they were taken into account for the calculation of the proportion of all oligosaccharide structures shown in Tables III and IV. In addition to the structures shown, some tetra-antennary glycans and some Asn-30-linked oligosaccharides lacking fucose were observed.
|
Site-specific O-glycan analysis of hCGß
The O-glycosylated tryptic peptides were difficult to detect by LC–MS analysis of the crude tryptic digest probably because of suppression by the enormous heterogeneity of N-glycan containing peptides. The C-terminal portion of hCGß was therefore isolated by RP-high performance liquid chromatography (HPLC) after cleavage with GluC protease. The N-terminus of the fragment was confirmed to be the expected Ser-66 by LC–MSMS analysis of the fragment tryptic digest (data not shown).
Direct ESI–MS analysis of the purified C-terminal fragment showed a wide variety of heterogeneous masses. The deconvulated masses of the C-terminal hCGß fragment from normal late pregnancy, invasive mole, and testicular cancer are shown in Figure 4. On the basis of the masses observed, it was evident that sialylated core-2 structures were present at all four O-glycosylation sites in hCGß purified from a testicular cancer patient (Figure 4D), whereas both core-1- and core-2-containing structures were observed in molar disease and pregnancy. The small masses of the C-terminal peptide in pregnancy hCG (Figure 4A) showed that variants with some unoccupied O-glycosylation sites also existed.
|
The site-specific O-glycan structures were studied in glycopeptides produced by tryptic digestion of the C-terminal fragment. Peptides P13 and P15, which contained Ser-121- and Ser-138-linked oligosaccharides, respectively, were almost exclusively present as [M + 2H]2+. The major oligosaccharide structures assigned by the mass-matching approach are shown in Figure 5. Ser-121 was constantly occupied by a core-2 glycan, whereas the core-1 structure always dominated at Ser-138, with some core-2 structure observed in malignancy. Both sites were also observed unoccupied by any glycan. The possibility that the exposure of small Tn and T antigens within these sites could be due to the desialylation in the mass spectrometric analysis cannot be ruled out.
|
, GalNAc; Analysis of P14 was complicated because of the presence of glycans on both Ser-127 and Ser-132 within the peptide. Figure 5C shows the observed masses, the major possible glycans, and combinations of these in four groups: (1) one core-1, (2) two core-1 or one core-2, (3) one core-2 and one core-1, or (4) two core-2 O-glycans. A single oligosaccharide was detected in most of the P14 from pregnancy and invasive mole; thus, either Ser-127 or Ser-132 was not glycosylated in almost half of the hCGß molecules. Either two core-1 glycans or one core-2 glycan made up most of the rest. These two had the same mass and could not be differentiated by the mass-matching approach. The largest mass of this group (2417.0) consists of either two intact core-1 structures with one sialic acid or one sialylated core-2 structure. The intense masses of 2126.0 and 1963.9 lacked either one sialic acid or, in addition, one hexose, respectively. In larger glycopeptides, visible mainly in P14 from testicular cancer, both glycosylation sites were occupied, but because different core-1 and core-2 structures could result in the same glycopeptide mass, the interpretation was difficult. However, the major structure (3073.3) contained a sialylated core-2 structure in one and an intact core-1 structure with one sialic acid in the other site. The glycopeptide with a mass of 3729.5 contained two sialylated core-2 oligosaccharides.
Differential glycan analysis of hCGß in malignancy and pregnancy
Characteristic site-specific differences and similarities in the glycan structures were observed between hCGß from patients with testicular cancer, choriocarcinoma, invasive mole, pregnant women of different gestational age, and a pharmaceutical hCG preparation and hCG from JEG-3 cells (Table III). The oligosaccharide structures linked to Asn-13 showed no significant increase of triantennary structures in malignancy, whereas both triantennary and monoantennary oligosaccharides attached to Asn-30 were enriched in hCG from cancer patients. The degree of fucosylation in the two N-glycosylation sites varied. The Asn-30-linked oligosaccharides were mainly fucosylated, whereas fucosylation of Asn-13-linked glycans varied (Table III). Less than 25% of the oligosaccharides were fucosylated in hCGß from late pregnancy and JEG cells, whereas the fucosylation degree exceeded 36% in hCGß from malignancies.
Ser-121 in P13 contained a core-2 O-glycan structure irrespective of the source of hCG. In contrast, core-1 oligosaccharide structures containing either one or two sialic acids dominated at Ser-138. Especially in hCGß from testicular cancer, but also to some extent in early pregnancy and JEG-3 cells, the core-2 structure was observed (Figure 5, Table III).
The two O-glycosylation sites of P14 displayed the largest variation between different hCGß preparations. In pregnancy and molar disease, a single core-1 structure was most common, with either one core-2 structure or two core-1 structures as the second most common form. In cancer patients and early pregnancy, core-2 oligosaccharide structures were most often present in either one or both sites. The glycan composition of the commercial hCG preparation resembled that in early pregnancy urine.
Fairly large individual differences were observed between the two samples from late pregnancy and the two testicular cancer patients. Thus, the patterns observed in this study are likely to represent only part of the true diversity.
Glycan analysis of hCGß variants isolated from a single patient
Various forms of hCG were isolated from the patient with an invasive mole for site-specific glycan analysis. hCG was purified both from urine and from serum, and the hCGß subunit was dissociated from the heterodimer. Very similar oligosaccharide structures and the same polypeptide chain cleavages in P5 and P9 occurred in hCG from serum and urine. In urinary hCGh isolated with monoclonal antibody (MAb) B152, core-2 O-glycans dominated in P14, and only 1% contained a single core-1 glycan. Interestingly, no N-glycan structure was enriched. Free hCGß isolated from urine of this patient contained more triantennary N-glycans in P3 and P4 and more core-2 O-glycans in P14 than that in hCG, and a higher proportion of the Asn-13 glycan was fucosylated (Table IV).
|
Discussion |
|---|
|
TopAbstractIntroductionResultsDiscussionMaterials and methodsAcknowledgmentsConflict of interest statementReferences |
|---|
This study confirms the results of earlier ones (Elliott et al., 1997
; Kobata and Takeuchi, 1999), showing differences in the carbohydrate structure of hCGß derived from pregnancy and malignant tumors, respectively. Furthermore, we demonstrate large differences in glycan structures between the different glycosylation sites and in hCG from various sources, which have not been characterized before.
We used an ESI–MS connected to an LC to analyze site-specific glycan structures in tryptic peptides of hCGß. Furthermore, we used novel differential expression analysis software, designed for proteomics (Skold et al., 2002), to study differential abundance of oligosaccharide structures in the protein. Glycopeptides were identified by comparing the tryptic digest of a protein with a digest of the same deglycosylated molecule. Digests of the same protein purified from different sources and different isoforms of the same protein could also be compared. The combination of LC–MS analysis with differential expression analysis is especially useful for glycoproteins that, such as hCG, are relatively easily digestible with proteases despite the oligosaccharides attached. This approach relies entirely on mass matching, and it does not provide exact linkage information about the glycans. However, the glycan structures of hCG have been fairly extensively studied by enzymatic deglycosylation combined with mass spectrometric analysis, and on the basis of this information, it was possible to deduce structures solely by mass matching.
The N-glycosylation of hCG has been reported to be different in choriocarcinoma than in pregnancy (Mizuochi et al., 1983; Elliott et al., 1997), but site-specific glycan differences of hCG have not been studied. In this study, the earlier observed increase in the content of triantennary complex-type N-glycans in choriocarcinoma hCG (Elliott et al., 1997) was seen only in the oligosaccharides attached to Asn-30 but not in that attached to Asn-13. The earlier described increase in monoantennary N-glycans (Elliott et al., 1997) was observed in sugar structures attached to both Asn-13 and Asn-30. The abnormal biantennary N-glycan structures bearing both antennae within the same mannose (Man) residue detected in choriocarcinoma by Kobata et al. (Mizuochi et al., 1983; Kobata and Takeuchi, 1999) could not be detected by the mass-matching approach because these structures carry the same mass as ordinary biantennary N-glycans. Another site-specific difference between the two N-glycans was the degree of fucosylation. Asn-30-linked glycans were nearly totally fucosylated, as earlier reported for pregnancy hCG (Kessler, Reddy, et al., 1979; Weisshaar et al., 1991), whereas the fucosylation degree of Asn-13-linked oligosaccharides varied between 2 and 56%. Furthermore, fucosylation on Asn-13-attached glycans was enriched in malignancies, and it was also higher in early than in late pregnancy. An increased fucose content of N-glycans in choriocarcinoma hCG with no site specificity has earlier been reported in one study (Mizuochi et al., 1983) but not in another one (Elliott et al., 1997).
Increased N-glycan fucosylation has previously been reported to occur in -fetoprotein produced by hepatocellular carcinoma, and this is of diagnostic value (Aoyagi et al., 1985). Fucose on the first GlcNAc residue of complex-type N-glycans can be specifically recognized by Lens culinaris agglutinin (LCA) lectin. Lectins have also been used to distinguish glycosylation differences in hCG from pregnancy and choriocarcinoma cell lines, but LCA was not among the lectins studied (Kelly et al., 2005).
Reports on the O-glycosylation of hCGß are controversial (Kessler, Mise, et al., 1979; Liu and Bowers, 1997; Birken et al., 2003; Gervais et al., 2003) and may be explained by the site-specific O-glycan difference observed in the present study. The core-1 structure has been found to dominate in pregnancy hCG, but we found that Ser-121 always contained a biantennary core-2 structure. This was previously suggested (Liu and Bowers, 1997) but was not found in the original reports (Kessler, Mise, et al., 1979). The structure of this glycan was virtually identical in all the conditions studied. The oligosaccharide attached to Ser-138 was, on the contrary, in all samples studied mainly of core-1 type, with either one or two sialic acids attached. Core-2 structures were observed only in hCGß purified from one testicular cancer patient, the choriocarcinoma cell line, and very early pregnancy.
The occurrence of two O-glycosylation sites in P14, which is not specifically cleavable by known proteases, complicated the analysis of the oligosaccharide structures in Ser-127 and Ser-132. Despite this, it was evident that in pregnancy, only a single glycosylation site was occupied in a significant proportion of hCG, as suggested earlier (Liu and Bowers, 1997). In most malignancy-derived hCG, two glycosylation sites were occupied, and two core-2 oligosaccharides were often detected. This is in agreement with the observation that MAb B152 recognizes a Ser-132-linked core-2 structure (Birken et al., 2003). The O-glycosylation of hCG purified from urine of a patient with an invasive mole resembled pregnancy-derived hCG, whereas the N-glycosylation pattern was more similar to that in cancer-derived hCG. In very early pregnancy, the glycosylation of P14 resembled that in cancer in that both sites were occupied. Some unique O-glycan structures have been observed in recombinant hCG (Gervais et al., 2003), including O-glycan fucosylation and increased content of N-acetylhexosamine. These structures were not observed in the human hCG preparations analyzed in this study. Thus, they appear to represent oligosaccharides specific to the Chinese hamster ovary cell line used to express recombinant hCG.
In addition to glycans observed in the study, it is interesting to note that some glycosylation sites, notably Ser-138, Ser-121, and Asn-13, are not glycosylated in some of the hCG variants studied.
Most studies on hCG glycosylation have been performed on hCG prepared from pooled pregnancy urine (Kessler, Mise, et al., 1979; Kessler, Reddy, et al., 1979; Weisshaar et al., 1991; Laidler et al., 1995; Liu and Bowers, 1997), and thus, the glycan structures observed can be expected to represent the average oligosaccharide composition of pregnancy hCG. In the commercial hCG preparation used in the present study, a fairly high proportion of core-2-associated P14 was observed, indicating that the preparation was mainly derived from first-trimester urine. The choriocarcinoma cell line JEG-3 is a potential source of an hCGh standard for assays based on MAb B152. The N-glycan structures of JEG-3-derived hCG resembled that from early pregnancy, including the degree of fucosylation and the content of triantennary glycan structures. The proportion of core-2 structures in P14 and P15 was as high as in hCG from cancer patients. This explains the strong reactivity of malignancy-related hCG with MAb B152.
Various isoforms of hCG, including free hCGß and hCGh, have been shown to be associated with malignancies (Alfthan, Haglund, Dabek, et al., 1992; Alfthan, Haglund, Roberts, et al., 1992; Elliott et al., 1997). We, therefore, isolated free hCGß and hCGh from urine and purified hCG from serum and urine of a patient with an invasive mole. Interestingly, the oligosaccharides of hCG isolated from serum displayed the same glycan pattern as hCG purified from urine. Therefore, the heterogeneous oligosaccharide structures of urinary hCG are not modified by additional degradation in urine. In hCGh isolated with MAb B152, an increase in core-2-containing P14 was seen, but no other difference in glycan composition, except a small increase in triantennary Asn-30-associated glycan, was observed. The degree of fucosylation in the Asn-13-linked glycan was not changed. Therefore, the hCGh detected by MAb B152 differs markedly from "normal" hCG only with respect to the Ser-132-associated glycan. This confirms the finding that B152 recognizes this core-2 O-glycan and surrounding polypeptide structures (Birken et al., 2003).
Interestingly, the oligosaccharide structures of free hCGß resembled that of hCGh with a high proportion of triantennary N-linked glycans and core-2 type O-glycans. In addition, the fucosylation of the Asn-13-linked glycan was markedly elevated. This observation could indicate that the large glycans might inhibit the association of hCGß with the -subunit. The observation of malignancy-related "hyperglycosylation" of free hCGß is in agreement with the fact that the proportion of hCGß is increased in patients with throphoblastic cancer (Alfthan, Haglund, Dabek, et al., 1992; Alfthan, Haglund, Roberts, et al., 1992). Because fucosylated N-glycans react with LCA lectin, the malignancy-associated hCGß could possibly be detected with a lectin assay.
In conclusion, we characterized six site-specific glycan structures of hCG by differential expression analysis of LC–MS data on hCG isolated from pregnant women and patients with malignant diseases. Major differences were observed in the fucosylation degree of Asn-13-linked glycans as well as in the occurrence of core-2 O-glycans in Ser-127, Ser-132, and Ser-138. Furthermore, larger glycan structures and a higher degree of fucosylation of Asn-13-linked glycans were observed in free hCGß than in heterodimeric hCG from the same patient. Analysis of hCGh isolated by affinity chromatography with MAb B152 confirmed that this antibody recognizes a core-2 glycan on Ser-132 and on the surrounding peptide structures.
|
Materials and methods |
|---|
|
TopAbstractIntroductionResultsDiscussionMaterials and methodsAcknowledgmentsConflict of interest statementReferences |
|---|
Patients and samples
hCG was isolated from the urine of two patients with nonseminomatous testicular cancer (stages 1 and 3, respectively) (Sobin and Wittekind, 2002
), one with choriocarcinoma (stage IIIA) and one with an invasive mole (stage IB) (Benedet and Pecorell, 2000). Urine from one pregnant woman was collected at 5 and 7 weeks and from two women at 35 weeks after fertilization. Informed consent was obtained from all patients, and the study was approved by the institutional ethics committee.
NaN3 was added to a concentration of 0.5 g/L into the urine samples, which were stored at 4°C. hCG in the samples was quantitated by a time-resolved immunofluorometric assay performed as described (Pettersson et al., 1983; Alfthan, Haglund, Dabek, et al., 1992; Alfthan, Haglund, Roberts, et al., 1992).
The choriocarcinoma cell line, JEG-3 [American Type Culture Collections (ATCC)], was cultured according to the guidelines provided. The hCG-containing medium was centrifuged to remove cells and stored at –20°C. Pregnyl was purchased from Organon Technica, and lot 167823 was used for analysis.
Purification of hCG
hCG was purified from 100 to 1000 mL of urine, containing 3–250 nmol/L of hCG. Urine was clarified by filtering through Nalgene disposable filter unit with 0.45-µm pore size, and the filtrate was applied to an anti-hCG affinity column (1 mL) with a flow rate of 200 µL/min. MAb 6G5, raised in the laboratory as described earlier (Alfthan, Haglund, Dabek, et al., 1992; Alfthan, Haglund, Roberts, et al., 1992), was coupled with CNBr-activated Sepharose 4B (Pharmacia) according to the instructions of the manufacturer. Before immunoaffinity chromatography, the samples were passed through a precolumn of equal size with an unrelated MAb. The columns were equilibrated with 50 mmol/L sodium phosphate (pH 7.4). After sample application, the column was washed with 40 mL of 10 mmol/L ammonium acetate (pH 4.5) and the bound protein eluted with 15 mL of 3 mol/L acetic acid. Fractions of 1 mL were collected and immediately neutralized with NH4OH. Fractions containing hCG were pooled, concentrated using a Centricon centrifugal device with a cutoff of 10 kDa (Millipore), and fractionated by gel filtration on a Sephacryl S-100 HR (Amersham Biosciences) 16 x 700 mm column in 0.1 mol/L ammonium bicarbonate (pH 8) with a flow rate of 200 µL/min. The content of intact hCG dimer and hCGß in 800-µL fractions was determined by time-resolved immunofluorometric assays, and the fractions containing hCG were freeze-dried and stored at –20°C.
To separate free hCGß from hCG, the urine was first applied to an affinity column containing MAb 9C11, raised in the laboratory, and shown to be specific to free hCGß (Alfthan, Haglund, Dabek, et al., 1992; Alfthan, Haglund, Roberts, et al., 1992). Bound hCGß was eluted and further treated as described above. The flowthrough fraction was further applied to the anti-hCG affinity column containing MAb 6G5. hCGh was purified from urine by immunoaffinity chromatography with MAb B152 (Birken et al., 1999) column.
hCG was purified from a 5-mL serum sample from a patient with invasive mole containing 1800 nmol/L of hCG. The sample was first purified by gel filtration on a 16 x 700 mm column packed with Sephacryl S-200 HR (Amersham Biosciences) in 0.1 mol/L ammonium bicarbonate (pH 8) at flow rate of 250 µL/min. Fractions containing proteins <70 kDa were collected and subjected to immunoaffinity purification as described above. hCG was purified from the growth medium of JEG-3 cell line using MAb 6G5.
In-liquid alkylation and digestion of hCGß
Purified hCG was reduced with dithiotreitol and alkylated with 4-vinyl pyridine (Aldrich) in 6 mol/L guanidine hydrochloride, 2 mmol/L EDTA, 0.5 M Tris (pH 7.5). The alkylated hCG with dissociated - and ß-subunits was desalted by RP-HPLC on a C4 column (symmetry C4, 3.9 x 20 mm, 300 Å, 5 µm, Waters) and eluted with a linear gradient of acetonitrile (0–80% in 30 min) in 0.1% trifluoroacetic acid (TFA). hCGß-containing fractions were pooled, dried, and subjected to trypsin digestion using 10% w/w sequencing grade trypsin (Promega) in 10 mmol/L ammonium bicarbonate at 37°C for 16 h.
For digestion with Glu-C (V8, Roche), 5% w/w protease was added in 50 mmol/L ammonium acetate (pH 4), and digestion was carried out for 4 h at room temperature. After each incubation hour, 2% w/w V8 was added. The C-terminal portion of hCG was separated by RP chromatography on a C4 column (symmetry C4, 2 x 50 mm, 300 Å, 3.5 µm, Waters). Proteolytic peptides were eluted with a linear gradient of acetonitrile (0–80% in 30 min) in 0.1% TFA with a flow rate of 0.2 mL/min. Fractions containing the C-terminal portion of hCGß were collected.
For the removal of N-glycans, 0.5 U recombinant PNGase F (Roche) was added to 10 µg of hCG in 0.1 M sodium phosphate buffer (pH 7.8). Deglycosylation was carried out for 48 h at 37°C. Another 0.5 U of PNGase F was added after 24-h incubation. The removal of N-glycans was confirmed by SDS–PAGE in 12% w/v gel (Laemmli, 1970).
MS
For LC–MS analysis, the peptides were separated by RP-HPLC on a CapLC instrument (Waters) with a 0.075 x 150-mm C18 column (Atlantis dC18, 100 Å, 3 µm, Waters), which was eluted with a linear gradient of acetonitrile (5–50% in 30 min) in 0.1% formic acid. The flow rate was 0.3 µL/min, and the eluent was directly injected into a quadrupole/time-of-flight hybrid mass spectrometer (Q-TOF Micro, Waters) equipped with an ESI source. The mass spectrometer was calibrated using 2 pmol/µL of glufibrinogenic peptide B fragments as a standard.
MSMS fragmentation spectra of the peptides were acquired by colliding the doubly or triply charged precursor ions with argon collision gas at accelerating voltages of 30–45 V.
For ESI–MS analysis of the chromatographically purified C-terminal fragment of hCGß, the polypeptide was injected into the mass spectrometer (Q-TOF Micro, Waters) via a nanoflow interface with a Hamilton-syringe pump with a flow rate of 0.3 µL/min. The mass spectrometer was calibrated using 400 fmol/µL of myoglobin (Sigma) as a standard.
Differential data analysis
Mass spectra collected during the LC–MS separation of hCGß peptides were exported into ASCII text files using the DataBridge of the MassLynx software (Waters). The text files were imported into DeCyder MS software (GE Healthcare), where different elution profiles were visualized as 2D maps with m/z on the y-axis and retention time on the x-axis as well as three-dimensional (3D) graphs with intensity based on ion count as z-axis. Peptide eluting at specific time point and represented by many different charge states was detected automatically, its charge states were assigned, MS signal ion counts of different charge states were integrated, and the actual mass of the peptide was deconvoluted. The ion counts of all different charge states of the same peptide were taken into account to calculate the total intensity of deconvoluted mass in the range of 500–6000 Da. PepMatch module of DeCyder MS software was used to align peptides from different LC–MS runs and for semiquantitative differential analysis of the peptides. Ion counts over the spot areas manually selected were integrated, and comparison of the integrated ion counts between samples was performed.
|
Acknowledgments |
|---|
|
TopAbstractIntroductionResultsDiscussionMaterials and methodsAcknowledgmentsConflict of interest statementReferences |
|---|
We thank Ms Anne Ahmanheimo, Ms Taina Grönholm, Ms Maarit Leinimaa, Ms Jenni Sievänen, and Ms Helena Taskinen for skillful technical assistance and Ms Ansa Karlberg for secretarial help. The financial support of the Finnish Academy of Sciences and European Union (LSHC-CT-2004–503011) is gratefully acknowledged.
|
Conflict of interest statement |
|---|
|
TopAbstractIntroductionResultsDiscussionMaterials and methodsAcknowledgmentsConflict of interest statementReferences |
|---|
None declared.
|
Abbreviations |
|---|
ESI, electrospray ionization; Fuc, fucose; hCG, human chorionic gonadotropin; hCGh, hyperglycosylated hCG; Hex, hexose; HexNAc, N-acetylhexosamine; HPLC, high performance liquid chromatography; LC, liquid chromatography; MAb, monoclonal antibody; MS, mass spectrometry; MSMS, tandem mass spectrometry; NeuAc, N-acetylneuraminic acid; PNGase F, peptide-N-glycosidase F; RP, reverse phase; SDS–PAGE, sodium dodecyl sulfate–polyacrylamide gel electrophoresis
|
References |
|---|
|
TopAbstractIntroductionResultsDiscussionMaterials and methodsAcknowledgmentsConflict of interest statementReferences |
|---|
Alfthan H., Haglund C., Dabek J., Stenman U.H. (1992) Concentrations of human choriogonadotropin, its beta-subunit, and the core fragment of the beta-subunit in serum and urine of men and nonpregnant women. Clin. Chem. 38:1981–1987.[Abstract]
Alfthan H., Haglund C., Roberts P., Stenman U.H. (1992) Elevation of free beta subunit of human choriogonadotropin and core beta fragment of human choriogonadotropin in the serum and urine of patients with malignant pancreatic and biliary disease. Cancer Res. 52:4628–4633.
Aoyagi Y., Isemura M., Suzuki Y., Sekine C., Soga K., Ozaki T., Ichida F. (1985) Fucosylated alpha-fetoprotein as marker of early hepatocellular carcinoma. Lancet 2:1353–1354.[Web of Science][Medline]
Benedet J. and Pecorell S. (2000) Staging Classifications and Clinical Practice Guidelines of Gynaecologic CancersElsevier FIGO Committe on Gynecologic Oncology.
Birken S., Krichevsky A., O’Connor J., Schlatterer J., Cole L., Kardana A., Canfield R. (1999) Development and characterization of antibodies to a nicked and hyperglycosylated form of hCG from a choriocarcinoma patient: generation of antibodies that differentiate between pregnancy hCG and choriocarcinoma hCG. Endocrine 10:137–144.[CrossRef][Web of Science][Medline]
Birken S., Yershova O., Myers R.V., Bernard M.P., Moyle W. (2003) Analysis of human choriogonadotropin core 2 o-glycan isoforms. Mol. Cell. Endocrinol. 204:21–30.[CrossRef][Web of Science][Medline]
Carlsen R.B., Bahl O.P., Swaminathan N. (1973) Human chorionic gonadotropin. Linear amino acid sequence of the beta subunit. J. Biol. Chem. 248:6810–6827.
Cole L.A., Kardana A., Andrade-Gordon P., Gawinowicz M.A., Morris J.C., Bergert E.R., O’Connor J., Birken S. (1991) The heterogeneity of human chorionic gonadotropin (hCG). III. The occurrence and biological and immunological activities of nicked hCG. Endocrinology 129:1559–1567.
Cole L.A., Shahabi S., Oz U.A., Bahado-Singh R.O., Mahoney M.J. (1999) Hyperglycosylated human chorionic gonadotropin (invasive trophoblast antigen) immunoassay: a new basis for gestational Down syndrome screening. Clin. Chem. 45:2109–2119.
Elliott M.M., Kardana A., Lustbader J.W., Cole L.A. (1997) Carbohydrate and peptide structure of the alpha- and beta-subunits of human chorionic gonadotropin from normal and aberrant pregnancy and choriocarcinoma. Endocrine 7:15–32.[Medline]
Gervais A., Hammel Y.A., Pelloux S., Lepage P., Baer G., Carte N., Sorokine O., Strub J.M., Koerner R., Leize E., Van Dorselaer A. (2003) Glycosylation of human recombinant gonadotrophins: characterization and batch-to-batch consistency. Glycobiology 13:179–189.
Hakomori S. (2002) Glycosylation defining cancer malignancy: new wine in an old bottle. Proc. Natl. Acad. Sci. U. S. A. 99:10231–10233.
Harvey D.J. (2001) Ionization and collision-induced fragmentation of N-linked and related carbohydrates using divalent cations. J. Am. Soc. Mass Spectrom. 12:926–937.[CrossRef][Web of Science][Medline]
Jacoby E.S., Kicman A.T., Laidler P., Iles R.K. (2000) Determination of the glycoforms of human chorionic gonadotropin beta-core fragment by matrix-assisted laser desorption/ionization time-of-flight mass spectrometry. Clin. Chem. 46:1796–1803.
Kelly L.S., Kozak M., Walker T., Pierce M., Puett D. (2005) Lectin immunoassays using antibody fragments to detect glycoforms of human chorionic gonadotropin secreted by choriocarcinoma cells. Anal. Biochem. 338:253–262.[CrossRef][Web of Science][Medline]
Kessler M.J., Mise T., Ghai R.D., Bahl O.P. (1979) Structure and location of the O-glycosidic carbohydrate units of human chorionic gonadotropin. J. Biol. Chem. 254:7909–7914.
Kessler M.J., Reddy M.S., Shah R.H., Bahl O.P. (1979) Structures of N-glycosidic carbohydrate units of human chorionic gonadotropin. J. Biol. Chem. 254:7901–7908.
Kobata A. and Takeuchi M. (1999) Structure, pathology and function of the N-linked sugar chains of human chorionic gonadotropin. Biochim. Biophys. Acta 1455:315–326.[Medline]
Kovalevskaya G., Birken S., Kakuma T., Ozaki N., Sauer M., Lindheim S., Cohen M., Kelly A., Schlatterer J., O’Connor J.F. (2002) Differential expression of human chorionic gonadotropin (hCG) glycosylation isoforms in failing and continuing pregnancies: preliminary characterization of the hyperglycosylated hCG epitope. J. Endocrinol. 172:497–506.[Abstract]
Laemmli U.K. (1970) Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 227:680–685.[CrossRef][Medline]
Laidler P., Cowan D.A., Hider R.C., Keane A., Kicman A.T. (1995) Tryptic mapping of human chorionic gonadotropin by matrix-assisted laser desorption/ionization mass spectrometry. Rapid Commun. Mass Spectrom. 9:1021–1026.[CrossRef][Web of Science][Medline]
Liu C. and Bowers L.D. (1997) Mass spectrometric characterization of the beta-subunit of human chorionic gonadotropin. J. Mass Spectrom. 32:33–42.[Medline]
Lustbader J.W., Lobel L., Wu H., Elliott M.M. (1998) Structural and molecular studies of human chorionic gonadotropin and its receptor. Recent Prog. Horm. Res. 53:395–424 discussion 424–395.[Medline]
Mizuochi T., Nishimura R., Derappe C., Taniguchi T., Hamamoto T., Mochizuki M., Kobata A. (1983) Structures of the asparagine-linked sugar chains of human chorionic gonadotropin produced in choriocarcinoma. Appearance of triantennary sugar chains and unique biantennary sugar chains. J. Biol. Chem. 258:14126–14129.
Pettersson K., Siitari H., Hemmila I., Soini E., Lovgren T., Hanninen V., Tanner P., Stenman U.H. (1983) Time-resolved fluoroimmunoassay of human choriogonadotropin. Clin. Chem. 29:60–64.
Puisieux A., Bellet D., Troalen F., Razafindratsita A., Lhomme C., Bohuon C., Bidart J.M. (1990) Occurrence of fragmentation of free and combined forms of the beta-subunit of human chorionic gonadotropin. Endocrinology 126:687–694.
Skold K., Svensson M., Kaplan A., Bjorkesten L., Astrom J., Andren P.E. (2002) A neuroproteomic approach to targeting neuropeptides in the brain. Proteomics 2:447–454.[CrossRef][Web of Science][Medline]
Sobin L. and Wittekind C. (2002) TNM Classification of Malignant Tumours. International Union Against Cancer (UICC)(Wiley-Liss, New York).
Stenman U.H., Alfthan H., Hotakainen K. (2004) Human chorionic gonadotropin in cancer. Clin. Biochem. 37:549–561.[CrossRef][Web of Science][Medline]
Weisshaar G., Hiyama J., Renwick A.G. (1991) Site-specific N-glycosylation of human chorionic gonadotrophin—structural analysis of glycopeptides by one- and two-dimensional, 1H NMR spectroscopy. Glycobiology 1:393–404.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  J. Krudysz-Amblo, M. E. Jennings II, K. G. Mann, and S. Butenas Carbohydrates and Activity of Natural and Recombinant Tissue Factor J. Biol. Chem., January 29, 2010; 285(5): 3371 - 3382. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 G. Zimmermann, W. Ackermann, and H. Alexander Epithelial Human Chorionic Gonadotropin Is Expressed and Produced in Human Secretory Endometrium During the Normal Menstrual Cycle Biol Reprod, May 1, 2009; 80(5): 1053 - 1065. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||