Clinical Chemistry
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

Clinical Chemistry 43: 1172-1181, 1997;
This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Web of Science (6)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Liu, C.
Right arrow Articles by Bowers, L. D.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Liu, C.
Right arrow Articles by Bowers, L. D.
Related Collections
Right arrow Drug Monitoring and Toxicology
Right arrow Endocrinology and Metabolism
(Clinical Chemistry. 1997;43:1172-1181.)
© 1997 American Association for Clinical Chemistry, Inc.

Articles

Mass spectrometric characterization of nicked fragments of the ß-subunit of human chorionic gonadotropin

Chuanliang Liu1 and Larry D. Bowersa

Athletic Drug Testing and Toxicology Laboratory, Department of Pathology and Laboratory Medicine, Indiana University Medical Center, 635 Barnhill Dr., Indianapolis, IN 46202-5120.
1 Present address: Pacific Northwest National Laboratory, Richland, WA.
a Author for correspondence. Fax (317)278-2018; e-mail lbowers{at}iupui.edu


   Abstract
Top
 Abstract
Introduction
 Materials and Methods
 Results
 Discussion
 References
 
We describe the use of an HPLC/MS technique for the characterization of nicked fragments of hCG ß-subunit. After reductive alkylation of the nicked hCG ß-subunit with vinylpyridine, endoproteinase Glu-C or trypsin was used to digest the protein to produce peptides that could be analyzed by HPLC/electrospray ionization MS. Human leukocyte elastase digestion was used to produce an experimentally nicked hCG. Two nicking sites were observed, between amino acids 42Thr and 43Arg and between 44Val and45Leu. The former site has not been previously reported for elastase digestion. The structures of the fragments were confirmed by HPLC/MS after removal of the oligosaccharide by direct mass measurement and by mass determination of their proteolytic digests. Without the glycopeptidase treatment, the microheterogeneity of the two N-linked oligosaccharides could be deduced from the spectra of the proteolytic fragments. Nicking with elastase was found to alter the oligosaccharide structures. Nicked ß-subunit samples isolated from the urine of choriocarcinoma patients were also analyzed and the location of the nicking site(s) agreed with that determined by classical techniques. Important differences in the oligosaccharide structures were also observed in these samples, including the presence of triantennary oligosaccharides not found in hCG from healthy subjects. These findings demonstrate the potential of HPLC/MS for characterization of glycoprotein standard preparations.


Key Words: indexing terms: chromatography, liquid • mass spectrometry • peptide sequencing • proline • oligosaccharide microheterogeneity


   Introduction
Top
 Abstract
 Introduction
Materials and Methods
 Results
 Discussion
 References
 
Human chorionic gonadotropin (hCG)1 is a glycoprotein hormone produced by the trophoblast in pregnancy and in trophoblastic disease (1). Measurement of hCG concentrations is currently used to examine the progress of pregnancy and diagnose trophoblastic diseases. In addition to intact hCG, free {alpha}-subunit, free ß-subunit, and ß-core fragment have also been found in urine and serum of different individuals (2)(3)(4)(5)(6). Nicked hCG ß-subunit has been commonly observed in the urine and serum of trophoblastic disease patients (7). Nicking is an in vivo proteolytic bond cleavage that results in relatively minor changes in protein structure, given the tightly folded tertiary structure or intrapeptide disulfide bonds. Nicking has been reported to occur at several different sites in the hCG ß-subunit, primarily in the region between amino acids ß40 and ß50, i.e., between amino acid positions 40 and 50 from the N-terminal end of the ß subunit. The site(s) and extent of nicking vary from sample to sample (7)(8). Birken et al. suggested that nicking occurred only between ß47 and ß48 in normal pregnancy, while nicking between ß44 and ß45 might be present in choriocarcinoma or abnormal pregnancies (9). They postulated that nicking could be caused by human leukocyte elastase (hLE). Subsequently, Kardana and Cole isolated and characterized other enzymes that could also be involved in the nicking process (10).

Nicking has substantial effects on the biological activity of hCG. Cole et al. reported that a single nick between ß44 and ß45 or between ß47 and ß48 could reduce the steroidogenic activity of hCG to ~24% of its original value (7). Similarly, antibodies directed toward intact hCG ß-subunit have a much lower affinity to nicked ß-subunit (11). These data suggest that nicked hCG has a substantial role in accurate quantification and may also provide some important information for clinical diagnosis.

Glycoprotein preparations used for calibration of immunoassays or as antigens for developing new immunoassays should be characterized with methods that can detect modifications to both the peptide and oligosaccharide portions of the molecule. We recently reported an HPLC/MS method for characterizing the ß-subunit of hCG, and have proposed its application for both rapid characterization of candidate peptide standards for hCG (12) and for forensic confirmation of hCG in presumptive cases of misuse for performance enhancement in sports (13). These earlier reports documented that the structural features of hCG could be readily detected on pyridylethylated tryptic fragments. We report here the extension of this technique to the characterization of nicked hCG.


   Materials and Methods
Top
 Abstract
 Introduction
 Materials and Methods
Results
 Discussion
 References
 
Reagents.
hLE (EC 3.4.21.37), endoproteinase Glu-C (cat. no. P-6181), glycopeptidase F (EC 3.2.2.18 and 3.5.1.52), hCG (cat. no. C-5297), and trypsin (EC 3.4.21.4; treated with L-1-tosylamide-2-phenylethylchloromethyl ketone to remove chymotrypsin) were purchased from Sigma (St. Louis, MO). HPLC-grade acetonitrile was obtained from Baxter (Muskegon, MI). 4-Vinylpyridine, dithiothreitol, and all the other chemicals (ACS-reagent grade or better) were also from Sigma and used as received. Water was purified through an Ultra-Pure water system (Millipore, Pleasanton, CA).

Instrumentation.
The HPLC system for isolation of peptides used a Hewlett-Packard (Little Falls, DE) Model 1090L solvent-delivery system equipped with a Spectra-Physics (San Jose, CA) Spectra 100 UV detector operated at 215 nm. A Vydac (Hesperia, CA) 218TP C18 column (4.6 x 150 mm) was used for the separations. The peptides were eluted with a linear gradient of 95/5 (by vol) solvent A/solvent B to 50/50 A/B in 60 min at a flow rate of 1 mL/min, where solvent A was trifluoroacetic acid (TFA), 1 mL/L in water, and solvent B was TFA, 1 mL/L in acetonitrile.

The mass spectrometer was a PE-Sciex API-IIIPlus triple quadrupole mass spectrometer equipped with an Ionspray interface (Thornhill, Ontario, Canada). The HPLC system for HPLC/MS analysis consisted of a Beckman (Palo Alto, CA) Model 126 solvent delivery module and a Model 166 variable UV-Vis detection module. A 1 x 150 mm Vydac 218TP C18 column was used for HPLC/MS analysis. The elution was carried out in 60 min with the same linear gradient of solvents A and B as above, but at a flow rate of 50 µL/min. A low-dead-volume Valco three-way tee was used postcolumn to split the column effluent 1:10, with the majority going through 127 µm (i.d.) polyetherether ketone tubing to the UV detector and a minor portion flowing through a 50 µm (i.d.) fused silica tubing to the nebulizing needle tip of the Ionspray interface. About 5 µL of each sample was injected for HPLC/MS analysis. The MS analysis was based on the "collisional-excitation scanning" method of Huddleston et al. (14) and was the same as described previously (12)(13). The Ionspray voltage was 4500 V with zero-grade air nebulizing gas flowing at 0.6 L/min at ~2.8–3.4 kPa (40–50 psi). The flow rate of high-purity curtain gas (N2, 99.999%) was 1.2 L/min. The scan rate was 3.78 s/scan at a step size of 0.5 Da. This system was used for all HPLC/MS analyses unless otherwise specified.

A microelectrospray system capable of sustaining stable electrospray at 100 nL/min was constructed as described by Covey (15) for some experiments, as indicated later. MS analysis was performed in the scan mode, from m/z 600 to m/z 870 with a dwell time of 5 ms and a step size of 0.2 Da. Formic acid was added to the Glu-C digest of peptide M1 (50 mL/L), and the digest was infused into the mass spectrometer at 100 nL/min. Four different orifice voltages were used: 50, 75, 100, and 120 V. The signal intensities of m/z 671, m/z 742, and m/z 866 were measured and studied as described in Results. MS/MS peptide sequencing of the m/z 742 ion was also performed with the microelectrospray system. In this experiment, the triply charged peptide ß48-ß65 (m/z 742) was isolated, with use of decreased mass resolution to obtain maximum intensity. The collision energy was 35 eV, and the collision gas thickness was 9.8 x 1014 molecules/cm1 . The MS/MS spectrum was obtained by scanning from m/z 200 to m/z 1500 with a dwell time of 2 ms and a step size of 0.5 Da.

Sample preparation.
About 500 µg of intact ß-subunit (separated from 1.25 mg of Sigma hCG by reversed-phase HPLC as previously described (12)) was dissolved in 200 µL of 0.1 mol/L Tris/HCl buffer, pH 8.0, containing 2 mmol/L CaCl2. To this solution was added 30 µL of 5 U/mL hLE (in the same buffer), and the solution was incubated at 37 °C to nick the hCG (9). The reaction approached completion after 48 h of incubation. The two major nicked species were separated from the reaction mixture by reversed-phase HPLC, collected, and dried in a Savant (Farmingdale, NY) SpeedVac.

The dried fraction of nicked materials from above was redissolved in 200 µL of pH 8.4 reduction buffer (6 mol/L guanidine hydrochloride, 0.5 mol/L Tris, and 2 mmol/L EDTA), reduced with dithiothreitol, and alkylated with 4-vinylpyridine (12). The pyridylethylated sample was desalted by reversed-phase HPLC and digested with either trypsin or endoproteinase Glu-C; the proteolytic digest was analyzed by HPLC/MS.

In some experiments, pyridylethylated nicked fragments were separated by HPLC and collected. Two N-terminal nicked fragments (see Results) were treated with glycopeptidase F, either with or without prior sialic acid removal. Sialic acid was removed by acidic hydrolysis in 0.01 mol/L HCl for 70 min at 90 °C. The glycopeptidase F-treated pyridylethylated N-terminal fragments were directly analyzed by HPLC/MS. Portions of these samples were also digested with trypsin and analyzed by HPLC/MS.

Two samples of nicked hCG isolated from human urine extracts, designated C5 and M1, were reduced, pyridylethylated, and digested with endoproteinase Glu-C as described above. The C5 sample was intact hCG, so we dissociated the subunits with TFA and separated them by HPLC (12). Three peaks corresponding to nicked ß-subunit were collected and analyzed. In some experiments a mobile phase containing formic acid was used to increase sensitivity (16). Sample M1 was hCGß and was analyzed without further purification. The Glu-C digest of C5 was analyzed by HPLC/MS and that of M1 was analyzed by microelectrospray MS and MS/MS.


   Results
Top
 Abstract
 Introduction
 Materials and Methods
 Results
Discussion
 References
 
The ß-subunit of hCG was isolated by HPLC and incubated with hLE. The reaction was monitored by reversed-phase HPLC, and >80% of the hCGß was hydrolyzed after 48 h of incubation at 37 °C, as indicated by the diminution of the intact hCGß peak (retention factor, k' = 19.5). A major peak, which from its shape appeared to consist of at least two components, was found in the chromatogram (k' = 15), beginning at 4 h of incubation. Attempts to resolve the two components by varying the gradient conditions were unsuccessful, so we collected the entire peak for further analysis. Several minor peaks (k' = 19, 19.5, and 20.5) were apparent only after 48 h of incubation. No further work was performed on the minor peaks.

After pyridylethylation of the major peak components, three peaks (P1, P2, and P3) were observed in the HPLC/MS chromatogram (Fig. 1 ), although the shape of the peak eluting at 40 min suggested the coelution of two peptides. The mass spectra of all three peaks were consistent with the presence of carbohydrates by the "collisional excitation scanning" method of Huddleston et al. (14). The pattern observed in the molecular ion region for P1 and P2 was consistent with the presence of multiple oligosaccharide moieties. A similar spectrum was obtained from the third peak, although the ions in the high-mass region were less intense. The predicted size of the nicked fragments, based on reported nicking positions (mainly in the ß40-ß50 range), suggested that the two early-eluting peaks should be the N-terminal fragments and that P3 should contain the two corresponding C-terminal fragments (9). The unknown peptide chain length and microheterogeneity of carbohydrates made it impossible to assign a structure to these peaks.



View larger version (14K):
[in this window]
[in a new window]
  		Figure 1. HPLC/MS chromatogram of reductively pyridylethylated nicked ß-subunit from elastase digestion.

Peaks P1 and P2 were identified as the fragments on the N-terminal side of the two nicking sites. Peak P3 was a mixture of the two fragments on the C-terminal side of the nicking sites.

Removal of the oligosaccharides from P1 and P2 by glycopeptidase F allowed the direct HPLC/MS detection of the peptide moieties of these two fragments (Fig. 2 ). The molecular masses, calculated from the different charge states produced by electrospray, matched the predicted masses of the sequences ß1–42 (Mr-obs = 5066.6 ± 1.4 amu, Mr-pred = 5065.1 amu) and ß1–44 (Mr-obs = 5322.3 ± 1.8 amu, Mr-pred = 5320.5 amu). When a tryptic digest of deglycosylated P1 was analyzed by HPLC/MS, peaks corresponding to the sequences ß1–8, ß9–20, and ß21–42 were found. Similarly, peaks corresponding to ß1–8, ß9–20, ß21–43, and ß21–44 were observed in the tryptic digest of deglycosylated P2. The presence of a tryptic cleavage site on the carboxylterminal side of ß43 precluded a definitive identification of the nicking site as between ß44 and ß45.



View larger version (19K):
[in this window]
[in a new window]
  		Figure 2. Mass spectra of the deglycosylated peptide collected from peaks P1 (A) and P2 (B) shown in Fig. 1Up .

Direct determination of the mass of the carboxylterminal portion of the nicked molecule was not done because the mass of the peptide exceeded the mass range of our instrument. HPLC/MS of the tryptic digest of P3 contained all of the expected tryptic fragments up to ß114. Similar to intact ß-subunit (12), trypsin cleavage at the C-terminus was not complete, and three predicted C-terminal tryptic fragments (ß115–122, ß123–133, and ß134–145) were not observed. Two tryptic fragments were observed around the nicking sites, ß44–60 and ß45–60, the latter being consistent with a nick between ß44 and ß45. Again, tryptic cleavage on the carboxylterminal side of 43Arg precluded definitive confirmation of the nicking site between ß42 and ß43.

Four peptide fragments around the nicking sites were observed from HPLC/MS of the Glu-C digest of nicked ß-subunits mixture (Fig. 3 ). As shown in Fig. 4 , the molecular masses of these four fragments corresponded to those of ß22–42, ß22–44, ß43–65, and ß45–65, respectively, suggesting the presence of two nicking sites, ß42–43 and ß44–45. The predicted masses of these peptides and the experimental results were in good agreement (Table 1 ). The intense carbohydrate fragment ions in the spectra of ß22–42 and ß22–44 indicated the presence of oligosaccharides in these two peptides. The microheterogeneity of the oligosaccharide attached to 30Asn could be determined from their spectra. Interestingly, the microheterogeneity of the oligosaccharide after the nicking reaction differed from that of intact hCG (12). In intact hCG ß-subunit, two Asn-linked oligosaccharides, N1 and N2 (see Fig. 5 ), were observed. The hLE-nicked ß-subunit displayed a substantial amount of asialyl oligosaccharide (structure N8, Fig. 5 ) along with N1 and N2 in the spectrum of ß22–44.



View larger version (15K):
[in this window]
[in a new window]
  		Figure 3. HPLC/MS chromatogram of Glu-C digest of pyridylethylated hLE-nicked ß-subunits, showing the identity of the sequence found in each peak.



View larger version (21K):
[in this window]
[in a new window]
  		Figure 4. Mass spectra of four fragments in Glu-C digest associated with the nicking site determination: (A) spectra of ß22–42 and ß22–44; (B) spectra of ß43–65 and ß45–65.

•, intact molecular ions; *, vinylpyridine adduct ions.


View this table:
[in this window]
[in a new window]
  		Table 1. Comparison of expected and observed masses for intact and Glu C-fragments of hLE-nicked hCGß.1



View larger version (18K):
[in this window]
[in a new window]
  		Figure 5. Proposed structures for N-linked oligosaccharides attached to hCGß (1)(2)(5)(7)(8)(9).

A sample from urine extract of trophoblastic disease patients, designated C5, was analyzed in the same manner. The HPLC/MS chromatogram of Glu-C digest of the reduced and pyridylethylated sample contained peaks corresponding to ß1–19, ß22–47, and ß48–65 (Fig. 6 ). The microheterogeneity of the oligosaccharide at 30Asn could be observed from the spectrum of ß22–47 (Fig. 7 A) and was much more heterogeneous than that of intact ß-subunit or hLE-nicked ß-subunit. Only biantennary oligosaccharides N1 and N2 were observed in intact ß-subunit, with N1 being the major component. In sample C5, however, ions consistent with triantennary (N6, N7; see Fig. 5Up ), asialyl (N8), and asymmetrical (N10) Asn-linked oligosaccharides were present, as well as the normally observed biantennary oligosaccharides (N1, N2). An even more complex spectrum was observed for ß1–19 of the Glu-C digest, indicating the extreme heterogeneity of the carbohydrate at 13Asn (data not shown).



View larger version (12K):
[in this window]
[in a new window]
  		Figure 6. HPLC/MS chromatogram of Glu-C digest of C5, labeled to indicate the identity of the sequences found in each peak.



View larger version (31K):
[in this window]
[in a new window]
  		Figure 7. (A) Mass spectrum of ß22–47 from Glu-C digest of nicked ß-subunit of C5 sample; (B) mass spectrum of ß48–65 from Glu-C digest of nicked ß-subunit of C5 sample.

The ions resulting from collision-induced dissociation of ß48–65 (precursor ion, •) in the interface region are ß50–65 (product ion y16, {infty}); and ß53–65 (product ion y13, {Delta}). The signal at m/z 577 could contain both quadruply charged ß48–65 and triply charged ß53–65. No quadruply charged ß53–65 is expected or observed. The ion peaks labeled with © and Ø are for vinylpyridine adduct ions.

In the spectrum of the doubly charged Glu-C fragment ß48–65 (Fig. 7BUp ), ions corresponding to intact ß48–65 were observed in high abundance. Two other series of ions were also observed in the spectrum. According to their masses, they could correspond to ß50–65 and ß53–65 of hCGß. HPLC/MS analysis of the tryptic digest of the pyridylethylated C5 sample was used to further confirm the above proposed structures. If there were nicking sites between 49Leu and 50Pro and between 52Leu and 53Pro that led to the formation of ß50–65 and ß53–65 in Glu-C digest, corresponding fragments ß50–60 and ß53–60 should be observed in the tryptic digest. Indeed, ions corresponding to fragments ß50–60 and ß53–60 were identified by HPLC/MS.

A second sample, M1, extracted from trophoblastic patient urine, was also analyzed. After pyridylethylation and Glu-C digestion, ß48–65 was observed. To investigate the source of additional ions from ß48–65, we directly infused the Glu-C digest of M1 into the mass spectrometer, using the microspray interface. The intensity ratios of m/z 671 [(ß50–65)3+] to m/z 742 [(ß48–65)3+] and of m/z 866 [(ß53–65)2+] to m/z 742 were examined as a function of orifice voltage to assess whether the doubly charged ions could be the product of collision-induced dissociation (CID) in the interface region. The intensity ratio of ß48–653+ to ß50–653+ decreased from 0.73 to 0.15 and the intensity ratio of ß48–653+ to ß53–652+ decreased from 0.58 to 0.17 as the orifice voltage was increased from 50 to 100 V, consistent with a greater amount of fragmentation with increased kinetic energy, and thus increased collision energy, in the interface region. When the m/z 742 ion was subjected to CID in the MS/MS mode, m/z 671 and m/z 866 ions were also observed (Fig. 8 ).



View larger version (18K):
[in this window]
[in a new window]
  		Figure 8. MS/MS spectrum of m/z 742 ion showing the fragmentation product ions: m/z 671 and m/z 866.

Shown at top is the amino acid sequence of ß48–65 along with the fragmentation sites that lead to the formation of ions at m/z 671 and m/z 866.


   Discussion
Top
 Abstract
 Introduction
 Materials and Methods
 Results
 Discussion
References
 
We previously reported an HPLC/MS method for characterization of the ß-subunit of hCG (12). The method involved pyridylethylation of the Cys residues after reduction of the disulfide bonds. The addition of the pyridine moiety increased the number of charges on the peptide. Digestion of the pyridylethylated protein with trypsin yielded peptides suitable for monitoring both the amino acid sequence and oligosaccharide composition with the mass spectrometer. N-linked oligosaccharides have been reported at 13Asn and 30Asn, and O-linked oligosaccharides have been reported at 122Ser, 127Ser, 132Ser, and 138Ser (1). Our earlier study (12) confirmed the location of the N-linked oligosaccharides but detected peptide masses consistent with only three O-linked sites. Work is at present underway to resolve this discrepancy. We report here on the extension of this approach to determinations of nicking sites on hCGß.

The ability to rapidly characterize nicked hCG would be of benefit in defining the composition of standard materials for immunoassay. Differentiation of nicked hCG, as well as nicked and native luteinizing hormone ß-subunit, from native hCG also has implications in confirmation methods for athletic drug testing (13). Nicked hCGß can be isolated from commercial preparations. To obtain sufficient amounts of a homogeneous pool of model compound for development of the methodology, we chose to prepare nicked hCGß by using hLE (9). Confirmation of the published hLE nicking reaction would also validate, and demonstrate the advantages of, the HPLC/MS approach, as well as provide an opportunity to evaluate the limits of detection of the technique. We then applied the methodology to analysis of nanomole amounts of two nicked hCGß mixtures isolated from urine.

Our initial approach to determining the nicking site was direct determination of the peptide mass. Because the presence of oligosaccharides can result in ambiguity in the structure determination of nicked fragments, we removed the oligosaccharides from 13Asn and 30Asn on the peptide isolated from peaks P1 and P2. Glycopeptidase F is commonly used to cleave Asn-linked carbohydrate and converts Asn to Asp while releasing the intact sugar chain. Although generally used after the terminal sialic acid residues are removed, the ability of the mass spectrometer to detect changes in oligosaccharide structure allowed us to directly monitor the progress of the reaction. Thus, we studied the reaction of glycopeptidase F with and without sialic acid removal on P1 and P2. After 4 h of glycopeptidase F digestion, most of the oligosaccharides from both P1 and P2 were removed. The main peak in the chromatogram represented the peptide backbone, given that no carbohydrate fragment ion was observed (Fig. 2Up ). A small peak was also observed earlier in the chromatogram, corresponding to the peptide backbone attached to one oligosaccharide moiety. By matching the mass of different carbohydrates reported, this oligosaccharide was found to have a mass corresponding to either N1 or N2, but not both (see Table 2). When this carbohydrate-containing peak was subjected to HCl treatment, the mass difference agreed with the loss of a sialic acid residue, which further confirmed our assignment of this peak. When the glycopeptidase F reaction time was extended to >24 h, the oligosaccharide-containing peak disappeared in both cases. We found no significant difference in glycopeptidase F digestion products with and without sialic acid removal.

As Fig. 2Up shows, a series of ions was produced that differed from the peptide backbone molecular ions by a mass of 105. Given that the molecular mass of vinylpyridine is 105, this increase in mass was thought to result from adduct formation between vinylpyridine and the peptide. However, the adduct proved too labile to determine the site of adduct formation by tandem MS; tryptic digestion of the peptide resulted in dissociation of the adduct; and reducing the reaction time from 10 h to 3 h also eliminated adduct formation. Moreover, no adducts were detected when iodoacetic acid was used as the alkylation reagent.

The position of the nicking site could be readily determined from the mass of the deglycosylated peptide on the N-terminal side of the cleavage point, but confirmation of the site on the C-terminal side was more difficult. The maximum number of charge sites and the expected presence of >100 amino acids and four O-linked oligosaccharides meant that the mass of the protein would exceed the mass range of our instrument. Second, from the microheterogeneity of the oligosaccharides, the same amino acid sequence would give rise to ions with many m/z values, greatly decreasing the sensitivity of the MS detection. Although no intact molecular ions were observed for P3, intense carbohydrate ions were detected, indicating the presence of the oligosaccharide chains.

As mentioned earlier, most of the nicking sites in hCGß were reported between ß40 and ß50. Application of our original method (using trypsin) to determination of the nicking site proved difficult, because the tryptic digestion site on the carboxylterminal side of 43Arg makes definitive identification of the nicking site impossible. For example, a nicking site between ß44 and ß45 would yield two fragments after the trypsin digestion: ß22–43 and ß45–60. Thus two interpretations are possible: a single nicking site between ß44 and ß45 and removal of ß44 during tryptic digestion; or in vivo cleavage between both ß43 and ß44 and ß44 and ß45, with ß44 missing from the peptide backbone. Endoproteinase Glu-C specifically cleaves proteins on the carboxylterminal side of a glutamic acid residue. In hCGß, residues ß3, ß19, ß21, and ß65 are glutamic acids. When intact ß-subunit is digested with Glu-C, a single fragment ß22–65 will be produced. However, if nicking occurs between ß44 and ß45, two fragments will be observed after Glu-C digestion: ß22–44 and ß45–65; this will result in two HPLC peaks rather than one and a mass consistent with each sequence on the mass spectrometer. Therefore, by using HPLC/MS to match the two complementary fragments before and after the nicking site, the nicking site can be determined.

Tryptic digestion of the isolated P3 peptide resulted in fragment sequences consistent with the carboxylterminal side of the nicking site, but the tryptic site precluded definitive assignment of the site. When Glu-C digestion was performed on the mixture of pyridylethylated peptides (P1, P2, and P3) isolated from the major HPLC peak after hLE hydrolysis of hCGß, four peptides were found. The presence of ß22–42 and ß43–65 and of ß22–44 and ß45–65 definitively proves the presence of two nicking sites between 42Thr and 43Arg and between 44Val and 45Leu. The presence of approximately equal amounts of the four peptides suggests that the sites are equally vulnerable to attack by hLE.

Birken et al. (9) hypothesized that hLE was responsible for in vivo nicking of hCGß. They observed specific nicking between 44Val and 45Leu after a short digestion time (2 h), based on N-terminal sequencing results. Digestion for 48 h produced additional nicking between 5Leu and 6Arg, 46Gln and 47Glu, and 48Val and 49Leu. Our results suggest an equal amount of nicking between 42Thr and 43Arg and between 44Val and 45Leu. Because HPLC/MS identified both the N-terminal and C-terminal peptides around the nicking sites, there seems to be little possibility for error. Possibly, the discrepancy reflects the presence of contaminating activity in the hLE. As mentioned above, incubation for 48 h yielded several small peaks with retention similar to the unmodified hCGß that had ~20% of the area of the primary peak. We did not characterize the peptides giving rise to those peaks—which could also partially explain the disagreement in the results. Interestingly, in neither case is the site of hLE cleavage specific to an amino acid, such as the carboxylterminus of Lys or Arg for trypsin, nor to a specific polarity of amino acid.

The ability of the HPLC/MS technique to determine the oligosaccharide distribution of a glycopeptide is also useful. In the study of hLE nicking, we observed that the oligosaccharide distribution changed during proteolysis. Both Glu-C and tryptic digests showed the presence of an ion corresponding to the mass of asialyl biantennary oligosaccharide N8 (see Table 2) on 30Asn that was not present in native hCG digests (12). Treatment of the glycopeptide mixture with HCl to remove sialic acid resulted in conversion to a single species, confirming our hypothesis of the presence of an asialyl oligosaccharide. Although the explanation for this observation is not clear, contaminating neuraminidase activity in the hLE preparation or bacterial contamination during the hLE incubation process, if present, could have caused the observed sialic acid cleavage. The ability of HPLC/MS to detect the changes, whether natural or experimentally induced, is important for verification of the experimental results.

In a second experiment, we observed the deglycosylation of ß1–42 by glycosidase F. Before introduction of the glycosylase, the ß1–42 glycopeptide contained oligosaccharide moieties at 13Asn and 30Asn. Based on their mass, combinations of biantennary oligosaccharide N1 with N2, N3, or N4 were observed as the major species. Glycosidase F is frequently used after removal of the terminal sialic acid residues. The fact that residues other than the disialyl oligosaccharide N1 were preferentially removed seems to support this practice, although the reaction will proceed to completion without removal of the sialic acid. We did not investigate whether one of the glycosylation sites was more accessible than the other, although 13Asn seemed more accessible in our earlier study (12).

The hCG isolated from the urine of a choriocarcinoma patient was 100% nicked between ß47 and ß48, as determined from N-terminal amino acid sequencing results (L. Cole, Yale University, personal communication). The nicking site in C5 was confirmed by the HPLC/MS technique. The chromatographic peak corresponding to ß22–47 was very broad and low in sensitivity. The peak shape reflects the heterogeneous carbohydrate contents at 30Asn (Fig. 7Up ). The presence of the triantennary oligosaccharides N6 and N7, and the extent of the desialylation are definitely unlike native hCG, in which small amounts of triantennary oligosaccharides were found only on 13Asn (12). The presence of extensive modification of the oligosaccharides has been reported earlier (8)(9).

The fragmentation of ß48–65 from C5 during the electrospray ionization process was somewhat unexpected. Fig. 7BUp shows the presence of three ions, corresponding to ß48–65, ß50–65, and ß53–65. Our initial explanation for this observation was nicking between ß49 and ß50 and between ß52 and ß53. The fact that all three species have the same HPLC retention time, however, is not consistent with this hypothesis: The retention times of these species would be expected to differ because of their size differences. Variation of the relative intensity of the ions with the orifice voltage supported the hypothesis that ß50–65 and ß53–65 were fragmentation products of ß48–65. Tandem MS experiments gave results similar to those of the orifice voltage experiment, further documenting the lability of the bond on the N-terminal side of Pro.

Dissociation of the peptide bond on the amino side of a Pro residue is highly favored, as shown in both high-energy tandem double-focusing CID studies and low-collision-energy CID with quadrupole and hybrid instrumentation (17)(18)(19)(20)(21). In general, dissociation of this bond is favored when the peptide contains predominantly neutral amino acids (Ala, Gly, lle, Phe, and Val) (18). Multiply charged peptides show somewhat different behavior, favoring fragmentation remote from charge sites, especially for X–Pro bond cleavage (21). In the present study, addition of an ethylpyridine group to all of the Cys residues produced additional charge sites. Thus fragmentation was not expected, or observed, for either of the internal Pro residues in the ß21–43 tryptic peptide, because on the N-terminal side of each Pro is a pyridylethylated Cys. Interestingly, Loo et al. (21) showed that not all Pro residues have the same propensity to induce cleavage; in peptides containing 2–4 Pro residues, only one may show significant dissociation (21). This suggests that the sequence or three-dimensional structure in the region of Pro may be important.

Inspection of the spectra obtained from the proteolytic digests of hCGß illustrates the complexity of the situation. The ß45–65 peptide from hLE nicking, with the sequence LQGVLPALPQVVCNYRDVRFE, displayed no significant fragmentation at either of the internal Pro residues. Nicking at the 48 position, however, gave rise to substantial dissociation at both of the Pro sites, although the only change to the sequence was the removal of the three N-terminal amino acids (LQG, LeuGlnGly). This dramatic change was responsible for our initial lack of recognition that the ß50–65 and ß53–65 peptides arose from collisionally induced fragmentation. The proximity of the Pro to the N-terminus also does not explain the fragmentation, as demonstrated by the lack of dissociation at any Pro of the deglycosylated tryptic fragment ß134–145 (LPGPSDTPILPQ). Tandem MS of the ß134–145 peptide shows a dominant y2/b10 complementary pair (data not shown), indicating the lability of the 143Leu–144Pro bond. Thus Pro bond lability cannot be predicted a priori, but this possibility should be considered when Pro is present.

The hCGß subunit isolated from the urine of a patient with trophoblastic disease was reported to have 5.3% nicking at ß43–44, 10.8% nicking at ß47–48, and 5% nicking at ß75–76, according to N-terminal amino acid sequencing results (L. Cole, personal communication). We were unable to confirm the nicking site between 75Gly and 76Val. Despite the absence of glutamic acid residues in the sequence after 65Glu, the presence of peptide ß66-ß75 (Mrpred 1161) could have been detected. The relatively small amount of material available, the fact that only 5% of the material was nicked between 75Gly and 76Val, and the use of the microelectrospray of the mixture of peptides all contributed to the lack of detection. Again, use of a tryptic digestion would have given ambiguous results. The other two nicking sites could be confirmed, although the use of the microelectrospray precluded quantitative estimates of the nicking.

The methodology presented here has several advantages over the classic N-terminal amino acid sequencing technique for determination of nicking site(s). The HPLC/MS technique is easy and reliable. Introduction of additional charge sites through pyridylethylation of Cys residues and selection of an appropriate endoproteinase are important factors for detection of the glycopeptides by electrospray MS. In the present case, endoproteinase Glu-C digestion provided cleavage sites that produced readily detectable differences in the nicked species. Although the removal of the oligosaccharides can provide more readily interpretable data, one appealing advantage of MS detection is its ability to demonstrate the microheterogeneity of the oligosaccharides. The difference in oligosaccharide distribution in the hCG from the urine of a patient with trophoblastic disease (C5 sample) demonstrates the utility of the technique. Although direct measurement of hCG from urine has been accomplished (13), the real value of the HPLC/MS technique should be in the rapid characterization of materials for antibody production and immunoassay standardization. The sensitivity of the technique was in the picomole range, comparable with that of the N-terminal amino acid sequencing technique.

The HPLC/MS approach also has limitations, which were apparent from this study. First, microheterogeneity of the oligosaccharides seriously deteriorates the limit of detection of the technique, through both broadening of the HPLC peak and decreased ionization efficiency in the MS interface. This could potentially be overcome by the use of capillary electrophoresis coupled to MS, and studies are underway in our laboratory to explore this technique. Second, the presence of multiple minor components, as was observed with the proteins isolated from human urine, makes detection difficult. This suggests that improved isolation and purification methods would enhance the performance of HPLC/MS. Finally, interpretation of MS results can sometimes be difficult, as exemplified by fragmentation adjacent to some Pro residues in the interface. Nevertheless, we were able to demonstrate that HPLC/MS is capable of characterizing nicked species of hCGß.


   Acknowledgments
 
We express our sincere gratitude to Laurence Cole of Yale University for providing samples of nicked hCG isolated from the urine of choriocarcinoma patients and for many useful scientific exchanges.


   Footnotes
 
1 Nonstandard abbreviations: hCG, human chorionic gonadotropin; hCGß, ß-subunit of hCG; hLE, human leukocyte elastase; MS, mass spectrometry; CID, collision-induced dissociation; and TFA, trifluoroacetic acid.


   References
Top
 Abstract
 Introduction
 Materials and Methods
 Results
 Discussion
 References
 

  1. Birken S, Krichevsky A, O'Connor J, Lustbader J, Canfield R. Chemistry and immunochemistry of hCG, its subunits, and its fragments. Glycoprotein hormones 1989:45-61 Serono Symposia Newport Beach, CA. .
  2. Endo T, Nishimura R, Saito S, Kanazawa K, Nomura K, Katsuno M, et al. Carbohydrate structures of ß-core fragment of human chorionic gonadotropin isolated from a pregnant individual. Endocrinology 1992;130:2052-2058. [Abstract/Free Full Text]
  3. Kato Y, Braunstein G. ß-Core fragment is a major form of immunoreactive urinary chorionic gonadotropin in human pregnancy. J Clin Endocrinol Metab 1988;66:1197-1201. [Abstract/Free Full Text]
  4. Birken S, Armstrong EG, Kolks MAG, Cole LA, Agosto GM, Krichevsky A, et al. Structure of the human chorionic gonadotropin ß-subunit fragment from pregnancy urine. Endocrinology 1988;123:572-583. [Abstract/Free Full Text]
  5. Skarulis MC, Wehmann RE, Nisula BC, Blithe AL. Glycosylation changes in human chorionic gonadotropin and free alpha subunit as gestation progresses. J Clin Endocrinol Metab 1992;75:91-96. [Abstract]
  6. Mann K, Saller R, Hoermann R. Clinical use of hCG and hCGß determinations. Scand J Clin Lab Invest 1993;53(Suppl 216):97-104. [Web of Science]
  7. Cole LA, Kardana A, Ying FC, Birken S. The biological and clinical significance of nicks in human chorionic gonadotropin and its free ß-subunit. Yale J Biol Med 1991;64:627-637. [Web of Science][Medline] [Order article via Infotrieve]
  8. Blithe DL, Akar AH, Wehmann RE, Nisula BC. Purification of ß-core fragment from pregnancy urine and demonstration that its carbohydrate moieties differ from those of native human chorionic gonadotropin-ß. Endocrinology 1988;122:173-180. [Abstract/Free Full Text]
  9. Birken S, Gawinowicz MA, Kardana A, Cole LA. The heterogeneity of human chorionic gonadotropin (hCG). II. Characteristics and origins of nicks in hCG reference standards. Endocrinology 1991;129:1551-1558. [Abstract/Free Full Text]
  10. Kardana A, Cole LA. Human chorionic gonadotropin ß-subunit nicking enzymes in pregnancy and cancer patient serum. J Clin Endocrinol Metab 1994;79:761-767. [Abstract]
  11. Cole LA, Kardana A. Discordant results in human chorionic gonadotropin assays. Clin Chem 1992;38:263-270. [Abstract/Free Full Text]
  12. Liu C, Bowers LD. Mass spectrometric characterization of human chorionic gonadotropin ß-subunit. J Mass Spectrom 1997;32:33-42. [Web of Science][Medline] [Order article via Infotrieve]
  13. Liu C, Bowers LD. Immunoaffinity trapping of urinary human chorionic gonadotropin and its high-performance liquid chromatographic–mass spectrometric confirmation. J Chromatogr B Biomed Appl 1996;687:213-220. [Medline] [Order article via Infotrieve]
  14. Huddleston MJ, Bean MF, Carr SA. Collisional fragmentation of glycopeptides by electrospray ionization LC/MS and LC/MS/MS: methods for selective detection of glycopeptides in protein digests. Anal Chem 1993;65:877-884. [Medline] [Order article via Infotrieve]
  15. Covey T. Characterization and implementation of micro-flow electrospray nebulizers [Abstract]. Proc., 43rd Am Soc Mass Spectrom Conf on Mass Spectrometry and Allied Topics, 1995:669..
  16. Medzihradszky KF, Maltby DA, Hall SC, Setteneri CA, Burlingame AL. Characterization of protein N-glycosylation by reversed-phase microbore liquid chromatography/electrospray mass spectrometry, complementary mobile phases and sequential exoglycosidase digestion. J Am Soc Mass Spectrom 1994;5:350-358. [Web of Science]
  17. Hunt DF, Yates JR, III, Shabanowitz J, Winston S, Hauer CR. Protein sequencing by tandem mass spectrometry. Proc Natl Acad Sci U S A 1986;83:6233-6237. [Abstract/Free Full Text]
  18. Gaskell SJ, Reilly MH. Hybrid tandem mass spectrometry of peptides above mass 1000. Rapid Commun Mass Spectrom 1988;2:188-191. [Medline] [Order article via Infotrieve]
  19. Schwartz BL, Bursey MM. Some proline substitution effects in the tandem mass spectrum of protonated pentaalanine. Biol Mass Spectrom 1992;21:92-96. [Web of Science][Medline] [Order article via Infotrieve]
  20. Bean MF, Carr SA, Thorne GC, Reilly MC, Gaskell SJ. Tandem mass spectrometry of peptides using hybrid and four-sector instruments: a comparative study. Anal Chem 1991;63:1473-1481. [Medline] [Order article via Infotrieve]
  21. Loo J, Edmonds C, Smith R. Tandem mass spectrometry of very large molecules. 2. Dissociation of multiply-charged proline-containing proteins from electrospray ionization. Anal Chem 1993;65:425-438. [Medline] [Order article via Infotrieve]



The following articles in journals at HighWire Press have cited this article:


Home page
 Clin. Chem.Home page
E. S. Jacoby, A. T. Kicman, P. Laidler, and R. K. Iles
Determination of the Glycoforms of Human Chorionic Gonadotropin {beta}-Core Fragment by Matrix-assisted Laser Desorption/Ionization Time-of-Flight Mass Spectrometry
Clin. Chem., November 1, 2000; 46(11): 1796 - 1803.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Web of Science (6)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Liu, C.
Right arrow Articles by Bowers, L. D.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Liu, C.
Right arrow Articles by Bowers, L. D.
Related Collections
Right arrow Drug Monitoring and Toxicology
Right arrow Endocrinology and Metabolism

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS