Analytical methods
PTMs and glycoprotein mass spectrometry, with HCG as the example
Post-translational modifications change mass by exact, computable amounts. This page tabulates them, then follows the heaviest case in our catalog, the glycoprotein hormone hCG, to show why glycoproteins are described by distributions and not by one mass.
The short version
A modification is a change in elemental composition, so it shifts the mass by an amount you can compute and look up. Creasy and Cottrell (2004) describe Unimod as a database of such mass differences derived from elemental compositions. For a short synthetic peptide with an acetyl cap or an amide, one shift explains the mass. For a glycoprotein the modification is a set of sugars that varies from molecule to molecule, so the intact hormone shows a family of masses. Our guide to peptide modifications covers the common synthetic ones; this page adds the biological ones and the glycans.
Mass shifts of common modifications
| Modification | Composition change | Monoisotopic shift (Da) | Average shift (Da) | Typical location |
|---|---|---|---|---|
| Phosphorylation | + HPO3 | +79.9663 | +79.98 | Ser, Thr, Tyr |
| Sulfation | + SO3 | +79.9568 | +80.06 | Tyr |
| Acetylation | + C2H2O | +42.0106 | +42.04 | N-terminus, Lys |
| Methylation | + CH2 | +14.0157 | +14.03 | Lys, Arg |
| Trimethylation | + C3H6 | +42.0470 | +42.08 | Lys |
| Oxidation | + O | +15.9949 | +16.00 | Met, Trp |
| Deamidation | + O − NH | +0.9840 | +0.98 | Asn, Gln |
| C-terminal amidation | − O + NH | −0.9840 | −0.98 | C-terminus |
| Pyroglutamate from Gln | − NH3 | −17.0265 | −17.03 | N-terminal Gln |
| Disulfide bond | − 2 H | −2.0157 | −2.02 | Cys pair |
| Hexose (Hex) | + C6H10O5 | +162.0528 | +162.14 | glycan residue |
| N-acetylhexosamine (HexNAc) | + C8H13NO5 | +203.0794 | +203.19 | glycan residue |
| Deoxyhexose (fucose) | + C6H10O4 | +146.0579 | +146.14 | glycan residue |
| N-acetylneuraminic acid (NeuAc) | + C11H17NO8 | +291.0954 | +291.26 | glycan residue |
Two pairs are worth noticing. Deamidation (+0.9840 Da) and C-terminal amidation (-0.9840 Da) are equal and opposite, so a low-resolution spectrum may not separate a mixture of an amide and an acid. And trimethylation (+42.0470 Da) and acetylation (+42.0106 Da) share a nominal mass but differ by 36.4 mDa. See how peptides degrade for the unintended modifications.
Phosphorylation versus sulfation
HPO3 and SO3 have the same nominal mass and differ by 9.52 mDa, which is 9.5 ppm at m/z 1000. Bossio and Marshall (2002) report resolving phosphorylated and sulfated forms of a CCK-8 peptide at a resolving power above 400,000, and state that the separation needs a minimum of about 118,000 for that pair. Chen et al. (2018) describe a second route: sulfotyrosine undergoes neutral loss under HCD and CID and is largely stable under ETD, while phosphotyrosine is stable under all of these, so fragmentation behaviour can distinguish them. The practical point for any COA is that a mass match at low resolution cannot tell the two apart.
HCG: a subunit pair with eight glycosylation sites
Human chorionic gonadotropin is a heterodimer of two non-covalently linked subunits. Al Matari et al. (2021) describe the alpha subunit as having two N-glycosylation sites and the beta subunit as having two N- and four O-glycosylation sites, which leads to numerous glycoforms; our HCG page lists the UniProt sequence data. Computing the polypeptide masses from the UniProt sequences (P01215 residues 25 to 116 with five disulfide bonds, and P0DN86 residues 21 to 165 with six) gives:
| Subunit | Residues | Composition | Monoisotopic (Da) | Average (g/mol) |
|---|---|---|---|---|
| Alpha | 92 | C437H672N122O134S13 | 10,188.589 | 10,195.7 |
| Beta | 145 | C668H1078N196O203S13 | 15,509.642 | 15,519.9 |
| Both, summed | 237 | 25,715.6 |
Camperi et al. (2020) describe the intact hormone as having a molecular weight of about 35 kDa. If that is right, the carbohydrate accounts for roughly 9,000 g/mol of it, about a quarter of the molecule; the quoted value is approximate, so treat that figure as an order of magnitude and not a measurement.
What the sugars add
A glycan attached to a peptide adds the sum of its monosaccharide residue masses. The compositions below are named by monosaccharide counts, as is conventional, and the masses are computed from the residue formulas. Valmu et al. (2006) report that the major N-glycan on hCG beta at Asn-13 and Asn-30 is a biantennary complex-type structure, with triantennary structures at Asn-30 and fucosylation at Asn-13 increased in cancer-derived hCG beta, and that O-glycans at Ser-121, Ser-127, Ser-132 and Ser-138 include core-2 glycans at Ser-121 and core-1 glycans at Ser-138 (numbering as in that paper, which corresponds to the UniProt precursor numbering minus 20).
| Glycan type | Composition | Added monoisotopic mass (Da) | Added average mass (Da) |
|---|---|---|---|
| Trimannosyl core | Hex3HexNAc2 | 892.317 | 892.81 |
| Biantennary, no sialic acid | Hex5HexNAc4 | 1622.582 | 1623.48 |
| Biantennary, one NeuAc | Hex5HexNAc4NeuAc1 | 1913.677 | 1914.74 |
| Biantennary, two NeuAc | Hex5HexNAc4NeuAc2 | 2204.772 | 2205.99 |
| Biantennary, core fucose, two NeuAc | Hex5HexNAc4dHex1NeuAc2 | 2350.830 | 2352.13 |
| Triantennary, three NeuAc | Hex6HexNAc5NeuAc3 | 2861.000 | 2862.58 |
| Core-1 O-glycan | Hex1HexNAc1 | 365.132 | 365.33 |
| Core-2 O-glycan | Hex2HexNAc2 | 730.264 | 730.67 |
A worked glycoform ladder (constructed)
To see why an intact spectrum is a family of peaks, take the alpha polypeptide (10,195.7 g/mol) and give each of its two N-glycan sites a biantennary glycan, Hex5HexNAc4 plus up to two sialic acids each. Removing one sialic acid at a time lowers the mass by 291.26 g/mol. The compositions are chosen for the exercise and are not a measurement of any product; Thakur et al. (2009) did observe more than 60 glycoforms of recombinant alpha hCG with up to nine sialic acids.
| Sialic acids | Description | Average mass (g/mol) |
|---|---|---|
| 4 | Both glycans disialylated | 14,607.7 |
| 3 | 1 sialic acid missing | 14,316.4 |
| 2 | 2 sialic acids missing | 14,025.2 |
| 1 | 3 sialic acids missing | 13,733.9 |
| 0 | No sialic acid | 13,442.6 |
The fully disialylated version of this combination is 14,608 g/mol, and real material also varies in neutral sugars, fucose and antenna number, so the observed distribution is wider than this ladder.
Subtracting the peptide: a check against the literature
Jacoby et al. (2000) analysed the hCG beta-core fragment, which comprises two disulfide-linked peptides (beta6 to beta40 and beta55 to beta92), by MALDI-TOF after reduction. They subtracted the mass of the peptide from the observed values to obtain carbohydrate masses, then subtracted monosaccharide residues to predict structures, and reported the nonglycosylated peptide beta55 to beta92 at m/z 4156.8. Computing the [M+H]+ of that reduced peptide from the UniProt beta sequence (mature numbering, residues 55 to 92) with average masses gives 4,156.9, which agrees. The same subtraction logic is how any glycopeptide or glycoprotein mass is read: observed mass minus the polypeptide mass leaves the carbohydrate mass, which you then match to residue combinations from the table above.
Three levels of analysis
| Level | What you learn | What you lose | hCG example |
|---|---|---|---|
| Intact protein or subunit | The distribution of whole-molecule masses, so the number and relative abundance of glycoforms | Which site carries which glycan | Al Matari et al. (2021): nano-RPLC-Orbitrap detected 42 alpha and 33 beta glycoforms. Besson-Magdelain et al. (2026): 76 alpha and 200 beta after switching the mobile phase additive from TFA to difluoroacetic acid and dissociating adducts. Camperi et al. (2020): HILIC-MS of intact hCG, complementary to RPLC. Thakur et al. (2009): capillary electrophoresis-MS found over 60 alpha glycoforms. |
| Glycopeptides after digestion | Site-specific glycan composition | Linkage between sites on the same molecule | Valmu et al. (2006): LC-ESI-MS of tryptic glycopeptides of hCG beta from pregnancy and malignancy. |
| Released glycans | The structures present across the sample, with fragmentation for structure | Where each glycan came from | Ibeto et al. (2020): MALDI-TOF MS and MS/MS of N-glycans from early-pregnancy, late-pregnancy and gestational trophoblastic disease hCG, finding bisected N-glycans and Lewis X antigens. |
Two practical points come from these papers. First, the LC conditions matter to the spectrum: Besson-Magdelain et al. report that TFA, which helps separate glycoforms, suppressed the MS signal and produced complex spectra with many adducts, and that difluoroacetic acid reduced adduct ratios from 30 to 10%. Second, sialic acids are labile and ionise unevenly; Zhang et al. (2018) review that sialoglycan analysis by MS remains challenging because of sialic acid instability, ionisation discrimination and mixed adducts.
What this means for a vial labelled in IU
Our HCG listing is labelled in International Units, an activity unit, and shows the COA as not yet provided. The intact-level literature above shows that hCG appears as many glycoforms and not as one peak, and de Medeiros and Norman (2009) review how the molecular forms of hCG matter for assays. For a COA, that suggests asking for the potency method and result, the unit, and an identity method suited to a glycoprotein. See how to read a COA for the general checklist.
Related
- N- and C-terminal modifications and mass spectrometry
- HCG: sizes, prices and identity data
- Glycoprotein hormone research compounds
- Peptide modifications explained
- HPLC vs LC-MS for peptides: identity and purity workflow
- Check the mass spectrometry on a COA yourself
References
- Al Matari A, Goumenou A, Combès A, et al. (2021). Identification and semi-relative quantification of intact glycoforms of human chorionic gonadotropin alpha and beta subunits by nano liquid chromatography-Orbitrap mass spectrometry. J Chromatogr A 1640:461945. PubMed 33556683 · DOI
- Besson-Magdelain P, Chendo C, Combès A, et al. (2026). Analysis of human chorionic gonadotropin glycoforms by nano-liquid chromatography-mass spectrometry: novel strategies based on mobile phase additive and adduct dissociation. Anal Bioanal Chem 418:5487-5503. PubMed 42321409 · DOI
- Camperi J, Combès A, Fournier T, et al. (2020). Analysis of the human chorionic gonadotropin protein at the intact level by HILIC-MS and comparison with RPLC-MS. Anal Bioanal Chem 412:4423-4432. PubMed 32377867 · DOI
- Thakur D, Rejtar T, Karger BL, et al. (2009). Profiling the glycoforms of the intact alpha subunit of recombinant human chorionic gonadotropin by high-resolution capillary electrophoresis-mass spectrometry. Anal Chem 81:8900-7. PubMed 19817480 · DOI
- Valmu L, Alfthan H, Hotakainen K, et al. (2006). Site-specific glycan analysis of human chorionic gonadotropin beta-subunit from malignancies and pregnancy by liquid chromatography--electrospray mass spectrometry. Glycobiology 16:1207-18. PubMed 16880503 · DOI
- Ibeto L, Antonopoulos A, Grassi P, et al. (2020). Insights into the hyperglycosylation of human chorionic gonadotropin revealed by glycomics analysis. PLoS One 15:e0228507. PubMed 32045434 · DOI
- Jacoby ES, Kicman AT, Laidler P, et al. (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 46:1796-803. PubMed 11067815
- Zhang Q, Li Z, Wang Y, et al. (2018). Mass spectrometry for protein sialoglycosylation. Mass Spectrom Rev 37:652-680. PubMed 29228471 · DOI
- Creasy DM, Cottrell JS (2004). Unimod: Protein modifications for mass spectrometry. Proteomics 4:1534-6. PubMed 15174123 · DOI
- Bossio RE, Marshall AG (2002). Baseline resolution of isobaric phosphorylated and sulfated peptides and nucleotides by electrospray ionization FTICR ms: another step toward mass spectrometry-based proteomics. Anal Chem 74:1674-9. PubMed 12033259 · DOI
- Chen G, Zhang Y, Trinidad JC, et al. (2018). Distinguishing Sulfotyrosine Containing Peptides from their Phosphotyrosine Counterparts Using Mass Spectrometry. J Am Soc Mass Spectrom 29:455-462. PubMed 29313205 · DOI
- de Medeiros SF, Norman RJ (2009). Human choriogonadotrophin protein core and sugar branches heterogeneity: basic and clinical insights. Hum Reprod Update 15:69-95. PubMed 18945715 · DOI
- Lapthorn AJ, Harris DC, Littlejohn A, et al. (1994). Crystal structure of human chorionic gonadotropin. Nature 369:455-61. PubMed 8202136 · DOI
- NIST, Atomic Weights and Isotopic Compositions with Relative Atomic Masses, physics.nist.gov: isotope masses and natural abundances used for the computed values on this page.
- UniProt Consortium, entries P01215 (glycoprotein hormones alpha chain) and P0DN86 (choriogonadotropin subunit beta 3): sequences, disulfide bonds and glycosylation sites.
Frequently asked questions
Why does HCG not have one molecular weight?
It is a heterodimer of two glycosylated subunits, and each glycosylation site carries a heterogeneous set of glycans. Al Matari et al. (2021) detected 42 alpha and 33 beta glycoforms at the intact level, and a later method detected 76 alpha and 200 beta glycoforms (Besson-Magdelain et al., 2026).
How heavy are the HCG subunits without their sugars?
Computed from the UniProt sequences with their disulfide bonds, the alpha polypeptide is 10,195.7 g/mol and the beta polypeptide 15,519.9 g/mol (average masses). Camperi et al. (2020) describe the intact hormone as about 35 kDa, so the carbohydrate accounts for roughly the remainder.
Can mass spectrometry tell phosphorylation from sulfation?
Only with high resolving power or a fragmentation test. The two differ by 9.5 mDa; Bossio and Marshall (2002) resolved them at a resolving power above 400,000, and Chen et al. (2018) used the fact that sulfotyrosine loses its group in HCD and CID while phosphotyrosine does not.
What should a COA for a glycoprotein report?
Potency in the stated unit and the method, and an identity method suited to a glycoprotein. A single sharp peak at one mass for intact hCG would not match the glycoform distributions reported in the intact-level literature.