Post-translational modification analysis
Capture phosphopeptides out of a digest using the chemistry of the phosphate group itself: titanium or zirconium dioxide, or immobilised Fe(III). The best developed of all modification workflows, and the one where the distinction between detecting a phosphopeptide and localising the phosphosite is sharpest.
Our ruleSimulations prepare you for the bench but don't replace hands-on practice.
Phosphopeptide enrichment
Capture phosphopeptides out of a digest using the chemistry of the phosphate group itself: titanium or zirconium dioxide, or immobilised Fe(III). The best developed of all modification workflows, and the one where the distinction between detecting a phosphopeptide and localising the phosphosite is sharpest.
How it differs from the standard method
- Loading buffer is high organic with a strong acid and a competing carboxylic acid (glycolic, lactic or DHB) so acidic peptides do not co-enrich; without it specificity falls from around 90% to near 50%.
- Fe-IMAC columns and TiO2 tips differ in bias: IMAC favours multiply phosphorylated peptides, metal oxides single ones, so the two give overlapping but not identical site lists.
- Sequential enrichment raises specificity and lowers recovery - a deliberate trade, not an improvement.
- Phosphotyrosine is about 1% of the phosphoproteome and is usually enriched with an anti-phosphotyrosine antibody rather than metal affinity.
- Phosphatase inhibitors must be in the lysis buffer from the start; sites lost at lysis leave no signature in the data.
- Enrichment is the technique
- The modification has to still be there
- Detecting a modification and localising it are different results
- Why collisional fragmentation loses the thing you are looking at
- What the three enrichment families give you
- Stoichiometry is a different experiment
- A changed modified peptide is not a changed modification
- Artefacts that look like biology
- Reading a site table honestly
- Key terms
- When to use it
- Sub-techniques
1Enrichment is the technique
A phosphopeptide is around one to two percent of a tryptic digest. An acetyl-lysine peptide is well under one percent.
2The modification has to still be there
Kinases stop working when the cell is lysed. Phosphatases do not: they are perfectly happy in a lysate, and within minutes of lysis a substantial share of phosphosites is gone.
3Detecting a modification and localising it are different results
A modified peptide has a mass shift (79.966 Da for phosphate, 42.011 for acetyl, 203.079 for a single HexNAc) and seeing that shift on a peptide is good evidence that the modification is on that peptide somewhere. Saying which residue carries it is a separate inference, and it needs a particular kind of evidence: fragment ions that cleave the backbone between the candidate residues.
4Why collisional fragmentation loses the thing you are looking at
In positive-mode collisional fragmentation the weakest bond breaks first, and for a phosphate ester or a glycosidic bond that is the bond holding the modification on. Phosphopeptides give a strong neutral-loss species 98 Da lighter than the precursor; glycopeptides give a spectrum dominated by oxonium ions from the sugar and very little peptide backbone.
5What the three enrichment families give you
Metal affinity
Titanium or zirconium dioxide and Fe(III)-IMAC bind the phosphate group itself, so they are modification-specific rather than sequence-specific. They also bind carboxylates, which is why the loading buffer contains a competing acid such as glycolic or lactic acid to keep acidic peptides off.
Hydrophilic interaction
Glycans are far more hydrophilic than peptide backbones, so loading in high organic solvent retains glycopeptides while unmodified peptides pass through. Broad across glycan classes, and the usual first choice for intact glycopeptide work.
Lectins
Carbohydrate-binding proteins that each recognise a structural motif. Useful when you want one glycan class, misleading if you treat a single lectin as a general glycoprotein enrichment, because what it misses is structured rather than random.
Antibodies against the modified residue
The only general route for acetyl-lysine, methyl-arginine, the di-glycine remnant left by ubiquitin after trypsin, and others. Used at the peptide level after digestion, because a short modified epitope is what the antibody was raised against. Needs milligrams of input and a validated antibody.
6Stoichiometry is a different experiment
A phosphoproteomics run tells you that a site exists and how its abundance changed between conditions. It does not tell you what fraction of the protein molecules carry the modification, and that fraction is usually what matters biologically: a site at one percent occupancy that doubles is a different thing from a site at fifty percent that doubles.
7A changed modified peptide is not a changed modification
If a phosphopeptide is twice as abundant after treatment, there are two explanations: the site became twice as modified, or there is twice as much of the protein. The enriched run cannot distinguish them, because the enrichment removed the unmodified peptides that would have told you the protein's level.
8Artefacts that look like biology
- Iodoacetamide in excess adds a mass that is indistinguishable from the di-glycine remnant used to detect ubiquitination, so over-alkylation manufactures ubiquitination sites. Use the recommended concentration and keep it in the dark.
- Asparagine deamidation adds 0.984 Da, the same shift PNGase F leaves behind on a formerly glycosylated asparagine, so a deamidation site can be mistaken for an occupied sequon.
- Methionine oxidation and N-terminal acetylation happen during handling as well as in the cell, and the proportions depend on how long the sample sat around.
- A modified peptide ionises differently from its unmodified counterpart, so intensities compare the same peptide between samples and never a modified peptide with an unmodified one.
- Trypsin does not cleave after a modified lysine, so acetyl, methyl and ubiquitin-remnant peptides always carry an internal lysine and a missed cleavage. That is expected, not a digestion failure.
9Reading a site table honestly
Ask four questions of any modification dataset. Was the modification preserved: were the right inhibitors in the lysis buffer?
Key terms
Practise these as flashcards →The fraction of identified peptides in the enriched sample that actually carry the modification. Below about 60% most of the instrument time went on unmodified peptides.
The confidence that the modification sits on one particular residue rather than another candidate in the same peptide. Separate from the confidence that the peptide was identified at all.
A fragment ion produced by cleavage between two candidate residues, so its mass differs depending on which one carries the modification. Without one, the site is ambiguous.
Loss of the modification as an uncharged molecule during collisional fragmentation, for example 98 Da of phosphoric acid from a phosphopeptide, which consumes signal that would have gone into backbone ions.
A low-mass sugar fragment (for example 204.087 for HexNAc) produced when a glycan fragments. Diagnostic that a spectrum came from a glycopeptide, and useless for locating the site.
The fraction of a protein's molecules carrying the modification at a given site. Rarely measured in discovery experiments and not what a fold change reports.
Electron-transfer fragmentation, alone or with supplemental collisional activation, which cleaves the backbone while leaving labile modifications attached.
The two glycine residues left on a ubiquitinated lysine after trypsin digestion, and the epitope the ubiquitin-site antibodies recognise.
A specific molecular form of a protein, with a defined set of modifications on the same chain. Bottom-up workflows average over proteoforms; they do not resolve them.
When to use it
- To map where a kinase, acetyltransferase or glycosyltransferase acts, across thousands of sites and without an antibody per site.
- To follow a signalling response over time, where the interesting change is in modification state rather than in protein abundance.
- To characterise a biologic or recombinant product, where glycan occupancy and heterogeneity are part of the specification.
- To find the modification responsible for a shift in migration or activity that a blot shows but cannot explain.
- To test whether a site identified by a phospho-specific antibody is the site the antibody is actually reporting on.
Compared with alternatives
The same instrument and the same digest, but without enrichment; it measures protein abundance, which is the control a modification experiment needs.
A modification-specific antibody on a blot is fast, cheap and quantitative enough for one known site, and it reports occupancy change on a defined protein, if the antibody is specific.
Immunoprecipitating the protein first and then looking for modifications is the targeted version, and avoids needing a modification enrichment at all.
Measures what the modification is usually a proxy for; a kinase activity assay answers a question a site table only implies.
Reports a modification in situ with spatial information in fixed cells, at one site at a time, where mass spectrometry reports thousands with none.
Resolves charge shifts caused by phosphorylation at the whole-protein level, which is one of the few ways to see proteoforms directly rather than inferring them.
Sub-techniques
Choosing one here switches the whole page, including the protocol and the lab, to it.
Capture phosphopeptides out of a digest using the chemistry of the phosphate group itself: titanium or zirconium dioxide, or immobilised Fe(III). The best developed of all modification workflows, and the one where the distinction between detecting a phosphopeptide and localising the phosphosite is sharpest.
- Loading buffer is high organic with a strong acid and a competing carboxylic acid (glycolic, lactic or DHB) so acidic peptides do not co-enrich; without it specificity falls from around 90% to near 50%.
- Fe-IMAC columns and TiO2 tips differ in bias: IMAC favours multiply phosphorylated peptides, metal oxides single ones, so the two give overlapping but not identical site lists.
- Sequential enrichment raises specificity and lowers recovery - a deliberate trade, not an improvement.
- Phosphotyrosine is about 1% of the phosphoproteome and is usually enriched with an anti-phosphotyrosine antibody rather than metal affinity.
- Phosphatase inhibitors must be in the lysis buffer from the start; sites lost at lysis leave no signature in the data.