Missed cleavages and ragged ends

A missed cleavage is a peptide that still contains a lysine or an arginine where trypsin was supposed to cut. A ragged end is a terminus that gained or lost a residue relative to a clean tryptic cut. The two look similar in a spreadsheet and they come from different causes. Bottom-up proteomics in plain language is the identification frame this page sits inside. The decision here is narrower: when a peptide list looks chewed or under-cut, do you change the search, or do you remake the digest?

Where trypsin actually stops

Trypsin is a serine protease that hydrolyses the bond on the carboxyl side of lysine and arginine. The cut fails, as a rule, when proline is the next residue. It often fails, or happens only partly, when an acidic residue sits next to the site, and it fails when the lysine itself has been modified so the side chain is no longer the structure the enzyme recognises. Sequencing-grade trypsin is treated to reduce cuts at other residues. It is still a protein preparation, not a theoretical scissors. A small population of missed sites is ordinary. A digest in which missed sites dominate the identified peptides is a sample-preparation result.

Ragged ends are a second population. One end looks tryptic and the other is a residue short or a residue long. Incomplete cleavage can leave that pattern. So can a chemical clip, such as a fragile bond that breaks in acid, and so can in-source fragmentation, where the peptide breaks in the ion source and the search treats the fragment as if it had been a peptide in the vial. A non-specific nick is a third pattern: the cut is not at lysine or arginine at all, often at an aromatic or hydrophobic residue, which points at residual chymotryptic activity or at a protease you did not mean to add.

What you allow the search to consider

A database search rebuilds peptides from a sequence database such as UniProt. If you tell it that trypsin never misses, every real peptide with an internal lysine becomes invisible or gets forced onto the wrong sequence. If you tell it that five missed sites are welcome, the number of candidate peptides explodes. More candidates mean more chances for a decoy sequence to score well, and the false discovery rate estimated for the list can get worse even when the spectra have not changed. One or two missed sites is the usual discovery setting. Semi-tryptic or non-specific searches are larger again. Use them when the biological question is about termini, and expect the threshold to be harder to satisfy.

Public builds such as PeptideAtlas show how often particular peptides are observed across experiments. They are a place to see that some missed-cleavage peptides are repeatedly real. They are not a licence to loosen your own search until a favourite protein appears.

Enzyme, unfolding, and inhibitors

The physical variables sit in the tube, before any spectrum exists. Protein that never unfolded hides cleavage sites in structure. Urea, guanidinium salts and selected surfactants are the usual denaturant classes, and each has a removal problem later. The enzyme-to-protein ratio is the next variable. Many in-solution protocols describe a modest ratio, commonly discussed as one part sequencing-grade trypsin to several tens of parts protein by mass, near 37 Celsius, for a time the supplier states. That is a planning range. The sheet that arrives with the enzyme is the instruction. Too little active enzyme, or enzyme that has lost activity, leaves internal lysines and arginines. Too much enzyme buys autolysis peptides and more non-specific nicks.

Inhibitors matter as much as the ratio. A protease-inhibitor cocktail that was sensible at lysis is poison if it is still present when trypsin is added. Residual serine-protease inhibitor, a strongly acidic or strongly chaotropic solution that was not diluted, or a detergent that unfolds the enzyme itself, all raise the missed-cleavage rate. The companion note on preparing peptides for mass spectrometry covers the cleanup after the enzyme has done its work. Cleanup cannot create the cuts that never happened.

A workflow with a branch when the control digest fails

Start from a protein amount you trust, or say plainly that the amount is an estimate. Digest a small known protein, or a previously successful lysate, beside the real samples. If that control shows the same high rate of missed sites, the enzyme, the buffer or the temperature is the branch. Remake the digest. Do not open the search settings and allow three missed cleavages so the control "passes".

If the control is clean and only the difficult sample is full of missed sites, look at unfolding and inhibitors in that matrix. Membrane proteins, disulphide-rich proteins and samples that went into the digest still cloudy are the usual cases. A Lys-C step before trypsin is a method class some protocols use when lysines are hard to reach. It changes the peptides you will see. Name it in the search.

If the missed-cleavage rate is ordinary and the oddity is a set of ragged ends on one protein, ask whether that protein is processed. A signal peptide leaves a new N-terminus that is biological. Initiator methionine removal does the same. Those ends deserve a semi-tryptic search, or a search that allows the known processing, aimed at that question. They do not deserve a global non-specific search of the whole proteome just to keep one peptide.

If ragged ends are everywhere, including on the control protein, think about in-source fragmentation and about chemical clips during a long acidic hold. Lowering source energy, within the instrument method you are allowed to change, and shortening the time peptides sit in strong acid, are the practical branches. A non-specific nick pattern, cuts at phenylalanine, tyrosine, tryptophan or leucine, points at enzyme quality rather than at the mass spectrometer.

Three peptide injuries, and what each implies

Read the pattern across the whole file, not on one favourite peptide. A single missed site next to glutamate can be the sequence. A proteome-wide rise in missed sites, judged against your own earlier digests of the same sample type, is the signal to act on. HUPO discussions of identification quality are a useful public frame for how the field talks about that restraint. They are not a certificate for any one laboratory's digest.

How to reason when the list looks wrong

Look at the digest before you look at the instrument. High missed-cleavage rates point at unfolding, enzyme ratio or inhibitors. The mass spectrometer reports the peptides it was given. Retuning it does not create a cut in a vial that never saw active trypsin.

Separate expected misses from unexpected ones. Sites followed by proline, sites beside aspartate or glutamate, and sites on a modified lysine are expected to be stubborn. If you count those in with the ordinary sites, you will scold a digest that behaved normally. Many search summaries can split the count. If yours cannot, inspect a sample of peptides by hand before you change the protocol.

Separate biological ragged ends from artefacts with a control that has no reason to be processed. A recombinant protein standard should not lose a signal peptide in your tube. If its termini are ragged, the cause is chemical or instrumental. If only the cellular protein shows the new N-terminus, and that terminus matches the processing annotated for the protein, you have a biological observation worth a careful semi-tryptic claim. Write the claim at that strength. A ragged end is not, by itself, evidence of a signalling pathway.

In-source fragments deserve suspicion when the "peptide" is a piece of a much more intense peak at nearly the same retention time. The chromatogram is the clue. Two molecular species that truly coexist usually do not share a perfect co-elution with a parent of higher mass in that simple way. Ask someone who knows the method to look before you add the fragment as a biological isoform.

Safety, and what a digest is allowed to mean

Trypsin buffers, urea, guanidine and the acids used to stop a digest are ordinary chemical hazards. Follow the safety data sheet for the acid and the denaturant you actually open. A research digest identifies peptide sequences under a stated false discovery rate. It does not diagnose a patient, and it does not prove that a protein does the job someone hopes it does. Biosafety containment for the tissue or the culture is an institutional decision. This page does not set it.

Warm urea and enzyme that sat through the heat

Urea that has been warm decomposes toward isocyanate, and isocyanate carbamylates lysines and N-termini. A carbamylated lysine is a poorer trypsin site and a mass the search may not expect. In a building that runs hot, a urea solution left on the bench is already a changed reagent. Make it cold, use it promptly, and do not store it in a drawer because the cold room is full. The same climate is unkind to enzyme aliquots. A trypsin aliquot that thawed during a power cut, or that was held at the temperature of the room while tubes were labelled, can look clear and still have lost activity. The next digest then shows a high missed-cleavage rate. Check that rate with a known protein before you book instrument time to "confirm" the biology. protocols.io is a public place to compare how published digestion methods talk about temperature and enzyme handling. Copy the controls, not someone else's microlitre table.

What to put in an enzyme enquiry

State the enzyme class, the ratio you plan to follow from a supplier protocol, the denaturant or detergent already present, the protein amount class, and whether biological termini must be preserved in the search. Say if a protease inhibitor was used at lysis and how you intend to be rid of it before trypsin. Those facts decide whether a sequencing-grade trypsin, a Lys-C then trypsin sequence, or a different protease class is the right conversation.

The method can be discussed against the shotgun discovery proteomics reference, the protein identification by LC-MS/MS reference and the differential abundance reference. Send the same facts with the quote request. A discussion of digestion conditions is a specification exercise. It is not a statement that a particular digest is already booked or that a reagent relationship exists.