Trimethylation?

Well, I labeled with these 2 dimethyl forms and cyanoborohydride.. it can then only be 4 Daltons apart or more, if there are more sites. But that would still be 8 Daltons apart, 12, etc. Not 6.

With every single methyl group that is incorporated, a difference of 2 Da (= 16-14 Da) is introduced.

So if you dimethylate, the mass difference is 4 Da (= 32-28 Da).

If for some reason, a trimethylation occured (of which I'm unsure this is possible under the reaction conditions you used), the difference would be (48-42 =) 6 Da, right?

Do you have MS/MS fragmentation data of the peptide(s) from the peak pairs that differ 6 Da?

Could those be peptides differing by 12 Da at charge 2? (12/2 = 6) because mass spec measures m/z

Thanks Eef, I have spectra enough, but am currently running still. :) It's a good suggestion though, I'll have a closer look at those. This is the method I'm using: https://www.ncbi.nlm.nih.gov/pubmed/19300442 and it states that the light label adds 28 Dalton (2x CH3) and the heavier one 32 Dalton (2x CHD2). I cannot really see how to get trimethylation, but that would make sense as a difference.

I'll upload a MS/MS spectrum asap.

Hi Eef, here's a spectrum of such a case. HCD fragmented and it's the heavy one from the pair (light is 711, heavier 714 with a charge of 2+).

edit: I guess you would need both. :) Let me have a look.. there you go

Well, I noticed why I got so confused. I didn't consider extra lysines in the sequence because of inefficient trypsinisation. Woops :)

Hi Vincent,

Thanks for providing the spectra. Does the fact that you didn't consider that trypsin could miss a cleavage site solve your '6 Da' problem..?

If not, I think there might be a clue in the MS/MS data, but I'd have to look into it in more detail to be sure: it seems that the peptide you selected here (STGGAPTFNVTVTK) is also singly methylated on the threonine closest to the C-terminus of the peptide. Does that ring a bell (in other words: is that a known side reaction..)?

Again, I'd have to spent a few more minutes to be sure -and maybe also get detailed information on some of the fragment ions in your spectra that are observed, but were not annotated with a mass.

EDIT: Now that I've given it some extra thought, the additional methyl group could also be (and this seems more likely) on the side chain of the C-terminal lysine..

Hope this helps!

Well, it could've been a good explanation, but I'm actually having trouble identifying a lot of spectra using Maxquant, indicating it might be a problem somehow with my search and thus maybe with the reaction. I've done these reactions before multiple times, but this was a few years ago. I'm starting to think it might be a too long incubation or so as I've never had this problem before.

An additional methylation on the lysine would then also make sense perhaps, I'll try to do a search with this in mind as a fixed modification, but obviously, if this is the reason, it will be a mixed bag of results. Thanks a lot for the help! I'll let you know what I come up with.

I'm pretty convinced that the example you provided contains a C-terminal methylated lysine: could you check whether the y1 ions (of the methylated lysine) are present in both spectra (161.1 for the light and 163.1 for the heavy-labeled species)? These could then be used as 'marker ions' (using an extracted ion chromatogram of these y1 ions from your MS/MS trace) to estimate the extent of unintentional lysine labeling you introduced in this experiment.

BTW, could arginine also be susceptible to methylation..?

You're correct and those Y1 ions are present in both spectra. Having labeled extra lysines is fine btw, the reaction should label any n-terminus and side-chains of lysines, so in the above case that makes perfect sense actually.

But searching for those gives me, not very bad results, but still a lot of cases where only 5% of 120.000 spectra are identified in a given run. This is the odd thing as the MS/MS spectra look great in general. Which is why I switched to "find-any-potential-problem-mode". :) 

I see :).

Did you include a control digest, that wasn't labeled? The search results of that should enable you to see whether the reaction conditions you applied are causing the apparently low success rate in your peptide ID efforts.

Assuming that you've included the correct cysteine (fixed) modification (if any) and the right mass accuracy for your precursor and product ions, you could also investigate if for a certain protein for which you retrieve numerous peptides (such as the Profilin-1, maybe), the peptides you find back have something in common, or if the peptides you don't find back have anything in common. In case you don't find any aspartate or glutamate-containing peptides, for example, these residues could potentially also have been subject to a side reaction.

Good luck!