TMTc (complementary) ions leveraged in Alzheimer's insoluble proteome analysis!

 

TMTc comes up once in a while, but relatively few groups have leveraged the complementary high mass fragment ions that are generating during fragmentation of TMT labeled peptides.

There are reasons for this, like a limitation on the number of tags you can use (because they either can not be resolved or because you can't resolve them, which is a different thing, but effectively the same).

But when you consider the likelihood of fragmenting 2 or more things that will produce a 126.012 and a 127.013 fragment ion vs two things that will produce a (for example) 1386.042 and 1387.043 it is pretty clear that you'd have loads less interference in the latter. (Everything you tagged with TMT will make those first 2 things, including lipids and small molecules with amine tags, but relatively few things will elute at exactly your same retention time and will produce the same high mass ions). 

More on TMTc here and here and - this new paper! 

What this group dissected was detergent-insoluble proteins. If you take a "normal" human brain and you solubilize the proteins in detergent and centrifuge it you'll get a little pellet. If you do the same thing with a Huntingin's disease brain or Alzheimer's (or more controversial -- a brain from someone with schizophrenia) you'll get a BIG PELLET. The question in a lot of these disease proteomics studies is WTF is in this pellet!

And the problem with doing it with TMT is that you're probably talking about some really big ratios. Presence/absence sort of things. And TMT ratio compression, plus the fact it has a somewhat limited dynamic range seriously confounds things. 

TMTc to the rescue! In this study they thoroughly optimize with TMTPro reagents using human/E.coli ratios to dial in their quan and use a multistep fractionation scheme to simplify the insoluble brain fractions from post-mortem human brains with Alzheimer's disease.

They use an HF or HF-X (I can't tell if I scribbled an X or if the picture I drew while trying to figure out how they combined their offline fractions just makes it look like and X and I'm not looking for the PDF again). The actual samples were fractionated by gel slice, but the libraries were made from a pool using offline HpH fractionation.

All the data was processed using JUMP, and I'm excited to see something outside of the top secret Harvard server that no one else can use can process these data and I will be exploing -- but I'd just use the free version of Proteome Discoverer with the IMP-Hyperplex nodes which could do this since it is Orbi/Orbi data. (www.pd-nodes.org) 

I don't understand what they found here. Something about "proteins with low complexity domains" but they seem excited about it, and with the prevalance of these awful diseases (and just terrible "diagnostics" for them), I bet they aren't the only ones.