Holiday peptides?

Sleepily just walking by these boxes on the counter has caused my brain to glitch every single time. 

Great. Sulfo-Tyrosines are important? Here's how you detect them!

 I think the picture above encapsulates the problem pretty well, right? There is a 0.01something difference in mass between a phospho- and sulfo-tyrosine.

Better to just ignore that either of those are things and move on, right? Unfortunately, a this new paper points out, the sulfome is important in viral infections and even knocking out sulfo-transferases in mice makes them skinny and have poor vision....

...and we don't have any idea what the real breadth of the sulfome is. How many proteins get "sulfomed" (sulfo-ded?).

Until now! 

There is a lot in this paper. But they basically work out a modified IMAC that can preferentially enrich peptides with sulfated tyrosine residues using acetic acid to charge them and titanium or zirconium IMAC enrichment (go acetic acid!) 

Also - even ignoring the sulfonation stuff - how cool is it to be able to enrich modified tyrosine peptides without one of those awful anti-phosphoY antibodies. I haven't done it in a while, but incubating in a cold room overnight while a mixture of antibody rocks is not my idea of a good time. Zirconium and vinegar? Sign me up (plus the antibody is expensive and I've heard hasn't improved much in the last dozen years). Okay, on a re-read I think I'm not getting this right. Maybe it is only enriching the sulfo-Y. In which case I think I'll just study sulfo-Ys next time. 

Cool - so now they have peptides with modified tyrosines - how do you tell a 0.01something Da mod apart? On my instruments you don't. Particularly on a week like this one where we don't have heat in our building. My TOFs are like +/- 0.03something, though they seem a little better between 1 and 4pm. 

At very low collision energies, they neutral loss different! They use an NCE of 10 and that is enough to neutral loss sulfoY but not phosphoY! Then they can fragment it again with HCD or ETD or ETHCD or whatever. 

End result? MASSIVE improvement in our understanding of what proteins get Y sulfo-ed! What a cool resource for people out there who are into sulfo- metabolism and have (rightly) thought that mass spectrometrists can't really help them!

QuantumESSI writes least biased app note in history - bye bye LCMS, Ploopinum time!

 

This week I was thrilled to find that 8 new emails from newcomer QuantumSi informing me about their head to head battle with LCMS.

I was even happier to find out I had to go to their site and fill out a bunch of stuff about myself in order to access this app note. 

Possibly the greatest moment of my life was reading through what appears to be a high school student's interpretation of the experiments that were performed. It wouldn't have been at the pinnacle of my life's experiences without the requirement to agree to be Added. To. The. Mailing. List. to download it when I clearly followed one of the 8 links handily provided to my inbox because - presumably - I'm on the mailing list....

9th time's the charm, right?

Here are the details on the mass spectrometry

I don't know who did this for them, but - bravo - 5 day turnaround - including however long it took them to ship it to you? 

The Ploopinum system takes about 4 days including digestion, the 10 hours the sample spent on the instrument in their lab, uploading the data to the secure cloud and - presumably - doing cloud things on this single purified and digested protein.

But it outputs prettier data than the core did! Look at this! 

Just about every LCMS core has different levels of data that you can provide to collaborators/clients right? The phenomenal core at Pitt has "data processing tiers" and I think that is a smart idea. Some of my collaborators want to install the Proteome Discoverer viewer on their PCs and actually look at their data. Others, however, want a protein list like the one at the top. I'm lying, of course, I always offer the viewer software and processed data and in 3 years at this university no one has wanted it. They want the spreadsheet. 

Takeaway number 2 ? The Ploopinum isn't fast. Even if you have one in your own lab, for your purchasing of this box, the reagents and the cloud subscription service and the dedicated operator to do the experiment, you're slightly faster than an LCMS core service.  (Takeaway #1 is that this company's marketing could use some work).

Okay - but why is the Ploopinum so superior? That's the whole reason you're reading this, right? 

The Quantum device identified 2 peptides that the LCMS analysis did not. 

The first was a 5 amino acid section RLYCK (this was a LysC digested protein) because the operators only considered a 6 amino acid or longer region to be worth reporting to the end user.

You could argue, I guess, that the reason we ignore chains of 5 amino acids is that they are generally quite likely to occur at random throughout the proteome of anything - like, for example, the amino acid combination RLY is found in 11 proteins in the cRAP database. So....I ain't reporting that I found an RLY peptide to anyone. 

RLYCK occurs in 4 human proteins in the tiny canonical (no isoform, no common population level genetic variants) UniProt/SwissProt database. (It occurs 14 times if you consider common human single amino acid variants deposited as of the last time I updated my library in 2020). UniProt matches shown below. 

So...if this was a "proteomics" service request, most facilities won't report this peptide. 

And the Ploopinum identified a peptide with a pyroglutamic acid in it that the LCMS core did not.

Of course, the LCMS core was able to identify that peptide when they were asked to look for it and you could argue that it seems a little suspicious that the core appears to have been provided a purified protein and was not asked to look for PTMs in the sequence - like...maybe this went to a proteomics core and not to a facility that specializes in single protein analysis? ....but it's too late. The damage has been done. They handed this data to Billy and he already got a C on his writeup for his 9th grade general science research paper.  Someone in this company who wears a suit every day said "looks great Billy....let's not put your name on it, though, and the best way to get this marketed to the right people would be to send it to people with LCMS systems!" 

It worked! I've already powered down my mass specs that weren't destroyed in our most recent flood, because clearly this little instrument that takes 10 hours to identify 8 LysC digested peptides from a purified protein is the future. (OMG. Because some people can't tell when I'm being saracastic, no I'm not powering down any system here. On my oldest system -- that is no longer even manufactured - an S-Trap digestion of this purified protein would be about 2 hours (about 12 minutes of work) and I'd run this on a 20 minute LCMS gradient. I'd run it through single protein analysis software - NOT proteomics software - and I bet you I'd have better coverage of this protein by the time I'd forget to eat lunch today. But not everyone has this stuff around. And you can't get a degree in proteomics. You have to learn it on the job. 

Now - I've been pretty negative about this, but I tell you what, there is at least one of these things on the way to my university now. Because this little box has a lot of potential for researchers with a lot of money and without a lot of patience for me and my whining that my I need consistent temperature, relatively consistent humidity and dry floors in my lab to operate my mass spectrometers. There is a market for this little box. Despite the 9 emails I'll get every time they write up every app note, it isn't marketed for me, but it certainly will find homes in labs out there in the world. 

What does the GET / TRC pathway do? Crazy extensive proteomics to the rescue!

 

Have you ever really really really needed to get to the bottom of things? Like...you needed to see literally every tryptic peptide that an organism can produce that is bigger than 7 amino acids and smaller than 60? 

Looks like a group at Harvard was really tired of the GET pathway (called TRC) in humans and getting questions about it, so they knocked every protein in the pathway out in yeast and then pooled 96 (NINETY- SIX) offline fractions for MS3 TMT (aided by RTS) and got stellar quantification across an absurd number of peptides and proteins. 

I didn't even think yeast had 4,580 protein coding regions! Didn't someone work it out to just less than 4k just a few years ago? 

The data is ridiculously pretty, as you might expect from that much work on a relatively simple organism. 

Overkill? Maybe, but there aren't any question marks here on what each member of this pathway contributes to yeast biology. The TRC proteins in humans are involved in all sorts of different diseases, as well, so this can only be a solid resource for anyone interested in these proteins and what they do. 

TP53 / P53 single amino acid variants completely alter cancer progression!

 

Not a lot of proteomics in this new paper, but it's a seriously big deal. 

P53 (or TP53) is one of best arguments for why we do proteomics. P53 is a bigtime central driver of cell cycle and in a normal cell it goes "STOP STOP STOP SOMETHING IS WRONG" when something is wrong. In a cancer cell, various mechanisms end up tricking it so it doesn't know thngs are wrong and it just lets a damaged cell keep on dividing and passing that damage on.

It's a big deal for us because P53 is regulated by degradation, not by transcription/translation. The number of transcripts are worse than pointless. They're misleading. Source: My first postdoc was trying to figure out what a recently discovered protein called the P53 binding protein number 1 actually did. That was a very long time ago, so please excuse if I forgot everything. 

BUT the point of that last ramble is that we've been studying P53 for decades and it is really really really important. So -- new single amino acid variants that dramatically alter how devastating a tumor is - is both a surprise and a very big deal. 

Proteomics Battle Royale Round 3 - Orbitrap vs TIMSTOF vs ZenoTOF vs Q/W-TOFs?

 

When this idea was suggested for the next Battle Royale, I'll admit I got a little nervous. Honestly, thinking about it very very very early on a Saturday morning (toddlers are always sick, apparently) makes me want to make a high pitched almost laughing sound. There has to be a 100 different reasons why we wouldn't carefully select people to defend their favorite instrument in different scenarios that we come up with, right? It's not like we tend to get....rather....attached....to our magic vacuum chambers... it probably wouldn't turn out extremely bad, right? 

Meh, if we don't do it no one else will! 

Probably in brainstorming stage at this point, but I think we'll really need to fine tune our rules on who our champions can be. Obviously, people who work for or consult for vendors couldn't do it, so that would rule out a lot of big names. I bet government employees and some big pharma companies would have stipulations for employees because championing a product might be construed as a conflict of interest. 

If we could still find people around those limitations, I'll reach out to a Bee Gees cover artist to see if he'll host it again. 

Your periodic reminder that hydroxylamine can cause chaos everywhere!

 

My next meeting is delayed by 15 mins? Time to knock this out!

Hydroxylamine is not harmless stuff. It messes with some amino acids and --- EEEK 

Knocks off Serine and Threonine phosphorylations? Fortunately, this group has some advice on how to minimize it. Yes, it's like 5 years old, but I didn't know this. 

Need to extract proteins and peptides from ancient bones - here is how the methods compare!

 

Not to weird you out or anything, but I know that you have been thinking "when the permafrost moves back to expose more old animals, how will I do proteomics on them"? 

Do I ever have the paper for you! 

Tracking metabolic flux (and proteomics) of chemotherapy combinations!

 

With possibly the new winner of longest title ever, I present the succintly named

This is a really solid study utilizing label free quantification DDA to investigate promising chemotherapeutic combinations (single treatments are rapidly falling out of favor due to general ineffectiveness, but some combinations can let us break out older therapies!) 

An Orbitrap Lumos is used for the study with a 100 um internal diameter columns x 100 cm. The stars of the study might be the integration of the SeaHorse metabolism system and the fact the proteomics results themselves, which indicate bigtime alterations in the MHC/HLA presentation pathway. Which -- during cancer treatment, which can sometimes turn into the world's most depressing game of "whack a mole" - could suggest some tertiary treatment options as CAR-T therapeutics and the like continue to mature. 

For data nerds out there, Qu lab's extremely consistent methods on the LCMS side of things make this yet another valuable dataset. There is very little out there where you know pretty much for certain they digested the samples the same way and didn't yield to the temptation of pushing every button on their instruments before starting the next study. 

The "you guys know each cell can be very different sizes, right?" paper is out!

 

The preprint of this paper is on the blog somewhere and I have to make sure to update all the citations when I mention it now.

If you get a chance look at some cells under a microscope sometime! Even if they're cancer cell lines that are 60 years old, you'll see some small ones and some big ones and some square ones and some that look like amoebas, etc., It's really jarring. 

This paper doesn't just point that out to people who are really really good at connecting 360 um glass tubes to other 360 um glass tubes and getting those to hold TENS OF THOUSANDS OF POUNDS PER SQUARE INCH of pressure, but might not know much biology.... it ALSO provides strategy for normalizing for cell size.