Tuesday, January 31, 2012

Protein Discovery Expedeon Merger

Last year I discovered one of those little gems -- a little biotech company nestled in the mountains of Tennessee with great products, prices and customer service.  Imagine my horror when I heard that Protein Discovery was merging with San Diego based Expedeon!
Not to fear.  We just received our first case of FASP kits from PD/Expedeon.  As before, they arrived in record time, and they don't seem to have changed at all.  What did change, however, was the price.  You now get a discount when you buy more than 1 box. 
Change doesn't have to be bad!

Saturday, January 28, 2012

ASMS Abstract Deadline

The deadline for ASMS abstracts is looming!  You have until February 3rd.  Get cracking!

Saturday, January 21, 2012

CID vs ETD for phosphopeptide identification

Experiment:  A new general phosphatase inhibitor was used to treat a SILAC heavy labeled cell line. Phosphopeptides were enriched by the three stage method outlined in my methods book.  LC-MS analysis was performed with CID neutral loss MS/MS with optional ETD activation of the phosphopeptides that were large (z=3, >900; z=4, >750, all ions z>=5; the default Xcalibur supplemental fragmentation method)
Post-processing we looked at the number of phosphopeptides that were identified by each method that were interesting (differentially regulated >2.0 fold).
Here's how it breaks down:
Striking?  Optional ETD fragmentation more than doubled the number of interesting proteins that were phosphorylated.  The two proteins that overlapped?  They were both unique phosphorylation events that happened to happen on the same protein.  At the peptide level there were ZERO overlapping phosphopeptides of interest.
Summary:  If you have ETD on your Orbitrap, even though it is a pain to clean and tune, using it to supplement CID may double the number of interesting phosphorylation events.

Thursday, January 19, 2012

Iniparib Nonselectively Modifies Cysteine-Containing Proteins in Tumor Cells and Is Not a Bona Fide PARP Inhibitor

This paper is both a relief and a disappointment.
Its a relief that someone finally published that iniparib is not a PARP inhibitor.
It is a disappointment because they published the molecular mechanism of iniparib before I could.  The wait for permission to publish scientific results at my old lab takes anywhere between several months and  a few years.  Unfortunately, while waiting for permission, we sometimes get scooped by other labs.  I have been told that we have withdrawn this study from the approval process in the hopes that management will get to one of the other studies we have pending.  Since this data will not be published, I see no reason why I shouldn't show these results here.
First of all, combining iniparib with glutathione (GSH) at a pH of 7 causes GSH to undergo two mass shifts to heavier species.
Fragmentation analysis suggested that the 2 species were the two structures shown in the bottom panel.  The lower mass appeared to be an addition of -OOH to the free sulfur, while the higher mass appeared to be the iniparib compound actually binding directly to the sulfur.  You'll notice that I ionized this mixture at a pH of 7, in my normal running buffer, 0.01% formic acid, the iniparib bound GSH disappears, as it seems to be cleaved to the -OOH version in the acid (this lines up with previous studies using radiolabeled drug)
GSH is a strange peptide.  In order to see if the drug would modify a more normal one (and to confirm the previous study that iniparib binds GAPDH), we had the cysteine-containing regions of GAPDH produced by a peptide synthesis company.
In this second experiment, we only saw the mass shift corresponding to the drug sticking directly to the cysteine of the synthetic peptide.  This modification was far less acid-labile than the GSH modification, suggesting that I could possibly detect this modification during a normal proteomics study.
First of all, we used the Cysteine-TMT system to label and pull down peptides with free cysteines (per manufacturer's instructions) from MCF-7 cells that were 1)control 2) iniparib treated.  We found a 26% drop in identificable TMT-modified cysteine peptides with iniparib treatment.
We then looked directly for the BSI static modification in a SILAC experiment (using a custom modification for the BSI mass shift) and identified 700 proteins that had picked up this modification.
I performed gene ontology analysis of these proteins using Panther GO --
-and saw that there was modification indiscriminate of pathway or cellular localization.  The drug selectively binds to any free cysteine it can find.  
The big question, then is why does this drug kill some cells and not others.  I know the answer to this question, as well, but I didn't do those experiments and I don't think I should write about it here without his permission.

Sunday, January 15, 2012

High resolution single ion monitoring (SIM) on an Orbitrap XL

We have reason to believe that a particular drug is sticking to a particular amino acid within a particular protein.  Due to the fact that this drug is patented property, I can't reveal any other details than this.
In order to directly test this, my preferred method would be using single reaction monitoring (SRM or MRM) on a triple quad instrument.
In this particular case, I don't have access to a triple quad, just an Orbitrap.  Therefore, we'll have to get creative.
On a synthetic peptide resembling the active site of our protein of interest, I know exactly the mass shift that occurs when the drug binds.
In silico digestion of the protein of interest + the drug mass shift gives me a z=2 mass of 1131.198
Here is the setup and results
Only the HeLa cells treated with the drug produce the a molecule that falls within my specific mass window.
Now to compare this to an MRM!

Sunday, January 8, 2012

Fixing Proteomics Campaign

I've spent a lot of time on this website and the links around it, recently.  This is due to my on-going obsession with applying quality control standards to proteomics experiments.  Fixing Proteomics is an organization dedicated to making experiments in our field more reproducible.  The main focus is to standardize techniques between labs.  They have this 4-part outline for how a lab can contribute to fixing proteomics:
If you are interested in quality control and reproducibility in proteomics, click here or on the screenshots from their website.

Wednesday, January 4, 2012

Maxquant 1.1 vs Proteome Discoverer 1.2

For proteomics processing, what should you use?  Proteome Discoverer 1.2 or Maxquant?

This evaluation is somewhat unfair to MaxQuant, due to the fact that I only have the 1.1 version operating.  This is due to the fact that the Nature Methods paper has instructions for this software and updated instructions for the new version using their in-house Perseus search engine are not available.
For this comparison, 22 files containing SILAC-labeled samples were downloaded from the Tranche database.
The samples were from human cancer cell lines and FASTA employed was the same IPI derived FASTA available at Maxquant.org in the useful files folder (which they very nicely update frequently!)

All parameters within my power to control were made equivalent, including exclusively using Mascot for both searches

Results:
1) Quantifiable protein IDs:  PD- 643, MQ, 311
2) Percent overlap (average):  ~23% (identification)
3) In matching protein IDs, direction of regulation overlap:  100%
3) Time to process:  PD ~ 9 hours, MQ, ~3 hours

Summary:  There are definite pros and cons to each software package.  PD takes longer, isn't free (unless it came with your equipment, which it often does), and comes up with more protein IDs.  MQ is faster, free, has open source code so you know what it is doing (except for when it comes to Mascot) but doesn't produce as many IDs as PD.  The best part is the number of unique proteins that come from using the two methods on the same dataset.  If you have PD, you might want to download MQ (its free, after all!) and run your SILAC samples on it as well.  It is fast and may give you more to work with!


Update:  If you are interested in a more recent comparison, please follow this link to my analysis of Proteome Discoverer 1.4 vs MaxQuant 1.3.0.5.


Monday, January 2, 2012

Moving this blog to blogger/blogspot

After seeing how much simpler the Google-owned blogger/blogspot service is for posting and linking items, I have decided to move this blog from its old location to this new address.  News in Proteomics Research is now here, at :  http://proteomicsnews.blogspot.com/.  I've copied the original site and pulled it off line completely.
I'll go back through the last 2 years of old entries and start copying over anything that got a lot of views or comments, as well as my favorite posts.  It make take me some time, as I've noticed that a few of the older links no longer work.  I'll update links, and maybe some of the images (because it is much easier to do it here!)
If there was an article or post that you would like me to move to the front of the queue, drop me an email at:  orsburn@vt.edu and I'll move it over here immediately.  Otherwise, the reappearance of articles will be pretty random.