Protein Crystallization Hits

11/22/2010 9:16:54 AM #

AcrB is also another MP that has been "serendipitously" crystallized and solved during the course of trying to study other MP's. Apparently you only need pg amounts for AcrB to make its way to ruining an experiment. www.sciencedirect.com/science

Bryan

11/22/2010 9:17:12 AM #

I'm wondering why it is that these stories are about membrane proteins rather than soluble proteins. Is that perception bias or real?

Peter Nollert

11/22/2010 9:17:53 AM #

Likely perception bias (or lack of reporting for soluble proteins).  
  
I've heard of a handful of stories about accidental crystallization of minor contaminants of non-membrane proteins. The stories usually ended with people drinking their sorrow away after spending months optimizing crystals and solving the structure of these contaminant. When it is a membrane protein, you can still get a publication out of it.  
  
My "favorite" anecdote came from a lab mate in graduate school. He was expressing a hyperthermophilic protein in E. coli. Purification was quite simple, just boil the E. coli lysate to precipitate all of the E. coli proteins and then either a quick ion exchange or size exclusion.  
  
Coomassie-stained gels looked great with a nice single "fat" band and enzymatic activity. Crystallization trials yielded crystals that didn't diffract too well, but optimization gave around 2-2.5Å diffraction. He couldn't solve the structure by MR, despite the fact that it should have been a slam dunk given sequence homology to other proteins.  
  
A few months later, an optimized selenomethionine crystals yielded a high-quality MAD data that gave excellent electron density. You could sequence the protein based on the electron density, but it was clear the sequence in the crystal did not match the recombinant hyperthermophilic protein.  
  
After a painful check at the PDB for matching cell dimensions and a BLAST query, the protein in the crystal was very likely E. coli inorganic pyrophosphatase. He then developed a silver-stained gel on the original protein solution and some washed crystals, and a very faint band corresponding to this contaminant was seen.  
  
Bam! One year gone working on an extremely minor contaminant. This taught me a few good lessons. First, essentially undetectable contaminants can ruin your week/month/career. Second, washing crystals and running them on a gel is very hard to do without carrying over a tiny amount of protein from the drop. Third, E. coli inorganic pyrophosphatase is a hearty little protein that can survive boiling, SeMet expression, crystallization at <0.01 mg/mL in 10 mg/mL of another protein, and diffracts well.  
  
Why don't we hear about these stories? Because they are fine for comments on a crystallization blog, but not worth the labor or expense of a low impact journal article.

Peter Nollert

11/22/2010 9:18:34 AM #

Thanks Bryan for the link! AcrB seems to be a big trouble maker.  
Together with Todd's story I suppose you'd want to have an early check of the protein crystals' identity. Big wake-up call for all of us who haven't checked yet.  
SDS-page of crystals and N-terminal sequencing of the major band would be a way to check.

Peter Nollert