Protein Crystallization Hits

As a trained biochemist I've always had a hard time finding information on small molecules. There are just too many different names and I hadn't been dealing with small molecule compounds often enough to get comfortable with their use. Parlez-vous smiles string?

It is easy to get confused in the jungle of systematic names, trade names, smiles strings, names with spelling mistakes, synonymous designations, CAS numbers etc. Nonetheless, small molecule compounds are a key component in protein crystallography-based drug discovery and, in membrane protein crystallization you're dealing with small molecule detergents and additives all the time. Need to co-crystallize or soak your protein crystals with small molecules - how do I find relevant data to construct a comound library? At the top of my frustration I had identified that there are >10 different names just for one single compound: Triton X-100. How can you make sense in this babylonian mess? Initially I got by with crutches: The sheer weight of the printed Merck index takes the fun out of finding any compound information. So Sigma had always been my friend and their online catalog does an OK job to aggregate data for particular compounds. And of course I have used popular online search engines only to become stuck with possibly dubious information. But none of that worked to my satisfaction.

So what a relief when I found the Google for small molecule compounds: ChemSpider. 

Finding small molecule information is a blast with the chemistry search engine ChemSpider.

Simply search by name, MW, chemical elements or chemical properties and find exact matches or expand your search to identify similar compounds. All nicely tied in with literature data bases and commercial sources in case you want to get the materials in the real world.

ChemSpider is actually better than Google because you're not restricted to search with typed words only, you can search starting from structures that you draw in a Java pop-up window.

ChemSpider tool to draw small molecule structures. Submit what you have in mind to a search and receive information about the small molecule compound from a variety of sources.

Designing small molecule compound libraries has never been easier.

Thanks to the folks at ChemSpider!
Peter

Sean Seaver over at P212121  recently blogged about the crystallization of his target protein in Emerald's new plug-based crystallization system, the MPCS (microcapillary protein crystallization system). Sean showed up with protein samples and crystallization formulations at  our booth at the ACA in Toronto more than a month ago and Cory set up a crystallization trial in CrystalCards with the new PlugMaker instrument. This protein crystallization experiment consumed only a couple of microliters (each crystallization plug has a total volume of ca. 15 nl) while screening a multitude of slightly varying precipitation concentrations. I was happy to see images of crystals that had formed (images of these crystals are posted on Sean's blog). Pretty good for a  crystallization trial that has traveled in a plane, took a long ride in his shirt pocket and made it to Alaska and back to the University of Toledo in Ohio.

Cory has left recommendations on how to proceed at P212121 - but since Sean has asked 2 questions that have come up previously, I'd like to address these here:
1. How can you tell which crystallization condition is in which drop?
This is a straight forward counting job: In a standard gradient run, a number of plugs are produced - let's say 200 individual crystallization experiments (=plugs), where the concentration of a precipitant is varied in a linear fashion (say from 0 to 100%). The start and end of the gradient are marked by longer plugs (with multiple gradients in a CrystalCard) or are defined by the length of the channel. In Sean's case we set up a shallow gradient of ca. 200 plugs (see a similar experiment in the figure below). Let's say you'd see a crystal in plug # 134, then the concentration of the precipitant in that particular plug is 68% (134/197) as compared to the maximum precipitation concentration used. If the maximum precipitation concentration was 2.5 M, the concentration in plug #134 is 1.70 M. Often crystals appear in neighboring plugs (this depends on the size of the 'crystallization slot'), which appears to be the case with Sean's experiment. 

Figure. Identifying the accurate precipitation concentration in a particular plug within a CrystalCard (protein used is Thaumatin). A shallow gradient of 197 plugs with precipitation concentrations ranging from 0-2.5 M Sodium Potassium Tartrate was prepared. Plug #134 has a protein crystal, corresponding to 68% (134/197) = 0.6802, or 1.70 M of Sodium Potassium Tartrate.

2. How do I scale up?
There are two answers:

a) You don't need to scale up at all! Just take the crystal out of the card by removing the plastic laminate and fish with a loop as shown in this video. Alternatively, you can place the entire card into the X-ray beam and collect data that way. Check out these links about further info on in-situ X-ray diffraction of protein crystals within CrystalCards.

b) The plugs prepared by the MPCS are essentially small versions of batch-under oil crystallizations. A good way to mimic a plug is to combine a portion of the protein solution with precipitant at the concentration as figured out in #1 above and cover with oil. If you want to eliminate any effects caused by the plastic bottom of the crystallization tray and closer imitate the oil layer that completely surrounds the plugs in the CrystalCards, you may want to try container-less crystallization.

Let us know if there's anything else you'd like to know about this new low-volume protein crystallization system.

Peter

What's the minimal purity level required to set up a protein crystallization trial? Simple question, no clear answer, really. Most crystallizers I asked are happy to work with protein samples that are more than 90-95% pure. With purity judged by the intensity of the Coomassie stained target band in an SDS-Polyacrylamide Gel (PAGE). Sometimes it's difficult to get to such purity levels though, and the question becomes: at what purity level should you refuse to subject protein samples to crystallization trials?
This is a question that I am very interested in, particularly for membrane proteins. Membrane proteins are difficult to purify and they tend to loose activity during the purification process (this is of course due to the detergent micelles being a poor substitute for biological membranes). So we addressed this question together with Phil Laible's group at Argonne National Lab and while we were at it, we also compared two fundamentally different protein crystallization regimes: crystallization in lipidic cubic phases (LCP) and the more regular crystallization as protein/detergent/lipid complexes. LCP won, hands down.
To our surprise, in LCP decent crystals of photosynthetic reaction center (this was the only protein we tried, I know just one datapoint) grew at up to 50% contamination levels. Check out the paper for details:

C. A. Kors, E. Wallace, D. R. Davies, L. Li, P. D. Laible and P. Nollert
Effects of impurities on membrane-protein crystallization in different systems
Acta Cryst. (2009). D65, 1062-1073

We attribute this feature to the lipid matrix. It forms a 3D network of diffusion pathways, providing diffusion channels and barriers for different molecular species.

What does it mean for the required sample purity standard?
1. Don't give up when all you've got are 'dirty' protein samples. Set them up but do check though the identity of the target crystal, you may end up with a surprise: crystals of contaminating proteins rather than your target.
2. Of course I'd like to see a similar study carried out on soluble proteins. The meshwork of the LCP should help as well (need any help setting up LCP-based crystallization trials? - send an email, I'll fill you in with details).

All the best,
Peter

Protein crystallization via vapor diffusion setups is by far the most commonly used crystallization technique. But which format? Hanging drops or sitting drops?

The BMCD shows that many researchers favor hanging drops: I found 9,083 entries for "hanging", and 1,534 for "sitting" drops (search results using a famous search engine confirmed this trend with 4.7:1). While sitting drops are simple to prepare by hand and by robots, hanging drops are more challenging to set up. There are several advantages though when dealing with hanging drops: often the protein crystals don't stick to the surface and collect in the middle of the drop. On the other hand, some dexterity is required when you flip the drop for harvesting protein crystals with a loop. These and many more reasons have split crystallizers into two factions, each claiming that hanging drops are better than sitting drops and vice versa.
And now guess what - you can have it both!

The Kim lab has just published an ingenious protocol to get the best of two worlds (Whon et. al., J.Appl.Cryst., 42, 975-976 "A simple technique to convert sitting-drop vapor diffusion into hanging-drop vapor diffusion by solidifying the reservoir solution with agarose"):
1. prepare sitting drop crystallization experiments in a protein crystallization tray
2. add agarose solution to reservoirs and let solidify
3. incubate, image and harvest upside down or downside up - as you whish.

A solid reservoir solution.

How simple is that? 

Figure: (B) Hanging sitting drop and (A) sitting hanging drop. 

BTW - this format should also be a nice way to transport 'living crystallization experiments' without the reservoir solution splashing into the crystallization drops.

Simple does it,
Peter

Protein crystals often form at air/water interfaces, on the container wall or on dust that's present in the crystallization drop. Why not expand this concept and throw some sand, horse hair or seaweed into crystallization setups? That's exactly what Thakur et al. have done and describe in their PLOS paper. While Allan d'Arcy has previously published on the use of natural seeding materials for nucleation of protein crystallization, this paper describes a systematic study to improve the success of sparse matrix protein crystallization screening with heterogenous nucleating agents. They find that hydroxyapatite, cellulose, horse hair and dried seaweed promote crystal formation. This apparently works even better once you combine all of these particles in a cocktail. The authors encourage crystallizers to further explore and identify even better materials for this purpose.
How about a few springkles of sawdust, dandruff or couscous? I suppose that this is one of those cases where "who crystallizes is right".

All the best,
Peter