Protein Cystallization: 10 Questions to Consider
Dec 27, 2009
10 questions to consider before setting up protein crystallization trials:
1) How pure is your protein sample?
2) What chemicals are present in the protein buffer?
3) Is the protein modified such as phosphorylated, glycosylated or methylated?
4) What temperature and pH is the protein stable?
5) What concentration causes the protein to precipitate out of solution?
6) Does the protein have any known substrates, metals, inhibitors or ligands?
7) Have related proteins been crystallized?
Is the protein sensitive to proteolysis?
9) Does the protein have cysteines?
10) How has the protein been stored?
Feel free to add any questions that you consider in the comments.
Dima
December 27th, 2009 at 2:58 PM #
… And if you don’t have an answer to any of these questions – or even to all of them – set up crystallization trial anyway. Does not take that much of an effort and you might get lucky. Having some crystallinity to begin with does wonders to ones determination to get answers to some of these questions in a hope to improve xtals. Particularly when everyone has all kind of sad stories when you do have very good answers to every single question about (and more) but the damn thing still won’t crystallize.
Dr_SNO
December 27th, 2009 at 10:45 PM #
I’ld go with Dima,
Set up the crystal trays and if nothing happens, start asking questions. Ignorance is bliss.
Although knowing the answer to “7) Have related proteins been crystallized?” could save you time. Is there a “real need” for another structure beyond your “intellectual” curiosity?
Artem
December 27th, 2009 at 11:51 PM #
Happy holidays! May the Arctic Badger bury many gifts under your Holiday Shrub.
General remarks:
You should know (7) before you set out to clone the damn thing.
You better know (2) and (9) at least if you work in my lab
(3) is what MS analysis is made for.
(4) good luck with that.
(5) ditto. You will know if it’s too concentrated – when the concentrator turns into a snow globe. Usually it’s too late at this point.
(6) be ready for surprises once you solve the structure.
(8) if it is – you will know way ahead because it will progressively break down during purification
(10) ideally it wasn’t stored for very long – purification is a single day affair these days, set up crystals at the end of the day. One should strive to make it so for every protein as it helps eliminate a major concern.
You should know (1) automatically, to some degree. If you’re philosophically inclined then I wish you luck figuring out purity taking into account the inherent micro-heterogeneity of protein samples. Don’t forget that wrong amino acids do get incorporated in small amounts instead of the right ones. Does it matter? No! Not to crystallization anyway. It does matter to the FDA (that’s why characterizing protein drugs is a major PAIN IN THE REAR).
So, in conclusion:
1. answering some of these questions consumes more protein than setting up crystallization.
2. crystallization is a way to answer some of these questions – for example pH tolerance patterns become quite apparent after some very basic screening.
Sean
December 29th, 2009 at 8:06 PM #
Sorry in getting back to you guys.
I am quite sick at the moment.
Artem
December 29th, 2009 at 8:12 PM #
Get well soon!
Incidentally, if you know someone who may be interested in a summer internship in my lab (at Monsanto in St. Louis), spread the word please
Artem
Paul
December 30th, 2009 at 3:06 PM #
The “shoot first and ask questions later” approach is good so long as you have protein to burn, or it’s easy to purify more (though for the latter, reproducibility can become a big factor); drops that don’t contain crystals can still provide valuable information. To answer questions 3), 6), 7), and 9) doesn’t necessarily require you to use protein; a lot of reliable data to answer these questions is already out there (assuming your expression system produces protein as expected), though it would be nice to be able to access them through a central portal. Hmmmmm…. that gives me an idea!
I’ve been trying to write some code to crudely answer questions related to 4) and 5) in a roundabout way, but unfortunately some of the companies that make crystal screening kits haven’t been too helpful in providing me the data I’d like to develop it.
roman
July 26th, 2010 at 1:40 PM #
Question: Recently I set up trays to crystallize a protein I purified and all the conditions were the same. Only a few of the trays had crystallized, yet I anticipated all of them to develop crystals. What could have gone wrong?
Artem
July 27th, 2010 at 8:56 PM #
This is not really a big surprise because crystallization of macromolecules can be influenced by subtle changes of parameters that are either not under your complete control, or that you don’t even realize
For instance – when you say that all conditions were ‘the same’ – that’s not entirely true. I’ve seen proteins that only crystallized on the edge wells of a 96-well plate (filled with ‘the same’ conditions) or would only crystallize if they were set up in the evening and left undisturbed over the week-end. Also, even under perfect conditions nucleation is a stochastic process, especially if you worked your conditions out so that you only expect to get a few big crystals per drop (i.e. nuclei form slowly/infrequently).
Now, even if all other stuff is truly ‘the same’ – is your crystallization solution exactly the same as the solution used to produce your previous hits? Things happen and even a few days can make a difference (especially if your solutions contain volatile components, oxidizable stuff, etc.).
Finally, is your protein completely preserved? Even if you’re taking fresh aliquots out of -80C storage *every time* (this is what I do) – there is still a chance for variation. My favorite example is a non-heme iron protein I used to work with (EGLN or PHD) that even at -80C used to lose activity with a half-life period of about 3 days. Even in liquid nitrogen the damn thing would go bad within a month or so. So we had to make it fresh for every crystallization. Interestingly it stored OK inside frozen E. coli pellets.
At any rate, what you’re experiencing is not entirely atypical nor even unexpected. If you would like to get lots of crystals reproducibly, I recommend working out an easy seeding protocol – 99% chances are that your troubles are caused by irregular nucleation.
Good luck,
A.
Sean
July 31st, 2010 at 3:10 PM #
Hi Roman,
I have had this problem as well. An area that I believe can be improved in protein crystallization is the purity of the chemicals used in screens. For example, we offer HEPES with a purity of 99.9 %… Currently, I don’t know of any company that makes crystallization screens that offers HEPES at this purity or better (if someone does feel free to link to it).