Hi David,

There are many ways to optimize crystallization conditions.

In your case, most likely in drops, where you have spherulites, the process is running too fast, in other terms your conditions are over-saturated. Most likely (you did not mention in your post) your crystals grow at high concentration of PEG. 

Several ways to control speed.

One way would be to increase size of the drop. If you use 2ul drop, make them 3ul and 4ul, try to increase the size of drops by 50% and 100%. In big drop processes go slower.

Another solution would be to decrease concentration of precipitant. You can build your new screens by optimizing those 3-4 conditions, which gave you spherulites, by varying the concentration of the precipitant.

Third solution would be to lower the temperature. If you are crystallizing at RT, try several lower temperatures (+18C, +12C, +4C).

One suggestion, try to optimize one parameter at the time. Do not vary them all at ones. That way you can find which one works the best.

Good luck!

Hi David,

All of George's suggestions are great. Definitely one variable at a time, or in a gradient of two variables, i.e. in your 24 or 96 well plate vary one component in the x-direction and one component in the y-direction. It's a good way to maintain control while seeing combination effects.

Also, you should definitely try seeding if you think nucleation is the problem. Take one of your drops with lots of crystals (spherulites) and vortex it vigorously to break the crystals (called a seed stock). This will produce micro-crystals that you can't see anymore by naked eye. Then do your normal drop size 2+2 for instance, and then add 1uL of seed stock. Try dilutions of this as one of your variables. Sometimes simply providing an already formed nucleation site will produce much larger, or even completely different, crystals.

Good luck!

Hi David,

I would lower the salt concentration in the protein buffer.

Hi David:

One final tip: try to change the Protein:PreciPitant ratio in your drops.  Typically a 1:1 ratio is used.  I would try 2P:1PP or even lower.  If after some time you get no crystals, move your plate to 4 C (by decreasing T you are affecting the solubility and the supersaturation).  If nothing happens, do not trash your plate: add salt (NaCl or MgSO4) to you reservoir solutions, reseal the system and continue the experiment (you are increasing the ionic strength of the reservoir, promoting the dehydration (i.e. supersaturation) of the drop).  Repeat this procedure as necessary (check Acta Cryst. (2005). F61, 435–438)

Can you share information on your protein?  It is possible to predict many useful parameters from the sequence to optimize the time consuming trial and error approach.  In any case do not give up and think that these conditions that yield spherulites may be only promising but not the good ones.

 Good luck.

I recommend you to use a high-throughput screening platform for optimization of the conditions your proteins to form the desired crystals. You can use a factorial design to test all the potential parameters affecting protein crystallization simultaneously.  Otherwise, you will waste much time if you are attempting to check individual parameters at a time.

Good luck!

If the sherulite occur at high precipitant concentration, but nothing happens at a lower concentration, you have in fact bracketed the condition under which crystallisation might occur. Try to explore the region in between.

Others suggested to change a variable at a time. It is, however, possible to change several factors at the same time  in a sparse-matrix approach originally suggested by Taguchi, you then need only 2k+1 instead of n^k experiments to explore the system (k = number of factors, n = number of values per factor). Key to this are balanced (orthogonal) matrices, these you can take from the literatur (e.g., http://www.itl.nist.gov/div898/handbook/pri/section5/pri56.htm)