Why do these crystals split up like this at their edges? This is an effect of lattice strain caused by an accumulation of impurities in the growing crystal. Filter your protein (and all the other solutions involved in the experiment) through a 0.22 micron filter. Always check the protein concentration again after you pass the protein through a filter.
If simple filtration doesn't help, check the purity of the protein on a gel. You may need to purify it more.
Another picture of crystals with obvious lattice strain problems.
These often begin as spherulites (see previous tutorial) or they grow out from a spherulite. They are extremely thin needles clustered around a single nucleation site. Often they appear “fuzzy”. The white arrows in this picture show three-dimensional crystals growing from independent nucleation sites in the same drop.
Extremely thin needles growing from a single nucleation center. Since these needles are much longer, I would not call this a “sea urchin” any longer. Notice the large 3-D crystal growing in the same drop.
Still too many but at least they are single needles. The nucleation rate is too high which is why they are too many and too thin. Try reducing the protein or precipitant concentration or both. Another method is to put a layer of oil over the reservoir in the vapor drop setup (see N. Chayen, J. Appl. Cryst 1997, vol. 30, p. 198-202). See also Tutorial 5 on seeding.
In a vapor diffusion experiment, the level of supersaturation increases fastest at the edges (i.e., on the inside perimeter) of the drop. That is because the rate of evaporation is fastest there. Faster evaporation means faster supersaturation. The needles shown here nucleated all along the inside perimeter of the drop and are growing inwards, towards the bulk volume of the drop. Take home lesson: Always check the edges of your drops carefully because if anything is going to happen, it usually happens here first.
Two-dimensional plates. It could be argued, but plates are usually considered an improvement over thin needles. The plates here are clearly growing from a single nucleation site and overlaying each other, which is far from optimal. The problem with plates is that they always diffract poorly in the thinnest direction. Thus you might get 2.5Å in one direction, and 8Å in the other, i.e., they are anisotropic.
Optimize to grow them separately and thicker. To grow them separately, seeding often helps. To grow them thicker, try an additive screen.
I call these the "Christmas trees". This is an example of tertiary dendritic growth. Dendritic means "tree-like".
Snowflakes are the best-known examples of dendritic crystal growth. This type of growth is also very common with metals and metal alloys, but these crystals shown here are definitely protein. They were on the cover the ICCBM 1993 conference program and come from Enrico Stura.
What to do about it? Do what he did: use them as seeds.
A three-dimensional crystal of P2 myelin, one of the first proteins I ever worked on. It is 0.5 mm in the longest direction. But check the diffraction before you get out the champagne. These did diffract, to 2.7Å.