Crystallisation principles

Dr Rupert C. Wilmouth School of Biological Sciences NTU

Protein crystals
Proteins are inherently hard to crystallise and protein crystals are fragile compared to crystals of inorganic compounds. Protein crystals are held together by weak forces, primarily hydrogen bonds. Crystals only diffract X-rays when wet. Initial experiments in the early 20th century using dried protein crystals gave no diffraction. Finally, in 1934, Bernal and Hodgkin measured diffraction from pepsin crystals that were still in their mother liquor, i.e. wet.

Growing crystals
In order to grow protein crystals, the protein must be pure (usually >95%), homogeneous, and in high concentration (> 5 mg mL-1). Since pH greatly affects crystallisation, a buffer must be present (e.g. MES, HEPES, TRIS). Finally, a precipitant is also required which causes the protein to precipitate out of solution. Under careful controlled conditions (especially temperature and evaporation rate), protein solutions will sometimes become supersaturated instead of precipitating. At this critical point, crystals may form. Nucleation, the formation of the first ordered aggregates, is a key event. Nucleation conditions are sometimes difficult to reproduce, and thus seeding procedures with preformed crystalline material may be invaluable to obtain reproducible results.

Vapour diffusion
A drop of protein solution is mixed with an approximately equal amount of crystallisation solution from a reservoir. The reservoir or well has a greased rim and is then sealed with a cover slide (either siliconized glass or plastic). In the resulting closed system, water vapour diffuses from the hanging drop into the reservoir, which contains about twice the precipitant concentration than the drop. This allows the protein to become more concentrated and potentially allows crystallisation to occur. The drop can either be placed on the underside of the cover slide (hanging drop) or placed on a plastic support above the surface of the reservoir (sitting drop). The method can be used for relatively large drops (2-20 L) suspended above 100 L 1 mL of reservoir solution in 24-well plates. It can also be used for micro-drops in small crystallisation strips. The smallest drop size is limited by the evaporation during the time delay from drop application to sealing of the well.

Hanging drop v. sitting drop

Hanging drop Cheaper Easier to setup Easier to harvest crystals

Sitting drop Able to use larger drops Easier for soaking Safer for transport

Multiple drops possible on one cover slide

Crytallisation kinetics phase diagrams

Vapour diffusion no crystals formed

Crytallisation kinetics phase diagrams

Vapour diffusion crystals formed

Microbatch
A 1-2 L drop of protein solution mixed with an aliquot of crystallisation solution (containing precipitant, possible additives, buffers, etc.) is pipetted onto the surface of a oil-covered microtitre plate well. The drop then sinks to the bottom of the well, and is isolated from the environment. Variations include placing the individual components under oil, with either protein drop or crystallisation solution first (creating two kinetically different scenarios), and using varying ratios of water permeable oils to allow water to diffuse into the environment. It should be noted that alcohols, detergents, and lipids can diffuse into the oil (and, to a much smaller degree, into the polymer of the well material). The micro-batch method is well suited for miniaturisation and automation, as there is no time delay between the application of the drop and the sealing. The ability to work with minute drops of 1 L or below makes the method very suitable for screening.

Crytallisation kinetics phase diagrams

Microbatch constant protein concentration

Other crystallisation methods

Another method used to bring about gradual change in the relative concentrations of the protein and precipitant is dialysis. This ranges from traditional dialysis tubing for milliliter samples to the use of specialised dialysis buttons which make it easy to dialyse as little as 10 L of sample. The use of dialysis permits one to increase the concentration of precipitant while essentially holding protein and other component concentrations steady. It is also possible to induce crystallisation through a change in pH by dialysing with a pH gradient. Hydrogels can be used for crystal growth with the crystallisation solution soaking a polymeric network. Two common examples are agarose and silica. A silica hydrogel can be formed by reacting sodium silicate with acetic acid. Gels can reduce nucleation and allow crystals to grow larger. They can also provide stability to fragile crystals. The main disadvantage is the extra hassle involved and it can also make crystal manipulation more difficult.

Sparse matrix screens

Sparse matrix screens are a common method for initially determining potential crystallisation conditions. A carefully chosen wide selection of buffers, pH values, salts and precipitants is screened. The most widely used sparse matrix screen is the one originally sold by Hampton and now copied by Molecular Dimensions and Sigma-Aldrich (all three are identical). Many other sparse matrix screens have been devised (see the Molecular Dimensions catalogue), and it is worth trying these if the Hampton screen does not yield interesting results. Remember to make sure that your protein solution is only weakly buffered, otherwise the pH value of the screen will be meaningless. Also, ensure that no other salts or additives have been added to your protein solution via the purification procedure.

Empty drop with some light skin

Empty drop with light precipitate

Empty drop with heavy precipitate

Heavy precipitate and crystals

Oil droplets phase separation

Needle cluster

Single needles

Needles from oil droplets

Twinned plates

Rectangular crystals

Hexagonal crystals (lysozyme)

Small thin crystals

Small rectangular crystals

Salt crystals

Dried out PEG droplet

Air bubble

Fungal growth

Crystal optimisation
Dr Rupert C. Wilmouth School of Biological Sciences NTU

Optimisation of conditions
Grid screens Temperature (e.g. 4C, 10C, 18C) Additives Cations (anions) Substrates/inhibitors Limited proteolysis Different protein constructs

Grid screens
Grid screens are useful for screening two conditions against each other. Three dimensional grid screens are also possible by using multiple trays. Grid screens take a fair amount of effort to make up, therefore choose your conditions carefully. Also, take extreme care not to make errors one mistake and you may miss the correct conditions. In general, screen precipitant conditions more widely than pH. Most proteins crystallise between pH 6 and pH 8.5. Make sure you use the correct buffer (i.e. within 0.5 pH of the pKa). For example, you find some promising looking precipitate in Hampton condition 36 (0.1 M TRIS pH 8.5, 8% (w/v) PEG 8000). What grid screen might one use next?

[PEG 8000] (w/v) 4% 0.1 M HEPES pH 7.0 6% 8% 12% 16% 20%

0.1 M HEPES pH 7.5

0.1 M TRIS pH 8.0

0.1 M TRIS pH 8.5

Additives
There are a huge number of additives that can be tried. Usually this is the last thing to be screened. Additives are most useful when your crystallisation conditions are almost correct, for example, when you want to change the morphology of your crystals or increase their size. Examples include cations (e.g. Ca, Fe, Cd, Zn, Ni etc.), detergents (e.g. n-octyl-,D-glucoside, NDSB-195 etc.), chaotropic agents (urea, guanidine), sugars (e.g. sucrose, glucose etc.), alcohols (e.g. methanol, ethanol, isopropanol, glycerol, 2methyl-2,4-pentanediol etc.).

Substrates/inhibitors
If there are substrates, substrate analogs or inhibitors available, then co-crystallisation should always be attempted. The presence of a molecule bound in the active site often makes the protein less flexible and hence easier to crystallise. When you solve the structure, having a substrate or inhibitor bound in the active can give immensely useful mechanistic information. Be careful when using substrates to ensure they are not turned over by the enzyme. You can use non-hydrolysable substrate analogs (e.g. ADPNP instead of ATP) or inactive conditions (e.g. low pH).

Different protein constructs

Often it is easier to try a different protein construct than trying the thousandth different additive in the hope of improving your crystal quality. Possible changes include moving the location of the 6His tag, deleting the 6His tag (native purification), using a different tag (e.g. GST), performing Nterminal or C-terminal truncations (try deleting residues, one at a time). Other possibilities may include changing the expression system, fungal, insect or mammalian cells are alternatives to bacterial expression.

Seeding
A seed provides a template for the assembly of molecules to form a crystal with the same characteristics as the crystal from which it originated. It is important to differentiate the process of crystal growth from nucleation. In general, the degree of supersaturation required for nucleation is higher than that required for crystal growth. Normally, during aggregation, there is an equilibrium between the formation of ordered nuclei (reversible) and the formation of precipitate (irreversible). When crystal seeds are added, the equilibrium can shift towards crystal formation, and avoids the random nature of spontaneous nucleation.

Seeding
There are three commonly used types of seeding: Microseeding adding very small pieces of crushed protein crystals to the crystallisation drop. Macroseeding adding an intact, already grown crystal to the crystallisation drop. Streak seeding similar to microseeding but uses a fine hair (often a cats whisker is best) to pick up small protein crystal fragments.

Microseeding
This is generally the easiest and most reliable form of seeding. A few small crystals, or one large crystal, is broken up to a fine powder using a needle. A small amount of mother liquor containing this fine powder is added to an aliquot of well solution and vortexed well. Serial dilutions of the microseed solution can then be carried out (e.g. 1:10, 1:100, 1:1000). Finally, a small quantity (e.g. 0.5 or 1 L) of the microseed solution is added to the protein crystallisation drop. Often the microseed solution is added at the same time that the crystallisation drop is set up. However, it is important to remember that adding the microseed solution to an already equilibrated drop can be advantageous.

Streak seeding
This is a quick method of seeding, and can be used to seed a series of drops rapidly. The main difficulty is to find a suitable fibre to use, cats whiskers are excellent (if you can find a willing cat!). For ease of use, superglue the whisker onto a plastic rod. Degrease the whisker with ethanol or methanol, and then rinse with distilled water. Either, touch an existing crystal with the fibre to dislodge a few seeds, or stroke the fibre through a microcrystalline precipitate. Then introduce the seeds into a fresh drop by stroking the fibre in a straight line through the drop.

Macroseeding
This is perhaps the most difficult seeding technique to use successfully. A single crystal, free from twinning or other crystalline deformities, is selected. This crystal is then washed in a slightly dissolving solution to remove the top layer which may contain invisible defects. The solution must not cause excessive etching or cracking. After washing several times, the crystal is then transferred to fresh drop, preferably pre-equilibrated. Extreme care has to be taken when handling the crystal to avoid damage and the generation of microseeds. Needles are particularly difficult to handle.

Cryo-conditions
These days, mounting of most protein crystals on an X-ray diffractometer is done at cryogenic temperatures, usually around 100K. A small nylon loop is used to pick up a crystal, immerse it a cryoprotectant solution for a short period of time, and then either plunge it in liquid nitrogen (or sometimes liquid propane), or place it directly on the goniometer head in a stream of cold nitrogen gas. The choice of cryoprotectant is critical. It must prevent ice formation and it must not damage the crystal. Always screen several cryoprotectants (e.g. glycerol, ethylene glycol, MPD, PEG 400, sucrose etc.) at several different concentrations, and at several different soaking durations. Do not forget the possibility of mounting your crystal at room temperature in a quartz capillary. This is a very difficult technique to master, but remains the best way of checking the diffraction of your crystals.

Cryo-annealing and dehydration

These are last ditch attempts to improve the diffraction of your crystals. Cryo-annealing is where your frozen crystal is (briefly) defrosted before being refrozen. Occasionally the mosaicity (and less often the resolution) can dramatically improve. Dehydration is the intentional drying out of your crystal. This can reduce the water content of your crystal, and hence perhaps improve the resolution. Two possibilities: stepwise increase of your precipitant concentration, or use a hanging drop suspended above a humectant solution (e.g. 50% or 75% glycerol).