X-ray crystallography
A drop of protein mixed with reservoir solution is sealed over a larger volume of the same reservoir. Water moves through the vapor from the drop to the more concentrated reservoir until their precipitant concentrations match, so the protein slowly concentrates into supersaturation.
Our ruleSimulations prepare you for the bench but don't replace hands-on practice.
Vapor-diffusion crystallization
A drop of protein mixed with reservoir solution is sealed over a larger volume of the same reservoir. Water moves through the vapor from the drop to the more concentrated reservoir until their precipitant concentrations match, so the protein slowly concentrates into supersaturation.
How it differs from the standard method
- Sitting drops (on a shelf) suit robots and nanolitre volumes; hanging drops (under a coverslip) are easier to harvest from.
- The drop ratio sets where the protein ends: 2:1 doubles the final protein concentration relative to 1:1.
- Start with sparse-matrix screens of 96 conditions, then optimise hits on fine grids of precipitant against pH.
1Why a crystal
A single protein molecule scatters far too few X-rays to measure. A crystal stacks around 10¹³ copies in identical orientations on a lattice, so their scattering adds up in a few sharp directions and cancels everywhere else.
Grow it
Bring a pure, monodisperse protein slowly into supersaturation so that a few ordered nuclei form and grow instead of amorphous precipitate.
Measure it
Cool the crystal to 100 K, rotate it in a synchrotron beam and integrate every spot. Resolution, completeness and dose are set here.
Solve it
Intensities lose the phase of each wave. Molecular replacement borrows phases from a similar structure; experimental phasing gets them from heavy atoms or anomalous scatterers.
2Supersaturation, not concentration
What matters in a crystallisation drop is how far above its solubility the protein sits. Just above the solubility curve a drop is metastable: an existing crystal would grow, but nothing nucleates, so the drop stays clear.
3Resolution is a property of the data
Spots far from the beam centre come from closely spaced planes and carry fine detail. Their intensity falls off with the crystal's disorder (the B-factor), so at some angle they sink into the noise.
4How it fails
- Precipitate or showers in every drop: the protein was pushed too far past the solubility curve, or aggregates in the stock nucleated everything at once.
- Clear drops forever: the drop never reached the nucleation zone. Raise protein or precipitant, or seed into the metastable zone.
- Ice rings at 3.90, 3.67 and 3.44 Å: the crystal was not cryo-protected and its solvent froze as crystalline ice.
- Fading outer spots during the sweep: too much dose. Beyond about 30 MGy a cryo-cooled crystal has lost a large fraction of its diffracting power.
- A molecular-replacement search with no clear peak: the model is too different, carries wrong loops, accounts for too little of the cell, or the space group is wrong.
- A low Rwork with no Rfree: the model may simply be fitted to noise, and there is nothing to show it.
Key terms
Practise these as flashcards →Protein concentration divided by its solubility under the drop's conditions. Crystals nucleate only when it is well above 1.
The smallest lattice spacing for which meaningful intensities were measured. Lower numbers mean finer detail.
Fraction of the unique reflections to a given resolution that were actually measured.
Correlation between intensities merged from two random halves of the data; used to decide the resolution cutoff.
Angular spread of the crystal's ordered blocks. High mosaicity smears spots over a wider rotation.
Energy deposited per kilogram of crystal. About 30 MGy is the accepted limit for cryo-cooled protein crystals.
The detector records intensities, the squares of the wave amplitudes, but not their phases; a map needs both.
Translation-function Z-score and log-likelihood gain from a molecular-replacement search; TFZ above 8 is usually an unambiguous solution.
Disagreement between observed and model amplitudes, for the reflections used in refinement and for a 5% test set that never was.
When to use it
- You need atomic detail, for example to see how a ligand or inhibitor sits in a binding pocket.
- The protein, or a well-behaved domain of it, can be made pure and homogeneous in milligram amounts.
- You will solve many structures of the same protein with different ligands, where a reproducible crystal form makes each new structure quick.
Compared with alternatives
No crystals needed, and better for large, flexible or membrane complexes; usually lower resolution for small proteins below about 50 kDa.
Structure and dynamics of small proteins in solution, without crystals; limited to roughly 30–40 kDa without special labelling.
Overall size and shape in solution in an afternoon; no atomic detail, but useful for checking that a crystal structure matches the protein in solution.
Sub-techniques
Choosing one here switches the whole page, including the protocol and the lab, to it.
A drop of protein mixed with reservoir solution is sealed over a larger volume of the same reservoir. Water moves through the vapor from the drop to the more concentrated reservoir until their precipitant concentrations match, so the protein slowly concentrates into supersaturation.
- Sitting drops (on a shelf) suit robots and nanolitre volumes; hanging drops (under a coverslip) are easier to harvest from.
- The drop ratio sets where the protein ends: 2:1 doubles the final protein concentration relative to 1:1.
- Start with sparse-matrix screens of 96 conditions, then optimise hits on fine grids of precipitant against pH.