In this review, Studying protein-ligand interactions using protein crystallography. Crystallography Reviews (2005) vol. 11 (1) pp. 61-71, McNae et al. discuss experimental approaches for ‘soaking-in’ ligands into protein crystals where the major problem is poor solubility of the ligand.
Extracted from the reference above:To maximize the formation of (P.L) complex we need as high a concentration of ligand as possible. Most biologically important complexes have dissociation constants (Kd) in the micromolar to nanomolar range. Using the law of mass action and the definition of occupancy, we can estimate that if the Kd for a protein–ligand complex in solution is say 10mM then the required concentration to give near full occupancy will be about 100mM (i.e. about ten times the Kd value)
Other useful readings:
Hassell et al. Crystallization of protein-ligand complexes. Acta Crystallogr D (2007) vol. 63 (Pt 1) pp. 72-9