doi:10.1016/j.ymeth.2004.03.021
Copyright © 2004 Published by Elsevier Inc.
Ions from the Hofmeister series and osmolytes: effects on proteins in solution and in the crystallization process
Kim D. Collins
, 
Department of Biochemistry and Molecular Biology, University of Maryland Medical School, 108 N. Greene Street, Baltimore, MD 21201-1503, USA Macromolecular NMR Section, Structural Biophysics Laboratory, National Cancer Institute, Frederick, MD 21702, USA
Accepted 24 March 2004.
Available online 20 August 2004.
References and further reading may be available for this article. To view references and further reading you must
purchase this article.
Abstract
Sephadex G-10 gel sieving chromatography, Jones-Dole viscosity B coefficients, and solution neutron and X-ray diffraction are used to show that small ions of high charge density (e.g., sulfate, phosphate, the carboxylate, sodium, and fluoride) are strongly hydrated (kosmotropes) whereas large monovalent ions of low charge density (e.g., ammonium, chloride, potassium, and the positively charged amino acid side chains) are weakly hydrated (chaotropes). The heats of solution of the crystalline alkali halides are then used to show that only oppositely charged ions of equal water affinity spontaneously form inner sphere ion pairs, and that this controls ion binding to proteins. The net charge on a protein is a major determinant of its solubility. Finally, the surface potential difference and surface tension at an air–salt solution interface are used to generate a simple model for how ions affect protein stability and solubility through indirect interactions at the protein–solution interface. A few comments about small neutral osmolytes are also included.
 |
Fig. 1. Gel sieving chromatography of salts and solutes on Sephadex G-10. 1.0-ml samples containing 0.1 M solute in 0.1 M NaCl plus
2 μCi 3H2O, and 0.5% dextran were chromatographed on a Sephadex G-10 column (1.5 × 85.5 cm) at 30 °C and a flow rate of 0.5 ml/min. The eluant was 0.1 M NaCl, pH 7.0. Anions were chromatographed as sodium salts; cations were chromatographed as chloride salts. The double line is our best estimate of ideal behavior for solutes on Sephadex G-10. The relative elution position Kd is defined as Kd=(Ve −Vo)/(Vi − Vo), where Vi is the included volume (measured with 3H2O), Vo is the excluded volume (measured with dextran), and Ve is the elution volume for a given solute. The points labeled 1–6 represent glycine and its homopolymers through hexaglycine. EG, ethylene glycol; TCA−, trichloroacetic acid; GUAN+, guanidinium; TRIS+, protonated Tris; and THO, 3H2O. All solutes were detected by scintillation counting or specific colorimetric assays. Data from Washabaugh and Collins [4].
Fig. 2. The radial distribution functions g10(r) for Li+ (curve A), Na+ (curve B), water–water (curve C), and K+ (curve D) in liquid water. These curves measure the density of the solution measured from the isotopically substituted ion, and effectively measure the distance from the monovalent cation to the nearest solvent oxygen. Curve C measures the oxygen–oxygen distance in liquid water. Reprinted from Enderby [11]. These are the data of Neilson, Enderby, and co-workers, and represent both neutron and X-ray diffraction experiments.
Fig. 3. The first-order difference function ΔGLi(r) for Li+ in D2O. This curve measures the distance from the isotopically labeled Li+ to the nearest solvent oxygen or deuteron. Reprinted from Enderby [11]. These are the neutron diffraction data of Newsome, Neilson, and Enderby.
Fig. 4. The first-order difference function ΔGAg(r) for Ag+ (an analog of Na+) in D2O. This curve measures the distance from the isotopically labeled Ag+ to the nearest solvent oxygen or deuteron. Reprinted from Enderby [11]. These are the neutron diffraction data of Neilson and co-workers.
Fig. 5. The first-order difference function ΔGK(r) for K+ in D2O. This curve measures the distance from the isotopically labeled K+ to the nearest solvent oxygen or deuteron. Reprinted from Enderby [11]. These are the neutron diffraction data of Neilson and Skipper.
Fig. 6. Division of the group IA cations and the VIIA halide anions into [strongly hydrated] kosmotropes (water structure makers) and [weakly hydrated] chaotropes (water structure breakers). The ions are drawn approximately to scale. A virtual water molecule is represented by a zwitterion of radius 1.78 Å for the anionic portion and 1.06 Å for the cationic portion. In aqueous solution, Li+ has 0.6 tightly attached water molecules, Na+ has 0.25 tightly attached water molecules, F− has 5.0 tightly attached water molecules, and the remaining ions have no tightly attached water [14].
Fig. 7. (A) Relationship between the standard heat of solution of a crystalline alkali halide (at infinite dilution) in kcal mol−1 and the difference between the absolute heats of hydration of the corresponding gaseous anion and cation, also in kcal mol−1. Source: Morris [33]. (B) Identification of ions as chaotropes (weakly hydrated) or kosmotropes (strongly hydrated). The enthalpy of solution of chaotrope–chaotrope and kosmotrope–kosmotrope salts is positive (takes up heat), whereas the enthalpy of solution of chaotrope–kosmotrope and kosmotrope–chaotrope salts is either negative (gives off heat) or positive (takes up heat).
Fig. 8. Ion size controls the tendency of oppositely charged ions to form inner sphere ion pairs. Small ions of opposite sign spontaneously form inner sphere ion pairs in aqueous solution; large ions of opposite sign spontaneously form inner sphere ion pairs in aqueous solution; and mismatched ions of opposite sign do not spontaneously form inner sphere ion pairs in aqueous solution. A large monovalent cation has a radius larger than 1.06 Å; a large monovalent anion has a radius larger than 1.78 Å.
 |
Fig. 9. Interfacial water near the polar surface of a test solute (protein molecule). Ions inserted into the third interfacial water layer modulate the interaction of the second interfacial water layer with the first interfacial water layer (arrows). While the number of hydrogen bonds between the second and first interfacial water layers must increase from the top of the figure to the bottom, the actual number shown is arbitrary. A [strongly hydrated] kosmotrope inserted into the third interfacial water layer makes the bulk solution a poorer solvent, and causes the protein molecule to minimize its solvent accessible surface area both by becoming more compact and also by forming crystal contacts (i.e., it decreases the solubility of the protein). A [weakly hydrated] chaotrope inserted into the third interfacial water layer makes the bulk solution a better solvent, and causes the protein molecule to maximize its solvent accessible surface area (i.e., it increases the solubility of the protein); however, the dominant mechanism by which chaotropes increase protein solubility is by interacting directly with the weakly hydrated portions of the protein. See the text for more details.
Table 1. Jones-Dole viscosity B coefficients

Sources. Phosphate, formate, and perchlorate from Krestov [108]; all others from Robinson et al. [109].