A model of protein-colloidal gold interactions.

We prepared homogeneous populations of colloidal gold particles of various sizes. These were analyzed for size distribution and number of particles per unit volume. On exposure to increasing concentrations of insulin, myoglobin, protein A, peroxidase, serum albumin, galactosylated serum albumin, lactoferrin, transferrin, catalase, low-density lipoprotein, ferritin, and polymeric IgA, protein binding was a saturable process. Using serum albumin, we verified that a reversible equilibrium was reached within 15 minutes. Scatchard analysis of the interactions between all of these proteins and the gold particles resulted in a single component, linear relation. For a given particle size, the number of binding sites for various proteins was inversely proportional to their molecular weight. Conversely, when the size of particles was varied, the number of binding sites was directly proportional to the average area of each gold particle. All results are compatible with a monomolecular shell of protein surrounding the particle at saturation, the binding capacity being inversely proportional to the projection area of the protein. We present direct morphological evidence for this model. The affinity of the various proteins for the colloid also increased with molecular weight, and was not related to the protein isoelectric point. For globular proteins, the monomolecular shell model makes possible prediction of the number of molecules that will saturate a gold particle, if the average diameter of the gold particles and the molecular weight of the protein are known.