We measured the free fraction of 25-hydroxyvitamin D (25OHD) in human serum and determined that 25OHD bound to a component with an affinity constant of 7 X 10(8) M-1 and a concentration of 4.5 X 10(-6) M. This concentration was equal to that of the vitamin D-binding protein (DBP) in the same serum sample. We removed DBP from the serum using actin affinity columns and found that the affinity for 25OHD of the remaining serum components was equivalent to that of human serum albumin (6 X 10(5) M-1). We then measured the free fractions of 25OHD, DBP, and albumin in normal and cirrhotic subjects. We calculated that 88 +/- 3% (+/- SD) and 83 +/- 8% of the 25OHD were bound to DBP in the serum of normal and cirrhotic subjects, respectively. We compared previously reported data for the free fraction and the free concentration of 1,25-dihydroxyvitamin D in these subjects with the current data for the free fraction and free concentration of 25OHD. The total concentrations and free fractions of both metabolites correlated to each other and to the DBP and albumin concentrations in these subjects, but the free concentrations of these metabolites did not. We conclude that 25OHD, like 1,25-dihydroxyvitamin D, is transported in blood bound primarily to DBP and albumin. Changes in the concentrations of DBP and albumin affected the total and free fractions of 25OHD in serum, but the actual free concentration of 25OHD was independent of such changes.

The mammalian plasma vitamin D binding protein (DBP), or Gc-globulin, is recognized to have at least two functional properties: sterol binding and G-actin sequestration. Affinity labeling of the sterol binding site with the radioactive electrophilic ligand, 3 beta-(bromoacetoxy)-25-hydroxycholecalciferol, followed by limited proteolysis, permitted the isolation and identification of three overlapping peptides in the amino terminus of the molecule. When G-actin affinity chromatography was applied to other proteolytic fragments, two fragments from the carboxy terminus of the molecule were isolated and identified. Another, large, tryptic fragment displayed both sterol- and actin-binding properties. The amino-terminal assignment of the sterol-binding domain was confirmed by demonstrating sterol-specific binding by an in vitro transcribed and translated product of a mutated rat DBP cDNA encoding a protein truncated in its carboxy terminus. The sterol-binding domain was localized to the region between the first-amino-terminal disulfide bond, and the actin-binding domain was found between residues 350 and 403. A high degree of sequence conservation in these regions was found among human, rat, and mouse DBP's. These functional domain assignments confirm the apparent independence of these two binding activities and help to explain the observed triprotein complex of DBP-actin-DNase I and the competition between DBP and profilin for G-actin binding. Our findings should facilitate more precise delineation of the binding domains by site-directed mutagenesis experiments.