Ultracentrifugation in swinging-bucket and fixed-angle rotors evaluated for isolation and determination of high-density lipoprotein subfractions HDL2 and HDL3.

We evaluated the density-gradient ultracentrifugation method in a swinging-bucket rotor (Anal Biochem 111, 149-157, 1981) in a slightly modified version for isolation and determination of high-density lipoprotein (HDL) subfractions. We prestained the serum with Coomassie Brilliant Blue R, which did not change the hydrated densities of the lipoproteins, and after only 2.2 X 10(8) gav . min obtained an equilibrium distribution of the lipoproteins along the gradient. The density distribution of the HDL of 120 sera obtained from apparently healthy persons and from patients with different types of hyperlipoproteinemia was bimodal. The HDL2 could be isolated in the density range 1.072-1.098 kg/L and the HDL3 at 1.100-1.176 kg/L, the latter fraction being more heterogeneous. At a solvent density of 1.100 we obtained similar results for HDL2-and HDL3-cholesterol by ultracentrifugation in two different fixed-angle rotors with tube angles of 15 degrees or 35 degrees. Independent of the rotor and the ultracentrifugation technique, subfractionation at d = 1.100 resulted in more distinct stained entities than ultracentrifugation at d = 1.125. In the swinging-bucket rotor procedure, interference by sinking pre-beta-lipoproteins was minimized because, having hydrated densities between 1.058 and 1.075, they could be removed without aspirating the HDL2. The method is both accurate and precise. For HDL2- and HDL3-cholesterol determined in a thawed frozen serum pool, CVs were 8.8 and 6.3%, respectively (n = 18).