Human plasma fibronectin structure probed by steady-state fluorescence polarization: evidence for a rigid oblate...

In order to more clearly define the structure of human plasma fibronectin (PFn) under physiologic buffer conditions, we determined the mean harmonic rotational relaxation times (rho H) of PFn and the thrombin-derived 190/170-kDa PFn fragment using steady-state fluorescence polarization. These measurements utilized the long lifetime emission (tau = 1.2 X 10(-7) S) exhibited by 1-pyrenebutyrate, which had been covalently attached to amino groups at random sites on the PFn subunit. Our data analysis assumed that two independent processes depolarize the fluorescence exhibited by the dansylcadaverine and 1-pyrenebutyrate conjugates of PFn: (A) rapid (rho H less than 10(-9) S) "thermally-activated" localized rotational motion of the protein side chains bearing the fluorescent probe [Weber, G. (1952) Biochem. J. 51, 145-154] and (B) slow (rho H approximately 10(-6) S) temperature-independent global rotational motion of the whole PFn molecule. Since only the rho H associated with the latter process is a true hydrodynamic parameter (i.e., sensitive to size and/or shape of the PFn molecule), we utilized isothermal polarization measurements to discriminate against the interfering signal arising from "thermally activated" probe rotation. The rho H (4.4 +/- 0.9 microseconds) derived from an experiment in which pyrene-PFn fluorescence polarization was monitored as a function of sucrose concentration at constant temperature is 7 (+/- 1.4) times longer than that predicted for an equivalent hydrated sphere. We propose that "thermally activated" probe rotation gives rise to the nearly 100-fold shorter PFn rho H values previously reported in the literature. Consequently, our data exclude all previous models which invoke segmental flexibility of the PFn peptide backbone. The simplest hydrodynamic model supported by our fluorescence data is an oblate ellipsoid with an axial ratio of 15:1. All prolate models can be unambiguously excluded by this result. We estimate that the disk-shaped PFn molecule has a diameter and thickness of 30 and 2 nm, respectively. Electron microscopy of negatively stained PFn specimens on carbon also showed PFn to have a compact rounded structure. The much faster rotational relaxation rate of the pyrene-190/170-kDa PFn fragment (rho H = 0.92 +/- 0.11 microseconds) compared to pyrene-PFn indicated that this monomeric PFn fragment, like native PFn, had an oblate shape under physiologic buffer conditions.