The purpose of the present study was to investigate the effect of secondary structure of three model polypeptides on their apparent permeability (P(app)) across a synthetic, microporous membrane. Poly-L-lysine (PL), poly-L-glutamate (PGlu), and poly-L-lysine-L-phenylalanine (1:1) (PLP) were selected because a solution environment in which their predominant secondary structure is random coil (RC), alpha-helix, and beta-sheet, respectively, is easily achieved. The conformation of each polypeptide was verified by circular dichroism (CD). Diffusion studies were conducted under sink conditions at 25 degrees C across a microporous polyester membrane using a donor concentration of 0.02 mM for each model polypeptide. NMR was utilized to obtain a second estimation of the diffusion coefficient for each polypeptide. The equivalent hydrodynamic radii (R(e)) of the three model polypeptides were calculated using the values of the diffusion coefficient obtained by both NMR and the classic in vitro diffusion studies. The viscosity of each polypeptide solution was also determined to investigate the effect of viscosity on the aqueous diffusion coefficient. Statistical analysis demonstrated a significant (P < 0.05) difference in both P(app) and the aqueous diffusion coefficient (D(aq)), as well as the calculated R(e) values, between all three model polypeptides and there was no significant (P > 0.05) difference in the viscosity of the polypeptide solutions. Values of D(aq) and R(e) calculated from the diffusion studies were in relatively close agreement to those obtained using NMR. The logarithm of P(app) was highly correlated (r = -0.961) with the values of R(e) calculated from NMR (R(e (NMR))) rather than the mw of the polypeptides (r = 0.681). Values of the Perrin or shape factor which deviate substantially from unity are suggestive of a non-spherical or ellipsoid shape and were 1.22 +/- 0.20, 1.55 +/- 0.11, and 2.38 +/- 0.20 for PGlu, PL, and PLP, respectively. In conclusion, the observed difference in the membrane transport/diffusion of the three model polypeptides is suggested to be due to the shape associated with the secondary structure of each macromolecule, rather than the polypeptide's mw or the viscosity of the dilute polypeptide solution.