Dilute solution properties of proteoglycan fractions from bovine nasal cartilage

Abstract

Abstract Fresh proteogycans (adult bovine nasal cartilage) isolated from the densest portion of a dissociative density gradient had a weight‐average molecular weight of ca. 10 6 in 4 M guanidine hydrochloride (GdnHCI) by light scattering. Fractions of such material obtained by elution with 4 M GdnHCI from 2% agarose gel, both normal and cross‐linkd, has proteoglycan subunit molecular weights ranging from 0.8 to 2.6 × 10 6 and root‐mean‐square radii ranging from 35 to 52 nm in the same solvent. The protein molecular weight per proteoglycan subunit was about 1.2 × 10 5 and that of keratan sulfate about 1.8 × 10 5 , both independent of total molecular weight. A random‐flight “graft copolymer” model having uniform side chains of chondroitin sulfate (40 disaccharides) and keratan sulfate (15 disaccharides) and a random‐coil polypeptide back bone was used to estimate the unperturbed radius, whihc was about 19 nm for a mol wt of 1.5 × 10 6 . Experimental light‐scattering data for fractions were fitted very well by theoretical curves for the particale scattering factor for both linear and appropriate branched polymers. Examination of coil expansion on the basis of perturbation calculations for branched polymer models suggested that expansion did not account for the experimentally observed radii in terms of unperturbed radii calculated from the model. A possible explanation is that substantial local stiffening of the polypeptide chain due to substitution of side‐chain clusters increases the unperturbed radii. The intrinsic viscosity [η] is 4 M GdnHCI ranged from 120 to 180 ml/g, and could be interpreted in terms of th eequivalent sphere model; the Flory number has approximately its normal value for flexible linear polymers. The treatment of the sedimentation coefficient by this is less successful, since the Man delkern‐Flory parameter β apparently increases with increasing molecular weight; average value are similar to those for flexible polymers, but the variation in β makes this method useful only for rough estimation of molecular weight of proteoglycans. Molecular weights of purified proteoglycans are the same in 0.2 M NaCI as in 4 M GdnHCI, while crude preparations gave higher molecular weights in 0.2 M NaCI, probably because of association due to incomplete removel of “linking” proteins.