Are molecular weights of proteins determined by superose 12 column chromatography correct?

Our research on several proteins indicates that accurate molecular weights cannot be determined by Superose 12 column chromatography. In support of this statement, we present data on molecular weights of purified red kidney bean alpha-amylase inhibitor (RKB alphaAI) and white kidney bean alpha-amylase inhibitor (WKB alphaAI) to document this problem. The molecular weight of purified RKB alphaAI determined by Sephadex G-100 gel filtration, polyacrylamide gel electrophoresis, Superose 12 gel filtration and cDNA were 49.0, 51.0, 22.9, and 49.805 kDa (not glycosylated), respectively. The molecular weights of WKB alphaAI by several methods were as follows: Sephadex G-100 gel filtration, 51.0 kDa; Superose 12 gel filtration in 0.2 M NaCl buffer, 23.1 kDa; polyacrylamide gel electrophoresis (PAGE), 51.0 kDa; sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE), 45.0 kDa; multiangle laser light scattering (MALLS), 49.940 kDa; laser-assisted time-of-flight mass spectrometry (LATOFMS), 56.714 kDa; and cDNA sequence (with 12.2% carbohydrate), 55.9 kDa. The data indicate there is ionic interaction between proteins and the matrix of Superose 12 in low ionic strength buffers and hydrophobic interaction at higher ionic strength buffers. Researchers should be cautious when using Superose 12 columns for molecular weight determinations.