A kinetic study in vitro of the reoxidation of interchain disulfide bonds in a human immunoglobulin IgGLk. Correlation between sulfhydryl disappearance and intermediates in covalent assembly of H2L2.

After reduction by dithiothreitol and removal of the reductant by molecular sieve chromatography, the four interchain disulfide bonds of the human IgGlk protein Fro reoxidize in the presence of oxygen and trace metal ions. The six molecular components of the reoxidation--L (light chain), H (heavy chain), HL, H2, H2L, H2L2--are quantitatively determined from polyacrylamide gels containing sodium dodecyl sulfate and the time-dependent sulfhydryl titer is measured with 5,5'-dithiobis-(2-nitrobenzoic acid). The rates of H2L2 covalent assembly depend on pH in an unexpected way: If the reduced protein is chromatographed at pH 3.2 and then adjusted to pH 7.5 (25 degrees, ionic strength equals 0.14), H2L2 formation proceeds rapidly, with half-times ranging between 20 and 40 min. In contrast, if chromatography is carried out at pH 5.5 before adjusting to the same final conditions, the half-times for H2L2 formation are considerably longer (120-180 min). The half-times in the former case approach the somewhat faster rates of H2L2 assembly observed in pulse-chase experiments with various types of mouse, IgG-producing cells [Baumal, et al. (1971) J. Exp. Med. 134, 1316-1334]. To facilitate comparison of experiments and models, we plot the concentrations of the six components against the corresponding number of sulfhydryl equivalents per mole of Fro. The respective plots for the pH 3.2 leads to 7.5 and 5.5 leads to 7.5 experiments are very similar despite the rate differences. Moreover, these plots differ significantly from the calculated plot for a hypothetical random reoxidation in which the intrinsic probability for formation of each correct HL and H2 disulfide bond is assumed equal and independent. It is concluded that the in vitro reoxidation of Fro (i) is other than random; (ii) involved a pathway of pathways with HL, H2, and H2L precursors; and (iii) involves at least some kinetic cooperativity in bond formation, since no model bases solely on independent bond formation adequately accounts for the results. The models were used also to examine the cellular assembly pathways of mouse IgG proteins.