Liu et al. (2010a) determined the relative susceptibility of all of the disulfide bonds of a human IgG1 to reduction, and found that inter-chain disulfide bridges between the heavy chain (HC) and the light chain (LC) were the most labile, followed by the two HC–HC disulfide bridges of the hinge (the most N-terminal of which was more susceptible than the other). L, CL, VH, and CH1 domains, and finally the CHstep 3 domain, which was most stable to reduction ( Liu et al., 2010a ). They also determined that the LC–HC inter-chain bonds of IgG1? were more labile than those of IgG1? ( Liu et al., 2010a ), a result similar to that found by Montano and Morrison (2002) with IgG2? vs IgG2?. The C-terminus of C? ends in the Cys residue that bridges to the HC, but with C?, the disulfide bridging Cys residue is followed by a Ser residue, which is the C-terminal residue of that chain ( Liu et al., 2011a ). The presence of that Ser residue (or even if changed to an Ala residue) was found to be the cause for the increase in lability of the IgG1? HC–LC disulfide bridge. Deletion of that C-terminal residue increased the stability of the IgG1? HC–LC disulfide bridge ( Liu et al., 2011a ).
Because indexed inside the Chapter nine , human IgG
2 antibodies have the ability to form multiple structural isoforms due to alternative disulfide paring ( Wypych et al., 2008; ; Liu et al., 2008c; Allen et al., 2009; Zhang et al., 2010; Lightle et al., 2010 ) (see Figure 16.3 ). These isoforms are found not only in recombinant IgG2 antibodies, but also in IgG2s isolated from myeloma patients and normal human serum. Thus they are “natural” and an intrinsic property of the human IgG2 isotype, rather than being the product of recombinant DNA manipulations and/or cell culture production ( Wypych et al., 2008 ). The root cause of this alternative disulfide bridging appears to be due mostly to the presence of two extra disulfide bonds in the hinge 
of IgG2s ( ). The three major isomers have been labeled IgG2-A (i.e. “normal” IgG2 disulfide bridging), IgG2-A/B, and IgG2-B ( Wypych et al., 2008; Lightle et al., 2010 ; see Figure 16.3 ). Additional minor isoforms have also been described, including IgG2-B1 and IgG2B2 ( ), and, more recently IgG2-B3 ( Zhang et al., 2010 ). These isoforms of IgG2 are generated because the Cys residue in the LC can form a disulfide bond with any of the three different Cys residues in the HC, Cys127 (Cys131; EU numbering) of the CH1 domain or Cys232 (Cys219; EU numbering) or Cys233 (Cys220; EU numbering) in the hinge region ( Allen et al., 2009 ).
Multiple investigators have demonstrated that the major species of recombinant IgG2 is not the “normal” IgG2-A ( Wypych et al., 2008; c; Zhang et al., 2010 ). Liu et al. (2008b) demonstrated that B cells produce IgG2-A, which is converted relatively rapidly in serum to IgG2-A/B, followed by slower conversion to IgG2-B, prompting the authors to state that the form might be an indicator of the “age” of an IgG2 in serum. Liu et al. (2008c) also demonstrated that the clearance of the different forms of IgG2 were essentially identical. The significance of this is that the different isoforms do not have the same binding activity of functionality. Dillon et al. (2008) compared an IgG1 isotype with the wild-type IgG2 and its various isoforms of an anti-IL-1? MAb. They found that IgG1 had about a threefold greater activity than classical IgG2 (IgG2-A) and its various isoforms, and that the IgG2-B isoform had about a threefold lower affinity and a threefold higher ICfifty than IgG2-A, indicating that there are significant differences in the in vitro biological activity of the various isoforms ( Dillon et al., 2008 ). Allen et al. (2009) generated Cys > Ser mutations of several of the IgG2 Cys residues involved in multiple isoform generation, and showed that the mutant IgG2 molecules were both stable and biologically active.