Written by Nick Gee - Conjugation Expert
Part 2 of 2: Random conjugation of labels to antibody lysines—is the concern over the impact of CDR modification on antigen binding overstated?
In part 1, I discussed the distribution of labels after random conjugation of a hypothetical antibody containing one hundred equally reactive lysines. Using probability theory, I showed that a high proportion of conjugate molecules possess at least one unmodified antigen-binding site.
In part 2, I review a paper by Hill et al. who determined the true susceptibility of lysine residues to modification using a clever tandem mass tagging (TMT) strategy. TMTs have identical masses but give different unique fragmentation patterns in MS analysis.
The authors conjugated an NHS ester of TMT-126 to the NIST IgG1(k) reference monoclonal antibody, resulting in degrees of incorporation of 3, 5, or 11. The conjugates were then fully denatured and heavily modified at all remaining positions using TMT-127. After proteolytic fragmentation and LC-MS/MS analysis of peptides, the susceptibility of each lysine to NHS ester modification was given by the proportion of TMT-126 at each site.
NIST mAb contains 49 lysines distributed across its H and L chains and just one N-terminal amine (the H chain is blocked with pGlu), providing one hundred conjugatable positions in an H2L2 structure. NIST mAb has two CDR lysines based on the Kabat numbering.
Interestingly, only 4 positions on the H and L chains were resistant to NHS ester conjugation, and most sites responded in a linear manner to an increase in the concentration of the TMT-126 label. K58/K59 in the H chain, K187 in the L chain, and the N-terminus of the L chain were very amenable to conjugation. A lower pKa of the N-terminus and hence higher reactivity is not unexpected.
While lysine side chains typically have relatively high pKa values, protonation and deprotonation of the side chains is a dynamic process, and evidently most lysines spend sufficient time in a reactive deprotonated state to incorporate some of the label.
While the pattern is not truly random, probably due to variation in pKa or steric factors, the Hill paper shows that most of the lysines in an antibody are conjugatable with NHS esters.
To conclude, random lysine conjugation creates millions of positional isomers. The competition by multiple lysines distributed over the antibody molecule means that the CDRs, even if they contain lysines, are likely to be afforded a measure of protection in NHS ester conjugation reactions.