Page 103 - Tyrosine-Based Bioconjugations - Jorick Bruins

Engineering of Tyrosine-Containing Peptide Loops Enables Non-terminal Protein Labeling
6.3. Loop antibodies
As a second option for non-terminal labeling of antibodies, we envisioned the incorporation of small tyrosine-containing peptide sequences into solvent-exposed loops in an antibody constant domains. These ‘loop antibodies’ would ideally display better pharmacokinetics than terminally labeled antibodies, in particular in combination with hydrophobic payloads, based on a potential partial covering of the payload by the antibody protein fragments. Using the two-step conjugation method described in chapter 5, these loop antibodies could be used to generate new bispecific antibody structures, which previously required the introduction of entire proteins in antibody loops.18
Our design of an antibody with an exposed tyrosine-bearing loop was based upon work performed by Grünewald et al.,19 on a total of 183 trastuzumab mutants containing either of two peptide insertions in a total of 110 distinct positions on the antibody. These peptide insertions (S6 = GDSLSWLLRLLN; YbbR = DSLEFIASKLA) contain a serine residue available for site-specific labeling with enzymes Sfp and AcpS phosphopantetheinyl transferases (PPTases).20 By using these tags and fusing them inside the solvent-exposed loops in trastuzumab, over 60 different ADCs with a drug-to-antibody ratio (DAR) >1.8 were produced and evaluated.21 It is noteworthy that the S6-tag has two serine residues, but only a single modification was detected. After mutating the serine residue at position 3 to an alanine (S6-S3A) the conjugate formation was completely abolished, confirming the specificity of the PPTase for that particular residue.19
Based on the finding by Grünewald et. al., a tyrosine-based insertion was envisioned. We selected insertion of a tyrosine loop after residue HC-P294,i based on the excellent DAR reported for conjugation of the drug monomethylauristatin F on tags following that position, as well as the high area-under-the-curve (AUC) with regards to pharmacokinetics, indicating long-term stability and low clearance of the conjugates. Thus, two distinct tyrosine-containing peptide loops were inserted after P294, both of which were based on the earlier reported S6 tag.20 The first tag, Y1, had the S3Y mutation (Figure 2), with the argument that the tyrosine would be located on the same site where conjugation with PPTase occurs.19 The second tag, Y2, was designed as a W6Y mutation (Figure 2), selected based to the comparable hydrophobicity and size for tyrosine and tryptophan. More importantly, a phage-displayed peptide library optimizing the tag for PPTases indicated that tyrosine was a common mutation of the W6.20 Tras-Y1 and Tras-Y2 antibodies were transiently expressed in CHO-K1 and purified, after which SPOCQ was attempted on both antibodies.
i Grünewald et. al. defined the residue as P291 by virtue of Eu numbering, however the correct numbering for trastuzumab is P294 (drugbank accession number DB00072).
101
6