Single scaffold antibody libraries created with high rates of mutagenesis or diversity focused for peptide recognition

This dissertation describes several strategies used to create diversity in non-immune antibody libraries. Two of the strategies were used to create two separate peptide focused libraries. Both of these strategies used to create these antigen-class focused libraries used a single scaffold antibody gene that contained diversity only in the variable heavy region. The scaffold antibody gene one of the libraries, the M:anti-pep library, was chosen based on hypervariable loop canonical structures that are characteristic of other anti-peptide antibodies. Additionally, all of the contact residues of this antibody are commonly used contact residues in other anti-peptide antibodies. These positions and others were varied to incorporate the natural diversity of other anti-peptide antibodies. The second library, the Hu:anti-pep, is based on a widely used, unique combination of human germline antibody segments that express well in bacterial expression. Positions were chosen for variation based on their usage as contact residues in both anti-peptide and anti-protein antibodies. The diversity was less focused than with the M:anti-pep library, incorporating all 20 amino acids at "high usage" positions and only four amino acids at "low usage" positions. Both libraries were validated by phage display selections against the peptide angiotensin (AT) and neuropeptide Y (NPY). The M:anti-pep library yielded specific antibodies to both peptides with dissociation constants as low as 14 nM against AT and 18 nM against NPY. The Hu:anti-pep library yielded specific clones with higher dissociation constants: 49 nM against NPY and 11 [mu]M against AT. The final strategy used to introduce diversity is widely used for affinity maturation of low affinity, previously selected antibodies. Extremely high rates of mutagenesis (2.2% of the gene to 2.7%) were used to create two libraries of the anti-digoxin antibody 26-10. The libraries had been screened by others in an attempt to examine the effects of highrates of mutagenesis on the directed evolution of an antibody. A total of 91 isolated clones from both libraries were sequenced. Several consensus mutations were identified near the CDRH3 in the isolated clones, indicating that they had a positive, selectable effect. This study confirmed that high-error rate antibody libraries contain more active clones than expected. Combinations of the selected consensus mutations from these libraries provide moderate enhancements to the kinetics and expression of the wild-type antibody in a non-synergistic manner.