Optimizing the generation of recombinant single-chain antibodies against placental alkaline phosphatase

Recombinant technologies to engineer ordinary hybridoma monoclonal antibodies (MAbs) to single-chain fragment variable (scFv) may cause loss of antibody affinity, increased tendency to aggregate, increased temperature sensitivity, and low yield of active protein. In the present investigation, the well-characterized MAb H7 against placental alkaline phosphatase (PLAP), used as a model antibody, was engineered to improve solubility and stability of scFv with retained high affinity. The original procedure to generate single-chain antibodies with a 10-amino acid linker between VH and VL yielded an almost insoluble product. By site-directed mutagenesis, four selective sequence substitutions were made in the VL fragment and one in the VH fragment to improve solubility. The importance of the linker length was investigated, and a 25/30 amino acid linker was found to improve solubility. In order to further increase the stability of the single-chain antibody, an additional covalent-S-S-bond was introduced between amino acid 100 in the VL fragment and amino acid 44 in the VH region, to make a single-chain disulphide stabilized variable fragment (scdsFv). Altogether five different antibody constructs were produced and compared in terms of solubility, stability, affinity, and production properties. Immunospecificity was tested by enzyme-linked immunosorbent assay (ELISA) against the target antigen, temperature sensitivity by exposing the purified scFv to higher temperatures. All the new constructs retained almost equal activity and high affinity for their target antigen, placental alkaline phosphatase (PLAP), compared to the intact MAb H7, up to +42 degrees C as evaluated by ELISA. The overall affinity K-A > 10(9) (M-1) of the new antibodies could be maintained in the same order of magnitude as the original one (H7), when evaluated by Biacore technology. The best final single-chain antibody was obtained by performing the specific site-directed mutations and introducing a linker of 30 amino acids, but not by additional stabilizing disulphide bonds. The yield of the final antibody was improved approximately 10-fold by the modifications. This antibody could easily be expressed in a bacterial system using the PET-32a TrxA vector and the Escherichia coli strain BL21 Origami B (DE3). Purified antibody, which could be kept at concentrations up to 0.8 mg/mL, was obtained, which is sufficient for clinical testing of therapeutic applications.