Improvement of the stability and yield of scFv immunotoxins by molecular modeling and protein engineering

E. S. Chowdhury, G. Vasmatzis, R. Fisas, B. Lsê, I. Easlaa

Research output: Contribution to journalArticlepeer-review


The conversion of monoclonal antibodies to scFvs results in problems with folding and yield. In many cases the usefulness of recombinant antibody fragments is severely limited. We describe a method that identifies problematic residues in Fv fragments of antibodies. By mutating these residues, the Fv portion is stabilized and the yield of Fv is increased. The method is confined to the framework regions since these regions are mainly responsible for the stability of Fvs. Initially we assign a character to each residue which indicates its state of exposure based on the known crystal structures of Fabs. This identifies residues not compatible with their environment in the folded state of the protein. Next we calculate the frequencies of different amino acids for each position of the Fvs based on the available database. This identifies residues which are not commonly present in the conserved positions. If these residues are compatible with their exposure profile they are left unaltered, otherwise, they are mutated to the most commonly occurring residue. We have identified four amino acids in the Fv region of anti-mesothelin MAb Kl. KlscFv when fused to a mutant of Pseudomonas exotoxin (KlscFv-PE38) to make an immunotoxin gives a very poor yield and is quite unstable. By mutating the four problematic residues into the most frequently observed residues, the yield and stability of KlscFv-PE38 increased by more than ten-fold. This approach could be a general method for improving other scFvs.

Original languageEnglish (US)
Pages (from-to)A1426
JournalFASEB Journal
Issue number8
StatePublished - Dec 1 1998

ASJC Scopus subject areas

  • Biotechnology
  • Biochemistry
  • Molecular Biology
  • Genetics


Dive into the research topics of 'Improvement of the stability and yield of scFv immunotoxins by molecular modeling and protein engineering'. Together they form a unique fingerprint.

Cite this