A process for producing a modified single chain Fv (scFv) antibody by improving the stability of an initial scFv comprising an initial heavy chain variable domain (V.sub.H) linked to an initial light chain variable domain (V.sub.L) by an intermolecular cysteine bridge or an oligopeptide linker, said process comprising,
- (a) providing an initial gene encoding an amino acid sequence of an initial variable domain to be modified,
- (b) comparing the amino acid sequence encoded by the initial gene with one of consensus tables 1-6, wherein
- (i) where the initial gene encodes the variable domain of a heavy chain of a human antibody, the amino acid sequence is compared with consensus table 1, SEQ ID NOS:
- 1-4,
- (ii) where the initial gene encodes the variable domain of a heavy chain of a mouse antibody, the amino acid sequence is compared with consensus table 2, SEQ ID NOS:
- 5-8,
- (iii) where the initial gene encodes the variable domain of a kappa light chain of a human antibody, the amino acid sequence is compared with consensus table 3, SEQ ID NOS:
- 9-12,
- (iv) where the initial gene encodes the variable domain of a kappa light chain of a mouse antibody, the amino acid sequence is compared with consensus table 4, SEQ ID NOS:
- 13-16,
- (v) where the initial gene encodes the variable domain of a lambda light chain of a human antibody, the amino acid sequence is compared with consensus table 5, SEQ ID NOS:
- 17-20,
- (vi) where the initial gene encodes the variable domain of a lambda light chain of a mouse antibody, the amino acid sequence is compared with consensus table 6, SEQ ID NOS:
- 21-24;
- (c) thereafter producing a modified gene which encodes a modified variable domain by modifying
- (1) at least one codon of each pair of codons in the initial gene, which pair together code for disulfide bridge-forming cysteines, so that all disulfide bridges present in the initial variable domain produced in a eukaryotic cell are absent from the modified variable domain as produced in said method, and
- (2) at least one additional codon in the initial gene which codes for an amino acid other than a disulfide bridge-forming cysteine, using the consensus table selected in step (b) as a guide, wherein each codon to be modified codes for an initial amino acid at a selected position, and
- (i) the codon coding for the initial amino acid is modified by substituting a codon coding for a substitute amino acid which is listed in the consensus table at a position corresponding to the selected position, when the initial amino acid is unlisted in the selected consensus table at the position corresponding to the selected position, or
- (ii) the codon coding for the initial amino acid is modified by substituting a codon coding for a substitute amino acid having higher frequency in the consensus table at a position corresponding to the selected position, when the initial amino acid is listed at a given frequency in the consensus table at the position corresponding to the selected table, the higher frequency being compared to the given frequency;
- (d) transforming a prokaryotic microorganism with a modified scFv gene comprising the modified gene and additional polynucleotide sequences to produce a transformed prokaryotic microorganism, wherein the modified gene and the additional polynucleotide sequences encode the heavy and light chain variable domains of the modified antibody, and wherein one of the heavy or light chain variable domains has been modified, wherein the modified gene encodes a modified variable domain;
- and
- (e) expressing in the transformed prokaryotic microorganism the modified scFv gene, wherein the modified variable domain expressed by the prokaryotic microorganism has improved stability when compared to the stability of the initial variable domain expressed by a eukaryotic microorganism.