Ionic liquids (ILs) are attractive (co-)solvents for biocatalysis. However, in high concentration (>10 % IL), enzymes usually show decreased activity. No general principles have been discovered to improve IL resistance of enzymes by protein engineering. We present a systematic study to elucidate general engineering principles by site saturation mutagenesis on the complete gene bsla. Screening in presence of four [BMIM]-based ILs revealed two unexpected lessons on directed evolution: 1) resistance improvement was obtainable at 50-69 % of all amino acid positions, thus explaining the success of small sized random mutant libraries; 2) 6-13 % of substitutions led to improved resistance. Among these, 66-95 % were substitutions by chemically different amino acids (e.g., aromatic to polar/aliphatic/charged amino acids), thus indicating that mutagenesis methods introducing such changes should, at least for lipases like BSLA, be favored to improve IL resistance.