Computational design of a single-chain galectin-1 yields a stable variant with retained glycan binding activity

Developing galectin-1-based therapeutics faces challenges that are specific for this lectin, such as its differing redox states and reversible monomer-dimer equilibrium. The generation of a cysteine-less mutant of galectin-1 (CSGal-1) stabilized its redox behavior, while the development of a covalently linked galectin-1 dimer avoided its concentration-dependent monomer-dimer equilibrium. However, the type and length of the linker can affect the therapeutic potency of the covalent construct, as well as the protein titer, solubility, domain folding or bioactivity. Therefore, the computational design of a linker with desired properties and predictable behavior is highly desired. In this study, we de novo designed three different linkers to covalently connect the two subunits of CSGal-1. The CSGal-1 subunits were genetically fused via the designed linkers and produced as single-chain (sc) proteins for the first time. The data showed that the sc construct with the longest and most acidic linker exhibited the highest expression titer, solubility, and even ability to refold from inclusion bodies, which was in accordance with the in silico prediction. Further, we characterized the best performing scCSGal-1 construct against the noncovalent dimer of CSGal-1. Data from differential scanning fluorimetry and isothermal titration calorimetry demonstrated that scCSGal-1 retained its functionality with a binding affinity and thermodynamic profile similar to that of noncovalently dimerized CSGal-1. These findings underscore the strength of the computational approach in linker design and encourage its extension to other lectins that require a stable dimeric form in their applications.