A stable fluorescent protein scaffold that can bind to any antigen of interest can be utilized as a biosensor. Such a protein-based biosensor has considerable application in cell biology, proteomics, and immunochemistry. The creation of a fluorescent protein scaffold capable of binding has been attempted by several laboratories but with limited success. In our approach, we selected green fluorescent protein (GFP) as the scaffold on which we proposed to graft loops to confer binding. To study the ability of GFP to accommodate binding loops, we selected CDRL3 and CDRH3 from the D1.3 hen egg lysozyme monoclonal antibody and inserted them between amino acids D102-D103 and E172-D173 of GFP respectively. The resulting protein was not expressed by yeast nor was it fluorescent likely the result of protein misfolding and chromophore instability. We therefore employed directed evolution using yeast surface display to engineer the GFP scaffold to be well expressed and fluorescent. To bias directed evolution towards scaffolds that fold correctly and have stable chromophores, selection pressures of elevated expression temperature (37°C) and thermal stability (upto 70°C) were employed. Among the scaffolds isolated using the above criteria, the one with the highest expression and chromophore stability has five mutations, four of which are proximal to the chromophore region. To determine if the scaffold can be expressed and remains fluorescent when endowed with other binding loops, we have randomized the inserted amino acids in both regions to create a library of GFP scaffolds offering binding loops. This library can be screened against antigens of interest to obtain fluorescent protein biosensors.