Protein fission, fusion, and circular permutation have been used to convert green fluorescent protein (GFP) family members into biosensors that dynamically report on cellular processes, ranging from protein expression and metabolite concentrations to protein solubility, protein-protein interactions, and ligand-binding. Unfortunately, GFP are unsuitable for deep tissue reporting in animal models because the wavelengths of light used with these reporters is highly absorbed by tissues. In contrast, near infrared fluorescent protein (IFP and iRFP) reporters derived from bacterial phytochrome proteins (BphP) are excited by light in the near-infrared spectrum (~700 nm, less absorptive) and are better suited for probing cellular processes within tissues. IFP and iRFP can report on biological processes under anaerobic conditions because it uses biliverdin (BV) as a chromophore and does not require oxygen for maturation, a requisite for GFP maturation. Unlike GFP, IFP and iRFP are not yet able to report on wide-range of biological processes beyond gene expression.
To report on gene expression, BphP must interlace its Per/ARNT/Sim (PAS) and cGMP phosphodiesterase/adenylcyclase/FhlA (GAF) domains into a topological knot. The extent to which this complex topology tolerates mutations (fission, fusion, and circular permutation) used to convert proteins into biosensors is not known. To better understand the tolerance of BphP to these types of mutational lesions, I have subjected IFP to random backbone fragmentation and iRFP to circular permutation using transposase mutagenesis. Screening a library of split IFP for fluorescent variants yielded thirteen unique fragmented IFP and with parent like spectral properties. These two-fragment IFP all required assistance from associating proteins for maximal fluorescence. These split sites displayed AND gate logic behavior when the ORFs encoding the different fragment are placed under distinct transcriptional regulation. In addition, screening a library of circularly permuted iRFP led to the discovery of twenty seven permuted iRFP variants with near infrared fluorescence. These variants arose from backbone fission in both the PAS and GAF domains, although the brightest permuted iRFP initiated at residues near the domain linker and termini. Biochemical analysis revealed that permuted iRFP display similar oligomerizatoin, quantum yield, and stability as native iRFP. These proteins also retained sufficient BV affinity serve as reporters of gene expression in mammalian cells without the addition of exogenous BV.
The results described in this thesis represent the first study to map the tolerance of a BphP to random fragmentation and circular permutation. These results demonstrate that knotted BphP retain the ability to fold as their contact order changes, suggesting that these proteins can be further developed as reporters of biological processes like GFP. The split IFP represent a suite of assays that will be useful for monitoring the dynamics of a protein-protein interactions under conditions where split GFP do not yield strong signals. These split IFP can also be used to report on protein-ligand interactions that regulate protein oligomerization. The permuted iRFP should be useful for building molecular switches through domain insertion, in which the set of permuted iRFP is randomly inserted into other protein domains to couple the ligand binding to iRFP fluorescence.