Fluorescent proteins (FPs) emerged in the mid-90s as a powerful tool for life science research. Among its numerous applications, FPs is best known as reporter genes to study the localization and dynamics of proteins. In this dissertation, we explore new uses of FPs as genetically encoded fluorescent probes to image cellular chemical species.

In chapter 2, we incorporated of p-azidophenylalanine (pAzF) into circularly permuted GFP (cpGFP), which gave rise to the first genetically encoded fluorescent probe for H2S. Further engineering of the cpGFP scaffold, followed by directed evolution, has resulted in an improved fluorescent probe for H2S–hsGFP.

In chapter 3, we incorporated p-boronophenylalanine (pBoF) into different cpGFP templates, which serendipitously resulted in pnGFP, the first genetically encoded fluorescent probe for the selective detection of peroxynitrite (ONOO–).

In chapter 4, we examined the molecular mechanism underlying the unusual chemoselectivity of pnGFP towards ONOO– over H2O2 by using site-directed mutagenesis, X-ray crystallography, 11B NMR, and computational simulation studies. Surprisingly, the data collectively support an sp3-hybridized boron in pnGFP through a water bridge mediated N-B interaction between His 9 and the boronic acid moiety on the chromophore.

In chapter 5, we described the engineering and biological application of “G-CaMP”-type indicators for zinc ion (Zn2+). Insertion of a naturally evolved zinc-binding domain into the β-barrel of single green or red FP, followed by directed protein evolution and functional screening, we identified single-FP based green and red Zn2+ indicators, ZnGreen and ZnRed. These optically compatible indicators enabled dual color imaging of Zn2+ at single cell resolution. We further demonstrated the use of ZnGreen1 in imaging of glucose-stimulated Zn2+ secretion in pancreatic β cells.

Finally, in chapter 6, we laid out challenges in the field and provide future outlooks on possible new directions.

The objective of my Ph.D. study is to develop novel genetically encoded fluorescent biosensors to image and dissect biological signaling pathways in the context of live cells. I utilized protein engineering techniques to convert fluorescent proteins into fluorescent biosensors that can actively respond to specific, spatiotemporally organized cellular changes.

In this thesis, we expanded the fluorescent protein toolkit by engineering one of the first red fluorescent probes⎯rxRFP1⎯for sensing general redox states in the live cells. To further extend the usage of this sensor in various subcellular domains, such as mitochondria, endoplasmic reticulum, and the cell nucleus, we developed a group of rxRFP1 mutants showing different midpoint redox potentials for studying compartmentalized redox dynamics under various pathophysiological conditions. We also developed the first genetically encoded fluorescent biosensor for thioredoxin (Trx) redox by engineering a redox relay between the active-site cysteines of human Trx1 and rxRFP1. We utilized the resultant biosensor⎯TrxRFP1⎯to selectively monitor the perturbations of Trx redox in various mammalian cell lines. We further combined TrxRFP1 with a green fluorescent Grx1-roGFP2 biosensor to simultaneously monitor the dynamics of the two major cellular antioxidant systems, Trx and glutathione, in live cells in response to chemically and physiologically relevant stimuli.

We exploit another strategy which introduces reactive functional groups into circular permutated fluorescent proteins (cpFPs) using a genetic code expansion technology. Through a powerful directed protein evolution process, we were able to modulate the reactivity and chemoselectivity of an introduced p-boronophenylalanine (pBoF) in a cpRFP scaffold, resulting in fluorescent probes selectively responsive to hydrogen peroxide (H2O2) and peroxynitrite (ONOO—). Furthermore, by using boronic acid and short peptides as synergistic recognition motifs, we were able to engineer a series of reversible probes for nucleotides and carbohydrates showing surprisingly high specificity and large dynamic ranges. We have successfully utilized this new family of fluorescent probes to visualize various cellular activities.

The objective of my Ph.D. study is to expand the unnatural amino acid (unAA) toolbox to genetically encode additional photocaging functional groups to achieve a precise control of proteins with light, to site-specifically label proteins with hapten moieties, and to further explore computational methods with an ultimate goal of using computers to design specific orthogonal aminoacyl-tRNA synthetases (aaRSes) for given unAAs.

In this thesis, we show that cellular biochemical processes can be spatiotemporally manipulated by light-activatable protein-splicing inteins. We genetically encoded a photocaged cysteine and introduced the photocaged cysteine into a highly efficient Nostoc punctiforme (Npu) DnaE intein, which is capable of excising itself and subsequently splicing adjacent N- and C-terminal extein flanks to form a new truncated peptide. The resulting photocaged intein was inserted into a red fluorescent protein (RFP) mCherry and a human Src tyrosine kinase, and a light-induced photochemical reaction was able to reactivate the intein and trigger protein splicing. The genetically encoded photocaged intein is a general optogenetic tool, allowing effective photocontrol of primary structures and functions of proteins.

Haptens, such as dinitrophenyl (DNP), are small molecules that induce strong immune responses when attached to proteins or peptides and, as such, have been exploited for diverse applications. In this thesis, we engineered a Methanosarcina barkeri pyrrolysyl-tRNA synthetase (mbPylRS) to genetically encode a DNP-containing unAA, N6-(2-(2,4-dinitrophenyl)acetyl)lysine (DnpK). This technique is a promising strategy for biological preparation of proteins containing site-specific DNP. This new capability is expected to find broad applications in biosensing, immunology, and therapeutics.

The experimental procedure to derive orthogonal aaRSes/aminoacyl tRNAs, which typically involves several rounds of positive and negative selection, is laborious and time-consuming, and requires considerable expertise. It is often not trivial to derive orthogonal aaRSes for unAA substrates that are very different from the enzymes’ native substrates. In this thesis, we compared several computational algorithms to evaluate the binding energies of unAA and previously developed orthogonal aaRSes. We hope to use these results to guide future designing and development of new aaRSes, and to extend the capability of the genetic code expansion technology to many new unAAs.

Phage display is a powerful screening method capable of potentially screening over 1011 unique peptide sequences for protein-protein or protein-DNA interactions. It has been extensively utilized in manufacturing, basic research, and therapeutics. With the recent creation of a genomically recoded organism, wherein all amber codons were replaced with ochre codons in the C321ΔA E. coli strain, this allows for an expansion the genetic code by reassigning the amber codon to designate a genetic incorporation of synthetic, non-canonical amino acid (ncAA) and opens an exciting avenue for further developing the phage display technology. My aims are to therefore: (1) develop and optimize a phage display system that can efficiently genetically encode multiple, different ncAAs, and subsequently demonstrate this application in a screen for (2) peptide inhibitors to Protein Tyrosine Phosphatase non-receptor type 1 (PTP1B), as well as (3) Src Homology 2 (SH2) domain. PTP1B is implicated in breast cancer, diabetes mellitus type 2, and obesity; and several mutant SH2 domains mediating aberrant protein-protein interactions have been implicated in human diseases, such as congenital heart disease, X-linked agammaglobulinemia and lymphoproliferative syndrome. Developing this technology will enable phage display to become an even more powerful screening method with its new, expanded chemistry set conferred from synthetic ncAAs.

Biosensors based on the principle of Förster resonance energy transfer (FRET) have proven to be powerful tools for biological research. Aided by the plethora of color variants of green fluorescent proteins, numerous FRET-based probes have been developed for study of molecular interactions, enzyme activities, and small molecules in live mammalian cells with high spatial and temporal resolution. In this dissertation, we focus on the development of genetically encoded FRET-based probes for cellular processes related to hydrogen sulfide and hypoxic signaling.

Hydrogen sulfide (H2S) has been recently identified as an important gasotransmitter playing crucial roles in cell signaling. We developed the first, genetically encoded, FRET-based biosensor, hsCY, for imaging live-cell H2S. We utilized a blue- (EBFP) and a green- (cpGFP) fluorescent proteins, and genetically modified the probe with p-azidophenylalanine as a H2S-sensory element. We validated hsCY in vitro and mammalian cells and demonstrated the use of hsCY for selective, ratiometric imaging of H2S. hsCY thus represents a valuable addition to the toolbox for H2S detection and imaging.

Oxygen is vital for all aerobic life forms. Oxygen-dependent hydroxylation of hypoxia-inducible factor (HIF)-1α by prolyl hydroxylase domain enzymes (PHDs) is an important step for controlling the expression of oxygen-regulated genes in metazoan species, thereby constituting a molecular mechanism for oxygen sensing and response. Herein, we report a genetically encoded dual-emission ratiometric fluorescent sensor, ProCY, which responds to PHD activities in vitro and in live cells. We demonstrated that ProCY could monitor hypoxia in mammalian cells. By targeting this novel genetically encoded biosensor to the cell nucleus and cytosol, we determined that, under normoxic conditions, the HIF-prolyl hydroxylase activity was mainly confined to the cytosol of HEK 293T cells. The results collectively suggest broad applications of ProCY on evaluation of hypoxia and PHD activities and understanding of pathways for the control of hypoxic responses.

As HIF-prolyl hydroxylases have emerged as promising drug targets for a variety of diseases, such as myocardial infarction, stroke, cancer, diabetes, and severe anemia, we also explored the use of ProCY in high-throughput assays to identify inhibitors or activators of PHDs. Our preliminary study supports the feasibility of this approach. A lentiviral system has been developed for transduction of HEK 293T and establishment of ProCY-expressing stable cells.