Structural Chemistry of a Green Fluorescent Protein Zn Biosensor David P. Barondeau, Carey J. Kassmann, John A. Tainer, and Elizabeth D. Getzoff* Department of Molecular Biology, The Skaggs Institute for Chemical Biology, The Scripps Research Institute, 10550 North Torrey Pines Rd., La Jolla, California 92037 Received December 6, 2001 Zinc metalloproteins influence DNA synthesis, microtubule polymerization, gene expression, apoptosis, and immune system function. 1 Furthermore, Zn(II) homeostasis is critical for complex neurobiological systems: Zn(II) acts in synaptic transmission, as a contributory factor in neurological disorders including epilepsy and Alzheimer’s disease, and as a neurotoxin under seizure or patho- logical conditions. 1 To fully understand Zn(II) functions in cell and neurobiology, probes are needed that are capable of monitoring in vivo spatial and temporal distributions of metal ions. Designed small molecule Zn(II) fluorescent probes have many advantages. 2 How- ever, protein-based biosensors 3 might complement these probes as proteins can be optimized for selectivity and affinity, added noninvasively to cells by transfection, and targeted to specific tissues, organelles, or cellular locations. Recently, proteins have been engineered to incorporate desired properties via structure- based 4 and directed evolution 5 strategies. Since protein metal site specificity and activity are exquisitely sensitive to precise ligand chemistry and geometry, structure-based design of metal ion biosensors would benefit from very high resolution structures of apo and distinct metal-bound protein states. We report here (i) the design and characterization of a green fluorescent protein (GFP) mutant that binds Zn(II) (enhancing fluorescence intensity) and Cu- (II) (quenching fluorescence) to a tridentate chromophore resem- bling half a porphyrin and (ii) very high-resolution crystallographic structures for these apo, Zn(II)-bound, and Cu(II)-bound proteins (Figure 1). These structures, reported with experimentally deter- mined standard uncertainty errors, define accurate metal-site geometric parameters for a novel GFP metal-binding site and provide critical structural chemistry for understanding the metal ion affinity, specificity, and activity of protein metal sites. Moreover, these results prompt structure-based hypotheses to explain the protein’s metal ion selectivity, metal binding kinetics, and fluo- rescence enhancement upon Zn(II) binding, with potential general implications for the design of protein metal ion biosensors. The GFP fluorophore is self-synthesized within the center of an antiparallel -barrel from protein residues Ser65, Tyr66, and Gly67. 6,7 The chromophore has a high fluorescence quantum yield (0.79) and an emission maximum at 508 nm. 6 Mutated proteins, such as Y66H Y145F, can still generate a chromophore, but with modified emission wavelengths (447 nm) and quantum yields (0.38). 8 The self-generated fluorophore makes GFP an ideal system for screening directed libraries of random mutants for desirable characteristics such as increased soluble expression or altered spectral properties. 6 Thus, the engineering of a specific metal binding site in GFP that couples metal binding to modified fluorescent signals provides a prototype suitable for optimization of kinetic and thermodynamic properties by directed evolution strategies. To create this prototype, we combined porphyrin-like metal ligand, protein solubility, and accessibility design criteria with high-resolution structure determination and analysis. To link metal ion binding to chromophore spectroscopic proper- ties, our first structure-based metal site design (BFPms1) includes the BFP chromophore (Y66H), which resembles a dipyrrole unit of porphyrin. 9a In addition, mutations were included for higher quantum yield (Y145F), 8 increased solubility (F64L, F99S, M153T, and V163A), 6 faster chromophore formation (S65T), 6 and creating a hole into the -barrel (H148G) 9b to improve metal binding kinetics. The metal site appears specific for Cu (24 μM K D ) and Zn (50 μM K D ), as no other divalent first row transition metal binds at metal concentrations up to 2 mM. Moreover, Zn(II) and Cu(II) each alter chromophore spectroscopic properties in distinct ways. Zn(II) binding increases fluorescence intensity 2-fold, but does not alter the absorbance spectra. In contrast, Cu(II) binding quenches * Address correspondence to this author. E-mail: edg@scripps.edu. Figure 1. Stereopairs of BFPms1 metal binding sites and 2Fo-Fc electron density maps contoured at 1 σ for (A) Zn-bound (1.44 Å), (B) Cu-bound (1.50 Å), and (C) apo (1.35 Å) structures. Images made with AVS. Published on Web 03/15/2002 3522 9 J. AM. CHEM. SOC. 2002, 124, 3522-3524 10.1021/ja0176954 CCC: $22.00 © 2002 American Chemical Society