Research

Cyanotryptophans (CN-Trp) are privileged multimodal reporters on protein structure. They are similar in size to the canonical amino acid tryptophan and some of them exhibit bright fluorescence which responds sensitively to changes in the environment. We selected aminoacyl-tRNA synthetases specific for 4-, 5-, 6-, and 7-CN-Trp for high-yield in vivo production of proteins with a single, site-specifically introduced nitrile label. The absorption maximum of 4-CN-Trp is distinct from Trp, allowing the selective excitation of its intense fluorescence. 4-CN-Trp fluoresces in the visible range with an intensity rivalling that of 7-hydroxy-coumarin. Crystal structures of maltose binding protein demonstrate near-complete structural conservation when a native buried Trp residue is replaced by 4-CN-Trp. Besides presenting an inconspicuous tag for live cell microscopy, the intense fluorescence of 4-CN-Trp enables measurements of subnanomolar ligand binding affinities in isotropic solution, as demonstrated by the complex between rapamycin and the peptidyl–prolyl isomerase FKBP12 furnished with a 4-CN-Trp residue in the substrate binding pocket. Furthermore, 4-CN-Trp residues positioned at different locations of a protein containing multiple tryptophan residues permits using fluorescence quenching experiments to detect the proximity of individual Trp residues to the binding site of aromatic ligands.

It has recently been shown that distances > 20 Å can be measured between Gd3+ spin tags and 19F labels in proteins using 19F-ENDOR at 94 GHz. Here we examine the precision with which this method can locate the positions of (aromatic) amino acid side chains and, with the help of established simulation tools, determine the conformational space sampled by solvent-exposed chains in solution. First, using a novel set of fluorinated phenylalanine amino acids incorporated into the metalloprotein Calbindin D9k binding Gd3+, we show that triangulation of the corresponding 19F-ENDOR determined distances can precisely identify the conformation of buried side chains. The obtained conformation agrees with the same (single) orientation seen in the crystal structure of the native protein. In a second set of proteins labeled with a Gd3+ spin tag and noncanonical 19F-labeled amino acids, the splittings, lineshapes, and integrated intensities of the 19F-ENDOR signals were used to constrain the conformational space of the side chains initially identified by comprehensive rotamer simulations. This work provides the first example where 19F-ENDOR constraints have been used to accurately pinpoint side chain conformations in a protein with the help of differently fluorinated amino acids.

Fluorine atoms are known to display scalar 19F–19F couplings in nuclear magnetic resonance (NMR) spectra when they are sufficiently close in space for nonbonding orbitals to overlap. We show that fluorinated noncanonical amino acids positioned in the hydrophobic core or on the surface of a protein can be linked by scalar through-space 19F–19F (TSJFF) couplings even if the 19F spins are in the time average separated by more than the van der Waals distance. Using two different aromatic amino acids featuring CF3 groups, O-trifluoromethyl-tyrosine and 4-trifluoromethyl-phenylalanine, we show that 19F–19F TOCSY experiments are sufficiently sensitive to detect TSJFF couplings between 2.5 and 5 Hz in the 19 kDa protein PpiB measured on a two-channel 400 MHz NMR spectrometer with a regular room temperature probe. A quantitative J evolution experiment enables the measurement of TSJFF coupling constants that are up to five times smaller than the 19F NMR line width. In addition, a new aminoacyl-tRNA synthetase was identified for genetic encoding of N6-(trifluoroacetyl)-l-lysine (TFA-Lys) and 19F–19F TOCSY peaks were observed between two TFA-Lys residues incorporated into the proteins AncCDT-1 and mRFP despite high solvent exposure and flexibility of the TFA-Lys side chains. With the ready availability of systems for site-specific incorporation of fluorinated amino acids into proteins by genetic encoding, 19F–19F interactions offer a straightforward way to probe the spatial proximity of selected sites without any assignments of 1H NMR resonances.