Designing the Selectivity of the Fluorescent Detection of Amino Acids: A Chemosensing Ensemble for Histidine Marta Ansa Hortala ´ , Luigi Fabbrizzi,* Nathalie Marcotte, Floriana Stomeo, and Angelo Taglietti UniVersita ` di PaVia, Dipartimento di Chimica Generale, Via Taramelli 12, I-27100 PaVia, Italy Received May 30, 2002 ; E-mail: luigi.fabbrizzi@unipv.it Fluorescent sensing of amino acids is a very important task in biochemistry and molecular biology, with a special regard to determinations which require both temporal and spatial resolution. 1 However, only a few fluorescent chemosensors for amino acids have been described thus far. They include a heteroditopic system, containing an NO 5 crown fragment and a guanidium subunit, suitable for linear recognition of analytes of formula NH 3 + -(CH 2 ) n -COO - , which displays selective behavior for n ) 3, 2 and signals the occurrence of the recognition through a useful off/on switching of fluorescence, and a dizinc(II) cryptate, which recognizes histidine through the formation of an imidazolate bridge between the two Zn II centers, 3 an event indicated by a less valuable on/off fluorescent response. In both cases, the fluorophore (a 1,9-anthracenyl fragment) acts as a spacer separating the two recognizing functionalities. We describe now a novel type of off/on fluorescent chemosensor for amino acids, suitable for selective detection of histidine, which does not require the establishing of any covalent linking between the fluorophore and the receptor but utilizes the fluorophore and the receptor as such. According to this approach, called the “chemosensing ensemble”, 4 the fluorescent indicator is bound through noncovalent interactions to the receptor, which quenches its emission; then, the added analyte displaces the indicator, which when released to the solution displays its full fluorescence. Thus, analyte recognition is signaled by the sharp appearance of the fluorescence of the indicator. Moreover, we intend to demonstrate how decisive is the choice of the indicator to achieve selectivity in recognition and sensing. Selective binding of histidine requires a receptor capable of interacting with the imidazole residue rather than with the car- boxylate group, which is common to all the amino acids and which cannot, therefore, induce any selectivity. The extremely weak acid imidazole (pK A ) 14.5), in the presence of two Cu II ions, prepositioned at the right distance within an appropriate ligand, deprotonates and bridges the two metal centers. 5 As an example, an imidazolate moiety bridges the two Cu II centers of the dimetallic complex of the bisdien macrocycle 1. In particular, each Cu II center in the ternary complex 2 becomes four-coordinate, according to a square geometry. 6 Thus, we decided to use the [Cu II 2 (1)] 4+ complex as a receptor for the recognition of histidine. Then, we chose a set of fluorescent indicators including: coumarine 343 (3), fluorescein (4), and eosine Y (5), which share high quantum yield, excitation, and emission wavelengths in the visible region, and the presence of a carboxylate group in the molecular structure. Titration of each indicator with the [Cu II 2 (1)] 4+ receptor complex at pH ) 7 resulted in a complete quenching of the emission, while nonlinear least-squares fitting of the titration profiles (fluorescence intensity, I F , vs equiv of [Cu II 2 (1)] 4+ ) 7 indicated formation of 1:1 adducts, whose association constants are, in log units: coumarine 4.5, fluorescein 5.9, eosine Y 7.2. Formation of a 1:2 adduct had to be ruled out, due to the very poor fitting of titration data. It is suggested that, in the receptor/ indicator 1:1 adduct, the two oxygen atoms of the carboxylate group of each dye bridge the two Cu II ions, which quench the proximate fluorophore through either an electron- or energy-transfer process. The capability of the -COO- group to encompass the two metal centers of the macrocyclic complexes has been confirmed through molecular modeling. Thus, the nonfluorescent chemosensing en- semble was generated by dissolving the [Cu II 2 (1)] 4+ receptor and the indicator in a solution buffered at pH ) 7 (HEPES 0.05 M). The concentration of the indicator was 10 -6 M, and that of the receptor was high enough to ensure quenching of the fluorescent probe. At this stage, each receptor/indicator pair was titrated with some representative L-amino acids, which included: His, Ala, Phe, Leu, Pro, plus Gly. In some cases, the amino acid was able to displace the indicator from the receptor, an event signaled by full fluorescence revival. In other cases, however, the amino acid was not able to dislodge the indicator, with no restoration of fluores- cence. Some typical titration profiles are reported in Figure 1. The [Cu II 2 - (1)] 4+ /coumarine ensemble (Figure 1a) does not discriminate His and Gly. In fact, both His and Gly displace the indicator and restore its full emission. The situation is more favorable with the fluorescein-containing ensemble (Figure 1b), which satisfactorily discriminates His (full recovery of fluorescence) from Gly, whose I F profile is distinctly less steep. However, the highest sensing selectivity is observed with the [Cu II 2 (1)] 4+ /eosine Y ensemble (Figure 1c), which sufficiently discriminates His from Gly and other investigated amino acids. Nonlinear least-squares treatment of competitive titration profiles shown in Figure 1 allowed the determination of the equilibrium constants for the interaction of [Cu II 2 (1)] 4+ with each amino acid, 9 at pH ) 7. Corresponding values are reported in the bar diagram in Figure 2; the highest binding constant is shown by His, which offers to the two Cu II centers the strongly donating imidazolate bridging group. Binding by the Published on Web 12/05/2002 20 9 J. AM. CHEM. SOC. 2003, 125, 20-21 10.1021/ja027110l CCC: $25.00 © 2003 American Chemical Society