Fluorescence kinetics of aqueous solutions of tetracycline and its complexes with Mg2+ and Ca2+

The fluorescence spectra of acidified aqueous solutions of tetracycline (tcH3+) exhibit three components with slightly different degrees of anisotropy: the 'blue' component (lambdad approximately equal to 475 nm) decays on the timescale of a few picoseconds; the second, most intense component (lambdad approximately equal to 530 nm) shows decay times of about 25 (H3O+) and 70 ps (D3O+); the third component (lambdad approximately equal to 650 nm) is longer lived (tau approximately equal to 200 ps). All three fluorescence components appear quasi-instantaneously, thus providing evidence that the relaxation processes which give rise to the unusually large Stokes shifts occur on a (sub-)picosecond timescale. The effect of H/D exchange suggests that these relaxation processes involve excited-state intramolecular proton transfer (ESIPT) of OH10 and/or OH12, but does not exclude a change in the hydrogen-bonding pattern to the solvent molecules. The low overall fluorescence yield of the fully protonated form must be correlated to the presence of a very fast decaying species. In alkaline aqueous solution, the fluorescence of the dianion (tc2-) essentially comprises two components; the decay time of the stronger, shorter-lived component is about 30 ps, that of the weaker, longer-lived one about 160 ps. The relative amplitude of the latter is larger at pH 11 than at pH 8.5, in accordance with the increase in the steady-state fluorescence intensity upon increasing the pH from 8.5 to 11. Complexation of the dianion with divalent metal ions like Mg2+ or Ca2+ leads to a strong enhancement of the steady-state fluorescence. In the time-resolved spectra, the decay time of the major fluorescence component exhibits approximately a five-fold increase in comparison to the major component of the dianion. It is about 150 ps in both types of complexes. The decay times of the minor component are increased to about 500 (Mg2+) and 320 ps (Ca2+). The absence of the ultra-fast component in the fluorescence of the dianion and its metal complexes provides evidence that a reaction of OH 12 must be responsible for the ultra-fast fluorescence component in tcH3+. The existence of a component with a lifetime of several tens of picoseconds in all samples suggests the involvement of hydrogen bonding at OH10 during the formation of the emitting species. DFT calculations for the isolated molecule provide evidence that ESIPT is indeed an energetically allowed relaxation process for those isomers that have only one intramolecular hydrogen bond to O11. The ESIPT process yields primary photoproducts that should emit at much longer wavelengths, thereby explaining the unusually large fluorescence Stokes shift.