Characterization of Fluorescent Dyes for Biochemical Applications. Chemometric Analysis of Spectroscopic Data
Doctoral thesis, 1999
Fluorescent dyes are used in a number of different biochemical applications to visualize and probe biomolecules. The aim of this work was to characterize the spectroscopic properties of two such dyes, fluorescein and thiazole orange, and to develop chemometric methods to extract further information from these data.
Fluorescein is currently one of the most commonly used fluorescent probes. Its four protolytic forms have the protolytic constants pK1=2.08, pK2=4.31 and pK3=6.43. Upon conjugation to DNA the molar absorptivity of the monoanion and dianion decreases by about 20 %, and pK3, relating their concentrations, increases to about 6.90 due to the local electrostatic potential around the nucleic acid. The quantum yield is 0.93 for the dianion and 0.37 for the monoanion free in solution, whereas the neutral and cationic species are converted to the anion in the excited state. The fluorescence quantum yield of the DNA conjugates depends, in a complex way, on temperature, local environment and oligonucleotide length, sequence and conformation.
Thiazole orange (TO) has a fluorescence quantum yield of about 2x10-4 when free in solution. It binds to double stranded DNA with high affinity, presumably by intercalation and as a monomer. The quantum yield then increases to around 0.1 due to restricted rotation around the internal bond of the fluorophore. TO binds 5-10 times more weakly to single-stranded polypurines and a further 10-1000 times weaker still to single-stranded polypyrimidines. This binding is more complex and involves dimer formation.
In this thesis, new chemometric approaches that utilize equilibrium expressions to resolve spectra of the fluorophores mixtures into spectra of their pure components are described. These methods also calculate concentrations and equilibrium constants and can be applied either to compounds in protolytic or dimerization equilibrium, or to compounds involved in ligand-DNA interactions. Two new ways of obtaining data suitable for analysis by multi-way methods, by extraction followed by fluorescence measurements and by fluorescence polarization measurements, are also described. Different approaches to predict the number of components in spectroscopic data are evaluated, and an automated procedure to carry out the prediction is presented.
principal component analysis