Oxygen fluorescence quenching studies with single tryptophan-containing proteins.

The work of Lakowicz and Weber [Biochemistry 12, 4161 (1973)] demonstrated that molecular oxygen is a powerful quencher of tryptophan fluorescence in proteins. Here we report studies of the oxygen quenching of several proteins that have a single, internal tryptophan residue. Among these are apoazurin (Pseudomonas aeruginosa), asparaginase (Escherichia coli), ribonuclease T1 (Aspergillus oryzae), and cod parvalbumin. Both fluorescence intensity and phase lifetime quenching data are reported. By comparison of these data we find that there is a significant degree of apparent static quenching in these proteins. The dynamic quenching rate constants,k q, that we find are low compared to those for tryptophan residues in other proteins. For example, for apoazurin we find an apparentk q of 0.59×10(9) M (-1) s(-1) at 25°C. This value is the lowest that has been reported for the oxygen quenching of tryptophan fluorescence.

Luminescence studies with trp repressor and its single-tryptophan mutants.

Time-resolved and steady-state fluorescence, low-temperature phosphorescence, and optically detected magnetic resonance (ODMR) measurements have been made to resolve the luminescence contributions of the two intrinsic tryptophan residues in the subunits of trp aporepressor from Escherichia coli. Assignments of spectral information have been confirmed by use of the single-tryptophan mutants W19F and W99F. Solute fluorescence quenching studies show that both Trp19 and Trp99 are exposed to acrylamide and iodide, with Trp99 being the more exposed. Time-resolved and steady-state fluorescence measurements show Trp19 to have a bluer emission, a longer mean fluorescence decay time, a higher quantum yield, and essentially no independent rotational motion with respect to the protein. Trp99 is found to have a redder emission, a shorter mean fluorescence decay time, a lower quantum yield, and a significant degree of rotational freedom. Phosphorescence studies show a clear resolution of 0-0 vibronic transitions for each type of residue, with maxima at 407 and 415 nm that are assigned to Trp19 and Trp99, respectively. ODMR measurements show the zero-field splitting parameters to be quite characteristically different for each tryptophan residue. The existence of resonance energy transfer from Trp19 to Trp99, in the wild-type protein, is indicated by three types of data: comparison of the long-lived decay time (attributed to Trp19) in the absence (W99F) and presence (wild type) of the acceptor Trp99, comparison of the fluorescence quantum yield of the wild-type and mutant proteins, and deviations from the expected phosphorescence intensities for Trp19 and Trp99 in the absence of energy transfer.

Fluorescence studies with human epidermal growth factor.

Steady-state and time-resolved fluorescence studies have been performed with human epidermal growth factor, a small globular protein having two adjacent tryptophan residues near its C-terminus. Based on the relatively red fluorescence and accessibility to solute quenchers, the two tryptophan residues are found to be exposed to solvent. Anisotropy decay measurements show the dominant depolarizing process to have a sub-nanosecond rotational correlation time indicating the existence of rapid segmental motion of the fluorescing tryptophan residues. From an analysis of the low-temperature excitation anisotropy spectrum of the protein (and in comparison with that of tryptophan, the peptide melittin, and the dipeptide trp-trp), it is concluded that homo-energy transfer and/or exciton interaction occurs between the adjacent tryptophan residues. A thermal transition in the structure of the protein, which is observed by circular dichroism measurements, is not sensed by the steady-state fluorescence of the protein. This result, in conjunction with the anisotropy decay results, indicates that the two tryptophan residues are in a highly flexible C-terminus segment, which is not an integral part of the three-dimensional structure of the protein. Fluorescence measurements with three site-directed mutants also show very little variation.

Photosensitization of aqueous model systems by hypericin.

Absorption and fluorescence measurements of purified hypericin (HY) were made in various media. Photosensitization of two aqueous systems was investigated: resealed red blood cell membranes (ghosts) and hen lysozyme (Lys). Solubilization of HY by ghost membranes was shown by means of diffuse reflectance spectroscopy. Visible light irradiation of the ghosts incorporating HY led to lipid peroxidation with evidence of singlet oxygen involvement. A binding model applicable for insoluble ligands is indicative of strong HY binding to HSA. The HY-HSA complex photosensitized inactivation of Lys. The pseudo-first-order reaction kinetics with protection by azide ion are consistent with a Type II mechanism mediated by singlet oxygen. The results are discussed in the context of the HY photodynamic and antiretroviral activities.

Fluorescence and conformational stability studies of Staphylococcus nuclease and its mutants, including the less stable nuclease-concanavalin A hybrids.

We report steady-state and time-resolved fluorescence studies with the single tryptophan protein, Staphylococcus aureus A, and several of its site-directed mutants. A couple of these mutants, nuclease-conA and nuclease-conA-S28G (which are hybrid proteins containing a six amino acid beta-turn substitute from concanavalin A), are found to have a much lower thermodynamic stability than the wild type. The thermal transition temperatures for nuclease-conA and S28G are 32.8 and 30.5 degrees C, which are about 20 degrees C lower than the Tm for wild-type nuclease A. These mutant proteins also are denatured by a much lower concentration of the denaturants urea and guanidine hydrochloride. We also show that an unfolding transition in the structure of the nuclease-conA hybrids can be induced by relatively low hydrostatic pressure (approximately 700 bar). The free energy for unfolding of nuclease-conA (and nuclease-conA-S28G) is found to be only 1.4 kcal/mol (and 1.2 kcal/mol) by thermal, urea, guanidine hydrochloride, and pressure unfolding. Time-resolved fluorescence intensity and anisotropy measurements with nuclease-conA-S28G show the temperature-, urea-, and pressure-perturbed states each to have a reduced average intensity decay time and to depolarize with a rotational correlation time of approximately 1.0 ns (as compared to a rotational correlation time of 11 ns for the native form of nuclease-conA-S28G at 20 degrees C).

Fluorescence studies with potato carboxypeptidase inhibitor.

Potato carboxypeptidase inhibitor (CPI) is a 39-residue globular protein whose X-ray structure is known. The protein's two tryptophan residues (W22 and W28) appear to be on the surface in the crystal structure. The fluorescence spectrum of CPI has a maximum at 344 nm. Acrylamide solute quenching yields an upward curving Stern-Volmer plot with KSV approximately 9 M-1. KI quenching yields a linear plot with KSV approximately 5.5 M-1. These studies indicate that emission occurs from a solvent exposed residue(s). Fluorescence lifetime measurements were fitted to a biexponential decay law with tau 1 = 0.9 ns. f1 = 0.22 and tau 2 = 3.9 ns. Anisotropy decay data were described by a single rotational correlation time of 1.2 ns. This value is somewhat small for the global rotation of a protein of this size. The low temperature phosphorescence of CPI shows a biexponential decay, with a rapidly decaying 0.5-1.0 second component. The triplet state of at least one of the two tryptophan residues must be perturbed by interaction with nearby disulfide bonds.

Fluorescence studies with malate dehydrogenase from Bradyrhizobium japonicum 3I1B-143 bacteroids: a two-tryptophan containing protein.

A number of fluorescence studies, both of trp residues and bound NADH, have been reported for porcine malate dehydrogenase (MDH). The large number of trp residues (six) complicates the interpretation of some studies. To circumvent this we have performed studies with a two-tryptophan (per subunit) MDH from Bradyrhizobium japonicum 3I1B-143 bacteroids. We have performed phase/modulation fluorescence lifetime measurements, as a function of temperature and added quencher KI, in order to resolved the 1.2-ns (blue) and 6.5-ns (red) contributions from the two classes of trp residues. Anisotropy decay studies have also been performed. The binding of NADH dynamically quenches the fluorescence of both trp residues, but, unlike mammalian cytoplasmic and mitochondrial MDH, there is not a large enhancement in fluorescence of bound NADH upon forming a ternary complex with either tartronic acid or D-malate.

Fluorescence lifetime studies with staphylococcal nuclease and its site-directed mutant. Test of the hypothesis that proline isomerism is the basis for nonexponential decays.

Using frequency domain methods, the fluorescence decay of Trp-140 in staphylococcal nuclease and its site-directed mutant (Pro-117----Gly) has been examined. Based on nuclear magnetic resonance (NMR) studies (Evans, P. A., C. M. Dobson, R. A. Kautz, G. Hatfull, and R. O. Fox. 1987. Nature [Lond.]. 329:266-268), it is believed that nuclease exists in two macroscopic, native conformations and that the slow interconversion of these conformations is controlled by the cis----trans isomerization of Pro-117. The above mutant shows only one native conformation in NMR experiments. To test the hypothesis that the biexponential fluorescence decay of Trp-140 of nuclease can also be related to the existence of these conformational states of the protein, we have compared the decay patterns of the wild type and mutant. Essentially no difference was observed, which indicates that there is some other basis for the nonexponential decay of Trp-140. We have used global nonlinear least squares analysis to link the fit of data at several temperatures.

Fluorescence studies of phosphoribulokinase, a light-regulated, chloroplastic enzyme.

Recent characterization of spinach phosphoribulokinase has revealed that the homodimeric molecule contains only two tryptophans per 44-kDa subunit. We have performed steady-state and frequency domain studies of the intrinsic fluorescence of this protein. The fluorescence properties reflect contributions from both types of tryptophan residues. One of these appears to be relatively exposed to solvent and the quencher, acrylamide; fluoresce with a lambda max of 345 nm; decay with a fluorescence lifetime of 6.3 ns; have a relatively red-shifted absorption spectrum; and have a certain degree of independent motional freedom, with respect to the protein. The other tryptophan residue appears to be more buried; fluoresce with lambda max of 325 nm; have a lifetime of 1.7 ns; have a relatively blue-shifted absorption spectrum; and not to enjoy independent motional freedom. On comparison of phase-resolved spectral data and solute quenching data, we suggest that resonance energy transfer between the blue and red tryptophan residues may occur. We also describe the strategy of simultaneously fitting Stern-Volmer quenching data collected at two emission wavelengths.

Does the fluorescence quencher acrylamide bind to proteins?

We have studied the protein concentration dependence of the acrylamide quenching of the fluorescence of the proteins, human serum albumin and monellin, and we have found no such dependence for the concentration range of 0.5-20 mg/ml. These quenching studies were performed by fluorescence lifetime measurements using phase/modulation fluorometry. We have also performed equilibrium dialysis studies, which show no large degree of association of acrylamide with serum albumin, and we have found that acrylamide has only a small effect on the activity of selected enzymes. These various studies do not indicate the existence of strong acrylamide-protein interactions and are in discord with a recent report by Blatt et al. in this journal (Blatt, E., Husain, A. and Sawyer, W.H. (1986) Biochim. Biophys. Acta 871, 6-13).

Phase-resolved spectral measurements with several two tryptophan containing proteins.

We have used frequency domain fluorescence techniques to resolve the component emission spectra for several two tryptophan containing proteins (e.g., horse liver alcohol dehydrogenase, sperm whale apomyoglobin, yeast 3-phosphoglycerate kinase, apoazurin from Alcaligenes denitricans). We have first performed multifrequency phase/modulation measurements and have found the fluorescence of each of these proteins to be described by a double exponential. Then, using phase-sensitive detection and the algorithm of Gratton and Jameson [Gratton, E., & Jameson, D. M. (1985) Anal. Chem. 57, 1694-1697], we have determined the emission spectrum associated with each decay time for these proteins. We have compared these phase-resolved spectra with the fractional contributions of the component fluorophores determined by selective solute quenching experiments. Reasonably good agreement is seen in most cases, which argues that the individual Trp residues emit independently. In the case of apoazurin, however, a negative amplitude is seen for the phase-resolved spectrum of the short-lifetime component. This pattern is consistent with the occurrence of energy transfer from the internal Trp residue to the surface Trp of this protein. We also present multifrequency lifetime measurements, phase-resolved spectra, and solute quenching data for a few protein-ligand complexes, to illustrate the utility of this approach for the study of changes in the fluorescence of proteins.

Frequency domain measurements of the fluorescence lifetime of ribonuclease T1.

Using multifrequency phase/modulation fluorometry, we have studied the fluorescence decay of the single tryptophan residue of ribonuclease T1 (RNase T1). At neutral pH (7.4) we find that the decay is a double exponential (tau 1 = 3.74 ns, tau 2 = 1.06 ns, f1 = 0.945), in agreement with results from pulsed fluorometry. At pH 5.5 the decay is well described by a single decay time (tau = 3.8 ns). Alternatively, we have fitted the frequency domain data by a distribution of lifetimes. Temperature dependence studies were performed. If analyzed via a double exponential model, the activation energy for the inverse of the short lifetime component (at pH 7.4) is found to be 3.6 kcal/mol, as compared with a value of 1.0 kcal/mol for the activation energy of the inverse of the long lifetime component. If analyzed via the distribution model, the width of the distribution is found to increase at higher temperature. We have also repeated, using lifetime measurements, the temperature dependence of the acrylamide quenching of the fluorescence of RNase T1 at pH 5.5. We find an activation energy of 8 kcal/mol for acrylamide quenching, in agreement with our earlier report.

Proximity relations between tyrosine and tryptophan residues in fd phage determined by luminescence measurements.

Resonance energy transfer from tyrosine to tryptophan residues was detected in the phage fd. The magnitude of the transfer efficiency was estimated by both a traditional and an alternative method. The latter involved comparison of the indole acceptor excitation spectrum full-width half-maximum with a set of standard values differing in the amount of absorbance contributed by tyrosine donor. Both methods lead to the same conclusion: essentially all the tyrosine residues of the viral coat are within 0.9 nm of a tryptophan residue. Also, fluorescence lifetime measurements provide additional support for the hypothesis that there are at least two different environments for the coat protein's sole indole side-chain. Little if any DNA phosphorescence was seen, consistent with the nucleic acid bases being stacked in the DNA core.

Evidence from pulse radiolysis and flash photolysis of the environment of tryptophan and tyrosine in the coat protein of fd phage.

Charge transfer has been observed between oxidised tryptophan-26 units and the tyrosine-21 or -24 of the coat protein of fd phage. The transfer is likely to be intramolecular. The rates suggest that the aromatic units are in a rigid region and that they may have at least two different environments. No apparent interaction occurs with the DNA, consistent with tryptophan and tyrosine units not being in contact with the bases.

A spectroscopic analysis of the thermally induced folding-unfolding transition of beta-trypsin.

Absorption and fluorescence changes were used to monitor the thermally induced folding-unfolding transition of beta-trypsin. These parameters reflect changes in the microenvironment of different subsets of the four tryptophanyl residues of this protein. The thermal transition was found to be sequential.

Transfer of singlet energy within trypsin.

Transfers of singlet energy within trypsin were investigated by measuring the fluorescence absorption anisotropy of its tryptophan residues. A ratio of the anisotropy of trypsin to that for N-acetyl-L-tryptophanamide was determined between 306 and 250 nm. The ratio had an average value of 0.7, whether the trypsin anisotropy was measured at 228 of 296 K. However, trypsin dissolved in 5 M guanidine hydrochloride showed little fluorescence depolarization at 228 K (the anisotropy ratio was approximately equal to 0.9). Thus, there is an extensive conformation-dependent energy transfer between tryptophans in trypsin. The ratio of anisotropies of tyrpsin at 304--270 nm was used to estimate energy transfer from tyrosine to tryptophan. Ratios of 1.8 and 1.7 were obtained at 296 K for the native and guanidinium-unfolded enzyme, respectively. The comparable value for N-acetyl-L-tryptophanamide was 1.7. This indicates that there is little transfer from tyrosine to tryptophan in trypsin at 296 K. As confirmation, the excitation wavelength dependencies of the indole fluorescence quantum yield were the same for native and unfolded trypsin. When experiments were performed at 228 K, the 304--270-nm anisotropy ratios were 2.6 for native and 2.1 for unfolded trypsin at pH2. This indicates that the efficiency of energy transfer from tyrosine to tryptophan increases at low temperatures. A photochemical source of error in the quantitation of the efficiency of energy transfer from tyrosine to tryptophan is also described.

Flash photolysis of human serum albumin: characterization of the indole triplet absorption spectrum and decay at ambient temperature.

The method of flash photolysis was used to identify the transient absorption spectrum and to characterize the decay kinetics of the indole triplet of human serum albumin. This protein was studied because it contains a single indole side chain which is deeply buried in an expandable oily region and because the phosphorescence of the homologous indole in bovine serum albumin could not be detected at ambient temperatures. The transient was identified on the following basis: (i) its triplet-triplet absorption spectrum was similar to those previously reported for indole and tryptophan; (ii) it was quenched by small quantities of oxygen; and (iii) it was photobleached by 370- to 700-nm light. In a nitrogen-saturated solution at room temperature, the indole triplet decays exponentially for more than a factor of 10 with a lifetime of 0.5 msec. These observations suggest that, because of its exponential decay and relatively long lifetime, the triplet will be more valuable than the indole singlet as an intrinsic reporter group for the study of the structure and dynamics of proteins in solution.

Indole to phenolate energy transfer, the alkaline quenching mechanism of beta-trypsin fluorescence.

Evidence that indole to phenolate energy transfer causes the decrease in beta-trypsin fluorescence with increasing alkalinity is obtained indirectly from viscosity, lifetime, wavelength dependence and chemical modification studies which either exclude or minimize the involvement of other possible mechanisms.