Principles of Fluorescence Spectroscopy

The successful application of fluorescence methods requires considerable attention to experimental details and a good understanding of the instrumentation. There are numerous potential artifacts which can distort the data. Fluorescence is a highly sensitive method. The gain or amplification of the instruments can usually be increased to obtain observable signals, even if the sample is nonfluorescent. These signals seen at high amplification may not originate with the fluorophore of interest. Instead, one may observe interference due to background fluorescence from the solvents, light leaks in the instrumentation, stray light passing through the optics, light scattered by turbid solutions, Rayleigh scatter, and/or Raman scatter, to name a few sources of interference. Furthermore, there is no ideal spectrofluorometer, and the available instruments do not yield true excitation or emission spectra. This is because of the nonuniform spectral output of the light sources and the wavelength-dependent efficiency of the monochromators and detectors (photomultiplier tubes). The polarization or anisotropy of the emitted light can also affect the measured fluorescence intensities. To obtain reliable spectral data, one needs to be aware of and control these numerous factors. In this chapter we will discuss the properties of the individual components in a spectrofluorometer and how these properties affect the observed spectral data.

Fluorescence probes represent the most important area of fluorescence spectroscopy. One can spend a great deal of time describing the instrumentation for fluorescence spectroscopy, including light sources, monochromators, lasers, and detectors. However, in the final analysis, the wavelength and time resolution required of the instruments are determined by the spectral properties of the fluorophores. Furthermore, the information available from the experiments is determined by the properties of the probes. Only probes with nonzero anisotropies can be used to measure rotational diffusion, and the lifetime of the fluorophore must be comparable to the correlation time of interest. Only probes which are sensitive to pH can be used to measure pH. And only probes with reasonably long excitation and emission wavelengths can be used with tissues which display autofluorescence at short excitation wavelengths.

Time-resolved measurements are widely used in fluorescence spectroscopy, particularly for studies of biological macromolecules. This is because time-resolved data frequently contain more information than is available from the steady-state data. For instance, consider a protein which contains two tryptophan residues, each with a distinct lifetime. Because of spectral overlap of the absorption and emission, it is not usually possible to resolve the emission from the two residues. However, the time-resolved data may reveal two decay times, which can be used to resolve the emission spectra and relative intensities of the two tryptophan residues. Then one can question how each of the tryptophan residues is affected by the interactions of the protein with its substrate or other macromolecules. Is one of the tryptophan residues close to the binding site? Is a tryptophan residue in a distal domain affected by substrate binding to another domain? Such questions can be answered if one measures the decay times associated with each tryptophan residue.

In the preceding chapter we described the theory and instrumentation for measuring fluorescence intensity decays using time-domain measurements. In the present chapter we continue this discussion, but we now consider the alternative method called frequency-domain (FD) fluorometry. In this method the sample is excited with light which is intensity-modulated at a high frequency comparable to the reciprocal of the lifetime. When this is done, the emission is also intensity-modulated at the same frequency. However, the emission does not precisely follow the excitation but rather shows time delays and amplitude changes which are determined by the intensity decay law of the sample. To be more precise, the time delay is measured as a phase-angle shift between the excitation and emission, as was shown in Figure 4.2. The peak-to-peak height of the modulated emission is decreased relative to that of the modulated excitation and provides another independent measure of the lifetime.

Solvent polarity and the local environment have profound effects on the emission spectra of polar fluorophores. These effects are the origin of the Stokes’ shift, which is one of the earliest observations in fluorescence. Emission spectra are easily measured, and as a result, there are numerous publications on emission spectra of fluorophores in different solvents and when bound to proteins, membranes, and nucleic acids. One common use of solvent effects is to determine the polarity of the probe binding site on the macromolecule. This is accomplished by comparison of the emission spectra and/or quantum yields of the fluorophore when it is bound to the macromolecule and when it is dissolved in solvents of different polarity. However, there are many additional instances where solvent effects are used. Suppose a fluorescent ligand binds to a protein. Binding is usually accompanied by a spectral shift due to the different environment for the bound ligand. Alternatively, the ligand may induce a spectral shift in the intrinsic or extrinsic protein fluorescence. Additionally, fluorophores often display spectral shifts when they bind to membranes.

In the preceding chapter we described the effects of solvents on emission spectra and considered how the solvent-dependent data could be interpreted in terms of the local environment. We assumed that the solvent was in equilibrium around the excited-state dipole prior to emission. Equilibrium around the excited-state dipole is reached in fluid solution because the solvent relaxation times are typically less than 100 ps whereas the decay times are usually 1 ns or longer. However, equilibrium around the excited-state dipole is not reached in more viscous solvents and may not be reached for probes bound to proteins or membranes. In these cases emission occurs during solvent relaxation, and the emission spectrum represents an average of the partially relaxed emission. Under these conditions, the emission spectra display time-dependent changes. These effects are not observed in the steady-state emission spectra but can be seen in the time-resolved data or the intensity decays measured at various emission wavelengths.

Fluorescence quenching refers to any process which decreases the fluorescence intensity of a sample. A variety of molecular interactions can result in quenching. These include excited-state reactions, molecular rearrangements, energy transfer, ground-state complex formation, and collisional quenching. In this chapter we will be concerned primarily with quenching resulting from collisional encounters between the fluorophore and quencher, which is called collisional or dynamic quenching. Static quenching is a frequent complicating factor in the analysis of quenching data, but it can also be a valuable source of information about binding between the fluorescent sample and the quencher. In addition to the processes described above, apparent quenching can occur due to the optical properties of the sample. For example, high optical densities or turbidity can result in decreased fluorescence intensities. This is a trivial type of quenching which contains little molecular information. Throughout this chapter, we will assume that such trivial effects are not the cause of the observed decreases in fluorescence intensity.

In the previous chapter we described the basic principles of quenching. A wide variety of quenchers are known. Quenching requires molecular contact between the fluorophore and the quencher. This contact can be due to diffusive encounters (dynamic quenching), or to complex formation (static quenching). Because of the close-range interaction required for quenching, the extent of quenching is sensitive to molecular factors that affect the rate and probability of contact, including steric shielding and charge—charge interactions.

In the preceding chapter we described the measurement and interpretation of steady-state fluorescence anisotropies. These values are measured using continuous illumination and represent an average of the anisotropy decay over the intensity decay. Measurement of steady-state anisotropies is simple, but interpretation of the steady-state anisotropies usually depends on an assumed form for the anisotropy decay, which is not directly observed in the experiment. Additional information is available if one measures the time-dependent anisotropy, that is, the values of r(t) following pulsed excitation. The form of the anisotropy decay depends on the size, shape, and flexibility of the labeled molecule, and the data can be compared with the decays calculated from various molecular models. Anisotropy decays can be obtained using the TD or the FD method.

In the preceding two chapters we described steady-state and time-resolved anisotropy measurements and presented a number of biochemical examples which illustrated the types of information available from these measurements. Throughout these chapters, we stated that anisotropy decay depends on the size and shape of the rotating species. However, the theory which relates the form of the anisotropy decay to the shape of the molecule is complex and was not described in detail. In the present chapter we provide an overview of the rotational properties of nonspherical molecules, as well as representative examples.

Fluorescence resonance energy transfer (FRET) is transfer of the excited-state energy from the initially excited donor (D) to an acceptor (A). The donor molecules typically emit at shorter wavelengths which overlap with the absorption spectrum of the acceptor. Energy transfer occurs without the appearance of a photon and is the result of long-range dipole-dipole interactions between the donor and acceptor. The term resonance energy transfer (RET) is preferred because the process does not involve the appearance of a photon. The rate of energy transfer depends upon the extent of spectral overlap of the emission spectrum of the donor with the absorption spectrum of the acceptor, the quantum yield of the donor, the relative orientation of the donor and acceptor transition dipoles, and the distance between the donor and acceptor molecules. The distance dependence of RET has resulted in its widespread use to measure distances between donors and acceptors.

In the previous chapter we described the principles of resonance energy transfer and how the phenomenon could be used as a “spectroscopic ruler” to measure distances between donor and acceptor sites on macromolecules. Energy transfer was described as a through-space interaction which occurred whenever the emission spectrum of the donor overlapped with the absorption spectrum of the acceptor. For a given donor—acceptor (D—A) pair, the efficiency of energy transfer decreases as r ⁻⁶, where r is the D—A distance. Each D—A pair has a characteristic distance, the Förster distance (R ₀), at which RET is 50% efficient. The extent of energy transfer, as seen from the steady-state data, can be used to measure the distance, to determine the extent of association based on proximity, or to determine the distance of closest approach between the D—A pair.

In the previous two chapters on energy transfer, we considered primarily covalently linked donor—acceptor pairs, or situations in which there was a single acceptor near each donor. However, there are numerous situations in which there exist multiple acceptors, such as the obvious case of donors and acceptors dissolved in homogeneous solutions. More interesting examples of the multiple-acceptor case occur in membranes and nucleic acids. Suppose that one has a lipid bilayer which contains both donors and acceptors (Figure 15.1, middle). Each donor will be surrounded by acceptors in two dimensions. Since the acceptor distribution is random, each donor sees a different acceptor population. Hence, the intensity decay is an ensemble average and is typically nonexponential. A similar situation exists for donors and acceptors which are intercalated into double-helical DNA (Figure 15.1, right), except that in this case the acceptors are distributed in one dimension along the DNA helix.

Discussions of biochemical fluorescence frequently start with the subject of protein fluorescence. This is because, among biopolymers, proteins are unique in displaying useful intrinsic fluorescence. Lipids, membranes, and saccharides are essentially nonfluorescent, and the intrinsic fluorescence of DNA is too weak to be useful. In proteins, the three aromatic amino acids—phenylalanine, tyrosine, and tryptophan—are all fluorescent. A favorable feature of protein structure is that these three amino acids are relatively rare in proteins. Tryptophan, which is the dominant intrinsic fluorophore, is generally present at about 1 mol % in proteins. A protein may possess just one or a few tryptophan residues, which facilitates interpretation of the spectral data. If all 20 amino acids were fluorescent, it is probable that protein emission would be too complex to interpret.

In the previous chapter we presented an overview of protein fluorescence. We described the spectral properties of the aromatic amino acids and how these properties depend on protein structure. We now extend this discussion to include time-resolved measurements of intrinsic protein fluorescence. Prior to 1983, most measurements of time-resolved fluorescence were performed using TCSPC. The instruments employed for these measurements typically used a flashlamp excitation source and a standard dynodechain-type PMT. Such instruments provided instrument response functions with a half-width near 2 ns, which is comparable to the decay time of most proteins. The limited repetition rate of the flashlamps, near 20 kHz, resulted in data of modest statistical accuracy, unless the acquisition times were excessively long. Given the complexity of protein intensity and anisotropy decays, and the inherent difficulty of resolving multiexponential processes, it was difficult to obtain definitive information on the decay kinetics of proteins.

In the preceding chapters we saw many examples of excited-state reactions. By an excited-state reaction we mean a molecular process which changes the structure of the excited-state fluorophore, and which occurs subsequent to excitation. Such reactions occur because light absorption frequently changes the electron distribution within a fluorophore, which in turn changes its chemical or physical properties. The best-known example of an excited-state reaction is that of phenol, which in neutral solution can lose the phenolic proton in the excited state. Deprotonation occurs more readily in the excited state because the electrons on the phenolic hydroxyl groups are shifted into the phenol ring, making this hydroxyl group more acidic.

Fluorescence sensing of chemical and biochemical analytes is an active area of research.^1–7 This research is being driven by the desire to eliminate radioactive tracers, which are costly to use and dispose of. Additionally, there is a need for rapid and low-cost testing methods for a wide range of clinical, bioprocess, and environmental applications. During the past decade, we have witnessed the introduction of numerous methods based on high-sensitivity fluorescence detection, including DNA sequencing, DNA fragment analysis, fluorescence staining of gels following electrophoretic separation, and a variety of fluorescence immunoassays. Historically, one can trace many of these analytical applications to the classic reports by Undenfriend and co-workers,^8,9 which anticipated many of today’s applications of fluorescence. More recent monographs have summarized the numerous analytical applications of fluorescence.^10–14

Throughout the previous chapters, we described fluorophores with decay times ranging from 1 to 20 ns. While this timescale is useful for many biophysical measurements, there are numerous instances where longer decay times are desirable. For instance, one may wish to measure rotational motions of large proteins or membrane-bound proteins. In such cases the overall rotational correlation times can easily be longer than 200 ns, and they can exceed 1 μs for larger macromolecular assemblies. Rotational motions on this timescale are not measurable using fluorophores which display nanosecond lifetimes. Processes on the microsecond or even the millisecond timescale have occasionally been measured using phosphorescence.^1–4 However, relatively few probes display useful phosphorescence in room-temperature aqueous solutions. Also, it is usually necessary to perform phosphorescence measurements in the complete absence of oxygen. Hence, there is a clear need for probes which display microsecond lifetimes. In this chapter we describe a family of metal—ligand probes which display decay times ranging from 100 ns to 10μs. The long lifetimes of the metal—ligand probes allow the use of gated detection, which can be employed to suppress interfering autofluorescence from biological samples and can thus provide increased sensitivity.⁵ Finally, the metal—ligand probes display high chemical and photochemical stability. Because of these favorable properties, we expect metal—ligand probes to have numerous uses in biophysical chemistry, clinical chemistry, and DNA diagnostics.

In Chapter 5 we described the use of the FD method to measure lifetimes and to resolve complex intensity decays. In FD fluorometers, the sample is excited with intensity-modulated light, and one measures the phase shift and modulation of the emission, both relative to the excitation. The FD method also allows several other types of measurement which can be useful in special circumstances. One method is measurement of phase-sensitive intensities and/or emission spectra. Another method is to use the measured phase and modulation values to resolve the components of species in a mixture based on known decay times.