In the past several years the utility of fluorescence decay measurements in biological and biochemical studies has been clearly demonstrated. The time correlated single photon counting (TCSPC) technique has been used extensively in these experiments. Improvements in the instrumentation and data analysis have allowed higher sensitivities with a higher temporal and spectral resolution of the data. The results have led to new insights into the understanding of conformational heterogeneity and dynamic processes in biochemical systems. The combination of picosecond high repetition rate pulsed laser excitation with micro-channel plate photomultiplier detectors has been a major advance in the instrumentation. Concomitant with these instrumental advances data analysis methods have also made great progress. These developments are reviewed and recent examples of applications to studies of biomolecules are presented.

We review multiplexed time-correlated single photon counting with respect Lo methods of implementation and advantages over simplex methods. The techniques are illustrated by recent results we have obtained with differential fluorometry, anisotropy studies, multiwavelength array detection and flow-cell fluorometry.

We describe our frequency-domain fluorometer which allows phase angle and demodulation measurements from 8 MHz to 2 GHz. This instrument using the harmonic content of a 7.59 MHz train of 5 ps pulses as the modulated excitation source. The detector is a microchannel plate PMT (R1564U), whose bandwidth extends to 2 GHz. Using this instrument we examined rapid rotational motions of indole in low viscosity solvents. To date, the shortest correlation time we have determined was 7 ps, for indole in methanol at 80°C. For indole in cyclohexane at 20°C we were able to resolve two rotational correlation times of 17 and 73 ps. It now appears that the frequency range can be extended to 8 GHz, so still faster processes should be observable.

A new fiber-optic-based fluorescence lifetime instrument is described. The new instrument consists of a multifrequency cross-correlation phase-modulation fluorometer and a fiber-optic sensor. Results are shown which demonstrate the resolution of complex decay kinetics (double and triple exponential decays of fluorescence) and remote sensing. In addition, results are shown for the accurate and precise determination of fluorescence lifetimes between 125ps and 225ns at locations up to 100 meters from the instrument.

Synchrotron radiation is an excellent excitation source for both the time-domain (pulse) and the frequency-domain (phase/modulation) methods of time-resolved fluorescence spectroscopy. All possible wavelengths are available to excite fluorescent biological probes since synchrotron radiation consists of an intense continuum of energies from the IR to X-rays. The pulsed nature of synchrotron radiation permits time-correlated single photon-counting techniques to collect a histogram of the probability of the decay of the excited state. The harmonics of the pulsed exciting light also provide a wide range of modulation frequencies that permit the frequency-domain method to match the current resolution for time-domain measurements.

The photoelectronic streak camera can clearly resolve picosecond electronic and molecular processes occurring in DNA and in proteins. Particularly appropriate examples are: fluorescence kinetics of nucleic acids which are sensitive to structure and base stacking interactions; fast rotational depolarization of emission from tyrosine sidechains in proteins; excited-state molecular dynamics as measured by time-resolved emission spectra.

Time resolved fluorescence spectroscopy is rapidly gaining popularity as a tool for studying the structure and dynamics of macromolecules. The multidimensional nature of luminescence has recently been exploited in a variety of data analysis innovations, all of which are intended to reveal the underlying heterogeneity of the emission from macromolecules. Global analysis (1,2) and association methods (3,4,5,6,7) have used multiple emission and/or excitation wavelengths, differing polarizations, concentration variations (especially global Stern Volmer quenching analysis and QDAS;7,8) and system-specific parameters (pH, temperature, and other conformational effectors, see 9) to spread analysis across more dimensions. When conditions change during collection, kinetic decay analysis (10,11) provides another new global view using the nanosecond and second time axes simultaneously ("kinetics of kinetics", see 9). The Laboratory of Technical Development, NIHB1, is dedicated to innovation in instrumental design, and we have focused attention on new, multidimensional approaches to time-resolved spectrophotofluorometry. This contribution will outline the current status of the LTD laser fluorometer along with plans being used for some unusual prototypes now under construction. Since rapid multidimensional data acquisition is likely to accelerate the growth of biochemical applications, we will judge current and planned methods by four criteria: 1) what is the expected limiting lifetime resolution in psec (fraction of TTS)? 2) what gain/sensitivity can be expected (relative to excellent current state of the art)? 3) what acquisition time will be needed for moderate SNR (and does SNR improv.e like (Tacq)**1/2)? 4) how many spatial, wavelength (etc.) channels can be acquired simultaneously? The current (and/or theoretically expected) answers to these questions are summarized in table I.

Picosecond relaxation processes of flavins and flavoproteins were investigated with mode-locked and synchronously pumped lasers as source of excitation and time-correlated single photon counting in detection. Free flavin rotational correlation times of 80-150 ps (values depending on the flavin derivative used) could be precisely determined. Picosecond-resolved fluorescence of the flavin bound in the electron-carrier protein flavodoxin from Desulfovibrio vulgaris yields a fluorescence lifetime component of 30 ps in the fluorescence decay. Time-resolved tryptophan fluorescence in flavodoxin exhibits a short lifetime component, which is attributed in part to energy transfer from tryptophan to flavin. Three-dimensional fluorescence spectroscopy and fluorescence anisotropy decay analysis of the two tryptophan residues in flavodoxin provide new evidence for specific flavin-tryptophan interaction. Finally, picosecond-resolved spectroscopy enables the direct measurement of energy transfer between two different chromophores in a protein, from which topographical details can be inferred.

A fluorescence spectroscopy global analysis environment is described. Within this analysis environment, multidimensional fluorescence decay data (time and frequency domains) can be analyzed in terms of a wide variety of photophysical models. A generalized compartmental analysis structure is utilized, where one can specify the functions used to link the various compartments together. All fitting parameters may be characterized by either discrete or distributed values. Multi-temperature global analysis of frequency domain fluorescence studies from a single tryptophan containing protein are described.

The application of global analysis to time-dependent fluorescence data provides a power-ful technique for investigating complex, heterogeneous systems by defining common parameters to elements of a data set where the data has been obtained as a function of an independent variable. Another, related approach is to incorporate information obtained from other physical techniques into the analysis of time-dependent fluorescence data. The information from the independent method is used to define a linkage among the parameters within one element of the time-dependent fluorescence data set; the linkage establishes certain parameters as dependent variables. We call this a linked-function analysis, which can be used in the analysis of a single data set element or globally in the analysis of multiple data set elemets. This approach has been applied to time-domain fluorescence spectroscopy combi n ned with 1H-NMR spectroscopy to resolve ground-state and excited-state processes in the fluorescence decay of aromatic amino residues in peptides.

This paper discusses the problem of fluorescent molecular systems where the complexity of the molecule or its environment give rise to distributions of decay times rather than discrete two- or three-component decay behavior. The Exponential Series Method for distribution recovery is examined and examples from recent work presented.

This paper examines the shape information available from measurements of fluorescence anisotropy decays. We report anisotropy decays for three systems and examine the kinds of information which may be obtained. Results for the dye rose bengal (~1000 dal-tons) suggest that an approximation used in rotational diffusion theory, that the solvent molecules are very small compared to the solute, is valid even for this relatively small molecule. Second, we examine the effects of calcium concentration on rotational diffusion of the protein calmodulin (~17,000 daltons) derivatized by crosslinking to form a dityrosine fluorophore. In the presence of sufficiently high calcium ion concentration, this crosslinked calmodulin shows a single exponential anisotropy decay which indicates rotational diffusion consistent with the extended dumbbell structure found for calmodulin by x-ray crystal-lography. At low calcium concentrations, th e crosslinked calmodulin rotates considerably faster, suggesting a much more compact shape; also, this anisotropy decay includes considerably shorter correlation times which are interpreted as arising from a segmental flexibility not evident for crosslinked calmodulin at high calcium concentration. Finally, we examine a transition which is observed at very low salt concentrations for nucleosome core particles, relatively large complexes (~200,000 daltons) derived from chromatin and comprised of eight histone molecules and 145 base pairs of DNA. The anisotropy decay of ethidium bound to the DNA indicates that core particles exposed to low ionic strength are considerably elongated relative to the shape at higher ionic strength.

Immunoglobulins are flexible proteins which display large-amplitude modes of motion on a nanosecond time scale. Different classes of antibodies differ markedly in their degree of segmental flexibility; a number of their essential biological functions are correlated with their nanosecond internal dynamics. These motions can be conveniently monitored by time-resolved fluorescence anisotropy measurements. An instrument built around a synch-pumped cavity-dumped dye laser and a fast time-to-digital convertor with histogramming memory has made it possible to obtain high-quality anisotropy in a few minutes on small amounts (ca. 100 pmol) of protein. Genetic engineering techniques have made it possible to construct a large number of immunoglobulins with identical binding sites for the fluorescent probe dansyllysine. These proteins differ in the heavy chain regions which are responsible for their biological effector functions and their segmental flexibility. We have analyzed a series of such constructs derived by genetic recombination between the genes coding for the mouse isotypes IgG1 and IgG2a. The results identify two regions responsible for their different degrees of segmental flexibility: the hinge region connecting the Fab and Fc portions of the antibodies, and a short stretch (residues 131-139) of sequence in the amino terminal half of the CH1 domain containing five amino acid substitutions.

Fluorescence decay and anisotropy studies of parinaric acid in phospholipid bilayers have demonstrated the presence of density fluctuations exhibiting critical behavior in these nominally single phase structures. The anisotropy behavior is used to extract the second and fourth rank order parameters <P2> and <P4> for comparison with models of acyl chain order. A recent study of parinaric acid in hexagonal urea inclusion complexes has demonstrated that the strongly allowed transition is polarized at an angle with respect to the chain axis. The implications of this for the interpretation of anisotropy experiments is discussed. The lysozyme from bacteriophage aq has been engineered in modified forms containing only one tryptophan residue and with substitutions near the buried tryptophan residue 138. Simple changes in the structure appear to result in large changes in the dynamics of this residue. These observations are compared with the results of molecular dynamics computations.

We calculated the fluorescence depolarization signal following polarized excitation using a model-independent approach where the time development of the molecular orientation density is equal to the operation of a linear time-development operator on the initial molecular orientation density. The time-development operator, T, is not explicitly specified and the time dependent fluorescence signal can be expressed in terms of well defined matrix elements of T. We discuss two strategies with which to calculate the matrix elements of T from the experimental time-resolved signal. We applied this formalism to time-resolved studies of fluorescent labeled myosin cross-bridges in relaxed muscle fibers to measure the steady-state cross-bridge angular probability density. We find that using the model of rotational diffusion in an angular potential we can estimate the rank six order parameters of the angular probability density. The rank six order parameters are shown to make a significant contribution to the proposed cross-bridge angular distribution for relaxed muscle fibers.

We investigated the 3-dimensional fluorescence of complex mixtures of bioloquids such as human serum, serum ultrafiltrate, human urine, and human plasma low density lipoproteins. The total fluorescence of human serum can be divided into a few peaks. When comparing fluorescence topograms of sera, from normal and cancerous subjects, we found significant differences in tryptophan fluorescence. Although the total fluorescence of human urine can be resolved into 3-5 distinct peaks, some of them. do not result from single fluorescent urinary metabolites, but rather from. several species having similar spectral properties. Human plasma, low density lipoproteins possess a native fluorescence that changes when submitted to in-vitro autoxidation. The 3-dimensional fluorescence demonstrated the presence of 7 fluorophores in the lipid domain, and 6 fluorophores in the protein. dovain- The above results demonstrated that 3-dimensional fluorescence can resolve the spectral properties of complex ,lxtures much better than other methods. Moreover, other parameters than excitation and emission wavelength and intensity (for instance fluorescence lifetime, polarization, or quenchability) may be exploited to give a multidl,ensio,a1 matrix, that is unique for each sample. Consequently, 3-dimensio:Hhal fluorescence as such, or in combination with separation techniques is therefore considered to have the potential of becoming a useful new H.ethod in clinical chemistry and analytical biochemistry.

Using time-resolved nanosecond fluorescence spectroscopy, we investigated the conformational changes of skeletal and cardiac troponin C. A thiol specific fluorescence probe N-(iodoacety1)-N'-(5-sulfo-1-naphthyl)- ethylenediamine (IAEDANS) was attached to cysteine 98 of skeletal troponin C (STnC) and 2-(4'-iodoacetamido-anilino)-naphthalene-6-sulfonic acid (IAANS) was linked to cysteine 35 and 84 of cardiac troponin C (CTnC). With excitation at 340 nm for STnC-IAEDANS and at 335 nm for CTnC-IAANS, apo-STnC and apo-CTnC exhibited biexponential decay kinetics. At 20°C and neutral pH, the following lifetimes were observed: (1) apo-STnC, 9 and 16 ns, and (2) apo-CTnC, 2.3 and 7 ns. The long lived component of the emission in STnC-IAEDANS comprised ~ 61% of observed intensity, however, the corresponding component in CTnC contributed only. A decrease of pH from 7.2 to 5.2 induced an increase of the lifetimes 20% (STnC) and 10% (CTnC). These results suggest that Cys-98 of STnC and Cys-35 and Cys-84 of CTnC became less quenched by their neighboring residues at low pH. Addition of guanidine hydrochloride to STnC resulted in a decrease of ~30% of both lifetimes. The lifetimes increased slightly when the temperature was lowered. Variation of solution viscosity by addition of sucrose did not affect the long component of the lifetimes of STnC. However, the short component did sense a viscosity effect. These results suggest that there was likely a chromophore heterogeneity which may arise from differences in conformation, environment, and or ionization of the excited state of the chromophore. At 20°C and neutral pH, two rotational correlation times were observed: (1) apo-STnC, - 1.2 ns, .2 1 11.3 ns, and (2) apo-CTnC, .~ 0.6 ns, .2 1 13.6 ns. The short rotational correlation times likely reflect rapid motions of the chromophores covalently attached to the side chain of the cysteine residues, and the long correlation times reflect the overall protein motions. The .2 values suggest that both proteins were not highly asymmetric at neutral pH. In the presence of Ca2+, at 20°C and neutral pH the rotational correlation times were (1) STnC, .1 ~ 2.1 ns, .2 1 13.9 ns, and (2) CTnC, .1 ~ 1.9 ns, .2 1 13.0 ns. A decrease of pH from 7.2 to 5.2 led to an increase of .2 ~ 20% for STnC and a reduction of .2 ~ 7% for CTnC. The two proteins appear to have different hydrodynamic properties in an acidic environment.

The affinity of Saccharomyces cerevisiae fructose-1,6-bisphosphate aldolase towards the metabolically related enzymes phosphofructokinase (PFK), triosephosphate isomerase (TPI) and glyceraldehyde-3-phosphate dehydrogenase (GPDH) was tested by using the signal of fluorescein isothiocyanate (FITC) attached covalently to the aldolase. The dissociation constants of the enzyme-enzyme complexes and rate constants of their formation and dissociation were measured and compared with the same parameters derived for enzymes from rabbit muscle. Hybrid complex formation between yeast aldolase and muscle GPDH and between muscle aldolase and yeast GPDH have also been observed. From the similarities in the determined parameters for the yeast and muscle enzymes we concluded that organization based on direct enzyme-enzyme interactions may be an ancient characteristic of the cytoplasm. The existence of in vitro hybrid complexes indicates that the recognition sites responsible for these interactions may have been conserved during the evolution.

Several well-characterized conformational changes in two proteins, bovine serum albumin and yeast enolase, were investigated using steady-state quenching and dynamic fluorescence measurements. These results were compared with published observations. Conformational changes involving domain or subunit separation are associated with increased Stern-Volmer quenching constants but no consistent change in emission maximum or average tryptophanyl-fluorescence lifetime. Rotational correlation times from tryptophanyl-fluorescence anisotropy decay are in agreement with expected values, showing the value of such measurements in analysis of conformational changes in proteins.

The fluorescence decay of sn-2 diphenylhexatrienylpropionyl- and parinaroyl-labeled choline glycerophospholipids as well as trimethylammonium-DPH (TMA-DPH) were determined by multifrequency phase fluorometry in single bilayer vesicles of diacylglycerophosphocholine (phosphatidylcholine) and its ether analog alkenylacylglycerophosphocholine (choline plasmalogen). The data were analyzed in terms of continuous lifetime distributions using Lorentzian distribution functions. Addition of cholesterol to the bilayers induced narrowing of the lifetime distributions for all labels and unlabeled matrix phospholipids tested. DPH propionyl phospholipids discriminated between ether (plasmalogen) and ester (phosphatidylcholine) phospholipid membranes in that they experienced a much larger lifetime heterogeneity when embedded in the ether lipid, even in the presence or absence of cholesterol.

We used time-dependent fluorescence energy transfer to determine the distribution of donor-to-acceptor distances in native and denatured troponin I(TnI). The single tryptophan residue (trp 158) of TnI served as the donor (D), and the acceptor (A) was a labeled cysteine residue (cys 133). The time-dependent intensity decays of the donor were measured by the frequency-domain method. The frequency-response of the donor emission, in the absence and presence of acceptor, was used to recover the distribution of D to Å distances, using an algorithm which accounts for the intrinsic multi-exponential decay of the donor. In the native state the D-A distribution is characterized by an average distance of 23 A and a half-width of 12 Å. Denaturation results in a modest increase in the average distance to 27 A, and a dramatic increase in half-width to 47 Å. We believe the ability to recover distance distributions will have numerous applications in the characterization of biological macromolecules.

We used GHz frequency-domain fluorometry to examine the tryptophan intensity decays of NATA (M-acetyl-L-tryptophanamide), gly-trp-gly, and the single tryptophan proteins ACTH, E. nuclease and ribonuclease T, (Rhase 1%). In all cases the intensity decays became more heterogeneous in the presence of quenching, which we attribute to a time-dependent rate constant for quenching (transient effects). The frequency-domain data were analyzed using the Smoluchowski model (exp(-t/t 2b/T)) and the radiation boundary condition (RBC) model. In contrast to the IT model, the RBC model does not assume the fluorophore-quencher pair is immediately deactivated, but rather assumes a rate constant for deactivation of the pair (K) as well as a mutual diffusion coefficient (D). The RBC model provides dramatically improved fits to the data. The values of both D and decreases progressively in the order listed above, which is with decreasing exposure to the aqueous phase. Because the RBC model may not be strictly correct in homogeneous solution, and probably less so in the hindered anisotropic environment of the proteins, the recovered values of D and IC should be regarded as apparent values. The recovered intensity decays can be compared with molecular dynamic calculations of quencher trajectories in proteins.

We use frequency-domain fluorometry to determine the anisotropy decays of the tryptophan emission from S. nuclease and from the model compound gly-trp-gly. Resolution of the rapid and complex anisotropy decays was enhanced by global analysis of the data measured in the presence of quenching by either oxygen or acrylamide. Data were obtained at four to six quencher concentrations, and the data analyzed globally to recover the anisotropy decay. Because the decay times were decreased by quenching, measurements were possible to a upper frequency limit of 2 GHz. The anisotropy decay of gly-trp-gly revealed 40 ps of the indole ring, which was resolved from the overall 150 ps correlation times of the tripeptide. The anisotropy decay of nuclease displayed a 90/ps component as well as a 10 ns component due to overall rotational diffusion. We believe these highly resolved anisotropy decays are suitably for comparison with molecular dynamic simulations.

The fluorescence intensity decay of many tryptophan and tyrosine analogues and single tryptophan- or tyrosine-containing peptides and proteins exhibits complex kinetic behavior. Many kinetic schemes have been proposed, invoking either ground-state or excited-state interactions, but the multi-exponentional decays have yet to be explained by a comprehensive model. Our approach has been to try to resolve ground-state from excited-state interactions. We have recently demonstrated that the complex fluorescence decay of tyrosine in model compounds and small peptides can be explained by a ground-state model which predicts a single exponential decay for each rotamer about the Ca-0 bond of the aromatic residue.1'2 This model also predicts that the interconversion between rotamers is much slower than the decay of the excited state of the aromatic side chain. Consequently, the weighting of the expo-nential components should correlate with the ground-state rotamer populations, which can be estimated from proton NMR results. To incorporate the ground-state rotamer populations directly into the analysis of the fluorescence decay data, we developed a linked-function analysis3 which restricts the parameter search space; as a result, proposed kinetic mechanisms can be more rigorously tested. The linked-function can be applied either to the analysis of a single decay curve or to a global analysis4 of a set of related decay curves.

The cytoskeletal architecture of a cell controls many cell processes and characteristics (cell shape, motility, endocytosis, cell division, organelle position and movement). Many of these processes involve the assembly and disassembly of cytoskeletal elements, but the highly cross-linked polymer system must, of necessity, possess flexibility and motional freedom. Components of the erythrocyte cytoskeleton (actin, spectrin and band 4.1) may be reconstituted into a ternary complex which forms a viscous cross-linked gel. It is unlikely that this structure is identical to that existing in vivo, however, it does provide a convenient experimental model system in which the rotational motion of the individual components may be studied. We have examined this system using time-resolved phosphorescence anisotropy which measures rotational diffusion in the microsecond to millisecond time window.

We describe steady state emission spectra and fluorescent decay times from highly purified solutions of human hemoglobin, and the precautions necessary to obtain reliable measurements on the highly quenched intrinsic tryptophan emission. Hemoglobin displays a wide range of decay times (10 ps to 8 ns). The ps component are probably due to Hb itself, and the ns components are probably due to impurities. The ps emission appears to have emission maxima near 320 nm, which is characteristic of tryptophan residues which are shielded from contact with water. The ns components appear to have emission maxima near 340 nm. Our results indicate that the intrinsic emission of Hb can be a useful probe for its functional behavior.

Two methods of analysis have been used to determine the lifetimes of human blood serum fluorescence by our fast analog technique. The first method models the serum fluorescence with up to two components, with lifetimes and pre-exponential coefficients of the multi-exponential decay determined by the Marquardt algorithm. Although this model provides a good fit, it is evident that not all components are discerned. The second method assumes an underlying distribution of lifetimes. We have applied James and Ware's series exponential method to the fluorescence decay of anthracene (discrete lifetime), a fluorescence probe [p-(N,N-dialkylamino)benzylidene]malononitrile in polymethyl methacrylate, and human serum (distribution of lifetimes). We found for our experimental technique that the percentage contributions (rather than the pre-exponential coefficients) of each lifetime to the total intensity should be set equal at the beginning of the iterations. Up to 22 lifetimes were used to fit the fluorescence data. This method was applied to a fluorescence assay to develop a calibration curve when a background signal was significant.

The fluorescence decay of tryptophan is complex. Various models have been proposed to account for the multiexponential decay, including ground-state rotamers, excited-state proton transfer, and lifetime distributions.^1-4 In order to test the rotamer model, we are studying a series of tryptophan derivatives in which rotation about the Ca-CB and CB-Cy, bonds is constrained. The ground-state conformations are determined by X-ray diffraction, molecular and 1H NMR. Or preliminary results for a tryptophan derivative with the a-amino group incorporated in a six-member ring are described below.

We report on new approaches to picosecond fluorescence spectroscopy, using commercial silicon avalanche photodiodes (APD's) as opto-electronic cross-correlators. By integrating lock-in techniques, fluorescence radiation can be detected with 0.001 photons per pulse sensitivity. In our method, the internal gain of the APD is modulated in synchronism with a syncpumped cw dye laser. Using pulsed gain-modulation, time-domain measurements with an apparatus response FWHM of 150 ps can be accomplished. In a second mode, sinusoidal gain-modulation is employed, enabling frequency-domain measurements up to about 3000 MHz at multiples of the laser pulse repetition frequency. This mode also allows high-resolution lifetime-selective detection and fluorescence lifetime direct indication.

Fluorescence anisotropies for single tryptophan-containing polypeptide hormones ACTH and glucagon, and a series of their fragments are studied. The data are discussed in the context of the theory of Perico and Guenza (J. Chem. Phys. 84, 510 (1986)) and a persistence length of about 7 to 10 residues obtained for the mobilities in the two hormones. Theory is able to account for chain length and probe location effects very well. The observed discrepencies between calculated and measured anisotropy decays very likely arise from approximations made in the calculation and from the limited time resolution of the experiments.

The absorption anisotropies of tryptophan (Trp) and some simple peptides are reported with a time resolution of 1 ps. The results are compared with those from fluoresence anisotropy, molecular dynamics simulations and Debye-Stokes-Einstein hydrodynamics. The rotational correlation time of tryptophan in aqueous solution (pH 7.0, 20°C) is 35 ps, obtained using time-resolved absorption anisotropy, which agrees with previous fluorescence anisotropy data. However, the rotational correlation time of Trp obtained from molecular dynamics simulations is three times faster than the experimental value, which indicates that the force field used in the simulation requires refinement. The data also show that polar interactions retard the reorientational motion significantly, which is qualitatively consistent with the results obtained from Debye-Stokes-Einstein hydrodynamic theory. Within the time resolution of the experiment (q,1 ps), all dipeptides show single exponen-tial anisotropy decays with rotational correlation times ranging from 40 to 60 ps. This suggests that the low initial value of the fluorescence anisotropy is attributable either to vibronic coupling or to one or more ultrafast processes in the excited state.

Our studies with quinacrine and benzo(a)pyrene suggest the formation of a multiorganelle detoxification complex (MODC) involving the endoplasmic reticulum, the Golgi apparatus, the lysosomes and the nuclear membrane. We have indications that not only is there a trapping of xenobiotics in the cytoplasmic components of the MODC, but there may also be a "nuclear pump" powered by postulated nuclear bioenergetic pathways involved in the ejection of these chemicals from the nucleus. We are using the microspectrofluorometric approach to study the extranuclear/nuclear energy metabolism related to these processes, and we have extended this method to investigate, in situ, within different components of the MODC, the blue/red spectral shifts associated with metabolites of fluorescent xenobiotics. Recording of fluorescence emission spectra, at different excitation wavelengths in L cells, allows the application of multivariate statistical methods to analyze complex (multicomponent spectra). Eluciadation of mechanisms, involved in the organization and activity of the MODC, can result in better targeting of gene modifiers and DNA-intercalating cancer chemotherapeutics towards their expected sites of action.

A new technique for the visualization of small molecular components in living cells is presented. It is based on the exploitation, by means of video techniques of the slight diffraction power of small metal particles (20-40 nm) in the light microscope. We will show that colloidal gold probes, coupled to selective antibodies or charge modulating molecules allow to follow and quantify dynamic processes in endocytosis, Saltatory Motion and membrane mobility.

Recent advances in fluorescent probe technology and image processing equipment have made available the measurement of calcium in living systems on a real-time basis. We present the use of the calcium indicator Fura-2 in intact normally stimulated rat heart cells for the spatial and dynamic measurement of the calcium excitation profile. After electric stimulation (1 Hz), the activation proceeds from the center of the myocyte toward the periphery. Within two frame times (80 ms), the whole cell is activated. The activation is slightly faster in the center of the cell than in the periphery. The mean recovery time is 200-400 ms. There is no difference along the cell's long axis. The effect of a beta-agonist and of a calcium antagonist is described.

The interactions of calf thymus DNA with the chemotherapeutic drug cis-diamminedichloroplatinum (II), and with the clinically ineffective trans-isomer, have been studied by time-dependent fluorescence depolarization of intercalated ethidium. The effect of binding these compounds on the flexibility of DNA has been determined by monitoring the rapid internal torsional and bending motions of the DNA via depolarization of the ethidium fluorescence. The depolarization data are analyzed with an elastic model of DNA dynamics and yield an estimate of the torsional rigidity of DNA. Binding of the cis-isomer to DNA has a pronounced effect on the torsional rigidity, causing increased rigidity at low binding levels and decreased rigidity at high levels. These results are discussed in terms of intrastrand cross-link formation and local melting of DNA duplex-structure. The torsional rigidity of DNA is unaffected by binding of the trans-isomer. The possible relevance of these alterations of DNA flexibility to the selective drug action of the cis-isomer is considered.

By use of steady state and time-resolved fluorescence techniques Forster energy transfer between the cofactors, FAD and FMN, in cytochrome-P450 reductase is observed. Red-edge spectroscopy allowed us to assign the actual rate constant of transfer as apparent in the anisotropy decay of the holoenzyme. The same approach was used in biflavinyl compounds entrapped in polymethylmetacrylate. As anticipated the fluorescence depolarisation of this system showed a great similarity with the enzyme system. From the results the interflavin distance in the enzyme was estimated between 1.6 and 2.4 nm.

Fluorescence emission anisotropy is a well established tool for measuring order in lipid bilayers. Coronene, a fluorescent membrane probe, with a mean fluorescence lifetime greater than 200ns, is sensitive to lipid chain disordering events that occur well after the decay of most other fluorescent probes. Further, its D6h planar symmetry provides exclusive detection of out-of-plane rotations. We have previously employed a compartmentalized gel-fluid equilibrium model to explain the time course of polarization for coronene in dimyristoylphosphatidylcholine (DMPC) and dipalmitoylphosphatidylcholine (DPPC) unilamellar vesicles (ULVs). We present here a more appropriate and complete model derived from Landau phase transition theory. A 'gating factor' is employed, which defines the number of lipid chains close to a coronene molecule which must disorder to allow the rotation of coronene. This model predicts a distribution of rotational correlation times (0i) that change with temperature. Our results indicate excellent agreement between theoretical anisotropy decays and data taken for coronene labelled DPPC large unilamellar vesicles (LUVs) measured at several temperatures up to the lipid melt transition temperature (Tc). The model can account for depolarization both at long times and in the first few nanoseconds of decay.

A microscope system has been constructed that enables digital imaging of a fluorescent cell under pulsed illumination. Each image is produced by a single laser pulse of duration less than 0.3 11 s. With this system, microsecond responses of a single cell to an externally applied electric field have been resolved temporally and spatially. The cell membrane was stained with a voltage-sensitive fluorescent dye. The induction of transsmembrane potential by the applied field, and the perforation (electroporation) of the cell membrane under an intense field, were seen as successive images. The major finding was a transient increase, at the moment of perforation, in the membrane permeability to an enormous level in localized regions of the cell membrane. Possible roles in cell technology, as well as other applications of the microscope system, are discussed.

We demonstrate that distance distributions between two sites on flexible molecules can be recovered using steady state measurements of fluorescence energy transfer. The characteristic Forster distance (R_°) for energy transfer was varied by collisional quenching of the donor, which decreases its quantum yield. The measured transfer efficiencies for each value of R. provide a different average of the distance distribution. The R.-dependent transfer efficiencies, when analyzed by non linear least- squares, were found to reliably determine the distance distribution, as shown by agreement with the frequency-domain method (Chem. Phys. Lett. 1987, 138:587-593).

We measured the frequency-response of the polarized emission of Yt-base in propylene glycol at 10°C. Data were obtained for excitation wavelengths of 290, 312 and 346 nm, for which the fundamental anisotropies are 0.05, 0.19 and 0.32, respectively. Additionally, data were obtained using CC14, to decrease the mean decay time from 9.1 to 4.2 ns. These data were analyzed globally to recover the anisotropy decay law. Three correlation times were needed to fit the data, 0.8, 3.0 and 5.6 ns, a range of only 7-fold. We believe this is the first reported detection of three correlation times for a rigid molecule.

We report measurements of time-resolved emission spectra of N-acetyl-L-tryptophanamide (NATA), adrenocorticotropic hormone (ACTH, residues 1-24), and of S. Nuclease. These spectra were calculated from the frequency-response of the emission, measured at several wavelengths across the emission spectra. Measurements were performed on samples not quenched and quenched by acrylamide, the latter providing additional information on the short time events. The time-resolved center-of-gravity does not decay as a single exponential. At least two spectral relaxation times are needed to account for the present data. NATA and ACTH each display relaxation times near 50 and 800 ps, which may be characteristic of exposed tryptophan residues. S. nuclease displayed slower relaxation times near 0.5 and 10 ns, which probably reflect the dynamic protein matrix which surrounds the residue.

A novel cholesterol-analogue probe1,2 with a diene-(2-naphthyl) fluorophore in the sidechain (Figure 1), hereafter referred to as DN-Chol, has had its steady-state and time-resolved fluorescence properties characterized in solvents and in various viscosity mineral oils.

The azurins are a family of homologous blue copper proteins which function as bacterial electron transferases (1). These proteins have been actively studied owing to several of their unusual and remarkable spectroscopic and chemical properties (2). In this study, homologous azurins from Pseudomonas fluorescens (ATCC 13525) and Pseudomonas aeruginosa (ATCC 10145) were purified and examined by a number of electrophoretic techniques and their copper:protein stoichiometry determined by atomic absorption and amino acid analysis. Provided that the spectral ratios (A620/A280) were above 0.50 and no evidence of a Soret band in the absorption spectrum existed, results showed that there was no contamination of these blue-copper proteins with either cytochrome or a "copper-less" apoazurin. Upon isoelectric focusing for example, apoazurin clearly migrated to a more acidic position relative to the holoazurin and only when the spectral ratio was below 0.50 could an extra band, co-migrating with apoazurin, be detected in holoazurin samples. Collectively the above results not only provided convincing evidence of protein homogeneity but also established, for the first time, definite criteria by which to judge azurin homogeneity.

We report on the development of a laser scanning fluorescence microscope possessing several features which facilitate its application to biological and biophysical analyses in living cells. It is built around a standard inverted microscope stand, enabling the use of standard optics, micromanipulation apparatus, and conventional (including video) microscopy in conjunction with laser scanning. The beam is scanned across the specimen by a pair of galvanometer-mounted mirrors, driven by a programmable controller which can operate in three modes: full raster scan, region of interest, and random-access. A full 512x512 pixel image can be acquired in one second. In region of interest mode, several subareas of the field can be selected for more rapid or detailed analysis. For those cases where the time scale of the observed phenomenon precludes full-field imaging, or where a full-field image is unnecessary, the random access mode enables an arbitrary pattern of isolated points to be selected and rapidly sequenced through. Via a graphical user interface implemented on the system's host computer, a user will be able to take a scout image either with video or a full-field laser scan, select regions or points on the scout image with a mouse, and set up experimental parameters such as detector integration times with a window-style menu. The instrument is designed to be a flexible testbed for investigating new techniques, without compromising its utility as a tool for biological research.

The discrimination of out of focus contributions in fluorescence microscopy possible in a confocal setup will establish itself as a supplement to conventional fluorescence microscopy. The improvement of the contrast compared with conventional fluorescence microscopy depends mainly on the density of the fluorescing material and the thickness of the sample. The term thickness, that which microscopists refer to as the size of the specimen along the optical axis, will gain a new quality since a confocal fluorescence microscope may reveal totally different features when recording data in planes that are 0.3μm apart. Differences that have in the past been neglected suddenly become important. The following article will outline important features in the application of confocal fluorescence microscopy in the biological sciences, point out its limitatk'ns, and draw attention to expected developments.

Any one of the spectroscopic properties of a fluorescent dye (excitation spectrum, emission spectrum, fluorescence lifetime, or quantum yield) could be used to report on its environment. However, pure fluorescence properties such as quantum yield or absolute spectral properties, while sensitive, require complex measurements and calibration procedures for their evaluation. Instead, it is much more common to use fluorescence flux over some part or all of the emission wavelength band as a monitor of a combination of effects including the composite information monitored by the absorption spectrum, the quantum yield and the emission and excitation spectra. The simplest methods use some ratiometric measurement monitoring either a separate reference dye or the signals at two or more wavelengths to read-out the disproportionation of fluorescent species on a relative scale, but other more complex approaches can be successful including spectral shape analysis and the use of the kinetics of photochemical changes. With the appropriate fluorophore, the fluorescence approach has the advantage that it allows both temporal and spatial resolution on the microscopic scale. However, in practice microspectrofluorometry has been difficult to use for quantitative analysis due to optical, physical and chemical problems inherent to microscopic samples and microspectrofluorometric systems. With the advent of very sensitive image detector systems, computer control and data acquisition, and the advanced optical components described in these proceedings, it has become possible to overcome many of these problems. This paper discusses the measurement strategies needed to effectively use advanced microspectrofluorometer systems to make quantitative measurements using fluorescent probes.

Algorithms for geometric correction of digital images, for registration of images acquired at different wavelengths and for sample repositioning after removal from the microscope stages are described. These image processing procedures increase significantly the analytic power of digital imaging fluorescence microscopy and the usefulness of immunocytochemistry.

We have used digital video microscopy to study the relationship of intracellular calcium ion concentration ([Ca2+]i) to the function of living cardiac and vascular smooth muscle cells. The technical goal of our work is to obtain, with high spatial and temporal resolution, "maps" of [Ca2+]i inside single living cells. To relate [Ca2+]i to cell function, such "maps" can be used in conjunction with measurements of cell electrical activity, contractile activity or biochemical assays.

An optically sectioning ocular fluorometer microscope is described with the capability of measuring the emission spectra of molecules in planes along the microscope axis. Its unique feature is that the objective is attached to a piezoelectric driver and scans from the tear film to the aqueous humor. This permits measurements on living animals and adoption for clinical use. The excitation light from a laser (nitrogen, dye, argon or helium cadmium) couples to the microscope via a quartz optical fiber. The light is projected through a 100 PM slit on the excitation side, through one half of the objective. The emitted light is collected by the second half of the objective and passes a second 100 pm slit in the conjugate plane of the eyepiece. The depth resolution is 6 um with an 100x objective, and 18 PM with a 50 power objective. The fluorescence is coupled by a quartz fiber to an optical spectrum analyzer. It consists of a monochromator with two microchannel plates attached to a linear diode array. The photocathode of the detector is gated for use with pulsed lasers or it can operate in the continuous mode. The applications include fluorescence measurements on thin layered structures. The present study involves the noninvasive measurement of oxidative metabolism of the component layers of the in vivo cornea. This is based on fluorescence measurements of the reduced pyridine nucleotide in the cornea. The fluorescence signals from the corneal epithelial (30 μm) and endothelial (4 μm) are clearly defined. Other applications to ophthalmology include studies of the fluorescence form the component layers of the ocular lens. Support from N.I.I. EY06958.

A time-resolved microfluorimeter with a synchroscan streak camera, which can be extended to a fluorescence lifetime imaging microscope (FLIM), has been developed for the use in biomedical research and clinical diagnosis. The schematic diagram of the instrument is given in Figure 1. The key idea for this instrument is to acquire molecular information on the structure of interest observed under the fluorescence microscope by using time-resolved fluorescence spectroscopy. The fluorescence lifetime is probably the only physical observable, the absolute value of which can be reliably determined in fluorescence microscopy.

Fluorescence Correlation Spectroscopy (FCS) and Fluorescence Photobleaching Recovery (FPR) are closely related technically but have an important difference in principle. The former extracts information from measurements of spontaneous concentration fluctuations, the latter, from measurements of the relaxation of a concentration gradient produced by a macroscopic perturbation. By monitoring fluorescence changes in an open region of the reaction system both methods can determine rates of molecular transport. Because of the high molecular specificity possible with fluorescence labels, the high spatial resolution attainable with laser-microscope excitation, and the relatively nonperturbing nature of the measurements, these methods have many potential uses. Although the two methods are fundamentally equivalent, FCS is simpler in concept and in analysis, but can be applied only to relatively stable systems. Typically the open observation region is defined by a laser beam which excites the measured fluorescence and is focused and detected using confocal microscope optics. Most previous applications have been to two-dimensional membrane systems. Although a number of interesting measurements have also been carried out on three-dimensional bulk solution systems, additional complications arise due to the variation in intensity and radius of the focused laser beam along the optic axis (z-axis). Because of the former the motion of particles in the z direction can be detected. Because of the latter the characteristic diffusion time, defined by the beam radius, differs for particles at different distances from the focal plane. Usually the latter is the dominant effect as confirmed both by theory and experimental measurements. An additional but essential complication arises from placing a field diaphragm before the photodetector to reduce the off-focus background signal. A natural extension of FCS is Fluorescence Distribution Sepctroscopy (FDS) which provides information about the distribution of aggregate sizes in systems of fluorescent molecules. This information is obtained from an analysis of the distribution of fluorescence photocounts emitted in a series of defined time intervals. The nonperturbing nature of FDS makes it especially useful for the analysis of reversible aggregation processes often found in biological systems. FDS measurements in three-dimensional systems are especially sensitive to variations in beam intensity and radius along the z direction.

Multifrequency phase-resolved fluorescence spectroscopy is used to generate a multidimensional data array in which phase-resolved fluorescence intensity is measured as a function of emission wavelength and excitation wavelength at a series of excitation modulation frequencies. The dependence of the phase-resolved fluorescence intensity on fluorescence lifetime, modulation frequency and detector phase angle setting is examined. A model two-component system of benzo(k)fluoranthene and benzo(b)fluoranthene has been used to demonsrate the ability of phase-resolved fluorescence spectroscopy to selectively enhance the spectral contribution of a component as a function of fluorescence lifetime in total luminescence spectroscopyl, and is also discussed here.

Commercial high-resolution phase and modulation fluorometers allow real time resolution of excitation and emission spectra of fluorophores with different lifetimes in a mixture. The technique can also be used for the separation of fluorescence and phosphorescence spectra of the fluorophore.

We describe an apparatus for the measurement of slow rotational diffusion based on the phosphorescent photoluminescence of the triplet probe, erythrosin, and using phase modulation techniques. The apparatus has two novel features over those previously described: up to 40 frequencies of modulation are collected simultaneously, and the superimposition of fast 0-100% modulation of the excitation laser beam enables contaminating prompt fluorescence to be removed by time domain resolution during the frequency domain resolution of phosphorescence.

Analysis of time-resolved fluorescence anisotropy measurements on TMA-DPH molecules in POPC vesicles in terms of the rotational diffusion model reveals two distinct X²r minima which are statistically equivalent. The ambiguity in the interpretation is resolved by carrying out angle-resolved fluorescence depolarization experiments on planar POPC multibilayers in the time domain. It is shown that the order parameters of the probe molecules are higher in the multibilayers than in the vesicles. Furthermore the reorientational motion of the probes is considerably slower in the multibilayer sample. These differences are ascribed to the effects of the curvature and hydration of the bilayers in the vesicle systems.