8506
J. Phys. Chem. B 2006, 110, 8506-8512
Generalized Two-Dimensional Heterocorrelation Analysis of Spectrally Resolved and Temporally Resolved Fluorescence of the 8-Anilino-1-naphthalenesulfonate-Apomyoglobin Complex with pH Perturbation Gufeng Wang, Yi Gao, and M. Lei Geng* Department of Chemistry, the Optical Science and Technology Center, and the Center for Biocatalysis and Bioprocessing, UniVersity of Iowa, Iowa City, Iowa 52242 ReceiVed: September 28, 2005; In Final Form: January 12, 2006
We demonstrate two-dimensional heterocorrelation analysis between spectrally resolved and temporally resolved fluorescence to investigate the decay dynamics of the 8-anilino-1-naphthalenesulfonate- (ANS-) apomyoglobin complex. The dynamic changes of the lifetime components are disclosed across the emission spectrum with an external pH-perturbation. Two different fluorescence lifetime schemes of the ANS-apomyoglobin complex are revealed. From pH 8.5 to 4.5, the transition of protein conformation from the native state to the folding intermediate, a short lifetime component is found to correlate with a short-wavelength emission whose population diminishes with decreasing pH. The lifetime components reflect the excited-state populations of the nascent and the charge-transfer species. From pH 4.2 to 1.0, the transition from the folding intermediate to the acid-unfolded state, the short lifetime is responsible for a long-wavelength emission and the fraction of this component increases when the solution becomes more acidic. In this pH range, the decay components reflect the ground-state populations of microenvironments. The relative decay dynamics across the emission spectrum are revealed without collecting decays at each wavelength. More importantly, these conclusions are reached without the necessity of statistical fitting of the decay data with an a priori decay model.
Introduction Two-dimensional correlation spectroscopy is a recently developed technique to enhance the spectral resolution, to correlate spectral bands and to visualize dynamic changes in spectral intensity.1-14 Two-dimensional correlation spectra are obtained by evaluating correlation between spectral bands according to their dynamic changes upon an external perturbation. Through dynamic analysis, 2D correlation provides enhanced spectral resolution, assignment of spectral bands or chemical species coupled by an interaction mechanism, identification of sequential orders of dynamic events caused by external perturbations, and detection of “spectroscopically invisible” species in a mass-balanced system. Because of these reasons, 2D correlation spectroscopy has been applied to various forms of spectroscopy and used in the studies of complex reaction kinetics. Of particular interest is the heterospectral 2D correlation spectroscopy, which extends the idea of dynamic correlation between spectral bands from the same type to two different types of spectroscopy.4-7,15-22 The combination of two simultaneous but separate measurements of a molecular system under a dynamic perturbation provides a multidimensional view of the dynamic changes in molecular structure. For example, correlation between dynamic circular dichroism and Raman spectra reveals the formation of the secondary structure of protein concurrent with dynamic changes in the backbone.16 A powerful application of hetero-2D-correlation spectroscopy is to understand not-so-well-characterized spectroscopic features in one * Corresponding author. Department of Chemistry, the Optical Science and Technology Center, and the Center for Biocatalysis and Bioprocessing, University of Iowa. Telephone: (319)335-3167. Fax: (319)335-1270. E-mail:
[email protected].
type of spectroscopy through dynamic spectra in another wellcharacterized spectroscopy, thus providing additional information about the system under investigation. Other advantages of hetero-2D-correlation spectroscopy include sequencing dynamic changes and enhancing spectral resolution. For example, Barton correlated mid-IR and near-IR spectroscopy to study the combination and overtone bands in the NIR region for complex samples.4 Jung used hetero-2D IR and Raman correlation to assign vibrational transitions in amide III region of β-lactoglobin (BLG).6 The sequential order of secondary structure changes induced by protein association was revealed. Choi observed temporal order of events in the electrochemical reaction of lithium with CoO through the heterocorrelation between soft X-ray absorption and Raman spectroscopy.7 In other applications, band assignments and event sequencing were achieved through 2D heterocorrelation between FT Raman and NIR,18 visible and NIR,21 attenuated total reflection/infrared spectroscopy and NIR,20 and IR and electron paramagnetic resonance.19 In this study, we report hetero-2D-correlation analysis between steady-state (spectrally resolved) and time-resolved fluorescence of the 8-anilino-1-naphthalenesulfonate (ANS)apomyoglobin complex to probe dynamics of protein unfolding. Steady-state fluorescence spectra are easily collected for biological molecules and are well characterized. Time-resolved fluorescence data, although containing additional information about the biomolecules, can be quite difficult to collect due to the propensity of photodegradation in biological system, especially when decay curves are collected at each wavelength. In addition, it is well documented that the recovery of decay kinetics from fluorescence decay data is fundamentally illconditioned and poses a difficult problem in interpretation of the dynamics of complex molecular systems.23-28 The aim of
10.1021/jp0555293 CCC: $33.50 © 2006 American Chemical Society Published on Web 04/01/2006
Two-Dimensional Heterocorrelation Analysis this paper is, by taking advantage of hetero-2D-correlation technique, to use the well-characterized spectrally resolved fluorescence to help understand the time-resolved fluorescence of ANS-apomyoglobin complex and, moreover, the dynamic changes of fluorescence lifetime components of ANS-apomyoglobin during the unfolding process. With the correlation of two types of fluorescence spectroscopy: (1) The relative decay dynamics across the emission spectrum are sequenced through the heterocorrelation analysis without measuring the fluorescence lifetime at each wavelength. (2) The dynamic changes of the lifetime components are disclosed across the emission spectrum with an external pH-perturbation, which tunes apomyoglobin conformation from the native state to the unfolded state. Two different fluorescence lifetime schemes of ANS-apomyoglobin complex are revealed in the protein conformation range. From pH 8.5 to 4.5, corresponding to the conformational transition from the native state (N) of apomyoglobin to the folding intermediate state I-1, a short-lifetime component correlates to a short-wavelength emission and the fraction of this component diminishes with the decreasing pH. To our best knowledge, this is the first report on this unusual dynamic change of the short lifetime component of ANS-apomyoglobin. From pH 4.2 to 1.0, corresponding to a conformational transition from I-1 to the acid unfolded state UA, a long lifetime is responsible for the short-wavelength region of emission and its fraction decreases when the solution becomes more acidic. No statistical model fitting is required to obtain these conclusions. The motivation to develop two-dimensional heterocorrelation technique between spectrally and time-resolved fluorescence in studying ANS-apomyoglobin dynamics is stated below. It is generally acknowledged that fluorescence lifetime contains a wealth of information on structure and dynamics of biological systems. It reveals not only the averaged microenvironmental properties around the probe but also the molecular populational distributions of the probed system. Because of this, fluorescence lifetime is often used to probe molecular structures and intermolecular interactions. However, the fluorescence decay of a host-guest system is usually very complicated because of the microheterogeneity of probe microenvironment and the complex relaxation of the microenvironment or the probe. Within experimental errors, the obtained fluorescence decay can usually be fitted with models varying from multiple discrete decays to multiple continuous distributions.23-28 Ample literature discussions have shown that the extracted decay models may have no physical meanings.23-28 To extract dynamic information from fluorescence lifetime data, the correct decay model must be derived first. However, without sufficient prior physical/ chemical information about the system, it could be very challenging to select the correct model from a number of statistically equivalent ones. Thus, extracting physical information directly from the original data without assuming decay models is especially desired in the fluorescence lifetime data analysis. Further more, in statistical analysis of fluorescence lifetime data, the mathematically feasible decay models often need to be validated by other spectroscopic techniques. In addition, to understand the complicated fluorescence lifetimes in multicomponent or photorelaxation systems, fluorescence decays sometimes need to be monitored at many wavelengths across the emission spectrum. These issues mandate lengthy, time-consuming data collection and difficult statistical analysis of experimental data in fluorescence lifetime studies. Here we develop 2D heterocorrelation analysis between spectrally resolved and temporally resolved fluorescence to extract dynamic information. The method is used to probe the
J. Phys. Chem. B, Vol. 110, No. 16, 2006 8507 excited-state relaxation of ANS probe in the ANS-apomyoglobin complex and the dynamic conformational changes of apomyoglobin during pH unfolding. Materials and Methods Horse skeletal apomyoglobin was obtained from Sigma (St Louis, MO) and used directly or after purification. The purchased lyophilized apomyoglobin was dissolved in pH 6.5 citric buffer and then purified either with a 0.22 µm filter, or with a Sephadex 25 column to remove the denatured aggregates. The preparation of apomyoglobin from myoglobin followed the procedures of Teale.29 The concentration of apomyoglobin was determined by UV absorbance in 6.0 M guanidine HCl solution, as described by Edelhoch.30 The final protein solution in the measurements contains 20.0 mM citric buffer and 10.0 mM NaCl. The ANS-apomyoglobin complexes were obtained by adding small aliquots of ANS stock solution into a solution of the protein. The dye/protein ratio was 1:1. Temperature was maintained constant at 5.0 ( 0.1 °C using an external bath circulator throughout the experiments. Spectroscopic measurements were made after equilibrium was reached. Several sets of experiments with apomyoglobin from different preparation procedures and the protein concentration ranging from 7.5 to 26.4 µM were conducted. The protein from different preparation procedures yields statistically identical results. All common chemicals were used as received. 1,8-ANS acid was acquired at the highest purity from Aldrich (Milwaukee, WI). The purity was checked using thin-layer chromatography (TLC). Guanidine hydrochloride was obtained from Amresco (Solon, OH). Citric acid and sodium citrate were purchased from Fisher (Fair Lawn, NJ). Buffer solutions were prepared with deionized ultrapure water generated by a Milli-Q system (MilliQ-Plus, MilliPore Corporation, Bedford, MA). All steady-state and frequency-domain data were collected with an SLM 48000MHF spectrofluorometer (Jobin Yvon, Edison, NJ). Multifrequency phase-modulation fluorescence lifetime data were collected using a multiharmonic Fourier transform (MHF) technique.31 A HeCd laser served as the excitation source (Liconix 3225N, Melles Griot, Carlsbad, CA, 325 nm). The emission was collected through a 345 nm longpass filter, or through a bandwidth defined by a monochromator. A solution of 20.0 µM 1,4-bis(5-phenyloxazol-2-yl)benzene (POPOP) in ethanol served as the lifetime reference. The excitation beam was vertically polarized and the emission was polarized at 54.7°. This magic angle polarization eliminates the effects of photoselection.32 The generalized 2D fluorescence correlation is calculated following literature in Matlab.2,13 In heterocorrelation analysis, the area-normalized steady-state spectra were aligned with the phases or modulations of the original frequency-domain lifetime data to form a master matrix for calculation. The spectrally and temporally resolved data were collected at the same excitation wavelength of 325 nm. Two types of cross-correlation were obtained: emission-phase and emission-modulation crosscorrelations. Background Perturbation-based 2D correlation spectroscopy evaluates correlation between spectral bands according to their time-course dynamic changes at corresponding wavelengths.1,11,33 A more generalized form, the generalized 2D correlation spectroscopy, was realized by extending the concept “time” to any physical variable upon which the spectral intensity varies.2 Correspond-
8508 J. Phys. Chem. B, Vol. 110, No. 16, 2006
Wang et al.
ingly, generalized 2D correlation spectroscopy can take arbitrary shape of perturbation waveforms and any perturbation methods. A synchronous correlation function Φ(λ1,λ2) and an asynchronous correlation function Ψ(λ1,λ2) are defined as
Φ(λ1,λ2) + iΨ(λ1,λ2) )
1 πT
∫0∞Y1(ω)Y/2(ω) dω
(1)
where Y1(ω) and Y/2(ω) are the Fourier transforms of the “time”-course dynamic spectra y(λ1,t) and y(λ2,t) at λ1 and λ2, respectively:
Y1(ω) )
∫-∞∞y(λ1,t)e-iωt dt
(2)
Y/2(ω) )
∫-∞∞ y(λ2,t)e+iωt dt
(3)
Practically speaking, Φ(λ1,λ2) and Ψ(λ1,λ2) are calculated through the discrete Hilbert-Noda transform:3
Φ(λ1,λ2) ) Ψ(λ1,λ2) )
k
1
∑yi(λ1)yi(λ2)
k - 1 i)1 1
k
(4)
k
Nijyj(λ2) ∑yi(λ1)∑ j)1
k - 1 i)1
(5)
where k is the total number of data points in the dynamic spectra; Nij is defined by
Nij ) 0 if i ) j, or Nij )
1 if i * j π(j - i)
(6)
The synchronous and asynchronous functions in 2D contour plots are the synchronous spectrum and the asynchronous spectrum, respectively. The synchronous spectrum contains onand off- diagonal peaks. The diagonal peaks are also termed auto-peaks. Cross-peaks located at off-diagonal positions of a synchronous spectrum indicates correlation between two spectral bands, which suggests a synchronous change, in turn, the possible existence of a coupled or related origin of the spectral variations. The asynchronous spectrum only has off-diagonal peaks indicating anticorrelation between two spectral bands. In other words, spectral intensities at the two bands change out of phase (i.e., changing at different rates) with each other, suggesting the two bands may have different origins. The diagonal elements of the asynchronous spectrum are all zeros because the spectral intensity of a species always changes in phase with itself. In generalized hetero correlation spectroscopy, the synchronous and asynchronous spectra are evaluated between spectral bands from different types of spectroscopy. In the spectrally and time-resolved heterocorrelation, the synchronous and asynchronous functions between wavelength λ and modulation frequency ω have the following forms:
Φ(λ,ω) ) Ψ(λ,ω) )
1
1
k
∑I(λ,pH)θ(ω,pH)
k - 1 pH k
∑
k - 1pH(i)
(7)
k
I(λ,pH)
∑ Ni,jθ(ω,pH)
(8)
pH(j)
where I is the steady-state fluorescence emission intensity and
Figure 1. Multifrequency phase-modulation data of ANS complexed with apomyoglobin. From the left to the right, the curves are for ANSapomyoglobin complex at pH 8.5, 7.0, 6.5, 6.0, 5.7, 5.4, 5.1, 4.8, 4.5, 4.2, 3.9, 3.6, 3.3, 3.0, 2.7, 2.4, 2.0, and 1.0, respectively. The apomyoglobin concentration is 20.0 µM and the apomyoglobin/ANS ratio is 1.0. The solution is in 20 mM citrate buffer and in the presence of 10 mM NaCl. The excitation is from a HeCd laser at 325 nm.
θ is one of the two frequency-domain parameters of phase angle φ or demodulation factor m, which are defined as25,26
tan φ ) A/B
(9)
m ) xA2 + B2
(10)
R(τ)ωτ2
A)
∞ dτ ∫τ)0 1 + ω2τ2 ∞ R(τ)τ dτ ∫τ)0
(11)
R(τ)τ
B)
∞ dτ ∫τ)0 1 + ω2τ2
∫τ∞) 0 R(τ)τ dτ
(12)
where R(τ)’s are the preexponential factors of continuously distributed fluorescence lifetime components of τ’s. The fluorescence lifetime information in frequency-domain fluorometry is carried in the excitation modulation frequency ω. The similarity and dissimilarity of dynamic changes of the system upon an external perturbation between wavelength λ and modulation frequency ω can be evaluated from the synchronous and asynchronous spectra. The disclosed correlation between wavelengths and modulation frequencies is equivalent to the correlation between wavelengths and lifetime components. An established correlation between a wavelength and a lifetime component indicates coupled or related origins for the spectral band and the lifetime component. Results and Discussion Steady-state and Time-resolved Fluorescence of ANSapomyoglobin Complex. Figure 1 shows the original frequencydomain fluorescence lifetime data of ANS complexed with apomyoglobin at different pH’s. The frequency-domain data contains two sets of data points collected at multiple modulation frequencies of the excitation: phase delays and demodulation factors (abbreviated as modulation). The phase delay increases while modulation decreases with frequency. From left to right, the curves stand for the ANS-apomyoglobin complex at pH 8.5, 7.0, 6.5, 6.0, 5.7, 5.4, 5.1, 4.8, 4.5, 4.2, 3.9, 3.6, 3.3, 3.0,
Two-Dimensional Heterocorrelation Analysis
J. Phys. Chem. B, Vol. 110, No. 16, 2006 8509 frequency-domain data, phase φ and modulation m, are related to the fluorescence lifetime τ:37
Figure 2. Fluorescence emission of the ANS-apomyoglobin complex. (A) Area-normalized emission spectra. From the left to the right, the curves stand for ANS-apomyoglobin complex at pH 8.5, 7.0, 6.5, 6.0, 5.7, 5.4, 5.1, 4.8, 4.5, 4.2, 3.9, 3.6, 3.3, 3.0, 2.7, 2.4, 2.0, and 1.0, respectively. (B) Emission maximum. Experimental conditions are the same as in Figure 1.
2.7, 2.4, 2.0, and 1.0, respectively. The lifetime data can be fitted with various models with nonlinear least squares (NLLS) analysis. Selection of the correct model is very complicated and will be presented elsewhere.34 Figure 2 shows the area-normalized fluorescence emission spectra and corresponding maxima of the ANS-apomyoglobin complex at the same pH’s. From pH 8.5 to 1.0, there is a continuous red-shift of the ANS emission (Figure 2A). Figure 2B clearly shows that the emission maximum of ANSapomyoglobin changes via a two-step transition as a response to pH change. This is because apomyoglobin unfolds in two steps through an intermediate I-1 in the pH-unfolding process.34-36 From pH 8.5 to 4.2, apomyoglobin unfolds from the native state (N) to I-1, with the disruption of helices BCDE out of a total of 8 helices named A-H, respectively. From pH 4.2 to 1.0, I-1 continuously unfolds to an acid-unfolded state UA, with the complete disruption of the hydrophobic core of the protein composed of helices AGH. To understand the dynamics of the fluorescent probe in the ANS-apomyoglobin complex and the dynamic conformational changes of apomyoglobin during unfolding, we performed the heterocorrelation analysis between the steady-state and timeresolved fluorescence. The 2D heterocorrelation was evaluated in the two transitional regions. Original frequency-domain data are used directly in the analysis. For a single component system containing a single fluorescent species, the two parameters of
tan φ ) ωτ
(13)
m ) 1/x1 + ω2τ2
(14)
The angular modulation frequency ω is equal to 2πf, where f is the linear modulation frequency. For multicomponent systems, the complete decay information needs φ and m to be collected at multiple frequencies. The phase φ and modulation m are related to the lifetime τi’s and the corresponding fractional contribution fi’s through Fourier transforms.37 In lifetime measurements, low modulation frequencies are required to obtain information on long lifetimes because the reciprocal relationship between ω and τ. Similarly, high modulation frequencies are required to obtain information on short lifetimes. In a multifrequency phase-modulation lifetime data set, low frequencies sample more efficiently on long lifetimes, thus are more related to the long lifetime components. Similarly, the high-frequency range bears more information about the short lifetimes, or is more related to the short lifetime components. N to I Transition. The generalized asynchronous heterocorrelation maps of ANS-apomyoglobin with pH-perturbation are displayed in Figures 3 and 4. Figure 3A shows the phaseemission correlation from pH 8.5 to pH 4.5. There are two peaks in Figure 3A. A positive peak centers at high frequencies (short τ) and long wavelengths (480-500 nm), which indicates that the short lifetime anticorrelates with the long-wavelength emission in this pH transition. Similarly, the negative peak in the asynchronous plot, which has an elongated shape but centers at short wavelengths (420-440 nm) and low frequencies (long τ), indicates the complementary information that long-τ component anticorrelates with the short-wavelength emission. Two conclusions can be drawn: (1) since the short-τ component anticorrelates with the long-wavelength emission, and the long-τ component with the short-wavelength emission, the fast-decaying component is responsible for a shortwavelength emission species, and the slow-decaying component corresponds to the long-wavelength emission species for ANSapomyoglobin in this pH range. (2) From pH 8.5 to 4.5, Figure 1 shows that the phase angle drop mainly occurs at high modulation frequencies, indicating a decrease in the fractional contribution of the short-τ component. Figure 2 shows a concurrent intensity decrease at short wavelength (red-shifting of the ANS emission maximum). It can be concluded that the fractional contribution of the short-τ component emitting at short-wavelengths is decreasing in the transition from the native state N to the unfolding intermediate I-1, with a concurrent elevation of the fractional contribution from long-τ and low energy emission. These conclusions are validated in a more comprehensive study.34 The fluorescence decays are measured at 31 wavelengths across the emission spectrum of ANS binding to native apomyoglobin at pH 6.5. The results in these experiments confirm that the averaged decay rates at short wavelengths are faster than those at long wavelengths. Decay-associated spectra show a short-τ component of 6.9 ns at short wavelengths centered at 440 nm, and a long-τ component of 16.6 ns at long wavelengths at 460 nm. The abnormal short-lifetime, shortwavelength emission is suggested to be emitted from a nonplanar excited-state of ANS restricted transiently in the protein media, while the dominant long-lifetime, long-wavelength emission is from the charge-transfer excited-state of ANS. From pH 8.5 to 4.5, there is a diminishing of the short τ component due to the
8510 J. Phys. Chem. B, Vol. 110, No. 16, 2006
Wang et al.
Figure 3. Asynchronous heterocorrelation maps of ANS-apomyoglobin with pH-perturbation. pH Value varied from 8.5 to 4.5. (A) Phaseemission correlation. (B) Modulation-emission correlation.
Figure 4. Asynchronous heterocorrelation maps of ANS-apomyoglobin with pH-perturbation. pH Value varied from 4.2 to 1.0. (A) Phaseemission correlation. (B) Modulation-emission correlation.
loosening of local ANS-binding environment, thus fast relaxation from the nonplanar state to the charge-transfer state. To our knowledge, this is the first observation of the unusual shortlifetime, short-wavelength emission component and the dynamic diminishing of this component during the conformation transition of apomyoglobin from N to I-1. In deriving the above conclusions, the fluorescence lifetimes are studied both with nonlinear least squares (NLLS) and maximum entropy method (MEM), and derived fluorescence lifetime models are validated with other spectroscopic methods.
I to UA Transition. Figure 4A shows the asynchronous phase-emission correlation map of ANS-apomyoglobin from pH 4.2 to 1.0. A completely different fluorescence decay scheme is depicted. There are also two correlation peaks, but at different locations in the map. A positive peak discloses the anticorrelation between long-lifetime component (low frequency) and the long-wavelength emission (460-540 nm). A negative peak discloses complementary anticorrelation information between the short-lifetime component (high frequency) and the shortwavelength emission (420-460 nm). Again, two conclusions
Two-Dimensional Heterocorrelation Analysis are drawn: (1) In the transition from pH 4.2 to 1.0, the longlifetime component are responsible for the short-wavelength emission, and the short-lifetime component should correspond to a long-wavelength emission. Clearly, for ANS-apomyoglobin in this pH range, the decay rates at the long-wavelength end should be faster than at the short-wavelength end. (2) From pH 4.2 to 1.0, Figure 1 shows that the phase angle drop mainly occurs at low modulation frequencies, indicating that the fractional contribution of the large-τ component is decreasing. Figure 2 shows again an intensity decrease at short-wavelength (still red-shifting of ANS-apomyoglobin emission maximum) from the folding intermediate to the unfolded state UA. Combining the spectral- and time-resolved information, it can be concluded that the fraction of the long-lifetime component at short-wavelength is decreasing, resulting in an increase of the short-lifetime, long-wavelength emission. Again these results are validated.34 The fast decay rates at long-wavelength end of the emission spectrum of ANS-apomyoglobin at pH 3.4 are observed directly. Decay-associated spectra disclose a shortlifetime component of ∼4.0 ns at long wavelengths centered at 500 nm, and a long-lifetime component of ∼14.0 ns at short wavelengths of 470 nm. The short-lifetime, long-wavelength emission component is assigned to ANS in a more polar environment, caused by either ANS binding to a newly opened surface binding site, or ANS binding to a population of unfolded apomyoglobin. With a decrease of the pH from 4.2 to 1.0, this fraction of short lifetime is increasing, indicating the increasing of both the extent of unfolding and unfolded fraction of apomyoglobin. The modulation-emission correlation maps in Figures 3B and 4B contain information similar to those from phase-emission correlation maps. The difference is that the maximum anticorrelation is located between the same wavelength region and shifted frequencies. This is the consequence of the different frequency responses of the phase and the modulation (see eqs 13 and 14). Advantages of Heterospectral Correlation. The results above demonstrate that with the assistance of steady-state fluorescence, the fluorescence decay of ANS-apomyoglobin can be much better understood through heterocorrelation analysis between the spectrally resolved and temporally resolved fluorescence. The relative decay dynamics across the emission spectrum is revealed. Although similar information can be obtained through decay-associated spectra (DAS) technique38,39 by measuring the fluorescence decay at each wavelength, there are two advantages of using heterocorrelation analysis: (1) The amount of experiments can be reduced. (2) There is a significant gain in the signal-to-noise ratio measuring the total fluorescence intensity decay than measuring the decays at selected wavelengths. For example, comparing the heterocorrelation analysis and decay-associated spectra, the former requires the measurements of the total fluorescence decay and the steady-state spectrum at each pH, both fast, high SNR measurements, while the latter mandates collecting fluorescence decays at 31 individual wavelengths. Each of these decays demands much longer experiment time compared to the total fluorescence decay, due to the low SNR. Compared with the necessity of monitoring decays at multiple wavelengths, heterocorrelation is clearly a superior methodology because it requires significantly shorter measurement time. This is especially important for biological samples where long laser irradiation can lead to photobleaching and sample degradation. Further, in this work, dynamic changes of the lifetime components of ANS-apomyoglobin are disclosed across the
J. Phys. Chem. B, Vol. 110, No. 16, 2006 8511 emission spectrum during pH unfolding for the first time in the literature. Two completely different fluorescence decay schemes for ANS-apomyoglobin at two different pH ranges are revealed using heterocorrelation analysis. In the pH range 8.5-4.5, a short lifetime component is found to correlate with a shortwavelength emission whose population diminishes with decreasing pH. The lifetime components reflect the excited-state populations of the nascent and the charge-transfer species. From pH 4.2 to 1.0, the short lifetime is responsible for a longwavelength emission and the fraction of this component increases when the solution becomes more acidic. In this pH range, the decay components reflect the ground-state populations of microenvironments. Conclusions In this example of ANS-apomyoglobin perturbed with pH, two completely different schemes depict how fluorescence lifetime varies upon pH changes in pH ranges of 8.5-4.5 and 4.2-1.0. From pH 8.5 to 4.5, the short-τ component corresponds to a short-wavelength emission while from pH 4.2 to 1.0, the short-τ component corresponds to a long-wavelength emission. In both cases, heterocorrelation analysis between spectrally and temporally resolved fluorescence discloses the relative decay rates across the emission spectra and the dynamic changes of the lifetime components. These results are not readily observed with traditional one-dimensional spectra or original fluorescence lifetime data. The relative decay dynamics across the emission spectra are resolved without measuring fluorescence lifetime at each wavelength as in the approach of decay-associated spectra. This results in a significant simplification of experiments and improvement in SNR and data quality. The consequent reduction in experimental time is especially beneficial to biophysical studies and bioimaging, where long laser irradiation leads to photodegradation of the sample and photobleaching of the fluorophores. More importantly, no statistical lifetime model is assumed to obtain these conclusions. This is especially beneficial to dynamic studies of complex molecular systems such as protein-ligand, protein-DNA binding, where it is wellknown that the selection of the correct decay models is very complicated, if possible. Fortunately, the dynamics of these systems can be revealed through recovering the relative decay rates across the emission wavelength region and the dynamic changes in the fractional contributions of the rate components. All the required information is obtained with the heterocorrelation analysis of two simple and high SNR experimentss measurements of the steady-state spectrum and the total fluorescence decay, without model fitting. Acknowledgment. We thank the University of Iowa (CIFRE grant) for supporting this work. The Center of Biocatalysis and Bioprocessing at the University of Iowa is gratefully acknowledged for graduate fellowships to G.W. and Y.G. References and Notes (1) Noda, I. J. Am. Chem. Soc. 1989, 111, 8116-8118. (2) Noda, I. Appl. Spectrosc. 1993, 47, 1329-1336. (3) Noda, I. Appl. Spectrosc. 2000, 54, 994-999. (4) Barton, F. E.; Himmelsbach, D. S.; Duckworth, J. H.; Smith, M. J. Appl. Spectrosc. 1992, 46, 420-429. (5) Wu, Y.; Jiang, J.-H.; Ozaki, Y. J. Phys. Chem. A 2002, 106, 24222429. (6) Jung, Y. M.; Czarnik-Matusewicz, B.; Ozaki, Y. J. Phys. Chem. B 2000, 104, 7812-7817. (7) Choi, H. C.; Jung, Y. M.; Noda, I.; Kim, S. B. J. Phys. Chem. B 2003, 107, 5806-5811. (8) Fabian, H.; Mantsch, H. H.; Schultz, C. P. Proc. Natl. Acad. Soc. U.S.A. 1999, 96, 13153-13158.
8512 J. Phys. Chem. B, Vol. 110, No. 16, 2006 (9) Graff, D. K.; Pastrana-Rios, B.; Venyaminov, Y. S.; Prendergast, F. G. J. Am. Chem. Soc. 1997, 119, 11282-11294. (10) Sasic, S.; Muszynski, A.; Ozaki, Y. Appl. Spectrosc. 2001, 55, 343349. (11) Wang, G.; Geng, L. Anal. Chem. 2000, 72, 4531-4542. (12) He, Y.; Wang, G.; Cox, J.; Geng, L. Anal. Chem. 2001, 73, 23022309. (13) Wang, G.; Geng, L. Anal. Chem. 2005, 77, 22-29. (14) Geng, L.; Cox, J. M.; He, Y. Analyst 2001, 126, 1229-1239. (15) Noda, I. Chemtracts-Macromol. Chem. 1990, 1, 89. (16) Pancoska, P.; Kubelka, J.; Keiderling, T. A. Appl. Spectrosc. 1999, 53, 655-665. (17) Wu, P.; Siesler, H. W. J. Mol. Struct. 2000, 521, 37-47. (18) Ren, Y.; Matsushita, A.; Matsukawa, K.; Inoue, H.; Minami, Y.; Noda, I.; Ozaki, Y. Vib. Spectrosc. 2000, 23, 207-218. (19) Nagai, N.; Yamaguchi, Y.; Saito, R.; Hayashi, S.; Kudo, M. Appl. Spectrosc. 2001, 55, 1207-1213. (20) Amari, T.; Ozaki, Y. Macromol. 2002, 35, 8020-8028. (21) Liu, Y. L.; Chen, Y. R. Appl. Spectrosc. 2000, 54, 1458-1470. (22) Czarnecki, M. A.; Czarnik-Matusewicz, B.; Ozaki, Y.; Iwahashi, M. J. Phys. Chem. A 2000, 104, 4906-4911. (23) James, D. R.; Ware, W. R. Chem. Phys. Lett. 1985, 120, 455459. (24) Siemiarczuk, A.; Wagner, B. D.; Ware, W. R. J. Phys. Chem. 1990, 94, 1661-1666. (25) Alcala, J. R.; Gratton, E.; Prendergast, F. G. Biophys. J. 1987, 51, 587-596.
Wang et al. (26) Lakowicz, J. R.; Cherek, H.; Gryczynski, I.; Joshi, N.; Johnson, M. L. Biophys. Chem. 1987, 28, 35-50. (27) He, Y.; Geng, L. Anal. Chem. 2001, 73, 5564-5575. (28) He, Y.; Geng, L. Anal. Chem. 2002, 74, 1819-1823. (29) Teale, F. W. J. Biochim. Biophys. Acta 1959, 35, 543. (30) Edelhoch, H. Biochemistry 1967, 6, 1948-1954. (31) Jameson, D. M.; Gratton, E.; Hall, R. D. Appl. Spectrosc. ReV. 1984, 20, 55-106. (32) Spencer, R. D.; Weber, G. J. Chem. Phys. 1970, 52, 1654-1663. (33) Noda, I. Appl. Spectrosc. 1990, 44, 550-561. (34) Wang, G.; Gao, Y.; Geng, L. Analysis of heterogeneous fluorescence decays in proteins. Using fluorescence lifetime of 8-anilino-1naphthalenesulfonate to probe apomyoglobin unfolding at equilibrium. Biochim. Biophys. Acta, in press. (35) Hughson, F. M.; Wright, P. E.; Baldwin, R. L. Science 1990, 249, 1544-1548. (36) Sirangelo, I.; Bismuto, E.; Tavassi, S.; Irace, G. Biochim. Biophys. Acta 1998, 1385, 69-77. (37) Lakowicz, J. R. Principles of fluorescence spectroscopy; Kluwer Academic/Plenum Publishers: New York, 1999. (38) Boo, B. H.; Kang, D. J. Phys. Chem. A 2005, 109, 4280-4284. (39) Sakuragi, Y.; Zybailov, B.; Shen, G. Z.; Bryant, D. A.; Golbeck, J. H.; Diner, B. A.; Karygina, I.; Pushkar, Y.; Stehlik, D. J. Biol. Chem. 2005, 280, 12371-12381.
