Role of Protein Stabilizers on the Conformation of the Unfolded State of Cytochrome c and Its Early Folding Kinetics

An insight into the conformation and dynamics of unfolded and early intermediate states of a protein is essential to understand the mechanism of its aggregation and to design potent inhibitor molecules. Fluorescence correlation spectroscopy has been used to study the effects of several model protein stabilizers on the conformation of the unfolded state and early folding dynamics of tetramethyl rhodamine-labeled cytochrome c from Saccharomyces cerevisiae at single molecular resolution. Special attention has been given to arginine, which is a widely used stabilizer for improving refolding yield of different proteins. The value of the hydrodynamic radius (rH) obtained by analyzing the intensity fluctuations of the diffusing molecules has been found to increase in a two-state manner as the protein is unfolded by urea. The results further show that the presence of arginine and other protein stabilizers favors a relatively structured conformation of the unfolded states (rH of 29 Å) over an extended one (rH of 40 Å), which forms in their absence. Also, the time constant of a kinetic component (τR) of about 30 μs has been observed by analyzing the correlation functions, which represents formation of a collapsed state. This time constant varies with urea concentration representing an inverted Chevron plot that shows a roll-over and behavior in the absence of arginine. To the best of our knowledge, this is one of the first applications of fluorescence correlation spectroscopy to study direct folding kinetics of a protein.

An insight into the conformation and dynamics of unfolded and early intermediate states of a protein is essential to understand the mechanism of its aggregation and to design potent inhibitor molecules. Fluorescence correlation spectroscopy has been used to study the effects of several model protein stabilizers on the conformation of the unfolded state and early folding dynamics of tetramethyl rhodamine-labeled cytochrome c from Saccharomyces cerevisiae at single molecular resolution. Special attention has been given to arginine, which is a widely used stabilizer for improving refolding yield of different proteins. The value of the hydrodynamic radius (r H ) obtained by analyzing the intensity fluctuations of the diffusing molecules has been found to increase in a two-state manner as the protein is unfolded by urea. The results further show that the presence of arginine and other protein stabilizers favors a relatively structured conformation of the unfolded states (r H of 29 Å ) over an extended one (r H of 40 Å ), which forms in their absence. Also, the time constant of a kinetic component ( R ) of about 30 s has been observed by analyzing the correlation functions, which represents formation of a collapsed state. This time constant varies with urea concentration representing an inverted Chevron plot that shows a roll-over behavior in the absence of arginine. To the best of our knowledge, this is one of the first applications of fluorescence correlation spectroscopy to study direct folding kinetics of a protein.
Protein aggregation has been well studied in recent literature because of its implication in multiple neurodegenerative diseases (1). Aggregation also poses serious complications in the manufacturing and formulation development of therapeutically relevant proteins (2,3). Extensive research efforts have been devoted to discover inhibitor molecules and to design formulation strategies to block protein aggregation (4 -6). The majority of these inhibitor molecules are targeted to the intermediate states, which occur prior to the formation of large aggregates (or amyloids under suitable solution conditions) (7). This is because of the fact that early intermediates (and often not the large aggregates or amyloids) have been shown to rep-resent the cytotoxic species (7). Use of small molecules of osmo-protecting ability (osmolytes) has generated widespread interests for their roles as protein stabilizers, particularly for industrial protein productions. (8 -11). Although the exact mechanism of their action is not known, these stabilizers have been shown to interact efficiently with the unfolded state of a protein (12,13).
The conformation and dynamics of the unfolded state may have important roles in the late stage of protein folding. Although native-like contact formation early in the unfolded state would lead to efficient folding, incorrect contacts may result in misfolding and subsequent aggregation. Early conformational dynamics are generally rapid and often fall in the s time scale, which is difficult to monitor by conventional techniques such as NMR and stopped flow because their time resolution is limited to ms. Computational techniques such as molecular dynamics simulation could offer limited help because enormous computational resources are needed for the simulation in the s time scale. Specialized experimental techniques have been designed to probe conformation (14 -16) and early contact formations of the unfolded states (17,18) of a number of proteins. There is, however, no report on how a protein stabilizer would affect the conformation of unfolded and early intermediate states of a protein. Any systematic understanding of the role of a stabilizer on the early stage of the protein-folding kinetics is also not available.
In this study, we have shown for the first time that fluorescence correlation spectroscopy (FCS) 2 can be used to monitor the conformation of the unfolded states and early folding kinetics of a protein at single molecular resolution. Tetramethyl rhodamine-maleimide-labeled cytochrome c from Saccharomyces cerevisiae (TMR-cytc) has been used as a model system for this study for the following reasons. First, different aspects of cytochrome c folding and related problems have been investigated (19 -24). Second, the cofactor heme, present intrinsically in the protein structure, and the extrinsically attached tetramethyl rhodamine maleimide (TMR) constitute a convenient fluorescence resonance energy transfer pair, which has already been explored using FCS (25) and photon-counting histogram methods (26). We have selected a number of protein stabilizers such as sucrose, NaCl, proline, and arginine to understand their roles * The work was supported by Council of Scientific and Industrial Research Network Project Grant NWP005 and by a grant from the acknowledges Indian National Science Academy (to S. M.). □ S The on-line version of this article (available at http://www.jbc.org) contains supplemental Figs. S1-S4. 1 To whom correspondence should be addressed. E-mail: krish@iicb.res.in. on the conformation of the unfolded state and on the early folding kinetics of cytochrome c. Special emphasis has been given to arginine as a model stabilizer as it has been used extensively to increase the refolding yield and to prevent protein self-association (27)(28)(29)(30).
Our results show that arginine and other protein stabilizers prefer a relatively compact conformation of the unfolded state. Variation of the time constant of a fast kinetic event representing the decrease in distance between TMR and heme with urea concentrations suggests a Chevron-like behavior. Measuring urea-induced unfolding of TMR-cytc in the absence and presence of arginine, we show that the use of high concentration of arginine influences formation of a partially folded intermediate. Other stabilizers, e.g. sucrose, NaCl, and proline, do not show this behavior.

EXPERIMENTAL PROCEDURES
Yeast Cytochrome c from S. cerevisiae (C2436) was obtained from Sigma. Urea, NaCl, and sucrose were obtained from Sigma-Aldrich in the highest available purity. TMR was obtained from Molecular Probes (Eugene, OR). DL-Arginine and DL-proline (for CD experiments) were obtained from MP Biomedicals (Cleveland, OH). All other reagents used were of high quality analytical grades.
Labeling of cytochrome c with TMR was carried out using a published procedure (31). Excess free dye was removed by extensive dialysis followed by column chromatography using a Sephadex G25 column equilibrated with 20 mM sodium phosphate buffer, pH 7.5.
Far UV CD spectra (in the range 200-nm) were recorded using a Jasco J715 spectropolarimeter (Japan Spectroscopic Ltd.). Experiments were performed typically with 20 M unlabeled protein using a cuvette of 1-mm path length. Steady state fluorescence experiments were carried out with 200 nM of labeled protein samples using a PTI fluorometer (Photon Technology International). The samples were excited at 550 nm, and the emission wavelengths were scanned between 560 and 650 nm; emission intensity at 576 nm was typically monitored for the unfolding experiments.
FCS experiments were carried out using a commercial instrument, ConfoCor 3 LSM (Carl Zeiss, Evotec, Jena, Germany) using a 40ϫ water immersion objective. Typically 500 l of the sample (dye or labeled protein) was placed into Nunc chambers (Nalge Nunc) and excited with an argon laser at 514 nm. The fluorescence signal was separated from the excited line using a main dichroic filter and collected using a pair of Avalanche photodiodes. The photo-current detected by the detectors was used to calculate single color cross-correlation function. FCS measurements were performed in 20 mM sodium phosphate buffer at pH 7.5 in the presence of 20 mM dithiothreitol. Typically, 50 -100 nM labeled protein concentration was used. To correct for the refractive index and viscosity of urea and arginine solutions, necessary correction measures were taken using microscope correction collar and height as described previously (32). Additionally, the experiments were carried out with free TMR at each solution condition to normalize the protein data.
Analysis of the Correlation Functions-For a single component system without any chemical reaction or conformational change event, the diffusion time ( D ) of a fluorophore and the number of particles in the observation volume (N p ) can be calculated by fitting the correlation function (G()) to Equation 1 (33) where S is the structure parameter, the depth-to-diameter ratio of the Gaussian observation volume.
The correlation function for a single component system with diffusion time, D , and an associated chemical reaction or conformational change event (A^B) with a relaxation time constant of R can be described by Equation 2 (34). where w is the beam radius of the observation volume, which can be obtained by measuring D of a fluorophore of known D.
The hydrodynamic radius (r H ) of a molecule can be calculated from D using the Stokes-Einstein equation (Equation 4) where k is the Boltzmann constant, T is the temperature, and corresponds to the viscosity of the solution.
Equilibrium Unfolding Transition-Protein samples in 20 mM sodium phosphate buffer at pH 7.5 were incubated overnight with different concentrations of urea at room temperature. The samples were analyzed using different spectroscopic techniques, such as steady state fluorescence, CD, and FCS. Experimental data were fit to a two-state unfolding transition using Equation 5 (35) ϭ where is the observed spectroscopic data (measured by steady state fluorescence, CD, or FCS) and N and D are the spectroscopic data for the native and completely unfolded proteins extrapolated to zero urea concentration. ⌬G 0 is the free energy of unfolding, and m corresponds to the cooperativity of unfolding transition. The use of the two-state approximations has been common in protein folding literatures including in the case of cytochrome c folding (36,37). However, deviation from the two-state behavior can result from significant accumulation of intermediate states in the unfolding pathway (see "Results"). The data were analyzed and fit using OriginPro 8.0 software (OriginLab Corp.).

RESULTS
In an FCS experiment (outlined using a schematic in Fig. 1), fluorescence fluctuations are measured in a small observation volume, keeping the system under thermodynamic equilibrium. These fluorescence intensity fluctuations can originate either from molecular diffusion of a labeled molecule inside the observation volume ( Fig. 1a) or through any chemical kinetics or conformational events (Fig. 1b) whose rate is faster than its molecular diffusion (32, 38 -40). Determination of molecular diffusion ( D ) using a suitable correlation function (Equation 1 for example) yields r H (31, 32) (Fig. 1a), which provides information about the conformation of the folded, unfolded, or intermediate states of a protein. Analysis of the intensity fluctuations due to any rapid chemical reaction or conformational events can, in addition, be performed using an exponential component in the correlation function (Equation 2 for example), yielding time constant values ( R ) in the s time scale (Fig.  1b). Applications of this methodology have been shown recently, when FCS in combination with fluorescence quenching and fluorescence resonance energy transfer has been used to monitor folding and conformational dynamics of different proteins (38,39,41,42).
The use of Equation 1 has been found inappropriate to fit the correlation functions obtained by the FCS experiments with TMR-cytc because the fit results in non-random behavior in the residual distribution (Fig. 1b). The data fit satisfactorily to Equation 2, and the goodness of the fit is verified by random residual distribution (Fig. 1b). Equation 2 contains a diffusional component with diffusion time of D and an exponential component ( R with an amplitude of F). Use of an equation containing one diffusional component ( D ) and two exponential components ( R1 and R2 with the amplitude of F 1 and F 2 and F 2 ϭ 1 Ϫ F 1 ) does not improve the fit (not shown). The hydrody-namic radius of the protein (r H ) calculated from the value of D obtained from the fit using Equation 2 (22 Å) has been found similar to the published results (43). These results establish that Equation 2 provides the simplest but sufficiently accurate model to fit the correlation functions obtained by the FCS experiments with TMR-cytc. The exponential component ( R of 28 s with the amplitude of ϳ20%) accounts for the s conformational dynamics probed by the changes in the distance between the fluorophore (TMR) and the heme group present in the protein. The values of R and F are similar to those reported earlier (25). The dye TMR and the visible region of heme absorption have been shown to constitute a fluorescence resonance energy transfer pair of Forster radius of 40 Å (25).
It has been reported recently that the use of imperfect Gaussian approximation at large pinhole diameters in FCS data analysis could lead to erroneous extra components (44). FCS experiments were carried out with the labeled protein using pinholes of different diameters and at variable laser powers. No systematic change in the values of R has been observed (Fig. 2). This indicates that R represents the time constant of a true physical motion, and it is not an artifactual component arising either from the blinking of the fluorophore or from imperfect Gaussian approximation of the confocal volume element (44).
Normalized correlation functions observed with TMR-cytc in the absence and presence of arginine are shown in Fig. 3 at three different urea concentrations, viz. 0, 3, and 10 M urea. As mentioned before, arginine has been used in this study as a model protein stabilizer. In the absence of arginine (Fig. 3a), the data clearly show that the amplitude (F) of the faster exponential component increases between 0 and 3 M. The amplitude subsequently decreases beyond 3 M urea and attains a minimum at 10 M urea. The decrease in F beyond 3 M urea coincides with the unfolding of the secondary structure of the protein as monitored by far UV CD (supplemental Fig. S1). In the presence of 500 mM arginine (Fig. 3b), exponential components are superimposable, and no change has been observed between 0 and 3 M urea. Beyond 3 M urea, the amplitude of the faster exponential process decreases (Fig. 3b), which is accompanied by the unfolding of the secondary structure of the protein. Fig. 4 shows the variation of different fit parameters with urea concentration in the absence (Fig. 4a) and presence of arginine (Fig. 4b). Both in the absence (Fig. 4a) and in the presence of arginine (Fig. 4b), r H increases with urea concentrations as the protein unfolds. The data have been fit to Equation 5 using the two-state approximation (35). The increase in r H with urea concentration suggests an extended structure of the unfolded state. In the absence of arginine, r H of the native protein (no urea) is observed to be 22 Å, which decreases to 17 Å in the presence of 500 mM arginine (Fig. 4, a and b). A large decrease in r H for the unfolded protein (from 40 to 29 Å) is also observed in the presence of 500 mM arginine. The decrease in r H of the unfolded state by arginine occurs due to the increase in the free energy of the unfolded state (12). Arginine-induced decrease in r H of the native state may arise from the ability of arginine to interact with the native or native-like states of the protein (see below).
In the absence of arginine, the variation of the amplitude (F) of the exponential component with urea concentration shows the presence of two steps (Fig. 4c). In the first step, the value of F increases at low urea concentration and reaches a maximum between 2 and 3 M urea (Fig. 4c). The second step is accompanied by a large decrease in F at high urea concentration. This step occurs simultaneously with the unfolding of the secondary structure of the protein (supplemental Fig. S1). In the presence of arginine, the initial increase in F is absent, and F decreases in a single step as the secondary structure of the protein unfolds (Fig. 4d). Both in the presence and in the absence of arginine, the variation of R with urea concentration represents the inverted Chevron behaviors. In the absence of arginine, the value of R deviates from linearity at low urea concentration, resulting in the roll-over behavior (Fig. 4e). No roll-over behavior is observed in the presence of arginine (Fig. 4f).
FCS has been used in a number of recent protein folding studies to measure s dynamics of native and unfolded states. For example, fluctuation dynamics between a fluorescent and a non-fluorescent conformer of the intestinal fatty acid binding protein have been measured, which yield a time constant of 35 s (39). A more complex analysis of the correlation functions involving more than one exponential component has been used for the Alexa Fluor 488 maleimide-labeled apo-myoglobin (41). Rapid conformational dynamics in the unfolded states of two different proteins at early s time regions have also been measured using a combination of FCS and fluorescence quenching techniques (38,42). Although the majority of these measurements show no change in the time constant of the exponential component (defined by R in this study) with denaturant concentrations (pH or urea or guanidinium hydrochloride), we see a distinctly different behavior. This is because of the fact that in case of these previous experiments, R was found to measure the internal conformational fluctuations between two conformers of a particular state (native or unfolded), and R did not correspond to the interconversion (or folding/unfolding kinetics) between the states. Because R is solely a representation of that particular state (folded or unfolded), it would not change as the protein folds or unfolds. The amplitude of R would of course change with denaturant concentration as the concentration of that state would change by the folding or unfolding events.
A significantly different behavior in the variation of R with urea concentration in the absence (Fig. 4e) and presence of (diffusion and exponential) leads to random residual distribution. arginine (Fig. 4f) is observed. The data represent an inverted Chevron plot. This behavior can be explained by assuming that R measures the time constant of the interconversion between the native (N) and the unfolded states (U). Assuming k U and k N to be the rate of formation of U (N to U conversion) and N (U to N conversion), respectively, in an ideal two-state transition (N^U), where K is the equilibrium constant of unfolding and [U] and [N] represent the concentration values for the unfolded and the native state of the protein. At low urea concentration, folded state (N) dominates the equilibrium and k N Ͼ Ͼ k U and hence R ϳ 1/k N . Alternatively, at high urea concentration, k U Ͼ Ͼ k N and hence R ϳ 1/k U . Essentially the dependence of R with urea concentration would have a profile of an inverted Chevron plot (as R is proportional to the inverse of the rate constant). The presence of intermediates at low denaturant concentration typically introduces non-linearity (roll-over behavior) at the low denaturant concentration of the Chevron plot. The observed data in the absence of arginine (Fig. 4e) are consistent with this analysis. The presence of arginine removes the non-linearity in the inverted Chevron plot, indicating the absence of any intermediate state (Fig. 4f). The initial increase in the values of F at low urea concentrations in the absence of arginine (Fig. 4c) is consistent with the presence of an intermediate state at low urea concentration. Because F in Equation 2 represents the average non-fluorescent fraction of cytochrome c molecules, an increase in F implies the presence of higher number of non-fluorescent molecules (where TMR and heme are close to each other) in the intermediate state. In the unfolded state, the protein is extended with the heme group and TMR well separated from each other, which results in a large decrease in F (Fig. 4c). Steady state fluorescence intensity change of the attached fluorophore  In the absence of arginine (a), the amplitude of the exponential component increases between 0 (black) and 3 M urea (red), which follows a large decrease at 10 M urea (blue). In the presence of 500 mM arginine (b), the data at 0 (black) and 3 M urea (red) are superimposable. FCS experiments were carried out at room temperature in 20 mM sodium phosphate buffer at pH 7.5 containing 20 mM dithiothreitol.
(TMR) in the absence of arginine compliments the behavior of F (supplemental Fig. S2). Fluorescence intensity of the attached fluorophore (TMR) decreases at low urea concentration (and the value of F increases) with the formation of the intermediate state. At high urea concentration, the steady state fluorescence intensity increases as the protein unfolds (and F decreases). In the presence of arginine, the initial increase in F (Fig. 4d) and the decrease in steady state fluorescence intensity (supplemental Fig. S2) have been found absent. This indicates a twostate unfolding process in the presence of arginine without any significant population of the compact and non-fluorescent intermediate, a conclusion well complimented by the lack of roll-over behavior in the Chevron plot (Fig. 4f).
Unfolding of cytochrome c proceeds through the dissociation of Fe-Met-80 bond at mild denaturant conditions (at low pH or in the presence of low concentration of chemical denaturants). Subsequently, one of the neighboring histidine residues (His-26, His-33, or His-39) coordinates to heme to form a misfolded intermediate with bis-His coordination geometry (23,46,47). To check whether the increase in F at low urea concentration is caused by Fe-Met-80 cleavage and subsequent Fe-His bond formation, FCS measurements were carried out at pH 4, and the results are described in Fig.  5. At pH below 5, histidine residues are protonated, and hence they would not coordinate to the heme iron to form bis-His adduct. As described in Fig. 5, unfolding of cytochrome c probed by FCS at pH 4 does not show any increase in F at low urea concentration. Unfolding at pH 4 has been found to be a twostate process accompanied by an increase in D occurring simultaneously with a single-phase decrease in F (Fig. 5).
To compare the effects of arginine with other common stabilizers, urea-induced unfolding transitions of TMR-cytc were carried out in the presence of sucrose, NaCl, and proline. Except in the case of arginine, an increase in F at low urea concentration has been observed for all other stabilizers (Fig. 6). The rollover behavior at low urea concentration in the inverted Chevron plot is also present with these stabilizers (Fig. 6, inset).
To understand the effect of the stabilizers on the conformation of the unfolded states, FCS experiments were carried out with cytochrome c unfolded by 10 M urea in the presence of different concentrations of stabilizing agents (arginine, NaCl, sucrose, and proline). Separate experiments were also carried out with the native protein in the absence of urea. The results are summarized in Fig. 7. For the unfolded protein (in the presence of 10 M urea), all the stabilizers behave similarly, and the values of r H decrease with the addition of stabilizers (Fig. 7a). For the folded protein (in the absence of urea), however, arginine behaves differently when compared with other stabilizers (Fig. 7b). Use of arginine results in significant decrease in r H for the folded protein, whereas NaCl, sucrose, and proline do not show any significant change (Fig. 7b).
The unfolded state of cytochrome c has been the subject of extensive studies (21,48,49). Small angle x-ray scattering experiments have shown that the unfolded state of cytochrome c contains at least two conformers (U1 and U2) and that the population of U2 is higher in the presence of high concentration of denaturants (50). Using fluorescence energy transfer between heme and an external DNS fluorophore placed at the different regions of cytochrome c, a detailed, complex, and heterogeneous description of the unfolded state has been achieved (51). This particular study (48) showed the presence of an equilibrium distribution of a compact conformer with heme-fluorophore distance of 30 Å and an extended conformer with hemefluorophore distance of 40 Å in the unfolded state, and their relative population varies with denaturant concentrations. Our data show that at 10 M urea, the extended structure is predominately present in the unfolded state (U2, r H is 40 Å). Protein stabilizers shift the equilibrium toward the more compact con-former (U1, r H in the presence of 10 M urea and 500 mM arginine is 29 Å).

DISCUSSION
In this report, we show an application of FCS to study the conformation of the unfolded state of a protein and its folding kinetics in the s time scale. Two important issues, however, need careful consideration. First, FCS, being a single molecular technique, requires labeling of a protein molecule by suitable dyes, which are bright and sufficiently photo-stable. Hence, care should be taken to rule out any significant conformational perturbation of the protein induced by the attached fluorophore. The effect of TMR labeling on cytochrome c has been studied by Perroud et al. (26) using CD. Their study confirmed the absence of any major structural change of the protein due to the labeling by TMR (26). Second, the kinetic events observed by FCS are limited by the diffusion time of the molecule. Cytochrome c folding kinetics is complex with observation of slow kinetics due to proline isomerization and other factors. FCS cannot monitor these slow events.   The model in Scheme 1 can be used to explain the FCS results described in the present report. In this model, the horizontal and vertical arrows indicate the increase in urea and arginine concentrations, respectively. In the absence of arginine, cytochrome c unfolds through the formation of a collapsed intermediate I, which is relatively non-fluorescent (as observed by the increase in F and also by the decrease in steady state fluorescence intensity of attached TMR) with conserved secondary structure (as observed by far UV CD). Experiments at low pH suggest that the intermediate I may be the result of misligations by a neighboring histidine residue (His-26, His-33, or His-39). A recent photon-counting histogram measurement on the TMR-labeled cytochrome c also indicates formation of a similar intermediate at alkaline pH with a Lys misligation (26). The alkaline intermediate is also observed to be accompanied by a large decrease in the fluorescence intensity of TMR (26). The unfolded state in the absence of arginine is relatively extended (U2) with r H of 40 Å.
Arginine, interacting with the native state of the protein (N) forms an intermediate, IЈ as observed by a decrease in r H to 17 Å. Although this behavior is in contrast to other protein stabilizers such as NaCl, sucrose, or proline (no change in r H at the native state of the protein), the ability of arginine to interact with the native state has been suggested before (28,52). An increase in F and a simultaneous decrease in the steady state fluorescence intensity at 576 nm (for the attached TMR) have been observed with the formation of IЈ (supplemental Fig. S3). The unfolded state of the protein in the presence of arginine (U1) is significantly compact with r H of 29 Å. The equilibrium between a compact (U1, 30 Å) and extended form (U2, 40 Å) in the unfolded state of cytochrome c has been shown previously (50,51).
The intermediates I and IЈ form early in the protein unfolding pathway. Steady state tryptophan fluorescence experiments suggest a small change in the tryptophan environments as the intermediates I and IЈ form (supplemental Fig. S4a). Near UV CD experiments on IЈ and I also show a small difference near 260 and 280 nm (shown in supplemental Fig. S4bby arrows). We speculate that alteration of the structure of the hydration layer may be predominately responsible for the formation of these two intermediates with possible minor change in the conformation of the aromatic region of the protein (53). It has been shown recently that arginine may influence the aromatic region of different proteins, although its affinity is weak (31). The absence of any significantly large change in the tryptophan fluorescence or near UV CD justifies the inference of weak interaction. Assuming the presence of an equilibrium population of IЈ and I in solution, arginine and urea may compete for the native state (N) of the protein. The ability of arginine to shift the equilibrium toward IЈ may arise from its competence to interact with I, which in turn stimulates its ability to block protein association and aggregation. It is interesting to note that arginine is different from other stabilizers in its ability to influence conformational fluctuations of a protein, its intermediate states, and presumably the stabilization of the side chains. Other stabilizers do not affect the native, or more importantly, native-like states of a protein. In contrast, the use of arginine affects their hydrodynamic behaviors. While common stabilizers such as trimethylamine-N-oxide and sarcosine have been shown to stabilize, although to different extents, early intermediate states of burstar (54), arginine has been shown to inhibit accumulation of early intermediate states (31). The uniqueness of arginine when compared with other osmolytes in its ability to interact with the side chains of the native or native-like states of a protein has been noted earlier (28). The dual nature of arginine of compacting the unfolded states of a protein (similar to other stabilizing osmolytes) and its ability to interact with the side chains of the native-like intermediates of a protein (different from other stabilizing osmolytes) makes it a potent protein stabilizer and a "special" molecule for the prevention of protein aggregation (28).
Microsecond events of cytochrome c folding have been studied in detail by several investigators and are subjects of intense debate. A 40-s folding kinetics was observed by Pascher et al.  (57,58), who suggested that the submillisecond process does not lead to a true barrierlimited intermediate; rather it is a manifestation of nonspecific compaction expected for any polymer as the denaturant gets diluted. Their conclusion was based on the fact that the amplitude of the burst phase observed for the cytochrome c refolding experiments is similar to the burst phase amplitude observed for two non-folding fragments of the protein. Qiu et al. (59) have addressed this controversy by showing that not only the native protein but even destabilized, non-foldable chains can contract and form a collapsed state and encounter significant barrier during the early process. In the present work, we observed a time constant of 30 s for the early collapse of cytochrome c in the presence of arginine by extrapolating R to zero urea concentration (from the plot of ln R versus urea concentration). Although similar analysis in the absence of arginine is not feasible because of the roll-over effect at low urea concentration, arginine does not seem to have any significant effect on the individual R values. The ability of arginine to inhibit aggregate formation may have direct relevance to its ability to influence early intermediate formation, although the time constants for the early events seem to be less crucial. In a previous study, the burst phase kinetics (measured by continuous flow fluores-SCHEME 1. Unfolding of TMR-cytc in the absence and presence of arginine. The horizontal and vertical arrows indicate the increase in urea and arginine concentrations, respectively. cence) and the relatively slower kinetics (measured by the stopped flow fluorescence method) of a number of turn mutations of the intestinal fatty acid binding proteins have been reported (35,45). Although the thermodynamic stability, tendency to form inclusion body, and the rates of slow kinetics of the mutant proteins were found to vary considerably, the time constants of their burst phase kinetics were remarkably similar (35).
In this study, we show that FCS can be used not only to accurately determine the hydrodynamics of the unfolded state of a protein but also to measure rapid folding kinetics. Although previous studies have shown the application of FCS to monitor internal s dynamics of the native and unfolded states of different proteins (38,39,41), direct measurement of the folding kinetics by FCS has been found difficult. To the best of our knowledge, this is the first report of any Chevron analysis using rapid kinetics measurements by FCS. We also believe that this is the first direct experimental observation of the role of protein stabilizers on the conformation of the unfolded states and early folding kinetics of any protein.