Stepwise unfolding of a multi-tryptophan protein MPT63 with immunoglobulin-like fold: detection of zone-wise perturbation during guanidine hydrochloride-induced unfolding using phosphorescence spectroscopy

Abstract

A combination of circular dichroism, steady-state and time-resolved fluorescence, and low temperature phosphorescence spectroscopy has been used to study guanidine hydrochloride (Gdn)-induced stepwise unfolding of MPT63, a small protein with immunoglobulin-like fold. A simple, rapid and inexpensive low temperature (77 K) phosphorescence (LTP) study of MPT63, having four Trp residues, in a suitable aqueous cryogenic solvent clearly indicates that the most exposed Trp 82 and the deeply buried Trp 129 exhibit a greater probability of being unfolded in comparison to the partially exposed Trp 48 during Gdn-induced unfolding. The partially exposed Trp 48 practically conserves its environment up to 1.6 M Gdn concentrations. The excitation wavelength-dependent LTP and lifetime of the triplet state further support the contention. Tyrosine residues which are silent in free MPT63 start exhibiting emission when the concentration of Gdn reaches 1.6 M. The overall solvent exposure of the Trp residues in the fully unfolded state of MPT63 is observed to be less than that observed for a Trp residue in a tripeptide and a heptapeptide in a 40% ethylene glycol matrix. The specific detection of perturbation of the environment of individual Trp residues during gradual unfolding has been achieved without using any Trp substituted mutant. This is possible since the wild-type MPT63 exhibits unusual optical resolution of all four Trp residues by LTP due to the special architecture of MPT63, where the four Trp residues are well distributed in its tertiary structure.

---

1. Introduction

Unfolding studies have promoted knowledge of the structure and stabilization of simple globular proteins^1–3 and of multi-domain proteins^4–6 and may provide clues to the pathway of folding. These investigations seek to answer how a protein folds into its unique, compact, highly ordered, and functionally active form. The field of protein folding–unfolding has seen essential advances in recent years because of growing interest in diseases that result from protein misfolding and aggregation.^7–11 One approach to this problem comes from studying equilibrium denaturation of proteins, since any intermediates formed in the presence of denaturants may contain stable elements of a secondary and/or tertiary structure. Unfolding induced by denaturants is modelled, in many cases, by a two-state mechanism involving the native and the unfolded state.^12,13 The native state is considered to comprise an ensemble of molecules, which share a common fold with the unique three-dimensional structure. The unfolded state is thought to be more or less a random coil, with little or no regular structure. Identifying the existence of folding initiation sites, as well as unfolded or partly folded intermediates, presents a difficult challenge. However, the application of cutting edge spectroscopic and biophysical techniques, including NMR, AFM and hydrogen exchange methods, has enabled successful detection and characterization of single and/or multiple intermediate states, which are populated in the unfolding pathways of multiple proteins.^14–22 For these cases, the unfolding pathways involve multiple steps (instead of two steps), which may be sequential or parallel. Recent understanding of protein folding often assumes the concept of foldons, small independent structural units, which fold at different time points. Many protein folds through a hierarchical fashion, in which different regions of a protein fold differently. All these results necessitate the use of multiple techniques, which could provide region-specific information of a protein.

For the present study, we have chosen MPT63 as the model protein. It is a secreted protein of unknown function that is specific to Mycobacterium tuberculosis and is a potential drug target.^23–25 Secreted proteins like MPT63 are needed for bacteria to survive in a hostile environment or to colonize a host successfully. The protein has been shown to stimulate humoral immune responses in guinea pigs infected with virulent M. tuberculosis.²³ MPT63 is a predominantly β-sheet-rich protein with a molecular weight of 17.5 kDa.²⁶ The X-ray crystal structure of MPT63 at 1.5 Å resolution has been solved. The protein is composed of two anti-parallel β-sheets differing in lengths juxtaposed to each other and a small 310 helix. The longer sheet contains four anti-parallel β-strands and the shorter sheet is made up of five β-strands and the helix²⁶ (Fig. 1). In our previous study²⁷ it was shown that the four tryptophan (Trp) residues located at 26, 48, 82, and 129 positions can be optically resolved using low temperature phosphorescence (LTP) spectroscopy due to the special inherent architecture of this protein (Fig. 1). One of the most important reasons for choosing this protein as our model system is that it provides an easy and convenient method to monitor the change in the local environment of each of these four tryptophans as a result of the unfolding stress. Furthermore, four tyrosine (Tyr) residues located at 3, 41, 78 and 106 (Fig. 1) are silent in the native state of MPT63. Thus it would be interesting to know the emissive features of Tyr residues during stepwise unfolding.

Fig. 1 (A) Structure of MPT63 showing the Trp and Tyr residues. (B) Distances (in Å) among the four Trp residues of MPT63. Inset shows the accessible surface area (ASA in Ų) of the four Trp residues.

Low temperature triplet state spectroscopy of biopolymers or biopolymer–substrate complexes in a proper cryosolvent often provides complementary structural data obtained from X-ray diffraction and NMR spectroscopy. Low temperature triplet state spectroscopic studies of wild-type T4 lysozyme and Trp substituted mutant molecules^28–30 showed that the local environments of the individual Trp residues are in complete agreement with those determined by X-ray crystallography.³¹ The structural properties using low temperature triplet state studies of a mutated (Y45W) RNase T1 and its complexes with two ligands, the specific 2′-GMP and the 2′-AMP pseudo-substrates,³² in an aqueous cryosolvent glass, agree very well with the crystallographic results.³³ There is also complete agreement of the low temperature triplet state studies with the NMR results of NCp7 binding with SL335 that Trp37 is in contact with G7 of the loop (2′) and that aromatic stacking is limited. It may be mentioned that a protein or protein–substrate complex in a proper cryosolvent upon cooling will seek lower energy conformations on the potential energy landscape. Some of these will represent conformations of the systems at ambient temperatures, but the higher energy representatives will not contribute at low temperature. This picture is also true in X-ray crystallographic structures where crystal packing forces restrict the number of conformations available in the natural environment.³⁴

The spectral properties of four Trp residues in MPT63 obtained from LTP studies²⁷ agree very well with the crystal structure data of the protein.²⁶ The conventional steady-state, time-resolved fluorescence, anisotropy decay and CD methods applied to multi-Trp MPT63 in this work suggest a two-state transition during Gdn-induced unfolding. LTP of MPT63 at a micromolar level concentration in a suitable aqueous cryosolvent glass has been shown to identify the perturbation of the local environment of individual Trp residues during the Gdn-induced unfolding process. Since these Trp residues are well distributed in the primary and tertiary structure, they could act as good spectral reporter groups during the unfolding process. Our results clearly show that inner buried Trp residues in this protein exhibit a greater probability of being unfolded in comparison to partially exposed Trp residues during Gdn-induced unfolding. Excitation wavelength-dependent LTP further confirms this contention. This kind of detection of zone-wise perturbation during unfolding without using a Trp substituted mutant is not available in the literature. This is possible in MPT63 because of unusual optical resolution of all four Trp residues by LTP in a suitable aqueous cryosolvent. LTP has also been exploited to detect the overall perturbation of Tyr residues during the unfolding process.

2. Materials and methods

2.1. Expression and purification of MPT63

MPT63 protein from Mycobacterium tuberculosis was expressed in E. Coli XL1, Blue pQE 30 (kind gift from Dr David Eisenberg, University of California, Los Angeles, USA) and purified as described by Ghosh et al.²⁷ RIP or heptapeptide (L-7880, Sigma-Aldrich, St. Louis, MO) was >99% pure (as assessed by SDS-PAGE and HPLC). Tripeptide Lys–Trp–Lys (Cat. no. L-3164) was purchased from Sigma. L-Tryptophan was purchased from Sigma.

2.2. Chemicals and buffers

Guanidine hydrochloride (Gdn) was obtained from Sigma-Aldrich at the highest available grade. All other reagents used were high grade analytical reagents. Protein solutions were prepared in pH 7.5, 20 mM sodium phosphate buffer.

2.3. Circular dichroism (CD) spectroscopy

Circular dichroism (CD) spectra were recorded using a Jasco J715 spectropolarimeter (Japan Spectroscopic Ltd.). Far-UV CD measurements (between 195 and 250 nm) were performed with 10 μM protein using a cuvette with the path length of 1 mm. Eight spectra were collected in continuous mode and averaged. For near-UV CD experiments (between 250 and 350 nm), 15 μM protein was used in a cuvette with the path length of 1 cm. Three spectra were recorded in continuous mode and averaged.

2.4. Steady-state and time-resolved study

The steady-state emission measurements of 5 μM MPT63 with varying concentrations of Gdn at pH 7.5 were carried out using a Hitachi Model F-7000 spectrofluorimeter equipped with a 150 W xenon lamp, at 298 K using a stopper cell of 0.5 cm path length. The emission measurements were carried out by exciting the samples at 290 nm and 295 nm (in order to minimize the contribution from Tyr). The excitation and emission band passes were 10 nm and 1 nm. Inner filter effects have been eliminated in all of the emission measurements.

The singlet state lifetime (τ) was measured by a Time Master fluorimeter from Photon Technology International (PTI, USA). The system consists of a pulsed laser driver of a PDL series i.e., PDL-800-B (from Picoquant, Germany) with interchangeable sub-nanosecond pulsed LEDs and pico-diode lasers (Picoquant, Germany) with a TCSPC set up (PTI, USA). The software Felix 32 controls all acquisition modes and data analysis of the Time Master system.³⁵ The sample of free 5 μM MPT63 with varying concentrations of Gdn at pH 7.5 was excited using PLS-290 (Pulse Width-700 ps) respectively at a repetition frequency of 10 MHz. The excitation and emission band passes were 10 nm and 5 nm. Instrument response functions (IRFs) were measured at the respective excitation wavelength (λ_exc), namely, 290 nm using slits with a band pass of 3 nm using Ludox as the scatterer. The decay of the sample was analyzed by a non-linear iterative fitting procedure based on the Marquardt algorithm. The deconvolution technique used can determine the lifetime up to 200 ps with sub-nanosecond pulsed LEDs and 100 ps with diode lasers. The quality of fit has been assessed over the entire decay, including the rising edge, and tested with a plot of weighted residuals and other statistical parameters, for example, the reduced χ2 ratio and the Durbin–Watson (DW) parameters.³⁵

Intensity decay curves were fitted as a sum of exponential terms:³⁶

F(t) = ∑α_i exp(−t/τ_i)

where α_i represents the pre-exponential factor to the time-resolved decay of the component with a lifetime τ_i. The amplitude average lifetime 〈τ〉 was calculated using the equation:³⁶

〈τ〉 = ∑(α_iτ_i)/∑α_i

2.5. Time-resolved anisotropy study

Anisotropy decay measurement was also carried out with a Time Master Fluorimeter (PTI, USA) using PLS-290 and a motorized Glen Thompson polarizer. The anisotropy, r(t) is defined as:

r(t) = [I_VV(t) − G × I_VH(t)]/[I_VV(t) + 2G × I_VH(t)] (1)

where the I(t) terms are defined as the intensity decay of emission of protein with the excitation polarizer orientated vertically and the emission polarizer oriented vertically and horizontally, respectively.

G = I_HV(t)/I_HH(t) (2)

where G is the correction term for the relative throughput of each polarization through the emission optics. The entire data analysis was done with the software Felix 32 which analyses the raw data I_VV and I_VH simultaneously by a global multi-exponential program and then the deconvolved curves (ID_VV and ID_VH) are used to construct r(t).³⁵ The anisotropy decay r(t) was considered to be a sum of discrete exponential functions:
where β_i and θ_i were the numerical parameters recovered to ensure the best fit of the r(t) function and r₀ is the anisotropy at t = 0. From the fitted curve the correlation time (θ_c) can be recovered.

2.6. Steady-state and time-resolved phosphorescence spectroscopy

Low temperature phosphorescence spectroscopic studies at 77 K were performed using a Hitachi Model F-7000 spectrofluorimeter (Hitachi High-Technologies Corp. Japan) equipped with a 150 W xenon lamp and other phosphorescence accessories. Measurements were made using a Dewar system having a 5 mm OD quartz tube. The freezing of the samples at 77 K was done at the same rate for all the samples. 5 μM of free protein and with Gdn at pH 7.5 was used for experiments and the samples were made in a 40% ethylene glycol (EG)–buffer mixture for measurements at 77 K. The samples were excited at different wavelengths (280–305 nm) using a 10 nm band pass, and the emission band pass was 1 nm except with λ_exc = 305 nm, where the emission band pass was 2.5 nm.

The low temperature phosphorescence decay times of 5 μM of free protein and with Gdn at pH 7.5 in the milliseconds or longer range were measured by phosphorescence time-based acquisition mode of the QM-30 from PTI, USA in which the emission intensity is measured as a function of time. The decay parameters were recovered using a non-linear iterative fitting procedure based on the Marquardt algorithm.

2.7. Data fitting

MPT63 samples were incubated at an increasing guanidine hydrochloride concentration for two hours at ambient temperature. Samples were then analysed using different structural probes like steady-state and time-resolved tryptophan fluorescence spectroscopy. The spectroscopic data shown in this paper were optimally fit using the two-state unfolding transitions. The data, when fitted to a more complex model (like the three-states transition model) did not improve the fits. In many cases, the fits using more complex models lead to physically unrealistic parameters. Spectroscopic data were fit to a simple two-state unfolding transition using the following equation:³⁷
where Y corresponds to the observed spectroscopic signal (fluorescence intensity/emission maxima/average lifetime/rotational correlation time/fluorescence area), Y_N and Y_D are the spectroscopic signals for the native and completely unfolded protein, respectively extrapolated to zero denaturant (here guanidine hydrochloride) concentration. ΔG⁰ stands for the free energy of unfolding and m is the cooperativity of the transition.

All the data were analyzed and fit using OriginPro version 8.5 (OriginLab Corp).

3. Results and discussion

3.1. Steady-state fluorescence study

The fluorescence of MPT63 in the presence of varying concentrations of Gdn is shown in Fig. 2A and Table 1. An indication of the changes in the overall micro-environment of Trp residues contributing to fluorescence upon increasing the Gdn concentration can be visualized from the analysis of the emission maximum, λ_max. Unfolding of MPT63 induced by Gdn resulted in a red shift of the emission spectrum and about a 2-fold drop in the fluorescence quantum yield (Fig. 2). The plot of fluorescence intensity and maxima against Gdn concentration indicates that the continuous decrease in fluorescence intensity and the resulting red shift of λ_max followed the unfolding in a monotonic way (Fig. 2B). The λ_max shifted over the range of 1.2–2.4 M Gdn from a value of 330 ± 1 nm in the native state to 350 ± 1 nm in the completely unfolded state. A further increase in the Gdn concentration to 6.0 M does not lead to any change in λ_max.

Fig. 2 Steady-state (A) corrected fluorescence and (B) normalized fluorescence spectra of 5 μM MPT63 with varying concentrations of Gdn at 298 K; curves (1–16) represent 0, 0.2, 0.4, 0.8, 1.0, 1.2, 1.4, 1.6, 1.8, 2.0, 2.2, 2.4, 2.6, 2.8, 3.0, and 3.4 M of Gdn respectively. Excitation wavelength = 295 nm; excitation and emission band pass = 5 nm and 1 nm respectively. Inset: variation of fluorescence (A) intensity at λ_max of fluorescence and (B) emission maxima as a function of concentration of Gdn.

Table 1 Fluorescence characteristics of MPT63 (WT) in the presence of different Gdn concentrations in aqueous buffer (pH 7.5) at 298 K. λ_exc = 290 nm. Concentration of protein in each case = 5 μM

Gdn used (M) with 5 μM protein	λmax (nm)	Quantum yield (ϕ)	Singlet state lifetime monitoring λmax				Rotational correlation time (θc) (ns) monitoring λmax
			τ1 (ns)	τ2 (ns)	τav (ns)	χ^2
0.0	330.0	0.232	4.46 (80.17%)	1.49 (19.83%)	3.87	1.02	6.5
0.2	330.0	0.231	4.41 (80.45%)	1.50 (19.55%)	3.84	1.11	6.5
0.4	329.8	0.254	4.42 (81.90%)	1.67 (18.10%)	3.92	0.84	6.5
0.8	330.0	0.285	4.37 (80.73%)	1.56 (19.27%)	3.82	0.91	6.4
1.0	330.2	0.301	4.40 (80.45%)	1.52 (18.85%)	3.82	0.95	5.5
1.2	330.6	0.260	4.29 (82.45%)	1.79 (17.55%)	3.85	1.04	4.8
1.4	330.8	0.221	4.26 (73.27%)	1.72 (26.73%)	3.58	1.06	4.5
1.6	336.2	0.165	3.92 (57.97%)	1.46 (42.03%)	2.88	0.99	3.0
1.8	339.6	0.126	3.86 (48.32%)	1.59 (51.68%)	2.68	0.85	2.5
2.0	343.8	0.130	3.72 (42.56%)	1.57 (57.44%)	2.48	0.80	1.7
2.2	347.2	0.114	3.58 (41.74%)	1.47 (58.26%)	2.35	1.06	1.3
2.4	348.4	0.116	3.31 (48.65%)	1.33 (51.35%)	2.29	1.00	1.3
2.6	348.2	0.115	3.35 (48.52%)	1.35 (51.48)	2.32	1.05	1.3
2.8	348.0	0.116	3.45 (48.71%)	1.29 (51.29%)	2.34	1.04	1.2
3.0	347.8	0.117	3.46 (48.32%)	1.35 (51.68%)	2.36	1.02	1.3
3.4	348.4	0.117	3.48 (47.96%)	1.39 (52.04%)	2.39	1.11	1.2

---

The value of λ_max observed for WT in the native state is indicative of the relatively less polar and more hydrophobic environment of the emitting Trp residues. The position of the emission maximum in the unfolded state reflects the highly polar environment and the increased exposure of the Trp residues in the denatured state. A similar position of emission maximum has been observed for low molecular weight peptides where the Trp residue is surrounded by polar groups.³⁶ The equilibrium dependencies of the Trp fluorescence intensity and fluorescence spectrum position revealed two state transitions (Fig. 2A inset and B inset).

3.2. Far-UV CD

In order to understand the effect of the denaturant guanidine hydrochloride (Gdn) on MPT63 at the tertiary and secondary structural level, far-UV CD spectra were measured in the presence of increasing denaturant concentrations. The far-UV CD spectra of the native and Gdn-induced unfolded MPT 63 are shown in Fig. 3. The shape of the curve of the native protein, in particular the maximum at 216 nm, is characteristic of a β-sheet-rich protein. A shoulder at 230 nm is observed which can be attributed to the effect of the tertiary structure of the protein. A mutant protein, Trp 26Phe, which has been shown to have distorted tertiary contacts, does not have the 230 nm band in the far UV CD spectrum.^27,38 Increasing the concentration of Gdn progressively alters the secondary structures (β-sheet) of MPT63 as is evident from the far-UV CD data (Fig. 3).

Fig. 3 Far-UV CD spectra of MPT63 with increasing concentrations of guanidine hydrochloride at 25 °C, pH 7.5. The far-UV CD spectra are shown up to 210 nm for some guanidine hydrochloride samples due to high sample absorption. Inset: variation of mdegree values obtained at 216 nm from CD data as a function of concentration of Gdn.

The chemical denaturation of MPT 63 was measured from the changes in intrinsic tryptophan fluorescence and far-UV CD on addition of increasing concentrations of guanidine hydrochloride at room temperature (25 °C). The fluorescence intensity at 331 nm (as the native folded protein shows its maximum fluorescence intensity at that wavelength) and CD value at 216 nm (as MPT63 is a β-sheet protein) were plotted along with the Gdn concentration, which results in cooperative transitions. Denaturation curves were fitted assuming two-state transitions. Fig. 3 shows the equilibrium unfolding transitions for the protein using Gdn. Although the unfolding curves show single, cooperative transitions, a small increase in the intrinsic tryptophan fluorescence intensity at the pre-transition regions was observed. The change in fluorescence intensity at low denaturant concentration suggests a subtle alteration in the local tryptophan environment before unfolding. Both the transitions coincide with each other when monitored by different spectroscopic probes indicating a strong evidence of the absence of any intermediate states during MPT63 unfolding. The midpoint of transition, C_M, is 1.4 ± 0.1 M for Gdn and the C_M value remains the same for unfolding transitions obtained using either tryptophan fluorescence or far-UV CD (Table 2).

Table 2 Thermodynamic parameters of Gdn-induced unfolding of wild-type MPT63 monitored by far-UV CD and tryptophan fluorescence

Measurement	ΔG^0 (kcal mol^−1)	m (kcal mol^−1 M^−1)	Cm (midpoint)^a
a Error in Cm = ±0.1.
Fluorescence intensity (Fig. 2A inset)	4.15	3.04	1.36
Average lifetime of fluorescence (Fig. 4)	5.30	4.0	1.33
Rotational correlation time (Fig. 5)	2.79	2.09	1.33
Area under fluorescence	3.30	2.50	1.32
Far-UV CD	—	—	1.40

---

3.3. Time-resolved fluorescence study

Time-resolved fluorescence spectroscopy reflects the structural complexity of proteins. Even a protein containing a single Trp residue exhibits several fluorescence lifetimes due to the existence of different rotameric states of the indole moiety. Table 2 lists the dependencies of the mean lifetimes of the excited states of MPT63 upon the Gdn concentration. Fluorescence intensity decays of WT MPT63 were adequately described by a bi-exponential fit in the native state at pH 7.5 (Fig. 4). The absence of a single lifetime, even when MPT63 is fully unfolded, indicates that Trp residues are not locked in a single conformation. The lifetime of the long-lived component, however, has a value of 4.42 ns in the native state and 3.48 ns in the fully unfolded state. A less significant change in the shorter lifetime (τ_short) was observed in the overall range of Gdn concentration used, where τ_short was recovered as 1.67 ns in the native state and 1.39 ns in the fully unfolded state. The percentage contribution of each component shows significant change in the range 1.2–1.8 M Gdn. Due to the complexity of the intrinsic Trp fluorescence decay, the average lifetime τ_av is used as the observable parameter at a different concentration of Gdn. The plot of τ_av as a function of denaturant concentration exhibits a continuous decrease of τ_av within 1.2–1.8 M Gdn (Fig. 4 inset). The Gdn-induced decrease in both the Trp fluorescence intensity and the fluorescence lifetime suggests an increase in the extent of fluorescence dynamic quenching caused by the proximity of positively charged amino acids upon protein unfolding and electron transfer from neighboring residues. The Gdn-induced changes in τ_av of MPT63 are found to fit a two-state transition, which begins at 0.3 M Gdn and ends at 2.1 M Gdn (Fig. 4).

Fig. 4 Time-resolved fluorescence spectra of 5 μM MPT63 with varying concentrations of Gdn at 298 K; λ_exc = 290 nm; excitation and emission band pass = 10 nm and 5 nm respectively. Inset: Gdn concentration dependence of fluorescence lifetime of MPT63 at room temperature (298 K).

3.4. Time-resolved fluorescence anisotropy

Time-resolved fluorescence anisotropy monitors the mobility and structural flexibility of proteins. Nanosecond dynamics of MPT 63 were measured in the presence of various concentrations of Gdn. A typical anisotropy decay with 1.6 M Gdn is shown in Fig. 5. All the anisotropy decays at various Gdn concentrations are found to fit with one exponential. The rotational correlation time, θ_c, associated with the Gdn-induced change in anisotropy decay of MPT63, is summarized in Table 2. The rotational correlation time thus recovered represents the total motion of the protein.

Fig. 5 Fluorescence anisotropy decays of WT MPT63 (5 μM) with 1.6 M Gdn concentration at 298 K; λ_exc = 290 nm; excitation and emission band pass = 10 nm each. Inset: Gdn concentration dependence of rotational correlation time of MPT63 at room temperature (298 K).

MPT63 exhibits θ_c = 6.5 ns in the native state at pH 7.5. The value matches with the total motion of the protein as expected from the globular structure calculated using the equation

θ_c = [ηM(ν + h)][1/RT]

where η is the viscosity of the solvent (0.94 × 10⁻³ Pa s), M is the molecular weight in Daltons, ν is the partial specific volume of the protein (assuming 0.72 mL g⁻¹ protein), and h is the degree of hydration (assuming 0.23 cm³ g⁻¹ protein).²⁷

Fig. 5, inset, displays the variation of the values of θ_c observed for the total motion of MPT63 as a function of Gdn concentration. The results show that unfolding leads to a relaxation of the slow motion. Initially, the value of the fluorescence anisotropy remains constant until 0.8 M Gdn. One can infer that in the range from 0.0 to 0.8 M Gdn the protein preserves a globular structure with greater restriction in mobility. In the range of 1.2–2.0 M, it sharply decreases, reaching the value characteristic of the unfolded protein at 3.0 M Gdn. This increase in the overall motion of MPT63 (decrease in rotational correlation time) reflects the cooperative loss of intramolecular interactions involving the fragment of the polypeptide chain. At the final stage, θ_c of MPT63 exhibits a value typical for a single Trp-containing protein with higher mobility.³⁶ Importantly, the final unfolding state observed by time-resolved fluorescence anisotropy is in good agreement with CD and the steady-state fluorescence properties. The time-resolved anisotropy data, which are similar to CD, steady-state fluorescence, and average decay time data, support a two-state model of unfolding.

There are several studies which showed that a significant red shift in the Trp fluorescence spectrum does not necessarily reflect the considerable increase in its environment mobility.^39,40 Furthermore, the red shift of the Trp fluorescence spectrum can be accompanied by an increase in rotational correlation time.^41,42 This phenomenon is explained by two opposite effects on the value of Trp correlation time: the increase in the flexibility of Trp residues in the open structure, which is accompanied by a decrease in θ_c, and the decrease in fluorescence lifetime (Table 1), which is accompanied by an increase in θ_c. The significant disruption of the secondary structure and increased exposure of Trp residues in MPT 63 in 0.8–2.0 M Gdn, suggested by the data of far-UV CD (Fig. 3) and steady-state fluorescence, prove the assumption that partial unfolding of MPT63 in this range of Gdn concentrations is accompanied by the increase in the mobility of Trp residues.

3.5. Phosphorescence study at 77 K in a suitable cryosolvent

When the structural aspects of a protein in the presence of denaturant are being investigated, it is necessary to address the overall change of protein conformation as well as the change of the micro-environment of individual Trp residues, if possible. The information about the conformational variation of proteins could be obtained from the steady-state and time-resolved fluorescence and the near-UV CD spectrum, but these methods could not be employed to characterize the zone-wise perturbation around each Trp residue for a multi-Trp protein. This is due to the fact that specific assignments of the individual Trp have been found to be almost impossible in the case of broad fluorescence spectra of a multi-Trp protein. Time-resolved fluorescence studies also can not be employed for this purpose as even a protein containing a single Trp residue exhibits two or more lifetime components. The use of decay-associated spectra (DAS) and time-resolved emission spectra (TRES) using time-resolved data at various wavelengths are not completely helpful as the contributions of other Trp residues could not be ignored. This problem could be solved by carrying out the experiments with Trp substituted mutated proteins where one or more Trp is replaced by either tyrosine (Tyr) or phenyl alanine (Phe), provided that site-directed mutation does not alter the protein structure.^43,45

Studies of the LTP of Trp residues in proteins revealed that several multi-Trp proteins exhibit overlapping phosphorescence spectra with resolved multiple (0,0) bands corresponding to different Trp residues present in the protein due to the widely different environment of Trp residues and inefficient energy transfer among Trp residues.²⁷ In general, the blue shifted (0,0) band results from solvent-exposed Trp residues with a polar environment.^{27 and references therein} The red shifted (0,0) band corresponds to a more rigid, hydrophobic and less solvent-exposed Trp residue.²⁷ The internal Stark effect arising from nearby charged residue/residues also affects the position of the (0,0) band.^{27 and references therein}

Thus LTP of protein can be exploited to study Gdn-induced unfolding in order to investigate the contribution of individual Trp residues toward phosphorescence along with the perturbation of the local environment of Trp residues. MPT63 in the native state at pH 7.5 exhibits four (0,0) vibronic bands at 406.0 nm, 412.2 nm, 417.4 nm and 421.6 nm in the overlapping LTP spectra corresponding to Trp 82, Trp 48, Trp 129 and Trp 26 respectively.²⁷ The assignment was made with the help of phosphorescence spectra of a mutant of MPT63 viz. W26F and theoretical calculation of the relative efficiency of energy transfer in the wild-type as well as in the generated structure of W26F. The four Tyr residues, however, were found to remain silent at pH 7.5 with λ_exc = 280 nm.²⁷

Phosphorescence spectra with the gradual change of Gdn from 0.0 M to 3.0 M are shown in Fig. 6 and Table 3. The spectra suggest that the phosphorescence remains unaltered in the Gdn range 0.0–0.8 M (Fig. 6). The highest energy (0,0) band at 406.0 nm observed for Trp 82 in the native state is not well resolved in the range 1.2–2.0 M Gdn. In the same range of concentration of Gdn, the second (0,0) band corresponding to Trp 48 is only slightly blue shifted from 412.2 nm to 411.4 nm with the relative intensity remaining practically unaltered (Fig. 6 and Table 3). This indicates that the micro-environment of Trp 48 is well conserved in comparison to the other Trp residues during Gd-induced unfolding. The slight blue shift is, however, suggestive of a somewhat more polar environment or solvent accessibility. The third (0,0) band corresponding to Trp 129 is well affected in this range of concentration of Gdn as the band at 417.4 nm gets gradually quenched. One could thus predict that Trp 129 is exposed to some residues, which can effectively quench tryptophan fluorescence. The lowest energy and the weakest fourth (0,0) band corresponding to Trp 26 remains unaltered up to 1.6 M. However the intensity completely diminishes at 2.0 M. The fully unfolded state is reached in the presence of 2.4 M Gdn where all four Trp residues give rise to a single blue shifted (0,0) band at 410.0 nm with larger bandwidths (bandwidths are shown in Table 3). The lifetime of Trp phosphorescence at various stages of unfolding is also indicated in Table 3.

Fig. 6 Phosphorescence spectra of 5 μM WT MPT63 with different Gdn concentrations in 40% EG–buffer mixture at 77 K. λ_exc = 280 nm. Excitation and emission band pass = 10 nm and 1 nm respectively.

Table 3 Phosphorescence data for WT with guanidine hydrochloride in 40% EG–buffer mixture at 77 K (λ_exc = 280 nm)

Medium	Width of the phosphorescence (0,0) band at half maxima (cm^−1)	Position of the (0,0) band (nm)^e	Phosphorescence lifetime (s) at 77 K monitoring (0,0) band				Assignment of (0,0) band	^eRatio of IW48 and IW129
			τ1 (s)	τ2 (s)	τav (s)	χ^2
a λexc = 280 nm.b λexc = 295 nm.c λexc = 300 nm.d λexc = 305 nm.e Error in the measurement of ±0.2 nm.
0 M Gdn	—	406.0 (sh)	6.4 (2.58%)	2.8 (97.40%)	2.9	1.01	Trp 82	1.26^a, 1.17^b, 1.02^c, 0.74^d
	140	412.2	6.4 (100%)	—	6.4	1.12	Trp 48
	170	417.4	6.4 (27.30%)	4.1 (72.70%)	4.7	1.03	Trp 129
	—	421.6	—	—	—	—	Trp 26
+1.6 M Gdn	200	411.4	5.25 (100%)	—	5.2	1.03	Trp 48	1.59^a, 1.58^b, 1.62^c, 1.65^d
	—	417.8	5.25 (18.1%)	2.5 (81.9%)	2.99	1.12	Trp 129
	—	421.8	—	—	—	—	Trp 26
+2.4 M Gdn	380	410.4	4.73 (100%)	—	—	1.19	—	—
+3.0 M Gdn	390	410.0	4.5 (100%)	—	—	1.10	—	—

---

In order to find the degree of overall solvent exposure of Trp residues in MPT63 in its fully unfolded state, we presented the phosphorescence spectra of pure tryptophan and that of a single Trp residue in a tripeptide and a heptapeptide under similar experimental conditions. The (0,0) band of pure tryptophan in a 40% EG–buffer matrix appears at 405.6 nm while the (0,0) bands of the tripeptide containing lysine (Lys)–Trp–lysine (Lys) and the heptapeptide containing Tyr–serine (Ser)–proline (Pro)–Trp–threonine (Thr)–asparagine (Asn)–Phe are somewhat red shifted and appear at 406.2 nm and 408.0 nm respectively at 77 K having a similar overall structure of the spectra (Fig. 7). In Lys–Trp–Lys the Trp residue is in between the two same polar residues lysine (Lys) whereas the Trp residue in heptapeptide resides in the midst of the six amino acid residues, viz., four polar residues of tyrosine (Tyr), serine (Ser), threonine (Thr), and asparagine (Asn) and two non-polar residues of proline (Pro) and phenylalanine (Phe). These results imply that the overall solvent exposure for all the Trp residues in MPT63 is less than that experienced by free Trp or the Trp residue in the said peptides in the 40% EG–buffer matrix. This is also evident from the fact that most of the solvent-exposed residue Trp 82 appears at 406.0 nm in the native state of MPT63.

Fig. 7 Phosphorescence spectra of L-Trp, Lys–Trp–Lys, heptapeptide in a 40% EG–phosphate buffer mixture at 77 K, λ_exc = 280 nm. Excitation band pass = 10 nm and emission band pass = 1.5 nm. Concentration of protein in each case = 10⁻⁵ M.

Since the phosphorescence of Tyr usually appears in the region 350–400 nm ³⁰ using λ_exc = 280 nm, LTP spectra of Gdn-induced MPT63 with λ_exc = 280 nm also gives an idea about the perturbation of Tyr residues. The four Tyr residues (Tyr 3, Tyr 41, Tyr 78 and Tyr 106) are found to be absent in the LTP spectra of the native state at pH 7.5 indicating efficient energy transfer from Tyr residues to Trp residues. However, when Gdn reaches 1.6 M, the emission in the region 350–400 nm due to Tyr residues is observed (Fig. 6). An increase in the distance between Trp residues and Tyr residues and/or a large separation between Arg residues and Tyr residues in the unfolding state may lead to prevention of efficient ET from Tyr to Trp. Furthermore, it could be also predicted that the electric field of charged Arg residues could not influence the electron transfer process from Tyr 78 and Tyr 106 in the unfolded state.

Fig. 8 demonstrates the LTP spectra of the partially unfolded state of MPT63 with 1.6 M Gdn as a function of excitation wavelength. The intensity ratio of Trp 48 and Trp 129 does not change significantly as one moves from λ_exc = 270 nm to λ_exc = 305 nm, whereas in the native state the intensity ratio changes drastically with the change in λ_exc²⁷ (Table 3). The intensity pattern observed for Trp 48 and Trp 129 as a function of excitation wavelength is consistent with the more conserved environment of Trp 48 compared to that of Trp 129 at a similar stage of unfolding induced by Gdn. The data also suggest the less rigid nature of the environment of Trp 129 compared to that in the native state.

Fig. 8 Phosphorescence spectra of 5 μM WT MPT63 with 1.6 M Gdn in 40% EG–buffer mixture at 77 K with different excitation wavelength. Excitation band pass = 10 nm and emission band pass = 1 nm in each case except with λ_exc = 305 nm, where the emission band pass = 2.5 nm.

The structure of the folding pathway of the cellular retinoic acid binding protein (CRABP1) containing three Trp residues located at 7, 87 and 109 has been studied using stopped flow fluorescence and CD with the help of three single Trp mutants where Trp residues have been replaced by Phe residues.^43,44 It is to be noted that the optical resolution of all four Trp residues in the phosphorescence spectra allows us to detect the perturbation of the environment of each Trp residue in MPT63 during the stepwise unfolding pathway without using any Trp substituted mutant.

4. Conclusions

The steady-state fluorescence, time-resolved fluorescence, the rotational correlation time obtained from anisotropy decay and the CD spectra indicate the two-state mechanism of unfolding of MPT63 at pH 7.4. Simple inexpensive LTP in a suitable cryosolvent at the micromolar level concentration is able to detect zone-wise perturbation of the environment surrounding each Trp residue during the unfolding process. It is most interesting to note that the buried Trp 129 residue in the native protein is perturbed while the partially buried and partially solvent-exposed residue (Trp 48) retains the conserved environment at the similar unfolding stage before complete denaturation. Trp 82 which is completely solvent-exposed, however, gets perturbed at the initial stage of unfolding. In the completely unfolded state, the Trp residues in MPT63 experience less solvent exposure compared to that for a free tryptophan or a Trp residue in the two shorter peptides studied here. The Tyr residues in MPT63, which are silent in the native state, exhibit emission from the lowest triplet state at 1.6 M Gdn implying a change of environment around Tyr residue/residues before complete unfolding. This kind of specific information regarding an individual Trp environment during stepwise unfolding in the case of a multi-Trp protein without using any Trp substituted mutant is not available in the current literature. We are able to address this problem because of optical resolution of all four Trp residues in the native state of MPT63 at pH 7.5 due to its inherent architecture.

Acknowledgements

S. G. gratefully acknowledges DST, Govt. of India (SR/S1/PC-57/2008 and SB/S1/PC-003/2013) and CSIR, Govt of India (No. 21(0871)/11/EMR-II) for financially supporting this work. M. M. and R. G. thank CSIR, GOI (08/155(0046)/2013-EMR-I), DST, GOI (No. SR/S1/PC-57/2008), for SRF fellowship. K. C. acknowledges CSIR for Network Project (UNSEEN, BSC-0113) grant from CSIR.

References and notes

1. N. C. Pace, Trends Biochem. Sci., 1990, 15, 14.
2. N. C. Pace and C. Tanford, Biochemistry, 1968, 7, 198.
3. C. Tanford, Adv. Protein Chem., 1968, 23, 122.
4. R. Jaenicke and R. Rudolph, Folding proteins, in Protein structure: a practical approach, ed. T. E. Creighton, IRL Press at Oxford University Press, Oxford, 1989, p. 117.
5. K. E. Neet and D. E. Timm, Protein Sci., 1994, 3, 2167.
6. D. Galani and R. K. O. Apenten, Food Res. Int., 1999, 32, 93.
7. A. L. Fink, Folding Des., 1998, 3, R9.
8. R. Wetzel, Trends Biotechnol., 1994, 12, 193.
9. R. W. Carrell and B. Gooptu, Curr. Opin. Struct. Biol., 1998, 8, 799.
10. J. D. Harper and P. T. Lansbury Jr, Annu. Rev. Biochem., 1997, 66, 385.
11. M. A. Speed, D. I. Wang and J. King, Nat. Biotechnol., 1996, 14, 1283.
12. P. L. Privalov, Protein Folding, T. E. Creighton, W. H. Freeman & Co., New York, p. 83, 1992.
13. M. Karplus and E. Shaknovich, Protein Folding, T. E. Creighton, W. H. Freeman & Co., New York, p. 127, 1992.
14. H. Maity, J. N. Rumbley and S. W. Englande, Proteins: Struct., Funct., Bioinf., 2006, 63, 349.
15. MINOES: a new approach to select a representative ensemble of structures in NMR studies of (partially) unfolded states. Application to Δ25-PYP, M. Krzeminski, G. Fuentes, R. Boelens and A. M. J. J. Bonvin.
16. V. L. Arcus, S. Vuilleumier, S. M. V. F. Bycroft and A. R. Fersht, J. Mol. Biol., 1995, 254, 305.
17. K. Sridevi and B. J. Udgaonkar, Biochemistry, 2003, 42, 1551.
18. A. Berglunda, A. C. Brorssonb, B. H. Jonssonc and I. Sethson, J. Magn. Reson., 2005, 172, 24.
19. L. Hoang, H. Maity, M. M. G. Krishna, Y. Lin and S. W. Englander, J. Mol. Biol., 2003, 331, 37.
20. J. M. J. Pérez, J. G. Renisio, J. J. Prompers, C. J. V. Platerink, C. Cambillau, H. Darbon and L. G. J. Frenken, Biochemistry, 2001, 40, 74.
21. A. L. Cortajarena, S. G. J. Mochrie and L. Regan, J. Mol. Biol., 2008, 379, 617.
22. V. L. Arcus, S. Vuilleumiert, S. M. V. Freund, M. Bycroft and A. R. Fersht, Proc. Natl. Acad. Sci. U. S. A., 1994, 91, 9412.
23. C. Manca, K. Lyashchenko, H. G. Wiker, D. Usai, R. Colangeli and M. L. Gennaro, Infect. Immun., 1997, 65, 16.
24. B. L. Wang, Y. Xu, Z. M. Li, Y. M. Xu, X. H. Weng and H. H. Wang, International Journal of Tuberculosis and Lung Disease, 2005, 12, 1327.
25. S. Nagai, H. G. Wiker, M. Harboe and M. Kinomoto, Infect. Immun., 1991, 59, 372.
26. C. W. Goulding, A. Parseghian, M. R. Sawaya, D. Cascio, M. I. Apostol, M. L. Gennaro and D. Eisenberg, Protein Sci., 2002, 11, 2887.
27. R. Ghosh, M. Mukherjee, K. Chattopadhyay and S. Ghosh, J.Phys. Chem. B, 2012, 116, 12489.
28. S. Ghosh, L. H. Zang and A. H. Maki, J. Chem. Phys., 1988, 88, 2769.
29. L. H. Zang, S. Ghosh and A. H. Maki, Biochemistry, 1989, 28, 2245.
30. S. Ghosh, L. H. Zang and A. H. Maki, Biochemistry, 1988, 27, 7816.
31. S. J. Remington, W. F. Anderson, J. Owen, L. F. Ten Eyck, C. T. Grainger and B. W. Matthews, J. Mol. Biol., 1978, 118, 81.
32. W. C. Lam, A. H. Maki, T. Itoh and T. Hakoshima, Biochemistry, 1992, 31, 6756.
33. T. Hakoshima, T. Itoh, K. Tomita, K. Gohda, S. Nishikawa, H. Morioka, W. Uesugi, E. Ohtsuka and M. Ikehara, J. Mol. Biol., 1992, 223, 1013.
34. M. Ferreras, J. G. Gavilanes, C. López-Otín and J. M. García-Segura, J. Biol. Chem., 1995, 270, 28570.
35. FELIX 32, Operation Manual, Version 1.1, Photon Technology International, Inc., Lawrenceville, NJ, 2003.
36. J. R. Lakowicz, Principles of Fluorescence Spectroscopy, Kluwer/Plenum Press, New York, 1999.
37. R. Ghosh, S. Sharma and K. Chattopadhyay, Biochemistry, 2009, 48, 1135.
38. M. Mukherjee, R. Ghosh, K. Chattopadhyay and S. Ghosh, J. Biomol. Struct. Dyn., 2015, 33(10), 2145.
39. S. P. McGlynn, T. Azumi and M. Kinosita, Molecular Spectroscopy of the Triplet State, Prentice-Hall, Englewood Cliffs, NJ, 1969.
40. W. E. Lee, Ph.D. thesis, McGill University, Montreal, Quebec, Canada, 1985.
41. R. Ghosh, Ph.D. thesis, University of Calcutta, 2013.
42. B. Somogyi, S. P. A. Rosenberg, I. Seres, J. Matkd, G. R. Welch and P. Nagyl, Biochemistry, 1985, 24, 6674.
43. P. L. Clark, Z.-p Liu, J. Zhang and L. M. Crierasch, Protein Sci., 1996, 5, 1108.
44. P. L. Clark, B. F. Watson and L. M. Crierasch, Protein Struct. Anal., Folding Des., 1998, 3, 401.
45. S. Ghosh, M. Petrin and A. H. Maki, Biophys. J., 1986, 49, 753.

---