NCS powder consisting apoNCS and NCS-Chr in a 1:1 molar ratio was a gift from Kayaku Co., Ltd., Itabashi-Ku, Tokyo, Japan. Extracted NCS-Chr was obtained after repeated methanol extractions of lyophilized NCS stock (0.5 mM, as determined by ε₃₄₀ = 10,800 M^-1 cm^-1) in 20 mM sodium citrate (pH 4), following the previously described method 18. Extracted NCS-Chr was stored at -80°C in amber glass vials. Integrity and concentration of NCS-Chr were examined by UV spectroscopy and HPLC analysis. Labeled ¹⁵NH₄Cl and D₂O were purchased from Cambridge Isotope Laboratories (Andover, MA, USA). All other chemicals used were of high quality analytical grade. All experiments were performed in 10 mM phosphate buffer (pH 3) at 25°C.

Recombinant apoNCS was overexpressed and purified by the procedure as described 19. Homogeneity of the protein was examined by UV, HPLC and SDS-PAGE. Protein yield was about 2 mg/L. Molecular mass of the purified protein was verified using ESI-Mass analysis.

Uniform ¹⁵N labeling was achieved by the established procedure 19. E. coli BL21 Codon Plus strain (Stratagene, La Jolla, CA, USA) carrying the apoNCS gene was cultured in M9 minimal medium containing ¹⁵NH₄Cl supplemented with vitamin B₁. Final yield of the ¹⁵N-labeled protein was about half of the corresponding unlabeled protein expressed in LB medium.

Fluorescence spectra were collected using a Hitachi F-4500 spectrofluorimeter at a resolution of 2.5 or 10 nm. For GdnHCl-induced unfolding study, excitation wavelength was set at 280 nm. Intrinsic fluorescence measurements were made at 25°C. Binding affinity of 1-anilino-8-napthalene sulfonate (ANS) to apoNCS at various concentrations of GdnHCl was monitored in a wavelength range of 375 to 625 nm using an excitation wavelength at 355 nm. Excitation and emission bandwidths were set at 5 nm. Concentration of ANS and protein was 100 μM and 10 μM respectively. All samples were prepared in 10 mM phosphate buffer at pH 3.

All CD measurements are carried out on a Jasco J-715 spectropolarimeter (Tokyo, Japan) equipped with a circulating water bath (Neslab, model RTE-140) (Portsmouth, NH, USA). Measurements were made using a 0.1 cm path-length water-jacketed quartz cell. Each spectrum represents an average of 30 scans with a scan speed of 50 nm/min. Concentration of the protein used was 15 μM. Background corrections were made in all spectra. Bandwidth was set to 1 nm and all spectra were acquired at 25°C.

Gel-filtration experiments were carried out at 25°C on a superdex-100 column using a Pharmacia AKTA FPLC chromatographic device (GE Healthcare Bio-Sciences AB, Uppsala, Sweden). Column was equilibrated with 2 bed volumes of 10 mM phosphate buffer (pH 3) containing appropriate concentrations of GdnHCl at a flow rate of 1 ml/min. Concentration of protein used for each analysis was approximately 250 μg/ml (dissolved in appropriate concentrations of GdnHCl). Protein elution was monitored by UV absorbance at 280 nm.

Thermal stability of apoNCS in 10 mM phosphate buffer (pH 3) in the presence or absence of 1.2 M GdnHCl was monitored by far-UV CD at 224 nm. Changes in ellipticity with temperature were followed from 5–91°C at an increment of 3°C. Experiments were performed using a water-jacketed cell connected to a thermal circulator equipped with a microprocessor and temperature sensor. Protein sample (25 μM) was allowed to equilibrate for 10 minutes at each temperature before data acquisition.

The NMR experiments were carried out on a Bruker DMX 600 MHz NMR spectrometer (Rheinstetten, Germany) at 25°C. A 5 mm inverse probe with a self-shielded z-gradient was used to obtain all gradient-enhanced ¹H-¹⁵N HSQC spectra 2021. ¹⁵N decoupling during acquisitions was achieved using the GARP sequence 22. Total of 2048 complex data points were collected in the ¹H-dimension of the ¹H-¹⁵N HSQC experiments. In the indirect ¹⁵N-dimension spectra, 512 complex data points were collected. The HSQC spectra were recorded by 32 scans at all concentrations of GdnHCl. ¹⁵N chemical shifts were referenced using consensus ratio of 0.0101329118. All spectra were processed on a Silicon Graphics workstation using XWINNMR and Sparky softwares.

Binding experiments were performed in 10 mM phosphate buffer (pH 3) by incubating 20 μM apoNCS (in the presence or absence of 1.2 M GdnHCl) with NCS-Chr at 1:1 molar ratio for 30 minutes at 25°C. The final methanol content introduced from NCS-Chr stock was kept minimal to about 4% (v/v). Analyses of the protein-bound NCS-Chr after binding experiments were performed through a Waters μ-Bondapak reverse phase C₁₈ column by a Waters Millennium HPLC equipped with a model 600E solvent delivery system, a 996 photodiode array detector and either a Waters 474 or a Jasco FP-1520 fluorescence detector following previously described method 23.

GdnHCl-induced equilibrium unfolding of apoNCS at pH 3 was monitored by changes in intrinsic tryptophan fluorescence and far-UV CD at 224 nm. Circular dichroism spectrum of the all β-sheet apoNCS protein is somewhat atypical 24 with a large positive maximum at 224 nm and small negative minimum centered at 212 nm 7. Fluorescence and far-UV CD spectral probes have been shown to reliably report the gross tertiary and secondary structural changes that possibly occur during the unfolding of apoNCS 11. Fig. 2 shows that GdnHCl-induced equilibrium unfolding of apoNCS under acidic condition (pH 3) monitored by changes in intrinsic tryptophan fluorescence is completely reversible. The fluorescence unfolding profile of apoNCS shows that the protein starts to unfold in 0.6 M GdnHCl. Unfolding of the protein is completed beyond 3 M GdnHCl (Fig. 2). C_m (concentration of GdnHCl at which 50% of the protein molecules are in denatured state) and 'm' (measure of cooperativity of unfolding process) are 1.1 ± 0.1 M and 7.3 ± 0.7 kJ mol^-1 M^-1 respectively. Change in free energy of unfolding obtained in the absence of the denaturant [ΔG (H₂O)] is estimated to be 8.5 ± 0.1 kJ mol^-1. Interestingly, the GdnHCl-induced unfolding curve obtained by measuring ellipticity changes at 224 nm is not superimposable with that obtained by monitoring the changes in intrinsic tryptophan fluorescence. These results possibly suggest that the unfolding of apoNCS may not follow a two-state model, and may involve accumulation of stable intermediate states (Fig. 2). C_m, 'm' and [ΔG (H₂O)] estimated from the unfolding profile obtained using far-UV CD are 1.8 ± 0.1 M, 4.4 ± 0.2 kJ mol^-1 M^-1and 8.6 ± 0.5 kJ mol^-1 respectively.

Size-exclusion chromatography (SEC) is a useful technique to acquire information on integral changes of molecular dimensions under the effect of the denaturant 2526. This technique has been successfully used to identify and obtain hydrodynamic data on stable intermediates in the folding/unfolding pathway of proteins 25. ApoNCS in 10 mM phosphate buffer (pH 3) in the absence of GdnHCl elutes as a single peak with an elution volume of ~129 ml (Fig. 3). On increasing the concentration of GdnHCl from 0 to 0.2 M, the elution volume of the protein peak sharply decreases to about ~111 mL. Interestingly, elution volume of the protein does not significantly change when the GdnHCl concentration is increased from 0.4 M to 1.2 M (Fig. 3). A clear plateau can be observed in the plot of elution volume versus concentration of GdnHCl suggesting the accumulation of a stable equilibrium intermediate in this range of denaturant concentration. Beyond 1.2 M GdnHCl, elution volume of the protein changes drastically and profile eventually reaches a plateau at 2 M GdnHCl. It should be mentioned that apoNCS elutes as a single peak at various concentrations of GdnHCl indicating that the protein exists as a single population at all concentrations of the denaturant (Fig. 3, inset). Uversky 26, using size-exclusion chromatography to study the denaturant-induced equilibrium unfolding profile of proteins, showed that proteins such as β-lactamase, bovine carbonic anhydrase and β-lactoglobulin, which denature through a MG-like intermediate, also elute as single peaks at all concentrations of the denaturant 26. The results of the SEC-FPLC experiment clearly show that the GdnHCl-induced unfolding of apoNCS proceeds through the accumulation of a stable intermediate at around 1.2 M GdnHCl at pH 3 (Fig. 3). At pH above 3 no stable intermediate was observed. The first phase of apoNCS unfolding between 0–1.2 M GdnHCl appears to represent an equilibrium transition between the native and the intermediate state.

Under optimal conditions (pH 3, 1.2 M GdnHCl) where the stable unfolding intermediate of apoNCS accumulates, considerable percentage (about 80%) of the secondary structure remains undisrupted, whereas 60% of the tertiary structure is disrupted (Fig. 2). This is a good indication that the stable intermediate could be in MG-like state. ANS is a fluorescent dye that binds to the hydrophobic regions of proteins 27 and this fluorescent probe has been immensely useful in the identification of equilibrium intermediates such as MG states. MG intermediates usually display significant exposure of hydrophobic cores to the solvent. Therefore, ANS binds strongly to MG states and fluoresces intensely 2829. The dye generally exhibits weak binding affinity to the native and denatured proteins 28. The binding affinity of apoNCS to ANS at various concentrations of GdnHCl was monitored by changes in the emission intensity at 475 nm (Fig. 4). The ANS emission intensity upon binding to apoNCS in 1.2 M GdnHCl is significantly higher than that observed with the protein in its folded state (Fig. 4). Further increase in the concentration of GdnHCl (beyond 1.2 M) results not only in the progressive decrease in the emission intensity but also in a continuous red shift in the emission maxima. The protein in its unfolded conformations, at and beyond 3 M GdnHCl, exhibits weak binding to ANS (Fig. 4). Thus, the results of ANS binding and SEC-FPLC experiments, analyzed in conjunction, clearly suggest that a MG-like intermediate accumulates maximally in 1.2 M GdnHCl.

It is important to understand the stability of the equilibrium unfolding intermediate states in relation to the folded state of the protein. In this context, we carried out thermal denaturation of apoNCS in the presence and absence of 1.2 M GdnHCl at pH 3 (Fig. 5). Thermal unfolding monitored by ellipticity changes at 224 nm shows that apoNCS at pH 3 unfolds to denatured state at temperatures beyond 62°C (Fig. 5). Apparent T_m (temperature at which 50% of the protein molecules exist in the denatured state) of the unfolding reaction is about 52 ± 1°C. In marked contrast, the MG-like state of the protein at the same pH level in 1.2 M GdnHCl unfolds almost completely beyond 50°C (Fig. 5). T_m of the MG-like state to denatured state of apoNCS is 35 ± 1°C. These results suggest that the protein in the MG-like state is significantly less stable than the folded state. It appears that some of the interactions stabilizing the compactly folded conformation are disrupted in MG-like state.

Two-dimensional ¹H-¹⁵N heteronuclear single quantum coherence (HSQC) NMR spectroscopy serves as a fingerprint of the conformational state of a protein. It provides residue level information on the structural changes that possibly occur during unfolding/folding processes. We synthesized and purified ¹⁵N-labeled apoNCS and monitored GdnHCl-induced unfolding by ¹H-¹⁵N chemical shift perturbation. Heteronuclear correlation experiments have been shown to be very sensitive because of high magnetization transfer between directly bonded nuclei. Fig. 6A shows that 2D ¹H-¹⁵N HSQC spectrum of apoNCS at pH 3 in the absence of GdnHCl is well-dispersed, which is a characteristic feature of a folded protein. Striking differences can be discerned in the HSQC spectra obtained in the absence and presence of 1.2 M GdnHCl. Many cross peaks in HSQC spectrum of the protein in 1.2 M GdnHCl are broadened, suggesting the overall flexibility of residues in the intermediate state (in 1.2 M GdnHCl) is significantly higher than that in the folded state (Fig. 6A). Though it is not easy to identify residues in the partially denatured state, we attempted to make an assignment by recording a series of HSQC NMR spectrum of apoNCS at pH 5 to 3 with 0.5 decrement of pH level. ¹H and ¹⁵N backbone chemical shifts of native apoNCS at pH 5 were assigned based on the reported values (105 of 113 residues of apoNCS were reported) 30. Once the cross peaks of apoNCS spectrum obtained at pH 3 were assigned, they were compared with those recorded at pH 3 in the presence of GdnHCl. By stepwise comparison, we identified 70 residues in the spectrum of MG-like state of apoNCS. The weighted average (of ¹⁵N and ¹H) chemical shift perturbation ((Δδ) = [(δH)² + 0.2 (δ¹⁵N)²)]^1/2) of residues in MG-like state of apoNCS is shown in Fig. 6B, in which the unassigned residues (residues 1–3, 5, 8, 9, 15, 19, 24, 28, 31, 38–44, 46, 48, 49, 53, 54, 56, 59, 62, 66, 75, 78, 79, 81, 82, 87, 89, 92, 93, 97, 99, 100, 103, 105, 106, 112) as well as the residues with perturbation below 0.1 ppm (the threshold of the average minimal perturbation) were eliminated. The result clearly shows that residues at N- and C-termini of the protein are significantly perturbed in the MG-like state (Fig. 6B). Most of the residues that show significant chemical shift changes or peak broadening appear to be those not involved in the secondary structure formation (Fig. 6A and 6B). HSQC spectra of apoNCS acquired beyond 4 M GdnHCl show significant decrease in chemical shift dispersion in ¹H dimension indicating complete unfolding of the protein (Fig. 6A). The results of the NMR experiments suggest that although the interactions at loop regions and both N- and the C-termini of the protein are perturbed, the secondary structural elements in the protein appear to be mostly unaffected in the equilibrium intermediate in 1.2 M GdnHCl.

NCS manifests its antineoplastic effects through damage to the cellular genome 1. The very potent cytotoxic activities come from the enediyne warhead NCS-Chr, which is very labile by itself 3. At pH 8, the life time of free NCS-Chr in an aqueous environment is only few seconds 4. Binding with apoNCS increases the life time of NCS-Chr to several thousand folds 4. Thus, the degree of protection against spontaneous NCS-Chr degradation can serve as a good probe to assess the binding affinity towards NCS-Chr 431. To examine whether apoNCS in MG-like state binds to NCS-Chr, an aliquot of NCS-Chr was incubated with 1:1 molar ratio of apoNCS at pH 3 and 25°C for 30 min in the presence and absence of 1.2 M GdnHCl. The remaining amount of intact NCS-Chr was analyzed by HPLC following the previously established method 23. The results clearly show that NCS-Chr was fully protected in the presence of either folded or MG-like state of apoNCS (Table 1). Evidently, the MG-like state of apoNCS in 1.2 M GdnHCl retains its full ability to bind to its enediyne chromophore.

Titration of NCS-Chr into native state apoNCS can produce drastic changes in the near-UV CD spectrum 1932. Binding of 1:1 molar ratio of NCS-Chr with native apoNCS produces a prominent negative ellipticity peak minimum at 255 nm, which is a characteristic feature of the native form of holoNCS. Interestingly, titration of 1:1 molar ratio of NCS-Chr into MG-like state of apoNCS in 1.2 M GdnHCl at pH 3 produces identical CD spectrum to that of the native holoNCS (Fig. 7). The results unambiguously suggest that NCS-Chr not only can bind to apoNCS in partially unfolded MG-like state but also help the protein fold back into its native state. Without NCS-Chr, the near-UV CD spectroscopy of the MG-like state of apoNCS shows that 60% of the tertiary structure is disrupted (Fig. 7, inset). The observation is consistent with that from intrinsic tryptophan fluorescence study (Fig. 2).