This Article Abstract Full Text (PDF) Supplemental All Versions of this Article: biophysj.106.092155v1 92/6/2090    most recent Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via Google Scholar Google Scholar Articles by Xie, T. Articles by Wang, J. Search for Related Content PubMed PubMed Citation Articles by Xie, T. Articles by Wang, J.

Folding Stability and Cooperativity of the Three Forms of 1–110 Residues Fragment of Staphylococcal Nuclease

National Laboratory of Biomacromolecules, Center for Structural and Molecular Biology, Institute of Biophysics, Chinese Academy of Sciences, Beijing 100101, People's Republic of China

Correspondence: Address reprint requests to Jinfeng Wang, Tel.: 86-10-6488-8490(O); Fax: 86-10-6487-2026; E-mail address: jfw{at}sun5.ibp.ac.cn.

	ABSTRACT

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION CONCLUSIONS SUPPLEMENTARY MATERIAL ACKNOWLEDGEMENTS REFERENCES

 
Folding stability and cooperativity of the three forms of 1–110 residues fragment of staphylococcal nuclease (SNase110) have been studied by various biophysical and NMR methods. Samples of G-88W- and V-66W-mutant SNase110, namely G-88W110 and V-66W110, in aqueous solution and SNase110 in 2.0 M TMAO are adopted in this study. The unfolding transitions and folded conformations of the three SNase fragments were detected by far- and near-ultraviolet circular dichroism and intrinsic tryptophan fluorescence measurements. The tertiary structures and internal motions of the fragments were determined by NMR spectroscopy. Both G-88W and V-66W single mutations as well as a small organic osmolyte (Trimethylamine N-oxide, TMAO) can fold the fragment into a native-like conformation. However, the tertiary structures of the three fragments exhibit different degrees of folding stability and compactness. G-88W110 adopts a relatively rigid structure representing a most stable native-like ß-subdomain conformation of the three fragments. V-66W110- and TMAO-stabilized SNase110 produce less compact structures having a less stable "ß-barrel" structural region. The different folding status accounts for the different backbone dynamic and urea-unfolding transition features of the three fragments. The G-20I/G-29I-mutant variants of the three fragments have provided the evidence that the folding status is correlated closely to the packing of the ß-strands in the ß-barrel of the fragments. The native-like ß-barrel structural region acts as a nonlocal nucleus for folding the fragment. The tertiary folding of the three fragments is initiated by formation of the local nucleation sites at two ß-turn regions, I-18–D-21 and Y-27–Q-30, and developed by the formation of a nonlocal nucleation site at the ß-barrel region. The formation of ß-barrel and overall structure is concerted, but the level of cooperativity is different for the three 1–110 residues SNase fragments.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION CONCLUSIONS SUPPLEMENTARY MATERIAL ACKNOWLEDGEMENTS REFERENCES

 
The question "how do proteins fold in vivo and in vitro" has been under intensive study for several decades. It was proposed (1) that small proteins of molecular mass 10–15 kDa fold the same way in vitro as they do in vivo. As has been well established, in vivo proteins are slowly biosynthesized on the ribosome from the N-terminus of the amino acid sequence, and there is a host of other molecular factors in the cell, such as large molecular chaperones and small chemical chaperones that play diverse supportive roles in protein folding. In vitro, the investigation of protein folding has been opened up to molecular and atomic level resolutions, centering on folding mechanisms, pathways, rates, and the driving forces of folding. Three classical hypotheses for the mechanisms of folding (2) have been proposed which put more emphasis on specific structures and pathways. The chemical physics view portrays folding as the motion of the polypeptide chain on a partially rugged, funnel-shaped energy landscape as it searches conformational space on the way to its unique native structure. On a funnel-like energy surface, proteins must pay an entropic price before the downhill tendency of the energy landscape can be manifested (3).

Staphylococcal nuclease (SNase) is composed of - and ß-subdomains and is a widely used model system for the study of protein folding. Many early studies focused on the thermodynamic and kinetic analysis of SNase folding and unfolding transitions and on the structural analysis of SNase in nonnative states. Abundant experimental data of the folding intermediates, folding stabilities, and residual structures of nonnative SNase have been accumulated. Various large SNase fragments and their mutants have been adopted for exploring the folding state and folding process of the molecule before folding to a stable structure (4–9). For the 1–102 residues SNase fragment, the G-88W and V-66W single mutation cannot drive the overall folding of the fragments G-88W102 and V-66W102, but only can make around 30–40% of the molecule to be possibly structured (4). However, the G-88V/V-66L double mutation can drive the 1–103 residues SNase fragment folding into a stable "OB-fold" subdomain (5). Fragment 131, consisting of residues 1–3 and 13–140, was found in a denatured but not random coil state (6). Chemical cleavage studies of SNase fragment 1–135 and its mutants revealed a native-like state of segment 1–104 and a disordered C-terminal region (7). Investigation of 136-amino acid polypeptide of SNase (SNase136) and V-66W-mutant SNase136 (V-66W136) showed the similar conformational features which exhibited the folded ß-subdomain and a disordered C-terminal region (8). Examination of the folding ability and stability of the fragments 1–140 and 1–141 indicated that both SNase fragments exhibited a native-like tertiary folding and W-140 is a key residue for the structure and stability of SNase fragments (9). The folding pathways of SNase with chain elongation from residue K-110 to the C-terminal region of the molecule, involving the generation of a "ß-barrel" hydrophobic core and overall architecture of the ß-subdomain, were analyzed by studying the N-terminal large fragments of SNase (10).

The 1–110 residues SNase fragment (SNase110) has a nearly complete sequence of the ß-subdomain, containing a main ß-barrel region and two -helices (helices 1 and 2). In aqueous solution, SNase110 exhibits an ensemble of species containing unfolded and partially folded states (10). We found that the G-88W and V-66W single mutations and a small organic osmolyte, Trimethylamine N-oxide (TMAO), can drive the 110-residue SNase fragment to form a native-like folded conformation. The focus of this study was to determine the folding status and folding mechanism of G-88W110, V-66W110, and SNase110 in 2.0 M TMAO. The tertiary structures, unfolding transitions, and internal motions of the three forms of 1–110 residues SNase fragments were detected using the circular dichroism (CD), fluorescence, and heteronuclear multidimensional NMR spectroscopy for studying the folding and stability of the fragments. The results ought to contribute to a better understanding of the folding mechanism of SNase in vitro.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION CONCLUSIONS SUPPLEMENTARY MATERIAL ACKNOWLEDGEMENTS REFERENCES

 
Sample preparation
The SNase110 fragment containing residues 1–110 of SNase from the V8 strain of Staphylococcus aureus and its mutant fragments, G-88W110 and V-66W110, were expressed and purified according to the procedures described previously (11). The mutant variants [G-20I/G-29I]G-88W110, [G-20I/G-29I]V-66W110, and [G-20I/G-29I]SNase110 were produced by G-20I/G-29I double mutation of the three 110-residue SNase fragments in the corresponding expression plasmids by the QuickChange method (12). Uniformly ¹⁵N- and ¹³C-labeled fragments were obtained through bacterial growth in M9 minimal media using ¹⁵NH₄Cl (1 g/l) and ¹³C-glucose (2 g/l) as the sole nitrogen and carbon sources. The purity of the proteins was checked by sodium dodecylsulfate-polyacrylamide gel electrophoresis to be a single band. Samples of G-88W110, V-66W110, [G-20I/G-29I]G-88W110, and [G-20I/G-29I]V-66W110 in 90% H₂O/10% D₂O containing 1–2.0 mM proteins and 50 mM acetate buffer (pH 4.9) were prepared for NMR experiments. The samples of SNase110 and [G-20I/G-29I]SNase110 for NMR studies were prepared by dissolving 1 mM protein in 90% H₂O/10% D₂O (pH 6.0) containing 2.0 M TMAO (the purity > 99.0%, purchased from Fluka, Milwaukee, WI). Samples of G-88W110, V-66W110, [G-20I/G-29I]G-88W110, [G-20I/G-29I]V-66W110, and samples of SNase110 and [G-20I/G-29I]SNase110 in 2.0 M TMAO containing 60 µM proteins and 20 mM Tris-HCl buffer (pH 7.4) were prepared for far-ultraviolet (UV) CD measurement; 120 µM G-88W110, V-66W110, [G-20I/G-29I]G-88W110, and [G-20I/G-29I]V-66W110 (pH 7.4) were used for near-UV CD measurements. Samples used for denaturation studies by CD measurements were 30 µM G-88W110, V-66W110, SNase110 in 2.0 M TMAO, and their G-20I/G-29I-double mutant variants, and those by fluorescence measurements were 6.0 µM G-88W110, V-66W110, [G-20I/G-29I]G-88W110, and [G-20I/G-29I]V-66W110 (pH 7.4) in the presence of urea at various concentrations ranging from 0 to 6.0 M with increments of 0.25 M. All the fragment samples prepared for denaturation studies were incubated overnight at 20°C before measurement. Protein concentrations were determined by UV absorption at 280 nm.

CD measurements
Far-UV CD (200–250 nm) and near-UV CD (240–300 nm) spectra of the six 110-residue SNase fragments were recorded on a Jasco (Tokyo, Japan) J-720 spectropolarimeter at room temperature. Employed were 1-mm and 10-mm cuvette pathlengths to record the far-UV and near-UV CD spectra, respectively. Five scans were averaged for each measurement. The far-UV CD spectra were used to estimate the contents of secondary structural components in the 110-residue SNase fragments by analysis of the ellipticities using programs in the software package CDPro (http://lamar.colostate.edu/sreeram/CDPro).

Fluorescence measurements
Intrinsic fluorescence emission spectra of the 110-residue SNase fragments were measured using a Hitachi (San Jose, CA) F-4500 fluorescence spectrophotometer at 20°C. The slit width for all the measurements was 5 nm. For the intrinsic fluorescence measurements, the excitation wavelength was 295 nm for the tryptophan residue. The emission intensity of tryptophan at 325 nm was recorded for G-88W110 and V-66W110 and their G-20I/G-29I-double mutant fragments.

Determination of unfolding free energy
The unfolding process was monitored by following the changes in ellipticity and intrinsic fluorescence at wavelengths 222 and 325 nm, respectively. The unfolding free energies of G-88W110, V-66W110, and SNase110 in the presence of urea were obtained using the intensities of these signals, which were normalized to the fraction of unfolded species using the standard relation: F_unf = (I – I_N)/(I_U – I_N), where N and U stand for the fluorescence intensity of the native and fully unfolded species (13). The F_unf values were calculated from the linear extrapolation of the pre- and postunfolding baselines. The unfolding free energy and m values in the relation G = G°(H₂O) + m[D] were obtained from the fitting of the denaturation data to a two-state model using the standard equation (13,14) for all the six fragments, except as otherwise indicated.

NMR spectroscopy
All NMR experiments were run on a Bruker (Billerica, MA) DMX 600 spectrometer equipped with a triple-resonance cryoprobe at 305 K for G-88W110 and at 298 K for V-66W110 and SNase110 in 2.0 M TMAO. For the resonance assignments and NOE and J-coupling constant determinations, the 3D ¹H-¹³C-¹⁵N HNCO, HN(CA)CO, CBCA(CO)NH, HNCACB, H(CCONH), and HNHB, and the 3D ¹H-¹⁵N total correlation spectroscopy (TOCSY)-heteronuclear single-quantum correlation (HSQC) and nuclear Overhauser effect spectroscopy (NOESY)-HSQC experiments (15,16) were performed with the samples containing 2.0 mM G-88W110, V-66W110, and with 1.0 mM SNase110 in 2.0 M TMAO. The mixing times for 3D ¹H-¹⁵N NOESY-HSQC and TOCSY-HSQC experiments of both G-88W110 and V-66W110 were 200 ms and 80 ms, respectively. Mixing times of 150 ms and 55 ms were set for the 3D ¹H-¹⁵N NOESY-HSQC and TOCSY-HSQC experiments of SNase110 in 2.0 M TMAO, respectively. The two-dimensional (2D) ¹H-¹⁵N HSQC experiments were carried out with G-88W110, V-66W110, and SNase110 in 2.0 M TMAO at protein concentrations of 0.05, 0.1, 0.2, 0.5, 1.0, and 2.0 mM, and with G-20I/G-29I-mutant fragments as well. The pulsed field gradient NMR experiment for translational diffusion measurement (17) was performed with G-88W110, V-66W110, and SNase110 in aqueous solution at protein concentrations of 0.05, 0.1, 1.0, and 2.0 mM and with SNase110 in the presence of TMAO (D-9, 98%; CIL, Lawrence, MA) at protein concentrations of 0.05, 0.1, 1.0, and 2.0 mM. All data were processed and analyzed using FELIX98 (MSI/Accelrys, San Diego, CA). The data points in each indirect dimension were usually doubled by linear prediction (18) before zero filling to the appropriate size. We used 90°–60°-shifted-square sine bell apodizations for all three dimensions before Fourier transformation. ¹H chemical shifts were referenced to internal 2,2-dimethyl-2-silapentane-5-sulfonate (DSS). ¹⁵N and ¹³C chemical shifts were referenced indirectly (19).

For the backbone dynamics studies, the 2D ¹⁵N R₁ and R₂ HSQC and ¹H-¹⁵N NOE experiments were performed with G-88W110, V-66W110, and SNase110 in 2.0 M TMAO at 298 K using standard methods (20). The concentration of the fragments was 1.0 mM in the relaxation experiments. The 2D ¹⁵N R₂ HSQC experiment was run also with the three 110-residues SNase fragments at concentration of 0.2 mM. For ¹⁵N R₁ measurements, the delay times were set to 12, 62, 142, 282, 382, 522, 822, 1202, and 1602 ms for the three fragment samples. For R₂ measurements, the relaxation delays of 8.5, 17.0, 25.4, 33.9, 50.9, 67.8, 84.8, 101.8, 118.7, 135.7, and 152.6 ms were used for the fragment samples. In the 2D ¹H-¹⁵N NOE experiments, a delay of 2 s was followed by ¹H saturation for 3 s, whereas the saturation period was replaced by a delay of equivalent duration in the control experiment. Two experiments were run in an interleaved manner.

Structure calculation
Based on the sequential assignments, the NOE crosspeaks between main-chain protons and between main-chain and side-chain protons were identified for G-88W110, V-66W110, and SNase110 in 2.0 M TMAO. For calculation of three-dimensional (3D) conformations of the fragments, only the medium- and long-range NOEs were used. Dihedral angle constraints were obtained using the program TALOS (21). Restraints of ±25° – ±60° were introduced for - and -angles. Stereospecific assignments were obtained for G-88W110 by analysis of the HNHB spectrum, J-coupling constants, and the NOEs built between the ¹H_N, ¹H, and ¹H_ß resonances. The structure calculations were carried out on a Silicon Graphics (Mountain View, CA) station Onyx 2 employing the program CNS 1.1 (22). A family of 50 structures was generated, from which a final set of the 12 best structures was considered for use in the analysis of structural statistics.

Determination of ¹⁵N relaxation parameters
Intensities for the amide ¹H-¹⁵N crosspeaks in the 2D ¹⁵N R₁, R₂, and ¹H-¹⁵N NOE spectra were determined by measuring peak heights using FELIX software. R₁ and R₂ values for each residue were determined by fitting the measured intensities to a two-parameter single-exponential decay function. The Levenberg-Marquartt algorithm in the program GNUPLOT (ftp://ftp.dartmouth.edu/pub/gnuplot/) was used to estimate the relaxation rates and the uncertainties of the relaxation parameters. The steady-state ¹H-¹⁵N NOEs were calculated from the ratio of peak heights in the NOE spectrum to those in the control spectrum. The root mean-square (RMS) value of the noise of background regions in the spectrum was used to estimate the standard deviation of NOE values.

The relaxation data (R₁, R₂, and ¹H-¹⁵N NOE) were analyzed according to the model-free approach of Lipari and Szabo (23,24), using the program FAST-Modelfree 1.2 (25). Before model-free interpretation of the ¹⁵N relaxation data, anisotropy in the rotational diffusion of the fragment was estimated. The rotational correlation time, _m, was initially estimated from R₂/R₁ ratios using the relaxation rates of a subset of residues chosen to be least affected by large amplitude fast and slow timescale motions (26). The diffusion tensor components (D_xx, D_yy, D_zz) were obtained by the program TENSOR 2.0 (27) to determine whether an isotropic or anisotropic diffusion model should be used in the analysis of the relaxation data. Residues with NOEs lower than 0.65 were first excluded in the above analysis. Besides, residues subject to conformational exchange were also excluded in analysis of _m and rotational diffusion parameters if the condition

(1)

applies, where T_2,n and T_1,n are the T₂ and T₁ values of residue n, and T₂ and T₁ are the average T₂ and T₁ values taken over residues that have not been excluded because of low NOE. SD is the standard deviation of the function calculated for these residues. The starting values of the parameters _m and D_||/D in the model-free analysis of the relaxation data were 8.35 ns and 1.15 for G-88W110, 9.70 ns and 1.25 for V-66W110, and 14.28 ns and 0.88 for SNase110 in 2.0 M TMAO, respectively. These calculated diffusion parameters and the R₁, R₂, and ¹H-¹⁵N NOE relaxation data were then used as the input for the model-free analysis. The axial symmetric model-free spectral density function was chosen to derive the dynamical parameters for the three fragments. The N-H bond length was assumed to be 1.02 Å and the N-H chemical shift anisotropy was taken as –160 ppm. A Lipari-Szabo analysis (23,24) was performed for all individual residues. Five model-free parameter sets were iteratively tested and selected as described by Mandel et al. (28). A total of 500 Monte Carlo simulations were performed to estimate the errors in the model-free parameters.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION CONCLUSIONS SUPPLEMENTARY MATERIAL ACKNOWLEDGEMENTS REFERENCES

 
Secondary and tertiary structures detected by CD spectra
The far-UV CD spectra in Fig. 1, A and B, were used to report the secondary structures of the 110-residue SNase fragment and its mutant variants. For comparison, the far-UV CD spectrum of full-length SNase was also presented in Fig. 1. Table 1 lists the estimated contents of secondary structural components in the 110-residue SNase fragments. The far-UV CD spectrum of G-88W110 resembles those of native SNase in Fig. 1 A, indicating an ordered conformation of G-88W110 having more ideal helical structures compared to the other fragments. Therefore, the apparent helix content of G-88W110 is higher than the other fragments (Table 1). SNase110 in aqueous solution provided a pronounced negative signal at 204 nm and a broad shoulder at 222 nm in the far-UV CD spectrum, which indicated a conformation exhibiting a transient equilibrium involving species containing partially folded and unfolded states (Fig. 1 A). However, 2.0 M TMAO facilitated SNase110 to produce a unique far-UV CD spectrum, which seems to be an additive result of the relatively high ß-strand and low -helix contents of the fragment. V-66W110 showed a negative signal at 207 nm and a broad shoulder at 222 nm in the far-UV CD spectrum (Fig. 1 A). This reflected a folded conformation having less ordered secondary structures in V-66W110 as compared with the far-UV CD spectra of G-88W110 and SNase110 in aqueous solution. The diminished intensity in the ellipticity at 222 nm in the far-UV spectrum of V-66W110 indicated much less -helix content of V-66W110 than those of G-88W110, since the far-UV CD signal is sensitive to the secondary structure and, particularly, to the -helix content of a polypeptide.

View larger version (17K): [in this window] [in a new window]   FIGURE 1  CD spectra measured for G-88W110, V-66W110, SNase110, and SNase110 in 2.0 M TMAO (A) and their G-20I/G-29I-double mutant variants (B) in far-UV region, and for G-88W110, V-66W110, [G-20I/G-29I]G-88W110, and [G-20I/G-29I]V-66W110 in near-UV region (C). For comparison, the far-UV and near-UV CD spectra of native SNase are provided.

Urea unfolding transitions of the fragments
The unfolding transitions of G-88W110, V-66W110, and SNase110 in 2.0 M TMAO and their G-20I/G-29I-mutant variants were detected using the far-UV CD and intrinsic tryptophan fluorescence measurements. The CD spectra of G-88W110, V-66W110, and SNase110 in 2.0 M TMAO upon unfolding measured after 40-fold dilution of the unfolded fragments in 6.0 M urea showed full restoration, indicating full reversibility of the unfolding transition (Fig. 1 in Supplementary Materials). The apparent free energy for unfolding in the presence of urea (G°(H₂O)) and the slope of the transition (m) were then calculated by fitting unfolding curves using a two-state mechanism model (13,14). Table 2 lists the estimated values of G°(H₂O) and m for these fragments. The unfolding profiles (Fig. 2) were different for the three fragments having different degrees of tertiary folding in native state as suggested by CD measurements. The nonsigmoidal unfolding transitions of [G-20I/G-29I]G-88W110, [G-20I/G-29I]V-66W110, and [G-20I/G-29I]SNase110 in 2.0 M TMAO (Fig. 2 in Supplementary Materials) indicated disruption of the tertiary folding of the three fragments by G-20I/G-29I double mutation.

View larger version (9K): [in this window] [in a new window]   FIGURE 2  Urea-induced unfolding transitions monitored by far-UV CD at 222 nm () and intrinsic fluorescence at 325 nm () for full-length SNase (A), G-88W110 (B), and V-66W110 (C). (C) () The unfolding transition of SNase110 in 2.0 M TMAO monitored by far-UV CD. The fitting of experimental data (continuous line) was accomplished according to the two-state mechanism model.

The equilibrium behaviors of the three 110-residue SNase fragments in the urea-induced unfolding process are quite different. As is indicated in Table 2, G-88W110 has relatively high unfolding free energy among the fragments. The m-values for G-88W110 estimated from CD and fluorescence measurements are nearly the same. At 2.0 M TMAO, the unfolding free energy of SNase110 is similar to those of G-88W110, but the m-value is lower than those of G-88W110. V-66W110 has the lowest unfolding free energy of the three 110-residue SNase fragments. However, the estimated m-values for V-66W110 from CD and fluorescence are close to SNase110 in 2.0 M TMAO and G-88W110, respectively.

Secondary chemical shifts
Fig. 3 shows the 2D ¹H-¹⁵N HSQC spectra of the three 110-residue SNase fragments. The spectra of G-88W110, V-66W110, and SNase110 in 2.0 M TMAO showed well-dispersed crosspeaks (Fig. 3, A– C). However, [G-20I/G-29I]V-66W110 and [G-20I/G-29I]SNase110 in 2.0 M TMAO provided the spectra with unstructured and aggregated features, respectively (Fig. 3 in Supplementary Materials). The spectrum of [G-20I/G-29I]G-88W110 (Fig. 3 D) is similar to those of SNase110 in aqueous solution (10). Sequential backbone assignments were determined for G-88W110, V-66W110, and SNase110 in 2.0 M TMAO using a series of 3D ¹H-¹³C-¹⁵N triple-resonance NMR experiments. Respectively, 100%, 86%, and 93% backbone resonances of G-88W110 (10), V-66W110, and SNase110 in 2.0 M TMAO were assigned. Most of the backbone resonances for helix 2 and a few for helix 1 of V-66W110 could not be assigned. This is consistent with the CD observations that V-66W110 has much less -helix content compared to the other two fragments. The assigned ¹⁵N and ¹H_N resonances of V-66W110 and SNase110 in 2.0 M TMAO are indicated in Fig. 3.

View larger version (21K): [in this window] [in a new window]   FIGURE 3  2D 1H-15N HSQC spectra of fragments: G-88W110 (A), V-66W110 (B), SNase110 in 2.0 M TMAO (C), and [G-20I/G-29I]G-88W110 (D). NMR resonance assignments were given by the one-letter amino acid code and the sequence positions.

ß31

ß31

View larger version (40K): [in this window] [in a new window]   FIGURE 4  (A) Histograms of the secondary chemical shifts for 1H resonances (H) of G-88W11010, V-66W110, and SNase110 in 2.0 M TMAO and the differences in H of the three fragments from those of native SNase(V8) (H). The secondary structures of the fragments are labeled at the top of the figure. (B) Ribbon representation of the native SNase structure. The regular secondary structural elements (three -helices: 1, 57–69; 2, 98–106; 3, 122–126; six ß-strands: ß1, 13–17; ß2, 22–26; ß3, 31–35; ß4, 8–11; ß5, 72–76; ß6, 90–93; and three ß-turns: 1, 18–21; 2, 27–30; 3, 94–97) are indicated. Arrows indicate the truncation and mutation sites.

1ß5

2ß6

View larger version (9K): [in this window] [in a new window]   FIGURE 5  NOE contact maps of G-88W110, V-66W110, and SNase110 in 2.0 M TMAO. Only the medium- and long-range NOEs are marked in the map. The secondary structures of the fragments are labeled at the top and side of the figure. NOEs within the ß-barrel and -helix are given above the diagonal, and those between them are given below.

View larger version (28K): [in this window] [in a new window]   FIGURE 6  Superposition of the 12 conformers of G-88W110, V-66W110, and SNase110 in 2.0 M TMAO, and native SNase(V8) obtained using MOLMOL. The N-terminal seven residues of all structures and the sequence region V-111–Q149 of SNase(V8) are not shown.

ß31

Apparent molecular size
To assess whether G-88W110, V-66W110, and SNase110 aggregate under the experimental conditions, the apparent molecular sizes of the three fragments were measured at various protein concentrations. The translational diffusion coefficient is commonly used to determine the apparent radius of a molecule under study. The relationship between diffusion coefficient D and solvated particle radius R is expressed by the Stokes-Einstein equation:

(2)

where K is the Boltzmann constant, T is the absolute temperature, and is the fluid viscosity. The diffusion coefficients of the 110-residue SNase fragments were obtained by translational diffusion measurements. For determining the apparent molecular radius of G-88W110 and V-66W110 in aqueous solution and SNase110 in 2.0 M TMAO, the diffusion coefficient ratio of the reference molecule dioxane to each fragment in aqueous solution or in the presence of TMAO was calculated. To convert the diffusion coefficient ratios of the three fragments to the apparent molecular radius, R, for each fragment, the diffusion coefficient ratio of dioxane to full-length SNase (149 a. a.), which has a known radius of 16.2 Å (33), was obtained under the same experimental conditions.

Variation of the calculated diffusion coefficient ratio and apparent molecular radius for the three SNase fragments at different protein concentrations are shown in Fig. 7. Approximately the same molecular radii were obtained for each of G-88W110 and V-66W110 in aqueous solution and SNase110 in the presence of TMAO at different protein concentrations. However, the concentration dependence of molecular size of SNase110 in aqueous solution was observed (Fig. 7), indicating the aggregation of SNase110 in aqueous solution at high concentration. This implies that G-88W110 and V-66W110 in aqueous solution and SNase110 in 2.0 M TMAO show no complex aggregate under study. The monomeric states of G-88W110, V-66W110, and SNase110 in 2.0 M TMAO were confirmed further by the 2D ¹H-¹⁵N experiments performed with the fragments at concentrations of 0.05–2.0 mM. The chemical shifts and line widths of crosspeaks in the 2D ¹H-¹⁵N HSQC spectra of 2.0-mM SNase fragments are very close to those of 0.05-mM fragments. (Fig. 6 in Supplementary Materials). Therefore, no aggregation occurs to each of the three fragments under the experimental conditions.

View larger version (11K): [in this window] [in a new window]   FIGURE 7  Translational diffusion coefficients and apparent molecular radii obtained for G-88W110 (), V-66W110 (), and SNase110 in aqueous solution () and for SNase110 in 2.0 M TMAO () at different protein concentrations.

Backbone dynamics
The T₁ and T₂ relaxation properties of G-88W110, V-66W110, and SNase110 in 2.0 M TMAO at 298 K are quite different (Fig. 8). However, such a large difference in R₁ and R₂ rates between the three 110-residue SNase fragments is not caused by aggregation of the fragments, but reflects the intrinsic relaxation properties of each fragment. The independence of molecular sizes of the three fragments on protein concentrations and the comparable R₂ values for each fragment at a concentration of 1.0 mM with those of 0.2 mM (Fig. 7 in Supplementary Materials) insured that the concern for aggregation effect can be excluded in the interpretation of relaxation data.

View larger version (32K): [in this window] [in a new window]   FIGURE 8  Sequence variation of 15N relaxation rates (R1 and R2) and 15N-NOEs and the values of the general order parameter, S2, effective correlation time, e, and the exchange rates, Rex, for G-88W110 (black square), V-66W110 (red circle), and SNase110 in 2.0 M TMAO (green triangle) at 298 K. The secondary structures are labeled at the top of the figure.

The general order parameter, S², and effective correlation time, _e, which describe the fast internal reorientation motions of the N-H bond vector were obtained for the fragments. The plot of the generalized order parameter, S², and internal correlation time, _e, against residue number for the three fragments are given in Fig. 8. A large number of residues in the disordered loop L_ß31 and in both N- and C-terminal regions (A1–K9 and L108–K-110 ) of all the three fragments have S² values well below 0.7 and _e values higher than 0.4 ns, indicating higher internal mobility and flexibility of these regions. The average S² value was calculated for all three fragments by averaging over most residues in secondary structural elements: -helix, ß-strand, and pdTp-binding loop, excluding the residues in the regions with large amplitude fast internal motions, namely the disordered L_ß31 loop and both terminal regions in the fragments. For the vast majority of the residues in the secondary structural regions of G-88W110, the S² values of backbone amide groups spread in a range from 0.87 ± 0.03 to 0.99 ± 0.03 with an average S² value of 0.93 ± 0.04. These data suggest a relatively rigid structure for most of the G-88W110 with the exception of the loop L_ß31. The larger spread order parameters with a lower average S² value (0.85 ± 0.13) were obtained for V-66W110 as compared with G-88W110. Residues in the ß-barrel region have the S² values spread from 0.78 ± 0.07 to 0.99 ± 0.02, and residues in the p-loop at sequence positions: 79, 80, 81, 83, 85, 86, 87, and 88 of V-66W110 display S² values < 0.7 and _e values > 1.2 ns, which indicates that V-66W110 has a highly flexible p-loop with very low motional restriction. The backbone amide groups in disordered loop L_ß31 of V-66W110 show largely reduced order parameters compared to G-88W110. Apparently, the observed backbone dynamic features for V-66W110 must correlate with its less compact structure of higher flexibility in comparison with the structure of G-88W110. The general order parameter values of SNase110 in 2.0 M TMAO spread in a range approximate to the S² for G-88W110, except the residues in p-loop which give general S² values in a range from 0.77 ± 0.03 to 0.99 ± 0.02 and _e values > 0.7 ns (Fig. 8).

The R_ex term was required in model-free analysis for some residues of the three fragments. The optimized R_ex values for subsets of residues in G-88W110, V-66W110, and SNase110 in 2.0 M TMAO are summarized in Fig. 8. For G-88W110 only a few amide groups display large R_ex contributions. Residues L-38 and V-39 with R_ex values around 30.58 ± 2.59 and 3.34 ± 0.73 Hz, respectively, occur at the beginning of the disordered loop L_ß31 of G-88W110, presumably due to the transition from rigid secondary ß-sheet structure to a completely flexible L_ß31 loop. In V-66W110, a contiguous group of residues: T-13, K-16, A-17, D-19, G-20, D-21, Y-27, and T-33 in the segment of ß_I-sheet experiences slow conformational exchange motions with R_ex > 3.0 Hz, indicating higher mobility on the millisecond-microsecond timescale of the ß_I-sheet region. Especially, the significant conformational exchange motions (R_ex > 9.0 Hz) are found for amide groups of residues K-16, A-17, D-19, G-20, and D-21 around ß-turn 1 in the ß_I-sheet of V-66W110. For SNase110 in 2.0 M TMAO, the slow exchange motions with R_ex > 3.0 Hz occur for a number of residues throughout the sequence region I-15–Q-106, among them residues I-15, D-19, G-20, D-21, V-39, and E-43 display large R_ex contributions (R_ex > 5.0 Hz). Residues V-39 and E-43 at the beginning of the disordered loop L_ß31 in TMAO-stabilized SNase110 show significant R_ex contributions (R_ex > 6.0 Hz), which is a feature common to G-88W110 in this region. Being different from G-88W110, residues D-19, G-20, D-21, and V-23 in the segment around ß-turn 1 of SNase110 in 2.0 M TMAO undergo the slow conformational exchange motions, whereas the R_ex contributions and low S² values for residues K-78, Q-80, and R-81 in the p-loop suggest some fast internal motions on a subnanosecond-picosecond timescale existing in this region of the fragment. The above described R_ex contributions indicate that the higher degree of slow conformational exchange motions occur to residues in the relevant secondary structural regions of V-66W110 and TMAO-stabilized SNase110, but not for G-88W110.

Therefore, the three 110-residue SNase fragments showing different internal motions exhibit different structure rigidity and compactness. G-88W110 adopts a relatively rigid structure having a higher motional restriction for the relevant N-H vectors in secondary structural regions. The less restricted backbone mobility on a nanosecond-picosecond timescale is observed for V-66W110, which exhibits a less compact structure of higher flexibility. The internal motions of SNase110 in 2.0 M TMAO are more complex, including the relatively high restricted fast internal motions and the contributions from slow conformational exchange which are displayed in G-88W110 and V-66W110, respectively. The fast internal motions dominate in the p-loop of V-66W110 and TMAO-stabilized SNase110. Loop L_ß31 and two terminal regions of all the three fragments are highly flexible.

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION CONCLUSIONS SUPPLEMENTARY MATERIAL ACKNOWLEDGEMENTS REFERENCES

 
Folding status of the three forms of 110-residue SNase fragment
As has been reported, the native SNase has a main ß-barrel hydrophobic core (30) which contains 40% of the total hydrophobic residues of SNase. Truncating SNase at K-110 disrupts the tertiary hydrophobic interactions in SNase and therefore also disturbs the parking interactions in the ß-barrel hydrophobic core of the molecule (30,11). As a result of disrupting the native structure, the retained driving force for folding SNase110 cannot balance with the opposing force from an entropic contribution which is principally responded by conformational freedom (34). This makes SNase110 adopt a conformational ensemble of coexistent partially folded and unfolded states in aqueous solution (10). However, G-88W and V-66W single-point mutations of 110-residue SNase fragment and an effective osmolyte TMAO can raise the unfolding free energies of the fragments (10,35) and facilitate the fragments folding into native-like conformations. The tertiary conformations of G-88W110, V-66W110, and SNase110 in 2.0 M TMAO demonstrate similar native-like ß-subdomain structures but exhibit different levels of folding stability, cooperativity, and compactness.

Different folding stabilities of the three 110-residue SNase fragments
Stabilities of G-88W- and V-66W-mutant SNase fragments having different chain lengths were examined previously by GuHCl-induced denaturation experiments. The obtained unfolding free energies revealed a more stable folding of G-88W110 than that of V-66W110 (11). In this study, the folding stabilities of SNase110 and its mutant variants are depicted by the urea-induced denaturation and backbone ¹⁵N relaxation data, which describe thermodynamic and structure stabilities, respectively, in the folding of the fragments. G-88W110 has the highest unfolding free energy of the three fragments (Table 2), indicating its high stability to unfolding. The restricted backbone internal motions of the ß-barrel and -helix regions in G-88W110 showing the higher average S² value reflect a tightly packed ß-barrel hydrophobic core and well-folded -helices in G-88W110. The ß-barrel hydrophobic core of G-88W110 is the most stable of the three fragments, since no explicit conformational exchange motions (R_ex > 3.0 Hz) can be observed in this structural region. V-66W110 experiences the less restricted fast internal motions and significant slow exchange motions revealed by the low average order parameter S² and the large R_ex value, respectively. The significant slow conformational exchange motions occur in the ß-barrel region of V-66W110, especially at ß-turn 1 of the ß-barrel, which shows a much higher degree of conformational exchange motions. This can be attributed to interconversion between different conformations in the ß-turn 1 region of the fragment, suggesting a dominant overall conformation with local conformational heterogeneity in V-66W110. The low unfolding free energy of V-66W110 provided by both CD and fluorescence measurements is associated largely with the unstable folding of V-66W110. The unfolding free energy and average order parameter, S², estimated from the unfolding transition and backbone relaxation, respectively, for SNase110 in 2.0 M TMAO are close to that for G-88W110 (Table 2 and Fig. 8). However, the R_ex contributions displayed in the ß-barrel region and at the site around the ß-turn 1 of TMAO-stabilized SNase110 are supposed to be associated with the lower level of packing in the ß-barrel. Apparently, the solvophobic effect of TMAO makes the backbone of SNase110 have restricted fast internal motions, but the ß-barrel hydrophobic core remains unstable. The estimated -helix content (Table 1) of both V-66W110 and SNase110 in 2.0 M TMAO is lower than those of G-88W110, implying the incomplete or less ordered folding of -helices in V-66W110 and TMAO-stabilized SNase110 compared to that in G-88W110, as displayed by 3D structure determination (Fig. 6).

The above analysis indicates the different folding stabilities of the three 110-residue SNase fragments, which accounts for the different backbone dynamic and urea-unfolding transitions of these fragments. Folding of G-88W110 exhibits a stable structure having a stable ß-barrel hydrophobic core. V-66W110 generates an unstable structure having an unstable ß-barrel in the tertiary conformation. TMAO-stabilized SNase110, exhibiting relatively high restricted fast internal motions, has an unstable folding of the ß-barrel structural region.

Different folding cooperativities of the three 110-residue SNase fragments
In the fully cooperative system undergoing a two-state unfolding, where the ensemble of proteins contains effectively only proteins that are completely folded or highly unfolded, losses of secondary and tertiary structures during unfolding are synchronous (36). The coincidence of fluorescence and CD data from equilibrium denaturation experiments has been taken as an indicator for the two-state folding of a protein. To analyze the cooperative transitions of the three 110-residue SNase fragments, the equilibrium denaturation transition of full-length SNase was monitored using both fluorescence and CD measurements. Native SNase has a relatively high level of folding cooperativity as shown by superimposability of equilibrium denaturation curves from CD and fluorescence experiments (Fig. 2 A) and by the coefficient of cooperativity, m (Table 2). The nearly same results of equilibrium denaturation experiments by fluorescence and CD for G-88W110 (Fig. 2 B) suggest that the folding of secondary and tertiary structures of G-88W110 is concerted. This implies that G-88W110 can be considered a fully folded 110-residue SNase fragment having native-like ordered secondary structures. In the tertiary structure of native SNase (30), the ß-barrel structural region is constructed by two antiparallel ß-sheets, namely ß_I- and ß_II-pleated sheets (Fig. 4 B). The 110-residue SNase fragment has a nearly complete sequence of the native ß-subdomain of SNase, containing an integrated ß-barrel hydrophobic core and two helices 1 and 2. Residue 88 is located at the C-terminal end of the p-loop or the N-terminal of strand ß6, which links up strand ß5 through the p-loop. The strand ß5 is antiparallel to strands ß6 and ß4 in the ß_II-pleated sheet. Obviously, the replacement of glycine by tryptophan at sequence position 88 improves the tertiary interactions in the structural region around residue W-88 and intensifies the hydrophobic packing interactions of residues from strand ß6 with the other residues in the ß-barrel region. As a result, G-88W110 has a well-packed, native-like ß-barrel hydrophobic core and -helices. The established network of interactions triggered by G-88W mutation should be responsible for the cooperative nature of the equilibrium unfolding transition of G-88W110.

In native SNase, V-66 is one of the residues forming helix 1. A peptide K-70–K-71 links helix 1 with strand ß5 of the ß_II-pleated sheet. The substitution of tryptophan for valine at sequence position 66 may intensify the hydrophobic interactions of W-66 from helix 1 with hydrophobic residues in the ß-barrel region of the protein, generating a native-like ß-subdomain structure (Fig. 4 B). However, the enhancement of the hydrophobic packing interactions in the ß-barrel region by V-66W mutation is not as effective as those by G-88W mutation. The CD and NMR experiments indicate that V-66W mutation can only generate a less ordered and unstable tertiary structure, which exhibits a decreased level of packing in the ß-barrel hydrophobic core and incomplete folding of -helices. Conformation of TMAO- stabilized SNase110 is composed of secondary structural elements that pack together in a manner similar to that seen in G-88W110 (Fig. 4 B). Nevertheless, the long-range hydrophobic interactions were disrupted in SNase110 and the effects of TMAO have little if anything to do with the hydrophobic interactions in the protein (35): the remaining hydrophobic interactions in the ß-barrel region of TMAO-stabilized SNase110 cannot generate tight packing of side chains in the ß-barrel. Thus, the loose packing of the side chains occurs in the ß-barrel hydrophobic core, and the secondary -helix structure is incompletely folded in TMAO-stabilized SNase110 similar to that in V-66W110. The loosely packed ß-barrel and incompletely folded -helical structures of both V-66W110 and TMAO-stabilized SNase110 suggest that both fragments cannot be regarded as fully folded fragments. This may account for the broadening of the unfolding transition obtained by CD measurements for V-66W110 and TMAO-stabilized SNase110 as compared with G-88W110. On the other hand, the CD experiments give approximately the same cooperativity coefficient, m, for V-66W110 and TMAO-stabilized SNase110, which is lower than those for G-88W110. Therefore, V-66W110 and TMAO-stabilized SNase110 have a lower level of folding cooperativity than that of G-88W110. The above analysis suggests that the hydrophobic packing interactions in the ß-barrel hydrophobic core play an important role in determining the folding cooperativity of the three fragments. Differences in cooperativity seen for the three 110-residue SNase fragments may reflect the different extent of formation of the native-like ß-barrel hydrophobic core and -helix of the fragment.

Different folding compactness of the three 110-residue SNase fragments
Compactness of a protein molecule is highly correlated with the extent of formation of secondary structures in protein and is one of the important properties characterizing the degree of protein folding. The denaturant m-value is roughly proportional to the change in solvent exposure during the unfolding transition and thus correlates also with the compactness of a protein (37). Translational diffusion measurement provides directly the apparent molecular size and thus an index of the folding compactness of the three fragments. The 110-residue SNase fragment has a shortage of 39 residues compared to the full-length SNase. Thus, the fully folded G-88W110, giving a molecular size smaller than the size of native SNase (Fig. 7), has a relatively high level of compactness in the structure. However, the low level of compactness is obtained for V-66W110 and TMAO-stabilized SNase110, since the apparent molecular radius of V-66W110 is close to native SNase and the size of TMAO-stabilized SNase110 is larger than those of native SNase (Fig. 7). The m-values in Table 2 indicate that the changes in solvent exposure during the unfolding transitions of V-66W110 and TMAO-stabilized SNase110 are smaller than that of G-88W110, implying also less compact structures of both V-66W110 and TMAO-stabilized SNase110. The compactness of G-88W110 can be seen in the packing of residues in the ß-barrel hydrophobic core. Compared to the tight packing of side chains of the relevant hydrophobic residues in the ß-barrel region of SNase(V8) (Fig. 9), the similar tight packing of the corresponding residues is observed for G-88W110. However, TMAO-stabilized SNase110 shows the large deviation in the relevant packing (Fig. 9). Therefore, G-88W110 is the most compact of the three 110-residue SNase fragments.

View larger version (18K): [in this window] [in a new window]   FIGURE 9  Stereo view, showing the packing of the side chains of those hydrophobic residues, which are located mostly in the upper part of the ß-barrel region in the tertiary structures of G-88W110, V-66W110, SNase110, and SNase(V8).

Analysis of the conformational states of the three forms of 110-residue SNase fragment reveals a relationship between stability, compactness, and cooperativity in the folding of fragments. The compactness of each conformational state is highly correlated with the extent of formation of a stable, tightly packed ß-barrel region. The fragment with higher stability tends to show the higher level of cooperativity in the unfolding transition and higher compactness in structure.

Folding mechanism of the 110-residue SNase fragment
In the partially folded state of SNase110 in aqueous solution (10), the native-like ß-turn conformations of the segments around I-18–D-21, Y-27–Q-30, and A-94–K-97, as well as the bend at peptide K-70–K-71 are supposed to be formed transiently based on the similarity of chemical shifts for the dispersed crosspeaks, A-17, G-20, M-26, Y-27, G-29, A-69, K-70, I-72, D-95, and A-96 in the 2D ¹H-¹⁵N HSQC spectrum of SNase110 to those observed for SNase(V8) (15,16). The preferences for formation of native-like ß-turn conformations in these segments of SNase110 are attributed to localized structural propensities that match native structural elements of the same segments in the ß-barrel region of native SNase. The localized propensity for turn-like conformations of the sequence regions, I-18–D-21 and Y-27–Q-30, was observed for the N-terminal short fragments of SNase, namely, SNase28, SNase36, and SNase79 containing 1–28, 1–36, and 1–79 residues of SNase, respectively (38,39). In the three-strand antiparallel ß<