3.2 Pressure- and temperature-stabilized intermediates
To further characterize the pressure- and temperature-stabilized intermediates, PRE effects on the 1H-transverse relaxation rate constant, R 2 were measured for the variant. Rather than directly measuring the relaxation rates from a series of two-dimensional NMR experiments, we obtained the ratios of peak heights for a paramagnetic (oxidized) sample to those of a diamagnetic (reduced) sample, i.e.,I para/I dia (Eq. 2).22 Figure 4A shows the ratio for each amide group in the E128C variant along with the residue number (i.e., 25–110) at 0.1, and 250 MPa at 313 K. TheI para/I dia of residues from 25 to 275 are shown in Figure S6. When the ratio was 1, no PRE effects were observed. In contrast, a lower ratio indicated greater PRE effects. Assuming an intrinsic line width at half-height (Δν(hertz) = R 2int/π) of 20 Hz and correlation time of 12 ns for the folded OspA, distances between the paramagnetic probe and nucleus of interest were calculated by a previously described method (Figure S7, Eq. 3).22 Note that the multiplicity of MTSL positions was not taken into account in the calculation. At 0.1 MPa, ratios of less than 0.8, corresponding to a distance of 15~20 Å from the paramagnetic probe, were observed for residues 82, 83, 101, and 102, while values < 0.3 corresponded to a distance of 12~15 Å for residues 103–110. These results indicate that residues 103–110 are closer to residue 128 than they are to residues 82, 83, 101, and 102, consistent with the structure resolved by X-ray crystallography at atmospheric pressure.6 At 250 MPa,I para/I dia showed both increased and decreased values at several sites. For example, the ratios for residues 82, 83, and 102 were increased by approximately 0.2, indicating that they drifted away from residue 128. In contrast, a remarkable decrease in the ratio for residue 59 indicated that it shifted toward residue 128. Figure 5A shows the locations of residues 59, 62, 82, 83, and 128, while the pressure dependence details for the ratios are shown in Figure 5B.
Figure 4B shows the intensity ratio,I para/I dia, for each amide group of the E128C variant for each residue at 303 and 318 K at 0.1 MPa. As the temperature increased, the ratios of residues 80, 82, and 83 increased by approximately 0.1–0.2, indicating they moved away from residue 128. Alternatively, slight decreases in the ratios for residues 33, 36, 37, 59, 62, and 105 indicated that these residues shifted closer to residue 128. Figure 5C shows the gradual increases and decreases inI para/I dia of residues 59, 62, 82, and 83 as temperature increases.
Different behaviors were observed at residues 36, 37, and 106 between increasing pressure and increasing temperature, as shown in Figures 6 and S8. As pressure increased, split cross-peaks were observed (Figure 6A). The second (new) cross-peak increased at the expense of the first (original) cross-peak with increasing pressure. In contrast, only a single peak was observed as the temperature was increased from 303 to 318 K (Figure 6B). These observations indicate that these amide groups have at least two folded conformations. Assuming a two-state exchange process, split peaks can be observed if the exchange rate, namelyk ex, is much smaller than the chemical shift difference (Δω ) between two states; however, a single averaged peak is observed if k ex is much larger than Δω . Figure 6C shows theI para/I dia values of original and new peaks for residues 36 and 37, and 106 at 200 MPa, where both original and new peaks were observed in the HSQC spectrum. For residues 36 and 37, new peaks showed 0.2–0.4-fold smaller values than the original peaks, indicating shorter distances in the intermediate than in the native conformations (Figure S6). Conversely, the new peak of residue 106 demonstrated a 0.6-fold larger value than the original peak, indicating much longer distance in the intermediate than in the native conformation. In the case of the averaged peak, the observed PRE effect is the weighted population average of the PRE effects for the two conformations, i.e., native and intermediate conformations. Since the average I para/I dia values of residues 36, 37, and 106 at 318 K and 0.1 MPa are similar to those of their new peaks (200 MPa and 313 K) (Figure 6D), the second conformation might be dominantly populated at 318 K.
Pressure and temperature effects onI para/I dia for each amide group were also investigated for the variants D118C and A140C, as shown in Figures S9 and S10, respectively. As pressure increased, theI para/I dia values of residues 59–61 of the D118C variant increased by approximately 0.1–0.2 (Figure S9A), indicating that D118 moved away from the β3–β4 turn in the pressure-stabilized intermediate. As the temperature increased, the values of residues 57–58 and 93–94 of the variant increased by approximately 0.2–0.3 (Figure S9B), indicating that residue 118 moved away from the β3–β4 and β6–β7 turns in the temperature-stabilized intermediate. Similarly, as pressure or temperature increased, theI para/I dia values of residues 94–97 of the A140C variant increased by 0.2–0.3, indicating that residue 140 moved away from the β7 in both of the intermediates (Figure S10A and B).I para/I dia values of residues 36, 37, and 106 for the A140C variant are summarized in Figure S11. A pair of cross-peaks was only observed for residue 36. Unlike E128C, the intensity ratio of the new residue 36 peak is similar to that of the original one (Fig. S11A). The ratios at higher temperature were significantly higher for residues 36 and 37, and slightly lower for residue 106 (Fig. S11B), showing opposite tendencies with the data from E128C. In addition, for the variants D118C and A140C, residues 33–39, including the β1–β2 turn, maintained lowerI para/I dia values (i.e., 0–0.6) in both intermediates.
These results indicate that the pressure- and temperature-stabilized intermediates shared a common structural feature in which residue 128 moved away from the β5–β6 turn (consisting of residues 82 and 83) while becoming more accessible to the β3–β4 turn region, including residues 59 and 62. Moreover, residues 118 and 140 moved away from the N-terminal domain, yet remained accessible to the β1–β2 turn. As residues 118, 128, and 140 are in the turn region between β8–β9, β9–β10, and β10–β11, respectively, these PRE results suggest that the central β-sheet, i.e., β8–β10, is partially disordered, and thus is located a longer average distance from the N-terminal domain in the intermediates than it is in the native conformation (Figure 1). Alternatively, the β-sheet region may retain weak specific contact with the β1–β2 turn, owing to a twist of the central β-sheet and increased flexibility of the polypeptide chain. Such specific long-range contact has also been observed in intrinsically disordered proteins such as α-synuclein.23,36
New peaks appeared in the central aspect of the spectrum, an area in which disordered polypeptide chains are typically observed, as pressure or temperature was increased. The original signals corresponding to the N-terminal domain and the new NMR signals showed gradual decreases in intensity after incubating the protein for a few days at 250 MPa, likely owing to protein aggregation. Therefore, the assignment of new peaks using the conventional triple-resonance NMR experiments has not been completed. However, it is useful to analyze theI para/I dia ratios for new peaks that likely correspond to amide groups in the central and C-terminal domains. Figure 7A shows the frequency ofI para/I dia values for all original cross-peaks that originated from the native protein conformation, as the paramagnetic probe was attached at residue 140 in the central domain. The ratios were widely distributed between 0 and 1; however, were primarily above 0.8, indicating that a folded conformation remained near the paramagnetic probe, and many amide groups were not accessible to the probe. Figures 7B and C show the frequencies ofI para/I dia values of new peaks in the pressure-stabilized and temperature-stabilized intermediates, respectively, of the A140C variant. In the temperature-stabilized intermediate (Figure 7C), the frequencyvs. I para/I dia plot had a Gaussian shape, with a center value of approximately 0.7; frequencies of less than 0.2 and more than 0.9 were not observed, indicating an entirely different shape from the native conformation. Since a Gaussian shape is typically obtained for random events, the PRE results may have resulted from an ensemble of disordered conformations of the polypeptide chain. The PRE provides 1/r 6ensemble-averaged distance information, and thus different conformations may show different PRE effects. Even amide groups more than 25 Å, apart from the paramagnetic probe in the native conformation, could transiently access the probe in the intermediates due to the increased flexibility of the polypeptide chain, while very few amide groups showed consistently strong PRE effects. Although theI para/I dia-frequency profile of the pressure-intermediate was similar to that of the temperature, the Gaussian shape was not as apparent, and the strong and weak I para/I dia ratios, i.e., > 0.9 and < 0.2, were still observed (Figure 7B). Since an intermediate population is retained in both conditions (Figure S2 and Figure S3), the differences in theI para/I dia-frequency profiles indicate that the pressure-intermediate has a more compact conformation. Similar frequency profiles were observed for the D118C and E128C variants, as shown in Figures S12 and S13, respectively.