Far-UV CD equilibrium measurement
The far-UV CD spectra obtained in the titration of acid-denatured cyt c (16 μM) with SDS at pH ∼2.0 were recorded as shown in Figure 2. Three phases can be found in this titration process, starting with the initial acid-denatured state, which can be viewed as dominantly a random coil showing a negative peak at ∼195 nm and a broad negative band at 200–220 nm (Woody 1992; Sreerama and Woody 2000; Shi et al. 2002; Xu and Keiderling 2004).
For [SDS] < 0.25 mM (Fig. 2A), formation of secondary structure from the acid-denatured state was observed as an ellipticity (negative) gain at both 222 nm and 207 nm, which is typical of α-helix (Woody 1996; Sreerama and Woody 2000). This helix formation is almost linearly dependent on the concentration of SDS up to 0.25 mM, as can be seen from the ellipticity change at 222 nm (inset), with a stable MG-like state being formed at [SDS]/[cyt] ∼14. The inset curve marked in diamonds represents the titration of acid-denatured cyt c (16 μM) with a solution having the same protein concentration at the same pH but including 4 mM SDS. A sigmoidal curve can be fit to this ellipticity change, but the breadth of the transition calls into question the possibility of a two-state transition.
For [SDS] values from 3 mM to 60 mM (Fig. 2B), the low pH cyt c far-UV CD spectra also show native-like secondary structure content, suggesting another MG-like state. These spectra seem independent of SDS concentration once it is above 3 mM. A surface tension test (data not shown) shows that the critical micelle concentration (CMC) for SDS solution at pH = 2.0 is around 3 mM. The fact that the same cyt c spectra were obtained for this range of [SDS] values suggests that micelle formation may affect the stability of this complex, but the number (concentration) of micelles is less important.
The third cyt c form occurs for [SDS] from 0.3 to 2.5 mM, between the above noted ranges, where the cyt–SDS solution becomes opaque and visual precipitation occurs as more SDS is added. The corresponding CD spectra first become distorted with a sharp ellipticity loss at 207 nm, and then appear to have β-structure characteristics before losing most of the intensity upon precipitation. These samples by contrast are not stable. A time-dependent experiment for the cyt–SDS complex solution (16 μM cyt c with 0.36 mM SDS at pH = 2) was carried out at T = 25°C, as shown in Figure 3A. With time, a spectrum suggesting formation of a large fraction of β-sheet structure is obtained. The ellipticity change at 222 nm versus time is plotted as an inset in Figure 3A, and can be fit to a single exponential curve. A separate FTIR-ATR spectrum was recorded on the precipitate collected from cyt c in 1 mM SDS, as shown in Figure 3B, and compared to solution data for the acid-denatured and SDS-induced MG-like forms. The FTIR-ATR spectrum shows three amide I features, a broad band at 1652 cm−1 representing the residual helix and turns, and a more intense one at 1625 cm−1 with a weaker shoulder at 1695 cm−1, corresponding to contributions from the newly formed β-sheet structure that are not seen in the denatured or MG-like spectra.
As a comparison, a titration of 16 μM cyt c at pH ∼2.0 with a positively charged surfactant, dodecyltrimethylammonium chloride (DTA), was recorded at the same pH with far-UV CD, as shown in Figure 4. Over the whole surfactant concentration range from 0 to 50 mM, no precipitation is observed, which differs from what is observed with SDS. The ellipticity change at 222 nm is plotted as an inset in Figure 4. Although both surfactants provide hydrophobic tails to aid solubility of the core, they differ because with DTA there is no charge neutralization. In the DTA case, ellipticity starts to increase at 6 mM (inset in Fig. 4), which is the CMC for DTA at pH ∼2 (data not shown), but a MG-like state is not obtained until [DTA] > 16 mM, much higher than for SDS, which offers charge neutralization as well as hydrophobic interactions.
Soret absorbance measurement
The UV-Vis absorption in the Soret region of the same protein–surfactant complex as in Figure 2 is shown in Figure 5. Three phases can also be found in this titration process. The acid-denatured cyt c has a peak located at 394 nm, which agrees well with previous studies (Brems and Stellwagen 1983). At low [SDS] (<0.2 mM) (Fig. 5A), the absorbance at 394 nm decreases sharply with little peak shift, suggesting no fundamental heme change in this concentration range but the possibility of multiple heme micro-environments broadening the Soret band. When [SDS] is larger than 0.2 mM (data not shown), the baseline increases, due to precipitation, and a red peak shift is observed. For [SDS] in the 3–60 mM range (Fig. 5B), the precipitate redissolves, with only a minor peak intensity decrease; however, especially at high [SDS] (20 mM or above), a band shift is observed, suggesting that a structural change at the heme occurs within this region even though the secondary structure is constant, as shown from far-UV CD.
Fluorescence spectra (Fig. 6) arising from the single Trp in cyt c were also recorded for the same cyt–SDS samples used in UV CD and Soret absorption experiments. In the native state structure, Trp 59 is buried in the hydrophobic core and almost no fluorescence is observed, presumably being quenched by Forster energy transfer to the heme group (Pinheiro et al. 1997). However, when the protein is denatured, the Trp residue becomes solvated and intense fluorescence results. As seen in Figure 6A, even 0.05 mM of SDS induces a big decrease in the fluorescence of acid-denatured cyt c, and about 80% of the fluorescence was quenched for [SDS] = 0.2 mM, suggesting a hydrophobic collapse even at this very low SDS concentration. There is a blue shift of the peak (∼10 nm) accompanying this decreased fluorescence intensity when [SDS] is varied from 0 to 0.2 mM. Between [SDS] = 0.2 mM and 3 mM, the turbidity distorts the observed fluorescence. When [SDS] is above 3 mM (Fig. 6B), an increase in the fluorescence is observed, and a small red shift occurs, especially at high SDS concentration (>20 mM), indicating that SDS starts to unfold the protein. Factor analysis of the fluorescence spectra as a function of bandshape also shows that there is a distinct maximum in the frequency shift of the band shape at around 10 mM SDS (data not shown), which agrees well with UV-Vis spectral results.
Thermal denaturation monitored by far-UV CD
In the above titration experiments, MG-like states were observed at both low [SDS] and above the CMC by use of far-UV CD, UV-Vis, and fluorescence. Thermal denaturation experiments monitored with far-UV CD were carried out to differentiate the MG-like states found above. Figure 7, A–C, shows the thermal denaturation behaviors of three MG-like states that have the same protein concentration but with different [SDS]/[Cyt] ratios, ∼12 (MG1), 240 (MG2), and 1500 (MG3), respectively. At low temperature, all three states show a typical α-helical pattern, similar to that of the salt-induced MG state at low pH (Xu and Keiderling 2004). For the MG1 state (Fig. 7A), the ellipticity at 222 nm is almost constant up to 20°C (inset in Fig. 7A), and then a loss in ellipticity that is linear with temperature is observed up to T = 85°C. Upon cooling back to 25°C, the spectral shape still shows typical α-helical character, and ∼64% of the ellipticity change at 222 nm is recovered. The thermal denaturation behavior of MG2 (Fig. 7B) is different, wherein the ellipticity at 222 nm (inset in Fig. 7B) does not have any low-temperature stable region. Instead, a linear loss of ellipticity with temperature is observed from 5°C up to 55°C, followed by a major transition above 60°C. This is probably due to aggregation, which is consistent with the CD band shape seen at high temperature. Only 24% of the ellipticity is recovered on cooling and the pattern remains β-sheet-like, indicating an irreversible transition. For the MG3 state, which has a high concentration of SDS (25 mM) and a high [SDS]/[cyt] ratio of ∼1500 (Fig. 7C), a linear loss of ellipticity at 222 nm (inset in Fig. 7C) is observed from low temperature (5°C) without any thermally stable region. No band-shape disruption is observed for MG3 even up to 80°C, and when the sample is cooled down to 20°C, 82% of the CD signal change at 222 nm is recovered. Meanwhile maintenance of the [θ]222/[θ]207 ratio before and after heating indicates that this process is mostly reversible. In addition, isodichroic points at 202 nm are observed in the thermal denaturation experiments for both MG1 and MG3, suggesting a two-state process, probably from α-helical to random coil conformation.
For comparison, the thermal behavior of native state cyt c denatured by SDS solutions at neutral pH is shown in Figure 7, D–F, under varying conditions. When only 0.2 mM SDS is involved, CD spectra at low temperature are just like that of native state, suggesting almost no unfolding occurs (Fig. 7D), which agrees with previous studies that unfolding by SDS at neutral pH starts at [SDS]/[cyt] ≈40 (Das et al. 1998). At high temperature, aggregation is apparent, with the Tm being ∼10°C lower than that observed for the native state (inset in Fig. 7D), suggesting that even low concentration SDS (submicellar) destabilizes cyt c. For the other two solutions with [SDS]/[Cyt] at ∼235 and 1400, MG-like CD spectra typical of helical structure were observed at both low and high temperatures, and no aggregation was observed for the whole temperature range. An isodichroic point at ∼204 nm is found for these two samples, which could be consistent with a two-state transition with an increase of temperature, but only a broad transition with no sigmoidal behavior is evident in the inset curves, partially due to a lack of a low temperature stable region, much as seen at low pH. Upon cooling, CD intensity at 222 nm was fully recovered for both samples, and a similar (θ)222/(θ)207 ratio is obtained, indicating that these two high [SDS] thermal unfolding processes are fully reversible.
Thermal denaturation monitored by FTIR
The thermal behaviors of cyt/SDS samples having the same protein concentration (∼1.6 mM) at low pH (∼2.0) but with different amounts of SDS were also studied with FTIR, as shown in Figure 8, A–C, corresponding to [SDS]/[Cyt] = 2, 12, and 270, respectively. When T = 5°C, the FTIR spectrum for cyt c with 3.8 mM SDS shows an amide I′ peak at ∼1649 cm−1 (Fig. 7A), which is consistent with a high content of α-helical structure. By T = 45°C, a shoulder at ∼1620 cm−1 appeared, suggesting some β structure or aggregation, but it did not increase much even at 80°C, although it remained after cooling, indicating a partially irreversible change.
For cyt c in 22 mM SDS (Fig. 8B), a similar result is obtained at low temperature, but the intensity sharply decreases with temperature increase, and a new peak at ∼1620 cm−1 and a shoulder at ∼1687 cm−1 appear at 25°C, indicating formation of β structure due to aggregation. Both peaks are nearly constant in intensity for T > 30°C, with no precipitation, and remain after cooling, suggesting irreversibility.
For cyt c in 0.51 M SDS at pH ∼2 (Fig. 8C), low temperature second derivative analysis shows that the amide I′ band is initially composed of two bands centered at 1652 cm−1 and 1635 cm−1, which could be assigned to interior α-helical and solvated α-helical structure, respectively (Haris and Chapman 1995; Williams et al. 1996). With an increase of temperature, the peak at 1635 cm−1 disappears at ∼40°C, while the intensity at 1641 cm−1 grows, suggesting a helix to coil transition. By contrast to samples in 3.8 mM and 22 mM SDS, cyt c in 0.51 M SDS does not show any aggregation features up to 80°C, and its spectral changes reverse on cooling.
The changes of A1620/A1649 with temperature for the above samples are plotted in Figure 9A. There is no obvious transition when [SDS]/[Cyt] is 270 (triangle), and only a small transition is observed for [SDS]/[Cyt] ∼2 (diamond). By contrast, a large transition is observed for [SDS]/[Cyt] around 12 (filled circle).
Figure 9B indicates singular-value decomposition (SVD) analysis results for the transitions observed in Figures 8, A–C, and 9A, as well as for an additional sample having 80 μM cyt c and 0.98 mM SDS ([SDS]/[Cyt] = 12, open circle). The small transition observed in Figure 9A for the [SDS]/[cyt] ratio = 2 is clearer with SVD (diamonds), yielding a transition with Tm = 37°C. When the ratio = 12, for both samples having 80 μM or 1.6 mM cyt c, big transitions are observed. However, the temperatures for the transitions are different, with Tm ∼ 25°C for high [Cyt] (closed circles), whereas no complete denaturation was observed for low [Cyt], indicating that the protein concentration as well as the ratio of [SDS]/[Cyt] are determining factors for stability of the protein–surfactant complex.
For comparison, the thermal behavior of native state cyt c denatured with SDS at a neutral pH was also studied by FTIR under varying conditions, as shown in Figure 8, D–F. For [SDS]/[cyt] ∼2 (Fig. 8D), at low temperature the amide I′ peak is located at 1649 cm−1, and intensity loss begins above 35°C with the appearance of a shoulder at ∼1618 cm−1 suggesting partial aggregation. When the [SDS]/[Cyt] ratio is 13 (Fig. 8E), the spectra show a sharp peak at 1616 cm−1 at higher temperatures, which remains on cooling, and precipitation can be observed visually, indicating irreversible aggregation.
At high SDS concentration (Fig. 8F), when T = 5°C the amide I′ peak was located at ∼1645 cm−1, which is normally assigned to random coil structure. However, second derivative analysis showed this amide I′ peak to be composed of components at 1650 cm−1 and 1635 cm−1, much as seen at low pH, which could be assigned to interior and solvated α-helical structure (Haris and Chapman 1995; Williams et al. 1996), respectively, and indicated that a large amount of secondary structure still existed. The component at ∼1635 cm−1 disappears and the one at 1650 cm−1 stays almost constant with increasing temperature, suggesting a helix to coil transition that is reversible on cooling. The changes of A1616/A1649 versus temperature for the above samples are plotted in Figure 9C. Two transitions atTm1 ∼ 37°C and Tm2 ∼70°C are observed for [SDS]/[cyt] = 2, and only one with Tm ∼40°C is seen for this ratio at 13, but no transition is observed for this ratio at 270.