Skip to main content
NCBI home page
Protein Science : A Publication of the Protein Society logoLink to Protein Science : A Publication of the Protein Society
. 2001 Aug;10(8):1539–1548. doi: 10.1110/ps.49401

High stability of a ferredoxin from the hyperthermophilic archaeon A. ambivalens: Involvement of electrostatic interactions and cofactors

Charmaine Moczygemba 1, Jesse Guidry 1, Kathryn L Jones 2, Cláudio M Gomes 3, Miguel Teixeira 3, Pernilla Wittung-Stafshede 1,2
PMCID: PMC2374097  PMID: 11468351

Abstract

The ferredoxin from the thermophilic archaeon Acidianus ambivalens is a small monomeric seven-iron protein with a thermal midpoint (Tm) of 122°C (pH 7). To gain insight into the basis of its thermostability, we have characterized unfolding reactions induced chemically and thermally at various pHs. Thermal unfolding of this ferredoxin, in the presence of various guanidine hydrochloride (GuHCl) concentrations, yields a linear correlation between unfolding enthalpies (ΔH[Tm]) and Tm from which an upper limit for the heat capacity of unfolding (ΔCP) was determined to be 3.15 ± 0.1 kJ/(mole • K). Only by the use of the stronger denaturant guanidine thiocyanate (GuSCN) is unfolding of A. ambivalens ferredoxin at pH 7 (20°C) observed ([GuSCN]1/2 = 3.1 M; ΔGU[H2O] = 79 ± 8 kJ/mole). The protein is, however, less stable at low pH: At pH 2.5, Tm is 64 ± 1°C, and GuHCl-induced unfolding shows a midpoint at 2.3 M (ΔGU[H2O] = 20 ± 1 kJ/mole). These results support that electrostatic interactions contribute significantly to the stability. Analysis of the three-dimensional molecular model of the protein shows that there are several possible ion pairs on the surface. In addition, ferredoxin incorporates two iron–sulfur clusters and a zinc ion that all coordinate deprotonated side chains. The zinc remains bound in the unfolded state whereas the iron–sulfur clusters transiently form linear three-iron species (in pH range 2.5 to 10), which are associated with the unfolded polypeptide, before their complete degradation.

Keywords: Hyperthermophiles, thermostability, iron-sulfur proteins, protein unfolding

Proteins from thermophilic organisms offer a unique opportunity to study the determinants of thermostability (Vogt and Argos 1997; Jaenicke and Bohm 1998). Although these proteins are often very similar in sequence and structure to their mesophilic homologs, they are much more resistant to thermal denaturation and inactivation. Efforts to determine the origin of this thermostability have led to several hypotheses, such as stabilization by an increased number of ionic interactions, an increased extent of hydrophobic surface burial, an increased number of prolines, and smaller surface loops (Vogt and Argos 1997). Although evidence for these and other modes of stabilization can be found in specific examples, none applies to all or even most thermostable proteins. If there are general rules for how thermophilic proteins attain their stability, it is clear that they do not lie exclusively in individual interactions; they may be based in properties of the whole molecule, such as how the stability is distributed and coupled throughout the structure or how it is divided by enthalpy and entropy (Szilagyi and Zavodszky 2000).

Three models have been proposed to explain the higher denaturation temperatures of thermophilic proteins (Rees and Adams 1995; Beadle et al. 1999); each has a different thermodynamic consequence. In the first model, compared with a mesophilic protein, the thermophilic one could be more thermodynamically stable throughout the temperature range, that is, have higher free energy of unfolding (ΔGU) than the mesophilic protein at every temperature. A second model predicts that the free energy profile of the thermophilic protein will be displaced horizontally to a higher temperature. In this model the maximal value for ΔGU would be equal for the mesophilic and thermophilic protein, but the maxima would occur at different temperatures. At high temperatures, the thermophilic protein would be more stable; at lower temperatures, the mesophilic protein would be more stable. Finally, a third model predicts that the free energy profile for the thermophilic protein would be a flattened version of that for the mesophilic protein. Thus, the thermophilic protein would have a more shallow dependence of ΔGU on temperature, but the maximal ΔGU would again be equal for the mesophilic and thermophilic proteins. Support for the different models, as well as combinations thereof, have been reported (McCrary et al. 1996; Beadle et al. 1999; Hollien and Marqusee 1999).

To gain better insight into mechanisms governing protein thermostability, simple models may be of great use. Ferredoxins are good examples of such proteins: They are small, monomeric polypeptides containing iron–sulfur centers whose integrity can be followed easily by spectroscopic methods. In addition, ferredoxins are widespread in the three domains of life; they are evolutionarily ancient proteins considered to have been the first bioinorganic catalysts (Otaka and Ooi 1987; Wachtershauser 1992). Because archaea are presumed to be the most ancient living organisms on contemporary Earth (Woese et al. 1990), the study of ferredoxins from such thermophilic organisms may elicit ancestral strategies for protein stabilization. Moreover, it may help to decide whether life originated at high or low temperatures (Pace 1991). A family of di-clusters, seven-iron-containing ferredoxins from various archaea belonging to the order Sulfolobales, has been characterized (Teixeira et al. 1995; Gomes et al. 1998). These organisms live optimally at pH 2–4.5 and temperatures of 65°C–80°C. Their ferredoxins are acidic, ∼12 kD proteins that contain one [3Fe-4S]1+/0 and one [4Fe-4S]2+/1+ center. The structure of the ferredoxin from Sulfolobus strain 7 (Fujii et al. 1997) revealed the presence of a zinc center, in addition to the iron clusters, in this protein. By monitoring the alteration of the spectral properties as a function of time at a high temperature, the folded forms of the Sulfolobus ferredoxins were shown to be extremely resistant to degradation (Teixeira et al. 1995; Gomes et al. 1998). We recently reported a preliminary study (Wittung-Stafshede et al. 2000) of the stability and unfolding of the Acidianus ambivalens ferredoxin, for which the complete amino acid sequence and a structural model are available, that revealed that this ferredoxin was highly resistant to both temperature (Tm of 122°C; pH 7) and chemical perturbation (addition of 8 M guanidine hydrochloride [GuHCl] did not unfold the ferredoxin at pH 7, 20°C). At pH 10, GuHCl-induced ferredoxin unfolding occurred with a midpoint at 6.3 M, and a ΔGU in water (pH 10; 20°C) of 70 ± 3 kJ/mole was reported (Wittung-Stafshede et al. 2000).

In the current investigation, we explore further the thermostability of the A. ambivalens ferredoxin (Fd). We show that the stability of Fd is extremely high, having a ΔGU near 80 kJ/mole at pH 7, 20°C. From thermal unfolding experiments a low heat capacity of unfolding (ΔCP) is estimated, indicating a rather flat stability versus temperature profile. Lowering the pH decreases the stability dramatically (both ΔGU at 20 °C and Tm), indicating that electrostatic interactions contribute favorably to the high stability of Fd at neutral pH. Upon polypeptide unfolding, at both high and low pH, the iron–sulfur clusters first rearrange into intermediate (potentially linear three-iron) species before dissociation and decomposition occur. Interestingly, the zinc ion remains coordinated to the unfolded protein.

Results

Temperature dependence of Fd unfolding free energy

The A. ambivalens Fd is a 103-residue β-sheet protein with two iron–sulfur clusters. The native state is characterized by negative CD intensity around 220 nm and a highly quenched tryptophan (Trp 62) emission (because of energy transfer to the iron centers) centered around 345 ± 2 nm (Fig. 1 ▶). The reported experiments are all performed with Fd in its fully oxidized form (the resting state); native oxidized Fd displays distinct visible absorption at 410 nm (ɛ410 = 30,400/M per cm; see Fig. 1B ▶). The melting temperature (Tm) for Fd in buffer (pH 7) was determined from temperature-induced unfolding experiments monitored by far-UV CD in the presence of various amounts of GuHCl, because the presence of GuHCl lowers the melting temperature. The Tms found in the presence of different GuHCl concentrations were used earlier to estimate, by linear extrapolation, a Tm of 122 ± 2°C for Fd at pH 7 in the absence of denaturant (Wittung-Stafshede et al. 2000). We now performed additional thermal unfolding experiments and analyzed all melting transitions according to a modified van't Hoff equation (see Materials and Methods).

Fig. 1.

Fig. 1.

Open in a new tab

Far-UV CD (A), visible absorption (B), and tryptophan emission (C) of folded (solid line) and unfolded (dashed line) Fd.

Fd unfolding is not reversible (Wittung-Stafshede et al. 2000). An irreversible unfolding reaction can be analyzed in terms of thermodynamics only under circumstances not involving a significant enthalpy change for the irreversible transition, such as an aggregation process (Privalov and Potekhin 1986; Pfeil et al. 1997). Unfolded Fd appears clear in solution and migrates as a monomer in gel electrophoresis, indicating that the unfolded polypeptide chains do not form covalent aggregates. Instead, the irreversibility of Fd unfolding is most probably due to decomposition of the iron–sulfur clusters. Several experimental findings (discussed throughout the text, listed in the Discussion), indicate that the thermodynamic parameters deduced here for Fd are close to their equilibrium values. This approximation has also been applied to other proteins that unfold irreversibly, such as phosphoglycerate kinase, lac repressor, and aspartate transcarbamoylase (Edge et al. 1985; Manly et al. 1985; Hu and Sturtevant 1987). Of relevance for the current study, thermodynamic parameters extracted from irreversible experimental data were published for another ferredoxin, the thermostable Thermotoga maritima ferredoxin that has one [4Fe-4S] cluster. It was suggested that, for T. maritima ferredoxin unfolding, the irreversible process was also that of cluster dissociation and decomposition. Because no aggregation was observed, extrapolated Tm values were identical in GuHCl and GuSCN experiments, calorimetric and van't Hoff derived ΔH values were identical, and ΔCP from calorimetric and ΔH versus Tm experiments agreed, the investigators argued that equilibrium thermodynamics had been obtained (Pfeil et al. 1997). It should be noted, however, that there was no measure of the possible enthalpy changes associated with cluster dissociation and decomposition in the T. maritima study; in addition, we have found no such estimates anywhere in the literature. Nevertheless, because the unfolding processes are likely to be similar for the two ferredoxins, and experimental results on A. ambivalens Fd add direct support, a thermodynamic treatment of our data appears valid as an approximation.

Table 1 lists the enthalpy of unfolding (ΔH[Tm]) values that were derived at the different Tms. A linear relationship between ΔH(Tm) and Tm was found (Fig. 2 ▶), indicating that ΔCP is independent of temperature; a linear fit provides a ΔCP value of 3.15 ± 0.1 kJ/(mole • K). No correction of the ΔH values at pH 7 for possible (in the presence of high GuHCl concentrations) contributions of GuHCl–protein interactions were made. However, ΔCP truly equals ΔCP(protein unfolding) + ΔCP(GuHCl interactions), so that the experimentally determined ΔCP value is an upper limit of ΔCP for protein unfolding (Makhatadze and Privalov 1992; Agashe and Udgaonkar 1995). Because the ΔH values for A. ambivalens Fd at pH 2.5 (at no, or low, GuHCl concentration) and at pH 7 correlate (Fig. 2 ▶), contributions to ΔH from GuHCl–protein interactions appear small. Moreover, potential errors in the ΔH values because of irreversibility are assumed negligible. If the ΔH values are affected by the reactions leading to irreversibility, the effect would be similar at all temperatures and, therefore, the slope of ΔH versus T (i.e., ΔCP) would not be perturbed (see Discussion). Strikingly, at pH 2.5 conditions, thermal unfolding takes place at an almost 60°C lower temperature than at pH 7: Tm is 64°C and 122°C in buffer at pH 2.5 and pH 7.0, respectively (Table 1).

Table 1.

Transition midpoints (Tm) and unfolding-enthalpies (ΔH(Tm)) for Fd thermal denaturation at the pH/denaturant conditions indicated

Solution pH [GuHCl] (M) Tm (K) ΔH(Tm) (kJ/mol)
7.0 7.2 308 (±1) 137 (±7)
7.0 6.5 317 (±1) 176 (±7)
7.0 5.5 329 (±1) 205 (±5)
7.0 5.0 336 (±1) 221 (±7)
7.0 0 395 (±2)a 430 (±10)b
2.5 2.0 318 (±1) 169 (±5)
2.5 0 337 (±1) 244 (±6)
Open in a new tab

a Determined by extrapolation in (Wittung-Stafshede et al. 2000).

b Determined by extrapolation in Figure 1 ▶.

Fig. 2.

Fig. 2.

Open in a new tab

The temperature dependence of the enthalpy of unfolding for Fd obtained from analyses of thermal melts in different conditions (data, see Table 1). A linear least-squares fit of the ΔH(Tm) versus Tm values provides a ΔCP of 3.15 ± 0.1 kJ/(mole • K).

At the organism's optimum growth temperature, 80°C, the stability is roughly 40 kJ/mole at pH 7 (calculated from ΔCP and ΔH[Tm] by a modified Gibbs-Helmholtz equation [Privalov et al. 1986; McCrary et al. 1996]), whereas at this temperature and pH 2.5 conditions the native state of Fd is not favored thermodynamically (Tm[pH 2.5] < 80°C). Although the organism grows optimally at low pH, the interiors of the cells do not necessarily adopt acidic pH. Indeed, for the closely related archaeon Thermoplasma acidophilum, also growing at pH 2, an internal pH of 6.6 has been measured (Hsung and Haug 1975).

Chemically induced Fd unfolding at neutral pH

A concentration of 8 M GuHCl did not promote Fd to unfold fully at pH 7 (Wittung-Stafshede et al. 2000), and no ΔGU(H2O) at 20°C, pH 7 has been reported. We now used guanidine thiocyanate (GuSCN; Cota and Clarke 2000), which is a stronger denaturant than GuHCl, to promote Fd unfolding at pH 7. GuSCN absorbs far-UV light to a significant extent, prohibiting the use of the change in Fd's far-UV CD signal to monitor GuSCN-induced unfolding. Instead, Fd unfolding was monitored by changes in Trp emission; upon protein unfolding the fluorescence increases in intensity, and the maximum shifts from 345 ± 2 to 355 ± 3 nm (Fig. 1C ▶). An identical equilibrium transition was observed by monitoring the disappearance of visible absorption (data not shown). Moreover, GuHCl-promoted unfolding of Fd at lower pHs (discussed below) revealed that, at each pH, identical transition midpoints and ΔGUs were found, independent of whether far-UV CD, emission maximum, or visible absorption was monitored. Also, Fd samples at conditions near the unfolding midpoints could be incubated for up to 24 h without changes in their spectroscopic properties. Taken together, these results support that equilibrium unfolding of Fd can be treated as an apparent two-state process and that estimations of thermodynamic values are appropriate.

Titration of Fd with GuSCN (pH 7) yields a sharp transition with a midpoint at 3.1 M GuSCN, shown in Figure 3 ▶. Using a two-state fit, ΔGU(H2O) at pH 7 (20°C) was calculated to be 79 ± 8 kJ/mole. Comparing GuHCl- and GuSCN-induced unfolding experiments with myoglobin provided a scaling factor of 2.3 for interconverting between GuSCN and GuHCl unfolding midpoint concentrations (an identical factor was derived from data in Cota and Clark [2000]). Using this scaling factor, the apparent midpoint for a GuHCl titration of Fd at pH 7 would occur at 7.1 M GuHCl.

Fig. 3.

Fig. 3.

Open in a new tab

GuSCN titration of Fd at pH 7.0 (degree of unfolding followed by tryptophan emission changes). Continuous line is a two-state fit that gives ΔGU(H2O) of 79 ± 8 kJ/mole.

pH dependence of Fd thermodynamic stability

The A. ambivalens Fd adopts the native structure in a wide pH range (20°C): There are no changes in visible absorption or far-UV CD from pH 2.0 to pH 12 (Wittung-Stafshede et al. 2000). However, the Tm is much lower at pH 2.5 than at pH 7 (Table 1). For many proteins, as also appears to be the case for the Fd, the highest thermodynamic stability is found around pH 7, regardless of whether the isoelectric point of the protein is acidic or basic (the A. ambivalens Fd is acidic). To characterize the pH dependence for this Fd thermodynamic stability further, GuHCl titrations were performed at pH 2.5, 4.0, 5.2, 8.5, and 10. The transitions were monitored by far-UV CD (220 nm), visible absorption (410 nm), and tryptophan emission (350 nm); all methods gave identical results for each condition. The transition midpoints and corresponding ΔGU(H2O) are listed in Table 2 and shown in Figure 4 ▶. It is clear from the data that there is a strong pH dependence of the Fd stability. The stability decreases drastically below pH 5, although it appears almost constant above pH 7 (only slightly lower at pH 10 than at pH 7). At pH 2.5, GuHCl-induced unfolding occurs with a midpoint of 2.3 M GuHCl, and a ΔGU(H2O) of 20 ± 1 kJ/mole was estimated (which is less than a third of the stability at pH 7).

Table 2.

Transition midpoints, [GuHCl]1/2, and unfolding-free energies in absence of denaturant, ΔGU(H2O), for Fd unfolding at the indicated pHs (20°C)

Solution pH [GuHCl]1/2 (M) ΔGU(H2O) (kJ/mol)
10* 6.3 70 (±3)
8.5 6.6 —b
7.0 7.1c 79 (±8)
5.2 6.9 —b
4.0 4.4 43 (±5)
2.5 2.3 20 (±1)
Open in a new tab

a Data from Wittung-Stafshede et al. (2000).

b Not determined; sufficient unfolded baselines were not available.

c GuSCN titration, GuHCl midpoint estimated by scaling (see text).

Fig. 4.

Fig. 4.

Open in a new tab

The pH dependence of Fd stability. The midpoints for GuHCl-induced unfolding transitions are shown as a function of pH (data, see Table 2). (Inset) ΔGU(H2O) plotted as a function of pH.

Fd unfolding mechanism: Cluster rearrangements

Upon Fd unfolding, the 410-nm absorption disappears, supporting the idea that the iron clusters disassemble in the unfolded state. Unfolding in 7 M GuHCl at pH 10 was shown earlier to promote conversion of the native brownish protein to a transient intermediate (black/purple) form with absorption features at 520 nm and 610 nm (Wittung-Stafshede et al. 2000). Based on comparisons with model complexes and the resemblance to beef-heart aconitase at high pH (Hagen et al. 1983; Kennedy et al. 1984), the Fd pH-10 unfolding intermediate was suggested to incorporate linear [3Fe-4S] clusters (Wittung-Stafshede et al. 2000). Formation of the linear clusters occurred in parallel with the disappearance of the secondary structure of the polypeptide, but the new clusters remained bound to the unfolded polypeptide (Wittung-Stafshede et al. 2000). A subsequent phase correlated with the conversion of the black/purple species into one lacking color, presumably the unfolded protein from which the irons had dissociated (Wittung-Stafshede et al. 2000).

To address if transient linear [3Fe-4S] intermediates (not detected in the equilibrium experiments) are also observed upon Fd unfolding at lower pH, we performed additional kinetic studies. GuHCl-induced unfolding kinetics of Fd at pH 2.5, using stopped-flow mixing because of the rapid reaction (Fig. 5A ▶), reveals the transient appearance of absorption at 610 nm (1/ku = 1.7 ± 0.1 sec, 4 M GuHCl, pH 2.5). This phase is followed by a slower decrease to zero absorption intensity (1/ku = 120 ± 5 sec, 4 M GuHCl, pH 2.5). Moreover, on adding 3.5 M GuSCN to Fd at pH 7, a 610-nm absorption band appears within the first minute after mixing. This process is followed by a slower (over several minutes) decrease, resulting in complete disappearance of visible absorption at both 410 and 610 nm. In identical pH 7 conditions, 520- and 610-nm absorption bands are also observed transiently with the Sulfolobus acidocaldarius seven-iron ferredoxin (Fig. 5B ▶). The S. acidocaldarius ferredoxin shares 89% amino acid identity with the A. ambivalens Fd. Thus, in a wide pH range (2.5 to 10), the unfolding path for the A. ambivalens Fd appears to involve a transient state in which the polypeptide coordinates rearranged, possibly linear [3Fe-4S], iron clusters.

Fig. 5.

Fig. 5.

Fig. 5.

Open in a new tab

(A) Unfolding kinetics monitored at 610 nm upon introducing native Fd into 4 M GuHCl (pH 2.5). A slow phase leads to a clear protein solution (monoexponential fit: 1/kU = 120 ± 5 sec). (Inset) A first (fast) phase (30% of total amplitude) correlates with an increase in the 610 nm absorption (monoexponential fit: 1/kU = 1.7 ± 0.1 sec). (B) Visible absorption of native, intermediate (after 30 sec in 3.5 M GuSCN) and unfolded (after 20 min in 3.5 M GuSCN) forms of ferredoxin from S. acidocaldarius at pH 7. (Inset) The absorption decrease at 410 nm and the appearance and disappearance of absorption at 610 nm as a function of time upon adding 3.5 M GuSCN to ferredoxin from S. acidocaldarius at pH 7.

Fd coordinates one zinc ion in the native structure. We used a fluorescent dye, APTRA-BTC (Molecular Probes) that is specific for zinc (absorption shifts from 380 to 460 nm upon zinc binding [Nowicki 1994]) to reveal if the zinc in Fd remains bound or not upon protein unfolding. There was no increase in dye excitation intensity at 460 nm (with emission monitored at 500 nm) above background for either folded or unfolded Fd, whereas the same concentration of free zinc added to the dye induced a large increase in excitation at this wavelength (data not shown). We also performed zinc analysis on a dialyzed (to remove any unbound zinc) unfolded Fd sample by inductively coupled plasma (ICP) atomic emission spectrometry: After dialysis, 50% of the original amount of zinc was still detected in the protein sample. Thus, both metal analysis and dye-binding experimental approaches give strong evidence that the zinc in Fd remains bound to the protein also after polypeptide unfolding.

Discussion

The experimental data on Fd from the archaeon A. ambivalens have been treated as thermodynamic equilibrium data, although the unfolding process of this protein is irreversible. The following observations justify this treatment empirically (to a first approximation): (1) no aggregation observed for unfolded protein, (2) no time dependence of the thermal melting profiles, (3) agreement between thermally and GuHCl-derived ΔGU(H2O) values (investigated at pH 2.5 and pH 7; data not shown), (4) identical transition midpoints and ΔGU(H2O) values when unfolding is monitored by different spectroscopic probes, and (5) protein incubation near unfolding midpoints, for up to 24 h, does not result in spectroscopic changes. It is not known if the irreversible steps during Fd unfolding, believed to be iron–sulfur cluster dissociation and decomposition, contribute to the estimated enthalpy changes. This causes concern with respect to the validity of the absolute ΔH values. However, because an enthalpic effect from the irreversible steps should be similar at each Tm, the slope of the ΔH versus Tm plot will still give a reliable ΔCP value.

Fd stability as a function of temperature

It is clear that A. ambivalens Fd is very stable toward both thermal and chemical denaturant unfolding near neutral pH. At pH 7 (20°C), we estimated the ΔGU for Fd to be 79 ± 8 kJ/mole. This high thermodynamic stability is in accord with the high thermal melting point estimated at pH 7 (Tm = 122°C). Ferredoxin from T. maritima (60 residues, one [4Fe-4S] cluster) was also reported to be stable beyond the boiling point of water (Pfeil et al. 1997). ΔCP relates to the amount of hydrophobic surface area exposed upon unfolding. The estimated ΔCP of 3.15 ± 0.1 kJ/(mole • K) for the A. ambivalens Fd, determined from thermal melts at different buffer/denaturant conditions (Fig. 2 ▶; Table 1), is low. Data sets for a large number of mesophilic proteins have yielded an average value for ΔCP of 59 J/(mole • K) per residue (Makhatadze and Privalov 1995; Myers et al. 1995; Robertson and Murphy 1997); this corresponds to a predicted ΔCP of 6.1 kJ/(mole • K) for unfolding of the 103-residue Fd. This predicted value is larger than the experimental one, indicating a more compact unfolded state (perhaps with local structure around the metal centers; see below) as compared with a random-coil polypeptide. For another iron–sulfur protein, the Ectothiorhodospira halophila high-potential iron–sulfur protein (similar size and structure as A. ambivalens Fd but from a mesophile), ΔCP was found to be 4.5 kJ/(mole • K) (Iwagami et al. 1995), comparable to the value for A. ambivalens Fd.

A low ΔCP corresponds to a shallow dependence of the ΔGU on temperature. In addition, the stability for Fd at 20°C, pH 7 (Fig. 3 ▶) appears very high. Taken together, these results indicate that Fd achieves its increased thermostability through a combination of a high maximal ΔGU (model one, see introduction) and a low ΔCP (model three, see introduction). This combination of mechanisms to attain thermostability was also observed for the Thermus thermophilus RNase H protein (Hollien and Marqusee 1999), and several other investigations have supported the `flattened' third model (McCrary et al. 1996; Beadle et al. 1999; Hollien and Marqusee 1999). In contrast to the very high ΔGU found at low temperatures, at the A. ambivalens optimal growth temperature around 80°C, the Fd stability at pH 7 is not more than 40 kJ/mole, a value similar to that for an average mesophilic protein at 20°C (Makhatadze and Privalov 1995).

Fd stability as a function of pH

The stability of A. ambivalens Fd is decreased at lower pHs (Fig. 4 ▶; Table 2), supporting the importance of electrostatic interactions for the overall stability. The abrupt decrease in stability at pH < 4 may indicate that interactions involving deprotonated Asp or Glu side chains are contributing to the high stability at neutral pH. At low pH, the free energy contribution associated with proton binding decreases the stability of the native conformation because of the favorable protonation of carboxylate groups (e.g., Asp or Glu) in the unfolded state (Luisi and Raleigh 2000). The most prominent pairwise ionic interactions in native Fd appear to be those between Asp 99–Arg 7 and Asp 41–Arg 10 (in the N-terminal region, the first pair directly linking the N and C termini) and also between Asp 64–Lys 73 and Asp 43–His 68 (see Fig. 6 ▶; Wittung-Stafshede et al. 2000). All of these residues are conserved among seven-iron ferredoxins from Sulfolobales (Gomes et al. 1998), perhaps indicating that such ion pairs, or the resulting overall electrostatic field, are common strategies to obtain thermostability in these proteins.

Fig. 6.

Fig. 6.

Open in a new tab

Three-dimensional model of native A. ambivalens Fd (Teixeira et al. 1995; Gomes et al. 1998) highlighting the residues (and distances between pairs) that are involved in salt bridges and the three metal centers (prepared with WebLab viewer).

That electrostatic interactions seem to be a common factor regulating the thermal tolerance of thermostable proteins has been shown for many proteins (Pace 2000). One obvious reason for this is the fact that the dielectric constant of H2O decreases at high temperatures, which makes electrostatic interactions stronger (and thus more important) at higher temperatures. For example, Sac7d from the hyperthermophile S. acidocaldarius contains an unusually large number of potential surface ion pairs and, accordingly, a strong pH dependence was observed for the thermodynamic stability (McCrary et al. 1996). For Sso7d from Sulfolobus solfataricus, a linkage model between proton binding and thermal unfolding, involving two surface salt bridges, could be derived (Graziano et al. 1999). Moreover, it was inferred from the structure of the rubredoxin from the hyperthermophile Pyrococcus furiosus that enhanced electrostatic contacts at the N terminus stabilize against fraying (Hiller et al. 1997). Additional support for the involvement (and higher frequency) of electrostatic interactions in thermostable proteins is given by a direct comparison of 14 thermophilic proteins and their mesophilic homologs (Spassov et al. 1995). In a recent comparison of four hyperthermophilic proteins with their mesophilic homologs, it was found that the optimum placement of the charged groups within the protein structure, and not isolated ion pairs, was more favorable in the thermophilic proteins (Xiao and Honig 1999).

Role of metal centers for Fd folding

Two iron–sulfur clusters are linked to the Fd polypeptide through eight cysteine sulfurs. Upon Fd unfolding at low pH, the cysteine sulfurs, involved in iron–sulfur cluster binding, may become protonated (and thus would contribute to the pH dependence of the stability). Using Ellman's assay we did not, however, detect any free sulfurs in the unfolded state of Fd, indicating that sulfur oxidation is promoted after Fd unfolding and metal cluster decomposition. Sulfur oxidation was shown to occur in oxidized azurin, a copper protein in which a cysteine sulfur is one metal ligand, upon GuHCl-induced protein unfolding and metal dissociation (Leckner et al. 1997), as well as upon iron release in desulforedoxin (in which iron coordinates four cysteine sulfurs) at low pH (Kennedy et al. 1998). The mechanism of cluster decomposition and subsequent cysteine oxidation in unfolded Fd may still occur through protonated cysteine intermediates. In fact, the strong pH dependence observed for the Fd stability indicates that at least some of the eight cysteine sulfurs become protonated after polypeptide unfolding and subsequent cluster decomposition. Ionic interactions on the protein's surface and the cysteine ligands of the iron–sulfur clusters may together explain the sharp pH dependence of Fd thermodynamic stability.

The residues involved in zinc coordination, three His and one Asp, are deprotonated in folded Fd but may become protonated upon low-pH unfolding, that is, if zinc dissociates. We revealed, however, that the zinc remains bound to the polypeptide upon ferredoxin unfolding (also at low pH). The same observation was made for zinc-substituted azurin: The zinc ion remained coordinated, presumably to some of the native-state ligands, also in the unfolded state (Leckner et al. 1997; Romero et al. 1998). According to the three-dimensional structural model (Wittung-Stafshede et al. 2000), the zinc in A. ambivalens Fd appears to hold the N-terminal region and the core fold together. In a recent study of the thermal behavior of truncated mutants of the Sulfolobus strain-7 ferredoxin (primary structure 95% identical to A. ambivalens Fd), also containing the zinc-binding site (Fujii et al. 1997), it was shown that the zinc was partly responsible for the thermal stabilization (Kojoh et al. 1999). It is possible that zinc coordination in the unfolded state (to some or all of the native-state ligands) decreases the entropy of this state; in this way the stability of the folded structure is increased. Residual structure, and/or a compact denatured state, is in accord with the low ΔCP found upon Fd unfolding.

That the iron–sulfur clusters stabilize the native structure of Fd is clear from the observation that chemical and thermal unfolding processes (leading ultimately to cluster dissociation and degradation) are irreversible. Fd refolding attempts at pH 2.5 (but not at higher pHs) showed formation of a large fraction of secondary structure. However, no characteristic color appeared, indicating the presence of only apoprotein under these conditions. We found the "refolded" apoprotein to enhance ANS emission significantly (11×), as well as to have a noncooperative far-UV CD thermal profile (data not shown). Both results are in accord with a nonnative, molten-globule-like structure for the apoprotein at low pH. It was shown for the [2Fe-2S] spinach ferredoxin (Pagani et al. 1986) that (enzymatic) cluster assembly and insertion drive the polypeptide toward full recovery of the native form. Thus, without inserted iron–sulfur clusters, ferredoxin proteins appear incapable of adopting their native folds. This indicates that for ferredoxin formation in vivo, cluster insertion may precede polypeptide folding, or chaperone proteins may be required. There are several reports supporting that cofactors (e.g., heme, copper, flavin, iron–sulfur clusters) can remain coordinated to their corresponding polypeptides upon unfolding (Leckner et al. 1997; Wittung-Stafshede 1999; Apiyo et al. 2000; Moczygemba et al. 2000). It is thus possible that for several cofactor-binding proteins in vivo, coordination of cofactor(s) to the unfolded polypeptide occurs before folding takes place. The cofactor(s) may in this way aid in shifting the equilibrium toward a native-like ordered structure.

In a wide pH range, Fd unfolding first leads (within seconds) to a denatured state in which the polypeptide coordinate rearranged, possibly linear [3Fe-4S], iron clusters. On a longer time scale, the new clusters decompose, yielding unfolded apo-Fd. Rearrangements into linear three-iron clusters were shown for beef-heart aconitase at high pH (pH > 9.5) and at neutral pH in the presence of urea (Kennedy et al. 1984), as well as for another seven-iron Fd at pH 7 (in the presence of GuSCN; Fig. 5B ▶). Unlike our results on the two thermostable ferredoxins, the linear [3Fe-4S] cluster in partly unfolded aconitase was stable for days at high pH and showed a half-time of 40 min in urea and pH 8 (Kennedy et al. 1984). Further studies may reveal if formation of this intermediate species is part of a general mechanism for cluster decomposition in iron–sulfur proteins.

Conclusion

A. ambivalens Fd is intrinsically extremely resistant to chemical and thermal perturbations around neutral pH. Electrostatic surface interactions and the metal centers (two iron–sulfur clusters and a zinc site) are factors contributing to the high stability of Fd. The unfolding pathway for Fd involves transient rearranged iron–sulfur clusters, which subsequently degrade, whereas the zinc ion remains coordinated to the polypeptide after complete denaturation.

Materials and methods

Chemicals

Chemical denaturants guanidine hydrochloride (GuHCl) and guanidine thiocyanate (GuSCN) were of highest purity, purchased from Sigma. All chemicals, including buffers and Ellman's reagent (thiosulfo) benzoate (DTNB), were from Sigma except the fluorescent dyes, APTRA-BTC (A-6895) and 1-anilino-naphtalene-8-sulfonate (ANS), which were obtained from Molecular Probes.

Protein preparation

A. ambivalens cells were grown in a 10-L fermentor as described previously (Teixeira et al. 1995). S. acidocaldarius cells were a kind gift from Prof. K.O. Stetter and H. Huber from Regensburg University, Germany. The A. ambivalens and S. acidocaldarius Fds were purified as detailed in Gomes et al. (1998). Protein purity was confirmed by a single band on 15% SDS–PAGE and a clean N-terminal sequence.

Denaturant-induced unfolding

Chemical denaturant GuHCl was used to promote protein unfolding (at 20°C) at all pHs except for pH 7; at the latter pH GuSCN was used. GuHCl/GuSCN titrations were performed at room temperature with 20 μM Fd in 5 mM phosphate (pH 7.0 and 5.2), Tris-HCl (pH 8.5), acetate (pH 4.0) or glycine (pH 2.5 and 10) buffer. There was no protein-concentration dependence for the unfolding transitions (in accord with apparent two-state transitions although irreversible processes; see text for discussion). Samples were incubated for 30 min before measurements (variation of the incubation time from 2 min to 24 h did not alter the observed transitions). Unfolding was monitored by far-UV CD (200–300 nm, 1-mm cell) on an OLIS spectropolarimeter, visible absorption (300–700 nm, 1-cm cell) on a Cary-50 spectrophotometer, and tryptophan emission (300–450 nm, excitation 280 nm, 1-cm cell) on a Varian Eclipse fluorometer.

Unfolding transitions were analyzed using a two-state model, fF + fU = 1, where fF and fU represent the fractions of total protein in the folded and denatured conformations, respectively. At each point in the reaction, the equilibrium constant, KU, and the free energy change, ΔGU, are related as follows:

graphic file with name M1.gif 1
graphic file with name M2.gif 2

ΔGU can be expressed as a function of the denaturant concentration:

graphic file with name M3.gif 3

In this equation, m describes the extent of hydrophobic surface exposure upon unfolding (Tanford 1970; Pace 1975) and ΔGU(H2O) is the free energy of unfolding in aqueous solution. Direct fits (in KaleidaGraph) to the experimental unfolding curves were performed using the following expression, derived from equations 1, 2, and 3:

graphic file with name M4.gif 4

Yobs, YU, and YF are the observed spectroscopic signal, denatured protein baseline, and folded protein baseline, respectively. From the fits, ΔGU(H2O) was determined at each condition (Table 2). The transition midpoints were calculated as ΔGU(H2O)/m or by direct inspection of the transitions. Reported uncertainties were obtained from the goodness of the fits.

Comparative GuHCl and GuSCN titrations were performed on sperm whale Met myoglobin; degree of unfolding was monitored by changes in heme absorption (409 nm) as reported previously (Moczygemba et al. 2000). The observed midpoints, [GuHCl]1/2 = 1.6 ± 0.1 M and [GuSCN]1/2 = 0.7 ± 0.1 M, were used to calculate a scaling factor of 2.3 for interconversion between GuHCl and GuSCN midpoint concentrations. This scaling factor is in excellent agreement with GuHCl and GuSCN data reported for barnase (Cota and Clarke 2000). The activity of GuSCN was shown to be linear up to at least 3.5 M; therefore, no correction of molar concentrations into activity units was performed (Pandya et al. 1999).

Thermally induced unfolding

Temperature-induced unfolding of Fd was monitored by far-UV CD (200–300 nm) in various pH/GuHCl conditions (see Table 1). CD spectra of 20 μM ferredoxin were recorded every 5°C from 20°C to 95°C, with 5 min of equilibrium at each temperature. Longer equilibration times (10 min) at each temperature did not change the melting profiles. In the end, the temperature was decreased to 20°C and a CD spectrum was recorded to check for refolding. The thermal reactions were monitored on the OLIS instrument connected to a digitally controlled water bath (Julabo). Analyses of thermal transitions according to a modified Van't Hoff equation (Makhatadze and Privalov 1995; Wittung-Stafshede et al. 1998), which accounted for the temperature dependence of ΔH and ΔS (first using a predicted ΔCp of 6 kJ/mole [Myers et al. 1995] followed by revision using a value of 3 kJ/mole), yielded ΔH(Tm) at each Tm (Table 1).

Unfolding kinetics

Kinetic unfolding measurements (20°C) were made on an Applied Photophysics SX.18MV stopped-flow reaction analyzer; absorption mode (410 nm and 610 nm detection); 1 : 1 mixing of native Fd and 8 M GuHCl (buffered to pH 2.5); pathlength 2 mm. No amplitude changes occurred in the dead time (<2 msec) of the instrument. For each time range, a minimum of eight kinetic traces was averaged and fit to monophasic decay equations using a nonlinear least-squares algorithm supplied by Applied Photophysics.

Irreversibility of unfolding reactions

Fd unfolding is irreversible upon both thermal and chemical denaturant perturbations, as also reported earlier (Wittung-Stafshede et al. 2000). Experimental results and related literature supporting a thermodynamic treatment of the collected data despite the irreversibility are discussed throughout the text. The unfolded protein does not aggregate. Tests with Ellman's reagent (DTNB) show that at least some of the cysteine sulfurs, which coordinate to the iron–sulfur clusters in native Fd, also remain unavailable in the unfolded state (indicating sulfur oxidation). At low pH (but not at neutral and high pHs), ∼80% of the native amount of secondary structure was regained in refolding attempts (by dilution from high to low denaturant conditions, 20°C, monitored by far-UV CD). No characteristic color was observed for this species, indicating the presence of only apoprotein. This nonnative (but not fully unfolded) apo form of Fd was characterized in terms of ANS binding (fluorescent probe for exposed hydrophobic surfaces) and thermal melting (see Discussion).

Acknowledgments

We thank Susanne Griffin for experimental help. The Louisiana Board of Regents (LEQSF[1999–02]-RD-A-39; P.W.-S.), the National Institutes of Health (GM59663–01A2), the Newcomb Foundation (Fellowship; C.M.) and the Keck Foundation (instrumentation) are acknowledged for financial support.

The publication costs of this article were defrayed in part by payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 USC section 1734 solely to indicate this fact.

Abbreviations

  • ΔGU, unfolding free energy

  • Tm, melting temperature

  • Fd, ferredoxin

  • GuHCl, guanidine hydrochloride

  • ΔH[Tm], enthalpy of unfolding

  • ΔCP, heat capacity of unfolding

  • GuSCN, guanidine thiocyanate

  • CD, circular dichroism

Article and publication are at http://www.proteinscience.org/cgi/doi/10.1101/ps.49401

References

  1. Agashe, V. and Udgaonkar, J. 1995. Thermodynamics of denaturation of barstar: Evidence for cold denaturation and evaluation of the interaction with guanidine hydrochloride. Biochemistry 34 3286–3299. [DOI] [PubMed] [Google Scholar]
  2. Apiyo, D., Guidry, J., and Wittung-Stafshede, P. 2000. No cofactor effect on equilibrium unfolding of Desulfovibrio desulfuricans flavodoxin. Biochim. Biophys. Acta 1479 214–224. [DOI] [PubMed] [Google Scholar]
  3. Beadle, B.M., Baase, W.A., Wilson, D.B., Gilkes, N.R., and Shoichet, B.K. 1999. Comparing the thermodynamic stabilities of a related thermophilic and mesophilic enzyme. Biochemistry 38 2570–2576. [DOI] [PubMed] [Google Scholar]
  4. Cota, E. and Clarke, J. 2000. Folding of β-sandwich proteins: Three-state transition of a fibronectin type III module. Prot. Sci. 9 112–120. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Edge, V., Allewell, N.M., and Sturtevant, J.M. 1985. High-resolution differential scanning calorimetric analysis of the subunits of Escherichia coli aspartate transcarbamoylase. Biochemistry 24 5899–5906. [DOI] [PubMed] [Google Scholar]
  6. Fujii, T., Hata, Y., Oozeki, M., Moriyama, H., Wakagi, T., Tanaka, N., and Oshima, T. 1997. The crystal structure of zinc-containing ferredoxin from the thermoacidophilic archaeon Sulfolobus sp. strain 7. Biochemistry 36 1505–1513. [DOI] [PubMed] [Google Scholar]
  7. Gomes, C.M., Faria, A., Carita, J.C., Mendes, J., Regalla, M., Chicau, P., Huber, H., Stetter, K.O., and Teixeira, M. 1998. Di-cluster, seven iron ferredoxins from hyperthermophilic Sulfobales. J. Biol. Inorg. Chem. 3 499–507. [Google Scholar]
  8. Graziano, G., Catanzano, F., and Nappa, M. 1999. Linkage of proton binding to the thermal unfolding of Sso7d from the hyperthermophilic archaebacterium Sulfolobus solfataricus. Int. J. Biol. Macromol. 26 45–53. [DOI] [PubMed] [Google Scholar]
  9. Hagen, K., Watson, A., and Holm, R. 1983. Synthetic routes to iron–sulphur clusters from the common precursor [Fe(SC2H5)4]2+. J. Am. Chem. Soc. 105 3905–3913. [Google Scholar]
  10. Hiller, R., Zhou, Z.H., Adams, M.W., and Englander, S.W. 1997. Stability and dynamics in a hyperthermophilic protein with melting temperature close to 200 degrees C. Proc. Natl. Acad. Sci. 94 11329–11332. [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Hollien, J. and Marqusee, S. 1999. A thermodynamic comparison of mesophilic and thermophilic ribonucleases H. Biochemistry 38 3831–3836. [DOI] [PubMed] [Google Scholar]
  12. Hsung, J. and Haug, A. 1975. Intracellular pH of Thermoplasma acidophila. Biochim. Biophys. Acta 389 477–482. [DOI] [PubMed] [Google Scholar]
  13. Hu, C.Q. and Sturtevant, J.M. 1987. Thermodynamic study of yeast phosphoglycerate kinase. Biochemistry 26 178–182. [DOI] [PubMed] [Google Scholar]
  14. Iwagami, S., Creagh, A., Haynes, C., Borsari, M., Felli, I., Piccioli, M., and Eltis, L. 1995. The role of a conserved tyrosine residue in high-potential iron–sulfur proteins. Prot. Sci. 4 2562–2572. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Jaenicke, R. and Bohm, G. 1998. The stability of proteins in extreme environments. Curr. Opin. Struct. Biol. 8 738–748. [DOI] [PubMed] [Google Scholar]
  16. Kennedy, M.C., Kent, T.A., Emptage, M., Merkle, H., Beinert, H., and Munck, E. 1984. Evidence for the formation of a linear [3Fe-4S] cluster in partially unfolded aconitase. J. Biol. Chem. 259 14463–14471. [PubMed] [Google Scholar]
  17. Kennedy, M., Yu, L., Lima, M.J., Ascenso, C., Czaja, C., Moura, I., Moura, J., and Rusnak, F. 1998. Metal binding to the tetrathiolate motif of desulforedoxin and related polypeptides. J. Biol. Inorg. Chem. 3 643–649. [Google Scholar]
  18. Kojoh, K., Matsuzawa, H., and Wakagi, T. 1999. Zinc and an N-terminal extra stretch of the ferredoxin from a thermoacidophilic archaeon stabilize the molecule at high temperature. Eur. J. Biochem. 264 85–91. [DOI] [PubMed] [Google Scholar]
  19. Leckner, J., Bonander, N., Wittung-Stafshede, P., Malmstrom, B.G., and Karlsson, B.G. 1997. The effect of the metal ion on the folding energetics of azurin: A comparison of the native, zinc and apoprotein. Biochim. Biophys. Acta 1342 19–27. [DOI] [PubMed] [Google Scholar]
  20. Luisi, D.L. and Raleigh, D.P. 2000. pH-dependent interactions and the stability and folding kinetics of the N-terminal domain of L9. Electrostatic interactions are only weakly formed in the transition state for folding [In Process Citation]. J. Mol. Biol. 299 1091–1100. [DOI] [PubMed] [Google Scholar]
  21. Makhatadze, G. and Privalov, P. 1992. Protein interactions with urea and guanidinium chloride. J. Mol. Biol. 226 491–505. [DOI] [PubMed] [Google Scholar]
  22. ———. 1995. Energetics of protein structure. Adv. Prot. Chem. 47 307–425. [DOI] [PubMed] [Google Scholar]
  23. Manly, S.P., Matthews, K.S., and Sturtevant, J.M. 1985. Thermal denaturation of the core protein of lac repressor. Biochemistry 24 3842–3846. [DOI] [PubMed] [Google Scholar]
  24. McCrary, B.S., Edmondson, S.P., and Shriver, J.W. 1996. Hyperthermophile protein folding thermodynamics: Differential scanning calorimetry and chemical denaturation of Sac7d. J. Mol. Biol. 264 784–805. [DOI] [PubMed] [Google Scholar]
  25. Moczygemba, C., Guidry, J., and Wittung-Stafshede, P. 2000. Heme orientation affects holo-myoglobin folding and unfolding kinetics. FEBS Lett. 470 203–206. [DOI] [PubMed] [Google Scholar]
  26. Myers, J., Pace, C., and Scholtz, J. 1995. Denaturant m values and heat capacity changes. Prot. Sci. 4 2138–2148. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Nowicki, J.L. 1994. Determination of zinc in seawater using flow injection analysis with fluorometric detection. Anal. Chem. 66 2732–2736. [Google Scholar]
  28. Otaka, E. and Ooi, T. 1987. Examination of protein sequence homologies: IV. Twenty-seven bacterial ferredoxins. J. Mol. Evol. 26 257–267. [DOI] [PubMed] [Google Scholar]
  29. Pace, C.N. 1975. The stability of globular proteins. CRC Crit. Rev. Biochem. 3 1–43. [DOI] [PubMed] [Google Scholar]
  30. ———. 2000. Single surface stabilizer. Nat. Struct. Biol. 7 345–346. [DOI] [PubMed] [Google Scholar]
  31. Pace, N.R. 1991. Origin of life—facing up to the physical setting. Cell 65 531–533. [DOI] [PubMed] [Google Scholar]
  32. Pagani, S., Vecchio, G., Iametti, S., Bianchi, R., and Bonomi, F. 1986. On the role of 2Fe-2S cluster in the formation of the structure of spinach ferredoxin. Biochim. Biophys. Acta 870 538–544. [Google Scholar]
  33. Pandya, M.J., Williams, P.B., Dempsey, C.E., Shewry, P.R., and Clarke, A.R. 1999. Direct kinetic evidence for folding via a highly compact, misfolded state. J. Biol. Chem. 274 26828–26837. [DOI] [PubMed] [Google Scholar]
  34. Pfeil, W., Gesierich, U., Kleemann, G.R., and Sterner, R. 1997. Ferredoxin from the hyperthermophile Thermotoga maritima is stable beyond the boiling point of water. J. Mol. Biol. 272 591–596. [DOI] [PubMed] [Google Scholar]
  35. Privalov, P.L. and Potekhin, S.A. 1986. Scanning microcalorimetry in studying temperature-induced changes in proteins. Meth. Enzymol. 131 4–51. [DOI] [PubMed] [Google Scholar]
  36. Privalov, P.L., Griko Yu, V., Venyaminov, S., and Kutyshenko, V.P. 1986. Cold denaturation of myoglobin. J. Mol. Biol. 190 487–498. [DOI] [PubMed] [Google Scholar]
  37. Rees, D.C. and Adams, M.W. 1995. Hyperthermophiles: Taking the heat and loving it. Structure 3 251–254. [DOI] [PubMed] [Google Scholar]
  38. Robertson, A.D. and Murphy, K.P. 1997. Protein structure and the energetics of protein stability. Chem. Rev. 97 1251–1267. [DOI] [PubMed] [Google Scholar]
  39. Romero, C., Moratal, J.M., and Donaire, A. 1998. Metal coordination of azurin in the unfolded state. FEBS Lett. 440 93–98. [DOI] [PubMed] [Google Scholar]
  40. Spassov, V.Z., Karshikoff, A.D., and Ladenstein, R. 1995. The optimization of protein–solvent interactions: Thermostability and the role of hydrophobic and electrostatic interactions. Prot. Sci. 4 1516–1527. [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Szilagyi, A. and Zavodszky, P. 2000. Structural differences between mesophilic, moderately thermophilic and extremely thermophilic protein subunits: Results of a comprehensive survey [In Process Citation]. Struct. Fold. Des. 8 493–504. [DOI] [PubMed] [Google Scholar]
  42. Tanford, C. 1970. Protein denaturation. C. Theoretical models for the mechanism of denaturation. Adv. Prot. Chem. 24 1–95. [PubMed] [Google Scholar]
  43. Teixeira, M., Batista, R., Campos, A.P., Gomes, C., Mendes, J., Pacheco, I., Anemuller, S., and Hagen, W.R. 1995. A seven-iron ferredoxin from the thermoacidophilic archaeon Desulfurolobus ambivalens. Eur. J. Biochem. 227 322–327. [DOI] [PubMed] [Google Scholar]
  44. Vogt, G. and Argos, P. 1997. Protein thermal stability: Hydrogen bonds or internal packing? Fold. Des. 2 S40–S46. [DOI] [PubMed] [Google Scholar]
  45. Wachtershauser, G. 1992. Groundworks for an evolutionary biochemistry: The iron–sulphur world. Prog. Biophys. Mol. Biol. 58 85–201. [DOI] [PubMed] [Google Scholar]
  46. Wittung-Stafshede, P. 1999. Effect of redox state on unfolding energetics of heme proteins. Biochim. Biophys. Acta 1432 401–405. [DOI] [PubMed] [Google Scholar]
  47. Wittung-Stafshede, P., Malmstrom, B.G., Sanders, D., Fee, J.A., Winkler, J.R., and Gray, H.B. 1998. Effect of redox state on the folding free energy of a thermostable electron-transfer metalloprotein: The CuA domain of cytochrome oxidase from Thermus thermophilus. Biochemistry 37 3172–3177. [DOI] [PubMed] [Google Scholar]
  48. Wittung-Stafshede, P., Gomes, C.M., and Teixeira, M. 2000. Stability and folding of the ferredoxin from the hyperthermophilic archaeon Acidianus ambivalens. J. Inorg. Biochem. 78 35–41. [DOI] [PubMed] [Google Scholar]
  49. Woese, C.R., Kandler, O., and Wheelis, M.L. 1990. Towards a natural system of organisms: Proposal for the domains Archaea, Bacteria, and Eucarya. Proc. Natl. Acad. Sci. 87 4576–4579. [DOI] [PMC free article] [PubMed] [Google Scholar]
  50. Xiao, L. and Honig, B. 1999. Electrostatic contributions to the stability of hyperthermophilic proteins. J. Mol. Biol. 289 1435–1444. [DOI] [PubMed] [Google Scholar]

Articles from Protein Science : A Publication of the Protein Society are provided here courtesy of The Protein Society

RESOURCES