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
. 2003 Jul;12(7):1483–1495. doi: 10.1110/ps.0302303

Solution structure and thermal stability of ribosomal protein L30e from hyperthermophilic archaeon Thermococcus celer

Kam-Bo Wong 1,2, Chi-Fung Lee 2, Siu-Hong Chan 1, Tak-Yuen Leung 1, Yu Wai Chen 3, Mark Bycroft 3
PMCID: PMC2323938  PMID: 12824494

Abstract

To understand the structural basis of thermostability, we have determined the solution structure of a thermophilic ribosomal protein L30e from Thermococcus celer by NMR spectroscopy. The conformational stability of T. celer L30e was measured by guanidine and thermal-induced denaturation, and compared with that obtained for yeast L30e, a mesophilic homolog. The melting temperature of T. celer L30e was 94°C, whereas the yeast protein denatured irreversibly at temperatures >45°C. The two homologous proteins also differ greatly in their stability at 25°C: the free energy of unfolding was 45 kJ/mole for T. celer L30e and 14 kJ/mole for the yeast homolog. The solution structure of T. celer L30e was compared with that of the yeast homolog. Although the two homologous proteins do not differ significantly in their number of hydrogen bonds and the amount of solvent accessible surface area buried with folding, the thermophilic T. celer L30e was found to have more long-range ion pairs, more proline residues in loops, and better helix capping residues in helix-1 and helix-4. A K9A variant of T. celer L30e was created by site-directed mutagenesis to examine the role of electrostatic interactions on protein stability. Although the melting temperatures of the K9A variant is ∼8°C lower than that of the wild-type L30e, their difference in Tm is narrowed to ∼4.2°C at 0.5 M NaCl. This salt-dependency of melting temperatures strongly suggests that electrostatic interactions contribute to the thermostability of T. celer L30e.

Keywords: NMR, helix capping, ribosome, RNA binding, protein structure

Proteins from hyperthermophilic organisms have to function at extreme temperatures (>80°C). Although some intracellular factors are reported to stabilize protein in vivo (Santos and da Costa 2001), most thermophilic proteins are intrinsically more stable at high temperatures. It is not only of academic curiosity to understand how these thermophilic proteins remain stable and active at elevated temperatures, the "rules" learned can also have potential applications in rational design of thermostable industrial enzymes (Bruins et al. 2001).

How thermophilic proteins achieve their extraordinary stability at elevated temperatures is still not fully understood. Sequence–structure comparison of homologous proteins from thermophilic and mesophilic origins provides insights on the structural basis of thermostability of proteins. A number of factors, for example, increased number of hydrogen bonds and salt bridges, better packing of hydrophobic core, stabilization of secondary structure, have been proposed (Vogt and Argos 1997; Vogt et al. 1997; Kumar et al. 2000; Szilagyi and Zavodszky 2000). It appears that different proteins may use different combinations of structural features to achieve thermostability (Petsko 2001). The only common trend for all thermophilic proteins is the increase in the number of ion pairs (Szilagyi and Zavodszky 2000; Petsko 2001). However, in a number of these comparative studies, the structural features that may contribute to thermostability were correlated with the optimal growth temperatures of the organism or with the thermal inactivation of enzymatic activities. These approaches are limited by the fact that thermal inactivation of enzymes is often complicated by secondary irreversible processes such as covalent modification. Moreover, although proteins from thermophilic organism are in general more thermostable than their mesophilic homologs (Kumar et al. 2001), the living temperature of the source organism is not a direct measurement of a protein's thermostability. To understand the intricate balance of noncovalent interactions that contribute to stability of thermophilic proteins, it is better to use the thermodynamics parameters, such as melting temperature or free energy of unfolding, in the sequence–structure comparison (for example, see Ladenstein and Antranikian 1998; Jaenicke and Bohm 2001). However, thermodynamic parameters for thermophilic proteins have been difficult to obtain, as many thermophilic proteins are large and oligomeric and they often denature irreversibly at high temperatures.

Ribosomal protein L30e from Thermococcus celer, a hyperthermophilic archaeon that grows optimally at 85°C, is a good model for the study of thermostability. It is a small (100 residues) monomeric protein without any cofactors or disulfide bonds. Both its guanidine and thermal-induced denaturation are reversible. T. celer L30e is extremely thermostable; we have shown in this study that its melting temperature is 94°C, which is among the most stable monomeric proteins reported. Ribosomal protein L30e, a component of the large subunit of ribosome, is highly conserved in eukaryotic and archaeal genomes. To our knowledge, the only L30e structure reported is that from yeast (Mao and Williamson 1999; Mao et al. 1999). In yeast, the L30e protein (formerly known as L32) can bind to its own mRNA and inhibit its splicing and translation (Eng and Warner 1991; Li et al. 1996; Vilardell and Warner 1997). Here, we report the solution structure of L30e from T. celer and demonstrate that the two homologous L30e proteins from T. celer and yeast differ greatly in their conformational stability. Determinants of the thermostability of T. celer L30e were discussed based on structural comparison of the thermophilic and mesophilic proteins.

Results

Ribosomal protein L30e from T. celer is highly thermostable

Thermal and guanidine-induced denaturations of the thermophilic and mesophilic L30e proteins were monitored by circular dichroism (CD) at 222 nm and were analyzed using a two-state model (see Fig. 1▶ legend for all fitting parameters). T. celer L30e is a highly thermostable protein; its unfolding transition started at ∼90°C and the melting temperature was 93.5° ± 0.3°C (Fig. 1A▶). The CD spectra acquired before and after heating were identical, suggesting the thermal denaturation of T. celer L30e was reversible. In contrast, the yeast L30e unfolded irreversibly and precipitated at temperatures >45°C. The irreversibility of thermal denaturation of yeast L30e was also observed in a previous study by Williamson and coworkers (Mao and Williamson 1999). Conformational stability of T. celer and yeast L30e proteins was compared by guanidine-induced denaturation, which was reversible for both homologous proteins. The difference in stability is large. At 25°C, the free energies of unfolding, ΔGu, were 44.9 ± 1.6 kJ/mole for T. celer L30e and 14.3 ± 0.3 kJ/mole for the yeast homolog (Fig. 1B▶).

Figure 1.

Figure 1.

Open in a new tab

T. celer L30e is more stable than the yeast homolog. (A) Thermal denaturation of T. celer L30e was followed by molar residual ellipticity at 222 nm. The melting temperature of T. celer L30e was 93.5° ± 0.3°C and the van't Hoff enthalpy, ΔHm, was 370 ± 30 kJ/mole. Yeast L30e was denatured irreversibly when the protein sample was heated to >45°C (data not shown). (B) (Open circles) Guanidine-induced denaturation of T. celer; (filled circles) yeast L30e proteins was reversible and followed an apparent two-state transition. The midpoint of transition and m-values were 4.45 ± 0.02 M, 10.1 ± 0.3 kJ/(mole M) for T. celer L30e and 1.46 ± 0.01 M, 9.8 ± 0.2 kJ/(mole M) for the yeast L30e. ΔGu were 44.9 ± 1.6 kJ/mole for T. celer and 14.3 ± 0.3 kJ/mole for yeast L30e proteins.

Structure determination

Due to limited dispersion on the 1H dimension for the methyl protons of T. celer L30e, the 13C,13C-HSQC-NOESY-HSQC experiment optimized for nuclear Overhauser effects (NOEs) between methyl groups (Zwahlen et al. 1998) was found to be very useful in obtaining an initial set of unambiguous NOEs (Fig. 2▶). Based on 992 manually assigned NOEs, initial structures were calculated by distance-geometry-simulated-annealing hybrid in the XPLOR program (A.T. Brünger, Yale University). The 25 lowest energy structures were used as template structures for automatic NOE assignment by the ARIA program (Linge et al. 2001). The final set of structural restraints contains 1328 unambiguous and 68 ambiguous NOEs (Table 1▶).

Figure 2.

Figure 2.

Open in a new tab

F1–F3 strip of the three-dimensional 13C,13C methyl NOESY spectrum (Zwahlen et al. 1998) acquired on a 15N,13C sample of T. celer L30e. The fact that methyl-methyl NOEs are well separated on the 13C dimension allows assignment of an initial set of unambiguous NOEs. Diagonal peaks are indicated by arrows and assignments of methyl NOEs are labeled.

Table 1.

Experimental restraints and structural statistics

Structural restraints
    Unambiguous NOEs 1328
        Intra-residue 509
        Sequential (|i − j| = 1) 308
        Medium range (|i − j| = 2–4) 215
        Long range (|i − j| > 4) 296
    Ambiguous NOEs 68
    Hydrogen bonds 48
    Dihedral angles 104
{SA}a ≤SA>rb
r.m.s. deviations from experimental restraints
    Distances (Å) 0.0097 ± 0.0011 0.0163
    Dihedral angles (°) 0.32 ± 0.06 0.54
r.m.s. deviation from idealized geometry
    Bonds (Å) 0.0031 ± 0.0001 0.0151
    Angles (°) 0.41 ± 0.02 0.77
    Impropers (°) 1.10 ± 0.08 0.43
Backbone stereochemistryc
    Most favorable regions 89.4% 95.7%
    Allowed regions 9.2% 4.3%
    Generously allowed regions 0.7% 0%
    Disallowed regions 0.7% 0%
r.m.s. deviations from the average structure (residues 3–96)
    Backbone 0.61 Å
    Heavy atoms 1.17 Å
Open in a new tab

a {SA} represents the ensemble of the 10 structures with the lowest energy.

b≤SA>r represents the restrained minimized average structure of {SA}.

c Quality of the structures was checked by PROCHECK-NMR (Laskowski et al. 1993, 1996).

The overall structure of T. celer L30e is well defined except the disordered amino- and carboxy- terminal residues (Fig. 3A▶). The r.m.s. deviations for the well-ordered region (residue 3–96) are 0.61 Å and 1.17 Å for backbone and heavy atoms, respectively. The structure agrees well with the experimental restraints and covalent geometry (Table 1▶). The quality of the structure was checked by PROCHECK (Laskowski et al. 1993, 1996); 96% and 89% of residues in the average and ensemble structures, respectively, are found in the most favorable regions (Table 1▶).

Figure 3.

Figure 3.

Open in a new tab

Solution structure of thermophilic L30e. (A) Stereo-diagram showing backbone trace of an ensemble of 10 simulated annealing structures with the lowest energy. (B) Ribbon diagram of the restrained minimized structure of T. celer L30e. Secondary structure elements are labeled.

Structure description

T. celer L30e consists of 5 helices (with helix-5 as a short 310 helix) and a mixed β-sheet with four strands. The alternating αβ patterns fold into a three-layer αβα sandwich (Fig. 3B▶). The mixed β-sheet, in which β1, β4, and β2 are antiparallel to each other and strand β3 runs parallel to β2, is sandwiched by two layers of helices, with α1, α4, and α5 on one side and α2 and α3 on the other. This topology is conserved within the L7Ae protein family of the Pfam database (http://www.sanger.ac.uk/Software/Pfam), which includes ribosomal proteins L30e, L7Ae, and a 15.5-kD spliceosomal protein. All of these proteins have been shown to bind a RNA kink-turn (Mao et al. 1999; Vidovic et al. 2000; Klein et al. 2001), and the L7Ae/L30e fold may be conserved for specific binding to this RNA motif.

Structural comparison of thermophilic and mesophilic L30e

Global fold, loops, and secondary structure

The r.m.s. deviation of backbone atoms between the structures of T. celer and the yeast L30e is 2.0 Å. Apart from the amino- and carboxy- terminal regions, the large deviation between the two homologous structures is confined to the loop regions, in particular the β3–α4 loop. In the yeast protein, the first turn of α4 is disordered (Fig. 4A▶) and only forms a stable helix when bound to the RNA substrate (Mao and Williamson 1999). In contrast, helix-4 is well defined in the structure of T. celer L30e, even in the absence of RNA (Fig. 4A▶). The increased stability of helix-4 in T. celer L30e can be explained by the presence of a helix capping interaction, in which the hydroxyl group of Thr-66 forms a hydrogen bond to the backbone amide of Glu-69 (Fig. 4B▶). This capping residue, Thr-66, is highly conserved in thermophilic L30e proteins but is absent in most of the eukaryotic L30e (Fig. 5▶). Another capping residue conserved in thermophilic L30e proteins is Asp-2, which caps helix-1 by forming hydrogen bonds to the backbone amide of Ala-4 and Phe-5.

Figure 4.

Figure 4.

Open in a new tab

Helix-4 of T. celer L30e is structured. (A) Ensembles of NMR structure of helix 4 of T. celer (black) and yeast (gray) L30e. Note that helix-4 is well defined in the structure of T. celer L30e but is more disordered in the yeast homolog. The arrow indicates the first turn of helix-4 of the yeast L30e, which is unstructured. (B) The amino -terminal of helix-4 of T. celer L30e is stabilized by capping. The hydroxyl group of Thr-66 forms a hydrogen bond to the backbone amide of Glu-69 at the amino- terminal of helix-4.

Figure 5.

Figure 5.

Open in a new tab

Sequence alignment of ribosomal protein L30e from T. celer, Pyrococcus horikoshii, Methanococcus jannaschii, yeast, rice, and humans. The optimal growth temperatures for the thermophilic archaea T. celer, P. horikoshii and M. Jannaschii are 85°C, 98°C, and 85°C, respectively. Sequences were aligned using CLUSTAL W (Thompson et al. 1994). Proline and helix capping residues conserved in thermophilic proteins are boxed. Secondary structure elements of T. celer and yeast L30e were shown above and below the sequences. The residues of yeast L30e were numbered according to those reported in the NMR structure (PDB code 1CN7).

Ion pairs

Sequence comparison between T. celer and the yeast L30e revealed that the thermophilic homolog has a dramatic increase in the number of acidic residues. The T. celer protein has 13 acidic (Glu+Asp) and 14 basic residues (Lys+Arg), whereas the yeast homolog has 7 acidic and 15 basic residues. The additional negative charges neutralize the net charge of T. celer L30e from +8 in the case of yeast sequence to +1 (or +2 dependent on the pKa value of the His-78 residue). The extra acidic residues may form more favorable charge–charge interactions in T. celer L30e that contribute to its thermostability. Here, we used the criteria of Szilagyi and Zavodszky (2000) to classify ion pairs using three distant limits of 4 Å, 6 Å, and 8 Å. Although the numbers of salt bridges (ion pairs within 4 Å) do not differ significantly between the two homologous proteins, T. celer L30e clearly has more long-range ion pairs (Table 2▶).

Table 2.

Number of ion pairs and hydrogen bonds in T. celer and yeast L30ea

T. celer L30e
Yeast L30e NMR Crystal
Ion pairsb
    Distance limits
        4 Å 2 ± 1 3 ± 2 4
        6 Å 3 ± 1 7 ± 2 6
        8 Å 6 ± 2 13 ± 2 12
Hydrogen bondsc
    Backbone–backbone 57 ± 4 60 ± 3 63
    Backbone–side chain 10 ± 2 9 ± 2 15
    Side chain–side chain 9 ± 2 4 ± 2 2
    Total 76 ± 4 73 ± 5 80
Open in a new tab

a Number of ion pairs and hydrogen bonds were calculated for the crystal and NMR structures of T. celer L30e and compared to those obtained for yeast L30e. The values of the mean and standard deviation calculated for the NMR ensembles are shown.

b Two oppositely charged residues were considered an ion pair if their charged atoms are closer to each other than a certain distance limits. Numbers of ion pairs were counted using three different limits: 4 Å, 6 Å, and 8 Å.

c Hydrogen bonds were calculated using the program HBPLUS (McDonald and Thornton 1994). Hydrogen bonds were classified into three groups: backbone-to-backbone, backbone-to-side chain, and side chain-to-side chain.

Proline

Proline residues located in loop regions may contribute to protein stability by entropically destabilizing the denatured state (Matthews et al. 1987). T. celer L30e has four proline residues. Pro-43, located at the amino terminus of helix-3, is highly conserved in all L30e proteins. The three other proline residues (Pro-59, Pro-77, Pro-88) are conserved in thermophilic L30e but are absent in the yeast sequence (Fig. 5▶). The extra proline residues at the loop regions of T. celer L30e may contribute to the stability of the thermophilic L30e.

Cavity and accessible surface area

Better packing has been proposed to be one of the factors contributing to thermostability of proteins (Querol et al. 1996; Szilagyi and Zavodszky 2000; Petsko 2001). To determine whether the two homologous proteins differ in their packing, internal cavity was detected by the program VOIDOO (Kleywegt and Jones 1994) with a probe radius of 1.2 Å. No internal cavity was detected in both T. celer and yeast L30e protein.

Hydrophobic interactions, one of the major driving forces for protein folding, can be correlated with the amount of accessible surface area buried with folding (Makhatadze and Privalov 1995; Pace 1995; Janin 1997). The solvent accessible surface area was calculated for the two homologous proteins by the NACCESS program. The area buried with folding was estimated by subtracting the surface area calculated for the folded state from those for the Ala-X-Ala tripeptide, which serves as a model for the unfolded state (Hubbard et al. 1991). The two homologous proteins buried a similar amount of total solvent accessible area (∼9900 Å) with folding (Table 3▶). However, there is a small difference in the relative amount of polar and nonpolar surface buried. T. celer L30e buried slightly less polar atoms and more nonpolar atoms with folding (Table 3▶).

Table 3.

Comparison of burial of accessible area a

Yeast L30e T. celer L30e NMR structure T. celer L30e crystal structure
All-atoms ΔASAb 9971 ± 129 Å3 9917 ± 134 Å3 9907 Å3
Non-polar ΔASAnpl 5953 ± 86 Å3 (60 ± 1%) 6185 ± 99 Å3 (62 ± 1%) 6058 Å3 (61%)
Polar ΔASApol 4019 ± 69 Å3 (40 ± 1%) 3732 ± 96 Å3 (38 ± 1%) 3849 Å3 (39%)
Open in a new tab

a Solvent accessible surface areas were calculated by the NACCESS program (Hubbard and Thornton 1993) for the crystal and NMR structures of T. celer and for the NMR structure of yeast L30e. A probe radius of 1.4 Å was used in all calculations. The values of the mean and standard deviation calculated for the NMR ensembles are shown.

b Buried surface areas (ΔASA) were obtained by subtracting the solvent accessible area calculated for the folded state from those for the tripeptide Ala-X-Ala, which serves as a model for the unfolded polypeptide (Hubbard et al. 1991). Accessible surface areas for atoms N and O are considered polar and all other atoms are nonpolar.

Hydrogen bonds

T. celer and the yeast L30e have a similar total number of hydrogen bonds (Table 2▶). T. celer L30e has slightly more backbone-to-backbone hydrogen bonds due to a more structured helix-4. Otherwise, the backbone-to-backbone hydrogen bonding patterns of the two proteins are similar because they have conserved secondary structure elements. More side chain-to-side chain hydrogen bonds were observed in the yeast L30e, mainly due to hydrogen bonds among buried polar residues. There are four polar residues, Thr-30, Ser-39, Thr-65, and Ser-91, whose side chains are completely buried in yeast L30e. The formation of side chain-to-side chain hydrogen bonds among these residues compensates for the desolvation penalty due to burial of polar groups (Fig. 6▶). In T. celer L30e, the only buried polar residue is Ser-23 (which corresponds to Thr-30 in yeast) whose OH group forms a hydrogen bond to the backbone carbonyl group of Gly-19. Other buried polar residues in the yeast protein (Ser-39, Thr-65, and Ser-91) are replaced by nonpolar one (Ala-32, Ile-58, and Ala-84) in T. celer L30e. The difference in the number of buried polar residues also accounts for the fact that T. celer L30e buries slightly less polar surface with folding (Table 3▶).

Figure 6.

Figure 6.

Open in a new tab

Stereo-diagram showing hydrogen bond networks among buried polar groups of T30, S39, T65, and S91 in yeast L30e.

Structure comparison with the crystal structure of T. celer L30e

In a parallel study, we have also determined the crystal structure of T. celer L30e (Chen et al. 2003) by molecular replacement using the NMR structure as the search model. In the crystal structure, no electron density was observed for the carboxy-terminal residues (97–100), which are also disordered in the NMR structure. The solution and crystal structures of T. celer L30e are essentially identical; the r.m.s. deviation for the backbone atoms of the well-ordered region (residues 3–96) is 0.65 Å. Both structures of T. celer L30e have similar number of ion pair, hydrogen-bonding pattern, and solvent accessible surfaces (Tables 2, 3▶▶). Noteworthy, the hydrogen bonds in the ensemble of NMR structures are more dynamic in nature. Hydrogen bonds found in the crystal structure are also found in at least one of the structures within the NMR ensemble. The slight decrease in the overall number of hydrogen bonds in the NMR structure of L30e is probably a result of thermal motions (the NMR experiments were performed at 310K, whereas the protein crystal was frozen at 100K during diffraction data collection), but not due to any real differences in their hydrogen-bonding pattern.

Salt dependency of the thermostability suggests the role of electrostatic interactions

If electrostatic interactions contribute to the thermostability of T. celer L30e, a higher salt concentration should destabilize the protein by screening the favorable electrostatic interactions. On the other hand, a higher salt concentration, in the case of NaCl, will stabilize the protein by the Hofmeister effect (Record et al. 1998). Thus, the salt dependency of the thermostability of a protein is a summation of these two counteracting effects. To dissect the contribution of these two effects, we have generated a K9A variant of T. celer L30e. The Lys → Ala substitution was designed to remove favorable electrostatic interactions among Lys-9 and its neighboring negatively charged residues (Asp-2, Glu-6, Asp-12, and Glu-90). Assuming the Hofmeister effect contributes similarly to the stability of wild-type L30e and the K9A variant, salt dependency of ΔTm (Tm(WT) − Tm(K9A)) will provide evidence for the role of electrostatic interactions to the thermostability. To this end, we have measured the salt dependency of Tm for wild-type T. celer L30e and the K9A variant (Fig. 7A▶). At 25–75 mM NaCl, the Tm of wild-type L30e was decreased by ∼1°C, suggesting that the protein was destabilized by the screening effect. On the other hand, wild-type T. celer L30e was stabilized at 0.2–0.5 M NaCl, where the Hofmeister effect dominates. In the case of the K9A variant, the Hofmeister effect dominates the salt dependency of thermostability; the Tm increased from at 85.4° ± 0.3°C at 0 M NaCl to 95.8° ± 0.2°C at 0.5 M NaCl.

Figure 7.

Figure 7.

Open in a new tab

Salt-dependency of Tm suggests that electrostatic interactions contribute to the thermostability of T. celer L30e. Melting temperatures (Tm) of wild-type T. celer L30e (open circles) and K9A (filled circle) of T. celer L30e at 0–0.5 M NaCl (in 10 mM sodium phosphate buffer at pH 7.4) were shown in A, suggesting that the K9A variant is less thermostable than the wild-type L30e at all concentrations of NaCl. (B) However, the differences in their Tm values (ΔTm = Tm(WT) − Tm(K9A)) are smaller at high concentrations of NaCl.

The differences between the thermostability of wild-type L30e and the K9A variant (measured by ΔTm) are salt dependent (Fig. 7B▶). The ΔTm decreased from a value of 8.3°C at 0 M NaCl to a plateau value of ∼4.2°C at 0.2–0.5 M NaCl. This salt dependency of ΔTm strongly suggests that the removal of favorable electrostatic interactions (by the Lys → Ala substitution) destabilizes the T. celer L30e.

Discussion

In this study, we use the thermophilic/mesophilic pair of ribosomal protein L30e from T. celer and yeast as a model to understand the thermostability of protein. Although the two proteins are homologous in sequences and in structure, they differ greatly in conformational stability. T. celer L30e has a melting temperature of 94°C whereas the yeast homolog unfolds irreversibly at 45°C. At 25°C, T. celer L30e is a highly stable protein with a ΔGu of 45 kJ/mole whereas the yeast L30e is only marginally stable with a ΔGu of 14 kJ/mole. Because the overall folds of the two proteins are similar, the large difference in stability is not due to major structural changes between the two homologs but due to subtle differences between the two homologous structures. The solution structure of T. celer L30e reported in this study allows a detailed structural comparison with the yeast L30e to identify the structural features that contribute to the thermostability of T. celer L30e.

First, our structural analyses show that T. celer L30e has more long-range ion pair interactions (Table 2▶). This observation is in agreement with previous structural comparisons of thermophilic and mesophilic homologous proteins (Vogt and Argos 1997; Vogt et al. 1997; Szilagyi and Zavodszky 2000). In a recent survey of 25 protein families, Szilagyi and Zavodszky (2000) concluded that the increased in the number of ion pairs is the only common structural feature found in thermophilic proteins. Although more ion pairs are found in most thermophilic proteins, their role in stabilizing proteins has been controversial (Fersht and Serrano 1993; Matthews 1993; Spek et al. 1998; Vetriani et al. 1998; Xiao and Honig 1999; Strop and Mayo 2000; Takano et al. 2000) since the pioneering observation of Perutz and Raidt (1975). It has been argued that solvent-exposed ion pair do not stabilize protein because the energy that is gained by the electrostatic interactions is offset by the desolvation energy and the entropic cost of fixing two charged side chains (Hendsch and Tidor 1994). However, ion pair interaction may be more favorable at high temperatures, because the desolvation penalty is reduced as water solvates charged groups less efficiently due to increased thermal motion (Elcock 1998; de Bakker et al. 1999). Moreover, clusters of ion pairs may be stabilizing due to synergetic effects among multiple ion pair interactions (Pappenberger et al. 1997; Lebbink et al. 1999).

We have modeled the RNA-binding site of T. celer L30e by fitting its structure to the structure of the yeast L30e-RNA complex, and found that extra ion pairs in T. celer L30e are clustered in regions far away from the putative RNA-binding site (Fig. 8A▶). There are 14 residues (Asp-2, Glu-6, Arg-8, Lys-9, Asp-12, Lys-22, Arg-39, Arg-42, Asp-44, Glu-47, Arg-54, Glu-62, Glu-64, and Arg-92) of T. celer L30e whose corresponding positions in yeast L30e are either uncharged or oppositely charged. Most of these extra charged residues form clusters of ion pairs in two regions: (1) helix-1 and helix-5, and (2) between strand-2,3 and helix-3 (Fig. 8B,C). It is likely that electrostatic interactions among these residues are evolved to improve the thermostability of T. celer L30e without affecting its binding of RNA substrate. The role of electrostatic interactions is also suggested by the dependency of Tm on both ionic strength (Fig. 7A▶) and pH. For example, the Tm of T. celer L30e is reduced to 79°C at pH 3.5 (C.F. Lee and K.B. Wong, unpubl.).

Figure 8.

Figure 8.

Open in a new tab

Ion pair clusters in T. celer L30e. (A) Stereo-diagram showing the putative RNA-binding site of T. celer L30e modeled by fitting its structure to the yeast L30e protein complexed with RNA (PDB code 1CN8). The 14 charged residues in T. celer L30e (D2, E6, R8, K9, D12, K22, R39, R42, D44, E47, R54, E62, E64, R92), whose corresponding positions are either uncharged or oppositely charged in the yeast sequence, are highlighted. Note that most of them are far from the RNA-binding site and are clustered in two regions. These extra-charged residues form an extensive ion pair network in T. celer L30e as illustrated in B and C.

To examine whether electrostatic interactions contribute to the thermostability, we have generated a K9A variant of T. celer L30e. Lys-9 is located in helix-1 and may form multiple stabilizing electrostatic interactions in ion pair cluster 1 (Fig. 8B▶). In the absence of NaCl, the Tm of the K9A variant is 8.3°C lower than that of the wild type (Fig. 7▶). Salt dependency of ΔTm values (Tm(WT) − Tm(K9A)) suggests that the destabilization due to the Lys → Ala substitution has two components. First, the K9A substitution removes favorable electrostatic interactions among Lys-9 and its neighboring acidic residues (Fig. 8B▶). This destabilizing effect can be screened by the addition of NaCl, which reduced the ΔTm value from 8.3°C at 0 M NaCl to ∼4.2°C at 0.2–0.5 M NaCl (Fig. 7B▶). Assuming most of the electrostatic interaction is effectively screened at 0.5 M NaCl, we estimate that electrostatic interactions removed by the K9A substitution contribute ∼4.1°C to the Tm of T. celer L30e. Second, the remaining ∼4.2°C difference in Tm, which is salt independent at 0.2–0.5 M NaCl, is probably due to the loss of hydrophobic interactions contributed by the hydrocarbon side chain of Lys-9. In summary, our results confirm the role that electrostatic interactions contribute in the thermostability of T. celer L30e.

Recent theoretical and experimental studies have highlighted the importance of optimization of long-range electrostatic interactions among charged residues on protein thermostability (Grimsley et al. 1999; Loladze et al. 1999; Perl et al. 2000; Spector et al. 2000; Martin et al. 2001; Perl and Schmid 2001). For example, substitution of two surface residues (R3E/L66E) is responsible for the differences in stability of a thermophilic cold shock protein and its mesophilic homologs (Perl et al. 2000). Their results suggest that the stabilizing effect is due to the optimization of surface electrostatic interactions (for example, avoiding repulsive contacts between same charges) but not the formation of specific salt bridges (Perl et al. 2000; Delbruck et al. 2001; Perl and Schmid 2001; Dominy et al. 2002). Moreover, electrostatic calculations, such as finite difference Poisson-Boltzmann procedure (Yang and Honig 1993, 1994) and the Tanford-Kirkwood model (Tanford and Kirkwood 1957; Bashford and Karplus 1991) have successfully predicted and improved stability in a number of proteins by optimization of their surface charges (e.g., ubiquitin [Ibarra-Molero et al. 1999; Loladze et al. 1999], myoglobin [Ramos et al. 1999], ribonuclease T1 and Sa [Grimsley et al. 1999; Shaw et al. 2001], a peripheral subunit-binding domain [Spector et al. 2000], and a cold shock protein [Dominy et al. 2002]).

Consistent with these observations, our structural comparison also supports the role of optimizing surface charges in thermostability of T. celer L30e. The overall charges of the two homologous proteins are very different. At neutral pH, yeast L30e has a net charge of +8 whereas T. celer L30e is near neutral, due to the extra acidic residues in the thermophilic protein. Surface potential calculation reveals that positive charges are prevalent over the whole yeast L30e molecule. It is likely that the destabilizing charge repulsion in yeast L30e is reduced by the extra acidic charged residues that form extensive ion pair network in T. celer L30e (Fig. 8▶). Reducing the net positive charges is a common trend in thermophilic L30e proteins. For example, thermophilic L30e from Pyrococcus horikoshii and Methanococcus jannaschii have net charges of +1 and +3, respectively, whereas mesophilic L30e proteins from human and rice have net charges of +10. Maintaining balanced surface charges is likely a common strategy for the L30e protein family to achieve thermostability.

Another notable structural feature is that T. celer L30e contains more proline residues. Three extra proline residues are found in loops connecting α3–β3 (Pro-59), α4–β4 (Pro-77), and β4–α5 (Pro-88). Due to their restricted N–Cα rotations, it has been proposed that proline residues, especially those in the loop regions, contribute to protein stability by decreasing the configurational entropy of the unfolded state (Matthews et al. 1987). Protein engineering studies on T4 lysozyme suggests that substitution of proline residues can contribute ∼4 kJ/mole to protein stability (Matthews et al. 1987). The three extra proline residues are all highly conserved in thermophilic L30e but are absent in the yeast protein (Fig. 5▶). It is likely that these proline residues play a role in the thermostability of T. celer L30e.

Better helix capping was also observed in T. celer L30e. In particular, the helix-4 of T. celer L30e is capped at the amino -terminus by Thr-66, which is highly conserved in L30e proteins of thermophilic origins. It is interesting to note that helix-4 of the yeast L30e is disordered and only becomes structured when bound to a RNA substrate (Mao and Williamson 1999) whereas helix-4 of T. celer L30e is well defined even in the absence of RNA. Protein engineering studies on barnase and T4 lysozyme (Fersht and Serrano 1993; Matthews 1993) suggest that Thr is among the best amino-capping residues. Another capping residue, Asn, is conserved at this position of helix-4 in some eukaryotic L30e (Fig. 5▶). Thr and Asn require different backbone conformations to form amino-capping hydrogen bonds (Bell et al. 1992; Matthews 1993). Thr is more preferable to Asn in this case because the backbone ξ dihedral angle of Thr-66 is 165°; protein engineering studies showed that when the ξ dihedral angle is 150°–180°, a substitution of Thr → Asn can cost ∼4 kJ/mole to protein stability (Bell et al. 1992; Fersht and Serrano 1993; Matthews 1993). It has been proposed that the flexibility in the region around helix-4 of yeast L30e is involved in its induced-fit binding of RNA (Mao and Williamson 1999). The differences in the amino acid composition of helix-4 may reflect different evolutionary pressure faced by the two homologous proteins; stability is selected for the thermophilic protein, whereas flexibility is selected for in the yeast protein.

Concluding remarks

In summary, we have determined the solution structure of the ribosomal protein L30e from T. celer, a hyperthermophilic archaeon that grows optimally at 85°C. Thermodynamics measurements show that T. celer L30e is an extremely thermostable protein with a melting temperature of 94°C. To our knowledge, it is among the most stable monomeric proteins reported to date. Structural comparison of thermophilic/mesophilic L30e proteins suggests that the two proteins do not differ in their packing, amount of buried accessible surface area, and their numbers of hydrogen bonds. T. celer L30e uses a combination of other structural features, including more long-range ion pairs, proline residues in loop regions, and better helix capping, to achieve thermostability. These structural features are conserved in other thermophilic L30e and they may contribute to a common strategy for the L30e protein family to increase thermostability. Our present studies have also shown that the guanidine and thermal-induced denaturation of T. celer L30e are reversible, which make this thermophilic protein an attractive model to understand how proteins remain stable at temperatures close to the boiling point of water. The structural features (ion pairs, proline residues, helix capping) identified in this study provide a testable hypothesis for subsequent experimental verification. In particular, by measuring the salt dependency of Tm for wild-type T. celer L30e and the K9A variant, we have shown that electrostatic interactions do play a role in the thermostability of the protein. Work is in progress to dissect the contribution of other residues to the thermostability of T. celer L30e by site-directed mutagenesis and thermodynamics measurements.

Materials and methods

Preparation of NMR sample

The sequence encoding the full-length T. celer L30e was cloned into a modified pRSETA vector (Invitrogen) without the polyhistidine tag. The vector was transformed in an Escherichia coli C-41 strain (Miroux and Walker 1996) for overexpression. To produce 15N or 15N/13C labeled sample, the bacterial culture was grown in M9 minimal media with 4 g/L of 13C glucose or 1 g/L of 15NH4 at 37°C until A600 = 0.4, when protein expression was induced by addition of isopropyl-β-d-thiogalactopyranose (IPTG) to a final concentration of 0.5 mM. Cells were harvested after 16 h and lysed by sonication in buffer A (50 mM Tris at pH 7.2, 0.5 mM phenylmethylsulfonyl fluoride, and 5 mM dithiothreitol). The soluble fraction was heated to 80°C for 10 min to denature and precipitate heat labile proteins from E. coli hosts. Supernatant collected by centrifugation was loaded to a heparin affinity column pre-equilibrated with buffer A. The protein was eluted using a linear gradient of 0–1.0 M NaCl in buffer A over a volume of 100 mL. The purified protein was dialyzed against the sample buffer (20 mM sodium acetate, 0.5 M Na2SO4 at pH 5.6) and concentrated to 1 mM for NMR spectroscopy. All NMR sample contains 10% D2O for lock signal and 0.05% sodium azide to prevent microbial growth.

Site-directed mutagenesis

The K9A mutation was introduced to the coding sequence of T. celer L30e by PCR using 5′-GCAATCCATGGTTGATTTT GCTTTCGAACTCCGTGCCGCTCAGGACACC-3′ as forward primer and 5′-TCGCGGATCCTCACTCTTTACCGCCCAACGC3′ as reverse primer. The coding sequence for the K9A variant was cloned into pET3d (Novagen) and the mutation was confirmed by sequencing.

Preparation of protein samples for CD measurements

The vectors containing the coding sequences of wild-type and K9A T. celer L30e were transformed to E. coli BL21 (DE3) pLysS (Novagen) for overexpression. The bacterial culture was grown in M9ZB medium until A600 reached 0.4, when protein expression was induced by the addition of 0.4 mM IPTG. Cells were harvested after 16 h, resuspended in 20 mM sodium acetate buffer at pH 5.4 (buffer B), and lysed by sonication. The lysate was centrifuged at 15,000g for 30 min and the supernatant was loaded to a Hi-Trap SP Sepharose HP column (Amersham Biosciences) pre-equilibrated with buffer B. The protein was then eluted at ∼0.4 M NaCl using a linear gradient of 0.2–0.7 M NaCl in buffer B over a volume of 225 mL. The eluted protein was loaded to a Hi-Trap Heparin HP column (Amersham Biosciences) pre-equilibrated with 0.2 M NaCl in buffer B. The protein was eluted at ∼0.4 M NaCl using a linear gradient of 0.2–0.7 M NaCl in buffer B over a volume of 120 mL. The eluted protein was then concentrated to ∼5 mL and loaded to a Superdex G-75 column HiLoad 26/60 (Amersham Biosciences) gel filtration column pre-equilibrated with 0.2 M Na2SO4 in buffer B. The purified T. celer L30e was eluted at ∼200 mL.

The sequence encoding the full-length yeast L30e was cloned in pET3d. The vector was transformed in an E. coli BL21 (DE3, pLysS) strain (Novagen) for overexpression. The bacterial culture was grown in M9ZB medium until A600 reach 0.4, when protein expression was induced by addition of 0.4 mM IPTG. Cells were harvested after 16 h and lysed by sonication in buffer A. The inclusion bodies were washed with 0.2 M NaCl, 1% deoxycholic acid, 1% NP-40, and then with 1% Triton X-100, and 1 mM EDTA solutions. Washed inclusion bodies were dissolved in 4 M guanidine hydrochloride in 10 mM sodium phosphate buffer at pH 7.4. Refolding of yeast L30e was achieved by stepwise dilution of guanidine hydrochloride to a final concentration of 0.25 M, followed by dialysis against buffer C (20 mM sodium acetate buffer at pH 5.4, 0.3 M NaCl). Refolded yeast L30e were loaded to a heparin affinity column pre-equilibrated with buffer C. The protein was then eluted using a linear gradient of 0.3–1.0 M NaCl in buffer C over a volume of 240 mL. The eluted protein was loaded to a Superdex G-75 HiLoad 26/60 (Amersham Biosciences) gel filtration column pre-equilibrated with 20 mM sodium acetate buffer at pH 5.4 with 0.2 M Na2SO4. Purified yeast L30e was eluted at ∼200 mL.

Guanidine-induced denaturation

Protein samples (20 μM) were equilibrated with 0–7.2 M of guanidine hydrochloride (GdnHCl) in 10 mM sodium phosphate bufferat pH 7.4 at 25°C for 30 min before CD measurements. Concentration of guanidine hydrochloride solution was determined from refractive index measurements (Pace and Scholtz 1997) using a Leica AR200 refractometer. Mean residue ellipticity at 222 nm was measured at 25°C using a 1-mm path length cuvette with a JASCO J810 spectropolarimeter equipped with a Peltier-type temperature control unit. The data were fitted by nonlinear regression to a two-state model using (Santoro and Bolen 1988): yobs = {(yn + mn [D]) + (yu + mu [D]) e−ΔG(D)/RT} / (1 + e−ΔG(D)/RT), where yobs is the observed mean residue ellipticity at 222 nm; yn and mn are the y-intercept and slope of the linear baseline before the transition; yu and mu are the y-intercept and slope of the linear baseline after the transition; R is the gas constant; T is the temperature in Kelvin; [D] is the concentration of GdnHCl; ΔG(D) is the free energy of unfolding at [D]. The free energy of unfolding without denaturant, ΔG(H2O), was obtained by the linear extrapolation model (Santoro and Bolen 1988): ΔG(D) = ΔG(H2O) − m [D]. Average values and standard deviations of ΔG(H2O), midpoint of transition, and m value over three independent experiments were reported.

Thermal denaturation

Thermal denaturation was followed by mean residue ellipticity at 222 nm using a JASCO J810 spectropolarimeter equipped with a Peltier-type temperature control unit. All protein samples were dialyzed in 10 mM sodium phosphate buffer at pH 7.4, and were thoroughly degassed before CD measurements. The samples were heated in a 1-mm path length cuvette from 25°–110°C at a heating rate of 1 K/min. Same results were obtained using heating rates at 0.5 K/min and 2 K/min. The cuvette was securely stoppered to ensure there was no loss in volume of protein solution due to evaporation. The thermal denaturation data were analyzed by a two-state model: K(T) = {yobs − (yn + mn T)} / {(yu + mu T) – yobs}, where K(T) is the equilibrium constant of unfolding at temperature T. K(T) values within the transition zone were used to obtain ΔG values by ΔG = −RT ln K(T). The melting temperature, Tm, was determined as the temperature at which ΔG = 0. The van't Hoff enthalpy, ΔHm, was derived from the slope of the plot lnK(T) versus 1/T. Average values and standard deviations of Tm and ΔHm over six independent experiments were reported.

To measure the salt dependency of Tm, the sample was dialyzed in 10 mM sodium phosphate buffer at pH 7.4 with 0.025, 0.05, 0.075, 0.1, 0.2, 0.3, 0.4, and 0.5 M NaCl. Tm measurements were repeated twice for both wild-type T. celer L30e and the K9A variant under different concentrations of sodium chloride.

NMR spectroscopy

All spectra were acquired at 37°C on Bruker AMX 500 and ARX 600 spectrometer equipped with triple resonance probes and pulse field gradient units. All NMR data were processed on a LINUX workstation using NMRPIPE (Delaglio et al. 1995) and analyzed with NMRVIEW (Merck Research Laboratories, NJ). Backbone assignments were obtained using triple resonance experiments: HNCACB, CBCA(CO)NH, HNCA, HN(CO)CA (Muhandiram and Kay 1994). Side chain assignments were obtained from HCCH-TOCSY (Kay et al. 1993) and 15N-TOCSY-HSQC experiments (Marion et al. 1989). Side chain assignments of aromatic residues were derived from homonuclear TOCSY acquired on a sample dissolved in D2O. Stereospecific assignments for the methyl groups of valine and leucine were obtained using a 10% fractionally labeled sample (Szyperski et al. 1992).

Structure calculation

Distance restraints were obtained from 15N-NOESY-HSQC (Marion et al. 1989), 13C-NOESY-HSQC (Muhandiram et al. 1993), and homonuclear NOESY experiments. A 13C,13C-HSQC-NOESY-HSQC (Zwahlen et al. 1998) was acquired to aid unambiguous assignment of NOEs between methyl groups. Mixing time of all NOESY experiments was set to 120 msec. Dihedral angle restraints were derived from 3JHNHα values from HNHA (Vuister and Bax 1993) experiments. TALOS (Cornilescu et al. 1999) derived dihedral angle restraints were included only if they were in agreement with the HNHA data. Hydrogen bonding restraints were identified by hydrogen–deuterium exchange experiments. Only hydrogen bonds in standard secondary structures (α-helix and β-sheet) were included in the structure calculation.

One hundred fifty initial structures were calculated by distance-geometry-simulated-annealing hybrid protocol implemented in XPLOR using 992 manually assigned NOEs, 48 hydrogen bonding, and 104 dihedral restraints. All structures calculated converged to the same fold. The 25 lowest energy structures were used as starting structures for structure refinement using ARIA (Linge et al. 2001). The frequency window for NOE assignment was ±0.05 ppm for proton and ±0.5 ppm for nitrogen and carbon shifts. The values of ARIA parameters were set as recommended (Linge et al. 2001). NOEs assigned by ARIA were inspected manually. All calculations were performed on a home-built LINUX cluster with 4 × PIII 800MHz CPUs. In the final iteration, 200 structures were calculated. The 10 structures with the lowest energy were subjected to refinement in explicit water with the CSDX/OPLS hybrid force field. No structure had NOE violations >0.5 Å or dihedral angle violations >5 degrees. The coordinates and restraints were deposited to the Protein Data Bank (ID code:1go0 and 1go1), and the NMR chemical shifts were deposited in BioMagBank (accession no. 5485).

Structural comparison of T. celer and yeast L30e

To ensure that any structural differences between the thermophilic and mesophilic L30e are due only to the experimental restraints observed but not to the differences in the refinement protocols (for example, different force fields used in the restrained molecular dynamics), we have recalculated the structure of yeast L30e based on the NMR restraints deposited in the Protein Data Bank (1cn7.mr.Z) using the same refinement protocols (CNS/ARIA) used for T. celer L30e. The 10 structures with the lowest energy were selected for analysis. None of them had NOE violations >0.5 Å or dihedral angle violations >5 degrees. The quality of the models were checked by the program PROCHECK (Laskowski et al. 1993, 1996). Certain percentages of the residues (80.7%, 16.8%, 1.4% and 1.1%) are found in the most favorable, allowed, generously allowed and disallowed regions, respectively. The structures of yeast L30e calculated were essentially identical to those reported by Mao and Williamson (1999), with r.m.s. deviations of 0.66 and 1.3 Å for backbone and heavy atoms, respectively, of the well-defined regions (residues 9–70, 89–100).

Ion pairs and hydrogen bonds

We used the definition of Szilagyi and Zavodszky (2000) to count the number of ion pairs. In brief, two oppositely charged residues were considered an ion pair if their charged atoms are closer to each other than certain distance limits. The number of ion pairs counted using three different distance limits, 4, 6, and 8 Å, were reported. The numbers of hydrogen bonds were determined using the program HBPLUS (McDonald and Thornton 1994).

Accessible surface area

Solvent accessible surface area (ASA) was calculated by the program NACCESS (Hubbard and Thornton 1993) using a probe radius of 1.4 Å. Surfaces of N and O atoms were considered polar, whereas the surfaces of other atoms were nonpolar. Surface area buried with folding (ΔASA) were calculated by: ΔASA = ASA(unfolding state) – ASA(folded state). ASA(unfolding state) was derived from the ASA calculated for the tripeptide, Ala-X-Ala (Hubbard et al. 1991), which serves as a model for the unfolded polypeptide. Similar results were obtained when extended polypeptide chains were used as a model for the unfolded state.

Modeling of RNA binding of T. celer L30e

The template used for modeling was the yeast L30e : pre-mRNA complex structure (PDB code 1cn8). The mean atomic model of T. celer L30e was superimposed on and replaced yeast L30e in the complex. The binary model was then subjected to three runs (each 200 cycles) of energy minimization performed with the CNS program (Brünger et al. 1998). The first run used large harmonic restraints (harmonic restraint constant, kharm = 10) imposed on protein atoms that are not involved in RNA binding. During the second run, atoms in loops close to the RNA-binding site were also allowed to move unrestrained, whereas medium harmonic restraints (kharm = 4) were applied to the remaining non-RNA-binding atoms. The last run used weak harmonic restraints (kharm = 2) on all atoms in the model. The final complex model has an r.m.s. deviation in bond lengths of 0.0022 Å and an r.m.s. deviation in bond angles of 0.56 degrees. This model is available upon request from the authors.

Acknowledgments

The work described in this paper was supported by grants from the Research Grants Council of the Hong Kong Special Administrative Region, China (Project No. CUHK4243/00M, CUHK4254/02M) and a grant from Research Committee of the Chinese University of Hong Kong (Project No. 2030253). The sabbatical visit of K.-B.W. in Cambridge was supported by a summer research grant from the Science Faculty of the Chinese University of Hong Kong. We thank Prof. Alan Fersht for his generous support during the visit of K.-B.W. in his laboratory.

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

  • L30e, ribosomal protein L30e

  • NMR, nuclear magnetic resonance

  • NOE, nuclear Overhauser effect

  • HSQC, heteronuclear single quantum correlation

  • CD, circular dichroism

  • r.m.s., root mean square

  • PDB, Protein Data Bank

  • ASA, solvent accessible surface area

  • GdnHCl, guanidine hydrochloride

  • Tm, melting temperature

  • IPTG, isopropyl-β-D-thiogalactopyranose

  • PCR, polymerase chain reaction

Article and publication are at http://www.proteinscience.org/cgi/doi/10.1110/ps.0302303.

References

  1. Bashford, D. and Karplus, M. 1991. Multiple-site titration curves of proteins: An analysis of exact and approximate methods for their calculation. J. Phys. Chem. 95 9556–9561. [Google Scholar]
  2. Bell, J.A., Becktel, W.J., Sauer, U., Baase, W.A., and Matthews, B.W. 1992. Dissection of helix capping in T4 lysozyme by structural and thermodynamic analysis of six amino acid substitutions at Thr 59. Biochemistry 31 3590–3596. [DOI] [PubMed] [Google Scholar]
  3. Bruins, M.E., Janssen, A.E.M., and Boom, R.M. 2001. Thermozymes and their application. Appl. Biochem. Biotech. 90 155–186. [DOI] [PubMed] [Google Scholar]
  4. Brünger, A.T., Adams, P.D., Clore, G.M., DeLano, W.L., Gros, P., Grosse-Kunstleve, R.W., Jiang, J.S., Kuszewski, J., Nilges, M., Pannu, N.S., et al. 1998. Crystallography & NMR system: A new software suite for macromolecular structure determination. Acta Cryst. D54 905–921. [DOI] [PubMed] [Google Scholar]
  5. Chen, Y.W., Bycroft, M., and Wong, K.B. 2003. Crystal structure of ribosomal protein L30e from the extreme thermophile Thermococcus celer: thermal stability and RNA binding. Biochemistry 42 2857–2865. [DOI] [PubMed] [Google Scholar]
  6. Cornilescu, G., Delaglio, F., and Bax, A. 1999. Protein backbone angle restraints from searching a database for chemical shift and sequence homology. J. Biomol. NMR 13 289–302. [DOI] [PubMed] [Google Scholar]
  7. de Bakker, P.I., Hunenberger, P.H., and McCammon, J.A. 1999. Molecular dynamics simulations of the hyperthermophilic protein sac7d from Sulfolobus acidocaldarius: Contribution of salt bridges to thermostability. J. Mol. Biol. 285 1811–1830. [DOI] [PubMed] [Google Scholar]
  8. Delaglio, F., Grzesiek, S., Vuister, G.W., Zhu, G., Pfeifer, J., and Bax, A. 1995. NMRPipe: A multidimensional spectral processing system based on UNIX pipes. J. Biomol. NMR 6 277–293. [DOI] [PubMed] [Google Scholar]
  9. Delbruck, H., Mueller, U., Perl, D., Schmid, F.X., and Heinemann, U. 2001. Crystal structures of mutant forms of the Bacillus caldolyticus cold shock protein differing in thermal stability. J. Mol. Biol. 313 359–369. [DOI] [PubMed] [Google Scholar]
  10. Dominy, B.N., Perl, D., Schmid, F.X., and Brooks 3rd, C.L. 2002. The effects of ionic strength on protein stability: The cold shock protein family. J. Mol. Biol. 319 541–554. [DOI] [PubMed] [Google Scholar]
  11. Elcock, A.H. 1998. The stability of salt bridges at high temperatures: Implications for hyperthermophilic proteins. J. Mol. Biol. 284 489–502. [DOI] [PubMed] [Google Scholar]
  12. Eng, F.J. and Warner, J.R. 1991. Structural basis for the regulation of splicing of a yeast messenger RNA. Cell 65 797–804. [DOI] [PubMed] [Google Scholar]
  13. Fersht, A.R. and Serrano, L. 1993. Principles of protein stability derived from protein engineering experiments. Curr. Opin. Struct. Biol. 3 75–83. [Google Scholar]
  14. Grimsley, G.R., Shaw, K.L., Fee, L.R., Alston, R.W., Huyghues-Despointes, B.M., Thurlkill, R.L., Scholtz, J.M., and Pace, C.N. 1999. Increasing protein stability by altering long-range coulombic interactions. Protein Sci. 8 1843–1849. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Hendsch, Z.S. and Tidor, B. 1994. Do salt bridges stabilize proteins? A continuum electrostatic analysis. Protein Sci. 3 211–226. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Hubbard, S.J. and Thornton, J.M. 1993. NACCESS. Department of Biochemistry and Molecular Biology, University College, London.
  17. Hubbard, S.J., Campbell, S.F., and Thornton, J.M. 1991. Molecular recognition. Conformational analysis of limited proteolytic sites and serine proteinase protein inhibitors. J. Mol. Biol. 220 507–530. [DOI] [PubMed] [Google Scholar]
  18. Ibarra-Molero, B., Loladze, V.V., Makhatadze, G.I., and Sanchez-Ruiz, J.M. 1999. Thermal versus guanidine-induced unfolding of ubiquitin. An analysis in terms of the contributions from charge–charge interactions to protein stability. Biochemistry 38 8138–8149. [DOI] [PubMed] [Google Scholar]
  19. Jaenicke, R. and Bohm, G. 2001. Thermostability of proteins from Thermotoga maritima. Methods Enzymol. 334 438–469. [DOI] [PubMed] [Google Scholar]
  20. Janin, J. 1997. Angstroms and calories. Structure 5 473–479. [DOI] [PubMed] [Google Scholar]
  21. Kay, L.E., Xu, G.Y., Singer, A.U., Muhandiram, D.R., and Forman-Kay, J.D. 1993. A gradient-enhanced HCCH-TOCSY experiment for recording side-chain H-1 and C-13 correlations in H2O samples of proteins. J. Magn. Reson. Ser. B 101 333–337. [Google Scholar]
  22. Klein, D.J., Schmeing, T.M., Moore, P.B., and Steitz, T.A. 2001. The kink-turn: A new RNA secondary structure motif. EMBO J. 20 4214–4221. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Kleywegt, G.J. and Jones, T.A. 1994. Detection, delineation, measurement and display of cavities in macromolecular structures. Acta Cryst. D 50 178–185. [DOI] [PubMed] [Google Scholar]
  24. Kumar, S., Tsai, C.J., and Nussinov, R. 2000. Factors enhancing protein thermostability. Protein Eng. 13 179–191. [DOI] [PubMed] [Google Scholar]
  25. ———. 2001. Thermodynamic differences among homologous thermophilic and mesophilic proteins. Biochemistry 40 14152–14165. [DOI] [PubMed] [Google Scholar]
  26. Ladenstein, R. and Antranikian, G. 1998. Proteins from hyperthermophiles: Stability and enzymatic catalysis close to the boiling point of water. Adv. Biochem. Eng. Biotechnol. 61 37–85. [DOI] [PubMed] [Google Scholar]
  27. Laskowski, R.A., MacArthur, M.W., Moss, D.S., and Thornton, J.M. 1993. PROCHECK: A program to check the stereochemical quality of protein structures. J. Appl. Crystallog. 26 283–291. [Google Scholar]
  28. Laskowski, R.A., Rullmannn, J.A., MacArthur, M.W., Kaptein, R., and Thornton, J.M. 1996. AQUA and PROCHECK-NMR: Programs for checking the quality of protein structures solved by NMR. J. Biomol. NMR 8 477–486. [DOI] [PubMed] [Google Scholar]
  29. Lebbink, J.H., Knapp, S., van der Oost, J., Rice, D., Ladenstein, R., and de Vos, W.M. 1999. Engineering activity and stability of Thermotoga maritima glutamate dehydrogenase. II: Construction of a 16-residue ion-pair network at the subunit interface. J. Mol. Biol. 289 357–369. [DOI] [PubMed] [Google Scholar]
  30. Li, B., Vilardell, J., and Warner, J.R. 1996. An RNA structure involved in feedback regulation of splicing and of translation is critical for biological fitness. Proc. Natl. Acad. Sci. 93 1596–1600. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Linge, J.P., O'Donoghue, S.I., and Nilges, M. 2001. Automated assignment of ambiguous nuclear overhauser effects with ARIA. Methods Enzymol. 339 71–90. [DOI] [PubMed] [Google Scholar]
  32. Loladze, V.V., Ibarra-Molero, B., Sanchez-Ruiz, J.M., and Makhatadze, G.I. 1999. Engineering a thermostable protein via optimisation of charge–charge interactions on the protein surface. Biochemistry 38 16419–16423. [DOI] [PubMed] [Google Scholar]
  33. Makhatadze, G.I. and Privalov, P.L. 1995. Energetics of protein structure. Adv. Protein Chem. 47 307–425. [DOI] [PubMed] [Google Scholar]
  34. Mao, H. and Williamson, J.R. 1999. Local folding coupled to RNA binding in the yeast ribosomal protein L30. J. Mol. Biol. 292 345–359. [DOI] [PubMed] [Google Scholar]
  35. Mao, H., White, S.A., and Williamson, J.R. 1999. A novel loop–loop recognition motif in the yeast ribosomal protein L30 autoregulatory RNA complex. Nat. Struct. Biol. 6 1139–1147. [DOI] [PubMed] [Google Scholar]
  36. Marion, D., Driscoll, P.C., Kay, L.E., Wingfield, P.T., Bax, A., Gronenborn, A.M., and Clore, G.M. 1989. Overcoming the overlap problem in the assignment of 1H NMR spectra of larger proteins by use of three-dimensional heteronuclear 1H-15N Hartmann-Hahn-multiple quantum coherence and nuclear Overhauser-multiple quantum coherence spectroscopy: application to interleukin 1 β. Biochemistry 28 6150–6156. [DOI] [PubMed] [Google Scholar]
  37. Martin, A., Sieber, V., and Schmid, F.X. 2001. In-vitro selection of highly stabilized protein variants with optimized surface. J. Mol. Biol. 309 717–726. [DOI] [PubMed] [Google Scholar]
  38. Matthews, B.W. 1993. Structural and genetic analysis of protein stability. Annu. Rev. Biochem. 62 139–160. [DOI] [PubMed] [Google Scholar]
  39. Matthews, B.W., Nicholson, H., and Becktel, W.J. 1987. Enhanced protein thermostability from site-directed mutations that decrease the entropy of unfolding. Proc. Natl. Acad. Sci. 84 6663–6667. [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. McDonald, I.K. and Thornton, J.M. 1994. Satisfying hydrogen bonding potential in proteins. J. Mol. Biol. 238 777–793. [DOI] [PubMed] [Google Scholar]
  41. Miroux, B. and Walker, J.E. 1996. Over-production of proteins in Escherichia coli: Mutant hosts that allow synthesis of some membrane proteins and globular proteins at high levels. J. Mol. Biol. 260 289–298. [DOI] [PubMed] [Google Scholar]
  42. Muhandiram, D.R. and Kay, L.E. 1994. Gradient-enhanced triple-resonance 3-dimensional NMR experiments with improved sensitivity. J. Magn. Reson. Ser. B 103 203–216. [Google Scholar]
  43. Muhandiram, D.R., Farrow, N.A., Xu, G.Y., Smallcombe, S.H., and Kay, L.E. 1993. A gradient C-13 NOESY-HSQC experiment for recording NOESY spectra of C-13-labeled proteins dissolved in H2O. J. Magn. Reson. Ser. B 102 317–321. [Google Scholar]
  44. Pace, C.N. 1995. Evaluating contribution of hydrogen bonding and hydrophobic bonding to protein folding. Methods Enzymol. 259 538–554. [DOI] [PubMed] [Google Scholar]
  45. Pace, C.N. and Scholtz, J.M. 1997. Measuring the conformational stability of a protein. In Protein structure 2nd ed. (ed. T.E. Creighton), pp. 299–321. Oxford University Press, Oxford, UK.
  46. Pappenberger, G., Schurig, H., and Jaenicke, R. 1997. Disruption of an ionic network leads to accelerated thermal denaturation of D-glyceraldehyde-3-phosphate dehydrogenase from the hyperthermophilic bacterium Thermotoga maritima. J. Mol. Biol. 274 676–683. [DOI] [PubMed] [Google Scholar]
  47. Perl, D. and Schmid, F.X. 2001. Electrostatic stabilization of a thermophilic cold shock protein. J. Mol. Biol. 313 343–357. [DOI] [PubMed] [Google Scholar]
  48. Perl, D., Mueller, U., Heinemann, U., and Schmid, F.X. 2000. Two exposed amino acid residues confer thermostability on a cold shock protein. Nat. Struct. Biol. 7 380–383. [DOI] [PubMed] [Google Scholar]
  49. Perutz, M.F. and Raidt, H. 1975. Stereochemical basis of heat stability in bacterial ferredoxins and in haemoglobin A2. Nature 255 256–259. [DOI] [PubMed] [Google Scholar]
  50. Petsko, G.A. 2001. Structural basis of thermostability in hyperthermophilic proteins, or "there's more than one way to skin a cat." Methods Enzymol. 334 469–478. [DOI] [PubMed] [Google Scholar]
  51. Querol, E., Perez-Pons, J.A., and Mozo-Villarias, A. 1996. Analysis of protein conformational characteristics related to thermostability. Protein Eng. 9 265–271. [DOI] [PubMed] [Google Scholar]
  52. Ramos, C.H., Kay, M.S., and Baldwin, R.L. 1999. Putative interhelix ion pairs involved in the stability of myoglobin. Biochemistry 38 9783–9790. [DOI] [PubMed] [Google Scholar]
  53. Record, M.T., Zhang, W., and Anderson, C.F. 1998. Analysis of effects of salts and uncharged solutes on protein and nucleic acid equilibria and processes: A practical guide to recognizing and interpreting polyelectrolyte effects, Hofmeister effects and osmotic effects of salts. Adv. Protein Chem. 51 281–353. [DOI] [PubMed] [Google Scholar]
  54. Santoro, M.M. and Bolen, D.W. 1988. Unfolding free energy changes determined by the linear extrapolation method. 1. Unfolding of phenylmethanesulfonyl α-chymotrypsin using different denaturants. Biochemistry 27 8063–8068. [DOI] [PubMed] [Google Scholar]
  55. Santos, H. and da Costa, M.S. 2001. Organic solutes from thermophiles and hyperthermophiles. Methods Enzymol. 334 302–315. [DOI] [PubMed] [Google Scholar]
  56. Shaw, K.L., Grimsley, G.R., Yakovlev, G.I., Makarov, A.A., and Pace, C.N. 2001. The effect of net charge on the solubility, activity, and stability of ribonuclease Sa. Protein Sci. 10 1206–1215. [DOI] [PMC free article] [PubMed] [Google Scholar]
  57. Spector, S., Wang, M., Carp, S.A., Robblee, J., Hendsch, Z.S., Fairman, R., Tidor, B., and Raleigh, D.P. 2000. Rational modification of protein stability by the mutation of charged surface residues. Biochemistry 39 872–879. [DOI] [PubMed] [Google Scholar]
  58. Spek, E.J., Bui, A.H., Lu, M., and Kallenbach, N.R. 1998. Surface salt bridges stabilize the GCN4 leucine zipper. Protein Sci. 7 2431–2437. [DOI] [PMC free article] [PubMed] [Google Scholar]
  59. Strop, P. and Mayo, S.L. 2000. Contribution of surface salt bridges to protein stability. Biochemistry 39 1251–1255. [DOI] [PubMed] [Google Scholar]
  60. Szilagyi, A. and Zavodszky, P. 2000. Structural differences between mesophilic, moderately thermophilic and extremely thermophilic protein subunits: Results of a comprehensive survey. Structure Fold. Des. 8 493–504. [DOI] [PubMed] [Google Scholar]
  61. Szyperski, T., Neri, D., Leiting, B., Otting, G., and Wuthrich, K. 1992. Support of 1H NMR assignments in proteins by biosynthetically directed fractional 13C-labeling. J. Biomol. NMR 2 323–334. [DOI] [PubMed] [Google Scholar]
  62. Takano, K., Tsuchimori, K., Yamagata, Y., and Yutani, K. 2000. Contribution of salt bridges near the surface of a protein to the conformational stability. Biochemistry 39 12375–12381. [DOI] [PubMed] [Google Scholar]
  63. Tanford, C. and Kirkwood, J.G. 1957. Theory of protein titration curves. I. General eqations for impenetrable spheres. J. Am. Chem. Soc. 79 5333–5339. [Google Scholar]
  64. Thompson, J.D., Higgins, D.G., and Gibson, T.J. 1994. CLUSTAL W: Improving the sensitivity of progressive multiple sequence alignment through sequence weighting, position-specific gap penalties and weight matrix choice. Nucleic Acids Res. 22 4673–4680. [DOI] [PMC free article] [PubMed] [Google Scholar]
  65. Vetriani, C., Maeder, D.L., Tolliday, N., Yip, K.S., Stillman, T.J., Britton, K.L., Rice, D.W., Klump, H.H., and Robb, F.T. 1998. Protein thermostability above 100 degrees C: A key role for ionic interactions. Proc. Natl. Acad. Sci. 95 12300–12305. [DOI] [PMC free article] [PubMed] [Google Scholar]
  66. Vidovic, I., Nottrott, S., Hartmuth, K., Luhrmann, R., and Ficner, R. 2000. Crystal structure of the spliceosomal 15.5kD protein bound to a U4 snRNA fragment. Mol. Cell. 6 1331–1342. [DOI] [PubMed] [Google Scholar]
  67. Vilardell, J. and Warner, J.R. 1997. Ribosomal protein L32 of Saccharomyces cerevisiae influences both the splicing of its own transcript and the processing of rRNA. Mol. Cell. Biol. 17 1959–1965. [DOI] [PMC free article] [PubMed] [Google Scholar]
  68. Vogt, G. and Argos, P. 1997. Protein thermal stability: Hydrogen bonds or internal packing? Fold. Des. 2 S40–S46. [DOI] [PubMed] [Google Scholar]
  69. Vogt, G., Woell, S., and Argos, P. 1997. Protein thermal stability, hydrogen bonds, and ion pairs. J. Mol. Biol. 269 631–643. [DOI] [PubMed] [Google Scholar]
  70. Vuister, G.W. and Bax, A. 1993. Quantitative J correlation: A new approach for measuring homonuclear three-bond J(HNHa) coupling constants in 15N-enriched proteins. J. Am. Chem. Soc. 115 7772–7777. [Google Scholar]
  71. Xiao, L. and Honig, B. 1999. Electrostatic contributions to the stability of hyperthermophilic proteins. J. Mol. Biol. 289 1435–1444. [DOI] [PubMed] [Google Scholar]
  72. Yang, A.S. and Honig, B. 1993. On the pH dependence of protein stability. J. Mol. Biol. 231 459–474. [DOI] [PubMed] [Google Scholar]
  73. ———. 1994. Structural origins of pH and ionic strength effects on protein stability. Acid denaturation of sperm whale apomyoglobin. J. Mol. Biol. 237 602–614. [DOI] [PubMed] [Google Scholar]
  74. Zwahlen, C., Gardner, K.H., Sarma, S.P., Horita, D.A., Byrd, R.A., and Kay, L.E. 1998. An NMR experiment for measuring methyl-methyl NOEs in 13C-labeled proteins with high resolution. J. Am. Chem. Soc. 120 7617–7625. [Google Scholar]

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

RESOURCES