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
. 2007 Feb;16(2):227–238. doi: 10.1110/ps.062448907

Thermodynamic stability of a cold-adapted protein, type III antifreeze protein, and energetic contribution of salt bridges

Olga García-Arribas 1, Roberto Mateo 1, Melanie M Tomczak 2, Peter L Davies 2, Mauricio G Mateu 1
PMCID: PMC2203292  PMID: 17189482

Abstract

A thermodynamic analysis of a cold-adapted protein, type III anti-freeze protein (AFP), was carried out. The results indicate that the folding equilibrium of type III AFP is a reversible, unimolecular, two-state process with no populated intermediates. Compared to most mesophilic proteins whose folding is two-state, the psychrophilic type III AFP has a much lower thermodynamic stability at 25°C, ∼3 kcal/mol, and presents a remarkably downshifted stability–temperature curve, reaching a maximum of 5 kcal/mol around 0°C. Type III AFPs contain few and non-optimally distributed surface charges relative to their mesophilic homologs, the C-terminal domains of sialic acid synthases. We used thermodynamic double mutant cycles to evaluate the energetic role of every surface salt bridge in type III AFP. Two isolated salt bridges provided no contribution to stability, while the Asp36–Arg39 salt bridge, involved in a salt bridge network with the C-terminal carboxylate, had a substantial contribution (∼1 kcal/mol). However, this contribution was more than counteracted by the destabilizing effect of the Asp36 carboxylate itself, whose removal led to a net 30% increase in stability at 25°C. This study suggests that type III AFPs may have evolved for a minimally acceptable stability at the restricted, low temperature range (around 0°C) at which AFPs must function. In addition, it indicates that salt bridge networks are used in nature also for the stability of psychrophilic proteins, and has led to a type III AFP variant of increased stability that could be used for biotechnological purposes.

Keywords: psychrophilic protein, cold adaptation, antifreeze protein, thermodynamic stability, salt bridge

Psychrophilic, or cold-adapted, proteins are widespread in organisms living in cold environments. These proteins constitute excellent models to study molecular adaptations to extreme conditions (Jaenicke 1990; Feller and Gerday 1997, 2003; Jaenicke and Böhm 1998; Smalas et al. 2000; Arnold et al. 2001; D'Amico et al. 2002; Marx et al. 2004), and may have many biotechnological applications (Feller and Gerday 1997; Russell 1998; Cavicchioli et al. 2002; Marx et al. 2004). Most studies on the stability of cold-adapted proteins (Ushakov 1964; Chen and Berns 1978; Cheng and DeVries 1989; Li and Hew 1991; Li et al. 1991; Bentahir et al. 2000; Thomas and Cavicchioli 2000; Collins et al. 2003) have not included a thermodynamic analysis, partly because many of the targeted proteins are large enzymes that unfold irreversibly. However, those studies did reveal a generally high susceptibility to denaturation and, when measured, a low transition temperature (Tm) and a low activation energy for the denaturation process. For a very few psychrophilic enzymes, quantitative thermodynamic analyses have been carried out, and these provided direct evidence for a low thermodynamic stability (Feller et al. 1999; Ásgeirsson et al. 2000; D'Amico et al. 2003b). The causes of the low stability of cold-adapted enzymes, which are being debated, are relevant for a better understanding of protein structure–function relationships and evolution. A long-standing view is that such low stability is a consequence of a presumed physico-chemical compromise between flexibility (for enzymatic activity) and rigidity (for stability). The selective pressure for efficient catalysis at low temperatures would favor a high flexibility, either global or local, and this would necessarily lead to low stability (Hochachka and Somero 1984; Feller and Gerday 1997; Zavodszky et al. 1998; D'Amico et al. 2001, 2002; Collins et al. 2003). An alternative view is that many enzymes and nonenzyme proteins from organisms living in cold environments may have little stability simply because of a relaxation of the selective pressure exerted by heat during the evolution of these proteins from mesophilic ancestors. The engineering by directed evolution of enzymes that are both active at low temperatures (and thus flexible enough for catalysis to occur), but highly stable at the same time, indicate that a high flexibility may not necessarily lead to low stability and has provided support for the latter proposal (Miyazaki et al. 2000; Wintrode et al. 2000; Arnold et al. 2001; Suzuki et al. 2001; D'Amico et al. 2003a).

A special class of cold-adapted proteins are the antifreeze proteins (AFP). Some bacteria, fungi, plants, invertebrates, and vertebrates living in cold environments prevent cell or tissue damage from freezing by producing AFPs. Several structurally unrelated AFP types have been identified and characterized (Davies and Hew 1990; Davies and Sykes 1997; Sönnichsen et al. 1998; Ewart et al. 1999; Fletcher et al. 2001; Jia and Davies 2002). Despite their very different structures, AFPs generally bind to the surface of ice crystal nuclei and prevent crystal growth, thus depressing the freezing point of the body fluids. Because of their generally small size and relatively simple structure AFPs may be ideally suited for providing further insights into the conformational stability of cold-adapted proteins. Because of their lack of enzymatic activity, the study of AFPs may allow us to differentiate between features of cold-adapted proteins that may have arisen as consequence of structural features, such as flexibility, that are needed for catalysis. In addition, AFPs have many potential applications, including their use as additives to improve the quality and shelf-life of frozen food, as cryoprotective agents for organ and cell cryopreservation, as chemical adjuvants to cancer cryosurgery, and in the development of transgenic plants and animals with increased tolerance to cold (Griffith and Ewart 1995; Breton et al. 2000; Kaiser 2002). For many of these applications it will be essential that the AFP remain active during prolonged periods of time. For some uses, such as food processing, any added AFP may also need to withstand heat treatment (pasteurization). Unfortunately, the denaturation temperature of different AFPs appears to be low (for example, see Cheng and DeVries 1989; Li et al. 1991). This may facilitate unwanted degradation or irreversible aggregation, leading to inactivation during storage (in some cases even at very low temperatures; Wang et al. 2002) and/or eventual heat treatment, such as pasteurization. Engineered AFPs with increased stability may be more desirable than natural AFPs for some industrial and biomedical purposes.

Type III AFPs are very small (∼65 amino acids long) proteins that are seasonally found at high concentrations in the blood of some fishes living in polar or high-latitude temperate seas (Hew et al. 1984; Fletcher et al. 2001). Type III AFPs show a globular fold that contains short β-strands and exhibit a flat ice-binding patch on their surface (Fig. 1). These proteins possess a remarkable sequence and structural similarity with the C-terminal domain of the enzyme sialic acid synthase (SAS) that occurs in many organisms, from bacteria to man (Baardsnes and Davies 2001; Gunawan et al. 2005; Hamada et al. 2006). Type III AFPs have been extensively studied, both structurally and functionally (for example, see Li et al. 1985, 1991; Ananthanarayanan et al. 1986; Schrag et al. 1987; Cheng and DeVries 1989; Li and Hew 1991; Sönnichsen et al. 1993, 1996; Chao et al. 1993, 1994; Jia et al. 1996; Madura et al. 1996; DeLuca et al. 1998; Yang et al. 1998; Graether et al. 1999; Miura et al. 2001; Antson et al. 2001; Ko et al. 2003; Du et al. 2003; Baardsnes and Davies 2002). However, the potential biotechnological applications of natural type III AFPs may be hampered by their low denaturation temperature (Cheng and DeVries 1989; Li et al. 1991; see above).

Figure 1.

Figure 1.

Open in a new tab

Model of the crystal structure of type III AFP (Jia et al. 1996). The atomic coordinates deposited in the Protein Data Bank (1MSI) and the PyMOL software (DeLano Scientific) were used. The ice-binding surface is colored magenta (top). The charged side chains, the C-terminal carboxylate, and the only Tyr side chain are shown as stick models and labeled. Salt bridges and hydrogen bonds between them are indicated by thin white lines. Glu25 is shown in two possible conformations deduced from the bifurcated electron density obtained for this residue.

In this study, we have carried out a quantitative thermodynamic analysis of the conformational stability of a type III AFP from an Arctic fish, the eel pout Macrozoarces americanus. To our knowledge, no such study had been done for any AFP type or for any other beta-clip domain, one of the smaller protein folds. In addition, we have used thermodynamic double mutant cycles to evaluate the energetic role of salt bridges on type III AFP stability. A refined NMR structure and a very high-resolution (1.25 Å) crystal structure are available for this type III AFP (Sönnichsen et al. 1993, 1996; Jia et al. 1996), and an extremely high-resolution (0.62 Å) crystal structure is available for a different isoform (Ko et al. 2003), which makes this protein an excellent model to analyze structure–stability relationships, including the energetic contribution of every solvent-exposed salt bridge.

Results

Spectroscopic characterization of type III AFP from eel pout and its homolog, the C-terminal domain of human sialic acid synthase

Purified recombinant type III AFP was spectroscopically characterized to determine which probes could be used to analyze its thermodynamic stability. The far-UV CD spectrum (Fig. 2A) resembled those previously obtained for other type III AFPs (Ananthanarayanan et al. 1986; Cheng and DeVries 1989). The intrinsic fluorescence emission spectrum yielded a maximum at 304 nm, corresponding as expected to that of a protein with Tyr but no Trp (Fig. 2B). Denaturation of type III AFP led to a substantial change in its far UV-CD spectrum and to a dramatic increase in the intrinsic fluorescence intensity (Fig. 2). Thus, the equilibrium denaturation of type III AFP could be analyzed by probing both its secondary structure by far UV-CD and its tertiary structure by fluorescence spectroscopy.

Figure 2.

Figure 2.

Open in a new tab

Spectroscopic characterization of type III AFP and C-SAS. (A) Far-UV CD spectra of native type III AFP (continuous line) or C-SAS (dashed line) or 4 M GdmHCl-denatured type III AFP (dotted line) or C-SAS (dashed-dotted line). (B) Fluorescence emission spectra of native type III AFP (continuous line), 4 M GdmHCl-denatured type III AFP (dashed line), or renatured type III AFP (dotted line). The experiments were carried out using 50 μM protein in 25 mM phosphate (pH 7.0) at 25°C.

In order to compare the conformational stability of type III AFP and a known mesophilic homolog, the C-terminal domain of the enzyme SAS from human (Hamada et al. 2006), the segment corresponding to residues 294 to the C terminus (residue 359) of human SAS was subcloned, and the isolated C-SAS domain expressed and purified. The sequence chosen contained the complete region of homology between C-SAS and type III AFPs, and included the SAS residues equivalent to the full-length AFP sequence, except a one- to three-residue N-terminal extension present in many, but not all, AFP variants (Davies and Hew 1990; Baardsnes and Davies 2001). As expected from structural comparisons (Hamada et al. 2006) the far-UV CD spectrum of human C-SAS was similar to that of type III AFP (Fig. 2A). The fluorescence spectrum revealed a maximum at 304 nm (data not shown), again as expected for a protein that contains Tyr but no Trp. Denaturation of C-SAS led to a change in its CD spectrum that followed closely that observed for type III AFP (Fig. 2A). However, the C-SAS fluorescence spectrum showed no significant variation in the presence of denaturant (data not shown), consistent with the only Tyr present (residue 29 in the C-SAS sequence, not equivalent to the Tyr contained in type III AFP at position 63) being already largely exposed to solvent in the native form (Hamada et al. 2006). Thus, the unfolding equilibrium of C-SAS could not be analyzed by intrinsic fluorescence, but it could be followed by far-UV CD.

Thermal denaturation of a type III AFP

Thermal denaturation of type III AFP probed by following the change in Tyr intrinsic fluorescence intensity at 304 nm in equilibrium experiments, using a protein concentration of 5 μM and a pH of 7, revealed a simple, cooperative process (Fig. 3) with a transition temperature (Tm) of 319.8 K (46.7°C) and an enthalpy change of ∼55 kcal/mol at the Tm. Cooling of a previously heated sample restored the initial fluorescence, showing the complete absence of irreversible aggregation, even at 70°C. Rescanning of the same sample yielded exactly the same Tm (Table 1). Use of a 10-fold higher protein concentration (50 μM) did not change the Tm either (Table 1). Aggregation at this concentration was observed, but only above 60°C, when the transition had already been completed. Thus, thermal denaturation of type III AFP occurred at a relatively low temperature and was protein-concentration independent and fully reversible.

Figure 3.

Figure 3.

Open in a new tab

Thermal denaturation equilibrium of type III AFP (5 μM in in 25 mM phosphate at pH 7.0) followed by intrinsic fluorescence. The experimental data (circles) were fitted to a unimolecular two-state transition using Equation 1 (continuous line).

Table 1.

Thermodynamic parameters for folding of type III AFP in different conditions

graphic file with name 227tbl1.jpg

Open in a new tab

Chemical denaturation of a type III AFP

To quantify the unfolding equilibrium of type III AFP, the tertiary environment of the Tyr residue and the secondary structure of this domain were probed in chemical denaturation experiments using fluorescence and far-UV CD spectroscopy, respectively. The fluorescence intensity change at 304 nm with increasing urea or guanidinium chloride (GdmHCl) concentration at neutral pH and 25°C fitted very well a cooperative two-state transition (Fig. 4A), and yielded m and ΔGuH2O values that were not significantly dependent on the type of denaturant or protein concentration (Table 1). Fitting of the data obtained using urea and 50 μM protein (pH = 7, 25°C) to a two-state unimolecular process yielded ΔGuH2O = 2.6 kcal/mol for the free energy of unfolding extrapolated to absence of denaturant; a very similar value was obtained using GdmHCl as denaturant (Table 1). The ellipticity change at 220 nm with increasing denaturant concentration also fitted a cooperative two-state transition very well (Fig. 4B) and yielded a ΔGuH2O value very similar to that obtained in the same conditions but using fluorescence (Table 1). Chemical denaturation, like thermal denaturation, was almost completely reversible, as probed by fluorescence (93%; Fig. 2B) and CD (data not shown) using 50 μM protein. The coincidence of the denaturation curves and the similar values of the thermodynamic parameters obtained using different denaturants and far-UV CD or Tyr fluorescence indicates that the type III AFP folding equilibrium is a unimolecular, two-state process, with no populated, stable intermediates.

Figure 4.

Figure 4.

Open in a new tab

Chemical denaturation equilibrium of type III AFP. The experiments were carried out using 50 μM protein in 25 mM phosphate (pH 7.0) at 25°C. The transition was followed by: (A) fluorescence spectroscopy, measuring the intrinsic fluorescence intensity at 304 nm. (a.u.) arbitrary units. In the experiment shown, urea was used as denaturant; (B) far-UV CD spectroscopy, measuring the ellipticity at 220 nm. In the experiment shown, GdmHCl was used as denaturant. In both cases the experimental data (circles) were fitted to a unimolecular two-state transition using Equation 2 (continuous line).

As the physiological temperature for antifreeze proteins is close to 0°C and not to 25°C, the unfolding equilibrium was also analyzed at 4°C, using GdmHCl as denaturant (Table 1). The ΔGuH2O value obtained was 4.9 kcal/mol. Thus, the thermodynamic stability is ∼2 kcal/mol higher at 4°C than at 25°C, although the absolute value is still remarkably low. The stability curve of type III AFP as a function of temperature was deduced (Pace and Laurents 1989) by fitting to Equation 4, the ΔGu values determined for the transition region of the thermal denaturation curves, around 46°C, and the ΔGuH2O values obtained at 25°C and 4°C in chemical denaturation experiments (Fig. 5). Relative to most mesophilic and thermophilic proteins (Kumar et al. 2002), the temperature of maximum stability for type III AFP is substantially shifted to lower values. Remarkably, the maximum stability occurs around 0°C, within the very restricted temperature range at which this protein must exert its antifreeze function.

Figure 5.

Figure 5.

Open in a new tab

Temperature–stability curve of type III AFP in 25 mM phosphate buffer (pH 7.0). The curve was obtained as described in Materials and Methods, using free energy values obtained at different temperatures in chemical and thermal denaturation equilibrium experiments and fitting them to Equation 4. Fitting using only a few data points for the thermal transition region (to avoid excessive weighing of those data) led to a very similar curve that yielded approximately the same temperature of maximum stability.

Chemical denaturation of a type III AFP homolog, the C-terminal domain of sialic acid synthase

In order to compare the conformational stability of the psychrophilic type III AFP with that of the isolated homologous C-terminal domain of the mesophilic human SAS, the unfolding equilibrium of C-SAS at pH = 7 and 25°C was followed by measuring the ellipticity change at 222 nm with increasing GdmHCl concentration (Fig. 6). A significant fraction of the C-SAS molecules was found to be already denatured in the absence of any denaturant; thus, the thermodynamic parameters could not be reliably determined. However, comparison of the denaturation curves and the GdmHCl concentration at which the protein is essentially denatured showed that this isolated form of C-SAS is, surprisingly, even less stable than type III AFP.

Figure 6.

Figure 6.

Open in a new tab

Chemical denaturation equilibrium of the human C-SAS domain, using GdmHCl. The experiments were carried out using 50 μM protein in 25 mM phosphate (pH 7.0) at 25°C. The transition was followed by far-UV CD spectroscopy, measuring the ellipticity at 222 nm.

Role of ionic interactions in type III AFP stability and analysis of the effect of individual salt bridges using thermodynamic double-mutant cycles

Sequence analysis revealed that C-SAS from different mesophilic organisms, from bacteria to man, contains between 18 and 22 charged groups. The structures of Neisseria C-SAS (Gunawan et al. 2005) and human C-SAS (Hamada et al. 2006) indicated the presence of one or two salt bridges and several medium-range attractive electrostatic interactions, but no like-charges with potentially repulsive effects located closer than 5.5 Å. In contrast, type III AFP domains have a substantially lower average charge density (between 8 and 15 charged groups). Despite this lower charge density, the structures of the type III AFP variant from Macrozoarces americanus used in this work (Jia et al. 1996; Sönnichsen et al. 1996) or of each of the two antifreeze domains in type III AFP from Lycodichthys dearborni (Miura et al. 2001), respectively, revealed four salt bridges or two medium-range interactions, but also several spatially close (<5.5 Å) like-charges (up to five pairs in L. dearborni AFP). This suggested that surface electrostatics is less optimized in type III AFPs compared to C-SAS.

If the global contribution of ionic interactions to type III AFP stability is small, it could be expected that both a decrease in pH, which would cause the protonation of carboxylates, or an increase in ionic strength, which would screen the interactions between charged groups, would not drastically decrease the thermodynamic stability of the protein. The equilibrium denaturation of type III AFP was thus analyzed at pH 2, or at 500 mM NaCl concentration, and the thermodynamic parameters compared to those obtained at neutral pH and no NaCl. A decrease in pH or an increase in ionic strength caused a small or negligible reduction in the stability of the protein, respectively yielding ΔGuH2O values 0.8 or 0.1 kcal/mol lower than that obtained at neutral pH and low ionic strength (Table 1).

In the type III AFP analyzed here, two solvent-exposed isolated salt bridges occur between Arg23 and Glu25 and between Arg47 and Asp58, and a solvent-exposed two-salt-bridge network is formed between Asp36, Arg39, and the terminal carboxylate of the protein (Fig. 1; Table 2). To investigate the individual role on type III AFP stability of each salt bridge between side chains, while excluding the positive or negative effects of other interactions of the charged side chains with neighboring residues, a thermodynamic double-mutant cycle analysis (Carter et al. 1984; Ackers and Smith 1985; Horovitz 1987, 1996) was carried out. Each side chain involved in a salt bridge was individually mutated to remove the charged group, and the two side chains involved in a same salt bridge were also mutated together (Table 2). The resulting six single mutants and three double mutants were purified, and their thermodynamic stability was compared to that of the nonmutated type III AFP in equilibrium unfolding experiments using GdmHCl as denaturant and the Tyr intrinsic fluorescence as a probe (Table 2). As the ionic strength contributed by GdmHCl could, in principle, have some effect on the results (Monera et al. 1994), the double mutant cycles were independently repeated, this time using urea as denaturant (Table 2). The results obtained using urea are summarized next (see Tables 2, 3); the results independently obtained using GdmHCl were very similar (Tables 2, 3) and led to the same conclusions.

Table 2.

Thermodynamic parameters of type III AFP single and double mutants

graphic file with name 227tbl2.jpg

Open in a new tab

Table 3.

Coupling energy between charged side chains in type III AFP, as obtained by thermodynamic double-mutant cycle analysis

graphic file with name 227tbl3.jpg

Open in a new tab

The Arg23–Glu25 salt bridge

Mutation Arg23Ala led to some decrease in stability, of ∼0.5 kcal/mol. Mutation Glu25Ala had no significant effect on AFP stability. Analysis of the AFP carrying the double mutation Arg23Ala/Glu25Ala and application of the thermodynamic double-mutant cycle indicate that the coupling energy between Arg23 and Glu25 side chains (beyond Cβ) is close to zero (Table 3). Thus, the Arg23–Glu25 salt bridge may not contribute per se to AFP stability.

The Arg47–Asp58 salt bridge

Mutation Arg47Ala caused a decrease in stability similar to that of Arg23Ala, of ∼0.5 kcal/mol. Mutation Asp58Ala completely prevented accumulation of type III AFP in the host cell, perhaps because of a folding defect or a substantial reduction in protein stability. Thus, the more conservative, isosteric mutation Asp58Asn was introduced in both the single and the double mutant to carry out the thermodynamic cycle. Mutation Asp58Asn did not prevent AFP expression and had no effect on its stability. Analysis of the double mutant Asp58Asn/Arg47Ala and application of the double-mutant cycle indicate that the coupling energy between the Arg47 and Asp58 groups is not significantly different from zero (Table 3). Thus, the Arg47–Asp58 salt bridge may not contribute to stability either.

The Asp36–Arg39 salt bridge

Mutation Arg39Ala destabilized the protein by ∼0.7 kcal/mol. Interestingly, mutation Asp36Ala was stabilizing by almost 1 kcal/mol, as found using either urea or GdmHCl as denaturant (Table 2). This mutation, both alone and in the double mutant Asp36Ala/Arg39Ala, led to a substantial increase in the fluorescence intensity of the native state, but not of the denatured state (data not shown). This indicates that removal of Asp36 leads to a change in the environment of the Tyr63 side chain in the native state that could be related to the observed increase in protein stability (see Discussion). This structural perturbation on mutation does not preclude the use of the double mutant cycle analysis, provided that the sum of the conformational changes in the single mutants approximates the change in the double mutant, which is consistent with the same change in fluorescence observed in both the single mutant Asp36Ala and the double mutant Asp36Ala/Arg39Ala. Indeed, in crystallographic studies using other proteins it was observed that some mutations may cause local structural shifts, but the majority of those found in the single mutants were also found in the double mutant (Vaughan et al. 2002). Their energetic effects in the double-mutant cycle are therefore cancelled, which highlights the very robust nature of the double-mutant cycle analysis. Analysis of the double mutant Asp36Ala/Arg39Ala and application of the double-mutant cycle yielded a significant coupling energy of 0.7 kcal/mol between Asp36 and Arg39. Use of GdmHCl instead of urea, in an independent experiment, led to a similar value for the coupling energy, 1.1 kcal/mol. Even though error propagation led to some fitting errors in the calculated coupling energy values, the fact that, for the three salt bridges, nearly the same values were reproducibly obtained in completely independent experiments using different denaturants (Table 3) indicates that the differences obtained in the coupling energies are significant.

To summarize the results of the double-mutant cycle analysis, the use of GdmHCl or urea as denaturant revealed a favorable contribution of the positively charged side chains of Arg 23, Arg47, and Arg39 to type III AFP thermodynamic stability, but this contribution is not due to the salt bridges they form with neighboring acidic side chains. The negatively charged carboxylates of Glu25 and Asp58 did not contribute to AFP stability, and the Asp36 carboxylate actually impaired stability. Of the three solvent-exposed salt bridges those three residue pairs formed, two provided no significant contribution to stability, while the Asp36–Arg39 salt bridge, a part of a two-salt-bridge network, had a significant contribution (close to 1 kcal/mol) to type III AFP stabilization.

Discussion

The low thermodynamic stability and downshifted temperature–stability curve of type III AFP

The above results indicate that the folding equilibrium of a type III AFP is a reversible, unimolecular, two-state process that involves only the denatured form and the native form of the polypeptide, with no populated intermediates. Most mesophilic monomeric proteins whose folding equilibrium is two-state have thermodynamic stabilities ranging approximately from 5 kcal/mol to 15 kcal/mol at 25°C (Pace 1990). In addition, it has been predicted that the maximum stability of globular proteins is reached around 20°C, as the hydrophobic effect, a major force in driving protein folding, is strongest around this temperature (Kumar et al. 2002). Indeed, by analyzing experimental thermodynamic data for 26 proteins with a two-state folding equilibrium, it was found that 20 of them are maximally stable around room temperature (20°C ± 8°C), irrespective of the melting temperature of the protein or the living temperature of its source organism (Kumar et al. 2002). The present study shows that, compared to most mesophilic proteins, the psychrophilic type III AFP analyzed has a much lower thermodynamic stability at 25°C, ∼3 kcal/mol, and, remarkably, presents a downshifted stability–temperature curve, reaching a maximum of ∼5 kcal/mol around 0°C. Thus, the stability of this cold-adapted protein reaches the lower end of the normal range for mesophilic proteins, but it only does so at a much lower temperature.

Two alternative possibilities that have been invoked to explain the low stability of cold-adapted enzymes (see the Introduction) could be also considered to rationalize the very low thermodynamic stability we found for type III AFP: The first is that, even though AFPs have no enzymatic activity, their low stability could be a physico-chemical consequence of a high structural flexibility required to achieve efficient binding to ice crystals at freezing temperatures, through conformational adaptation of the binding epitope. However, NMR analyses revealed no evidence of an unusually high global or local flexibility in this AFP (Madura et al. 1996; Sönnichsen et al. 1998). Structural flexibility does not seem a property of the insect AFPs, either (Daley and Sykes 2003). Moreover, the above possibility would not explain the observed downshift in the stability curve, which in fact does not occur in psychrophilic enzymes like α-amylase from an Antarctic bacterium (D'Amico et al. 2003b).

Alternatively, the low stability and downshifted stability curve of type III AFP could be a consequence of a relaxation in the selective pressure for thermostability exerted on this protein during its evolution from a C-SAS-related mesophilic protein. Some observations consistent with this possibility are: (1) Type III AFPs are seasonally expressed (mainly in winter) (Fletcher et al. 1985), are only required to function when the water temperature falls close to the freezing point of blood and other body fluids, and cannot function at higher temperatures, where no ice is formed. Thus, no selective pressure to maintain an acceptable AFP stability at temperatures substantially higher than 0°C may be envisaged. (2) Type III AFP is just as stable at its working temperature (around 0°C) as the least stable mesophilic proteins are at their much higher working temperatures (around 20°C–37°C). (3) Type III AFP concentration in the blood has been observed to decrease in the summer months. A lower stability at temperatures substantially higher than 0°C, where AFP is not needed, would facilitate AFP degradation, which could be advantageous because of a decrease in fluid viscosity. (4) C-SAS, the mesophilic homologs of type III AFPs, appear to have a better distribution of charges than type III AFPs; coulombic interactions do contribute little to type III AFP stability. (5) A single amino acid substitution, caused by a single-nucleotide mutation (a very frequent occurrence during evolution) and located far away from the ice binding site was enough, when introduced in the laboratory, to significantly increase the very low natural stability of this protein at 25°C. This mutation was not observed in natural type III AFPs, which is consistent with a lack of a selective pressure for its fixation. It is tempting to tentatively propose a scenario in which most globular proteins may have evolved under a selective pressure by heat to reach a minimal stability threshold (∼5 kcal/mol) at working temperatures. As type III AFPs (and, perhaps, many other cold-adapted proteins) must function only at very low temperatures, they may be under selective pressure by heat to reach the 5 kcal/mol stability threshold at close to 0°C. However, unlike mesophilic proteins, they would be subjected to no selection pressure to preserve such a minimum stability at higher temperatures.

The thermodynamic comparison of psychrophilic type III AFP and mesophilic C-SAS homologs could have provided some insight on the molecular reasons for the low stability and downshifted stability curve of type III AFP. Unfortunately, rationalization of the surprisingly low stability we found for the isolated form of the human C-SAS domain has now revealed that a comparison between the stability of C-SAS and type III AFP may be not adequate: The recently determined crystal structure of SAS from Neisseria meningitidis (Gunawan et al. 2005) shows that this protein is a domain-swapped homodimer, where each C-SAS domain, and the linker between domains, interacts with the N-terminal domain of the other monomer. Interdomain interactions may contribute to the stability of C-SAS as a part of the complete, dimeric protein. In addition, the last seven residues located at the end of the linker segment between the N- and C-terminal domains, immediately preceding the C-SAS-type III AFP homology region, are involved in hydrophobic interactions that could also contribute to stabilize C-SAS in the full-length protein. These seven additional residues have no equivalent in type III AFPs, but could be structurally considered part of the C-SAS domain itself. If the same structural organization occurs in human SAS, its type III antifreeze-like domain, unlike type III AFP, may be conformationally further stabilized through partial thermodynamic coupling between folding and interdomain and/or intermonomer association.

Electrostatic contributions to the thermodynamic stability of type III AFP

In addition to widely divergent contributions to conformational stability in type III AFPs and C-SAS discussed above, other more subtle contributions based on surface electrostatics may be also proposed. The small decrease in type III AFP stability we observed upon increasing the ionic strength or lowering the pH does suggest that the global contribution of electrostatic interactions is very small. This is also consistent with the preservation of type III AFP activity over a wide pH range (2–11) (Chao et al. 1994). An increase in favorable ionic interactions and improved surface electrostatics has been observed in many thermophilic proteins compared to their mesophilic counterparts, and this has been invoked as the result of an increase in selective pressure for thermostabilization (Xiao and Honig 1999). Along the same lines, an attractive possibility is that the poor distribution of charges in type III AFP and the small contribution of charge–charge interactions to its stability could be a consequence of the proposed relaxation of the selective pressure by heat for stability in the very cold environment endured by fish living in high-latitude seas in winter, where type III AFPs evolved and must function.

The available evidence indicates that a substantial number of isolated surface salt bridges in proteins contribute little to stability (for example, see Sali et al. 1991); in contrast, networks of salt bridges may have a substantial favorable effect on stability, and appear to be the basis of one evolutionary strategy for increased stability in thermophilic proteins (Xiao and Honig 1999). Our thermodynamic double-mutant cycle analysis showed that the two isolated solvent-exposed salt bridges contribute negligibly to type III AFP stability. In contrast, the Asp36–Arg39 salt bridge, a part of a solvent-exposed two-salt-bridge network, contribute ∼1 kcal/mol to type III AFP stability. This provides evidence of the use of salt bridge networks for the stabilization also of psychrophilic proteins.

Generation of a type III AFP of increased stability

In type III AFP, the stabilizing effect of the Asp36–Arg39 salt bridge was more than compensated by the destabilizing effect of Asp36 itself. As a consequence, the Asp36Ala mutant was substantially more stable than the nonmutated AFP. We observed that for both mutants D36A and D36A/R39A the intrinsic fluorescence of Tyr63 in the native state was much higher than for the nonmutated protein and all other mutants. In contrast, in the denatured state the fluorescence of Tyr63 was very similar in every case. This indicates a change in the environment of Tyr63 in the native state when the carboxylate of Asp36 is removed. Inspection of the type III AFP crystal structure suggests that removal of this carboxylate not only eliminates the salt bridge, but also removes a hydrogen bond with the Tyr63 hydroxyl group (Fig. 1). The subsequent local rearrangement could have contributed to the observed increase in stability.

Mutational studies by different groups have shown that changing type III AFP residues that are located away from the ice-binding site, as are the residues we have mutated, has no effect on AFP activity provided the protein's folding is not destabilized. Thus, our generation of a type III AFP mutant with a substantially increased stability, and the stabilizing effect of the 36–39 networked salt bridge by itself, encourage further engineering of type III AFP, including the adequate introduction or extension of salt bridge networks away from the ice-binding site, for the generation of hyperstable variants for biotechnological applications.

Materials and methods

DNA amplification and cloning and site-directed mutagenesis

The DNA segment coding for type III AFP from the eel pout Macrozoarces americanus (variant rQAE m1.1) in plasmid pT7-7f (Chao et al. 1993) was subcloned in expression vector pET21b+ (Novagen), and the recombinant plasmid was amplified in Escherichia coli DH5α and purified using the Wizard Plus SV minipreps kit (Promega). Mutagenesis was carried out by the inverse PCR method using the QuikChange site-directed mutagenesis kit (Stratagene). The DNA segment coding for C-SAS was amplified from a construct containing full-length human SAS cDNA (a gift from Dr. Michael J. Betenbaugh, Department of Chemical Engineering, Johns Hopkins University) using the primers SAS5 (CCGATCATATGTCTGTGGTGGCCAAAGTGA) and SAS3 (GCTTGCTGCAGTTAAGACTTGATTTTTTTGCCATG). This insert was digested with NdeI and PstI and ligated into similarly digested pT7-7f. The parental AFP, mutant AFP, and SAS sequences were confirmed by automated sequencing of the corresponding DNA segments and flanking sequences.

Protein expression and purification

BL21(DE3) E. coli cells were transformed with AFP recombinant plasmids. Parental and mutant AFP proteins were expressed in 4 × 500 mL cultures of E. coli BL21(DE3) grown at 37°C until the optical density at 600 nm reached 0.7–0.8. Expression of the proteins was then induced by addition of IPTG to 1 mM, and the cultures were further incubated at 37°C for 2 h. Cells were harvested by centrifugation, and kept at −20°C until use. The procedure for purification of AFP from inclusion bodies was derived from that previously described (Chao et al. 1993). All steps were carried out at 0°C–4°C. The cell pellets were resuspended in 70 mL buffer A (10 mM Tris-HCl at pH 8.0, 0.1 mM EDTA), and the cells were disrupted by sonication. The suspension was centrifuged at 15,000g for 20 min and the pellet, containing AFP inclusion bodies, was resuspended in 50 mL buffer A and centrifuged at 15,000g for 10 min. The washed pellet was resuspended in 20 mL buffer B (10 mM Tris-HCl at pH 7.5), and the inclusion bodies were dissolved by adding 20 mL of a 8 M urea solution. The resultant solution was centrifuged at 15,000g for 15 min, and the clear supernatant was thoroughly dialyzed against buffer B, and then acidified by further dialysis against buffer C (50 mM sodium acetate at pH 3.7). The solution was centrifuged at 15,000g for 10 min, and the supernatant was applied to a 5-mL SP-Sepharose (Amersham Biosciences) cation-exchange column equilibrated with buffer C. The protein was eluted with a 0–0.5 M NaCl gradient in the same buffer, and fractions containing pure AFP were pooled. Purified AFP was analyzed by SDS-PAGE and Coomassie Blue staining of overloaded gels, and found free of detectable contaminants. For some preparations a further purification step by size-exclusion chromatography in a Superdex 75 (Amersham Biosciences) column was introduced. The protein solution was extensively dialyzed against buffer D (25 mM sodium phosphate at pH 7.0) and filtered through 0.2 μM Millex GV-4 membranes (Millipore). The protein concentration was determined by UV spectrophotometry using an extinction coefficient ɛ280 = 1480 M−1 cm−1, corresponding to the single tyrosine in the AFP molecule.

The C-SAS construct was transformed into E. coli JM83 containing a pGP1-2 plasmid, and cells were incubated on a 4-L scale at 30°C until the optical density at 600 nm reached 0.7–0.8. Protein expression was induced by a 30-min heat shock at 42°C followed by incubation at 37°C for 4 h. Cells were harvested by centrifugation, and kept at −20°C until use. The cell pellets were resuspended in 70 mL of 5 mM Tris-HCl (pH 8.0), 0.1 mM EDTA, 0.1 mM PMSF, and the cells were disrupted by sonication. The suspension was centrifuged at 15,000g for 20 min and loaded on a Sephadex G-75 size-exclusion column equilibrated in 5 mM Tris-HCl (pH 8.0), 0.1 mM EDTA at 1.2 mL/min. Fractions containing C-SAS were pooled and acidifed by the addition of TFA to 0.1% final concentration. This suspension was centrifuged to remove precipitated proteins. The clarified supernatant was applied to a semipreparative C18 HPLC column equilibrated in 65% Buffer A (0.1% TFA), 35% Buffer B (2:1 acetonitrile:isopropanol, 0.1% TFA). The column was washed at 35% Buffer B and then a gradient was run from 35% to 65% Buffer B. C-SAS eluted around 50% Buffer B. Fractions containing C-SAS were lyophilized.

Fluorescence spectroscopy

A Varian Cary Eclipse luminescence spectrophotometer equipped with a computer-interfaced temperature control unit was used. AFP solutions in different buffers containing no or different concentrations of GdmHCl or urea as denaturant were allowed to reach chemical and thermal equilibrium. Samples in a 2-mm pathlength (2 × 10 mm) cell were excited at 280 nm using 5 nm excitation and emission slit widths, and the emission spectrum was registered between 290 and 360 nm at 60 nm/min using acquisition times of 1 sec. Temperature was kept constant at 25°C or 4°C. Temperature scans were performed at 48°C/h, and the intensity of fluorescence emission at 304 nm was registered using acquisition times of 1 sec.

CD spectroscopy

CD measurements were carried out using a Jasco-600 spectropolarimeter equipped with a temperature control unit interfaced to a computer. The recorded far-UV spectra were the average of five scans obtained at a rate of 50 nm/min, a response time of 2 sec, and a bandwidth of 1 nm. Chemical denaturation equilibrium experiments were carried out by measuring the ellipticity at 220 nm of AFP solutions containing different concentrations of GdmHCl, using a 1-mm pathlength cell. Temperature was kept constant at 25°C, and each sample was allowed to reach chemical and thermal equilibrium. Each ellipticity value was obtained by averaging 10 2-sec time measurements.

Equilibrium denaturation data analyses

The data were fitted to unimolecular N (native)–U (denatured) two-state transitions as previously described (Mateu 2002). Thermodynamic parameters for thermal unfolding and baseline values were obtained by direct nonlinear fitting of the experimental ellipticity or fluorescence intensity values I at any temperature T to the equation

graphic file with name 227equ1.jpg

which gave the fitting values for Tm, the transition temperature; ΔHuTm, the enthalpy of unfolding at the Tm; In0 and Iu0, the ellipticity or fluorescence intensity corresponding respectively to the native (n) or denatured (u) states, extrapolated to T = 0; and mn and mu, the linear increase in ellipticity or fluorescence intensity for the native or denatured state, respectively, as a function of T (i.e., the slopes of the baselines preceding or following the transition region). The value of ΔCp was given a fixed value of 1.2 kcal/mol, which was obtained by fitting the experimental data to Equation 4, as described below. The fitting values obtained for Tm and ΔHuTm showed very little variation when different ΔCp values were used, and were very similar to those obtained by using Equations 3 and 4, which assumed no particular value for ΔCp (see below).

Thermodynamic parameters for chemical unfolding followed by fluorescence and baseline values were obtained by direct nonlinear fitting of the experimental ellipticity or fluorescence intensity values I at any denaturant concentration D to the equation

graphic file with name 227equ2.jpg

which gave the fitting values for ΔGuH2O, the free energy of unfolding extrapolated to absence of denaturant; m, the variation in the free energy of unfolding with D; In0 and Iu0, the ellipticity or fluorescence intensity values corresponding respectively to the native (n) or denatured (u) state, extrapolated to D = 0; and mn and mu, the linear increase in ellipticity or fluorescence intensity for the native or denatured state, respectively, as a function of D (i.e., the slopes of the baselines preceding or following the transition region). For the data obtained at 4°C or by CD in chemical denaturation experiments, fitting had to be carried out using the indirect (linear) approach as previously described (Mateu 2002). However, a comparison of the two methods using data obtained for type III AFP in the same experiment yielded values that were not significantly different (ΔGuH2O values 3.35 ± 0.36 and 3.40 ± 0.23, m values −2.27 ± 0.21 and −2.17 ± 0.13, respectively), as previously observed for other proteins (Mateu 2002).

The dependency of the free energy of unfolding ΔGu with temperature was calculated as indicated (Pace and Laurents 1989) by first experimentally determining ΔGu at any T within the thermal transition region, using the experimental fluorescence intensity values I obtained at temperatures T in thermal denaturation experiments, and the equation

graphic file with name 227equ3.jpg

where In0, Iu0, mn, and mu were previously obtained by linear regression analysis of the baselines. Then, the ΔGu values thus obtained, and the ΔGuH2O values obtained at 4°C and 25°C in chemical denaturation experiments, were fitted to the equation

graphic file with name 227equ4.jpg

where ΔGu (T) is the free energy of unfolding at any temperature T, ΔHuTm is the enthalpy of unfolding at the transition temperature Tm, and ΔCp is the heat capacity, which is assumed to be constant over the temperature range of interest. Fitting values for ΔHuTm, Tm, and ΔCp could thus be obtained. All fittings were carried out using the program Kaleidagraph (Abelbeck Software).

Double mutant cycle analysis

For two side chains, X and Y, that interact with each other in a protein P-XY, an assumption that has proved reasonable is that if X or Y is altered by mutation to A without introducing new interactions or steric constraints (e.g., by mutation to Ala), the structure of the protein will not be perturbed or, if perturbed, the perturbations induced by mutating residue X in P-XY or in the single mutant P-XA will be similar, and the perturbations induced by mutating residue Y in P-XY or in the single mutant P-AY will be similar (Carter et al. 1984; Ackers and Smith 1985; Horovitz 1987, 1996; Vaughan et al. 2002). Then, ΔΔGintXY, the coupling energy (the interaction energy) between X and Y can be calculated from the differences in free energy between each single mutant and the nonmutated protein, ΔΔGP-XY→P-XA and ΔΔGP-XY→P-AY, and the difference in free energy between the double mutant P-AA and the nonmutated protein, ΔΔGP-XY→P-AA, as follows (Carter et al. 1984; Ackers and Smith 1985; Horovitz 1987, 1996):

graphic file with name 227equ5.jpg

Acknowledgments

We thank Dr. Michael J. Betenbaugh, Department of Chemical Engineering, Johns Hopkins University for the gift of human SAS cDNA, and Dr. Andy Scotter and Dr. José Luis Neira for comments on the manuscript. This work was supported by grants from the Spanish Ministerio de Ciencia y Tecnología (BIO2003-04445) to M.G.M. and from the Canadian Institutes for Health Research to P.L.D., and by an institutional grant from Fundación Ramón Areces to the Centro de Biología Molecular. M.G.M. is an associate member of the Instituto de Biocomputación y Física de los Sistemas Complejos, Zaragoza, Spain. P.L.D. holds a Canada Research Chair in Protein Engineering.

Footnotes

Reprint requests to: Mauricio G. Mateu, Centro de Biología Molecular “Severo Ochoa” (UAM-CSIC), Universidad Autónoma de Madrid, Cantoblanco, 28049 Madrid, Spain; e-mail: mgarcia@cbm.uam.es; fax: +34-914974799.

Abbreviations: AFP, antifreeze protein; CD, circular dichroism; GdmHCl, guanidinium hydrochloride; Tm, transition temperature; ΔHuTm, enthalpy of unfolding at the transition temperature; ΔGuH2O, free energy of unfolding extrapolated to absence of denaturant; m, variation in the free energy of unfolding with the denaturant concentration; ΔΔGuH2O, the change in ΔGuH2O upon mutation; ΔΔGintXY, the coupling energy between groups X and Y in a protein.

Article published online ahead of print. Article and publication date are at http://www.proteinscience.org/cgi/doi/10.1110/ps.062448907.

References

  1. Ackers, G.K. and Smith, F.R. 1985. Effects of site-specific amino acid modification on protein interactions and biological function. Annu. Rev. Biochem. 54: 597–629. [DOI] [PubMed] [Google Scholar]
  2. Ananthanarayanan, V.S., Slaughter, D., and Hew, C.L. 1986. Antifreeze proteins from the ocean pout, Macrozoarces americanus: Circular dichroism spectral studies on the native and denatured states. Biochim. Biophys. Acta 870: 154–159. [Google Scholar]
  3. Antson, A.A., Smith, D.J., Roper, D.I., Lewis, S., Caves, L.S., Verma, C.S., Buckley, S.L., Lillford, P.J., and Hubbard, R.E. 2001. Understanding the mechanism of ice binding by type III antifreeze proteins. J. Mol. Biol. 305: 875–889. [DOI] [PubMed] [Google Scholar]
  4. Arnold, F.H., Wintrode, P.L., Miyazaki, K., and Gershenson, A. 2001. How enzymes adapt: Lessons from directed evolution. Trends Biochem. Sci. 26: 100–106. [DOI] [PubMed] [Google Scholar]
  5. Ásgeirsson, B., Hauksson, J.B., and Gunnarsson, H. 2000. Dissociation and unfolding of cold-active alkaline phosphatase from atlantic cod in the presence of guanidinium chloride. Eur. J. Biochem. 267: 6403–6412. [DOI] [PubMed] [Google Scholar]
  6. Baardsnes, J. and Davies, P.L. 2001. Sialic acid synthase: The origin of fish type III antifreeze protein? Trends Biochem. Sci. 26: 468–469. [DOI] [PubMed] [Google Scholar]
  7. Baardsnes, J. and Davies, P.L. 2002. Contribution of hydrophobic residues to ice binding by fish type III antifreeze protein. Biochim. Biophys. Acta 1601: 49–54. [DOI] [PubMed] [Google Scholar]
  8. Bentahir, M., Feller, G., Aittaleb, M., Lamotte-Brasseur, J., Himri, T., Chessa, J.-P., and Gerday, C. 2000. Structural, kinetic, and calorimetric characterization of the cold-active phosphoglycerate kinase from the antarctic Pseudomonas sp. TACII18. J. Biol. Chem. 275: 11147–11153. [DOI] [PubMed] [Google Scholar]
  9. Breton, G., Danyluk, J., Ouellet, F., and Sarhan, F. 2000. Biotechnological applications of plant freezing associated proteins. Biotechnol. Annu. Rev. 6: 59–101. [DOI] [PubMed] [Google Scholar]
  10. Carter, P.J., Winter, G., Wilkinson, A.J., and Fersht, A.R. 1984. The use of double mutants to detect structural changes in the active site of the tyrosyl-tRNA synthetase (Bacillus stearotermophilus). Cell 38: 835–840. [DOI] [PubMed] [Google Scholar]
  11. Cavicchioli, R., Siddiqui, K.S., Andrews, D., and Sowers, K.R. 2002. Low-temperature extremophiles and their applications. Curr. Opin. Biotechnol. 13: 253–261. [DOI] [PubMed] [Google Scholar]
  12. Chao, H., Davies, P.L., Sykes, B.D., and Sönnichsen, F.D. 1993. Use of proline mutants to help solve the NMR solution structure of type III antifreeze protein. Protein Sci. 2: 1411–1428. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Chao, H., Sönnichsen, F.D., DeLuca, C.I., Sykes, B.D., and Davies, P.L. 1994. Structure–function relationship in the globular type III antifreeze protein: Identification of a cluster of surface residues required for binding to ice. Protein Sci. 3: 1760–1769. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Chen, C.H. and Berns, D.S. 1978. Comparison of the stability of phycocyanins from thermophilic, mesophilic, psychrophilic and halophilic algae. Biophys. Chem. 8: 203–213. [DOI] [PubMed] [Google Scholar]
  15. Cheng, C.-H.C. and DeVries, A.L. 1989. Structures of antifreeze peptides from the Antarctic eel pout, Austrolycicthys brachycephalus . Biochim. Biophys. Acta 997: 55–64. [DOI] [PubMed] [Google Scholar]
  16. Collins, T., Meuwis, M.-A., Gerday, C., and Feller, G. 2003. Activity, stability and flexibility in glycosidases adapted to extreme thermal environments. J. Mol. Biol. 328: 419–428. [DOI] [PubMed] [Google Scholar]
  17. Daley, M.E. and Sykes, B.D. 2003. The role of side chain conformational flexibility in surface recognition by Tenebrio molitor antifreeze protein. Protein Sci. 12: 1323–1331. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. D'Amico, S., Gerday, C., and Feller, G. 2001. Structural determinants of cold adaptation and stability in a large protein. J. Biol. Chem. 276: 25791–25796. [DOI] [PubMed] [Google Scholar]
  19. D'Amico, S., Claverie, P., Collins, T., Georlette, D., Gratia, E., Hoyoux, A., Meuwis, M.-A., Feller, G., and Gerday, C. 2002. Molecular basis of cold-adaptation. Philos. Trans. R Soc. Lond. B Biol. Sci. 35: 917–925. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. D'Amico, S., Gerday, C., and Feller, G. 2003a. Temperature adaptation of proteins: Engineering mesophilic-like activity and stability in a cold-adapted α-amylase. J. Mol. Biol. 332: 981–988. [DOI] [PubMed] [Google Scholar]
  21. D'Amico, S., Marx, J.-C., Gerday, C., and Feller, G. 2003b. Activity–stability relationships in extremophilic enzymes. J. Biol. Chem. 278: 7891–7896. [DOI] [PubMed] [Google Scholar]
  22. Davies, P.L. and Hew, C.L. 1990. Biochemistry of fish antifreeze proteins. FASEB J. 4: 2460–2468. [DOI] [PubMed] [Google Scholar]
  23. Davies, P.L. and Sykes, B.D. 1997. Antifreeze proteins. Curr. Opin. Struct. Biol. 7: 828–834. [DOI] [PubMed] [Google Scholar]
  24. DeLuca, C.I., Davies, P.L., Ye, Q., and Jia, Z. 1998. The effects of steric mutations on the structure of type III antifreeze protein and its interaction with ice. J. Mol. Biol. 275: 515–525. [DOI] [PubMed] [Google Scholar]
  25. Du, N., Liu, X.Y., and Hew, C.L. 2003. Ice nucleation inhibition: Mechanism of antifreeze by antifreeze protein. J. Biol. Chem. 278: 36000–36004. [DOI] [PubMed] [Google Scholar]
  26. Ewart, K.V., Lin, Q., and Hew, C.L. 1999. Structure, function and evolution of antifreeze proteins. Cell. Mol. Life Sci. 55: 271–283. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Feller, G. and Gerday, C. 1997. Psychrophilic enzymes: Molecular basis of cold-adaptation. Cell. Mol. Life Sci. 53: 830–841. [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Feller, G. and Gerday, C. 2003. Psychrophilic enzymes: Hot topics in cold adaptation. Nat. Rev. Microbiol. 1: 200–208. [DOI] [PubMed] [Google Scholar]
  29. Feller, G., D'Amico, D., and Gerday, C. 1999. Thermodynamic stability of a cold-active α-amylase from the Antarctic bacterium Alteromonas haloplanctis . Biochemistry 38: 4613–4619. [DOI] [PubMed] [Google Scholar]
  30. Fletcher, G.L., Hew, C.L., Li, X.M., Haya, K., and Kao, M.H. 1985. Year-round presence of high levels of plasma antifreeze proteins in a temperate fish, ocean pout (Macrozoarces americanus). Can. J. Zool. 63: 488–493. [Google Scholar]
  31. Fletcher, G.L., Hew, C.L., and Davies, P.L. 2001. Antifreeze proteins of teleost fishes. Annu. Rev. Physiol. 63: 359–390. [DOI] [PubMed] [Google Scholar]
  32. Graether, S.P., DeLuca, C.I., Baardsnes, J., Hill, G.A., Davies, P.L., and Jia, Z. 1999. Quantitative and qualitative analysis of type III antifreeze protein structure and function. J. Biol. Chem. 274: 11842–11847. [DOI] [PubMed] [Google Scholar]
  33. Griffith, M. and Ewart, K.V. 1995. Antifreeze proteins and their potential use in frozen food. Biotechnol. Adv. 13: 375–402. [DOI] [PubMed] [Google Scholar]
  34. Gunawan, J., Simard, D., Gilbert, M., Lovering, A.L., Wakarchuk, W.W., Tanner, M.E., and Strynadka, N.C. 2005. Structural and mechanistic analysis of sialic acid synthase NeuB from Neisseria meningitidis in complex with Mn2+, phosphoenolpyruvate, and N-acetylmannosaminitol. J. Biol. Chem. 280: 3555–3563. [DOI] [PubMed] [Google Scholar]
  35. Hamada, T., Ito, Y., Abe, T., Hayashi, F., Güntert, P., Inoue, M., Kigawa, T., Terada, T., Shirouzu, M., and Yoshida, M., et al. 2006. Solution structure of the antifreeze-like domain of human sialic acid synthase. Protein Sci. 15: 1–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Hew, C.L., Slaughter, D., Joshi, S.B., Fletcher, G.L., and Ananthanarayanan, V.S. 1984. Antifreeze polypeptides from the Newfoundland ocean pout, Macrozoarces americanus: Presence of multiple and compositionally diverse components. J. Comp. Physiol. [B] 155: 81–88. [Google Scholar]
  37. Hochachka, P.W. and Somero, G.N. 1984. Biochemical adaptation. Princeton University Press, Princeton, NJ.
  38. Horovitz, A. 1987. Non-additivity in protein–protein interactions. J. Mol. Biol. 196: 733–735. [DOI] [PubMed] [Google Scholar]
  39. Horovitz, A. 1996. Double-mutant cycles: A powerful tool for analyzing protein structure and function. Fold. Des. 1: R121–R126. [DOI] [PubMed] [Google Scholar]
  40. Jaenicke, R. 1990. Protein structure and function at low temperatures. Philos. Trans. R. Soc. Lond. B Biol. Sci. 326: 551–553. [DOI] [PubMed] [Google Scholar]
  41. Jaenicke, R. and Böhm, G. 1998. The stability of proteins in extreme environments. Curr. Opin. Struct. Biol. 8: 738–748. [DOI] [PubMed] [Google Scholar]
  42. Jia, Z. and Davies, P.L. 2002. Antifreeze proteins: An unusual receptor-ligand interaction. Trends Biochem. Sci. 27: 101–106. [DOI] [PubMed] [Google Scholar]
  43. Jia, Z., DeLuca, C.I., Chao, H., and Davies, P.L. 1996. Structural basis for the binding of a globular antifreeze protein to ice. Nature 384: 285–288. [DOI] [PubMed] [Google Scholar]
  44. Kaiser, J. 2002. New prospects for putting organs on ice. Science 295: 1015. [DOI] [PubMed] [Google Scholar]
  45. Ko, T.P., Robinson, H., Gao, Y.G., Cheng, C.H., De Vries, A.L., and Wang, A.H. 2003. The refined crystal structure of an eel pout type III antifreeze protein RD1 at 0.62-Å resolution reveals structural microheterogeneity of protein and solvation. Biophys. J. 84: 1228–1237. [DOI] [PMC free article] [PubMed] [Google Scholar]
  46. Kumar, S., Tsai, C.J., and Nussinov, R. 2002. Maximal stabilities of reversible two-state proteins. Biochemistry 41: 5359–5374. [DOI] [PubMed] [Google Scholar]
  47. Li, X. and Hew, C.L. 1991. Structure and function of an antifreeze polypeptide from ocean pout, Macrozoarces americanus: Role of glutamic acid residues in protein stability and antifreeze activity by site-directed mutagenesis. Protein Eng. 4: 1003–1008. [DOI] [PubMed] [Google Scholar]
  48. Li, X.-M., Trinh, K.-Y., Hew, C.L., Buettner, B., Baenziger, J., and Davies, P.L. 1985. Structure of an antifreeze polypeptide and its precursor from the ocean pout, Macrozoarces americanus . J. Biol. Chem. 260: 12904–12909. [PubMed] [Google Scholar]
  49. Li, X., Trinh, K.Y., and Hew, C.L. 1991. Expression and characterization of an active and thermally more stable recombinant antifreeze polypeptide from ocean pout, Macrozoarces americanus, in Escherichia coli: Improved expression by the modification of the secondary structure of the mRNA. Protein Eng. 4: 995–1002. [DOI] [PubMed] [Google Scholar]
  50. Madura, J.D., Taylor, M.S., Wierzbicki, A., Harrington, J.P., Sikes, C.S., and Sönnichsen, F. 1996. The dynamics and binding of a type III antifreeze protein in water and on ice. J. Mol. Struct. 388: 65–77. [Google Scholar]
  51. Mateu, M.G. 2002. Conformational stability of dimeric and monomeric forms of the C-terminal domain of human immunodeficiency-1 capsid protein. J. Mol. Biol. 318: 519–531. [DOI] [PubMed] [Google Scholar]
  52. Marx, J.C., Blaise, V., Collins, T., Damico, S., Delille, D., Gratia, E., Hoyoux, A., Huston, A.L., Sonan, G., and Feller, G., et al. 2004. A perspective on cold enzymes: Current knowledge and frequently asked questions. Cell. Mol. Biol. 50: 643–655. [PubMed] [Google Scholar]
  53. Miura, K., Ohgiya, S., Hoshino, T., Nemoto, N., Suetake, T., Miura, A., Spyracopoulos, L., Kondo, H., and Tsuda, S. 2001. NMR analysis of type III antifreeze protein intramolecular dimer. Structural basis for enhanced activity. J. Biol. Chem. 276: 1304–1310. [DOI] [PubMed] [Google Scholar]
  54. Miyazaki, K., Wintrode, P.L., Grayling, R.A., Rubingh, D.N., and Arnold, F.H. 2000. Directed evolution study of temperature adaptation in a psychrophilic enzyme. J. Mol. Biol. 297: 1015–1026. [DOI] [PubMed] [Google Scholar]
  55. Monera, O.D., Kay, C.M., and Hodges, R.S. 1994. Protein denaturation with guanidine hydrochloride or urea provides a different estimate of stability depending on the contributions of electrostatic interactions. Protein Sci. 3: 1984–1991. [DOI] [PMC free article] [PubMed] [Google Scholar]
  56. Pace, C.N. 1990. Conformational stability of globular proteins. Trends Biochem. Sci. 15: 14–17. [DOI] [PubMed] [Google Scholar]
  57. Pace, C.N. and Laurents, D.V. 1989. A new method for determining the heat capacity change for protein folding. Biochemistry 28: 2520–2525. [DOI] [PubMed] [Google Scholar]
  58. Russell, N.J. 1998. Molecular adaptations in psychrophilic bacteria: Potential for biotechnological applications. Adv. Biochem. Eng. Biotechnol. 61: 1–21. [DOI] [PubMed] [Google Scholar]
  59. Sali, D., Bycroft, M., and Fersht, A.R. 1991. Surface electrostatic interactions contribute little to stability of barnase. J. Mol. Biol. 220: 779–788. [DOI] [PubMed] [Google Scholar]
  60. Schrag, J.D., Cheng, C.-H.C., Panico, M., Morris, H.R., and DeVries, A.L. 1987. Primary and secondary structure of antifreeze peptides from Arctic and Antarctic zoarcid fishes. Biochim. Biophys. Acta 915: 357–370. [DOI] [PubMed] [Google Scholar]
  61. Smalas, A.O., Leiros, H.K., Os, V., and Willassen, N.P. 2000. Cold adapted enzymes. Biotechnol. Annu. Rev. 6: 1–57. [DOI] [PubMed] [Google Scholar]
  62. Sönnichsen, F.D., Sykes, B.D., Chao, H., and Davies, P.L. 1993. The nonhelical structure of antifreeze protein type III. Science 259: 1154–1157. [DOI] [PubMed] [Google Scholar]
  63. Sönnichsen, F.D., DeLuca, C.I., Davies, P.L., and Sykes, B.D. 1996. Refined solution structure of type III antifreeze protein: Hydrophobic groups may be involved in the energetics of the protein–ice interaction. Structure 4: 1325–1337. [DOI] [PubMed] [Google Scholar]
  64. Sönnichsen, F.D., Davies, P.L., and Sykes, B.D. 1998. NMR structural studies on antifreeze proteins. Biochem. Cell Biol. 76: 284–293. [DOI] [PubMed] [Google Scholar]
  65. Suzuki, T., Yasugi, M., Arisaka, F., Yamagishi, A., and Oshima, T. 2001. Adaptation of a thermophilic enzyme, 3-isopropylmalate dehydrogenase, to low temperatures. Protein Eng. 14: 85–91. [DOI] [PubMed] [Google Scholar]
  66. Thomas, T. and Cavicchioli, R. 2000. Effect of temperature on stability and activity of elongation factor 2 proteins from Antarctic and thermophilic methanogens. J. Bacteriol. 182: 1328–1332. [DOI] [PMC free article] [PubMed] [Google Scholar]
  67. Ushakov, B. 1964. Thermostability of cells and proteins in poikilotherms. Physiol. Rev. 44: 518–560. [DOI] [PubMed] [Google Scholar]
  68. Vaughan, C.K., Harryson, P., Buckle, A.M., and Fersht, A.R. 2002. A structural double-mutant cycle: Estimating the strength of a buried salt bridge in barnase. Acta Crystallogr. D Biol. Crystallogr. 58: 591–600. [DOI] [PubMed] [Google Scholar]
  69. Wang, L.H., Wusteman, M.C., Smallwood, M., and Pegg, D.E. 2002. The stability during low-temperature storage of an antifreeze protein isolated from the roots of cold-acclimated carrots. Cryobiology 44: 307–310. [DOI] [PubMed] [Google Scholar]
  70. Wintrode, P.L., Miyazaki, K., and Arnold, F.H. 2000. Cold adaptation of a mesophilic subtilisin-like protease by laboratory evolution. J. Biol. Chem. 275: 31635–31640. [DOI] [PubMed] [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, D.S., Hon, W.C., Bubanko, S., Xue, Y., Seetharaman, J., Hew, C.L., and Sicheri, F. 1998. Identification of the ice-binding surface on a type III antifreeze protein with a “flatness function” algorithm. Biophys. J. 74: 2142–2151. [DOI] [PMC free article] [PubMed] [Google Scholar]
  73. Zavodszky, P., Kardos, J., Svingor, A., and Pestko, G.A. 1998. Adjustment of conformational flexibility is a key event in the thermal adaptation of proteins. Proc. Natl. Acad. Sci. 95: 7406–7411. [DOI] [PMC free article] [PubMed] [Google Scholar]

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

RESOURCES