- Email: [email protected]
Contribution of hydrophobic residues to ice binding by fish type III antifreeze protein
Biochimica et Biophysica Acta 1601 (2002) 49 – 54 www.bba-direct.com
Contribution of hydrophobic residues to ice binding by fish type III antifreeze protein Jason Baardsnes, Peter L. Davies * Department of Biochemistry, Queen’s University, Stuart St., Kingston, ON, Canada K7L 3N6 Received 24 June 2002; received in revised form 29 August 2002; accepted 29 August 2002
Abstract Type III antifreeze protein (AFP) is a 7-kDa globular protein with a flat ice-binding face centered on Ala 16. Neighboring hydrophilic residues Gln 9, Asn 14, Thr 15, Thr 18 and Gln 44 have been implicated by site-directed mutagenesis in binding to ice. These residues have the potential to form hydrogen bonds with ice, but the tight packing of side chains on the ice-binding face limits the number and strength of possible hydrogen bond interactions. Recent work with alpha-helical AFPs has emphasized the hydrophobicity of their ice-binding sites and suggests that hydrophobic interactions are important for antifreeze activity. To investigate the contribution of hydrophobic interactions between type III AFP and ice, Leu, Ile and Val residues on the rim of the ice-binding face were changed to alanine. Mutant AFPs with single alanine substitutions, L19A, V20A, and V41A, showed a 20% loss in activity. Doubly substituted mutants, L19A/V41A and L10A/I13A, had less than 50% of the activity of the wild type. Thus, side chain substitutions that leave a cavity or undercut the contact surface are almost as deleterious to antifreeze activity as those that lengthen the side chain. These mutations emphasize the importance of maintaining a specific surface contour on the ice-binding face for docking to ice. D 2002 Elsevier Science B.V. All rights reserved. Keywords: Antifreeze protein; Site-directed mutagenesis; Thermal hysteresis; van der Waals interaction
1. Introduction Antifreeze proteins (AFPs) exist in many different organisms. In fish, they protect the host from freezing at temperatures below the colligative freezing point [1,2]. Type III AFP is one of five different fish AFPs. This 7kDa, compact, globular protein is produced by the ocean pout (Macrozoarces americanus) and other members of the subclass Zoarcoidei [3]. It protects the fish from freezing by adsorbing to the surface of nucleating ice crystals and inhibiting their growth. Thus, the freezing point of the fish is depressed below the ocean temperature by a non-colligative process [4]. There are numerous isoforms of type III AFP from ocean pout separable by HPLC and ion-exchange chromatography [5– 7]. Of these, the HPLC-12 isoform has been extensively studied and was the first to have its 3-D structure solved [8– 10]. Structures are also available for the HPLC-3 isoform [11], and for a natural dimer [12]. Mutagenic studies * Corresponding author. Tel.: +1-613-533-2983; fax: +1-613-533-2497. E-mail address: [email protected] (P.L. Davies).
on HPLC-12 have helped identify ice-binding residues. Candidate residues for site-directed mutagenesis were originally selected on the assumption that hydrogen bonding was the primary interaction with the ice surface [13]. By combining information from the initial NMR solution structure [14], with a comparison of type III isoform sequences, conserved hydrophilic residues forming a cluster on one face of the globular protein were identified as residues potentially involved in hydrogen bonding to the ice surface. Five of these residues, Gln 9, Asn14, Thr 15, Thr 18, and Gln 44, were targeted for substitution [13]. Replacement of these residues by amino acids with longer or shorter side chains caused significant loss of antifreeze activity and served to identify the ice-binding face of the AFP. The ice-binding face centered on Ala 16 is quite flat, and this attribute has been suggested to be important for antifreeze activity [11]. After the HPLC-12 structure was ˚ resolution, the five hydrophilic residues solved to 1.25 A and the carbonyl of Ala 16 were modeled to form hydrogen bonds with water molecules in the ice lattice on the {1010} plane of ice [8,9]. However, recent ice etching studies by Antson et al. [10] suggest that type III AFP adsorbs to other planes in addition to {10-10}. Their ice-docking
1570-9639/02/$ - see front matter D 2002 Elsevier Science B.V. All rights reserved. PII: S 1 5 7 0 - 9 6 3 9 ( 0 2 ) 0 0 4 3 1 - 4
50
J. Baardsnes, P.L. Davies / Biochimica et Biophysica Acta 1601 (2002) 49–54
simulations with a high-resolution X-ray structure of HPLC-12 indicate that type III AFP interacts with multiple ice planes, which helps explain the complex etch pattern. In the studies by Sonnichsen et al. [8] and Yang et al. [11], type III ice-binding specificity based solely on hydrogen bonding to the ice surface was questioned. The hydrophobicity of the ice-binding face was noted, and non-polar groups on this face, including the g-methyl groups of Thr 15 and Thr 18, were modeled to form favorable van der Waals contacts with the {10-10} plane of ice. Also, the number of predicted hydrogen bonds was quite low because donor side chain orientation and packing typically favored the formation of only one H-bond residue to the ice surface. Additionally, hydrogen bonds formed between the AFP and solid water are predicted to be less energetically favorable than hydrogen bonds between the AFP and liquid water because water is more constrained in ice [8,15]. For these reasons, other energetic components were considered for their potential contributions to ice binding. Computer simulations have been used to gauge type III AFP binding to ice. In one modelling approach, the proposed ice-binding surface was considered the most likely one to interact with ice [16]. In another study, Chen and Jia [17] analyzed 11 different surface patches in icedocking simulations for their potential to bind to ice. Each ˚ of either prism ice or a patch was brought to within 2 – 4 A
random ice slab, and binding energies were determined by energy minimization followed by molecular dynamics in vacuo. The more favorable interactions were solvated and underwent further optimization. The ice-binding face of type III AFP identified by mutagenesis was independently verified by this simulation as the surface that most favorably interacts with the prism plane of ice. This simulation suggested that the ice-binding surface is larger than previously thought and might incorporate surrounding hydrophobic residues like Leu 19, Val 20, Ile 13 and Val 41 (Fig. 1). A third simulation study used a different approach to identify the ice-binding surface; a neural network was constructed to separately calculate the contributions made by the accessible surface area of the protein, as well as hydrophobic, polar and charged residues [18]. The network was trained to predict thermal hysteresis activities based on previous structural and activity data of various mutants. It was tested by examining the correlation coefficient after leaving out one of the prediction parameters. Interestingly, the largest drop in the correlation coefficient occurred after the hydrophobic properties were left out of the calculations, the second largest drop occurred after the accessible surface area parameters were left out. These results are consistent with the emerging idea that shape complementarity and van der Waals interactions are important for ice binding.
Fig. 1. Three views of the type III AFP HPLC-12 isoform (PDB 1MSI). The ‘‘side’’ views (top) are rotated F 90j with respect to the front view of the icebinding surface (bottom). The hydrophobic residues on the periphery of the ice-binding face are shown in green, and the corresponding Ala substitutions are shown in yellow. The central residues colored by atom type have been previously identified as ice-binding residues by mutagenesis analysis (except Pro12) [13,30,31]. The van der Waals surfaces were generated by Swiss-PDBViewer (v. 3.6), and the images were rendered using POV-Ray (v. 3.1).
J. Baardsnes, P.L. Davies / Biochimica et Biophysica Acta 1601 (2002) 49–54
51
Here, we have assessed the importance of hydrophobic residues (Leu 10, Ile 13, Leu 19, Val 20, Val 41) that lie on the periphery of the ice-binding face for their contribution to binding. These residues were largely ignored until the recent results with type I AFP down-graded the importance of hydrophilic residues in the ice adsorption process [19 – 21], and redefined the ice-binding site as being relatively hydrophobic [22,23]. We have examined the contributions of these hydrophobic residues to ice binding by substituting them with alanine to reduce potential van der Waals contacts with the ice surface.
2. Materials and methods 2.1. Primer-directed mutagenesis Amino acid substitutions, L10A/I13A, L19A, V20A, L19A/V41A and V41A were made by primer-directed mutagenesis [24] as previously described [25]. The synthetic oligonucleotides used for mutagenesis are as follows: L 1 0 A / I 1 3 A 5 V- G T T G T G G C C A A C C A G G C GATCCCGGCTAATACTGCTCTGACTC-3V (PvuII knock-out), L19A 5V-TAATACTGCTCTGACCGCGGTT T A T G A T G C G T A G T G - 3 V( S s t I I ) , V 2 0 A 5 VGCTCTGACTCTGGCCATGATGCGTAGTG-3V (MscI), V41A 5V-ATATCCCGCGTCTGGCTAGCATGCAGGTTTAAC-3V(NheI). The underlined regions contain unique restriction sites used to select for positive clones, except for the L10A/I13A primer, which removed the wild-type PvuII restriction site. Sequences in bold indicate changed bases compared to the original type III sequence. To make L19A/ V41A, a double annealing reaction was carried out with the two single mutagenic primers. The template for the reactions was the recombinant Type III AFP (rQAE m1.1) [25], in the vector pT7-7(f) [26], which allows the same vector to be used for Kunkel mutagenesis and expression. All positive sequences were verified by automated DNA sequencing using the T7 forward primer (CORTEC, Inc.). 2.2. Expression and purification of the recombinant protein Expression plasmids for mutant or wild-type AFPs were co-transformed into Escherichia coli JM-83 with the vector
Table 1 Molecular weight verification of wild-type and mutant proteins by ES-MS Type III protein
MW theoretical
MW ES-MS
Wild type L19A V20A V41A L10A/I13A L19A/V41A
7035.3 6993.3 7007.3 7007.3 6951.2 6965.2
7035.7 6993.5 7007.8 7007.8 6951.5 6965.8
ProtParam was used to determine the theoretical molecular weight (MW) of the wild-type and mutant AFPs.
Fig. 2. Retention times of wild type and variant proteins on C18 reversedphase analytical HPLC column chromatography using an acetonitrile/ isopropanol (2:1) gradient. Absorbance was measured at 230 nm. Numbers designate the AFP peaks, and the percentage B for elution is in brackets. Peak 1: L10A/I13A (44%); peak 2: L19A/V41A (45%); peak 3: V41A (46%); peak 4: V20A (47%); peak 5: L19A (48%); peak 6: wild type (51%).
pGP1-2 [27]. Expression and purification of AFP from 4-l cultures were carried out as previously described [13]. The protein preparations including wild type, V41A and L19A/ V41A were refolded from inclusion bodies and purified by S-Sepharose FPLC. The remaining protein preparations L19A, V20A and L10A/V41A were fractionated from the soluble fraction of E. coli supernatant by the addition of 50 mM sodium acetate (pH 3.7) and purified by S-Sepharose FPLC [25]. Further purification by HPLC was carried out on L19A, V20A, V41A and L10A/I13A using a preparative C18 column (Vydac Inc.). Samples were loaded at a flow rate of 8.0 ml/min onto a column preequilibrated in 35% acetonitrile/isopropanol (2:1) containing 0.05% TFA and eluted by a linear acetonitrile/isopropanol (2:1) gradient increasing at 1%/min. Protein concentrations were determined by spectrophotometry using the known extinction coefficient of type III AFP m1.1. 2.3. Analytical HPLC An HPLC C18 analytical column (Vydac) was used to analyze the homogeneity and integrity of the purified proteins. Approximately 20 Ag of mutant or wild-type AFP was loaded at a flow rate of 1.0 ml/min onto a column preequilibrated in 35% acetonitrile/isopropanol (2:1) containing 0.05% TFA. Samples were eluted by
52
J. Baardsnes, P.L. Davies / Biochimica et Biophysica Acta 1601 (2002) 49–54
Analyses were performed at the Biological Mass Spectrometry Laboratory, University of Western Ontario. 2.5. Thermal hysteresis assays and ice morphology determination Thermal hysteresis is defined as the temperature difference (jC) between the melting point and non-equilibrium freezing point of an AFP solution. Mutant and wild-type AFPs were assayed for thermal hysteresis activity as described previously [28]. All measurements were made in 100 mM NH4HCO3. Ice growth of more than 0.2 Am/s signifies that the solution freezing point has been reached or exceeded. For ice morphology analyses, 0.5 mg/ml samples were undercooled by 0.1 jC over 75 s. Ice crystals were observed using a Leitz 22 microscope and recorded by a Panasonic CCTV camera linked to a JVC Super VHS video recorder. Still images were obtained from a Silicon Graphics INDY terminal using IRIS Capture version 1.2. Fig. 3. Thermal hysteresis activity of wild-type and mutant AFPs as a function of protein concentration. Each data point represents the average of at least three measurements: wild type (closed circle), L19A (open square, dashed line), V20A (closed square), V41A (open circle), L10A/I13A (closed triangle), and L19A/V41A (open triangle). Standard deviations are shown as vertical bars.
increasing the acetonitrile/isopropanol (2:1) concentration at a rate of 1%/min. 2.4. Mass spectrometry Mutant and wild-type proteins were analyzed by electrospray mass spectrometry to verify their masses. Samples were desalted using ZipTipC18k micropipette tips (Millipore) following standard protocols with a modification only to the final elution step. Due to the hydrophobic nature of the proteins, samples were eluted using 4 Al of 50% HPLC buffer (2:1 acetonitrile/isopropanol) containing 0.05% TFA.
3. Results 3.1. Purification and validation of the type III AFP mutants When recombinant type III AFP is overexpressed in E. coli, the wild-type protein is equally distributed between the inclusion bodies and soluble fraction [25]. We have previously observed that mutations influence this distribution, with hydrophilic substitutions driving more of the AFP into the supernatant fraction and hydrophobic substitutions resulting in more AFP in the inclusion bodies. As expected, the alanine-substitution mutants (with reduced hydrophobicity at one or more surface-exposed residues) were enriched in the soluble fraction during overexpression. This was particularly evident for the doubly substituted mutants (data not shown). The mass of each mutant determined by electrospray mass spectrometry was precisely that predicted for the
Fig. 4. Ice morphology of wild-type and mutant AFPs. Ice crystals were formed in the presence of 0.5 mg/ml AFP (except for 1.0 mg/ml L19A/V41A) in 0.1 M NH4HCO3 (pH 7.9). The images were taken at 0.1 jC of undercooling.
J. Baardsnes, P.L. Davies / Biochimica et Biophysica Acta 1601 (2002) 49–54
specific mutation (Table 1). This confirmed the amino acid substitution and demonstrated that there were no other changes to the AFP. Each mutant protein eluted from the C-18 reversed-phase HPLC as a single, sharp peak, which demonstrated its purity and confirmed that the AFP was well folded (Fig. 2). The order of elution was L10/I13A, L19A/V41A, V41A, V20A, L19A followed by wild type. In each case, the mutant AFPs had reduced retention times compared to the wild type. Reduced retention times were to be expected because the surface hydrophobicity of the variant proteins had been decreased. The mutants should consequently desorb from the stationary phase and partition into the solvent more readily compared to the wild type. That the double mutants L10A/I13A and L19A/V41A eluted before the single mutants is consistent with this expectation. 3.2. Effect of ice-binding surface reduction on antifreeze activity Each of the mutant type III AFPs showed decreased thermal hysteresis activity in comparison to the wild type (Fig. 3). Thermal hysteresis values were measured over a range of AFP concentrations and the relative activities were consistent over the entire range. At 1 mg/ml, L19A and V20A showed a 23% loss of activity compared to wild type, and V41A had a 28% loss of activity. The double mutants were considerably less active, with L10A/I13A and L19A/ V41A losing approximately 55% and 75% activity, respectively, compared to wild type at 1 mg/ml. The large drop in activity of the double mutants is approximately double that of the single mutants, suggesting that the effect of each Ala substitution is additive. These activity losses were reflected in changes to the ice crystal morphology. Wild-type AFP produced ice crystals with a characteristic hexagonal – bipyramid shape [29]. L19A and V20A produced ice bipyramids that were similar to the wild-type crystals in size and shape (Fig. 4) with a cto a-axis ratio of f 2:1 (Table 2). However, V41A produced a crystal that was noticeably elongated along the caxis with a ratio of 3.2:1. The double mutants (L10A/I13A and L19A/V41A) generated ice bipyramids with significantly larger c- to a-axis ratios of 5.5:1 and 4.5:1, respectively (Fig. 4 and Table 2). An increase in the c- to a-axis Table 2 c:a axis length ratios of ice crystals formed in 0.5 mg/ml AFP solution (or 1.0 mg/ml L19A/V41A) undercooled by 0.1 jC Protein
Wild type L19A V20A V41A L10A/I13A L19A/V41A
c:a Ratio Trial 1
Trial 2
Trial 3
Average
2.3:1 2.2:1 1.8:1 3.1:1 5.2:1 4.4:1
2.1:1 1.8:1 2.1:1 3.5:1 5.7:1 4.8:1
2.0:1 1.8:1 1.7:1 2.9:1 5.7:1 4.4:1
2.1 F 0.2 1.9 F 0.2 1.9 F 0.2 3.2 F 0.3 5.5 F 0.3 4.5 F 0.2
53
ratio was previously seen with type III AFP mutants that have decreased antifreeze activity [30].
4. Discussion The ice-binding face of type III AFP was first inferred by the conservation of a cluster of surface-exposed hydrophilic residues (Asn 14, Thr 18, Gln 44, Asn 46) on one face of the protein [13]. The involvement of these residues in ice binding was supported by site-directed mutagenesis. Originally, the mutations that caused loss of activity (e.g. T18N, N14S and Q44T) were interpreted as being due to a hydrogen-bonding mismatch between the side chain and ice. In retrospect, we suggest that mutations such as T18N, which caused a 90% loss of activity, are deleterious due to steric interference with the docking of the AFP to ice. This was demonstrated by the mutation series where Ala16 was replaced by amino acids with larger side chains [30,31], and for a different AFP (type I) by substitutions of Ala by Leu [22], and Ala by Lys [23]. Mutations that did not change the size or shape of the side chain (N14D, Q44E) were neutral, even though a charge was introduced and the role of hydrogen bond donor/acceptor was reversed [13]. Mutations that decreased the size of the side chain of these hydrophilic residues were also quite deleterious, with N14S, Q44T and T15A causing 75%, 50% and 30% loss of activity [9,13]. In a parallel system, there was a drastic loss of activity on replacing two ice-binding Thr in type I AFP with Ser but not when Val was substituted [19]. These results were extended and confirmed by other groups [20,21], and argue for a role of the Thr g-methyl group in ice binding. In both systems, the loss of hydrogen bonding potential may be secondary to the loss of surface –surface complementarity, and with it, van der Waals contacts. In addition, the hydrophobic character of the antifreeze ice-binding face might form a clathrate of solvating water molecules [32]. Burial of this face on the ice surface could desolvate the constrained water molecules that form the clathrate, thereby increasing the entropy of the system [8]. We have extended these studies by mutating hydrophobic surface residues on the ice-binding face (Leu 10, Ile 13, Leu 19, Val 20 and Val 41) to a smaller, but still hydrophobic, amino acid. These mutations, where one hydrophobic residue replaces another, remove direct hydrogen bonding from consideration. The observation that smaller hydrophobic side chains cause a loss of thermal hysteresis activity is consistent with a role for these residues in hydrophobic and van der Waals interactions for which optimal surface– surface complementarity is required. At the same time, these mutations might also alter the distribution of ordered water molecules on the ice-binding face. Indeed, shortening an aliphatic side chain should slightly reduce the amount of clathrate-like water. This in turn would reduce the net gain in entropy on AFP binding and slightly weaken the overall AFP –ice interaction.
54
J. Baardsnes, P.L. Davies / Biochimica et Biophysica Acta 1601 (2002) 49–54
Our view of how AFPs bind to ice, which has been shaped by structure – function studies on type I AFP [19 – 23], is supported by these studies on type III AFP. AFPs have three-dimensional shapes that allow them to present a significant percentage of their surface area for contact with ice. A key attribute of binding appears to be an intimate surface-to-surface complementarity between ice and the AFP, which is readily spoiled by mutations that extend or truncate the length of an ice-binding amino acid residue. The latter mutations detailed in this report would reduce the overall van der Waals interactions between the two surfaces. For any one Ala substitution, there would only be a small reduction in the van der Waals contribution to binding. However, these surrounding hydrophobic residues could be working in concert with the central hydrophilic residues. Loss of contact with the ice on the periphery of the icebinding face might weaken the enthalpic contribution of hydrogen bonds from residues in the centre. Although hydrogen bonding would be preferred between the more hydrophilic AFP surfaces and bulk solvent [15], once the AFP was optimally oriented to maximize hydrogen bonding to the solvent, any such bonds that could form within the ice –AFP contact zone would be expected to contribute to the enthalpy of binding. Reducing the volume of the outer residues may disrupt a ‘‘hydrophobic seal’’ around the potential hydrogen-bonding residues and allow competing water to disrupt the hydrogen-bonding network between adsorbed AFP and ice.
Acknowledgements We thank Drs. G. Lajoie and D. Brewer from the Biological Mass Spectrometry Laboratory, University of Western Ontario, for ES-MS, and Dr. F. So¨nnichsen of Case Western Reserve University for helpful discussions. This work was funded by a grant from the Canadian Institutes for Health Research (CIHR). P.L.D. is a Killam Research Fellow. J. B. was supported by an Ontario Graduate Scholarship and a Queen’s Graduate Fellowship.
References [1] G.L. Fletcher, C.L. Hew, P.L. Davies, Annu. Rev. Physiol. 63 (2001) 359 – 390. [2] K.V. Ewart, Q. Lin, C.L. Hew, Cell. Mol. Life Sci. 55 (1999) 271 – 283. [3] M. Shears, M.H. Kao, G. Scott, P.L. Davies, G.L. Fletcher, Mol. Mar. Biol. Biotechnol. 2 (1993) 104 – 111.
[4] J.A. Raymond, A.L. DeVries, Proc. Natl. Acad. Sci. U. S. A. 74 (1977) 2589 – 2593. [5] X.M. Li, K.Y. Trinh, C.L. Hew, B. Buettner, J. Baenziger, P.L. Davies, J. Biol. Chem. 260 (1985) 12904 – 12909. [6] C.L. Hew, N.C. Wang, S. Joshi, G.L. Fletcher, G.K. Scott, P.H. Hayes, B. Buettner, P.L. Davies, J. Biol. Chem. 263 (1988) 12049 – 12055. [7] J.D. Schrag, C.H. Cheng, M. Panico, H.R. Morris, A.L. DeVries, Biochim. Biophys. Acta 915 (1987) 357 – 370. [8] F.D. Sonnichsen, C.I. DeLuca, P.L. Davies, B.D. Sykes, Structure 4 (1996) 1325 – 1337. [9] Z. Jia, C.I. DeLuca, H. Chao, P.L. Davies, Nature 384 (1996) 285 – 288. [10] A.A. Antson, D.J. Smith, D.I. Roper, S. Lewis, L.S. Caves, C.S. Verma, S.L. Buckley, P.J. Lillford, R.E. Hubbard, J. Mol. Biol. 305 (2001) 875 – 889. [11] D.S. Yang, W.C. Hon, S. Bubanko, Y. Xue, J. Seetharaman, C.L. Hew, F. Sicheri, Biophys. J. 74 (1998) 2142 – 2151. [12] K. Miura, S. Ohgiya, T. Hoshino, N. Nemoto, T. Suetake, A. Miura, L. Spyracopoulos, H. Kondo, S.J. Tsuda, Biol. Chem. 276 (2001) 1304 – 1310. [13] H. Chao, F.D. Sonnichsen, C.I. DeLuca, B.D. Sykes, P.L. Davies, Protein Sci. 3 (1994) 1760 – 1769. [14] F.D. Sonnichsen, B.D. Sykes, H. Chao, P.L. Davies, Science 259 (1993) 154 – 157. [15] A.D.J. Haymet, L.G. Ward, M.M. Harding, J. Am. Chem. Soc. 121 (1999) 941 – 948. [16] J.D. Madura, M.S. Taylor, A. Wierzbicki, J.P. Harrington, C.S. Sikes, F.J. So¨nnichsen, Mol. Struct. 388 (1996) 65 – 77. [17] G. Chen, Z. Jia, Biophys. J. 77 (1999) 1602 – 1608. [18] S.P. Graether, C.I. DeLuca, J. Baardsnes, G.A. Hill, P.L. Davies, Z. Jia, J. Biol. Chem. 274 (1999) 11842 – 11847. [19] H. Chao, M.E. Houston Jr., R.S. Hodges, C.M. Kay, B.D. Sykes, M.C. Loewen, P.L. Davies, F.D. Sonnichsen, Biochemistry 36 (1997) 14652 – 14660. [20] A.D. Haymet, L.G. Ward, M.M. Harding, C.A. Knight, FEBS Lett. 430 (1998) 301 – 306. [21] W. Zhang, R.A. Laursen, J. Biol. Chem. 273 (1998) 34806 – 34812. [22] J. Baardsnes, L.H. Kondejewski, R.S. Hodges, H. Chao, C. Kay, P.L. Davies, FEBS Lett. 463 (1999) 87 – 91. [23] J. Baardsnes, M. Jelokhani-Niaraki, L.H. Kondejewski, M.J. Kuiper, C.M. Kay, R.S. Hodges, P.L. Davies, Protein Sci. 10 (2001) 2566 – 2576. [24] T.A. Kunkel, J.D. Roberts, R.A. Zakour, Methods Enzymol. 154 (1987) 367 – 382. [25] H. Chao, P.L. Davies, B.D. Sykes, F.D. Sonnichsen, Protein Sci. 2 (1993) 1411 – 1428. [26] C.I. DeLuca, P.L. Davies, J.A. Samis, J.S. Elce, Biochim. Biophys. Acta 1216 (1993) 81 – 93. [27] S. Tabor, C.C. Richardson, Proc. Natl. Acad. Sci. U. S. A. 92 (1995) 6339 – 6343. [28] A. Chakrabartty, C.L. Hew, Eur. J. Biochem. 202 (1991) 1057 – 1063. [29] H. Chao, C.I. DeLuca, P.L. Davies, FEBS Lett. 357 (1995) 183 – 186. [30] C.I. DeLuca, H. Chao, F.D. So¨nnichsen, B.D. Sykes, P.L. Davies, Biophys. J. 71 (1996) 2346 – 2355. [31] C.I. DeLuca, P.L. Davies, Q. Ye, Z.J. Jia, Mol. Biol. 275 (1998) 515 – 525. [32] T.E. Creighton, Proteins: Structures and Molecular Properties, 2nd ed., W.H. Freeman and Company, New York, 1993.
Contribution of hydrophobic residues to ice binding by fish type III antifreeze protein Jason Baardsnes, Peter L. Davies * Department of Biochemistry, Queen’s University, Stuart St., Kingston, ON, Canada K7L 3N6 Received 24 June 2002; received in revised form 29 August 2002; accepted 29 August 2002
Abstract Type III antifreeze protein (AFP) is a 7-kDa globular protein with a flat ice-binding face centered on Ala 16. Neighboring hydrophilic residues Gln 9, Asn 14, Thr 15, Thr 18 and Gln 44 have been implicated by site-directed mutagenesis in binding to ice. These residues have the potential to form hydrogen bonds with ice, but the tight packing of side chains on the ice-binding face limits the number and strength of possible hydrogen bond interactions. Recent work with alpha-helical AFPs has emphasized the hydrophobicity of their ice-binding sites and suggests that hydrophobic interactions are important for antifreeze activity. To investigate the contribution of hydrophobic interactions between type III AFP and ice, Leu, Ile and Val residues on the rim of the ice-binding face were changed to alanine. Mutant AFPs with single alanine substitutions, L19A, V20A, and V41A, showed a 20% loss in activity. Doubly substituted mutants, L19A/V41A and L10A/I13A, had less than 50% of the activity of the wild type. Thus, side chain substitutions that leave a cavity or undercut the contact surface are almost as deleterious to antifreeze activity as those that lengthen the side chain. These mutations emphasize the importance of maintaining a specific surface contour on the ice-binding face for docking to ice. D 2002 Elsevier Science B.V. All rights reserved. Keywords: Antifreeze protein; Site-directed mutagenesis; Thermal hysteresis; van der Waals interaction
1. Introduction Antifreeze proteins (AFPs) exist in many different organisms. In fish, they protect the host from freezing at temperatures below the colligative freezing point [1,2]. Type III AFP is one of five different fish AFPs. This 7kDa, compact, globular protein is produced by the ocean pout (Macrozoarces americanus) and other members of the subclass Zoarcoidei [3]. It protects the fish from freezing by adsorbing to the surface of nucleating ice crystals and inhibiting their growth. Thus, the freezing point of the fish is depressed below the ocean temperature by a non-colligative process [4]. There are numerous isoforms of type III AFP from ocean pout separable by HPLC and ion-exchange chromatography [5– 7]. Of these, the HPLC-12 isoform has been extensively studied and was the first to have its 3-D structure solved [8– 10]. Structures are also available for the HPLC-3 isoform [11], and for a natural dimer [12]. Mutagenic studies * Corresponding author. Tel.: +1-613-533-2983; fax: +1-613-533-2497. E-mail address: [email protected] (P.L. Davies).
on HPLC-12 have helped identify ice-binding residues. Candidate residues for site-directed mutagenesis were originally selected on the assumption that hydrogen bonding was the primary interaction with the ice surface [13]. By combining information from the initial NMR solution structure [14], with a comparison of type III isoform sequences, conserved hydrophilic residues forming a cluster on one face of the globular protein were identified as residues potentially involved in hydrogen bonding to the ice surface. Five of these residues, Gln 9, Asn14, Thr 15, Thr 18, and Gln 44, were targeted for substitution [13]. Replacement of these residues by amino acids with longer or shorter side chains caused significant loss of antifreeze activity and served to identify the ice-binding face of the AFP. The ice-binding face centered on Ala 16 is quite flat, and this attribute has been suggested to be important for antifreeze activity [11]. After the HPLC-12 structure was ˚ resolution, the five hydrophilic residues solved to 1.25 A and the carbonyl of Ala 16 were modeled to form hydrogen bonds with water molecules in the ice lattice on the {1010} plane of ice [8,9]. However, recent ice etching studies by Antson et al. [10] suggest that type III AFP adsorbs to other planes in addition to {10-10}. Their ice-docking
1570-9639/02/$ - see front matter D 2002 Elsevier Science B.V. All rights reserved. PII: S 1 5 7 0 - 9 6 3 9 ( 0 2 ) 0 0 4 3 1 - 4
50
J. Baardsnes, P.L. Davies / Biochimica et Biophysica Acta 1601 (2002) 49–54
simulations with a high-resolution X-ray structure of HPLC-12 indicate that type III AFP interacts with multiple ice planes, which helps explain the complex etch pattern. In the studies by Sonnichsen et al. [8] and Yang et al. [11], type III ice-binding specificity based solely on hydrogen bonding to the ice surface was questioned. The hydrophobicity of the ice-binding face was noted, and non-polar groups on this face, including the g-methyl groups of Thr 15 and Thr 18, were modeled to form favorable van der Waals contacts with the {10-10} plane of ice. Also, the number of predicted hydrogen bonds was quite low because donor side chain orientation and packing typically favored the formation of only one H-bond residue to the ice surface. Additionally, hydrogen bonds formed between the AFP and solid water are predicted to be less energetically favorable than hydrogen bonds between the AFP and liquid water because water is more constrained in ice [8,15]. For these reasons, other energetic components were considered for their potential contributions to ice binding. Computer simulations have been used to gauge type III AFP binding to ice. In one modelling approach, the proposed ice-binding surface was considered the most likely one to interact with ice [16]. In another study, Chen and Jia [17] analyzed 11 different surface patches in icedocking simulations for their potential to bind to ice. Each ˚ of either prism ice or a patch was brought to within 2 – 4 A
random ice slab, and binding energies were determined by energy minimization followed by molecular dynamics in vacuo. The more favorable interactions were solvated and underwent further optimization. The ice-binding face of type III AFP identified by mutagenesis was independently verified by this simulation as the surface that most favorably interacts with the prism plane of ice. This simulation suggested that the ice-binding surface is larger than previously thought and might incorporate surrounding hydrophobic residues like Leu 19, Val 20, Ile 13 and Val 41 (Fig. 1). A third simulation study used a different approach to identify the ice-binding surface; a neural network was constructed to separately calculate the contributions made by the accessible surface area of the protein, as well as hydrophobic, polar and charged residues [18]. The network was trained to predict thermal hysteresis activities based on previous structural and activity data of various mutants. It was tested by examining the correlation coefficient after leaving out one of the prediction parameters. Interestingly, the largest drop in the correlation coefficient occurred after the hydrophobic properties were left out of the calculations, the second largest drop occurred after the accessible surface area parameters were left out. These results are consistent with the emerging idea that shape complementarity and van der Waals interactions are important for ice binding.
Fig. 1. Three views of the type III AFP HPLC-12 isoform (PDB 1MSI). The ‘‘side’’ views (top) are rotated F 90j with respect to the front view of the icebinding surface (bottom). The hydrophobic residues on the periphery of the ice-binding face are shown in green, and the corresponding Ala substitutions are shown in yellow. The central residues colored by atom type have been previously identified as ice-binding residues by mutagenesis analysis (except Pro12) [13,30,31]. The van der Waals surfaces were generated by Swiss-PDBViewer (v. 3.6), and the images were rendered using POV-Ray (v. 3.1).
J. Baardsnes, P.L. Davies / Biochimica et Biophysica Acta 1601 (2002) 49–54
51
Here, we have assessed the importance of hydrophobic residues (Leu 10, Ile 13, Leu 19, Val 20, Val 41) that lie on the periphery of the ice-binding face for their contribution to binding. These residues were largely ignored until the recent results with type I AFP down-graded the importance of hydrophilic residues in the ice adsorption process [19 – 21], and redefined the ice-binding site as being relatively hydrophobic [22,23]. We have examined the contributions of these hydrophobic residues to ice binding by substituting them with alanine to reduce potential van der Waals contacts with the ice surface.
2. Materials and methods 2.1. Primer-directed mutagenesis Amino acid substitutions, L10A/I13A, L19A, V20A, L19A/V41A and V41A were made by primer-directed mutagenesis [24] as previously described [25]. The synthetic oligonucleotides used for mutagenesis are as follows: L 1 0 A / I 1 3 A 5 V- G T T G T G G C C A A C C A G G C GATCCCGGCTAATACTGCTCTGACTC-3V (PvuII knock-out), L19A 5V-TAATACTGCTCTGACCGCGGTT T A T G A T G C G T A G T G - 3 V( S s t I I ) , V 2 0 A 5 VGCTCTGACTCTGGCCATGATGCGTAGTG-3V (MscI), V41A 5V-ATATCCCGCGTCTGGCTAGCATGCAGGTTTAAC-3V(NheI). The underlined regions contain unique restriction sites used to select for positive clones, except for the L10A/I13A primer, which removed the wild-type PvuII restriction site. Sequences in bold indicate changed bases compared to the original type III sequence. To make L19A/ V41A, a double annealing reaction was carried out with the two single mutagenic primers. The template for the reactions was the recombinant Type III AFP (rQAE m1.1) [25], in the vector pT7-7(f) [26], which allows the same vector to be used for Kunkel mutagenesis and expression. All positive sequences were verified by automated DNA sequencing using the T7 forward primer (CORTEC, Inc.). 2.2. Expression and purification of the recombinant protein Expression plasmids for mutant or wild-type AFPs were co-transformed into Escherichia coli JM-83 with the vector
Table 1 Molecular weight verification of wild-type and mutant proteins by ES-MS Type III protein
MW theoretical
MW ES-MS
Wild type L19A V20A V41A L10A/I13A L19A/V41A
7035.3 6993.3 7007.3 7007.3 6951.2 6965.2
7035.7 6993.5 7007.8 7007.8 6951.5 6965.8
ProtParam was used to determine the theoretical molecular weight (MW) of the wild-type and mutant AFPs.
Fig. 2. Retention times of wild type and variant proteins on C18 reversedphase analytical HPLC column chromatography using an acetonitrile/ isopropanol (2:1) gradient. Absorbance was measured at 230 nm. Numbers designate the AFP peaks, and the percentage B for elution is in brackets. Peak 1: L10A/I13A (44%); peak 2: L19A/V41A (45%); peak 3: V41A (46%); peak 4: V20A (47%); peak 5: L19A (48%); peak 6: wild type (51%).
pGP1-2 [27]. Expression and purification of AFP from 4-l cultures were carried out as previously described [13]. The protein preparations including wild type, V41A and L19A/ V41A were refolded from inclusion bodies and purified by S-Sepharose FPLC. The remaining protein preparations L19A, V20A and L10A/V41A were fractionated from the soluble fraction of E. coli supernatant by the addition of 50 mM sodium acetate (pH 3.7) and purified by S-Sepharose FPLC [25]. Further purification by HPLC was carried out on L19A, V20A, V41A and L10A/I13A using a preparative C18 column (Vydac Inc.). Samples were loaded at a flow rate of 8.0 ml/min onto a column preequilibrated in 35% acetonitrile/isopropanol (2:1) containing 0.05% TFA and eluted by a linear acetonitrile/isopropanol (2:1) gradient increasing at 1%/min. Protein concentrations were determined by spectrophotometry using the known extinction coefficient of type III AFP m1.1. 2.3. Analytical HPLC An HPLC C18 analytical column (Vydac) was used to analyze the homogeneity and integrity of the purified proteins. Approximately 20 Ag of mutant or wild-type AFP was loaded at a flow rate of 1.0 ml/min onto a column preequilibrated in 35% acetonitrile/isopropanol (2:1) containing 0.05% TFA. Samples were eluted by
52
J. Baardsnes, P.L. Davies / Biochimica et Biophysica Acta 1601 (2002) 49–54
Analyses were performed at the Biological Mass Spectrometry Laboratory, University of Western Ontario. 2.5. Thermal hysteresis assays and ice morphology determination Thermal hysteresis is defined as the temperature difference (jC) between the melting point and non-equilibrium freezing point of an AFP solution. Mutant and wild-type AFPs were assayed for thermal hysteresis activity as described previously [28]. All measurements were made in 100 mM NH4HCO3. Ice growth of more than 0.2 Am/s signifies that the solution freezing point has been reached or exceeded. For ice morphology analyses, 0.5 mg/ml samples were undercooled by 0.1 jC over 75 s. Ice crystals were observed using a Leitz 22 microscope and recorded by a Panasonic CCTV camera linked to a JVC Super VHS video recorder. Still images were obtained from a Silicon Graphics INDY terminal using IRIS Capture version 1.2. Fig. 3. Thermal hysteresis activity of wild-type and mutant AFPs as a function of protein concentration. Each data point represents the average of at least three measurements: wild type (closed circle), L19A (open square, dashed line), V20A (closed square), V41A (open circle), L10A/I13A (closed triangle), and L19A/V41A (open triangle). Standard deviations are shown as vertical bars.
increasing the acetonitrile/isopropanol (2:1) concentration at a rate of 1%/min. 2.4. Mass spectrometry Mutant and wild-type proteins were analyzed by electrospray mass spectrometry to verify their masses. Samples were desalted using ZipTipC18k micropipette tips (Millipore) following standard protocols with a modification only to the final elution step. Due to the hydrophobic nature of the proteins, samples were eluted using 4 Al of 50% HPLC buffer (2:1 acetonitrile/isopropanol) containing 0.05% TFA.
3. Results 3.1. Purification and validation of the type III AFP mutants When recombinant type III AFP is overexpressed in E. coli, the wild-type protein is equally distributed between the inclusion bodies and soluble fraction [25]. We have previously observed that mutations influence this distribution, with hydrophilic substitutions driving more of the AFP into the supernatant fraction and hydrophobic substitutions resulting in more AFP in the inclusion bodies. As expected, the alanine-substitution mutants (with reduced hydrophobicity at one or more surface-exposed residues) were enriched in the soluble fraction during overexpression. This was particularly evident for the doubly substituted mutants (data not shown). The mass of each mutant determined by electrospray mass spectrometry was precisely that predicted for the
Fig. 4. Ice morphology of wild-type and mutant AFPs. Ice crystals were formed in the presence of 0.5 mg/ml AFP (except for 1.0 mg/ml L19A/V41A) in 0.1 M NH4HCO3 (pH 7.9). The images were taken at 0.1 jC of undercooling.
J. Baardsnes, P.L. Davies / Biochimica et Biophysica Acta 1601 (2002) 49–54
specific mutation (Table 1). This confirmed the amino acid substitution and demonstrated that there were no other changes to the AFP. Each mutant protein eluted from the C-18 reversed-phase HPLC as a single, sharp peak, which demonstrated its purity and confirmed that the AFP was well folded (Fig. 2). The order of elution was L10/I13A, L19A/V41A, V41A, V20A, L19A followed by wild type. In each case, the mutant AFPs had reduced retention times compared to the wild type. Reduced retention times were to be expected because the surface hydrophobicity of the variant proteins had been decreased. The mutants should consequently desorb from the stationary phase and partition into the solvent more readily compared to the wild type. That the double mutants L10A/I13A and L19A/V41A eluted before the single mutants is consistent with this expectation. 3.2. Effect of ice-binding surface reduction on antifreeze activity Each of the mutant type III AFPs showed decreased thermal hysteresis activity in comparison to the wild type (Fig. 3). Thermal hysteresis values were measured over a range of AFP concentrations and the relative activities were consistent over the entire range. At 1 mg/ml, L19A and V20A showed a 23% loss of activity compared to wild type, and V41A had a 28% loss of activity. The double mutants were considerably less active, with L10A/I13A and L19A/ V41A losing approximately 55% and 75% activity, respectively, compared to wild type at 1 mg/ml. The large drop in activity of the double mutants is approximately double that of the single mutants, suggesting that the effect of each Ala substitution is additive. These activity losses were reflected in changes to the ice crystal morphology. Wild-type AFP produced ice crystals with a characteristic hexagonal – bipyramid shape [29]. L19A and V20A produced ice bipyramids that were similar to the wild-type crystals in size and shape (Fig. 4) with a cto a-axis ratio of f 2:1 (Table 2). However, V41A produced a crystal that was noticeably elongated along the caxis with a ratio of 3.2:1. The double mutants (L10A/I13A and L19A/V41A) generated ice bipyramids with significantly larger c- to a-axis ratios of 5.5:1 and 4.5:1, respectively (Fig. 4 and Table 2). An increase in the c- to a-axis Table 2 c:a axis length ratios of ice crystals formed in 0.5 mg/ml AFP solution (or 1.0 mg/ml L19A/V41A) undercooled by 0.1 jC Protein
Wild type L19A V20A V41A L10A/I13A L19A/V41A
c:a Ratio Trial 1
Trial 2
Trial 3
Average
2.3:1 2.2:1 1.8:1 3.1:1 5.2:1 4.4:1
2.1:1 1.8:1 2.1:1 3.5:1 5.7:1 4.8:1
2.0:1 1.8:1 1.7:1 2.9:1 5.7:1 4.4:1
2.1 F 0.2 1.9 F 0.2 1.9 F 0.2 3.2 F 0.3 5.5 F 0.3 4.5 F 0.2
53
ratio was previously seen with type III AFP mutants that have decreased antifreeze activity [30].
4. Discussion The ice-binding face of type III AFP was first inferred by the conservation of a cluster of surface-exposed hydrophilic residues (Asn 14, Thr 18, Gln 44, Asn 46) on one face of the protein [13]. The involvement of these residues in ice binding was supported by site-directed mutagenesis. Originally, the mutations that caused loss of activity (e.g. T18N, N14S and Q44T) were interpreted as being due to a hydrogen-bonding mismatch between the side chain and ice. In retrospect, we suggest that mutations such as T18N, which caused a 90% loss of activity, are deleterious due to steric interference with the docking of the AFP to ice. This was demonstrated by the mutation series where Ala16 was replaced by amino acids with larger side chains [30,31], and for a different AFP (type I) by substitutions of Ala by Leu [22], and Ala by Lys [23]. Mutations that did not change the size or shape of the side chain (N14D, Q44E) were neutral, even though a charge was introduced and the role of hydrogen bond donor/acceptor was reversed [13]. Mutations that decreased the size of the side chain of these hydrophilic residues were also quite deleterious, with N14S, Q44T and T15A causing 75%, 50% and 30% loss of activity [9,13]. In a parallel system, there was a drastic loss of activity on replacing two ice-binding Thr in type I AFP with Ser but not when Val was substituted [19]. These results were extended and confirmed by other groups [20,21], and argue for a role of the Thr g-methyl group in ice binding. In both systems, the loss of hydrogen bonding potential may be secondary to the loss of surface –surface complementarity, and with it, van der Waals contacts. In addition, the hydrophobic character of the antifreeze ice-binding face might form a clathrate of solvating water molecules [32]. Burial of this face on the ice surface could desolvate the constrained water molecules that form the clathrate, thereby increasing the entropy of the system [8]. We have extended these studies by mutating hydrophobic surface residues on the ice-binding face (Leu 10, Ile 13, Leu 19, Val 20 and Val 41) to a smaller, but still hydrophobic, amino acid. These mutations, where one hydrophobic residue replaces another, remove direct hydrogen bonding from consideration. The observation that smaller hydrophobic side chains cause a loss of thermal hysteresis activity is consistent with a role for these residues in hydrophobic and van der Waals interactions for which optimal surface– surface complementarity is required. At the same time, these mutations might also alter the distribution of ordered water molecules on the ice-binding face. Indeed, shortening an aliphatic side chain should slightly reduce the amount of clathrate-like water. This in turn would reduce the net gain in entropy on AFP binding and slightly weaken the overall AFP –ice interaction.
54
J. Baardsnes, P.L. Davies / Biochimica et Biophysica Acta 1601 (2002) 49–54
Our view of how AFPs bind to ice, which has been shaped by structure – function studies on type I AFP [19 – 23], is supported by these studies on type III AFP. AFPs have three-dimensional shapes that allow them to present a significant percentage of their surface area for contact with ice. A key attribute of binding appears to be an intimate surface-to-surface complementarity between ice and the AFP, which is readily spoiled by mutations that extend or truncate the length of an ice-binding amino acid residue. The latter mutations detailed in this report would reduce the overall van der Waals interactions between the two surfaces. For any one Ala substitution, there would only be a small reduction in the van der Waals contribution to binding. However, these surrounding hydrophobic residues could be working in concert with the central hydrophilic residues. Loss of contact with the ice on the periphery of the icebinding face might weaken the enthalpic contribution of hydrogen bonds from residues in the centre. Although hydrogen bonding would be preferred between the more hydrophilic AFP surfaces and bulk solvent [15], once the AFP was optimally oriented to maximize hydrogen bonding to the solvent, any such bonds that could form within the ice –AFP contact zone would be expected to contribute to the enthalpy of binding. Reducing the volume of the outer residues may disrupt a ‘‘hydrophobic seal’’ around the potential hydrogen-bonding residues and allow competing water to disrupt the hydrogen-bonding network between adsorbed AFP and ice.
Acknowledgements We thank Drs. G. Lajoie and D. Brewer from the Biological Mass Spectrometry Laboratory, University of Western Ontario, for ES-MS, and Dr. F. So¨nnichsen of Case Western Reserve University for helpful discussions. This work was funded by a grant from the Canadian Institutes for Health Research (CIHR). P.L.D. is a Killam Research Fellow. J. B. was supported by an Ontario Graduate Scholarship and a Queen’s Graduate Fellowship.
References [1] G.L. Fletcher, C.L. Hew, P.L. Davies, Annu. Rev. Physiol. 63 (2001) 359 – 390. [2] K.V. Ewart, Q. Lin, C.L. Hew, Cell. Mol. Life Sci. 55 (1999) 271 – 283. [3] M. Shears, M.H. Kao, G. Scott, P.L. Davies, G.L. Fletcher, Mol. Mar. Biol. Biotechnol. 2 (1993) 104 – 111.
[4] J.A. Raymond, A.L. DeVries, Proc. Natl. Acad. Sci. U. S. A. 74 (1977) 2589 – 2593. [5] X.M. Li, K.Y. Trinh, C.L. Hew, B. Buettner, J. Baenziger, P.L. Davies, J. Biol. Chem. 260 (1985) 12904 – 12909. [6] C.L. Hew, N.C. Wang, S. Joshi, G.L. Fletcher, G.K. Scott, P.H. Hayes, B. Buettner, P.L. Davies, J. Biol. Chem. 263 (1988) 12049 – 12055. [7] J.D. Schrag, C.H. Cheng, M. Panico, H.R. Morris, A.L. DeVries, Biochim. Biophys. Acta 915 (1987) 357 – 370. [8] F.D. Sonnichsen, C.I. DeLuca, P.L. Davies, B.D. Sykes, Structure 4 (1996) 1325 – 1337. [9] Z. Jia, C.I. DeLuca, H. Chao, P.L. Davies, Nature 384 (1996) 285 – 288. [10] A.A. Antson, D.J. Smith, D.I. Roper, S. Lewis, L.S. Caves, C.S. Verma, S.L. Buckley, P.J. Lillford, R.E. Hubbard, J. Mol. Biol. 305 (2001) 875 – 889. [11] D.S. Yang, W.C. Hon, S. Bubanko, Y. Xue, J. Seetharaman, C.L. Hew, F. Sicheri, Biophys. J. 74 (1998) 2142 – 2151. [12] K. Miura, S. Ohgiya, T. Hoshino, N. Nemoto, T. Suetake, A. Miura, L. Spyracopoulos, H. Kondo, S.J. Tsuda, Biol. Chem. 276 (2001) 1304 – 1310. [13] H. Chao, F.D. Sonnichsen, C.I. DeLuca, B.D. Sykes, P.L. Davies, Protein Sci. 3 (1994) 1760 – 1769. [14] F.D. Sonnichsen, B.D. Sykes, H. Chao, P.L. Davies, Science 259 (1993) 154 – 157. [15] A.D.J. Haymet, L.G. Ward, M.M. Harding, J. Am. Chem. Soc. 121 (1999) 941 – 948. [16] J.D. Madura, M.S. Taylor, A. Wierzbicki, J.P. Harrington, C.S. Sikes, F.J. So¨nnichsen, Mol. Struct. 388 (1996) 65 – 77. [17] G. Chen, Z. Jia, Biophys. J. 77 (1999) 1602 – 1608. [18] S.P. Graether, C.I. DeLuca, J. Baardsnes, G.A. Hill, P.L. Davies, Z. Jia, J. Biol. Chem. 274 (1999) 11842 – 11847. [19] H. Chao, M.E. Houston Jr., R.S. Hodges, C.M. Kay, B.D. Sykes, M.C. Loewen, P.L. Davies, F.D. Sonnichsen, Biochemistry 36 (1997) 14652 – 14660. [20] A.D. Haymet, L.G. Ward, M.M. Harding, C.A. Knight, FEBS Lett. 430 (1998) 301 – 306. [21] W. Zhang, R.A. Laursen, J. Biol. Chem. 273 (1998) 34806 – 34812. [22] J. Baardsnes, L.H. Kondejewski, R.S. Hodges, H. Chao, C. Kay, P.L. Davies, FEBS Lett. 463 (1999) 87 – 91. [23] J. Baardsnes, M. Jelokhani-Niaraki, L.H. Kondejewski, M.J. Kuiper, C.M. Kay, R.S. Hodges, P.L. Davies, Protein Sci. 10 (2001) 2566 – 2576. [24] T.A. Kunkel, J.D. Roberts, R.A. Zakour, Methods Enzymol. 154 (1987) 367 – 382. [25] H. Chao, P.L. Davies, B.D. Sykes, F.D. Sonnichsen, Protein Sci. 2 (1993) 1411 – 1428. [26] C.I. DeLuca, P.L. Davies, J.A. Samis, J.S. Elce, Biochim. Biophys. Acta 1216 (1993) 81 – 93. [27] S. Tabor, C.C. Richardson, Proc. Natl. Acad. Sci. U. S. A. 92 (1995) 6339 – 6343. [28] A. Chakrabartty, C.L. Hew, Eur. J. Biochem. 202 (1991) 1057 – 1063. [29] H. Chao, C.I. DeLuca, P.L. Davies, FEBS Lett. 357 (1995) 183 – 186. [30] C.I. DeLuca, H. Chao, F.D. So¨nnichsen, B.D. Sykes, P.L. Davies, Biophys. J. 71 (1996) 2346 – 2355. [31] C.I. DeLuca, P.L. Davies, Q. Ye, Z.J. Jia, Mol. Biol. 275 (1998) 515 – 525. [32] T.E. Creighton, Proteins: Structures and Molecular Properties, 2nd ed., W.H. Freeman and Company, New York, 1993.