Báo cáo khoa học: The crystal structure of the tryptophan synthase b2 subunit from the hyperthermophile Pyrococcus furiosus Investigation of stabilization factors

doi:10.1111/j.1432-1033.2004.04191.x

Eur. J. Biochem. 271, 2624–2635 (2004) (cid:3) FEBS 2004

The crystal structure of the tryptophan synthase b2 subunit from the hyperthermophile Pyrococcusfuriosus Investigation of stabilization factors

Yusaku Hioki1,2, Kyoko Ogasahara1, Soo Jae Lee1, Jichun Ma1, Masami Ishida3, Yuriko Yamagata4, Yoshiki Matsuura1, Motonori Ota5, Mitsunori Ikeguchi6, Seiki Kuramitsu2 and Katsuhide Yutani7,8 1Institute for Protein Research, Osaka University, Japan; 2Department of Biology, Graduate School of Science, Osaka University, Japan; 3Tokyo University Marine Science and Technology, Japan; 4Graduate School of Pharmaceutical Sciences, Kumamoto University, Japan; 5Global Scientific Information and Computing Center, Tokyo Institute of Technology, Japan; 6Graduate School of Integrated Science, Yokohama City University, Japan; 7Kwansei Gakuin University, Graduate School of Sciences, Hyogo, Japan; 8RIKEN Harima Institute, HTPF, Hyogo, Japan

the reported values from mesophiles at (cid:1) pH 8. On the basis of structural information on both proteins, the analyses of the contributions of each stabilization factor indicate that: (a) the higher stability of Pfb2 is not caused by either a hydrophobic interaction or an increase in ion pairs; (b) the number of hydrogen bonds involved in the main chains of Pfb is greater by about 10% than that of Stb, indicating that the secondary structures of Pfb are more stabilized than those of Stb and (c) the sequence of Pfb seems to be better fitted to an ideally stable structure than that of Stb, as assessed from X-ray structure data.

Keywords: calorimetry; crystal structure; hyperthermophile; tryptophan synthase b2 subunit; stability.

The structure of the tryptophan synthase b2 subunit (Pfb2) from the hyperthermophile, Pyrococcus furiosus, was deter- mined by X-ray crystallographic analysis at 2.2 A˚ resolu- tion, and its stability was examined by DSC. This is the first report of the X-ray structure of the tryptophan synthase b2 subunit alone, although the structure of the tryptophan synthase a2b2 complex from Salmonella typhimurium has already been reported. The structure of Pfb2 was essentially similar to that of the b2 subunit (Stb2) in the a2b2 complex from S. typhimurium. The sequence alignment with secon- dary structures of Pfb and Stb in monomeric form showed that six residues in the N-terminal region and three residues in the C-terminal region were deleted in Pfb, and one residue at Pro366 of Stb and at Ile63 of Pfb was inserted. The denaturation temperature of Pfb2 was higher by 35 (cid:1)C than

a reaction

indole-3-glycerol phosphate $ indole þ d-glyceraldehyde 3-phosphate

ð1Þ

b reaction

ð2Þ

l-serine þ indole ! l-tryptophan þ H2O

Prokaryotic tryptophan synthase (EC 4.2.1.20) is an a2b2 complex composed of nonidentical a and b subunits [1,2]. The a2b2 complex with an abba arrangement [3] can be isolated as the a monomer and b2 subunits. The a and b2 subunits catalyse inherent reactions, termed the a and b reactions (Eqns 1 and 2), respectively. The physiologically important reaction catalysed by the a2b2 complex, termed the ab reaction (Eqn 3), is the sum of the a and b reactions:

ab reaction

l-serine þ indole 3-glycerol phosphate !

l-tryptophan þ d-glyceraldehyde 3-phosphate þ H2O ð3Þ

When the a and b2 subunits associate to form the a2b2 complex, the enzymatic activity of each subunit is syn- chronically enhanced by one to two orders of magnitude [2]. The tryptophan synthase is a typical allosteric enzyme whose activity is affected by the ligands [3–6]. Prokaryotic tryptophan synthase has been studied extensively as an excellent model system for investigating protein–protein interaction mechanisms [2,7–10].

In order to elucidate the structural basis of the subunit communication and mutual activation of the functions of each subunit resulting from the formation of the a2b2 complex, it is necessary to determine the three-dimensional

Correspondence to K. Yutani, RIKEN Harima Institute, HTPF, Kouto, Mikazuki-cho, Sayo-gun, Hyogo 679-5148, Japan. Fax: +81 791 58 2917, Tel.: +81 791 58 2937, E-mail: yutani@spring8.or.jp Abbreviations: ASA, accessible surface area; Eca, tryptophan synthase a subunit from Escherichia coli; Ecb2, tryptophan synthase b2 subunit from E. coli; Pfa, tryptophan synthase a subunit from Pyrococcus furiosus; Pfb2, tryptophan synthase b2 subunit from P. furiosus; Pfb, monomer of tryptophan synthase b2 subunit from P. furiosus; PLP, pyridoxal 5¢-phosphate; Sta, tryptophan synthase a subunit from Salmonella typhimurium; Stb2, tryptophan synthase b2 subunit from S. typhimurium; Stb, monomer of tryptophan synthase b2 subunit from S. typhimurium; RMSD, root mean square deviation. Enzymes: prokaryotic tryptophan synthase (EC 4.2.1.20). (Received 21 January 2004, revised 25 March 2004, accepted 28 April 2004)

(cid:3) FEBS 2004

Crystal structure of tryptophan synthase b2 subunit alone (Eur. J. Biochem. 271) 2625

tometric assay ordinarily used [33], because temperature control of the spectrophotometer was difficult above 80 (cid:1)C. The assay was carried out in the presence of a 3 : 1 molar excess of the a subunit over the b subunit monomer. One unit of activity is defined by the formation of 0.1 lmol of product in 20 min at the indicated temperature [33].

DSC

structures of the a or b2 subunits alone as well as that of the complex. The three-dimensional structure of the tryptophan synthase a2b2 complex from Salmonella typhimurium was determined by X-ray analysis in 1988 [3]. However, the determination of the structure of the a or b2 subunit alone has not yet succeeded, although much effort has expended on obtaining good quality crystals of the subunits from mesophiles. Recently, the structure of the a subunit alone of tryptophan synthase from a hyperthermophile, Pyrococcus furiosus, was determined by X-ray analysis [11]. In this report we describe the crystal structure of the b2 subunit alone of tryptophan synthase from P. furiosus.

DSC was carried out using an adiabatic differential microcalorimeter, VP-DSC (Microcal) at a scan rate of 1 (cid:1)CÆmin)1. Before making measurements, the protein solution was dialysed against buffer with the composition 10 mM Gly/KOH, 1 mM EDTA, 0.02 mM pyridoxal 5¢-phosphate (PLP) (as described in Fig. 1). The dialysed sample was filtered through a 0.22-lm pore size membrane and then degassed in a vacuum. Protein concentrations during the measurements were 0.5–1.5 mgÆmL)1.

Protein crystallization and data collection

Proteins from hyperthermophiles are remarkably stable compared with homologous proteins from mesophiles [12,13]. Three-dimensional structures of many proteins from hyperthermophiles have been analysed to determine the structural bases of unusually high stability [14–17]. Structural features of hyperthermophile proteins compared with their mesophilic homologues vary depending on the individual proteins. Hydrophilic factors such as ion pairs and hydrogen bonds are superior in some proteins [11,14,18–23], and hydrophobic interaction is favoured in others [24,25]. The internal cavity decreases in hyperthermo- phile proteins [25]. An entropic effect has been reported to be important for enhanced stability [11].

The crystals were grown by a hanging drop vapour diffusion at 10 (cid:1)C, by mixing 2 lL of the protein solution with 2 lL of a reservoir solution containing 12% (w/v) PEG 20 000 and 100 mM Mes, pH 6.5. The concentration of Pfb2 was 10–12 mgÆmL)1 in 20 mM Tris/HCl pH 8.5 containing 100 lM dithioerythritol and 20 lM PLP.

Diffraction experiments with the Pfb2 crystal were performed at the beam line, BL44XU and BL411XU at SPring8. The crystal belonged to the orthorhombic space group of P212121 with unit cell dimensions of a ¼ 84.8, b ¼ 110.5, c ¼ 160.0 A˚ . The value of the Matthews coefficient is 2.2 A˚ 3ÆDa)1 for two Pfb2 per asymmetric unit, correspond- ing to a solvent content of 44.0%. The crystals were flash- cooled in a cold nitrogen gas stream immediately after cryoprotection by addition of the reservoir solution con- taining 25% (w/v) glycerol to the crystallization buffer at

However, the cause of the extremely high stabilization of proteins from hyperthermophiles still remains unclear. Elucidating the structural basis of the ultra-thermostability of proteins is an important for understanding protein folding problems, aspects of biotechnological applications, and progress in structural genomics. Using mutant human lysozymes Funahashi et al. [26,27] have proposed the parameters of various stabilization factors estimated by a unique equation, considering the relationship between stability and conformational changes due to the mutations. Using these parameters, the stabilization mechanism of pyrrolidone carboxyl peptidase from P. furiosus has been elucidated on the basis of its X-ray structure [17]. In this report, the stabilization mechanism of the hyperthermophilc b2 subunit will be discussed on the basis of the crystal structures, compared with the structural features of the hyperthermophile and mesophile proteins.

Materials and methods

Purification of proteins

The b2 subunit of tryptophan synthase from P. furiosus (Pfb2) was overproduced in Escherichia coli strain JM109 [28]. Pfb2 and the a-subunit of (pb1837) tryptophan synthase from P. furiosus (Pfa) were purified as described [29,11]. The equivalent subunits from E. coli (Eca, Ecb2) were purified also [10,30,31]. All of the purified proteins showed a single band on SDS/PAGE.

The protein concentrations were determined from the 1 cm ¼ 6.92 for Pfa and

absorbance at 278.5 nm using A1% 10.18 for Pfb2 [29], 4.4 for Eca [32] and 6.5 for Ecb2 [33].

Enzymatic activity assay

The b activity was measured by the disappearance of indole using a phenol reagent [1] instead of the direct spectropho-

Fig. 1. pH dependence of the denaturation temperature of Pf b2. The denaturation temperature, Td, represents the peak temperatures of DSC curves observed at a scan rate of 1 (cid:1)CÆmin)1. d, s and m represent Pfb2, Ecb2, and Stb2, respectively. The buffer conditions were 10 mM Gly-KOH with 1 mM EDTA and 0.02 mM PLP. The pH indicates the values after DSC measurements. The data for Ecb2 and Stb2 are those reported in [29] and [35] respectively.

2626 Y. Hioki et al. (Eur. J. Biochem. 271)

(cid:3) FEBS 2004

have been deposited in the Protein Data Bank (PDB accession no. 1V8Z).

Table 1. Data collection and refinement statistics of the tryptophan synthase b subunit from P. furiosus.

Characteristics of the crystals

Results

Thermal stability and enzymatic activity of Pfb2

P212121 Space group Cell parameters

)1)

a (A˚ ) b (A˚ ) c (A˚ )

84.8 110.5 160.0 16 2.2 44 Z Vm (A˚ 3ÆDa Solvent content (%) Data collection

Resolution (A˚ ) No. of unique reflections Average redundancy Rmerge (%)a,b Completeness (%)a 2.2 75 098 9.2 5.3 (27.6) 98.7 (97.0)

Figure 1 shows the pH dependence of the denaturation temperatures of Pfb2 measured by DSC. The heat denatur- ation of Pfb2 was not reversible. The peak temperatures of the DSC curves above pH 6.5 were around 115 (cid:1)C independent of pH, which were higher by about 35 (cid:1)C than those reported for mesophilic proteins [29,35]. DSC meas- urements could not be carried out between pH 6 and 4, because the protein became turbid on heating. Below pH 4, the denaturation temperatures decreased markedly. Ultra- centrifugation analysis of Pfb2 indicates that the apparent molecular weight of the protein, which exists in a dimeric form in solution around pH 7, decreases with decreasing pH below 4.0, resulting in dissociation to a monomer at pH 3.0 [29]. This suggests that the decreased denaturation temperature below pH 4.0 is correlated with the dissociation from a dimer to a monomer. The mesophilic protein of E. coli (Ecb2) was denatured in the acidic region.

67–2.2 75 065 20.8 26.3 Refinement statics Resolution (A˚ ) No. of reflections Rfactor (%)c Rfree (%)d RMSDs

a Values within parentheses are for the last shell of data. b Rmerge ¼ ShSi|(Ih – Ihi)|/ShSiIhi* 100. c Rfactor ¼ S||Fobs| – |Fcalc||/S|Fobs| * 100. d Rfree ¼ S||Fobs| – |Fcalc||/S|Fobs| * 100 where |Fobs| are test set amplitudes (10%) not used in refinement.

100 K. This crystal diffracted to a maximum of 2.2 A˚ and was suitable for structure determination.

The data collected were processed and integrated by DENZO and scaled by SCALEPACK [34]. Data collection statistics are summarized in Table 1.

Structure determination and refinement

The enzymatic activities of Pfb2 and Ecb2 were measured at various temperatures in the presence of excess a subunit from P. furiosus or E. coli (Fig. 2B). The activity for Ecb2 rapidly decreased at temperatures above 55 (cid:1)C. This decrease might be due to thermal denaturation of Eca in a2b2 complex, because the denaturation temperature of Eca is around 55 (cid:1)C, although Ecb2 denatures at 80 (cid:1)C [29]. It has also been reported that Sta in the complex is inactivated by 50% at 55 (cid:1)C, whereas 50% inactivation of Stb2 occurs at 80 (cid:1)C [4]. The activity of Pfb2 at the physiological temperature of mesophiles was negligible, although the specific activity for Pfb2 around 90 (cid:1)C was comparable with that of Ecb2 around 50 (cid:1)C. This was in marked contrast to the result with a hyperthermophilic pyrrolidone carboxyl peptidase from P. furiosus, which exhibits higher specific activity over a broad range of temperature than the corresponding mesophilic protein [13].

complex

(Sta2b2)

The dimeric structure (Stb2) of the b subunit in the tryptophan synthase a2b2 from S. typhimurium (1BKS) [3] provided the initial model for molecular replacement solutions using AMORE. The cross- rotation function showed two peaks for the two-dimer molecules. The model was subjected to cycles of rigid body refinement using noncrystallographic symmetry (NCS): the four b subunit molecules in the asymmetric unit were refined using NCS restraints. The experimental map at 2.2 A˚ was of high quality and allowed unambiguous modelling of all residues 1–388. The model was built using O and refined by energy minimization, simulated annealing and restrained B-factor refinement procedures with NCS. Successive refinement with temperature factors and addition of solvents resulted in an R-value of 22.0% and an Rfree of 26.4% for all reflections in the resolution range 100–2.2 A˚ . Rfree was calculated with 10% of the reflections. The current model consists of four chains of residues 1–388 of Pfb and 193 water molecules per asymmetric unit. All residues are within the most favoured (89.7%) and additional allowed regions (10.3%) of the Ramachandran plot. Refinement statistics are summarized in Table 1. The final coordinates

The Arrhenius plots of the activity for Pfb2 were clearly divided into two lines at a boundary around 45 (cid:1)C (Fig. 2A). The low-temperature portion showed a much higher slope than the high-temperature portion. The Arrhenius activation energies (Ea) of Pfb2 calculated from the slopes were 215.4 and 54.6 kJÆmol)1 for the low- and high-temperature portions, respectively. The Ea values of Pfb2, especially in the low-temperature portion, were higher than that for Ecb2 (135.7 and 43.0 kJÆmol)1, respectively) which also showed biphasic Arrhenius plots (Fig. 2A). The Ea values for Ecb2 were similar to those of b activity of tryptophan synthase (Sta2b2) from S. typhimurium reported [5,36]. Based on the effect of temperature on the catalytic properties for Sta2b2 and Stb2 in the presence of monova- lent cations and an allosteric ligand, Fan et al. have shown that biphasic Arrhenius plots are caused by a temperature- dependent conformational change from a low-activity (cid:2)open(cid:3) conformation to a high-activity (cid:2)closed(cid:3) conformation [36]. It seems that the Pfb2 is also converted from a low activity conformation to a highly active one by increasing temperature.

0.006 1.3 RMSD lengths (A˚ ) RMSD angles ((cid:1))

(cid:3) FEBS 2004

Crystal structure of tryptophan synthase b2 subunit alone (Eur. J. Biochem. 271) 2627

Table 2. Comparison of the amino acid compositions of the tryptophan synthase b subunit monomers from P. furiosus and S. typhimurium. Values within parentheses are for the percentage of residue per total number of residues.

Pfba Stbb D(Pfb–Stb)

Residue number Residue number Differences in residue number Residue

Hydrophobic

Gly Ala Val Leu Ile Met Phe Trp Pro Neutral Ser Thr Asn Gln Cys Hydrophilic

Asp Glu Lys His Arg Tyr )4 ()0.22) 1 ()0.01) )5 ()1.04) 11 (2.94) )3 ()0.55) )1 ()0.12) )4 ()0.94) 0 (0.08) 2 (0.52) )3 ()0.66) )16 ()3.70) )4 ()0.92) )5 ()1.17) 2 (0.58) )5 ()1.19) )4 ()1.00) 11 (3.48) )1 ()0.15) 2 (0.68) 11 (2.94) )3 ()0.69) )3 ()0.67) 5 (1.36) 210 (54.12) 42 (10.82) 38 (9.79) 30 (7.73) 35 (9.02) 23 (5.93) 11 (2.84) 13 (3.35) 3 (0.77) 15 (3.87) 57 (14.69) 15 (3.87) 16 (4.12) 13 (3.35) 12 (3.09) 1 (0.26) 121 (31.19) 17 (4.38) 30 (7.73) 30 (7.73) 11 (2.84) 16 (4.12) 17 (4.38) 388 214 (53.90) 43 (10.83) 43 (10.83) 19 (4.79) 38 (9.57) 24 (6.05) 15 (3.78) 13 (3.27) 1 (0.25) 18 (4.53) 73 (18.39) 19 (4.79) 21 (5.29) 11 (2.77) 17 (4.28) 5 (1.26) 110 (27.71) 18 (4.53) 28 (7.05) 19 (4.79) 14 (3.53) 19 (4.79) 12 (3.02) 397 )9 Total number of residues

Amino acid composition of the b subunit from P.furiosus

Table 2 shows the amino acid compositions of both b monomers of Pfb2 and Stb2 (Pfb and Stb, respectively). Pfb consists of 388 residues, but Stb has 397. The content (%) of hydophobic residues for Pfb was similar to that for Stb, although the number of hydrophobic amino acid residues of Pfb was slightly lowered. The number of hydrophilic residues increased from 110 (27.71%) to 121 (31.19%) in Pfb, compared with that of Stb. The number of neutral residues of Pfb was largely reduced from 73 (18.39%) to 57 (14.69%). In the case of the a subunit of tryptophan synthase from P. furiosus, hydrophobic residues were remarkably reduced from 58.58% to 53.93%, compared with those from S. typhimurium [11].

Overall structure of the b subunit from P.furiosus

of two domains, N (residues 1–46, 81–200) and C (residues 47–80, 201–388) domains of almost equal size. The N-terminal (1–200), and the C-terminal (201–388) residues are coloured red and blue, respectively. The core of the N domain is formed from four strands which are surrounded by seven helices. The core of the C domain constitutes six strands with five parallel strands and one antiparallel strand. A short piece (residues 47–80) of the N-terminal residues intrudes into the C domain, forming the first two strands of a b-sheet at the centre of the C domain. A helical structure (residues 58–64) between the first two strands is clearly observed in Pfb although it is not reported in Stb. Arrows in Fig. 3A point to the first two strands and one helical structure (residue 58–64) that intrude into the C domain. The coenzyme PLP is located in the deep cleft between the two domains. PLP forms a Shiff base with the e-amino group of Lys82 in Pfb, corresponding to Lys87 in an active site of Stb. The overall topology of Pfb was equivalent to the b subunit monomer in the Sta2b2 complex reported by Hyde et al. [3].

Structural comparison of Pfb and Stb

The structure of the b subunit from P. furiosus was observed as a dimeric form in which the two b subunits are tightly associated over a broad surface. The buried surface at the interface between the two subunits was estimated to be 3945 A˚ 2 (Table 3). The dimer structure is depicted by the ribbon drawing in Fig. 3A. The subunit structure consists

Fig. 4 shows the secondary structure-based sequence align- ment using the secondary structure elements assigned by

Amino acid compositions were taken from: a Ishida et al. [28] and b Hyde et al. [3]. Fig. 2. Temperature dependence of the specific enzymatic activities of the b reaction for the Pfb2 subunit (d) and the Ecb2 subunit (s) at pH 7.0. Activities for Pfb2 and Ecb2 were measured in the presence of an excess of the a subunit from P. furiosus and E. coli, respectively. (A) Arrhenius plots of the data from (B). (B) Comparison of the activities of Pfb2 and Ecb2.

2628 Y. Hioki et al. (Eur. J. Biochem. 271)

(cid:3) FEBS 2004

Table 3. Estimate of the difference in stability between tryptophan synthase b subunits from P. furiosus and S. typhimurium on the basis of structural information. ASA values were calculated for Pfb and Stb without PLP. DDGHP, DDGHB, DDGCAV, and DDGENT represent the difference of DG values between Pfb and Stb, due to hydrophobic interaction, hydrogen bond, cavity volume, and entropic effect, respectively. (cid:2)Monomer/dimer(cid:3) represents the values calculated using monomer and dimer forms of b subunit, respectively. The positive value of DG means that the protein from P. furiosus is more stable than the other.

Pfb Stb D(Pfb–Stb)

388 397 )9 Total number of residues ASA value (N-state)

C/S atoms (monomer/dimer) N/O atoms (monomer/dimer) 8095/13 700 A˚ 2 7404/13 353 A˚ 2 7788/13 208 A˚ 2 7259/13 266 A˚ 2 307/419 A˚ 2 145/87 A˚ 2 ASA value (D-state)

C/S atoms (monomer) N/O atoms (monomer) 33 838 A˚ 2 19 935 A˚ 2 34 093 A˚ 2 20 161 A˚ 2 )255 A˚ 2 )226 A˚ 2 DASA value (D–N)

25 743/53 976 A˚ 2 12 531/26 517 A˚ 2 26 305/54 905 A˚ 2 12 902/27 056 A˚ 2 )562/)929 A˚ 2 )371/)539 A˚ 2 C/S atoms (monomer/dimer) N/O atoms (monomer/dimer) Surface area buried at b/b interface

C/S atoms N/O atoms 2490 A˚ 2 1455 A˚ 2 292/595 A˚ 3 2295 A˚ 2 1252 A˚ 2 343/734 A˚ 3 195 A˚ 2 203 A˚ 2 )51/)139 A˚ 3 Cavity volume (monomer/dimer) Secondary structure content (a-helix/b-sheet) 44.3/19.6% 40.3/19.4% 4.0/0.2% Contribution of various factors to the stability

in a turn region, and there is also no sequence identity between residues 360 and 367 of Pfb (Fig. 4). The deviations of the other two peaks I and III are not great, less than about 3 A˚ . These regions are slightly decreased in sequence identity compared with the others.

DSSP [37]. The sequence homology between Pf b and Stb is 58.5%. The alignment indicates that six residues in the N-terminal domain and three residues in the C-terminal domain were deleted in Pfb. Pro366 of Stb and Ile63 of Pfb were inserted in each protein. Fig. 5 shows a schematic stereo view of superimposed b monomer structures of the tryptophan synthase b2 from P. furiosus and S. typhimu- rium. The most different part is an a-helical structure around position 60 of Pfb in place of a turn structure in Stb (an arrow in Fig. 5).

The core region of the N domains of Stb has been reported to have a conformation similar to that of the C domain [3]. To estimate the structural similarity between the N and C domains in Pfb, the RMSD values of the structurally homologous region of the two domains were calculated using 73 Ca pairs corresponding to the residues of Stb, which are reported to deviate by less than 4.0 A˚ between both domains. The values were 2.7 and 2.4 A˚ for Pfb and Stb, respectively. That for Stb was quite similar to that reported (2.2 A˚ ) [3]. As shown in Fig. 3B, the overall topology of the N and C domains in Pfb is similar, and especially, a four-stranded b-sheet structure is well super- imposed. In order to superimpose the C domain on the N domain, the C domain had to be rotated 165.2(cid:1) about an axis and moved by 26.0 A˚ between the centroids of the two domains for Pfb and 160.5(cid:1) and 26.6 A˚ for Stb, respectively. This slight difference might be due to the differences in the structures of the b2 subunit alone and the a2b2 complex, although the complex structure from P. furiosus has not yet been solved.

Structures of N and C domains. The structures of Pfb and Stb (1BKS) could be superimposed with a root mean square deviation (RMSD) of 1.181 A˚ between 385 equivalent Ca atoms in both monomers (Fig. 6). The RMSD values of only the N domain (168 residues) and C domain (185 residues) were 0.596 and 1.003 A˚ , respectively. These results indicate that the structures of both b monomers show a smaller deviation compared with that of the Pfa subunit (RMSD ¼ 2.82 A˚ ) [11], especially for the N domain of the b subunits, because of higher sequence identity. The sequence identities between Pfb and Stb in the N and C domains are 64.5 and 54.1, respectively, while that of Pfa and Sta is 31.5%. As shown in Fig. 6, two large deviations are found in peaks II and IV. In the case of peak II, Ile63 is inserted in Pfb and the region from Lys57 to Ile63 of Pfb clearly forms the a-helix, although the corresponding region of Stb is judged to be in a turn. There is no sequence identity except for one residue (Thr) in this region (Fig. 4). At peak IV, one residue of Pfb at Pro366 of Stb is deleted

Active site. The X-ray crystal structure of the Sta2b2 complex indicates the presence of a 25-A˚ long hydrophobic tunnel connecting a and b active sites through which the metabolic intermediate of the a reaction, indole, would be transferred from the a subunit to the b subunit. The residues

)76.9/)129.1 kJ mol)1 24.7 kJ mol)1 291.0 kJ mol)1 2.7/7.2 kJ mol)1 101.7/127.3 kJ mol)1 DGHP (monomer/dimer) DDGHP b/b interface DDGHB (monomer) DDGCAV (monomer/dimer) DDGENT (–TDS) (25/100 (cid:1)C)

(cid:3) FEBS 2004

Crystal structure of tryptophan synthase b2 subunit alone (Eur. J. Biochem. 271) 2629

PLOT). The distances between the Ca atom of Lys82 and those of the above residues interacted with PLP were calculated and compared with the distance between the corresponding residues of Stb. They agreed within 0.1 A˚ , indicating that the conformation of the active site in Pfb alone is the same as that in Stb in the a2b2 complex.

Complex formation of the b2 subunit with an a sub- unit. Four residues of Lys167, Asn171, Arg175, and Ser178 in the region of Stb, which interact with the a subunit, correspond to Lys162, Asn166, Arg170, and Val173 in Pfb, respectively. The substitution with Val for Ser might not affect a hydrogen bond forming between the a and b subunits because a peptide backbone atom of Ser178 forms a hydrogen bond with the N atom of Gly181 in Sta. From titration calorimetry [38], it has been reported that the formation of the a2b2 complex from subunits of E. coli follows local folding coupled to the subunit association, corresponding to an (cid:2)induced fit(cid:3) with a large conforma- tional change. However, in the case of P. furiosus, the conformational change coupled to the subunit association is slight, resembling a rigid body association [29]. These results suggest that the b subunit from P. furiosus in the complex form might be similar to the structure of the b2 subunit solved in this study. On the other hand, the a and/or b2 subunits from S. typhimurium, which have not yet been solved, might have a much more flexible region than the structure of the a and/or b subunits in the complex form reported.

Discussion

Structure of Pfb2 and mutual activation

Tyr279 and Phe280 in Stb, having a gating function in the tunnel are substituted by Phe274 and His275 in Pfb.

The active site (Lys82) of Pfb corresponding to Lys87 in Stb is located at a flexible region between the two topologically similar domains of the b subunit (Fig. 3A). The phosphate group of PLP covalently bonded with the e- amino group of Lys82 in Pfb is highly ligated through hydrogen bonds with the peptide backbone atoms of residues Gly227, Gly228, Gly229, Ser230, and Ala232 and with the side chains of Ser230 and Asn231, which are likely binding sites for the substrate, L-serine [3]. These residues are completely conserved in Stb, and also the GGGSN sequence is conserved in all the b subunits of the tryptophan synthase reported (protein sequence data bank in a SWISS

Since the first report in 1988 [3], the crystal structures of the a2b2 complex from S. typhimurium have been determined for several forms with bound allosteric ligands [39–45]. These results provide important information for under- standing the allosteric mechanism of the tryptophan synthase complex. The structure of the isolated subunit alone should be determined in order to understand the structural basis of the subunit communication and the allosteric mechanism. In the case of tryptophan synthase from P. furiosus, the structure of the a2b2 complex is not determined yet, although the a [11] and b2 (present work) subunits alone have already been solved. Therefore, we compared the crystal structures of Pfa and Pfb2 alone with those of Sta and Stb2 in the Sta2b2 complex, respectively. The overall structures were quite similar, suggesting that the stimulation effects of enzymatic activities due to a2b2 complex formation are not involved with drastic conform- ational changes. From the isothermal titration calorimetry, the number of residues of local folding coupled to the subunit association in tryptophan synthase from P. furiosus is postulated to be slight, although large conformational changes occur coupled to the subunit association in tryptophan synthase from E. coli [29]. This agrees with present structural results. However, the mechanism of the mutual activation of tryptophan synthase complexes from hyperthermophile and mesophiles could not be understood in detail without the complete set of structures for the two subunits alone and the complex.

Fig. 3. Crystal structure of b2 subunit alone of tryptophan synthase from P. furiosus. (A) The overall structure of the tryptophan synthase b2 dimer from P. furiosus. The N-terminal (1–200) and the C-terminal (201–388) residues are coloured red and blue, respectively. Arrows point to the first two strands and one helical structure (residue 58–64) that intrude into the C domain. The PLP molecule is represented as a CPK model, coloured gold. Drawings were prepared using MOLSCRIPT [71]. (B) Two similar N and C domains of Pfb were superimposed using 69 Ca pairs fitted well among the 73 residues of Stb, which are reported to deviate by less than 4.0 A˚ between both domains [3]. The N and C domains are depicted in gold and green, respectively. Fitting used program LSQKAB [72].

2630 Y. Hioki et al. (Eur. J. Biochem. 271)

(cid:3) FEBS 2004

ð4Þ

DDGHP ¼ aDDASAnonpolar þ bDDASApolar

Stabilization mechanism of Pfb2 on the basis of the structure

where, DDASAnonpolar and DDASApolar represent the differ- ence in the change in accessible surface area (ASA) of nonpolar and polar atoms of all residues in a protein, respectively, upon denaturation between the wild-type and mutant proteins. The parameters a and b have been determined to be 0.154 and )0.026 kJÆmol)1ÆA˚ )2, respect- ively, using the stability/structure database upon denatur- ation of mutant human lysozymes [27]. For calculation of the ASA value, carbon and sulfur atoms in the residues were assigned to ASAnonpolar, and nitrogen and oxygen atoms to ASApolar.

In order to elucidate the stabilization mechanism, the structure of a protein should be analysed in detail, because the conformation of a protein is marginally maintained by many positive and negative factors for stabilization. Using mutant human lysozymes with systematic and comprehen- sive substitutions, changes in stabilities and structures due to mutations have been analysed by DSC and X-ray crystal structures, respectively. It has been proposed that changes in the stability of each mutant human lysozyme are represen- ted by a unique equation, considering the conformational changes due to the mutations [26,27]. The obtained parameters of the relationship between changes in stability and structure should be useful in elucidating the stabiliza- tion mechanism of Pfb on the basis of structural differences between Pfb and Stb.

Hydrophobic interaction. A hydrophobic effect is one of the most important stabilizing forces of a folded structure. The change in unfolding Gibbs energy (DG) due to a hydrophobic effect between the wild-type and mutant proteins (DDGHP) can be expressed as follows:

The contribution of hydrophobic interaction in Pfb and Stb to stabilization was estimated using Eqn 4. The ASA values in the native state were calculated by the procedure of Connolly [46] using the X-ray structures of the two proteins. The values in the denatured forms were estimated using extended structures of each protein, which were generated from the native structures using INSIGHT II. As shown in Table 3, the DGHP values due to hydrophobic interaction of Pfb were less than those of Stb, and the differences between them (DDGHP) were )76.9 and )129.1 kJÆmol)1 in

Fig. 4. Sequence alignments based on secondary structures of the b monomers of tryptophan synthase from P. furiosus and S. typhimurium. The first and sixth lines shown residue numbers of Stb and Pfb, respectively. The second and fifth lines represent secondary structural elements of the Stb subunit (1BKS) and the Pfb subunit, respectively, as judged from the secondary structure definition established by DSSP [37]. H, E, B, G, T, and S in the secondary structure elements represent the a-helix, b-strand, b-bridge, 3-helix, turn, and bend, respectively. The third and fourth lines represent the amino acid sequences of Stb and Pfb, respectively.

(cid:3) FEBS 2004

Crystal structure of tryptophan synthase b2 subunit alone (Eur. J. Biochem. 271) 2631

Fig. 5. Schematic stereo view of the superimposed b monomer structures of the tryptophan synthase b2 from P. furiosus and S. typhimurium. Blue and red lines represent the coordinates of Pfb and Stb (1BKS), respectively. Drawings were prepared using MOLSCRIPT [71]. Residual numbers are shown with an increase of 10 for the Pfb. An arrow indicates the most different part between the proteins around position 60 of Pfb.

interaction compared with that of Stb. It has been reported that subunit–subunit interaction and higher order organ- ization contribute to the enhanced stability of hyperthermo- phile proteins [13,47].

a monomeric form and a dimeric form, respectively. This means that the higher stability of Pfb is not caused by the hydrophobic interaction. The number of hydrophobic amino acid residues of Pfb was slightly decreased compared with that of Stb (Table 2). This trend has been observed in the comparison of the a subunit of tryptophan synthase from P. furiosus with that from S. typhimurium [11]. The hydrophobic effects at the interface of the b/b interaction were also examined. DDGHP at the interface between Pfb2 and Stb2 was 24.7 kJÆmol)1, indicating that the subunit interaction of Pfb is more stabilized due to hydrophobic

Ion pairs (salt bridges) and hydrogen bonds. Ion pairs (salt bridges) seem to play important roles in the stabiliza- tion of hyperthermophile proteins because they occur frequently in hyperthermophile proteins [11,14,17–19,22, 48–53]. Table 4 lists the numbers of ion pairs for Pfb2 and Stb2. The number of ion pairs in Pfb was less than that in

Fig. 6. RMSDs in Ca atoms between Pf b and Stb after a least-squares fit of the corresponding Ca atoms. The residue number represents the value for Pfb. I to IV represents a discrimination mark for large differences.

2632 Y. Hioki et al. (Eur. J. Biochem. 271)

(cid:3) FEBS 2004

Table 4. Number of ion pair and hydrogen bond in the tryptophan synthase b2 subunit from P. furiosus and S. typhimurium.

cavity volume (52 JÆmol)1ÆA˚ )3) [27]. Using this parameter, the increment in stabilization of the Pfb in a dimer state due to the decrease in cavity volume could be calculated to be 7.2 kJÆmol)1 (Table 3), compared with that of the Stb.

Pfb Stb D(Pfb–Stb)

Number of ion pairs (monomer)

17 42 73. 18 46 72 )1 )4 1 (cid:1) 3 A˚ (cid:1) 4 A˚ (cid:1) 5 A˚ Number of ion pairs at b/b interface

0 8 14 2 6 16 )2 2 )2 (cid:1) 3 A˚ (cid:1) 4 A˚ (cid:1) 5 A˚

Number of hydrogen bonds within 3.2 A˚

Entropic effect. An entropic effect is one of the important stabilizing factors (DG ¼ DH ) TDS). When the conform- ational entropy of a protein is decreased in the denatured state due to substitution(s) or deletion(s) of an amino acid residue, the stability is increased. We can calculate the entropic effects of denaturation from the amino acid compositions using thermodynamic parameters proposed by Oobatake and Ooi [60]: the denaturation entropies for Pfb and Stb were 1.00 and 1.34 kJÆmol)1ÆK)1, respectively. This indicates that Pfb is stabilized by 101.7 kJÆmol)1 at 25 (cid:1)C and 127.3 kJÆmol)1 at 100 (cid:1)C due to its entropic effect (Table 3).

284 405 256 371 28 34

Stb, although the number of the charged (hydrophilic) residues in Pfb was higher than that in Stb (Table 2). The number of ion pairs at the b/b interface was also less in Pfb than in Stb. These results suggest that the higher stability of Pfb2 is not caused by the increase in charged residues.

Aromatic–aromatic interaction. Aromatic–aromatic inter- action of the side chains of Phe, Tyr, or Trp has been reported to contribute to the conformational stability of a protein [61]. In the case of the small ribonuclease from Bacillus amyloliquefacience, the edge of the aromatic ring of Tyr17 interacts with the face of that of Tyr13, and the interaction energy is estimated to be )5.4 kJÆmol)1 using the double-mutant cycle analysis [62]. As shown in Table 2, the numbers of aromatic residues of Pfb increase by two residues for Trp and by five for Tyr compared with those of Stb. However, there seem to be no aromatic–aromatic interaction with suitable angles to contribute to the stabilization in Pfb.

Many studies of mutant proteins connected with hydro- gen bonds have shown that hydrogen bonds contribute to stabilizing the conformation of a protein [54–57]. The number of hydrogen bonds involved in the main chains of Pfb was greater by about 10% than that of Stb (Table 4). This increase in Pfb mainly comes from the extra a-helix (Helix 2¢) from Lys57 to Ile63 and the extension of the a-helix in Leu344–Ser346 (Fig. 4). The net contribution of intramolecular hydrogen bonds has been estimated to be 8.56 kJÆmol)1 for a 3 A˚ hydrogen bond [27]. Using this parameter, the contribution due to hydrogen bonds (of the main and side chains; Table 4) to the stability of Pfb was estimated to be greater by 291 kJÆmol)1 than that of Stb (Table 3). This suggests that hydrogen bonds remarkably contribute to enhancing the stability of Pfb.

Analysis by knowledge-based potential. Using the know- ledge-based potential derived from PDB, several methods have been developed to estimate the stability of mutant proteins [63–65]. These methods are computationally rapid and look very robust for the parameters of use. Correlations between the experiment and calculations are expected within 0.5–0.9 depending on the samples. A method developed by Ota et al. [65] estimates the changes in conformational stability due to all of the single amino acid substitutions and represents them in SPMP (Stability Profiles of Mutant Protein). A pseudo-energy potential (DDGSPMP) consisting of four elements is used: side-chain packing (DDGSP), hydration (DDGHyd), local structure (DDGLC) and backbone side-chain repulsion (DDGBR) [66]:

Further extensive analyses of the electrostatic interaction, i.e. solving the Poisson–Boltzmann equation may be promising [58]. However the results appear to be very sensitive to the dielectric constant, other parameters used and the assumed denatured state. Hence we should leave this for the future work.

DDGSPMP ¼ DDGSP þ DDGHyd þ DDGLC þ DDGBR ð5Þ

sites of enzymes

This method had been applied to the mutants of several proteins, e.g. Ribonuclease HI [65], human lysozyme [67,68], as well as the evaluating structure–sequence com- patibility, i.e. threading [69]. Recently, locating the func- tional to identify the structurally destabilizing residues was included in the method [70].

The conformational stabilities of both b subunit struc- tures were analysed by SPMP. The individual stability scores for four terms are summarized in Table 5. The total score of the b subunit from the hyperthermophile clearly shows higher stability than that from the mesophile. All of the score terms contribute to the higher stabilization of Pfb in a monomer state. The scores for a dimeric form in both

Cavity volume. Changes in the cavity size in the interior of a protein affect the conformational stability [59]. Therefore, the cavity volume was determined by attempting to insert a probe sphere of radius 1.4 A˚ (assuming a water molecule) [46]. In the case of Pfb2 in a dimer state, 19 cavities were found and the total volume was 595 A˚ 3 (Table 3). These cavities with a small volume were distributed throughout the molecule. In the case of Stb2, 14 cavities were found with a total volume of 734 A˚ 3 which included two newly intro- duced cavities (total: 48 A˚ 3) when associated. The cavity volumes for the monomer and the dimer were lower in Pfb2 than in Stb2, suggesting a more rigid packing of the Pfb2 molecule. The energy term for protein stability (DG) due to changes in the cavity size can be expressed in terms of the

6 14 )8 Main chain (monomer) Main chain and side chain (monomer) Number of hydrogen bonds at b/b interface

(cid:3) FEBS 2004

Crystal structure of tryptophan synthase b2 subunit alone (Eur. J. Biochem. 271) 2633

Table 5. SPMP stability scores of a monomer and a dimer of the b subunit from P. furiosus and S. typhimurium: stability scores. Units are kJÆmol)1. Positive values show stabilization in Pfb and Pfb2. Total, SP, Hyd, LC, and BR represent the value of DDGSPMP, DDGSP, DDGHyd, DDGLC, and DDGBR, respectively, for each subunit.

of both proteins are more stable than the monomeric forms (Table 5). Only the contributions of hydrogen bonding and ion pairs were comparable. As shown in Fig. 1, the denaturation temperatures of Ecb2 and Stb2 from meso- philes were considerably high, 80 (cid:1)C around pH 8.0, although that of Pfb2 is higher, 115 (cid:1)C. The dimeric forms strongly contribute to the higher thermal stability in the case of the b subunits from both mesophiles and hypertherm- ophiles. The unusual stability of Pfb2 might be caused by the contribution of the intensive b/b subunit interaction in addition to the enhanced stability in a monomeric form.

Total SP Hyd LC BR

Monomer Pfb Stb Difference(Pf-St) 728.6 670.1 58.5 584.6 535.0 49.6 55.7 52.0 3.7 206.8 204.9 1.9 )118.6 )121.4 2.8

Conclusions

Dimer Pfb2 Stb2 Difference(Pf-St) 765.0 698.0 67.0 621.1 559.8 61.3 63.7 58.9 4.8 206.8 204.9 1.9 )126.9 )125.5 )1.4

Table 6. SPMP stability scores of a monomer and a dimer of the b subunit from P. furiosus and S. typhimurium: average ranking order. 20 aa and 56 rotamers mean the ranking order out of 20 amino acids and 56 kinds of rotamers, respectively.

The structure of the tryptophan synthase b2 subunit (Pfb2) from the hyperthermophile, Pyrococcus furiosus, was deter- mined by X-ray crystallographic analysis at 2.2 A˚ resolu- tion, which was the first report of the X-ray structure of the tryptophan synthase b2 subunit alone, although the struc- the tryptophan synthase a2b2 complex from ture of Salmonella typhimurium has already been reported. The structure of Pfb2 was essentially similar to that of the b2 subunit in the a2b2 complex from S. typhimurium.

Monomer Dimer

20 aa 56 rotamers 20 aa 56 rotamers

proteins were higher than those in a monomeric form, coinciding with experimental results showing that both proteins stably exist as a dimeric form in solution.

Stability was examined by DSC. Denaturation temper- atures above pH 6.5 are around 115 (cid:1)C independent of pH; this is about 35 (cid:1)C higher than those reported for mesophilic proteins. On the basis of structural information on Pfb and Stb, it could be concluded that: (a) the higher stability of Pfb is not caused by either a hydrophobic interaction or an increase in ion pairs; (b) the number of hydrogen bonds involved in the main chains of Pfb (monomeric form) is greater by about 10% than that of Stb, and the contribution due to hydrogen bonds (of the main and side chains) to the stability of Pfb was estimated to be greater by 291 kJÆmol)1 than that of Stb, suggesting that hydrogen bonds remarkably contribute to enhancing the stability of Pfb; (c) the dimeric form of Pfb is stabilized due to hydrophobic interaction and a decrease in cavity volume at the b/b interface; and (d) in total, the sequence of Pfb seems to be more fitted to an ideally stable structure (lower energy level) than that of Stb, as judged from X-ray structure data.

pf b Stb Difference(Pf–St) 5.47 5.88 )0.41 9.2 10.17 )0.97 5.33 5.76 )0.43 8.82 9.85 )1.03

References

SPMP provides stability scores for each residue at every site of an amino acid sequence. In the case of Pfb (388 residues solved by X-ray analysis), the DDG value for 388 · 19 mutants can be predicted by SPMP using the crystal structure, resulting in ranking of the native residues of the Pfb. The average ranking of all native residues among 20 amino acids for Pfb and Stb was 5.47 and 5.88, respectively, in a monomeric form, and 5.33 and 5.76, respectively, in a dimeric form (Table 6), indicating that the hyperthermophile protein in both monomer and dimeric forms adopts (selects) the residues with lower ranking. The average ranking among 56 rotamers [66] including side- chain conformations of both proteins also showed the same trend (Table 6). These results indicate that the conformation of a hyperthermophile protein (Pfb) is more fitted to an ideal structure (lower energy level) than that of a mesophilic protein (Stb).

1. Yanofsky, C. & Crawford, I.P. (1972) Tryptophan synthase. In The Enzymes (Boyer, P.D., ed.), 3rd edn, pp. 1–31. Academic Press, New York. 2. Miles, E.W. (1995) Tryptophan synthase. Structure, function, and protein engineering. Subcell. Biochem. 24, 207–254.

(1988) Three-dimensional structure of

3. Hyde, C.C., Ahmed, S.A., Padlan, E.A., Miles, E.W. & Davies, D.R. the tryptophan synthase a2b2 multienzyme complexfrom Salmonella typhimurium. J. Biol. Chem. 263, 17857–17871.

4. Ruvinov, S.B. & Miles, E.W. (1994) Thermal inactivation of tryptophan synthase: Stabilization by protein–protein interaction and protein–ligand interaction. J. Biol. Chem. 269, 11703–11706. 5. Fan, Y.-X., McPhie, P. & Miles, E.W. (2000) Thermal repair of tryptophan synthase mutations in a regulatory intersubunit salt bridge. J. Biol. Chem. 275, 20302–20307.

Dimeric form of Pfb. Pfb has been reported to exist in a dimeric form in solution like prokaryotic tryptophan synthase b subunit from mesophiles [29]. The surface areas buried at the b/b interface of Pfb for C/S and N/O atoms were increased by 195 and 203 A˚ 2, respectively, com- pared with that of Stb, indicating that the b/b interface of Pfb is more stabilized by 24.7 kJÆmol)1 due to hydrophobic interaction (Table 3). As shown in Table 3, the decrease in the cavity volume at the b/b interface contributes to the stabilization of the dimeric form of Pfb. Stability profiles of mutant protein analyses also suggest that the dimeric forms

6. Pan, P., Woehl, E. & Dunn, M.F. (1997) Protein architecture, dynamics and allosteryin tryptophan synthase channeling. TIBS 22, 22–27.

2634 Y. Hioki et al. (Eur. J. Biochem. 271)

(cid:3) FEBS 2004

dehydrogenase monomers. Proteins Struct. Func. Gen. 38, 368–383. 7. Creighton, T.E. & Yanofsky, C. (1966) Association of the alpha and beta-2 subunits of the tryptophan synthetase of Escherichia coli. J. Biol. Chem. 241, 980–990.

23. Szilagyi, A. & Zavodszky, P. (2000) Structural differences between mesophilic, moderately thermophilic and extremely thermophilic protein subunits: results of acomprehensive survey. Struct. Fold Des. 8, 493–504. 8. Wiesinger, H., Bartholmes, P. & Hinz, H.J. (1979) Subunit inter- action in tryptophansynthase of Escherichia coli: calorimetric studies on association of alpha and beta 2 subunits. Biochemistry 18, 1979–1984.

3-isopropylmalate dehydrogenase of 9. Lane, A.N., Paul, C.H. & Kirschner, K. (1984) The mechanism of self-assembly of the multi-enzyme complex tryptophan synthase from Escherichia coli. EMBO J. 3, 279–287.

24. Kirino, H., Aoki, M., Aoshima, M., Hayashi, Y., Ohba, M., Yamagishi, A., Wakagi, T. & Oshima, T. (1994) Hydrophobic interaction at the subunit interface contributes to the thermo- stability from an extremethermophile, Thermus thermophilus. Eur. J. Biochem. 220, 275–281.

10. Ogasahara, K., Hiraga, K., Ito, W., Miles, E.W. & Yutani, K. (1992) Origin of the mutual activation of the a and b2 subunits in the a2b2 complex of tryptophansynthase. Effect of alanine or glycine substitutions at proline residues in the a subunit. J. Biol. Chem. 267, 5222–5228.

25. Auerbach, G., Ostendorp, R., Prade, L., Korndorfer, I., Dams, T., Huber, R. & Jaenicke, R. (1998) Lactate dehydrogenase from the hyperthermophilic bacterium Thermotoga maritima: the crystal structure at 2.1. A resolution reveals strategies forintrinsic protein stabilization. Structure 6, 769–781.

11. Yamagata, Y., Ogasahara, K., Hioki, Y., Lee, S.J., Nakagawa, A., Nakamura, H., Ishida, M., Kuramitsu, S. & Yutani, K. (2001) Entropic stabilization of thetryptophan synthase a-subunit from a hyperthermophile, Pyrococcus furiosus: X-ray analysis and calori- metry. J. Biol. Chem. 276, 11062–11071. 26. Funahashi, J., Takano, K., Yamagata, Y. & Yutani, K. (1999) Contribution of aminoacid substitutions at two different interior positions to the stability of humanlysozyme. Protein Eng. 12, 841–850.

27. Funahashi, J., Takano, K. & Yutani, K. (2001) Are the para- meters of variousstabilization factors estimated from mutant human lysozymes compatible with otherproteins? Protein Eng. 14, 127–134. 12. Ogasahara, K., Nakamura, M., Nakura, S., Tsunasawa, S., Kato, I., Yoshimoto, T. & Yutani, K. (1998) Unusual slow unfolding rate causes the high stability ofpyrrolidone carboxyl peptidase from a hyperthermophile, Pyrococcus furiosus: Equili- brium and kinetic studies of guanidine hydrochloride-induced unfolding and refolding. Biochemistry 37, 17535–17544.

28. Ishida, M., Oshima, T. & Yutani, K. (2002) Overexpression in Escherichia coli ofthe AT-rich trpA and trpB genes from the hyperthermophilic archaeon Prococcus fruriosus. FEMS Microbiol. Lett. 216, 179–183. 13. Ogasahara, K., Khechinashvili, N.N., Nakamura, M., Yoshi- moto, T. & Yutani, K. (2001) Thermal stability of pyrrolidone carboxyl peptidases from hyperthermophilicarchaeon, Pyrococcus furiosus. Eur. J. Biochem. 268, 3233–3242.

29. Ogasahara, K., Ishida, M. & Yutani, K. (2003) Stimulated inter- action between a andb subunits of tryptophan synthase from hyperthermophile enhances its thermalstability. J. Biol. Chem. 278A, 8922–8928. 14. Tahir, H.T., Oki, H., Tsukihara, T., Ogasahara, K., Yutani, K., Ogata, K., Izu, Y., Tsunasawa, S. & Kato, I. (1998) Crystal structure of methionine aminopeptidase fromhyperthermophile, Pyrococcus furiosus. J. Mol. Biol. 284, 101–124.

15. Backmann, J. & Schafer, G. (2001) Thermodynamic analysis of hyperthermostableoligomeric proteins. Methods Enzymol. 334, 328–342. 16. Rees, D. (2001) Crystallographic analyses of hyperthermophilic 30. Yutani, K., Ogasahara, K., Tsujita, T., Kanemoto, K., Matsu- moto, M., Tanaka, S., Miyashita, T., Matsushiro, A., Sugino, Y. & Miles, E.W. (1987) Tryptophansynthase a subunit glutamic acid 49 is essential for activity. Studies with 19 mutantsat position 49. J. Biol. Chem. 262, 13429–13433. proteins. Methods Enzymol. 334, 423–437.

31. Zhao, G.-P. & Somerville, R.L. (1992) Genetic and biochemical characterization ofthe trpB8 mutation of Escherichia coli trypto- phan synthase. An amino acid switch atthe sharp turn of the trypsin-sensitive (cid:2)hinge(cid:3) region diminishes substrate binding and- alters solubility. J. Biol. Chem. 267, 526–541. 17. Tanaka, H., Chinami, M., Mizushima, T., Ogasahara, K., Ota, M., Tsukihara, T. & Yutani, K. (2001) X-ray crystalline structures of pyrrolidone carboxyl peptidase from a hyperthermophile, Pyrococcus furiosus, and its Cys-free mutant. J. Biochem. Tokyo 130, 107–118.

32. Ogasahara, K., Yutani, K., Suzuki, M., Sugino, Y., Nakanishi, M. & Tsuboi, M. (1980) State of Tyr49 in a mutant tryptophan synthase alpha-subunit substituted atposition 49. J. Biochem. (Tokyo) 88, 1733–1738.

33. Miles, E.W., Bauerle, R. & Ahmed, S.A. (1987) Tryptophan synthase from Escherichia coli and Salmonella typhimurium. Methods Enzymol. 142, 398–414.

18. Hennig, M., Darimont, B., Sterner, R., Kirschner, K. & Janso- nius, J.N. (1995) 2.0 A˚ structure of indole-3-glycerol phosphate synthase from the hyperthermophile Sulfolobus solfataricus: pos- sible determinants of protein stability. Structures 3, 1295–1306. 19. Knapp, S., de Vos, W.M., Rice, D. & Ladenstein, R. (1997) Crystal from the structure of glutamate dehydrogenase hyperthermophilic eubacterium Thermotogamaritima at 3.0 A˚ resolution. J. Mol. Biol. 267, 916–932. 34. Otwinowski, Z. & Minor, W. (1997) Processing of X-ray diffrac- tion data collected inoscillation mode. Methods Enzymol. 276, 307–326.

20. Hennig, M., Sterner, R., Kirschner, K. & Jansonius, J.N. (1997) Crystal structure at 2.0 A˚ resolution of phosphoribosyl anthrani- late isomerase from the hyperthermophile Thermotoga maritima: Possible determinants of protein stability. Biochemistry 3, 6009– 6016. 35. Remeta, D. P., Miles, W. W. & Ginsburg, A. (1995) Thermally induced unfoldingof the tryptophan synthase a2b2 multienzyme complex from Salmonella typhimurium. Pure Appl. Chem. 67, 1859–1866.

36. Fan, Y.-X., McPhie, P. & Miles, E.W. (2000) Regulation of tryptophan synthase bytemperature, monovalent cations, and allosteric ligand. Evidence from Arrhenius plots, absorption spectra, and primary kinetic isotope effect. Biochemistry 39, 4692– 4703. 21. Britton, K.L., Yip, K.S., Sedelnikova, S.E., Stillman, T.J., Adams, M.W., Ma, K., Maeder, D.L., Robb, F.T., Tolliday, N., Vetriani, C., Rice, D.W. & Baker, P.J. (1999) Structure determination of the glutamate dehydrogenase from thehyperthermophile Thermo- coccus litoralis and its comparison with that from Pyrococcus furiosus. J. Mol. Biol. 293, 1121–1132.

22. Kumar, S., Ma, B., Tsai, C.-J. & Nussinov, R. (2000) Electrostatic strengths of saltbridges in thermophilic and mesophilic glutamate 37. Kabsch, W. & Sander, W. (1983) Dictionary of protein secondary structure: pattern recognition of hydrogen-bonded and geo- metrical features. Bioplolymers 22, 2577–2637.

(cid:3) FEBS 2004

Crystal structure of tryptophan synthase b2 subunit alone (Eur. J. Biochem. 271) 2635

54. Shirley, B.A., Stanssens, P., Hahn, U. & Pace, C.N. (1992) Con- tribution of hydrogen bonding to the conformational stability of ribonuclease T1. Biochemistry 31, 725–732. 38. Hiraga, H. & Yutani, K. (1996) Thermodynamic analysis of conformational changedue to the a2b2 complex formation of tryptophan synthase. Eur. J. Biochem. 240, 63–70.

55. Byrne, M.P., Manuel, R.L., Lowe, L.G. & Stites, W.E. (1995) Energeticcontribution of side chain hydrogen bonding to the stability of staphylococcalnuclease. Biochemistry 34, 13949– 13960. 39. Rhee, S., Parris, K.D., Ahmed, S.A., Miles, E.W. & Davies, D.R. (1996) Exchange of K+ or Cs+ for Na+ induces local and long- range changes in thethree-dimensional structure of the tryptophan synthase a2b2 complex. Biochemistry 35, 4211–4221.

to the conformational

56. Yamagata, Y., Kubota, M., Sumikawa, Y., Funahashi, J., Tak- ano, K., Fujii, S. & &Yutani, K. (1998) Contribution of hydrogen bonds stability ofhuman lysozyme: Calorimetry and X-ray analysis of six Tyr fi Phe mutants. Biochemistry 37, 9355–9362. 40. Rhee, S., Parris, K.D., Hyde, C.C., Ahmed, S.A., Miles, E.W. & Davies, D.R. (1997) Crystal structures of a mutant (bK87T) tryptophan synthase a2b2 complex with ligands bound to the active sites of the a- and b-subunits reveal ligand-induced con- formationa changes. Biochemistry 36, 7664–7680.

57. Takano, K., Yamagata, Y., Funahashi, J., Hioki, Y., Kuramitsu, K. & Yutani, K. (1999) Contribution of intra- and intermolecular hydrogen bonds to theconformational stability of human lyso- zyme. Biochemistry 38, 12698–12708.

41. Schneider, T.R., Gerhardt, E., Lee, M., Liang, P.-H., Anderson, K.S. & Schlichting, I. (1998) Loop closure and intersubunit com- munication in tryptophan synthase. Biochemistry 37, 5394–5406. 42. Weyand, M. & Schlichting, I. (1999) Crystal structure of wild-type tryptophansynthase complexed with the natural substrate indole- 3-glycerl phosphate. Biochemistry 38, 16469–16480. 58. Xiao, L. & Honig, B. (1999) Electrostatic contributions to the stability ofhyperthermophilic proteins. J. Mol. Biol. 289, 1435– 1444.

43. Weyand, M., Schlichting, I., Marabotti, A. & Mozzarelli, A., (2002) Crystal structures of anew class of allosteric effectors complexed to tryptophan synthase. J. Biol. Chem. 277, 10647– 10652. 59. Eriksson, A.E., Baase, W.A., Zhang, X.J., Heinz, D.W., Blaber, M., Baldwin, E.P. & Matthews, B.W. (1992) Response of a protein structure to cavity-creating mutations and its relation to the hydrophobic effect. Science 255, 178–183.

60. Oobatake, M. & Ooi, T. (1993) Hydration and heat stability effects on protein unfolding. Prog. Biophys. Molec. Biol. 59, 237–284. 61. Burley, S.K. & Petsko, G.A. (1988) Weakly polar interactions in 44. Weyand, M. & Schlichting, I. (2000) Structural basis for the impaired channelingand allosteric inter-subunit communication in the bA169I/bC170W mutant of tryptophan synthase. J. Biol. Chem. 275, 41058–41063. proteins. Adv. Protein Chem. 39, 125–189.

62. Serrano, L., Bycroft, M. & Fersht, A.R. (1991) Aromatic– aromatic interactions andprotein stability investigation by double- mutant cycles. J. Mol. Biol. 218, 465–475. 45. Kulik, V., Weyand, M., Seidel, R., Niks, D., Arac, D., Dunn, M.F. & Schlichting, I. (2002) On the role of aThr183 in the allosteric regulation and catalytic mechanism oftryptophan synthase. J. Mol. Biol. 324, 677–690. 46. Connolly, M.L. (1993) The molecular surface package. J. Mol. Graphics 11, 139–141. 63. Gilis, D. & Rooman, M. (1996) Stability changes upon mutations of solvent-accessible residues in proteins evaluated by database- derived potentials. J. Mol. Biol. 257, 1112–1126.

64. Miyazawa, S. & Jernigan, R.L. (1994) Protein stability for single substitution mutants and the extent of local compactness in the denatured state. Protein Eng. 7, 1209–1220.

65. Ota, M., Kanaya, S. & Nishikawa, K. (1995) Desk-top analysis of the structural stability of various point mutations introduced into ribonuclease H. J. Mol. Biol. 248, 733–738. 47. Jaenicke, R. & Bohm, G. (1998) The stability of proteins in extreme environments. Curr. Opin. Struct. Biol. 8, 738–748. 48. Day, M.W., Hsu, B.T., Joshua-Tor, L., Park, J.-B., Zhou, Z., Adams, M.W.W. & &Ree, D.C. (1992) X-ray crystal structures of the oxidized and reduced forms of therubredoxin from the marine hyperthermophilic archaebacterium Pyrococcus furiosus. Protein Sci. 1, 1494–1507.

ferredoxinoxidoreductase. Science aldehyde 66. Ota, M., Isogai, Y. & Nishikawa, K. (2001) Knowledge-based potential defined for a rotamer library to design protein sequences. Protein Eng. 14, 557–564. 49. Chan, M.K., Mukund, S., Kletzin, A., Adams, M.W.W. & Rees, (1995) Structure of a hyperthermophilic tungstopterin 267, D.C. enzyme, 1463–1469.

67. Takano, K., Ota, M., Ogasahara, K., Yamagata, Y., Nishikawa, K. & Yutani, K. (1999) Experimental verification of the (cid:2)stability profile of mutant protein(cid:3) (SPMP) data using mutant human lysozymes. Protein Eng. 12, 663–672.

50. Yip, K.S.P., Stillman, T.J., Britton, K.L., Artymiuk, P.J., Baker, P.J., Sedelnikova, S.E., Engel, P.C., Pasquo, A., Chiaraluce, R., Consalvi, V., Scandurra, R. & Rice, D.W. (1995) The structure of Pyrococcus furiosus glutamate dehydrogenase reveals a key role for ion-pair networks in maintaining enzyme stability at extreme temperatures. Structures 3, 1147–1158.

68. Yutani, K., Takano, K. & Funahashi, J. (1999) Evaluation of some factors that contribute to conformational stability of a protein using database of stability/structure. In Old and New Views of Protein Folding (Kuwajima, K. & Arai, M., eds), Elsevier Science, Amsterdam. crystal 69. Ota, M. & Nishikawa, K. (1999) Feasibility in the inverse protein folding protocol. Protein Sci. 8, 1001–1009. 51. DeDecker, B.S., O’Brien, R., Fleming, P.J., Geiger, J.H., Jackson, structure of a (1996) The S.P. & Sigler, P.B. hyperthermophilic archaeal TATA-box bindingprotein. J. Mol. Biol. 264, 1072–1084.

70. Ota, M., Kinoshita, K. & Nishikawa, K. (2003) Prediction of catalytic residues in enzymes based on known tertiary structure, stability profile, and sequence conservation. J. Mol. Biol. 327, 1053–1064. 52. Russell, R.J., Ferguson, J.M., Hough, D.W., Danson, M.J. & Taylor, G.L. (1997) The crystal structure of citrate synthase from the hyperthermophilic archaeon pyrococcus furiosus at 1.9 A resolution. Biochemistry 36, 9983–9994.

the glyceraldehyde-3-phosphatedehydrogenase 71. Kraulis, P.J. (1991) MOLSCRIPT: a program to produce both detailed and schematic plots of protein structures. J. Appl. Crys- tallogr. 24, 946–950. 72. Kabsch, W. (1996) A solution for the best rotation to relate two sets of vectors. Acta. Crystallogr. A32, 922–923. 53. Isupov, M.N., Fleming, T.M., Dalby, A.R., Crowhurst, G.S., (1999) Crystal structure Bourne, P.C. & &Littlechild, J.A. of from the hyperthermophilic archaeon Sulfolobus solfataricus. J. Mol. Biol. 291, 651–660.