- Email: [email protected]
The First Crystal Structure of l -Threonine Dehydrogenase
doi:10.1016/j.jmb.2006.11.060
J. Mol. Biol. (2007) 366, 857–867
The First Crystal Structure of L-Threonine Dehydrogenase Kazuhiko Ishikawa 1 ⁎, Noriko Higashi 1 , Tsutomu Nakamura 1 Takanori Matsuura 2 and Atsushi Nakagawa 2 ⁎ 1
National Institute of Advanced Industrial Science and Technology (AIST), 1-8-31 Midorigaoka, Ikeda, Osaka 563-8577, Japan 2
Institute for Protein Research, Osaka University, 3-2 Yamadaoka, Suita, Osaka 565-0871, Japan
L-Threonine
dehydrogenase (TDH) is an enzyme that catalyzes the oxidation of L-threonine to 2-amino-3-ketobutyrate. We solved the first crystal structure of a medium chain L-threonine dehydrogenase from a hyperthermophilic archaeon, Pyrococcus horikoshii (PhTDH), by the single wavelength anomalous diffraction method using a selenomethioninesubstituted enzyme. This recombinant PhTDH is a homo-tetramer in solution. Three monomers of PhTDHs were located in the crystallographic asymmetric unit, however, the crystal structure exhibits a homo-tetramer structure with crystallographic and non-crystallographic 222 symmetry in the cell. Despite the low level of sequence identity to a medium-chain NAD (H)-dependent alcohol dehydrogenase (ADH) and the different substrate specificity, the overall folds of the PhTDH monomer and tetramer are similar to those of the other ADH. Each subunit is composed of two domains: a nicotinamide cofactor (NAD(H))-binding domain and a catalytic domain. The NAD(H)-binding domain contains the α/β Rossmann fold motif, characteristic of the NAD(H)-binding protein. One molecule of PhTDH contains one zinc ion playing a structural role. This metal ion exhibits coordination with four cysteine ligands and some of the ligands are conserved throughout the structural zinc-containing ADHs and TDHs. However, the catalytic zinc ion that is coordinated at the bottom of the cleft in the case of ADH was not observed in the crystal of PhTDH. There is a significant difference in the orientation of the catalytic domain relative to the coenzyme-binding domain that results in a larger interdomain cleft. © 2006 Elsevier Ltd. All rights reserved.
*Corresponding authors
Keywords: archaea; Pyrococcus horikoshii; threonine dehydrogenase; alcohol dehydrogenase; X-ray crystallography
Introduction A hyperthermophilic anaerobic archaeon, Pyrococcus horikoshii, was isolated from a hydrothermal volcanic vent in the Okinawa Trough in the Pacific Ocean.1 An open-reading frame (ORF) (PH0655: GenBank accession no. O58389) in the P. horikoshii genome database has been classified as an LPresent address: N. Higashi, Graduate School of Engineering Science, Osaka University, 1-3 Machikaneyama, Toyonaka, Osaka 560-8531, Japan. Abbreviations used: TDH, L-Threonine dehydrogenase; ADH, alcohol dehydrogenase; Ph, Pyrococcus horikoshii; Ec, Escherichia coli; Bs, Bacillus subtilis; Hl, horse liver; Ss, Sulfolobus solfataricus; Ap, Aeropyrum pernix. E-mail addresses of the corresponding authors: [email protected]; [email protected]
threonine dehydrogenase (TDH; EC 1.1.1.103), 2 with sequence homology to an alcohol dehydrogenase (ADH; EC 1.1.1.1). We have cloned PH0655 and found it to be a hyperthermophilic TDH of P. horikoshii (PhTDH).3 TDH can catalyze the NAD (H)-dependent oxidation of L-threonine to L-2amino-3-ketobutyrate4–7 and requires zinc for its catalytic activity. PhTDH was classified as a member of the medium-chain, NAD(H)-dependent, and zinccontaining alcohol/polyol dehydrogenase family.3,8 PhTDH has a broad substrate specificity and is expected to be useful for the organic synthesis of several chiral compounds (L-2-amino-3-ketobutyrate, L-2-amino-3-ketopropionate and their carbonic acid derivatives). However, there has been no report for the catalytic mechanism of TDH. Many studies on the structure of ADHs have been carried out and the catalytic mechanism has been discussed in detail. ADHs are divided according to
0022-2836/$ - see front matter © 2006 Elsevier Ltd. All rights reserved.
858 their size and cofactor (NAD(H) or NADP(H)) requirement.9 In particular, medium-chain NAD (H)-dependent ADHs mostly containing two zinc atoms per subunit are widely distributed, being found in bacteria, eukarya and archaea. Between medium-chain ADHs and TDHs, some sequence identity and a conserved sequence region have been observed. The detailed crystal structures for the homo-tetrameric structures of ADHs from hyperthermophilic archaea Sulfolobus solfataricus and Aeropyrum pernix have also been reported.10–12 The sequence of PhTDH is homologous to the hyperthermophilic archaeon ADH's from S. solfataricus (29.5 % identity) (SsADH) and A. pernix (29.1 % identity) (ApADH). The PhTDH gene has been cloned and the recombinant enzyme overexpressed in Escherichia coli, and its characteristics have been reported.3 However, there has been no report on the crystal structure of TDHs. Analysis of the structural and functional similarities between TDH and ADH will increase our understanding of the different functions of these enzymes towards the structural origin
Crystal Structure of Threonine Dehydrogenase
of the hyperthermostable protein in this group of archaea. Furthermore, the structure of the active site and the catalytic residues of TDH are informative. Here, we prepare a crystal of PhTDH (SePhTDH) using the selenomethionine-labelled protein13 and report its crystal structure.
Results and Discussion In the genome database of P. horikoshii, a TDH homologue gene was found.2,3 The gene product (PhTDH) was composed of 348 amino acid residues (1044 base-pairs), being homologous with TDHs from E. coli (EcTDH; GenBank accession no. P07913, 44.0% identity) and Bacillus subtilis (BsTDH; GenBank accession no. O31776, 49.1% identity), and ADHs from horse liver (HlADH; GenBank accession no. P00327, 28.3% identity), S. solfataricus (SsADH; GenBank accession no. P39462, 29.5% identity), and A. pernix (ApADH; GenBank accession no. O9Y9P9, 29.1% identity), as shown in Figure 1. PhTDH was
Figure 1. Alignment of amino acids compared among TDHs and ADHs from various origins. The abbreviations are as follows: Ph, P. horikoshii (PH0655); Ec, E. coli (ORF no. b3616); Bs, B. subtilis (BSU16990); Hl, horse liver; Ss, S. solfataricus; and Ap, A. pernix. Identical residues are shown by asterisks. In the PhTDH sequence (upper row), the main secondary structure is shown: residues highlighted in blue and red (or green) are β-strands and α-helices, respectively. NAD(H) binding motifs (GxGxxG) are shaded. The sequence segment in parentheses corresponds to the cofactor-binding domain. The lobe loop is shown in blue letters. The speculative conserved residues coordinated by the catalytic zinc are shown in red letters.
Crystal Structure of Threonine Dehydrogenase
cloned and expressed in E. coli. The purified PhTDH protein exhibits a hyperthermophilic TDH activity3 and gives a single band corresponding to a molecular mass of about 35,000 daltons on SDS–PAGE. The molecular mass of the protein in solution determined on high performance gel chromatography (data not shown) was approximately 125,000 daltons, suggesting that this protein exists as a tetramer, as observed for TDHs from E. coli and chicken liver.4–6 The crystal of PhTDH was prepared14 but the structure of PhTDH could not be solved by molecular replacement 15 with the ADH molecule. This first reported that the structure of PhTDH was determined by the single wavelength anomalous diffraction (SAD) method, and refined to the crystallographic R and Rfree factors of 0.204 and 0.239 in the resolution range of 130.19 Å–2.05 Å, respectively. Overall fold In the crystallographic asymmetric unit, three monomer PhTDHs were identified. The final model contains three monomers with 346 amino acid residues each. Each subunit is composed of two domains: a nicotinamide cofactor-binding domain (residues 153–291) and a catalytic domain (residues 2–152 and 292–346) (Figures 1 and 2). Superimposition of all Cα atoms of one monomer PhTDH on the other PhTDHs in the asymmetric unit exhibited an RMS deviation of less than 0.11 Å, showing that the structures of the three monomers are the same. Cispeptide bonds occur preceding Pro62, Pro194 and Trp297. Pro62 is well conserved in TDH and ADH, and a cis-peptide bond at this position is also conserved in ADH. A rare non-prolyl cis-peptide at Trp297 is observed at the surface of the catalytic
859 domain. Trp137, Trp297, Trp300 and Tyr301 form a hydrophobic cluster, suggesting an alternate conformation of Trp297 is adopted at this position in PhTDH. The result that some of these hydrophobic residues are conserved in TDH (Figure 3) suggests that TDH has a hydrophobic cluster at this position. The cofactor-binding domain contains the α/β Rossmann fold motif, characteristic of the nicotinamide cofactor (NAD(H))-binding protein,16–18 and comprises a six-stranded parallel β-sheet surrounded by six α-helices. The catalytic domain consists of a five-stranded β-sheet of a mixed type with four α-helices on the surface. Electron density of NAD(H) was observed at the cofactor-binding domain (Figure 4) indicating that the structure was solved as the holo-form. Superimposition of the Cα traces of the cofactor-binding domains of PhTDH monomer and the holo form of the A. pernix ADH (ApADH;1H2B) monomer matches 129 atoms with an RMS deviation of 1.40 Å (Figure 2). Superimposition of the Cα traces of the catalytic domains matches 116 atom with an RMS deviation of 1.44 Å. If both domains are used for structural alignment, 280 Cα atoms are superimposed with an RMS deviation of 1.65 Å. These values show that the overall fold of the two enzymes is very similar, in spite of the low level of sequence identity between TDH and ADH, and the most prominent difference involves the substrate specificity. The recombinant PhTDH exhibited a homo-tetramer structure. The asymmetric unit contained three subunits, and the solvent content of the crystals has been estimated to be 64.7% (v/v), VM = 3.5.14 However, a homo-tetramer with a 222 symmetry structure was observed in the cell (Figure 5). The result that PhTDH exists as a homo-tetramer structure in
Figure 2. PhTDH monomer, shown as (a) ribbon diagram with the structural zinc ion (gray), NAD(H) (yellow), and α1-helix (green). (b) Stereo view of the superimposed Cα traces of the cofactor-binding domains of PhTDH monomer (blue) and the holo form of ApADH monomer (green). The Figure was prepared using MOLSCRIPT.41
860
Figure 3. Structure of the hydrophobic cluster. The cispeptide at Trp297 (orange) is in the catalytic domain. Trp137, Trp297, Trp300 and Tyr301 are forming a hydrophobic cluster. The Figure was prepared using MOLSCRIPT.41
solution indicates that PhTDH forms the homotetramer structure with crystallographic and noncrystallographic 222 symmetry (Figure 5) as a biological molecule. Despite the low level of sequence identity to SsADH (29.5%) and ApADH (29.1%), the homo-tetramer structure of PhTDH is very similar to that of SsADH and ApADH.10,12 The tetramer seems to be essentially a dimer of dimers that have been designated as AB and CD in analogy
Crystal Structure of Threonine Dehydrogenase
with the previous structures (Figure 5).10,12 The formation of either the AB or CD dimers buried 1712 Å2 of the 15,667 Å2 accessible surface of each monomer, giving a solvent-accessible surface area of 27,911 Å2 for the dimer. Association of the dimers into a tetramer gives a final accessible surface area of 49,568 Å2, burying a further 3127 Å2 of each dimer. These values are similar to those for ApADH.12 The interaction between monomers A and B (or C and D) of PhTDH is similar to for ApADH and SsADH. The most extensive point of contact between monomers A and B (or C and D) is located at two C-terminal βstrands (residues 272–276 (βS) and residues 285–291 (βF) (Figure 1)) of the Rossmann fold of each monomer. As found in ADHs10,12 the formation of an extended 12-stranded β-sheet across the dimer interface involved the β-strands of both subunits (Figure 5). In addition, there is also a second point of contact via the lobe loop surrounding the structural zinc ion (residues 97–111) (Figure 6). The lobe loop contains some charged residues forming ionic bonds with the other subunits. Arg105 of monomer A forms an ionic bond with Thr258 and the mainchain of Pro259 of monomer B (the symmetryrelated equivalent was observed). Tyr101 of monomer A forms a hydrogen bond with Glu238 of monomer B (the symmetry-related equivalent was observed). Lys99 of monomer A forms an ionic bond with the main-chain of Glu81 of monomer C (the symmetry-related equivalent was observed). Arg104 of monomer A forms an ionic bond with the main-chain of Ile136 of monomer C (the symmetryrelated equivalent was observed). Tyr101 of monomer A undergoes a hydrophobic interaction with Tyr301 located in the hydrophobic cluster of monomer C (the symmetry-related equivalent was
Figure 4. Stereo view showing the additional electron density at the NAD(H) binding site. The Fo–Fc map was calculated using phases from the model in which NAD(H) was omitted. Dotted lines are hydrogen or ionic bonds drawn between either Glu199 or Arg204, and NAD(H). The sequence motif GxGxxG (residues 175–180) is shown in gray. The Figure was prepared using MOLSCRIPT.41
861
Crystal Structure of Threonine Dehydrogenase
Figure 5. Ribbon diagram of the PhTDH homo-tetramer generated from crystallographic and non-crystallographic symmetry. The tetramer is a dimer of dimers (A/B and C/D). The structural zinc ion is shown in gray. The Figure was prepared using MOLSCRIPT.41
observed). Between subunits A and D, and B and C, significant hydrophobic and ionic interactions were not observed. Structural zinc ion
Figure 6. Structural zinc site and lobe loop (residues 97–111). The structural zinc ion is shown in gray. The coordinate bonds are represented by pink lines. The Figure was prepared using MOLSCRIPT.41
Most NAD(H)-dependent ADHs have two zinc ions in each monomer. One functional catalytic zinc ion is present at the active site of ADH. The other is the structural zinc ion located in the lobe loop of the catalytic domain that is important for formation of the tetramer structure.10,12 The NADP(H)-dependent ADHs from Clostridium beijerinckii and Thermoanaerobacter brockii have the tetrameric structure and have the a similar lobe loop, but contain no zinc at the lobe loop in each monomer.19 The structural zinc ion is present in mammalian type I and archaeal NAD(H)-dependent ADHs. The structural zincbinding site (residues 97–111) located at the lobe loop of PhTDH is homologous to those of archaeal
862
Crystal Structure of Threonine Dehydrogenase
Table 1. Data-collection statistics Space group Unit-cell parameters (Å) Matthews coefficient (Å 3 Da−1) Solvent content (%) Subunits per asymmetric unit Resolution range (Å) Number of observed reflections Total number of unique reflections Multiplicity Rmergea Completeness (%)
Table 2. Interactions of PhTDH with cofactor NAD(H) I4122 a = b = 143.84, c = 304.13 3.5 64.7 3 84.82–2.05 (2.16–2.05) 717, 762 (105, 866) 99, 566 (14, 305) 4.8 (1.7) 7.4 (7.4) 0.105 (0.427) 100.0 (100.0)
Values for the last resolution shell are given in parentheses. a Rmerge = ∑hkl∑i|Ii(hkl)−〈I(hkl)〉|/∑hkl∑iIi(hkl), where Ii(hkl) is the i-th intensity measurement of reflection hkl, including symmetry-related reflections, and 〈I(hkl)〉 is their average.
and mammalian ADHs,10,12,20 although there are differences in the coordinating residues. One molecule of PhTDH contains one zinc ion.3 From the structure of PhTDH, it was shown that PhTDH coordinates the zinc ion with four cysteine residues (Cys97, Cys100, Cys103 and Cys111) and the ideal tetrahedral geometry (Figure 6), as observed for human and horse liver ADHs.20 From the high sequence homology at the lobe loop in TDHs (Figure 1), TDHs seem to coordinate one zinc ion with four cysteine residues, as the structural zinc ion,5 while SsADH and ApADH coordinate the zinc ion with three cysteine residues and one carboxyl residue of Glu or Asp. In the case of ApADH12 a disulfide bond with two cysteine residues distorting the tetrahedral geometry was observed. However, such distortion was not observed and the tetrahedral geometry for the zinc ion seems to be stable in the structure of PhTDH. Removal of the zinc ion from the molecule of PhTDH using EDTA decreased the thermostability of PhTDH,3 suggesting that the tetrahedral geometry containing the zinc ion at this position contributes to the hyperthermostability of PhTDH. Active site The two domains forming a PhTDH monomer are separated by a cleft containing a deep pocket that is able to accommodate both the threonine substrate and NAD(H). In the case of ADHs, the catalytic site can be identified from the location of the zinc ion that is essential for catalysis. The active site of PhTDH is also centered around the catalytic zinc ion and to be situated in the pocket containing the cofactor NAD(H), as observed for NAD(H)-dependent ADHs.10,12 The NAD(H) moiety is bound by the protein predominantly through hydrogen or ionic bonds occurring between NAD(H) and amino acid residues, as shown in Table 2. In ADHs and TDHs, the primary determinations of NAD(H)specificity are the GxGxxG sequence motif and the presence of Asp or Glu residues at the C-terminal end of the second β-sheet in cofactor binding domain (Figure 1).3,17 In PhTDH, Glu199 is present
NAD(H) atom NO7 NN7 NO3⁎ NO1 AO1 AO3⁎ AO2⁎
Interacting residues Thr292 (OG1) Ile291 (O), Leu266 (O) Leu268 (N) Leu179 (N) Arg204 (NH1) Arg204 (NH1) Glu199 (OE1) Glu199 (OE2)
Interactions shown in bold text are ionic, others are hydrogen bonds. The nomenclature for the atoms within NAD(H) is consistent with the REFMAC5 dictionary for the molecule.21
at the C-terminal end of the second β-sheet in cofactor binding domain (Figure 1) and forms hydrogen bonds with the two hydroxyl groups of the adenosine ribose in NAD(H) (Table 2 and Figure 4). This residue is equivalent to Asp218 of ApADH, Asp203 of SsADH, and Asp223 of HlADH (Figure 1), and is thought to determine cofactor specificity for ADH.16 This carboxyl residue is highly conserved in NAD(H)-dependent ADHs and TDHs, but is replaced by Gly in NADP(H)-dependent enzymes.9,19 In PhTDH, the highly conserved sequence motif GxGxxG (residues 175–180) (Figure 1) is observed at the NAD(H) binding site near the active cleft (Figure 4), and allows NAD(H) to come close to the mainchain, as observed for ADHs.11,12,19,22 In PhTDH, Arg204 forms an ionic bond with the phosphate group in NAD(H) (Table 2 and Figure 4). This Arg is well conserved in TDH but not in ADH (Figure 1). In HlADH as well as in other dehydrogenases, a Lys residue (Lys228 in HlADH and Lys223 in ApADH) corresponding to Arg204 in PhTDH provides an additional positive charge in the NAD(H) binding site.23–25 When Lys228 was replaced by Arg in HlADH, Arg228 fitted well into the NAD(H) binding site, but a small steric and electrostatic effect decreased the affinity for NAD(H).24,25 Glu199 and Arg204 seem to be the most important residues for NAD(H) binding in PhTDH (Figure 4). To explore the role of these two residues (Glu199 and Arg204), we prepared the mutant enzymes of which Glu199 and Arg204 were changed to Ala. The mutant E199A and R204A were prepared and kinetical parameters were determined (Table 3) with the previous method.3 The increase of Km values for NAD+ without significant difference for kcat values (Table 3) indicates that these residues are important Table 3. Kinetical parameters of wild-type PhTDH (WT) and mutants Enzyme
Km for threonine (mM)
Km for NAD+ (mM)
WT E199A R204A
(1.18 ±0.23) × 10−2 2.50 ± 0.24 1.51 ± 0.30
(9.91 ±1.21) × 10−3 0.304 ± 0.065 0.291 ± 0.11
kcat (min−1) 66 ± 1 40 ± 7 33 ± 10
Values for Km and kcat were determined at pH 7.5 and 65 °C by a non-linear regression analysis fit to the Michaelis–Menten equation.3
Crystal Structure of Threonine Dehydrogenase
for NAD(H) binding. The structural data indicate that these residues are not located at the threonine binding site. Therefore, the increase of Km values for threonine by the mutations (Table 3) suggests that these residues are related to NAD(H) binding and controlling the substrate (threonine) binding of PhTDH. For ApADH, another ionic bond between Arg354 and NAD(H) was observed.12 This residue is equivalent to Arg369 in HlADH and Arg342 in SsADH. From the sequence homology between ADHs and TDHs, the Arg residue corresponds to Lys342 of PhTDH. This Lys residue is well conserved in TDHs (Figure 1). In PhTDH, Lys342 is located over 7 Å from NAD(H) and undergoes no interaction with NAD(H), but forms an ionic bond with the carboxyl group of Glu68 (Figure 4). The role of conserved Lys342 (Arg342 in SsADH, Arg354 in ApADH, and Arg369 in HlADH) in PhTDH seems to be different from these in ADHs. The sequence motif (residues 41–49 forming the α1-helix in PhTDH) (Figure 1) conserved in TDHs and ADHs is located at the cleft of the active center (Figure 2), but exhibits no interaction with NAD(H) in PhTDH. In the case of ApADH, however, some residues in the above sequence motif interact with NAD(H) and are important for the activity.12 These residue and their positions seem to be important for the substrate specificity between TDH and ADH. TDH was reported to contain one zinc ion per subunit,26,27 but was found to be an enzyme that needs two zinc ions per subunit to exhibit activity like ADH. 5 Analysis of the metal in PhTDH indicated that it contained one zinc ion per subunit and exhibited its activity with zinc ion.3 However, the electron density of the zinc ion was not observed at the active cleft of the crystal of PhTDH. As observed for EcTDH, PhTDH seems to contain a functional catalytic zinc ion at the active center coordinated by Cys42, His67, Glu68
863 and Glu152, but the zinc ion seems to be less tightly bound at the site.5 The active site of apo-SsADH coordinates the zinc ion with four amino acid residues (Cys38, His68, Glu69, and Cys154).10 These four amino acids are conserved in ADHs (Figure 1). The active site of the complex of holoApADH and an inhibitor coordinates the zinc ion with three amino acid residues (Cys54, His79, and Asp168) and the inhibitor.12 Glu80 in ApTDH is at a distance of 5.7 Å from the catalytic zinc ion. These four residues are also conserved in TDHs. The equivalent residues are Cys42, His67, Glu68 and Glu152, and are located in the pocket near the reactive site of NAD(H) in PhTDH (Figure 7). The distance between the reactive site of NAD(H) and the above conserved residues is about 7 Å. For ADH, an Arg residue (Arg354 in ApADH and Arg369 in HlADN) interacts with the conserved Glu residues (Glu80 in ApADH and Glu68 in HlADN) and phosphate group in NAD(H). For PhTDH, however, the Arg residue is replaced by Lys342 and interacts with Glu68 without interaction with NAD (H) (Figures 4 and 7). In addition, Arg294 (conserved in TDH) forms an ionic bond with Glu152, and the carboxyl residue Glu152 is located over 6 Å from Cys42 and Glu68. Compared with the active site of ApADH and SsADH, the positions of the side-chains for Cys42, His67 and Glu68 are well conserved in PhTDH (Figure 7). When PhTDH was treated with EDTA and soaked in 5 mM nickel chloride or gadolinium chloride solution (Materials and Methods), the electron density of the ions was observed and was coordinated to Cys42, His67 and Glu68 at the active cleft in the crystal of PhTDH (data not shown). These results suggest that Cys42, His67 and Glu68 coordinate to a functional catalytic zinc ion at the active center of PhTDH with low affinity. 5 In the case of ADHs, the residues coordinating to the catalytic zinc ion are different
Figure 7. The active site of PhTDH. (a) The conserved residues at the active site. Dotted lines are ionic bonds drawn between Glu68 and Lys342, and Glu152 and Arg294. (b) Stereo view of superimposed PhTDH (blue) and ApADH (green) in the active site. The catalytic zinc ion of ApADH is shown in pink. The Figure was prepared using MOLSCRIPT.41
864 from that for the above equivalent residues of PhTDH. In the case of ADHs, Ser48 and His51 in HlADH and Thr56 and His59 in ApADH located in the conserved sequence motif (forming α1-helix) form hydrogen bonds with hydroxyl groups of NAD(H), and set up the proton relay system that connects the water or substrate bond with the catalytic zinc with the bulk solvent during the enzyme reaction.25,28,29 Rearrangement of the α1-helix and flexible loop caused by NAD(H) and substrate was observed in ApADH.29 These residues are conserved in ADHs and TDHs. In the case of PhTDH, the equivalent residues are Thr44 and His47 in the α1-helix, and are located over 6 Å from the 2′-hydroxyl groups of the nicotinamide ribose of NAD(H) (Figure 8). In the case of PhTDH, the substrate seems to rearrange the region (α1-helix) into a closed conformation and to relieve the hydrogen transfer, as observed in the case of HlADH.25,29 As is distinct from the narrow substrate specificity of ADH, the broad substrate specificity of TDH3–5 may be related to the large conformational change by the substrate. The hydrogen transfer mechanism of TDHs seems to be similar to that of ADHs. Compared with ADHs, the substrate-binding cleft of PhTDH does not provide a significant hydrophobic environment. Only, Leu46, Trp54, Ile116 and
Crystal Structure of Threonine Dehydrogenase
Leu268 are observed at the entrance of the cleft of PhTDH and are conserved in TDHs. In the case of EcTDH, it has been reported that Glu88 and His90 are important for the substrate specificity.30,31 In particular, His90 was reported not to coordinate the catalytic zinc ion but to modulate the substrate specificity of EcTDH.31 In the case of PhTDH, Glu92 (Glu88 in EcTDH) forms an ionic bond with Arg294 and is not located in the cleft. Glu92, His94 (His90 in EcTDH), Glu152 and Arg294 forms hydrogen or ionic bonds near the electron transfer site of NAD(H) and seem to be related to the substrate specificity. Between these residues and NAD(H), electron density for SO42− derived from the purification procedure was observed. SO42− is located 4.2 Å from the reactive site in NAD(H) and seems to form hydrogen bond to His94 (His90 in EcTDH) and Arg294 (Figure 9). The distance between SO42− and the catalytic center is 5.5–6.5 Å. This position corresponding to SO42− seems to be important for the substratebinding site (carboxyl residue and amino residue in threonine), as observed in the case of EcTDH.31 Most of the ADH structures have been solved as the apo-form. The finding that PhTDH was crystallized in the absence of NAD(H) indicates that NAD (H) in the recombinant enzyme is derived from E. coli cells and the binding affinity of PhTDH to NAD(H) is high, as suggested by the kinetical data in the previous paper.3 In the case of formaldehyde dehydrogenase (nicotinoprotein), the high binding affinity between enzyme and NAD(H) was observed around the moiety of adenine in NAD(H) as the hydrogen bonds.32,33 Without the significant hydrogen bonds, the high affinity of PhTDH to NAD(H) was observed. These results strongly support the proposed ordered Bi Bi mechanism for TDH,3,31 i.e. NAD+ combines with the enzyme prior to the binding of threonine, and 2-amino-3-oxobutyrate leaves from the enzyme before the release of NADH.
Materials and Methods Expression and purification
Figure 8. Structure around the α1-helix (green) in the active site of PhTDH. Thr44 and His47 are located over 6 Å (dotted lines) from the 2′-hydroxyl groups of the nicotinamide ribose of NAD(H) (yellow). The Figure was prepared using MOLSCRIPT.41
Expression and purification of the recombinant PhTDH were carried out as described,3 except that a modified M9 medium was used instead of LB medium.13 From the genome DNA of P. horikoshii, the gene encoding the PhTDH was amplified by PCR. The amplified gene was inserted into the pET-11a vector. The host, E. coli BL21 (DE3), was transformed using the plasmid for overexpression of PhTDH. Expression of selenomethioninelabeled PhTDH (SePhTDH) was performed in a modified M9 medium13 containing 0.05 mg/ml of ampicillin. The production of the recombinant protein was induced by adding 1 mM isopropyl β-D-thiogalactopyranoside, followed by further cultivation for 4 h. The culture of the transformant E. coli was subjected to centrifugation at 7000g and the collected cells were frozen at 253 K. The crude enzyme was extracted from the cells by sonication in 50 mM Tris–HCl buffer (pH 8.0). The extract was heated at 358 K for 30 min and then centrifuged to precipitate thermolabile proteins. The supernatant fraction was
865
Crystal Structure of Threonine Dehydrogenase
Figure 9. Structure of the active site showing the SO2− 4 and NAD(H) binding site. Glu92, His94 Glu152 and Arg294 are forming hydrogen or ionic bonds. Dotted lines are hydrogen bonds drawn between SO2− 4 and His94, and SO2− 4 and Arg294. The Figure was prepared using MOLSCRIPT.41 dialyzed against 50 mM Tris–HCl buffer (pH 8.0), and then purified on HiTrap Q and HiTrap Phenyl HP columns (GE Healthcare, USA). The purity of PhTDH fractions was confirmed by the single band on SDS–PAGE. The protein concentrations were determined with Coomassie protein reagent (Pierce Chemical Company, USA), using bovine serum albumin as the standard protein. Sitedirected mutagenesis was performed by overlap extension PCR method, employing the wild-type PhTDH gene as the template and the mutagenesis primer containing the desired mutation. Purification of the mutant enzymes was performed by the same method described above. To substitute the metal ions, the enzyme was dialyzed against 50 mM sodium phosphate buffer (pH7.5) containing 50 mM EDTA for 12 h and dialyzed against 50 mM sodium phosphate buffer (pH7.5) containing 5 mM of some metals (nickel chloride and gadolinium chloride) for 12 h. Enzyme assay Enzyme activity was determined by the previous method.3 The assay mixture contained variable concentrations of threonine in the presence of several fixed concentrations of NAD+ in 25 mM sodium phosphate buffer (pH 7.5). The mixture was pre-incubated for 5 min at 65 °C and the reaction was initiated by addition of the enzyme solution. NADH formation was monitored by measuring the increase in absorbance at 340 nm spectrophotometry. Values for Km and kcat were determined by a non-linear regression analysis fit to the Michaelis–Menten equation.
Crystallization was performed by the hanging-drop vapor-diffusion method at 277 K. The drops consisted of equal volumes (1.5 μl) of the protein and reservoir solutions. Crystals of SePhTDH were obtained by using a reservoir solution consisting of 0.2 M sodium chloride, 0.1 M Hepes buffer (pH 7.5), and 40% (v/v) PEG-400 over a period of five days.14 Data collection and processing The crystals were scooped up with a Cryo-loop (Hampton Research, Aliso Viejo, CA, USA) and immediately flashcooled at 100 K in a nitrogen cryo-stream. Diffraction data were collected at BL41XU, SPring-8 (Harima, Japan), and processed and scaled with MOSFLM 6.2.434 and SCALA15 from the CCP4 package.15 Selenium sites were searched with SOLVE35 and then the atomic parameters of selenium sites were refined by SHARP.36 Initial phases were calculated by the single wavelength anomalous diffraction (SAD) method by SHARP36 and improved by SOLOMON37 and dm38 implemented in SHARP. The space group and cell parameters of crystal PhTDH were the same as those of PhTDH.14 Refinement up to 2.05 Å was performed with REFMAC539 in the CCP4 package. Superimpositions of the TDH and ADH structures was performed with the Turbo-Frodo program.40 The diffraction data statistics and the crystallographic refinement statistics are summarized in Table 1. The figures of the structures were prepared using MOLSCRIPT.41 Protein Data Bank accession code
Crystallization The purified protein was dialyzed against 50 mM Tris– HCl buffer (pH 7.5) and then concentrated to 10 mg ml−1.
The structure factors and refined coordinates of PhTDH have been deposited with Protein Data Bank; the accession code is 2DFV.
866
Crystal Structure of Threonine Dehydrogenase
Acknowledgements We thank Professor emeritus Noritake Yasuoka of Himeji Institute of Technology and Professor Masahito Taya, Division of Chemical Engineering, Graduate School of Engineering Science, Osaka University for the fruitful discussions and helpful advice. We also thank Dr Mitsuo Ataka, our group leader, for the encouragement to carry out this work. This work was supported by the National Project on Protein Structural and Functional Analyses of the Ministry of Education, Culture, Sports, Science and Technology of Japan. The X-ray diffraction experiments were carried out with the approval of Japan Synchrotron Radiation Research Institute (proposal No.2004B0831-NL1-np-P3k, Japan).
References 1. Gonzalez, J. M., Masuchi, Y., Robb, F. T., Ammerman, J. W., Maeder, D. L. & Yanagibayashi, M. (1998). Pyrococcus horikoshii sp. nov., a hyperthermophilic archaeon isolated from a hydrothermal vent at the Okinawa Trough. Extremophiles, 2, 123–130. 2. Kawarabayasi, Y., Sawada, M., Horikawa, H., Haikawa, Y., Hino, Y. & Yamamoto, S. (1998). Complete sequence and gene organization of the genome of a hyper-thermophilic archaebacterium, Pyrococcus horikoshii OT3. DNA Res. 5, 55–76. 3. Higashi, N., Fukada, H. & Ishikawa, K. (2005). Kinetic study of thermostable L-threonine dehydrogenase from an archaeon Pyrococcus horikoshii. J. Biosci. Bioeng. 99, 175–180. 4. Epperly, B. R. & Dekker, E. E. (1991). L-Threonine dehydrogenase from Escherichia coli. Identification of an active site cysteine residue and metal ion studies. J. Biol. Chem. 266, 6086–6092. 5. Johnson, A. R., Chen, Y. W. & Dekker, E. E. (1998). Investigation of a catalytic zinc binding site in Escherichia coli L-threonine dehydrogenase by sitedirected mutagenesis of cystein-38. Arch. Biochem. Biophys. 358, 211–221. 6. Aoyama, Y. & Motokawa, Y. (1981). L-Threonine dehydrogenase of chicken liver. Purification, characterization, and physiological significance. J. Biol. Chem. 256, 12367–12373. 7. Ray, M. & Ray, S. (1985). L-Threonine dehydrogenase from goat liver. Feedback inhibition by methylglyoxal. J. Biol. Chem. 260, 5913–5918. 8. Aronson, B. D., Somerville, R. L., Epperly, B. R. & Dekker, E. E. (1989). The primary structure of Escherichia coli L-threonine dehydrogenase. J. Mol. Biol. 264, 5226–5232. 9. Reid, M. F. & Fewson, C. A. (1994). Molecular characterization of microbial alcohol dehydrogenases. Crit. Rev. Microbiol. 20, 13–56. 10. Esposito, L., Sica, F., Raia, C. A., Giordano, A., Rossi, M., Mazzarella, L. & Zagari, A. (2002). Crystal structure of the alcohol dehydrogenase from the hyperthermophilic archaeon Sulfolobus solfataricus at 1.85 Å resolution. J. Mol. Biol. 318, 463–477. 11. Esposito, L., Bruno, I., Sica, F., Raia, C. A., Giordano, A., Rossi, M. et al. (2003). Crystal structure of a ternary complex of the alcohol dehydrogenase from Sulfolobus solfataricus. Biochemistry, 42, 14397–14407. 12. Guy, J. E., Isupov, M. N. & Littlechild, J. A. (2003). The
13. 14.
15. 16. 17. 18. 19.
20.
21.
22.
23.
24. 25.
26. 27. 28.
29.
30.
structure of an alcohol dehydrogenase from the hyperthermophilic archaeon Aeropyrum pernix. J. Mol. Biol. 331, 1041–1051. Doublie, S. (1997). Preparation of selenomethionyl proteins for phase determination. Methods Enzymol. 276, 523–530. Higashi, N., Matsuura, T., Nakagawa, A. & Ishikawa, K. (2005). Crystallization and preliminary X-ray analysis of hyperthermophilic L-threonine dehydrogenase from archaeon Pyrococcus horikoshii. Acta Crystallog. sect. F, 61, 432–434. Evans, P. R. (1997). Scala joint CCP4 and ESF-EACBM. Newsletter Protein Crystallography, 33, 22–24. Rossmann, M. G., Moras, D. & Olsen, K. W. (1974). Chemical and biological evolution of nucleotidebinding protein. Nature, 250, 194–199. Bellamacina, C. R. (1996). The nicontinamide dinucleotide binding motif: a comparison of nucleotide binding proteins. FASEB J. 11, 1257–1269. Fan, F. & Plapp, B. V. (1999). Probing the affinity and specificity of yeast alcohol dehydrogenase I for coenzymes. Arch. Biochem. Biophys. 367, 240–249. Korkhin, Y., Kalb (Gilboa), A. J., Peretz, M., Bogin, O., Burstein, Y. & Frolow, F. (1998). NADP-dependent bacterial alcohol dehydrogenases: crystal structure, cofactor-binding and cofactor specificity of the ADHs of Clostridium beijerinckii and Thermoanaerobacter brockii. J. Mol. Biol. 278, 967–981. Eklund, H., Nordstrom, B., Zeppezauer, E., Soderlund, G., Ohlsson, I. & Boiwe, T. (1976). Threedimensional structure of horse liver alcohol dehydrogenase at 2–4 Å resolution. J. Mol. Biol. 102, 27–59. Murshudov, G. N., Vagin, A. A. & Dodson, E. J. (1997). Refinement of macromolecular structures by the maximum-likelihood method. Acta Crystallog. sect. D, 53, 240–255. Eklund, H. & Branden, C. I. (1987). Clystal structure, coenzyme conformations, and protein interactions. In Pyridine Nucleotides (Dolphin, D. et al., eds), pp. 51–98, Wiley & Sons, New York. Cho, H., Ramaswamy, S. & Plapp, B. V. (1997). Flexibility of liver alcohol dehydrogenase in stereoselective binding of 3-butylthiolane 1-oxides. Biochemistry, 36, 382–389. LeBrun, L. A. & Plapp, B. V. (1999). Control of coenzyme binding to horse liver alcohol dehydrogenase. Biochemistry, 38, 12387–12393. LeBrun, L. A., Park, D. H., Ramaswamy, S. & Plapp, B. V. (2004). Participation of histidine-51 in catalysis by horse liver alcohol dehydrogenase. Biochemistry, 43, 3014–3026. Vallee, B. L. & Hoch, F. L. (1955). Zinc a component of yeast alcohol dehydrogenase. Proc. Natl Acd. Sci. USA, 41, 327–338. Veillon, C. & Sytkwski, A. J. (1975). The intrinsic zinc atoms of yeast alcohol dehydrogenase. Biochem. Biophys. Res. Commun. 67, 1494–1500. Eklund, H., Plapp, B. V., Samama, J. P. & Branden, C. I. (1982). Binding of substrate in a ternary complex of horse liver alcohol dehydrogenase. J. Biol. Chem. 257, 14349–14358. Ramaswamy, S., Park, D. H. & Plapp, B. V. (1999). Substitutions in a flexible loop of horse liver alcohol dehydrogenase hinder the conformational change and unmask hydrogen transfer. Biochemistry, 38, 13951–13959. Chen, Y. W., Dekker, E. E. & Somerville, R. L. (1995). Functional analysis of E. coli threonine dehydrogenase by means of mutant isolation and characterization. Biochim. Biophys. Acta, 1253, 208–214.
Crystal Structure of Threonine Dehydrogenase
31. Johnson, A. R. & Dekker, E. E. (1998). Site-directed mutagenesisi of Histidine-90 in Escherichia coli Lthreonine dehydrogenase alters its substrate specificity. Arch. Biochem. Biophys. 351, 8–16. 32. Piersma, S. R., de Vries, S. & Duine, J. A. (1996). In Enzymology and Molecular Biology of Carbonyl Methabolism 6 (Weiner H. et al. eds), pp. 4425–4434, Plenum Press, New York. 33. Tanaka, N., Kusakabe, Y., Ito, K., Yoshimoto, T. & Nakamura, K. T. (2002). Crystal structure of formaldehyde dehydrogenase from Pseudomonas putida: the structural origin of the tightly bound cofactor in nicotinoprotein dehydrogenases. J. Mol. Biol. 324, 519–533. 34. Leslie, A. G. W. (1992). Jnt CCP4/ESF-EACBM. Newsletter Protein Crystallog, 26. 35. Terwilliger, T. (2003). SOLVE and RESOLVE: automated structure solution and density modification. Methods Enzymol. 374, 22–37.
867 36. La Fortella, E. & Bricogne, G. (1997). Maximum-likelihood heavy-atom parameter refinement in the MIR and MAD methods. Methods. Enzymol. 276, 472–494. 37. Abrahams, J. & Leslie, A. (1996). Methods used in the structure determination of bovine mitochondrial F1 ATPase. Acta Crystallog. sect. D, 52, 30–42. 38. Cowtan, K. (1999). Error estimation and bias correction in phase-improvement calculations. Acta Crystallog. sect. D, 55, 1555–15567. 39. Murshudov, G., Vagin, A. & Dodson, E. (1997). Refinement of macromolecular structures by the maximum-likelihood method. Acta Crystallog. sect. D, 53, 240–255. 40. Roussel, A. & Cambillau, C. (1991). Turbo-Frodo. Silicon Graphics Geometry Partners Directory Silicon Graphics, Mountain View, CA. 41. Kraulis, P. J. (1991). MOLSCRIPT: a program to produce both detailed and schematic plots of protein structures. J. Appl. Crystallog, 24, 946–959.
Edited by K. Morikawa (Received 21 July 2006; received in revised form 10 November 2006; accepted 17 November 2006) Available online 21 November 2006
J. Mol. Biol. (2007) 366, 857–867
The First Crystal Structure of L-Threonine Dehydrogenase Kazuhiko Ishikawa 1 ⁎, Noriko Higashi 1 , Tsutomu Nakamura 1 Takanori Matsuura 2 and Atsushi Nakagawa 2 ⁎ 1
National Institute of Advanced Industrial Science and Technology (AIST), 1-8-31 Midorigaoka, Ikeda, Osaka 563-8577, Japan 2
Institute for Protein Research, Osaka University, 3-2 Yamadaoka, Suita, Osaka 565-0871, Japan
L-Threonine
dehydrogenase (TDH) is an enzyme that catalyzes the oxidation of L-threonine to 2-amino-3-ketobutyrate. We solved the first crystal structure of a medium chain L-threonine dehydrogenase from a hyperthermophilic archaeon, Pyrococcus horikoshii (PhTDH), by the single wavelength anomalous diffraction method using a selenomethioninesubstituted enzyme. This recombinant PhTDH is a homo-tetramer in solution. Three monomers of PhTDHs were located in the crystallographic asymmetric unit, however, the crystal structure exhibits a homo-tetramer structure with crystallographic and non-crystallographic 222 symmetry in the cell. Despite the low level of sequence identity to a medium-chain NAD (H)-dependent alcohol dehydrogenase (ADH) and the different substrate specificity, the overall folds of the PhTDH monomer and tetramer are similar to those of the other ADH. Each subunit is composed of two domains: a nicotinamide cofactor (NAD(H))-binding domain and a catalytic domain. The NAD(H)-binding domain contains the α/β Rossmann fold motif, characteristic of the NAD(H)-binding protein. One molecule of PhTDH contains one zinc ion playing a structural role. This metal ion exhibits coordination with four cysteine ligands and some of the ligands are conserved throughout the structural zinc-containing ADHs and TDHs. However, the catalytic zinc ion that is coordinated at the bottom of the cleft in the case of ADH was not observed in the crystal of PhTDH. There is a significant difference in the orientation of the catalytic domain relative to the coenzyme-binding domain that results in a larger interdomain cleft. © 2006 Elsevier Ltd. All rights reserved.
*Corresponding authors
Keywords: archaea; Pyrococcus horikoshii; threonine dehydrogenase; alcohol dehydrogenase; X-ray crystallography
Introduction A hyperthermophilic anaerobic archaeon, Pyrococcus horikoshii, was isolated from a hydrothermal volcanic vent in the Okinawa Trough in the Pacific Ocean.1 An open-reading frame (ORF) (PH0655: GenBank accession no. O58389) in the P. horikoshii genome database has been classified as an LPresent address: N. Higashi, Graduate School of Engineering Science, Osaka University, 1-3 Machikaneyama, Toyonaka, Osaka 560-8531, Japan. Abbreviations used: TDH, L-Threonine dehydrogenase; ADH, alcohol dehydrogenase; Ph, Pyrococcus horikoshii; Ec, Escherichia coli; Bs, Bacillus subtilis; Hl, horse liver; Ss, Sulfolobus solfataricus; Ap, Aeropyrum pernix. E-mail addresses of the corresponding authors: [email protected]; [email protected]
threonine dehydrogenase (TDH; EC 1.1.1.103), 2 with sequence homology to an alcohol dehydrogenase (ADH; EC 1.1.1.1). We have cloned PH0655 and found it to be a hyperthermophilic TDH of P. horikoshii (PhTDH).3 TDH can catalyze the NAD (H)-dependent oxidation of L-threonine to L-2amino-3-ketobutyrate4–7 and requires zinc for its catalytic activity. PhTDH was classified as a member of the medium-chain, NAD(H)-dependent, and zinccontaining alcohol/polyol dehydrogenase family.3,8 PhTDH has a broad substrate specificity and is expected to be useful for the organic synthesis of several chiral compounds (L-2-amino-3-ketobutyrate, L-2-amino-3-ketopropionate and their carbonic acid derivatives). However, there has been no report for the catalytic mechanism of TDH. Many studies on the structure of ADHs have been carried out and the catalytic mechanism has been discussed in detail. ADHs are divided according to
0022-2836/$ - see front matter © 2006 Elsevier Ltd. All rights reserved.
858 their size and cofactor (NAD(H) or NADP(H)) requirement.9 In particular, medium-chain NAD (H)-dependent ADHs mostly containing two zinc atoms per subunit are widely distributed, being found in bacteria, eukarya and archaea. Between medium-chain ADHs and TDHs, some sequence identity and a conserved sequence region have been observed. The detailed crystal structures for the homo-tetrameric structures of ADHs from hyperthermophilic archaea Sulfolobus solfataricus and Aeropyrum pernix have also been reported.10–12 The sequence of PhTDH is homologous to the hyperthermophilic archaeon ADH's from S. solfataricus (29.5 % identity) (SsADH) and A. pernix (29.1 % identity) (ApADH). The PhTDH gene has been cloned and the recombinant enzyme overexpressed in Escherichia coli, and its characteristics have been reported.3 However, there has been no report on the crystal structure of TDHs. Analysis of the structural and functional similarities between TDH and ADH will increase our understanding of the different functions of these enzymes towards the structural origin
Crystal Structure of Threonine Dehydrogenase
of the hyperthermostable protein in this group of archaea. Furthermore, the structure of the active site and the catalytic residues of TDH are informative. Here, we prepare a crystal of PhTDH (SePhTDH) using the selenomethionine-labelled protein13 and report its crystal structure.
Results and Discussion In the genome database of P. horikoshii, a TDH homologue gene was found.2,3 The gene product (PhTDH) was composed of 348 amino acid residues (1044 base-pairs), being homologous with TDHs from E. coli (EcTDH; GenBank accession no. P07913, 44.0% identity) and Bacillus subtilis (BsTDH; GenBank accession no. O31776, 49.1% identity), and ADHs from horse liver (HlADH; GenBank accession no. P00327, 28.3% identity), S. solfataricus (SsADH; GenBank accession no. P39462, 29.5% identity), and A. pernix (ApADH; GenBank accession no. O9Y9P9, 29.1% identity), as shown in Figure 1. PhTDH was
Figure 1. Alignment of amino acids compared among TDHs and ADHs from various origins. The abbreviations are as follows: Ph, P. horikoshii (PH0655); Ec, E. coli (ORF no. b3616); Bs, B. subtilis (BSU16990); Hl, horse liver; Ss, S. solfataricus; and Ap, A. pernix. Identical residues are shown by asterisks. In the PhTDH sequence (upper row), the main secondary structure is shown: residues highlighted in blue and red (or green) are β-strands and α-helices, respectively. NAD(H) binding motifs (GxGxxG) are shaded. The sequence segment in parentheses corresponds to the cofactor-binding domain. The lobe loop is shown in blue letters. The speculative conserved residues coordinated by the catalytic zinc are shown in red letters.
Crystal Structure of Threonine Dehydrogenase
cloned and expressed in E. coli. The purified PhTDH protein exhibits a hyperthermophilic TDH activity3 and gives a single band corresponding to a molecular mass of about 35,000 daltons on SDS–PAGE. The molecular mass of the protein in solution determined on high performance gel chromatography (data not shown) was approximately 125,000 daltons, suggesting that this protein exists as a tetramer, as observed for TDHs from E. coli and chicken liver.4–6 The crystal of PhTDH was prepared14 but the structure of PhTDH could not be solved by molecular replacement 15 with the ADH molecule. This first reported that the structure of PhTDH was determined by the single wavelength anomalous diffraction (SAD) method, and refined to the crystallographic R and Rfree factors of 0.204 and 0.239 in the resolution range of 130.19 Å–2.05 Å, respectively. Overall fold In the crystallographic asymmetric unit, three monomer PhTDHs were identified. The final model contains three monomers with 346 amino acid residues each. Each subunit is composed of two domains: a nicotinamide cofactor-binding domain (residues 153–291) and a catalytic domain (residues 2–152 and 292–346) (Figures 1 and 2). Superimposition of all Cα atoms of one monomer PhTDH on the other PhTDHs in the asymmetric unit exhibited an RMS deviation of less than 0.11 Å, showing that the structures of the three monomers are the same. Cispeptide bonds occur preceding Pro62, Pro194 and Trp297. Pro62 is well conserved in TDH and ADH, and a cis-peptide bond at this position is also conserved in ADH. A rare non-prolyl cis-peptide at Trp297 is observed at the surface of the catalytic
859 domain. Trp137, Trp297, Trp300 and Tyr301 form a hydrophobic cluster, suggesting an alternate conformation of Trp297 is adopted at this position in PhTDH. The result that some of these hydrophobic residues are conserved in TDH (Figure 3) suggests that TDH has a hydrophobic cluster at this position. The cofactor-binding domain contains the α/β Rossmann fold motif, characteristic of the nicotinamide cofactor (NAD(H))-binding protein,16–18 and comprises a six-stranded parallel β-sheet surrounded by six α-helices. The catalytic domain consists of a five-stranded β-sheet of a mixed type with four α-helices on the surface. Electron density of NAD(H) was observed at the cofactor-binding domain (Figure 4) indicating that the structure was solved as the holo-form. Superimposition of the Cα traces of the cofactor-binding domains of PhTDH monomer and the holo form of the A. pernix ADH (ApADH;1H2B) monomer matches 129 atoms with an RMS deviation of 1.40 Å (Figure 2). Superimposition of the Cα traces of the catalytic domains matches 116 atom with an RMS deviation of 1.44 Å. If both domains are used for structural alignment, 280 Cα atoms are superimposed with an RMS deviation of 1.65 Å. These values show that the overall fold of the two enzymes is very similar, in spite of the low level of sequence identity between TDH and ADH, and the most prominent difference involves the substrate specificity. The recombinant PhTDH exhibited a homo-tetramer structure. The asymmetric unit contained three subunits, and the solvent content of the crystals has been estimated to be 64.7% (v/v), VM = 3.5.14 However, a homo-tetramer with a 222 symmetry structure was observed in the cell (Figure 5). The result that PhTDH exists as a homo-tetramer structure in
Figure 2. PhTDH monomer, shown as (a) ribbon diagram with the structural zinc ion (gray), NAD(H) (yellow), and α1-helix (green). (b) Stereo view of the superimposed Cα traces of the cofactor-binding domains of PhTDH monomer (blue) and the holo form of ApADH monomer (green). The Figure was prepared using MOLSCRIPT.41
860
Figure 3. Structure of the hydrophobic cluster. The cispeptide at Trp297 (orange) is in the catalytic domain. Trp137, Trp297, Trp300 and Tyr301 are forming a hydrophobic cluster. The Figure was prepared using MOLSCRIPT.41
solution indicates that PhTDH forms the homotetramer structure with crystallographic and noncrystallographic 222 symmetry (Figure 5) as a biological molecule. Despite the low level of sequence identity to SsADH (29.5%) and ApADH (29.1%), the homo-tetramer structure of PhTDH is very similar to that of SsADH and ApADH.10,12 The tetramer seems to be essentially a dimer of dimers that have been designated as AB and CD in analogy
Crystal Structure of Threonine Dehydrogenase
with the previous structures (Figure 5).10,12 The formation of either the AB or CD dimers buried 1712 Å2 of the 15,667 Å2 accessible surface of each monomer, giving a solvent-accessible surface area of 27,911 Å2 for the dimer. Association of the dimers into a tetramer gives a final accessible surface area of 49,568 Å2, burying a further 3127 Å2 of each dimer. These values are similar to those for ApADH.12 The interaction between monomers A and B (or C and D) of PhTDH is similar to for ApADH and SsADH. The most extensive point of contact between monomers A and B (or C and D) is located at two C-terminal βstrands (residues 272–276 (βS) and residues 285–291 (βF) (Figure 1)) of the Rossmann fold of each monomer. As found in ADHs10,12 the formation of an extended 12-stranded β-sheet across the dimer interface involved the β-strands of both subunits (Figure 5). In addition, there is also a second point of contact via the lobe loop surrounding the structural zinc ion (residues 97–111) (Figure 6). The lobe loop contains some charged residues forming ionic bonds with the other subunits. Arg105 of monomer A forms an ionic bond with Thr258 and the mainchain of Pro259 of monomer B (the symmetryrelated equivalent was observed). Tyr101 of monomer A forms a hydrogen bond with Glu238 of monomer B (the symmetry-related equivalent was observed). Lys99 of monomer A forms an ionic bond with the main-chain of Glu81 of monomer C (the symmetry-related equivalent was observed). Arg104 of monomer A forms an ionic bond with the main-chain of Ile136 of monomer C (the symmetryrelated equivalent was observed). Tyr101 of monomer A undergoes a hydrophobic interaction with Tyr301 located in the hydrophobic cluster of monomer C (the symmetry-related equivalent was
Figure 4. Stereo view showing the additional electron density at the NAD(H) binding site. The Fo–Fc map was calculated using phases from the model in which NAD(H) was omitted. Dotted lines are hydrogen or ionic bonds drawn between either Glu199 or Arg204, and NAD(H). The sequence motif GxGxxG (residues 175–180) is shown in gray. The Figure was prepared using MOLSCRIPT.41
861
Crystal Structure of Threonine Dehydrogenase
Figure 5. Ribbon diagram of the PhTDH homo-tetramer generated from crystallographic and non-crystallographic symmetry. The tetramer is a dimer of dimers (A/B and C/D). The structural zinc ion is shown in gray. The Figure was prepared using MOLSCRIPT.41
observed). Between subunits A and D, and B and C, significant hydrophobic and ionic interactions were not observed. Structural zinc ion
Figure 6. Structural zinc site and lobe loop (residues 97–111). The structural zinc ion is shown in gray. The coordinate bonds are represented by pink lines. The Figure was prepared using MOLSCRIPT.41
Most NAD(H)-dependent ADHs have two zinc ions in each monomer. One functional catalytic zinc ion is present at the active site of ADH. The other is the structural zinc ion located in the lobe loop of the catalytic domain that is important for formation of the tetramer structure.10,12 The NADP(H)-dependent ADHs from Clostridium beijerinckii and Thermoanaerobacter brockii have the tetrameric structure and have the a similar lobe loop, but contain no zinc at the lobe loop in each monomer.19 The structural zinc ion is present in mammalian type I and archaeal NAD(H)-dependent ADHs. The structural zincbinding site (residues 97–111) located at the lobe loop of PhTDH is homologous to those of archaeal
862
Crystal Structure of Threonine Dehydrogenase
Table 1. Data-collection statistics Space group Unit-cell parameters (Å) Matthews coefficient (Å 3 Da−1) Solvent content (%) Subunits per asymmetric unit Resolution range (Å) Number of observed reflections Total number of unique reflections Multiplicity Rmergea Completeness (%)
Table 2. Interactions of PhTDH with cofactor NAD(H) I4122 a = b = 143.84, c = 304.13 3.5 64.7 3 84.82–2.05 (2.16–2.05) 717, 762 (105, 866) 99, 566 (14, 305) 4.8 (1.7) 7.4 (7.4) 0.105 (0.427) 100.0 (100.0)
Values for the last resolution shell are given in parentheses. a Rmerge = ∑hkl∑i|Ii(hkl)−〈I(hkl)〉|/∑hkl∑iIi(hkl), where Ii(hkl) is the i-th intensity measurement of reflection hkl, including symmetry-related reflections, and 〈I(hkl)〉 is their average.
and mammalian ADHs,10,12,20 although there are differences in the coordinating residues. One molecule of PhTDH contains one zinc ion.3 From the structure of PhTDH, it was shown that PhTDH coordinates the zinc ion with four cysteine residues (Cys97, Cys100, Cys103 and Cys111) and the ideal tetrahedral geometry (Figure 6), as observed for human and horse liver ADHs.20 From the high sequence homology at the lobe loop in TDHs (Figure 1), TDHs seem to coordinate one zinc ion with four cysteine residues, as the structural zinc ion,5 while SsADH and ApADH coordinate the zinc ion with three cysteine residues and one carboxyl residue of Glu or Asp. In the case of ApADH12 a disulfide bond with two cysteine residues distorting the tetrahedral geometry was observed. However, such distortion was not observed and the tetrahedral geometry for the zinc ion seems to be stable in the structure of PhTDH. Removal of the zinc ion from the molecule of PhTDH using EDTA decreased the thermostability of PhTDH,3 suggesting that the tetrahedral geometry containing the zinc ion at this position contributes to the hyperthermostability of PhTDH. Active site The two domains forming a PhTDH monomer are separated by a cleft containing a deep pocket that is able to accommodate both the threonine substrate and NAD(H). In the case of ADHs, the catalytic site can be identified from the location of the zinc ion that is essential for catalysis. The active site of PhTDH is also centered around the catalytic zinc ion and to be situated in the pocket containing the cofactor NAD(H), as observed for NAD(H)-dependent ADHs.10,12 The NAD(H) moiety is bound by the protein predominantly through hydrogen or ionic bonds occurring between NAD(H) and amino acid residues, as shown in Table 2. In ADHs and TDHs, the primary determinations of NAD(H)specificity are the GxGxxG sequence motif and the presence of Asp or Glu residues at the C-terminal end of the second β-sheet in cofactor binding domain (Figure 1).3,17 In PhTDH, Glu199 is present
NAD(H) atom NO7 NN7 NO3⁎ NO1 AO1 AO3⁎ AO2⁎
Interacting residues Thr292 (OG1) Ile291 (O), Leu266 (O) Leu268 (N) Leu179 (N) Arg204 (NH1) Arg204 (NH1) Glu199 (OE1) Glu199 (OE2)
Interactions shown in bold text are ionic, others are hydrogen bonds. The nomenclature for the atoms within NAD(H) is consistent with the REFMAC5 dictionary for the molecule.21
at the C-terminal end of the second β-sheet in cofactor binding domain (Figure 1) and forms hydrogen bonds with the two hydroxyl groups of the adenosine ribose in NAD(H) (Table 2 and Figure 4). This residue is equivalent to Asp218 of ApADH, Asp203 of SsADH, and Asp223 of HlADH (Figure 1), and is thought to determine cofactor specificity for ADH.16 This carboxyl residue is highly conserved in NAD(H)-dependent ADHs and TDHs, but is replaced by Gly in NADP(H)-dependent enzymes.9,19 In PhTDH, the highly conserved sequence motif GxGxxG (residues 175–180) (Figure 1) is observed at the NAD(H) binding site near the active cleft (Figure 4), and allows NAD(H) to come close to the mainchain, as observed for ADHs.11,12,19,22 In PhTDH, Arg204 forms an ionic bond with the phosphate group in NAD(H) (Table 2 and Figure 4). This Arg is well conserved in TDH but not in ADH (Figure 1). In HlADH as well as in other dehydrogenases, a Lys residue (Lys228 in HlADH and Lys223 in ApADH) corresponding to Arg204 in PhTDH provides an additional positive charge in the NAD(H) binding site.23–25 When Lys228 was replaced by Arg in HlADH, Arg228 fitted well into the NAD(H) binding site, but a small steric and electrostatic effect decreased the affinity for NAD(H).24,25 Glu199 and Arg204 seem to be the most important residues for NAD(H) binding in PhTDH (Figure 4). To explore the role of these two residues (Glu199 and Arg204), we prepared the mutant enzymes of which Glu199 and Arg204 were changed to Ala. The mutant E199A and R204A were prepared and kinetical parameters were determined (Table 3) with the previous method.3 The increase of Km values for NAD+ without significant difference for kcat values (Table 3) indicates that these residues are important Table 3. Kinetical parameters of wild-type PhTDH (WT) and mutants Enzyme
Km for threonine (mM)
Km for NAD+ (mM)
WT E199A R204A
(1.18 ±0.23) × 10−2 2.50 ± 0.24 1.51 ± 0.30
(9.91 ±1.21) × 10−3 0.304 ± 0.065 0.291 ± 0.11
kcat (min−1) 66 ± 1 40 ± 7 33 ± 10
Values for Km and kcat were determined at pH 7.5 and 65 °C by a non-linear regression analysis fit to the Michaelis–Menten equation.3
Crystal Structure of Threonine Dehydrogenase
for NAD(H) binding. The structural data indicate that these residues are not located at the threonine binding site. Therefore, the increase of Km values for threonine by the mutations (Table 3) suggests that these residues are related to NAD(H) binding and controlling the substrate (threonine) binding of PhTDH. For ApADH, another ionic bond between Arg354 and NAD(H) was observed.12 This residue is equivalent to Arg369 in HlADH and Arg342 in SsADH. From the sequence homology between ADHs and TDHs, the Arg residue corresponds to Lys342 of PhTDH. This Lys residue is well conserved in TDHs (Figure 1). In PhTDH, Lys342 is located over 7 Å from NAD(H) and undergoes no interaction with NAD(H), but forms an ionic bond with the carboxyl group of Glu68 (Figure 4). The role of conserved Lys342 (Arg342 in SsADH, Arg354 in ApADH, and Arg369 in HlADH) in PhTDH seems to be different from these in ADHs. The sequence motif (residues 41–49 forming the α1-helix in PhTDH) (Figure 1) conserved in TDHs and ADHs is located at the cleft of the active center (Figure 2), but exhibits no interaction with NAD(H) in PhTDH. In the case of ApADH, however, some residues in the above sequence motif interact with NAD(H) and are important for the activity.12 These residue and their positions seem to be important for the substrate specificity between TDH and ADH. TDH was reported to contain one zinc ion per subunit,26,27 but was found to be an enzyme that needs two zinc ions per subunit to exhibit activity like ADH. 5 Analysis of the metal in PhTDH indicated that it contained one zinc ion per subunit and exhibited its activity with zinc ion.3 However, the electron density of the zinc ion was not observed at the active cleft of the crystal of PhTDH. As observed for EcTDH, PhTDH seems to contain a functional catalytic zinc ion at the active center coordinated by Cys42, His67, Glu68
863 and Glu152, but the zinc ion seems to be less tightly bound at the site.5 The active site of apo-SsADH coordinates the zinc ion with four amino acid residues (Cys38, His68, Glu69, and Cys154).10 These four amino acids are conserved in ADHs (Figure 1). The active site of the complex of holoApADH and an inhibitor coordinates the zinc ion with three amino acid residues (Cys54, His79, and Asp168) and the inhibitor.12 Glu80 in ApTDH is at a distance of 5.7 Å from the catalytic zinc ion. These four residues are also conserved in TDHs. The equivalent residues are Cys42, His67, Glu68 and Glu152, and are located in the pocket near the reactive site of NAD(H) in PhTDH (Figure 7). The distance between the reactive site of NAD(H) and the above conserved residues is about 7 Å. For ADH, an Arg residue (Arg354 in ApADH and Arg369 in HlADN) interacts with the conserved Glu residues (Glu80 in ApADH and Glu68 in HlADN) and phosphate group in NAD(H). For PhTDH, however, the Arg residue is replaced by Lys342 and interacts with Glu68 without interaction with NAD (H) (Figures 4 and 7). In addition, Arg294 (conserved in TDH) forms an ionic bond with Glu152, and the carboxyl residue Glu152 is located over 6 Å from Cys42 and Glu68. Compared with the active site of ApADH and SsADH, the positions of the side-chains for Cys42, His67 and Glu68 are well conserved in PhTDH (Figure 7). When PhTDH was treated with EDTA and soaked in 5 mM nickel chloride or gadolinium chloride solution (Materials and Methods), the electron density of the ions was observed and was coordinated to Cys42, His67 and Glu68 at the active cleft in the crystal of PhTDH (data not shown). These results suggest that Cys42, His67 and Glu68 coordinate to a functional catalytic zinc ion at the active center of PhTDH with low affinity. 5 In the case of ADHs, the residues coordinating to the catalytic zinc ion are different
Figure 7. The active site of PhTDH. (a) The conserved residues at the active site. Dotted lines are ionic bonds drawn between Glu68 and Lys342, and Glu152 and Arg294. (b) Stereo view of superimposed PhTDH (blue) and ApADH (green) in the active site. The catalytic zinc ion of ApADH is shown in pink. The Figure was prepared using MOLSCRIPT.41
864 from that for the above equivalent residues of PhTDH. In the case of ADHs, Ser48 and His51 in HlADH and Thr56 and His59 in ApADH located in the conserved sequence motif (forming α1-helix) form hydrogen bonds with hydroxyl groups of NAD(H), and set up the proton relay system that connects the water or substrate bond with the catalytic zinc with the bulk solvent during the enzyme reaction.25,28,29 Rearrangement of the α1-helix and flexible loop caused by NAD(H) and substrate was observed in ApADH.29 These residues are conserved in ADHs and TDHs. In the case of PhTDH, the equivalent residues are Thr44 and His47 in the α1-helix, and are located over 6 Å from the 2′-hydroxyl groups of the nicotinamide ribose of NAD(H) (Figure 8). In the case of PhTDH, the substrate seems to rearrange the region (α1-helix) into a closed conformation and to relieve the hydrogen transfer, as observed in the case of HlADH.25,29 As is distinct from the narrow substrate specificity of ADH, the broad substrate specificity of TDH3–5 may be related to the large conformational change by the substrate. The hydrogen transfer mechanism of TDHs seems to be similar to that of ADHs. Compared with ADHs, the substrate-binding cleft of PhTDH does not provide a significant hydrophobic environment. Only, Leu46, Trp54, Ile116 and
Crystal Structure of Threonine Dehydrogenase
Leu268 are observed at the entrance of the cleft of PhTDH and are conserved in TDHs. In the case of EcTDH, it has been reported that Glu88 and His90 are important for the substrate specificity.30,31 In particular, His90 was reported not to coordinate the catalytic zinc ion but to modulate the substrate specificity of EcTDH.31 In the case of PhTDH, Glu92 (Glu88 in EcTDH) forms an ionic bond with Arg294 and is not located in the cleft. Glu92, His94 (His90 in EcTDH), Glu152 and Arg294 forms hydrogen or ionic bonds near the electron transfer site of NAD(H) and seem to be related to the substrate specificity. Between these residues and NAD(H), electron density for SO42− derived from the purification procedure was observed. SO42− is located 4.2 Å from the reactive site in NAD(H) and seems to form hydrogen bond to His94 (His90 in EcTDH) and Arg294 (Figure 9). The distance between SO42− and the catalytic center is 5.5–6.5 Å. This position corresponding to SO42− seems to be important for the substratebinding site (carboxyl residue and amino residue in threonine), as observed in the case of EcTDH.31 Most of the ADH structures have been solved as the apo-form. The finding that PhTDH was crystallized in the absence of NAD(H) indicates that NAD (H) in the recombinant enzyme is derived from E. coli cells and the binding affinity of PhTDH to NAD(H) is high, as suggested by the kinetical data in the previous paper.3 In the case of formaldehyde dehydrogenase (nicotinoprotein), the high binding affinity between enzyme and NAD(H) was observed around the moiety of adenine in NAD(H) as the hydrogen bonds.32,33 Without the significant hydrogen bonds, the high affinity of PhTDH to NAD(H) was observed. These results strongly support the proposed ordered Bi Bi mechanism for TDH,3,31 i.e. NAD+ combines with the enzyme prior to the binding of threonine, and 2-amino-3-oxobutyrate leaves from the enzyme before the release of NADH.
Materials and Methods Expression and purification
Figure 8. Structure around the α1-helix (green) in the active site of PhTDH. Thr44 and His47 are located over 6 Å (dotted lines) from the 2′-hydroxyl groups of the nicotinamide ribose of NAD(H) (yellow). The Figure was prepared using MOLSCRIPT.41
Expression and purification of the recombinant PhTDH were carried out as described,3 except that a modified M9 medium was used instead of LB medium.13 From the genome DNA of P. horikoshii, the gene encoding the PhTDH was amplified by PCR. The amplified gene was inserted into the pET-11a vector. The host, E. coli BL21 (DE3), was transformed using the plasmid for overexpression of PhTDH. Expression of selenomethioninelabeled PhTDH (SePhTDH) was performed in a modified M9 medium13 containing 0.05 mg/ml of ampicillin. The production of the recombinant protein was induced by adding 1 mM isopropyl β-D-thiogalactopyranoside, followed by further cultivation for 4 h. The culture of the transformant E. coli was subjected to centrifugation at 7000g and the collected cells were frozen at 253 K. The crude enzyme was extracted from the cells by sonication in 50 mM Tris–HCl buffer (pH 8.0). The extract was heated at 358 K for 30 min and then centrifuged to precipitate thermolabile proteins. The supernatant fraction was
865
Crystal Structure of Threonine Dehydrogenase
Figure 9. Structure of the active site showing the SO2− 4 and NAD(H) binding site. Glu92, His94 Glu152 and Arg294 are forming hydrogen or ionic bonds. Dotted lines are hydrogen bonds drawn between SO2− 4 and His94, and SO2− 4 and Arg294. The Figure was prepared using MOLSCRIPT.41 dialyzed against 50 mM Tris–HCl buffer (pH 8.0), and then purified on HiTrap Q and HiTrap Phenyl HP columns (GE Healthcare, USA). The purity of PhTDH fractions was confirmed by the single band on SDS–PAGE. The protein concentrations were determined with Coomassie protein reagent (Pierce Chemical Company, USA), using bovine serum albumin as the standard protein. Sitedirected mutagenesis was performed by overlap extension PCR method, employing the wild-type PhTDH gene as the template and the mutagenesis primer containing the desired mutation. Purification of the mutant enzymes was performed by the same method described above. To substitute the metal ions, the enzyme was dialyzed against 50 mM sodium phosphate buffer (pH7.5) containing 50 mM EDTA for 12 h and dialyzed against 50 mM sodium phosphate buffer (pH7.5) containing 5 mM of some metals (nickel chloride and gadolinium chloride) for 12 h. Enzyme assay Enzyme activity was determined by the previous method.3 The assay mixture contained variable concentrations of threonine in the presence of several fixed concentrations of NAD+ in 25 mM sodium phosphate buffer (pH 7.5). The mixture was pre-incubated for 5 min at 65 °C and the reaction was initiated by addition of the enzyme solution. NADH formation was monitored by measuring the increase in absorbance at 340 nm spectrophotometry. Values for Km and kcat were determined by a non-linear regression analysis fit to the Michaelis–Menten equation.
Crystallization was performed by the hanging-drop vapor-diffusion method at 277 K. The drops consisted of equal volumes (1.5 μl) of the protein and reservoir solutions. Crystals of SePhTDH were obtained by using a reservoir solution consisting of 0.2 M sodium chloride, 0.1 M Hepes buffer (pH 7.5), and 40% (v/v) PEG-400 over a period of five days.14 Data collection and processing The crystals were scooped up with a Cryo-loop (Hampton Research, Aliso Viejo, CA, USA) and immediately flashcooled at 100 K in a nitrogen cryo-stream. Diffraction data were collected at BL41XU, SPring-8 (Harima, Japan), and processed and scaled with MOSFLM 6.2.434 and SCALA15 from the CCP4 package.15 Selenium sites were searched with SOLVE35 and then the atomic parameters of selenium sites were refined by SHARP.36 Initial phases were calculated by the single wavelength anomalous diffraction (SAD) method by SHARP36 and improved by SOLOMON37 and dm38 implemented in SHARP. The space group and cell parameters of crystal PhTDH were the same as those of PhTDH.14 Refinement up to 2.05 Å was performed with REFMAC539 in the CCP4 package. Superimpositions of the TDH and ADH structures was performed with the Turbo-Frodo program.40 The diffraction data statistics and the crystallographic refinement statistics are summarized in Table 1. The figures of the structures were prepared using MOLSCRIPT.41 Protein Data Bank accession code
Crystallization The purified protein was dialyzed against 50 mM Tris– HCl buffer (pH 7.5) and then concentrated to 10 mg ml−1.
The structure factors and refined coordinates of PhTDH have been deposited with Protein Data Bank; the accession code is 2DFV.
866
Crystal Structure of Threonine Dehydrogenase
Acknowledgements We thank Professor emeritus Noritake Yasuoka of Himeji Institute of Technology and Professor Masahito Taya, Division of Chemical Engineering, Graduate School of Engineering Science, Osaka University for the fruitful discussions and helpful advice. We also thank Dr Mitsuo Ataka, our group leader, for the encouragement to carry out this work. This work was supported by the National Project on Protein Structural and Functional Analyses of the Ministry of Education, Culture, Sports, Science and Technology of Japan. The X-ray diffraction experiments were carried out with the approval of Japan Synchrotron Radiation Research Institute (proposal No.2004B0831-NL1-np-P3k, Japan).
References 1. Gonzalez, J. M., Masuchi, Y., Robb, F. T., Ammerman, J. W., Maeder, D. L. & Yanagibayashi, M. (1998). Pyrococcus horikoshii sp. nov., a hyperthermophilic archaeon isolated from a hydrothermal vent at the Okinawa Trough. Extremophiles, 2, 123–130. 2. Kawarabayasi, Y., Sawada, M., Horikawa, H., Haikawa, Y., Hino, Y. & Yamamoto, S. (1998). Complete sequence and gene organization of the genome of a hyper-thermophilic archaebacterium, Pyrococcus horikoshii OT3. DNA Res. 5, 55–76. 3. Higashi, N., Fukada, H. & Ishikawa, K. (2005). Kinetic study of thermostable L-threonine dehydrogenase from an archaeon Pyrococcus horikoshii. J. Biosci. Bioeng. 99, 175–180. 4. Epperly, B. R. & Dekker, E. E. (1991). L-Threonine dehydrogenase from Escherichia coli. Identification of an active site cysteine residue and metal ion studies. J. Biol. Chem. 266, 6086–6092. 5. Johnson, A. R., Chen, Y. W. & Dekker, E. E. (1998). Investigation of a catalytic zinc binding site in Escherichia coli L-threonine dehydrogenase by sitedirected mutagenesis of cystein-38. Arch. Biochem. Biophys. 358, 211–221. 6. Aoyama, Y. & Motokawa, Y. (1981). L-Threonine dehydrogenase of chicken liver. Purification, characterization, and physiological significance. J. Biol. Chem. 256, 12367–12373. 7. Ray, M. & Ray, S. (1985). L-Threonine dehydrogenase from goat liver. Feedback inhibition by methylglyoxal. J. Biol. Chem. 260, 5913–5918. 8. Aronson, B. D., Somerville, R. L., Epperly, B. R. & Dekker, E. E. (1989). The primary structure of Escherichia coli L-threonine dehydrogenase. J. Mol. Biol. 264, 5226–5232. 9. Reid, M. F. & Fewson, C. A. (1994). Molecular characterization of microbial alcohol dehydrogenases. Crit. Rev. Microbiol. 20, 13–56. 10. Esposito, L., Sica, F., Raia, C. A., Giordano, A., Rossi, M., Mazzarella, L. & Zagari, A. (2002). Crystal structure of the alcohol dehydrogenase from the hyperthermophilic archaeon Sulfolobus solfataricus at 1.85 Å resolution. J. Mol. Biol. 318, 463–477. 11. Esposito, L., Bruno, I., Sica, F., Raia, C. A., Giordano, A., Rossi, M. et al. (2003). Crystal structure of a ternary complex of the alcohol dehydrogenase from Sulfolobus solfataricus. Biochemistry, 42, 14397–14407. 12. Guy, J. E., Isupov, M. N. & Littlechild, J. A. (2003). The
13. 14.
15. 16. 17. 18. 19.
20.
21.
22.
23.
24. 25.
26. 27. 28.
29.
30.
structure of an alcohol dehydrogenase from the hyperthermophilic archaeon Aeropyrum pernix. J. Mol. Biol. 331, 1041–1051. Doublie, S. (1997). Preparation of selenomethionyl proteins for phase determination. Methods Enzymol. 276, 523–530. Higashi, N., Matsuura, T., Nakagawa, A. & Ishikawa, K. (2005). Crystallization and preliminary X-ray analysis of hyperthermophilic L-threonine dehydrogenase from archaeon Pyrococcus horikoshii. Acta Crystallog. sect. F, 61, 432–434. Evans, P. R. (1997). Scala joint CCP4 and ESF-EACBM. Newsletter Protein Crystallography, 33, 22–24. Rossmann, M. G., Moras, D. & Olsen, K. W. (1974). Chemical and biological evolution of nucleotidebinding protein. Nature, 250, 194–199. Bellamacina, C. R. (1996). The nicontinamide dinucleotide binding motif: a comparison of nucleotide binding proteins. FASEB J. 11, 1257–1269. Fan, F. & Plapp, B. V. (1999). Probing the affinity and specificity of yeast alcohol dehydrogenase I for coenzymes. Arch. Biochem. Biophys. 367, 240–249. Korkhin, Y., Kalb (Gilboa), A. J., Peretz, M., Bogin, O., Burstein, Y. & Frolow, F. (1998). NADP-dependent bacterial alcohol dehydrogenases: crystal structure, cofactor-binding and cofactor specificity of the ADHs of Clostridium beijerinckii and Thermoanaerobacter brockii. J. Mol. Biol. 278, 967–981. Eklund, H., Nordstrom, B., Zeppezauer, E., Soderlund, G., Ohlsson, I. & Boiwe, T. (1976). Threedimensional structure of horse liver alcohol dehydrogenase at 2–4 Å resolution. J. Mol. Biol. 102, 27–59. Murshudov, G. N., Vagin, A. A. & Dodson, E. J. (1997). Refinement of macromolecular structures by the maximum-likelihood method. Acta Crystallog. sect. D, 53, 240–255. Eklund, H. & Branden, C. I. (1987). Clystal structure, coenzyme conformations, and protein interactions. In Pyridine Nucleotides (Dolphin, D. et al., eds), pp. 51–98, Wiley & Sons, New York. Cho, H., Ramaswamy, S. & Plapp, B. V. (1997). Flexibility of liver alcohol dehydrogenase in stereoselective binding of 3-butylthiolane 1-oxides. Biochemistry, 36, 382–389. LeBrun, L. A. & Plapp, B. V. (1999). Control of coenzyme binding to horse liver alcohol dehydrogenase. Biochemistry, 38, 12387–12393. LeBrun, L. A., Park, D. H., Ramaswamy, S. & Plapp, B. V. (2004). Participation of histidine-51 in catalysis by horse liver alcohol dehydrogenase. Biochemistry, 43, 3014–3026. Vallee, B. L. & Hoch, F. L. (1955). Zinc a component of yeast alcohol dehydrogenase. Proc. Natl Acd. Sci. USA, 41, 327–338. Veillon, C. & Sytkwski, A. J. (1975). The intrinsic zinc atoms of yeast alcohol dehydrogenase. Biochem. Biophys. Res. Commun. 67, 1494–1500. Eklund, H., Plapp, B. V., Samama, J. P. & Branden, C. I. (1982). Binding of substrate in a ternary complex of horse liver alcohol dehydrogenase. J. Biol. Chem. 257, 14349–14358. Ramaswamy, S., Park, D. H. & Plapp, B. V. (1999). Substitutions in a flexible loop of horse liver alcohol dehydrogenase hinder the conformational change and unmask hydrogen transfer. Biochemistry, 38, 13951–13959. Chen, Y. W., Dekker, E. E. & Somerville, R. L. (1995). Functional analysis of E. coli threonine dehydrogenase by means of mutant isolation and characterization. Biochim. Biophys. Acta, 1253, 208–214.
Crystal Structure of Threonine Dehydrogenase
31. Johnson, A. R. & Dekker, E. E. (1998). Site-directed mutagenesisi of Histidine-90 in Escherichia coli Lthreonine dehydrogenase alters its substrate specificity. Arch. Biochem. Biophys. 351, 8–16. 32. Piersma, S. R., de Vries, S. & Duine, J. A. (1996). In Enzymology and Molecular Biology of Carbonyl Methabolism 6 (Weiner H. et al. eds), pp. 4425–4434, Plenum Press, New York. 33. Tanaka, N., Kusakabe, Y., Ito, K., Yoshimoto, T. & Nakamura, K. T. (2002). Crystal structure of formaldehyde dehydrogenase from Pseudomonas putida: the structural origin of the tightly bound cofactor in nicotinoprotein dehydrogenases. J. Mol. Biol. 324, 519–533. 34. Leslie, A. G. W. (1992). Jnt CCP4/ESF-EACBM. Newsletter Protein Crystallog, 26. 35. Terwilliger, T. (2003). SOLVE and RESOLVE: automated structure solution and density modification. Methods Enzymol. 374, 22–37.
867 36. La Fortella, E. & Bricogne, G. (1997). Maximum-likelihood heavy-atom parameter refinement in the MIR and MAD methods. Methods. Enzymol. 276, 472–494. 37. Abrahams, J. & Leslie, A. (1996). Methods used in the structure determination of bovine mitochondrial F1 ATPase. Acta Crystallog. sect. D, 52, 30–42. 38. Cowtan, K. (1999). Error estimation and bias correction in phase-improvement calculations. Acta Crystallog. sect. D, 55, 1555–15567. 39. Murshudov, G., Vagin, A. & Dodson, E. (1997). Refinement of macromolecular structures by the maximum-likelihood method. Acta Crystallog. sect. D, 53, 240–255. 40. Roussel, A. & Cambillau, C. (1991). Turbo-Frodo. Silicon Graphics Geometry Partners Directory Silicon Graphics, Mountain View, CA. 41. Kraulis, P. J. (1991). MOLSCRIPT: a program to produce both detailed and schematic plots of protein structures. J. Appl. Crystallog, 24, 946–959.
Edited by K. Morikawa (Received 21 July 2006; received in revised form 10 November 2006; accepted 17 November 2006) Available online 21 November 2006