A catalytic mechanism that explains a low catalytic activity of serine dehydratase like-1 from human cancer cells: Crystal structure and site-directed mutagenesis studies

Biochimica et Biophysica Acta 1780 (2008) 809–818

Contents lists available at ScienceDirect

Biochimica et Biophysica Acta j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / b b a g e n

A catalytic mechanism that explains a low catalytic activity of serine dehydratase like-1 from human cancer cells: Crystal structure and site-directed mutagenesis studies ☆ Taro Yamada a, Junichi Komoto a, Tatsuo Kasuya a, Yoshimi Takata a,b, Hirofumi Ogawa b, Hisashi Mori b, Fusao Takusagawa a,⁎ a b

Department of Molecular Biosciences, University of Kansas, 1200 Sunnyside Avenue, Lawrence, KS 66045-7534, USA Department of Molecular Neuroscience, Graduate School of Medicine, University of Toyama, Sugitani, Toyama 930-0194, Japan

a r t i c l e

i n f o

Article history: Received 19 November 2007 Received in revised form 29 January 2008 Accepted 30 January 2008 Available online 19 February 2008 Keywords: Serine dehydratase Isoform Crystal structure Mutagenesis Catalytic mechanism PLP-dependent β-elimination

a b s t r a c t SDH (L-serine dehydratase, EC 4.3.1.17) is a pyridoxal-5′-phosphate (PLP)-dependent enzyme that catalyzes dehydration of L-Ser/Thr to yield pyruvate/ketobutyrate and ammonia. A SDH isoform (cSDH) found in human cancer cell lines has relatively low catalytic activity in comparison with the liver enzyme (hSDH). The crystal structure of cSDH has been determined at 2.8 Å resolution. A PLP is covalently attached to K48 by Schiff-base linkage in the active site. The ring nitrogen of PLP is involved in a H-bonding with C309, but is apparently not protonated. Twenty-three amino residues that compose the active site surfaces were identified. The human and rat liver enzymes (hSDH and rSDH) have the same residues, while residues G72, A172, and S228 in cSDH are replaced with A66, S166, and A222, respectively, in hSDH. These residues in hSDH and cSDH were mutated to make complementary pairs of mutated enzymes, and their kinetic parameters were determined. C303 of hSDH and C309 of cSDH which are H-bonding partner of the ring nitrogen of PLP were mutated to alanine and their kinetic parameters were also determined. The crystal structures and the mutation data suggest that having a glycine at residue 72 of cSDH is the major reason for the reduction of catalytic activity of cSDH. Changing alanine to glycine at residue 72 increases the flexibility of the substrate binding-loop (71S(G/A)GN74), so that the bound substrate and PLP are not pushed deep into the active cleft. Consequently, the proton transfer rate from SG of C309 to N1 of the bound PLP is decreased, which determines the rate of catalytic reaction. © 2008 Elsevier B.V. All rights reserved.

1. Introduction Serine dehydratase (SDH, EC 4.3.1.17; formerly 4.3.1.13) catalyzes the pyridoxal-5′-phosphate-dependent conversion of L-serine (Lthreonine) to pyruvate (α-ketobutyrate) and ammonia [1]. Mammalian SDH is distributed mainly in liver [2]. Pyruvate formation by SDH is a two-step reaction in which the hydroxyl group of serine is cleaved to produce aminoacrylate, and then the aminoacrylate is deaminated by non-enzymatic hydrolysis to produce pyruvate, giving the overall reaction shown below [3].

☆ The atomic coordinates and structure factors have been deposited with the Protein Data Bank (entry name: 2RKB). ⁎ Corresponding author. Tel.: +1 785 864 4727; fax: +1 785 864 5321. E-mail address: [email protected] (F. Takusagawa). Abbreviations: cA172, A172 in cSDH; cG72, G72 in cSDH; cS228, S228 in cSDH; cSDH, SDH-like-1 in cancer cells; cSDH-PLP, PLP linked cSDH; hA222, A222 in hSDH; hA65, A65 in hSDH; hS166, S166 in hSDH; hSDH, human liver SDH; hSDH-PLP, PLP linked hSDH; OMS, O-methylserine; PLP, pyridoxal 5′-phosphate; rSDH, rat liver SDH; rSDH [PLP-OMS], PLP-OMS aldimine-bound rSDH; SDH, serine dehydratase 0304-4165/$ – see front matter © 2008 Elsevier B.V. All rights reserved. doi:10.1016/j.bbagen.2008.01.020

PLP-dependent enzymes are classified into at least three families (α, β, and γ) on the basis of their primary sequences [4]. SDH, which catalyzes α,β-elimination, belongs to the β family. Rat liver contains a relatively high concentration of SDH among the mammalian livers examined so far [5–9]. The enzyme activity fluctuates depending on the available nutrients and in response to the levels of various hormones [2]. The purified rat liver SDH is a dimer with a Mr 34,200 subunit [8,10–12]. PLP binds to K41 to form a Schiff base, and the amino acid sequence, 39SXKIRG44, is well conserved among SDHs from rat [13], human [14], tomato [15], Escherichia coli [16,17], and yeast [18]. The crystal structures of apo-rat SDH (rSDH) and PLP-methylserine aldimine-bound rSDH (rSDH[PLP-OMS]) have been determined [19]. The crystal structure of PLP-bound human liver SDH (hSDH-PLP) was also reported recently [20], and shows that the rat and human enzymes

810

T. Yamada et al. / Biochimica et Biophysica Acta 1780 (2008) 809–818

have similar tertiary structures, as would be expected since they have 84.1% identity in the amino acid sequence [19–21]. The crystal structure of the biosynthetic E. coli SDH (threonine deaminase) was solved previously [22]. The E. coli enzyme is composed of catalytic and regulatory domains, and is regulated by feed-back regulation (i.e., inhibited by the end-product isoleucine). The overall amino acid sequence identity between the prokaryotic and eukaryotic SDHs is very low (25%). However their folding patterns are quite similar [19–22]. A SDH-like-1 gene has been found through the human genome sequence [23]. The mRNA has been found in human cancer cell lines (lung, kidney and brain cancer cells), and the cDNA has been constructed from the isolated mRNA from lung cancer cell lines [24]. The gene product of SDH-like-1 gene is named cSDH (SDH in cancer cells). The cSDH and hSDH have 329 and 328 amino residues, respectively. A sequence alignment indicates that cSDH has 7 extra amino residues in the N-terminal region while hSDH has 5 extra amino residues in the C-terminal and one extra residue (P128) in the middle section. The nucleotide and amino acid sequences between cSDH and hSDH are 36.4% and 58.8% identical, respectively. cSDH has been expressed in E. coli to examine its catalytic activity. cSDH exhibited L-serine and L-threonine dehydratase activity, demonstrating that it is an isoform of SDH. However, the KM and Vmax parameters of cSDH were different from those of the hSDH by two orders of magnitude [24]. The SDH-like-1 gene mRNA occurs in vertebrates across fish, frog, chicken, rat, mouse and human [25] and the physiological role of the protein remains to be determined. Here we report the crystal structure of PLP-bound cSDH and the kinetic parameters of complementary pairs of mutated enzymes of cSDH and hSDH. On the basis of the results, a possible catalytic mechanism that explains the reduction of the catalytic activity of cSDH is proposed. 2. Materials and methods 2.1. Purification and crystallization procedures cSDH used in this study was obtained from human lung cancer cell lines, and the recombinant enzyme was produced in E. coli BL21 transformed with the pET28-SDH plasmid that contains the coding sequence of cSDH cDNA [24]. The enzyme was purified to homogeneity from E. coli extracts by ammonium sulfate precipitation, dialysis in 10 mM potassium phosphate buffer (pH 8.0), followed by DEAE-cellulose chromatography and a Ni-chelating resin column as described previously [24]. The yield was relatively poor, and ~ 1 mg of enzyme was obtained from one liter culture. The hanging-drop vapor-diffusion method was employed for crystallization of the enzyme. Crystals were grown in a solution containing 50 mM Na-citrate buffer (pH 5.6), 10 mM dl-O-methylserine (OMS), 200 mM potassium acetate, 5 mM dithiothreitol, 15% (w/v) PEG-8000 at a protein concentration of 10 mg/mL at 21 °C. Crystals suitable for Xray diffraction studies (~ 0.5 × 0.05 × 0.05 mm) were grown for one week. 2.2. Data measurement A crystal in a hanging-drop was scooped up with a nylon loop and was dipped into a cryo-protectant solution containing 20% (v/v) glycerol, 50 mM Na-citrate buffer (pH 5.6), 200 mM potassium acetate, and 17% (w/v) PEG 8000 for 5 min before the crystal was frozen in cold nitrogen gas (−180 °C) on a Rigaku RAXIS imaging plate X-ray diffractometer with a rotating anode X-ray generator as an X-ray source (CuKα radiation operated at 50 kV and 100 mA). The X-ray beam was focused to 0.3 mm by confocal optics (Osmic, Inc., USA). Although the diffraction data were measured at −180 °C, the crystals decayed relatively quickly so that four crystals were used. The diffraction data were measured up to 2.8 Å resolution at − 180 °C. The individual data were processed with the program DENZO/SCALEPACK [26], and the four data sets were merged by using a locally developed multiple-scaling program [27]. The data statistics are given in Table 1. 2.3. Crystal structure determination of cSDH The unit cell dimensions and space group indicated that an asymmetric unit contained five (or six) subunits with VM = 3.3 Å3 (or 2.7 Å3). The crystal structure was determined by a molecular replacement procedure using CNS [28]. A subunit of hSDH was converted to a poly-alanine model, and was used as the search model. Five subunits were found in the asymmetric unit, in which four subunits form two dimers related with non-crystallographic 2-fold symmetry and the fifth subunit is related by the crystallographic 2-fold symmetry to form the third dimer. The main and side chains

Table 1 Experimental details and refinement parameters of crystal structure analyses Experimental details No. crystals ω-scan angle (°) per frame Resolution range (Å) No. reflections measured Redundancy No. unique reflections Completeness (%) Rmerge b I/σ(I) Refinement parameters No. residues No. PLP molecules No. K+ ions No. water molecules Rcryst c Free R d Mean B-value (Å2)

4 0.5 20.0–2.8 277,357 4.1 57,075 (5,662) a 99.9 (99.9) a 0.157 (0.281) a 14.3 (5.1) a 1590 5 5 282 0.213 (0.267) a 0.230 (0.304) a 25.7

rmsd from the ideal values Bond (Å) 0.007 Angle (°) 1.22 Torsion angle (°) 21.9 Ramachandran plot Residues in most favored region Residues in additional allowed region

91.4% 8.6%

Space group: C2221; cell dimension: a = 97.21 Å, b = 154.74 Å, c = 306.37 Å, Mr of subunit: 34652; no. subunits in the unit cell: 40; VM = 3.32 Å3; percentage of solvent content: 63% (v/v). a Highest resulution shell = 2.8–2.9 Å resolution. b Rmerge = ∑h∑i|Ihi − bIhN| / ∑h∑i|Ihi|. c Rcryst = Σ|Fo − Fc| / Σ|Fo|. d Free R was calculated with 10% reflections randomly selected.

were fit into the resulted electron density. The structure was refined with the simulated annealing procedure of CNS. (2Fo–Fc) maps calculated after refinement showed that several sections of the polypeptide had different conformations from those of hSDH. (Fo–Fc) maps showed a large significant residual electron density peak in the region of the active site. Since the crystal of cSDH had a yellowish color, PLP was fitted into the electron density peak. One peak in each subunit that was significantly higher than those of water molecules was assigned to a potassium ion since the crystals were grown in the solution containing potassium ions (200 mM potassium acetate). Other well-defined residual electron density peaks in (Fo–Fc) maps were assigned to water molecules if peaks were able to bind the protein molecules with H-bonds. The final model was refined by the simulated annealing procedure, and followed by the individual B-factor refinement procedure of CNS. During the refinement, the five subunits related by a non-crystallographic symmetry were tightly restrained to have the same structure in order to increase the accuracy of coordinates. The final crystallographic R-factor was 0.213 for observed data (no σ cut off) from 20 to 2.8 Å resolution. The free-R for 10% randomly selected data was 0.230. Crystallographic parameters are listed in Table 1. 2.4. Site-directed mutagenesis, enzyme purification, and assay Site-directed mutagenesis was carried out using a QuickChange site-directed mutagenesis kit (Stratagene) as described previously [24]. The nucleotide sequences obtained were examined by an ABI-Hitachi 3100 DNA sequencer. The cDNA and hSDH in a pET28 vector (Novagen) were expressed in E. coli BL21 (DE3). The mutated enzymes were purified as described previously [24]. 2.5. Enzyme assay and kinetics The SDH activity was measured colorimetrically as described previously [24]. The reaction mixture (250 μL) consisted of 50 mM borate-KOH (pH 8.3), 100 μM PLP, x mM substrate (x = 1.0–50), and 50 nM enzyme. Enzyme kinetics was performed at least three times using the mutated enzyme from the same or different preparations.

3. Results and discussion 3.1. Overall crystal structure The crystallographic refinement parameters (Table 1), final (2Fo–Fc) and (Fo–Fc) maps and conformational analysis by PROCHECK [29] indicate that the structure of cSDH has been determined with acceptable statistics. A crystallographic asymmetric unit contains five subunits. Two sets of two subunits interact strongly and form two non-crystallographic dimers, while the fifth subunit is related by a crystallographic 2-fold

T. Yamada et al. / Biochimica et Biophysica Acta 1780 (2008) 809–818

811

Fig. 1. A dimer structure of cSDH showing two subunits related by non-crystallographic 2-fold symmetry. Each subunit contains a PLP illustrated with ball and stick presentation. The large domains (residues 1–42 and 143–329) in each subunit are colored aquamarine and light pink, while the small domains (residues 43–142) are illustrated by blue and magenta. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

symmetry to form the third dimer (Fig. 1). Each subunit is composed of two domains (large domain (residues 1–42 and 143–329), and small domain (residues 43–142)). The active site is located between the small and large domains, and a PLP is attached to K48 by a Schiff-base linkage (Fig. 2). In the dimer of cSDH, the large domain in each subunit interacts with each other and forms the core while the small domain does not interact with the partner subunit so that the small domain is movable but the large domain is not. The active site is widely opened, indicating that cSDH has an open conformation. The N-terminal residues 1–10 and the C-terminal residue 329 of each subunit are not defined due to a heavily disordered structure. Since X-ray crystal structures at 2.8 Å resolution do not provide the proton positions, the protonation/nonprotonation status of the PLP N1 is predicted by carefully examining the N1 involved H-bond networks and the pKa values of the surrounding amino acid residues. 3.2. The active site contents determine the enzyme conformation The crystal structures of human liver and rat liver SDH (hSDH and rSDH) have been determined, and their coordinates are available for comparison with cSDH [19,20]. cSDH and hSDH are composed of 329 and 328 residues, respectively, and a sequence alignment indicates that cSDH and hSDH align well [24]. cSDH has seven extra amino residues in the N-terminus, while hSDH has five extra amino residues in the C-terminus. P128 is an extra residue in hSDH and there is one gap between residue 134 and 135 of cSDH. The peptide folds of the known SDH's were compared by using the program SARF [30]. As shown in Fig. 3, the peptide folds of cSDH and hSDH are very similar although the sequence homology is relatively low (60%). Except for

N126 of hSDH, the 317 CA atoms of cSDH and hSDH are superimposable with the rmsd of 0.96 Å, suggesting that the extra residue in hSDH is N126 but P128 predicted by a sequence analysis. The sequence homology between hSDH and rSDH is much higher (85%) than that between hSDH and cSDH (60%). However, the tertiary structures of hSDH and cSDH are quite similar while those of hSDH and rSDH are significantly different (rmsd of CA's are 0.96 Å vs. 1.55 Å). The bound PLP in the cSDH and hSDH structures is covalently linked to a lysine residue in the active site, and thus the crystal structures of cSDH and hSDH represent a holo-SDH structure (i.e., SDH-PLP). The small domain of cSDH is in the open position so that the active site cleft is largely opened (Fig. 4). On the other hand, the bound PLP in the rSDH structure is covalently connected to an inhibitor OMS to form a PLP-OMS aldimine, and thus the crystal structure represents a ternary complex (rSDH[PLP-OMS]). The small domain is moved in order to close the active site cleft, and thus rSDH has a closed conformation. In the structure of apo-rSDH, a loop of the large domain moves significantly (Fig. 4), while the small domain movement is minimal in comparison with holo-cSDH, so that apo-rSDH has an open conformation. From the structures of cSDH, hSDH, and rSDH, upon the substrate binding, a loop (residues 69–75) changes the conformation and the small domain moves to close the catalytic cleft. Therefore, the SDH structures are altered by the content in the active site rather than the amino residue sequences. 3.3. Active site geometry The phosphate moiety of the K48 linked PLP is surrounded by five amide groups of the conserved amino residues 174GGGGL178 (phosphate

Fig. 2. A (Fo–Fc) map showing the residual electron density of the Schiff-base linked PLP to K48. The final model of the PLP-K48 is superimposed, and the contour is drawn at the 2.0 σ level.

812

T. Yamada et al. / Biochimica et Biophysica Acta 1780 (2008) 809–818

Fig. 3. A superimposed view of cSDH (magenta) and hSDH (aquamarine). N126 and P128 in hSDH are marked. As shown below, sequence and structural alignments suggest the different gap and extra residues:

(For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

binding-loop). The two oxygens (O1P, O2P) of the phosphate group form three H-bonds with the amide groups of G174, G176, G177, and L178, and two bound waters, and the third oxygen (O3P) has two H-bonds with the amide group of G175 and a bound water (w1). O3 of PLP is apparently deprotonated and forms two H-bonds with the linked NZ of K48 and ND2 of N74. The pyridine ring of PLP is fit into a “pyridine ring pocket” and N1 of PLP is involved in a H-bond with SG of C309. The same PLP-binding mode is observed in the hSDH-PLP holo-enzyme and rSDH[PLP-OMS] ternary structures [19,20]. A potassium ion is coordinated by five oxygen and one sulfur atoms (O[G173], O[G174], O[A204], O[L229], OE2[E200], and SG[C206]) near the active site (Fig. 5A). The major function of this ion is apparently to build the framework of the phosphate binding-loop but is not involved in catalysis. A K+ ion is located at the same site in the rSDH [PLP-OMS] ternary structure [19], but not reported in the hSDH-PLP holo-enzyme structure [20]. Similar K+ ions were found in the structures of β-elimination enzymes, tryptophan synthase [31] and tryptophanase [32]. On the basis of the rSDH[PLP-OMS] ternary structure, the substrate (serine/threonine) bound model of cSDH was built (Fig. 5B). The hydroxyl group (OG) of serine/threonine is predicted to occupy the water

binding site (w1), where OG is able to form two H-bonds with the PLP phosphate oxygen and the carbonyl oxygen (O) of S228. In the catalytic pathway, the substrate serine forms the geminal diamine intermediate, and then PLP-Ser aldimine. Since the –O–CH3 of the inhibitor OMS cannot become a water to be released from the PLPOMS aldimine, OMS was trapped in the active site of rSDH as PLP-OMS aldimine [19]. Although cSDH was crystallized in the presence of 10 mM OMS, the bound PLP did not form the PLP-OMS aldimine as observed in rSDH[PLP-OMS]. This observation indicates that the Schiff-base linkage between the bound PLP in cSDH is relatively stable, and thus the bound PLP in cSDH has a weaker reactivity than those in hSDH and rSDH. The differences of the bound PLP reactivity can be explained on the basis of the active site structures of cSDH, hSDH, and rSDH. 3.4. Substituted residues on the surface of the active site cleft Amino residues within 4.0 Å from the linked PLP or the bound PLPOMS are assumed to compose the active site cleft, and those residues

Fig. 4. Three SDH structures showing that the contents of the active site change the conformation of enzyme. (A) apo-rSDH, (B) holo-cSDH-PLP, and (C) ternary rSDH[PLP-OMS]. The substrate binding-loops are illustrated thick red coils. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

T. Yamada et al. / Biochimica et Biophysica Acta 1780 (2008) 809–818

813

Fig. 5. (A) Possible interactions between cSDH and PLP in the active site. The bound PLP is connected to K48 with a Schiff-base linkage. Possible hydrogen bonds between PLP and cSDH are illustrated by thin lines. The altered residues G72, A172, and S228 are marked along with the other important residues. OG of S166 of hSDH is shown with open bond at A172. O[L229], OE2[E200], and SG[C206] that coordinate to K+ are omitted for clarity. (B) A model structure of PLP-Thr aldimine-bound cSDH deduced from the rSDH[PLP-OMS]. The threonine moiety is built by attaching a methyl group to CB of OMS and removing the methyl group attached to OG of OMS. The CA–CB bond is rotated until OG fits into the water binding site (w1). (C) A superimposed view of the bound PLP in cSDH-PLP (aquamarine) and rSDH[PLP-OMS] (light pink). The 217 CA atoms of large domains of rSDH are superimposed on the corresponding CA atoms of cSDH by least-squares method (rmsd = 0.65 Å), and the obtained rotation and translation parameters are applied to the bound PLPOMS aldimine in rSDH. The N1···SG distances in cSDH and rSDH are 3.6 and 2.9 Å, respectively. The pyridine ring of PLP of rSDH is significantly tilted from that of cSDH (inter-plane angle = 26°). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

814

T. Yamada et al. / Biochimica et Biophysica Acta 1780 (2008) 809–818

Table Table 22 Amino acid residues that compose of of the the surface surface of of active activesite sitecleft, cleft,which whichare areless lessthan than 4.0 Å from the bound PLP-OMS PLP-OMS in in the the rSDH[PLP-OMS] rSDH[PLP-OMS] complex complex or or the the linked linked PLP PLP in in cSDH-PLP and hSDH-PLP

were identified in the cSDH-PLP, hSDH-PLP, and rSDH[PLP-OMS] structures (Table 2). Twenty-three residues were identified, and all of them in hSDH and rSDH are the same (100% homology), while three residues (G72, A172, and S228) in cSDH are different from those in hSDH (87% homology). The amino acid homology of the active site cleft is significantly higher than that in the other parts (for example, 87% vs. 57% for cSDH–hSDH). The tertiary ternary structure of active site cleft is also well conserved in cSDH and hSDH. For example, the rmsd of CA positions of these residues between cSDH-PLP and hSDHPLP is 0.47 Å, and the side chains of cSDH and hSDH are also superimposable within experimental errors. Since all amino acid residues forming the active site surface are the same in hSDH and rSDH, the catalytic activities of hSDH and rSDH are expected to be similar. Indeed, the reported kinetics parameters of hSDH (KM = 67 mM and kcat = 61 s− 1) [24] and rSDH (KM = 57 mM and kcat = 66 s− 1) [6] are the same, suggesting that the active sites of hSDH and rSDH have the same characteristic nature. On the other hand, the kinetics parameter of cSDH (KM = 32 mM and kcat = 4.5 s− 1) [24] is significantly different from those of hSDH and rSDH, suggesting that the three altered amino acid residues on the surface of the active site might be responsible for the reduction of the catalytic activity of cSDH. 3.5. Possible roles of three altered amino acid residues Amino acid residues 64SAGNA68 in hSDH and 71SGGNA75 of cSDH form a loop, and are on surface of the active site (Fig. 5). The amide groups of residues 65–67 of hSDH and residue 72–74 of cSDH point their all amide hydrogens on the inside surface of the cleft. Furthermore, the hydroxyl OG of cS71 (hS64) that forms a H-bond with the amide group of cA75 (hA68) points the hydroxyl hydrogen on surface of the cleft (Fig. 5A). As a result, four partially positively charged hydrogens are concentrated in the small loop and exposed on the active site surface. The negatively charged carboxyl group of the substrate is expected to interact with the loop. Indeed, the carboxyl group of OMS forms H-bonds with these hydrogens of the loop in rSDH [PLP-OMS] [19]. So, the 64SAGNA68 loop in hSDH and 71SGGNA75 loop in cSDH are substrate binding-loops. On the basis of the structures of cSDH, hSDH, and rSDH, SDH assumes a closed conformation upon the substrate binding to the substrate binding-loop. The loop apparently changes conformation upon binding the substrate in order to maximize H-bonding with the substrate (Figs. 5B and 6). The bound

substrate is brought into the active site by movement of the substrate binding-loop as if it is a “lid” on the active site. Therefore, the loop has to have an exact conformation and flexibility to bring the substrate to the correct location in the active site, in order to form a geminal diamine intermediate. Obviously, the 71SGGNA75 loop of cSDH is more flexible than the 64SAGNA68 loop of hSDH because the ϕ and ψ angles of glycine have less restriction than those of alanine. Therefore, the loop of cSDH might still be dynamic and relatively flexible even when cSDH has a closed conformation, and consequently, the catalytic efficiency (kcat/KM) of cSDH would be lower than that of hSDH. As shown in Fig. 5A, S166 of hSDH (A172 of cSDH) is located deep in the active site cleft and the hydroxyl group of S166 apparently participates in H-bonding with SG of C303. On the other hand, the corresponding A172 of cSDH does not have a side chain to form a Hbond with SG of C309. SG of cC309/hC303 participates in H-bonds with N1 of the bound PLP and SG of cC276/hC270, which is also involved in a H-bond with the amide group of cG310/hG304 (Figs. 5 and 7). In general, the pKa value of the sulfhydryl group (-SH) of cysteine is lowered by a H–SG···H–SG′ H-bonding. In hSDH, since SG of C303 is surrounded by two partially positively charged H-bond donors (S166 and C270), the SG tends to be deprotonated, and consequently, the Hbond partner N1 of the bound PLP would be protonated. The protonated PLP increases the reactivity to form a geminal diamine intermediate and the protonated PLP-Ser/Thr aldimine can form a resonance-stabilized carbanion by releasing the α-H, which stimulates β-elimination of OH− (Fig. 7C). On the other hand, since A172 in cSDH cannot polarize SG of C309, the pKa value of SG might not be as low as that of C303 in hSDH. Consequently, the dehydration reaction would be slowed down. The protonation/deprotonation of PLP should be strongly related to the reactivity of the Schiff-base linked PLP. In the crystallization experiments, OMS forms a PLP-OMS aldimine with the bound PLP in rSDH/hSDH but not in cSDH. This observation might be due to the change in the amino acid residue (S → A) at 172. In cSDH, the pyridine ring of PLP binds deep in the active site cleft, and a loop (residues 227–229) is located above the ring, which participates in building the framework of the pyridine ring pocket (Figs. 5 and 8). OG of S228 forms a H-bond with OE1 of E273, which also forms a H-bond with ND2 of N74 in the small domain. On the other hand, the corresponding residue in hSDH/rSDH is alanine (A222), which is not able to form a H-bond with E266. Therefore, the [S228] OG–H···OE2[E273] H-bonding in cSDH might increase the rigidity of the pyridine ring pocket of cSDH. 3.6. Kinetics parameters of the mutated enzymes As described above, three pairs of the altered residues, (hA65 vs. cG72), (hS166 vs. cA172), and (hA222 vs. cS228), could explain the difference in catalytic activity between hSDH and cSDH. In order to determine the relative contributions to the catalytic activity, sitedirected mutagenesis experiments were performed. Six single mutated

Fig. 6. Two different loop conformations of 71S(G/A)GNA seen in cSDH-PLP and rSDH[PLP-OMS]. Possible H-bonds are illustrated by thin lines.

T. Yamada et al. / Biochimica et Biophysica Acta 1780 (2008) 809–818

815

Fig. 7. Schematic diagrams showing the serine to alanine change at residue 172 in cSDH. (A) Possible H-bonding networks around C309 of cSDH, (B) Possible H-bonding networks around C303 of hSDH/rSDH, and (C) Stable resonance structures of PLP-Ser aldimine during the dehydration reaction.

enzymes and two double mutated enzymes were constructed and purified, and their kinetics parameters were determined by using both serine and threonine as the substrate. In addition to these mutated enzymes, hC303A and cC309A mutated enzymes were prepared. The kinetics parameters of four complementary pairs of mutated enzymes are listed along with those of the wild type enzymes and the hC303A and cC309A mutated enzymes in Table 3. A large effect was only observed in the complementary pair of hA65G and cG72A mutation for the substrate threonine. The cG72A mutation brought the catalytic efficiency (kcat/KM) up to the level observed with the hSDH. Likewise the hA65G mutation reduced the catalytic efficiency to that observed for the cSDH, suggesting that changing alanine to glycine at residue 72 in cSDH is responsible for the reduced catalytic activity of cSDH. However, for the substrate serine, the complementary pair of hA65G and cG72A mutation does not

Fig. 8. A schematic drawing of H-bond networks around S228, which participates in building the framework of the pyridine ring pocket. The loop over the pyridine ring of PLP is illustrated by a thick line.

exhibit significant effect. As will be discussed below, this discrepancy might be due to the different molecular sizes of the substrates. The similar effect was observed in the double mutation pair (cG72A/S228A and hA65G/A222S), confirming the results of the hA65G and cG72A mutations.

Table 3 Kinetics parameters of hSDH, cSDH, and their mutated enzymes Enzyme

KM (S) a mM

KM (T) a mM

kcat (S) a s− 1

kcat (T) a s− 1

kcat/KM (S) a

kcat/KM (T) a

hSDH cSDH

23 ± 1 30 ± 2

31 ± 17 7±1

79 ± 7 10 ± 3

71 ± 13 1.3 ± 0.7

3.5 × 103 (1.00) b 3.3 × 102 (1.00)

2.3 × 103 (1.00) b 1.7 × 102 (1.00)

hA65G cG72A

49 ± 5 30 ± 8

43 ± 3 4±1

11 ± 4 4±1

4±2 15 ± 6

2.2 × 102 (0.06) 1.4 × 102 (0.43)

8.7 × 10 (0.04) 3.5 × 103 (20.0)

hS166A cA172S

56 ± 3 40 ± 3

50 ± 2 8±1

105 ± 6 7±2

82 ± 9 1.3 ± 0.6

1.9 × 103 (0.54) 1.6 × 102 (0.49)

1.6 × 103 (0.71) 1.5 × 102 (0.88)

hA222S cS228A

22 ± 6 33 ± 6

23 ± 3 19 ± 5

62 ± 21 1.5 ± 0.4

61 ± 21 0.3 ± 0.1

2.9 × 103 (0.82) 4.6 × 101 (0.14)

2.6 × 103 (1.14) 1.5 × 101 (0.09)

hA65G/ A222S cG72A/ S228A

22 ± 5

22 ± 4

24 ± 8

10 ± 1

1.0 × 103 (0.30)

4.1 × 102 (0.19)

35 ± 8

9±3

13 ± 1

12 ± 3

3.8 × 102 (1.15)

1.4 × 103 (8.0)

hC303A cC309A

133 ± 13 105 ± 7

7±2 0.1 ± 0.1

1.7 ± 0.9 0.1 ± 0.1

50 (0.014) 1.3 (0.004)

16 (0.007) 0.7 (0.004)

108 ± 4 125 ± 9

The complementary pairs are separated by lines, and significant complementary effects are underlined. a S and T in parenthesis indicate substrate serine and Thr, respectively. b Relative catalytic efficiency on the basis of the wild type enzyme (hSDH or cSDH).

816

T. Yamada et al. / Biochimica et Biophysica Acta 1780 (2008) 809–818

The hC303A and cC309A mutations abolish the catalytic activity, indicating that N1 protonation of the bound PLP by hC303/cC309 is an essential process to initialize the β-elimination reaction. The hS166A and cA172S mutation did not change significantly the catalytic activity for both the substrates, suggesting that the [S166]OG–H··· SG[C303] H-bonding in hSDH does not decrease the pKa of C303 to stimulate the N1 protonation of the bound PLP. The mutation results indicate that there is an unknown mechanism to protonate N1 during the catalytic reaction. The hA222S mutation did not alter the catalytic activity for either substrate, indicating that increasing the rigidity of the pyridine ring pocket does not increase the catalytic activity. However, the cS228A mutation significantly decreased the catalytic activity, suggesting that a rigid pyridine ring pocket is necessary for cSDH having a flexible substrate binding-loop. In the double mutation (cG72A-S228A), the catalytic activities for both the substrates were substantially recovered, suggesting that a proper substrate binding-loop of the active site compensates for the loosed pyridine ring pocket. 3.7. The PLP-dependent β-elimination enzymes are divided into two groups Crystal structures of PLP-dependent β-elimination enzymes are known. Some of those are: SDH [19,20], threonine deaminase [22,33], cystathionine β-synthase [34,35], tryptophan synthase [31,36–48], Oacetylserine sulfhydrylase [49–57], O-phosphoserine sulfhydrylase [58], Sep-tRNA:Cys-tRNA synthase [59], cystathionine β-lyase [60–63], cystine lyase [64–68], tyrosine phenyl-lyase [69,70], 4-amino-4deoxychorismate lyase [71], cysteine desulfurase/selenocysteine lyase [72–77], and tryptophanase [32,78]. The overall structures of the PLP-binding domains of these enzymes are quite similar to each other. However, these enzymes are clearly divided into two groups by the H-bond partner residue of N1 of the bound PLP. Cystathionine βlyase, cystine lyase, tyrosine phenyl-lyase, 4-amino-4-deoxychorismate lyase, cysteine desulfurase/selenocysteine lyase, and tryptophanase belong to the first group that has either aspartic acid or glutamic acid for the N1 H-bond partner. In this group, N1 is protonated and forms a +N1–H···O− H-bond. The bound PLP can have a quinone form before the catalytic reaction. On the other hand, threonine deaminase, cystathionine β-synthase, tryptophan synthase, O-acetyleserine sulfhydrylase, O-phosphoserine sulfhydrylase, Sep-tRNA:Cys-tRNA synthase, and SDH belong to the second group that has a neutral amino acid residue (serine, asparagine, or cysteine) for the N1 H-bond partner. The pKa value of N1 of PLP aldimine is recently determined to be 5.8 [79] by NMR measurement, suggesting N1 of the bound PLP would not be protonated by the neutral amino acid residues (serine (pKa N 14), cysteine (pKa ≈ 8.5), asparagine (pKa N 14)) under a normal condition. Therefore, we proposed that the catalytic reaction of SDH does not require the N1 protonation of the bound PLP [19]. However, the hC303A and cC309A mutations abolish the catalytic activity, indicating that N1 protonation of the bound PLP by hC303/cC309 is an essential process to initialize the β-elimination reaction, and thus our proposed mechanism is wrong. 3.8. A possible catalytic mechanism that explains the reduction of the catalytic activity of cSDH The hS166A and cA172S mutations indicate that the [S166]OG–H···SG [C303] H-bonding in hSDH does not polarize SG of hC303 to produce a thiolic anion (S−G) and to protonate N1 of the bound PLP. Therefore, N1 of the bound PLP is protonated during the catalytic reaction by an unknown mechanism. The hydrogen attached to SG can be transferred to N1 in the [PLP]N1···H–SG[Cys] H-bonding if the [PLP]N1···SG[Cys] distance becomes short [80–82]. In the SDH case, the bound PLP and substrate are apparently pushed deep into the active site cleft by a movement of the small domain, so that [PLP]N1···SG[cC303/cC309]

distance becomes short and the proton on SG can be transferred to N1 of the bound PLP. In the cSDH-PLP and rSDH[PLP-OMS] structures, the bound PLP in rSDH is indeed moved by 0.7 Å toward SG of C303 in comparison with the PLP in cSDH (Fig. 5C). Dynamical movements of the small domains of cSDH and hSDH/ rSDH are slightly different because their amino acid sequences are different, and flexibilities of the substrate binding-loops of cSDH and hSDH/rSDH are also different. Since the catalytic activity of cSDH is weaker than those of hSDH/rSDH, the small domain movement of cSDH is smaller than those of hSDH/rSDH (i.e., the PLP pushing power of cSDH is weaker than those of hSDH/rSDH). With this dynamical mechanism, we can explain why the complementary pair of the hA65G and cG72A mutation exhibits a large effect for the substrate threonine but not for the substrate serine. The substrate (serine or threonine) binds between the small domain and the PLP-bound large domain of cG72A mutated enzyme that has an open conformation. Since threonine is slightly bulkier than serine (because it has an additional methyl group at CB), the serine bound active site cleft has a larger gap than the threonine bound cleft does. As described above, the small domain movement of cSDH is assumed to be smaller than those of hSDH/rSDH. Therefore, even the substrate binding-loop recovered its rigidity by G72A mutation, the small domain of cG72A mutated enzyme cannot push the bound PLP deep enough into the active site cleft when the small substrate (serine) is bound. As a result, the degree of the N1 protonation is reduced and resulting in a lower catalytic activity compared to what is obtained with threonine as the substrate. In summary, the crystal structure and mutagenesis studies suggest that substitution of alanine to glycine at residue 72 of cSDH decreases the rigidity of the substrate binding-loop (residues 69–75), and substitutions of unidentified residues in the small domain reduce movement of the small domain during the catalytic reaction. The following chain events would occur in cSDH: The flexible substrate binding-loop absorbs a fraction of the pushing power generated by the movement of the small domain, the bound substrate and PLP are not pushed deep into the active site cleft, N1 of the bound PLP is not brought near SG of cC309, the rate of proton transfer from SG to N1 is reduced, and consequently, the dehydration rate of the substrate is decreased. Acknowledgments The work has been partially supported by a grant from the Arthritis Foundation (Grant No. 50538). We express our thanks to Professor Richard H. Himes for the critical reading of this manuscript and very valuable comments. References [1] H. Holzer, C. Cennamo, M. Boll, Product activation of yeast threonine dehydratase by ammonia, Biochem. Biophys. Res. Commun. 14 (1964) 487–492. [2] K. Snell, Enzymes of serine metabolism in normal, developing and neoplastic rat tissues, Adv. Enzyme Regul. 22 (1984) 325–400. [3] H. Reiber, Proceedings: alpha, beta-elimination of serine: mechanism and catalysis of a model reaction and the consequences for the active centre of serine dehydratases, Hoppe-Seyler Z. Physiol. Chem. 355 (1974) 1240. [4] F.W. Alexander, E. Sandmeier, P.K. Mehta, P. Christen, Evolutionary relationships among pyridoxal-5′-phosphate-dependent enzymes. Regio-specific α, β, and γ families, Eur. J. Biochem. 219 (1994) 953–960. [5] H. Ogawa, Structure and function relationships of serine dehydratases from various sources, Trends Comp. Biochem. Physiol. (2000) 1–19. [6] H. Nakagawa, H. Kimura, The properties of crystalline serine dehydratase of rat liver, J. Biochem. 66 (1969) 669–683. [7] H. Inoue, C. Kasper, H.C. Pitot, Studies on the induction and repression of enzymes in rat liver: VI. Some properties and the metabolic regulation of two isozymic forms of serine dehydratase, J. Biol. Chem. 246 (1971) 2626–2632. [8] H. Ogawa, T. Gomi, F. Takusagawa, T. Masuda, T. Goto, T. Kan, N.H. Huh, Evidence for a dimeric structure of rat liver serine dehydratase, Int. J. Biochem. Cell Biol. 34 (2002) 533–543. [9] T. Masuda, H. Ogawa, T. Matushima, S. Kawamata, M. Sasahara, K. Kuroda, Y. Suzuki, Y. Takata, M. Yamazaki, F. Takusagawa, H.C. Pitot, Localization and hormonal control of serine dehydratase during metabolic acidosis differ markedly from those of phosphoenolpyruvate carboxykinase in rat kidney, Int. J. Biochem. Cell Biochem. 35 (2003) 1234–1247.

T. Yamada et al. / Biochimica et Biophysica Acta 1780 (2008) 809–818 [10] H. Ogawa, D.A. Miller, T. Dunn, Y. Su, J.M. Burcham, C. Peraino, M. Fujioka, K. Babcock, H.C. Pitot, Isolation and nucleotide sequence of the cDNA for rat liver serine dehydratase mRNA and structures of the 5′ and 3′ flanking regions of the serine dehydratase gene, Proc. Natl. Acad. Sci. U. S. A. 85 (1988) 5809–5813. [11] H. Ogawa, M. Fujioka, T. Date, M. Mueckler, Y. Su, H.C. Pitot, Rat serine dehydratase gene codes for two species of messenger RNA of which only one is translated into serine dehydratase, J. Biol. Chem. 265 (1990) 14407–14413. [12] R. Leoncini, A. Henschen, K. Krieglstein, J. Calvete, R. Pagani, E. Marinello, Primary structure of rat liver L-threonine deaminase, Ital. J. Biochem. 39 (1990) 228–234. [13] H. Ogawa, K. Konishi, M. Fujioka, The peptide sequences near the bound pyridoxal phosphate are conserved in serine dehydratase from rat liver and threonine dehydratases from yeast and Escherichia coli, Biochim. Biophys. Acta 996 (1989) 139–141. [14] H. Ogawa, T. Gomi, K. Konishi, T. Date, H. Nakashima, K. Nose, Y. Matsuda, C. Peraino, H.C. Pitot, M. Fujioka, Human liver serine dehydratase: cDNA cloning and sequence homology with hydroxyamino acid dehydratases from other sources, J. Biol. Chem. 264 (1989) 15818–15823. [15] A. Samach, D. Hareven, T. Gutfinger, S. Ken-Dror, E. Lifschitz, Biosynthetic threonine deaminase gene of tomato: isolation, structure, and upregulation in floral organs, Proc. Natl. Acad. Sci. U. S. A. 88 (1991) 2678–2682. [16] P. Datta, T.J. Goss, J.R. Omnaas, R.V. Patil, Covalent structure of biodegradative threonine dehydratase of Escherichia coli: homology with other dehydratases, Proc. Natl. Acad. Sci. U. S. A. 84 (1987) 393–397. [17] E. Eisenstein, H.D. Yu, K.E. Fisher, D.A. Iacuzio, K.R. Ducote, F.P. Schwarz, An expanded two-state model accounts for homotropic cooperativity in biosynthetic threonine deaminase from Escherichia coli, Biochemistry 34 (1995) 9403–9412. [18] C. Bornaes, J.G.L. Petersen, S. Holmberg, Serine and threonine catabolism in Saccharomyces cerevisiae: the CHA1 polypeptide is homologous with other serine and threonine dehydratases, Genetics 131 (1992) 531–539. [19] T. Yamada, J. Komoto, Y. Takata, H. Ogawa, H.C. Pitot, F. Takusagawa, Crystal structure of serine dehydratase from rat liver, Biochemistry 42 (2003) 12854–12865. [20] L. Sun, M. Bartlam, Y. Liu, H. Pang, Z. Rao, Crystal structure of the pyrodxal-5′phosphate-dependent serine dehydratase from human liver, Protein Sci. 14 (2005) 791–798. [21] T. Kashii, T. Gomi, T. Oya, Y. Ishii, H. Oda, M. Maruyama, M. Koabayashi, T. Masuda, M. Yamazaki, T. Nagata, K. Tsukada, A. Nakajima, K. Tatsu, H. Mori, F. Takusagawa, H. Ogawa, H.C. Pitot, Some biochemical and histochemical properties of human liver serine dehydratase, Int. J. Biochem. Cell Biol. 37 (2005) 574–589. [22] D.T. Gallagher, G.L. Gilliland, G. Xiao, J. Zondlo, K.E. Kisher, D. Chinchilla, E. Eisenstein, Structure and control of pyridoxal phosphate dependent allosteric threonine deaminase, Structure 15 (1998) 465–475. [23] J.C. Venter, M.D. Adams, E.W. Myers, W. Li, P.R.J. Mural, et al., The sequence of the human genome, Science 291 (2001) 1304–1351. [24] H. Ogawa, T. Gomi, M. Nishizawa, Y. Hayakawa, S. Endo, K. Hayashi, H. Ochiai, F. Takusagawa, H.C. Pitot, H. Mori, H. Sakurai, K. Koizumi, I. Saiki, H. Oda, T. Fujishita, T. Miwa, M. Maruyama, M. Kobayashi, Enzymatic and biochemical properties of a novel human serine dehydratase isoform, Biochim. Biophys. Acta 1764 (2006) 961–971. [25] A.J. Edgar, Mice have a transcribed L-threonine aldolase/GLY1 gene, but the human GLY1 gene is a non-processed pseudogene, BMC Genomics 6 (2005) 32–44. [26] Z. Otwinowski, W. Minor, Processing of X-ray diffraction data collected in oscillation mode, Methods Enzymol. 276 (1997) 307–326. [27] F. Takusagawa, A combined isotropic and multiple s-shell anisotropic scaling method for multiple data sets, J. Appli. Cryst. 25 (1992) 26–30. [28] A.T. Brünger, P.D. Adams, G.M. Clore, W.L. DeLano, P. Gros, R.W. Grosse-Kunstleve, J.S. Jiang, J. Kuszewski, M. Nilges, N.S. Pannu, R.J. Read, L.M. Rice, T. Simonson, G.L. Warren, Crystallography and NMR system: a new software suite for macromolecular structure determination, Acta Crystallogr. D54 (1998) 905–921. [29] R.A. Laskowski, M.W. MacArthur, D.S. Moss, J.M. Thornton, ROCHECK: a program to check the stereochemical quality of protein structures, J. Appl. Cryst. 26 (1993) 283–291. [30] N.N. Alexandrov, SARFing the PDB, Protein Eng. 9 (1996) 727–732. [31] C.C. Hyde, S.A. Ahmed, E.A. Padlan, E.W. Miles, D.R. Davies, Three-dimensional structure of the tryptophan synthase α2β2 multienzyme complex from Salmonella typhimurium, J. Biol. Chem. 263 (1988) 17857–17871. [32] M.N. Isupov, A.A. Antson, E.J. Dodson, G.G. Dodson, I.S. Dementieva, L.N. Zakomirdina, K.S. Wilson, Z. Dauter, A.A. Lebedev, E.H. Harutyunyan, Crystal structure of tryptophanase, J. Mol. Biol. 276 (1998) 603–623. [33] D.K. Simanshu, H.S. Savithri, M.R. Murthy, Crystal structures of Salmonella typhimurium biodegradative threonine deaminase and its complex with CMP provide structural insights into ligand-induced oligomerization and enzyme activation, J. Biol. Chem. 281 (2006) 39630–39641. [34] M. Meier, M. Janosik, V. Kery, J.P. Kraus, P. Burkhard, Structure of human cystathionine β-synthase: a unique pyridoxal 5′-phosphate-dependent heme protein, EMBO J. 20 (2001) 3910–3916. [35] S. Taoka, B.W. Lepore, O. Kabil, S. Ojha, D. Ringe, R. Banerjee, Human cystathionine β-synthase is a heme sensor protein. Evidence that the redox sensor is heme and not the vicinal cysteines in the CXXC motif seen in the crystal structure of the truncated enzyme, Biochemistry 41 (2002) 10454–10461. [36] S. Rhee, K.D. Parris, S.A. Ahmed, E.W. Miles, D.R. Davies, Exchange of K+ or Cs+ for Na+ induces local and long-range changes in the three-dimensional structure of the tryptophan synthase α2β2 complex, Biochemistry 35 (1996) 4211–4221. [37] T.R. Schneider, E. Gerhardt, M. Lee, P.H. Liang, K.S. Anderson, I. Schlichting, Loop closure and intersubunit communication in tryptophan synthase, Biochemistry 37 (1998) 5394–5406.

817

[38] M. Weyand, I. Schlichting, Crystal structure of wild-type tryptophan synthase complexed with the natural substrate indole-3-glycerol phosphate, Biochemistry 38 (1999) 16469–16480. [39] A. Sachpatzidis, C. Dealwis, J.B. Lubetsky, P.H. Liang, K.S. Anderson, E. Lolis, Crystallographic studies of phosphonate-based α-reaction transition-state analogues complexed to tryptophan synthase, Biochemistry 38 (1999) 12665–12674. [40] M. Weyand, I. Schlichting, Structural basis for the impaired channeling and allosteric inter-subunit communication in the βA169L/βC170W mutant of tryptophan synthase, J. Biol. Chem. 275 (2000) 41058–41063. [41] M. Weyand, I. Schlichting, A. Marabotti, A. Mozzarelli, Crystal structures of a new class of allosteric effectors complexed to tryptophan synthase, J. Biol. Chem. 277 (2002) 10647–10652. [42] M. Weyand, I. Schlichting, P. Herde, A. Marabotti, A. Mozzarelli, Crystal structure of the βSer178 to Pro mutant of tryptophan synthase. A “knock-out” allosteric enzyme, J. Biol. Chem. 277 (2002) 10653–10660. [43] V. Kulik, M. Weyand, R. Siedel, D. Niks, D. Arac, M.F. Dunn, I. Schlichting, On the role of αThr183 in the allosteric regulation and catalytic mechanism of tryptophan synthase, J. Mol. Biol. 324 (2002) 677–690. [44] Y. Hioki, K. Ogasahara, S.J. Lee, J. Ma, M. Ishida, Y. Yamagata, Y. Matsuura, M. Ota, M. Ikeguchi, S. Kuramitsu, K. Yutani, The crystal structure of the tryptophan synthase β subunit from the hyperthermophile Pyrococcus furiosus. Investigation of stabilization factors, Eur. J. Biochem. 271 (2004) 2624–2635. [45] V. Kulik, E. Hartmann, M. Weyand, M. Frey, A. Gierl, D. Niks, M.E. Dunn, I. Schlichting, On the structural basis of the catalytic mechanism and the regulation of the α subunit of tryptophan synthase from Salmonella typhimurium and BX1 from maize, two evolutionarily related enzymes, J. Mol. Biol. 352 (2005) 608–620. [46] S.J. Lee, K. Ogasahara, J. Ma, K. Nishio, M. Ishida, Y. Yamagata, T. Tsukihara, K. Yutani, Conformational changes in the tryptophan synthase from a hyperthermophile upon α2β2 complex formation: crystal structure of the complex, Biochemistry 44 (2005) 11417–11427. [47] H. Ngo, R. Harris, N. Kimmich, P. Casino, D. Niks, L. Blumenstein, T.R. Barends, V. Kulik, M. Weyand, I. Schlichting, M.F. Dunn, Synthesis and characterization of allosteric probes of substrate channeling in the tryptophan synthase bienzyme complex, Biochemistry 46 (2007) 7713–7727. [48] L. Blumenstein, T. Domratcheva, D. Niks, H. Ngo, R. Seidel, M.F. Dunn, I. Schlichting, βQ114N and βT110V mutations reveal a critically important role of the substrate α-carboxylate Site in the reaction specificity of tryptophan synthase, Biochemistry 46 (2007) 14100–14116. [49] P. Burkhard, G.S. Rao, E. Hohenester, K.D. Schnackerz, P.F. Cook, J.N. Jansonius, Three-dimensional structure of O-acetylserine sulfhydrylase from Salmonella typhimurium, J. Mol. Biol. 283 (1998) 121–133. [50] P. Burkhard, C.H. Tai, C.M. Ristroph, P.F. Cook, J.N. Jansonius, Ligand binding induces a large conformational change in O-acetylserine sulfhydrylase from Salmonella typhimurium, J. Mol. Biol. 291 (1999) 941–953. [51] P. Burkhard, C.H. Tai, J.N. Jansonius, P.F. Cook, Identification of an allosteric anionbinding site on O-acetylserine sulfhydrylase: structure of the enzyme with chloride bound, J. Mol. Biol. 303 (2000) 279–286. [52] E.R. Bonner, R.E. Cahoon, S.M. Knapke, J.M. Jez, Molecular basis of cysteine biosynthesis in plants: structural and functional analysis of O-acetylserine sulfhydrylase from Arabidopsis thaliana, J. Biol. Chem. 280 (2005) 38803–38813. [53] M.T. Claus, G.E. Zocher, T.H. Maier, G.E. Schulz, Structure of the O-acetylserine sulfhydrylase isoenzyme CysM from Escherichia coli, Biochemistry 44 (2005) 8620–8626. [54] G. Zocher, U. Wiesand, G.E. Schulz, High resolution structure and catalysis of Oacetylserine sulfhydrylase isozyme B from Escherichia coli, FEBS J. 274 (2006) 5382–5389. [55] J.A. Francois, S. Kumaran, J.M. Jez, Structural basis for interaction of O-acetylserine sulfhydrylase and serine acetyltransferase in the Arabidopsis cysteine synthase complex, Plant Cell 18 (2006) 3647–3655. [56] A. Chattopadhyay, M. Meier, S. Ivaninskii, P. Burkhard, F. Speroni, B. Campanini, S. Bettati, A. Mozzarelli, W.M. Rabeh, L. Li, P.F. Cook, Structure, mechanism, and conformational dynamics of O-acetylserine sulfhydrylase from Salmonella typhimurium: comparison of A and B isozymes, Biochemistry 46 (2007) 8315–8330. [57] R. Schnell, W. Oehlmann, M. Singh, G. Schneider, Structural insights into catalysis and inhibition of O-acetylserine sulfhydrylase from Mycobacterium tuberculosis. Crystal structures of the enzyme alpha-aminoacrylate intermediate and an enzyme-inhibitor complex, J. Biol. Chem. 282 (2007) 23473–23481. [58] Y. Oda, K. Mino, K. Ishikawa, M. Ataka, Three-dimensional structure of a new enzyme, O-phosphoserine sulfhydrylase, involved in L-cysteine biosynthesis by a hyperthermophilic archaeon, aeropyrum pernix K1, at 2.0 Å resolution, J. Mol. Biol. 351 (2005) 334–344. [59] R. Fukunaga, S. Yokoyama, Structural insights into the second step of RNAdependent cysteine biosynthesis in Archaea: crystal structure of Sep-tRNA:CystRNA synthase from Archaeoglobus fulgidus, J. Mol. Biol. 370 (2007) 128–141. [60] T. Clausen, R. Huber, B. Laber, H.D. Pohlenz, A. Messerschmidt, Crystal structure of the pyridoxal-5′-phosphate dependent cystathionine β-lyase from Escherichia coli at 1.83 Å, J. Mol. Biol. 262 (1996) 202–224. [61] T. Clausen, R. Huber, A. Messerschmidt, H.D. Pohlenz, B. Laber, Slow-binding inhibition of Escherichia coli cystathionine β-lyase by L-aminoethoxyvinylglycine: a kinetic and X-ray study, Biochemistry 36 (1997) 12633–12643. [62] U. Breitinger, T. Clausen, S. Ehlert, R. Huber, B. Laber, F. Schmidt, E. Pohl, A. Messerschmidt, The three-dimensional structure of cystathionine β-lyase from Arabidopsis and its substrate specificity, Plant Physiol. 126 (2001) 631–642. [63] L.J. Ejim, J.E. Blanchard, K.P. Koteva, R. Sumerfield, N.H. Elowe, J.D. Chechetto, E.D. Brown, M.S. Junop, G.D. Wright, Inhibitors of bacterial cystathionine β-lyase: leads for new antimicrobial agents and probes of enzyme structure and function, J. Med. Chem. 50 (2007) 755–764.

818

T. Yamada et al. / Biochimica et Biophysica Acta 1780 (2008) 809–818

[64] T. Clausen, A. Schlegel, R. Peist, E. Schneider, C. Steegborn, Y.S. Chang, A. Haase, G.P. Bourenkov, H.D. Bartunik, W. Boos, X-ray structure of MalY from Escherichia coli: a pyridoxal 5′-phosphate-dependent enzyme acting as a modulator in mal gene expression, EMBO J. 19 (2000) 831–842. [65] T. Clausen, J.T. Kaiser, C. Steegborn, R. Huber, D. Kessler, Crystal structure of the cystine C–S lyase from Synechocystis: stabilization of cysteine persulfide for FeS cluster biosynthesis, Proc. Natl. Acad. Sci. U. S. A. 97 (2000) 3856–3861. [66] E.B. Kuettner, R. Hilgenfeld, M.S. Weiss, The active principle of garlic at atomic resolution, J. Biol. Chem. 277 (2002) 46402–46407. [67] J.T. Kaiser, S. Bruno, T. Clausen, R. Huber, F. Schiaretti, A. Mozzarelli, D. Kessler, Snapshots of the cystine lyase “C-DES” during catalysis: studies in solution and in the crystalline state, J. Biol. Chem. 278 (2003) 357–365. [68] L.J. Shimon, A. Rabinkov, I. Shin, T. Miron, D. Mirelman, M. Wilchek, F. Frolow, Two structures of alliinase from Alliium sativum L.: apo form and ternary complex with aminoacrylate reaction intermediate covalently bound to the PLP cofactor, J. Mol. Biol. 366 (2007) 611–625. [69] B. Sundararaju, A.A. Antson, R.S. Phillips, T.V. Demidkina, M.V. Barbolina, P. Gollnick, G.G. Dodson, K.S. Wilson, The crystal structure of Citrobacter freundii tyrosine phenollyase complexed with 3-(4′-hydroxyphenyl)propionic acid, together with sitedirected mutagenesis and kinetic analysis, demonstrates that arginine 381 is required for substrate specificity, Biochemistry 36 (1997) 6502–6510. [70] D. Milic, D. Matkovic-Calogovic, T.V. Demidkina, V.V. Kulikova, N.I. Sinitzina, A.A. Antson, Structures of apo- and holo-tyrosine phenol-lyase reveal a catalytically critical closed conformation and suggest a mechanism for activation by K+ ions, Biochemistry 45 (2006) 7544–7552. [71] T. Nakai, H. Mizutani, I. Miyahara, K. Hirotsu, S. Takeda, K.H. Jhee, T. Yoshimura, N. Esaki, Three-dimensional structure of 4-amino-4-deoxychorismate lyase from Escherichia coli, J. Biochem.(Tokyo) 128 (2000) 29–38.

[72] T. Fujii, M. Maeda, H. Mihara, T. Kurihara, N. Esaki, Y. Hata, Structure of a NifS homologue: X-ray structure analysis of CsdB, an Escherichia coli counterpart of mammalian selenocysteine lyase, Biochemistry 39 (2000) 1263–1273. [73] H.I. Krupka, R. Huber, S.C. Holt, T. Clausen, Crystal structure of cystalysin from Treponema denticola: a pyridoxal 5′-phosphate-dependent protein acting as a haemolytic enzyme, EMBO J. 19 (2000) 3168–3178. [74] H. Mihara, T. Fujii, S. Kato, T. Kurihara, Y. Hata, N. Esaki, Structure of external aldimine of Escherichia coli CsdB, an IscS/NifS homolog: implications for its specificity toward selenocysteine, J. Biochem. (Tokyo) 131 (2002) 679–685. [75] C.D. Lima, Analysis of the E. coli NifS CsdB protein at 2.0 Å reveals the structural basis for perselenide and persulfide intermediate formation, J. Mol. Biol. 315 (2002) 1199–1208. [76] J.R. Cupp-Vickery, H. Urbina, L.E. Vickery, Crystal structure of IscS, a cysteine desulfurase from Escherichia coli, J. Mol. Biol. 330 (2003) 1049–1059. [77] B. Tirupati, J.L. Vey, C.L. Drennan, J.M. Bollinger Jr., Kinetic and structural characterization of Slr0077/SufS, the essential cysteine desulfurase from Synechocystis sp. PCC 6803, Biochemistry 43 (2004) 12210–12219. [78] S.Y. Ku, P. Yip, P.L. Howell, Structure of Escherichia coli tryptophanase, Acta Crystallogr. Sect. D 62 (2006) 814–823. [79] S. Sharif, M.C. Huot, P.M. Tolstoy, M.D. Toney, K.H.M. Jonsson, H.H. Limbach, 15 N nuclear magnetic resonance studies of acid-base properties of pyridoxal-5′phosphate aldimines in aqueous solution, J. Phys. Chem., B 111 (2007) 3869–3876. [80] P.A. Frey, S.A. Whitt, J.B. Tobin, A low-barrier hydrogen bond in the catalytic triad of serine proteases, Science 264 (1994) 1927–1930. [81] C.L. Perrin, J.B. Nielson, “Strong” hydrogen bonds in chemistry and biology, Annu. Rev. Phys. Chem. 48 (1997) 511–544. [82] W.W. Cleland, P.A. Frey, J.A. Gerlt, The low barrier hydrogen bond in enzymatic catalysis, J. Biol. Chem. 273 (1998) 25529–25532.