Mutational and structural analysis of cobalt-containing nitrile hydratase on substrate and metal binding Akimasa Miyanaga 1 , Shinya Fushinobu 1 , Kiyoshi Ito 2 , Hirofumi Shoun 1 and Takayoshi Wakagi 1 1 Department of Biotechnology, The University of Tokyo, Japan; 2 Life Science Laboratory, Mitsui Chemicals Inc., Togo, Mobara-shi, Chiba, Japan Mutants of a cobalt-containing nitrile hydratase (NHase, EC 4.2.1.84) from Pseudonocardia thermophila JCM 3095 involved in substrate binding, catalysis and formation of the active center were constructed, and their characteristics and crystal structures were investigated. As expected from the structure of the substrate binding pocket, the wild-type enzyme showed significantly lower K m and K i values for aromatic substrates and inhibitors, respectively, than alipha- tic ones. In the crystal structure of a complex with an inhibitor (n-butyric acid) the hydroxyl group of bTyr68 formed hydrogen bonds with both n-butyric acid and aSer112, which islocatedintheactivecenter.ThebY68F mutant showed an elevated K m value and a significantly decreasedk cat value. The apoenzyme, which contains no detectable cobalt atom, was prepared from Escherichia coli cells grown in medium with- out cobalt ions. It showed no detectable activity. A disulfide bond between aCys108 and aCys113 was formed in the apoenzyme structure. In the highly conserved sequence motif in the cysteine cluster region, two positions are exclusively conserved in cobalt-containing or iron-containing nitrile hydratases. Two mutants (aT109S and aY114T) were con- structed, each residue being replaced with an iron-containing one. The aT109S mutant showed similar characteristics to the wild-type enzyme. However, the aY114T mutant showed a very low cobalt content and catalytic activity compared with the wild-type enzyme, and oxidative modifications of aCys111 and aCys113 residues were not observed. The aTyr114 residue may be involved in the interaction with the nitrile hydratase activator protein of P. thermophila. Keywords: cobalt-containing nitrile hydratase; imidate; Pseudonocardia thermophila; noncorrin cobalt; claw setting. Nitrile hydratase (NHase, EC 4.2.1.84) catalyzes the hydra- tion of nitriles to the corresponding amides [1,2]. NHase has been used in the industrial production of acrylamide and nicotinamide from the corresponding nitriles. NHase is a metalloenzyme that contains iron or cobalt in its catalytic center. Iron-containing (Fe-type) NHase contains a nonheme iron ion [3,4], and cobalt-containing (Co-type) NHase contains a noncorrin cobalt ion [5–7]. NHase consists of a and b subunits, the amino acid sequences of which do not exhibit homology. In all known NHases, each subunit has a highly homologous amino acid sequence. In particular, three cysteine residues and one serine residue in the cysteine cluster region, which co-ordinate the iron or cobalt ion of the a subunit, and two arginine residues of the b subunit, are fully conserved (Fig. 1). Fe-type NHase shows photoreactivity and binds a nitric oxide (NO) molecule, whereas Co-type NHase does not [8–10]. Fe-type NHase preferentially hydrates small aliphatic nitriles [11], whereas Co-type NHase exhibits a high affinity for aromatic nitriles [12,13]. The crystal structures of Fe-type NHases, in the active form at 2.65 A ˚ resolution, and in the NO-bound inactive state, at 1.7 A ˚ resolution, have been reported [14,15]. Two cysteine residues, coordinated to the iron ion, are post- translationally modified to cysteine-sulfinic acid and cysteine-sulfenic acid, yielding a claw setting structure. These modifications enable a photoreaction and associ- ation with NO, and are essential for the catalytic activity [15–18]. NHase from Pseudonocardia thermophila JCM 3095 is an a 2 b 2 heterotetramer and contains cobalt [6]. Recently, we determined the crystal structure of this Co-type NHase at 1.8 A ˚ resolution [19]. In this structure, two cysteine residues, coordinated to the cobalt ion, were modified and had the claw setting, as in the Fe-type NHase. From studies on the structure and function of NHase, a possible reaction model was proposed [2]. In this model, imidate is produced as a reaction intermediate beforeitisconvertedtoanamide. Two arginine residues, bArg52 and bArg157, formed four hydrogen bonds with the modified oxygen atoms. It is thought that these bonds also stabilize the claw setting. In the Fe-type NHase, mutants of the two arginine residues exhibited sharply reduced stability and enzymatic activity [17,18]. Of the residues of P. thermophila NHase participating in the recognition of a substrate, three (bLeu48, bPhe51 and bTrp72) form a hydrophobic pocket [19]. This hydrophobic Correspondence to S. Fushinobu, Department of Biotechnology, The University of Tokyo, 1-1-1 Yayoi, Bunkyo-ku, Tokyo 113-8657, Japan. Fax: + 81 3 5841 5337, Tel.: + 81 3 5841 5151, E-mail: asfushi@mail.ecc.u-tokyo.ac.jp Abbreviations: De, absorption coefficient; ICP-AES, inductive coupled plasma–atomic emission spectroscopy; Co-type NHase, cobalt- containing nitrile hydratase; Fe-type NHase, iron-containing nitrile hydratase; NHase, nitrile hydratase; NO, nitric oxide. Enzymes: nitrile hydratase (EC 4.2.1.84) (Received 23 October 2003, revised 17 November 2003, accepted 25 November 2003) Eur. J. Biochem. 271, 429–438 (2004) Ó FEBS 2003 doi:10.1046/j.1432-1033.2003.03943.x pocket is thought to accommodate the alkyl chain or aromatic ring of a nitrile substrate. bTrp72 of the Co-type NHase from P. thermophila replaces the tyrosine residue of Fe-type NHase, and the substrate binding pocket in the Co-type NHase was larger than that in the Fe-type NHase. This difference seems to be the cause of the different substrate preferences of Co-type and Fe-type NHases. Although the structures of Fe-type NHase and Co-type NHase are very similar, these NHases specifically bind their own metals [20]. It is unknown why NHase selects only a single metal: cobalt or iron. There has only been one report on metal substitution in NHase [20]. When an Fe-type NHase from Rhodococcus sp. N-771 was expres- sed in Escherichia coli (cultured in cobalt-supplemented Fig. 1. Alignment of the amino acid sequences of nitrile hydratases (NHases). Three cysteine residues and one serine residue in the cysteine cluster region, two conserved arginine resi- dues, and one conserved tyrosine residue, are highlighted. Triangles indicate residues involved in the formation of the substrate binding pocket. Arrows indicate the residues that were mutated in this study. Asterisks indicate completely conserved residues. P. thermo, Pseudonocardia thermophila JCM 3095; R. rhoJ1L, Rhodococcus rhodochrous J1 low-molecular-mass; P. put, Pseudomon- as putida NRRL-18668; R. rhoJ1H, Rhodo- coccus rhodochrous J1 high-molecular-mass; R. sp. Rhodococcus sp.N-771;P.chlo,Pseu- domonas chlororaphis B23. 430 A. Miyanaga et al.(Eur. J. Biochem. 271) Ó FEBS 2003 medium), without coexpression of the Fe-type NHase activator, a cobalt ion was incorporated into the catalytic center of the NHase. However, the cobalt-substituted Fe-type NHase showed low enzymatic activity. There have been no reports on metal substitutions in Co-type NHase. Co-type NHase contains threonine and tyrosine in the -C(T/S)LCSC(Y/T)- sequence of the active center, whereas Fe-type NHase contains serine and threonine (Fig. 1). The differences in the side-chains, especially at the threonine/serine position, have been thought to be important [7]. However, there have been no mutational studies on these residues. In this article, we report mutational and structural analysis of the substrate binding and metal specificity of a Co-type NHase. Experimental procedures Kinetic study NHase activity was determined by measuring the hydration of acrylonitrile, methacrylonitrile, benzonitrile, 3-cyano- pyridine, or 4-cyanopyridine, in 100 m M potassium phos- phate buffer, pH 7.6. The rate of nitrile hydration was determined from the increase in absorbance at 25 °C, using the absorption coefficients (De) of the corresponding amides. Acrylamide (De 225 ¼ 2.9 m M Æcm )1 ) and methacryl- amide (De 225 ¼ 3.2 m M Æcm )1 ) were measured at 225 nm. Benzamide (De 242 ¼ 5.5 m M Æcm )1 )wasmeasuredat 242 nm. Nicotinamide (De 235 ¼ 3.2 m M Æcm )1 ), which is produced from 3-cyanopyridine (De 235 ¼ 0.8 m M Æcm )1 ), was measured at 235 nm. Isonicotinamide (De 233 ¼ 2.6 m M Æcm )1 ), which is produced from 4-cyano- pyridine (De 233 ¼ 0.6 m M Æcm )1 ), was measured at 233 nm. At least 10 data points were collected for each substrate. The inhibition constants for n-butyric acid, propionic acid and benzoic acid were determined using Dixon plots [21]. Methacrylonitrile was used as a substrate, the activity being measured using various inhibitor concen- trations (0–10 m M for n-butyric acid, 0–50 m M for propionic acid, and 0–100 l M for benzoic acid) and two substrate concentrations (0.5 and 5 m M metacrylo- nitrile). Site-directed mutagenesis An expression plasmid, pUC18-NHase [19], which contains the genes for the b and a subunits of NHase and for the NHase activator, was used for mutagenesis. Site-directed mutagenesis was carried out by using the megaprimer PCR method [22]. The primers used were as follows: 5¢-GAGC TC GAATTCTGAGAGGAGCTC-3¢ (bF), 5¢-GGTCAT GCC GCGGCCGCCTTCGTG-3¢ (bR), 5¢-CACGAAGGCG GCCGCGGCATG-3¢ (aF), 5¢-GCATGCAAGCTTGCA TGCCGGTG-3¢ (aR), 5¢-CTCGAGTCGCCGTTCTACT GGCACTGGATC-3¢ (68f), 5¢-GATCCAGTGCCAGTA GAACGGCGACTCGAG-3¢ (68r), 5¢-CACGTCGTCGT GTGCTCGCTCTGCTCCTGC-3¢ (109f), 5¢-GCAGGAG CAGAGCGAGCACACGACGACGTG-3¢ (109r), 5¢-CTC TGCTCCTGCACCCCATGGCCGGTGCTG-3¢ (114f), and 5¢-CAGCACCGGCCATGGGGTGCAGGAGCAGAG-3¢ (114r). Restriction enzyme recognition sites are underlined and mutated residues are shown in bold. The r primer had a sequence completely complementary to that of the corres- ponding f primer. The first PCR amplification was performed with KOD- plus DNA polymerase (Toyobo, Japan), using the bFand 68r primers, bR and 68f primers, aF and 109r primers, aR and 109f primers, aF and 114r primers, or aR and 114f primers. Following an initial denaturation at 98 °Cfor 2 min, 35 PCR cycles were carried out: each cycle comprised incubation at 98 °C for 15 s, followed by a 30 s incubation at 55 °C, and a 1 min incubation at 68 °C. The second PCR was also performed with KOD-plus DNA polymerase, using the two megaprimers and the bFandbR primers, or the two megaprimers and the aFandaR primers. The same PCR program, as described above for the first reaction, was used. The 0.72 kb DNA fragment and pUC18-NHase were digested with EcoRI and NotI, and the 1.08 kb DNA fragment and pUC18-NHase were digested with NotIand HindIII. The PCR product was ligated to a pUC18 plasmid. The resulting plasmid was then sequenced to confirm the presence of the mutation. Expression and purification of each mutant and the apoenzyme The mutants were expressed in E. coli JM109. Cells were grown at 37 °C, in 1 L of Luria–Bertani broth containing ampicillin (100 mgÆL )1 ). When the attenuance (D) reached 0.3 at 600 nm, the cells were induced by the addition of 0.1 m M isopropyl thio-b- D -galactoside and the metal sources (0.25 m M cobalt chloride, 0.25 m M ferric chloride, or nothing). The cells were then cultured for an additional 16 h. All subsequent manipulations were performed at 5 °C. After cell harvesting by centrifugation, the pellet was resuspended in 10 mL of 50 m M Tris/HCl (pH 7.6). From a cell-free extract prepared by sonication, the protein was purified by ammonium sulfate precipitation (40–70%) and anion-exchange chromatographies (DEAE–Sepharose and MonoQ columns). Each mutant and the apoenzyme were purified in a soluble form. Crystallization, data collection and crystallographic refinement Crystals of the mutants and apoenzyme were grown under the same conditions as for the wild-type NHase [19]. Crystals of a complex with n-butyric acid were prepared by soaking a native crystal in a reservoir solution containing 15 m M n-butyric acid for 3 h. Before flash-freezing, the crystals were equilibrated with the reservoir solution con- taining 20% (v/v) glycerol. Data were collected using a CCD camera at the BL6A station of the Photon Factory (Tsukuba, Japan) and the BL38B2 and BL40B2 stations of SPring-8 (Harima, Japan), at 100 K. Diffraction images were indexed, integrated, and scaled using the DPS/MOSFLM [23] or the HKL [24] program suites. The crystal structure of Co-type NHase (Protein Data Bank code 1IRE) was used as the first model. At the first stage of the crystallographic refinement, the models had the following removed: the side-chain of the mutated residue in the aT109S and aY114T structures, and a cobalt atom and three oxygen atoms in the apoenzyme and aY114T structures. Several rounds of refinement and model Ó FEBS 2003 Mutants of Co-type NHase (Eur. J. Biochem. 271) 431 correction were carried out using programs CNS [25] and XFIT [26]. At the final stage of refinement of the complex structure, n-butyric acid was added to the model according to the F o –F c map. At the final stage of refinement of the aY114T structure, a cobalt atom was added to the model, according to the F o –F c map. The coordinates and structure factors have been deposited in the Protein Data Bank (codes: 1UGP, 1UGQ, 1UGR and 1UGS). Analytical procedures Protein concentrations were determined by using the bicinchoninic assay (Pierce Chemical Co.), with BSA as the standard. The cobalt and iron contents were determined by ICP-AES (SPS-1200 V; Seiko Instruments, Chiba, Japan) using sample solutions, and standard solutions of cobalt and iron (Wako Chemical Co.). The thermostabilities of the mutants and wild-type enzymes were investigated by measuring the activity at 25 °C after incubation at different temperatures for 30 min. The T m was defined as the temperature at which the activity remaining was 50% of that without any incubation. Figures 2–5 were prepared using programs XFIT [26] and RASTER 3 D [27]. Results Kinetic parameters and inhibitor studies The kinetic parameters of P. thermophila NHase for acrylonitrile, methacrylonitrile, benzonitrile, 3-cyanopyri- dine and 4-cyanopyridine, were measured (Table 1). In general, the enzyme exhibited significantly smaller K m values for aromatic nitriles than for small aliphatic nitriles. n-Butyric acid competitively inhibited P. thermophila NHase, which showed a K i value of 1.3 m M .ThisK i value is similar to that of Rhodococcus sp. N-771 NHase (1.6 m M ) [17]. The inhibition by propionic acid and benzoic acid was also competitive, the K i values being 9.9 m M and 33 l M , respectively. Therefore, the enzyme exhibited significantly stronger inhibition by aromatic organic acids than by small aliphatic organic acids. n-Butyric acid is known not only as a competitive inhibitor but also as a stabilizer of Fe-type NHase [11,28]. The activity of Fe-type NHase gradually decreases under conditions with a lack of n-butyric acid. For Co-type NHases, a native NHase is more stable. P. thermo- phila NHase retained complete activity when stored at 4 °C for 1 year, and the catalytic center did not change, as confimed by analysis of the crystal structure (data not shown). Complex structure with n -butyric acid An NHase crystal soaked in 15 m M n-butyric acid for 3 h diffracted up to 1.63 A ˚ (Table 2). The refined structure obtained using this data set closely resembled the native structure. Electron density, similar to n-butyric acid in shape, was observed in the active site, although it was unclear (Fig. 2). The electron density of the alkyl end was weak, whereas that of the carboxylic group was strong. Moreover, two types of conformations seemed to be present in a mixed state. In one state (designated as type I), one oxygen atom of the carboxylic group formed a hydrogen bond (2.90 A ˚ )with bTyr68. On the other hand, one carboxylic oxygen atom of another state (type II) appeared to be positioned at a short distance ( 1.4 A ˚ ) from the oxygen atom of Table 1. Kinetic parameters and cobalt contents of the wild-type and mutant proteins. ND, not determined. Wild-type Apoenzyme bY68F aT109S aY114T Acrylonitrile k cat (s )1 ) 1910 <0.1 15.2 621 33.9 K m (m M ) 3.6 ND 58 107 9.9 k cat /K m (s )1 Æm M )1 ) 537 ND 0.26 5.8 3.4 Methacrylonitrile k cat (s )1 ) 1000 <0.1 14.1 482 34.6 K m (m M ) 0.49 ND 2.6 4.7 2.3 k cat /K m (s )1 Æm M )1 ) 2040 ND 5.4 103 15 Benzonitrile k cat (s )1 ) 123 <0.1 7.4 132 4.9 K m (m M ) 0.020 ND 0.23 0.025 0.11 k cat /K m (s )1 Æm M )1 ) 6150 ND 32 5290 45 Nicotinonitrile k cat (s )1 ) 131 – – – – K m (mM) 0.12 – – – – k cat /K m (s )1 Æm M )1 ) 1090 – – – – Isonicotinonitrile k cat (s )1 )90–––– K m (m M ) 0.079 – – – – k cat /K m (s )1 Æm M )1 ) 1140 – – – – Cobalt content Co/ab 0.81 a <0.01 0.83 0.51 0.035 a Value taken from reference [19]. 432 A. Miyanaga et al.(Eur. J. Biochem. 271) Ó FEBS 2003 aCys-SOH113. It is possible that a covalent bond is formed between the two oxygen atoms, but the precise chemical species of this state has yet to be elucidated. In the Fe-type NHase, 2-cyano-2-propyl hydroperoxide irreversibly inac- tivates the enzyme, probably by the oxidation of aCys-SOH to aCys-SO 2 H [29]. Carboxylic oxygen atoms of both types were directly coordinated to a cobalt ion (2.6 and 2.3 A ˚ in types I and II, respectively), instead of the water molecule Table 2. Data collection and refinement statistics. n-Butyric acid complex Apoenzyme aT109S aY114T Data collection statistics Beam line KEK-PF SPring-8 SPring-8 SPring-8 BL6A BL40B2 BL38B2 BL40B2 Wavelength (A ˚ ) 0.978 1.00 1.00 1.00 Space group P3 2 21 P3 2 21 P3 2 21 P3 2 21 Cell constant (A ˚ ) a ¼ b ¼ 65.564 a ¼ b ¼ 65.362 a ¼ b ¼ 65.437 a ¼ b ¼ 65.517 c ¼ 184.994 c ¼ 184.099 c ¼ 184.257 c ¼ 183.874 Resolution limit (A ˚ ) 1.63 2.0 1.8 2.0 Unique observations 57,837 31,332 43,459 30,513 Completeness (%) 99.0 98.5 99.9 94.7 I/r (highest shell) 10.9 (3.2) 11.1 (4.7) 8.6 (3.0) 14.0 (7.3) R merge (highest shell) (%) 5.4 (24.0) 5.7 (20.1) 5.9 (32.2) 4.7 (14.1) Refinement statistics Resolution range (A ˚ ) 24.6–1.63 19.4–2.0 19.4–1.8 19.5–2.0 R (highest shell) (%) 18.2 (20.5) 18.4 (18.1) 18.9 (21.2) 18.1 (16.7) R free (highest shell) (%) 19.7 (22.2) 22.0 (23.9) 21.3 (23.1) 22.5 (22.7) R.m.s. deviations from the native structure (1IRE) (A ˚ ) Ca atoms 0.067 0.126 0.128 0.149 All atoms 0.272 0.138 0.311 0.257 No. of amino acid residues 429 431 431 431 No. of water molecules 397 268 330 280 No. of cobalt atoms (occupancy) 1 (1.01) 0 1 (0.71) 1 (0.29) No. of n-butyric acid 0.5 + 0.5 0 0 0 Average B-factors (A ˚ 2 ) Protein 11.7 19.3 19.4 17.4 Solvent 21.7 24.7 29.4 23.3 n-Butyric acid (type I) 27.0 n-Butyric acid (type II) 26.9 Fig. 2. Complex structure with n-butyric acid. (A) F o –F c electron density map for the catalytic center of the complex structure with n-butyric acid is shown at the 3 r contour level, using a stereographic representation. The map was constructed prior to the incorporation of n-butyric acid. The type I n-butyric acid molecule is shown in light grey and the type II molecule in dark grey. The two hydrogen bonds with bTyr68 are represented by broken lines in light gray, and the short contact between type II n-butyric acid and aCys-SOH113 by a broken line in dark grey. Ó FEBS 2003 Mutants of Co-type NHase (Eur. J. Biochem. 271) 433 that coordinated in the native structure. Moreover, this carboxylic oxygen atom was trapped by three oxygen atoms of the claw setting (aCys-SO 2 H111, aSer112, and aCys- SOH113). We conducted crystallographic refinement of the complex structure with two alternative states, fixing the position of the oxygen atom forming a short contact in the type II state. The occupancies of both alternative states were fixed at 0.5, because the values did not change on the refinement. Average temperature factors of the butyric acid molecule in the two states are shown in Table 2. On the basis of the type I complex structure, we focused on bTyr68 as being the key residue in substrate binding and/or catalysis. The tyrosine residue at this position is fully conserved in all NHases (Fig. 1). bTyr68 formed another hydrogen bond (2.55 A ˚ )withanoxygenatomoftheside- chain of aSer112 (Fig. 2). This hydrogen bond is also present in the native NHase [19] and Fe-type NHase structures [15]. The alkyl group of n-butyric acid was extended in the direction of the hydrophobic pocket. bPhe37, bLeu48 and bPhe51 are involved in the hydropho- bic environment around the alkyl group of n-butyric acid. We also attempted to prepare co-crystals with other inhibitors, benzoic acid and propionic acid. The crystals diffracted up to high resolution (benzoic acid, 1.7 A ˚ ;and propionic acid, 1.5 A ˚ ). Although some ÔblobsÕ of electron densities were seen in the substrate binding sites of these structures instead of the water molecule present in the native structure, these ÔblobsÕ could not be interpreted (data not shown). Mutagenic analysis of bTyr68 To evaluate the importance of bTyr68, we replaced it with phenylalanine. The k cat and K m values of the bY68F mutant are shown in Table 1. The mutant showed significantly decreased activity compared with the wild-type enzyme. The K m value of the mutant was about 10 times higher, and the k cat value 100 times lower, respectively, than those of the wild-type enzyme, when acrylonitrile was used as the substrate. The K i value for n-butyric acid was also  10 times higher than that of the wild-type enzyme, being 18 m M . The crystal structure of the bY68F mutant was deter- mined at 2.0 A ˚ resolution. The structure of the bY68F mutant was almost identical to that of the wild-type enzyme, including the formation of a normal claw setting, except for the mutation site (data not shown). Apoenzyme of NHase The ÔapoenzymeÕ of NHase, which contains no metal ion, was prepared by expressing the protein in E. coli in a medium lacking cobalt chloride. The apoenzyme did not show any detectable enzymatic activity. Moreover, analysis (by ICP-AES) of the metal content of the apoenzyme, revealed that it did not contain a cobalt or an iron ion. Expression of the wild-type enzyme in iron- Fig. 3. The metal centers of (A) the apoenzyme, and (B) aT109S and (C) aY114T mutants. (A) 2F o –F c electron density map at the 2 r contour level. The disulfide bond is shown by a dotted line. (B) F o –F c electron density map (6 r) constructed prior to incorporation of the three modified oxygen atoms and a cobalt atom into the model structure. (C) F o –F c electron density map (6 r) constructed prior to incorporation of a cobalt atom into the model structure. 434 A. Miyanaga et al.(Eur. J. Biochem. 271) Ó FEBS 2003 supplemented medium produced a similar ÔapoenzymeÕ (data not shown). The crystal structure of the apoenzyme was also deter- mined at 2.0 A ˚ resolution (Table 2). A cobalt atom was not observed in the catalytic center of the apoenzyme, and modifications of cysteine residues and the formation of a claw setting were not observed (Fig. 3A). Instead, electron density connecting aCys108 and aCys113 was observed, indicating the formation of a disulfide bond between these residues. Incubation of the apoenzyme with cobalt ions did not recover the NHase activity, the incorporation of a cobalt ion probably being blocked by the disulfide bond. The conformation of the cysteine cluster region was significantly closed compared with that of the wild-type enzyme (Fig. 4). Structural differences between Co-type and Fe-type NHases In the highly conserved cysteine cluster region, two amino acid residues, corresponding to positions a109 and a114 in P. thermophila NHase, showed significant conservation in Co-type and Fe-type NHases. aThr109 and aTyr114 were conserved in Co-type NHase, whereas, in Fe-type NHase, these residues were replaced with serine and threonine, respectively (Fig. 1). aThr109 is located in the cysteine cluster region. In Co-type NHase, the side-chain of aThr109 undergoes a hydrophobic interaction with the side-chain of aVal136 (Fig. 5A). The distance between the Cc2atomofaThr109 and the Cc1atomofaVal136 is 3.7 A ˚ . On the other hand, in Fe-type NHase, the side-chain of the corresponding serine residue does not interact with the valine residue. aTyr114 is also located near the cysteine cluster region. In Co-type NHase, the hydroxyl group of aTyr114 forms hydrogen bonds with the main-chain oxygen atoms of aLeu119 and aLeu121, via a water molecule. Moreover, the aTyr114 residue undergoes hydrophobic interactions with its surroundings. In Fe-type NHase, the corresponding threonine residue (aThr115) forms a hydrogen bond with the main-chain oxygen atom of a serine residue (aSer113) in the cysteine cluster (Fig. 5B). The conformation of the cysteine cluster region of Fe-type NHase is slightly open, compared with the Co-type NHase (Fig. 5B) [19], although this difference ( 0.1 A ˚ ) was smaller than the coordinate errors estimated from the Luzzati plot (0.17 A ˚ for 1IRE, and 0.19 A ˚ for 2AHJ). The hydrogen bond seems to pull aSer113 closer to aThr115, and thus makes the conforma- tion of the cysteine cluster region open. Mutagenic analysis of aThr109 To evaluate the effect of aThr109 on metal selectivity, we replaced it with serine to produce the aT109S mutant. Judging from the results of ICP-AES, the aT109S mutant contained 0.51 cobalt ions per ab heterodimer. The cobalt content of this mutant was slightly decreased compared with that of the wild-type enzyme (Table 1). The K m value was 30 times higher than that of the wild-type enzyme, when acrylonitrile was used as the substrate (Table 1). On the other hand, the K m value was similar to that of the wild-type enzyme, when benzonitrile was used. To confirm the existence of a cobalt atom in the active center of the aT109S mutant, the crystal structure of the aT109S mutant was determined at 1.8 A ˚ resolution (Table 2). In the F o –F c map, an electron density peak of a cobalt atom (10 r), as well as three oxygen atoms of the modified cysteine residues, were clearly observed (Fig. 3B). The occupancy of the cobalt atom was found to be 0.71. Therefore, the structure of this mutant was almost identical to that of the wild-type enzyme, except for the mutation site. Mutagenic analysis of aTyr114 To evaluate the effect of aTyr114 on metal selectivity, we replaced it with threonine to produce the aY114T mutant. Judging from the results of ICP-AES, the aY114T mutant contained only 0.035 cobalt ions per ab heterodimer (Table 1). The mutant exhibited only slight activity, probably as a result of the low cobalt content. When the mutant was expressed in medium without a cobalt ion, or in iron-supplemented medium, neither the cobalt content nor the catalytic activity was detectable. The aT109S/aY114T double mutant showed characteristics similar to those of the aY114T mutant (data not shown). The crystal structure of the aY114T mutant was deter- mined at 2.0 A ˚ resolution (Table 2). The structure of the aY114T mutant was almost identical to that of the wild- type enzyme, except for the mutation site and the catalytic Fig. 4. Superimposition of the metal centers of the wild-type enzyme, the apoenzyme, and the aT109S and aY114T mutants. The wild-type enzyme is shown in yellow, the apoenzyme in black, the aT109Smutantinred,andthe aY114T mutant in purple. Purple arrows indicate the movement of atoms in the aY114T mutant compared with that observed in the wild-type enzyme. The disulfide bond of the apoenzyme is shown as a black dotted line. Ó FEBS 2003 Mutants of Co-type NHase (Eur. J. Biochem. 271) 435 center. The electron density map around the metal center was not clear, probably because of the low cobalt ion content. A low density peak of cobalt atom was observed at the catalytic center (Fig. 3C). The occupancy and tempera- ture factor of the cobalt atom were determined to be 0.29 and 24.3, respectively. Modifications of the cysteine residues aCys111 and aCys113 were not observed (Fig. 3C); how- ever, no disulfide bond was detected. The aThr114 residue of the mutant formed a weak hydrogen bond (3.1 A ˚ ) with an oxygen atom of the main- chain of aSer112, as in the Fe-type NHase (Fig. 5B). However, the conformation of the cysteine cluster region was not open, as in the Fe-type NHase, but closed, as found in the wild-type Co-type NHase (Fig. 5B). The atoms of the side-chains of aCys111 and aCys113 were moved to the outside, and the cobalt atom was also slightly moved (Fig. 4). As a result, the distances between the cobalt atom and its ligands in the aY114T mutant became greater, compared the wild-type enzyme (Table 3). Discussion Substrate specificity comparison with other NHases The kinetic parameters of P. thermophila NHase, with regard to the size of substrates, were similar to those of the low-molecular-mass Co-type NHase from R. rhodochrous J1 [13]. Three putative residues that determine the substrate specificity of P. thermophila NHase, i.e. bLeu48, bPhe51, and bTrp72, are fully conserved in the low-molecular-mass Co-type NHase from R. rhodochrous (Fig. 1). In the high- molecular-mass Co-type NHase from R. rhodochrous [12], the kinetic parameters for small aliphatic nitriles were similar to those found for P. thermophila, but the K m values for aromatic nitriles were 1000 times higher. This enzyme contains tryptophan and serine residues at the positions corresponding to bLeu48 and bPhe51 of P. thermophila NHase, respectively. The Fe-type NHase from Pseudo- monas chlororaphis B23 shows only minimal hydration of aromatic nitriles as substrates [11]. The bLeu48, bPhe51 and bTrp72 residues of this enzyme are replaced with valine, valine, and tyrosine, respectively. These three residues are fully conserved in Fe-type NHases, and form a narrower substrate-binding pocket [19]. In summary, differences in the form and size of the hydrophobic pocket Fig. 5. Superimpositioning of the metal centers of the wild-type enzyme and mutants of the cobalt-containing nitrile hydratase (Co-type NHase) and the iron-containing (Fe-type) NHase from Rhodococcus sp. N-771. (A) Vicinity of aThr109 of the Co-type NHase. Thewild-typeenzyme,aT109S mutant, and Fe-type NHase are shown in yellow, red and cyan, respectively. The Cc2atomofaThr109 and the Cc1atomofaVal136 of the Co-type NHase are connected by a yellow line. (B) Vicinity of aTyr114 of the Co-type NHase. The wild-type, aY114T mutant, and Fe-type NHase are shown in yellow, purple, and cyan, respectively. The hydrogen bonds in the aY114T mutant and Fe-type NHase are shown as purple and cyan lines, respectively. Table 3. Distances in the metal centers of the wild-type and mutants. Wild-type a Apoenzyme aT109S aY114T Atoms and distances (A ˚ ) Co-Sc (aCys108) 2.28 – 2.38 2.51 Co-Sc (aCys111) 2.14 – 2.13 2.38 Co-Sc (aCys113) 2.28 – 2.45 2.66 Co-N (aSer112) 2.09 – 2.07 2.15 Co-N (aCys113) 1.96 – 1.92 2.45 Sc (aCys108)-Sc (aCys111) 3.28 3.27 3.39 3.61 Sc (aCys108)-Sc (aCys113) 3.16 2.04 3.15 3.98 Sc (aCys111)-Sc (aCys113) 3.09 4.28 3.26 3.77 a Values taken from reference [19]. 436 A. Miyanaga et al.(Eur. J. Biochem. 271) Ó FEBS 2003 seem to produce the various substrate preferences among NHases. Proposed role of bTyr68 The bY68F mutant showed not only an increased K m value, but also a greatly decreased k cat value. This indicates that the hydroxyl group of the bTyr68 residue plays an important role, not only in substrate binding but also in catalysis. In the proposed reaction model for NHase [2], an imidate intermediate is formed during the reaction. Organic acids, such as n-butyric acid, may inhibit the enzyme as an analogue of the imidate intermediate. bTyr68 is probably involved in the stabilization of this intermediate, as well as in the claw setting, through the hydrogen bond with aSer112 of the claw setting. Difference in metal center between Co-type and Fe-type NHases Payne et al. suggested the importance of the difference between threonine and serine in the cysteine cluster [7]. However, the mutant at this position (aT109S) had a normal active center with a certain amount of a cobalt ion, indicating that the residue is not critical for metal selectivity. The mutant showed a high K m value for a small substrate, acrylonitrile (Table 2), probably owing to an increase in the flexibility of the metal center. When the thermostability of the aT109S mutant was examined, its T m value was found to be  10 °C lower than that of the wild-type enzyme (data not shown). On the other hand, the K m value for benzonitrile of the aT109S mutant was similar to that of the wild-type enzyme (Table 2). The binding affinity of large aromatic substrates seems to originate mainly from hydrophobic interactions around the aromatic ring. In contrast to the aT109S mutant, the cobalt content of the aY114T mutant was decreased dramatically (Table 1), and its two cysteine residues were not modified (Fig. 3C). Therefore, the aTyr114 residue is clearly important for the formation of the active center of P. thermophila Co-type NHase. What kind of structural factor causes such variance? One possibility is that the enzyme is converted to an Fe-type NHase by this mutation. A hydrogen bond between the mutated aThr114 residue and the main-chain oxygen atom of aSer112 was certainly formed. However, the cysteine cluster region was not open, like that of Fe-type NHase(Fig.5B),andnoironatomwasincorporatedwhen the mutant was expressed in iron-supplemented medium. A co-expression experiment on the mutant with a Fe-type NHase activator will confirm this hypothesis. Another possibility is that the aTyr114 residue is involved in the interaction with the Co-type NHase activator of P. thermophila. Co-type NHase activators are not homo- logous with Fe-type NHase activators, and their functions are believed to be different [30–32]. Although Fe-type NHase activators show significant similarity to an ATP- dependent ion transporter, Co-type NHase activators show a slight similarity to the b subunit of NHase. The Co-type NHase activator would assist the formation of the active center during the maturation steps. aTyr114 is located on the molecular surface, near the active center, in the a subunit monomer structure. On the other hand, in the active center of the apoenzyme, a disulfide bond (not cysteine-sulfinic acid or cysteine-sulfenic acid) was formed through oxida- tion. In addition to cysteine residues, aCys108 and aCys113, the aCys111 residue is located in this vicinity. The active center of NHase seems to form a disulfide bond easily under oxidative conditions, if the site contains no metal ion. The metal activator of NHase may incorporate a metal ion to prevent the formation of a disulfide bond, and may facilitate the correct oxidative modification of the cysteine residues. Acknowledgements We wish to thank the staff of the Photon Factory and SPring-8 for their assistance with the data collection. This work was supported by the National Project on Protein Structural and Functional Analysis. References 1. Yamada, H. & Kobayashi, M. (1996) Nitrile hydratase and its application to industrial production of acrylamide. Biosci. Biotechnol. Biochem. 60, 1391–1400. 2. Kobayashi, M. & Shimizu, S. (1998) Metalloenzyme nitrile hydratase: structure, regulation, and application to biotechnology. Nat. Biotechnol. 16, 733–736. 3. Ikehata, O., Nishiyama, M., Horinouchi, S. & Beppu, T. (1989) Primary structure of nitrile hydratase deduced from the nucleotide sequence of a Rhodococcus species and its expression in Escherichia coli. Eur. J. Biochem. 181, 563–570. 4. Nishiyama, M., Horinouchi, S., Kobayashi, M., Nagasawa, T., Yamada, H. & Beppu, T. (1991) Cloning and characterization of genes responsible for metabolism of nitrile compounds from Pseudomonas chlororaphis B23. J. Bacteriol. 173, 2465–2472. 5. Kobayashi, M., Nishiyama, M., Nagasawa, T., Horinouchi, S., Beppu, T. & Yamada, H. (1991) Cloning, nucleotide sequence and expression in Escherichia coli of two cobalt-containing nitrile hydratase genes from Rhodococcus rhodochrous J1. Biochim. Bio- phys. Acta 1129, 23–33. 6. Yamaki, T., Oikawa, T., Ito, K. & Nakamura, T. (1997) Cloning and sequencing of a nitrile hydratase gene from Pseudonocardia thermophila JCM3095. J. Ferment. Bioeng. 83, 474–477. 7.Payne,M.S.,Wu,S.,Fallon,R.D.,Tudor,G.,Stieglitz,B., Turner, I.M. & Nelson, M.J. (1997) A stereoselective cobalt- containing nitrile hydratase. Biochemistry 36, 5447–5454. 8. Odaka, M., Fujii, K., Hoshino, M., Noguchi, T., Tsujimura, M., Nagashima, S., Yohda, M., Nagamune, T., Inoue, Y. & Endo, I. (1997) Activity regulation of photoreactive nitrile hydratase by nitric oxide. J. Am. Chem. Soc. 119, 3785–3791. 9. Bonnet,D.,Artaud,I.,Moali,C.,Petre,D.&Mansuy,D.(1997) Highly efficient control of iron-containing nitrile hydratases by stoichiometric amounts of nitric oxide and light. FEBS Lett. 409, 216–220. 10. Endo, I., Odaka, M. & Yohda, M. (1999) An enzyme controlled by light: the molecular mechanism of photoreactivity in nitrile hydratase. Trends Biotechnol. 17, 244–248. 11. Nagasawa, T., Nanba, H., Ryuno, K., Takeuchi, K. & Yamada, H. (1987) Nitrile hydratase of Pseudomonas chlororaphis B23. Eur. J. Biochem. 162, 691–698. 12. Nagasawa, T., Takeuchi, K. & Yamada, H. (1991) Character- ization of a new cobalt-containing nitrile hydratase purified from urea-induced cells of Rhodococcus rhodochrous J1. Eur. J. Bio- chem. 196, 581–589. 13. Wieser, M., Takauchi, T., Wada, Y., Yamada, H. & Nagasawa, T. (1998) Low-molecular-mass nitrile hydratase from Rhodococcus rhodochorous J1: purification, substrate specificity and comparison Ó FEBS 2003 Mutants of Co-type NHase (Eur. J. Biochem. 271) 437 with the analogous high-molcular-mass enzyme. FEMS Microbiol. Lett. 169, 17–22. 14. Huang,W.,Jia,J.,Cummings,J.,Nelson,M.,Schneider,G.& Lindqvist, Y. (1997) Crystal structure of nitrile hydratase reveals a novel centre in a novel fold. Structure 5, 691–699. 15. Nagashima, S., Nakasako, M., Dohmae, N., Tsujishima, M., Takio, K., Odaka, M., Yohda, M., Kamiya, N. & Endo, I. (1998) Novel non-heme iron center of nitrile hydratase with a claw setting of oxygen atoms. Nat. Struct. Biol. 5, 347–351. 16. Murakami, T., Nojiri, M., Nakayama, H., Odaka, M., Yohda, M., Dohmae, N., Takio, K., Nagamune, T. & Endo, I. (2000) Post-translational modification is essential for catalytic activity of nitrile hydratase. Protein Sci. 9, 1024–1030. 17. Piersma, S.R., Nojiri, M., Tsujimura, M., Noguchi, T., Odaka, M., Yohda, M., Inoue, Y. & Endo, I. (2000) Arginine 56 mutation in the b subunit of nitrile hydratase: importance of hydrogen bond- ing to the non-heme iron center. J. Inorg. Biochem. 80, 283–288. 18. Endo, I., Nojiri, M., Tsujishima, M., Nakasako, M., Nagashima, S., Yohda, M. & Odaka, M. (2001) Fe-type nitrile hydratase. J. Inorg. Biochem. 83, 247–253. 19. Miyanaga, A., Fushinobu, S., Ito, K. & Wakagi, T. (2001) Crystal structure of cobalt-containing nitrile hydratase. Biochem. Biophys. Res. Commun. 288, 1169–1174. 20. Nojiri, M., Nakayama, H., Odaka, M., Yohda, M., Takio, K. & Endo, I. (2000) Cobalt-substituted Fe-type nitrile hydratase of Rhodococcus sp. N-771. FEBS Lett. 465, 173–177. 21. Dixon, M. & Webb, E.C. (1979) Enzymes, 3rd edn. Academic Press, New York. 22. Sarkar, G. & Sommer, S.S. (1990) The megaprimer method of site- directed mutagenesis. Biotechniques 8, 404–407. 23. Powell, H.R. (1999) The Rossmann Fourier autoindexing algo- rithm in MOSFLM. Acta Crystallogr. D 55, 1690–1695. 24. Otwinowski, Z. & Minor, W. (1997) Processing of X-ray diffrac- tion data collected in oscillation mode. Methods Enzymol. 276, 307–326. 25. Brunger,A.T.,Adams,P.D.,Clore,G.M.,DeLano,W.L.,Gros, P.,Grosse-Kunstleve,R.W.,Jiang,J S.,Kuszewski,J.,Nilges, M., Pannu, N.S., Read, R.J., Rice, L.M., Simonson, T. & Warren, G.L. (1998) Crystallography & NMR System: a new software suite for macromolecular structure determination. Acta Crystal- logr. D 54, 905–921. 26. McRee, D.E. (1999) XtalView/Xfit – A versatile program for manipulating atomic coordinates and electron density. J. Struct. Biol. 125, 156–165. 27. Merritt, E.A. & Murphy, M.E.P. (1994) Raster3d, Version 2.0. A program for photorealistic molecular graphics. Acta Crystallogr. D 50, 869–873. 28. Kopf, M.A., Bonnet, D., Artaud, I., Petre, D. & Mansuy, D. (1996) Key role of alkanoic acids on the spectral properties, activity, and active-site stability of iron-containing nitrile hydratase from Brevibacterium R312. Eur. J. Biochem. 240, 239–244. 29. Tsujimura, M., Odaka, M., Nakamura, H., Dohmae, N., Koshi- no, H., Asami, T., Hoshino, M., Takio, K., Yoshida, S., Maeda, M. & Endo, I. (2003) A novel inhibitor for Fe-type nitrile hydratase: 2-cyano-2-propyl hydroperoxide. J. Am. Chem. Soc. 125, 11532–11538. 30. Wu, S., Fallon, R.D. & Payne, M.S. (1997) Over-production of stereoselective nitrile hydratase from Pseudomonas putida 5B in Escherichia coli: activity requires a novel downstream protein. Appl. Microbiol. Biotechnol. 48, 704–708. 31. Nojiri, M., Yohda, M., Odaka, M., Matsushita, Y., Tsujishima, M., Yoshida, T., Dohmae, N., Takio, K. & Endo, I. (1999) Functional expression of nitrile hydratase in Escherichia coli: requirement of a nitrile hydratase activator and post-translational modification of a ligand cysteine. J. Biochem. 125, 696–704. 32. Lu, J., Zheng, Y., Yamagishi, H., Odaka, M., Tujimura, M., Maeda, M. & Endo, I. (2003) Motif CXCC in nitrile hydratase activator is critical for NHase biogenesis in vivo. FEBS Lett. 553, 391–196. 438 A. Miyanaga et al.(Eur. J. Biochem. 271) Ó FEBS 2003 . Mutational and structural analysis of cobalt-containing nitrile hydratase on substrate and metal binding Akimasa Miyanaga 1 ,. no mutational studies on these residues. In this article, we report mutational and structural analysis of the substrate binding and metal specificity of