Purification, Molecular Cloning, and Catalytic Activity of Schizosaccharomyces pombe Pyridoxal Reductase A POSSIBLE ADDITIONAL FAMILY IN THE ALDO-KETO REDUCTASE SUPERFAMILY*

Pyridoxal reductase (PL reductase), which catalyzes reduction of PL by NADPH to form pyridoxine and NADP, was purified from Schizosaccharomyces pombe. The purified enzyme was very unstable but was stabilized by low concentrations of various detergents such as Tween 40. The enzyme was a monomeric protein with the native molecular weight of 41,000 6 1,600. The enzyme showed a single absorption peak at 280 nm (E 5 10.0). PL and 2-nitrobenzaldehyde were excellent substrates, and no measurable activity was observed with short chain aliphatic aldehydes; substrate specificity of PL reductase was obviously different from those of yeast aldo-keto reductases (AKRs) so far purified. The peptide sequences of PL reductase were identical with those in a hypothetical 333-amino acid protein from S. pombe (the DDBJ/EMBL/GenBank accession number D89205). The gene corresponding to this protein was expressed in Escherichia coli, and the purified protein was found to have PL reductase activity. The recombinant PL reductase showed the same properties as those of native PL reductase. PL reductase showed only low sequence identities with members of AKR superfamily established to date; it shows the highest identity (18.5%) with human Shaker-related voltage-gated K channel b2 subunit. The elements of secondary structure of PL reductase, however, distributed similarly to those demonstrated in the three-dimensional structure of human aldose reductase except that loop A region is lost, and loop B region is extended. Amino acid residues involved in substrate binding or catalysis are also conserved. Conservation of these features, together with the major modifications, establish PL reductase as the first member of a new AKR family, AKR8.

Pyridoxal reductase (PL 1 reductase) (formerly designated as PN dehydrogenase) catalyzes reduction of PL with NADPH and oxidation of PN with NADP ϩ as the reverse reaction. The enzyme was for the first time found in a budding yeast, Saccharomyces cerevisiae, i.e. bakers' (1) and brewers' (2) yeasts. Guirard and Snell (3) have purified it to homogeneity from bakers' yeast and showed that the enzyme is a monomeric protein with the molecular weight of about 33,000. They designated the enzyme as PL reductase because of the equilibrium of the enzyme reaction lying so far to formation of PN, and the substrate specificity, molecular weight, and the monomeric structure of the enzyme. The enzyme resembled chlordecone reductase (4) in optimal pH, molecular weight, specific requirement of for NADPH, and behavior toward sulfhydryl reagents, barbital, and common ketone substrates.
The chlordecone reductase from human liver belongs to family 1 of the AKR superfamily. 2 The AKRs form an expanding oxidoreductase superfamily classified into seven families containing a variety of monomeric oxidoreductases such as aldehyde and aldose reductases, hydroxysteroid dehydrogenases, chalcone reductases, Shaker-related voltage-gated K ϩ channel ␤2 subunits, and aflatoxin B 1 -aldehyde reductases (5). Because the nomenclature system is based on identities in amino acid sequence (5), elucidation of primary structure of PL reductase is required to classify the enzyme in the family.
We recently found that a fission yeast, Schizosaccharomyces pombe, also contained PL reductase activity, suggesting that the enzyme also plays an important role in other eukaryotes. Here we show the purification and properties of PL reductase from S. pombe. Primary structure of the enzyme was determined based on cloning and expression of its gene in Escherichia coli. The results showed that the enzyme is the founding member of the 8th AKR family.

Materials
NAD ϩ , NADP ϩ , NADH, NADPH, and a Cosmosil C 18 column were purchased from Nacalai Tesque, Kyoto, Japan. PL, PN, and PLP were from Wako Chemicals, Osaka, Japan. Matrix Orange A was purchased from Grace Japan, Amicon, Tokyo, Japan. Butyl-Toyopearl and DEAE-Toyopearl were from Tosoh Corp., Tokyo, Japan. An Ultra-Free concentrator was from Milipore, Tokyo, Japan. A lysyl endopeptidase was from Roche Molecular Biochemicals, Mannheim, Germany. Primer DNAs for PCR were from Kurabo, Osaka, Japan. Restriction enzymes, E. coli JM109 competent cells, a Takara LA-PCR buffer II (Mg 2ϩ plus), a Takara LA Taq polymerase, and a SUPREC-01 tube were from Takara Shuzo, Kyoto, Japan. A Ligation High was from Toyobo, Osaka, Japan. A plasmid pTrc99A was from Amersham Pharmacia Biotech, Uppsala, Sweden. A plasmid SY1115 carrying the PL reductase gene of S. pombe * The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
The nucleotide sequence reported in this paper has been submitted to the DDBJ/GenBank TM /EBI Data Bank with the accession number D89205.

Microorganism and the Culture Conditions
S. pombe IFO 0346 cells were grown at 30°C for 72 h with shaking in a synthetic medium containing 1% D-glucose, 100 mM PN, and yeast nitrogen base, containing neither amino acids nor thiamine. The cells were harvested by centrifugation at 10,000 ϫ g for 10 min, washed twice with 0.9% NaCl, and then stored at Ϫ20°C.

Enzyme Assays
PL reductase was assayed by measuring the initial decrease in A 366 of NADPH at 37°C in 1.0 ml of the reaction mixture. Each reaction mixture contained 0.2 mM PL, 0.2 mM NADPH, 0.1 M MOPS/KOH buffer (pH 7.5), and the enzyme. The reaction was started by addition of PL. One unit of activity was defined as the amount of enzyme required to reduce 1 mol of PL per min.

Enzyme Purification
All the steps were performed at 4 -10°C.
Step 1-Frozen cells (195 g, wet weight) were thawed and suspended in 390 ml of Buffer C, 100 mM sodium citrate (pH 6.5) containing stabilizing reagents (1 mM PMSF, 0.01% 2-mercaptoethanol, 1 mM EDTA, and 0.1 mM PL). The suspension (100 ml, each) was vigorously vortexed for 1 min and then centrifuged at 10,000 ϫ g for 10 min to separate the supernatant solution. The precipitated cells were resuspended in 390 ml of Buffer C, vortexed, and centrifuged. The step was repeated again. The combined supernatant solutions (1165 ml) were used as a crude extract.
Step 2-NaCl (final 2 M) was added to the crude extract, and the solution was applied to a butyl-Toyopearl column (1.8 ϫ 18 cm) equilibrated with Buffer T (10 mM Tris acetate buffer, pH 6.0, containing the stabilizing reagents and 2 M NaCl). The column was washed with Buffer T; the enzyme activity was eluted as a broad single peak from about 700 to 1770 ml of eluate. The active fractions (1070 ml) were combined and concentrated to 65 ml with butyl-Toyopearl by an adsorption-desorption cycle. Tween 40 (0.005%) was immediately added to the concentrated solution to stabilize the enzyme.
Step 3-The concentrated enzyme solution was dialyzed at 4°C overnight against 2 liters of Buffer M (10 mM MOPS/KOH buffer, pH 7.5, containing the stabilizing reagents, 0.005% Tween 40, and 0.1 M NaCl). The dialyzed solution was applied to an Orange A column (0.9 ϫ 3.2 cm) equilibrated with Buffer M. After the column was washed with 150 ml of Buffer M, the enzyme was eluted with Buffer M containing 1 mM NADPH.
Step 4 -The active fractions obtained by Orange A column chromatography (8.5 ml) were dialyzed at 4°C overnight against 1 liter of Buffer A (10 mM MOPS/KOH buffer, pH 7.5, containing the stabilizing reagents and 0.005% Tween 40). The dialyzed solution was applied to a DEAE-Toyopearl column (1.0 ϫ 40 cm). The enzyme was eluted with Buffer A. The active fractions (53.6 ml) were concentrated to 2.1 ml with an Ultra-Free concentrator. The concentrated solution was used as the purified enzyme.

Molecular Weight Determination
The purity of the enzyme and the subunit molecular weight were estimated by SDS-PAGE by the method of Laemmli (6). The molecular weight of the enzyme was estimated by gel filtration at 4°C, using an Ultrogel AcA 34 column (1.2 ϫ 110 cm) equilibrated with 10 mM MOPS/ NaOH buffer (pH 7.5) containing 1 mM EDTA, 1 mM PMSF, 0.02% 2-mercaptoethanol, 0.005% Tween 40, and 0.1 M NaCl at a flow rate of 0.4 ml/min. A calibration curve was made from the elution pattern of bovine liver catalase (M r ϭ 240,000), pig heart mitochondrial aspartate aminotransferase (90,000), malate dehydrogenase (67,000), and horse heart cytochrome c (12,400). The protein concentration was determined by means of a protein assay kit (Bio-Rad) with bovine serum albumin as the standard.

Substrate Specificity of PL Reductase
K m and V max values were determined by measuring initial reaction velocities with concentrations of substrate between 0.1ϫ K m and 10ϫ K m , whereas NADPH concentration was held at 0.2 mM. Kinetic parameters were determined by employing the curve-fitting software (Kalei-daGraph) to fit the Michaelis-Menten equation using the Levenberg-Marquardt algorithm.

Amino Acid Sequencing
The enzyme (1.95 nmol) was digested at 37°C for 6 h with a lysyl endopeptidase (23 pmol). The digested peptides were isolated by reversed-phase high performance liquid chromatography on a Cosmosil C 18 column (4.6 ϫ 150 mm) with a linear gradient from 0 to 90% acetonitrile containing 0.1% trifluoroacetic acid in a Shimadzu LC-10A system. The amino acid sequences were determined by an automated Edman degradation with an Applied Biosystems 492 protein sequencer.

Construction of Plasmids Carrying PL Reductase Gene
The 1.2-kbp fragment carrying the PL reductase gene (plr ϩ ) was cut out from a plasmid SY1115 with both EcoRI and KpnI, collected using a SUPREC-01 tube, and then ligated into the EcoRI-KpnI site of pTrc99A. The recombinant plasmid was designated as pPLR1. The PL reductase-coding sequence in pPLR1 was amplified by PCR using the two oligonucleotides 5Ј-GCGAAGCTTGGTACCAGGAGGAGTGAAAAGAT-GCCTATCGTTAGCGGATTTAA-3Ј (primer NF, sense) and 5Ј-GCGG-GATCCTTAAACGGAAAGAGTGCCCGCAAGCTGTTCATT-3Ј (primer CR, antisense) to introduce BamHI (underlined in primer NF), ribosome-binding (boldface types in primer NF), and HindIII sites (underlined in primer CR). The reaction mixture (50 l) of PCR consisted of Takara LA-PCR buffer II (Mg 2ϩ plus), 20 nmol of each dNTP, 2.5 units of Takara LA Taq polymerase, 0.5 mg of a plasmid pPLR1 (as a template), and 100 pmol of each primer. The mixture was heated at 98°C for 20 s and then incubated at 68°C for 20 min. The programmed temperature shift was repeated 16 times, and then, without delay, "autosegment extension" program was done as follows: heating at 98°C for 20 s and incubation at 68°C for 20 min ϩ t s, where t denotes the segment extension time that increases by 15 s at each cycle; the temperature shift was repeated 14 times. Finally, the mixture was held at 72°C for 10 min. The amplified DNA fragments (about 1.0 kbp) were digested by BamHI and HindIII and then ligated into the BamHI-HindIII site of pTrc99A. The constructed plasmid was designated as pPLR2. A plasmid pPLR3 carrying a modified plr, in which a possible internal palindrome structure in plr ϩ was disrupted by substitutions of certain nucleotides without changing the amino acid sequence of PL reductase, was prepared as follows. The 669-bp fragment of plr ϩ , encoding the amino-terminal 223 amino acids of PL reductase, was amplified by PCR using the two oligonucleotides as the primers: primer NF (sense) and 5Ј-GCGGAATTCTTTAAGGTCTTCGACAGTCTTAATGCG-ACC-3Ј (primer MR, antisense) to introduce BamHI and ribosome-binding sites and EcoRI site (underlined in primer MR). PCR was performed  The plasmids pPLR1, pPLR2, or pPLR3 were introduced into Takara E. coli JM109-competent cells. The clone cells were used for analysis of the enzyme production in E. coli.

Purification of Recombinant PL Reductase
Cells of E. coli JM109 clone harboring pPLR3 were grown in 5 liters of LB medium (7) containing ampicillin (50 g/ml) and isopropyl-␤-Dthiogalactopyranoside (1 mM) at 37°C for 16 h. The cells (wet weight, 18.9 g) were harvested and disrupted by sonication to prepare crude extracts. PL reductase (5.3 mg of total protein; 116.0 units of total activity) was purified from the crude extracts (2,100 mg of total protein; 1,281.6 units of total activity) by the same purification procedure as that used for purification of the enzyme from S. pombe.

Homology Search of PL Reductase with Other Proteins
DNA and protein data bases in the DDBJ/EMBL/GenBank TM were searched for proteins homologous with PL reductase by use of the BLAST algorithm (8).

RESULTS
Purification and Some Properties of PL Reductase-The enzyme was purified to homogeneity by three steps of column chromatography (Table I). Almost all of the enzyme activity was leached from the frozen yeast cells, but not from intact yeast cells, by the vigorous mixing. On the contrary, lactate dehydrogenase (a cytosolic enzyme) activity was not extracted from the frozen yeast cells, suggesting different cellular localization of the two enzymes. PL reductase was stable in a buffer without a detergent before the butyl-Toyopearl column chromatography. However, after the step, the enzyme became quite unstable; 95.2% of original activity was lost after 24 h when stored at 4°C. Very low concentrations of several detergents stabilized the enzyme; 0.005% of Tween 40 was more effective at this concentration than Triton X-100, Tween 20, or Tween 80.
The purified enzyme showed a single protein band with a molecular weight of 37,600 Ϯ 370 (average and S.D. of four experiments) on an SDS-PAGE gel (Fig. 1). The molecular weight of the native enzyme was 41,000 Ϯ 1,600 (average and S.D. of three experiments) by gel filtration. The enzyme is thus a monomeric protein.
The optimum pH for PN formation was 6.5-7.5; PL reductase was not active at pH 5.0, but 17% of the maximal activity was displayed at pH 8.5. The optimum pH for PL formation was 7.5-8.5; the enzyme was not active at pH 6.5, but 18% of the maximal activity was displayed at pH 10.0. NADPH (K m ϭ 16 Ϯ 0.8 mM at pH 7.5) and NADP ϩ were required for PN and PL formations, respectively; NADH and NAD ϩ were inactive as coenzymes.  the enzyme may be involved in the interaction of the enzyme and NADPH. Substrate Specificity-Steady-state kinetic parameters for PL reductase are given in Table II. All substrates showed Michaelis-Menten kinetics in the concentration range studied. PL and 2-nitrobenzaldehyde were excellent substrates. Although PL was a substrate with the highest k cat , K m was 9.4-fold higher than that of 2-nitrobenzaldehyde; the hemiacetal form of PL, which is predominant at around pH 7.0, may not be used as a substrate. Thus, 2-methyl and 3-hydroxyl groups and ring nitrogen in PL are not essential for high substrate activity. In contrast, the 5-hydroxymethyl group in PL and the nitro group (but not the carboxyl group) adjacent to the formyl group in 2-benzaldehyde were essential: pyridine-4-aldehyde was a poor substrate, and benzaldehyde or 2-carboxybenzaldehyde was not substrate. 2-Phthalaldehyde was a moderate substrate and gave Michaelis-Menten kinetics when initial reaction rates were measured. However, PL reductase was inactivated during reduction of 2-phthalaldehyde by an unknown reason. No measurable activity was observed with short chain aliphatic aldehydes.
Amino Acid Sequences of Peptides of PL Reductase and Search for the Enzyme Gene-The amino-terminal sequence could not be determined by Edman degradation probably because an amino-terminal amino acid residue is modified. Then, we determined the amino acid sequences of lysylendopeptides and cydnogen bromide peptides of the enzyme (shown by underlines in Fig. 2). Homology search was performed with these sequences. Sequences of a hypothetical protein of S. pombe composed of 333 amino acids (the DDBJ/EMBL/GenBank TM accession number D89205) were found to be identical. The predicted molecular weight (M r ϭ 36,933) agreed well with that of PL reductase.
Expression of PL Reductase Gene in E. coli-The presumed PL reductase gene (plr ϩ ) was ligated into pTrc99A, and recombinant plasmids (pPLR1, pPLR2, and pPLR3), schematic structures of which are shown in Fig. 3A, were used to transform E. coli JM109 cells. Fig. 3B shows that all the recombinant plasmids expressed PL reductase activity, indicating plr ϩ encodes PL reductase. The enzyme activity in the crude extract of E. coli cells transformed by pPLR1 was very low (an average specific activity of 0.023 unit/mg) even though the host plasmid, pTrc99A, is a high expression vector. A typical bacterial ribosome-binding sequence was bound to the 5Ј-end of plr ϩ , and the modified gene was ligated into pTrc99A to increase an expression level. However, the obtained plasmid, pPLR2, did not express higher amount of PL reductase than pPLR1 (Fig.   3B). As shown in the scheme, plr ϩ contained a 47-bp palindrome structure (nucleotides 752-805), which may restrict effective expression of the enzyme. When the palindrome structure was disrupted by site-directed mutagenesis to prepare pPLR3 (see "Experimental Procedures"), the derived clone harboring pPLR3 (E. coli JM109/pPLR3) produced high levels of PL reductase; an average specific activity was 0.78 unit/mg. Thus, the palindrome structure in plr ϩ appears responsible for the low expression in E. coli.
Properties of Recombinant PL Reductase-PL reductase from E. coli JM109/pPLR3 showed the same enzymatic properties, such as the molecular weight, specific activity, and aminoterminal 12 amino acids sequence, as the enzyme from S. pombe.
Sequence Comparisons-The amino acid sequence of PL reductase was compared with other protein sequences. Unexpectedly, no proteins belonging to the AKR superfamily showed higher than 20% homology. However, 12 proteins, whose functions are unknown, were found to have higher identity than 20% (Table III). The F8A5.2 protein of Arabidopsis thaliana showed the highest identity; this protein is an auxin-induced protein. YAKC protein of S. pombe was similar to PL reductase, suggesting it is an isoenzyme of PL reductase. IolS protein of Bacillus subtilis is encoded by a gene which is a member of operon of myo-inositol degradation in the bacterium (9), but its function is unknown.
Secondary Structure Comparisons-Although the primary structure of PL reductase showed low identity with the AKRs so far reported, a secondary structure predicted with a Genetyx software (Software Development, Tokyo, Japan) was similar to that of human aldose reductase (10,11), a typical member of AKRs. Fig. 4 shows the alignment of PL reductase, human Shaker-related voltage-gated K ϩ channel ␤2 subunit (12,13), human 2-CBA reductase (14), and human aldose reductase and secondary elements of the former two proteins. The human Shaker-related voltage-gated K ϩ channel ␤2 subunit and human 2-CBA reductase were aligned because they showed high identity (18.3 and 13.5%, respectively) with PL reductase relative to most AKRs. A region corresponding to loop A in human aldose reductase is not found in PL reductase, human Shakerrelated voltage-gated K ϩ channel ␤2 subunit, or human 2-CBA reductase, and all three proteins contain the loop B region, which is not found in human aldose reductase. However, the elements of secondary structure are distributed similarly as in human aldose reductase, and amino acid residues involved in substrate binding or catalysis are conserved among these proteins as discussed below (10, 11, 14). colonies of E. coli JM109/pTrc99A, JM109/pPLR1, JM109/pPLR2, and JM109/pPLR3 clones were isolated. They were individually grown in 100 ml of LB medium (7) containing 50 g/ml ampicillin and 1 mM isopropyl-␤-D-thiogalactopyranoside. PL reductase activity in the crude extracts of the cells (0.2 g, wet weight) was measured. An average and standard deviation of the specific activity is shown. 47, Tyr-53, and Lys-83 may be the active site residues, making spatially conserved arrangement as seen in human aldose reductase. Ala-123 in PL reductase corresponds to His-117 in human aldose reductase but cannot fill the role of proton donor suggested (10,11) for His-117 in human aldose reductase. An alternative candidate for this role, Tyr-53, may be involved in the catalytic action of PL reductase. It has been noted that His-117 in human aldose reductase is also less likely to function as the proton donor at physiological pH (21).
Jez et al. (5) have proposed a system for nomenclature of the AKR superfamily based on amino acid sequence identities. They distinguish seven families based on amino acid identities of 40% or more. Since PL reductase shows less than 20% identity with these other AKR proteins, we suggest that it represents a new AKR family, 3 AKR8. Among several other proteins with a similar amino acid sequence but unknown function, we have cloned and expressed the iolS gene of B. subtilis in E. coli and found that the IolS protein indeed shows PL reductase activity but not aldose reductase activity. 4 Because the iolS gene is a structural gene of the myo-inositol operon (9), there may be some relation between vitamin B 6 and myo-inositol metabolism.
Although PL reductase from S. pombe showed properties similar to those of the enzyme from S. cerevisiae (3), the former enzyme did not catalyze reduction of PLP, an important fact in ascribing a possible function to the enzyme. PLP, a coenzyme form of vitamin B 6 , is mainly synthesized from PN through pyridoxine 5Ј-phosphate in cells (22). If PL reductase catalyzes reduction of PLP, the enzyme by itself without cooperation with a phosphatase may control the synthesis of PLP by reversing the synthetic pathway. However, this may be not the case. As Guirard and Snell (3) have suggested, PL reductase may be involved in salvage of vitamin B 6 and excretion of PN from yeast cells by reduction of PL to PN. Actually, we have found that S. pombe excretes fairly high amounts of PN into cultivation medium during its cultivation (23). The results that PL reductase leaked easily from the frozen cells and was stabilized by detergents having hydrophobic side chains and localized in a cellular compartment different from that occupied by lactate dehydrogenase suggest that PL reductase may be an extrinsic protein which is weakly bound to the membrane lipids inside the cell wall. This also supports the possibility that PL reductase play an important role in the excretion of PN from the yeast cell. We are examining the localization of PL reductase.