Cloning, Sequencing, and Enzymatic Activity of an Inducible Aldo-Keto Reductase from Chinese Hamster Ovary Cells*

Treatment of Chinese hamster ovary (CHO) cells by the aldehyde containing calpain inhibitor I resulted in the induction of a 35-kDa protein that was partially sequenced and shown to be a member of the aldo-keto reductase superfamily (Inoue, S., Sharma, R. C., Schimke, R. T., and Simoni, R. D. (1993) J. Biol. Chem. 268, 5894–5898). Using rapid amplification of cDNA ends polymerase chain reaction, we have sequenced the cDNA for this protein (CHO reductase). This enzyme is a new member of the aldo-keto reductase superfamily and shows greatest amino acid sequence identity to mouse fibroblast growth factor-regulated protein and mouse vas deferens protein (92 and 80% sequence identity, respectively). The enzyme exhibits about 70% sequence identity with the aldose reductases (ALR2; EC 1.1.1.21) and about 47% with the aldehyde reductases (ALR1; EC 1.1.1.2). Northern analysis showed that it is induced in preference to either ALR1 or ALR2 and RNase protection assays showed gene expression in bladder, testis, jejunum, and ovary in descending order of expression. The cDNA for this inducible reductase was cloned into the pET16b vector and expressed in BL21(DE3) cells. Expressed CHO reductase showed kinetic properties distinct from either ALR1 or ALR2 including the ability to metabolize ketones. This protein joins a growing number of inducible aldo-keto reductases that may play a role in cellular regulation and protection.

To characterize the calpain inhibitor I-sensitive protease(s) involved in the degradation of 3-hydroxy-3-methylglutaryl-CoA reductase, Simoni's group (1) attempted to isolate Chinese hamster ovary (CHO) 1 cells resistant to this peptide aldehyde (N-acetyl-leucyl-leucyl-norleucinal (ALLN)). Instead of inducing a protease, a 35-kDa protein was overexpressed which gave tryptic peptide fragments with a high degree of sequence iden-tity to members of the aldo-keto reductase superfamily. 2 This superfamily is a rapidly growing group of monomeric oxidoreductases containing at least 40 members at present (2). This group includes the aldehyde and aldose reductases, a number of hydroxysteroid dehydrogenases, Shaker channels, and plant chalcone reductases. They are characterized by a TIM-barrel structure (3) and the preferential use of NADPH over NADH. Substrates include aliphatic and aromatic aldehydes, monosaccharides, steroids, prostaglandins, polycyclic aromatic hydrocarbons, and isoflavinoid phytoalexins.
Several members of this family have been shown to be induced in response to hormonal or chemical factors. These include fibroblast growth factor-regulated protein (FR-1) (4), mouse vas deferens protein (MVDP) (5), aldose reductase (ALR2) (6 -8), and dihydrodiol dehydrogenase (9). To determine the relationship of this new reductase to the other members of the family we have used RACE PCR to isolate and sequence its mRNA from CHO cells. It shows highest sequence identity to FR-1 and MVDP at 92 and 80%, respectively. FR-1 and MVDP have not been extensively characterized kinetically but CHO reductase showed a greater ability to reduce ketones than the aldehyde or aldose reductases and a strong preference for aromatic aldehydes.

EXPERIMENTAL PROCEDURES
Materials-Calpain inhibitor I was purchased from Calbiochem (La Jolla, CA). Daunomycin was purchased from Rhône-Poulenc Rorer (Montreal, Quebec, Canada). NADPH and NADH were from Boehringer Mannheim (Laval, Quebec). All steroids were from Steraloids (Wilton, NH). p-Nitroacetophenone was from Aldrich while all other enzyme substrates were from Sigma. T7 RNA polymerase, Moloney murine leukemia virus reverse transcriptase, and all restriction enzymes were from Promega (Madison, WI), while RNase A, RNase T1, and DNase I were from U. S. Biochemical Corp. (Cleveland, OH).
Cell Culture-Induced CHO-K1 cells, kindly provided by Dr. Simoni, were grown essentially as described by Inoue et al. (1) to a final ALLN concentration of 70 M in the presence of 20 M verapamil. Cells were collected and stored at Ϫ70°C until assayed. Control CHO-K1 cells were obtained from the American Type Culture Collection and grown without ALLN or verapamil.
cDNA Generation and Sequencing-Total RNA was isolated from induced cells by the TRIzol method (10). First strand synthesis was primed with the RACE adapter T 17 primer using Moloney murine leukemia virus reverse transcriptase (11). An initial cDNA fragment for CHO reductase was generated using 3Ј-RACE PCR with a primer based on the human aldose reductase sequence (5Ј-CTCAACAACGGCGC-CAAGATGCCCA corresponding to position 19 to 43 of the coding region). Subsequently, gene specific primers were used with 5Ј-RACE PCR to obtain the 5Ј-end cDNA. The 5Ј-RACE procedure was improved by titrating the amount of gene-specific primer over the range 0.001 to 10 pmol per reaction (12). The entire cDNA was sequenced in both directions using overlapping gene specific primers. Sequencing was performed by the dye terminator method using an Applied Biosystems model 377A sequencer. * The work was supported by a grant from the Medical Research Council of Canada. 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.
Total RNA from noninduced cells was used with 3Ј-RACE PCR to generate cDNA fragments for hamster aldehyde reductase (ALR1; EC 1.1.1.2) and aldose reductase (ALR2; EC 1.1.1.21). A coding region fragment for aldehyde reductase of 735 bp was generated using the primers 5Ј-TGTCCTCCTGCACACTG (sense) (position 15 to 31 of the rat coding region) and 5Ј-GGACCTGCCACCTGAGC (antisense) (position 734 to 750 of the pig coding region). A 3Ј-end fragment of hamster aldose reductase was generated using the RACE adapter primer and the primer 5Ј-GCCTTGATGAGCTGTG (sense) (position 898 to 913 of the mouse coding region). These fragments were sequenced and aligned with known members of the AKR family to ensure their correct assignment for use as probes in Northern analysis. 3Ј-RACE cDNA from noninduced total RNA was used to generate a Hamster glyceraldehyde-3-phosphate dehydrogenase probe using a sense (5Ј-GTTGCCATCAAC-GACCCCTT) and antisense (5Ј-AGCATCAAAGGTGGAAGAATG) primer derived from the rat sequence.
Cloning, Expression, and Purification of CHO Reductase-The entire CHO reductase cDNA was amplified using primers in the 5Ј-and 3Ј-untranslated region containing restriction sites for EcoRI and XhoI, respectively. This fragment was digested with EcoRI and XhoI and ligated into pCRII (Invitrogen) similarly digested. The plasmid CHOpCRII was transformed into CJ236 cells. Introduction of a NdeI site at the N-terminal methionine was performed using the Kunkel method (13). Successful mutagenesis was screened by NdeI digestion and positive clones were resequenced to ensure introduction of no other mutations. The NdeI-XhoI fragment was subcloned into the pET16b vector (Novagen) to generate an in-frame histidine leader sequence for HisTag purification. The CHOpET16b plasmid was transformed into BL21(DE3) cells containing pLysS. One-half liter cultures were grown in LB supplemented with 50 g/ml ampicillin and 34 g/ml chloramphenicol at 30°C to an absorbance of 0.5 at 600 nm. Isopropyl-1-thio-␤-D-galactopyranoside was added to a final concentration of 1 mM and the cultures were grown for a further 3 h. The cells were harvested by centrifugation at 1000 ϫ g for 15 min. The pellets were resuspended at 30 ml/liter of culture in bind buffer (5 mM imidazole, 0.5 M NaCl, 20 mM Tris, pH 7.9) and stored at Ϫ20°C. Cells were thawed and phenylmethylsulfonyl fluoride added to a final concentration of 1 mM. Cells were disrupted by sonication and the insoluble fraction separated by centrifugation at 12,000 ϫ g for 20 min at 4°C. The supernatant was applied to a 5-ml chelating Sepharose column (Pharmacia) previously charged with nickel sulfate and washed with bind buffer. The loaded column was washed with 60 mM imidazole, 0.5 M NaCl, 20 mM Tris, pH 7.9 (wash buffer), and CHO reductase was eluted using a linear gradient from wash buffer to elute buffer (1 M imidazole, 0.5 M NaCl, 20 mM Tris, pH 7.9). The pooled protein was dialyzed against 10 mM sodium phosphate buffer, pH 8.0, containing 1 mM EDTA and stored at 4°C or Ϫ20°C. Protein samples from the purification of CHO reductase were run on 12% SDS-polyacrylamide minigels and stained with Coomassie Brilliant Blue R.
Northern Blot Hybridization-Total RNA was isolated by the TRIzol method as described above. Total RNA was glyoxalated (13) and separated on a 1% agarose gel in 10 mM sodium phosphate, pH 7.0, and transferred to a nylon membrane (Zeta-Probe, Bio-Rad). Prehybridization was done at 37°C for at least 1 h in 6 ϫ SSC, 5 ϫ Denhardt's, 0.5% SDS, and 100 g/ml salmon sperm DNA. Hybridization was performed overnight at 42°C in the same solution containing at least 1 ϫ 10 7 cpm/ml of each probe at 1 ϫ 10 8 cpm/g. The probes were either an 800-bp hamster glyceraldehyde-3-phosphate dehydrogenase cDNA, a 735-bp hamster ALR1 cDNA, a 450-bp hamster ALR2 cDNA, or the full-length CHO cDNA prepared by nick-translation (Nick Translation System, Life Technologies, Inc.). Final membrane washes were at 0.1 ϫ SSC, 0.5% SDS, 60 min, 70°C with exposure for 16 h at 25°C with DuPont Reflection NEF film and one intensifying screen.
Ribonuclease Protection Assay of Chinese Hamster Tissues-A radiolabeled riboprobe was prepared by in vitro transcription of the 3Ј portion of the hamster CHO reductase gene with T7 RNA polymerase in the presence of [ 32 P]CTP. The plasmid CHOpCRII was linearized at the unique AccII site to generate a 200-bp cRNA transcript. The polyacrylamide gel-purified riboprobe (0.5 ϫ 10 6 cpm) and tissue RNA (20 g) or CHO-K1 cell-derived RNA (2-5 g) were denatured at 85°C for 10 min and annealed overnight at 55°C in hybridization buffer (80% formamide, 40 mM PIPES, pH 6.5, 0.4 M NaCl, 1 mM EDTA). Sample processing and display were as described previously except that the ribonuclease digestion buffer contained 50 g/ml RNase A in addition to 2 g/ml RNase T1 (14).
Enzyme Kinetics-All substrates were assayed at 25°C in 1 ml of 0.1 M sodium phosphate, pH 7.0, using a Hewlett-Packard 8452A diode array spectrophotometer, except for daunomycin which was assayed in 0.1 M HEPPS, pH 8.5. Apparent K m values for the substrates were assayed in the presence of 200 M NADPH. Apparent K m values for the cofactors were assayed in the presence of 1 mM pyridine-3-aldehyde. NADH was assayed in a 5-mm path-length cell at 370 nm with correction for the different ⌬⑀ at that wavelength. All steroids were assayed in the presence of 2.5% acetonitrile. Kinetic constants were determined using the Marquardt algorithm (15) supplied with the spectrophotometer.

RESULTS
Nucleotide Sequence of CHO Reductase cDNA-Sequence information obtained from the initial 3Ј-RACE reaction performed on induced CHO-K1 total RNA using the primer to human ALR2 was different from known ALR2 sequences. Specific primers were used to sequence the sense and antisense strands and to obtain the 5Ј-untranslated region using 5Ј-RACE. The full-length cDNA sequence generated by RACE The deduced protein sequence of CHO reductase showed high identity to members of the aldo-keto reductase superfamily. The highest identity was to the mouse fibroblast growth factor regulated (FR-1) protein at 92% (4) (Fig. 2). CHO reductase was 80% identical to the MVDP (5). The aldose reductases were approximately 70% identical while the aldehyde reductases were approximately 47% identical. Multiple sequence alignment analysis showed that CHO reductase, FR-1, and MVDP form a distinct subgroup within the superfamily and should be classified in the AKR1B group using the nomenclature system proposed by Jez et al. (2). CHO reductase contained the same residues implicated in the catalytic mechanism as other members of the aldo-keto reductase family. These included Asp-44, Tyr-49, Lys-78, and His-111.
Induction of CHO Reductase mRNA in CHO-K1 Cells-RNA gel blot hybridization analysis from cells treated with ALLN at a concentration of 70 M confirmed the induction of CHO re-ductase (Fig. 3). This induction was instead of aldose reductase or aldehyde reductase. The selectivity of the aldose reductase probe was maximized by using the 3Ј-end of its cDNA, which contained the greatest differences between the hamster aldose reductase and CHO reductase sequences.
CHO Reductase mRNA Levels in Chinese Hamster Tissues-Analysis of a range of CHO tissues revealed no detectable levels of CHO reductase mRNA when probed by Northern blot using the full-length CHO cDNA as probe. These blots did confirm the integrity of the RNA. RNase protection assay analysis of the same RNA stocks showed transcript in the bladder, testis, jejunum, and to a lesser extent in the ovary (Fig. 4). The basal level in CHO-K1 cells was higher than in all tissues tested comparing the signal from 5 g of noninduced cells total RNA with 20 g of tissue RNA. As indicated by the Northern analysis, cells induced by ALLN showed a markedly elevated transcript level.
Expression and Kinetic Analysis-The recombinant protein yield was 45-50 mg/liter of cells after chromatography on the nickel-Sepharose column. The protein was greater than 98% pure as judged by Coomassie staining of SDS-polyacrylamide gels. The expressed CHO reductase was active when tested with a variety of common substrates for the aldo-keto reductases. The K m values for the cofactors NADPH and NADH were very similar to those obtained for aldehyde and aldose reductase (Table I). ALLN, which was used to induce the protein, was a substrate although it bound poorly as indicated by its millimolar K m . Overall aliphatic substrates such as the commonly used DL-glyceraldehyde were poor substrates. The only kinetic values for FR1 or MVDP are a K m of 0.92 mM for FR1 Br, brain. Panel A was exposed for 24 h. Panel B was exposed for 30 min to highlight the induced CHO reductase band.

TABLE II
Kinetic constants for CHO reductase with aromatic aldehydes ALR1 and ALR2 sources as indicated in Table I. with DL-glyceraldehyde (16). This value falls between that obtained for CHO reductase and what has been observed with ALR2 (Table I). The poor ability to catalyze aliphatic substrates also included glucose which was only converted at a level of 2-3 times background up to a concentration of 1.2 M. Small aromatic aldehydes were better substrates (Table II), with micromolar K m values and k cat values 2-10-fold higher than for the aliphatic substrates. A number of ketones were substrates for the CHO reductase (Table III). They exhibited low k cat values (Ͻ1 s Ϫ1 ) and tighter binding for aromatic or heterocyclic substrates. The diagnostic ketone substrate for the carbonyl reductases, menadione (K m of 0.045 mM (17)), was not a substrate for CHO reductase. CHO reductase was also able to catalyze the reduction of a number of steroid aldehydes and ketones (Table  IV). The affinities were high (low M) while the turnover rates were again less than 1 s Ϫ1 . The enzyme showed selectivity for aldehydes or ketones on ring D (e.g. 17␣-hydroxyprogesterone) and no reactivity toward ketones on ring A (e.g. testosterone). The enzyme showed a pH optimum of 6 to 6.5 with all substrates except daunomycin which was maximally reduced at pH 8.5 (data not shown). CHO reductase was inhibited 35% by 1 mM sodium barbitone and 31% by 1 mM sodium valproate. DISCUSSION We have isolated and sequenced a full-length cDNA from Chinese hamster ovary cells that encodes CHO reductase, a new member of the aldo-keto reductase superfamily. Initial characterization of this protein was described by Inoue et al. (1) and the sequence presented here confirms their assignment to the aldo-keto reductase superfamily. The enzyme showed highest protein sequence identity to the fibroblast growth factor-regulated protein (FR-1) and to MVDP. FR-1, MVDP, and now CHO reductase comprise a subgroup of the aldo-keto reductases that are inducible. FR-1 is induced in cells by exposure to acidic fibroblast growth factor-1 (4) and thus may play a role in the mitogenic action of fibroblast growth factor-1. MVDP expression in the vas deferens was dependent on the presence of testosterone (5) and androgen response elements have been isolated in the MVDP gene (18). Both these murine aldo-keto reductases appear to be expressed instead of the more common aldehyde or aldose reductases. CHO reductase induction was also instead of the Chinese hamster aldehyde or aldose reductases.
Besides FR-1 ((AKR1B8); nomenclature proposed by Jez et al. (2)), MVDP (AKR1B7) and now CHO reductase (AKR1B9), many other members of the AKR superfamily have been shown to be induced by chemical, hormonal, or extracellular factors. AKR inducing agents have included ethacrynic acid (9), ethoxyquin (19), 3Ј-methyl-4-dimethyl-aminoazobenzene (20), and N-methyl-N-nitrosourea (21), while spontaneous induction has been observed during the development of hereditary hepatitis leading to hepatocarcinogenesis (6). Hormonal regulation of AKR members has been observed during the estrous cycle (7), in the cell cycle of Leishmania major (22) and due to the effects of basic fibroblast growth factor on astrocytes (8). An-   Table I.

TABLE IV
Kinetic constants for CHO reductase with steroids ALR1 and ALR2 sources as indicated in Table I. drogen response elements have been found for ALR2 (23) in addition to that mentioned for MVDP. ALR2 was also regulated by extracellular osmolytes (24,25) and an osmotic response element has been found in the rabbit ALR2 gene (26). The different members of the aldo-keto reductase superfamily are clearly important in a number of cellular defense and developmental mechanisms.
RNase protection assay analysis showed that the CHO reductase was variably expressed in the organs of the Chinese hamster. The highest levels were seen in the bladder with transcript also present in the jejunum, testis, and to a lesser extent, ovary. This is in contrast to the FR-1 and MVDP genes that were expressed in testis, heart, ovary (FR-1), and adrenal (both) (27). The low level of CHO reductase transcript in the ovary is puzzling since a high basal level was clearly evident by RNase protection assay analysis of the CHO-K1 cells. This may indicate that the protein is developmentally regulated or that transcriptional regulation in the CHO-K1 cells line has been altered. These cells may thus be primed for the tremendous induction that occurred when the aldehyde ALLN was introduced.
Expressed CHO reductase was functional and displayed kinetics distinct from other members of the aldo-keto reductase family. It is interesting that despite its close sequence identity to the aldose reductases (approximately 70%) it showed an affinity for aliphatic substrates more closely comparable to the aldehyde reductases (Table I). Conversely, for aromatic substrates, its affinity was generally comparable with aldose reductase (Table II). In contrast to the work of Vander Jagt et al. (28), on human aldose reductase, which reported no substituent effects for aromatic aldehydes, CHO reductase showed a 10-fold lower K m for p-nitrobenzaldehyde versus benzaldehyde. The preference for aromatic substrates likely accounts for the ability of CHO reductase to reduce the anthracycline-based ketone daunomycin. CHO reductase showed similar kinetic constants to the previously reported daunorubicin reductase (29) (K m , 0.043 versus 0.08 mM; k cat , 0.47 versus 1.5 s Ϫ1 , respectively at pH 8.5). However, it is not daunorubicin reductase since CHO reductase, as with other aldose reductase members (30), was only slightly inhibited by 1 mM sodium barbitone. Daunorubicin reductase and other aldehyde reductases are almost completely inhibited at this concentration (31,32). Daunomycin also bound tighter to CHO reductase than it did to the ketone preferring carbonyl reductase (0.043 versus 0.13 mM) (17). However, CHO reductase was not able to reduce menadione. The observed preference for aromatic and other heterocyclic substrates could also explain why CHO reductase was able to metabolize ketone steroids, even though aldehyde containing steroids were preferred. This ability to use ketones makes this enzyme kinetically distinct from the aldehyde and aldose reductases.
The underlying structural reasons for these subtle kinetic differences between CHO reductase and the aldose reductases can be explained by analysis of the crystal structures of both these proteins. CHO reductase has not been crystallized. However, the FR-1 crystal structure can be used in this comparison because of the high degree of sequence identity to CHO reductase, particularly around the substrate site. Comparison of the ternary complexes of FR-1 (16) and human aldose reductase (33) complexed with NADP and the inhibitor zopolrestat show that there are many aromatic residues involved in the binding of zopolrestat. These include Tyr-49, Trp-80, Trp-112, Phe-123, and Tyr-310. These residues would provide the primary binding energy for aromatic substrates such as pyridine-3-aldehyde or steroids through -bond stacking interactions. This would account for the similar K m values using aromatic substrates for CHO reductase and aldose reductase. However, the substrate site is bounded by two regions of considerable sequence difference between FR-1 (and thus CHO reductase) and aldose reductase (Fig. 5). These differences result in the active site pocket in FR-1 being wider due to the more remote position of the loop containing Cys-299. In addition, more space is seen around the trifluoromethyl group of zopolrestat in FR-1. This larger substrate pocket in CHO reductase may not clamp small aliphatic substrates as firmly as does aldose reductase, resulting in the lower observed affinity. This larger substrate pocket may also allow for more movement of larger substrates such as daunomycin or the ketone containing steroids thus allowing for the correct orientation and reduction of these ketones.