The occurrence of carbonyl reduction in continuous cell lines emphasizes the essentiality of this metabolic pathway

Using the ketone compound metyrapone (MPON) as a substrate for carbonyl reduction it has been verified for the first time that various permanent cell lines in culture express carbonyl reducing activity. This is even true for the dedifferentiated and fibroblastoid cell line V79, emphasizing the essentiality of this metabolic pathway. MPON reducing enzyme activities are located in the endoplasmic reticulum as well as in the cytoplasm of the cells. Compared to MPON-reductase in rat liver microsomes, no immunological homology to microsomal C2REV7 rat liver hepatoma cell MPON-reductase could be detected, indicating differences in antigenic determinants between the enzymes of the solid organ and respective cells in continuous culture.     

Localization of pulmonary carbonyl reductase in guinea pig and mouse: enzyme histochemical and immunohistochemical studies.

We studied the localization of carbonyl reductase (E.C. 1.1.1.184) in guinea pig and mouse lung by enzyme histochemistry and immunohistochemistry, using antibodies against the guinea pig lung enzyme which crossreacted with the lung enzymes of both animals. Carbonyl reductase activity was detectable in the bronchiolar epithelial cells of small airways and in alveolar cells. In the immunohistochemical staining for carbonyl reductase, the reaction was strongest in the non-ciliated bronchiolar cells (Clara cells) and was weak in the ciliated cells and type II alveolar pneumocytes. Injection of a single dose of naphthalene led to significant impairment of carbonyl reductase activity and of microsomal mixed-function oxidase activities in mouse lung, with a marked decrease in both activity and immunoreactive staining in the bronchiolar epithelial cells. The results indicate that carbonyl reductase is localized primarily in the Clara cells, which are known to be sites of pulmonary drug metabolism.     

In Situ Association of Calvin Cycle Enzymes,Ribulose-1,5-Bisphosphate Carboxylase/Oxygenase Activase,Ferredoxin-NADP+ Reductase,and Nitrite Reductase with Thylakoid and Pyrenoid Membranes of Chlamydomonas reinhardtii Chloroplasts as Revealed by Immunoelectron Microscopy

The in situ localization of the chloroplast enzymes ribulose-1,5-bisphosphate carboxylase (Rubisco), Rubisco activase, ribose-5-phosphate isomerase, glyceraldehyde-3-phosphate dehydrogenase, aldolase, nitrite reductase, ferredoxin-NADP+ reductase, and H+-ATP synthase was studied by immunoelectron microscopy in Chlamydomonas reinhardtii. Immunogold labeling revealed that, despite Rubisco in the pyrenoid matrix, Calvin cycle enzymes, Rubisco activase, nitrite reductase, ferredoxin-NADP+ reductase, and H+-ATP synthase are associated predominantly with chloroplast thylakoid membranes and the inner surface of the pyrenoid membrane. This is in accord with previous enzyme localization studies in higher plants (K.H. Suss, C. Arkona, R. Manteuffel, K. Adler [1993] Proc Natl Acad Sci USA 90: 5514-5518). Pyrenoid tubules do not contain these enzymes. The pyrenoid matrix consists of Rubisco but is devoid of the other photosynthetic enzymes investigated. Evidence for the occurrence of two Rubisco forms differing in their spatial localization has also been obtained: Rubisco form I appears to be membrane associated like other Calvin cycle components, whereas Rubisco form II is confined to the pyrenoid matrix. It is proposed that enzyme form I represents an active Rubisco when assembled into Calvin cycle enzyme complexes, whereas Rubisco form II may be part of a CO2-concentrating mechanism. Pyrenoidal Calvin cycle complexes are thought to be highly active in CO2 fixation and important for the synthesis of starch around the pyrenoid.     

Substrate specificity of three prostaglandin dehydrogenases

Studies on the substrate specificity, kcat/Km, and effect of inhibitors on the human placental NADP-linked 15-hydroxyprostaglandin dehydrogenase (9-ketoprostaglandin reductase) indicate that it is very similar to a human brain carbonyl reductase which also possesses 9-ketoprostaglandin reductase activity. These observations led to a comparison of three apparently homogeneous 15-hydroxyprostaglandin dehydrogenases with varying amounts of 9-ketoprostaglandin reductase activity: an NAD- and an NADP-linked enzyme from human placenta and an NADP-linked enzyme from rabbit kidney. All three enzymes are carbonyl reductases for certain non-prostaglandin compounds. The placental NAD-linked enzyme, which has no 9-ketoprostaglandin reductase activity, is the most specific of the three. Although it has carbonyl reductase activity, a comparison of the Km and kcat/Km for prostaglandin and non-prostaglandin substrates of this enzyme suggests that its most likely function is as a 15-hydroxyprostaglandin dehydrogenase. The results of similar comparisons imply that the other two enzymes may function as less specific carbonyl reductases.     

The inducible trimethylamine N-oxide reductase of Escherichia coli K12: its localization and inducers

We used an anti-trimethylamine-N-oxide reductase (EC 1.6.6.9) serum and different immunological techniques (Ouchterlony, rocket immunoelectrophoresis, immunoblotting) to show that dimethylsulphoxide (DMSO), tetrahydrothiophene 1-oxide (THTO) and pyridine N-oxide (PNO) were effective inducers of the inducible form of trimethylamine N-oxide reductase. We confirmed this genetically and biochemically using a strain in which phage MudII 1734 carrying lacZ was inserted into torA, the structural gene for inducible trimethylamine-N-oxide reductase. By subcellular fractionation and quantitation with rocket immunoelectrophoresis, we showed that the enzyme was principally localized in the periplasmic fraction. Constitutive trimethylamine-N-oxide reductase was localized in the membrane fraction and, like the inducible enzyme showed a broad specificity with respect to various compounds such as DMSO, THTO and PNO. Apart from their immunological properties, the two enzymes could be clearly differentiated by their temperature stability.     

Galactosyltransferases involved in galactolipid biosynthesis are located in the outer membrane of pea chloroplast envelopes

The galactosylation steps in the biosynthesis of galactolipids involve two different enzymes; a UDP-Gal:diacylglycerol galactosyltransferase and a galactolipid:galactolipid galactosyltransferase. Previous localization studies have shown that in spinach these enzymes are located in the chloroplast envelope. Our results with peas (Pisum sativum var Laxton's Progress No. 9) confirm these results and extend the localization by providing evidence that the galactosyltransferases are in the outer membrane of the envelope. The specific activity of UDP-Gal:diacylglycerol galactosyltransferase in outer membrane preparations was 6 to 10 times greater than that exhibited by inner membrane preparations. In addition, using quantitative sodium dodecyl sulfate-polyacrylamide gel electrophoresis, it was possible to show that the UDP-Gal:diacylglycerol galactosyltransferase activity associated with inner membrane preparations could be accounted for by outer membrane contamination. It is concluded from these results that this enzyme is located predominantly, if not exclusively, in the outer membrane of the envelope. An analysis of the galactolipid products synthesized by the highly purified outer membrane showed that the galactolipid:galactolipid galactosyltransferase is also present, suggesting that this enzyme is also an outer membrane enzyme. The implication of these results is that the final assembly of galactolipids is carried out on the outer membrane of the chloroplast envelope.     

Cloning and sequence analysis of D-erythrulose reductase from chicken: its close structural relation to tetrameric carbonyl reductases

Sequence analysis of a cDNA for D-erythrulose reductase from chicken liver showed that the deduced open reading frame encodes the protein with a molecular mass of 26 kDa consisting of 246 amino acids. Although the reductase shares more than 60% identity in the amino acid sequence with the mouse tetrameric carbonyl reductase, these two enzymes have many biochemical differences; their substrate specificity, subcellular localization, organ distribution, etc. A three-dimensional structure of D-erythrulose reductase was predicted by comparative modeling based on the structure of the tetrameric carbonyl reductase (PDB entry = 1CYD). Most of the residues at the active site (within 4 A from the ligand) of the carbonyl reductase were also conserved in the D-erythrulose reductase. Nevertheless, Val190 and Leu146 in the active site of the tetrameric carbonyl reductase were substituted in the D-erythrulose reductase by Asn192 and His148, respectively. The substitutions in the active sites may be related to the difference in substrate specificity of the two enzymes. The phylogenic analysis of D-erythrulose reductase and the other related proteins suggests that the protein described as a carbonyl reductase D-erythrulose reductase.     

Homologies between enzymes involved in steroid and xenobiotic carbonyl reduction in vertebrates, invertebrates and procaryonts.

Evidence is reported for the existence of a structurally and functionally related and probably evolutionarily conserved class of membrane-bound liver carbonyl reductases/hydroxysteroid dehydrogenases involved in steroid and xenobiotic carbonyl metabolism. Carbonyl reduction was investigated in liver microsomes of 8 vertebrate species, as well as in insect larvae total homogenate and in purified 3 alpha-hydroxysteroid dehydrogenase preparations of the procaryont Pseudomonas testosteroni, using the ketone compound 2-methyl-1,2 di-(3-pyridyl)-1-propanone (metyrapone) as substrate. The enzyme activities involved in the metyrapone metabolism were screened for their sensitivity to several steroids as inhibitors. In all fractions tested, steroids of the adrostane or pregnane class strongly inhibited xenobiotic carbonyl reduction, whereas only in the insect and procaryotic species could ecdysteroids inhibit this reaction. Immunoblot analysis with antibodies against the respective microsomal mouse liver metyrapone reductase revealed strong crossrections in all fractions tested, even in those of the insect and the procaryont. A similar crossreaction pattern was achieved when the same fractions were incubated with antibodies against 3 alpha-hydroxysteroid dehydrogenase from Pseudomonas testosteroni. The mutual immunoreactivity of the antibody species against proteins from vertebrate liver microsomes, insects and procaryonts suggests the existence of structural homologies within these carbonyl reducing enzymes. This is further confirmed by limited proteolysis of purified microsomal mouse liver carbonyl reductase and subsequent analysis of the peptide fragments with antibodies specifically purified by immunoreactivity against this respective crossreactive antigen. These immunoblot experiments revealed a 22 kDa peptide fragment which was commonly recognized by all antibodies and which might represent a conserved domain of the enzyme.     

Subcellular localization and characterization of chorismate synthase in the apicomplexan Plasmodium falciparum

The resurgence of drug-resistant apicomplexa, in particular Plasmodium falciparum, the most fatal human malarial parasite, has focused attention on the recent discovery of the shikimate pathway in these organisms, as it may provide the urgently required, novel drug targets resulting from the absence of this pathway in mammals. The direction of a parasiticidal drug design programme obviously requires knowledge of the subcellular localization and indeed full characterization of the possible enzyme targets. Here, we report the cloning and characterization of chorismate synthase from P. falciparum and present the first biochemical and immunological studies of an enzyme of the shikimate pathway from an apicomplexan parasite. We show that this chorismate synthase does not possess an intrinsic flavin reductase activity and is therefore monofunctional like the plant and bacterial chorismate synthases. Highest immunological cross-reactivity was found with a plant chorismate synthase. However, in contrast to the plant enzyme, which is located to the plastid, P. falciparum chorismate synthase is found in the parasite cytosol, akin to the fungal enzymes that possess an intrinsic flavin reductase activity (i.e. are bifunctional). Thus, P. falciparum chorismate synthase has a combination of properties that distinguishes it from other described chorismate synthases.     

The distal heme center in Bacillus subtilis succinate:quinone reductase is crucial for electron transfer to menaquinone

Succinate:quinone reductases are membrane-bound enzymes that catalyze electron transfer from succinate to quinone. Some enzymes in vivo reduce ubiquinone (exergonic reaction) whereas others reduce menaquinone (endergonic reaction). The succinate:menaquinone reductases all contain two heme groups in the membrane anchor of the enzyme: a proximal heme (heme b(P)) located close to the negative side of the membrane and a distal heme (heme b(D)) located close to the positive side of the membrane. Heme b(D) is a distinctive feature of the succinate:menaquinone reductases, but the role of this heme in electron transfer to quinone has not previously been analyzed. His28 and His113 are the axial ligands to heme b(D) in Bacillus subtilis succinate:menaquinone reductase. We have individually replaced these His residues with Leu and Met, respectively, resulting in assembled membrane-bound enzymes. The H28L mutant enzyme lacks succinate:quinone reductase activity probably due to a defective quinone binding site. The H113M mutant enzyme contains heme b(D) with raised midpoint potential and is impaired in electron transfer to menaquinone. Our combined experimental data show that the heme b(D) center, into which we include a quinone binding site, is crucial for succinate:menaquinone reductase activity. The results support a model in which menaquinone is reduced on the positive side of the membrane and the transmembrane electrochemical potential provides driving force for electron transfer from succinate via heme b(P) and heme b(D) to menaquinone.     

Subcellular localization of ribonucleotide reductase in Novikoff hepatoma and regenerating rat liver

The subcellular localization of ribonucleotide reductase was ascertained in Novikoff heptoma and normal and regenerating rat tissue. Over 90% of the cellular ribonucleotide reductase is found to be associated with a membrane fraction derived from the postmicrosomal supernatant after centrifugation at 78,000g for 18 hr which bands at 1.3 m sucrose in a discontinuous sucrose gradient. The properties of this particular ribonucleotide reductase are similar to those reported for mammalian ribonucleotide reductase. This membrane fraction, which contains ribonucleotide reductase, had been previously shown to contain a DNA polymerase whose activity is related to cell proliferation. The association of these two enzymes involved in DNA synthesis leads to the suggestion that there may exist a complex of enzymes involved in deoxynucleotide and DNA synthesis in this membrane fraction.     

Localization of the deoxyribonucleotide biosynthetic enzymes ribonucleotide reductase and thymidylate synthase in mouse L cells

Two different approaches were used to define the intracellular localization in mouse L929 cells of two deoxyribonucleotide biosynthetic enzymes: ribonucleoside diphosphate reductase (EC1.17.4.1) and thymidylate synthase (EC2.1.1.45). The first involved treatment with saponins, which render the plasma membrane permeable to proteins without disrupting intracellular organelles. Under conditions where nuclear DNA synthesis and the activity of the nuclear enzyme NMN adenylyltransferase were unaffected, the entire cellular complements of a cytosolic enzyme, glucose-6-phosphate dehydrogenase, and of ribonucleotide reductase and thymidylate synthase were released at the same rate and with similar dependence on saponin concentration. The second approach involved centrifugal enucleation of cells treated with cytochalasin B (CB) and measurement of the distribution of enzyme activities in the resulting cytoplast and karyoplast fractions. Whereas most NMN adenylyltransferase activity remained with the karyoplasts, glucose-6-phosphate dehydrogenase, ribonucleotide reductase, and thymidylate synthase were almost exclusively associated with the enucleated cytoplasts. These results indicate that, under conditions where nuclear DNA synthesis is apparently unperturbed, the intracellular distribution of the deoxyribonucleotide biosynthetic enzymes studied is the same as that of glucose-6-phosphate dehydrogenase, a typical cytosol enzyme, and clearly differs from that of NMN adenylyltransferase, a nuclear enzyme.     

Subcellular distribution and properties of carbonyl reductase in guinea pig lung

On subcellular fractionation, carbonyl reductase (EC 1.1.1.184) activity in guinea pig lung was found in the mitochondrial, microsomal, and cytosolic fractions; the specific activity in the mitochondrial fraction was more than five times higher than those in the microsomal and cytosolic fractions. Further separation of the mitochondrial fraction on a sucrose gradient revealed that about half of the reductase activity is localized in mitochondria and one-third in a peroxidase-rich fraction. Although carbonyl reductase in both the mitochondrial and microsomal fractions was solubilized effectively by mixing with 1% Triton X-100 and 1 M KCl, the enzyme activity in the mitochondrial fraction was more highly enhanced by the solubilization than was that in the microsomal fraction. Carbonyl reductases were purified to homogeneity from the mitochondrial, microsomal, and cytosolic fractions. The three enzymes were almost identical in catalytic, structural, and immunological properties. Carbonyl reductase, synthesized in a rabbit reticulocyte lysate cell-free system, was apparently the same in molecular size as the subunit of the mature enzyme purified from cytosol. These results indicate that the same enzyme species is localized in the three different subcellular compartments of lung.     

Characterization of multiple forms of carbonyl reductase from chicken liver.

Three enzyme forms (CR1, CR2 and CR3) of carbonyl reductase were purified from chicken liver with using 4-benzoylpyridine as a substrate. CR1 was a dimeric enzyme composed of two identical 25-kD subunits. CR2 and CR3 were monomeric enzymes whose molecular weights were both 32 kD. CR1 exhibited 17 beta-hydroxysteroid dehydrogenase activity as well as carbonyl reductase activity in the presence of both NADP(H) and NAD(H). CR2 and CR3 had similar properties with regard to substrate specificity and inhibitor sensitivity. They could exhibit the activity only with NADPH and had no hydroxysteroid dehydrogenase activity. CR2 and CR3 cross-reacted with anti-chicken kidney carbonyl reductase antibody, though CR1 did not. The results suggest that CR1 is a hydroxysteroid dehydrogenase, and CR2 and CR3 are similar to each other and to the kidney enzymes.     

Extramitochondrial localization of NADH-fumarate reductase in trypanosomatids

Trypanosoma brucei procyclic trypomastigotes and T. cruzi epimastigotes (both Tulahuen and Y strains) were permeabilized by incubation with increasing amounts of digitonin, causing enzymes to be released from different intracellular compartments. After 10 min incubation with digitonin, the cells were centrifuged and the activity of marker enzymes (aspartate-dependent malic enzyme for cytoplasm, hexokinase for glycosomes and either isocitrate dehydrogenase or citrate synthase for mitochondria) was analyzed in the supernatant. The results were compared with the release of NADH-fumarate reductase in order to determine if this enzyme was preferentially released with a specific intracellular marker. Fumarate reductase was released at lower digitonin concentration than those required to either release isocitrate dehydrogenase or citrate synthase. Similarly, Leishmania donovani promastigotes (S-2 strain) were exposed to a single concentration of digitonin (200 micro M) but in this case we monitored the release of fumarate reductase and hexokinase, while monitoring the mitochondrial membrane potential (using safranine O). Again, substantial fumarate reductase and hexokinase activities were released without loss of mitochondrial membrane potential indicating that part of the enzyme was released while the inner mitochondrial membrane remained intact. These results suggest that, in the three species of trypanosomatids the enzyme fumarate reductase is, at least in part, located outside the mitochondrial matrix.     

Induction of a human carbonyl reductase gene located on chromosome 21

Carbonyl reductase (EC 1.1.1.184) belongs to the group of enzymes called aldo-keto reductases. It is a NADPH-dependent cytosolic protein with specificity for many carbonyl compounds including the antitumor anthracycline antibiotics, daunorubicin and doxorubicin. Human carbonyl reductase was cloned from a breast cancer cell line (MCF-7). The cDNA clone contained 1219 base paires with an open reading frame corresponding to 277 amino acids encoding a protein of Mr 30,375. Southern analysis of genomic DNA digested with several restriction enzymes and analyzed by hybridization with a labeled cDNA probe indicated that carbonyl reductase is probably coded by a single gene and does not belong to a family of structurally similar enzymes. Southern analysis of 17 mouse/human somatic cell hybrids showed that carbonyl reductase is located on chromosome 21. Carbonyl reductase mRNA could be induced 3-4-fold in 24 h with 10 microM 2,(3)-t-butyl-4-hydroxyanisole (BHA), beta-naphthoflavone or Sudan 1.     

Identification of Pig Brain Aldehyde Reductases with the High-Km Aldehyde Reductase, the Low-Km Aldehyde Reductase and Aldose Reductase, Carbonyl Reductase, and Succinic Semialdehyde Reductase

Four NADPH-dependent aldehyde reductases (ALRs) isolated from pig brain have been characterized with respect to substrate specificity, inhibition by drugs, and immunological criteria. The major enzyme, ALR1, is identical in these respects with the high-Km aldehyde reductase, glucuronate reductase, and tissue-specific, e.g., pig kidney aldehyde reductase. A second enzyme, ALR2, is identical with the low-Km aldehyde reductase and aldose reductase. The third enzyme, ALR3, is carbonyl reductase and has several features in common with prostaglandin-9-ketoreductase and xenobiotic ketoreductase. The fourth enzyme, unlike the other three which are monomeric, is a dimeric succinic semialdehyde reductase. All four of these enzymes are capable of reducing aldehydes derived from the biogenic amines. However, from a consideration of their substrate specificities and the relevant Km and Vmax values, it is likely that it is ALR2 which plays a primary role in biogenic aldehyde metabolism. Both ALR1 and ALR2 may be involved in the reduction of isocorticosteroids. Despite its capacity to reduce ketones, ALR3 is primarily an aldehyde reductase, but clues as to its physiological role in brain cannot be discerned from its substrate specificity. The capacity of succinic semialdehyde reductase to reduce succinic semialdehyde better than any other substrate shows that this reductase is aptly named and suggests that its primary role is the maintenance in brain of physiological levels of gamma-hydroxybutyrate.     

A novel NADH-dependent carbonyl reductase with unusual stereoselectivity for (R)-specific reduction from an (S)-1-phenyl-1,2-ethanediol-producing micro-organism: purification and characterization

AIMS: To purify and characterize the (R)-specific carbonyl reductase from Candida parapsilosis; to compare the enzyme with other stereospecific oxidoreductases; and to develop an available procedure producing optically active (R)-1-phenyl-1,2-ethanediol (PED). METHODS AND RESULTS: An (R)-specific carbonyl reductase was found and purified from C. parapsilosis through four steps, including blue-sepharose affinity chromatography. The relative molecular mass of the enzyme was estimated to be 35 kDa on gel-filtration chromatography and 37.5 kDa on Sodium dodecyl sulfate-polyacrylamide gel electrophoresis. The purified enzyme catalysed the reduction of various ketones, including alkyl and aromatic ketones, and was specific to short-chain and medium-chain alkyl ketones. The enzyme activity was inhibited by divalent ion of CuSO(4) and FeSO(4), whereas zincum ion stimulated its activity. For catalysing reduction, the enzyme performed maximum activity at pH 6.0 and the optimum temperature was 45 degrees C. The carbonyl reductase catalysed asymmetric reduction of beta-hydroxyacetophenone to the corresponding (R)-PED with the optical purity of 100% enantiomeric excess (e.e.). By analysing its partial amino acid sequences, the enzyme was proposed to be a novel stereospecific carbonyl reductase. CONCLUSIONS: The purified carbonyl reductase showed unusual stereospecificity and catalysed the NADH-dependent reduction of beta-hydroxyacetophenone to (R)-PED. The enzyme was different from other stereoselective oxidoreductases in catalytic properties. SIGNIFICANCE AND IMPACT OF THE STUDY: The discovery of (R)-specific oxidoreductase exhibiting unusual stereospecificity towards hydroxyl ketone is valuable for the synthesis of both enantiomers of useful chiral alcohols, and provides research basis for the achievement of profound knowledge on the relationship between structure and catalytic function of (R)-specific enzymes, which is meaningful for the alteration of stereospecificity by molecular methods to obtain the enzymes with desired stereospecificity.     

NADH- and NADPH-linked aquacobalamin reductases occur in both mitochondrial and microsomal membranes of rat liver

To elucidate the mechanism for the synthesis of the coenzyme forms of cobalamin in mammals, rat liver aquacobalamin reductase was partially characterized. Rat liver contained both NADH- and NADPH-linked aquacobalamin reductases. The NADH-linked enzyme was distributed in the mitochondria (approx. 40%) and microsomes (60%), identical to the distribution of the NADPH-linked enzyme. The two mitochondrial NADH- and NADPH-linked enzymes were located inside of the outer membrane.     

Inactivation of carbonyl reductase from human brain by phenylglyoxal and 2,3-butanedione: a comparison with aldehyde reductase and aldose reductase

Aldehyde reductase (alcohol:NADP+ oxidoreductase, EC 1.1.1.2), aldose reductase (alditol:NAD(P)+ 1-oxidoreductase, EC 1.1.1.21) and carbonyl reductase (secondary-alcohol:NADP+ oxidoreductase, EC 1.1.1.184) constitute the enzyme family of the aldo-keto reductases, a classification based on similar physicochemical properties and substrate specificities. The present study was undertaken in order to obtain information about the structural relationships between the three enzymes. Treatment of human aldehyde and carbonyl reductase with phenylglyoxal and 2,3-butanedione caused a complete and irreversible loss of enzyme activity, the rate of loss being proportional to the concentration of the dicarbonyl reagents. The inactivation of aldehyde reductase followed pseudo-first-order kinetics, whereas carbonyl reductase showed a more complex behavior, consistent with protein modification cooperativity. NADP+ partially prevented the loss of activity of both enzymes, and an even better protection of aldehyde reductase was afforded by the combination of coenzyme and substrate. Aldose reductase was partially inactivated by phenylglyoxal, but insensitive to 2,3-butanedione. The degree of inactivation with respect to the phenylglyoxal concentration showed saturation behavior. NADP+ partially protected the enzyme at low phenylglyoxal concentrations (0.5 mM), but showed no effect at high concentrations (5 mM). These findings suggest the presence of an essential arginine residue in the substrate-binding domain of aldehyde reductase and the coenzyme-binding site of carbonyl reductase. The effect of phenylglyoxal on aldose reductase may be explained by the modification of a reactive thiol or lysine rather than an arginine residue.