Rat Liver Daunorubicin Reductase AN ALDO-KETO REDUCTASE

Abstract Daunorubicin, a cancer chemotherapeutic antibiotic, is converted by rat liver extracts to daunorubicinol by daunorubicin reductase. This enzyme is purified from rat liver by ammonium sulfate fractionation, DEAE-cellulose, hydroxyapatite, and Bio-Gel P-200 column chromatography with an overall recovery of 46%. On the basis of polyacrylamide electrophoresis with and without sodium dodecyl sulfate, gel filtration chromatography, and high speed ultracentrifugation, the final preparation is judged homogeneous. Sedimentation equilibrium analysis and sodium dodecyl sulfate polyacrylamide electrophoresis give molecular weights of 38,796 ± 369 and 39,800, respectively, and suggest a single polypeptide chain. In agreement is the molecular weight of 39,541 calculated from the proposed amino acid composition. The enzyme has an isoelectric point of 6.3 and a strict requirement for NADPH. The daunorubicin-NADPH interaction occurs with a 1:1 stoichiometry. The apparent equilibrium constant at pH 8.5 for reduction is 2.91 (± 0.26) x 109 m-1, and the apparent Km for daunorubicin is 57 µm. Several common sulfhydryl reactants inhibited the enzyme. A closely related antibiotic, adriamycin, is reduced at about 5% the rate of daunorubicin reduction. In addition to the antibiotics, the enzyme reduces some aldo sugars, particularly, d-glucuronolactone, d-glyceraldehyde, and the isomers d-glucuronate and d-galacturonate. d-Glucuronolactone and daunorubicin were the preferred substrates. This reactivity with d-glucuronolactone suggests a normal role of daunorubicin reductase in ascorbic acid synthesis or in the glucuronic acid cycle, or both.


SUMMARY
Daunorubicin, a cancer chemotherapeutic antibiotic, is converted by rat liver extracts to daunorubicinol by daunorubicin reductase.
This enzyme is purified from rat liver by ammonium sulfate fractionation, DEAE-cellulose, hydroxyapatite, and Bio-Gel P-ZOO column chromatography with an over-all recovery of 46%.
On the basis of polyacrylamide electrophoresis with and without sodium dodecyl sulfate, gel filtration chromatography, and high speed ultracentrifugation, the final preparation is judged homogeneous. Sedimentation equilibrium analysis and sodium dodecyl sulfate polyacrylamide electrophoresis give molecular weights of 38,796 =t 369 and 39,800, respectively, and suggest a single polypeptide chain.
In agreement is the molecular weight of 39,541 calculated from the proposed amino acid composition.
The enzyme has an isoelectric point of 6.3 and a strict requirement for NADPH.
The apparent equilibrium constant at pH 8.5 for reduction is 2.91 ( f 0.26) x log &I-', and the apparent K,,, for daunorubicin is 57 PM. Several common sulfhydryl reactants inhibited the enzyme. A closely related antibiotic, adriamycin, is reduced at about 5 % the rate of daunorubicin reduction.
In addition to the antibiotics, the enzyme reduces some aldo sugars, particularly, D-glucuronolactone, D-glyceraldehyde, and the isomers Dglucuronate and D-galacturonate. D-Glucuronolactone and daunorubicin were the preferred substrates. This reactivity with D-glucuronolactone suggests a normal role of daunorubicin reductase in ascorbic acid synthesis or in the glucuronic acid cycle, or both.
The cancer chemotherapeutic antibiotic, daunorubicin, is metabolized by various mammalian cells and cell extracts (l-4). One step in the metabolic sequence is the conversion of daunorubicin to daunorubicinol (Fig. 1). Occurring primarily in the cellular cytosol, this enzymatic reaction is sustained by NADPH (2).
The in vivo activity of the reductase is well documented by the large amounts of the enzymatic product, daunorubicinol, that are excreted by humans and rats treated with daunorubicin (5,6). In fact, daunorubicinol is the major excretion form of daunorubicin and is a major form of the drug in tissues. One significant aspect of this metabolic conversion is that the enzymatic product, daunorubicinol, has cytotoxic properties and is an important component in the action of daunorubicin (7). Since daunorubicin reductase appears to be a constitutive and ubiquitous mammalian enzyme, even occurring in erythrocytes and platelets, it is doubtful that the natural substrate for this enzyme is an antibiotic arising from a mold.
In order to answer, ultimately, the basic questions of the enzyme's identity and its role in drug metabolism we have purified the enzyme to apparent homogeneity from rat liver and have determined a number of its physical and chemical properties.
These studies are the subject of this report. METHODS AND All operations were carried out at O-5" unless otherwise specified.
Step activity were determined on each 4.ml fraction, and tubes having a constant activity to absorbance ratio were combined and concentrated by ultrafiltration (Amicon Uhf-2 membrane) (Fig. 3).

Storage
The purified enzyme obtained from ultrafiltration in Step 5 was separated into small aliquots and stored frozen at -15".

Enzyme
Putification-The purification of daunorubicin reductase presented no insurmountable problems.
In fact, the purification was facilitated by the low retention of this protein to DEAE-cellulose and hydroxyapatite (Table I). Purity of the enzyme was established by a number of criteria. One protein band resulted from polyacrylamide electrophoresis at pH 8.4 (Fig. 4A) and pH 9.5 (Fig. 4B) and after polyacryl-   4 (left). Polyacrylamide electrophoresis of daunorubicin reductase from Bio-Gel P-200 column (A), 7.2 pg of enzyme subjected to electrophoresis at pH 8.4; (B) 6.3 pg of enzyme subjected to electrophoresis at pH 9.5; and (C) 11 pg of enzyme subjected to electrophoresis in the presence of 0.1% sodium dodecyl sulfate. Electrophoresis was from top (cathode) to boitom (anode) and the dye fronts are marked with wire. B C amide electrophoresis in the presence of sodium dodecyl sulfate (Fig. 4C). After polyacrylamide electrophoresis at pH 9.5, the coincidence of enzyme activity and protein was established by assaying unstained gel slices and comparing them to a gel run at the same time but stained with Coomassie brilliant blue (Fig. 5). The protein used for the experiment in Fig. 5 was obtained from an earlier slightly different purification procedure. Enzyme purity was further suggested by the single symmetrical protein and coincident activity peaks which eluted from Bio-Gel P-200 chromatography and the constant specific activity of the pooled fractions across the peak (Fig. 3).
Ultracentrifugation provided strong additional evidence for size homogeneity.
The plots of reciprocal number-and weightaverage molecular weights versus protein concentration gave nearly straight lines throughout the cell channels for each analysis (Fig. 8).
Physical and Chemical Characteristics-Total amino acid analysis was performed on two different purified preparations of daunorubicin reductase after acid hydrolysis at 24, 48, and 72 hours. The proposed amino acid composition is summarized in Table II. No neutral sugars (<l%) were detected when the protein was tested by the phenolsulfuric acid reaction. When the purified enzyme was electrofocused on a pH gradient, a single peak of enzyme activity occurred at pH 6.3. This activity was coincident with the single protein peak from the isoelectric column.
The absorption spectrum of the purified enzyme exhibited a maximum at 280 nm with a shoulder at 292 nm. The 280:260 nm ratio was 1.84.
The molecular weight was estimated by several methods. The results from gel filtration on calibrated columns of Bio-Gel P-150 and Sephadex G-150 are presented in Fig. 6. Rat liver daunorubicin reductase elutes as a protein of molecular weight 43,500 and 30,500, respectively.
When rat liver daunorubicin reductase was similarly chromatographed on Sephadex G-100, a molecular weight of 29,700 was obtained.
Polyacrylamide electrophoresis of rat liver daunorubicin reductase in the presence of sodium dodecyl sulfate yielded a single protein band of molecular weight 39,800 (Fig. 7).
The molecular weight of daunorubicin reductase was also determined by high speed sedimentation equilibrium. Fig. 8 is an example of a typical computer-plotted reciprocal number-, weight-and z-average molecular weights versus protein concentration (in millimeters of fringe displacement) across one cell channel.
The weight-average molecular weights from such data were obtained by linear regression analysis of the reciprocal weight-average molecular weight plot followed by extrapolation to zero protein concentration.
The molecular weight obtained from five similar analyses at two different rotor speeds and three different initial protein concentrations was 28,786 f 369. In the equilibrium experiments the reciprocal molecular weights were nearly constant across each cell channel.
The average slope of the best fit reciprocal weight-average molecular weight versus protein concentration plots was -0.0458 f 0.0086. The molecular weight calculated from the proposed amino acid composition was 39,541. A partial specific volume of 0.742 ml per g was calculated.
Effect of pH on Enzyme Activity-Optimal enzyme activity with daunorubicin as substrate was observed between pH 8.5 and 9.0. Tris-HCl buffer exhibited a slight inhibition relative to glycine and phosphate buffers (Fig.  9).  The apparent thermodynamic equilibrium constant, Keq, and stoichiometry for the reaction in Fig. 1  and 2.1 X 10" M-l, respectively. Despite the increase in K,, with increased pH, the reverse reaction was detected spectrophotometrically only above pH 9.0.' Under the standard assay condit.ions, the apparent K, of the reductase for daunorubicin was 57 PM (Table III). The apparent K, for NADPH at 1.5 mM daunorubicin was about 1 pM.
The closely related anthracycline antibiotic, adriamycin, was also reduced by the enzyme; but only at about 5% of the daunorubicin rate (Table III).

Endogenous
Substrates-In addition to daunorubicin and adriamycin, a number of naturally occurring biological compounds were tested as possible alternative substrates for daunorubicin reductase.
The reductase readily reduced n-glucuronolactone, n-glyceraldehyde, n-glucuronate, and n-galacturonate and a lower level of reactivity was displayed toward n-glucose B-sulfate, n-ribose, n-galactose, n-glucose, and n-glucose 6-phosphate (Table III).
The enzyme was tested for dehydrogenase activity with isocitric acid and malic acid as substrates under previously described assay conditions (8,9). No reaction with these substrates was detected.
Inhibitors-A number of chemicals were tested for their ability to inhibit daunorubicin reductase . (Table IV).
In addition to being sensitive to the classical sulfhydryl reagents, pchloromercuribenzoylsulfonate, mercaptoethanol, and dithiothreitol, the enzyme also was inhibited by cyanide. Stability-In concentrated protein solutions, the reductase was stable to storage at -15" or to dialysis at each purification step.
The purified enzyme was similarly stable for at least 4 months at -15" and a concentration of 3 to 5 mg per ml. However, at low protein concentrations, the enzyme was unstable either at -15" or at 4". DISCUSSION With the purification of daunorubicin reductase from rat liver cytosol, we have found this enzyme capable of cat,alyzing the reduction of several endogenous substrates. Depending on which compounds are compared, both loose and broad specificity might be inferred.
Since n-glucuronolactone, n-glyceraldehyde, n-glucuronate, and daunorubicin all function as substrates, it appears that the enzyme is a general aldo-keto reductase.
However, several other closely related aldoses, such as D-ribose, n-glucose, n-galactose, and n-glucose 6-phosphate, do not func-tion well as substrates. Also, adriamycin which differs from daunorubicin only by the replacement of a hydrogen with a hydroxyl is a poor substrate.
Thus, the enzyme displays a reactivity for substrates which have little similarity, i.e., Dglucuronolactone versus daunorubicin whereas compounds closely related to both of these active substrates are less or unreactive.
Besides n-glucuronolactone and n-glyceraldehyde, the endogenous substrates which have a high degree of reactivity are n-glucuronate and n-galacturonate, both of which have acidic moieties at the sixth position.
However, the alcohol conjugers of these uranic acids, n-glucose and n-galactose, have little reactivity.
Two other substituted acid sugars, D-glUCOSe 6sulfate and o-glucose B-phosphate, also have very low substrate activity.
It therefore appears that the aldose activity is enhanced with a carboxylic acid group at the sixth position but, as is also illustrated by n-glucuronolactone, charge alone is not enough to determine substrate activity.
The most effective endogenous substrate tested was n-glucuronolactone.
The ability of daunorubicin reductase to reduce n-glucuronolactone as well as n-glucuronate suggests that the enzyme may be involved in L-ascorbic acid synthesis and in t,he glucuronic acid cycle in vivo.  1.1.20).
A partially purified preparation catalyzing both of these reactions was previously described from rat liver and was characterized extensively with respect to substrate specificity (22). Our purified enzyme catalyzes both reactions and may normally function at the bifurcation of the L-ascorbic acid synthesis and the glucuronic acid cycle.
Purification of daunorubicin reductase activity yielded a single protein with enzymatic activity. This conclusion was supported by the results from electrophoresis, gel filtration, and sedimentation equilibrium data. Confidence that the major protein isolated represented the active enzyme was obtained from the coincidence of reductase activity with protein stain after polyacrylamide electrophoresis. Since all criteria examined indicated homogeneity, the reactivity of n-glucuronolactone, n-glyceraldehyde, and the uranic acids with daunorubicin reductase apparently represent true multiple substrate cysteines and the other two are linked via a disulfide bond. Cyanide inhibition might reflect the presence of susceptible metal(s) as part of the active enzyme structure.
If metal(s) are present, the stability of the enzyme to dialysis or treatment with chelators imply a very t.ight or restricted association between metal and enzyme.
Conjugation of cyanide with oxidized pyridine nucleotide would not normally be expected to effect the initial rates used as a measure of inhibition.
The precise mechanism of the cyanide inhibition is unknown.
When the in vivo metabolism of daunorubicin and adriamycin in rats and other mammals were studied, it was apparent that the aldo-keto reductase funct,ioned in situ. In the 8 hours after treatment, daunorubicin-treated rats excreted from 50 to 60% of their drug as daunorubicinol in urine and bile (6) whereas adriamycin-treated rats excreted about 15 to 20% of their drug as adriamycinol?
Similarly, rat tissues contained high levels of the reduced product of daunorubicin and lesser amounts of reduced adriamycin.
This degree of in viva conversion agrees with the reactivity of these substrates with the purified rat liver enzyme.
The in vivo experiments also suggest that this reversible reaction favors the reductive pathway.
The apparent thermodynamic equilibrium constant obtained from the purified enzyme in the present work verifies this conclusion.
Since daunorubicin reductase has been observed in all of the tissues of several mammalian species tested, it is likely that the characteristics of this enzyme differ both on an interspecies basis as well as on an interorgan basis. More importantly, we wonder if this broad ability of mammalian tissues to reduce these compounds may indicate active cytoplasmic systems which have a yet wider substrate receptivity for exogenous aldehydes and ketones.
We are presently investigating this possibility. consists of a single polypeptide chain.
Chromatography of the reductase on a calibrated Rio-Gel P-150 column gave a molecular weight of 43,500 in reasonable agreement with electrophoresis and ultracentrifugation data. However, gel filtration on Sephadex columns consistently yielded lower molecular weight values around 30,000.
Since the enzyme had reactivity with certain sugars, this slower elution of reductase from Sephadex relative to Bio-Gel columns probably reflected specific interactions of the enzyme with the dextran support.
Inhibition of the reductase by sulfhydryl reagents demonstrates both reactive free sulfhydryl and disulfide bonds. Since amino acid analysis indicated 4 half-cystine residues per mole of enzyme, this suggests that 2 of the half-cystines exist as free