THE JOURNAL OF BIOLOGICAL CHEMISTRY Vol. 264, No. 14, Issue of May 15, pp. 8066-8073,1989 Printed in U.S.A.

The Biosynthesis of Tetrahydrobiopterin in Rat Brain

(Received for publication, December 19, 1988)

S. Milstienz and S. Kaufman From the Laboratorv of Neurochemistrv. National Institute of Mental Health, National Institutes of Health,

I,

Bethesda, Maryland 20892

An enzyme with 6-pyruvoyl tetrahydropterin (6PPH4) (2’-oxo)reductase activity was purified to near homogeneity from whole rat brains by a rapid method involving affinity chromatography on Ciba- cron blue F3Ga-agarose followed by high performance ion exchange chromatography and high performance gel filtration. The enzyme has a single subunit of M, 37,000 and has a similar amino acid composition to previously described aldoketo reductases. The reduc- tase activity is absolutely dependent on NADPH, will only catalyze the reduction of the C-Z’-oxo group of 6PPH4, and is inactive towards the C-1‘-oxo group. However, the enzyme also shows high activity towards nonspecific substrates, such as 4-nitrobenzaldehyde, phenanthrenequinone, and menadione.

The role of this 6PPH4 reductase in the formation of tetrahydrobiopterin (BH4) was investigated. Meas- urements were made of the rate of conversion of 6PPH4, generated from dihydroneopterin triphosphate with purified 6PPH4 synthase, to BH4 in the presence of mixtures of pure sepiapterin reductase and the 6PPH4 (2’-0xo)reductase purified from rat brains. The results suggest that when sepiapterin reductase activ- ity is limiting, a large proportion of BH4 synthesis proceeds through the 6-lactoyl intermediate. However, when sepiapterin reductase is not limiting, most of the BH4 is probably formed via reduction of the other mono-reduced intermediate which is produced from 6PPH4 by sepiapterin reductase alone.

Tetrahydrobiopterin (BH,)’ is the cofactor required for the hydroxylation of the aromatic amino acids (1). It has long been known that the de nouo biosynthesis of BH4 from GTP occurs in many organs and cells (2). The initial and rate- limiting step in the pathway is the conversion of GTP to

* The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “adoertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

$ T o whom correspondence should be addressed Laboratory of Neurochemistry, National Institute of Mental Health, Bldg. 36, Rm. 3D-30, Bethesda, MD 20892.

The abbreviations used are: BH4, 6-[dihydroxypropyl-(~.-erythro)- 5,6,7&tetrahydropterin]; (6R)BH4 is the natural isomer; BHz, 6- [dihydroxypropyl-(~-erythro)-7,8-dihydropterin]; NHzTP, 6-[trihy- droxypropyl-(~-erythro)-7,8-dihydropterin-3’-triphosphate; GLPHa, 6-[2’-hydroxy-l’-oxopropyl-(~-erythro)-5,6,7,8-tetrahydropterin]; GPPH,, 6-[1’,2’-dioxopropyl-5,6,7,8-tetrahydropterin]; 1’-hydroxy- 2’-oxo-pH4, 6-[l’-hydroxy-2’-oxopropyl-(~-erythro)-5,6,7,8-tetrahy- dropterin]; HPLC, high performance liquid chromatography; SDS, sodium dodecyl sulfate; PAGE, polyacrylamide gel electrophoresis; FPLC, fast protein liquid chromatography; 2’-0xo, 2’-oxopropyl.

dihydroneopterin triphosphate (NH2TP), a reaction catalyzed by the enzyme, GTP cyclohydrolase (see Fig. 1) (3). It has now been established in a number of different systems that the next intermediate on the pathway is a tetrahydropterin, 6-pyruvoyl tetrahydropterin (6PPH4), which is formed from NHZTP by an intramolecular redox reaction coupled to elim- ination of the triphosphate group (4-6). The nature of the pathway beyond 6PPH4 has not yet been completely charac- terized, although it has been definitively demonstrated that NADPH is necessary for the formation of BH4 (7).

In 1983, we found that sepiapterin reductase, an enzyme whose only previously described function was the NADPH- dependent conversion of sepiapterin to dihydrobiopterin (8) and the NADP-dependent reverse reaction (9), was required for the formation of tetrahydrobiopterin and suggested that 6-lactoyl tetrahydropterin (6LPH4) could be the intermediate which was the substrate of sepiapterin reductase (10). This hypothesis was supported by the demonstration that not only was synthetic 6LPH4 reduced to BH, in a reaction catalyzed by sepiapterin reductase, but 6LPH4 was indeed formed from NH4TP by rat brain extracts and accumulated when sepiap- terin reductase activity was blocked by inhibitors (11). How- ever, Heintel et al. (12) and Smith (13) demonstrated that both side chain carbonyls of 6PPH4 could be reduced in uitro to form BH4 by the addition of a large excess of sepiapterin reductase alone, in the absence of any other reductase activity (12, 13). Thus, as shown in Fig. 1, there are apparently two potential routes for the formation of BH4 from 6PPH4, both utilizing sepiapterin reductase for the terminal step.

In order to better characterize the pathway for BH4 synthe- sis, we have purified an enzyme from rat brain to near homogeneity which catalyzes the NADPH-dependent reduc- tion of 6PPH4 to 6LPH4. This enzyme is similar to a family of related aldehyde reductases, yet is distinct from sepiapterin reductase which catalyzes essentially the same carbonyl re- duction reaction with the same substrate. During the prepa- ration of this report, a brief description of the purification of an apparently similar reductase from human liver has ap- peared (14).

EXPERIMENTAL PROCEDURES

Materials Glucose oxidase, catalase, phenylmethylsulfonyl fluoride, apro-

tinin, leupeptin, and pepstatin were from Sigma. HPLC-grade chem- icals were from Fisher. BH4 and sepiapterin were from B. Schircks (Jona, Switzerland). All other chemicals were of the highest grade purity. Frozen rat brains were obtained from Rockland Farms (Gil- bertsville, PA).

Analysis of Tetrahydropterins by HPLC with Coulometric Detection The HPLC apparatus consisted of a Gilson model 302 pump, a

Rheodyne injector with a 100-pl loop, an ESA coulometric detector

8066

Tetrahydrobiopterin Biosynthetic Enzymes

GTP cyclohydrolase

8067

FIG. 1. The biosynthesis of BH4 via dual reductase pathways for the reduction of 6PPH4. Sepiapterin re- ductase is 6PPH4 (1’,2’-dioxo)reductase and 6PPH4 reductase is 6PPH4 (2’-0~0) reductase.

I GPPH, synthase

I GPPH, reductase reductase

I sepiapterin reductase sepiapterin reductase

0

equipped with the high sensitivity cell set at +O.O volts, a Spectra- Physics 4270 recording integrator, and an Altex C-18 ultrasphere IP column (4.6 X 150 mm, 5 pm). The solvent was 0.1 M KHzPO4,2 mM octanesulfonic acid, 0.1 mM EDTA, 8% methanol (adjusted to pH 2.5 with H3P04). The flow rate was usually 0.7-0.9 ml/min, and the buffer was recycled and used for several weeks. The flow rate was adjusted so that (6R)BH4, (6S)BH4, and 6PPH4 eluted at approxi- mately 7, 8, and 13 min, respectively. With these conditions, 6LPH4 and the 2’-oxo isomer were eluted at approximately 10 and 9 min,

when not in use, and the detector and pump were left on continuously. respectively. A valve was used to divert the flow around the column

The detector response was calibrated with (6R)BH4. With the ESA detector set a t maximum gain, 10 pmol of (6R)BH4 gave full-scale recorder response. Prior to assays each day, as suggested by Bailey and Ayling (15), the system was depleted of oxygen by the injection of 50 pl of 1 M dithionite. The column was detached from the detector and washed for 25 min at 0.8 ml/min after the dithionite treatment.

Measurement of Total Biopterin in Rat Brain Regions by HPLC with Fluorometric Detection

ment with HC104. Aliquots (200 11) of each extract were mixed on ice Extracts, prepared as described below, were deproteinized by treat-

with 50 pl of 3 M HC104/2 M H3P04 and then centrifuged in a microcentrifuge for 2 min. The supernatants were removed and mixed with 10 mg MnOz to oxidize reduced pterins to their fully aromatic, fluorescent forms. After 10 min at room temperature, the samples were centrifuged and filtered to remove the MnOz. The HPLC appa- ratus used for the reverse phase analysis of the oxidized pterins with fluorescent detection has been described (16).

Measurement of the NADPH-dependent Reduction Of 6PPH4 to 6LPH

6PPH4 to 6LPH4 will be referred to as 6PPH4 reductase, although For convenience, the reductase which catalyzes the conversion of

sepiapterin reductase is also capable of catalyzing the reduction of 6PPH4 (see “Discussion”). Since 6PPH4 is unstable, it was usually generated in situ from NH2TP, in the presence of glucose oxidase and glucose to remove oxygen, with partially purified preparations of 6PPH4 synthase isolated from rat brain (see below). Assays contained, in a total volume of 0.1 ml (in micromoles): Tris-HC1, pH 7.4 (10); dithiothreitol (80); MgS04 (0.2); NADPH (0.02); N-acetylserotonin (0.1); glucose (10); dihydroneopterin triphosphate (0.035); catalase (20 pg); glucose oxidase (110 units); 6PPH4 synthase (50 units). The 6PPH4 concentration generated in this way is only saturating for low concentrations of 6PPH4 reductase activity. However, this assay is convenient for routine use, since 6PPH4 is eluted as a broad peak from the HPLC at the end of a run, and the presence of larger amounts of 6PPH4 in the assay requires longer times between injec- tions until the detector returns to near base line. In order to determine maximal activities, 6PPH4 was generated in a prior reaction, assayed by HPLC, and used as soon as possible at a final concentration of 25 p ~ . Reaction mixtures were usually incubated for 15 min at 37 “C, stopped by the addition of 10 pl of 30% trichloroacetic acid (w/v), and then centrifuged for 2 min at 4 “C in a microcentrifuge. The supernatant was analyzed by HPLC with coulometric detection as described above. The amounts of reductase assayed with this amount of 6PPH4 synthase activity were varied empirically so that the assays were carried out in a linear activity range.

Sepiapterin Reductase Assay

Sepiapterin reductase activity was measured essentially as de- scribed by Katoh (17). Assays contained in a total volume of 0.2 ml (micromoles): potassium phosphate, pH 6.5 (10); NADPH (0.04); sepiapterin (0.01). Reaction mixtures were prepared in microcentri- fuge tubes and transferred to microcuvettes thermostated at 37 “C. The NADPH-dependent disappearance of sepiapterin was monitored

8068 Tetrahydrobiopterin Biosynthetic Enzymes at 410 nm in a Beckman Acta MVI spectrophotometer. 1 unit of activity is equal to the formation of 1 nmol of BHz/min.

Preparation of Dihydroneopterin Triphosphate

NHZTP was prepared from GTP with the use of pure Escherichia coli GTP cyclohydrolase, purified essentially as described by Yim and Brown (18). Reaction mixtures contained in a total volume of 1.2 ml (in micromoles): Tris-HC1, pH 8.0 (60); GTP (1.2); KC1 (180); bovine serum albumin (1.5 mg); and GTP cyclohydrolase (0.4 mg). Reactions were incubated at 42 “C for 3 h. Aliquots were removed, oxidized with MnOz to convert NH,TP to neopterin triphosphate, and analyzed by HPLC with fluorescent detection essentially as described by Blau and Niedenvieser (19). Reactions were run to 75-90% completion. The bovine serum albumin and GTP cyclohydrolase were usually removed by ultrafiltration, although this was not necessary for routine assays. The NHzTP solutions were stored in aliquots a t -70 “C.

Rat Brain Dissection Rats were killed by decapitation, and brains were quickly removed

and placed on ice. Brain regions were dissected according to Glow- insky and Iversen (20) and homogenized in 3-5 volumes of 50 mM Tris-HC1, pH 7.4, 1 mM dithiothreitol, 0.1 mM EDTA, 0.25 mM phenylmethylsulfonyl fluoride.

Preparation of 6PPH4 Synthase A highly active, partially purified preparation of 6PPH4 synthase

was isolated from whole rat brains as a side fraction during the purification of 6PPH4 reductase described below. The synthase activ- ity is not retained by Cibacron blue-agarose columns. The 30-60% ammonium sulfate fraction is applied to the Cibacron blue column and the run-through is then brought to 65 “C by heating in a boiling water bath. After cooling and centrifugation to remove denatured proteins, the activity is concentrated by precipitation with ammonium sulfate (0.41 g/ml). The concentrated enzyme fraction is dissolved in 50 mM Tris-HC1, pH 7.4, 1 mM dithiothreitol, 0.1 mM EDTA, 0.25 mM phenylmethylsulfonyl fluoride and the activity adjusted to 10,000 units/ml (1 unit = 1 pmol/min). This preparation has no detectable sepiapterin reductase or 6PPH4 reductase activity and has been used for the further purification of the synthase to homogeneity (to be reported elsewhere). The 6PPH4 synthase activity was measured as described previously (11), except that glucose, glucose oxidase, and catalase, in the amounts described above for the 6PPH4 reductase assay, were added to remove oxygen, and 6PPH4 was measured by HPLC with coulometric detection as described above.

Purification of 6PPH4 (2’-Oxo)reductase 1. Extract-Unless otherwise indicated, all procedures were carried

out in the cold room, and buffers were prepared fresh daily, kept on ice, and contained 50 mM Tris-HC1, pH 7.4, 1 mM dithiothreitol, 0.1 mM EDTA, 0.25 mM phenylmethylsulfonyl fluoride, 5 pg/ml apro- tinin, 2 pg/ml pepstatin, and 2 pg/ml leupeptin. Frozen rat brains (300 g) were suspended in 900 ml of cold buffer for 30 min or until defrosted. The brains were then homogenized two times at maximum speed in a Waring blender for 60 s with a 5-min interval in between. The homogenate was then centrifuged at 15,000 X g for 60 min in a Sorvall RC-5B centrifuge at 4 “C. The supernatant was carefully decanted, the loose pellets were re-extracted by homogenizing as above in 600 ml of the same extraction buffer, and then centrifuged as above and the two extracts combined. 2. Ammonium Sulfate Precipitation-Finely powdered ammonium

sulfate (0,199 g/ml) was added over a period of 20 min to the extract while stirring in an ice bath. After stirring for an additional 30 min, the sample was centrifuged for 25 min at 15,000 X g. The pellets, which contained most of the GTP cyclohydrolase activity, were frozen at -70 “C and used for the purification of this enzyme (to be reported elsewhere). Ammonium sulfate (0.156 g/ml) was added to the super- natant as above. After centrifugation, pellets were dissolved in %o the original extract volume in extract buffer.

3. Cibacron Blue F3GA-Agarose Column-Immobilized Cibacron blue F3GA (50-ml bed, Pierce Chemical Co.) was washed with 100 ml of 10% sodium dodecyl sulfate (SDS) (w/v), 1000 ml of distilled water, and then equilibrated with 500 ml of extract buffer. All of the ammonium sulfate fraction was applied to the column at a flow rate of ”1 ml/min and washed with the same buffer until A280 was almost 0. The first 250 ml through the column contained all of the 6PPH4 synthase activity which was further purified as described above. The

column was then washed with 500 ml of phosphate-buffered saline (10 mM potassium phosphate, pH 7.4, 137 mM NaC1, 27 mM KC1) and the reductases eluted with 250 ml of phosphate-buffered saline containing 0.2 mg/ml of NADPH. It was not possible to separate 6PPH4 (2’-oxo)reductase from sepiapterin reductase by elution with NADPH gradients. The NADPH eluate was concentrated to -5 ml by ultrafiltration and stored at -70 “C.

4. Anion Exchange Chromatography-Preliminary experiments demonstrated that the 6PPH4 reductase activity could be completely separated from sepiapterin reductase activity on several different anion exchange media, including DEAE-Sephacel (Pharmacia LKB Biotechnology Inc.), Accel Q (Waters), and Mono Q FPLC (Phar- macia LKB Biotechnology Inc.). Fast flow Q-Sepharose (Pharmacia LKB Biotechnology Inc.) was chosen for preparative purifications since separations could be scaled up directly from small scale Mono Q FPLC separations which could be run rapidly. All protein chro- matography separations were run on a Varian model 5000 liquid chromatograph equipped with a Valco injector and large loop. Detec- tion was by absorbance at either 280 or 210 nm with a Gilson model 116 UV detector. Peak areas were determined with a Hewlett-Packard 3392A integrator. Chromatography was usually done at room tem- perature. Fractions were collected at room temperature, removed from the fraction collector, and placed on ice as soon as possible. A 1 X 20-cm column of fast flow Q-Sepharose was equilibrated with 20 mM Tris-HCI, pH 7.4, containing 1 mM dithiothreitol, 0.1 mM EDTA, and 0.25 mM phenylmethylsulfonyl fluoride. The NADPH eluate from the previous step was first desalted on Sephadex G-25 columns (Pharmacia LKB Biotechnology Inc.) equilibrated in starting buffer and then injected at a flow rate of 1 ml/min. The column was washed until the A280 was =O. Then a linear gradient to 0.4 M NaCl in starting buffer was run at a flow rate of 2 ml/min over 60 min, collecting 1- min fractions. Fractions were assayed for sepiapterin reductase and 6PPH4 reductase activity, activity peaks pooled, and concentrated by ultrafiltration. As indicated in Fig. 2, there is a complete separation between 6PPH4 (2’-oxo)reductase activity and sepiapterin reductase activity.

5. Mono Q FPLC-The concentrated Q-Sepharose pools of the two reductases were desalted as above and injected in separate runs onto an FPLC Mono Q HR5/5 column in 20 mM Tris-HC1, pH 7.4, containing 1 mM dithiothreitol, 0.1 mM EDTA, and 0.25 mM phenyl- methylsulfonyl fluoride at a flow rate of 0.5 ml/min. The column was washed for 10 min and then eluted with a linear gradient to 0.2 M NaCl in the same buffer over 60 min at a flow rate of 1 ml/min. 1- min fractions were collected, placed on ice, and assayed as described under “Experimental Procedures.” The 6PPH4 reductase activity was eluted at 20 min or at about 0.06 M NaCl (Fig. 3). Sepiapterin reductase was eluted at 30 min or 0.1 M NaCI. The peak fractions of each reductase were pooled and concentrated with a centrifugal concentrator (Centricon 30, Amicon) to =250 pl.

6. High Performance Gel Filtration-To optimize separations as well as to increase loading capacity, two high performance gel filtra- tion columns with different properties were connected in series. Samples were injected with a 250-rl loop onto a Zorbax GF-250 column (4.6 X 300 mm, Du Pont) equipped with a Zorbax guard column. This was connected to a Superose 12 column (Pharmacia LKB Biotechnology Inc.). The buffer used was 0.1 M NaH2PO4, pH 7, at a flow rate of 0.5 ml/min. The Mono Q pool of 6PPH4 reductase was essentially pure after this step (Fig. 4). The native molecular mass as determined by high performance gel filtration was 37 kDa. SDS-polyacrylamide gel electrophoresis (SDS-PAGE) revealed a sin-

A ”O

MINUTES

FIG. 2. Separation of 6PPH4 (2’-oxo)reductase and sepiap- terin reductase by fast flow Q-Sepharose. The NADPH eluate from the Cibacron blue F3GA-agarose column was concentrated, buffer exchanged on Sephadex G25-M, and chromatographed as described under “Experimental Procedures.” 6PPH4 (2’- oxo)reductase and sepiapterin reductase activity peaks are indicated by the arrowheads and were at 22 and 36 min, respectively.

Tetrahydrobiopterin Biosynthetic Enzymes 8069

B I

MINUTES MINUTES

FIG. 3. Purification of 6PPH4 (2’-oxo)reductase ( A ) and sepiapterin reductase ( B ) by Mono Q FPLC. Pooled activity peaks from the previous fast flow Q-Sepharose step were concen- trated, desalted, injected onto a Mono Q H5/5 column, and eluted with the linear NaCl gradients (0-0.2 M in 60 min) as described under “Experimental Procedures.” The arrowheads indicate the peaks of enzyme activity.

30 10 20 30 MINUTES

FIG. 4. Purification of 6PPH4 (2’-oxo)reductase ( A ) and sepiapterin reductase ( B ) by high performance gel filtration. Peak fractions from the Mono Q FPLC step were concentrated and injected into the tandem high performance gel filtration columns as described under “Experimental Procedures.” The arrowheads indicate the peaks of enzyme activity.

-21 -DF

1 2 3 4 Distance (cm)

FIG. 5. SDS-PAGE of 6PPH4 (2‘-oxo)reductase. Peak frac- tions from the high performance gel filtration step were concentrated. 10 pg was electrophoresed on a 12% acrylamide gel (42) and stained with Coomassie Blue (lune A ) . The standards used ( l u n e B ) were phosphorylase b (97.4 kDa), bovine serum albumin (66.2 kDa), oval- bumin (42.7 kDa), carbonic anhydrase (31 kDa), and soybean trypsin inhibitor (21.5 kDa). Lysozyme (14.4 kDa) migrated with the dye front (DF).

gle subunit with a molecular mass of 36.5 kDa (Fig. 5B) . There was a trace amount of an M, 33,000 contaminant.

Sepiapterin reductase was very unstable after the mono Q step, and even though there was a single major protein peak which corre- sponded with the sepiapterin reductase activity after high perform- ance gel filtration, SDS-PAGE analysis showed multiple bands (not shown). On the basis of specific activity, this preparation was only about 3% pure, compared to the specific activity of homogeneous erythrocyte sepiapterin reductase. Further experiments are necessary to determine if brain sepiapterin reductase is identical to sepiapterin reductase in other tissues.

Amino Acid Analysis Samples of 6PPH4 reductase (10 pg) were hydrolyzed in 6 N HCl

in a Waters work station and the amino acids analyzed as their phenylthiohydantoin derivatives on a Waters Picotag HPLC System. Tryptophan was determined spectrophotometrically in 6 M guanidi- nium hydrochloride as described by Edelhoch (21).

RESULTS

Distribution of 6PPH4 (2’-0xo)reductase and Sepiapterin Reductase Activities in the Rat Brain-BH4 and GTP cyclo- hydrolase are widely distributed in the rat brain with the highest concentrations in the striatum and midbrain, areas with the greatest concentrations of adrenergic and serotoner- gic neurons. Our previous studies of the biosynthesis of BH4 in rat brain were carried out on extracts prepared from striatum (ll), the region of the brain with the greatest con- centration of BH4. Before proceeding with large scale purifi- cations of the BH4 biosynthetic enzymes, it was necessary to determine the distributions of these enzymes in the brain and also whether the distributions paralleled those of BH4 and GTP cyclohydrolase in order to determine if it was necessary or advantageous to use specific brain regions. Table I shows the distribution of 6PPH4 reductase activity and BH4 content in seven major regions of the rat brain. As can be seen, there is not a good correlation between 6PPH4 reductase activity and BH4 content, and there is also less than a %fold range in reductase specific activity over the entire rat brain. Interest- ingly, there are also no major differences in sepiapterin re- ductase activities in these same brain regions (data not shown). Since the activities of both 6PPH4 reductase and sepiapterin reductase show the same noncorrelation with BH4 content, and both reductases have nearly constant specific activities throughout the brain (22), whole rat brains were used as the starting material for the purification of 6PPH4 reductase which is described in detail under “Experimental Procedures.”

Purification of 6PPH4 Reductase Actiuities-As shown in Table 11, the enzyme catalyzing the reduction of 6PPH4 to 6LPH4 was purified approximately 2000-fold to near-homo- geneity from whole rat brains with a good recovery. This can probably be attributed to the great separation power and rapidity of the HPLC methodology that was used extensively in this purification.

TABLE I Distribution of 6PPH4 (2’-oro)reductose activity

and BH4 in rat brain GPPH, (Z’-oxo)reductase BHI

Striatum Cerebellum Medulla Hypothalamus Hippocampus Cortex Midbrain Whole brain

pmollminlg 430 470 580 410 330 270 520 440

nmollg 0.8 0.2 0.4 0.8 0.3 0.4 0.6 0.4

8070 Tetrahydrobiopterin Biosynthetic Enzymes

TABLE I1 Summary of the purification of rat brain GPPH, (2‘-oxo)reductase

Step Volume Protein Activity Specific activity

rnl rng units unitslmg 1. Extract 1,285 5,783 128,000 22 2. Ammonium sul- 128 2,394 110,000 46

3. Cibacron blue 9 36 96,000 2,670 4. Q-Sepharose 3 4.7 77,000 16,400

6. HPGF” 5. Mono Q 0.25 2.6 75,000 28,800

2.0 1.5 74,000 49,300

fate

HPGF, high performance gel filtration.

Sepiapterin reductase was not homogeneous after the high performance gel filtration step and was unstable. Further studies are necessary to better characterize this enzyme from brain and to determine the reasons for the instability. How- ever, analysis by SDS-PAGE and immunoblotting with anti- sepiapterin reductase serum indicated that the enzyme in rat brain is identical with the well characterized reductase present in rat erythrocytes (23) and with the major form of sepiapterin reductase present in rat liver.’ The sepiapterin reductase fraction did not catalyze the formation of any detectable 6LPH4 from 6PPH4.

Role of 6PPH4 (2‘-Oxo)reductase in BH4 Biosynthesis-The conversion of 6PPH4 to BH4 requires the stereospecific re- duction of the two side chain carbonyl groups of 6PPH4 to alcohols of the opposite configuration to those present in NHZTP (Fig. 1). It has recently been discovered that sepiap- terin reductase has dicarbonyl reductase activity (24) and can catalyze the NADPH-dependent reduction of both side chain carbonyls of 6PPH4, proceeding with reduction of the C-1’ carbonyl first (13). However, when NADPH and a sepiapterin reductase inhibitor are added to crude extracts from rat brain or from bovine adrenal medulla, another tetrahydropterin, 6LPH4, is produced from 6PPH4 (10). Furthermore, of the two isomers resulting from reduction of only one carbonyl group of 6PPH4 (6LPH4 and 1’-hydroxy-2’-oxo-pH4), 6LPH4 is the preferred substrate for reduction by sepiapterin reduc- tase (25). In order to distinguish between the two possible reduction routes from 6PPH4 to BH4 (Fig. l), an in vitro experimental system was designed which utilized the purified reductases to convert 6PPH4 to BH4. As shown in Fig. 6, when pure sepiapterin reductase (from rat erythrocytes) was used to reduce 6PPH4, both 1’-hydroxy-2’-oxo-pH4 and BH4 were formed, with a quantitative conversion to BH4 requiring the addition of a large amount of sepiapterin reductase. How- ever, in the presence of a constant amount of added 6PPH4 (2’-oxo)reductase and at low concentrations of sepiapterin reductase, there was a significant increase in the yield of BH4 with no change in the amount of 1‘-hydroxy-2’-oxo-PH4 produced. It should be noted that at the lower concentrations of sepiapterin reductase used in this experiment, the ratio of GPPH, (2’-oxo)reductase activity to sepiapterin reductase activity was similar to that found in crude brain extracts. When the ratio of 6PPH4 (2’-oxo)reductase activity to se- piapterin reductase activity was increased, there was a de- crease in the amount of 2‘-0xo compound formed due to competition for the diketo intermediate (data not shown). Thus, the reduction of 6PPH4 to BH4 by the mixture of carbonyl reductases appears to proceed by both possible routes shown in Fig. 1 simultaneously, with a greater fraction going through the 6PPH4 (2’-oxo)reductase route when sepia-

* R. A. Levine, G. K. Kapatos, s. Kaufman, and s. Milstien, manuscript in preparation.

*O r

1 1 I 1 10 20

SEPIAPTERIN REDUCTASE (mu) FIG. 6. The effect of 6PPH4 (2’-oxo)reductase on the se-

piapterin reductase-dependent conversion of NHzTP to BH4. All assays contained 60 units of 6PPH4 synthase activity. 6PPH4 (2’- 0xo)reductase (1.5 units) was added where indicated with 5, 10, and 20 milliunits of sepiapterin reductase activity. Tetrahydropterins produced from NHzTP (45 PM) were measured after a 15-min incu- bation by HPLC with electrochemical detection as described under “Experimental Procedures.” BH4 was produced in the absence (0) or in the presence (0) of GPPH, (2’-oxo)reductase; 6LPH4 was produced in the presence of sepiapterin reductase (m); and l’-hydroxy-2’-oxo- PH, was produced in the absence (A) or in the presence (A) of 6PPH4 (2’-oxo)reductase. For clarity, lines were omitted from the figure for the data points that indicate the formation of the 2’-oxo-PH, isomer, since its formation is independent of 6 P P R (2’-oxo)reductase (see Fig. 1). The amount of 6PPH4 remaining in the reaction mixtures after addition of the two reductases is also not indicated.

TABLE I11 Effect of some aldehyde reductase inhibitors and substrates on 6PPH4

(2’-ono)reductase activity A. Inhibitors Concentration 7% inhibition

Salicylate Valproate Phenobarbital LiCl

mM 10 0 1 0 1 0

10 0 B. Substrates Concentration Relative activity” K,

P M P M

6PPHI 1.5 1 1.0 BH, 5 0 Sepiapterin 10 0 4-Nitrobenzaldehyde 200 6 100 Menadione 250 0.4 Phenanthreneouinone 100 2

“GPPH, (2’-oxo)reductase activity was measured as described un- der “Experimental Procedures.” The activity with 6PPH4 was 28 nmol/min/mg. Activity with other substrates was measured spectro- photometrically by following the disappearance of NADPH at 340 nm. Products of the reactions with 4-nitrobenzaldehyde, menadione, and phenanthrenequinone have not been identified.

pterin reductase activity is limiting. The function of this apparently redundant system which utilizes 6PPH4 (2’- oxo)reductase may be to convert any 6PPH4 that is formed in uiuo to the more stable 6LPH4 form where the conversion to BH4 might then be regulated by the activity of sepiapterin reductase.

Properties of 6PPH4 Reductase-No activity could be de- tected with sepiapterin as a substrate for 6PPH4 (2’- oxo)reductase. Sepiapterin at concentrations as high as 1 mM also did not inhibit the reduction of 6PPH4. BH4, up to 50 PM, did not inhibit the reductase, suggesting that there is probably no feedback control of BH4 biosynthesis at this step. The reductase also had no detectable activity with the 2’-0xo- tetrahydropterin isomer. Thus, no BH4 can be produced from 6PPH4 in the absence of sepiapterin reductase activity. The K,,, for 6PPH4 is approximately 2 PM.* The reductase is

Tetrahydrobiopterin Biosynthetic Enzymes 8071

TABLE IV Amino acid composition of GPPH4 (2‘-oxo)reductase, sepiapterin

reductase, and dihydropteridine reductase Residues per Subunit

Amino acid GPPH, Sepiapterin Dihydropteridine reductase“ reductaseb reductase‘

ASP 34 20 19

G ~ Y 17 21 26

Glu 39 23 20 Ser 15 19 20

His 9 1 5 Thr 21 10 16 Ala 26 16 29 Arg 13 13 8 Pro 28 7 9 TY r 15 4 3 Val 28 12 18 Met 8 4 8 CY s 6 4 8 Ile 19 4 9 Leu 41 31 23 Phe 15 5 6 LY s 28 6 16 TW 4 11 4

This work, per 37-kDa subunit. Rat erythrocytes, per 28-kDa subunit (23). Rat liver, per 25.5-kDa subunit (36).

TABLE V Amino acid composition comparisons of some reductases

Reductase Source Reference Subunit Difference index” M, x 10-3

6PPH4 reductase Rat brain This work 37 Sepiapterin reduc- Rat red blood 23 28 21.6

Aldehyde reduc- Pig kidney 37 33 9.0

Diacetyl reductase Hamster liver 38 23.5 23.0 Ketone reductase Chicken kidney 39 29.5 12.0 Aldose reductase Rat lens 40 38 18.8 Dihydropteridine Rat liver 36 25.5 18.5

Glucose-6-phos- Mouse liver 41 62 1.5

tase cells

tase

reductase

phate dehydro- genase

Relative to GPPH, (2’-oxo)reductase composition (Table IV). Calculated according to Metzger et al. (35).

absolutely specific for NADPH. NADH, NADP, and NAD were neither inhibitors nor substrates. No NADP-dependent formation of 6PPH4 from 6LPH4 at pH 7.4, 8.0, or 8.5 could be detected indicating that unlike sepiapterin reductase, 6PPH4 (2’-oxo)reductase does not catalyze the reverse reac- tion to any significant extent. It should be noted that sepia- pterin reductase is able to catalyze only the single step reversal reaction from BH4 to the 2’-oxo-tetrahydropterin isomer (13) and does not catalyze the formation of any detectable 6PPH4 from the 2‘-0xo isomer (data not shown).

This 6PPH4 reductase is, like sepiapterin reductase (24), a nonspecific carbonyl reductase. However, it has neither activ- ity with, nor is inhibited by, sepiapterin or BHz. The activity towards a number of substrates is listed in Table 111. It catalyzes the reduction of a wide variety of carbonyl com- pounds, with highet activity towards p-nitrobenzaldehyde. Unlike other keto reductases (26), 6PPH4 (2’-oxo)reductase is not inhibited by sodium salicylate, sodium valproate, phe- nobarbital, or LiCl.

Molecular Mass and Amino Acid Composition of GPPH, (2’- 0xo)reductase-Unlike sepiapterin reductase, which is a di-

mer of 27.5-kDa subunits (23), 6PPH4 (2’-0xo)reductase is apparently a monomer of 37 kDa, since the subunit molecular mass determined by SDS-PAGE and the native molecular mass, determined by high performance gel filtration (Fig. 4), are both 37 kDa.

The amino acid composition is given in Table IV. As can be seen, the composition is very different from that of sepia- pterin reductase (23). Especially notable is the large number of proline residues that are present in 6PPH4 (2’- 0xo)reductase. This is a characteristic shared by a number of other aldehyde reductases prepared from a variety of sources (see “Discussion”).

DISCUSSION

The de nouo biosynthesis of tetrahydrobiopterin has long been known to proceed from GTP, following the demonstra- tion that eight of the nine carbon atoms of GTP were incor- porated into biopterin (27). One carbon atom is lost in the first step of the pathway, the conversion to NHzTP with the elimination of C-8 as formate, catalyzed by GTP cyclohydro- lase (27). Since Kaufman (28) originally isolated biopterin from rat liver in the 7,8-dihydro form and demonstrated that it could be converted into the active tetrahydro cofactor form for the phenylalanine hydroxylase reaction by reduction with dihydrofolate reductase, a logical pathway for BH4 biosyn- thesis evolved which proeceeded from NHzTP to BH2 via dihydro intermediates (29). The final step of this proposed pathway required the reduction of BH2 to BH4 by dihydrofo- late reductase. A key intermediate in this pathway was se- piapterin, 6-lactoyl-7,8-dihydropterin. Several groups dem- onstrated that addition of sepiapterin to crude biopterin syn- thesis systems inhibited [14C]BH2 synthesis and trapped [“C] sepiapterin when starting with [I4C]GTP (29, 30). These results appeared to provide strong support for the dihydro pathway in the absence of further information on the struc- tures of the intermediates.

The unexpected finding by Duch et al. (31) that dihydro- folate reductase was not necessary for the biosynthesis of BH4 suggested that the original dihydro pathway leading to BH2 could not be correct, since there was no other known mecha- nism for the formation of BH4 from BHZ (32). However, any other possible pathway still had to account for the observa- tions that addition of sepiapterin inhibits in vitro BH4 syn- thesis and that it apparently traps a labeled sepiapterin inter- mediate. In 1983, we (10) suggested an alternative pathway which utilized tetrahydro intermediates and which seemed to explain all of the known facts. A key intermediate in this proposed pathway was 6LPH4. We showed that sepiapterin reductase was essential for BH4 synthesis and that 6LPH4 could readily be reduced to BH4 by sepiapterin reductase, an enzyme which had previously been thought to only catalyze the reversible interconversion of the dihydro pterins, sepiap- terin and dihydrobiopterin. It now seems reasonable that competitive inhibition of sepiapterin reductase by added se- piapterin would lead to accumulation of 6LPH4. Then, se- piapterin would be formed spontaneously by oxidation of 6LPH4 during the isolation procedures and thus explain the results of the trapping experiments, since [14C]6LPH4 would be converted to [14C]sepiapterin. The formation of the puta- tive intermediate, 6LPH4, required the postulation of a new keto reductase which could catalyze the specific reduction of the C-2‘ carbonyl of 6PPH4. Subsequently, it was demon- strated that there was a distinct NADPH-dependent reduc- tase present in rat brain and liver as well as in bovine adrenal medulla, which could catalyze the reduction of GPPH, to 6LPH4 and, furthermore, that the kinetics of the formation

8072 Tetrahydrobiopterin Biosynthetic Enzymes

and disappearance of GLPH, were consistent with the pro- posal that it could be a precursor of BH4 (11).

Examination of the tetrahydropterins produced from NHzTP in adrenal medulla extracts (6) and in rat brain extracts (11) demonstrated the formation of another tetra- hydropterin which has since been identified as 1’-hydroxy- 2‘-oxo-tetrahydropterin (33), the product of the reduction of the 1’-carbonyl group of 6PPH4. The formation of this com- pound is catalyzed by sepiapterin reductase (33), which can also catalyze the subsequent reduction of its 2”carbonyl group to form BH4. Sepiapterin reductase has a more favorable K,,, and a higher V,,, with GLPH, than with the other isomer (25), and it is possible that the formation of the other isomer could be a result of the in uitro assays where high concentra- tions of GPPH, allow sepiapterin reductase to catalyze the NADPH-dependent reduction of some of the 6PPH4 to the other isomer. However, even in experiments where there is a greater concentration of the 2’-0xo isomer than 6LPH4, as shown in Fig. 6, some GLPH, appears to be converted to BH4 by sepiapterin reductase.

GPPH, reductase activity is widely distributed in various rat tissues and organs. Both the relative amounts and specific activities of GPPH, reductase and sepiapterin reductase in brain, liver, kidney, spleen, and uterus are nearly constant (data not shown). GPPH, reductase activity has also been found in human liver (15), brain: and bovine adrenal medulla (33).

There are at least three distinct carbonyl reductase activi- ties in rat brain: sepiapterin reductase (17); carbonyl reduc- tase (34); and 6PPH4 (2’-oxo)reductase (11). GPPH, (2’- 0xo)reductase may belong to the group of aldoketo reductases which are characterized as having a single subunit of 35-40 kDa, high specificity for NADPH, high activity with 4-nitro- benzaldehyde, and a high proline content (34). In the absence of protein sequence data, it is not possible to determine definitively if 6PPH4 (Y-oxo)reductase is related to or iden- tical with previously described reductases. A useful compari- son can be made based on amino acid composition alone with the use of a composition difference index. In Table V are shown comparisons of amino acid compositions for a number of reductases based on the method of Metzger et al. (35). Difference index values of less than 10 are considered to indicate a high degree of relatedness. As can be seen, 6PPH4 (2’-oxo)reductase is clearly not related to sepiapterin reduc- tase or other known keto reductases by this criterion. How- ever, the difference index of 1.5 suggests that there is a striking degree of homology between glucose-6-phosphate de- hydrogenase from mouse liver and 6PPH4 (2’-oxo)reductase. The high molecular mass of glucose-6-phosphate dehydrogen- ase (120 kDa) might indicate that it could arise by gene duplication of the 6PPH4 (2’-oxo)reductase gene or that 6PPH4 (2’-oxo)reductase could be derived from glucose-6- phosphate dehydrogenase. However, preliminary experiments with purified glucose-6-phosphate dehydrogenase from bovine liver and yeast indicate that these enzymes have no detectable 6PPH4 reductase activity. In contrast to aldehyde reductase (38) and aldose reductase (40), 6PPH4 (2’-oxo)reductase ac- tivity is not inhibited by valproate, phenobarbital, or salicy- late, Thus, GPPH, (2’-oxo)reductase appears to be a unique type of keto reductase.

Antibodies to GPPH, (2’-oxo)reductase have been raised in rabbits and are being used to explore further the role of this 6PPH4 reductase in BH, biosynthesis. Preliminary experi- ments with crude rat brain extracts treated with this antibody indicate that a substantial fraction of BH4 produced from

,’S. Milstien, unpublished observations.

NHzTP in vitro arises via the 6PPH4 (2’-oxo)reductase route. However, since both sepiapterin reductase and 6PPH4 (2’- oxo)reductase are ubiquitous in the rat brain, it is not possible at the present time to definitively determine the in vivo route of reduction of 6PPH4 to BH4.

Acknowledgment-We wish to thank Dr. Peter S. Backlund, Jr. for performing amino acid analyses.

REFERENCES

1. Kaufman, S. (1974) in Amino Acids in the Brain (Wolstenholme, G. E. W., and Fitzsimons, D. W., eds) pp. 85-115, Elsevier Scientific Publishing Co., Inc., New York

2. Kaufman, S. (1967) Annu. Reo. Biochem. 36,171-185 3. Rembold, H., and Gyure, L. (1972) Angew. Chem. 22,1088-1099 4. Switchenko, A. C., Primus, J. P., and Brown, G. M. (1984)

Biochem. Biophys. Res. Commun. 120,754-760 5. Heintel, D., Ghisla, S., Curtius, H.-Ch., Niederwieser, A., and

Levine, R. A. (1984) Neurochem. Int. 1, 145-151 6. Smith, G. K., and Nichol, C. A. (1984) Biochem. Bwphys. Res.

Commun. 120,761-766 7. Curtius, H.-Ch., Heintel, D., Ghisla, S., Kuster, T., Leimbacher,

W., and Niederwieser, A. (1985) Eur. J. Biochem. 148, 413- 419

8. Matsubara, M., Katoh, S., Akino, M., and Kaufman, S. (1966) Biochim. Bwphys. Acta 122 , 202-212

9. Kaufman, S. (1967) J. Biol. Chem. 242,3934-3943 10. Milstien, S., and Kaufman, S. (1983) Biochem. Biophys. Res.

Commun. 115,888-893 11. Milstien, S., and Kaufman, S. (1985) Biochem. Biophys. Res.

Commun. 128,1099-1107 12, Heintel, D., Leimbacher, W., Redweik, U., Zagalak, B., and

Curtius, H.-Ch. (1985) Biochem. Biophys. Res. Commun. 127, 213-219

13. Smith, G. K. (1987) Arch. Biochem. Biophys. 255 , 254-266 14. Curtius, H.-Ch., Steinerstauch, P., Leimbacher, W., Redweik, U.,

Takikawa, S. I., and Ghisla, S. (1987) in Unconjugated Pterins and Related Biogenic Amines (Curtius, H. Ch., Blau, N., and Levine, R. A., eds.) pp. 89-98, W. de Gruyter, Berlin

15. Bailey, S. W., and Ayling, J. E. (1978) J. Biol. Chem. 253, 1598- 1605

16. Milstien, S. (1983) in Biochemical and Clinical Aspects of Pteri- dines (Curtius, H.-Ch., Pfleiderer, W., and Wachter, H., eds) Vol. 2, pp. 65-70, W. de Gruyter, Berlin

17. Katoh, S. (1971) Arch. Biochem. Biophys. 146,202-214 18. Yim, J. J., and Brown, G. M. (1976) J. Biol. Chem. 251, 5087-

19. Blau, N., and Niederwieser, A. (1983) Anal. Biochem. 128 , 446-

20. Glowinsky, J., and Iversen, L. L. (1966) J. Neurochem. 13 , 655-

21. Edelhoch, H. (1967) Biochemistry 6,1948-1954 22. Milstien, S. (1987) in Unconjugated Pterins and Related Biogenic

Amines (Curtius, H. Ch., Blau, N., and Levine, R. A., eds) pp. 49-65, W. de Gruyter, Berlin

23. Sueoka, T., and Katoh S. (1982) Biochim. Biophys. Acta 7 1 7 ,

24. Katoh, S., and Sueoka, T. (1984) Biochem. Biophys. Res. Com- mun. 118,859-866

25. Milstien, S., and Kaufman, S. (1986) in Chemistry and Biology of Pteridines, 1986 (Cooper, B. A., and Whitehead V. M., eds) pp. 169-181, W. de Gruyter, Berlin

26. Wermuth, B. (1985) in Enzymology of Carbonyl Metabolism 2: Aldehyde Dehydrogenose (Flynn, G. F., and Weiner, H., eds) pp. 209-230, Alan R. Liss, New York

27. Foor, F. W., and Brown, G. M. (1975) J. Biol. Chem. 250,3545- 3551

28. Kaufman, S. (1963) Proc. Natl. Acad. Sci. U. S. A. 50,1085-1093 29. Tanaka, K., Akino, M., Hagi, Y., Doi, M., and Shiota, T. (1980)

30. Kapatos, G., Katoh, S., and Kaufman, S. (1982) J . Neurochem.

31. Duch, D. S., Lee, C. L., Edelstein, M. P., and Nichol, C. A. (1983) Mol. Pharmaeol. 24,103-108

32. Milstien, S., and Kaufman, S. (1983) in Biochemical and Clinical Aspects of Pteridines (Curtius, H.-Ch., Pfleiderer, W., and

5094

452

669

49-65

J. Biol. Chem. 256,2963-2972

39,1152-1162

33.

34.

35.

36.

Tetrahydrobiopterin Biosynthetic Enzymes 8073

Wachter, H., eds) Vol. 2, pp. 133-137, W. de Gruyter, Berlin Smith, G. K., and Nichol, C. A. (1986) J. Biol. Chem. 261,2725-

2737 Sawada, H., Hara, A., Nakayama, T., Usui, S., and Hayashibara,

M. (1982) in Enzymology of Carbonyl Metabolism: Aldehyde Dehydrogenase and Aldo/Keto Reductase (Weiner, H., and Wer- muth, B., eds) pp. 275-289, Alan R. Liss, New York

Metzger, H., Shapiro, M., Mosimann, J. E., and Vinto, J. E. (1968) Nature 219, 1166-1168

Webber, S., and Whiteley, J. M. (1981) Arch. Biochem. Biophys. 206,145-152

37.

38.

39.

40.

41. 42.

Flynn, T. G., Shires, J., and Walton, D. J. (1975) J. Siol. Chem.

Sawada, H., Hara, A., Nakayama, T., and Seiriki, K. (1985) J. Biochem. (Tokyo) 98, 1349-1357

Hara, A., Deyashiki, Y., Nakagawa, M., Nakayama, T., and Sawada, H. (1982) J . Biochem. (Tokyo) 92, 1753-1762

Shiono, T., Sato, S., Reddy, V. N., Kador, P. F., and Kinoshita, J. H. (1987) in Enzymology and Molecular Siology of Carbonyl Metabolism (Weiner, H., and Flynn, T. G., eds) pp. 317-324, Alan R. Liss, New York

250, 2933-2940

Hizi, A., and Yagil, G . (1974) Eur. J. Biochem. 45,201-209 Laemmli, U. K. (1970) Nature 227, 680-685