Structural Studies and Isolation of cDNA Clones Providing the Complete Sequence of Rat Liver Dihydropteridine Reductase*

The cleavage of reductively alkylated rat liver di- hydropteridine reductase with cyanogen bromide afforded a mixture of peptides, six of which (CB-1 to CB- 6) were isolated and purified by Cs reverse-phase high performance liquid chromatography. Portions of pep- tides CB-1, CB-4, and CB-6 were sequenced by automated Edman degradation and high performance liq- uid chromatography and the carboxyl-terminal region by conventional procedures. Further proteolytic diges- tion of CB-6 and isolation of the products afforded a seven-amino acid peptide. A low degeneracy probe comprising 20 nucleotides was synthesized from the sequence of this peptide and was used to screen a rat liver cDNA expression library constructed in the vec- tor XgtlO. Positive clones were isolated, and detailed examination of five of these by restriction endonu- cleases and dideoxy sequence analyses allowed identification of the entire coding region for dihydropteri- dine reductase. The gene was found to code for a protein of 240 amino acids (excluding the methionine initiator) of M, = 25,420. Each of the sequences corre- sponding to the peptides CB-1, CB-4, CB-6, and the carboxyl terminus were identified in the deduced pro- tein sequence. The rat enzyme is highly homologous to the human dihydropteridine reductase; the two pro-

Structural Studies and Isolation of cDNA Clones Providing the Complete Sequence of Rat Liver Dihydropteridine Reductase* (Received for publication, June 17, 1987) Manoucher Shahbaz, James A. Hoch, Kathleen A. Trach, John A. Hural, Stephanie Webber, and John M. WhiteleyS  The cleavage of reductively alkylated rat liver dihydropteridine reductase with cyanogen bromide afforded a mixture of peptides, six of which (CB-1 to CB-6) were isolated and purified by Cs reverse-phase high performance liquid chromatography. Portions of peptides CB-1, CB-4, and CB-6 were sequenced by automated Edman degradation and high performance liquid chromatography and the carboxyl-terminal region by conventional procedures. Further proteolytic digestion of CB-6 and isolation of the products afforded a seven-amino acid peptide. A low degeneracy probe comprising 20 nucleotides was synthesized from the sequence of this peptide and was used to screen a rat liver cDNA expression library constructed in the vector XgtlO. Positive clones were isolated, and detailed examination of five of these by restriction endonucleases and dideoxy sequence analyses allowed identification of the entire coding region for dihydropteridine reductase. The gene was found to code for a protein of 240 amino acids (excluding the methionine initiator) of M, = 25,420. Each of the sequences corresponding to the peptides CB-1, CB-4, CB-6, and the carboxyl terminus were identified in the deduced protein sequence. The rat enzyme is highly homologous to the human dihydropteridine reductase; the two proteins differ in only 10 amino acids, and all are conservative substitutions. In contrast, the sequence shows little homology with that of mammalian dihydrofolate reductase: reduced pyridine nucleotide-requiring enzymes with superficial mechanistic similarities.
Dihydropteridine reductase (EC 1.6.99.7) promotes the NADH-mediated reduction of "quinonoid" dihydrobiopterin to tetrahydrobiopterin in mammalian tissue. The tetrahydroderivative is a required cofactor in the enzymatic hydroxylation of phenylalanine, tyrosine, and tryptophan, and the integrity of the coupled reductase-hydroxylase cycles is vital to the host as they provide essential steps in the biosyntheses of dopamine, epinephrine, and serotonin (1-4). Disturbances in these pathways can lead to either elevated catecholamine production and hypertension ( 5 ) or cause phenylketonuria, a genetic defect in children which, if undetected, can lead to serious irreversible brain damage. This latter problem can * This investigation was supported by United States Public Health Service Grants GM 22125 and CA 11778. Publication 4884-BCR from the Research Institute of Scripps Clinic. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "aduertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
The nucleotide sequencefs) reported in thispaper has been submitted to the GenBankTM/EMBL Data Bank with accession numberfs) 503481.
$ To whom reprint requests should be addressed.
result from an absence of a functional phenylanine hydroxylase or dihydropteridine reductase and in some cases by deficiencies in pteridine cofactor biosynthesis (6, 7). Dihydropteridine reductase is of mechanistic interest as it requires an isomeric, structurally unstable quinonoid dihydropteridine as its substrate (8) in contrast to dihydrofolate reductase, an NADPH-requiring enzyme, which acts upon the 7,8-dihydrofolate/pteridine isomer (9). Because of the clinical importance of dihydropteridine reductase and its superficial mechanistic similarities to dihydrofolate reductase, this laboratory has published several reports characterizing the enzyme isolated from rat liver (10, 11) and has presented preliminary x-ray crystallographic data (12). Recently, to assist in physical and mechanistic comparisons between the two reductases and to allow full three-dimensional characterization of the crystal structure, it was considered necessary to obtain the complete enzyme amino acid sequence. Cloning techniques were considered most propitious; therefore, this report describes the identification of a suitable peptide sequence from the purified enzyme from which a DNA probe was synthesized and then used to identify complementary sequences in a rat liver cDNA library constructed in hgtl0. This approach ultimately allowed the isolation, characterization, and sequencing of the full dihydropteridine reductase coding region (720 base pairs), thus enabling deduction of the amino acid sequence of the enzyme. Synthesis of the Nucleotide Probe-This was kindly carried out by personnel at the Agouron Institute, La Jolla, CA, using an Applied Biosystems automatic synthesizer.

Rat Liver
Dihydropteridine Reductase Sequence 16413 procedure (14). Rat liver (600 g) was homogenized in 0.01 M acetic acid (1500 ml) containing 10 p~ protease inhibitors phenylmethylsulfonyl fluoride, TPCK, and TLCK. The mixture was centrifuged (10' X g), dialyzed against buffer, and then applied consecutively to two Cibacron blue agarose columns (3.5 X 20 cm). Recovery from the first column was achieved with 2 M NaCl and combined active fractions dialyzed against buffer prior to application to the second column where elution was achieved with 100 p M NADH in buffer. Application to phenyl-Sepharose (2.5 X 12 cm) pre-equilibrated with 1 M (NH&SO, in buffer and elution with buffer alone afforded a product which, after concentration, was passed through Sephadex G- 150 (2 X 120 cm) employing the same buffer. The standard buffer solution used throughout was 50 mM potassium phosphate, pH 6.8, containing 1 mM 8-mercaptoethanol. The average yield of enzyme, homogeneous by SDS-polyacrylamide gel electrophoresis, was 25 mg of specific activity >300 units/min/mg. Reductive Alkylation and Cyanogen Bromide Cleavage-According to the procedure of Crestfield et al. (15), enzyme (10 mg in 2 ml of standard buffer) was dialyzed against 0.1 M Tris-HC1 containing 0.5 mM EDTA and then treated with urea to a final concentration of 6 M. A 10-fold molar excess of 8-mercaptoethanol (relative to halfcystine content) was added, and the stirred solution was incubated under argon, in the dark, at 25°C for 2 h. A 10-fold molar excess of iodoacetamide was added, and the mixture was incubated for a further 2 h followed by the addition of a 10-fold excess of B-mercaptoethanol, dialysis against water overnight, and then lyophilization. Cleavage was carried out by the method of Prahl and Porter (16) by dissolving the sample in 70% formic acid (2 ml) containing a 100-fold molar excess of cyanogen bromide (relative to methionine content), and the stirred solution was incubated in the dark at 4 "C for 16 h. The sample was then diluted 10-fold with water and lyophilized.
Amino Acid Analyses-Purified enzyme or peptide fragments were hydrolyzed with constant boiling HCl in vacuo at 110 "C for 20-60 h. The amino acid compositions of the hydrolysates were analyzed with automatic amino acid analyzers, Beckman Models 119C and 6300.
SDS-Polyacrylamide Gel Electrophoresis-This procedure was carried out according to the method of Weber and Osborn (17)  Peptides were separated on 12.5% gels containing 1.25% bisacrylamide and 6 M urea, which were run at a constant current of 3 mA/gel, fixed in 10% trichloroacetic acid, and stained with Coomassie R-250.
High Performance Liquid Chromatography-HPLC was carried out with a Beckman Model 332 instrument equipped with a variable wavelength UV detector and a Model 427 recorder and integrator. To purify the CNBr peptides, reverse-phase HPLC was carried out on a 15-cm IBM Cs column employing a gradient of water to 2:l acetonitrile/2-propanol, each solvent containing 0.1% trifluoroacetic acid. From 0 to 60 min, the gradient was linear, from 60 to 70 min, isocratic, and from 70 to 100 min, linear.
Amino-terminal Analyses of Peptides-According to the procedure of Chang (18,19), DABITC was coupled to free amino-terminal groups by incubating the peptide or protein with 1 mM reagent in pyridine for 1 h at 70 "C. The excess reagents were removed from the reaction mixture by successive extractions with 2:l heptane/ethyl acetate, and the remaining material was dried under vacuum. The derivatized amino acid was cleaved from the peptide by a 1-h incubation with 38% trifluoroacetic acid at 54 "C after which the sample was dried under vacuum. The resulting mixture was dissolved in water (50 pl) and the released derivative extracted with butyl acetate (200 pl). The solvent was evaporated, and the residue was dried under vacuum, dissolved in ethanol, and then identified on silica TLC, by comparison with authentic standards, employing a two-dimensional solvent system of (a) 2:l (v/v) acetic acid/water and (b) 2:l:l (v/v) tolueneln-hexanelacetic acid.
Determination of the Carboxyl-terminal Peptide Sequence-The carboxyl-terminal residues of peptides were analyzed using carboxypeptidase Y according to the procedure outlined by Klenum (20).
Dialyzed protein was lyophilized and dissolved in 200 p1 of 0.1 M pyridine acetate, pH 5.6, containing 1% SDS and 0.1 mM norleucine and then incubated at 60 "C for 20 min to denature the protein. After cooling, an aliquot (25 pl) was removed to provide a zero time reading, and then carboxypeptidase Y was added and the mixture incubated at room temperature. Aliquots (25 pl) were extracted at 1-, 2-, 5-, lo-, 20-, and 30-min intervals, glacial acetic acid (5 pl) was added, and samples were lyophilized. Amino acids were then identified and quantitated by analytical HPLC.
Hydrolysis of CNBr Peptide 6-The peptide (-10 nmol dissolved in 50 mM ammonium bicarbonate containing 2 mM EDTA) was incubated with 50 pg of S. aureus V8 protease for 3 h at 37 "C, pH -8. The product was then subjected to reverse-phase HPLC analysis on the C8 column as described previously.
Oligonucleotide Labeling-According to the procedure of Sgaramella and Khorana (21), oligonucleotides were labeled by phosphate transfer from [Y-~'P]ATP in the presence of polynucleotide kinase. The mixture was subjected to electrophoresis on a 16% polyacrylamide gel and the location of the derivative identified by autoradiography for 15-60 s. The radioactively labeled band was then extracted by the crush and soak method of Maxam and Gilbert (22), filtered through glass wool, and then was used directly in hybridization experiments.
Isolation of Rat Dihydropterine Reductase cDNA Clones-The X g t l O library was screened (250,000 plaques) for dihydropteridine reductase inserts using the 32P-labeled oligonucleotide probe. Filters for hybridization were prepared according to the method of Benton and Davis (23), and the process was carried out by standard techniques for 16 h employing Denhardt's solution and SSC (0.3 M sodium citrate, 3 M sodium chloride buffer solution, pH 7) (24). 38 positive plaques were identified and were purified to 100% homogeneity by three subcloning steps. DNA from all 38 plaques (XrDHPR 1-38) was isolated from plate lysates and was purified by two extractions with phenol and one with chloroform. The purified samples of DNA were then precipitated with ethanol, dried, and resuspended in TE buffer (10 mM Tris-C1-, 1 mM EDTA, pH 8) prior to digestion with EcoRI (2-4 pg of DNA, 8-10 units of EcoRI) at 37°C for 1 h. Insert sizes were estimated by electrophoresis on agarose by comparison with a 1-kilobase marker. The electrode buffer contained 50 mM Tris-C1-, pH 8.1, 57 mM boric acid, 2 mM EDTA, and 50 pg/ml ethidium bromide.
DNA Sequencing-DNA containing cDNA inserts was digested with EcoRI, electroeluted from 4% polyacrylamide gels, and collected by ethanol precipitation. The purified material was then subcloned into the EcoRI site of M13mp18 and M13mp19 and then sequenced by the dideoxynucleotide method of Sanger et al. (25) and Williams et al. (26).

Enzyme Degradation and Peptide
Isolation-Purified dihydropteridine reductase was reductively alkylated in the presence of 6 M urea and then cleaved with CNBr in 70% formic acid. SDS-polyacrylamide gel electrophoresis showed the presence of eight major bands ranging in M, from 2,500 to 20,000 (27). Quantitative HPLC analysis of the products on a C, reverse-phase column afforded the profile shown in Fig.  1 and allowed the isolation of six peptides (CB-1 to CB-6). Each of the peptides demonstrated a free amino-terminal amino acid, which were Ile/Leu, Ala, Pro, Trp, and Pro, respectively, whereas the holoenzyme contained a blocked residue indicating that the amino-terminal peptide was not isolated from the HPLC column. CB-1, CB-4, and CB-6 were sequenced via automated Edman degradation. The sequences, published elsewhere (27), served both to confirm the aminoterminal assignments and also provide information by which a suitable nucleotide probe could be designed.
Probe Synthesis-The peptide designated CB-6 gave the following sequence from the amino terminus: Pro-Glu-Ala-

Asp-Phe-Ser-Ser-Trp-Thr-Pro-Leu-Glu-Phe-Leu-VaI-Glu-Thr-Phe-His-Asp-Trp-Ile-Lys-Gly-Asn-Lys-Gly-Pro-Asn.
Because some uncertainty often occurs with extended sequencing, further cleavage of this peptide was carried out using a protease from S. aureus strain V8 (28), and the product was subjected to HPLC on the Cs reverse-phase column. Elution with an aqueous 2:l acetonitrile/l-propanol linear gradient afforded five principal peaks (Fig. 2). Amino acid analysis showed that the second peak contained desirable amino acids of limited codon degeneracy and suggested that

Rat Liver Dihydropteridine
Reductase Sequence this region would be suitable for synthetic probe purposes. Its sequence was determined as Thr-Phe-His-Asp-Trp-Ile, with cleavage having occurred as expected adjacent to glutamic acid, and was identical to that contained in CB-6 above. As a prelude to cloning and sequencing the entire enzyme, a nucleotide probe 20 bases in length was therefore synthesized complementary to the expected messenger RNA sequence as follows.
mined by conventional protein sequencing as indicated in Fig.  4. The figure also illustrates the region complementary to the synthetic probe. Table I shows the predicted amino acid composition of rat liver dihydropteridine reductase, which agrees well with the previously published values for the enzyme. The amino acid composition of the human enzyme is also very similar, reflecting a high degree of sequence homology between the enzymes from the two sources.
The oligonucleotide mixture was purified and phosphorylated with 32P as described under "Experimental Procedures." Isolation and Sequence Analysis of Dihydropteridine Reductase Clones-A X g t l O library of cDNA from rat liver messenger RNA was screened by hybridization using the labeled probe. Hybridization was observed to 38 plaques of approximately 250,000 screened. These were purified until 100% of the plaques hybridized with the probe. One of the isolates, Xr DHPR38, was physically mapped by restriction enzyme analysis. The entire EcoRI fragment in this phage was inserted into M13 and sequenced using the oligonucleotide probe as a primer. Translation of the sequence corresponded exactly to that predicted from the sequence of the peptide confirming the identity of the isolate. The entire sequence of the dihydropteridine reductase structural gene was compiled from five separate bacteriophage isolates having different length cDNA inserts (Fig. 3). Amino Acid Sequence of Rat Dihydropteridine Reductase-Translation of the open reading frame of the nucleotide sequence has provided the first complete amino acid sequence of rat dihydropteridine reductase (Fig. 4). The predicted molecular weight of the enzyme subunit (without the initiator methionine) is 25,420 (240 amino acids), which is very close to the literature reports for the rat enzyme of 25,500 (14) and almost identical to the human enzyme ( M , = 25,760 (30)).
The predicted sequence contains each of the regions deter-

DISCUSSION
This paper reports the isolation and sequencing of cDNA clones for rat dihydropteridine reductase and includes the determination of the complete coding sequence for the holoenzyme. CNBr cleavage of the reductively alkylated reduc-   . Nucleotide and predicted amino acid sequence of dihydropteridine reductase cDNA. The nucleotide sequence was determined using the dideoxy sequencing procedure of Sanger et al. (25). The deduced amino acid sequence is shown below. The amino acid sequence numbers start from the aminoterminal alanine. Amino acids that differ in human dihydropteridine reductase are shown below the line, in particular three additional alanines and a serine-alanine conversion are inserted in the aminoterminal region. Sequences corresponding to peptides sequenced by conventional automated Edman degradation are underlined, and the probe region is boxed. tase afforded peptides that could be identified by polyacrylamide gel electrophoresis and isolated by C8 reverse-phase HPLC. Previous analyses had suggested both six and eight methionines per subunit; however, the final sequence showed seven. Therefore eight peptides should have been recovered after CNBr cleavage. In fact, only six peptides were isolated and identified. One peptide lost was the amino-terminal fragment, since no peptide with a blocked amino group was ever identified. Peptide CB-6 was degraded further with S. aureus strain V8 protease and a peptide isolated by HPLC whose composition, reflecting low degeneracy of the genetic code, allowed a 20-base nucleotide probe to be synthesized preparatory to screening the X g t l O rat liver cDNA library. The sequence of dihydropteridine reductase deduced from the nucleotide sequence of the cloned gene is very similar to that of the human dihydropteridine reductase whose sequence has been recently reported from two laboratories (30,31). Rat dihydropteridine reductase differs in 10 residues from one of the human enzyme sequences (30). The positions of substitution are randomly distributed throughout the protein, and large regions are completely conserved, e.g. residues 4-35,75-168, and 170-214. The most significant difference between the amino acid sequences of the rat and human enzymes lies at the amino-terminal position, there being an additional three alanines and an alanine-serine replacement in the human protein, i.e. Ala-Ala-Ala-Ala-Ala-Ala (human) versus Ala-Ala-Ser (rat). Preliminary structural evidence (27) has suggested a pyroglutamate-blocked terminal, but the current results support the concept of an acetylalanine block as has been suggested by other workers (29). Interestingly, none of the residues that differ are located in the conserved regions identified between dihydropteridine reductase and dihydrofolate reductase (30). The two human sequences reported contain a serine/threonine polymorphism at residue 50 (excluding the methionine initiator); this residue is serine (amino acid 47) in the rat enzyme.

CGAATTC
In addition to possible site-directed mutagenesis experiments, the cloning and sequencing of the gene will allow complete resolution of the x-ray crystallographic structure (12). Specific information on the active sites might further be deduced from the sequence, and comparison can also be made with the related dihydrofolate reductase, although preliminary analysis has suggested little structural similarity to this latter enzyme despite its superficially similar mechanism of action (32). However, a region of homology between human dihydropteridine reductase and dihydrofolate reductase (residues 100-109 and 16-25, respectively), noted by Dahl et al. (30) as being receptive to the binding of a nicotinamide moiety, is identical in the rat enzyme (residues 96-105) and thus also suggests the presence of a preferred nucleotide-binding site. It is of interest to note the remarkable conservation of se-

Rat Liver
Dihydropteridine Reductase Sequence quence between rat and human reductases, suggesting a crucial integrity of structure pertinent to function. This also 13. reinforces the probability that dihydropteridine reductase is an enzyme whose activity is essential to host survival. 14.