John M. Whiteley et uf. : Dihydropteridine reductase

Pteridines Vol. 4, 1993, pp. 159-173

Review

Dihydropteridine Reductase

John M. Whiteley§, Kottayil I. Varughesej:, Nguyen H. Xuong#, David A. Matthews, t and Charles E. Grimshawf

§The Scripps Research Institute, La Jolla, CA 92037, USA., #University of California at San Diego, La Jolla, CA 92093-0317, U.SA., tAgouron Pharmaceuticals, Inc., San Diego, CA 92121, USA, and fThe Whittier Institute, La Jolla, CA 92037, USA.

(Received August lO, 1993)

Summary

During the past decade numerous advances have been made in understanding the structure, mechanism and clinical properties of dihydropteridine reductase. An attempt is made here to delineate the current status of this essential enzyme by describing its structural features, its kinetic mechanism, the cloning and expression of both rat and human enzyme forms, the solution of their crystal structures, their classifica­ tion as members of a large family of short chain dehydrogenases, and finally a brief description is included indicating how current molecular biological applications have allowed the clinical definition of the aberrant form of phenylketonuria caused by a defective reductase.

Key words: Dihydropteridine reductase, Quinonoid dihydrobiopterin, Crystal structure, Aberrant PKU, Gene expression, Mutagenesis

Introduction and history in their heterocyclic nucleus of many centers for protonation, which often influence binding and reac­ Naturally occurring pteridines, which usually con­ tivity. Important metabolic functions of conjugated tain 2-amino and 4-hydroxyl substituents can be sep­ pteridine-mediated biological reactions include the arated into two distinct classes. One class contains one-carbon insertion reactions fundamental to pu­ the pterins of the folic acid series which possess rine biosynthesis (4. 5), the vitamin B,Tmediated syn­ a p-aminobenzoylglutamate (or polyglutamate) moie­ thesis of methionine (6), and the synthesis of thymi­ ty substituted into the heterocyclic nucleus via a 6- dylate (7). The most important unconjugated pteri­ methylene group, and the second class contains mo­ dine-containing enzymatic systems are associated lecules possessing other substituents at the 6-position, with the metabolism of the aromatic amino acids e.g., biopterin, neopterin and xanthopterin. Since (8-10). the discovery and classification of pteridines in Dihydropteridine reductase and phenylalanine. ty­ the early part of this century (1 , 2) a large number rosine and tryptophan hydroxylases play vital roles of pteridine-or folate-requiring enzymes have been in the synthesis of the catecholamines dopamine. identified and their mechanisms of action examined epinephrine and serotonin and indirectly can also (3). Many investigators have been intrigued by the intluence the generation of the melanin pigments. diverse functions of pteridines when they act as co­ Tetrahydrobiopterin is an essential cofactor in these factors or substrates in enzymatic reactions, by their metabolic pathways and facilitates the monooxygen­ unusual organo-chemical properties. by their pref­ ase activity which ultimately leads to the hydroxyl­ erence for hydroxylic solvents and by the presence ation of the aromatic amino acid substrate (Figure

Plcridines / Vol.4 / No. 4 160 John M. Whiteley el al.: Dihydropteridine reductase

protocol ineffective treatment for this group of pa­ tients and emphasizes the importance of characterising Phenylalanine Tryptophan ~~ NAO" these alternate defects. Currently, diagnosis is usually Tyrosine ~: achieved by further serological analyses (17), but H,N H H clearly a more precise understan·ding of DHPR ac­ T etrahydro bi opterln Aromatic tion could be of value in interpreting clinical situa­ Amino Acid Hydroxylases tions relating to a malfunctioning reductase. For example, if the genetic defect in DHPR deficient + H 0 2 patients can be identified, as has in fact been recent­ Tyrosine 01 hyd roxy p h e ny lal ani ne ly reported for several patients (18, 19), the lesion 5-hyd roxytryptophan H,NLa: H H can be related to the structure, and the cause for Hqulnonoid" dihydrobiopterin diminished enzymatic activity may be identified at Figure I. The participation of DHPR in the conversion of the molecular level. In addition, as the data base quinonoid dihydrobiopterin (R=dihydroxypropyl) to tetra­ of defective structures increases, a possible pattern hydrobiopterin, which is used as a cofactor in the aromatic of common causal features may evolve via correla­ amino acid hydroxylations. tion with the X-ray crystallographic s~ructure and targets may he created for molecular corrective ac­ 1). In these reactions phenylalanine, tyrosine and tion. tryptophan are converted to tyrosine, dihydroxyphen­ The purpose of this review article is to outline ylalanine and 5-hydroxytryptophan, and the cofac­ current knowledge relating to DHPR with a particu­ tor undergoes oxidation to a quinonoid dihydro lar emphasis on more recent developments in the structural and mechanistic understanding of this en­ form of biopterin (q-BH2). q-BH2 then becomes the substrate for DHPR and in an NADH-mediated zyme. Excellent reviews by Annarego et al. (20) and Shiman (8) have covered many of the salient fea­ reaction is recycled to BH4. The principal metabolic end products of these BH4-dependent amino acid tures of this enzyme and therefore areas covered in hydroxylations have important cellular functions. depth by these earlier reviews will receive only limit­ Dopamine (3,4-dihydroxyphenyIethylamine) is a cen­ ed coverage in the current report. tral nervous system neurotransmitter representing more than 50% of the total catecholamine content in the brain and spinal cord of many species (II). Source, purification and properties of DHPR and norepinephrine (3,4-dihydroxyphenylethanol­ amine) is the chemical transmitter in sympathetic DHPR has been obtained from several mamma­ neurons and is an essential neurotransmitter in lian sources that include sheep brain and beef adre­ brain tissue. Epinephrine (N-methyl-l(3-4-dihydroxy­ nal medulla (21), the liver of sheep (22), beef (23). phenyl)ethanolamine) is best known as an important rat (24, 25), monkey and human (26, 27), and hormone of the peripheral atKonomic system, al­ also from various murine (28) and human cell sam­ though it also functions as a transmitter in the olfac­ ples (29, 30). In addition, the enzyme has been isolat­ tory system and as a central nervous system neuro­ ed from species of Pseudomonas (31), E~cherichia coli transmitter (12). Serotonin (5-hydroxytryptamine) is (32), and Crithidia fasciculata (33). From mammalian an important brain neurotransmitter that exerts wi­ sources the tissue is usually diced and homogenised despread influence over arousal, sensory perception, in the presence of buffer or dilute acetic acid contain­ emotion and higher cognitive functions (13). ing a cocktail of protease inhibitors and solid ma­ The defective function of the hydroxylation pro­ terial is separated by centrifugation. The active com­ cess has long been recognized clinically in the auto­ ponent is then concentrated by ammonium sulfate somal recessive disease hyperphenylalaninemia or fractionation and after dialysis is subjected to a va­ phenylketonuria (PKU) (14). Originally this disease riety of chromatographic separation procedures. The was considered to correlate only with a defective earlier methods relied heavily on DEAE-cellulose phenylalanine hydroxylase, but a significant number or-sephadex fractionation but more recent efforts of cases have now been identified whose cause lies have employed specific affinity matrices or dye con­ in a defective DHPR function or aberrant biosyn­ taining matrices such as Procion Red or Cibacron thesis of the tetrahydrobiopterin cofactor (15,16). The Blue. In some instances a further hydrophobic chro­ enhanced complexity resulting from an additional matographic step followed by Sephadex or Sepha­ requirement for the cofactor in catecholamine biosyn­ cryl sizing columns has allowed isolation of a com­ thesis makes the standard phenylalanine-free diet pletely homogeneous product (34). A typical protocol

P :~ri di llcS / Vol.4 / NO.4 John M. Whiteley er al.: Dihydropteridine reductase 161

Table I . The Purification of the E. coli expressed human DHPR

Volume Enzy me Protein Specific Activity Recovery Purification Step (ml) (units) (mg) (units/ mg) (%) ------,,-,,-- Acetic Acid Extract 310 4200

(N H4)2S0.(21-45%) 112 1900 1st Cibacron Blue 282 21.630 725 29.8 100 2nd Cibacron Blue 186 19,400 69 281 90 Sephadex G-150 72 16,900 41 412 78

is shown in Table 1. It is interesting to note the progression of enhanced specific activities (from 2.5- -:>400 units/mg protein) over the years with the improvement in purification procedures. It is evident that the blue dye column materials, known to recog­ .2m(b) nise dinucleotide protein binding sites, are particu­ larly effective in ensuring a product of higher specific Figure 2. The comparative structures of the quinonoid dihy­ activity is obtained (35). This observation correlates dropteridine (a) and the 7,8-dihydropteridine (b). well with the structural and kinetic properties of the enzymes, outlined elsewhere in this report, that illustrate the enhance stability of the protein when (ku-O.02 f,lM) over NADPH (kct-2.2 f,lM), although bound to its dinucleotide cofactors. The isolation the possibility does exist that in vivo the relative con­ of the bacterial and flagellate reductases essentially centrations of reduced pyridine dinucleotides might requires the large scale growth of cultures, separation allow NADPH to function as the reductant in this and cell lysing followed by centrifugation and am­ situation. The substrate is invariably a 'quinonoid­ monium sulfate fractionation of the supernatant. dihydropteridine and not the more stable 7,8-dihy­ The dialysed active fractions are then purified by droisomer favored by dihydrofolate reductase (DHFR) various ion-exchange and sizing matrices. Since E. (Figure 2). The specific activities of highly purified coli has also become an optimal vehicle for express­ samples range from 300 to 400 units/mg protein for ing the wild type and mutant mammalian enzymes, both rat and human samples. The affinity for the a procedure has been adopted very similar to that oxidised dinucleotide is very low (k,-O.l mM) and described for the wild type mammalian enzyme (34) a single broad pH optimum in the vicinity of pH for purifYing these products to homogeneity (36, 37) 7 is observed. The enzyme is somewhat hydrophobic (Table 1). and is therefore susceptible to purification by phenyl Of the non-mammalian enzymes, that derived sepharose chromatography. The apo enzyme can be from E. coli is best characterised and has bee n titrated with p-chloromercuribenzoate, 5,5' -dithiobis shown to contain bound FAD (32). It exhibits simi­ (2-nitrobenzoic acid) and N-bromosuccinimide with lar properties to those of th e reductases isolated complete loss of activity in each case, however. from Pseudomonas and C. fasciculata. However, as bound NADH will offer protection against inactiva­ a class they are sufficiently different to those of the tion but neither quinonoid dihydro substrate nor mammalian enzymes, particularly in chromatogra­ tetra hydro product provide any protection (41). phic behavior, that they will be excluded from this review. In addition the NADPH-requiring pteridine reductases, also occurring in mammalian liver (38) Assay have sufficiently differing properties to allow their exclusion. Several assays have been developed for the estima­ Common features of the mammalian NADH-re­ tion of DHPR (22, 23, 42). All can be criticised qui ring reductases include a molecular weight of ap­ because of the instability of the enzyme substrate proximately 51,000 composed from two identical sub­ and the difficulty of generating the substrate in the units. Earlier isoelectric focusing studies had suggest­ assay. The general aim of the assay is to create a ed some disparity in subunit composition (39), but high concentration of the quinonoid dihydropteri­ more recent cloning and expression of the enzyme dine substrate, usually from the tetrahydro form, and cDNA would favor identity of the two subunits (40). observe its re-reduction by NADH. The change in The enzyme shows a marked preference for NADH absorbance at 340 nm can then be related to enzy-

Pteridines / Vol. 4 / No. 4 162 John M. Whiteley er af.: Dihydropteridine reductase

matic activlty. Unfortunately the quinonoid pterin The second substrate for DHPR is NADH. Evi­ substrates have varied stabilities with products fonn­ dence is derived from a twenty-fold magnitude distinc­ ing that have differing spectral properties. In addi­ tion in KM value between it and NADPH for the tion non-enzymic oxidation of NADH occurs· at an enzyme (42, 49), a marked difference in 1<...1 values appreciable rate and can also be a source of error. from fluorescence titration experiments and the ob­ The favored methods for generating the quinonoid servation, in the crystallographic structure of the rat form are to use peroxide and peroxidase, ferricy­ binary complex, of tight hydrogen bonding between anide, dichlorophenol indophenol and in earlier pa­ Asp 37 and the adenine ribose 2',3' -hydroxyl substi­ pers the associated amino acid hydroxylating sys­ tuents (50). The enzyme also shows strong affinity tems for which the tetrahydropterin is a substrate. for the inactive thio-NADH (kd-0.OO5 11M) and An alternate spectral procedure is to observe the adenine-uracil (kd-0.35 11M). The tight binding of reduction of ferri-cytochrome c at 550 nm by the NADH to DHPR appears essential for retaining tetrahydropterin fonned in the reaction. This assay protein stability and for creating the enzymatic site depends on initial incubation to generate the pterin as will be described further in the kinetic and cry­ dihydro quinonoid fonn to create a concentration stallographic sections of this review. plateau prior to enzyme addition, but suffers from similar problems to those occurring in the other as­ say procedures. A" there is clearly no completely Kinetics satisfactory approach it is probably advantageous to choose a procedure, buffer, temperature and pH Limitations in the assay procedures discussed earl­ convenient to a particular laboratory and ensure that ier have restricted the precision with which the me­ all assays are then self consistent. In this way com­ chanism can be analysed kinetically, however sev­ parative assays performed under these conditions eral studies favor an ordered mechanism in which can be readily interpreted. NADH binds first to thc enzyme followed by the quinonoid dihydropteridine substrate. The tetrahy­

dropteridine product then leaves followed by NAD j Substrates (51-53). It is suggested a change in enzyme confor­ mation at some stage in the sequence provides a The natural substrate for the enzyme in mamma­ rate limiting step. Several lines of evidence support lian tissue has been identified as quinonoid dihy­ the initial nucleotide binding. For example, the enzy­ drobiopterin (43) (Figure 1). The substrate is generated me can bind pyridine nucleotides in the absence from (6R)-L-erythro-5,6,7,8-tetrahydrobiopterin by of pterins, whereas the reverse is not true. and the various oxidising agents in vitro or during the course enzyme can be purified by NAD+-agarose, AMP-aga­ of the aromatic amino acid hydroxylations in vivo. rose or Cibacron blue agarose also in the absence Until recently specific chiral reduced biopterin sub­ of pterin substrate or product. Probably the most strates were not readily available but fortunately compelling evidence stems from the recent crystallo­ DHPR is tolerant of alternate substituents at the graphic work that demonstrates the pterin binding 6-and 7-positions and therefore the methyl analogues site is partially created by the presence of the bound have been used extensively. Substrates contain­ dinucleotide cofactor (54). ing a proton as the second substituent at C6 are Most reports with enzymes from various mamma­ relatively unstable and are therefore generated in situ lian sources show NADH to be the favored cofactor when required. Their mode of change is either via with consistent KM values of 5 -- > 30 11M and k.1 isomerisation to the more stable 7,S-dihydro isomer, values 01'-0.05 11M: NADPH shows a consistently or by substituent loss (44). The 7,S-dihydro isomer higher KM of-IOO 11M. NAD+ has a kd-IOO 11M is not a substrate for DHPR. More stable quinonoid with the rat liver enzyme, illustrating its propensity dihydro structures have been created in which the for dissociation from the protein. KM values of the C6 proton is replaced by either a deuterium or sec­ pterin substrate vary with the alkyl ring substituent ond alkyl or aryl substituent (45, 46). Unfortunately from values of-I J.1M for the 6R (-) isomer of the enzyme is not so tolerant of the disubstitution, quinonoid dihydrobiopterin up to-40 11M for the as the KM values increase markedly for such deriva­ unresolved isomers of the 6-methyl or 6}-dimethyl tives. The 4-hydroxy or keto substituent is also im­ quinonoid dihydropteridines, commonly used as sub­ portant for activity as the 4-amino analogs are not strate analogs in routine assays because of their substrates (47), however. again some license is offer­ being more readily accessible. It is interesting to : d to 2-amino substituent replacements (4S). note that 1<....", is consistently higher with the methyl

.. li nes / Vol.4 / No.4 John M. Whiteley et af.: Dihydropteridine reductase 163

Table 2. Values of typical kinetic constants for rat and human DHPR substrates and co factors

Source of DHPR Cofactor Substrate KM kc", kd

11M S- l 11M NADH 13 152 0.025 NADPH ISO 37 2.19 Rat" NAD+ 70.0

q 6,7-diMePtH2 27 170 q 6R ( ~)BH2 0.32 22 NADPH <) 17 0.034 NADPH 135 20 2.02 Human NAD+ 60

q 6,7-diMePtH2 18 162

q 6R(-)BH2 0.15 17

"The rat enzyme was isolated from liver a nd the human enzyme was isolated from a cl oned human liver eDNA expressed in E. culi. KM values for pterins were obtained with a fixed concentratio n of 100 11M NADH and KM values for the differing dinucleotides with a fixed concentration of 10 11M pterin. Measurements were carried out at pH 6.8 and 20°e. Dissociation constants were obtained by fluorescence titrations in the usual way observing the decrease in protein fluorescence at 340 nm with increasing concentrations of cofactor. substituted substrate analogs than is shown by the Considerable effort has heen expended towards natural substrate. Other pertinent factors include ob­ determining inhibitors with specificity for DHPR. servations that the reaction has a single somewhat Unfortunately, despite the superficial similarity to flat pH optimum around 7 (50), that both KM and DHFR no compound has been discovered with the kcat values increase with temperature (55), that the affinity amethopterin (methotrexate) displays for this kJ value for NADH suggests an on rate commensu­ latter enzyme. In fact, aminopterin (21 , 47), iodoami­ rate with the diffusion controlled limit for this sub­ nopterin, and amethopterin (61), are all inhibitors strate (52), and that the transfer of hydrogen from of DHPR (Ki-25 11M), however, the kinetic analyses NADH in the reaction is exclusively from the l3-face are inconsistent regarding competitive, noncompeti­ (i.e., the pro-S hydrogen atom) (56). tive or uncompetitive interactions with pterin sub­ The recently reported characterisation of DHPR strate. It is of interest to note in this context that cDNAs and their expression in Escherichia coli has whenever the DHPR-NADH binary crystal is soak­ allowed accurate sequencing of the enzyme and mu­ ed in solutions containing one of the above inhibi­ tagenesis experiments to be carried out (57-59). It tors prior to crystallography, the structural pattern is apparent that the enzyme contains a characteris­ emerging indicates a loss of the dinucleotide, suggest­ tic l3al3 Rossman fold (60) in the nucleotide binding ing the inhibitor may somehow disrupt the dinu­ region and that in the clone obtained from a rat cleotide binding site. The crystallographic ternary com­ liver source, aspartate-37 is crucial to the binding plex suggests that both extremities of the quin­ of the adenine ribose, via close hydrogen bonding onoid pterin nucleus may be solvent accessible (54), (50). It is this interaction that creates the preference therefore inhibitor design may restrict substitution the enzyme exhibits for NADH. An isoleucine mu­ to the immediate vicinity of the pterin ring and mech­ tant has been formed and as expected the differen­ anistically may require a quinonoid bond distri­ tial affinity of the enzyme for NADH rather than bution. The most interesting compounds falling into NADPH has been diminished. In fact kcaJ KM for this category (Figure 3) have been reported by Ar­ the NADH catalysed reaction falls from 12 f..lM-IS- 1 marego and associates (62). In these structures, the to 1.2 11M- l s- 1 with the mutant, whereas that for crucial 6-position of the pterin is replaced by a sub­ NADPH rises from 0.25 --> 0.6 IlM-IS-I , which is stituted nitrogen atom, thus locking the structure now close to half the figure for NADH, demonstrat­ into the quinonoid form. These compounds do in ing a similar stickiness for either dinucleotide. Se­ fact appear to be good inhibitors with Ki-IO --> lected examples of kinetic constants are illustrated in 100 IlM. Table 2. At one time it was considered that products of the catecholamine biosynthetic pathways might offer feedback inhibitory control of DHPR and thus afford Inhibition regulation of the enzyme in vivo (63). It was an

Pteridines / Vol. 4 / No.4

164 John M. Whiteley er aJ.: Dihydropteridine reductase

current molecular biological procedures to the eluci­ dation of its sequence from both rat and human sources (57, 59). In both cases oligonucleotide probes were used to identify the enzyme cDNA in phage libraries derived from rat and human sources respec­ tively. Expression in both E. coli and COS cells has led to both sequence identity (Figure 4) and Figure 3. 6-Methyl-6-azapterin designed as an inhibitor of the ability to express the respective proteins in good DHPR. The 6-nitrogen substituent ensures the structure re­ yields, particularly from the E. coli source. Moreover, ta ins a quinonoid double-bond distribution. the application of mutagenesis techniques has fur­ ther allowed specific amino acid replacements to be made such that the functionality of these specific interesting idea but it would appear the compounds units can be identified with regard to their binding would rapidly degrade to quinone structures that of substrate, cofactor and their influence on reaction interfered with accurate estimates of NADH concen­ rate and mechanism (36, 50). Table 3 illustrates mu­ trations in assays and thus gave erroneous figures, tant enzyme sequences that have been created with suggesting inhibition (64). the rat enzyme and the effect on kcat is shown. It is interesting to note (Figure 4) that the transition from rat to human enzyme requires merely the inser­ Antibodies tion of three extra alanines and deletion of one se­ rine at the N-terminal plus an additional ten conser­ Antibodies, both polyclonal and monoclonal, have vative random replacements in a protein of approx­ been raised against DHPR samples from various imately 240 amino acids. sources (65-67) and against pterins bound to albu­ As was indicated earlier, naturally occurring mu­ min (68). The resultant immunoglobulins have been tational changes in the sequence of human DHPR used for various purposes. Polyclonal antibodies rais­ can lead to aberrant forms of PKU. Many of these ed by standard techniques have been used to de­ errors have been identified (Table 4) particularly by monstrate the lack of DHPR in patients suffering Cotton and his associates (71). The usual procedure from an aberrant PKU in which the reductase is is to obtain a fibroblast culture or white cell pellet. missing, as a means of quantitating DHPR by en­ then generate a cDNA library that allows amplifica­ zyme immunoassay techniques (69), and for estimat­ tion of the DHPR component by PCR techniques ing the enzyme in human blood cells. A novel ap­ proach pursued by Cotton and his associates (70) has led to the generation of anti-idiotypic monoclonal 10 20 * ) 0 ~ A ,..;-;. - ,;, -",·-,.,-;c E A II. R V L V Y G G R G AL G S R ev Q ;. F R A R N W W V A antibodies, which have then been used quite effec­ H A ... : 5 __ -:C E A II. R V L V Y G G R G A [.. G S Re v Q II F II. A II. N W W V A ----- •• - 10 20 3D tively to probe differing folate and pterin binding u Y R ~ 5:"V": 0 v v ENE E A S A S;i-:l V ):; M T 0 S F TEO A 0 Q v T A E V G" L L C enzymes' to help delineate pterin binding sites on s:r;o v v E to E E A S A s:v:x v K K 'l; 0 S F T E Q Pt. D Q v T A £ V C J(; L L C '.. 4 0 .••• 50 60 70 the protein surfaces. This approach has proven use­ ful, however, conclusions should be reviewed with some caution as the contours of the anti-idiotypic 120 : 1 30 14 0 antibody determinant may not be sufficiently distinc­ STIS SHLPt.TK H LKECC L LTLAC J.KPt."t.. DCT PGMI G¥G H STISSH L ATKHt..KECC L LTL"GPt.K A" LOGTP GM ICYC H tive to categorically distinguish the pterin binding 120 130 1 4 0 • 160 " 170 180 site from sites for other similar molecules such as H ~ mlln A KG" V H Q L C Q S L " G K N 5 G K p !'p : C A A Pt. I A V L P V T LOT P H N Rat A K C A v H Q L C Q 5 LAG K N S G M P ;?:C A It 11 r ,., V 1.. P V T L D T P H N pyrimidines or nucleotides. Thus deductions drawn 160 }70 180 190 2 00 ... 210 · ... _3 20 , .. ~ regarding the similarity of DHFR and DHPR (68) Ft I( S H P t " 0 F S S iii T P L E F L V t T F H 0 i0oi I T C: I( N: II. P : S; S 0 S L I R J(; S H PEA D F 5 5 \01 T P L E F L V Ii: T F H D W ! T G ~. _K:: R P ~ ~ : S C S L I could perhaps be considered a little premature, since 190 2 00 2 10 2 2 0 240 when the structures were subsequently compared cry­ Q V V T T :'i:';C:'R':T E L T P " ¥ r stallographically (54), there appeared to be rather o V V T \.j~G:.~. : T E L T P JiI. \:0 limited similarity between the two enzymes. Figure 4. The comparative a mino-acid sequences of human and rat DHPRs. The numbering excludes the initial me­ thionine in both cases. Differences in sequence between Cloning, expression, mutagenesis and PCR species are highlighted by the dashed boxes. The positions of the reported naturally occurring errors are indicated by Progress in structurally characterising DHPR '*' for mutants. by '=' for insertion. and by ' + ' for chain made significant advances with the application of termination.

Pteridines / Vol.4 ! NO.4 John M. Whiteley et al. : Dihydropteridine reductase 165

Table 3. Comparative activities for rat-liver DHPR and va­ X-ray crystallography rious E. coil expressed .mutants

Enzyme kat The purified rat liver DHPR has been crystallized from a polyethylene glycol-ethanol mixture using a (sec·') vapor diffusion method (34, 54). It crystallized in Rat liver DHPR 156 E. coli DHPR 158 two different crystal forms. The first crystal fonn E. coli Ala-6-Val 159 amenable to X-ray diffraction studies was monocli­ E. coli Asp-37-I1e 43 nic, space group C2 with a=224.3 A, b=46.5 A, E. coli Trp-I04-Phe ISO c=94.9 A, 13=99.0° and contained two dimeric mo­ E. coli Tyr-146-Phe 0.89 lecules in the asymmetric unit (Figure 6); however, E. coli Lys-150-Gln attempts to find good isomorphous heavy atom de­ rivatives were unsuccessful. Fortunately, an ortho­

rhombic fornl, C2221' a=50.0 A, b= 139.2 A, c=64.8 o Table 4. Naturally occurring errors recorded for human A was obtained with one monomer per asymmetric DHPR unit. Diffraction data were measured on the native Amino acid replacement, insertion or excision (2.1 A) and four heavy atom derivatives-(3.5 A to 2.5 A). The heavy atom parameters were refined Gly 22 Asp (3 ) His 157 Tyr Trp 107 Gly Gly 169 Thr using the program package HEAVY and the initial Thr 122 insert Leu 204 stop MIR map to 5 A revealed the molecular boundary. Pro 144 Leu Phe 211 Cys A 2.8 Asolvent flattened map using B.c. Wang's pro­ Gly 150 Ser Arg 220 stop gram package showed identifiable features such as a-helix and 13-strands. Finally, the crystal structure The numbering of amino acids in this table excludes the of a binary complex of DHPR with its cofactor. ;nitial methionine as in Figure 4. NADH, was solved and refined to a final R-factor of 15.4% using 2.3 A diffraction data. As stated ear­ lier, DHPR is an am protein with a Rossman type dinucleotide fold for NADH binding. It is a dimer as is shown in Figure 6, and this is mediated by a four helix bundle motif (two helices from each protomer having an unusual right-handed twist). Surprisingly, DHPR is structurally and mechan­ istically distinct from DHFR appearing to more closely resemble certain nicotinamide dinucleo­ tide requiring flavin-dependent enzymes such as glutathione reductase. Attempts to generate a crystalline ternary complex have so far been unsuccessful. For example. it is Figure S. The superposItion of the equivalent Ca-carbon known that methotrexate. aminopterin and iodoami­ .;lums of human and rat DHPR. In particular the a-carbons nopterin are inhibitors of DHPR however, attempts ,I' residucs 8-243 from the human enzyme are superimposcd to diffuse these compounds into the binary structure "ith the a-carbons of residues 5-240 of the rat liver enzyme. The trace of the human structure is shown by the heavier II ne. Specific amino acid numbers are shown to illustrate the rat-human amino acid replacements using the human numbering system (see Fig. 4). and subsequent sequencing identifies the errors. Fig­ ure 5 illustrates the human DHPR structure obtain­ ed from X-ray crystallography. The occurrence of the currently known errors can be determined from the a-carbon backbone of this structure. As might he expected a priori the errors appear to interfere \\ith substrate or cofactor binding, dimer formation or the active site (see next section). Figure 6. The dimeric crystal structure of rat liver DHPR.

Pteridines / Vol.4 / No.4 166 John ~1. \\hit.:k~ del i. Dihydropteridine reductase have been fruitless. In fact, as was stated previously, xybenzoate hydroxylase. Such analogies are parti­ evidence from this and kinetic work suggest these cularly attractive, as the quinonoid dihydropteridine inhibitors may actually compete for the dinucleotide substrate contains the crucial conjugated chromo­ rather than pterin binding site. The orthorhombic phore of the flavins and thus one might expect DHPR

C2221 crystal form is tightly packed with a low sol­ to have many mechanistic and structural sim­ vent content compared to other proteins of similar ilarities to flavin-mediated hydroxylation reactions. size. This could be an additional reason for the dif­ Comparable reactivities have been suggested pre­ ficulties experienced forming the ternary complex. viously, but only with the enhanced clarity of the The C2 monoclinic crystal form has also now been emerging more definitive DHPR structure' can hy­ characteri<;ed to an R factor of 19.7% by a molecular potheses now engender a seI11blance of proven cre­ replacement method using the program MERLOT. dibility. Fortunately, this crystal is more loosely packed with The human DHPR has also been crystallized and two dimers per asymmetric unit and offers large sol­ the structure was solved using 2.5 A data and has vent entry ports to thc protein. It is possible this been refined to an R value of 16.9%. It also crystallis­ structure may be more amenable to formation of , ed in both the monoclinic and orthorhombic space the ternary complex. groups and the latter was more favorable to rapid Using molecular graphics the quinonoid dihydro­ solution. The superposition of the structural back­ biopterin substrate has been docked at the suspected bone of human and rat DHPRs are shown in Fig­ active site and using energy minimisation procedures ure 5. It is interesting to note that despite the se­ a tentative interpretive working model for the reduc­ quential amino acid changes there are only small tive reaction that conforms to known mechanistic differences in the two structures, in fact the r.m.s. restrictions has been created (Figure 7). It is appa­ deviation for all the Ca atoms is only 0.27 A. The rent that DHPR bears little resemblance to DHFR ten replacements occurring in the transition from in its nucleotide binding region. A schematic com­ rat to human DHPR are also shown in this figure. parison is shown in Figure 8. In the latter case the The average temperature factor changes associated nicotinamide component of the dinucleotide resides with the amino acid changes occurring in the tran­ close to the N-terminus of the enzyme, whereas sition from rat to human DHPR are delineated in in the case of DHPR the reverse is true and the Table 5. Matthews and coworkers (72) have studied nicotinamide component is towards the center of the effect of mutations on the thermal stability and the molecule and creates one face of the active site. folding of proteins. They have shown that substitu­ Such a configuration is consistent with enzymes in tion of residues that are solvent accessible and which hydride transfer from a reduced dinucleotide which have a high temperature factor have very little to a flavin can occur in reactions such as the NAD PH-mediated glutathione reductase and p-hydro- DHPR

DHFR

Figure 7. Stereoview of the active site of DHPR showing a possible binding mode for quinonoid dihydrobiopterin and amino acids pertinent to the binding site and mechan­ ism of action. Numbering of amino acids uses the rat no­ Figure 8. Comparison of the backbone folding of DHPR menclature (see Fig. 4) and distances between atoms is in with DHFR. The a helices are represented as cylinders and angstroms. the 13 strands as triangles.

Pteridines / Vol. 4 / No.4 John M. Whiteley et at.: Dihydropteridine reductase

Table 5. Temperature factors associated with the amino acid value of the average temperature factor for the rat changes occurring in the transition from rat to human DHPR molecule. In the human enzyme this awrage DHPR is 22.9 ;\2 which is more than twice the value of Rat Human the average for all residues.

Residue B(A') Residue B(A')

l1e 36 8.4 Val 39 2.8 Mechanism and classification of DHPR Val 48 9.3 l1e 51 4.8 Asp 73 20.8 Glu 76 26.5 Dihydropteridine reductase converts quinonoid di­ GIn 74 26.9 Glu 77 27.4 hydrobiopterin to tetrahydrobiopterin in an NADH­ Asp 169 19.0 Pro 172 5.1 mediated reaction. Dihydrofolate reductase generates Asn 215 37.4 Lys 218 31.7 tetrahydrofolate (FH4) from 7,8-dihydrofolate (FH ) Lys 216 30.9 Asn 219 27.5 2 Asn 219 19.0 Ser 222 8.7 in an NADPH-mediated reaction. Both products Asp 230 31.5 Glu 233 30.7 have the 5,6,7,8 tetrahydropteridine structure but have Lys 232 29.6 Arg 235 26.0 differing substituents at the 6-position; BH4 hav­ ing a dihydroxypropyl group, whereas FH4 has a para-aminobenzoylglutamate component (Figure 9). There is clearly a superficial similarity insofar as effect on enzyme stability. On the other hand, when each enzyme employs a reduced dinucleotide cofac­ huried residues with low temperature factors are tor to convert a dihydropteridine to its tetrahydro mutated, they reduce the stability of the protein. For counterpart. However, upon further examination this example with T4 lysozyme such a mutation typically initial observation proves to be deceptive. For exam­ caused the loss of stability by 4 kcal/mole. ple, DHPR is a dimer of M r -5l,000 daltons (58), The average temperature factors for all the resi­ whereas DHFR usually exists as a monomer with dues in human and rat DHPR enzymes are 10.6 M r -18 --> 22,000 daltons (74) and DHPR has a ..\ 2 and 14.2 ,.\2 respectively. Of particular note is clearly enhanced specificity for NADH (56) which that two of the random insertions Val 39 (lie 36 contrasts with the preference for NADPH exhibited in the rat) and lie 51 (Val 48 in the rat) have low by DHFR (74). The specificity for the dihydropteri­ temperature factors suggesting the replacement could dine substrate also differs. Although a quinonoid lead to significant energy changes. However, the two dihydrofolate can be generated that can be a sub­ changed residues occur on adjacent J3-strands (B strate for DHPR the quinonoid form canot be a sub­ and C) and the nature of the change and position strate for DHFR nor can the 7,8-dihydro form be ensures that this is both small and compensatory. a substrate for DHPR. This is realy not surprising In contrast, all of the other eight random substitu­ as the ground state distribution of electron density tions are highly exposed to solvent and are accom­ around the two dihydropteridine rings is distinctly panied by large thermal parameters. For example, different (1. Gready, personal communication) and two aspartate residues (73 and 230) in the rat are would suggest a priori an altered receptor site might replaced by glutamate residues (76 and 233) in the be necessary for reactivity. Moreover, the specific human enzyme and the asparagine and lysine resi­ activities of the two enzymes are very different (ap­ dues (215 and 216) in the rat become lysine and prox. 300 and 25 units/mg protein for DHPR (56) asparagine residues (218 and 219) in the human si­ and DHFR (75) respectively), suggesting both a dif­ tuation. Probably the most drastic change is that ferent mechanism and site of reaction for the two of serine 169 to proline 172 in the human protein. enzymes. This is further substantiated by the clear The introduction of the proline residue imposes a restriction on the rotation at this residue. However, the (, \jI) values of Ser 169 are - 58° and l3r respec­ tively and these values fall in the allowed region for a proline residue (73). Substitutions at the N-ter­ minal region also cause little stability change in the two structures as this region of the molecule can not be clearly resolved, presumable because of a lack of structural constraints and free interaction with the solvent. It is interesting to note that the average temperature factor for the eight residues Figure 9. The biosynthetic reductive pathways converting with limited constraints is 27 ;\2; almost twice the qBH, and 7,8-FH, to BH4 and FH4 respectively.

Pteridines / Vol.4 / NO.4 168 John M. Whiteley eT uf.: Dihydropteridine reductase lack of affinity exhibited by DHPR for the oxidised by analogy with flavin mediated reductive processes dinucleotide product of reaction, NAD+ (76) which (79). In this instance. however. the transfer of hy­ is a very different situation to that of DHFR, where dride must again be followed or preceded by a proton NADP+ has been identified as a participant in the transfer to achieve a fully reduced pteridine. Several kinetic turnover of the enzyme (77) and has also possibilities exist for the site of protonation (R I) (Fig­ been used in the generation of ternary complexes ure 10). that have sufficient stability to be examined by X­ Experiments conducted with tritium-labeled ray crystallography (78). Other experimentally observ­ NADH have indicated that hydride is transferred ed differences would include a single pH optimum to a solvent exchangeable position on the pteridine for DHPR (54) compared to the two for DHFR ring (56) and mechanistically this type of center also (74) the hydrophobic nature of DHPR, exhibited accommodates the proton, therefore labeling experi­ by its binding to phenyl sepharose (34) and the lack ments cannot define the reductive centers. Reports of unique titrCltable amino acids which can cause have also indicated little energy difference between abrupt changes in specific activity (41). This con­ the exocyclic and endocyclic isomeric forms of the trasts markedly with the characteristic enhancement of quinonoid dihydropteridines (81) therefore a priori eukaryotic DHFR specific activity when this enzyme it is difficult to assign a certain course of reductive is titrated with mercury salts (74). Moreover, there action. Nevertheless, specific features of the DHPR is no sequence correlation between the two enzymes active site containing a "docked" qBH2 suggest a (58, 74). course of reaction in which the Tyr 146 XXX Lys The crystal structures of DHPR and DHFR show 150 motif palticipates in proton donation. A structure that each enzyme contains eight l3-sheets with seven can be realised in which the hydroxyl substituent parallel and one antiparallel, however DHPR has of Tyr 146 is within 3 A of the 4-oxo group on a higher a-helical content with helices E and F form­ the pteridine, hence direct protonation could occur. ing an interface for the dimeric structure. As desc­ However, more recent high resolution X-ray struc­ ribed earlier (Figure 8) the dinucleotide binding of tures suggest a water molecule is situated such that the two enzymes is significantly different. Moreover, proton donation could be translocated to N3 of the in DHPR the active site for pteridine binding does pteridine. Both the lysine £-amino groups and tyro­ not exist until NADH is bound. The site is then sine hydroxyl groups have pKs-lO, thus it is pos­ created from the connecting regions between I3D and sible that a protonated Lys 150 influences Tyr 146 131' and the crossover that occurs between I3E and sufficiently so that it or the adjacent water molecule I3G. Thus rather than having a separate binding do­ is the source of the proton. Compelling supportive main for substrate such as is obscIyed with lactate evidence stems from the observation that a Lys ISO dehydrogenase, DHPR uses a strategy of simply ela­ GIn mutant has a specific activity of only 50 borating connecting loops to facilitate substrate bind­ units/mg protein compared to-300 units/mg for the ing that is intimately connected to a requirement natural enzyme and that a Tyr 146 Phe mutant has for prior NADH interaction. In fact the quinonoid virtually no activity, yet has an equally strong affin­ structure of the substrate pteridine has distinct simi­ ity for NADH as the wild-type enzyme, and also larities to that of the l1avin molecule and thus me­ responds to the rat DHPR polyclonal antibody. How­ chanisms of reduction such as those described for ever, Lys 150 is also close to the nicotinamide ri­ the NADPH-mediated glutathione reductase (79) bose hydroxyl groups therefore an additional func­ and p-hydroxybenzoate hydroxylase (80) appear to tion of this amino acid could be to aid in the cor­ be better analogies for understanding the DHPR rect orientation of NADH relative to substrate. mechanism of action. In crystallographic analyses It has been reported that the Tyr XXX Lys pattern of glutathione reductase the reduced nicotinamide ring is aligned parallel to the l1avin such that a hydride transfer is facilitated to N5 of the l1avin. A similar situation can be created using graphic tech­ niques with DHPR (Figure 7). In this figure, the reduced nicotinamide is oriented such that transfer of the C4 hydride can readily exocyclic qBH2 endocyclic qBH2 occur to the N5 position of the pteridine. Such a transfer has been suggested previously on structural Figure 10. Differing possible NADH-mediated reductive grounds (57), by the low ground state electron den­ routes for the quinonoid dihydropteridine when bound to sity at N5 (J. Gready, personal communication), and DHPR.

Pteridines I Vol. 4 I No.4 - - \ 1 Whiteley et a'-: Dihydropteridine reductase 169 is present throughout a diverse group of enzymes in phenyalanine metabolism caused by an autosom­ classified as the short chain dehydrogenases (82) al recessive gene defect in the enzyme phenylala­ (Figure 11). Each has a binding domain for a dinu­ nine hydroxylase (84). The usual treatment for pa­ cleotide cofactor and each has a second domain tients having this defect was a phenylalanine free diet. containing the sequences outlined in the figure. In However, it became apparent that a percentage of every case, and more than twenty examples are now patients (approx. 1 --> 5%) did not respond to this known, the Tyr XXX Lys motif appears essential diet and demonstrated progressive neurological pro­ for enzymatic activity. One other crystal structure blems. Through the efforts of Isabel Smith in Lon­ for an enzyme of this class has been reported, that don (85), Seymour Kaufman at NIH (86) and CUf­ of 20~DH (83) and here again it is possible that tius and Niederwieser in Zurich (87), it became ap­ the tyrosine participates, directly or indirectly via parent that both the biosynthetic route to the hydro­ water molecules, in proton donation. xylase cofactor, tetrahydrobiopterin, and the enzyme A unique distinctive feature between the dehydro­ DHPR that recycles this cofactor could be defective genases and Dl-fPR is that the alanine group, thir­ (Figure 12). Alternate therapies based on these con­ teen amino acids upstream from the tyrosine, is re­ clusions were therefore developed (88), and additional placed by serine. Structural analysis demonstrates it methods for clinical analyses and defect detection is sufficiently close to the active site to interact with were discovered (89). In several cases a lack of, or the tyrosine hydroxyl. This interaction could be cru­ deficiency in DHPR was confirmed. The DHPR cial to the ambivalent direction of reaction occurring gene has been identified with chromosome 4 (90) with the dehydrogenases. Nevertheless, it is clear and RFLP procedures have been employed in pre­ that the relevant energy differentials between the re­ natal screening of DHPR deficiency (91). Using cur­ spective substrates and products is also instrumental rent biological techniques Cotton and various asso­ in driving the reaction to completion. ciates have additionally been able to detect the spe­ Currently it would appear that DHPR has only cific gene defects occurring in several patients exhi­ superficial similarities to DHFR that in terms of biting DHPR deficiency (71, 92). Table 4 illustrates hydride transfer it has common mechanistic features the effect these defects have on the complementary to flavin reduction pathways and yet its overall struc­ protein and where these changes occur in the X-ray tural assembly suggests a close familial resemblance crystallographic structure can be traced in Figure to the short chain dehydrogenases. 5. It is interesting to note that the changes interfere with dimer formation, dinucleotide binding and the active site. Clearly the possibilities afforded by the Clinical developing areas of gene therapy (93) could perhaps very soon be applied to correct this debilitating me­ During the course of the past two decades it has tabolic defect. hecome apparent that PKU has a rather complex etiology. Earlier it had been identified as a defect

135 14 0 15 0 (1 ) J fiPR G ~ L L T LAG A K A A L D G T P G MI G Y G M A K .- PAH ?G DH G G I I I N M S S LAG L M P V A Q Q P v Y GAS K -­ Phe Tyr

17 PDH S G I R v L V T G S V G G L M G L P F N D v y e A S K -­

J ADH GIG i I leN I G S V T G F N A I Y Q v P v Y S G T K -.

20PDH G S I V N I S S A A G L M G L A L T S S Y GAS K • - f\ ~ GTP BH4 'q' BH, (3) Figure 11. Alignment of a selected region of fi ve short chain dehydrogenases including DHPR. Strictly conserved resi· dues are outlined. Residue numbers at the start of each ~ line refer to each sequence and those above to the rat liver DHPR DHPR. Abbreviations are: DHPR. rat liver dihydropteri­ (2) dine reductase; PGDH, human 15-hydroxyprostaglandin dehydrogenase; 1 7 ~DH , human 17~-h y droxysteroid dehy­ Figure 12. The three pathways where genetic mutations can drogenase; DADH. Drosophila melanogaster alcohol dehyd­ occur causing altered protein structures that affect enzyma­ rogenase and 20~DH. Streptomyces hydrogenans 20~-hydroxy­ tic action and ultimately give rise to symptoms of PKU steroid dehydrogenase. in patients.

Pteridines / Vol. 4 / No.4 170 John M. Whiteley el al.: Dihydropteridine reductase

Conclusion eds. 1990; Berlin: Walter de Gruyter. 4. Inglese J, Johnson DL, Shiau A, Smith JM, Benkovic S. DHPR has come a long way from being the pa­ Subcloning, characterisation, and affinity labeling of Es­ rochial "sheep liver enzyme". It has been isolated cherichia coli glycinamide ribonucleotide transformylase. from a great many other sources and both human Biochemistry 1990; 29: 1436-1443. 5. Smith GK Mueller WT. Slieker U , DeBrosse CWo Benko­ and rat enzymes have been cloned and expressed vic SJ. Direct transfer of one-carbon units in the transform­ in E. coli. The crystal structures from each of these ylations of de novo purine biosynthesis. Biochemistry 1982; sources have been obtained at high resolution and 21: 2870-2874. it is clear that any resemblance to DHFR is superfi­ 6. Fujii K., Galivan JH, Huennekens FM. Activation of me­ cial. Mechanistic analogies are more related to flavin thionine synthase: Further characterisation of the flavo­ reduction and there is clearly an interesting struc­ protein system. Arch Biochem Biophys 1977; 178: 662-670. tural and mechanistic resemblance to a larger family 7. Santi DV. Danenberg PY. Folates in pyrimidine nucleotide biosynthesis. In: Folates and Pterins, Vol. 2. Blakley RL, of short chain dehydrogenases. Complete elucidation Benkovic SJ, eds. 1984; 345-398, New York: John Wiley of the en0'me mechanism still awaits both the de­ and Sons. velopment of a natural ternary complex followed 8. Shiman R. Phenylalanine hydroxylase and dihydropterin by its X-ray solution, and the resolution of the pre­ reductase. In: Folates and Pterins, Vol. 2. Blakley RL, Ben­ cise interplay of amino acids, identified via both kovic SJ, eds. 1984; L79-249, New York: John Wiley and sequence alignment and mutational experiments, as Sons. participating at the enzyme active site. The similarity 9. Kaufman S, Kaufman EE. Tyrosine hydroxyLase. In: Fola­ tes and Pterins, Vol. 2. Blakley RL, Benkovic SJ, eds. 1984; exhibited by DHPR to the dehydrogenase family 251-352, New York: John Wiley and Sons. may go some way towards explaining the somewhat 10. Kuhn DM, Lovenberg W. Tryptophan hydroxylase. In: puzzling ubiquitous distribution of the reductase in Folates and Pterins, Vol. 2. Blakley RL Benkovic SJ, eds. tissues, where it appears to have no purpose, in that 1984; 353-382, New York: John Wiley and Sons. it may be a refinement of a common gene that ori­ 11. Dopamine. In: Advances in Biochemical Psychopharma­ ginally had a more general purpose. The unique cology, Vol. 19. Roberts PJ, Woodruff GN, Iverson LL, feature of the TyrXXXLys sequence is clearly an cds. 1978; New York: Raven Press. important conserved evolutionary feature for dealing 12. HoktTelt T, Fuxe K Goldstein M. Application of immuno­ histochemistry to studies on monoamine cell systems with with the reduction of a polarised double-bond, or special reference to nerve tissues. Ann NY Acad Sci 1975; the dehydrogenation of the complementary reduced 254: 407-43 2. species. It is possible that this new interpretation 13. Serotonin in biological psychiatry, In: Advances in Bio­ of DHPR at the molecular level might assist the chemical Psychophamlacology, Vol. 34. Ho BT, Schoolar clinical understanding of aberrant PKU caused by Je. Usdin E. cds. 1982; New York: Raven Press. a reductase deficiency and that it might in the fore­ 14. FoIling A. iller Ausscheidung von Phenylbreztrauben­ seeable future contribute to improved therapy. saure in den Ham als StoflWechselanomalic in Verbin­ dung mit 1mbezillitiit. Hoppe-Seyler's Z Physiol Chern 1934; 277: 169-176. 15. Blau N, Curtius H-Ch. Inborn errors in tetrahydrobiopte­ Acknowledgements rin metabolism. In: Chemistry and Biology of Pteridines. Curtius H-Ch, Ghisla S, Blau N, eds. 1989; 383-388, Berlin, The authors are indebted to the following agencies New York: Walter de Gruyter. for their support. Grant RROl644 (UCSD) and CA 16. Kaufman S, Holtzman NA, Milstien S, Butler IJ, Kre­ 11778 (TSRI) from the National Institutes of Health, mholz A. Phenylketonuria due to a deficiecy of dihydro­ grant DIR 88-22385 from the National Science Foun­ pteridine reductase. N Engl J Med 1975; 293: 785-790 17. Ponzone A, Ferrero GB, Guordamagna 0 , Ferraris S. Cur­ dation (UCSD), and the Lucille P. Markey Foun­ tius H-Ch, Blau N. Screening and treatment of tetrahy­ dation (UCSD). drobiopterin deficiency. In: Chemistry and Biology of Pteri­ dines 1989. Curtius H-Ch, Ghisla S, Blau N, eds. 1990; 394-401, Berlin, New York: Walter de Gruyter. References 18. Howells DW, Forrest SM, Dahl ' H-HM. Cotton RGH. In­ sertion of an extra codon for threonine is a cause of dihy­ 1. Wieland H, Schopf C. iller den gelben Rugelfarbstoff dropteridine reductase deficiency. Am J Hum Genet 1990; des Citronentallers (Gonepleryx rhamni). Ber dt Chern Ges 47: 279-285. 1925; 58: 2178-2183. 19. Smooker PM, Howells DW, Dianzani I, Cotton RGH. 2. Purrmann R. iller Rugelpigmente der Schmetterlinge. VII. Mutations causing dihydropteridine reductase deficiency Justus Liebigs Ann Chern 1940; 544: 182-190. detected by the chemical cleavage method. In: Pteridines 3. Chemistry and Biology of Pteridines 1989. Pteridines and and Related Biogenic Amines and Folates. Blau N, Cur­ Folic Acid Derivatives. Curtius H-Ch, Ghisla S, Blau N, tius H-Ch, Levine R Yim J, eds. 1992; 138-139, Korea;

Pteridines I Vol. 4 I No. 4 John M. Whiteley et al.: Dihydropteridine reductase 171

Hanrim Publishing Co. in Escherichia coli. Biochem J 1989; 261: 265-268. 20. Armarego WLF, Randles D, Waring P. Dihydropteridinc 38. Nakanishi N, Hasegawa H, Yamada S, Alcino M. Purifi­ reductase (DHPR), its cofactors. and its mode of action. cation and physiochemical properties of NADPH-specific Med Res Rev 1984; 4: 267-321. dihydropteridine reductase from bovine and human livers. 21. Cheema S. Soldin SJ. Knapp A Hofmann T, Scrimgeour J Biochem 1986; 99: 635-644. KG. Properties of purified quinonoid dihydropteridine re­ 39. Webber S, Hural JA, Whiteley 1M. Multiple forms of rat­ ductase. Canad J Biochem 1973; 51: 1229-1239. liver dihydropteridine reductase identified by their ditTering 22. Craine JE. Hall ES, Kaufman S. The isolation and chara­ isoelectric points. Arch Biochem Biophys 1986; 248: 358- cterisation of dihydropteridine reductase from sheep liver. 367. J Bioi Chern 1972; 247: 6082-6091. 40. Hoch J, Trach K, Schneider M, Whiteley 1M. Expression 23. Hasegawa H. Dihydropteridine reductase from bovine li­ of rat-liver dihydropteridine reductase in Escherichia coil. ver. J Biochem 1977; 81: 169-177. Chemistry and Biology of Pteridines. Curtius H-Ch. Ghisla 24. Webber S. Deits TL, Snyder WR Whiteley JM. The purifi­ S, Blau N, cds. 1990: 612-615. cation of rat and sheep liver dihydropteridine reductases 41. Webber S, Whiteley 1M. The etTect of specific amino acid by affinity chro,natography on methotrexate-scpharose. modifications on the catalytic properties of rat liver dihy­ Anal Biochem 1978; 84: 491-503. dropteridine reductase. Arch Biochem Biophys 1981 ; 206: 25. Nakanisi N, Hirayama K, Yamada S. A simple procedure 145-152. for purification of NADH-specific dihydropteridine reduc­ 42. Nielsen KH, Simonsen V, Lind KE. Dihydropteridine re­ tase from mammalian liver. J Biochem 1982; 92: 1033-1040. ductase: A method for the measurement of activity and 26. Cotton RGH, Jennings I. A naphthaquinone adsorbent investigations of the specificity for NADH and NADPH. for affinity chromatography of human dihydropteridine re­ Eur J Biochem 1969; 9: 497-502. ductase. Eur J Biochem 1978; 83: 319-324. 43. Kaufman S. The structure of the phenylalanine-hydroxyla­ 27. Firgaira FA, Cotton RGH, Jennings I. Danks DM. Use tion cofactor. Proc Nat! Acad Sci USA 1963; 50: 1085-1093. of a naphthaquinone adsorbent for the isolation of human 44. Randles D, Armarego WLF. Reduced 6.6,8-trimethyl­ dihydropteridine reductase. Methods in Enzymol 1987; pterins. Eur J Biochem 1985; 146; 467-474. 142: 116-126. 45. Armarego WLF, Randles D, Taguchi H. KM and kca, va­ 28. Nakanisi N, Kobayashi H, Yamada S. NADH-specific di­ lues for [6,6,7,7-2HJ7,7(6Hrdihydropterin and 2,6-diamino- hydropteridine reductase in mastocytoma P-815 cells. 5-iminopyrimidin-4-0ne with dihydropteridine reductase. Comp Biochem Physiol 1984; 78B: 515-518. Biochem J 1983; 211: 357-361. 29. Nakanishi N, Yoshida A, Ozawa K, Yamada S. Purifica­ 46. Bailey SW, Ayling E. 6,6-Dimethylpterins: Stable quinoid tion and properties of NADH-specific dihydropteridine re­ dihydropterin substrates for dihydroperidine reductase and ductase from human erythrocytes. Enzyme 1986; 35: 42-52. tetrahydropterin cofactors for phenylalanine hydroxylase. 30. Shen R-s, Abell CWo Purification of dihydropteridine re­ Biochemistry 1983: 22: 1790-1798. ductase from human platelets. J Neurosci Res 1981; 6: 193- 47. Lind KE. Dihydropteridine reductase: Investigation of the 201. specificity for quinoid dihydropteridine and the inhibition 31. Williams CD, Dickens G, Letendre CH, GurotT G, Haines by 2,~iaminopteridines . Eur J Biochem 1972; 25: 560-562. C Shiota T. isolation and characterisation of dihydropteri­ 48. Armarego WL, Ohnishi A, Taguchi H. New pteridine sub­ dine reductase from Pseudomonas species. J Bacteriol strates for dihydropteridine reductase and horseradish per­ 1976; 127: 1197-1207. oxidase. Biochem J 1986; 234: 335-342. 32. Vasudevan SG, Shaw DC Armarego WLF. Dihydropteri­ 49. Webber S, Whiteley JM. Pyridine nucleotide interaction dine reductase from Escherichia coli Biochem. J 1988; 255: with rat liver dihydropteridine reductase. J BioI Chern 581-588 . 1978; 253: 6724-6729 .'3. Hirayama K, Nakanisi N, Sueoka T, Katoh S, Yamada 50. Grimshaw CE, Matthews DA, Varughese KI, Skinner M, S. Dihydropteridine reductase and tetrahydropterin in Cn'­ Xuong NH. Bray T, Hoch , Whiteley M. Characterisation thidia jasciculata cells. Biochim et Biophys Acta 1980; 612: and nucleotide binding properties of a mutant dihydro­ 337-343. pteridine reductase containing an aspartate 37-isoleucine re­ 34. Matthews DA, Webber S, Whiteley 1M. Preliminary X-ray placement. J BioI Chern 1992; 267: 15334-25339. ditTraction characterisation of crystalline rat liver dihydro­ 51. Korri KL, Chippel D, Chauvin MM, Tirpals A, Scrim­ pteridine reductase. J Bioi Chern 1986; 261: 3891-3893. geour KG. Quinonoid dihydropteridine reductase from 35. Chauvin M, Korri KK, Tirpak A, Simpson RC Scrim­ beef liver. Canad J Biochem 1977; 55: 1145-1152. geour KG. Purification of dihydropteridine reductase using 52. Poddar S, Henkin J. Isotope, pulse-chase, stopped-flow. immobilised Cibacron Blue. Canad J Biochem 1979; 57: and rapid quench studies on the kinetic mechanism of 178-187. bovine dihydropteridine reductase. Biochemistry 1984; 23: 36. Matthews DA, Varughese Kl, Skinner M, Xuong NH, 3143-3148. Hoch J, Trach K, Schneider M, Bray T, Whiteley 1M. 53. Aksnes A, Ljones T. Steady state kinetics of dihydropteri­ Role of aspartate-37 in determining cofactor specificity and dine reductase: Initial velocity and inhibition studies. Arch binding in rat-liver. Arch Biochem Biophys 1991; 287 : 234- Biochem 1980; 202: 342-347. 239. 54. Varughese KI, Skinner MM, Whiteley 1M, Matthews DA, 37. Armarego WLF. Cotton RGH, Dahl H-HM, Dixon NE. Xuong NH. Crystal structure of rat liver dihydropteridine High-level expression of human dihydropteridine reductase reductase. Proc Natl Acad Sci USA 1992; 89: 6080-6084. (EC 1.6.99 . ~1. without N-terminal amino acid protection, 55. Randles D. Temperature dependence of dihydropteridine

Pteridines / Vol. 4 / No.4 172 John M. Whiteley el al.: Dihydropteridine reductase

reductase activity. Eur 1 Biochem 1986; 155: 301-304. 72. Alber T, Dao-pin S, Nye JA, Muchmore DC, Matthews 56. Annarego WLF. Hydrogen transfer from 4-R and 4-S (4- BW. Temperature sensitive m}ltations of bacteriophage T4 3H)NADH in the reduction of d./-cil·-6, 7-dimethyl-6,7(8H)­ lysozyme occur at sites with low mobility and low solvent dihydropterin with dihydropteridine reductasc from human accessibility in the folded protein. Biochemistry 1987; 26: liver and sheep liver. Biochem Biophys Res Commun 3754-3758. 1979; 89: 246-29. 73. Ramachandran ON, Sasisekharan Y. Conformation of poly­ 57. Dahl H-HM, Hutchinson W, McAdam W, Wake S, Mor­ peptides and proteins. Adv Protein Chern 1968; 23 : 283- gan F, Cotton ROH. Human dihydropteridine reductase: 457. Characterisation of a cDNA clone and its use in Analysis 74. Blakley RL. Dihydrofolate reductase. In: Folates and Pte­ of patients with dihydropteridine reductase deficiency. Nucl rins. Blakley RL, Benkovic SJ, eds. 1984; 191-253. New Acids Res 1987: 15: 192H932. York: John Wiley and Sons. 58. Shahbaz M, Hoch A, Trach K Hural JA, Webber S, Whi­ 75 . Kaufman BT. Kemerer YF. Purification and characterisa­ teley 1M. Structural studies and isolation of cDNA clones tion of beef liver dihydrofolate reductase. Arch Biochem providing the complete sequence of rat liver dihydropteri­ Biophys 1976; 172: 289-300. dine reductase 1 Bioi Chern 1987; 262: 16412-16416. 76. Lind KE. Dihydropteridine reductase. Measurement of dis­ 59. Lockyer J. Cook RO, Milstien S, Kaufma n S, Woo SLC, sociation constants for binary complexes of enzyme and Ledley FD. Structure and expression of human dihydrop­ ligands. Eur J Biochem 1973: 33: 67-70. teridine reductase. Proc Nat! Acad Sci USA 1987; 84: 3329- 77. Fierke CA, Johnson KA, Benkovic S. Construction and 3333. evaluation of the kinetic scheme associated with dihydro­ 60. Wierenga RK Terpstra P, Hoi WGJ, Prediction of the folate reductase from Escherichia coli. Biochemistry 1987: occurrence of the ADP-binding J3a(3-fold in proteins using 26: 4085-4D92. an amino acid sequence fingerprint. 1 Mol 1986; 187 : 1D1- 78. Bystroff C, Oatley SJ. Kraut J. Crystal structures of Escheri· 1D7. chia coil dihydrofolate reductase: The NADP+ holoenzyme 61. Whiteley, 1M. 1993; Unpublished information. and the folate-NADP+ ternary complex. Biochemistry 62. Randles 0, Taguchi H , Armarego WLF. New inhibitors 1990; 29 : 3263-3277. of dihydropteridine reductase (human brain). Tenth Inter­ 79. Pai EF, Schulz OE. The catalytic mcchanism of gluta­ nationl Syposium on the Chemistry and Biology"of Pteri­ thione reductase as derived from X-ray diffraction analyses dines and Folates 1993; in press. of reaction intermediates. J Bioi Chem 1983; 258: 1752- 63. Abell CW, Shen R-S, Gessnor W, Brossi A Inhibition 1757. of dihydropteridine reductase by novel l-methyl-4-phenyl- 80. Schneider WGH. Hoi WOH, Drenth J. Analysis of the 1,2,3,6-tetrahydropyridine analongs. Science 1984; 224: 405- active site of the flavoprotein p-hydroxybenzoate hydroxy­ 407. lase and some ideas with respect to its reaction mecha­ 64. Milstien S. Kaufman S. The oxidation of apomorphine nism. Biochemistry 1990; 29: 31 ()] -3108. and other catechol compounds by horseradish peroxidase: 8!. Gready E. Theoretical studies on pteridi nes. 3. Geometries. Relevance to the measurement of dihydropteridine reduc­ tautomer and ionisation energies, and rearrangement and tase activity. Biochim et Biophys Acta 1987: 923: 333-338. reduction mechanisms of the quinonoid dihydropteridinc 65. Milstien S, Kaufman S. Preparation of an antiserum to substrates of dihydropteridine reductase. J Amer Chern sheep liver dihydropteridine reductase. Methods in Enzy­ 1985; 107: 6689-6695. mol 1980; 66: 723-725. 82. Persson B, Krook M , Jornvall H. Characteristics of short­ 66. Webber S. Hural lA, Whiteley 1M. The estimation of di­ chain alcohol dehydrogenases and related enzymes. Eur hydropteridine reductase in human blood cells. Clin Chim J Biochem 1991; 200: 537-543. Acta 1988; 173: 117-126. 83. Ghosh 0, Weeks CM. Grochulski P. Duax WL, Erman 67. Figaira FA, Choo KH, Cotton RGH, Danks OM. Mole­ M, Rimsay RL. Orr C. Three-dimensional structure of cular and immunological comparison of human dihydrop­ holo 3(3,20(3-hydroxysteroid dehydrogenase: A member of teridine reductase in liver, cultured fibroblasts and contin­ a short-chain dehydrogenase family. Proc Natl Acad Sci uous lymphoid cells. Biochem J 1981; 197: 45-53. USA 19991 : 88: 10064-10068. 68. Ratnam S, Ratnam M. Cotton ROH, Jennings JO, Frei­ 84. Jervis OA The genetics of phenylpyruvic oligophrenia. J sheim IH. Anti-idiotypic antibodies elicited by pterin recog­ Ment Sci 1939; 85: 719. nise active site epitopes in dihydrofolate reductases and 85. Smith I, Clayton BE. Wolff OH. New variant of phenylke­ dihydropteridine reductase. Arch Biochem Biophys 1989: tonuria with progressive neurological illness unresponsive 275: 344-353. to phenylalanine restriction. Lancet 1975; I: 1108. 69. Kwan S-W, Shen R-S, Abell CWo An enzyme immunoas­ 86. Kaufman S, Holtzman NA, Milstien S. Butler 1], Krum­ say for the quantitation of dihydropteridine reductase. holz A Phenylketomuria due to a deficiency of dihydro­ Anal Biochem 1987; 164: 391-396. pteridine reductase. N Eng! J Med 1975; 293: 785-790. 70. Jennings l. Cotton ROH. Structural similarities among en­ 87. Curti us H-eh, Niederwieser A, Viscontini M. Atypical zyme pterin binding sites as demonstrated by a monoclo­ phenylketonuria due to tetrahydrobiopterin deficiency: nal anti-idiotypic a ntibody. J Bioi Chern 1993: Diagnosis and treatment with tetrahydrobiopterin, dihy­ 71. Smooker PM, Howells OW, Dianzani I, Cotton ROH. drobiopterin and sepiapterin. Clin Chim Acta 1979; 93: 251. The spectrum of mutations causing DHPR deficiency. Ten­ 88. Barlow S. Progress in phenylketonuria: Defects in the me­ th International Symposium on the Chemistry and Bio­ tabolism of biopterin. Pediatrics 1980; 65: 837-839. logy of Pteridines and Folates 1993; in press. 89. Howells OW, Smith I, Hyland. K. Estimation of tetrahy-

Pteridines I Vol. 4 I No.4 John M. Whiteley et at.: Dihydropteridine reductase 173

drobiopterin and other pterins in cerebrospinal fluid using 92. Ponzone A Guardamagna O. Dianzani L Ponzone R. reversed-phase and high-performance liquid chromatogra­ Ferrero GB. Spada M. Cotton RGH. Catalytic activity of phy with electrochemical and fluorescence detection. J Chro­ tetrahydrobiopterin in dihydropteridine reductase deficien­ matog 1986; 381: 28S-294. cy and indications for treatment. Ped Res 1993; 33: 12S­ 'Xl. Brown RM, Dahl H-HM. Localisation of the human di­ Wl. hydropteridine reductase gene to band p1S.3 of chromosome 93. Mikami H. Matsubara Y. Hayasaka K. Narisawa K. Oli­ 4 by in situ hybridisation. Genomics 1987: I: 67-70. mata M. Watanabe A Haginoya K. Miyabayashi S. Tada 91. Dahl H-HM, Wake S. Cotton RGH, Danks OM. The K. Dahl H-HM. Molecular analysis of dihydropteridine use of restriction fragment length polymorphisms in pre­ reductase deficiency and restoration of the enzyme activity natal diagnosis of dihydropteridine reductase deficiency. by gene transfer. J Inher Metab Dis 1990; 787-791. J Med Genet 1988; 2S; 2S-28.

Ptelidines I Vol.4 I NO.4