Kang et aZ.: Biosynthetic enzymes of tetrahydrolimipterin

Pteridines VoL 9,1998, pp. 69 - 84

Biosynthetic Enzymes of Tetrahydrolimipterin from Green Sulfur Bacterium Chlorobium Limicola

Dongmin Kang, Sangjoon Kim and J eongbin Yim §

Department of Microbiology and Institute for Molecular Biology and Genetics, Seoul National "l'ni\'ersity, Seoul 151-742, Korea

(Received December 1, 1997)

Summay

Based on the structure of limipterin (Cha, Pfleiderer, and Yim, Helv, Chim. Acta, 78: 600-61-1-. 199~ '. the biosynthetic pathway for the newly identified pterin glycoside was investigated. It \\'as dcmonstr.ltcd that tetrahydrolimipterin (H4 -limipterin) can be synthesized from GTP by the enzymes GTP C~ ' ­ c1ohydrolase I, 6-pyruvoyltetrahydroptcrin (PTP) synthase, sepiaptcrin reductase and lilTlipterin synthase, ;:,rcsent in the extract of Chlorobium limicola. Limipterin synthase (UDP-N -acetylglucosamine:5 ,6,7 ,8-tetrahydro-L-biopterin 2'-0-\)-N-acetylglucosa­ m.inyl transferase) catalyzed the condensation of tetrahydrobiopterin (H4-biopterin) with UDP-N-a­ cetylglucosamine in the presence of dithiothreitol and MnClz. It could also produce D-tepidopterin, [1- 0-(D-threo-biopterin-2'-yl)-J)- N -acetylglucosamine] when 5,6,7 ,8-tetrahydro-D-threobiopterin and UDP-N-acetylglucosamine were used as substrates. Substrate analogues such as UTP, UDP and UDP-N­ acetylgalactosamine inhibit the enzyme activity. The K.n values for tetrahydrobiopterin and UDP-N-a­ cetylglucosamine were 42.2 ~M and 124.3 ~M, respectively. Optimum pH and temperature were pH 8.0 and 37"C. The molecular weight of the enzyme was calculated to be 46,300 dalton from a calibrated Su­ perdex 75 and the subunit molecular weight was estimated at 46,000 dalton by SDS-PAGE. These results suggest that limipterin synthase exists as a monomer. Biosynthetic intermediates ofH4-limipterin such as H zNTP, 6-PTP, and H 4-biopterin were identified in vi­ tro using purified GTP cyclohydrolase I, PTP synthase, sepiapterin reductase, and limipterin synthase. From the HPLC and TLC analyses of the enzymic intermediater, it could be concluded that H4-Jimipterin comes from GTP by way of H 4 -biopterin in Chlorobium limicola.

Key words: Tetrahydrolin1ipterin, Green sulfur bacteria, Chlorobium limicola, GTP cyolohydrolase, PTP synthase, Sepiapterin reductase, Limipterin synthase

Introduction been cloned from mammalian and insect sources.

Three enzymes are involved in synthesizing H 4 - In mammalian cells, the de novo biosynthesis of biopterin (10,11,12,13,25). GTP cyclohydrolase H 4- biopterin has been elucidated over the last 15 I produces Hrneopterin triphosphate (HzNTP) years, and recently biosynthetic enzymes have from GTP (25). The sequential action of 6-py­ ruvoyltetrahydropterin synthase (PTPS) and sepi­

§ Author to whom correspondence should be addressed. apterin reductase yields the final product H 4 - This article is dedicated to Prof. Wolfgang Pfleiderer. biopterin. Besides or instead of sepiapterin reduc-

Pteridines/ VoL 9 / No. 2 ::-0 Kang et al.: Biosynthetic enzymes of tetrahydrolimipterin

tase, aldose reductase and lor carbonyl reductase, o HN:t.Ny ~ R ~ _ can also catalyze the reduction steps (15). Until GTP H2NJ.."N ~OSH2 0 -~=-0 - ~:'0-~:,0 recently, it was thought that the reduction of H 2 - i)----{1 0 0 0 biopterin to H 4 -biopterin by dihydrofolate reduc­ OH OH tase was the final reaction of the pathway. I GTP Cyclohydrolase I However, since the biosynthesis of H -biopterin 4 ,. occurs in the absence of dihydrofolate reductase H NAyN~%~%~CH20 - ~-O-~-O - ~-O- H.2 NTP or in the presence of dihydrofolate reductase in­ J.." J( ..-l .. M M 0- 0-·0- H2N N N H H hibitors, this Hrpterin pathway must be in­ H correct. It was then proposed, and evidence was H.2 + !6-PTP Synthas~ presented, that the true intermediates between ,.

H 2 NTP and H 4 -biopterin are actually H 4 -pterins, namely pyruvoyl-H4-pterin (PTP) and lactoyl-H4 - pterin (LTP). Pteridine biosynthesis in microorganism had 2NAO"K ~ Sepiapterin Reductase not been studied as much as pteridine biosyn­ 2NADP++ 2H ~ thesis in mammalian cells. The activity of biosyn­ HN~1 :+.~-~-CH3 the tic enzymes found in animal cells was detected ~ Jl. ..JcN OK OK H2N N N H in some microorganisms. GTP cyclohydrolase I, K the enzyme catalyzing the first step of pteridine UDP-G1UNAC~ Mn2 + Transferase synthesis from several bacteria has been described, uo. and the enzyme has been purified to apparent homogeneity from E. coli and Bacillus subtilis (20, 25). In E. coli, the gene which encodes GTP cy­ TETRAHYDROLIMIPTERIN clohydrolase was cloned and analyzed in relation to the regulation of pteridine biosynthesis. A new pteridine was purified and identified Figure 1. Proposed pathway for the biosynthesis of tetra­ from Chlorobium limicola, a photosymthetic Green hydrolimipterin in Chlorohium limicola. Sulfur bacterium. This pterin compound is named as limipterin after the species name limicola (4). From these results, the overall biosynthetic path­

The complete structure of limipterin is proposed way for H 4 -lirnipterin could be proposed (Fig. 1) as [I-0-(L-erythro-biopterin-2 r -yl)-I3-N-acetyl glu­ (9). In this report we described purification and cosamine]. The molecular formula is C17H240~N6 characterizations of the enzymes involved III and it contains a biopterin and a glucosarnine limipterin biosynthesis in Chlorobium limicola. moiety. The attachment of N-acetylglucosamine was confirmed by acid hydrolysis as well as en­ Materials and Method zymatic hydrolysis experiment using 13- N -ace­ tylglucosaminidase. Lirnipterin was susceptible to Microorganism and cultivation this enzyme, which is highly specific for l3-glyco­ side linkage of N-acetylglucosamine. The oxida­ Chlorobium limicola f. thiosulfatophilum NCIB tion state of native limipterin was established by 8327 was supplied by S. Sirevag (University of the differential iodine oxidation experiments. It Oslo, Norway) and grown photoautotrophically could be estimated that >93% of limipterin is main­ under the light intensity of 50-80 par. for 7 to 10 tained as tetrahydro form in vivo (4). days at 30°C. The organism was cultured in nar­ Based on the constructed structure, the path­ row neck polycarbonate carboys each containing way leading to the biosynthesis of lirnipterin was 19.8 liter of modified Pfenning medium (18) investigated. GTP cyclohydrolase, PTP synthase, which contained KH2 P04 (1.0 g), NH4 Cl (0.5 g), and sepiapterin synthase were all present in the 30- Na2S203 · 5H2 0 (1.0 g), MgS04 • 7H20 (0.4 gL

80% ammonium sulfate fraction of Chlorobium NaCI (0.5 g), CaCl2 . 2HzO (0.05 g), 1.0 ml of extract. The enzyme activity that produce tetrahy­ Vitamin Bl2 solution (2.0 mg/IOO ml), 1.0 ml drolimipterin using tetrahydrobiopterin and UDP­ of trace element solution (EDTA 5.2 g, FeCl2 .

N-acetyl-glucosamine as substrates existed in 40- 4HzO 1.5 g, CoCl2 . 6H20 190 mg, MnCh· 4H20 60% ammonium sulfate fractions. This enzyme ab­ 100 mg, H 2B03 6 mg, CuCl2 • 4H20 17 mg, Na2 solutely requires manganese ion and dithiothreitol. Mo04 . 2H2 0 188 mg, NiCI2 · 6HzO 25 mg,

Pteridines/ Vol. 9/ No. 2

72 Kang et at.: Biosynthetic enzymes of tetrahydrolimipterin

Sepiapterin reductase (SR) act:J.vlty was deter­ of 100 mM PIPES buffer (pH 6.8), containing 2.S mined spectrophotometrically by the method of Ka­ mM DTT. The cell suspension was sonicated toh (22). The standard assay mixture contained the anaerobically at successive 1 second sonication following components: O.OS M potassium pho­ and S seconds intervals for 30 min and then cen­ sphate buffer (pH 6.8), 0.1 mM NADPH, O.S trifuged at lS,OOO x g for 40 min. Saturated am­ mM sepiapterin, S mM dithiothreitol, and enzyme monium sulfate solution containing 2.5 mM DTT solution in a final volume of 0.7 ml. The reaction was added to the supernatant (950 ml) fo r frac­ was initiated by adding enzyme at 30"C. The rate tionation. of disappearance of sepiapterin was monitored at 420 nm. One unit of SR was defined as the am­ Purification of GTP cyclohydrolase J ount of enzyme that caused the disappearance of 1 pmol sepiapterin per min wlder the assay condition. The 20-40% ammonium sulfate precipitate was dissolved in SO mM Tris-HCI buffer (pH 8 .0), Limipterin synthase assay containing 10 mM l3-mercaptoethanol. The ma­ terial (24 ml) was dialyzed twice against SOO The reaction mixture containing 0 .1 M PIPES volumes of 20 mM Tris-HCl buffer (pH 8.0) con­ (pH 6.8), 10 mM MnClz, 10 mM Glucose, 12 U taining 10 mM l3-mercaptoethanol (buffer G). Glucose oxidase, 100 IlM UDP-GlcNAc, 200 IlM The dialyzed sample (26 ml) was applied to a co­

H 4 -biopterin, 10 mM DTT, and enzyme fraction lumn of DEAE-sephadex (Vt=80 ml) equilibrated was incubated at 3TC for 1 hr. The final reaction in advance with buffer G. The column was wash­ volume was 40 Ill. The reaction was terminated ed with the same bufter and developed with a by adding 20 III of acidic iodine solution (1 % 121 linear gradient of 0.0-1.0 M NaCI in a total 2% KI in 1 M HC1) and left for 2 hrs at room volume of 300 ml at a flow rate of 40 mljh. Ac­ temperature in the dark. The precipitate was re­ tive fractions (0.6-0.7 M NaCl) were co llected moved by microcentrifugation at lS,OOO x g for S and dialyzed twice against 500 ml of 20 mM Tris­ min, followed by adding IS III of 2% ascorbic acid. HCl (pH 8.0) buffer containing 2 .S mM EDTA, After microcentrifugation, S III aliquot was analyz­ and 10 mM l3-mercaptoethanol. Dialyzed frac­ ed with C18 NovaPak HPLC column using 10 tions (13 ml) were loaded onto a column of GTP­ m~l potassium phosphate (pH S.S), 2 .S mM agarose (0.8 x 6 cm, Vt=3 ml) which had been e­ EDTA as mobile phase at a flow rate of 1.0 m1/ quilibrated with the dialysis buffer. The column min. The fluorescence of limipterin was monitored was washed with the same buffer and developed at the wavelength of 4S0 nm upon excitation at successively with (1) 20 mM Tris-HCI (pH 8 .0), 350 nm. One unit of LS was defined as the am­ 2.5 mM EDTA, 10 mM l3-mercaptoethanol, 0.2 ount of enzyme which produced 1 nmol limipte­ M NaCl; (2 ) the equilibration bufter; (3) 20 mM rin per min under the standard assay condition. Tris-HCI buffer (pH 8.0), containing 2 .5 mM EDTA, 10 mM l3-mercaptoethanol, and 1 mM Determination of protein GTP. Most of the enzyme activity was recovered in the final elution buffer. The protein concentration was determined ac­ cording to the method proposed by Bradford (2 ), Partial Purification of PTP synthase and sepiap­ using bovine serum albumin as a standard protein. terin reductase

Purification of the biosynthetic enzymes of H 4 - PTP synthase- The 50-70% ammonium sulfate limipterin from Chlorobium limicola precipitate was dissolved in 100 mM PIPES buff­ er (pH 6 .8 ) containing 2.5 mM DTT, and was ap­ Unless otherwise indicated, all procedures were plied to a column of Ultrogel AcA 44 (3 .0 x 85 carried out in the cold room, and buffers were em, Vt=SOO ml). The column was developed with prepared fresh. All buffers contained 2.S mM di­ 20 mM PIPES buffer (pH 6.8) containing 2 .S thiothreitolor 10 mM l3-mercaptoethanol to main­ mM DTT (buffer P ). The fractions containing en­ tain anaerobic condition. zyme activity were combined and loaded onto a Reactive-Yellow column (3.0x 4 .S cm, Vt=25 ml) Preparation of crude extract that had been equilibrated with bufier P. The column was washed with the same buffer and The cells ( 100 g) were suspended in 8 volumes the proteins were eluted with a linear gradient of

Pteridines/ Vol. 9 / N o. 2 Kang e t at.: Biosynthetic enzymes of tetrahydrolimipterin 73

0 .0-1.0 M NaCI in a total volume of 80 mI. The (separation range=5,000-75,000 dalton, Vt=24 ml) active fractions were collected and stored at -80°C equilibrated with 50 mM PIPES buffer (pH 6.8) until used. containing 2.5 mM DTT at a flow rate of 0.4 mll Sepiapterin reductase- The 30-50% ammonium min. The column was developed with the same sulfate precipitate was dissolved in 100 mM buffer at the same flow rate. The active fractions PIPES buffer (pH 6.8) containing 2.5 mM DTT were pooled and stored at -80°C until used. and applied to a column of Ultrogel AcA 44 (3. Ox 85 cm, Vt=500 ml). The column was develo­ Molecular weight and subunit size of limipterin syn­ ped with buffer P . The active fractions were col­ thase lected and subjected to chromatography on Reac­ tive-Brown Agarose (3.0 x 5 em, Vt=30 ml) pre-e­ The molecular weight of the enzyme was de­ quilibrated with buffer P . The column was wash­ termined by gel filtration chromatography on a ed with the same buffer and developed with a gra­ Superdex 75 (Vt=24 ml, Pharmacia). The column dient of 0.0-1.0 M NaCI in a total volume of 120 was calibrated with Blue dextran (2,000 kd), al­ ml at a flow rate of 40 ml/ h. bumin from bovine serum (68 kd), albumin from hen egg (45 kd), chymotrypsinogen A (25 kd) Purification of limipterin synthase and cytochrome c (12.5 kd). SDS-PAGE was per­ formed on 12.5% polyacrylamide gels. As standard Saturated ammonium sulfate solution contain­ marker proteins, myosin (220 kd), phosphorylase ing 2.5 mM DTT was added to the crude extract b (97.4 kd), BSA (66 kd), ovalbumin (46 kd), car­ (940 ml) to obtain 40-60% fractions. The am­ bonic anhydrase (30 kd), trypsin inhibitor (21. 5 monium sulfate precipitate was dissolved in 100 kd) and lysozyme ( 14.3) were used. After elec­ mM PIPES buffer (pH 6.8) containing 2.5 mM trophoresis, protein band was visualized with the DTT and applied to Sephacryl S100 (3.0 x 90 em, silver staining method according to Wray (23). \ Tt =550 ml) pre-equilibrated with buffer P and de­ \'eloped with the same buffer at a flow rate of 35 Enzymatic synthesis of Hrlimipterin from GTP mljh. The active fractions were pooled and load­ ed onto an activated Methotrexate-agarose. The Preparation of dihydroneopterin triphosphate­ column (0.8 x 10 em, Vt=5 ml) which had been GTP-agarose eluant was used as enzyme source. equilibrated with buffer P was washed with the The reaction mixture contained 0.1 M Tris-HCI same bufier and developed successively with 1) buffer (pH 8.0), 0.1 M NaCl, 3 mM GTP, 5 0.2 M NaCl, and 2) a linear gradient of 0.2-1.0 mM EDTA and purified GTP CHase I in a total M NaCl. Solid NaCI was added to the pooled ac­ volume of 500 f.ll and was incubated at 42"C t()f tive fractions until 1 M concentration was reached. 2 .5 hours. The dihydroneopterin triphosphate pro­ The sample was then applied to Phenyl-Sepharose duced was used as substrate for the subsequent (1.8 X 5.8 em, 10 ml) that had been equilibrated biosynthesis experiments. At intervals of 0.5, 1 with 20 mM PIPES buffer (pH 6.8) containing and 2.5 hours, 20 f.ll of reaction mixture was used 2.5 mM DTT, and 1 M NaCl. The column was for product analysis. Ten f.ll of 5 N TCA and 25

washed with the equilibration bufier and de­ f.ll of iodine solution (1% 12/ 2% KI) were added veloped with a gradient of 1.0-0.0 M NaCI in a to the mixture. After 30 min of oxidation, 13 total volume of 40 m!. The enzyme fractions f.ll of 2% ascorbic acid was added to destroy ex­ were collected and dialyzed with 2 liter of 20 cess iodine. The aliquots were analyzed with mM Tris-HCI (pH 7.4), 2.5 mM DTT. Dialyzed f.lBondaPak C18 column using 2% methanol as flow-through fractions (the active fractions) were solvent. To confirm the reaction product, each 20 applied to Mono-Q column (Vt=1 ml) pre-e­ f.ll sample was mixed with 1 U alkaline phos­ quilibrated with 20 mM Tris-HCl (pH 7.4), 2.5 phatase and incubated for 1 hour. Ten f.ll of 5 N mM DTT using FPLC system. The column was TCA and 25 f.ll of iodine solution (1% Iz/2% KI) washed with the equilibration buffer and de­ were added to the mixture. After 30 min ox­ veloped with a gradient of 0 .0-1.0 M NaCI in a idation, 13 J..lI of 2% ascorbic acid was added and total volume of 25 m!. Salt eluted active fractions analyzed with f.lBondaPak C18 column using 2% were concentrated to a small volume (400 f.ll) us­ methanol as solvent. ing Centricon (M. W. cutoff 30,000 dalton). Two Preparation of PTP from HjVTP- PTP synthase hundred f.ll of the concentrated sample (total was partially purified through Ultrogel AcA34 volume=400 f.ll ) was then applied to Superdex 75 and Reactive-Yellow dye-binding gel and sepi-

Pteridines/ Vol. 9 / No. 2 Kang et al.: Biosynthetic enzymes of tetrahydrolimipterin apterin reductase was isolated by using Ultrogel reductase (Afti-Brown fraction), and limipterin AcA44 and Reactive-Brown agarose gel as des­ synthase (Superdex 75 fractions) in a total volume ciebed previously. Partially purified PTP synthase of 80 f...Ll was incubated at 37"C for 2hrs. The pro­ was mixed with 0.1 M Tris-HCl buffer, pH 7 .4 duct was analyzed by the same method described containing 25 mM MgCI2 , 420 f...LM H 2NTP, 10 previously. mM DTT in a total volume of 20 f...L1. The mix­ ture was incubated at 37"C for 2 hours. Ten f...ll of Results

5 N TCA and 50 f...Ll of 1% 12/ 2% KI solution were added to convert 6-pyruvoyl tetrahydrop­ Localization of enzymes involved in the biosynthe­ terin to pterin in acidic conditions. After 60 min, sis of Hrlimipterin the insoluble material was removed by centrifuga­ tion at 13,000x g for 10 min and excessive iodine Based on the constructed structure, the biosyn­ was destroyed by the addition of 20 f...ll of 2% as­ thetic pathway for limipterin was investigated. En­ corbic acid. The acidic supernatant was applied to zyme activity of GTP cyclohydrolase I, PTP syn­ f...LBondaPak C18 HPLC column using 20 mM po­ thase, and sepiapterin reductase were detected in tassium phosphate (pH 5.5), 2.5 mM EDTA as the 30-80% ammonium sulfate fractions. Each en­ solvent. zyme was partially purified from Chlorobium lim­

Preparation of Hrbiopterin from HflTP-Stan­ icola. The enzyme activity that produced H 4 - dard reaction mixture included 0.1 M Tris-HCI limipterin using H 4 -biopterin and UDP-N-acetyl­ buffer (pH 7 .4) containing 25 mM MgCI2 , 420 glucosamine as substrates was present in 40-60% 1-1,\1 H2NTP, 10 mM DTT, 1 mM NADPH, ammonium sulfate fractions. In order to de­ PTP synthase (Affi-Yellow fractions), and SR (Affi­ termine in vivo localization, each enzyme activity Brown fractions) in a total volume of 40 f...L1. Twen­ was examined in cell wall-membrane fraction and ty 1-11 of 5 N TCA and 100 f...ll of 1% 12/ 2% KI cytosolic fraction. As shown in Table 1, all the

'-olution were added to the mixture. After 60 min, biosynthetic enzyme actIvlUes of H 4 -limipterin the insoluble material was removed by cen­ and pteridine compounds were found in cytosolic trifugation at 13,000 x g for 10 min and excessive fractions. Limipterin synthase activity also was iodine was destroyed by the addition of 40 f...LI of found in the cytosol unlike the other well-known 2% ascorbic acid. The acidic supernatant was ap­ sugar transferases that are present in the mem­ plied to f...lBondaPak C18 HPLC column using 20 brane fraction (14,19,21). mM potassium phosphate buffer (pH 5 .5) con­ taining 2.5 mM EDTA or H 20 as solvent. Purification and biochemical characterization of In vitro synthesis of tetrahydrolimipterin from H z GTP cydohydrolase I NTP-The reaction mixture containing 0.1 M Tris-HCI buffer (pH 7 .4), 20 mM MgClz, 20 Majority of GTP cyclohydrolase actiVIty was de­ mM MnCI2 , 276 f...LM H 2 NTP, 1 mM UDP-N-a­ tected in 0-40% ammonium sulfate precipitates. cetylglucosamine, 10 mM DTT, 1 mM NADPH, Further purification was achieved by DEAE-sep­ PTP synthase (Alli-Yellow fractions), sepiapterin hadex and GTP-agarose chromatography. A SWTI-

Table 1. Biosynthetic enzyme activity and pteridine contents in the cytosolic and membrane i'i-actions of Chlorobium Enzyme activity (Unit) *Membrane fraction * Membrane fraction #Cytosolic (Detergent-soluble) (Suspension) Fraction GTP cyclohydrolase I 0 .03 0.08 1.45 PTP synthase ND 0 .03 0 .53 Sepiapterin reductase 0.14 1.07 11.2 Limipterin synthase ND 0 .11 36.0 pteridine content ND 0 .069 1.70 One Unit: one pmol per min/ mg protein Pteridine content: J.lmol!g dry weight ND: not-determined * Membrane fractions were obtained as precipitates after ultracentrifugation of ciude extract at 50,000 x g: tlJr -to min. Detergent soluble fractions were dissolved in 0 .1 ;\\ PIPES (pH 6 .8) containing 1% Triton XIOO and the tT.1-: tions were suspended in 0.1 M PIPES, pH 6.8. #Cytosolic fractions were prepared as supernatant .1fter <:enrrih.lgation at 50,000 x g for 40 min.

Ptcridines/ Vol. 9 / No. 2 Kang et al.: Biosynthetic enzymes of tetrahydrolimipterin 75

Table 2 . Summary of the purification of GTP cyclohydrolase I from Chlorobium limicola Total protein Total activity Specific activity Purification procedure Purification fold Yield (% ) (mg) (U) (Ujmg) Crude extract 2,450 3,552 l.45 1 100 Ammonium sulfate 210 674 3.21 2 .21 8 .5 DEAE-sephadex 53 472 8.92 6 .15 13.3 GTP-agarose 0.32 102 320 220 2 .9 One Unit: one pmol per min under standard reaction condition

mary of the purification is presented in Table 2. perature for the GTP CHase I of Chlorobium lim­ An overall purification of 220 fold was obtained icola were found to be pH 7.5-S.0 (Fig. 2A) and with the most effective step being GTP-agarose. 60°C (Fig. 2B). Unlike the E. coli and mammalian GTP cyclohy­ drolases, the Chlorobium enzyme did not bind Purification of PTP synthase to UTP-agarose. The dialyzed ammonium sulfate fractions were applied to a DEAE-Sephadex col­ PTP synthase was purified from 50-70% am­ umn equilibrated in advance with 20 mM Tris­ monium sulfate precipitate through Ultrogel AcA HCI (pH 8.0), 10 mM I)-mercaptoethanol. Active 44 and Reactive Yellow-agarose chromatography. fractions were eluted at 0 .6-0.7 M KCI. After di­ A total purification of 33 fold was achieved, as alysis, the sample was loaded to a column of GTP­ shown in Table 3 . The ammonium sulfate pre­ agar- ose pre-equilibrated with the Tris, I)-mer­ cipitate was dissolved in 100 mM PIPES (pH 6.S), captOethanol buffer containing 2.5 mM EDTA. 2 .5 mM DTT and applied to a column of UI­ The enzyme was eluted with 1 roM GTP in the trogel AcA 34. The fractions containing enzyme equilibration buffer. The optimum pH and tem- activity were pooled and loaded onto a Reactive-

120 B 120 A MES. 100 PIPES· 100 HEPES. TAPS y :t!:- ~ .;; 80 .2: 80 ~ ~ ..E ~ >- >- ::! 60 60 .. .,~ .2: (;j ~ 40 1& 40 a::C6 a::

20 20

0 01----.----.----.---.----~--_.----r_--~ 4 5 6 7 8 9 o 10 20 30 40 50 60 70 80 pH Temp (OC) Figure 2 . Effect of pH and temperature on the activity of GTP cyclohydrolase I from Chlorobium /imico/a. A) Effect of pH. The pH dependence was measured at the pH values indicated. 0.1 M MES butters at pH 5-7, 0.1 M PIPES buff­ ers at pH 6-7, 0.1 M phosphate buffers at pH 6-7, 0.1 M HEPES buffers at pH 7-7.4, and 0.1 M TAPS butters at pH 7.4-8.6. B) Effect of temperature. Reactions were carried out in 100 mM Tris-HCl buffer, pH 8.0, containing 100 j.lM GTP,5 mM EDTA, 2 mM dithiothreitol, and 0.05 M NaCl, at indicated temperature for 30 min.

Table 3 . Summary of the purification of PTP synthase from Chlorobium limicola Total protein Total activity Specific activity Purification proccdure Purification fold Yield (%) (mg) (U) (Ujmg) C rude extract 3,890 2,956 0 .76 1 100 Ammonium sulfate 681 946 1.39 1.8 32.0 Ultrogel AcA44 320 602 1.88 7 20.4 Reactive YeUow ll.5 293 25.4 33 9.9 One Unit: one pmol per min under standard reaction condition

Pteridines/ Vol. 9 / No. 2 b Kang eT at.: Biosynthetic enzymes of tctrahydrolimipterin

Table 4. Purification of sepiapterin reductase from Chlorobium limicola Total protein Total activity Specific activity Purification procedure Purification fold Yield (%) (mg) (U) (U/ mg) Crude extract 1,322 11,237 8.5 1 100 Ammonium sulfate 779 18,851 24.2 2.85 168 Ultrogel AcA 44 47.7 4,580 96.1 11.3 40.8 Reactive Brown 2.86 2,160 754 88.7 19.2 One Unit: one pmol per min under standard reaction condition

Table 5 . Summary of the purification of limipterin synthase from Chlorobium limicola Total protein Total activity Specific activity Purification procedure Purification tc)ld Yield (%) (mg) . (u) (U / mg) Crude extract 2,370 440 0.19 100 Ulracentrifugation 990 394 0.40 2.11 89.5 Ammonium sulEne 532 283 0 .53 2.79 64.4 Sephacryl S 1 00 136 213 1.56 8.21 48.4 MTX agarose 11.5 105 9.15 48.1 23.9 Phenyl sepharose 7.8 98 12.5 66.1 22.3 Mono-Q 1.6 75 46.6 245 17.0 Superdex-75 0 .06 41 686 3614 9 .4 One Unit: one nmol per min under standard reaction condition

Yellow column. The active fractions were eluted a Sephacryl S100 column which had been pre­ in a salt concentration of 0.4-0.5 M. viously equilibrated with 10 mM PIPES buffer (pH 6.8) containing 2.5 mM DTT. The active Purification of sepiapterin reductase fractions which were separated from major pro­ teins and pigments were pooled and loaded to a Sepiapterin reductase activity was detected main­ pre-activated Methotrexate-agarose. The enzyme h' at 30-50% ammonium sulfate precipitates. Fur­ activity appeared at the 0.2 M NaCI. (Fig. 3). ther purification was achieved by Ultrogel AcA 44 Mter the addition of solid NaCl to 1 M, the en­ and Reactive Brown-agarose chromatography. A zyme solution was applied to phenyl-sepharose. summary of the purification is presented in Table The enzyme was eluted in the flow-through frac- 4 . An overall purification of 89-fold was achieved. The ammonium sulfate precipitate was dissolved 200 ~------.- .6 in 100 mM PIPES bufier (pH 6.8) containing L E 2.5 mM DTT and applied to a column of 1M NaCI .5 Ultrogel AcA 44. The fractions containing en­ 150 zyme activity were collected and loaded to a Reac­ .4

tive-Brown column. The active fractions were 100 S :§ .3 co eluted at 0.5-0.6 M NaCl and separated from the :::> c:~ major protein peak. :§ .. 2 e I 50 c.. Purification of limipterin synthase . 1 a [ A summary of purification is shown in Table 5 . 0 .0 The enzyme was purified by ultracentrifugation, ammonium sulfate precipitation (40-60%), Sepha­ J cryl S100, methotrexate-agarose, phenyl-sephar­ 0 5 10 15 20 25 30 35 ose, mono-Q and Superdex 75 chromatography. Fraction No. By this procedure, the enzyme was purified to Figure 3. Affinity chromatography of limiptcrin synthase from Chlorobium limicola on methotrexate-agarose column. 3610-fold with an overall recovery of 9.4%. After Column dimensions: 0.8 X 10 cm, Vt=5 ml. The column dialysis against 20 mM PIPES buffer (pH 6.8) which had been equilibrated with 10 mM PIPES buffer containing 10% glycerol and 2 .5 mM DTT, the (pH 6.8) containing 2.5 mM DTT was washed with the n purified enzyme was stable for months at -80 C. same buffer and developed successively with 1 ) 0 .2 M The ammonium sulfate precipitate was applied to NaCl, and 2 ) a linear gradient of 0.2-1.0 M NaC!.

Pteridines/ Vol. 9 / No. 2 Kang et at.: Biosynthetic enzymes of tetrahydrolimipterin 77

30 1.0 50 mM PIPES buffer (pH 6 .8) contammg 2.5 E 1 M mM DTT using FPLC system. The activity ap­ 25 W 0.8 peared at the molecular weight range of about 0.8M/ 46,000 dalton (Fig. 5). Active fractions were col­ 20 0.6 M n 0.6 lected and stored at -80 C until used.

15 § 0 Characteristics of purified limipterin synthase co , 0.4 '"

-0.2 products, substrate specificity, enzyme kinetics, 0 10 20 30 40 50 60 etc. were investigated for this enzyme. Fraction No. :-:igure 4. Ion exchange chromatography of limipterin " :nthase on a Mono-Q column. The column (Vt=I 11'11 ) .35 .,·:'i(h had been equilibrated with 20 mM Tris-HCI (pH 7 . A 2 . 2 .5 mM DTT was washed with the same bufier and de­ .30 ~1 '~ red with a gradient of 0.0-1.0 M NaCI in a total ·.ume of 25 ml.

.. .25 tions. The Phenyl-Sepharose How-through frac­ I 1 tions were collected and dialyzed against 20 mM !!.. 0 T ris-HCI buffer (pH 7.4) containing 2.5 mM u:::> .20 DTT and applied to a mono-Q column. Eg. 4: shows that active fractions were eluted at 0 .6-0.7 .15 1\1 NaCl. Active enzyme solution was con­ centrated to a small volume (400 f..ll) using Cen­ tricon (cutoff MW=30,000 dalton) and applied .10 0 2 3 4 5 7 to Superdex 75. The column was developed with Time (min)

1 2 3 4 5 .50 14 .6 B 1 ~'V ~ 'V ..s 12 .5 ~ .40 10 '\ .35 '\ 4 , \ a g E 1 \ 00 .30 8 , N 2- :~ :'5. § f- .3 c: .z:.;; I \ ,\ 'v ! .25 6 , \ 1\ ~ ~ , \ , \ (L ~ I Li- .2 .20 4 I \ f- [;. 2 • I \ .15 I \ .1 2 , , .10 r 0.0 0 .05 0 2 3 4 5 6 7

10 15 20 25 30 35 40 45 50 Time (min) Fraction No. Figure 6. HPLC analysis of the reaction product of Figure 5. Gel permeation chromatography of limipterin limipterin synthase from Chlorobium limicola. Column: synthase on Superdex 75 (Vt=24 ml). The column was NovaPak CIS (2.5 mm x IS cm), mobile phase: 20 mM previously equilibrated with 50 mM PIPES, pH 6,8, 2.5 KH2 P04 (pH 5.5), 2 .5 mM EDTA. Flow rate: 1.0 1111/ mM DTT using a FPLC system. Mol. wt. standards: min. Standard pteridines: 1 ) biopterin 2) limipterin, A :

1 . Blue dextran 2,000 kd, 2 . Albumin from bovine H 4 -limipterin synthesized by limipterin synthase from H 4 - serum 68 kd, 3. Albumin from hen egg 45 kd, 4. Chy­ biopterin, B : TFA-hydrolyzed limipterin that was syn­ motrypsinogen A 25 kd, 5. Cytochrome C 12.5 kd. thesized in panel A.

Pteridines/ Vol. 9 / No. 2 78 Kang et al.: Riosynthetic enzymes of tetrahydrolimipterin

Reaction product The fully oxidized-form was not the substrate for this enzyme. We also studied the regulation of To confirm the catalytic aCtlVIty of limipterin the activity by substrate analogues such as UDP­ synthase, the enzymatic products were analyzed N-acetylgalactosamine, UDP-galactosamine etc. by reversed phase HPLC. The retention time of and other structurally related compounds. UDP the reaction product was the same as that of inhibited the enzyme activity to over 70% at the limipterin purified from Chlorohium. The en­ concentration of 300 I-J.M and other product analo­ zymatic product was hydrolyzed in cone. tri­ gues including UTP exhibited inhibitory efkcts fluoroacetic acid for 3 hours at room temperature on limipterin synthase (Table 7). UDP-N-acet:yl­ and analyzed again by HPLC. As shown in Fig. 6, galactosamine and UDP-galactose also had in- the hydrolytic product was biopterin that was a 08 substrate in the reaction mixture before it was IA treated with limipterin synthase. 07 J os J Substrate specificity and regulators ., g 05 ~ Limipterin synthase catalyzes the condensation ~ .04 i4: 2 of H 4 -biopterin with UDP-GlcNAc to form the pterin glycoside. We examined the ability of the .03 enzyme to use various pteridines, and other forms .02 of UDP-sugars. Table 6 shows that UDP-N-a­ .01 l cetylglucosamine was the only choice for the 10 sugar moiety whereas both dihydrobiopterin and Time (min) its tetrahydro-form could be used for the pterin. 30 B 1 25 Table 6 . Substrate specificity of limipterin synthase .20 Biopterin RH2 BH. PH. 6-MPH4 ~ § UDP-GlcNAc 57.5 100 .15 ~ UDP-GalNAc i4: UDP-Glc . 10 UDP-Xyl .05 UDP-GluA

Concentration of pteridines: 100 f.!M 0 .00 Concentration of UDP-sugars: 300 f.!M 10 Time (m.n) BH2 : dihydrobiopterin BH4 : tetrahydrobiopterin PH. : tetrahydropterin 6-MPH. : 30 6-methyltetrahydropterin C 1 UDP-GalNAc: UDP-N-acetylgalactosamine .25 UDP-Glc: UDP-glucose UDP-Xyl: UDP-xylose .20 UDP-GluA: UDP-glucuronic acid .15

Table 7 . Effects of limipterin synthase activity by variolls . 10 substrate analogues 2 05 Substrates Activity (% ) Substrates Activity (%) \Jl UDP-GlcNAc 100 UDP-GlaNAc 51.0 0 .00 UDP 21.3 UDP-Gal 63.8 0 10 UTP 50.0 UDP-Xyl 73.6 Time (min) ATP 86.7 UDP-GluA 87.0 Figure 7. Enzymatic s\'nthesis of H.-D-tepidopterin from

AMP 80.0 DUP-Glc 87.0 H 4 -D-threo-bioptcrin and UDP-N-acetylglucosamine by GTP 80.0 GlcNAc 87.0 limipterin svnthase trom Chlorobium limicola. Column:

GMP 80.0 GalNAc 100 f.!RondaPak. CIS. m l)hilc phase: 20 mM KH2 PO. (pH 5 . CDP 73.3 ManNAc 100 5 ), 2.5 m]\ 1 E D '1':\ . rl()\\' rate: 1.0 ml/min. A: pteridine Concentration of H.-biopterin: 100 f.!M standards I , d i,n'npcrin 2 ) tepidopterin, R: no enzyme Concentration of nucleotides and substrate analogues: control. c : r::~7\'l~l.lri, product from H.-D-threobiopterin 300 f.!M bv limiptcn:-: "''-:1:r.J<.C.

Pteridines/ Vol. 9 / No. 2 Kang et at.: Biosynthetic enzymes of tetrahydrolimipterin 79 hibitory effects on the enzyme . We examined if acetylglucosamine calculated irom double-re­ limipterin synthase could utilize 5,6,7,8-tetrahydro­ ciprocal (Lineweaver-Burk) plot were 42 ~M and D-threobiopterin in place of 5,6,7,8-tetrahydro-L­ 124 ~M respectively (Fig. 8). Considering a set erythrobiopterin as a substrate. As shown in Fig. 7, of intersecting lines patterns in the double-re­ the enzyme was able to produce D-tepidopterin, ciprocal plot, limipterin synthase is supposed to [l-O-(D-threo-biopterin-2 '-yI)-f3-N -acetyI-gIu­ form a ternary complex during the course of the cosamine], when 5,6,7 ,8-tetrahydro-D-threobiop­ reaction. terin and UDP-GlcNAc were used as substrates. Optimum pH and temperature Estimation of Km Fig. 9A shows that the activity was the highest Initial velocities of the enzyme reaction were at pH 8.0 and declined rapidly about pH 9. The measured with varying concentrations of tetrahy­ drobiopterin and UDP-N-acetylglucosamine. The K.n values for tetrahydrobiopterin and UDP-N- A 100

100 UDP-GlcNAce 10uM ~ 80 A .,-'=' ~ 60 80 4(

UDP-GIcNAe 20 uM '"> ~ ''0 ::- 60 a::'" 2- 20 >- .:; • :g « 40 ° 3 4 5 6 7 8 9 10 pH -- UDP-GleNAc 1 mM 1==

-40 -20 o 20 40 60 80 100 120 B BH. (mM-') 100

~ 80 Z"' B 150 ~ u 60 4(

'"> ~ 40 c; 0:: 100 20

0 BH.50uM 10 20 3C 40 50 60 50 Temperature (·e) BH4 lOOuM Figure 9. Effect of pH and temperature. A) Effect of pH. The pH dependence of the activity was measured at

BH~ lOOOuM the pH values indicated. 0.1 ~M acetate buffers were used at pH 4-5, 0.1 M MES buffers at pH 5-6, 0.1 M -100 -50 o 50 100 PIPES buffers at pH 6-7, 0.1 M phosphate buffers at UDP-N-AcetyIGlc (mM-') pH 6-7, 0.1 M HEPES butters at pH 7-8, and 0.1 M Figure 8. Initial velocity patterns of limipterin synthase Tris-Hel buffers at pH 8-9. B) Effect of temperature. with various concentrations of substrates. (A) with Reactions were carried out in 100 mM PIPES bufier tetrahydrobiopterin as a fixed amount of UDP-N-acetyl­ (pH 6.8), including 10 mM Mne!" 300 IJ.M UDP­ glucosamine, (B) with UDP-N- acetylglucosamine as a GIcNAc, 100 IJ.M BH4 , 10 mM DIT, and enzyme frac­ fixed tetrahydrobiopterin. tion at each temperature tor 30 min.

Pteridines/ Vol. 9 / No. 2 Kang et al.: Biosynthetic enzymes of tetrahydrolimipterin

Table 8. Effect of metal ions on limipterin synthase activity A .'-letal ion added (5 mM) Relative activity (%) 2 Mn + 100 Mg2+ 23 ~L l B Ca + 0.07 2 C0 + o 2 Cu + o 2 c Ni + o Zn2+ o 2 Fe + o EDTA o Cations were added in the form of chloride o activity was higher with organic buffers such as PIPES, HEPES and Tris-HCI than potassium Figure 11. SDS/ PAGE of limipterin synthase purified phosphate buffer. The optimum temperature for trom Chlorobium limicola. Electrophoresis was perfor­ the reaction was 3rC (Fig. 9B). med on 12.5% polyacrylamide gels and the bands were visualized by silver staining. Marker proteins: A) Bovine Effect of metal ions and dithiothreitol serum albumin (66,000), B) Ovalbumin (46,000), C) Carbonic anhydride (30,000), D) Lysozyme (14,300), L: The enzyme required manganese ions for max­ Limipterin synthase (46,000). imal activity, as shown in Table 8 . The mag­ nesium ion can replace the role of manganese ion was calculated to be 46,000 dalton by comparing to some extent. This result agreed to the fact that its mobility with those of standard proteins on a mammalian sugar transferases require manganese calibrated Superdex 75 column (Fig. 10). In SDS­ ions to bind with sugars. The activity, however, PAGE, the enzyme fraction generated no visible was inhibited by MnCl2 at the concentration high­ protein bands of lower molecular weight except er than 10 mM. Limipterin synthase was most ac­ one having apparent molecular weight of 45,000 tive in the presence of 10 mM DTT. dalton (Fig. 11). The results indicated that limip­ terin synthase of Chlorobium limicola most likely Molecular weight and subunit size exists as a monomer.

Analysis of biosynthetic intermediates of H 4 -limipte­ The molecular weight of limipterin synthase rin in vitro

.50 Identification of H~TP-It is known that all na­ turally occurring pteridines originate from GTP .45 D by the action of GTP cyclohydrolase 1. In order to identify the enzymatic product of GTP CHase .40 • I, the mixture of Chlorobium GTP CHase I reac­ C .35 tion was analyzed by fluorescence spectroscopy • and the spectrum was compared with the em­ ~~ .30 L ission spectra of authentic D-erythro-neopterin B (Fig. 12: A,B,C). As shown in Fig. 12: E, G, the .25 • I iodine oxidation product of Chlorobium GTP CH .20 I under the acidic condition was neopterin tri­ A phosphate, and the alkaline-phosphatase treated .15 • reaction product was neopterin. The results in­ dicate that the reaction product of the Chloro­ bium limicola GTP CHase I is D-erythro-dihy­ 4.0 4.1 4.2 4.3 4.4 4.5 4.6 4 .7 4.8 4.9 droneopterin triphosphate. Log MW PTP- Figure 10. Determination of molecular weight of limipte­ Detection of Using H lNTP as a substrate, rin synthase by Superdex 75 chromatography. Standard the reaction product of Chlorobium PTP synthase

proteins: A) albumin, bovine, B) albumin, egg, C I ch\­ was analyzed by HPLC, using the method of Ha­ motrypsin A, D) cytochrome C. L: limipterin synthase takeyama. Because 6-PTP is converted to pterin

Pteridines/Vol. 9 / No. 2 Kang et at.: Biosynthetic enzymes of tetrahydrolimipterin 81

. ~r8---- - ' A ~ c A 2

40' 0le ·.... 1....

'. !z ~~';: :-"-- ~'i ~~ r;. ..,'" '----7------,~;-. ~~­ Wavelength Wavelength 8 c 1:

~ """'-----

r_ [min) .. D I" ~ 2

" 1 i o , 2 1 .. , • , • rome (",",) ' ...... (""'"') figure 12. Fluorescence analyses of the reaction product Figure 13. Enzymatic conversion of H ,~TP to PTP lw of Chlorobium and E. coli GTP cyclohydrolase 1. A: em­ PTP synthase purified from Chlorobium limicola in vitro. ission spectra of standard neopterin, B: emission spectra Column: !lBondaPak C IS, mobile phase: 20 mM KH, of reaction product by E. coli GTP CHase I, C: emission P04 (pH 5.5), 2.5 mM EDTA, How rate: 1.0 mljmin. spectra of reaction product by Chlorobium GTP CHase I, A: pteridine standards 1) neopterin 2) pterin, B: no en­ 0: HPLC analysis of 0 min control, E: HPLC analysis zyme control, C: no MgC\" D: reaction product by Dro­ of reaction product of Chlorobium GTP CHase 1, F: sophila PTP synthase, E : reaction product by Chlorobium HPLC analysis of pteridine standards I ) neopterin tri­ PTP synthase. phosphate 2) neopterin 3 ) biopterin 4) pterin, G : HPLC analysis of the reaction mixture upon alkaline phos­ phatase treatment of E fraction. Column: I-lBondaPak C When all biosynthetic enzymes and cofactors 18, mobile phase: 2% MtOH, flow rate: 1.0 mljrnin. such as GTP CHase I, PTPS, SR and limipterin synthase, two metal ions (Mn2+, Mg2+), an­ In acidic iodine oxidation condition, the reaction tioxidant (DTT) and NADPH existed in the reac­ product of Chlorobium PTP synthase was iden­ tion mixture, the final product, limipterin could tified as pterin. As shown in Fig. 13, this product be synthesized efficiently upon incubation at 3TC. was the same as the one that was produced by (Fig. 14, D). The intermediates as well as the fi­ Drosophila PTP synthase (purified from Dro­ nal product, limipterin, could also be identified us­ sophila melanogaster by the method of Park et al.). ing thin layer chromatography (Fig. 15). The The enzyme absolutely requires magnesium ions results of the TLC analysis of the biosynthetic in­ for its activity. termediates of limipterin are summarized in Table 9 . Biosynthesis of Tetrahydrobiopterin- When H 2 NTP was incubated with Chlorobium PTP syn­ Discussion thase and Chlorobium sepiapterin reductase in the

presence of MgCl2 and NADPH, biopterin was Little information is available concerning the found to be one of the products after acidic biosynthesis of unconjugated pterins in microor­ iodine oxidation (Fig. 14, C) ganisms. It was reported that various microor­ ganisms, especially a number of photosynthetic Biosynthesis of Hr limipterin in vitro bacteria contain relatively high amount of pterin

Pteridines/ Vol. 9 / No. 2 82 Kang et al.: Biosynthetic enzymes of tetrahydrolimipterin

4 A B

02 1 •

~ .el1

Time (mir.) Tkne (min)

c o 4 l

Qt. 3 .. ,: rllnfl(min) Time (min) Figure 15. Thin layer chromatography (Cellulose-plate) Figure 14. Enzymatic synthesis of H.-limipterin from H 2 of biosynthetic intermediates of H .-Iimipterin using pu­ NTP by PTP synthase, sepiapterin reductase and limipte­ rified GTP CHase I , PTPS, SR and LS from Chlorobium rin synthase from Chlorobium limicola in vitro. Column: limicola. Lane A : standard pteridines 1 ) neopterin (R,= J..lBondaPak C18, mobile phase: 20 mM KH2PO. (pH 5. 0.39),2) pterin (0.41), 3) biopterin (0 .52), 4) limipterin 5), 2.5 mM EDTA, flow rate: 1.0 ml/ min. A : pteridine (0.57), Lane B: reaction product from GTP b y GTP standards 1) carboxy pterin 2) neopterin 3) pterin 4) C Hase (R.=0.21), Lane C : reaction product from H 2 limipterin, B : no PTPS, C : reaction product by Chloro­ NT!' by PTPS (R,=0.41), Lane D : alkaline phosphatase bium PTPS and SR for 60 min, D : reaction product by treatment of Lane B product (R,=O .38), Lane E : reaction Chlorobium PTPS, SR and limipterin synthase for 120 min. product from H 2NTP by PTPS and SR (R,=0.39, 0 .51 ), Lane F : No MnCl2 control of limipterin biosynthetic

reaction from H 2 NTP (R,=0.39, 0 .51), Lane G : reaction compounds (6). Recently, a new pterin com­ product from H 2NTP by PTPS, SR and Iimipterin syn­ pound named limipterin was isolated from green thase (R,=0.56). sulfur bacteria, Chlorobium limicola (4). The struc­ ture of limipterin was established to be [1-0-(L­ thase is a monomeric protein with a molecular erythro-biopterin-2' -yl)-I3-N-acetyl glucosamine] weight of about 46,000 dalton. from IH-NMR, CD-spectra as well as from vari­ It was demonstrated that L-threobiopterin gly­ ous mass spectrometric techniques and chemical coside of N -acetylglucosamine is present in Chloro­ cleavage method. bium tepidium in relatively large amount (2-5 The pathway leading to the biosynthsis of /-Lmol per gr dry weight). The compound, [1-0- limipterin in Chlorobium limicola was first pro­ (L-threo-biopterin-2 '-yl)-I3-N -acetyl glucosamine], posed by Kang and others (9). It was shown that was designated as tepidopterin after the species Chlorobium, like the case of mammalian systems, name of the Chlorobium genus (5). Limipterin uses the same set of enzymes, i.e, GTP cy­ synthase, the final enzymes of the H+-limipterin clohydrolase I, 6-PTP synthase and sepiapterin synthesis could produce H 4-tepidopterin from D­ reductase for the synthesis of biopterin part of the threo-5,6,7,8-tetrahydrobiopterin and UDP-N-a­ pterin glycoside. Biosynthetic intermediates of the cetylglucosamine. The epimerization can occur pathway such as Hz-NTP, 6-PTP, and H 4 -limipte­ either at the biopterin level or at the pterin glyco­ rin were identified by HPLC and TLC methods. side level. We are currently investigating the ep­ The enzyme that condenses H4 -biopterin with N­ imerization enzymes that convert H4 -L-erythro­ acetylglucosamine was purified and some of the biopterin to its threo form, H 4-L-threobiopterin properties were examined. The enzyme uses H 4- ( dictyopterin), or H4 -limipterin to H 4- tepidop­ biopterin and UDP-N-acetylglucosamine as sub­ terin. A plausible pathway for the enzymatic syn­ strate and requires manganese ions and a reduc­ thesis of H 4 -limipterin and H 4-tepidopterin is dep­ ing agent, dithiothreitol to exhibit maximal ac­ icted in Fig 16. tivity. It is interesting to see that some other Although tetrahydro form of limiptcrin can be sugar transferases <.. ;0 prefer manganese ion rath­ used efficiently as a cofactor for mammalian aro­ er than magnesiUIJl for catalysis. Limipterin syn- matic amino acid hydroxylases (9), the biological

Pteridines/ Vol. 9 / No. 2 Kang et al.: Biosynthetic enzymes o f tetrahydrolimipterin 83

T able 9 . T L C analysis of biosynthetic intermediates of H. -limipterin Developing solvent ( Rt) Samples A B C Neopterin 0 .39 0.51 0.22 Biopterin 0 .52 0.47 0.44 Pterin 0 .41 0 .41 0 .34 Limipterin 0.57 0.62 0.38 Reaction product from GTP by Chlorobium limicola GTP CHase I 0 .21 0 .83 0.19

Reaction product from H 2 N TP by Chlorobium limicola PTPS 0 .41 0.40 0 .32 Alkaline phosphatase treatment of Lane B product 0 .38 0.51 0 .21

Reaction product from H 2NTP by Chlorobium PTPS and SR 0.39 0 .40 0.33 0 .51 0.48 0.44 ~o MnC b control o f limipterin biosynthetic reaction from H zNTP 0.39 0 .41 0.32 0 .51 0.48 0.44 Reaction product from H zNTP by Chlorobium limicola PTPS, SR and 0 .56 0 .63 0.38 Ii mipterin synthase Solvent A : isopropanol-2% amlTIonium acetate :;olvent B: 4 % sodium citrate Solvent C : iso propanol: water=7:3 function of the biopterin glycoside in Chlorobium high concentration led u s to suspect that it might ~ ::,ecies is not clear. The in vivo co ncentration of h ave another impo rtant roles in the pho­ .::Tlipterin was estimated to be 1 .85 J..lmol/g dry tosynthetic species. We are investigating whe ther weight which is at least 10 times higher than the H 4 -limipterin is involved in photoreception, pho ­ biopterin content in pineal gland of rats. The totransduction, re dox process in photosynthetic dark reactions or in the biosynthesis of cell wall component of the Chlorobium species. GTP Acknowledgment I GTP cyolohy","ol". I This work was supported in part b y the Korean

7,8-Dihydroneopterin triphosphate (HzNTP) Ministry of Health and Welfare, Project No. HMP-96-D-4-1038 to J.Y. I PTP 'Ynth". References 6-Pyruvoyltetrahydr9pterin (PTP) 1 . Ahn, C., Byun, C. and Yim, J. Purification, clo ning, and functio n al expressio n of dihydro neopterin tri­ S.piapt""" reOo",,,. / \ S.piapterin ,.doc"". phosphate 2' -epimerase from E scherichia coli. J. Bio I. C hem. 1997; 272: 15323-15328. 2 . Bradford, M.M. A rapid and sensitive method fo r 5,6,7,8-Tetrahydrobiopterin _ 5,6,7,8-Tetrahydrodictyopterin the quantitation of microgram quantities o f protein (H4- L-erythrobiopterin) (H.-L-threobiopterin) utilizing the principle of protein-dye binding. Anal. ---( ? ) Biochem. 1976; 7 2 : 248-254. 3. Cha, K., Jacobson , W . and Yim, JI Iso latio n and characterization o f GTP cyclo hydrolase T trom Limipterin Limipterin mouse liver: Compariso n of normal and t he hph- l synthase synthase mutant, J. BioI. Chern. 1991; 266: 12294-12300. 4. C ha, K.W., Pt1eidere r, W. and Yim, J. Iso latio n and characterization of limipte rin .( 1 -O-(L-e rythro ­ biopterin-2'-yl)-13-N-acetylglucosamine) and its 5,6,7, 5,6,7,8-Tetrahydrolimipterin ____ 5,6,7,8-Tetrahydrotepidopterin 8-tetrahydro d erivative trom green sulfur bacterium (H4-L-erythrobiopterin --- (H.-L- threobiopterin C hlorobium limicola F. thiosulfatophilum NCIB 8327. glycoside) ( ? ) glycoside) Helv. C him. Acta. 1995; 7 8 : 600-614. Figure 16. Plausible pathway for the biosynthesis o f tetra­ 5. C ho, S.H ., Na, J.D., Youn, H. and Kang s.o. A n ov­ hydrolimipterin and tetrahydrotepidopterin from GTP in el tepidopterin of a green sulfur bacterium Chloro ­ Chlorobium limicola. bium tepidum. Vol 2, p . 859-862. In Photosynthesis:

Pteridines/ Vol. 9 / No. 2 84 Kang et al.: Biosynthetic enzymes of tetrahydrolimipterin

from light to biosphere P. Mathis (ed.), Academic chern. Biophys. Res. Commun. 1991; 175: 738-7 44. Publishers. Netherlands 1995. 16. Park, Y.S., Kim, J.B., Jacobson, K.B. and Yim, J.J. 6. Forrest, H.S. and van Baalen C. Microbiology of un­ Purification and characterization of 6-pyruvoylte­ conjugated pteridines. Ann. Rev. Microbiol. 1970; trahydropterin synthase from Drosophila m elano­ 24: 91-108. gaster. Biochim. Biophys. Acta, 1990; 1038: 186- 7 . Hatfield, D .L., van Baalen C. and Forrest H .S. Pter­ 194. idines in blue green algae. Plant. Physiol. 1961 ; 36: 17. Park, Y.S., Kim, N., Kim, H ., Park, D. and Yim, J. 240-243. Expression and characterization of recombinant Dro­ 8. Iwai, K., Okinaka, 0 ., Ikeda, M. and Suzuki, N . sophila 6-pyruvoyltetrahydropterin synthase. Pter­ Biosynthesis of folic acid compounds in plants; In idines 1995; 6(2): 58-62. Chemistry and Biology of Pteridines. p . 281 -293. 18. Pfennig, N . and Truper, H.G. An oxygenic pho­ Iwai K, Akino M, Goto M and Iwanami Y, (eds.) tosynthetic bacteria, In Bergey's manual of systematic International Academic Printing Co. 1970. bacteriology. Vol. 3, p. 1635-1709. Baltimore: Wil­ 9 . Kang, D ., Kim, S. and Yim, J. Purification and liams and Wilkins 1984. characterization of enzymes in the biosynthesis of 19. Preuss, U. and Ro bert, K.Y. Purification and charact­ tetrahydrolimipterin Pteridines 1995; 6(3): 93-96. erization of C MP-N-acetylneuraminic acid: Lac­ 10. Maier, J. and Ninnemann H. Biosynthesis of pter­ tosylceramide (a2-3 ) sialyltransferase (GM3-synthase) idines in Neurospora crassa, Phycomyces blakesleea­ from rat brain. J. BioI. Chem. 1993; 268: 26273- nus and Euglena gracilis: Detection and charact­ 26278. erization of biosynthetic e nzymes. Photochem. Pho­ 20. Saizieu, A ., Vankan, P . and Loon. A.P. Enzymatic to bioI. 1995; 61(1): 43-53. characterization of Bacillus subtilis GTP cyclo­ 11 . Masada, M., Akino, M., Sueoka, T . and Katoh, S. hydrolase 1. Biochem J. 1995; 306: 371 -377 . Dyspropterin, an intermediate formed from dihy­ 21. Shoreibah, M.G., Hindsgaul, O. and Pierce, M. Pu­ droneopterin triphosphate in the biosynthetic path­ rification and characterization of rat kidney UDP-N­ way of tetrahydrobiopterin. Biochim. Biophys. Acta. acetylglucosamine a-D-mannoside- l ,6-N -acetyl-glu­ 1985; 840: 235-244. cosaminyltransferase. J. BioI. Chern. 1992; 267: 12. Milstein, S. and Kaufman, S. The biosynthesis of 2920-2927. tetrahydrobiopterin in rat brain. J. BioI. Chern. 1989; 22. Sueoka, T. and Katoh, S. Purification and charact­ 264(14): 8066-8073. erization of sepiapterin reductase from rat erythro­ 13. Nichol, C.A., Smith, G .K. and Ouch, D.S. Biosyn­ cyte. Biochim. Biophys. Acta. 1981; 717: 265-271. thesis and metabolism of tetrahydrobiopterin and 23. Wray, W ., Bo ulikas, T ., Wray, V.P. and Hancock, B. molybdopterin. Annu. Rev. Biochem. 1985; 54: 729- Silver staining of proteins in polyacrylamide gels. 764. Anal. Biochem. 1981; 118: 197. 14. Park, Y.S. and Kent, C. Expression, purification, and 24. Yamamoto, H ., Hanaya, T., Harada, K., Kawamoto, characterization of CTP: Glycerol-3-phosphate cyti­ H. and Pfleiderer, W . A efficient synthesis of limipte­ ddtransterase from Bacillus subtilis., J. BioI. Chern. rin. Pteridines 1996; 7: 110-112. 1993; 268: 16648-16654. 25. Yim, J.J. and Brown, G .M. Characteristics of guano­ 1:'i. Park, Y.S ., Heizmann, C.W., Wermuth, B., Levine, sine triphosphate cyclohydrolase I purified from R.A., Steinerstauch, P . J. Guzman and N. Blau. Hu­ Escherichia coli. 1. BioI. Chern. 1976; 251: 5087- man carbonyl and aldose reductase-new catalytic 5094. functions in tetrahydrobiopterin biosynthesis. Bio-

Pteridines/ Vol. 9 / No. 2