THE JOURNAL OF Bm.nCrcfi CHEMISTRY Vnl. No. Issue June … · THE JOURNAL OF Bm.nCrcfi CHEMISTRY 0 1994 by The American Society for Biochemistry and Molecular Biology, Inc. Vnl. 269,

THE JOURNAL OF Bm.nCrcfi CHEMISTRY 0 1994 by The American Society for Biochemistry and Molecular Biology, Inc.

Vnl. 269, No. 24, Issue of June 24, pp. 16726-16732, 1994 Printed in U.S.A.

Indolepyruvate Ferredoxin Oxidoreductase from the Hyperthermophilic Archaeon Pyrococcus firiosus A NEW ENZYME INVOLVED IN PEPTIDE FERMENTATION*

(Received for publication, February 7, 1994, and in revised form, April 15, 1994)

Xuhong Mai and Michael W. W. AdamsS From the Department of Biochemistry and the Center for Metalloenzyme Studies, University of Georgia, Athens, Georgia 30602

Pyrococcus furiosus is a strictly anaerobic archaeon that grows optimally at 100 "C by a fermentative-type metabolism in which complex peptide mixtures such as yeast extract and Tryptone, and also certain sugars, are oxidized to organic acids, H, and CO,. Enzymes involved in the utilization of peptides such as proteases, aromatic amino transferases, and glutamate dehydrogenase have been previously purified from this organism. It is shown here that E? furiosus also contains significant cytoplasmic concentrations of a new enzyme termed indolepyruvate ferredoxin oxidoreductase (IOR). This catalyzes the oxidative decarboxylation of aryl pyruvates, which are generated by the transamination of aromatic amino acids, to the corresponding aryl acetyl-coA. IOR is a tetramer (a2&) of two identical subunits (66,000 and 23,000 Da) with a molecular weight of 180,000. The enzyme contains one molecule of thiamine pyrophosphate and four [4Fe-4SI2+~'+ and one [3Fe-4Sl0~'+ cluster, as determined by iron analyses and EPR spectroscopy. Sig- nificant amounts of other metals such as copper and zinc were not detected. IOR was virtually inactive at 25 "C and exhibited optimal activity above 90 "C (at pH 8.0) and at pH 8.5-10.5 (at 80 "C). The enzyme was sensitive to inactivation by 0,, losing 50% of its activity after exposure to air for 20 min at 23 "C, and was quite ther- mostable, with a half-life of activity at 80 "C (under anaerobic conditions) of about 80 min. The K,,, values (in p ~ ) for indolepyruvate, p-hydroxyphenylpyruvate, phenylpyruvate, CoASH, and €? furiosus ferredoxin, the physiological electron carrier, were 250, 110,90, 17, and 48, respectively. IOR was inhibited by KCN (apparent Ki = 7.5 mM), but not by CO (1 atm). An enzyme analogous to IOR has not been reported previously. Curiously, it has few properties in common with the pyruvate ferredoxin oxidoreductase of €? furiosus, even though the two enzymes catalyze virtually identical reactions. In fact, of known ketoacid oxidoreductases, the catalytic mechanism of IOR appears to be most similar to that of the pyruvate ferrodoxin oxidoreductase from the hyperthermophilic bacterium Thermotoga maritima.

Microorganisms have been isolated recently from shallow and deep sea volcanic environments that have the remarkable

* This research was supported by Grant N00014-90-5-1894 from the Office of Naval Research and Grant FG09-88ER13901 from the Depart- ment of Energy. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "aduertisernent" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

$ To whom correspondence should be addressed: Dept. of Biochemis- try, Life Sciences Bldg., University of Georgia, Athens, GA 30602. Tel.: 706-542-2060; Fax: 706-542-0229; E-mail: [email protected].

property of growing at temperatures near and even above 100 "C (Stetter, 1982, 1986; Kelly and Deming, 1988; Adams, 1990; Stetteret al., 1990;Adams and Kelly, 1992;Adams, 1993). Virtually all of these so-called hyperthermophiles are classified as archaea (formerly archaebacteria: Woese et al., 1990). They include some methanogenic and sulfate-reducing species, but the majority are "sulfur-dependent" organisms that reduce el- emental sulfur (So) t o H,S. Most of the So-dependent hyperthermophiles are strictly anaerobic heterotrophs that utilize peptides and proteins as their sole sources of carbon and nitro- gen. Some are also able to utilize certain carbohydrates as a carbon source, and an unusual pathway for carbohydrate metabolism has been proposed (Mukund and Adams, 1991; Scha- fer and Schonheit, 1992; Adams, 1994). The pathways of peptide utilization are less understood. Several hyperthermophilic archaea have been shown to contain high intracellular proteolytic activities (Adams, 1993; Eggen et al., 1990; Blumentals et al., 1990; Snowden et al., 1992; Connaris et al., 1991) together with high concentrations of glutamate dehydrogenase (Con- salvi et al., 1991; Robb et al., 1992; Ohshima and Nishida, 1993; DiRuggiero et al., 1993; Ma et al., 1994) and pyruvate ferredoxin oxidoreductase (POR)' (Blarney and Adams, 1994). A novel tungsten-containing formaldehyde-oxidizing enzyme has also been proposed to play a role in peptide metabolism (Mukund and Adams, 1993). However, the specific pathways by which amino acids are metabolized by hyperthermophiles are not known.

We have focused on the metabolism of aromatic amino acids by the hyperthermophilic archaea and recently reported (Andreotti et al., 1994)' on the characterization of two distinct aromatic aminotransferases from Pyrococcus furiosus (Fiala and Stetter, 1986) and Thermococcus litoralis (Neuner et al., 1990). Using 2-ketoglutarate as the amino acceptor, these enzymes catalyze the transamination of the aromatic acids to generate glutamate and the corresponding 2-keto acid. These enzymes show virtually no activity toward aspartate, alanine, valine, or isoleucine. For tryptophan, phenylalanine, and tyrosine, the products of these transamination reactions are indolepyruvate, phenylpyruvate, andp-hydroxyphenylpyruvate, respectively. The further metabolism of these keto acids is the subject of this paper. It has been shown previously that some mesophilic anaerobes, such as Clostridium bifermentans and Clostridium sticklandii (Elsden et al., 19761, are able to use aromatic amino acids as a sole carbon source (Fujioka et al.,

tase; IOR, indolepyruvate ferredoxin oxidoreductase; TPP, thiamine py- The abbreviations used are: POR, pyruvate ferredoxin oxidoreduc-

rophosphate; EPPS, N-(2-hydroxyethyl)piperazine-N"3-propanesulfonic acid; CAPS, 3-(cyclohexylamino)-l-propanesulfonic acid; MES, 2-(N-morpholino)ethanesulfonic acid; D m , dithiothreitol.

G. Andreotti, M. V. Cubellis, G. Nitti, G. Sannia, X. Mai, M. W. W. Adams, and G. Marino, submitted for publication.

16726

P. furiosus Indolepyruvate Ferredoxin Oxidoreductase 16727

NAD(P) NAD(P)H

Ar.CH2.CH0 A Ar.CH2.COOH cy (3) MESOPHILIC BACTERIA

Ar.CH2.CH(NH2).COOH 2-Kx Ar.CH,.CO.COOH ( 1 )

Fd,, + CoASH

Ar.CH,.CO.SCoA . - - + Ar.CO.COOH.--*Ar.CO.SCoA ? ?

(5) (6)

HYPERTHERMOPHILIC ARCHAEA SCHEME 1. Proposed pathways of aryl pyruvate oxidation in mesophilic bacteria and hyperthermophilic archaea. The enzymes

involved are: 1, aromatic aminotransferases; 2, indolepyruvate decarboxylase; 3, arylacetaldehyde dehydrogenase; 4, indolepyruvate ferredoxin oxidoreductase; 5, hypothetical arylacetyl CoA oxidoreductase; and 6, hypothetical arylglyoxylate oxidoreductase. Fd,, and Fd,,, are the oxidized and reduced forms of ferredoxin. See text for details.

1970; Barker, 1981). These organisms convert the aryl 2-keto acids generated by transamination reactions to the corresponding aryl acetates either by direct oxidative decarboxylation or by decarboxylation to the corresponding aldehyde followed by its oxidation. Thus, as shown in Scheme 1, mesophilic anaerobes contain aromatic aminotransferases, decarboxylases, and dehydrogenases, which convert phenylalanine, tyrosine, and tryptophan to their corresponding aryl acetates in NAD-dependent reactions. However, very little information is available on the enzymes that metabolize aryl pyruvates, in fact, little is known about similar enzymes from aerobic bacteria. A phenylpyruvate decarboxylase has been purified from the phenylalanine-utilizing aerobe, Achromobacter eurydice (Fujioka et al., 1970), and indolepyruvate decarboxylase was recently characterized from Enterobacter cloacae (Koga et al., 19921, an enzyme which is apparently involved in the synthesis of indole- 3-acetic acid.

Herein we report on the purification and characterization from the hyperthermophile, r! furiosus, of a new enzyme involved in the metabolism of aromatic amino acids. Termed indolepyruvate ferredoxin oxidoreductase or IOR, it catalyzes the CoASH-dependent conversion of indolepyruvate, phenylpyruvate, orp-hydroxyphenyl pyruvate to indoleacetyl-CoA, phenyl- acetyl-coA, or p-hydroxyphenylacetyl-CoA. Thus, J? furiosus differs from mesophilic bacteria in directly generating the activated keto acid from the transaminated aromatic amino acid. Moreover, it uses the redox protein ferredoxin (Aono et al., 1989) as the electron acceptor rather than nicotinamide nucle- otides. The relationship between IOR and I? furiosus POR, a copper-iron-sulfur-containing enzyme which catalyzes a virtually identical reaction (Smith et al., 19941, is discussed.

EXPERIMENTAL PROCEDURES Growth of Organism--I! furiosus (DSM 3638) was grown in a 600-

liter fermentor using maltose as a carbon source as described previously (Bryant and Adams, 1989).

Enzyme and Protein Assays-IOR activity was routinely determined spectrophotometrically using serum-stopped cuvettes under argon a t 80 "C (Bryant and Adams, 1989) by the indolepyruvate-dependent reduction of methyl viologen. The standard assay mixture (2.0 ml) contained indolepyruvate (5 mM), MgCl, (2.5 mM), TPP (0.4 mM), coenzyme A (CoASH: 0.1 mM), and methyl viologen (1 mM) in 50 mM EPPS buffer, pH 8.4. Absorbance changes at 600 nm were measured using a DMS 200 spectrophotometer (Varian Associates) equipped with a thermostatted cuvette holder and a thermoinsulated cell compartment. Results are expressed as units/mg of protein where 1 unit equals the oxidation of 1 pmol of indolepyruvate/min. The activity of IOR with other 2-keto acids was carried out under the same conditions, except indolepyruvate was replaced as indicated. -I! furiosus POR was purified and its activity was determined as described previously (Blarney and Adams, 1993). Protein

concentrations were routinely estimated by the Lowry method using bovine serum albumin as standard (Lowry et al., 1951). The protein content of samples of pure IOR was also determined by the quantitative recovery of amino acids from compositional analyses (see below). Amounts of protein in all samples were 95 f 5% of those measured by the colorimetric protein assay (from three separate determinations). All analytical values for the pure protein that were based on the Lowry method have therefore been corrected by a factor of 0.95.

Purification oflOR-The enzyme was routinely purified from 500 g (wet weight) of frozen cells. All procedures were carried out a t 23 "C under strictly anaerobic conditions (see Bryant and Adams, 1989). All buffers were repeatedly degassed and flushed with argon, prepurified with heated BASF catalyst (BASF, Kontes, NJ) and were maintained under a positive pressure of argon. All buffers also contained sodium dithionite (2 mM) to protect against trace 0, contamination. The buffer used throughout the purification was 50 mM Tris/HCI, pH 8.0, containing glycerol (lo%, v/v), sodium dithionite (2 mM), and dithiothreitol (DTT 2 mM). All chromatography columns were run using a fast protein liquid chromatography system (Pharmacia Biotech Inc.). The procedures for lysing the cells and preparing the cell-free extract were the same as for the purification of I? furiosus hydrogenase (Bryant and Adams, 1989), as was the first chromatography step which used a column (7.5 x 21 cm) of DEAE-Sepharose Fast Flow (Pharmacia). IOR activity eluted at 170-198 mM NaCl using a gradient (9 liters) from 0 to 500 mM NaCl in buffer. Fractions (100 mlJ from this column with IOR activity above 1.5 unitdmg were combined (600 ml) and loaded directly onto a column (8 x 30 cm) of hydroxylapatite (high resolution, Behring Diagnostics) equilibrated with buffer. The absorbed proteins were eluted with a gradient (1.2 liters) from 0 t o 0.2 M potassium phosphate in the same buffer at a flow rate at 3 mumin. Fractions of 50 ml were collected and IOR activity eluted as 0.17-0.20 M phosphate was applied. Fractions with IOR activity above 2.5 unitdmg were combined (200 ml) and concentrated to approximately 25 ml by ultrafiltration (Amicon PM-30).

The concentrated sample of IOR from the previous step was applied to a column (6 x 60 cm) of Superdex 200 (Pharmacia) equilibrated at 5 mumin with buffer containing 200 mM NaCI. Fractions (25 mlf with IOR activity above 9 unitdmg were combined (200 ml), diluted with an equal volume ofbuffer containing ammonium sulfate (2.0 M), and were applied to a column (3.5 x 10 cm) of phenyl-Sepharose (Pharmacia) previously equilibrated with buffer containing ammonium sulfate (1.0 M) at 2 mL' min. The absorbed protein was eluted with a decreasing gradient (2 liters) of ammonium sulfate (1.0 M to 0). IOR activity was detected in the eluent as 300 mM ammonium sulfate was applied. Fractions (15 ml) with IOR activity above 30 unitdmg were separately analyzed by nondenaturing and SDS-gel electrophoresis. Those judged pure were combined (75 ml), concentrated by ultrafiltration to approximately 8 mg/ml, and were stored as pellets under liquid N,.

NH,-terminal Sequencing and Amino Acid Analyses-The NH,-terminal sequences of the subunits of IOR were determined using an Applied Biosystems model 477 sequencer. The subunits of the enzyme were separated by SDS-gel electrophoresis and electroblotted onto PVDP protein-sequencing membranes using a Bio-Rad electroblotting system. Electroblotting was carried out in 10 mM CAPS buffer, pH 11.0, containing methanol (1070, v/v) for 1 h at 50 V. Amino acid analyses were

16728 l? furiosus Indolepyruvate Ferredoxin Oxidoreductase

carried out on a n Applied Riosystems model 4240A analyzer after the hydrolysis of IOR samples under argon at 165 "C for 1 h in the presence o f 6 \I HCI, phenol (lei, wh l , and thioglycolic acid (874, w/v). Serine and threonine were corrected for destruction. TOR apoprotein was prepared and reduced with WIT using the methods described for P: fitriosrrs ferredoxin (Aono et nl., 1989). Tryptophan was determined from the A,,,, value of the apoprotein after correction for tyrosine (Edelhoch, 19671. The cysteine content of the reduced apoprotein was estimated by the reaction with 5,5'-dithiohis(2-nitrobenzoic acid) (Riddles 6.t al., 1983).

Moleculnr "eight D~tcrt~1ir1otio17.s-Electrophoresis in the presence and absence of SIIS using 6,12. or 20'7 (wivl acrylamide was performed as descrihrd previously (Weher e / al., 19721. Samples were prepared by hea t ing a t 95 "C for 30 min in the presence of SDS (2.5%, w/v) and mercaptoethanol (1'7, vh]. Subunit molecular weights were estimated hy comparison with the following standard proteins (their molecular weights are given in parentheses): (3-amylase (200,0001, (3-galactosidase (116,0001, phosphorylase h (97,400), bovine serum alhumin (66,2001, ovalhumin (45,0001, and carbonic anhydrase (29.0001. The molecular weight of the holoenzyme was estimated by gel filtration using a column (2.6 x 100 cm) of Superdex 200. The proteins used to calibrate the column (and their molecular weights) were ferritin (450,000), catalase (240,000), bovine serum albumin (66,2001. ovalbumin (45,0001, and myoglobin (16,8001. The column was run at 1.5 ml/min using 50 m\l TrisMCI huffrr, pH 8.0, containing NaCl (200 m\ll as the e luent .

Metnl. Sulfide. T h i n r n i ~ ~ ~ ~ v r o ~ ~ h o . s p h a / ~ . nnd Spectroscopic Analwes-The iron and acid-labile sulfide content of IOR were measured using o-phenanthrnline (Lovenberget al., 19631 and hy methylene blue formation (Chen and Mortenson, 1977), respectively. A complete metal analysis (forty elements) was carried out by plasma emission spectroscopy using a Jarrel Ash Plasma Comp 750 instrument within the Department of Ecology of the University of Georgia. The amount of TPP in IOR was determined by fluorescence spectroscopy (Penttinen, 1979). Thionine-oxidized samples of IOR in the presence and absence of suhstrates were prepared as described previously for I? furiosrrs POR (Smith et nl., 19941 except that indolepyruvate was used as the substrate instead of pyruvate. EPR spectra were recorded on an IRM- Rruker ER2OOD spectrometer interfaced to an IBM 9001 microcom- puter and equipped with an Oxford Instruments ESR-9 flow cryostat. Spin quantitations were determined by double integration of spectra recorded at 8 K using 20-UW microwave power. These were compared with spectra of copper (1 m\l)/EDTA(100 m\ll recorded under the same conditions.

RESULTS

Purification of IOR-Significant amounts of IOR activity were detected in cell-free extracts of l? furiosus. The cytoplasmic fractions from five different batches of cells contained between 0.3 and 0.4 uni tdmg of IOR activity. No activity was detected in the particulate fraction of cell-free extracts, indicating that IOR is a cytoplasmic enzyme. It was quite sensitive to inactivation by O,, as approximately half of the IOR activity was lost after incubating the cytoplasmic extract in air for 20 min (at 23 "C), whereas there was no decrease in activity even after several days if anaerobic conditions were maintained. Thus, the purification procedure was carried out under argon and all buffers contained sodium dithionite (2 mu) to protect against trace 0, contamination. All buffers also contained DTT (2 mal) and glycerol (10%). Neither seemed to stabilize or de- stabilize IOR during purification, but they do stabilize other ferredoxin-linked oxidoreductases of l? furiosus (Blarney and Adams, 1993; Mukund and Adams, 1991), and they are included in the buffers used to purify these enzymes. These reagents were also used for IOR as all of these enzymes are routinely obtained from the same batch o f P furiosus cells. The results of a typical purification are shown in Table I. The enzyme was purified about 120-fold with a yield of activity of 38%. Approximately 100 mg of purified IOR was obtained from 500 g (wet weight) of cells.

Molecular Composition and Stability of IOR-The purified enzyme gave rise to a single protein band on nondenaturing electrophoresis gel (6%) (data not shown). Two protein bands were observed after SDS-gel electrophoresis (12.5% acrylamide), and these corresponded to M, values of 66,000 2 4,000 and

Step Activity I'rotrin z,!:!iy; Rrcovcry Purificatinn

u n i t s tnp i f l f i t s / l f f p r; -fold

Cell extract 9,800 30,300 0.32 100 1.0 DEAE-Sepharose 6.200 4,300 1.44 63 4.5 Hydroxylapatite 8,400 3,000 2.8 86 8.8 Superdex S-200 5,900 568 10.4 60 32 Phenyl-Sepharose 3,800 100 38.0 39 119

r - l

I - 23,000 I 1

FK:. 1. SDS-polyacrylamide electrophoresis of P. furiosus IOR. The right Inrw contained purified IOR ( 5 pg), and the left lane Contained

97,400, 66,000, 45.000, and 29,000. marker proteins with molecular weights (from top to hottorn) of 116,000,

23,000 2 2,000 (Fig. 1). No lower molecular weight protein bands were evident after electrophoresis using 20% acrylamide gels. The pure enzyme eluted from a Superdex 200 gel filtration column as a single protein peak with an apparent M, of 160,000 2 25,000. These data indicate that IOR probably has a tet- rameric structure (a,&) with a native molecular weight of approximately 180,000 (this value was used in all subsequent calculations, see below). The NH,-terminal amino acid sequences of the two subunits are shown in Fig. 2. Amino acid compositional analysis of the holoprotein indicated a high content of glycine (10.7% of the total residues), alanine (9.7%), and valine (8.6%) (data not shown). Also of note was the presence of approximately 20 cysteinyl residues/mol of the enzyme. Esti- mations of the iron and acid-labile sulfide content of IOR using colorimetric assays gave values of 21.0 2 1.9 and 15.7 -+ 0.9 g a t o d m o l , respectively (values from at least three separate determinations). A similar iron content was determined by plasma emission spectroscopy. The results from this technique also showed that purified IOR contained magnesium, zinc, and copper in insignificant amounts (the values, in g atoms/mol, were 0.12, 0.18, and 0.025, respectively), and all other metals except iron were present a t concentrations <0.1 g a todmol . IOR also contained 0.24 2 0.03 TPP/mol (from three determinations of the same enzyme preparation). Thus, like l? fiwiosus POR (Smith et al., 1994), IOR is a TPP-containing Fe/S enzyme but in distinct contrast it does not contain copper.

IOR in its purified state was as sensitive to inactivation by O2 as it was in the cell-free extract. Using either indolepyruvate and phenylpyruvate as the ketoacid substrate, less than 50% of the original activity was recovered when the enzyme (8.0 mg/ml in 50 mM Tris/HCI, pH 8.0, containing 2 mM sodium dithionite and 2 mu DTT) was shaken in air to oxidize the dithionite and then left exposed to air at 23 "C for 20 min. There was no recovery of activity when the same enzyme solution was repeatedly degassed and flushed with argon and then reduced by the addition of sodium dithionite to 2 m ~ . When IOR was incubated under anaerobic conditions a t 80 "C (8.0 mg/ml in 50 mM EPPS, pH 8.4, containing 2 mu sodium dithionite, and 2 mM

I? furiosus Indolepyruvate Ferredoxin Oxidoreductase 16729

PfIORa - V q p T D I Y L W D E P P E B P L L L ~ N Q A I V R G A L E G N L A V F A X Y P PfPORa - - q y - - - - - - M B - P N E A A A W TmPORa M E R Y V E R V A V T G A - g A Y A N A M R Q I E P D V V A A Y P L T P Q T P I V HhPORa T D D E L I W R I A G G S P D G I D S T S Q N F A K A L M R S G L D V F T H R H ECIPD - M R T P Y C P - A g Y L L D E L T D C G A D H L F G V P G D Y N L Q F L D H v I

PfIORP X L K E Y N X V U G W - G Z L T A A N I L G X L A - L R A G Y X V X V G Pf PORP - - A V R K P P U T R E Y W A P G H U X A G G TmPORP P V N X K Q L A Q D E F D K K E X T Q G H R L X P G X G A P I T V K F V M M I A H h P O R P S K A F S A I D E R E V D R D A F T P G V E P Q P T W C P G G D F G V L K A L K G Tm PORY - - - - - X L R K V M K A N E A A A W A A K - L A K P K V I A A F P X X P X

PfPORy - M I E - - V A F H G R G s K A V T A A N I L A E L A F L G

FIG. 2. Comparison of the amino-terminal amino acid sequences of indolepyruvate- and pyruvate-oxidizing enzymes. The abbreviations are as follows: Pf, FI furiosus; Tm, Thermotoga maritima; Hh, H. halobium; Ec, E. cloacae; IDC, indolepyruvate decarboxylase. a, p, and

refer to different subunits. The subunit composition of each enzyme and the source ofthe data are: Ec IDC (a4, Koga et al., 1992); Hh POR (a2& Plaga et al., 1992); and Pf POR and T m POR (both apy8, Blamey and Adams, 1994). Gaps have been inserted to maximize homology with the (y

and p subunits of Pf IOR, and identical residues are underlined.

DTT), the time required for a 50% loss of activity was about 80 min (residual activity was determined by the direct transfer of the enzyme to assay cuvettes also at 80 "C). The effect of temperature on IOR activity was similar to that observed for other enzymes from P furiosus (see Adams, 1993). That is, IOR was virtually inactive a t ambient temperature and showed a dra- matic increase in activity above 70 "C, with an optimum above 90 "C (at pH 8.0, using either indolepyruvate of phenylpyruvate as substrates: data not shown). At 80 "C using the same substrates, the enzyme was not active at pH 5.2 (using 50 mM MES as buffer) or a t 11.0 (using 50 mM CAPS as buffer) and showed a broad optimum between pH 8.0 and 10.0 (using as buffers either MES, EPPS, or glycine, each at 50 mM concentration: data not shown).

Catalytic Properties of IOR-The activity of IOR was abso- lutely dependent upon CoASH: no activity was detected unless CoASH was added to the assay solution. This cofactor has a half-life at 80 "C of about 90 min (Blamey and Adams, 19941, so its thermal instability is not a limiting factor in the assay of IOR. As shown in Table 11, significant IOR activity was detected in the absence of added Mg" ions and TPP. The addition of Mg2' ions had only a minor effect, whereas IOR activity increased about 8-fold if TPP (with and without Mg2' ions) was added. A linear double-reciprocal plot was obtained when the concentration of TPP was varied in the standard assay medium, and the apparent K, value for TPP was 4 p ~ . Pure IOR contained only about 0.2 mol of TPP and 0.1 g atom of magnesiudmol. These kinetic data therefore suggest that Mg2' ions, whereas stimu- latory, are not obligatory for catalysis, whereas TPP is essential and dissociates during the purification procedure.

In addition to indolepyruvate, p-hydroxyphenylpyruvate and phenylpyruvate were oxidized by IOR in a CoASH-dependent reaction. Linear double-reciprocal plots were obtained when the concentration of each of these substrates was varied (in the presence of 0.1 mM CoASH, 2.5 mM MgCl,, 1.0 mM methyl viologen, and 0.4 mM TPP). The calculated K,,, and V,,, values were 250 PM and 54 units/mg for indolepyruvate (concentration range, 0.1-1.0 mM), 110 PM and 86 units/mg for hydroxyphenylpyruvate (0.05-0.5 mM), and 95 p~ and 101 units/mg for phenylpyruvate (0.05-0.4 mM), respectively. Thus, the enzyme exhibited highest affinity for phenylpyruvate. A linear double- reciprocal plot was also obtained when the concentration of CoASH was varied (0.02-0.2 mM, using either 5 mM indolepyruvate or 5 mM phenylpyruvate as substrate). For CoASH, the K, and V, values were 17 p~ and 62 units/mg using indolepyruvate and 21 PM and 142 units/mg using phenylpyruvate, respectively. I? furiosus ferredoxin replaced methyl viologen as the electron acceptor for indolepyruvate oxidation catalyzed by

The effects of TPP and Mg2' ions on the activity of E! furiosus IOR T~BLE I1

Substrate added" Specific activity

unitslmg

None 4.0 MgCl, (2.5 mM) 5.7 TPP (0.4 mM) 28.6 MgC1, (2.5 mM) + TPP (0.4 mM) 35.2

The assay mixture contained CoASH (0.1 mM), indolepyruvate (5.0 mM), and methyl viologen (1.0 mM).

IOR. Using indolepyruvate (5 mM) and CoASH (0.1 mM), a double-reciprocal plot of indolepyruvate oxidation activity uer- sus ferredoxin concentration was linear over the range 7.8-80 PM, and the K,,, and V,,, values were 48.4 p~ and 35 unitdmg. These data indicate that ferredoxin is an efficient electron carrier for the enzyme and presumably this is the physiological reaction.

As shown in Table 111, IOR also utilized as substrates in CoASH-dependent reactions 2-ketoisocaproate and 2-keto-y- methylthiobutyrate, which are the products of transamination reactions of leucine and methionine, respectively. However, pyruvate and the transaminated forms of several other amino acids (aspartate, valine, glutamate, and alanine) were not oxidized by IOR. For comparison, the activity of P furiosus POR was also tested with the same potential substrates. The two enzymes have very complementary substrate specificities, as POR did not utilize the arylpyruvates, but did oxidize 2-ketobutyrate and oxaloacetate, in addition to pyruvate. The transaminated forms of valine and glutamate were not oxidized by either enzyme. These data are consistent with preliminary results which show that I? furiosus contains two other 2-ketoacid ferredoxin oxidoreductases, one specific for 2-ketoglu- tarate3 and one specific for 2-ketoi~ovalerate.~

Copper and zinc ions were potent inhibitors of IOR activity. No activity could be detected when either (CuSO, or ZnSO, at a final concentration of 1 mM) was included in the standard assay mixture. IOR was also quite sensitive to inhibition by cyanide when assayed in its presence. Using various concentrations of KCN from 1 to 50 mM in the assay mixture, approximately 50% of the original activity was lost in the present of 7.5 mM KCN. Carbon monoxide (CO), which is the potent inhibitor of I? furiosus POR (Blamey and Adams, 1993), had no inhibi- tory effect on IOR either when various amounts of CO were added to the assay solution (up to 1 atm, equivalent to 320 PM in solution), or when the enzyme was incubated with CO (up to

X. Mai and M. W. W. Adams, unpublished data. H. Heider and M. W. W. Adams, unpublished data.

16730 I? furiosus Indolepyruvate Ferredoxin Oxidoreductase

TABLE I11 Substrate specificity of ZOR and POR from I? furiosus

Substrate" IORb PORb

Phenylpyruvate (Phe)" 87 0 p-Hydroxyphenylpyruvate ( T y r ) 70 0 Indolepyruvate (Trp) 38 0 2-Keto-y-methylthiobutyrate (Met) 26 0 2-Ketoisocaproate (Leu) 15 5 2-Ketobutyrate 0 11 Pyruvate (Ala) 0 21 Oxalacetate (Asp) 0 13 2-Ketoisovalerate (Val) 0 0 2-Ketoglutarate (Glu) 0 0 2-Ketomalonate 0 0 Phenylglyoxylate 0 0

a Each substrate was issued at a final concentration of 5 m ~ . * The activities of IOR and POR were measured at 80 "C and are

e The potential amino acid source (via transamination reactions) of expressed in micromoles of substrate oxidizedminlmg of protein.

the substrates are indicated.

1 atm) for 5 min at 23 "C prior to assaying in the absence of CO. IOR was also unaffected by the presence of sodium nitrite (up to 125 mM) in the standard assay mixture. Again, in contrast, I? furiosus POR is inhibited by nitrite (Blarney and Adams, 1993). POR was not inhibited by sodium fluoride or sodium azide (each at 5 mM final concentration: Blarney and Adams, 19941, and this was also true with IOR: the presence of these reagents in the assay medium did not affect its catalytic activity. Analogs of the substrates of IOR were also tested as potential inhibitors using both indolepyruvate and phenylpyruvate as substrates. However, sodium benzoate, phenylacetate, pyruvate, 2-ketoglutarate, and oxamate (at concentrations up to 5 mM) had no effect on IOR activity.

Electron Paramagnetic Resonance Properties of ZOR-IOR as purified in 50 mM Tris/HCl buffer, pH 8.0, containing sodium dithionite (2 mM) and DTT (2 mM) gave rise to a simple rhombic- type EPR spectrum signal at 8 K with g-values (gz, g,,, g,) of 2.05, 1.94, and 1.90 (Fig. 3b). The shape of the spectrum re- mained unchanged upon varying either the microwave power (from 0.02-20 milliwatts at 8 K) or the temperature (from 4.2 to 15 K using 10-milliwatt power). The signal was not observed upon raising the temperature above 20 K, and no additional resonances were evident a t lower magnetic fields (at 6 K, 20 milliwatts). The g-values and temperature dependence of this signal suggest that it arises from a S = 1/2 [4Fe-4SI1+ cluster. A virtually identical spectrum was obtained when the enzyme was prepared at pH 10.0 in dithionite-containing buffer (Fig. 3c). The spectra observed at the two pH values both represented 3.7 0.6 spins/mol, indicating that IOR contains at least three and more likely four magnetically isolated and noninter- acting S = 1/2 [4Fe-4SI1+ clusters. The effective reduction potential of sodium dithionite increases with increasing pH (from approximately -500 mV at pH 8.0 to -620 mV at pH 10: May- hew, 1978), yet no change was observed in the EPR properties of the dithionite-reduced enzyme (at pH 8.0 and 10.0). This suggests that the 4Fe-type centers are fully reduced in the sample prepared at pH 8.0.

The EPR spectrum of the reduced 4Fe-clusters of IOR was lost after anaerobic oxidation of the enzyme with excess thionine ( E , (midpoint redox potential) = +60 mV) and was replaced by a broad axial-type signal centered at g = 2.02 (Fig. 4a). This signal was not seen above 20 K, and temperature and power saturation analyses suggested that it arose from a single paramagnetic species. The signal represented 0.68 f 0.15 spin/ mol at 8 K. Its lineshape and temperature dependence are typical of the EPR properties of an oxidized S = 1/2 [3Fe-4SI1+ cluster. Such clusters are frequently generated during oxidative procedures by the degradation of [4Fe-4S11+ clusters (Bein-

a)

C)

I I I 0.30 0.35 0.40

Magnetic Field (T)

FIG. 3. Electron paramagnetic resonance spectra of reduced forms of €? furiosus IOR The samples were: a , thionine-oxidized IOR (9.9 mg/ml in 50 m~ TridHC1 buffer, pH 8.0) treated with CoASH (20 mM) and indolepyruvate (20 mM) at 70 "C for 2 min prior to rapid freezing; b, IOR as purified in 50 mM TrisiHC1 buffer, pH 8.0, containing sodium dithionite (2 mM) and D l T (2 m ~ ) ; and c, thionine-oxidized IOR (9.6 mg/ml in 50 mM CAPS buffer, pH 10.0) treated with sodium dithionite (5 mM) for 5 min at 23 "C prior to freezing. The conditions of measurement were: microwave power, 1 milliwatt; modulation ampli- tude, 2 G microwave frequency, 9.46 GHz; receiver gain, 5.0 x lo4; temperature, 8 K.

0.300 0.325 0.350 Magnetic Field (T)

oxidized €? furiosus IOR treated with substrates. The samples are: FIG. 4. Electron paramagnetic resonance spectra of thionine-

a, thionine-oxidized IOR (9.9 mg/ml in 50 mM Tris/HCl buffer, pH 8.0); b, as in a, except the enzyme was incubated with CoASH (20 mM) at 70 "C for 2 min prior to rapid freezing; and c, as in b, except the enzyme was incubated with indolepyruvate (20 mM) rather than CoASH. The conditions of measurement were as described in the legend to Fig. 3.

ert and Thomson, 1983). For example, the hyperthermophilic PORs from I? furiosus and Thermotoga maritima in their oxidized forms both exhibit EPR signals of this type, e.g. Blarney

I? furiosus Indolepyruvate

and Adams, 1993, 1994. However, in these cases, the oxidized 3Fe-center is a minor species representing <0.1 spidmol. Since there was no detectable loss of IOR activity after the oxidative procedure and the EPR signal of its oxidized 3Fe-cluster represented a significant spin concentration, these data suggest that the [3Fe-4SI1+ cluster is part of the active enzyme. Thus, IOR appears to contain four [4Fe-4Sl and one 13Fe-4SI cluster. Such a cluster content (equivalent to 19 iron atoms/mol) is consistent with the measured values for iron (approximately 21 iron atoms/mol) and cysteine (21 residues/mol, sufficient for complete coordination to all clusters). The acid-labile sulfide content (approximately 16 atoms/mol) is somewhat lower than would be expected, but the inherent difficulties in accurately determining this form of sulfur (Beinert, 1983) means that the measured value is a minimum one and therefore not inconsist- ent with the proposed cluster composition.

Intermediates in the catalytic cycle of indolepyruvate oxidation by IOR were generated by adding excess CoASH or indolepyruvate, or both, to the thionine-oxidized enzyme, incubating the sample at 70 “C for 2 min, and then rapidly freezing it for EPR analysis. As shown in Fig. 4b, the addition of CoASH did not change the lineshape of the EPR spectrum of the thionine- oxidized enzyme but caused a more than 50% decrease in signal intensity, as the signal from the putative [3Fe-4SI1+ cluster represented 0.25 2 0.05 spidmol (compare Figs. 4, a and b ) . In contrast, the addition of indolepyruvate to the oxidized enzyme had more than one effect. First, the signal from the oxidized 3Fe-center appeared to be lost or at least was greatly reduced in intensity, and second, it was replaced by a sharp isotropic- type signal with a g-value of 2.01 (Fig. 4c). The latter represented 0.15 2 0.05 spidmol and was observed even at 70 K, suggesting that it arose from an organic radical-type center. The addition of both substrates to oxidized IOR generated a g = 1.94-type spectrum typical of reduced [4Fe-4SI1+ clusters (Fig. 3a). This was similar, although not identical, to the EPR spectrum of the dithionite-reduced enzyme, and it also represented a significant spin concentration (3.0 2 0.5 spins/mol). The EPR signals seen upon the separate addition of CoASH and indolepyruvate were not apparent. Thus, the effect of CoASH and indolepyruvate in combination are to reduce the 4Fe-type centers. In essence, the enzyme undergoes a single turnover as the catalytic cycle cannot continue until the 4Fe-clusters are re- oxidized, which does not occur in this experiment in the absence of an exogenous electron carrier.

DISCUSSION

The hyperthermophilic proteolytic archaeon, l? furiosus, contains in its cytoplasm significant amounts of a new type of enzyme, IOR, which catalyzes the oxidative decarboxylation of the transaminated forms of the aromatic amino acids to the corresponding aryl acetyl CoA derivative at temperatures above 90 ”C. This enzyme is distinct from the analogous CoASH-dependent POR (Blarney and Adams, 1993) previously purified from this organism. POR, together with a novel CoASH-independent tungsten-containing aldehyde ferredoxin oxidoreductase (Mukund and Adams, 19911, are proposed to catalyze two of the oxidation steps in the conversion of glucose to acetyl CoA, H, and CO,, via a new type of nicotinamide- independent “pyrosaccharolytic” pathway (Mukund and Ad- ams, 1991; for review, see Adams, 1994). Like these enzymes, IOR utilizes P. furiosus ferredoxin as an electron carrier, and the excess reductant generated during the oxidation of aryl pyruvates is presumably coupled directly to H, production via the ferredoxin-dependent hydrogenase in this organism (Bryant and Adams, 1989).

In spite of the similarity in the reactions that they catalyze, IOR and POR from l? furiosus differ in many of their proper-

Ferredoxin Oxidoreductase 16731

ties. POR is a heterotetramer (apyG) with an apparent M , value of approximately 115,000 (Blarney and Adams, 1993), whereas IOR is larger (M, - 180,000) and contains only two types of subunits (a,p,). There may be some structural relationship, however, as the p-subunit of IOR ( M , 23,000) shows some NH,- terminal sequence identity with the y-subunit of POR ( M , 24,000). There was much less homology with the y-subunit of the heterotetrameric POR (a&@ from the hyperthermophilic bacterium, l! maritima (Blarney and Adams, 19941, and no significant identity between any combinations of the other subunits of these three enzymes, nor with the two subunits of the POR from the mesophilic and halophilic archaeon, Halobacte- rium halobium (Plaga et al., 1992: see Fig. 2). IOR and POR of l? furiosus have similar sensitivities to inactivation by 0, and resistance to high temperature, but differ in their catalytic properties. For example, IOR is more active (V, = 101 unitslmg with phenylpyruvate) and has a much higher affinity for CoASH (K, = 17 PM) and its primary substrates (K,,, = 95 PM for phenylpyruvate) than POR (V, = 23 units/mg with pyruvate, with K,, values for CoASH and pyruvate of 110 and 460 PM, respectively: Blarney and Adams, 1993), although they have comparable affinities for l? furiosus ferredoxin (K, values of 48-95 PM). Similarly, they have a mutually exclusive range of keto acid substrates, although both show some activity toward the transaminated form of leucine (Table 111).

Perhaps the most striking difference between IOR and POR stems from the fact POR is a copper-containing enzyme (Smith et al., 1994), whereas only insignificant amounts of this ele- ment were present in IOR. Moreover, the copper site in POR is intimately involved in catalyzing pyruvate oxidation (Smith et al., 1994). It is thought to stabilize an unusual hydroxyethyl- TPP radical intermediate, and to bind and oxidize CoASH, thereby facilitating its reaction with the activated acetyl derivative to generate acetyl-coA. In essence, copper in l? furiosus POR has been proposed in part to replace the role of lipoic acid in the well characterized multienzyme pyruvate dehydrogenase complex (Smith et al., 1994). Similarly, the electron- accepting role of FAD in pyruvate dehydrogenase (with subsequent electron transfer to NAD) is assumed by two ferredoxin- type [4Fe-4S] centers in l? furiosus POR (with subsequent electron transfer to ferredoxin). The question therefore arises as to the mechanism of aryl pyruvate oxidation by IOR? The results presented herein suggest that this enzyme contains one [3Fe-4Sl cluster and four [4Fe-4Sl clusters. As in POR, the 4Fe-centers are reduced by the addition of both substrates to the oxidized enzyme, consistent with their role in electron transfer from the site of pyruvate oxidation to the external electron carrier. On the other hand, the addition of CoASH led to the partial reduction of the oxidized 3Fe-center (in POR it is the copper site that is reduced in the analogous reaction), whereas the addition of indole pyruvate generated a radical- type EPR spectrum (the same occurs in POR in the analogous reaction). The latter is assigned to the same hydroxyethyl-TPP radical intermediate seen in POR (Smith et al., 1994). Thus, these preliminary results suggest that the copper site in POR is replaced by a [3Fe-4Sl cluster in IOR. Interestingly, the POR of the hyperthermophilic bacterium l! maritima also lacks copper, and we have postulated that in this enzyme an unusual FeS center (probably of the 4Fe-type, but not of the 3Fe-type), is involved in CoASH binding and acetyl transfer (Smith et al., 1994). However, pyruvate oxidation by T maritima POR in- volves more conventional TPP chemistry (there is no evidence for a radical-type intermediate) similar to that of the pyruvate dehydrogenase complex. It therefore appears that l? furiosus IOR catalyzes the oxidation of aryl pyruvates utilizing the radical-type mechanism of P. furiosus POR in combination with a

16732 I? furiosus Indolepyruvate Ferredoxin Oxidoreductase

novel function for a FeS center previously seen only in l! maritima POR. Further spectroscopic analyses of IOR are underway to confirm this usual and somewhat speculative mechanism for the oxidative decarboxylation of keto acids. As far as we aware, an enzyme analogous to P. furiosus IOR

that converts aryl pyruvates to the aryl acetyl-coA has not been reported from any other organism. Decarboxylase-type enzymes, which yield the corresponding aldehyde from aryl pyruvates rather than the CoASH-derivative, have been found in plants and certain bacteria, and these are involved in the synthesis of the plant hormone indole-3-acetic acid (see Koga et al., 1992). Similar enzymes are present in some aerobic and anaerobic bacteria which use aromatic amino acids as their carbon and energy sources (Fujioka et al., 1970; Eldsen et al., 1976; Barker, 1982). In these cases, nicotinamide-linked aldehyde dehydrogenases convert the aryl aldehyde to the aryl acetate. Only two aryl pyruvate decarboxylases have been characterized in any detail: these are indolepyruvate decarboxylase from E. cloacae (Koga et al., 1991, 1992) and phenylpyruvate decarboxylase from Acromobacter eurydice (Fujioka et al., 1970). Both are homotetrameric enzymes and contain TPP, but no other prosthetic groups. Their activity is stimulated by TPP (K, values - 3 p~), and they require TPP to maintain their quarternary structures. Like IOR, phenylpyruvate decarboxylase utilized indolepyruvate, phenylpyruvate, 2-ketocaproate, and 2-keto-y-methylthiobutyrate as substrates and had the highest affinity for phenylpyruvate (K, = 51 p) but did not oxidize pyruvate. In contrast, indoleppvate decarboxylase specifically oxidized indolepyruvate (K, = 15 p~), showed no reaction with phenylpyruvate, and was inhibited by pyruvate and 2-ketoglutarate. Thus, from the available information, IOR is more similar to phenylpyruvate decarboxylase than indolepyruvate decarboxylase in its catalytic properties. The complete amino acid sequence of indolepyruvate decarboxylase has been deduced from the gene sequence, and it shows consider- able homology to pyruvate decarboxylase. Perhaps not surpris- ingly, its NH,-terminal sequence does not exhibit any homology to either of the subunits of IOR (Fig. 2).

Finally, we turn to the physiological role of IOR in P. furiosus. This organism contains two types of aminotransferases which convert aromatic amino acids to the aryl pyruvate derivatives (Andreotti et aZ., 19941.2 IOR is therefore proposed to generate the corresponding aryl acetyl-coA, in contrast to mesophilic organisms which produce the corresponding aldehyde by decarboxylation and the aryl acetate by dehydrogenation (Scheme 1). The question then becomes: what is the fate of this aryl acetyl- CoA in P. furiosus? Although the degradation of aryl acetates has been well studied in aerobic organisms, e.g. Dagley (19861, little information is available on their anaerobic degradation. However, elegant studies by Fuchs and co-workers (Mohamed and Fuchs 1993; Mohamed et al., 1993) have established that benzoyl-CoAis a central metabolite in the utilization of phenyl- acetic acid. Thus, the CoA derivative is oxidized to phenylglyoxylate by a mechanistically unusual dehydrogenase system (step 5 in Scheme l), and phenylglyoxylate is then oxidatively decarboxylated to benzoyl-CoA by a hypothetical phenylglyoxylate: acceptor oxidoreductase (step 6 in Scheme 1). As indicated in Table 111, neither IOR nor POR are able to utilize phenylglyoxylate as a substrate. However, cell-free extracts of P. furiosus do contain significant amounts of CoASH-dependent

phenylglyoxylate oxidoreductase activity, as measured by the reduction of viologen dyes.3 It is therefore possible that €? furiosus metabolizes the products of the IOR reaction via phenylglyoxylate and benzoyl-CoA, and studies to substantiate these preliminary results are underway.

Acknowledgments-We thank Zhi Hao Zhou for growing i? furiosus and for preparing cell-free extracts and Jenny Blarney for helpful discussions.

REFERENCES

Adams, M. W. W. (1990) FEMS Microbiol. Rev. 75, 219-238 Adams, M. W. W. (1993) Annu. Rev. Microbiol. 47, 627-658 Adams, M. W. W. (1994) FEMS Microbiol. Rev., in press Adams, M. W. W., and Kelly, R. M. (eds) (1992) Biocatalysis at Extreme Tempera-

tures: Enzyme Systems Near and Aboue 100 "C, Series No. 498, American Chemical Society, Washington D. C.

Andreotti, G., Cubellis, M. V, Nitti, G., Sannia, G., Mai, X., Marino, G., andAdams, M. W. W. (1994) Eur. J. Biochem. 220,543-549

Aono, S., Bryant, F. O., and Adams, M. W. W. (1989) J. Bacteriol. 171,3433-3439 Barker, H. A. (1981) Annu. Reu. Biochem. 6 0 , 2 3 4 0 Beinert, H. (1983) Anal. Biochem. 131,373-378

Blarney, J. M., and Adams, M. W. W. (1993) Biochim. Biophys. Acta 1161, 19-27 Beinert, H., and Thomson, A. J . (1983) Arch. Biochem. Biophys. 222,333-361

Blarney, J. M., and Adams, M. W. W. (1994) Biochemistry 33, 1000-1007 Blumentals, I. I., Itoh, M., Olson, G., and Kelly, R. M. (1990) Appl. Enuiron.

Bryant, F. O., and Adams, M. W. W. (1989) J. Biol. Chem. 264, 5070-5079 Chen, J.-S., and Mortenson, L. E. (1977) Anal. Biochem. 79,157-165 Connaris, H., Cowan, D. A,, and Sharp, R. J. (19911 J. Gen. Microbiol. 137, 1193-

Consalvi, V., Chiaraluce, R., Politi, L., Vaccaro, R., De Rosa, M., and Scandurra, R.

Dagley, S. (1986) in TheBacteria (Sokatch, I. R., ed)Vol. 10, pp. 527655,Academic

DiRuggiero, J., Robb, F. T., Jagus, R., Klump, H. K., Borges, K. M., Kessel, M., Mai,

Edelhoch, H. (1967) Biochemistry 6, 1948-1954 Eggen, R., Geerling, A,, Watts, J., and de VOS, W. M. (1990) FEMS Microbiol. Lett.

Elsden, S. R., Hilton, M. G., and Waller, J. M. (1976lArch. Microbiol. 107,283-288 Fiala, G., and Stetter, K. 0. (1986) Arch. Microbiol. 145, 5-1 Fujioka, M., Morino, Y., and Wada, H. (1970) Methods Enzymol. 1 7 k 585-596 Kelly, R. M., and Deming, J. W. (1988) Biotech. Prog. 4 , 4 7 4 2 Koga, J., Adachi, T., and Hidaka, H. (1991) Mol. & Gen. Genet. 226, 10-16 Koga, J., Adachi, T., and Hidaka, H. (1992) J. Biol. Chem. 267, 15823-15828 Lovenberg, W., Buchanan, B. B., and Rabinowitz, J. C. (1963) J. Biol. Chem. 238,

Lowry, 0. H., Rosebrough,N. J., Farr,A. L., andRandal1, R. J . (195115. Biol. Chem.

Ma, K., Robb, F. T., and Adams, M. W. W. (1994) Appl. Enuiron. Microbiol. 60,

Mayhew, S. G. (1978) Eur. J . Biochem. 85, 535-547 Mohamed, M. E.-S., and Fuchs, G. (1993) Arch. Microbiol. 159, 554-562 Mohamed. M. E.&. Sevfried, S., Tschech.A., and Fuchs, G. (1993)Arch. Mirrohid

Microbiol. 56, 1992-1998

1199

(1991) Eur. J. Biochem. 202, 1189-1196

Press, New York

X., and Adams, M. W. W. (1993) J. Biol. Chem. 268, 17767-17774

71, 17-20

3899-3913

193,265-275

562-568

.~ ~. . . . . . .

Mukund, S., and Adams, M. W. W. (1990) J. Biol. Chem. 266, 11508-11516 Mukund, S., and Adams, M. W. W. (1991) J. Biol. Chem. 266, 14208-14216 Mukund, S., and Adams, M. W. W. (1993) J. Biol. Chem. 268, 13592-13600 Neuner, A., Jannasch, H. W., Belkin, S., and Stetter, K. 0.11990) Arch. Microbiol.

159,563-575

Ohshima, T., and Nishida, N. (1993) Biosci. Biotechnol. Biochem. 57, 945-951 Penttinen, H. K. (1979) Methods Enzymol. 62,58-69 Plaga, W., Lottspeich, F., and Oesterhelt, D. (1992) Eur. J. Biochem. 205,391-397 Riddles, P. W., Blakeley, R. L., and Zerner, B. (1983) Methods Enzymol. 9 1 , 4 9 4 0 Robb, F. T., Park, J.-B, and Adams, M. W. W. (19921 Biochim. Biophys. Acta 1120,

Schafer, T., and Schonheit, P. (1992) Arch. Microbiol. 158, 188-202 Smith, E. T., Blarney, J. M., and Adams, M. W. W. (1994) Biochemistry 33, 1008-

Snowden, L. J., Blumentals, I. I., and Kelly, R. M. (1992)Appl. Microbiol. Enuiron.

Stetter, K. 0. (1982) Nature 300,25%260 Stetter, K 0. (1986) in The Thermophiles: General, Molecular and Applied Micm-

Stetter, K. O., Fiala, G., Huber, R., and Segerer, G. (1990) FEMS Microbiol. Reu. 75,

Weber, K., Pringle, J. R., and Oshorn, M. (1972) Methods Enzymol. 26, 2-27 Woese, C. R., Kandler, O., and Wheelis, M. L. (1990) Proc. Natl. Aead. Sci. U. S. A.

153,205-207

267-272

1016

58, 1134-1141

biology (Brock, T. D., ed) pp. 39-74, John Wiley & Sons, Inc., New York

117-124

87,45764579

Documents

THE JOURNAL OF Bm.nCrcfi CHEMISTRY Vnl. No. Issue June … · THE JOURNAL OF Bm.nCrcfi CHEMISTRY 0 1994 by The American Society for Biochemistry and Molecular Biology, Inc. Vnl. 269,