nad(p)h:flavin oxidoreductase of escherichia coli

THE JOURNAL OF BIOLOGICAL C~eaosncv 0 1987 by The American Society for Biochemistry and Molecular Biology, Inc.

VOL . 262, NO. Isem of September 5, pp. 1232&12331,1987 printed in U.S.A.

NAD(P)H:Flavin Oxidoreductase of Escherichia coli A FERRIC IRON REDUCTASE PARTICIPATING IN THE GENERATION OF THE FREE RADICAL OF RIBONUCLEOTIDE REDUCTASE*

(Fkeived for publication, March 16,1987)

Marc Fontecave+, Rolf Eliasson, and Peter Reichard From the Department of Biochemistv, Medical Nobel Institute, Karoiinska Institutet, S-104 01 Stockholm, Sweden

The active form of one subunit of Escherichia coli ribonucleotide reductase (protein B2) contains an organic free radical localized to tyrosine 122 of its polypeptide chain. When this radical is scavenged, e.g. by treatment with hydroxyurea, the enzyme is inactivated (protein B2/HU). E. coli contains an enzyme system consisting of at least three proteins that in the presence of NADPH, FMN, dithiothreitol, and oxygen introduce the tyrosyl radical into B2/HU (Eliasson, R., Jornvall, H., and Reichard, P. (1986) Proc. Nd. A d . Sci. U. S. A. 83, 2373-2377). One of the three proteins was identZ1ed as superoxide dismutase. We now identify a second protein, previously provisionally named Frac- tion c, as an NAD(P)Hflavin oxidoreductase (flavin reductase). After 4,000-fold purifkation the protein moved as a single band on sodium dodecyl sulfate gel electrophoresis with a molecular weight of 28,000- 29,000. The enzyme contained no flavin but reduced riboflavin, FMN, and FAD by NADH, or riboflavin and FMN by NADPH. It is a powerful ferric iron reductase. We propose that its complementing activity during radical generation involves participation in the reduction of the femc iron center of protein B 2 m . Radical formation is then linked to the reoxidation of iron by oxygen. The flavin reductase may also participate in other aspects of iron metabolism of E. coli.

Ribonucleotide reductase is an essential enzyme of all living cells (1-3). It catalyzes the reduction of the four common ribonucleoside diphosphates to the corresponding deoxyribo- nucleotides and provides the cell with the proper supply of precursors for DNA. The enzyme from Escherichia coli which is the prototype of all known eucaryotic and virus-coded reductases consists of two nonidentical subunits, called proteins B1 and B2, each consisting of two identical polypeptide chains. B1 contains two binding sites for substrates and four sites for allosteric effectors and provides the redox-active sulfhydryl groups that participate in the reduction of the OH group of ribose. Protein B2 contains an iron center consisting of two antiferromagnetically coupled Fe3+ ions, linked by a p- oxo bridge, and one stable organic radical, identified as a one- electron oxidation product of tyrosine 122 in one of the two

* This work was supported by a grant from the Swedish Medical Research Council (to P. R) and by a postdoctoral research fellowship from the Centre National de la Recherche Scientifique (to M. F.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in aceordance with 18 U.S.C. Section 1734 solely to indicate this fact.

$ Present address: Laboratoire de Chimie et Biochimie Pharma- cologiques et Toxicologiques, Universiti Renb Descartes, 45 Rue des Saints-Peres. 75270 Paris 06, France.

polypeptide chains (4). This radical is postulated to participate in the catalytic process (5, 6). Treatment of B2 with hydroxyurea and some other chemicals destroys the radical resulting in the formation of an inactive form of B2 that still contains the intact iron center but has lost the radical (7,8). This form of enzyme was called B2/HU (9). A second inactive form of protein B2, called apoB2, is obtained by removal of iron with chelating reagents (8). ApoB2 lacks not only iron but also the tyrosyl radical, since the iron center is required for its stabiition. Witroduction of iron, in the form of Fez+, in the presence of oxygen regenerates both the iron center and the tyrosyl radical resulting in the formation of active B2 (10).

Extracts from E. coli contain enzyme activities that catalyze the interconversion of B2 and BS/HU (9, 11). It seems pos- sible that these activities form part of a biological mechanism that regulates the activity of ribonucleotide reductase by setting the content of tyrosyl radical in the enzyme. During purification of the radical introducing activity we separated three protein fractions that, in concert, and in the presence of NADPH, FMN, a dithiol, and oxygen catalyze the transformation of B 2 m to B2. One protein was identified as superoxide dismutase (11) while the function of the other two, provisionally named Fractions b and c, was unknown. Here we report the purification of Fraction c to near homogeneity and identify the protein as an NAD(P)H.flavin oxidoreductase (for brevity called flavin reductase) catalyzing the following reaction:

NAD(P)H + H+ + fla“, NAD(P)+ + flak The enzyme also acts as a femc reductase, reducing Fe3+

ions to Fez+ ions. We suggest that its function in the radical introducing reaction is linked to a reduction of the iron center of ribonucleotide reductase.

EXPERIMENTAL PROCEDURES’

RESULTS

Purification of “Fraction c”-Fraction c is one of the three proteins that together introduce the tyrosyl radical into B2/ HU. The other two proteins are superoxide dismutase and Fraction b (11). During purification from a bacterial extract superoxide dismutase was first separated from the other two

Portions of this paper (including “Experimental procedures” and

is easily read with the aid of a standard magnifying glass. Full aize Fig. 1) are presented in miniprint at the end of this paper. Miniprint

photocopies are available from the Journal of Biological Chemistry, 9650 Rockville Pike, Bethesda, MD 20814. Request Document No. 87M-0807, cite the authors, and include a check or money order for $2.80 per set of photocopies. Full size photocopies are also included in the microfilm edition of the Journal that is available from Waverly PreaS.

12325

12326 Flavin Reductase of E. coli

FIG. 2. Separation of Fraction b from Fraction c by chromatography on AcA 54. Protein (675 mg) from the DEAE step was added to a 5 X 90-cm column of AcA 54 equilibrated with 50 mM Hepes buffer, pH 8.0, and eluted with the same buffer. Fractions (20 m1/20 min) were collected and analyzed for protein (absorption at 280 nm, solid line) and Fraction b and c activity (5 pl/assay) as described under “Experimental Procedures.” Fraction b activity was assayed in the presence of 1 pg of helper Fraction c; Fraction c assays contained 6 pg of helper Fraction b. Helper enzymes were obtained from a previous similar AcA 54 chromatogram in which material corresponding to Fractions b and c had been pooled as shown in Fig. 2 (Fraction b = tubes 27-33, Fraction c = tubes 47-54). The amount of Fraction b became limiting in the assays for Fraction c in tubes 47-54 (cf. peak in Fig. 1B).

fractions by stepwise elution from DEAE. The subsequent separation of Fractions b and c by chromatography on AcA 54, a column that separates molecules according to size, is shown in Fig. 2. When portions from each chromatographic fraction were analyzed for their ability to complement superoxide dismutase we found no activation of B2/HU. However, complementation was obtained by combining fractions present early and late in the chromatogram. In Fig. 2 two peaks of activity appear, one early in the chromatogram (Fraction b), found when a protein present in late fractions was added in excess, and vice versa, one late peak (Fraction c) found when complemented with protein from early fractions. During the further purification of Fraction c, detailed in the Miniprint Section, assays involving activation of B2/HU were always made in the presence of an excess of Fraction b.

Fraction c Is an NAD(P)H:Flavin Oxidoreductase-A requirement for NADPH and FMN during B2/HU activation became apparent when the reaction was carried out with Fractions b and c after the AcA 54 step. This led to the discovery that Fraction c by itself catalyzes the oxidation of reduced pyridine nucleotides by flavins and provided us with a simpler assay for Fraction c activity. Table I shows the results from the two assays run side by side during the consecutive purification steps of Fraction c described in the Miniprint Section. Up to the last step, both assays gave an identical 1200-1300-fold purification. After the final FPLC’ step the protein was purified 4000-fold, judged from NADPH oxidation. Activity was not simultaneously measured by B2/ HU activation, but a later experiment showed a 4-fold purification compared to the previous step also by this assay. We will show below that the 4000-fold purified enzyme gave a

The abbreviations used are: FPLC, fast protein liquid chromatography; HPLC, high pressure liquid chromatography; SDS, sodium dodecyl sulfate; DTT, dithiothreitol; PMSF, phenylmethanesulfonyl fluoride; SOD, superoxide dismutase; Hepes, 4-(2-hydroxyethyl)-l- piperazineethanesulfonic acid.

TABLE I Parallel purification of complementing activity for B2 fHU activation

and NAD(P)H:fivin oxidoreductase BP/HU NADPH

activation oxidation

Units Specific Units Specific X 10“ activity X IO-’ activity

Protein

mg Extract 8,500 2.4 0.28 2.4 28 DEAE 1,320 3.1 2.34 1.3 98

Phenyl-Seph- 0.90 0.3 360 0.33 37,000

FPLC“ arose

0.045 NDb 0.055 122,000

AcA-54 32 0.8 25 0.77 2,400

“Only half the material from the phenyl-Sepharose step was purified by FPLC.

Not determined.

single band during gel electrophoresis and on isoelectric focusing. There can be little doubt that one and the same protein has the ability to complement B2/HU activation and to catalyze the reduction of riboflavin by NADPH. In different preparations the overall yield of the protein varied from 5 to 10%.

Physical Properties of the Flavin Reductase-The enzyme was stable at pH 7.4 except at low protein concentration (below 0.02 mg/ml). It was stored either at 0°C or in the frozen state. Dilute solutions could be stabilized by the addition of bovine serum albumin (0.1 mg/ml) and/or by glycerol. The sensitivity to dilution became problematic during chromatography, in particular during the FPLC step. Care was then taken to concentrate by centrifugation the fractions immediately as they emerged from the column. Alternatively, fractions were collected in tubes containing serum albumin.

The elution volume of the enzyme during AcA 54 chromatography (Fig. 2) suggested a molecular weight of approximately 30,000. A value of 28,000 was found when material after AcA 54 and FPLC was chromatographed by HPLC on a TSK-2000 SW column (Fig. 3). A similar value was obtained by gel electrophoresis under nondenaturing conditions (data not shown). During the last two purification steps enzyme activity copurified with material that on sodium dodecyl sulfate gel electrophoresis gave a band corresponding to a molecular weight of 28,000-29,000. After the FPLC step the enzyme moved as a single band on an SDS gel with this mobility (Fig. &I). We conclude that the enzyme is a monomer with a molecular weight of approximately 28,500. The enzyme was also analyzed by one-dimensional isoelectric focusing by Pharmacia P-L Biochemicals’ Phast technique (Fig. 4B). A single band was focused at a position corresponding to an isoelectric point of 4.9.

The reductase gave a typical protein spectrum at pH 7.4 with no particular chromophore visible at wavelengths above 300 nm at a protein concentration of 0.5 mg/ml (data not shown). When FMN or FAD were added to the enzyme the flavins were not bound tightly and could easily be separated from the protein by dialysis.

Catalytic Properties of the Flavin Reductase-The enzyme catalyzes the following reaction.

NAD(P)H + H+ + flavib,+ NAD(P)’ + flavin&

The experiments described in this paper were all done under aerobic conditions when reduced flavins were rapidly reoxi- dized by oxygen. The above reaction was monitored by following the disappearance of the absorbance of NAD(P)H at 340 nm. With this assay we measured the dependence of the reaction on various flavins and on reduced pyridine nucleo-

Flavin Reductase of E. coli 12327

> Y -

1 2 4 # a 1 0

Fraction 110

FIG. 3. High performance liquid gel chromatography of flavin reductase. The enzyme (50 pg) purified by AcA 54 chromatography and FPLC, together with 150 pg of bovine serum albumin, in 0.1 ml was chromatographed on a 7.5 X 600-mm TSK-2000 SW column equilibrated with 100 mM NaCl in 50 mM Tris-HCI, pH 7.5, with a HPLC machine from LKB products, Stockholm, Sweden. The same buffer was used for elution. Fractions (0.25 m1/0.5 min) were analyzed for BZ/HU activation (o"--o) and ferric citrate reduction (X---X), with the latter assay under nonstandard conditions (1 p~ FMN and 50 p~ NADPH). The solid line gives the recorded absorbance at 280 nm. The arrows indicate positions of separately chromatographed standard proteins (a, bovine serum albumin; 0, ovalbumin; m, myoglobin). The inset shows a semilog plot for molecular weights of the standards uersw their elution volume and positions the flavin reductase in the diagram.

4.55- - B L

FIG. 4. Analysis of flavin reductase by SDS gel electrophoresis ( A ) and isoelectric focusing ( B ) . Material after FPLC (1.5 pg for gel electrophoresis, stained with Coomassie Blue, and 56 ng for isoelectric focusing, silver stained) was used. Markers of known molecular weight or isoelectric point were run in parallel as indicated.

tides (Fig. 5). With NADH as reducing agent, all three flavins tested (riboflavin, FMN, and FAD) served as substrates. While their apparent K,,, values were nearly identical M) their Vmar values clearly were different with riboflavin giving the highest value. The differences between flavins were

I

I

C

0,02 , 0,04

FIG. 5. Lineweaver-Burk plots for substrates of the flavin reductase. Panel A , dependence on flavins with NADH (0.25 mM); p a n e l B , dependence on flavins with NADPH (0.25 mM); p a n e l C, dependence on reduced pyridine nucleotides with 15 p~ riboflavin. The reaction was started by addition of the reductase (0.31-1.58 pg/ mi). The rate of the reaction was calculated from the decrease in the absorption at 340 nm and is expressed as specific activity.

TABLE I1 Kinetic parameters for flnvins and reduced pyridine nucleotides

Substrate Second Apparent E':. substrate* K"l rateb

Riboflavin FMN FAD

Riboflavin FMN FAD

NADH NADPH

NADH NADH NADH

NADPH NADPH NADPH

Riboflavin Riboflavin

CM 0.6 0.8 0.8

1.6 2.0

25 43

1.0 0.35 0.35

1.5 0.12 0

1.1 1.7

a When pyridine nucleotides served as second substrate their concentration was 0.25 mM; with riboflavin as second substrate, the concentration was 15 p ~ .

* Relative initial rates were obtained from Vm.= values. 1.0 corre- sponds to a specific activity of 25,000.

still more apparent with NADPH as electron donor. Now FAD gave no reaction, and the Vmax value for FMN was much lower than for riboflavin. The two pyridine nucleotides had identical K,,, values. Table I1 summarizes the values for V,,, and K,,, obtained with a preparation of the enzyme after the phenyl-Sepharose step. The turnover number of the enzyme was low. Assuming a purity of 25% of this preparation (see Table I) and a molecular weight of 28,500 the turnover number of the enzyme with NADPH and riboflavin was approximately 5,000 calculated from the VmaX value of Table 11.

Reduced flavins formed by the action of the enzyme are efficient reductants for various compounds. Of particular


interest was the reduction of femc ions since the enzyme was discovered as part of a system acting on a protein containing a femc iron center. Reduction of femc citrate was demon- strated by the appearance of a strongly absorbing chromophore (X, = 562 nm) in the presence of ferrozine used as a ferrous ion trap. During purification of Fraction c, whatever the chromatographic method used, we found that femc iron reductase activity cochromatographed with B2/HU activation. This is exemplified by Fig. 3 which shows copurification of the two activities during HPLC on TSK-2000 SW. We conclude that the two activities reside in the same protein. The reduction of ferric citrate required addition of both a reduced pyridine nucleotide and flavin, with specificities identical to those found in the absence of Fe3+ and documented in Table I1 (data not shown). In the coupled reaction, the amount of pyridine nucleotide oxidized in air exceeded the amount of ferric iron reduced. The K,,, for FMN was 3 4 times higher for the reduction of Fe3+ compared to the reaction in the absence of iron. A closer analysis of the kinetics of Fe3+ reduction was complicated by their nonlinearity with time and awaits experiments under anaerobic conditions. The results obtained so far suggest that during aerobic conditions oxygen and Fe3+ compete for reduced flavin. A direct dem- onstration of this effect and its quantitation is given for ferricytochrome c, another Fe3+ substrate, in the experiments shown in Fig. 3 of the accompanying paper (16). In this case approximately 40% of the electron flow from reduced FMN was used for direct reduction of the ferric center of cytochrome c while the rest was lost by reduction of 0, to superoxide radicals. It is likely that our enzyme is closely related to femc iron reductases described by others. We will return to this point under “Discussion.”

Several dyes and other compounds were also reduced by the enzyme in the presence of NADH and FMN. Dichlorophen- olindophenol, an often used electron acceptor for flavopro- teins, was reduced directly by NADH in the absence of added flavin. Menadione and butyl peroxide did not behave in this way.

The enzyme was inhibited strongly by SH inhibitors such as N-ethylmaleimide ‘or iodoacetic acid (Table 111). EDTA, desferal, and dicoumarol did not inhibit the reaction.

Participation of Flavin Reductase in BZ/HU Activatwn- The results shown in Table I demonstrate a parallel purifi-

TABLE I11 Inhibition of flavin reductase by SH reagents

The enzyme (0.12 pg after the phenyl-Sepharose step) was incubated at room temperature for the indicated times in 0.055 ml of 0.1 M Tris acetate, pH 6.5 (N-ethylmaleimide experiment) or 0.1 M Tris- HCl, pH 7.5 (iodoacetate experiment). At zero time 3 pl of the 0.1 M inactivating reagent (final concentration 6 mM) was added. After 30 or 60 min, 5 pl of 10 mM NADPH, 5 p l of 0.15 mM riboflavin, and 135 pl of buffer were added, and flavin reductase activity was determined spectrophotometrically.

Addition Incubation time rnin

None 0 30 60

N-Ethylmaleimide 0 30 60

Iodoacetate 0 30 60

Units

2.40 2.25 2.55

2.10 0.80 0.10

2.25 0.95 0.10

cation of flavin reductase with B2mU activation. In this section we describe some of the requirements for complementation for B2/HU activation by the highly purified flavin reductase.

The need for three separate proteins is shown in Fig. 6. Panel A gives the effect of increasing concentrations of the flavin reductase at two fixed concentrations of Fraction b in the presence of an excess of superoxide dismutase. Both curves are S-shaped, but the sigmoidicity is more pronounced at the lower concentration of Fraction b. At higher concentrations of flavin reductase the rate of the reaction decreases (not shown in Fig. 6). These results demonstrate that complementation of BB/HU activation can be used as an assay for flavin reductase only within a limited concentration range and that the amount of Fraction b used in such assays is critical. This amount varied and had to be determined for each preparation of Fraction b. Panel B of Fig. 6 shows the effect of increasing concentrations of Fraction b at a saturat- ing level of flavin reductase, both with and without superoxide dismutase. In this and many other experiments proportion- ality between the reaction rate and amount of Fraction b added was limited to a small concentration range. This experiment also demonstrates a strong dependence of B2mU activation on the presence of superoxide dismutase. Panel C of Fig. 6 finally shows the time course of the reaction at two concentrations of flavin reductase. Again S-shaped curves were found. For practical reasons we chose an incubation time of 30 min for our assays. It should be realized that Fraction b is very crude and might contain several activities required for the overall reaction. Attempts to purify this fraction have so far met with little success.

BZ/HU activation required the presence of a reduced pyridine nucleotide and a flavin (Fig. 7). The flavin specificity mirrored earlier results found in the NAD(P)H:flavin oxidoreductase reaction (Fig. 5 and Table 11) in that it could be satisfied by either riboflavin, FMN, or FAD in the presence of NADH (panel A of Fig. 7), while with NADPH (panel B ) FAD was inactive. Optimal results were again obtained with riboflavin and NADPH. These data strongly indicate that the flavin reductase reaction as such is required for BB/HU activation. The reaction had an absolute requirement for reduced

B I

FIG. 6. B2 /W activation requires the cooperation of three protein fractions. Panel A, requirement for flavin reductase. Incu- bations were carried out under standard conditions with increasing amounts of flavin reductase (FPLC fraction) in the presence of either 15 pg (0) or 45 fig (X) of Fraction b; panel B, requirements for Fraction b and superoxide dismutase. Incubations were under standard conditions with 0.07 pg of flavin reductase with (0) or without (0) 0.1 pg of superoxide dismutase; panel C, time curve. Incubations were with 0.11 pg (m); or 0.055 pg (A) of flavin reductase and 45 pg of Fraction b for the indicated time periods.

Flavin Reductase of E. coli 12329

A I 20

[NAD(P)n]mM 2 i

FIG. 7. Requirements of B 2 D activation for flavins and reduced pyridine nucleotides. In all experiments B2/HU (15 pg) was incubated with 7 pg of Fraction b, 0.06 pg of flavin reductase, and 0.1 pg of superoxide dismutase under s t a n d a r d conditions but without an NADPH-regenerating system and at the indicated concentrations of reduced pyridine nucleotides and flavins. Incubations in p a n e l A were made with 2 mM NADH, in panel B with 2 mM NADPH, and in p a n e l C with 10 pM riboflavin. The ordinates give units as defined for the B2/HU assay.

TABLE IV B2/HU activation requires the presence of thiols

Incubations were under standard conditions with 0.07 pg of flavin reductase (phenyl-Sepharose fraction) and 45 pg of Fraction b except for the specified additions of thiol compounds. The experiments marked thioredoxin were 6 p~ thioredoxin and 0.3 p~ thioredoxin reductase; those marked glutaredoxin were 6 p~ glutaredoxin, 0.3 p~ glutathione reductase, and 2 mM glutathione.

Additions Units

Experiment 1 None DTT

1 mM 10 mM

Glutathione 10 mM 100 mM

10 mM 100 mM

Mercaptoethanol

Experiment 2 None Thioredoxin Glutaredoxin 1 mM DTT 1 mM DTT + thioredoxin 1 mM DTT + elutaredoxin

1.4

9.2 52

1.2 24

3.8 16

1.2 2.6 1.2

10.4 44 21

pyridine nucleotides which was not saturated even at 4 mM NAD(P)H (Fig. 7, panel C). We found, however, that when an NADPH-regenerating system was included, 0.5 mM NADPH sufficed for optimal results.

The activation of B2/HU depended critically on the presence of thiols (Table IV). Dithiols such as DTT were most

efficient giving optimal results at concentrations between 5 and 10 mM, while higher concentrations resulted in lower activity. Some activity was found with monothiols also, but then much higher concentrations were required, and not even 100 mM glutathione or mercaptoethanol could fully substitute for DTT. We could not identify a biological dithiol equivalent. Reduced thioredoxins (14) or glutaredoxins (15) were inactive by themselves but increased the effects of suboptimal concentrations of nonprotein thiols (Table IV). Addition of antibod- ies to either thioredoxin or glutaredoxin to the complete system (assuming their presence in Fraction b) did not inhibit B2/HU activation (data not shown).

DISCUSSION

The enzyme described in this communication was discovered as a component of a complex multiprotein system catalyzing the transformation of an inactive form of ribonucleotide reductase (BS/HU) into an active one by changing a tyrosine residue of the polypeptide chain into a tyrosyl free radical (9, 11). In this context the enzyme complements the activity of two other protein fractions. One of these was identified as superoxide dismutase, and its function in the overall reaction is discussed in the accompanying paper (16). The function of the other, Fraction b, is unknown. At first, complementation of B2/HU activation formed the basis of our purification procedure. We then found that radical intro- duction required addition of NADPH and FMN, suggesting the possibility of a reductive step in the overall reaction. This led to the discovery that the enzyme by itself catalyzed the reduction of various flavins by reduced pyridine nucleotides and thus can be classified as an NAD(P)H:flavin oxidorectu- rase.

The more than 4,000-fold purified enzyme gave on SDS gel electrophoresis a single band, with a mobility corresponding to a molecular weight of 28,000-29,000. The migration of the undenatured protein on gel exclusion chromatography and during gel electrophoresis suggested that the enzyme exists as a monomer in solution. The visible spectrum of the protein gave no evidence for a chromophore and excluded the presence of flavins in the isolated enzyme. When added, FMN or FAD did not bind tightly, and the enzyme should not be classified as a flavoprotein.

Flavins (including riboflavin) are substrates and are reduced by the enzyme. The enzyme apparently juxtaposes a pyridine nucleotide and a flavin such that electron transfer can take place. Steric hindrance might then explain why the bulky FAD is reduced by NADH but not by NADPH, while the other flavins are reduced by both pyridine nucleotides. Inhibition of the enzyme by SH reagents suggests that SH groups might participate in the catalytic process. As an alter- native explanation, SH groups may fulfill a structural function.

Many of the properties of our enzyme from E. coli are similar to those of the earlier described flavin reductases from luminous marine bacteria (17-21). The activity of these enzymes provides the reduced flavins required for the luciferase reaction responsible for bioluminescens. Two reductases from Beneckea harueyi, one specific for NADH and the other for NADPH, have been characterized (18, 19). The best studied NADH enzyme is inhibited by SH reagents (20) and uses riboflavin as the preferred flavin substrate (17). A major difference from the E. coli reductase is the pronounced (but not complete) specificity for NADH. In contrast, another marine bacteria, Photobacterium fischeri, apparently has a single flavin reductase using both reduced pyridine nucleotides (19,21). Flavin reductases have also been reported from


other organisms, including Bacillus subtills (22), Entameba histolytica (23), and human erythrocytes (24). These enzymes showed quite a pronounced preference for NADPH.

The B. harveyi enzyme has the ability to reduce and release iron from ferritin in the presence of NADH and a flavin (25). The similar ability of our enzyme to reduce Fe3+ ions raises the question of its relation to other ferric iron reductases. Enzymes capable of reducing substrates containing Fe3+ have been purified from several microorganisms, and indirect evidence points to their occurrence also in mammals. Such enzymes were given various names such as ferric iron reductase (26), ferrisiderophore reductase (27), or ferrichrome reductase (28). They were proposed to function in the release of Fez+ iron from storage forms of Fe3+ iron and thereby make iron available for insertion into various iron-containing proteins. These enzymes have in common requirements for a reduced pyridine nucleotide and a flavin. A highly purified preparation of ferric reductase from Rhodopseudomom sphaeroides (26) had a reported molecular weight of 32,000, close to the value for our E. coli enzyme. The two enzymes also show similar behavior during DEAE chromatography, and both have a strong affinity for phenyl-Sepharose. Also in view of the above mentioned ability of the B. harveyi flavin reductase to reduce ferric iron it appears likely that our E. coli enzyme is closely related to or even identical with earlier described ferric reductases from other sources.

What is the function of the flavin reductase during B2/HU activation? It was previously suggested that reduction of the iron center might be a prerequisite for radical formation (11). One strong reason for this comes from results concerning the chemical transformation of apoB2 to B2. In this case, the inactive form of B2 lacks both iron and the radical. Nonen- zymatic activation occurred on treatment of apoB2 with ferrous iron in the presence of oxygen (10). Apparently the radical is formed “automatically” when ferrous iron is oxidized in situ in the protein by oxygen. In contrast, BP/HU activation requires radical generation in a protein that already contains a ferric iron center. In analogy with the apoB2 reactivation we postulated (11) that the iron center first must be reduced and that the tyrosyl radical is formed nonenzymatically during reoxidation of iron by oxygen.

In consequence we now suggest that the flavin reductase provides the means for the reduction of the iron center of B2/ HU. The enzyme produces reduced flavins, and as such these are capable of reducing both Fe3+ ions and certain forms of stored ferric iron. The enzyme alone is, however, not able to activate B2/HU but requires complementation with superoxide dismutase and Fraction b. In the accompanying paper (16) we show that the requirement for superoxide dismutase probably depends on its ability to protect the system from

harmful oxygen radicals, generated by the flavin reductase. We cannot explain the requirement for Fraction b. Neverthe- less, this requirement demonstrates that reduced flavins by themselves are not able to reduce the ferric iron center of B2/ HU, at least not to such a state that on reoxidation regenerates the tyrosyl radical. Understanding of this apparently complex situation awaits purification and characterization of Fraction b.

REFERENCES 1. Thelander, L., and Reichard, P. (1979) Annu. Reu. Biochem. 4 8 ,

2. Lammers, M., and Follmann, H. (1983) Struct. Bonding 5 4 , 27-

3. Reichard, P., and Ehrenberg, A. (1983) Science 221,514-519 4. Larsson, A., and Sjoberg, B-M. (1986) EMBO J. 5 , 2037-2040 5. Stubbe, J., and Ackles, D. (1980) J. Biol. Chem. 255,8027-8030 6. Sjoberg, B-M., Graslund, A., and Eckstein, F. (1983) J. Biol.

Chem. 258,8060-8067 7. Ehrenberg, A., and Reichard, P. (1972) J. Biol. Chem. 247,3485-

3488 8. Atkin, C. L., Thelander, L., Reichard, P., and Lang, G. (1973) J.

Biol. Chem. 248 , 7464-7472 9. Barlow, T., Eliasson, R., Platz, A., Reichard, P., and Sjoberg, B-

M. (1983) Proc. Natl. Acad. Sci. U. S. A. 80, 1492-1495 10. Petersson, L., Graslund, A., Ehrenberg, A., Sjoberg, B-M., and

Reichard, P. (1980) J. Biol. Chem. 255,6706-6712 11. Eliasson, R., Jornvall, H., and Reichard, P. (1986) Proc. Natl.

Acad. Sci. U. S. A. 83.2373-2377 12. Bradford, M. M. (1976) Anal. Biochem. 72, 248-254 13. Fuller, R. S., Kaguni, J. M., and Kornberg, A. (1981) Proc. Natl.

14. Holmgren, A. (1985) Annu. Rev. Biochem. 5 4 , 237-271 15. Holmgren, A. (1979) J. Biol. Chem. 254 , 3664-3671 16. Fontecave, M., Graslund, A., and Reichard, P. (1987) J. Biol.

17. Michaliszyn, G. A., Wing, S. S., and Meighen, E. A. (1977) J.

18. Gerlo, E., and Charlier, J. (1975) Eur. J. Biochem. 57,461-467 19. Jablonski, E., and DeLuca, M. (1977) Biochemistry 16 , 2932-

20. Jablonski, E., and DeLuca, M. (1978) Biochemistry 17,672-678 21. Duane, W., and Hastings, J. W. (1975) Mol. Cell. Bioehern. 6,53-

22. Hasan, N., and Nestor, E. W. (1978) J. Biol. Chem. 2 5 3 , 4987-

23. Lo, H-S., and Reeves, R. E. (1980) Mol. Biochem. Purasitol. 2 ,

24. Yubisui, T., Matsuki, T., Takeshita, M., and Yoneyama, Y. (1979)

25. Jones, T., Spencer, R., and Walsh, C. (1978) Biochemistry 17 ,

26. Moody, M. D., and Dailey, H. A. (1985) J. Bacteriol. 163 , 1120-

27. Lodge, S. J., Gaines, C. G., Arceneaux, J. E. L., and Byers, B. R.

28. Straka, J. G., and Emery, T. (1979) Biochim. Biophys. Acta 569,

133-158

91

Acad. Sci. U. S. A. 78,7370-7374

Chem. 262,12332-12336

Biol. Chem. 252,7495-7499

2936

64

4992

23-30

J. Biochem. (Tokyo) 85,719-728

4011-4017

1125

(1982) J. Bacteriol. 149, 771-774

277-286

Flavin Reductase of E. coli 12331 Supplementary Material to:

NAU1PIH:FLAVIN OXIDOREDUCTASE OF E . COLI: A FERRIC IRON REDUCTASE PARTICIPATING IN THE GENERATION

OF THE FREE RADICAL OF RIBONUCLEOTIDE REDUCTASE

Marc Fonrecave, Rolf EliassOn and Peter Relchard

EXPERIMENTAL PROCEDURES

were from LKB-Pradukter. Stockholm. Sweden. phenyl-Sepharose and the Mono Q HR 5/5 column vlth the FPLC equipment were frbm Pharrnacra, Vppsala, Sweden. DEAE coIumns were regenerated by vashlng wlrh at least 20 v01umes of M NaCl. fol-

and 1 m M PMSF ln 1 0 8 glycerol [buffer AI. DEAE could be reused at least 10 lowed by equ~llbrat~an with a buffer contalnlng 25 mH Hepes, pH 8.0. 1 mM EDTA

times. AI1 solUt10nS used for enzyme purlf~cat~on and assays were treated vlth

Materlals - OEAE-Tr~saccyi 12. Ultrogel AcA 54 and Ultropac TSK 2000 SW

. . paste of 1.5 kg. The bacterla were stored frozen before use.

Assay of flavln reductase by complemenLaLiOn of 82/HU activation ~ ThlS assay ~ n v o l v e s Sllght modifications Of an e a r l i e r descrlbed procedure 1111. It IS carried ou t ~n two Steps. The actual B2/HU aCtIvatlon occurs during the

""its per mg of protein. Puc~f~catlon of flavin reductase - Extcactlun of bacterla - The extrac-

t l o n procedure followed the prlnclples g i v e n I" 1 1 3 1 . A frozen pellet of cen-

pH E . O . I mM EDTA and I mH PMSFI and allowed to thaw l n an Ice bath 12 h 0 . trifuged bacterla 1180 g1 was suspended 1" 180 m l of buffer 1 0 . 1 M Trls-HCl.

The pH of the SOIUtlon should now be between 7 and 8. The bacterial EUSpenSlOn was transferred to 60 ml centrifuge tubes of the Beckman TL 45 rotor and lysed by addltlon of 2 M KC1 (flnal C O ~ C 80 mM1 and 0.3 mg/ml of egg vhlte lYsOlYme + 0.2 ug/ml Of TI-lysozyme. The rubes were flrst Immersed I" an ice bath foc 45 mln, frozen ~n llquld nitrogen. thawed a g a ~ n 1" ~ c e and flnally centrifuged at 45,000 rpm for 90 min in a Beckman ulrracehtrlfuge. The supernatant SOIU- tlon 1230 ml, 37 mg proLeln/rnlI was decanted carefully and u5ed for chromatography on DEAE.

ChromaLography on U E A E . The extract was dlluted wlth redlstllled water to

acryl M a t a r a t e of 240 ml/hr. The column was eluted at a rate Of 80 m1/20 1000 ml and pumped onto the top Of a 8 x 14 cm 1700 mll column of DEAE-Tc1S-

mln. first wlth 0.07 M NaCi in buffer A Untll the absorbance at 280 nm was below 0.2, fallowed by 0.18 M NaCl Ln buffer A IF19 I A I , Mn-SOU could be recovered from the washthrough fractlon. Fe-SOD and the hybrld between Mn-SOU and Fe-SOU was present i n the materlal eluted w ~ t h 0.07 M NaCl. The last peak in Flg. IA 1340 m l , 6.2 mg prateln/mll contained the flavln reductase.

arnrnonlum sulfate 10.277 gm/rnli to g1Ye a flnal saturation Of 45 8 . Tbe Suspen- To thls was added 40 m l of fi phosphate buffer, pH 7.4, followed by solld

slon was stirred for I hr 1" the cold room and then centrrfuged In two 250 m l cenrrlfuge tubes for 30 m ~ n a t 1 0 0 0 rpm. The preckplrate was suspended ~n 3 to 4 m l of 25 mM Henes buffer. OH 8.0. ln 1 0 8 alvcerol-1 nM PMSF and dialyzed

Thls gave a 22 ml Of a fractlon Containing 6 0 ~ m g Of proteln/ml.

added from the bottonl to a 5 x 90 cm column of ACA 54 and displaced upwards vlth a buffer Contdlnlng 50 nM Hepes. pH 8.0. 10 8 glycerol and 1 mM PMSF. The column had been pre-equilibrated vlth the same buffer. FraCtlOnS 120 ml/2O m l n ] were collected from the t o p and analyzed for proteln labsorbance at 280 n m ] and flavln reductase actlvlty 1 F l g . 181. The flavln reductase appeared at the t a l l end Of the last proteln peak of the chromatogram. co~~espondlng LO a

Chromatography an ACA 54. The dlalyzed ammOnlUm Sulfate fraCtlOn Was

position Of a proteln wlth a molecular weight Of appraxlmately 30.000. Frac-

trated by ultradialysls agalnst 2 5 mM Hepes, pH 8.0 - 1 nM PMSF In ID 8 glyce- tlOnS were pooled Itotal volume 80 mlI as Indicated 1" F l g . 10 and concen-

To1 to a flnal volume Of 10 ml 13.2 mg proteLn/mlI. FraCtLOn b aCLlYlty was present as a broad peak early I" the chromatogram lcf F l g . 2 ~n Resultsl. When thls acLiYltv was recovered we oooled 5 to 7 oeak fractlons and concentrated

p"Cp05eS. Chromatography on phenyl-sepharose. Solid KC1 10.375 gm, flnal concen-

tratlon 0.5 M I was added to the enzyme from the ACA Step and the SolYtlOn was s l o w l y 15 m1/30 nin and fractlonl pumped onto the top of a 4 ml column Of phenyl-Sepharose, prewashed wlth 0.5 H KC1 - 2 5 mll Trls-HCI, pH 7.5 ~n 10 8 glycerol. The column was eluted wlth the same buffer and flow rate until the db- sorbance was below 0.05 a t 28O~nm. at which t;& Iflrst arrow i n F l q IC1 the KC1 COncentrarlOn was chanaed to 0.2 M . When the absorDtlOn at 280 nrn aoaln

made at I ml/rnln vlth a gradient from 0 to 0.7 H NaCl i n 10 mM Tcls-HCI. pH

durlng the first 5 m ~ n . followed by 0.07 M to 0.21 n between 5 and 40 mln. and 7.5, contalnlng 10 8 glycerol. The gradlent was run from 0 to 0.07 M NaCl

fmally from 0.21 to 0.35 M between 40 and 46 mln. Protein was detected by absorption at 280 nm. enzyme activity was determlned on portions Of the 1 ml fractions. ACtlYe fcactlons 13 nll wece irnmedlately pooled and concentrated by centrlfugatzon ~n centrlcon 10 mrcroconcentrators ln the SS 34 rotor Of Sor- "all centrlfuge to a flnal volarne Of 0.11 ml (0.4 mg proteln/mll The enzyme SOlUtlOn was frozen rapldly ~n llquld nitrogen and stored a t -60'.

B I I

6 R I C T l O N NUMBER Flg. 1. Chrornatagraphlc s t e p s far the purlflcatlon of flavln reductase.

protelnl was chromatographed as described in the text. The graph Shows the Panel A: Chromatography on DEAE-TrISBCry1. The bacterial extract 18.5 gm

absorbance at 260 (broken llnel and 280 nm Isolld l l n e l of the Chromato- graphic fractlons vlth exclu~lon of the runthrough fractlon. At the flrst arcow elutlon was Started vlth 0 . 0 7 U NaCl in buffer A, at the second arrow the NaCl Concentration Of buffer A was raised to 0.18 M . The runthrough frac- t m n can be used for preparatlon Of In-SOD, the flrst proteln peak labsorption at 280 nm 260nml for preparatlon Of Fe-SOD and the hybrld enzyme 1111. It was lmportant to elute nucleotkdes and nucleic acids present ~n the next two Peaks lab50rptlon a t 260 nm > 280 nml before changlng to the hlgher NaCl con- Centration. Flavln 1edUctaSe and Fraction b emerged together I" the peak eluted by 0.18 n NaC1. r l a l from the p<ev~ous s tep 11.32 g protein1 was chromatographed as described

Panel B: Chromatography on AcA 54. The arnmonlum sulfate preclpltated mate-

~n the text. The 4caDh Shows the oratein Oattern lsolld line. absorbance a t 280 nml and flavln reductase actlblty lbrbken line).

flavln reductase aCtlYILY. ElULlOn was started vlth 0 . 5 M NaCl and changed to Step was used. The solid llne gLve5 protein concentration, the broken llne

0.2 M a t the flcst arrow. At the second arrow NaCl was omltted completely and the reductase was eluted v l t h 25 rnM TriS-HCI, pH 7.5, ~n 10 a glycerol.

Panel U: FPLC chromatography On Mono Q 5 / 5 . The materlal from the previous step 10.45 mgI was pumped Onto the column and chromatographed as described i n the t e x t . The graph Only gLVes the data Obtained from the run between 5 and 40 mi", vlrh the solid llne Showing the absorbance a t 280 nm lprotelnl and the broken llne flavln reductase activlty of the lndivldual 1 m l fractions.

Panel C: Chromatography on phenyl-Sepharose. The materlal from the previous

nad(p)h:flavin oxidoreductase of escherichia coli

Documents