New mechanistic insights into the reactivity of the R2 protein of E. coli ribonucleotide reductase (RNR)

www.elsevier.nl/locate/jinorgbio Journal of Inorganic Biochemistry 79 (2000) 59–65

New mechanistic insights into the reactivity of the R2 protein of E. coli ribonucleotide reductase (RNR) M.B. Twitchett, A.M. Dobbing, A.G. Sykes * Department of Chemistry, University of Newcastle, Newcastle upon Tyne NE1 7RU, UK Received 17 April 1999; received in revised form 10 January 2000; accepted 14 January 2000

Abstract Further to a linear free-energy correlation of cross-reaction rate constants k12 for the reaction of eight organic radicals (OR), e.g. MVxq, from methyl viologen, with cytochrome c(III), we consider here similar studies for the reduction of the R2 protein of Escherichia coli ribonucleotide reductase, which has FeIII2 and Tyrx redox components. The same two techniques of pulse radiolysis and stopped-flow were used. Cross-reaction rate constants (22 8C) at pH 7.0, Is0.100 M (NaCl), were determined for the reduction of active-R2 with the eight ORs, reduction potentials E01 from y0.446 to q0.194 V. Samples of active-R2 have an FeIII2 met-R2 component, which in the present studies was close to 40%. Concurrent reactions have to be taken into account for the five most reactive ORs, corresponding to reduction of the FeIII2 of met-R2 and then of active-R2. Separate experiments on met-R2 reproduced the first of these rate constants, which on average is ;66% larger than the second rate constant. A single Marcus free-energy plot of log k12y0.5 log10 f versus yE01/0.059 describes all the data and the slope of 0.54 is in satisfactory agreement with the theoretical value of 0.50. Such behaviour is unexpected since the Tyrx is a much stronger oxidant (E0;1.0 V versus NHE) as compared to FeIII2 (E0 close to zero). X-ray structures of the met- and red-R2 states have ˚ buried active site is maintained. Proton transfer is therefore proposed as a rapid sequel to electron indicated that electroneutrality of the ;10 A transfer. Other reactions considered are the much slower conventional time-range reductions of active-R2 with hydrazine and dithionite. For these reactions one and/or two-equivalent changes are possible. With both reductants, met-R2 reacts about four-fold faster than active-R2, and as with the ORs the less strongly oxidising FeIII2 component is reduced before the Tyrx. q2000 Elsevier Science Inc. All rights reserved. Keywords: Ribonucleotide reductase; R2 protein of RNR; FeIII2 enzyme; Tyrosyl radical enzyme; RNR mechanism; Dithionite reactivity; Hydrazine reactivity

1. Introduction The enzyme ribonucleotides (RNRs) catalyse the reduction of the four biological ribonucleotides to the corresponding deoxyribonucleotides, formation of which is a prerequisite to de novo DNA synthesis [1–6]. The catalytic reaction is indicated in Eq. (1):

(1)

Three major classes of RNR have been defined [5], all of which appear to make use of a radical mechanism at some stage of the catalytic cycle. The most extensively studied is * Corresponding author. Tel.: q00-44-91-222-6700; fax: q00-44-91261-1182.

the class with a binuclear FeIII2 centre and stable tyrosyl free radical (Tyrx). This class is found in mammalian cells, viruses of the herpes group, and some prokaryotes, e.g. E. coli. Inactivation of RNR is a target in the context of new drugs against pathogenic organisms and certain forms of tumours. To date, E. coli RNR has received most attention. The enzyme consists of two proteins with identical subunits (aa and bb) referred to as R1 (Mrs2=85.5 kDa) and R2 (Mrs2=43.5 kDa) [7]. The R1 component binds substrate and allosteric effectors and contains a number of redox active cysteine residues [8,9]. The R2 subunit in its active form contains a tyrosyl radical at Tyr-122, which is stabilised by (but not coordinated to) a m-oxo antiferromagnetically coupled FeIII2 centre. From X-ray crystallography the phenolate ˚ from the nearest iron FeA [10]. O-atom of Tyr-122 is ;5.3 A The mechanism by which the tyrosyl radical is generated in R2 involves binding of O2 to the FeII2 state, with oxidation yielding FeIII2 as well as Tyrx [5,11,12]. A mechanism has also been proposed for the reduction of substrate bound to

0162-0134/00/$ - see front matter q2000 Elsevier Science Inc. All rights reserved. PII S 0 1 6 2 - 0 1 3 4 ( 0 0 ) 0 0 0 0 8 - 8

Friday Apr 07 10:49 AM

StyleTag -- Journal: JIB (Journal of Inorganic Biochemistry)

Article: 6350

60

M.B. Twitchett et al. / Journal of Inorganic Biochemistry 79 (2000) 59–65

R1 [11,13,14]. This involves hydrogen abstraction at the ribose C-39 position of the substrate by the thiyl Cys-439 radical. Reduction of the OH to H2O at C-29 follows, giving an intermediate a-keto radical. The latter is then reduced to a 39-deoxy-nucleotide radical with concurrent oxidation of two cysteines to a disulfide. Finally, the H-atom abstracted from C-39 is returned to the C-39 position as indicated in Eq. (1). The full mechanism is believed to involve regeneration of the thiyl radical at Cys-439 by the Tyrx on R2, a process ˚ involving electron transfer over 35 A. The three oxidation states of R2 identified by UV–Vis spectroscopy are defined in Eq. (2): ey III 2 active-R2

Fe

«Tyr ™Fe x

2 ey III 2 met-R2

«(TyrH)™Fe II 2«(TyrH)

(2)

red-R2

Here Tyrx is the radical of the deprotonated phenolate form, and TyrH is normal tyrosine. X-ray crystal structures of metR2 and red-R2 have been determined, and the active centres are shown in Fig. 1 [10,15]. It is noted that in the 3ey conversion of active-R2 to red-R2 electroneutrality of the active site is maintained, and 3Hq ions are also taken up. The phenolate O-atom of Tyr-122 and the FeIII2 centre are ;10 ˚ from the surface of R2 (broken line in Fig. 2) and the A immediate environment is hydrophobic. In addition to the three forms in Eq. (2), EPR spectra of a semimet-R2 with an FeIIFeIII centre have been observed under carefully controlled conditions [16]. This paper reviews kinetic behaviour relating to the reduction of the FeIII2 and Tyr. of R2 by eight organic radicals (OR). A Marcus free-energy relationship for the same eight organic radicals as reductants for horse-heart cytochrome c(III) has already been established [17]. Whether a similar free-energy relationship exists for R2 is now considered [18]. Other results for the reduction of R2 with N2H4 and S2O42y are considered [19,20]. In previous studies [2,5] a specific route for R2 electron-transfer has been proposed, which is believed to be a part of an extended pathway for communication with substrate bound to R1. The amino acids involved are conserved, and are connected by H-bonds (Fig. 2) [5]. E. coli and mouse R2 variants, in which relevant amino acids are replaced by residues unable to H-bond, have been shown to severely impair redox activity [21,22].

Fig. 1. Active-site structures of E. coli R2: (a) met-R2 form, and (b) redR2 which is four-coordinate with (in addition) the second carboxylate ˚ from the Fe-atoms [10,15]. O-atoms of Asp-84 and Glu-204 at ;2.7 A

2. Materials and methods 2.1. Preparation of R2 protein Wild-type E. coli R2 protein in the bb form was prepared from an E. coli over-producing strain by procedures already described [23]. Protein was made air-free by dialysing against deaerated buffer. Samples of met-R2 were prepared by reducing active-R2 (0.5 mM) with hydroxyurea (50 mM) for 30 min at 25 8C [24]. Met-R2 has UV–Vis absorbance peaks l/nm (´/My1 cmy1) at 325 (1.0=104) and 370

Friday Apr 07 10:49 AM

Fig. 2. The pathway proposed for long-range electron transfer from the substrate-binding site in R1 to the tyrosyl radical in R2 of E. coli ribonucleotide reductase [2,5].

(8.4=103) [25,26]. Active-R2 has an additional sharp Tyrx peak at 410 nm (6600) [26]. Fully reduced-R2 protein has no significant absorbance at )300 nm. No preparation of active-R2 to date has -25% of the met-R2 component which, in the present studies, was close to 40% [27]. Concentrations of enzyme were determined using the difference in absorption

StyleTag -- Journal: JIB (Journal of Inorganic Biochemistry)

Article: 6350

M.B. Twitchett et al. / Journal of Inorganic Biochemistry 79 (2000) 59–65

61

coefficients at 280 and 310 nm, ´280y´310s1.2=105 My1 cmy1 [27]. 2.2. Buffers The following buffers (from Sigma Chemicals) were used: tris-(hydroxymethyl)amino-methane (Tris, pH 7.1–8.9); sodium hydrogen phosphates (pH 5.7–8.0); N-(2-hydroxyethyl)piperazine-N9-ethanesulfonic acid (Hepes, pH 6.8– 8.2); 2-morpholinoethane-sulfonic acid (Mes, pH 5.5–6.7). 2.3. Reductants The following organic reagents (Fig. 3) were used to generate one-electron reduced organic radicals (ORs), where trivial, systematic, and abbreviated names are indicated: Methyl Viologen, C12H14N2Cl2 (1,19-dimethyl-4-49-bipyridinium dichloride), MV2q; Benzyl Viologen, C24H22N2Cl2 (1,19-dibenzyl-4,49-bipyridinium dichloride), BV2q; Phenosafranin, C18H15N4Cl (3,7-diamino-5-phenylphenazinium chloride), Pfq; Riboflavin, C17H20N4O6 (vitamin B2), Rb; Resorufin, C12H6NO3Na (7-hydroxy-3H-phenoxazine-3one), Rfy; Methylene Blue, C16H18N3SCl (3,7-bis(dimethylamino)phenazothionium chloride), MBq; Toluidine Blue, (C15H6N3SCl)2ZnCl2 (approximate formula only), TBq; Indo-Phenol, C12H6Cl2NO2Na (phenolindo-2,6-dichlorophenol, Naq salt), IPy. These reagents were obtained from Sigma Chemicals except MBq and IPy (BDH), and TBq (Fluka). Toluidine Blue was purified by recrystallisation as described [17]. Reduction potentials for the oneequivalent process involving the OR and parent vary over the range y0.446 (for MVxq) to q0.194 V (for IPx2y). The purity of sodium dithionite (Na2S2O4, Fluka) was determined by titration with [Fe(CN)6]3y, the UV–Vis peak at 420 nm (´s1010 My1 cmy1) using a glove-box (O2-2 ppm), and found to be 84% of the formula indicated [19]. 2.4. Kinetic studies The more reactive ORs were generated in situ using pulse radiolysis techniques, and other ORs by prior reduction of the parent with dithionite followed by stopped-flow studies. In both procedures it was necessary to include consideration of doubly reduced parent forms. Pulse radiolysis studies were carried out on a Van de Graaff accelerator at the University of Leeds, using a triple-pass cell (6.9 cm light path length) and a 2.5 MeV (;4=10y16 J) beam of electrons [28]. Airfree solutions were saturated with N2O containing R2 protein (;10 mM), organic reagent (100 mM) at pH 7.0 (40.5 mM phosphate) with 10 mM sodium formate, Is0.100 M. Strongly reducing formate radicals CO2xy (y1.9 V) were generated in situ in a rapid stage, and the ORs (written here as Xxy) were formed in a fast process, CO2xyq X™CO2qXxy, rate constants )5=109 My1 sy1 [17]. Subsequent reactions of Xxy with the R2 protein were monitored at UV–Vis wavelengths at which absorption of the OR

Friday Apr 07 10:49 AM

Fig. 3. Formulae (pH;7) and abbreviations for the eight organic reagents as parent forms (X) of the ORs (Xxy) studied in this paper.

or parent is dominant. No direct reaction of CO2xy or eyaq with R2 is observed. Absorbance at 410 nm was dominated by the parent or radical OR, and it was not possible to monitor decay of the R2 components at this or nearby wavelengths. Stopped-flow studies were carried out using an Applied Photophysics SX-17ML UV–Vis stopped-flow spectrophotometer. Reduction of the parent forms MBq, TBq and IPy was achieved by dropwise addition of concentrated sodium dithionite (10 mM) under air-free conditions in the glovebox until little colour remained. Excess of S2O42y was avoided because it is able to react directly with the protein [19]. The solution generated is substantially in the doubly reduced form, from stoichiometry measurements (see below) in which the amount of dithionite consumed was determined. Reaction conditions adopted were as follows: R2 protein (30–47 mM), organic reagent for OR (4.1–9.9 mM), pH 7.0 (45 mM phosphate), Is0.100 M. Reactions were monitored at wavelengths corresponding to reformation of the parent, details as in [17]. Appropriate fitting programs and procedures [29] were also used. For the reduction of R2 by hydrazine [20] and dithionite [19] UV–Vis absorbance changes were monitored by conventional time-range spectrophotometry, with the reductant in )10-fold excess of the R2 protein taking account of the stoichiometry. 2.5. Marcus free-energy relationship In order to investigate the correlation of cross-reaction rate constants, k12, for OR reductants with E. coli R2 (k1 and k2

StyleTag -- Journal: JIB (Journal of Inorganic Biochemistry)

Article: 6350

62

M.B. Twitchett et al. / Journal of Inorganic Biochemistry 79 (2000) 59–65

values in the text below), the following Marcus Eqs. (3) and (4) were used: k122sk11 k 22 K 12 f log fs

(3)

log K 122 4 log(k11 k 22/Z 2)

(4)

where k11 and k22 are self-exchange rate constants for OR and R2 couples respectively, and the collision frequency Z is assumed to be 1011 My1 sy1 [30]. The Nernst equation, log K12s(E02yE01)/0.059, was used to obtain equilibrium constants K12 from the reduction potentials for the OR (E01) and protein (E02) couples. Eq. (3) can accordingly be written as Eq. (5): (log k12y0.50 log f )s0.50 (log k11qlog k 22 0

(5)

0

and the existence of a free-energy correlation explored by plotting (log k22y0.50 log f) against (yE01/0.059).

3. Results 3.1. Reduction of R2 by organic radicals Reduction of active R2 samples by the radicals MVxq, BVxq, Pf x, Rbxy and Rf x2y (written here as Xxy), generated in situ by pulse radiolysis, can be expressed as in Eqs. (6) and (7): X qFe

II

III

«Tyr ™XqFe Fe «Tyr x

x

(6) (7)

Rate constants k2 for the reaction of met-R2 with the ORs were determined by separate experiments on met-R2 [18]. The disproportionation of OR (Eq. (8)): kD 2y

or X 2 2y

(8)

also contributes to absorbance changes in the case of Pf x, Rbxy and Rf x2y. No steps corresponding to the further reaction of X2y were required. From fitting procedures to Eqs. (6)–(8) values of k1 could be determined; see listings in Table 1. The similarity of k1 and k2 values (k2)k1 on average 66%) suggests that reduction of FeIII2 is taking place in both cases. Thus, the Tyrx of active-R2 plays a passive role, with subsequent changes triggered by the arrival of a proton and intramolecular ET giving FeIII2«TyrH. In the case of metR2 the initial product is FeIIFeIII«TyrH, which has been detected by EPR at ;5% levels [16]. This appears to undergo rapid further reduction to FeII2«TyrH. Kinetic experiments on the three less strongly reducing ORs (TBx, MBx and IP2y) were carried out by the stoppedflow method, using dithionite reduced solutions. With TBq and MBq, stoichiometry experiments indicate reduction to

Friday Apr 07 10:49 AM

k1 (My1 sy1)

k2 (My1 sy1)

MVxq a BVxq a Pf x a Rbxy a Rf x2y a MBx b TBx b IPx2y b

y0.446 y0.355 y0.243 y0.200 y0.059 q0.006 q0.015 q0.194

1.1(1)=109 3.3(4)=108 2.2(4)=107 1.1(1)=107 2.7(5)=106 7.3(3)=105 6.9(3)=105 1.1(2)=104

12.0(1)=109 3.5(2)=108 3.4(5)=107 2.6(2)=107 3.4(5)=106 – – –

a

Pulse radiolysis results. Stopped-flow results.

the doubly reduced X22y or X2y state (Eq. (8)). In the case of IPy, only one-electron reduction to the OR form is observed, and the treatment is simpler. No further reduction of met-R2 was observed, which is in keeping with a twoequivalent reduction potential of y0.115 V determined for the FeIII2/FeII2 couple [31]. It is possible to study the reduction of active-R2, since the thermodynamically unfavourable reduction of FeIII2«Tyrx to FeIIFeIII«Tyrx is followed by the favourable intramolecular ET to Tyrx giving met-R2. In the case of TBq and MBq reaction steps involving X2y and Xxy can be written as in Eqs. (9)–(11): k3

X 2yqFe III2«Tyr™X xyqFe II Fe III«Tyrx

(9)

2X xy™XqX 2y

(10) k1

k2

X xyqFe III2«TyrH™XqFe II Fe III«TyrH

2X xy™XqX

E01 (V)

kD

k1 III 2

OR

b

qE 2/0.059)y0.50 E 1/0.059

xy

Table 1 Results from pulse radiolysis and stopped-flow experiments. Rate constants (22 8C) for the reduction of FeIII2 in active-R2 (k1) and met-R2 (k2) with organic radicals (OR). These are plotted as k12 in Fig. 4

X xyqFe III2«Tyrx™Fe II Fe III«Tyrx

(11)

Values for k1 only are listed in Table 1. In the case of IPy only the single reduction by dithionite is observed, and kD and k3 do not contribute. 3.2. Marcus free-energy plot The procedure has been tested previously for the reduction of cytochrome c(III) with the same eight ORs [17], when the assumption that the ORs have a common self-exchange k11 value of 1.0=106 My1 sy1 proved acceptable. It is therefore possible to proceed with the same value of k11 in the present studies, and comment on the magnitude of E02 and k22 (both for R2) contributing to the intercept of the yE01/ 0.059 axis in Eq. (5). At the outset it was not certain which R2 reduction potential (for FeIII2 or Tyr.) was applicable. Fig. 4 shows a plot of rate constants for all eight ORs with active-R2, and for five of the ORs with met-R2, as log10k12y0.5 log10f against yE01/0.059. The inset indicates the corresponding behaviour observed in the free-energy plot for cytochrome c(III) [17]. The combined R2 data give a single linear plot. Separate consideration of the active- and met-R2 data has a negligible effect on the slope and intercept

StyleTag -- Journal: JIB (Journal of Inorganic Biochemistry)

Article: 6350

M.B. Twitchett et al. / Journal of Inorganic Biochemistry 79 (2000) 59–65

63

fold variation), it is proposed that this reaction also proceeds by reduction of FeIII2. As in the OR reactions, subsequent intramolecular reduction of Tyrx may well become easier on arrival of a proton. Relatively rapid reduction of FeIII2 or FeIIFeIII through to the FeII2 product then occurs. Active-R2 requires 3Hq as well as 3ey to maintain electroneutrality. It is difficult to specify the order of different ey and Hq steps. Two protons are taken up by the m-oxo of met-R2, which is not present in red-R2, Fig. 1. From the effects of pH little or no reaction of N2H5q (pKas8.2) with R2 is observed. Reactions with other cationic CoII complexes are also unfavourable [33]. 3.4. Summary of reduction of R2 by dithionite [19]

Fig. 4. Marcus free-energy plot for reduction of active-R2 (m) and met-R2 (n) from E. coli ribonucleotide reductase, by eight organic radicals listed in Table 1 [18]. The inset shows the correlation of data for the same ORs with cytochrome c(III) [17].

(;2%). The slope of 0.54 is in satisfactory agreement with the theoretical value of 0.50. The intercept of 5.94(9) is equal to log k11qlog k22qE02/0.059 in Eq. (5). 3.3. Summary of the reduction of R2 by hydrazine [20] Relevant features with hydrazine as a reductant for R2 are described. Hydrazine can undergo single or multiple electron steps to a maximum of four (Eqs. (12) and (13)) [20,32]: ™

1/2 N 2qNH 3

y2 ey/y2 Hq

N2H4

™

S 2 O 42yq2H 2 O™2HSO 3yq2Hqq2ey

(14)

again indicating the availability of an equal number of electrons and protons as long as it is possible to utilise 2H2O molecules. Biphasic kinetics are observed for the reduction of active-R2 samples containing ;40% met-R2. Both stages exhibit saturation kinetic behaviour, the only example of saturation kinetics so far observed with R2. There is no dependence on pH. The kinetics of the first phase are reproduced in separate experiments using met-R2 only. A mechanism involving association followed by electron transfer is proposed (Eqs. (15)–(18)). This reads for k1obs K

yey/yHq

N2H4

Dithionite generally reacts as a one-equivalent reagent following rapid equilibration S2O42y|2SO2xy, which gives a rate law dependence [S2O42y]1/2. At one stage it was thought that mediators (such as MV2q) were essential for reduction of R2. The present studies are for the direct reaction without any mediators. The rate law indicates a dependence on [S2O42y], which unusually is the only contributing step. The overall reaction can be written as in Eq. (14):

(12)

Fe III2«TyrHqS 2 O 42y|Fe III2«TyrH, S 2 O 42y

(13)

Fe III 2«TyrH, S 2 O 42y™Fe II 2«TyrHq2SO 2

k et

y2 ey/y2 Hq

N2H2

™

N2

Reduction of R2, studied by conventional time-range spectrophotometry, occurs in a biphasic process involving metand active-R2 components. First-order rate constants k1obs and k2obs (hydrazine in )10-fold excess) exhibit linear dependences on [N2H4]T, from which second-order rate constants k1s55.0 My1 sy1 and k2s15.3 My1 sy1 are obtained at pH 7.10 (50 mM Tris/HCl), Is0.100 M (NaCl). From separate studies on met-R2 a uniphasic process is observed which gives the same k1 value. Reactions of hydrazine as in Eqs. (12) and (13) result in the release of an equal number of electrons and protons. The rate constant k1 can be assigned as a one- or two-electron reduction of the met-R2 component of active-R2 giving FeIIFeIII or FeII2. The second stage k2 is assigned to the reduction of activeR2. Because of the similarity of rate constants k1 and k2 (;4-

Friday Apr 07 10:49 AM

(15) (16)

and for k2obs K

Fe III2«TyrxqS 2 O 42y| Fe III2«Tyrx, S 2 O 42y

(17)

k et

Fe III2«Tyrx, S 2 O 42y™Fe II 2«Tyrxq2SO 2

(18)

Applying mass balance the expression Eq. (19) can be derived: k obss

Kk et [S 2 O 42y] 1qK[S 2 O 42y]

(19)

For k1obs association Ks330(30) My1 occurs prior to kets4.8(5)=10y4 sy1; and for k2obs, Ks800(50) My1 and kets5.6(2)=105 sy1. To make comparisons with k1 and k2 for N2H4 the ketK products are used.

StyleTag -- Journal: JIB (Journal of Inorganic Biochemistry)

Article: 6350

64

M.B. Twitchett et al. / Journal of Inorganic Biochemistry 79 (2000) 59–65

4. Discussion Bearing in mind the substantially smaller reduction potential (;1 V) for FeIII2 as compared with Tyr., the observation that the reductant ORs, N2H4, and S2O42y, give similar rate constants for reaction with active- and met-R2 is quite remarkable. In all cases it is met-R2 that reacts the faster, by an average of 66% in the case of the ORs, and by a factor of four for N2H4 and S2O42y. The similarity of rate constants for the two R2 forms (Table 1, Fig. 4) indicates that the reduction of the FeIII2 centre is a common rate-limiting step, with Tyrx of active-R2 playing a passive role. The small differences are attributed to minor FeIII2 structural changes in the met- and active-R2 forms. Subsequently, the Tyrx is reduced in a rapid intramolecular step. The Tyrx of R2 is a much stronger oxidant than the FeIII2 (y115 mV versus NHE, for 2ey change [31]). Estimates of the reduction potential for Tyrx vary from 940 mV in small peptides [34] to 400 mV for the coordinated and modified (39-S-cysteinyl-tyrosine) radical in galactose oxidase at pH 7.5 [35,36]. In the case of the ORs all the reactions conform to the Marcus equation, which therefore suggests an electron transfer process. To bring about the Tyrx™TyrH conversion both Hq and an electron are required. Electron transfer from FeII to Tyrx may require and be triggered by the arrival of Hq. The fast electron transfer from FeIIFeIII to Tyrx is supported by the high driving force, and close proximity of FeA to the ˚ phenolate O-atom of Tyrx (5.3 A). The observation that Tyrx is not reduced in the initial step becomes more significant on examining the distance of the radical from the protein surface, which is similar to that of the FeIII2 centre. Thus, the Cg of the Tyr-122 aromatic ring ˚ from Arg-127 at the surface, while FeA is 11.9 A ˚ is 12.6 A ˚ from the residue Phe-46 at the surface. Howand FeB 12.7 A ever, a specific electron-transfer pathway has been proposed (Fig. 2) to allow transfer of electrons to the Tyr-122. This pathway starts at Trp-48, as lead-in group, or alternatively at Tyr-356 which is a component of the C-terminal tail of R2. The latter, consisting of 35 residues, is disordered in the R2 crystal structure, but becomes ordered on association of R2 with R1 [27,37]. The pathway in the case of Tyr-356 then proceeds by residues Trp-48, Asp-237, His-118, FeA of the binuclear 13 centre, and via Asp-84 through to Tyr-122 (all of which are conserved). It is particularly relevant that FeA lies on the pathway to Tyr-122. The pathway can be regarded as a vehicle for both electron and proton transfer. Thus, variant R2 forms, in which amino acids in the pathway are replaced by residues which are unable to form H-bonds, severely diminish enzyme redox activity [21,22]. Furthermore, theoretical calculations indicate unfavourable energetics for electron transfer without the coupled proton transfer to maintain electroneutrality at the active site [38]. It is proposed that proton transfer is a rapid process subsequent to electron transfer, and can be envisaged as a process involving the proton travelling to the active site in the ‘slipstream’ of the electron. A similar mechanism has been proposed for the

Friday Apr 07 10:49 AM

reduction of the [3Fe–4S] cluster of Azotobactor vinelandii [39], where the electron-transfer process appears to drive the subsequent proton transfer. In the case of the OR reactions the proton is presumably provided by solvent H2O rather than the reductant. Further insight into the reduction potential for the FeIII2/ FeIIFeIII couple is provided by the inability of the organic radicals MBx, TBx and IPx2y, with reduction potentials G0.006 V, to reduce met-R2. Reduction of met-R2 with Rfx2y (y0.051 V) is observed however, leading to the conclusion that the reduction potential for the FeIII2/FeIIFeIII couple is between y0.051 and q0.006 V. This range is consistent with an estimate of E02 for the mouse R2 FeIII2/ FeIIFeIII couple of 0.010 to 0.150 V [40]. A value of E02;0 is therefore assumed as a contribution to the intercept at yE01/0.059s0 (Fig. 4). Larger E02 values may be considered with estimates of k22 decreasing by an order of magnitude for each 0.059 V. The intercept in Fig. 4 is 5.94. With k11 fixed at 1.0=106 y1 y1 M s the self-exchange rate constant k22 for the FeIII2/ FeIIFeIII self-exchange process is found to be 7.6=105 My1 sy1. This is surprisingly large for an intermolecular electron˚ (=2) separation of bintransfer process involving a 10.9 A uclear Fe centres on two R2s. The R2 pathway for electron ˚ pathway between transfer (k22) is relevant as part of a 35 A the active site of R2 and the substrate binding site of R1 (Fig. 2). Studies on the reduction of R2 solutions by hydrazine and dithionite provide further information on the reactivity of active-R2. The similarity in the rate constants for the reduction of both active- and met-R2 (;4-fold variation) again suggests a common rate-determining step. For both reductants there is the possibility that a two-electron process is contributing. Just how feasible this might be requires further evaluation, since there are no examples of long-range twoelectron transfer reactions involving metalloproteins. For inorganic reactions between two metal centres, when a closer interaction between the donor and acceptor sites is possible, only a limited number of examples have been identified. A major problem in considering long-distance metalloprotein two-electron transfer reactions remains the high reorganisation requirements, with some advantages resulting from two successive one-electron steps. In the case of R2 there is the additional problem of coupled electron and proton transfer to address. The only evidence for a 2ey step in these reactions is the rate-law dependence on [S2O42y] instead of [S2O42y]1/2. This may however relate to the involvement of H2O as a source of Hq. Further points to make are that, for the reduction FeIII2™FeII2, protonation and loss of the m-oxo group are required to give red-R2 (Fig. 1(b)). The one- and two-equivalent reactions of N2H4 give different amounts of N2 and NH3 as products (Eqs. (12) and (13)) [32]. A final answer may therefore have to await the quantitative determination of reaction products. Whereas the organic radicals conform to a linear freeenergy relationship (Fig. 4), the reactions with hydrazine

StyleTag -- Journal: JIB (Journal of Inorganic Biochemistry)

Article: 6350

M.B. Twitchett et al. / Journal of Inorganic Biochemistry 79 (2000) 59–65

and dithionite give smaller than expected rate constants. The strong one-equivalent cobalt(II) reductants [Co(sep)]2q (y300 mV) and [Co(9-aneN3)2]2q (y400 mV) (seps1,3,6,8,10,13,16,19-octaazabicyclo[6.6.6]-eicosane; 9-aneN3s1,4,7-triazacyclononane) also give rate constants some 108 smaller than required by the free-energy plot in Fig. 4 [33]. The favourable rate constants for the organic radicals suggest that a special feature such as stacking of aromatic rings with the lead-in residue (either Trp-48 or Tyr-356) may be occurring, thus allowing a greater orbital overlap for electron transfer. This type of interaction is not possible with the other reductants. In conclusion, the kinetic studies described provide evidence for a common rate-determining step (reduction of the FeIII2) in the OR reductions of met- and active-R2. The Marcus free-energy correlation establishes the electron-transfer step as rate-limiting. Arrival of a proton at the FeIIFeIII centre may then initiate reduction of the Tyrx. In the reactions of N2H4 and S2O42y it has been established that FeIII2 reduction also occurs as the first step. In the absence of stronger evidence one- rather than two-electron transfer is more likely. Cationic reductants including N2H5q and two CoII complexes react only slowly with R2, and there are large differences in rate constants as compared to those for the OR reactions. An advantage of the ORs may be their ability to stack with aromatic R2 residues prior to electron transfer.

Acknowledgements We are grateful to the UK Engineering and Physical Sciences Research Council, and the North of England Cancer Research Campaign for financial support, and to Dr G.A. Salmon (University of Leeds) for help with the pulse radiolysis experiments.

References ¨ [1] B.-M. Sjoberg, Struct. Bonding 88 (1997) 139. ¨ [2] K.K. Anderson, A. Graslund, Adv. Inorg. Chem. 43 (1995) 359. ¨ [3] A. Graslund, M. Sahlin, Annu. Rev. Biophys. Biomol. Struct. 25 (1996) 259. [4] B.J. Wallar, J.D. Lipscomb, Chem. Rev. 96 (1996) 2625. ¨ [5] B.-M. Sjoberg, Structure 2 (1994) 7930. [6] M. Fontecave, P. Nordlund, H. Eklund, P. Reichard, Adv. Enzymol. 65 (1992) 147. ¨ [7] O. Nilsson, T. Lundqvist, S. Hahne, B.-M. Sjoberg, Biochem. Soc. Trans. 16 (1988) 91.

Friday Apr 07 10:49 AM

65

[8] S.S. Mao, T.P. Holler, G.X. Yu, J.M. Bollinger, S. Booker, M.I. Johnston, J. Stubbe, Biochemistry 31 (1992) 9733. ˚ ¨ A. Ahgren, B.-M. [9] A. Aberg, S. Hahne, M. Karlsson, M. Ormo, ¨ Sjoberg, J. Biol. Chem. 264 (1989) 12249. [10] P. Nordlund, H. Eklund, J. Mol. Biol. 232 (1993) 123. [11] J. Stubbe, Annu. Rev. Biochem. 58 (1989) 257. ¨ ¨ [12] T.E. Elgren, J.B. Lynch, C. Juarezgarcia, E. Munck, B.-M. Sjoberg, L. Que Jr., Biol. Chem. 266 (1991) 19265. [13] U. Uhlin, H. Eklund, Nature 370 (1994) 533. ˚ ¨ [14] O. Nilsson, A. Aberg, T. Lundqvist, B.-M. Sjoberg, Nucleic Acids Res. 16 (1988) 4174. ˚ ¨ J. Hajdu, H. Eklund, [15] D.T. Logan, X.-D. Su, A. Aberg, K. Regnstrom, P. Nordlund, Structure 4 (1996) 1053. [16] C. Gerez, M. Fontecave, Biochemistry 31 (1992) 780. [17] C.D. Borman, A.M. Dobbing, G.A. Salmon, A.G. Sykes, J. Phys. Chem. B 103 (1999) 6605. [18] A.M. Dobbing, C.D. Borman, M.B. Twitchett, D.N. Leese, G.A. Salmon, A.G. Sykes, J. Am. Chem. Soc., in press. [19] A.M. Dobbing, J.-Y. Han, A.G. Sykes, J. Biol. Inorg. Chem. 3 (1998) 620. [20] J.-Y. Han, J.C. Swarts, A.G. Sykes, Inorg. Chem. 35 (1996) 4629. [21] B.O. Persson, M. Karlsson, I. Climent, J.S. Ling, J.S. Loehr, M. Sahlin, ¨ B.-M. Sjoberg, J. Biol. Inorg. Chem. 1 (1996) 247. ¨ D.T. Logan, A. [22] B. Katterle, M. Sahlin, P.P. Schmidt, S. Potsch, ¨ ¨ Graslund, B.-M. Sjoberg, J. Biol. Chem. 272 (1997) 10414. ¨ ¨ ¨ S. Hahne, M. Karlsson, H. Jornvall, M. Goransson, [23] B.-M. Sjoberg, B.E. Uhlin, Biol. Chem. 261 (1986) 5658. ¨ Proc. [24] T. Barlow, R. Eliasson, A. Platz, P. Reichard, B.-M. Sjoberg, Natl. Acad. Sci. USA 80 (1983) 1492. [25] G.D. Armstrong, T. Ramasami, A.G. Sykes, Inorg. Chem. 24 (1985) 3230. ¨ ¨ A. Ehrenberg, B.-M. Sjoberg, P. Reichard, [26] L. Petersson, A. Graslund, J. Biol. Chem. 255 (1980) 6706. ¨ C.Y. Huang, Biochemistry 31 (1992) [27] I. Climent, B.-M. Sjoberg, 4801. [28] G.A. Salmon, A.G. Sykes, Methods Enzymol. 227 (1993) 522. [29] A.R. Curtis, W.P. Sweetenham, FACSIMILE/CHECKMAT Users Manual, UKAEA, Atomic Energy Research Establishment, Harwell, UK, 1987, p. R12805. [30] R.A. Marcus, N. Sutin, Biochim. Biophys. Acta 811 (1985) 265. [31] K.E. Silva, T.E. Elgren, L. Que Jr., M.T. Stankovich, Biochemistry 34 (1995) 14093. [32] D.M. Stanbury, Prog. Inorg. Chem. 47 (1998) 511. [33] J.C. Swarts, A.G. Sykes, Inorg. Chim. Acta 242 (1996) 165. [34] M.R. DeFelippis, C.P. Murthy, F. Broitman, D. Weinraub, M. Faraggi, M. Klapper, J. Phys. Chem. 95 (1991) 3416. [35] G.A. Hamilton, P.K. Adolf, J. de Jersey, G.C. Dubois, G.R. Dyrkacz, R.D. Libby, J. Am. Chem. Soc. 100 (1978) 1899. [36] C.G. Saysell, C.D. Borman, A.J. Baron, M.J. McPherson, A.G. Sykes, Inorg. Chem. 36 (1997) 4520. ¨ [37] I. Climent, B.-M. Sjoberg, C.Y. Huang, Biochemistry 30 (1991) 5164. [38] P.E.M. Siegbahn, M.R.A. Blomberg, R.H. Crabtree, Theor. Chem. Acc. 97 (1997) 289. [39] J. Hirst, J.L.C. Duff, G.N.L. Jameson, M.A. Kemper, B.K. Burgess, F.A. Armstrong, J. Am. Chem. Soc. 120 (1998) 7085. ¨ [40] M. Atta, K.K. Anderson, R. Ingelmarson, L. Thelander, A. Graslund, J. Am. Chem. Soc. 116 (1994) 6429.

StyleTag -- Journal: JIB (Journal of Inorganic Biochemistry)

Article: 6350