Iron-sulphur clusters in electron transfer, catalysis

Biochemical Society Transactions

Iron-sulphur clusters in electron transfer, catalysis and control

594

Andrew J. Thomson, Jacques Breton, Simon J.George, Julea N. Butt,* Fraser A. Armstrong* and E. Claude Hatchikiant School of Chemical Sciences, University of East Anglia, Norwich NR4 7TJ,U.K., *Department of Chemistry, University of California at Irvine, Irvine, CA 927 17, U.S.A., and tCNRS-LCB., 3 I , Chemin Joseph,Aiguier-PB 781, I3277 Marseille Cedex 9, France

Introduction Iron-sulphur clusters are the redox active cofactors of proteins which function as water soluble electron carriers, exemplified by bacterial ferredoxins and high-potential iron proteins. They are also found as electron storage centres in high-molecular-mass multi-component enzymes such as hydrogenase, nitrogenase, sulphite and nitrite reductases [ 13. Typical examples contain a core of iron and acidlabile sulphide, [2Fe-2SI2+/l+and [4Fe-4S]3+/2+/I+, liganded to the polypeptide chain by four thiolate groups of cysteine. Examples are now known in which the [2Fe-2S] core is liganded by two thiols and two imidazole groups of histidine residues [2]. This appears to be the case for the Rieske protein of mitochondria1 and photosynthetic electrontransport chains. Recently examples of noncysteinyl co-ordination of the [4Fe-4S] core have also appeared [3, 41. Each of these electrontransfer/storage clusters undergo only a singleelectron redox cycle within a given protein and within a potential range of about 1.5 V. Since all coordination positions to the core cluster are occupied by protein ligands reaction is possible only with an electron and sometimes with a proton; but not with more bulky ligands, as would be required typically for direct catalysis. In 1981 [5] it was shown that some simple bacterial ferredoxins, such as the 2[4Fe-4S] ferredoxin from Clostridium pasteurziznum, under oxidative stress release one iron atom to generate a novel product identified by e.p.r. and low-temperature magnetic c.d. (m.c.d.) spectroscopy as a three-iron cluster, proposed to contain the core [ 3Fe-4SI I +/'. This was the first evidence to suggest that release of iron from a [4Fe-4S] cluster might occur readily under oxidizing conditions. The discovery led rapidly to the realization that the well-known ferrous ion activation of the enzyme aconitase corresponds to the reverse process, that is, to the reductive gain of iron(I1) to generate a [4Fe-4SI2+

core which becomes the catalytically-active centre [6]. After several years of conflict between spectroscopic and X-ray diffraction evidence, over the structure of the three-iron cluster, several recent X-ray structure determinations have vindicated the conclusions originally drawn from spectroscopic data [7]. The three-iron cluster is a cube with one iron atom missing from the corner [8, 91. Only three cysteine thiols are required to bind this core to the polypeptide. The structural basis of the three-four iron interconversion is shown in Fig. 1 and can be expressed chemically as

[ 3Fe-4SI1

+

+e-

[ 3Fe-4SI" t l WW +e[4Fe-4SI2+ + [4Fe-4SI1+

The X-ray structure determination of pig heart aconitase shows that the catalytic centre is a [4Fe-4SI2+ core ligated by only three thiolate groups and that the labile fourth iron is not co-ordinated by a protein side-chain [lo, 111. Extensive spectroscopic evidence shows that the substrate can interact directly with one iron atom of the cube; in so doing the co-ordination number of the iron atom increases by simultaneous ligation of one or two substrate oxygen atoms in addition to an oxygen atom (OH/H,O) present in the resting enzyme [ 121. Since this catalytic iron atom is readily lost, the activity of the enzyme may be regulated by the Fe(I1) activity within the mitochondrion and, possibly, by the prevailing electrochemical potential.

Fig. I

lnterconversion between [3Fe-4S] and [4Fe-4S] clusters

- Fe d 5 -

Abbreviations used: IKE-BP, iron-regulated mRNAbinding protein; IRE,iron-responsive element; Fd, ferredoxin; PGE, pyrolytic graphite edge; CV, cyclic voltammetric scan; SHE, standard hydrogen electrode.

Volume 19

0

Sulphur

0 Iron

+ Fe

Respiratory Electron Transfer Complexes

A recent report of the sequence of an ironregulated mRNA-binding protein (IRE-BP) shows regions of homology with that of aconitase [13]. This has led to the proposal that IRE-BP is an ironsulphur protein containing a [ 3Fe-4SI cluster that has a high-affinity binding site for iron(I1). By undergoing a structural change dependent upon iron binding, such a 'dynamic' iron-sulphur centre is responsive to the iron(I1) status of the cell. This is the first example of regulation of protein expression at the mRNA level. Cellular iron metabolism is controlled by regulation of mRNA molecules that encode ferritin and the transferrin receptor [ 141. Excess iron(I1) results in an increase in the translation of ferritin mRNA and a decrease in the stability of the transferrin-receptor mRNA [ 151. These regulatory events are mediated by similar sequence motifs referred to as iron-responsive elements (IRES) that exist within the 5' untranslated region of the ferritin mRNA and the 3' untranslated region of the transferrin receptor mRNA. The IRES from both mRNAs interact with a cytoplasmic protein, the IRE-BP, whose binding affinity is dependent upon the iron status of the cell. Thus it appears that the [3Fe-4SI * [4Fe-4S] transformation is involved also in control of protein expression. Such cluster interconversions have also been identified in a number of bacterial ferredoxins (Fds), particularly, Fd 111 from Desulfovibrio uficunus [3], in Fd I1 from Desulfovibrio &us [ 161 and the Fd from the extreme thermophile Pyrococmfitriosus [4]. It has been shown that metal ions beside iron(I1) can also be taken up by the vacant corner position of the [3Fe-4S] cluster. These include cobalt(I1) [ 171, zinc(I1) [ 181, nickel(I1) [ 191 and cadmium(I1) [20]. In the cases of Fd 111 (D. ufrzcanus) [21] and Fd (Pfuriosus) the amino acid sequence shows insufficient cysteine ligands for the cluster [M3Fe-4S] to be ligated by four thiolate groups. In place of the expected cysteine residue an aspartic acid group is found suggesting that carboxylate may be a ligand to the added metal ion, M. These discoveries raise a host of new questions about the chemistry of these processes and also about their possible biological roles. What are the features of the protein structure that lead to metal-ion lability from [4Fe-4S] clusters? What is the affinity of the [3Fe-4S] ligand for metal ions, M, and does iron(I1) have the highest binding affinity? How are the redox properties of the clusters affected by interconverison? Does iron(I1) lability have a role in control of enzymatic function, of electron transfer activity or of gene regulation?

Some of these problems are addressed in this paper by means of studies of two bacterial Fds, Fd I, Azotobucter chroococcum [22] and Fd 111, D. afh'cunus [23]. Both proteins contain one [3Fe4S]'+/' and one [4Fe-4SI2+/'+ core when aerobically isolated [24]. However, in the case of A. chroococcum Fd I the sequence contains nine cysteines and in the latter only seven cysteine residues. A wealth of spectroscopic evidence [25] shows that A. chroococcum Fd I is similar if not identical to that of Azotobucter vinelandii, Fd I , whose X-ray structure has recently been re-determined by two different groups [8, 91. Fd 111, D. afriunus undergoes rapid and facile uptake of various metal ions [3] at the [3Fe-4SIo cluster whereas Fd I from A. chroococcum is unreactive when highly purified. W e have developed combined electrochemical and spectroscopic methods to enable these transformation processes to be studied quantitatively [26, 271.

Experimental methodology While identification of the type and redox status of iron-sulphur clusters is difficult to achieve, due to the broad, rather featureless u.v.-visible spectra, e.p.r. and m.c.d. spectroscopy under conditions of low-temperature provide useful diagnostic signatures of the electronic spin, g-values and zero-field splitting parameters of the various cluster oxidation states [28-301. However, these methods require substantial sample and it can be difficult to manipulate proteins under critical conditions of controlled low potential. W e have discovered that specific cluster transformations in Fds can be identified and investigated by a rapid voltametric procedure that uses, in practice, a fraction of a nanomole per experiment and permits equilibrium and kinetic parameters to be obtained [20, 26, 271. The technique extends our recent discovery that Fds coabsorb with amino-cyclitols at a pyrolytic graphite edge (PGE) electrode giving a stable electroactive film [26, 261. The PGE surface is rich in acidic oxides (pK, 5.6) and interacts with proteins bearing negatively-charged surface domains in the presence of multi-charged cations such as aminosugars which provide an 'ionic bridge'. Fast, direct electron transfer can take place between the protein and the electrode surface [31]. To prepare electroactive films about 2 PI, of an ice-cold solution of Fd, typically about 100 PM in 0.1 M NaCI, mixed buffer containing about 2 mM neomycin sulphate and 100 PM EGTA were spread evenly across the surface (area 0.2 cm2) of a freshly polished, chilled PGE electrode. When transferred promptly into an

-

-

1991

595


Fig. 2 Direct, unmediated cyclic voltammograms of an adsorbed film of Fd Ill from D. afiicanus

596

( a ) in EGTA and ( b ) in presence of Zn(ll) showing the dis-

appearance of waves A‘ and c’ and the appearance of a new owing to the newly formed [Zn3Fe-4S] cluster. wave D’,

electrolyte solution also containing 2 mM-neomycin and with 10 ~ M - E G T Acyclic , voltametry over the region 0 to - 8.50 mV [versus standard hydrogen electrode (SHE)] reveals several redox couples, as shown in Fig. 2a.

-

Results and discussion

(4

Fd I (A. chroococcum) Azotobacter contain a ferredoxin, Fd I, which, when isolated aerobically from A. chroococcum and A. vinelandii, contains one [3Fe-4S] ‘ + I oand one [4Fe4SI2+ / l + cluster. Extensive spectroscopic comparisons between Fds from A. vinelandii and A. chroococcum show them to be very similar. The X-ray structure of the protein from A. vinelandii is

’

C’

B’

1[3Fe-4S]0’2-?

now solved and reveals that the [3Fe-4S] cluster is ligated by three cysteine ligands, residues 8, 16 and 49 (see Scheme 1). Cys 11 is not bound to the cluster but is involved in a salt bridge with Lys 100. The redox properties of Fd I have been almost as controversial as the X-ray structure. However, using direct electrochemistry at a PGE electrode, both as a thin film and in bulk solution, we have shown that the [3Fe-4S] cluster exhibits a pH independent E”’ value, E = - 460 k 10 mV versus SHE, above p H 8.0 [26]. Below this pH the

A’

I

[4Fe-4S]Z+oi+ [3Fe-4SIi+”

A. Chroococcum Fd I t

l

-800

l

l

-600

l

l

-400

l

i

1

-200

I

+

0

8

16

11

E/mV versus SHE

(b)

Film of 7Fe protein into 10 pM-Zn2+ 3Fe

4Fe

49 Desulphovibrio africanus Fd III

11

14

17

21

scans, each (N) CYS-X-X-ASP-X-X-CYS .......... CYS-PRO

I

3Fe

I

i

(C) GLN-GLU-CYS 51

I

Volume 19

I

4Fe

/I\

.........

CYS-X-X-CYS-X-X-CYS 47

44 Scheme 1

41


value is pH-dependent corresponding to the uptake of a proton in the reduced state with a pK, = 7.8. The relevant equilibria are

El)'

[.3Fe-4SI1++ e -

s= 112 [3Fe-4SIo+H S=2

+

+

[3Fe-4S]" s=2

E,,L=-460~10mV

[3Fe-4SIo ... H

+

pK, = 7.8

s=2

Marked changes between the m.c.d. spectra of the alkaline and acidic forms indicate that protonation occurs at the cluster or is linked directly to cluster reorganization [24]. Apart from the cluster in Rzotobucter Fd no other [3Fe-4SI1+'" redox couple has been discovered to be pH dependent. The [4Fe=4SIm+ cluster in Fd I, A chroococcum, undergoes a one-electron reduction from m = 2 to m = l at a value of E 0 ' = - 6 4 5 k 1 0 mV versus SHE at pH 8.3. There is only a slight pH dependence. This represents one of the lowest potentials determined directly for a biological ironsulphur cluster. The presence of two clusters in a relatively low-molecular-mass Fd with E" values separated by 200 mV raises the question of the function of this protein and its form in the organism. Most small iron-sulphur proteins have either a single [4Fe-4S] cluster or two [4Fe-4S] centres of almost identical EO'values (within 10 mV in the case of the 2[4Fe-4S] containing Fd from Clostridium pusteurianum).

-

-

Fd Ill, 0. ofricanus Fd 111 is one of three Fds expressed by D.ufriunus. The amino-acid sequence (6600 kDa) shows the presence of seven cysteine residues, distributed (see Scheme 1) so as to be capable of binding one [4Fe4S] cluster and one [3Fe-4S] centre [21]. Between the cysteine residues at positions 11 and 17 I'ies an aspartate group, position 14, instead of cysteine as would be found in a typical 2[4Fe-4S] Fd. The [ 3Fe-4SI" cluster incorporates Fe(I1) rapidly to form a [4Fe-4S] cluster that has novel magnetic properties. It is most likely that the entering Fe occupies the sub-site ajdacent to Asp 14 [3]. The e.p.r. and m.c.d. spectra of the 7Fe form of Fd 111 show that the [3Fe-4S] cluster has in the oxidized state a ground state electronic spin S = 1/2, with a g-value of 2.01 and an m.c.d. spectrum diagnostic of the three-iron centre in several crystallographically characterized proteins [23]. In the reduced state the [3Fe-4SIn cluster has a spin S= 2 subject to a negative, predominantly axial zero-field splitting [23]. The [4Fe-4SIm+cluster is diamag-

netic when m = 2 and has spin S = 1/2 when m = 1. On addition of iron(I1) to the three-iron cluster the resulting [4Fe-4SIZ+cluster can have co-ordination by only three cysteine and either aspartate or water. On reduction of this cluster by one electron the resulting state [4Fe-4S] has an unusual spin-state S = 3/2 subject to an axial zero-field splitting. This state has g-values of 5.27, (2.34), (1.62). The m.c.d. spectra are little altered in form by this spin-state flip. Cyclic voltammetric scans (CV) of a preformed film of D. ufkicunus Fd 111, transferred to an EGTA-containing buffer solution at pH 7, reveal well-defined waves due to three redox couples. Figure 2u. One of these, B', corresponds to the stable [4Fe-4SI2+/'+cluster. The other two, A' and C', are associated with the [3Fe-4S] cluster and assigned respectively as the normal 1 + /O couple and a chemically-reversible two-electron process of as-yet-unestablished identity. By comparative integration, and from the effect of pH upon wave positions, C' corresponds to a two-electron transfer coupled to a two-proton uptake [22]. If the coated electrode is then transferred to stirred solutions devoid of EGTA but containing low concentrations of Fe(II), Zn(I1) or Cd(II), reductive passage through couple A'( [ 3Fe-4SI +/'I) initiates rapid changes. During subsequent cycles over the course of several seconds, waves A' and C' disappear together and are replaced by a new couple D'. Values of En' (versus SHE) in the film and in bulk solution are compared in Table 1. The positions of the new couples correspond closely with cyclic voltammograms of Fd 111 undergoing transformations in the solution phase. Characterization by e.p.r. and m.c.d. spectroscopy identifies the species to be [4Fe-4SI2+/If,[Zn3Fe-4SI2+"+ and [Cd3Fe-4SI2+/'+.The clusters [Zn3Fe-4SI2+ and [Cd3Fe-4SI2+ are shown to be isoelectronic with [3Fe-4SI0 and to have a ground electronic state S = 2 subject to a negative, axial zero-field splitting. The reduced states, [M3Fe-4S]'+, have a ground state spin S = 5/2 with characteristic e.p.r. spectra. Rates and equilibria of metal-ion uptake for protein molecules confined to the electrode surface depend upon the identity and concentration of the metal ion. The height of each CV peak is proportional to the concentration of the redox active species present. Hence a study of these peaks as a function of Fe(I1) concentration enables cluster-metal dissociation constants to be determined. Values of Kd. the equilibrium dissociation constant given by {M2+){ 3Fe-4S]o)/{M3Fe-4S]2+} (see Table 1) establish an affinity order Cd2+t Zn2+%Fez+,as

'

+

'

1991

597


Table I

598

The dissociation constants, K,, and formal reduction for the heterometal clusters potentials, E,, [M3Fe-4SIZ+"+ in Fd 111, of D. ofN'conus

Eo (mV) versus SHE M2+

Kd

Film

Solut ion

Fe

30f 15 I .6 f I .O 0.8 f 0.5

-393f 10 -492f 10 -569f 10

-400f 15 -490f 15 -580f 10

Zn Cd

expected for a bivalent metal ion co-ordinated to a sulphide-rich site.

Conclusions W e have compared the cluster lability of a[4Fe-4S] centre and the metal-ion uptake ability of a [3Fe-4S] core in two low molecular mass proteins isolated from bacterial sources. In both cases the protein contains a second cluster of the [4Fe-4S] type which is liganded by the thiolate groups of four cysteine residues and which is remarkably inert to mild oxidative stress and the presence of chelators such as EGTA. It is well known that in the absence of iron-sulphur clusters the polypeptide chains of Fds d o not fold. Hence one obvious function of the second cluster is to cause the polypeptide chain to provide a conformation appropriate for the [3Fe4S] core. W e have shown that this core in itself can be a good ligand for the further binding of cations. W e attribute this to the open unoccupied face of three bridging sulphide groups. In the case of A. chroococcum Fd I, it is possible that the cluster can bind H + , or possibly H,O+, capping the triangular sulphide face in a redox-linked proton uptake. No other cations can be induced to bind to this cluster without some major perturbation of the protein structure. W e note that the cysteine residue, Cys 11, which does not ligate the [3Fe-4S] cluster is salt-bridged to the E-NH,' side-chain of Lys 100. This residue is, however, invariant in a number of Fds. If Cys 11 were required for the uptake of divalent cations it would be necessary to is break this salt-bridge. Fd 111, from D.u*unus, very different in its reactivity towards bivalent cations. W e have shown here that the cations Cd2+2 Zn2+9Fez+ bind in that stability sequence to the three-iron cluster only in its reduced state, [ 3Fe-4SIo. Hence the cation uptake process is redox linked. The role played by the aspartic acid residue

Volume 19

at position 14 is not clear at this stage. Little is known about the relative available concentrations of Fe(I1) and Zn(1I) within cells of obligate anaerobic bacteria such as D.u$-mnus. However, the possibility must be borne in mind that heteroatom-metal sulphur clusters of the type described here can have greater stability than the all-iron forms and may play a role within cells. The persistent question posed by these results is the following. Are [3Fe-4S] clusters, which have the capacity to bind additional cations in a redox linked process, simply electron-transfer proteins? In the case of 2[4Fe-4S] Fds loss of a metal ion from one centre could lead to a change from a twoelectron to one-electron carrier by switching the potential of one of the clusters out of range. Alternatively they may not be electron-transfer proteins, but could be involved in metal-ion-mediated cellular control such as gene expression. Low molecular mass ferredoxins provide interesting and tractable model systems in which to explore the chemistry of redox-linked metal-ion exchange. This work was supported by Molecular Recognition Initiative of the S.E.R.C., by the University of California and by grants from N.A.T.O. (CRG 900302), by an Exxon Education Foundation Award (FAA) and by C.N.R.S., France.

1. Thomson, A. J. (1985) in Metalloproteins (Harrison, P. M., ed.), Part I, pp. 79- 120, Verlag Chemie, Weinheim, F.R.G. 2. Cline, J. F., Hoffman, B. M., Mims, W. B., Lattaie, E., Ballou, D. P. & Fee, J. A. (1985) J. Biol. Chem. 260, 325 1-3254 3. George, S. J., Armstrong, F. A., Hatchikian, E. C. & Thomson, A. J. (1989) Biochem. J. 264,275-284 4. Conover, R. C., Park, J.-B., Adams, M. W. W. &Johnson, M. K. (1990)J. Am. Chem. SOC.112,4562-4565 5. Thomson, A. J., Robinson, A. E., Johnson, M. K., Cammack, R., Rao, K. K. & Hall, D. 0. (1981) Biochim. Biophys. Acta 637,423-432 6. Beinert, H. & Thomson, A. J. (1983) Arch. Biochem. Biophys. 222,333-361 7. George, G. N. & George, S. J. (1988) Trends in Biochem. Sci. 13,369-370 8. Stout, G. H., Turley, S., Sieker, I,. C. & Jensen, 1,. H. (1988) Proc. Natl. Acad. Sci. U.S.A. 85, 1020-1022 9. Stout, C. D. (1988)J. Biol. Chem. 263,9256-9260 10. Robbins, A. J. & Stout, C. D. (1989) Proc. Natl. Acad. Sci. U.S.A. 86,3639-3643 11. Robbins, A. H. & Stout, C. D. (1989) Proteins 5, 280-3 12 12. West, M. M., Kennedy, M. C., Beinert, H. & Hoffman, B. M. (1990) Biochemistry 29, 10526-10532


13. Roualt, T. A., Stout, C. D., Kaptain, S., Harford, J. B. & Klausner, R. D. (1991) Cell (Cambridge, Mass.) 64, 881-883 14. Theil, E. C. (1990) J. Biol. Chem. 265,477 1-4774 15. Klausner, R. D. & Harford, J. B. (1989) Science 244, 357-3 59 16. Huynh, B. H., Moura, J. J. G., Moura, I., Kent. T. A., LeGall, J.? Xavier, A. V. & Munck, E. (1980) J. Riol. Chem. 255,3242-3244 17. Moura, I., Moura, J. J. G., Munck, E., Papaefthymiou, V. & LeGall, J. (1986) J. Am. Chem. SOC. 108, 349-351 18. Surerus, K. K., Munck, E., Moura, I., Moura, J. J. G. & LeGall, J. (1987)J. Am. Chem. SOC.109,3805-3807 19. Conover, R. C., Kowal, A. T., Fu, W., Park, J.-B., Aano, S., Adams, M. W. W. &Johnson, M. K. (1990) J. Biol. Chem. 265,8533 20. Butt, J. N.. Armstrong, F. A., Breton, J., George, S. J., Thomson, A. J. & Hatchikian, E. C. (1991) J. Am. Chem. SOC.in the press 21. Bovier-Lapierre, G., Bruschi, M., Bonicel, J. & Hatchikian, E. C. (1987) Biochim. Biophys. Acta 913,20-26 22. Yates, M. G. (1970) FEBS Lett. 8,281-285 23. Armstrong, F. A., George, S. J., Cammack, R., Hatchikian, E. C. & Thomson, A. J. (1989) Biochem. J. 264, 265-273

24. George, S. J., Richards, A. J. M., Thomson, A. J. & Yates, M. G. (1984) Biochem. J. 224,247-25 1 25. George, S. J. (1986) Ph.D. Thesis, University of East Anglia, Nonvich 26. Armstrong, F. A., George, S. J., Thomson, A. J. & Yates, M. G. (1988) FEHS Lett. 234, 107-1 10 27. Armstrong, F. A., Butt, J. N., George, S. J., Hatchikian, E. C. & Thomson, A. J. (1989) FEBS Lett. 259, 15- 18 28. Thomson, A. J., Robinson, A. E., Johnson, M. K., Moura, J. J. G., Moura, I., Xavier, A. V. & LeGall, J. (1981) Biochim. Biophys. Acta 670,93-100 29. Hatchikian, E. C., Cammack, R., Patel, D. S., Robinson, A. E., Richards, A. J. M., George, S. J. & Thomson, A. J. (1984) Biochim. Biophys. Acta 784, 40-47 30. Johnson, M. K., Robinson, A. E. & Thomson, A. J. (1982) in Iron-Sulfur Proteins (Spiro, T. G., ed.), vol. 4, chapter 10, Wiley, New York 31. Armstrong, F. A,, Cox, P. A,, Hill, H. A. O., Lowe, V. J. & Oliver, B. N. (1987) J. Electroanal. Interfacial Electrochem. 217,331-366

Received 18 April 1991

Electron transfer in succinate :ubiquinone reductase and quinol :fumarate reductase J. C. Salerno Biology Department, Rensselaer Polytechnic Institute, Troy, NY I 2 180, U.S.A.

Introduction Succinate :ubiquinone reductase (SQR) catalyses the oxidation of succinate to fumarate, resulting in the reduction of ubiquinone to ubiquinol. In mitochondria and aerobic bacteria, ubiquinol donates electrons through an electron transfer system to molecular oxygen, producing water, and driving the synthesis of A T P through the creation of a proton electrochemical gradient. In some bacteria growing with fumarate as the terminal oxidant, menaquinol produced via the oxidation of low-potential substrates in turn reduces fumarate to succinate to a reaction catalysed by menaquinol :fumarate reductase (QFR). This also results in the generation of a proton electrochemical gradient of the same sign as that produced by succinate oxidation. Considering the vectorial nature of energy conservation, it is Abbreviations used: SQR, succinate:ubiquinone reductase; QFR, menaquino1:fumarate reductase; FP, a 70 kDa flavoprotein; IP, 30 kDa catalytic subunit; Qs, bound ubisemiquinone.

clear that the QFR system is not merely a version of the SQR system kinetically optimized to run in the reverse direction. SQR and QFR have been the subject of a number of recent reviews [ 1-31. Both consist of two large, relatively soluble subunits, often termed the catalytic subunits, and one or two smaller hydrophobic subunits. T h e catalytic subunits consist of FP, which is a flavoprotein of molecular mass of about 70 kDa and contains the succinate/fumarate catalytic site, and IP, which contains three ironsulphur clusters and has a molecular mass of about 30 kDa. These subunits are highly homologous in SQR and QFR enzymes. T h e hydrophobic anchor polypeptide components are not conserved between systems. In Escherichzh coli both SQR and FQR enzymes can be synthesized; both have two anchor polypeptides of similar size (about 14 m a ) , but n o significant sequence identity exists. In E. coli, the hydrophobic peptides of SQR carry a low-potential b-type cytochrome. No such cytochrome is associated with the QFR enzyme in

1991

599