Oxidation-Reduction Properties of the Two Fe4S4 ... - Semantic Scholar

3 downloads 0 Views 453KB Size Report
the oxidation-reduction properties of bacterial ferredoxins have been studied in some detail (4-lo), but the sum of the data is confused. Most of the experiments ...
Vol. 262, No. 11, Issue of April

THEJOURNAL OF BIOLOGICAL CHEMISTRY 01987 by The American Society of Biological Chemists, Inc.

pp. 5125-5128,1987 Printed in U.S.A.

Oxidation-Reduction Properties of the Two Fe4S4Clusters in Clostridium pasteurianumFerredoxin* (Received for publication, November 17,1986)

Roger C. Prince and Michael W. W. Adams From Exxon Research and Engineering, Annandale, New Jersey 08801

(16)) as redox mediators. Sodium dithionite was used as thereductant and potassium ferricyanide as the oxidant. Dithionite has a pHdependent E,,, (17), but at reasonable concentrations this is significantly below the E,,, of the H2/H+ couple. It is not a t all clear that “equilibrium” potentials measured below, or even near, the HZ/H+ couple can be considered truly at equilibrium, for they are clearly not at equilibrium with the aqueous protons. For this reason, we have made no efforts to use alternative reductants, such as electrochemically generated reduced viologens, as reductants at acid pH. All titrations involved samples taken in both oxidative and reductive Bacterial ferredoxin was first isolated from the anaerobic directions, and no significant hysteresis was detected. Neither was N,-fixing bacterium Clostridium pasteurianum in 1962 and any degradation of the Fe& clusters to Fe3S, forms by ferricyanide found to be an essential electron carrier in several electron detected. Potentials were measured with respect to a saturated calomel electrode but are reported with respect to thehydrogen electrode transfer processes (1, 2). Since then such proteins have been by assuming that the calomel electrode had a potential of +247 mV found in a variety of microorganisms, shown to contain Fe4S4 (18). clusters, and have been the subject of extensive research (Ref. Electron spin resonance spectroscopy used a Varian E-109 spec3 and references therein). Because of their vital role as elec- trometer equipped with an Oxford Instruments liquid helium cryostat.

The ferredoxin from Clostridium pasteurianumcontains two Fe4S4 clusters. In this paper we determine their oxidation-reduction midpoint potentials; we find them to be essentially identical (within 10 mV) and to have pH-independentE,,, values of -412 f 11 mV from pH 6.3 to 10.0.

tron mediators in and between various metabolic processes, the oxidation-reduction properties of bacterial ferredoxins have been studied in some detail (4-lo), but the sum of the data is confused. Most of the experiments have involved titrations of the optical absorption, which unfortunately is broad and of low extinction coefficient. Furthermore, many of the “titrations” in fact reflect only a single “equilibration” with the H2/2H+ couple via hydrogenase. Only one paper has specifically addressed the question of whether the two Fe4S4 clusters are identical in their oxidation-reduction properties, and that only at a single pH (11). In thispaper we address the oxidation-reduction properties of the two Fe4S4clusters of C. pasteurianum ferredoxin using redox potentiometry and electron spin resonance spectroscopy. This ferredoxin has proven to be typical of the class, a small protein ( M , about 6000) containing two Fe4S4clusters with low oxidation-reduction potentials (2, 3). The two Fe4S4 clusters have spin = 1h in the reduced state and are spin coupled to yield a complex esr spectrum (12). This hasallowed us to quantitate the fraction of molecules with none, one, or two reduced Fe4S4clusters at various redox potentials and thereby determinethe oxidation-reduction midpoint potential (E,,,)of each cluster independently. We find the two clusters to have essentially identical E,,, values (-412 f 11 mV), which are independent of pH over the range pH 6.3-10. MATERIALS AND METHODS

C. pasteuriunum W5 wasgrown

under Nn-fixing conditions as described previously (13). The ferredoxin was purified, essentially as described by Rabinowitz (14), except that the final crystallization step was omitted. The protein hadA3*/AZm= 0.79. Redoxpotentiometry followed the methods described by Dutton (15), using 40 p~ methyl and benzyl viologens (E,,,= -430 and -350 mV, respectively

* The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “aduertisernent” in accordance with 18 U.S.C. Section 1734 solely to indicate t.his fact.

RESULTS

Fig. 1 shows the results of a typical titration, in this case at pH 9.8 to allow complete equilibrium reduction of both clusters of the ferredoxin. The titration displays the well known spin coupling of the two clusters (12) in the fully reduced protein (& - 533 mV), but as the potential is raised this is replaced by the rhombic spectrum of a single cluster. The two spectra are shown more clearly in Fig. 2, where the spectrum of the proteincontaining only a single reduced cluster (E), -347 mv) is shown at an amplification 20-fold that of the spectrum of the fully reduced protein ( E h -495 mV). The spectra of Figs. 1 and 2 show that the spectral feature at 3097 mT’ is due solely to the fully reduced form, while that at 3194 mT is due to the singly reduced protein. This allows the independent assessment of the amount of each form at any given ambient Eh, as shown in Fig. 3. Note that theprotein cannot be fully reduced at equilibrium at the lower pH values of Fig. 3, although the datamay still be used to estimate the E,,, values for the clusters. For a situation where two redox couples are in the same protein, we may consider three thermodynamic models. In the simplest case the two centersare independent, andtheir behavior isno different from the case of twice as many molecules each bearing a single redox center. Alternatively, the reduction of one might alter the redox properties of the second, either positively or negatively; with positive cooperativity, the electrochemistry will tend toward n = 2; with negative cooperativity, the two E,,, values will become separated. The case of C. pasteurinnum ferredoxin seems to be a minor variant of the “independent centers” model, that is, the two clusters interact strongly enough to altertheir esr spectra but not strongly enough to alter their redox potentials. In such a The abbreviations used are: mT, millitesla; Mes, 2-(N-morpho1ino)ethanesulfonic acid; Tes, N-tris[hydroxymethyl]methyl-2-aminoethanesulfonic acid.

5125

Redox Properties ofFerredoxin Clostridial

5126 *Value 2.2

I

- 377mV

1.8

2.0

l

l

gValue 2.0

2.2

I

!

I

I

I

1.8 I

I

I

v I

I 0.29

0.35

0.41

I

I

I

0.30

0.35

0.40

Yagnotlc FIold (Torla)

FIG. 1. Redox titration of C. pasteurianurn ferredoxin at pH 9.8. 230 p M ferredoxin in 5 mM Tris, 50 mM cyclohexylaminopropane sulfonate, 100 mM KCl, 40 p~ methyl and benzyl viologens, pH 9.8. Spectra were measured at 11,000 with 10 milliwatts of applied power and a modulation amplitude of 1 mT. Microwave frequency 9.226 GHz.

situation, the fraction of proteincontaining only a single reduced cluster may be calculated. The Nernst equation, as it is used inredox chemistry (see Refs. 18 and 19), may be written as Eh = E ,

+ RT -.InnF

[oxidized] [reduced]

Magnotlc FIeld (Tomla)

FIG. 2. ESR spectra of the fully (doubly) reduced and singly reduced ferredoxin. Conditions as in Fig. 1, except that the buffer was 50 mM Tricine, and the pHwas 8.2.

:

s r

1 m

pH 10.0

pH 8.2

where Eh is the ambientredox potential, E,,, is the oxidationreduction midpoint potentialof the couple under study, R, T, and F have their usual meanings, IZ is the numberof electrons involved in the redox process, and [oxidized] and [reduced] are the concentrations of the oxidized and reduced species, respectively. At room temperature and pressure, this may be simplified as follows. Eh = E ,

59 + --.lo n

[oxidized] [reduced]

Let one cluster in the ferredoxin be named A, the other B, with E,,, values of EmAand Ern*,respectively, and bothbe n = 1Nernstian species. The fractionof A reduced a t any Eh may then be calculatedas [R,] = [10(Eh-Ed/59 + 11-1

while the fraction of B reduced at anyEh is [RBI = [10(Eh-E&)/59+

11-1

The fraction of protein that has both clustersreduced will then be RA.RB, while the fraction that has neither reduced will be (1 - RA). (1 - R B ) .The fraction that has only one cluster reduced will then be 1 - [ R A . R B ]- [(l- RA).(1 R B ) ] Such . behavior is described for a range of E,,,Aand EmB

-350

-550

EhmV

-550

-350

EhmV

FIG. 3. Redox titrations of ferredoxin. The datawere obtained from titrations similar to that of Fig. 1, except that the buffers were 50 mM Mes (pH 6.3), Tes (pH 7.2), Tricine (pH 8.2), or cyclohexylaminopropane sulfonate (pH 10.0). The fraction doubly reduced (F=, 0 )was measured at 3097 mT; the fraction singly reduced (F-, A) at 3194 mT.

values (expressed as their difference, EA - E B )in Fig. 4. Note that this behavior, where twoindependent centerseach accept a single electron, is quite distinct from that where a single redox species accepts two electrons, as for example, aquinone (see Fig. 5 of Ref. 20). The data reported here are from esr experiments. This spectroscopy is an excellent comparative technique, but absolute quantitationsof mixtures of spectra as complicated as those of Figs. 1 and 2 are difficult. For this reason, Fig. 4 concentrates on those featuresof esr redox titrations that are most sensitive to the difference in midpoints of two redox species such as those found in C. pasteurianum ferredoxin. The upper portion of the figure shows a typical calculation, in this case where EmA = E,B = 0 mV. The fraction of the protein that hasa single reducedcluster is maximal, at 0.5, at

Redox Properties of C ~ s t Ferredoxin ~ ~ d ~ ~ lLV0 independent n = 1 centers f

I

i

1

I

1

I

5127

any significant dependence on the concentration of the ferredoxin (Fig. 5 ) . There is a slight dependence on the ionic strength (Fig. 6), with the E, increasing with the ionic strength. The data aretabulated in Table I. DISCUSSION

The data reported in this paper show that the two clusters of the ferredoxin fromC. pasteurianum have essentially identical oxidation-reductionproperties (within 10 mV) and that these do not vary significantly as a function of pH. As an appropriate corollary of thepH independence, the E , is

I

I

-200

I

I

0

200

I

I

I

-

EmA EmB

FIG. 4. The behavior of two independent R = 1 couples as a function of the difference between their E, values. This figure is discussed in the text. The upper portion is a simulation of the fractions of the protein that are singly reduced (F-, the hump)or doublyreduced (F-, the curve that assymptoticallyapproaches 1) when both E,a and EmB are 0 mV. Three readily measured parameters, A, B , and C, are shown. The lower portion of the figure shows how these three parametersvary as the difference between em^ and EmBis varied.

FIG. 5. Redox titrations of ferredoxin as a function of its concentration. Ferredoxinattheindicatedconcentrationwas titrated as in Fig. l, except that the buffer was 50 mM sodium pyrophosphate,pH 8.9. Symbols as in Fig. 3.

0 mV, while the fraction of the protein that has two reduced clusters rises with an apparent “E,,,” at -22 mV.Of course r I I I I 1 this latter value is not the E, of either ciuster, but it is a value readily measured in redox titrations such as those in Fig. 3. Two other features are readily obtained in such experiments: the Eh of maximal singly reducedprotein, and hence the difference between this point and the apparent E,,, of the doubly reduced form ( A in Fig. 4), and the “width at halfheight” of the titration of the singly reduced protein ( B in Fig. 4). Also readily calculated, although difficult to quantitate in absolute terms in these experiments, is the fraction of protein c o n ~ i n i n gonly a singly reducedcluster (C in Fig. 4). c I I I I I 1 The threeparameters discussed aboveturn out to be sensitive .-e0 indicators of the difference in E, between the two clusters, I = 1075 m d EY as shown in the lower part of Fig. 4. Note that these calculations haveassumed that both clusters have identical esr properties, so that curue C is symmetric about 0 mV. Similar calculations can readily model alternative possibilities, such as the possibility that only one of the clusters is detectable in the singly reduced protein. The data of Fig. 3, together with those to be discussed in Figs. 5 and 6, are all fit very well by assuming that the two -500 -400 clusters of C. posteurianum ferredoxinhave identical E , EhmV values and thatthe maximal differencebetween the E, values, FIG. 6. Redox titrations of ferredoxinas a function of ionic if there is one, must be less than 10 mV. strength. 228 FM ferredoxin wastitrated as in Fig.1,except that the Fie. 3 shows that there is no sismificantDH deuendence of buffer was50 mM glycine,. UH 9.4. with and without 1 M KCl. Symbols the E,,, of the clostridial ferredoxin clusters. Neither is there as in Fig. 3. ”

Redox Properties of ~ l o s ~Fre~rdr ie~dlo x ~ ~

5128 -.

TABLE I E,,, values of Clostridium pasteurianum ferredoxin PH [Ferredoxin] I E, 6.3 7.2 8.2 8.9 8.9 8.9 9.4 9.4 9.8 125 10.0

BM

mM

mV

228 228 228 36 261 2150 228 228 228 228

125 125 125 845 845 845 75 1075 125

-415 -420 -420 -400 -403 -400 -434 -405 -405 -409

the potentials of the two clusters. They concluded that the clusters differed by 10 f 5 mV in C. pasteurianum ferredoxin and by 4 0 mV in C.acidi-urici ferredoxin. They only measured these values at pH 7.0 (actually in D20 at pD 7.4) so again were unable to fully reduce the ferredoxins at equilibrium. Our findings on the oxidation-reduction properties of the two Fe4S4clusters of ferredoxin now need to be extended to other systems where similar esr spectra suggest a pair of interacting clusters, such as in the 12-iron bidirectional hydrogenase of C. pasteurianum (21) and nitrate reductase of E s c ~ r ~ ccoli, ~ i which a contains threeor four ferredoxin Fe& clusters (22). Is the equivalence of electrochemical properties a general feature of such systems?

slightly dependent on the ionic strength, with the protein REFERENCES becoming easier to reduce as theionic strength increases. 1. Mortenson, L. E., Valentine, R. C., and Carnahan, J. E. (1962) The determined E,,, of C.pasteurianum ferredoxin (-412 & Biochem. Biophys. Res. Comrnun. 7,448-460 11 mV) is in reasonable agreement with the values of others 2. Mortenson, L. E., and Nakos, G. (1973) in Iron Sulfur Proteins (Lovenberg,W., ed) Vol. 1,pp. 37-64, Academic Press, Orlando, (4-lo), although our conclusions are somewhat different from FL those of Magliozzo et aL (10). They found that theE,,, varied 3. Spiro, T. G . (ed) (1982) Iron Sulfur Proteins Vol. 4, John Wiley from -375 mV at pH 6.5 to -438 mV at pH 8.8. These values & Sons, Ltd., Chichester, Great Britain are not very different from ours (mean of their data = -401 4. Sobel, B. E.,and Lovenberg, W.(1966) Biochemistry 6.6-13 & 24 mV, mean of ours = -412 11 mV), but ours show no 5. Tagawa, K.,and Arnon, D. I. (1968) Bhchim. Biophys. Acta 153, 602-613 convincing pH dependence. There are several differences in 6. Eisenstein, K. K., and Wang, J. H. (1969) J. Biol. Chem. 2 4 4 , the experimental approaches used in thetwo laboratories. We 1720-1728 have used equilibrium redox potentiomet~ and esr spectros7. Lode, E.T., Murray, C. L., and Rabinowitz, J. C. (1976) J. Bioi. copy; Magliozzo et al. (10) used the method of Lode et al. (7), Chem. 251,1683-1687 which involves equilibration with the hydrogen electrode via 8. Stombaugh, N. A., Sundqvist, J. E., Burris, R. H.,and Ormehydrogenase, and measuring the reduction optically. One Johnson, W. H. (1976) Bwchemlstry 1 5 , 2633-2641 9. Moulis, J-M., and Meyer, J. (1982) Biochemistry 21,4762-4771 problem with such a technique is that essentially full reduction (>99%) of the protein would only be expected when its 10. Magliozzo, R. S., McIntosh, B. A., and Sweeney, W. V. (1982) J. Bwl. Chem. 257,3506-3509 E , is more than 118 mV positive of that of the hydrogen 11. Packer, E. L., Sternlicht, H., Lode, E. T., and Rabinowitz, J. C. electrode. This was not achieved in the pH range they used, (1975) J. Bid. Chem. 2 5 0 , 2062-2072 so estimates of the fully reduced spectrum may have been 12. Mathews, R., Charlton, S., Sands, R. H., and Palmer, G. (1974) J. Biol. Chem. 249,4326-4328 significantly underestimated. Only at pHvalues above pH 8.2 do the dataof Magliozzo et al. (Fig. 1 in Ref. 10) suggest that 13. Daesch, G., and Mortenson, L.E. (1972) J. Bacteriol. 110, 103109 they reduced the ferredoxin by more than SO%, and their data 14. Rabinowitz, J. C. (1972) Methods Enzymol. 24,431-446 set only includes three pH values in this region. In contrast, 15. Dutton, P. L. (1978) Methods Enzyml. 64, 411-435 the approach to fitting the data obtained here, outlined in 16. Prince, R. C., Linkletter, S. J. G., and Dutton, P. L. (1981) Fig. 4, allows reasonable extrapolations even at pH values Bwchim. Bwphys. Acta 635,132-148 where the protein can be only partially reduced at eq~librium.17. Mayhew, S. G. (1978) Eur. J. Biochem. 85,535-547 Nevertheless, Magliozzo et al. (10) provided additional evi- 18. Clark, W. M. (1972) Oxidation Reduction Potentials of Organic Systems, Robert E. Kruger Publishing Co., Huntington, NY dence for the pHdependence of the E , by measuring proton 19. Wood, P. M. (1985) Trends Biochem. Sei. 1 0 , 106-107 binding during reduction, which although small, seemed to be 20. Robertson, D. E., Prince, R. C., Bowyer, J.R., Matsuura, K., appropriate for the pH dependence measured in their equiliDutton, P. L.,and Ohnishi, T. (1984) J. Biol. Chem. 259,17581763 bration experiments. We are at a loss to explain this part of 21. Chen, J-S., Mortenson, L. E., and Palmer, G. (1976) in Iron and their data. Copper Proteins (Yasunobu, K. T., Mower, H. F.,and Hayaishi, Our finding that thetwo clusters have essentially identical O., eds) pp. 68-82, Plenum Publishing Corp., New York E , values under all the conditions of pH, concentration, and 22. Johnson, M. K.,Bennett, D. E., M o m i n g s t ~ J. , E., Adams, M. ionic strength measured here is in reasonable agreement with W. W., and Mortenson, L. E. (1985) J. Biol. Chem. 260,54565463 the finding of Packer et aZ. (11),who used nmr to estimate

*