Proton n.m.r. of ferrodoxin from Clostridium pasteurianum

Proton n.m,r. of ferredoxin from Cbsttidium pasteurianum Lars SkjeMal, Jostein Kraoe and Torbjern Ljooes Depurtmentof Chemistry, University of Trondheim, N-705s Draguoll, Norwuy

(Received 13 April 1989; revised 18 June 1989) Proton magnetic resonancespectra at 500 MHz are reported for the oxidized and reduced forms of the q[4Fe4s ]-ferredoxin jiom Clostridium pasteurianum. The reduced protein showed additional peaks in the lOdO ppm region, which were preuiously unobserved, and there were sisnifcant diflerences between oxidized and reduced states in the whole region. The electron exchange rate in partially reduced ferredoxin is slow on the n.m.r. time scale when reduced with sodium dithionite, 6ut fat when zinc reduced methyl vioiogen is used as reducing agent. We explain the difference between fart and slow exchangeas being due to the different chemical propertiesof the two reducing agents. Keywords:Ferredoxin;Clostridiumpmteurtanunt; n.m.r. spectroscopy; electrontransfer

Introduction Ferredoxin from CIostridium pusteuriamtm contains two [4Fe-4S] clusters coordinated to cysteine residues in a cubane-like structure. The protein consists of 55 amino acids and transfers electrons among the hydrogenase, pyruvate dehydrogenase, and nitrogenase systems, undergoing reversible oxidation and reduction with a potential of about - 0.420 V. The iron-sulphur clusters transfer one electron each’v2. The presence of paramagnetic centres in ironsulphur proteins makes them attractive candidates for n.m.r. spectroscopy 3*4. Both ‘H and 13C n.m.r. studies have been performed on a number of different iron-sulphur proteins, including the C. pasteurianum ferredoxin’-I*. The ‘H spectra of C. pusteurianum and the closely related Clostridium acidi-urici ferredoxins reveal temperaturedependent, paramagnetic-shifted resonancea that were tentatively assigned to P-CH, hydrogens bound to the iron-sulphur cluster’1.12. The resonances arising from groups near the ironsulphur clusters are shifted downfield because of the paramagnetism of the clusters. This effect is observed even with the residual paramagnetism of oxidized clusters, but it is dramatically increased upon oneelectron reduction of the cluster2*4.“. Few details have been reported on the ‘H n.m.r. properties of reduced clostridial ferredoxins, and the only published ‘H spectrum was that of C. acidi-urici ferredoxin, which was 50% reduced by zinc reduced methyl viologcn’ ’ . Spectra of the fully reduced proteins have not been published, and information on ‘H n.m.r. properties of sodium dithionite reduced C. pasteurianum ferredoxin, which was reported to exhibit broad contactshifted resonances at 220 MHz, was not substantiated by published spectra”. We now report ‘H n.m.r. spectra at 500 MHz of oxidized and reduced C. pasteurianum ferredoxin. These spectra reveal much detail previously unobserved. We also report results on the dynamic n.m.r. properties of

partially reduced ferredoxin that clarify what previously 0141-Rl30/89/060~22-~~03.00 0 1989 Butterworlh & Co. (Publishers)Ltd 322

Int. J. Biol. Macromol., 1989, Vol. 11, December

appeared as a fundamental difference between interprotein electron exchange rates of various ferredoxins2.

Experimental W5 ferredoxin was purified according to the previously reported procedure”*14, and it was subjected to diafiltration against 100 mM K,DPO, pH’= 7.6 in an Amicon ultrafiltration cell using a UM2 membrane. Quoted pH* values are meter readings uncorrected for the isotope effect.

C. pasteurianum

N.m.r.

spectroscopy

n.m.r. experiments were performed at the MR centre in Trondheim. An n.m.r. sample consisted of 0.6 ml 2mM ferredoxin in 0.1 M K,DPO, pH*=7.6, sealed under an argon atmosphere with a rubber septum. Reduction ofthe protein was carried out anaerobically by injecting small amounts of a 50mt.4 stock solution of sodium dithionite through the septum, using a syringe. The stock solution of sodium dithionite was prepared by dissolving solid dithionite in deaerated and argon flushed lOOmhi K,DPO, pH* = 7.6. ‘H n.m.r. spectra were obtained on an AM 500 Bruker instrument with a resonance frequency for ‘H of 500.14 MHz. The residual solvent signal was suppressed by a presaturation pulse from the decoupler, which was turned off during acquisition of the n.m.r. signal. About 600 transients were collected for each spectrum to achieve an adequate signal-to-noise ratio. Chemical shifts were’measured relative to the HDO signal, which was assumed to be 4.76 ppm downtield from 4,4dimethyI-4-silapentancsuIphonate (DSS) at 25°C. The chemical shifts reported are relative to DSS. The spectra were baseline corrected with a spline lit procedure, which is part of the processing routine. The

Proton n.m.r, of ferredoxin: L. Skjeldal et al. d.

|

19.5

1~4 PPM

Figure 1 A, 1H n.m.r, spectrum at 500 MHz of 2 mM oxidized ferredoxin, showing the downfield contact-shifted region, 8.519.5 ppm. Temperature is 297 K. B, Half reduced ferredoxin, 1 mM sodium dithionite is added to the sample in A. C, Fully reduced ferredoxin, 2 mM sodium dithionite is added to the sample in A

Results and discussion The proton n.m.r, spectra of the fully oxidized protein contain eight resonances between 11 and 18 ppm. These eight resonances observed downfield of 11 ppm in the oxidized ferredoxin at 500 MHz (Figure 1A), appear in the same region as do resonances that are observed earlier at 220 MHz for the oxidized forms of ferredoxins from C. pas teurianum ~2, ~5 and C. acidi-uricia 6. These resonances for the oxidized clostridial ferredoxins have been assigned, tentatively, to eight of the fl-CH 2 protons of the cysteinyl groups that are responsible for attachment of the iron-sulphur clusters to the polypeptide chain. Resonances in the same region from oxidized Bacillus polymyxa ferredoxin 7 and of the reduced form of the high potential protein from Chromatium strain D 17 have also been assigned to the flCH 2-prOtOns, largely dispelling any doubts about this assignment. It is interesting to note that the general features of the contact-shifted resonances of the oxidized clostridial 214Fe-4S] ferredoxin, the oxidized [4Fe-4S] ferredoxin of B. polymyxa, and the reduced [4Fe-4S] high potential iron-sulphur protein from Chromatium D, indicate that geometrically, electronically, and magnetically these clusters are similar v,1T. The two sharp signals with smaller linewidths at 9.8 ppm (a) and 10.1 ppm (a') (Figures 1A and 2) have previously been assigned to ~t-CH protons I x. When ferredoxin was anaerobically titrated with sodium dithionite, the eight well resolved resonances

between 11 and 18 ppm were diminshed in intensity, but remained at the same positions, and new resonances appeared (Figure 1B,C). After the first few additions of sodium dithionite, the reduced state was reoxidized during the accumulation of the spectra. This was probably due to an initial amount of residual oxygen in the n.m.r, tube, but it was no problem upon further anaerobic additions. Many of the new resonances grew up close to the old ones, and some new resonances even grew up at the same chemical shifts as the old ones. This is most easily seen in Figure 1B, when the ferredoxin was about 50Yo reduced. Both the chemical shifts and the linewidths of resonances belonging to each particular state remained constant throughout the redox titration at this temperature (24°C). When ferredoxin was reduced with equimolar amounts of sodium dithionite, at least 20 resonances were observed between 8 and 65 ppm with different intensities and line widths in the 500 MHz spectrum (Figure 2). The resonances obtained upon reduction are more broadened than those observed in the oxidized state, as expected from the increased magnetic susceptibility of the reduced state 12. New resonances also appeared between 8.8 and 9.6 ppm upon reduction, probably shifted out of the aromatic and aliphatic envelope. We did not integrate the proton signals in this state because of line broadening, which increased as expected when going downfield. All of the eight resonances seen in the oxidized state in the 8-11 ppm region, as well as the two signals at 9.8 and 10.1 ppm, disappeared upon reduction with one mole of dithionite per mole ferredoxin. In the high field region there are two signals at 0.5 ppm and 0.7 ppm that also disappear upon reduction (Figure 3). In the aliphatic and, especially, the aromatic regions there were significant pertubations when the protein was reduced, but this complex region will not be considered here. The three most downfield shifted resonances that we were able to detect in the reduced protein at 24°C were found at 53, 55 and 59 ppm in the 500 MHz spectrum (Figure 2). Upon careful aeration, the n.m.r, spectrum returned to that of the oxidized state, showing that the protein was still intact and not denatured. When a large excess of sodium dithionite was added to the reduced state, the protein decomposed, and it was not possible to reoxidize the sample. We cannot be sure that the ferredoxin is 100Yo reduced, because of the uncertainty in the redox potential of sodium dithionite. Few quantitative studies on the redox potential have been published, and the agreement among the values is poor Is. A large excess of sodium dithionite had to be avoided because of the difficulties due to irreversible decomposition, but we interpret the disappearance upon reduction of the two highfield resonances at 0.5 and 0.7 ppm, together with the disappearance of the signals at 9.8 and 10.1 ppm, as an indication of full reduction of the ferredoxin. Theoretically there should be 24 contact-shifted resonances; 16 fl-CH 2 protons and 8 a-CH protons. But if the earlier assumption that the signals at 9.8 and 10.1 ppm derive from ct-CH protons 1~ is correct, then clearly some fl-CH 2 protons must be less influenced by the paramagnetic centre, at least in the oxidized state. Because both intrinsic shifts (contact and pseudo-contact

Int. J. Biol. Macromol., 1989, Vol. 11, December

323

Proton n.m.r, of ferredoxin: L. Skjeldal et al.

60 '

50 ~

4b

'

3'0

20 '

10 '

PPM

Figure 2 1H n.m.r, spectrum of fully reduced ferredoxin at 500 MHz, showing the downfield contact-shifted region, 2 mM sodium dithionite was added anaerobically to 2 mM ferredoxin. Temperature is 297K

origin) and extrinsic shifts (pseudo-contact origin) are possible in this 2[-4Fe-4S]-protein, assignments of all of the downfield shifted resonances require further experiments. Previous reports on the n.m.r, spectra of partially reduced ferredoxins containing one or two [-4Fe-4S] clusters reveal a difference in the rate of exchange of electrons between reduced and oxidized clusters (reviewed in Ref. 2). As pointed out by Sweeney and Rabinowitz 2, this cannot be explained by the previous suggestion 7 that two-cluster ferredoxins show n.m.r. spectra in the fast exchange limit due to electron transfer between clusters in the same protein molecule. Thus, the fast exchange previously observed with clostridial twocluster ferredoxins a'9'11 must be due to fast electron transfer between different protein molecules, as opposed to the much slower electron transfer between molecules of the one-cluster ferredoxin of Bacillus polymyxa 7. This effect could be due to differences between the protein environments of the clusters. However, closer scrutiny of the experimental conditions reveals an important distinction. The two-cluster proteins were reduced with zinc reduced methyl viologen, whereas the one-cluster

324

Int. J. Biol. Macromol., 1989, Vol. 11, December

protein was reduced with sodium dithionite. Reduced and oxidized methyl viologen is a well-known mediator, donating electrons to oxidized clusters and accepting electrons from reduced clusters 19, but such a mediator function is unknown for the S202-/502- couple. Moreover, the high levels of zinc used 11 could possibly also contribute to a higher rate of electron transfer, as it has been reported that divalent cations can promote electron transfer between redox-active proteins 2°. Our present report confirms the need to consider the different experimental conditions before interpretation of the dynamic n.m.r, properties of these types of ferredoxins. When sodium dithionite is used as the sole reductant and cations are excluded, n.m.r, spectra of partially reduced C. pasteurianum ferredoxin (Figure 1B) show superimposed spectra of those deriving from the fully oxidized and fully reduced forms, indicating slow exchange on the n.m.r, time scale. Thus, when the experimental conditions are similar, there is no striking difference in the rates of intermolecular electron transfer between the one-cluster and two-cluster ferredoxins. This also agrees with the published observations on the different forms of Desulphovibrio gigas ferredoxins, which

P r o t o n n.m.r, o f f e r r e d o x i n : L . S k j e l d a l et al.

have three a n d four [ 4 F e - 4 S ] clusters per aggregate, b u t nevertheless the exchange between the oxidized a n d r e d u c e d states is slow when s o d i u m d i t h i o n i t e is used 6. W e have p e r f o r m e d p r e l i m i n a r y experiments with p r o t o n n.m.r, s p e c t r a of C. p a s t e u r i a n u m p a r t i a l l y r e d u c e d with zinc r e d u c e d m e t h y l viologen a n d have in this case, a n d in a g r e e m e n t with 11, o b s e r v e d downfield resonances with p o s i t i o n s i n t e r m e d i a t e between their fully oxidized a n d fully r e d u c e d positions, i n d i c a t i n g fast exchange (data n o t shown).

i

References 1 2 3 4 5 6 7 8 9 I

10

I

11

I

I

12

I

j

13 14 15

I

16 I

17

k.

18 19 20

1'2.

' 1.0

.'8

.'6

. "4

.' 2

0

Orme-Johnson, W. H. Annu. Rev. Biochem. 1973, 42, 159 Sweeney,W. V. and Rabinowitz, J. C. Annu. Rev. Biochem. 1980, 49, 139 Ho, C., Fung, L. W.-M. and Wiechelman, K. J. Methods Enzymol. 1978, 54, 192 Philips, W. D. and Poe, M. in 'Iron-Sulfur Proteins' (Ed. W. Lovenberg), Academic Press, New York, 1973, II, p. 255 Chan, T.-M. and Markley, J. L. Biochemistry 1983, 22, 5982 Moura, J. J. G., Xavier, A. V., Brushi, M. and Le Gall, J. Biochim. Biophys. Acta 1977, 459, 278 Philips, W. D., McDonald, C. C., Stombaugh, N. A. and OrmeJohnson, W. H. Proc. Natl. Acad. Sci. USA, 1974, 71, 140 Packer, E. L., Sternlicht, H. and Rabinowitz, J. C. Proc. Natl. Acad. Sci. USA 1972, 69, 3278 Packer, E. L., Sternlicht, H. and Rabinowitz, J. C. Ann. N.Y. Acad. Sci. 1973, 222, 824 Packer, E. L., Sternlicht, H., Lode, E. T. and Rabinowitz, J. C. J. Biol. Chem. 1975, 2,50, 2062 Packer, E. L., Sweeney, W. V., Rabinowitz, J. C., Sternhcht, H. and Shaw, N. J. Biol. Chem. 1977, 2,52, 2245 Poe, M., Philips, W. D., McDonald, C. C. and Lovenberg, W. Proc. N.Y. Acad. Sci. 1970, 65, 797 Rabinowitz, J. C. Methods Enzymol. 1972, 24, 431 Skjeldal, L. and Ljones, T. J. lnorg. Biochem. 1988, 33, 227 McDonald, C. C., Philips, W. D., Lovenberg, W. and Holm, R. H. Ann. N.Y. Acad. Sci. 1973, 222, 789 Poe, M., Philips, W. D., McDonald, C. C. and Orme-Johnson, W. H. Biochem. Biophys. Res. Comm. 1971, 42, 705 Philips, W. D., Poe, M., McDonald, C. C. and Bartsch, R. G. Proc. Natl. Acad. Sci USA 1970, 67, 682 Mayhew, S. G. Eur. J. Biochem. 1978, 85, 535 van Dijk, C., van Eijs, T., van Leeuwen, J. W. and Veeger, C. FEBS Lett. 1984, 166, 76 Armstrong, F. A., Hill, H. A. O. and Walton, N. J. FEBS Lett. 1982, 145, 241

PPM Figure 3 Upper: 1H n.m.r, spectrum at 500 MHz of 2mM oxidized ferredoxin, showing the highfield region. Lower: 1.8 mM sodium dithionite added to the sample. Temperature is 297K

Int. J. Biol. M a c r o m o l . , 1989, Vol. 11, D e c e m b e r

325