Nitrogenase. VIII. Mössbauer and EPR spectroscopy. The MoFe protein component from Azotobacter vinelandii OP

Biochimica et Biophysiea Acta, 400 (1975) 32-53 © Elsevier Scientific Publishing Company, Amsterdam - - Printed in The Netherlands

BBA 37091 N I T R O G E N A S E . VIII. MOSSBAUER A N D EPR SPECTROSCOPY. THE MoFe PROTEIN C O M P O N E N T FROM A Z O T O B A C T E R

V I N E L A N D I I OP

E. MONCK', H. RHODESb, W. H. ORME-JOHNSON ~, L. C. DAVISc, W. J. BRILLd and V. K. SHAH a "Freshwater Biological Institute, University of Minnesota, Navarre, Minn., bDepartment of Physics, University of Illinois, Urbana, IlL 61801, CDepartment of Biochemistry, aDepartment of Bacteriology, and c.~The Center for Studies of N2 Fixation, College of Agricultural and Life Sciences, University of Wisconsin-Madison, Madison, Wisc. 53706 (U.S.A.) (Received January 3rd, 1975)

SUMMARY We have studied the molybdenum-iron protein (MoFe protein, also known as component 1) from Azobacter vinelandii using MiSssbauer spectroscopy and electron paramagnetic resonance on samples enriched with 57Fe. These spectra can be interpreted in terms of two EPR active centers, each of which is reducible by one electron. A total of four different chemical environments of Fe can be discerned. One of them is a cluster of Fe atoms with a net electronic spin of 3/2, one of them is high-spin ferrous iron and the remaining two are iron in a reduced state (probably in clusters). The results are as follows: Chemical analysis yields !1.5 Fe atoms and 12.5 labile sulfur atoms per molybdenum atom; the molecule contains two Mo atoms per 300 000 daltons. The EPR spectrum of the MoFe protein exhibits g values at 4.32, 3.65 and 2.01, associated with the ground state doublet of a S = 3/2 spin system. The spin Hamiltonian /c/ = D ( S 2 _ 5/4 ÷ 2(S~ -- $2)) + go~ S . / t fits the experimental data for go -- 2.00 and 3. -- 0.055. Quantitative analysis of the temperature dependence of the EPR spectrum yields D / k ---- 7.5 °K and 0.91 spins/molybdenum atom, which suggests that the MoFe protein has two EPR active centers. Quantitative evaluation of M6ssbauer spectra shows that approximately 8 iron atoms are associated with the EPR active centers. At temperatures above 20 °K these iron atoms give rise to one quadrupole doublet; at lower temperatures magnetic spectra, associated with the ground electronic doublet, are observed; at least two magnetically inequivalent sites can be distinguished. Taken together the data suggest that each EPR center contains 4 iron atoms. The EPR and M6ssbauer data can only be reconciled if these iron atoms reside in a spin-coupled (S = 3/2) cluster. Under nitrogen fixing conditions the magnetic M6ssbauer spectra disappeared concurrently with the EPR signal and quadrupole doublets are observed at all temperatures. The data suggest that each EPR active center is reduced by one electron. The M6ssbauer investigation reveals three other spectral components characteristic of iron nuclei in an environment of integer or zero electronic spin, i.e. they reside in complexes which are "EPR-silent". One of the components (3-4 iron atoms)

33 has M6ssbauer parameters characteristic of the high-spin ferrous iron as in reduced rubredoxin. However, measurements in strong fields indicate a diamagnetic environment. Another component, representing 9-11 iron atoms, seems to be diamagnetic also. It is suggested that these atoms are incorporated in spin-coupled clusters.

INTRODUCTION Nitrogenase is an enzyme complex capable of reducing dinitrogen (N2) to ammonia. Both a molybdenum-iron protein (MoFe protein) and an iron protein (Fe protein) are required for this reduction. The former binds the reducible substrate, while the latter supplies low potential electrons in the presence of MgATP to effect the reduction. There are a number of recent reviews covering the chemical composition and enzymology of nitrogenase and its components [1-4]. Low temperature EPR spectroscopy [3] has been an essential tool in many of the recent studies of the nitrogenase mechanism. In particular, the characteristic EPR signal of the MoFe protein (with g values at 4.3, 3.7 and 2.0) has proven useful. This signal decreases during the functioning of nitrogenase, i.e. when both components are incubated anaerobically in the presence of MgATP and low potential electrons from $2042-, to a steady-state level determined by temperature of incubation, ratio of protein components, and the ATP/ADP ratio. Because the nitrogenase protein components contain iron, M6ssbauer spectroscopy of isotopically substituted proteins should yield information on the involvement of iron in the catalytic mechanism of nitrogenase. In M/Sssbauer spectroscopy the S7Fe nucleus is a probe whose interactions with surrounding electrons give structural and chemical information. The strength of the method is that a M~ssbauer spectrum can be observed regardless of the spin or valence state of the iron. This is of particular advantage for investigations of the MoFe protein since this technique allows simultaneous observation of all iron present in the sample. EPR spectroscopy is limited to paramagnetic states which do not account for all of the iron present in the MoFe protein. We will show that a combined EPR and M6ssbauer investigation of the same sample yields more information than is obtainable by either technique alone. Two M/Sssbauer investigations of nitrogenase of Klebsiella pneumoniae have been reported [5, 6] using the MoFe protein of Azotobaeter vinelandii we have obtained two classes of spectra. There are spectral components showing paramagnetic hyperfine structure at low temperature, which we will show are due to the EPR visible centers, representing approximately 40 ~ of the total iron. In addition there are three distinct spectral components with integer or zero electronic spin which are only accessible by MiSssbauer spectroscopy. Using the combined results of EPR and M6ssbauer spectroscopy of the MoFe protein as isolated, and in the fixing state, it is possible to develop a spin Hamiltonian for the paramagnetic center(s) and to make an approximate quantitative assignment of iron to the different classes of observed M~Sssbauer-detectable centers. Our assignments can only be approximate because of limitations in chemical analytical methods and uncertainties in molecular weight and subunit composition of the MoFe protein, not because of limitations in either the EPR or M6ssbauer techniques. Considering

34 the difficulties in purification and characterization of MoFe protein from all sources, it is encouraging that there is good agreement between our spectra and those obtained by Smith and Lang using K. pneumoniae [5]. We first present the chemical and EPR characterization of the MoFe protein, next the relevant spin Hamiltonian for the EPR active center and then the results of M/Sssbauer investigation. From the combination of these three, we finally make most probable assignments ot the iron in the MoFe protein. MATERIALS AND METHODS Proteins The MoFe and Fe protein components of nitrogenase from A. vinelandii OP were prepared as previously described [10]. When proteins enriched in 57Fe and 9SMo were required, the bacteria were cultured in acid washed 201 carboys, using media prepared from analyzed reagents and providing an isotopic dilution of less than 10 %. The 95Mo and 57Fe were purchased from Union Carbide Corporation and were enriched in excess of 90 %. The Mo was supplied to the bacteria as molybdate and the Fe as ferric chloride. Iron [11 ] and molybdenum [12] analyses were conducted by a wet ashing procedure, labile sulfur was determined according to Siegel [13], and protein concentration was determined by the biuret reaction [14]. Activity was determined as described [10] and specific activity was expressed as nmol C2H2 reduced per min per mg protein. Methods EPR spectroscopy was performed at X-band with a modified [15] Varian V4500, with provisions for 100 K H z modulation and operation in the temperature range 6-40 °K using helium boiloff gas as the coolant. Temperatures were routinely monitored via a calibrated carbon resistor below the sample in the gas stream. When accurate temperatures (+0.1 °K) were needed a 0.057% Fe + Au vs chromel P thermocouple [16] was employed in contact with the top of the sample in the EPR cavity. In experiments aiming to quantitate the g -----4.32, 3.65, 2.01 signal complex, spectra were obtained at fixed temperatures, and integrated to yield the absorption curve. The absorption curve was fitted to a linear baseline and integrated a second time to yield the integrated intensity. A similar procedure was applied to a 2 mM solution of Cu 2÷.EDTA complex. Both integrated intensities were corrected for amplifier gain and g value anisotropy by the method of Aasa and Vanngard [17], and the numbers were compared to yield the net number of spins in the MoFe protein. The EPR and M6ssbauer samples were prepared by incubating the required mixture in serum capped bottles rendered anaerobic by flushing with N2 or Ar (02 < 0.5 ppm) and evacuation in alternating cycles. The mixtures were transferred either to EPR tubes [18] and then frozen in a stirred isopentane bath at 130 °K or to a glass chamber containing the M6ssbauer cuvette which was seated in a brass cooling bar communicating through a rubber sleeve to a liquid nitrogen bath. In either case, the apparatus was rendered anaerobic by gas flushing and evacuation prior to introduction of the sample. Freezing took less than 10 s in either case and parallel samples were obtained for critical experiments so that EPR and M6ssbauer spectra could be obtained on the same mixture.

35 The M~Sssbauer spectrometer was of the constant acceleration type. A 45 mCi source of STCo in rhodium was used which gave a minimum observable line width of 0.22 mm/s. The system was calibrated with a metallic iron absorber; all isomeric shifts are listed relative to this standard. A variable-temperature cryostat (Model 8DT from Janis Research Co.) was used for most of the measurements. The temperature of the sample was measured with either a carbon resistor or a thermistor; both sensors were embedded in the frozen protein solution. Magnetic fields of 360 G were applied by mounting a permanent magnet around the tail section of the cryostat; a transverse field of 6.6 kG was obtained with a Varian electromagnet. Measurements in 13.3 k G parallel field were performed in separate system containing a superconducting magnet; for these measurements both the source and the absorber were kept at 4.2 °K. RESULTS AND DISCUSSION Before developing the results of the M/Sssbauer investigations it is relevant to describe in some detail the chemical composition and EPR properties of the MoFe protein with which we are working (Table I). TABLE I COMPOSITION OF THE 57Fe MoFe PROTEIN USED FOR MOSSBAUER AND EPR QUANTITATIVE EVALUATIONS Specific activity = 1650 nmol of acetylene reduced per min (excess Fe protein); the observed range is 1640-1670 for crystalline preparations. See ref. 10. Component

Content

Mo Fe S= EPR

2 gatom/3 •10s g protein 11.5 gatom/gatom Mo 12.8 gatom/gatom Mo 0.91 eq (S = 1/2) spins/gatom Mo

The metal composition of the MoFe protein component is not exactly known, for a variety of reasons which stem from the oxygen lability of the protein and from the fact that nitrogenase components have to be isolated from mixtures rich in other iron-sulfur proteins, including the ferredoxins. Such proteins associate relatively strongly with the component being isolated. These experimental facts have made it difficult to estimate accurately either molecular weight or Fe, Mo or labile sulfur contents. Recent published reports on MoFe proteins from K. pneumoniae [7], Clostridium pasteurianum [8] and A. vinelandii [9] suggest that a molecular weight of approx. 250 000 and 2 Mo, 24 Fe and 24 labile S per molecule is reasonable. However, the analytical methods used certainly will not determine metal or labile S composition to better than one part in 20, and the possibility of contamination by other iron proteins or adventitiously bound iron makes this an upper limit in precision for the iron analysis. We can somewhat simplify the problem by considering properties as being related to one Mo atom, assuming that one complete set of metal centers lies in each half of the molecule.

36 Because oxygen inactivated species of the MoFe protein do not show the g = 4.32, 3.65, 2.01 signal complex, we can expect that unless stringent exclusion of oxygen is maintained during the purification of the protein, less than stoichiometric amounts of spins will be recovered from EPR integration experiments. When this approximate number is compared to the similar approximations for metal contents and decomposition of M6ssbauer spectra, we will present our results as whole numbers in spite of uncertainties in analytical methods, so that we can present the most probable assignments of species according to the present data.

1. The perthlent spin Hamihonian describing the EPR results The rather unique g values of the MoFe protein can be derived from an S = 3/2 spin system [17]. We will show that our results can be described adequately in the framework of the spin Hamiltonian

/~e(S=3/2)=

I

D[S~--~S(S t l) t-2(S~-- S~)]--gofl/q.3"

(i)

The first term describes the fine structure of the spin quartet, the zero-field splitting. The second term expresses the interaction of the electronic magnetic moment gofl~S with an applied magnetic field /4; we assume this interaction to be isotropic and take the g value of a free spin, go ~- 2.00. In zero-applied field the spin quartet is split in two Kramers doublets (see Fig. 1), separated by an energy A = 2D x/1 + 3J.-2 (without loss of generality one can choose 0 -~< ), :~ 1/3). For applied fields H such that

gx

g~j

gz

0.34

0.32

6.0

s=y

Fig. 1. Energy level d i a g r a m resulting from Eqn (I) for D . - 0. The effective g values for each K r a m e r s doublet are given for ;t = 0.055.

goflH ~ JA ] it is convenient to describe the magnetic properties of each doublet separately by using a spin Hamiltonian with an effective spin S' = 1/2,

Be (S' = 1/2) = fl ~"~'/4

(2)

The principal components of the g-tensor in Eqn (2) may be computed from Eqn (1) for each doublet. For small values of 2, i.e. for small rhombic distortions, the two

37 d o u b l e t s have d r a s t i c a l l y different E P R properties. O n e d o u b l e t gives rise to an intense E P R signal, the o t h e r d o u b l e t is a l m o s t E P R silent. F o r the f o r m e r d o u b l e t the g values m a y be c o m p u t e d with gx = g0(2 32 - 3/23. 2) gy = g0(2 + 33. -- 3/23. 2) gz = g0(l -- 33.2)

(3)

( F o r 2 < 0.15 these a p p r o x i m a t e f o r m u l a s give the g values to within 0.03.) The values go = 2.0 and 2 -- 0.055 yield g , = 3.66, gy = 4.32 and gz = 1.98 is r e a s o n a b l e agreement with the observed g values for the M o F e protein from A. vinelandii'. F o r the q u o t e d value o f 3. the o t h e r K r a m e r s d o u b l e t has g values o f g~ = 0.34, gy = 0.32 and gz = 6.0. F o r this d o u b l e t the resonances at gx and gy occur at too high a magnetic field to be accessible with n o r m a l EPR s p e c t r o s c o p y ; the transition p r o b a b i l i t y at gz = 6 is p r o p o r t i o n a l to g~ and g~ and thus vanishingly small c o m p a r e d to the t r a n s i t i o n p r o b a b i l i t y o f the other doublet. The effective g values o f the two K r a m e r s d o u b l e t s d e p e n d only on one o f the coefficients (2) describing the zero-field splitting and as long as the c o n d i t i o n goflH ~ I A I is fulfilled, the g values are insensitive to the m a g n i t u d e o f D. E P R spectroscopy, however, can be used in an indirect way to measure D. The EPR a b s o r p t i o n arising from each o f the K r a m e r s d o u b l e t s o f an S = 3/2 system is p r o p o r t i o n a l to the spin p o p u l a t i o n in each K r a m e r s d o u b l e t and these p o p u l a t i o n s are governed by a Boltzm a n n d i s t r i b u t i o n . Thus by measuring the intensity o f the E P R spectrum as a function o f t e m p e r a t u r e , the zero-field splitting m a y be d e t e r m i n e d provided the EPR signal can be observed at t e m p e r a t u r e s (kT) c o m p a r a b l e to the zero-field splitting. F r o m o u r M t i s s b a u e r investigation (see below) it was a p p a r e n t that the two d o u b l e t s are s e p a r a t e d by at least 6 °K and that D > 0; i.e. the d o u b l e t giving rise to the observed E P R signal is the g r o u n d state.

2. EPR results W h e n we d e t e r m i n e d the integrated signal intensity o f the 57Fe M o F e protein over the t e m p e r a t u r e range 8-25 °K, we were able to fit d a t a from 8-15 °K to a B o l t z m a n n d i s t r i b u t i o n for w h i c h A = 15 °K. This is illustrated in Fig. 2. A b o v e 15 °K, the spectra b r o a d e n r a p i d l y and are not integrable (they do not return to the baseline). W h e n we c a l c u l a t e d the p o p u l a t i o n o f the g r o u n d state d o u b l e t at 0 °K from this B o l t z m a n n curve, we o b t a i n e d a spin c o n c e n t r a t i o n in the s a m p l e equal to 0.91 times the measured M o c o n c e n t r a t i o n . Thus, we t h i n k t h a t there is 1 E P R center per M o a t o m in the M o F e p r o t e i n and therefore two centers per molecule.

3. M6ssbauer spectroscopy T h e p r i m a r y objective o f the M/Sssbauer investigation was to establish whether iron a t o m s are associated with the EPR active centers. Before presenting the d a t a we briefly discuss how such an a s s o c i a t i o n can be a c c o m p l i s h e d u n a m b i g u o u s l y . In M 6 s s b a u e r s p e c t r o s c o p y o f S7Fe-containing p r o t e i n s one generally encounters * Agreement can be improved by assuming a small anisotropy in the Zeeman term (gl'., g.t) in Eqn (1); for gl = 2.03 and g , = 2.0 the expressions (3) yield g, - 3.66, g~ - 4.32 and gz = 2.01.

38 IO0

--- ~ ' ~ * ~ ..... ~

~--~->>T

~'''~.~.~-~

= 14,8°K

75

ZX /- --<
Ii.

0

uJ 50 x 25

$ 5

%

t5

~_o

TEMPERATURE °K Fig. 2. Temperature dependence of the integrated EPR signal of 57Fe substituted MoFe protein of A. vinelandii. The enzyme was purified as indicated in Methods and transferred to an EPR tube and

frozen in liquid N~ in the presence of 1 mM dithionite. Spectra were obtained under the following conditions: microwave frequency 9.21 GHz, microwave power 1 mW, modulation frequency 100 KHz, modulation amplitude 10 G, magnetic field sweep rate 500 G/min, time constant 0.25 s. The temperature was varied from 7 to 20 °K by varying the flow rate. The temperature of the sample was monitored with a gold-iron thermocouple and integrations of signal area were performed as indicated in Materials and Methods, using 2 mM CuEDTA solution as a standard. Above 15 °K, the signal could not be integrated accurately.

two types o f spectra: q u a d r u p o l e d o u b l e t s or m a g n e t i c spectra. F o r i r o n in a p a r a m a g netic c o m p l e x with a half-integer electronic spin one can observe, at low temperatures, magnetically split M 6 s s b a u e r spectra associated with a K r a m e r s doublet. Associated with each K r a m e r s d o u b l e t is an electronic g-tensor and a m a g n e t i c hyperfine tensor A. F o r magnetically isotropic systems or for systems with m o d e r a t e a n i s o t r o p i c s o f the g- a n d A - t e n s o r s the intensities o f the spectra with magnetic splitting d e p e n d quite strongly on the direction o f a weak a p p l i e d field relative to the M 6 s s b a u e r r a d i a t i o n . (In practice one applies a magnetic field o f a few h u n d r e d gauss either parallel or p e r p e n d i c u l a r to the observed ),-radiation.) In this context the observed g values o f the M o F e p r o t e i n can be considered to be m o d e r a t e l y a n i s o t r o p i c and the intensities o f a M 6 s s b a u e r s p e c t r u m associated with this K r a m e r s d o u b l e t should be quite sensitive to the direction o f an a p p l i e d field. F o r K r a m e r s d o u b l e t s for which two g values are very small c o m p a r e d to the third one the M 6 s s b a u e r spectra are quite insensitive to the direction o f the a p p l i e d field. F o r such d o u b l e t s it is extremely difficult to observe an E P R signal since the m a t r i x elements o f Sx and Sy, i.e. the E P R transition probabilities, are quite small within the electronic doublet. This a r g u m e n t w o u l d a p p l y for the u p p e r K r a m e r s d o u b l e t o f the S = 3/2 state o f the M o F e p r o t e i n (for details a b o u t m a g n e t i c spectra see L a n g [19] and M f n c k et al. [20]). In c o n t r a s t to the spectra o f half-integer spin states the M 6 s s b a u e r spectra o f iron in complexes with zero or integer spin systems ( n o n - K r a m e r s systems) lack any magnetic features (unless a strong magnetic field is a p p l i e d ) ; they consist o f simple q u a d r u p o l e doublets. N o n - K r a m e r s systems are in general n o t a m e n a b l e to E P R

39 spectroscopy and the M6ssbauer effect becomes a particularly useful tool capable of probing the environment of the iron atoms. (EPR signals for the high-spin ferrous system (S = 2) have been observed only in a few instances.) Here, Mtissbauer spectroscopy can serve two functions. It can elicit information about iron which is EPR silent in all steps of a catalytic reaction, or it can tell us about the state of iron after it has become EPR silent as a consequence of a redox reaction. In the MoFe protein we will encounter both situations. If iron is involved in the EPR active center of the MoFe protein, one expects a M6ssbauer spectrum with a magnetic pattern constrained by the properties of the electronic spin system. To describe such a spectrum we may augment the spin Hamiltonian Eqn (1) by terms describing the magnetic hyperfine interaction and the electric quadrupole interaction of the 5VFe nucleus, [1 (S -- 3/2) :.. [1e -- .40 5. i l- [1o -- g, ft, H. 7

(4)

with eQVz~

[1o-- 4 1 ( 2 I - - 1)

•(31~z -- l(I--i- l)-. ~?(12 - 12)).

In Eqn (4) we have assumed that the magnetic hyperfine interaction is isotropic (A0). H o describes the interaction of the electric field gradient tensor (principal components Vxx, Vyr, Vzz) with the nuclear quadrupole moment Q; ~ -- (v,,x - Vyy)/Vzz is the asymmetry parameter. In the absence of magnetic hyperfine interaction the quadrupole splitting d E o = (eQ vzJ2)( vii + -} ~2) is observed. The last term in Eqn (4) describes the interaction of the applied field with the nuclear magnetic moment g, ft, 7. As for the electronic Zeeman interaction we describe the magnetic hyperfine interaction in the effective spin formalism. This yields [1(S' ---- 1/2) ---- f l S ' . ~ . f f l

~,- ~ ' . . 4 . 1 -

[t 0 - g , fl, ~1.-]

(5)

From Eqns (4) and (5) it follows that the following relations hold within the doublet gx

gy

gz

go

A~

Ay

A~

Ao

(6)

Thus, since the principal components of the g-tensor are known from EPR, the magnetic hyperfine tensor in Eqn (5) is determined except for a scaling factor. This factor can easily be determined from the total magnetic splitting of the M6ssbauer spectrum. Eqn (6) is widely used for the high-spin ferric ion; the relation holds when both the Zeeman term and the magnetic hyperfine interaction ill Eqn (4) are isotropic, and when f l H << D. One might also expect to detect such nuclear hyperfine effects in the EPR spectrum of the MoFe protein. Fig. 3 shows that this is indeed the case. Comparison of the g = 2.01 region of the 56Fe and 57Fe enriched MoFe proteins shows a broadening of 7 G in the STFe substituted protein. The broadening effect was not observed in the

40 56 Fe

' ZB-Go~s;-7

/

35

Gauss _.~

Fig. 3. EPR spectra of MoFe protein of A. vinelandii with either S6Fe or STFe. Enzyme purified as indicated in Materials and Methods, transferred to EPR tubes and frozen in liquid N2 in the presence of I mM dithionite. EPR spectra obtained as indicated in Fig. 2 except that microwave power was 3 roW, modulation amplitude was 5 G, magnetic field sweep rate was 100 G/rain, time constant was 0.5 s. This is an expanded view of the g = 2 region of the spectrum.

g = 4.32 or g - 3.65 regions o f the spectrum, nor did a 9SMo enriched protein show any detectable b r o a d e n i n g in any region o f the spectrum.

4. M6ssbauer results for the native MoFe protein We have studied a STFe enriched s a m p l e o f the M o F e protein with M 6 s s b a u e r s p e c t r o s c o p y in the t e m p e r a t u r e range from 1.5 to 150 °K. We recorded spectra in zero magnetic field and in magnetic fields o f 360 G applied parallel and p e r p e n d i c u l a r to the 7 - r a d i a t i o n ; a total o f 15 spectra were recorded. Before and after the M 6 s s b a u e r investigation the s a m p l e was checked with EPR spectroscopy. As expected for a protein c o n t a i n i n g so m a n y iron a t o m s the M 6 s s b a u e r spectra are quite complex. T h e spectra taken at 1.5 ~'K (Fig. 4) consist o f a s u p e r p o s i t i o n o f three q u a d r u p o l e d o u b l e t s and a spectral c o m p o n e n t showing p a r a m a g n e t i c hyperfine structure. Two o f the d o u b l e t s are readily recognized. One a p p e a r s in the center o f the spectra (we label it D); it has a q u a d r u p o l e s p l i t t i n g A E o = 0.81 mm/s and an isomeric shift 6 = 0.64 mm/s (all isomeric shifts are quoted relative to Fe metal). The o t h e r d o u b l e t (we label it Fe z+ d o u b l e t ) has a b s o r p t i o n lines at --0.9 mm/s and +2.1 mm/s. The values A E o = 3.02 m m / s and ?~ 0.69 mm/s suggest high-spin ferrous iron. A l t h o u g h 6 is quite small for a high-spin ferrous iron there is strong s u p p o r t for this a s s i g n m e n t from M 6 s s b a u e r studies on the p l a n t - t y p e ferredoxins (/) = 0.58 mm/s at 4.2 °K for p u t i d a r e d o x i n [20]) and reduced r u b r e d o x i n (6 = 0.65 mm/s at 77 °K [21]). F o r the latter proteins high-spin ferrous iron has been established u n a m b i g u o u s l y .

41 0.0

i

i

l

i

i illll~lliiiiiiiikllillllllll~l~I iIi i i it111}LIIIIIIIII~II IIit II I1"*111

i

, iI

1.0

i

i,",l, i

q

0

ii

i

I1,~,I ,

iI I~ i i I )111 illl II Ii I i

i II

i i i

2.0

li I il I i i

ii ii

oJ u

Z

i i i

O.O '

',.,,,%

b LO

i

i~l,bllllllL"~"l

i

1.0

1

'%,

,,,,,

2.0

II

¢,,,,," ,, ,1,

Eill~ I

3.0

u..o

5.0

i

,

6.0 I

I-q.o

I-3.0

-2.0

I

-1.0

VELOCITY

O.O

, , 'I 1.0

' 2.0

'3 .0

' q.O

IN (MM/SEC)

Fig. 4. M6ssbauer spectra of the native MoFe protein taken at 1.5 °K in a magnetic field of 360 G applied (a) parallel and (b) perpendicular to the M6ssbauer radiation. The velocity scales in Figs 4 and 5 are plotted relative to a S7Co (Rh) source, kept at room temperature.

Moreover, in these proteins iron atoms reside in a distorted tetrahedral environment. Taking into account that the MoFe protein contains an equal number of iron and acid labile sulfur atoms, the quoted M6ssbauer parameters suggest that the Fe E+ doublet represents iron coordinated (probably tetrahedrally) to sulfur. Both quadrupole doublets, with the same values for AEQ and 6, were found for the M o F e proteins from K. pneumoniae (labeled M4 and M5 by Smith and Lang [5]). There is a third quadrupole doublet in the spectra displayed in Figs 4a and 4b. We found this component when we substracted the doublet D from the data; two aborption lines appear at --0.2 mm/s and -91.2 mm/s. This spectral component (labeled S) accounts for one iron atom with M~ssbauer parameters AEQ ~ 1.4 mm/s and 6 ~ 0.6 mm/s. It is tempting to assume that S represents an impurity, however, we noticed this species also in the spectra from K. pneumoniae, although Smith and Lang do not mention S explicitly. The remainder of the low temperature spectra in Figs 4a and 4b is a spectral component showing paramagnetic hyperfine structure. The intensities of this spectrum depend on the orientation of the applied magnetic field relative to the 7-radiation; this is quite apparent at velocities of -- 1.3 mm/s and + 1.8 mm/s. For future reference

42 we label this spectrum M and we will show that it represents iron associated with clusters of iron atoms giving rise to the observed EPR signal. Fig. 5 shows a M6ssbauer spectrum taken at 30 °K. As in the spectra taken at 1.5 °K, the doublets labeled D, Fe 2÷ and S are present, and fortunately for our further analysis, the absorption lines of these spectral components occur at the same velocities at both temperatures. In addition, a strong quadrupole doublet has appeared at velocities of v = --0.1 mm/s and v = +0.7 mm/s; this doublet is M with the magnetic hyperfine interaction averaged out due to fast electron spin relaxation. This is in accord with the EPR results (see Fig. 2); above 15 °K the EPR lines are significantly broadened indicating that the electron spin relaxation rate is considerably faster than the nuclear precession frequency ( ~ 10 MHz). The total absorption in the M doublet 0.0

. ~j i

i

i

I

I

I Fe +z

S

I--

1.0

lZ w

2.0

n~ u o_ Z

3.0

F-

W "~

q.O

S.O

MEPR DI

I._

I

I

1. . . .

L__

I

I

I ....

[

-q.O

-3.o

-z.O

-i.0

0.0

1.0

z.O

3.0

q.O

VELOCITY IN (MMISEC)

Fig. 5. M6ssbauer spectrum of the native MoFe protein taken at 30 °K. The solid line is the result of least-squares fitting four quadrupole doublets to the data; the results are quoted in Table II. accounts for about 8 iron atoms and it is amazing that so many iron atoms should give rise to one well-defined doublet. We have subtracted the spectrum taken at 30 °K from the 1.5 °K data. The difference spectrum taken this way should contain only the component M, since the contributions of the doublet D, Fe 2÷ and S cancel. The difference spectrum shows the magnetic M spectrum superimposed with an "inverted" quadrupole doublet, the M spectrum at 30 °K. The "inverted" doublet is symmetric and has reasonably sharp lines (about 0.3 mm/s full width at half maximum); we find AEQ-~ 0.76 mm/s and ~ = 0.40 mm/s. We will give further evidence for the ap-

43 pearance o f only one quadrupole doublet associated with M when discussing the results obtained for the fixing mixture. The M6ssbauer spectrum in Fig. 5 exhibits four quadrupole doublets, three o f which are present in the 1.5 °K spectra, while M shows a magnetic pattern at 1.5 °K. Since we k n o w the positions and widths o f all absorption lines quite well, a leastsquares fitting procedure is appropriate for further quantitation. We have fitted four quadrupole doublets to the spectrum shown in Fig. 5, allowing the intensities, positions and line widths to be free parameters but requiring that b o t h lines of a doublet have the same intensities. The results o f the fitting procedure are shown in Table I I ; the theoretical curve is shown as a solid line in Fig. 5. The errors quoted for the intensities are estimated; the fitting procedure, o f course, gives unreasonably small errors since enough parameters are available for a satisfactory fit. In translating these intensities into the n u m b e r o f iron atoms contributing to each spectral c o m p o n e n t we assume that the recoilless fraction (the Debye-Waller factor) is the same for each component.

TABLE II MOSSBAUER RESULTS ON THE NATIVE MoFe PROTEIN AT 30 °K The numbers in parentheses give the error in the last significant digit. Isomeric shifts are quoted relative to Fe metal. Spectral component

AEQ (mm/s)

6 (mm/s)

Percent of total absorption

D Fe 2+ M S

0.81 (3) a 3.02 (2) b 0.76 (3) c 1.4 (1)

0.64 (3) 0.69 (2) 0.40 (3) 0.6 (1)

42.5 4- 3 14 ::k 1 38.5 :k 3 ~ 5

a AEQ is temperature independent up to 160 °K. b AEQ = 2.94 mm/s at 160 °K. c AEQ decreases to 0.67 mm/s at 160 °K.

Chemical analysis o f the sample gave a ratio o f m o l y b d e n u m :iron, 1:11.5; i.e. 23 iron atoms per molecule. Using the results f r o m Table II we have 8.8 iron atoms associated with spectral c o m p o n e n t M. The true n u m b e r o f iron atoms should be even, since the E P R results imply the presence o f two centers. Within the margins o f error 8 or 10 iron atoms are compatible with our data. We favor 8 iron atoms for two reasons. First, the chemical analysis might overestimate the iron content (possible S6Fe contaminants). Secondly, an assignment o f eight iron atoms aesthetically m o r e appealing, since it suggests 4 iron atoms per EPR-active center. (The k n o w n ironsulfur clusters contain two or four iron atoms.) One conclusion regarding the E P R centers can already be drawn at this stage o f the analysis: the E P R and M6ssbauer data can only be reconciled if the iron atoms are spin-coupled, otherwise the E P R signal should quantitate to 4 spins/molybdenum, atom. Furthermore, the presence o f a half-integer electronic spin (S = 3/2) implies an odd n u m b e r o f electrons in each cluster. If we have four iron atoms, they must be inequivalent or another paramag-

44 netic atom (Mo?) must be involved. Figs 4a and 4b clearly show the presence of magnetically inequivalent iron atoms. To analyze the magnetic components in the spectra taken at 1.5 °K we have subtracted the spectral components D, Fe z+ and S from the data shown in Fig. 4 using the results of the least-squares fitting procedure. The remaining magnetic spectra are displayed in Fig. 6. These spectra are a superposition of at least two components, reflecting inequivalent magnetic sites. One of the magnetic components in M is identified rather easily; its shape and the change of the intensities with the direction of the applied magnetic field is predicted from Eqn (5) and Eqn (6). This spectrum extends from --1.3 mm/s to -51.7 mm/s; the overall magnetic splitting (which is quite small), determines A0 in Eqn (4) and therefore the A-tensor in Eqn (5). The quadrupole splitting is known. Using a

l%llll4 t lllll4bllllllllllt ,

13.0

i

IIlU~

IllirlIIllll# !

0.q

j1I lI'

0.8

i I

1

1! I

It

1.2

ftllljFI

~lllI i

f

0.0

# Illillllllllllllllllllll# II llllll I 0.u,

i

IIIIll '

i

0.13

i el 4 I I ~Olll~Ilrlllll llil IIpll rI

1.2

II1~111 III

1.6

II~Jiillt

I

|

t

1

I

T

r

I

r

-q.o

-3.0

-2.0

-1.0

0.0

1.0

2.0

3.0

q.O

VELOCITY

IN (MM/SEC)

Fig. 6. Low temperature spectra o f spectral component M in (a) parallel and (b) perpendicular applied field. Spectra were obtained from those shown in Fig. 4 by removing components D, Fe 2+ and S using the results quoted in Table II. The sharp little peak at --0.1 mm/s is probably due to an incorrect removal of the left line of component S. The displayed spectra contain at least two distinct components. For one of them a computer simulation was attempted (see text). The second magnetic spectrum shows distinct features around --2 mm/s and for velocities > 2 mm/s.

45 computer program for the calculation of paramagnetic Mi3ssbauer spectra for polycrystalline samples (for a detailed discussion see Mi~nck et al. [22]), we have generated the theoretical curves shown in Figs 6a and 6b. The following set of parameters was used for the computation: gx = 3.65, gy ---- 4.32, gz = 2.01, A0 = 4.0.10-4 c m - ' (12.1 MHz), AE o -- --0.78 ram/s, and ~ = 0.6. The value for A0 is unusually small suggesting that the electron spin is delocalized; the magnetic splitting of this spectral component is about half the splitting bound for the plant ferredoxins [20, 23], for which the electronic spin is delocalized over two iron atoms. The computer simulations also show that the quadrupole coupling constant is positive, i.e./1E o :> 0. Resonance absorption at --2.0 mm/s and at velocities ~ 2 mm/s shows the presence of a second magnetic spectrum. However, this component does not show enough features to attempt a computer simulation, but we think that its computation requires a fairly anisotropic magnetic hyperfine interaction in Eqn (4). The fact that the magnetically different iron sites in the EPR observable complex have the same quadrupole splittings is probably a coincidence. A similar situation is found for the oxidized high-potentiai iron-sulfur protein from Chromatium (a paramagnetic complex consisting of four iron atoms). At 77 °K only one quadrupole doublet is observed; the data taken at 1.3 °K, however, suggest inequivalent iron sites [24]. The presence of magnetically inequivalent iron sites was recently confirmed by electron nuclear double resonance spectroscopy [25]. It is quite clear that magnetic M6ssbauer spectra convey more detailed information than spectra exhibiting quadrupole doublets only. We have evidence from M/3ssbauer spectroscopy that the zero-field splitting of the spin S - 3/2 multiplet is positive and larger than 6 °K. At temperatures for which the upper doublet is appreciably populated one would observe two magnetic (electron spin relaxation time assumed to be long) M/~ssbauer spectra, one spectrum associated with each Kramers doublet. The intensities of these spectra are governed by the Boltzmann factor e -a/kT where A is the zero-field splitting of the spin multiplet. To check whether the two magnetic spectra shown in Fig. 6 originate from two Kramers doublets we measured Mi3ssbauer spectra at 1.5 °K and 4.2 °K. No changes were observed for the relative intensities of the two spectral components. This observation yields d / k ~ 6 °K. We noticed, however, that the absorption lines are a little bit sharper at 1.5 °K than at 4.2 °K, suggesting that the electron spin relaxation time at 4.2 °K is not sufficiently long to observe a fully developed magnetic spectrum. To summarize, the M6ssbauer data show that approximately four iron atoms are associated with each of the two EPR active centers. The spectra taken at 1.5 °K show two paramagnetic M/3ssbauer spectra, which are associated with the lowest Kramers doublet of the S = 3/2 spin system, implying that the spin-coupled iron cluster contains magnetically inequivalent iron atoms. So far we have described the S = 3/2 spin system assuming that flH <
46 dicular to the 7-radiation. C o m p a r i s o n of the spectrum with the spectrum shown in Fig. 4b showed that the c o n d i t i o n flH ~ A is still well fulfilled at H : 6.6 kG. This finding is in accord with E P R observations by Smith et al. [26] (on the M o F e protein from K. pneumoniae) who state that the g values at 35 G H z differ little from those observed at 9 GHz. These experiments rule out the possibility that A/k < 1.5 ~'K, thus strengthening our belief that the observed decrease of the EPR signal at temperatures above 4.2 °K is caused by the p o p u l a t i o n of the upper K r a m e r s doublet of the S = 3/2 system and not due to some other low lying electronic state. One crucial experiment can further establish the correctness of the association of the magnetic M spectrum with the E P R active centers. As published earlier by O r m e - J o h n s o n et al. [27] and as mentioned above, the EPR signal of the M o F e protein vanishes in the fixing mixture, i.e. when M g A T P , a proper reductant and the Fe protein are added to the M o F e protein. (We refer to these conditions as fixing conditions (Fig. 7)). If the disappearance of the EPR signal implies that the EPR active centers have been transformed by a one electron reduction process, then the iron clusters will have integer or zero electronic spin and the associated M6ssbauer spectrum will exhibit q u a d r u p o l e interactions only. This is indeed observed.

A I

U

Fig. 7. EPR spectra of STFe MoFe protein of A. vinelandiiduring fixing conditions. A fixing reaction was prepared by adding 900,ul of 25 mg/ml STFe MoFe protein, to 400td purified Fe protein at 18 mg/ml. One half of this mixture plus 10td 1 M dithionite was added to 200/~1 of the fixing mixture containing 210 mM creatine phosphate, 21 mM ATP, 40 mM magnesium chloride and 8.5 mg/ml creatine kinase in Tris/chloride, pH 7.4. After quickly mixing these at 4 ~C in a serum bottle, the entire sample was withdrawn into a Hamilton syringe. 250 tll was added to an anaerobic EPR tube and the remainder to a M6ssbauer cuvette, as described in Materials and Methods. A: Samples were frozen approximately 45 s after preparation. EPR spectra obtained as in Fig. 2 except microwave power was 3 mW and magnetic-field sweep rate was 1000 G/rain. Sample temperature was 13 'K. B: Essentially the same as A but the sample was allowed to incubate with no added dithionite 30 min at room temperature before transfer to EPR tube or M6ssbauer cuvette. Spectrum obtained under same conditions as A.

47

5. MB spectroscopy for the MoPe protein in the fixing mixture M6ssbauer spectra of MoFe protein in N2 fixing conditions were recorded in the temperature range from 1.5 °K to 150 °K. At 4.2 °K the samples were measured in applied fields of 360 G, 3.3 kG, 6.6 kG and 13.3 kG. Fig. 8 shows a spectrum taken at 4.2 °K in a magnetic field of 360 G applied perpendicular to the ),-radiation. Comparison of the spectrum in Fig. 8 with those shown in Fig. 4 yields two results: First, that the magnetic spectrum M has vanished and it appears as a quadrupole doublet in Fig. 8 and second, that the spectral components labeled D and Fe 2+ present themselves in the same spectroscopic state as in the native protein. It appears that component S is also present as in the native preparation; however, the spectrum in Fig. 8 is quite complex and a decomposition into individual doublets is not as reliable as described for the spectra taken on the native protein. There is a close similarity of the spectrum shown in Fig. 8 with the corresponding spectrum obtained by Smith and Lang [5] for the MoFe protein from K. pneumoniae.

i

;

i

i

.....

1-----

i

i

i

r 3.0

I q.O

0.0

O°S

I

iI

~1 I

I~

I

i I

lil

I I I

I I I

1.0

I I

1.5

I

I z.o

thl

L

i I I I

tI I

I

I

I

I

-q.0

-3.0

-2.0

-1.0

I 0.0

I 1.0

I 2.0

VCI_~CITY

IN [Mr~/sECI

Fig. 8. M6ssbauer spectrum of the S~Fe enriched MoFe protein in the presence of the Fe protein, MgATP and dithionite (fixing mixture). The spectrum was taken at 4.2 °K ; the velocity scale is plotted relative to sTCo (Rh), kept at 4.2 °K.

It is quite obvious that the spectral components D and FC ÷ did not change when the MoPe protein was added to the components comprising the fixing mixture. As a first step in decomposing the spectrum, we subtracted the Fe '÷ component and found that Fe 2+ has to have 14~o of the total intensity for a complete removal. This percentage agrees quantitatively with the result found for the protein as isolated (see Table II). Next we subtracted D and S using the results from Table II. The remainder

48 should be M as it a p p e a r s (in the E P R silent state) in the fixing m i x t u r e ' . After subtraction we are left with a well-defined a b s o r p t i o n line at --0.25 mm/s and an a p p r e ciably b r o a d e n e d a d s o r p t i o n band at T 0 . 8 mm/s. This observation suggests that M a p p e a r s in the fixing mixture as a s u p e r p o s i t i o n o f q u a d r u p o l e doublets with splittings ranging from 0.95 to 1.3 mm/s. The isomeric shift has increased slightly suggesting again that each iron cluster associated with the E P R signal has been reduced by one electron. (A two electron reduction would yield again a c o m p l e x with h a l f - i n t % e r electronic spin; the associated M 6 s s b a u e r spectrum would exhibit p a r a m a g n e t i c hyperfine structure at low t e m p e r a t u r e s and an E P R signal would very likely be observed.) We were intrigued by the o b s e r v a t i o n that the M o F e protein a p p a r e n t l y has three times as m a n y iron a t o m s associated with D as with Fe 2÷ (see Table ll). Does this mean that the M o F e protein contains a n o t h e r type o f 4Fe-4S cluster consisting o f three iron a t o m s o f the D type and one iron o f the Fe 2÷ t y p e ? We have p o i n t e d out a b o v e that the q u a d r u p o l e splitting A E o and the isomeric shift 6 o f c o m p o n e n t Fe z+ closely match those found for r u b r e d o x i n (especially the value found for 6 suggests that iron is b o u n d to sulfur ligands). If spectral c o m p o n e n t Fe z÷ represents isolated single iron a t o m s like in rubredoxin, we can expect a Mt~ssbauer spectrum showing p a r a m a g n e t i c hyperfine structure u p o n a p p l y i n g a strong magnetic field. (The highspin ferrous iron has S = 2 and the term ~ . , ~ . T c a n be observed in a strong magnetic field; for a detailed discussion see E. MiJnck and P. C h a m p i o n [28].) F o r r u b r e d o x i n the a p p l i c a t i o n of a 1 k G field induces, at 4.2 °K, an internal magnetic field o f app r o x i m a t e l y 200 k G (see Fig. 8 o f reference [21]). The Fe 2÷ c o m p o n e n t o f the M o F e protein behaves quite differently when a strong field is applied. W e have studied the s a m p l e in the fixing mixture at 4.2 °K in fields o f 6.6 and 13.3 kG. The following conclusions can be d r a w n from these m e a s u r e m e n t s : Both Fe 2÷ and D behave like iron in a d i a m a g n e t i c complex (S = 0); we see no evidence for an induced internal magnetic field. W e have c o m p u t e r simulated the spectra for D and Fe 2÷ assuming /~e = 0 and S = 0 in Eqn (4). The predicted line b r o a d e n i n g (due to the nuclear Zeeman term) agrees perfectly with the observed spectra. We emphasize that these measurements d o not conclusively prove d i a m a g n e t i s m ; a very large electronic zero-field splitting o r small c o m p o n e n t s o f the magnetic hyperfine tensor can give the a p p e a r a n c e o f diamagnetism. However, the indication that Fe 2. and D reflect iron a t o m s in a d i a m a g n e t i c e n v i r o n m e n t suggests that these a t o m s reside in p o l y n u c l e a r complexes, and that the observed d i a m a g n e t i s m is a result o f spin-coupling**. * In substracting Fe2÷, D and S we took into account that the EPR signal had decreased to about 10~, i.e. 90% of the intensity of species M appears as doublets in the fixing mixture while the remaining 10~ reflect the ERP active state giving rise to a spectrum like that shown in Fig. 7. The intensity of the magnetic component is too weak to be observable in Fig. 8. However, at 30 °K the magnetic spectrum has collapsed into a doublet as discussed in the preceding section, and the corresponding absorption lines can be observed. We have measured the sample at 30 ~K and subtracted the spectrum from the data shown in Fig. 8. The result is a difference M6ssbauer spectrum containing two components --the paramagnetic M spectrum as shown in Fig. 7 and the same iron atoms observed as a doublet at 30 °K. This doublet, representing the eight iron atoms in the EPR active centers, has two fairly sharp absorption lines, reinforcing our claim that all iron atoms associated with the EPR active centers display the same quadrupole splitting. *" We have some evidence that D and Fe2÷ can be brought into a more oxidized state. Both components seem to oxidize concomitantly and appear as magnetic M6ssbauer spectra with an overall splitting of about 8.6 mm/s. These spectra seem to be identical with spectral component M 1 of K. pneumoniae obtained when the sample was oxidized by Lauth's Violet (see Fig. 3 of ref. 5).

49 In contrast to the spectral components D and Fe z+ the species M, in the fixing mixture, seems to be paramagnetic, i.e. the application of an external magnetic field induces an internal magnetic field. This suggests that the iron atoms in the EPRactive center reside in a complex (in the N2 fixing mixture) with integer electronic spin (S ~> I); i.e. S = 3/2 c- S >~ 1 However, the decomposition o f t h e spectra taken in strong applied fields is extremely difficult and further studies are required. The concentration of enzyme in the sample used for Fig. 8 was only half that of the native protein (legend, Fig. 4), hence the decrease in relative intensity. Allowing the nitrogenase to exhaust its source of electrons (Fig. 7b) showed complete restoration of the EPR signal and the reappearance of M at the appropriate intensity for enzyme at this concentration. The spectral components Fe z+, D and S were also present in the reductant depleted reaction mixture. SUMMARY AND CONCLUSIONS Our M~Sssbauer investigation has uncovered four species of iron in the MoFe protein. In descending order of concentration we have D, M, Fe e+ and S. After the rather detailed discussion in the preceding section we summarize our findings (Table

III). For the EPR-active centers a reasonably consistent picture has emerged. The observed g values at 4.32, 3.65 and 2.01 arise from the ground state doublet of an S = 3/2 spin system. The energy of the associated excited state doublet was inferred from the temperature dependence of the integrated EPR signal intensity of the ground state doublet. The knowledge of the position of the excited state allowed a quantitative estimate of the EPR signal; we found 1.82 spins/mol. This number is substantially higher than the value 0.5 spins/mole reported for the MoFe protein from Clostridium pasteurianum [18]. However, the signal from Clostridium was quantitated at 20 °K assuming that only the ground doublet is populated; this assumption leads to a sizable underestimation of the spin concentration assuming that A for this protein is near that observed for the A. vinelandii protein. From our results we conclude that the MoFe protein contains two EPR active centers. Our M~Sssbauer results show that iron is involved in the EPR-active centers. Taking into account the uncertainties connected with a quantitative evaluation of the M6ssbauer data and allowing a reasonable uncertainty in the total number of Fe atoms we found 8-10 iron atoms associated with the EPR active centers. These data are consistent with the suggestion that each center consists of four iron atoms (one of the "magic" numbers so far found for simple iron-sulfur proteins). Above 20 °K all eight iron atoms appear as a single, well-defined quadrupole doublet. Thus, no different environments can be distinguished from quadrupole and isomeric shift data alone. However, at lower temperatures paramagnetic hyperfine structure is observed and at least two magnetically inequivalent iron environments can be distinguished. The small magnetic splittings indicate that the unpaired spin (S = 3/2) must be delocalized over many atoms. The magnitudes of the observed splittings agree quite well with those found for the high-potential iron-sulfur [29] protein, a protein having a 4Fe-4S cluster. It is possible that the EPR-active centers in nitrogenase are a modification of those fundamental structures found in high potential iron protein [30] and clostridial ferredoxin [31]. The combined information obtained here from the quanti-

50 TABLE III SPECTRAL COMPONENTS OBSERVED IN THE MoFe PROTEIN The fact that spectral component S quantitates to one iron atom does not imply that S represents a single isolated iron atom. It could represent, for instance, one iron atom in a cluster of inequivalent iron atoms. It is clear from the data that spectral component M represents two isolated clusters of iron atoms; such a statement, however, cannot be made, at the present time, about components D, Fe2+ and S. The absorption of the three spectral components could result either from one kind of iron cluster with inequivalent iron atoms or from physically separated structural units. Spectral component

Number of Fe atoms/molecule

Conditions of observation

Assignments and remarks

M

8-10

Native protein and reductant depleted fixing mixture

Reduced form of M

8-10

Fixing mixture

D

9-11

All conditions investigated

Fe 2÷

3--4

All conditions investigated

S

~1

Probably present under all conditions investigated

Most probable number of Fe atoms is 8; associated with two EPR active centers of spin S = 3/2; spin-coupled clusters consisting of 4 iron atoms; irons are indistinguishable from quadrupole interactions, however, at least two magnetically inequivalent species; magnetic spectra associated with ground doublet of S = 3/2 spin system. Quadrupole doublets appear concurrently with the disappearance of the EPR signal. Spectra represent more reduced form of spectral component M. Clusters of four iron atoms spincoupled to S ~> 1. Almost temperature independent quadrupole splitting; isomeric shift suggests reduced iron; irons seem to be in diamagnetic complex (S = 0). Could be low-spin ferrous iron, but participation in spin-coupled clusters more likely. AEQ and S very similar to reduced rubredoxin; high-spin ferrous in character, with sulfur coordination. However, behaves unlike rubredoxin in strong magnetic fields. Seems to be incorporated in diamagnetic complex. Possibly involved in spin-coupled clusters, either among themselves or with irons associated with component D. AEQ ~ 1.4 mm/s and S ~ 0.6 mm/s indicates low-spin ferrous iron. Could be contaminant, but present also in preparations from K. pneumoniae [5].

t a t i o n o f the E P R m e a s u r e m e n t s and f r o m M 6 s s b a u e r spectroscopy implies a spinc ou p l ed system (S = 3/2). Th e m a g n e t i c hyperfine interactions observed in the M 6 s s b a u e r spectra can be used to estimate the line b r o a d e n i n g o f the E P R line at g = 2.01 for 57Fe enriched material. U s i n g the q u o t e d value for A0 o f the c o m p u t e r simulated spectrum (the one with the smaller m a g n e t i c splitting) we expect a line b r o a d e n i n g o f a b o u t 5 G at g = 2.01, in good ag ree m e n t with the E P R results (7 G). Th e lack o f hyperfine b r o a d e n i n g f or 95Mo enriched material suggests that m o l y b d e n u m is n o t involved in the E P R active centers. (Participation o f m o l y b d e n u m in spin-coupling, however, could reduce

51

the 95Mo hyperfine coupling sufliciently to make the observation of hyperfine broadening in an EPR spectrum of a polycrystalline specimen virtually impossible.) The magnetic M6ssbauer spectrum, M, associated with the EPR active centers disappeared concurrently with the EPR signal of the MoFe protein when the Fe protein, MgATP, and dithionite (N2-fixing mixture) were present. In the fixing mixture only quadrupole doublets are observed at temperatures down to 1.5 °K implying that the iron atoms associated with the EPR active centers reside in complexes with integer electronic spin. Data taken in strong magnetic fields suggest S >~ 1. Besides the iron associated with the EPR active centers, the M~ssbauer investigation reveals three additional iron species, which we have labeled D, Fe z+ and S. These iron atoms did not undergo spectroscopic changes when the protein was brought from the native state into the fixing mixture. From the isomeric shifts associated with these spectral components we conclude that these iron atoms are in a reduced form. Quadrupole splitting and isomeric shifts of component Fe 2÷ strongly suggest high-spin ferrous iron (S = 2) coordinated to sulfur ligands. Yet, components Fe 2÷ and D seem to be incorporated into a diamagnetic complex (S --- 0). Furthermore, D and Fe 2÷ appear to be in the ratio of 3 :I. Do D and Fe 2÷ reflect another type of 4Fe4S cluster consisting of 3Fe atoms of spectral type D and one Fe atom of spectral type Fe 2+ ? Spin-coupling within such a cluster could explain the observed diamagnetism. Spectral component S represents a single iron atom, possibly in a low-spin ferrous state. We would have been inclined to associate S with an impurity. However, this component is present also in the MoFe protein from K. pneurnoniae [5]. Whether S is an essential feature of the MoFe protein or whether it is a common impurity in both preparations remains to be established. To date, only the M species has unambiguously been shown to undergo reversible spectroscopic changes when passing from one enzymatically relevant state to another. We do not believe that D, Fe z+ and S are inert ingredients of the protein but only that they remain reduced in the states which we have examined. We need to find reversible redox processes by which to alter these iron atoms. One approach may be the use of some potent inhibitors of N2 fixation such as CO, NO, N 3 or C N - to interdict electron flow and partition the electron transferring centers into oxidized and reduced centers by analogy to the Chance crossover theorem. Such experiments are in progress and indeed, the use of CO has revealed new paramagnetic centers in inhibited, functioning nitrogenase (Davis, OrmeJohnson, Bmris, unpublished, and in ref. 4). If we assume sequential electron transfers with rate constants varying by orders of magnitude, the lack of apparent changes in the majority of iron species during the enzyme turnover is reasonable. That is, the rate limiting step in nitrogenase function is something other than a redox change of the majority of iron atoms being observed in M6ssbauer spectroscopy. Another approach is to try to uncover chemical differences rather than physical ones in the environment of the irons being observed here. It is clear from our results that unmodified clusters of the type found in simple ferredoxins, or synthetically prepared by Holm and coworkers [32, 33] are not likely to be the origin of all the iron species observed by ourselves and Smith and Lang [5].

52 ACKNOWLEDGMENTS

We would like to thank Dr Helmut Beinert for the use of his EPR apparatus, and Mr W. D. Hamilton for assistance with the EPR spectroscopy. We thank Drs J. Peisach and W. Blumberg for originally suggesting the interpretation that the MoFe protein signal arises from an S = 3/2 system, and for an early experiment aimed at measuring D/k in the i.5-8 °K region. We also thank Dr Peter Debrunner for critical comments about the physics of these systems. This work was supported by U.S. Public Health Service Grants GM 16406, GM 17170, by Research Career Development Award K04-GM 70683 (E.M.), and by National Science Foundation Grant GB 36787. REFERENCES I Dalton, H. and Mortenson, L. E. (1972) Bacteriol. Rev. 36, 231-260 2 Eady, R. R. and Postgate, J. R. (1974) Nature 249, 805-810 3 Burris, R. H. and Orme-Johnson, W. H. (1974) Bacterial Iron Metabolism (Nielands, J. B., ed.), pp. 187-209, Academic Press, New York 4 International Symposium on Nz Fixation: Interdisciplinary Discussions June 3-7, 1974, Pullman, Washington 5 Smith, B. E. and Lang, G. (1974) Biochem. J. 137, 169-180 6 Kelly, M. and Lang, G. (1970) Biochim. Biophys. Acta 223, 86-101 7 Eady, R. R., Smith, B. E., Cook, K. A. and Postgate, J. R. (1972) Biochem. J. 128, 655-675 8 Tso, M.-Y. W. (1974) Arch. Microbiol. 99, 71-80 9 Kleiner, D. and Chen, C. H. (1974) Arch. Microbiol. 98, 93-100 10 Shah, V. K. and Brill, W. J. (1973) Biochim. Biophys. Acta 305, 445-454 11 Van de Bogart, M. and Beinert, H. (1967) Anal. Biochem. 20, 325-334 12 Clark, L. J. and Axley, J. H. (1955) Anal. Chem. 27, 2000-2003 13 Siegel, L. M. (1965) Anal. Biochem. 11, 126-132 14 Gornall, A. G., Bardawill, C. J. and David, M. M. (1949) J. Biol. Chem. 177, 751-766 15 Beinert, H. and Palmer, G. (1965) Adv. Enzymol. 27, 105-198 16 Rosenbaum, R. N. (1965) Rev. Sci. lnstrum. 39, 890-899 17 Aasa, R. and Vanngard, T. (1970) J. Chem. Phys. 52, 1612-1615 18 Palmer, G., Multani, J. S., Cretney, W. C., Zumft, W. G. and Mortenson, L. E. (1972) Arch. Biochem. Biophys. 153, 325-332 19 Lang, G. (1970) Q. Rev. Biophys. 3, 1 20 Miinck, E., Debrunner, P. G., Tsibris, J. C. M. and Gunsalus, I. C. (1972) Biochemistry 1 I, 855863 21 Rao, K. K., Evans, M. C. W., Cammack, R., Hall, D. O., Thompson, L. C., Jackson, P. J. and Johnson, C. E. (1972) Biochem. J. 129, 1063-1070 22 Miinck, E., Groves, J. L., Tumolillo, T. A. and Debrunner, P. G. (1973) Comput. Phys. Comm. 5, 225-233 23 Dunham, W. R., Bearden, A. J., Salmeen, 1. T., Palmer, G., Sands, R. H., Orme-Johnson, W. H. and Beinert, H. (1971) Biochim. Biophys. Acta 253, 134-152 24 Evans, M. C. W., Hall, D. O. and Johnson, C. E. (1970) Biochem. J. 119, 289-291 25 Anderson, R. E., Anger, G., Petersson, L., Ehrenberg, A., Cammack, R., Hall, D. O., Mullinger, R. and Rao, K. K. (1975) Biochim. Biophys. Acta 376, 63-71 26 Smith, B. E., Lowe, D. J. and Bray, R. C. (1973) Biochem. J. 135, 331-341 27 Orme-Johnson, W. H., Hamilton, W. D., Ljones, T., Tso, M.-Y. W., Burris, R. H., Shah, V. K. and Brill, W. J. (1972) Proc. Natl. Acad. Sci. U.S. 69, 3142-3145 28 Mfinck, E. and Champion, P. M. (1974) International Conference on the Application of the M6ssbauer Effect, Bendor, France, Sept. 2-6, J. de Physique, in the press 29 Dickson, D. P. E., Johnson, C. E., Cammack, R., Evans, M. C. W., Hall, D. O. and Rao, K. K. (1974) Biochem. J. 139, 105-108

53 30 Carter, C. W., Kraut, J., Freer, S. T. and Alden, R. A. (1974) J. Biol. Chem. 249, 6339-6346 31 Sieker, L. C., Adman, E. and Jensen, L. H. (1972) Nature 235, 40-42 32 Herskovitz, T., Averill, B. A., Holm, R. H., lbers, J. A., Phillips, W. D. and Weiher, J. F. (1972) Proc. Natl. Acad. Sci. U.S. 69, 2437-2441 33 Frankel, R. B., Averill, B. A. and Holm, R. H. (1974) International Conference on the Application of the M6ssbauer Effect, Bendor, France, Sept. 2-6, J. de Physique, in the press