192

Biochimica et Biophysica Acta, 527 (1979) 192--203

© Elsevier/North-Holland Biomedical Press

BBA 38084 NITROGENASE XI: MOSSBAUER STUDIES ON THE COFACTOR CENTERS OF THE MoFe PROTEIN FROM A Z O T O B A C T E R V I N E L A N D H OP B.H. HUYNH a, E. M~NCK a and W.H. ORME-JOHNSON b a Freshwater Biological Institute, Department of Biochemistry, University of Minnesota, P.O. Box 100, Navarre, MN 55392 and b Department of Biochemistry, University of Wisconsin, Madison, W153706 (U.S.A.)

(Received June 12th, 1978) Key words: Nitrogenase co factor centers; MiSssbauer spectroscopy; Spin-coupled metal clusters; MoFe protein; (A. vinelandii OP)

Summary We have studied the MoFe protein from A z o t o b a c t e r vinelandii OP with MSssbauer spectroscopy in applied magnetic fields up to 50 kG. The results are as follows. (1) The MSssbauer spectra of the S = 3/2 centers, which reside on the cofactor of nitrogenase, have been decomposed into six subcomponents. This suggests that each center contains 5--7, most probably 6, Fe atoms, thus confirming our earlier conclusions which were based on the quantitation of EPR data and on the assumption that the MoFe protein contains (30 + 2) Fe atoms. (2) Analysis of the high-field data shows that three subsites are characterized by a positive magnetic hyperfine coupling constant, Ao, while Ao is negative for the other three sites. This observation demonstrates t h a t the S = 3/2 centers are spin-coupled structures. (3) The zero-field splitting parameter D = +(6 ± 1.5) cm -1 obtained from the M~ssbauer data is in good agreement with our earlier EPR results, D ~ +5.5 cm -1. (4) The resolution of the MSssbauer spectra of the MoFe protein can be dramatically increased by employing Fourier transform deconvolution techniques. This allows a clear demonstration of spectral component S. I. Introduction For the past few years we have investigated the m o l y b d e n u m and iron containing protein (MoFe protein) of the nitrogen fixation system with MSssbauer and EPR spectroscopy. The results obtained for the proteins from A z o t o b a c t e r vinelandii and C l o s t r i d i u m . p a s t e u r i a n u m suggest the following picture: 1. The holoprotein (four subunits, M~ ~ 220 000) contains (30 -+ 2) Fe atoms and two Mo atoms. This conclusion is based on the assumption that the Debye-Wailer factor is the same for each iron site, and on the observation that t h e smallest

193 discernible components in the MSssbauer spectra account for 6.5% of the total iron absorption {for details see ref. 1). 2. The data suggest [1,2] that 12 iron atoms belong to two, apparently identical, novel iron clusters (called M) which we have shown recently [3l to be structural components of the iron and m o l y b d e n u m containing cofactor [4] of nitrogenase. The M-clusters can be stabilized in three oxidation states [1,2], M ° x ~ M N ~ M R . In the resting state, M s , the electronic groundstate has a system spin S = 3/2; this state gives rise to the prominent EPR signal at g = 4.32, 3.65 and 2.01 [2,3]. Oxidation of the M-centers yields the diamagnetic (S = 0) state M °x. Under nitrogen fixation conditions [2] M N is reduced into the state M R which has integer electronic spin. 3. Evaluation of the MSssbauer spectra of thionine oxidized MoFe protein has led to the conclusion that 16 iron atoms are associated with four iron centers which we have called the P-clusters. In our interpretation the P-clusters are made up of those iron atoms which, in the native protein, give rise to those MSssbauer spectra which we have labeled previously 'components D and Fe 2÷'. The MSssbauer data prove conclusively that the P-clusters are spin-coupled structures [ 1 ]. Our previous conclusions are partly based on the assumption that the DebyeWaller factor, f, is the same for all iron sites. This seems to be a reasonable assumption considering the recent results of Dwivedi et al. [5] that the f values are the same to within 5% for myoglobin and rubredoxin, proteins which furnish quite dissimilar ligand environments for the iron. The assumption of constant f values has led to the conclusion [3] that each cofactor center M contains six iron atoms. In this paper we will provide further evidence for this assertion: the magnetically split MSssbauer spectrum observed for the state M N appears to consist of six subcomponents. II. Methods STFe-enriched protein from A. vinelandii was prepared a n d handled as described previously [2]. The sample studied here was from the same batch as those used for our thionine titration studies [1 ]. The MSssbauer spectra were taken in a superconducting magnet system in vertical transmission geometry [6]. Both the source (60 mCi STCo in rhodium) and the absorber were immersed in liquid helium. Temperatures of 1.5 K were achieved by pumping on the He bath; the temperature was measured b y a carbon resistor and a diaphragm vacuum gauge. Data reductions were performed as described previously

[11. III. Results Fig. 1 shows MSssbauer spectra of the MoFe protein from A. vinelandii taken at 4.2 K in magnetic fields of 600 G applied parallel and perpendicular to the observed 7-radiation, We have marked in the figure those q u a d r u p o l e doublets which we have previously called components D, Fe 2÷ and S. Components D and Fe 2÷, originate from the P-clusters; in the oxidation state considered here these clusters .,are diamagnetic. The remainder of the spectrum is a magnetic pattern originating from two identical S = 3/2 centers, the M-clusters.

194 I

0.0

I~11~t11~11~1111~1~1~11~11

A

I ii

Fe 2+

I

,,,,,,,,,,,,,

S 1.0

I,~ o i~ ~1 ,-,

,~. (;3

0.0

~J~1111~1111~1~111~11~11~11111~11I

I t

1.0

I-,-, o I~1

2.0

i

3.0

I

I

I

-3

-2

-1

JD 0 VELOCITY

I IH

E~

3

4

H~z'S

Fig. 1. M6ssbauer spectra o f native M o F e p r o t e i n f r o m A. v i n e l a n d i i t a k e n at 4 . 2 K i n a magnetic field of 6 0 0 G applied ( A ) parallel and (B) transverse to the observed T-radiation.

We have previously discussed in detail [2] h o w the absorption lines of components D (41% of total Fe), Fe 2÷ (13%), and minority c o m p o n e n t S (6%) can be subtracted from the data to yield a fairly good representation of the shape of c o m p o n e n t M. We have repeatedly [1,2] made reference to c o m p o n e n t S, a quadrupole doublet barely discernible in tile raw data. Its presence, however, can be clearly demonstrated by using the following procedure. A MSssbauer spectrum can be viewed as a convolution of a source lineshape function (a Lorentzian) with a transmission function of the absorber. By removing the linewidth contribution of the source, by means of a deconvolution technique [7,8], the resolution can be significantly increased. In order to resolve component S optimally one can additionally use the following trick. At 10 K, the magnetically split spectrum of the S = 3/2 centers is broad and featureless, due to relaxation broadening (Fig. 2A). Under these conditions only components D and Fe 2÷ are resolved; c o m p o n e n t S is indicated by shoulders at velocities --0.07 m m / s and +1.30 mm/s. A deconvoluted spectrum is shown in Fig. 2B (we have used a Gaussian filter [7] with 0.07 m m / s full width). A further enhancement in resolution is achieved by manipulating the data in a way which is referred to as 'deconvolut~on' in Fourier transform nuclear magnetic resonance spectroscopy (for details, see ref. 9}. This amouts effectively to subtracting a fraction K (K is the ratio of the linewidths Of the spectra in Figs. 2A and

195 i

0.0

f

~

,

........... ~

I

'~"~"~ m. ~,'~~'~'~,.,~. #.,~,.,~,,~

~

A

S ~

~ II i1~11~1~ ~ ~ ~ ~ i~ ~

1.0

~ ~ "

1

"~'~' ~

~ ~

~ ~

/%~

-

~'

~1

2.0

.~

i i

0.0

~'~"~"%~"V~""~'~ ',~'^'~"V'%~,,~

~

.

' '~L

i

~N

',

,

,

2.0

,W~,w~ ." ~,' ', ~ , ,

i i

f~,% ,

,~,~,,,~ ~,,¢~h,~~ ~

,

, , ,,,'

,' ,~

,

~,'

4.0

~'m%'~,,,~##",~,~'~'@~,~;, I C~

I

#~',~ ,~,~,,,~,,,~,/~,,,~';,,"~

IIIi

I

~

iI

~

~i

lllll

II

i~ i II iI

I

II

II iI I I

I

-4

-3

I

I

-2

-1

, I

I

I

i

I

0

1

~

3

4

YELOCZTYZ~ ~ Fig. 2. M 6 s s b a u e r s p e c t r u m o f t h e M o F e p r o t e i n t a k e n a t I 0 K i n z e r o m a g n e t i c field (A). U n d e r t h e e x p e r i m e n t a l c o n d i t i o n s t h e s p e c t ~ a m o f c o m p o n e n t M is b r o a d a n d featureleg~; its a b s o r p t i o n r a n g e s f r o m - - 2 m m / s t o +3 m m / s . C o m p o n e n t S is i n d i c a t e d b y t h e b r a c k e t . T h e s p e c t r u m i n B is o b t a i n e d by removing the linewidth contribution of the $TCo(Rh)sottrce according to procedures de~Ibed by D i b a r - U r e a n d F l i n n [ 6 ] . T h e s p e c t z t t m in C is o b t a i n e d b y s u b t r a c t i n g 809b o f t h e specttnam o f A f r o m t h a t o f B. T h e M ~ s ~ b a u e r p a r a m e t e r s o f c o m p o n e n t S ~re f o u n d t o b e ~ E Q = ( 1 . 3 7 + 0 . 0 4 ) m m / s a n d ~ = + ( 0 . 6 4 + 0 . 0 4 ) m m / s ( q u o t e d w i t h r e s p e c t t o F e m e t a l a t 2 9 8 K).

B) o f the spectrum of Fig. 2A from that of Fig. 2B. The result is shown in Fig. 2C. The resolution is further increased, albeit with a loss of signal to noise ratio and a distortion of the line shape. The presence o f component S is clearly indicated in the spectra of Figs. 2B and C. The transformed spectra of Fig. 2B

196

o.o

A

//

,,.--,, ,~,,,, /."'"",,, :,

0.5

"'

0.0

w

B

~~o

~. IIg m

0.5

-3

-2

-1

0 VELOCITY IM

1

2

3

4

Mtl/S

F i g . 3. L o w - t e m p e r a t u r e s p e c t r a o f t h e c o f a c t o r c e n t e r s M in p a r a l l e l ( A ) a n d t r a n s v e r s e ( B ) m a g n e t i c field. T h e s p e c t r a w e r e o b t a i n e d f r o m t h o s e s h o w n in F i g . 1 b y s u b t r a c t i n g the c o n t r i b u t i o n s o f c o m p o n e n t s D , F e 2+ a n d S. T h e solid lines in F i g s . 3 a n d 5 are t h e o r e t i c a l s p e c t r a , g e n e r a t e d w i t h t h e p a r a m e t e r set l i s t e d in T a b l e I. I n F i g . 3 A t h e t h e o r e t i c a l s p e c t r a o f t h e s u b c o m p o n e n t s A 1 ( ), A 2 ( . . . . . . ), A 3 ( . . . . . ), a n d B ( . . . . . . ) are indicated.

also reveal that the absorption lines of components D (12 Fe atoms) and Fe z÷ (four Fe atoms) are extremely narrow. This observation implies for component D that its 12 iron atoms exhibit to ~vhitin 5% the same quadrupole splittings and isomeric shifts. The remarkable sharpness of components D, Fe 2÷ and S allows us to describe the absorption lines of the three quadrupole doublets by simple Lorentzians and thus prepare the low temperature spectrum of component M quite reliable * The spectra of component M are displayed in Fig. 3. Before we discuss the decomposition of these spectra into subcomponents we briefly review some results from our earlier work [2]. The spectra in Fig. 3 result from an electronic system with spin S = 3/2. We have described the EPR results with the

* R a t h e r t h a n u s i n g t h e s p e c t r a o l i s o l a t e d c o f a c t o r m a t e r i a l w e have c h o s e n to d e c o m p o s e t h e data t a k e n o n t h e h o l o e n z y m e b e c a u s e t h e l a t t e r e x h i b i t s m u c h sharper s p e c t r a o f c o m p o n e n t M (see d i s c u s s i o n and F i g . 2 o f ref. 3 ) .

197 spin Hamiltonian = DIS

+

(Sl

+go

S. H

(1)

For the A. vinelandii protein, D ~ +5.5 cm -~, k = 0.055, and go = 2.0. The o~e~ EPR s i ~ a l at g~ = 3.65, g~ = 4.32 ~ d g~ = 2.01 ~ s u l ~ from the M = ~1/2 ~ o u n d Kramem doublet. To e v £ u a ~ the M ~ b a u e r d a ~ the Hamilt o n i ~ of Eqn. 1 is a u ~ e n ~ by

~ , = AoS" I + efl~,~ [ 3 ~ -- I(I + 1) + Q(~ -- Q ) ] - g . ~ . H " I 12

(2)

We will a ~ u m e that each iron s i ~ of the M - c l u s ~ is c h ~ a c ~ z e d by a l ~ h y ~ r f i n e coupling c o n s ~ t A,, (see below). The elec~onic spin in Eqn. 2 is the clus~r spin S = 3/2; i~ e x e c r a t i o n v~ues c ~ be c o m p u ~ from Eqn. 1. The quadmpole splitting ~ and the isome~c shift, 5, for each iron si~ ~ k n o ~ faffly well (see below). As long as we ~ e c o n c e m ~ with ~ e low-field M S , b a u e r s p ~ ~ k e n at 4.2 K we c ~ i ~ o r e the p r e ~ n c e of the M = ~3/2 doublet; the s ~ a in Fig. 3 result from the EPR-active M = ~1/2 doublet. I n s p ~ t i o n of the s p e c ~ in Fig. 3 readily m y e r s at l e ~ t two m a ~ e t i c s u ~ o m p o n e n ~ . Component A h ~ the ~ghtmost a ~ o ~ t i o n lines around +3 mm/s while component B D v ~ ~ ~ strong absorption ~ o u n d +1.8 mm/s. In the following we give ~ m e n ~ that the a b s o ~ t i o n ~ o u n d +3 mm/s resul~ from two s u ~ o m p o n e n ~ . Fig. 3A shows that ~ e r e is ~ abso~tioQ line at +2.7 mm/s and a shoulder at +3.1 mm/s. It follows from the p r o ~ R i e s of ~ e elec~onic ~ o u n d doublet that the in~nsities of the a b ~ t i o n lines d e ~ n d q u i ~ strongly on the direction of the a p p l i ~ field relative to the ~-mdiation (following roughly a 6 : 0 : 2 : 2 : 0 : 6 p a t ~ m in p ~ l l e l field, and a 3 : 4 : 1 : 1 : 4 : 3 p a t ~ m in p e ~ e n d i c u l ~ field). The a b s o ~ t i o n in the v e l ~ i t y r ~ g e from +2.5 to +3.5 mm/s o b ~ in t r a n s v e ~ field is about half of that o~e~ in p ~ l e l field. This s u ~ e s ~ that the peak at +2.7 mm/s and the shoulder at +3.1 mm/s ~ e the rightmost a b s o ~ t i o n lines of two s u ~ o m p o nent s p ~ . It could ~ ~ e d that some of the absorption ~ o u n d +3 mm/s could ~ a t t ~ b u ~ to a minor p ~ a m a ~ e t i c impurity. This is q u i ~ unlikely b ~ a u s e the field d e ~ n d e n c e implies a K r ~ e m doublet with either isotropic or m o d e r a ~ l y anisotropic g v£ues (see C h a p ~ r IV of inf. 10). Thus the s u ~ e s ~ d impurity would ~ve ~se to an EPR s i ~ . As c ~ be ~ e n from Fig. 1A of ref. 3, MoFe protein ~ p l e s exhibit no EPR s i ~ £ other than that attmbu~ble to the S = 3/2 c e n t r e . Also the p r o ~ i n from C. pasteur~num yields s p u t a pmctic~ly identic~ to those displayed in Fig. 3 ( ~ e Fig. 2C of ref. 3). Thus we conclude that the absorption features around +3 mm/s reflect two different ffon e n v i r o n m e n t , which we l a ~ l c o m p o n e n ~ A1 ~ d A2. ~ e c o m p u ~ r simulations d e w , b e d ~ l o w requ~e that a ~ u t 30--35% of the total a b s o ~ t i o n ~ l o n g s to A1 and A2. This s u ~ e s ~ that the M ~ e n ~ m c o n ~ i n sm kon atoms, in excellent a ~ e e m e n t with our e~lier, independent, q u ~ t i m t i o n s [3]. In o ~ e r to o b ~ i n i n f o ~ a t i o n about the nature of the spin-coupling for the M ~ e n ~ m we have studied the MoFe p r o ~ i n in strong e x ~ m ~ m a ~ e t i c fields. An a p p l i ~ field c ~ influence the s ~ c ~ of component M in ~ m e d~tinct ways. Fimt, a strong field p r ~ u c e s a large Z e e m ~ splithng of the e l ~ t r o n i c

198 ground doublet; for the studies reported here only the M = --1/2 state needs to be considered (see caption of Fig. 5). Second, the nuclear Zeeman term has to be included; the nucleus will sense an effective field Hef~ = Hint + H. For the M = --1/2 state the components of the internal field are given by

I-l~i~)t= giAol(4gnfin)

i = x, y, z

(3)

where the gi are the experimentally observed g values. Third, the applied field will mix the M = --3/2 state into the M = --1/2 ground state. To give the reader a feeling for the magnitude of the mixing term we have computed, for the simplified case of axial symmetry (~ = 0), the effective field along the x direction _

gxAo

~3 go~ Ao

)

4gn~n + \'~ g ~ n D + 1 H x

Heff

(4)

The first term in parantheses results from mixing the M = --3/2 state into the M = --1/2 level. For the iron subsites of the M cluster the numerical magnitude of this term is approximately equal to 1. Thus for Ao < 0 the contribution of the mixing term almost cancels the effect of the applied field. On the other hand, for Ao > 0 the mixing term adds to the nuclear Zeeman term and the magnetic splitting increases as the applied field is increased. Note that the presence of a mixing term provides a means to determine experimentally the magnitude of the zero-field splitting parameter, D. In Fig. 4 we have displayed a MSssbauer spectrum taken at 4.2 K in a parallel field of 50 kG. The solid line is a spectral simulation of component D, assuming diamagnetism (S = 0). Note that the assumption of diamagnetism produces exactly the experimentally observed magnetic splitting of component D. Com1

0.0

I

I --

I

~-----I

~

I

~

r

I IIIIIIII1!111 h/lllllllllllll IIIIIIIIfllll~ll IIIII I

IIIIIIIIIIIIllllllllllllllllllllllllllllllllllllllltlllllllt iI

III o

~0 ,-#

IhI Illlllllll

1.O

IIIi

~, .~

,,!!It111 I

~/~ i/

iiii I

IiiII IIt ~.0

II I~

"0 m J~

m 7

r

II t Iii

iI

II

~.0

II I11111111

"-I

f',"7,' I

I

I

I I

Ill

4.0 I

-S

_

I

I

-4

-3

~

-2

~

-1

~

I

I

0

1

2

VELOCITY

--

I

I

i - -

3

4

S

IN M M x S

Fig. 4. M 6 s s b a u e r s p e c t r u m o f native M o F e p r o t e i n t a k e n at 4 . 2 K in a parallel field o f 50 kG. T h e solid line is a c o m p u t e r s i m u l a t i o n o f c o m p o n e n t D a s s u m i n g d i a m a g n e t i s m (S = 0 ) , A E Q > 0, and ~1 = 0.

199

ponent Fe 2" reflects a diamagnetic environment also. This c o m p o n e n t is best resolved, when the protein is prepared under nitrogen fixing conditions (i.e., in the presence of Fe protein and Mg. ATP, see ref. 2). High-field studies performed at 4.2 K in fields up to 55 kG implicate a diamagnetic environment; moreover these studies have shown unambiguously that the cofactor centers M are paramagnetic (S > 1) in the reduced state, M R (B.H. Huynh and E. Miinck, unpublished). Fig. 5 shows the high-field spectra of c o m p o n e n t M, obtained by removing the contributions of components D, Fe 2" and S. For these subtractions we have

~° ~ ~4~

~! iI~

° ..I

~~" ~

~ J~',,~'i

,,~,/", r"~"

:.

~

~

M

..

o

o o~

,

oo

~~. ~ , ~ ,:.,~ ~

o.~

~~ /

l 0

!

j.

5

i

|

;

-5

-4

-3

~ -~

,

-1

0

L

E

|

~

;

3

4

5

VELOCITY I N M ~ ' $ FIE. 5. H ~ h - f / e l d spectra o f c o m p o n e n t M, o b t ~ n e d by ~ m o ~ ~m ~e ~w ~ ~e con~bu~o~ of c o m p o n ~ n ~ D, Fe 2., a ~ S ~ d ~ d ~ the t e x t . The raw ~ ~ n ~ p~el f i ~ d o f 14 EG w e ~ ~co~ at I . S K w ~ e ~ e 3 5 kG ~ d ~ kG d a ~ w e ~ ~ k e n a t 4.2 K. F o r f l e ~ e r ~ 8 0 k G ~e 8ve~e ~man ~ t ~ of ~ e ~ u ~ K~e~ doublet of ~ e S - 3/2 s y ~ m ~ ~ e e n o u ~ ~ e h ~ a t at 4.2 K o ~ y ~ e M = - - I / 2 r o t e ~ ~ n ~ y popu~d.

200 assumed that c o m p o n e n t S is diamagnetic and that all three components ~ 0 and ~? = 0. The choice of ~ and sign (AE¢~) is not critical, since the shapes of high-field spectra of diamagnetic compounds, in absolute terms, depend very little on these parameters. We do not know whether c o m p o n e n t S reflects paramagnetic or diamagnetic iron sites. However, since it accounts for only 6% of the total absorption, only a small error is made if the assumption of diamagnetism is incorrect. Inspection of the spectra of Fig. 5 reveals that the magnetic splittings of components AI and A2 are practically independent of the applied field (i.e., Ao ~ 0) while components B exhibit an increased splitting as the strength of the applied field is increased. have ~ Q

IV. Analysis Eqn. 3 shows that the largest components of Hi,t are in the electronic x-y plane. Consequently, the MSssbauer spectra of a polycrystalline specimen are essentially sensitive to the components of the hyperfine interactions which are in the x-y plane. (The physics of the system is quite similar to that of high-spin ferric hemes, see Miinck and Champion, ref. 11). In particular, the magnetic anisotropy projects out those components of the electric field gradient (EFG) tensor which are in the x-y plane (this is the main reason w h y the spectra of the M-centers have sharper features than those observed for 4Fe-4S clusters}. In order to keep the number of unknowns manageably small we have made the following assumptions for our data analysis: 1. The A-tensor of each iron site has axial symmetry around the electronic z-axis (since the properties of the electronic system render the MSssbauer spectra quite insensitive to As we can treat the A-tensor as a scalar Ao). 2. We assume that the local EFG tensors are collinear with the zero-field splitting tensor. Inspection of the spectra readily reveals that subcomponents A 1 and A2 have positive field gradients along the largest c o m p o n e n t of Hi,~t while those of components B are negative. These observations can be accomodated by allowing 90 ° rotations of the local field gradients relative to the frame defined by the zero-field splitting (which refers to the cluster as a whole). With these assumptions the MSssbauer spectrum of each s u b c o m p o n e n t has essentially three unknowns: Ao, which is obtained from the total magnetic splitting, sign (AEQ), a n d s . One further assumption was made for IAEQ I and for the isomeric shifts of the subcomponents. At 40 K all subcomponents have collapsed into one symmetric quadrupole doublet with linewidth P = 0.33 mm/s and an average isomeric shift 5AV = +0.41 mm/s [2]. The analysis of the low temperature data suggests that the isomeric shifts of the B-components are about 0.1 mm/s smaller than those of the A-components. This observation is compatible with the high-temperature data if one considers the d o u b l e t observed at 40 K as a superposition of two equally intense doublets, each having t A E Q I = 0.76 mm/s with linewidth 0.25 mm/s, but with isomeric shifts 5 A = +0.47 mm/s and 5 B = +0.35 mm/s. The solid lines in Figs. 3 and 5 are the results of computer simulations based on Eqns. 1 and 2. The parameter set used is listed in Table I. Considering the complexity of the problem the experimental data are well represented by the theory. The reader should keep in mind that the 'experimental' spectra are n o t

201 TABLE I •

Sni~ H a r n i l t o n i a n p a r a m e t e r s u s e d {o generate the theoretical c u r v e s in Figs. 3 a n d 5 f r o m Eqns. 1 a n d 2. Fo~ all s i m u l a t i o n s , D = +6 c m ~ , ~ = 0 . 0 5 5 , a n d F = 0 . 3 0 m m / s (fu~ w i d t h a t h a l f m a x i m u m ) w a s used.

A0

eQ Vzz

Spectral component

Number o f Fe atoms

gn~n (kG)

2 (ram/s) *

A1 A2 A3 B

1 1 1 3

--144 --124 --96 +76

~0.73 --0.76 +0.38 --0.38

~1

~(mm/s) **

•

0.5 0 3 --3

+0.47 +0.47 +0.47 -'+ 0 . 3 5

* T h e E F G t e n s o r s are referzed to the f r a m e (x, y, z) d e f i n i n g t h e e l e c t r o n i c zero-field splitting: Vzz is n o t necessaxit~" the largest c o m p o n e n t ~ ~ ffi ( V x x - - Vyy)/Vzz. ** i s o m e r i c shift at 4.2 K w:~th ~espect to Fe m e t a l at r o o m t e m p e r a t u r e .

very precisely defined in the velocity range from --0.5 mm/s to +1 mm/s (component D, which contains 12 iron atoms, has been removed). We do not claim that the parameter set listed in Table I is unique. Some important results, however, have emerged from our analysis. 1. A good decomposition of the data has been obtained by assuming that each of the M-clusters contains six iron atoms. 2. The zero-field splitting parameter D = +(6 + 1.5) cm -~ determined from the high-field data is in good agreement with our earlier results obtained from EPR spectroscopy, D ~ +5.5 cm -~. 3. The high-field data have yielded the sign of Ao for each subcomponent. We have partitioned the spectrum into three A-components (Ao < 0) and three B-components (Ao > 0). The subcomponents Al and A2 give rise to the absorption pattern around +3 mm/s (see Fig. 3A). In weak applied fields component A3 is practically indistinguishable from components B; the high-field data, however, reveal that it is characterized by a negative hyperfine coupling constant. The values quoted for the principal axes values of the EFG tensors are, at best, tentative. The high-temperature data show that IAEQI is the same for all subsites. At low-temperatures the frame of reference is essentially the x-y plane defined by the zero-field splitting term of Eqn. 1. We suppose that the different signs of V~ and Vyy reflect an 0rientational inequivalence, i.e., the local EFG tensors are oriented differently relative to the zero-field splitting frame which refers to the cluster as a whole. V. Discussion We have shown above that the magnetically split MSssbauer spectrum which originates from the S = 3/2 centers can be decomposed into six subcomponents. Unless we have grossly misinterpreted the physics of the system' the M-centers cannot contain four iron atoms because components A1 and A2 would have to accout for 50% of the total absorption, in clear disagreement with the data (30--35%). A cluster with eight iron atoms is improbable from the following line of evidence. From a detailed analysis of the MSssbauer ~pectra of thionine oxidized MoFe protein and taking into account oxidative ~itrations, we have recently proposed [1] that the D and Fe 2÷ irons are organized in four iron clusters (the P-clusters), most probably of the 4Fe:4S type.

202 The appropriate number of 4Fe-4S centers is also found by chemical analysis (Rawlings, J., Averill, B.A. and Orme-Johnson, W.H., in preparation). The P-clusters account for (54 + 2.5)% of the total MSsbauer absorption while c o m p o n e n t M quantitates to (40 + 2)%. If the P-clusters are indeed a variant of familiar 4Fe-4S centers, then it follows that 12 iron atoms belong to component M. Quantitative EPR analysis has shown that there are two S = 3/2 centers per molecule, and thus six Fe atoms per M-center. This line of evidence is based on the assumption that the Debye-Waller factors, f, are the same for all iron sites. On the other hand, the assumption of eight iron atoms per M-center would yield the conclusion that the f-values of the M-center sites are 25% smaller than those associated with the P-cluster sites. In the light of the findings by Dwivedi et al. [5] we consider this improbable. The assumption that the M-centers contain five (or seven) Fe atoms would require that 40% (or 28%) of the total absorption of c o m p o n e n t M is associated with subcomponents A1 and A2. This is n o t far from the percentage (30--35%) required to obtain the best fits to the data set and thus n o t conclusively excluded by our analysis. However, it is apparent from two independent lines of evidence that six Fe atoms per M-center is in fact the most probable number. As judged from the quadrupole splittings, AEQ, the iron environments of the M-centers would appear to be quite equivalent. However, when the magnetic hyperfine parameters are taken into account it is apparent that the sites are remarkably different; the magnetic hyperfine interaction experienced by a nucleus of the Al-site is almost twice that for a B-site nucleus. Moreover, the coupling constants Ao for the A-sites are negative (as for isolated high-spin ferric or ferrous ions) while the B sites are characterized by a positive Ao. This observation demonstrates unambiguously that the M-centers are antiferromagnetically coupled structures. It is premature, however, to speculate a b o u t the details of a spin~oupling s c h e m e , since very little is known a b o u t the structural details. Recent results (Orme-Johnson, W.H., Mims, W.B., OrmeJohnson, N.R., Peisach, J., Henzl, M. and Rawlings, J., unpublished) on the nuclear modulation (by 9SMo) of the spin echo signals suggest strongly that m o l y b d e n u m is a structural c o m p o n e n t of the M-centers. This observation is s u p p o r t e d by studies of the extended X-ray absorption fine structure (EXAFS) of nitrogenase [12]. The EXAFS data implicate 1--2 iron atoms in close proximity of a m o l y b d e n u m atom. The observation that clearly less than six Fe atoms contribute to the m o l y b d e n u m EXAFS pattern is not in conflict with our findings because the EXAFS is essentially sensitive to the first and second coordination spheres around the m o l y b d e n u m atom. From the MSssbauer data we obtain a less local view: c o m p o n e n t M contains all iron sites which experience paramagnetic hyperfine interactions due to the delocalized S = 3/2 spin. Whatever the structural details of M-centers may be, present evidence suggests a novel structure containing one Mo atom (the holoenzyme contains one Mo per S = 3/2 center [3]), six Fe atoms, and a b o u t 6--8 labile sulfurs.

Acknowledgements W e thank Dr. R. Z i m m e r m a n n for his m a n y contributions to the analyses of nitrogenase spectra, Drs. W.J. Brill and V.K. Shah for providing STFe enriched

203

Azotobacter M o F e protein, and M. Henzl and Dr. J. Rawlings for aid in preparing these samples for MiSssbauer spectroscopy. This work was supported by the National Science Foundation grant PCM-08522, by the National Institutes of Health through grant GM17170, by the Graduate Research Committee of the University of Wisconsin, and by Research Career Development Award K 0 4 - G M 70683 {E.M.). W.H. O-J. is an I.H. Romnes Faculty Fellow of the University of Wisconsin. References I Z i m m e r m a n n . R.. M 0 n c k , E.. Brt~. W.J.. S h a h . V.K.. H e n : l . M.T.. R a w l m ~ , J. a n d Orme-Johnso~1. W.H. ( ] 9 ? 8 ) B i o c h l m . B i o p h y s . A c ( , 5 3 ? . 1 8 5 - - 2 0 7 2 M 0 n c k . E.. R h o d e s . H.. O r m e - J o h n s o n , W.H.. Divis, I,.C.. Brill. W.J. i n d S h a h . V.K. (197,5) B~o~him. B i o p h y s . Act~ 4 0 0 . 3 2 - - 5 3 3 R a w l i n ~ s . J.. S h a h . V . K . . C h ~ e U . J . R . , BriJ]. W.J.. Z i m m e r m a n n . R.. M ~ n c k . F:. ~nd O r m e - J o h n s o n , W.H. ( ] 9 7 g ) J . Biol. C h e m . 2 5 3 . 1 0 0 1 - - 1 0 0 4 4 S h a h . V.K. I n d Brt~. W.J. ( ] 9 ? ? ) P r o c . N i t L A c i d . Scl. U.S. ? 4 . 3 2 4 9 - - 3 2 ~ 3 5 Dwlvedi. A.. P e t e r s o n . T. a n d D e b r u n n e r . P.G. ( 1 9 7 8 ) J. de Phys/