Catalytic reduction of cis-dimethyldiazene by the [MoFe3S4]3+ clusters. The four-electron reduction of a N = N bond by a nitrogenase-relevant cluster and implications for the function of nitrogenase

Article abstract of DOI:10.1021/ja963475s

The catalytic reduction of cis-dimethyldiazene by the (Et₄N)₂[(Cl₄-cat)(CH₃CN)MoFe₃S₄Cl₃] cluster (Cl₄-cat = tetrachlorocatecholate) is reported. Unlike the reduction of cis-dimethyldiazene by the Fe/Mo/S center of nitrogenase, which yields methylamine, ammonia, and methane (the latter from the reduction of the C-N bond), the reduction of cis-dimethyldiazene by the synthetic cluster yields exclusively methylamine. In separate experiments, it was shown that the C-N bond of methylamine is not reduced by the [MoFe₃S₄]³⁺ core, perhaps accounting for the differences observed between the biological and abiological systems. 1,2-Dimethylhydrazine, a possible partially reduced intermediate in the reduction of cis-dimethyldiazene, was also shown to be reduced to methylamine. Interaction of methylamine with the Mo atom of the cubane was confirmed through the synthesis and structural characterization of (Et₄N)₂[(Cl₄-cat)(CH₃NH₂)MoFe₃S₄Cl₃]. Phosphine inhibition studies strongly suggest that the Mo atom of the [MoFe₃S₄]³⁺ core, which has a Mo coordination environment very similar to that in nitrogenase, is responsible for the binding and activation of cis-dimethyldiazene. The reduction of a N = N bond exclusively at the heterometal site of a nitrogenase-relevant synthetic compound may have implications regarding the function of the nitrogenase Fe/Mo/S center, particularly in the latter stages of dinitrogen reduction.

Full text of DOI:10.1021/ja963475s

1662
J. Am. Chem. Soc. 1997, 119, 1662-1667
3
+
Catalytic Reduction of cis-Dimethyldiazene by the [MoFe S ]
3
4
Clusters. The Four-Electron Reduction of a NdN Bond by a
Nitrogenase-Relevant Cluster and Implications for the Function
of Nitrogenase
1
2
3
Steven M. Malinak, Anton M. Simeonov, Patrick E. Mosier,
2
,1
Charles E. McKenna, and Dimitri Coucouvanis*
Contribution from the Department of Chemistry, The UniVersity of Michigan,
Ann Arbor, Michigan 48109-1055, and Department of Chemistry,
The UniVersity of Southern California, Los Angeles, California 90089-0744
X
ReceiVed October 4, 1996
Abstract: The catalytic reduction of cis-dimethyldiazene by the (Et4N)2[(Cl4-cat)(CH3CN)MoFe3S4Cl3] cluster (Cl4-
cat ) tetrachlorocatecholate) is reported. Unlike the reduction of cis-dimethyldiazene by the Fe/Mo/S center of
nitrogenase, which yields methylamine, ammonia, and methane (the latter from the reduction of the C-N bond), the
reduction of cis-dimethyldiazene by the synthetic cluster yields exclusively methylamine. In separate experiments,
3
+
it was shown that the C-N bond of methylamine is not reduced by the [MoFe3S4] core, perhaps accounting for
the differences observed between the biological and abiological systems. 1,2-Dimethylhydrazine, a possible partially
reduced intermediate in the reduction of cis-dimethyldiazene, was also shown to be reduced to methylamine. Interaction
of methylamine with the Mo atom of the cubane was confirmed through the synthesis and structural characterization
of (Et4N)2[(Cl4-cat)(CH3NH2)MoFe3S4Cl3]. Phosphine inhibition studies strongly suggest that the Mo atom of the
3
+
[
MoFe3S4] core, which has a Mo coordination environment very similar to that in nitrogenase, is responsible for
the binding and activation of cis-dimethyldiazene. The reduction of a NdN bond exclusively at the heterometal site
of a nitrogenase-relevant synthetic compound may have implications regarding the function of the nitrogenase Fe/
Mo/S center, particularly in the latter stages of dinitrogen reduction.
Introduction
Nitrogenase is capable of reducing many small, unsaturated
molecules and ions, the most important of which is atmospheric
N2. The enzymatic synthesis of ammonia by reduction of
atmospheric N2 under ambient conditions supplies nitrogen in
the form needed in the biosynthesis of proteins and nucleic acids.
It is not surprising therefore that much research has been
undertaken in order to achieve a better understanding of this
remarkable reduction process.4 In recent X-ray structure
determinations of nitrogenase the structure of the Fe/Mo/S active
site has been obtained at near-atomic resolution.5 This MoFe7S9
cluster of nitrogenase, referred to as the FeMo-cofactor, is
comprised of MoFe3S3 and Fe4S3 site-voided cubanes linked
by three bridging sulfides (Figure 1). The Mo atom of the
Figure 1. The FeMo-cofactor of nitrogenase.
FeMo-cofactor appears to be coordinatively saturated, with a
terminal imidazole from a histidine residue and a bidentate (R)-
homocitrate ligand completing the coordination sphere.
The synthesis of exact analogs for the FeMo-cofactor has
not been accomplished. Nevertheless, partial analogs exist in
6
synthetic clusters. Outstanding among these clusters are simple
7
Fe/Mo/S cubanes that contain the [MoFe3S4] cores and a Mo
X
atom with first and second coordination spheres nearly identical
to those found for the Mo atom in nitrogenase. In addition to
Abstract published in AdVance ACS Abstracts, February 1, 1997.
(
(
(
1) The University of Michigan, Ann Arbor.
2) The University of Southern California.
3) Present Address: Department of Macromolecular Science, Case
7
the structural and electronic similarities, it has recently been
n+
shown that these [MFe3S4] clusters (M ) Mo, n ) 3 or M )
Western Reserve University, Cleveland OH 44106-7202.
4) (a) Molybdenum Enzymes, Cofactors and Model Systems; Stiefel, E.
(
V, n ) 2, the latter of which are relevent to the alternative
I., Coucouvanis, D., Newton, W. E., Eds.; ACS Symposium Series 535;
American Chemical Society: Washington, DC, 1993; Chapters 10-23. (b)
Eady, R. R.; Leigh, G. J. J. Chem. Soc., Dalton Trans. 1994, 2739. (c)
Kim, J.; Rees, D. C. Biochemistry 1994, 33, 389. (d) Evans, D. J.;
Henderson, R. A.; Smith, B. E. In Bioinorganic Catalysis; Reedjik, J., Ed.;
Marcel Decker, Inc.: New York, 1993; Chapter 5.
8
V-nitrogenase ) are also partial functional models for the
enzyme. Substrate reductions that have been investigated with
9
these clusters include the reduction of hydrazine to ammonia
10
and acetylene to ethylene and ethane. The most significant
results from these studies include the observation that the
heterometal atom (Mo or V) in these clusters is either
(
5) (a) Kim, J.; Rees, D. C. Science 1992, 257, 1677. (b) Kim, J.; Rees,
D. C. Nature 1992, 360, 553. (c) Chan, M. K.; Kim, J.; Rees, D. C. Science
993, 260, 792. (d) Bolin, J. T.; Campobasso, N.; Muchmore, S. W.; Minor,
1
W.; Morgan, T. V.; Mortenson, L. E. In New Horizons in Nitrogen Fixation;
Palacios, R., Mora, J., Newton, W. E., Eds.; Kluwer Academic Publishers:
Dordrecht, 1993; pp 89-94. (e) Bolin, J. T.; Campobasso, N.; Muchmore,
S. W.; Morgan, T. V.; Mortenson, L. E. In Molybdenum Enzymes, Cofactors
and Model Systems; Stiefel, E. I., Coucouvanis, D., Newton, W. E., Eds.;
ACS Symposium Series 535; American Chemical Society: Washington,
DC 1993; pp 186-195.
(6) Coucouvanis, D. Acc. Chem. Res. 1991, 24, 1.
(7) (a) Holm, R. H. AdV. Inorg. Chem. 1992, 38, 1, and references therein.
(b) Holm, R. H.; Simhon, E. D. In Molybdenum Enzymes; Spiro, T. G.,
Ed.; John Wiley & Sons, Inc.: New York, 1985, pp 1-87 and references
therein.
(8) (a) Ciurli, S.; Holm, R. H. Inorg. Chem. 1989, 28, 1685. (b) Kovacs,
J. A.; Holm, R. H. Inorg. Chem. 1987, 26, 702.
S0002-7863(96)03475-0 CCC: $14.00 © 1997 American Chemical Society

Reduction of cis-Dimethyldiazene by the [MoFe3S4]3+
J. Am. Chem. Soc., Vol. 119, No. 7, 1997 1663
9
10
exclusively or principally involved in the activation of the
substrate toward reduction.
were dependent on both the specific isomer used as substrate
and (as demonstrated for the cis isomer) the Fe:FeMo protein
ratio. The strained-ring diazene, diazirine, is also reduced by
The synthetic clusters most efficient in their ability to act as
catalysts have been the (Et4N)2[(L)(L′)MoFe3S4Cl3] single
cubanes, with either a labile solvent molecule weakly coordi-
15a
nitrogenase to the same products.
In order to investigate what
role, if any, the Mo atom may play in reduction of a NdN bond,
cis-dimethyldiazene was investigated as a potential substrate
nated to the Mo atom (L ) bidentate tetrachlorocatecholate, L′
7
3+
)
CH3CN ) or, paradoxically, a tridentate polycarboxylate
for the [MoFe3S4] cubanes. Herein, we report on the catalytic
11
ligand (L, L′ ) polycarboxylato ligand ) that results in a
coordinatively saturated Mo atom. When the structure of the
Fe/Mo/S center in nitrogenase revealed that the Mo atom of
the cluster was coordinatively saturated in the resting state, the
direct involvement of the Mo atom in substrate binding appeared
reduction of cis-dimethyldiazene by the (Et4N)2[(Cl4-cat)-
(CH3CN)MoFe3S4Cl3] cubane. Results show that the only
detectable product from the reduction of cis-dimethyldiazene
by the synthetic cluster is methylamine. Additionally, it has
been demonstrated through inhibition studies that activation and
reduction of cis-dimethyldiazene occurs exclusively at the Mo
site. The implications of these observations on the possible
function of the cofactor of nitrogenase will also be discussed.
3+
unlikely. The ability of the polycarboxylate-ligated [MoFe3S4]
model clusters to serve as catalysts suggested a special role for
the coordinated carboxylate ligands. Indeed, the observation
that carboxylate-bound clusters with coordinatively saturated
Mo atoms are also catalytically active in substrate reduction
implies their ability to generate coordination sites for the
substrate by displacing one of the “arms” of the carboxylate
Experimental Section
General Considerations. All manipulations were performed under
an inert atmosphere using standard glove box and Schlenk techniques.
9,11
9
ligand through protonation.
In some cases, these polycar-
Solvents were distilled under N
diethyl ether and THF from sodium/benzophenone, and CH
) or stored over 3 Å molecular sieves (absolute ethanol) and
thoroughly degassed with N or Ar prior to use. Reagent-grade chemi-
cals were purchased from Aldrich Chemical Company (cobaltocene
CoCp ), 99% 2,6-lutidine (Lut), anhydrous DMF, methylamine,
ethylamine, dimethylhydrazine dihydrochloride, NaBPh , 1.0 M ethereal
2
from the appropriate drying agents
boxylate-ligated clusters are actually better catalysts than the
catecholate precursors, presumably due to the protonated arm
of the ligand that may serve as a “shuttle”, returning a proton
to the reduced substrate. While N2 is not reduced by the
(
3
CN from
2 3
B O
2
n+
synthetic [MFe3S4] clusters, the catalytic reactivity of the latter
(
2
suggests the possibility of partially reduced substrates interacting
and being reduced at the heterometal atom of the Fe/M/S center
in nitrogenase.
With regard to dinitrogen reduction, it has been suggested
that dinitrogen is activated in the Fe3(µ-S)3Fe3 “cage” created
4
HCl, PEt
3
) and used without further purification. Freshly prepared
16
solutions of cis-dimethyldiazene in distilled, degassed CH
3
CN (typi-
cally 0.1-0.2 M as determined by UV spectroscopy, ꢀ367 ) 266) were
stored at -196 °C until immediately prior to use.
Physical Measurements. Infrared spectra (CsI disks) were obtained
by the six three-coordinate Fe atoms in the central part of the
-
1
cofactor.12 The mechanism of dinitrogen reduction is believed
using a Nicolet 740 FT-IR spectrometer (far-IR 500-150 cm ) or a
-
1
13
5DXB FT-IR spectrometer (mid-IR 4000-400 cm ). Quantification
of methylamine and ammonia was performed using an HPLC technique
previously described.17 An HP 5890 Series II gas chromatograph
equipped with either a porapak N column (Supelco) or a 4% carbowax
to proceed through diazene-like intermediates, although di-
azene has not yet been demonstrated to interact with the
nitrogenase cofactor.14 At present, it is not known where the
reduction of dinitrogen to ammonia occurs. It could take place
entirely at the six-Fe “cage” or it may undergo the initial two-
or four-electron reduction at the six-Fe “cage” and the final two-
to four-electron reduction and cleavage of the N-N bond at
the Mo atom.
2
2
0 m column (Supelco) was used in order to detect methane or EtNH ,
respectively. EPR studies and elemental analysis were performed by
the Biophysics Research Division and the Analytical Services Division,
respectively, at the University of Michigan. Integration of the EPR
18
signal was accomplished according to previously discussed techniques.
Analysis samples were routinely kept under dynamic vacuum for 12 h
before submission.
Recently, it has been reported that both cis- and trans-
dimethyldiazene are substrates for nitrogenase, and as such
Preparation of Compounds. Analytically pure 2,6-lutidinium
hydrochloride (Lut‚HCl) was prepared from the reaction between
represent the first example of reduction of an unstrained NdN
bond by the Fe/Mo/S center of nitrogenase.15 Products detected
lutidine and ethereal HCl. Lut‚HBPh
4
was prepared from the metathesis
included methylamine, methane, and ammonia in ratios that
reaction between Lut‚HCl and NaBPh
4
in ethanol. (Et
4
N)
2 4
[(Cl -
19
20
(
9) (a) Demadis, K. D.; Malinak, S. M.; Coucouvanis, D. Inorg. Chem.
996, 35, 4038. (b) Coucouvanis, D.; Demadis, K. D.; Malinak, S. M.;
Mosier, P. E.; Tyson, M. A.; Laughlin, L. J. J. Mol. Cat. A: Chemical
996, 107, 123. (c) Malinak, S. M.; Demadis, K. D.; Coucouvanis, D. J.
Am. Chem. Soc. 1995, 117, 3126. (d) Demadis, K. D.; Coucouvanis, D.
Inorg. Chem. 1995, 34, 3658. (e) Coucouvanis, D.; Mosier, P. E.; Demadis,
K. D.; Patton, S.; Malinak, S. M.; Kim, C. G.; Tyson, M. A. J. Am. Chem.
Soc. 1993, 115, 12193.
10) Laughlin, L. J.; Coucouvanis, D. J. Am. Chem. Soc. 1995, 117, 3118.
11) (a) Demadis, K. D.; Coucouvanis, D. Inorg. Chem. 1995, 34, 436
and references therein. (b) Demadis, K. D.; Coucouvanis, D. Inorg. Chem.
995, 34, 3658 and references therein.
12) (a) Stavrev, K. K.; Zerner, M. C. Chem. Eur. J. 1996, 2, 83. (b)
3
cat)(CH CN)MoFe
3
S
4
Cl
3
]
and (Ph
4
P)
2
[Fe
4
S
4
Cl
4
]
were obtained by
1
procedures similar to those previously reported.
Et N) [(Cl -cat)(RNH )MoFe Cl ] (R ) Me or Et). An amount
of RNH (0.10 mL of a 2 M EtOH solution) was added to an CH CN
solution (30 mL) of (Et N) [(Cl -cat)(CH CN)MoFe Cl ] (0.21 g, 0.20
(
4
2
4
2
S
3 4
3
1
2
3
4
2
4
3
S
3 4
3
mmol) in one portion. After approximately 1 h of stirring, the solution
was filtered and ether (150 mL) was layered on the filtrate. After
overnight standing, a near-quantitative yield of brown crystals was
isolated by filtration and washed well with ether.
(
(
1
45 3 2 7 4 3
(A) R ) Me (I). Analysis calculated for C23H N O Cl S Fe Mo:
C, 26.68; H, 4.38; N, 4.06. Found: C, 27.01; H, 4.22; N, 4.10. Mid-
IR (CsI disks): 3290(w), 3250(w) from amine. Far-IR (CsI disks):
(
Dance, I. G. Aust. J. Chem. 1994, 47, 979. (c) Deng, H.; Hoffman, R. Angew.
Chem., Int. Ed. Engl. 1993, 32, 1062. (d) Chan, M. K.; Kim, J.; Rees, D.
C. Science 1993, 260, 792.
4
08(m), 351(vs). Single crystals of this complex were the subject of
an X-ray structure determination.
(
13) Thorneley, R. N. F.; Lowe, D. J. Biochem. J. 1984, 224, 887.
(
14) Burris, R. H.; Winter, H. C.; Munson, T. O.; Garcia-Rivera, J. in
Intermediates and Cofactors in Nitrogen Fixation; San Pietro, A., Ed.;
Antioch Press: Yellow Springs, OH, 1965; p 315.
(16) Simeonov, A. M.; McKenna, C. E. J. Org. Chem. 1995, 60, 1897.
(17) Bravo, M.; Eran, H.; Zhang, F. X.; McKenna, C. E. Anal. Biochem.
1988, 175, 482.
(15) (a) McKenna, C. E.; Simeonov, A. M.; Eran, H.; Bravo-Leerab-
handh, M. Biochemistry 1996, 35, 4502. (b) McKenna, C. E.; Simeonov,
A. M. In Nitrogen Fixation: Fundamentals and Applications; Tikhonovich,
I. A., Provorov, N. A., Romanov, V. I., Newton, W. E., Eds.; Kluwer
Academic Publishers: Dordrecht, 1995; p 158. (c) Simeonov, A. M.;
McKenna, C. E. submitted for publication.
(18) Hearshen, D. O.; Hagen, W. R.; Sands, R. H.; Grande, H. J.;
Dunham, W. R. J. Magn. Reson. 1986, 69, 440 and references therein.
(19) Palermo, R. E.; Holm, R. H. J. Am. Chem. Soc. 1983, 105, 4310.
(20) Wong, G. B.; Bobrik, M. A.; Holm, R. H. Inorg. Chem. 1978, 17,
578.

1664 J. Am. Chem. Soc., Vol. 119, No. 7, 1997
Malinak et al.
Table 1. Summary of Crystal Data for
µmol), Lut‚HCl (10.0 mL, 109 µmol) and cluster (0.15 mL, 0.72 µmol)
analogous to those described in section B above were placed in a flask
(
Et N) [(Cl -cat)(CH NH )MoFe Cl ] (I)
4
2
4
3
2
3
S
4
3
3
containing predetermined amounts of CH CN. With the concentrations
3 4 7 23 45 3 2
chemical formula MoFe S Cl C H N O
of these reagents held constant, the appropriate amount of a cis-
dimethyldiazene solution (3-167 equiv based on cluster concentration)
was added to the flask, marking t ) 0 h. At t ) 1 h, 3.0 mL aliquots
of the reaction solution were obtained and worked up as described
above.
molecular weight, g/mol
color, morphology
space group
a, Å
b, Å
1034.52
brown, needles
Pbcm
11.635(6)
32.14(1)
33.06(1)
12506
c, Å
V, Å
(
D) Attempted Reductions of 1,2-Dimethylhydrazine, Methyl-
amine, and Ethylamine. To a 125 mL reaction flask filled with an
appropriate amount of CH
CN (total solution volume was 40 mL) was
added 4.2 mL of a 7.2 mM CoCp solution (30 µmol), 0.03 g of
Lut‚HBPh (70 µmol) and 0.2 mL of a 4.8 mM (Et N) [(Cl -cat)-
CH CN)MoFe Cl ] solution. An excess of the appropriate substrate
MeNH , EtNH , or 1,2-dimethylhydrazine as CH
5 equivalents based on cubane concentration) was then added, marking
3
Z
12
1.63
1.65
Mo K
2.01
3° e ω e 45°
4016
2762
399
3
density (observed), g/mL
density (calculated), g/mL
radiation
absorption coefficient
data collected
unique data
data used (F
number of parameters
phasing techniques
R, R
2
4
4
2
4
R
3
3
S
2
4
3
21
(
1
2
3
CN solutions, 10-
2
2))
t ) 0 h. At t ) 1 h, aliquots were obtained and quenched with HCl as
described and analyzed for products. In the case of the amines, the
headspace gas in the reaction solutions was analyzed for CH , and the
4
solution analyzed for ammonia by the indophenol method as previously
described.9 In the case of 1,2-dimethylhydrazine, the vials were
analyzed for methylamine by HPLC.
The reactions described in sections A, B, and D were performed in
duplicate, while those in section C were performed in triplicate.
Additionally, each reaction aliquot was analyzed for products at least
twice. While repeat measurements on each vial typically did not vary
by more than 10%, absolute product yield varied slightly between
identical reactions. Regardless, the trends observed for the time course
of the reaction, the phosphine inhibition and the ν vs [cis-dimethyl-
diazene] curves were consistent between sets of experiments and are
o
> 3σ(F
o
direct methods
0.0717, 0.0840
w
(
B) R ) Et (II). Analysis calculated for C24
3
47 3 2 7 4 3
H N O Cl S Fe Mo‚
1/
4
2
CH CN: C, 28.06; H, 4.58; N, 4.58. Found: C, 28.45; H, 4.53; N,
.86. Mid-IR (CsI disks): 3297(w), 3239(w) from amine; 2249(w)
from lattice CH CN. Far-IR (CsI disks): 421(m), 410(m), 392(w),
50(vs), 343(sh), 298(w).
X-ray Structural Determination. Crystals suitable for diffraction
study were obtained by vapor diffusion of ether into an CH CN solution
3
3
3
of I. Brown needlelike crystals were mounted in glass capillaries and
flame sealed under argon. Diffraction experiments were performed at
ambient temperature with a Nicolet R3m four-circle automated dif-
fractometer using graphite-monochromatized Mo KR radiation. The
orientation and unit cell parameters were determined from 18 machine-
centered reflections with 10° < 2θ < 25°. A total of 5234 reflections
were collected in the range 3° < ω < 45° with 4016 unique reflections.
The structure was solved using direct methods. Anisotropic temperature
factors were used for all non-carbon atoms of the anion. The final
cycle full-matrix least squares refinement was based on 2762 reflections
reported as such.
+
]3 Catalyst. A
(E) Recovery and Identification of the [MoFe
S
3 4
1
2
25 mL flask was charged with 23 mL of a 7.2 mM CoCp
3 mL of a 10.9 mM Lut‚HCl solution, and 3.0 mL of a 4.8 mM
N) [(Cl -cat)(CH CN)MoFe Cl ]. Addition of 0.42
2
solution,
solution of (Et
4
2
4
3
S
3 4
3
mL of a 0.1 M cis-dimethyldiazene solution marked t ) 0 min. After
stirring for 90 min, the solution was taken to dryness. The resulting
residue was washed well with THF to remove any unreacted CoCp ,
2
and subsequently dissolved in 10.0 mL DMF. An aliquot of this DMF
solution (1.3 mM cubane) was then obtained and subjected to a
quantitative EPR analysis.
(
I > 3σ(I)) and 399 variables using the SHEXTL Plus crystallographic
software package. The refinement converged with R ) 0.0717 and
) 0.0840. Crystal data are summarized in Table 1.
Reduction of cis-Dimethyldiazene. (A) Time-Course Studies. A
R
w
Results and Discussion
1
0
25 mL flask was charged with 0.030 g (160 µmol) of CoCp
.030 g (210 µmol) of Lut‚HCl. Acetonitrile was then added (35.5
CN solution of
] (0.45 mL, 2.2 µmol) was then
2
and
Synthesis and Crystallographic Results. After some 20
mL) to form a slurry. An aliquot of a 4.8 mM CH
Et N) [(Cl -cat)(CH CN)MoFe Cl
3
years of very thorough studies, the synthesis, ligand-substitution
(
4
2
4
3
3
S
4
3
3+
properties, and reactivity of clusters with the [MoFe3S4] core
added, followed immediately by an aliquot of the cis-dimethyldiazene
solution (32 µmol). The addition of the substrate marked t ) 0 h. At
t ) 0.5, 1.0, 2.0, 3.5, and 5 h, a 100 µL sample of head space gas was
7
are well established. It has been demonstrated that (a) the
catecholate moiety on the Mo atom of the cubane can be
2
2
removed through protonation by acidic catechols and poly-
obtained from the reaction flask and analyzed for CH
4
. In addition,
1
1
3
.0 mL samples were removed from the reaction flask and placed in 5
carboxylic acids and (b) the single labile solvent molecule on
the Mo atom in (Et4N)2[(Cl4-cat)(CH3CN)MoFe3S4Cl3] and
related clusters can be readily replaced by any number of
mL septa-capped vials. These aliquots were immediately quenched
with 0.2 mL of a 1.0 M aqueous HCl solution. The vials were analyzed
for methylamine, ethylamine, and ammonia. Reactions were performed
7
σ-bases but generally not by π-acids except in the case of
under both N
conditions but only sampled at 3.0 h.
B) Phosphine-Inhibition Studies. In a 20 mL vial fitted with a
septa, 8.3 mL of a 7.2 mM CoCp solution (CH CN, 60 µmol), 8.3
mL of a 10.9 mM Lut‚HCl solution (CH CN, 90 µmol), and an
appropriate amount of a 9.5 mM PEt solution (CH CN, from 0-10
equiv based on cubane concentration) were combined. A 0.2 mL
aliquot of a 4.8 mM cluster solution (CH CN, 0.96 µmol) was then
added, followed immediately by a 0.08 mL aliquot of the substrate
solution (CH CN, 15 µmol). Addition of the substrate marked t ) 0
2
and Ar. Also, “blanks” were run under the same
23
reduced cores. It is not unexpected, therefore, that amines
will readily coordinate to the Mo atom of the cluster through
the N-lone pair, replacing the relatively poor CH3CN ligand.
The sole purpose in verifying this ligand substitution syntheti-
cally and crystallographically is to establish unambiguously the
Mo-amine interaction in light of the result that neither MeNH2
(
2
3
3
3
3
3
+
nor EtNH2 are reduced by the [MoFe3S4] core (Vide infra).
The methylamine single cubane I, shown in Figure 2,
crystallizes with two distinct anions within the asymmetric unit,
one in which all atoms are located on general positions (A)
3
3
h. A 2.0 mL aliquot of the reaction solution was obtained at t ) 1 h
and immediately quenched with 0.2 mL of a 1.0 M aqueous HCl
solution in a 5 mL septa-capped vial. Vials were subsequently analyzed
(21) In order to prepare a standard solution of 1,2-dimethylhydrazine in
CH3CN, a slight excess of Et3N was used to neutralize and hence solubilize
the dihydrochloride salt.
for methylamine and ammonia. In some reactions, (Ph
was used in place of the (Et N) [(Cl -cat)(CH CN)MoFe
C) Investigation of Kinetics. Reactions were prepared in 125 mL
flasks. The total solution volume was brought up to 40 mL with
acetonitrile. Aliquots of acetonitrile solutions of CoCp (15.0 mL, 108
4
P)
2
[Fe
4 4 4
S Cl ]
4
2
4
3
3 4 3
S Cl ] cluster.
(
22) Palermo, R. E.; Singh, R.; Bashkin, J. K.; Holm, R. H. J. Am. Chem.
Soc. 1984, 106, 2600.
23) Mascharak, P. K.; Armstrong, W. H.; Mizobe, Y.; Holm, R. H. J.
Am. Chem. Soc. 1983, 105, 475.
(
(
2

Reduction of cis-Dimethyldiazene by the [MoFe3S4]3+
J. Am. Chem. Soc., Vol. 119, No. 7, 1997 1665
Figure 2. An ORTEP plot (30% probability ellipsoids) showing the
anion of (Et N) [(Cl -cat)(CH NH )MoFe S Cl ] (IA).
4 2 4 3 2 3 4 3
Figure 4. Product distributions (plotted as HPLC response) after 3 h
for reaction systems with all reagents present except (A) cis-dimeth-
3+
yldiazene, (B) the [MoFe
sources, and (D) with 10 equiv of PEt
the catalyst. (E) A typical product distribution for the reduction of cis-
dimethyldiazene (15-fold excess) by the (Et N) [(Cl -cat)(CH CN)-
MoFe Cl ] catalyst. Ethylamine production results from the apparent
3
S
4
]
catalyst, and (C) proton and electron
3
added to inhibit the Mo site of
4
2
4
3
S
3 4
3
reduction of acetonitrile. See Experimental Section for reaction
conditions. (Bars are from left to right: ammonia, ethylamine,
methylamine/10.)
the substrate, (b) the [MoFe3S4]3 catalyst, and (c) proton and
electron sources and (d) with 10 equiv of PEt3 added to inhibit
the Mo site. As shown in Figure 4, while methylamine is
present in all systems, amounts of methylamine from these
blanks is typically e10% of the amounts detected from reactions
where all reagents are present and the [MoFe3S4]3 cluster is
present as a catalyst. These results taken together strongly
suggest that cis-dimethyldiazene is catalytically reduced by the
+
Figure 3. Time course of methylamine production from the reduction
of cis-dimethyldiazene, plotted as HPLC response (area) vs t.
+
[
(Et
4
N)
o
2
[(Cl
4 3 3 4 3
-cat)(CH CN)MoFe S Cl ] ] ) 54 µM, [cis-dimethyldi-
azene]
) 800 µM. Conversion at 5 h is approximately 90%.
3
+
and the second which is bisected by a crystallographic mirror
plane through Mo1, Fe2, S2, and S4 (B). As a result, there are
[MoFe3S4] cuboidal core in the presence of externally added
protons and electrons.
1
2 anions per unit cell. The structure of I shows the Mo single
The integrity of the (Et4N)2[(Cl4-cat)(CH3CN)MoFe3S4Cl3]
catalyst at the end of the reaction was verified through EPR
spectroscopy. After removal of excess CoCp2 (S ) /2) from
cubane ligated by methylamine (Figure 2). The nitrogen atom
of methylamine in IB is located on the mirror plane. As a result,
the methyl carbon (C1) is somewhat disordered over two
positions but is modeled sufficiently by a 50% occupancy on
both positions. The Mo-N distance in both anions is very
nearly the same at 2.28(2) and 2.35(4) Å. The average
Mo- -Fe (2.737(6) Å), Mo-S (2.346(9) Å), Mo-O (2.09(2)
Å), Fe-S (2.26(1) Å), Fe-Cl (2.22(1) Å), and Fe- -Fe
1
the reaction mixture (THF washings were essentially colorless),
an EPR spectrum of the reaction solution was obtained. The
3
3+
characteristic S ) /2 spectrum for the [MoFe3S4] core was
observed, and the signal integrated to 1.1 ( 0.3 mM, suggesting
no cluster decomposition during the first 90 min of reaction
time. Previous studies on the identification of the recovered
(Et4N)2[(Cl4-cat)(CH3CN)MoFe3S4Cl3] cluster after catalytic
cycles included elemental analysis, electronic and infrared
spectroscopy, and demonstration of further catalysis by the
(2.727(7) Å) distances are unexceptional. Given the abundance
7
of Mo-cubane structures in the literature, no further description
of this structure is warranted.
9,10
Substrate Reductions. As a reference point, catalytic
reductions were generally performed with the substrate in
approximately 15-fold excess relative to the catalyst, with
externally added sources of electrons (CoCp2, g4 equiv) and
protons (Lut‚HCl, g6 equiv). As shown in Figure 3, cis-
dimethyldiazene is catalytically reduced to methylamine under
these conditions. There is an initial burst to approximately 50%
conversion (based on 2 mol of methylamine for every mole of
cis-dimethyldiazene) in the first hour of reaction time for the
recovered cubane.
These results taken together demonstrate
the robust nature of (Et4N)2[(Cl4-cat)(CH3CN)MoFe3S4Cl3]
under reaction conditions and identifies the cubane as the active
catalyst for the reduction of a variety of substrates.
Ammonia yields for the reactions shown in Figure 4 were
(1) exceedingly small, only reaching approximately 1% of the
total yield of methylamine in the catalytic systems, and (2)
essentially equivalent for all systems, catalytic reactions and
blanks alike. It was therefore concluded that the reduction of
cis-dimethyldiazene with these synthetic clusters led to forma-
tion of methylamine but not ammonia, unlike what is observed
in the enzyme system. No change in product distribution was
observed when the reaction was performed under Ar instead of
N2. Given the inability for N2 to serve as a ligand for the Mo
cluster in either the 3+ or 2+ state, this result was expected.
The lack of reduction products derived from C-N bond
reduction in cis-dimethyldiazene (the only route leading to NH3)
was verified in two individual experiments. First, GC analysis
of the headspace gases in the catalytic systems routinely showed
1
5:1 substrate/catalyst ratio, with conversion slowly leveling
off. This behavior may be due to a slow precipitation of the
catalyst as cations are generated in solution (i.e., CoCp2 ,
+
+
MeNH3 ), a behavior typically observed in the reduction
systems previously investigated.9
In order to verify that (1) the substrate was being activated
3+
and reduced by the [MoFe3S4] core exclusively and (2)
methylamine was obtained only from reduction of cis-dimeth-
yldiazene, a number of “blanks” were performed. These blanks
included reaction systems with all reagents present except (a)

1666 J. Am. Chem. Soc., Vol. 119, No. 7, 1997
Malinak et al.
no CH4 up through a 5 h period. Methane would necessarily
have to be present in a 1:1 ratio with any ammonia that was
produced in the system. Second, in experiments using methyl-
3+
amine as a “substrate” for the [MoFe3S4] cluster, no ammonia
or methane were observed at any time as a result of potential
reduction of the C-N bond. Given that methylamine is a better
ligand for the [MoFe3S4]3 cluster than CH3CN on the basis of
synthetic and crystallographic results (Vide supra), the lack of
reactivity observed with methylamine is due to its inability to
be reduced by the [MoFe3S4]3 core and not to its inability to
interact with the catalyst. The inability of the model systems
to reduce methylamine may be due to the lack of a free lone
pair needed for protonation once the substrate is coordinated to
the catalyst. In model systems employing hydrazine as a
substrate, protonation of the bound hydrazine was shown to be
a necessary first step in making the reduction potential of the
cluster accessible.9 No turnover of hydrazine to ammonia was
observed in systems which lacked sufficiently acidic protons
to protonate hydrazine.
+
+
3
Figure 5. Inhibition of cis-dimethyldiazene reduction by PEt . See
Experimental Section for reaction conditions.
The ability of the [MFe3S4]n+ (M ) Mo, n ) 3; M ) V, n
)
2) cores to catalytically reduce hydrazine to ammonia has
9
been established previously. As a matter of principle, however,
it was important to verify that 1,2-dimethylhydrazine (a possible
reduced intermediate in the reduction of cis-dimethyldiazene)
reacted in a similar manner. Under catalytic conditions (15-
fold excess of 1,2-dimethylhydrazine relative to catalyst),
approximately 80% of the substrate was reduced to methylamine
within the first hour, a yield essentially identical to that reported
for the unsubstituted hydrazine.9 A “blank” reaction which
contained no catalyst showed the typical background levels of
methylamine (6%) established previously. In these reactions,
Lut‚HBPh4 was used instead of Lut‚HCl to insure that a majority
of the substrate stayed soluble initially, given that the dihydro-
chloride salt of 1,2-dimethylhydrazine is quite insoluble in
CH3CN.
e
o
Figure 6. Reaction velocity (ν) vs [cis-dimethyldiazene] . Reaction
velocities are based on the first 60 min of reaction time. See
Experimental Section for reaction conditions.
In addition to methylamine, it was established by HPLC and
GC that ethylamine is also a product in these systems (Figure
4
). Yields of ethylamine (a) are largest when there is no cis-
dimethyldiazene in the system, (b) are essentially nonexistent
in the absence of catalyst or protons/electrons, (c) are present
in lower amounts when cis-dimethyldiazene is present, and (d)
are not observed when 10 equiv of phosphine (per cluster) are
3+
present in the system. These results suggest that the [MoFe3S4]
core may reduce the solvent, CH3CN, to ethylamine at a very
slow rate. When “better” ligands are present (phosphine,
hydrazine, cis-dimethyldiazene), the yields are either decreased
or nonexistent. In separate experiments, it was verified that
while ethylamine binds to the Mo atom of the cluster (II), it is
not reduced to ammonia and methane.
Figure 7. Percent conversion of cis-dimethyldiazene to methylamine
vs [cis-dimethyldiazene] .
o
catalysis. Additionally, as shown in Figure 5, it is clear that as
phosphine is added in greater amounts to the reaction system
(from 0 to 10 equiv of phosphine), the turnover of substrate to
Role of the Mo Atom in Substrate Reduction. It has been
products drops dramatically. At 1 equiv of phosphine, the
amount of product in 1 h drops to about half of the amount
obtained when [PEt3] ) 0. Between 5 and 6 equiv of phosphine,
the amount of methylamine present in the reaction solutions
was determined to be well within background limits, indicating
complete inhibition. These results taken with those results
obtained previously with other substrates demonstrates the
importance of the Mo atom in substrate binding and activation
toward reduction.
Investigation of Kinetics. From initial time-course studies,
it was determined that the conversion of substrate to product
with time was essentially linear within the first 60 min.
Therefore, catalytic systems were performed over a number of
cis-dimethyldiazene concentrations and analyzed at t ) 1 h. As
shown in Figure 6, it is clear that typical saturation kinetics
observed for enzyme systems are not applicable in this model
well-established in the reduction of other nitrogenase substrates
n+
by the synthetic [MFe3S4] clusters that the heterometal (either
9
Mo or V) is either exclusively (in the case of hydrazine) or
predominantly (in the case of C2H2)10 involved in the activation
of the substrate toward reduction. The role of the heterometal
in the cis-dimethyldiazene system was investigated by (1) using
2-
3+
the [Fe4S4Cl4] as a potential catalyst in place of the [MoFe3S4]
core and (2) observing the effect of the addition of phosphine
to the catalytic system. Phosphine is known to bind strongly
21
to the Mo atom and presumably dramatically affects the ability
of other, weaker σ-bases to coordinate.
By using the all-Fe cluster as a potential catalyst, the amounts
of methylamine observed after 1 h were well within back-
ground limits established with the blanks, suggesting that the
Fe atoms in the [MoFe3S4]3 cluster are not involved in
+

Reduction of cis-Dimethyldiazene by the [MoFe3S4]3+
J. Am. Chem. Soc., Vol. 119, No. 7, 1997 1667
+
]3 cuboidal cores.
Figure 8. Possible reaction pathways for the reduction of cis-dimethyldiazene by synthetic [MoFe
S
3 4
system. This observed trend may be due in part to insolubility
Summary and Conclusions
+
of the catalyst as cations are generated in solution (CoCp2 ,
+
MeNH3 ). At lower substrate:catalyst ratios, the catalyst
The nitrogenase enzyme has recently been shown to catalyze
the reduction of cis-dimethyldiazene to methane, ammonia, and
methylamine. The nitrogenase model compound (Et N) [(Cl -
remains in solution to effect near quantitative reduction within
the first hour. At higher concentrations of substrate, the reaction
velocity drops to a minimum, followed by a gradual increase.
As shown in Figure 7, however, the percent conversion of cis-
dimethyldiazene to methylamine (based on 2 equiv of product
per mole of substrate) is essentially constant at high [cis-
dimethyldiazene].
4
2
4
cat)(CH CN)MoFe S Cl ] also reduces the NdN bond in cis-
3
3
4
3
dimethyldiazene, but methylamine is the only obserVable
product. The inability of the model compound to reduce
methylamine, even though it has been unambiguously demon-
strated that methylamine binds to the Mo atom in the cluster,
may account for the lack of methane and ammonia formation.
Insofar as the cuboidal clusters can be considered reactivity
models for the cofactor of nitrogenase, the ability of the
A Comparison between the Synthetic and the Enzymatic
Systems. It has been reported15 that as a substrate for the A.
Vinelandii nitrogenase, cis-dimethyldiazene is reduced to NH3,
CH4, and CH3NH2 in a ratio of 1.0:1.0:1.6 ( 0.1 (Fe:FeMo
protein ratio of 6.6), in contrast to the synthetic system
3
+
[
MoFe3S4] cores to reduce both NdN and NsN exclusiVely
employing [MoFe3S4]3+ as catalyst, where methylamine is
at the Mo site may suggest the Mo-atom in the cofactor is, in
1
2
fact, not “innocent” as suggested in the past, but rather plays
a vital role in the later stages of dinitrogen reduction. We have
previously considered the possibility of initial activation of
dinitrogen at the unique, three-coordinate Fe atoms of the
cofactor, followed by “migration” of a partially -reduced species
(hydrazine) to the peripheral Mo atom where reduction to
produced exclusively. In the synthetic system, methylamine is
not reduced and hence no methane or ammonia are observable
as products. It seems likely that the reaction proceeds by one
of two pathways (assuming mononuclear activation), as depicted
in Figure 8. The coordinated cis-dimethyldiazene may be
reduced in a “symmetric” way, forming dimethylhydrazine as
an intermediate, which has been shown to yield methylamine
upon protonation and reduction. Alternatively, the substrate may
be reduced in an “unsymmetric” manner, forming methylamine
and a Mo-bound imine after the first two-electron, two-proton
step. Either way, it is clear in the model system that methyl-
amine is the only product and that the Mo atom seems to be of
primary importance.
In the enzymatic system, it seems likely that either (a) the
NdN bond and the CsN bond are is some way reduced
concomitantly or (b) methylamine is an initial product, which
is activated in some way by the nitrogenase cofactor. If the
latter, the methylamine would have to bind to a center that was
capable of removing significant electron density from the C-N
bond given that only one lone pair of electrons are available
and, when bound to a metal center, there would be no place for
protonation to occur on methylamine. Either way, it is clear
that a different mechanism is occurring in the biological and
abiological systems, specifically with regard to C-N bond
cleavage.
9a
ammonia proceeds. This work not only supports this pos-
sibility, but also suggests that the Mo atom may be involved in
the reduction process at the diazene level.
Acknowledgment. The support of this work by grants from
NIH (GM-33080), NRICGP/CREES/USDA (95-37305-2068)
and by a Dean’s Innovative Research Award to C. E. McKenna
is gratefully acknowledged. We also thank Nam Doo Moon
and W. R. Dunham for obtaining EPR spectra.
Supporting Information Available: Tables of fractional
atomic coordinates, calculated positions of H atoms and
anisotropic displacement parameters for I (6 pages). See any
current masthead page for ordering and Internet access
instructions.
JA963475S