Stereospecificity of acetylene reduction catalyzed by nitrogenase

Article abstract of DOI:10.1021/ja003662x

In addition to catalyzing the reduction of dinitrogen to ammonia, the metalloenzyme nitrogenase catalyzes the reduction of a number of alternative substrates, including acetylene (C₂H₂) to ethylene (C₂H₄) and, in certain cases, to ethane (C₂H₆). The stereochemistry of proton addition for C₂D₂ reduction to C₂D₂H₂ catalyzed by the Mo-dependent nitrogenase has been used to probe substrate binding and proton addition mechanisms. In the present work, the C₂D₂ reduction stereospecificity of altered MoFe proteins having amino acid substitutions within the active site FeMo-cofactor environment was examined by Fourier transform infrared (FTIR) spectroscopy. Altered MoFe proteins examined included those having the α-subunit 96^Arg residue substituted by Gln, Leu, or Ala, the α-subunit 69^Gly residue substituted by Ser, and the α-subunit 195^His residue substituted by Asn. The stereochemistry of proton addition to C₂D₂ does not correlate with the measured K_m values for C₂H₂ reduction, or with the ability of the enzyme to reduce C₂H₂ by four electrons to yield C₂H₆. Instead, the electron flux through nitrogenase was observed to significantly influence the ratio of cis- to trans- 1,2-C₂H₂D₂ formed. Finally, the product distribution observed for reduction of C₂H₂ in D₂O is not consistent with an earlier proposed enzyme-bound intermediate. An alternative model that accounts for the stereochemistry of C₂H₂ reduction by nitrogenase based on a branched reaction pathway and an enzyme-bound η²-vinyl intermediate is proposed.

Full text of DOI:10.1021/ja003662x

1822
J. Am. Chem. Soc. 2001, 123, 1822-1827
Stereospecificity of Acetylene Reduction Catalyzed by Nitrogenase
†
‡
,‡
,†
Paul M. C. Benton, Jason Christiansen, Dennis R. Dean,* and Lance C. Seefeldt*
Contribution from the Department of Chemistry and Biochemistry, Utah State UniVersity,
Logan, Utah 84322, and Department of Biochemistry, Virginia Tech, Blacksburg, Virginia 24061
ReceiVed October 11, 2000
Abstract: In addition to catalyzing the reduction of dinitrogen to ammonia, the metalloenzyme nitrogenase
catalyzes the reduction of a number of alternative substrates, including acetylene (C2H2) to ethylene (C2H4)
and, in certain cases, to ethane (C2H6). The stereochemistry of proton addition for C2D2 reduction to C2D2H2
catalyzed by the Mo-dependent nitrogenase has been used to probe substrate binding and proton addition
mechanisms. In the present work, the C2D2 reduction stereospecificity of altered MoFe proteins having amino
acid substitutions within the active site FeMo-cofactor environment was examined by Fourier transform infrared
Arg
(
FTIR) spectroscopy. Altered MoFe proteins examined included those having the R-subunit 96 residue
Gly
His
substituted by Gln, Leu, or Ala, the R-subunit 69 residue substituted by Ser, and the R-subunit 195 residue
substituted by Asn. The stereochemistry of proton addition to C2D2 does not correlate with the measured Km
values for C2H2 reduction, or with the ability of the enzyme to reduce C2H2 by four electrons to yield C2H6.
Instead, the electron flux through nitrogenase was observed to significantly influence the ratio of cis- to trans-
1
,2-C2H2D2 formed. Finally, the product distribution observed for reduction of C2H2 in D2O is not consistent
with an earlier proposed enzyme-bound intermediate. An alternative model that accounts for the stereochemistry
2
of C2H2 reduction by nitrogenase based on a branched reaction pathway and an enzyme-bound η -vinyl
intermediate is proposed.
Biological nitrogen fixation is catalyzed by nitrogenase, a
two-component metalloenzyme responsible for the MgATP-
observed by monitoring the stereoisomers of C2D2H2 formed
when C2D2 is used as the substrate and reduction is catalyzed
in the presence of protons. For wild-type nitrogenase, the
reductive addition of protons to C2D2 occurs via a highly
stereospecific mechanism as revealed by the production of
∼96% of the product as the cis isomer of dideuterated ethylene
(cis-1,2-C2H2D2), with the remaining ∼ 4% being the trans
1-3
dependent reduction of dinitrogen to ammonia.
The Mo-
dependent nitrogenase is composed of an Fe protein component,
which possesses the site of MgATP hydrolysis and is the
electron donor to the MoFe protein component. Substrate
binding and reduction occurs within the MoFe protein at a
metallocluster site called FeMo-cofactor (7Fe:9S:Mo:homoci-
8
-12
isomer.
In addition, little or no mono- or trideuterated
4
trate). In addition to its physiological substrate dinitrogen (N2),
ethylene products are detected. The stereospecificity of proto-
nation of C2D2 can be perturbed in certain altered MoFe proteins
that have amino acid substitutions within the first shell of
nitrogenase also catalyzes the reduction of a variety of other
5,6
multiply bonded substrates, the best known being acetylene
7
12
(
C2H2), which is usually reduced by two electrons to yield
noncovalent interactions with the active site FeMo-cofactor.
ethylene (C2H4). Insights into the catalytic mechanism of
nitrogenase have come from analysis of the stereochemistry of
protonation of C2D2 upon its reduction.8-11 This feature can be
Further, in addition to C2H4, some of these altered proteins
1
3
reduce C2H2 by four electrons to yield C2H6. Thus, analysis
of these substituted proteins provides a way to elucidate further
details of the C2H2 reduction mechanism of nitrogenase.
Recently, three such altered MoFe proteins, one having the
*
Address correspondence to these authors: LCS, phone (435) 797-3964,
fax (435) 797-3390, e-mail seefeldt@cc.usu.edu; DRD, phone (540) 231-
Gln
Lys
R-subunit 191 residue substituted by Lys (R-191 ) and the
5
895, fax (540) 231-7126, e-mail deandr@vt.edu.
His
†
others having the R-subunit 195 residue substituted by either
Utah State University.
‡
Gln
Asn
Virginia Tech.
Gln(R-195 ) or Asn (R-195 ), were examined for their
(
1) Burgess, B. K.; Lowe, D. J. Chem. ReV. 1996, 96, 2983-3011.
12
stereospecific protonation of C2D2 and production of C2H6.
(2) Dean, D. R.; Bolin, J. T.; Zheng, L. J. Bacteriol. 1993, 175, 6737-
From those studies it was concluded that loss of cis proton
addition to C2D2 is correlated with a change in the Km for C2H2
reduction and with the production of C2H6. A model was
proposed where an increase in Km results in a longer active site
residence time of a common ethylenic intermediate, leading to
the further reduction of C2H2 to yield C2H6 and the loss in
stereospecificity in protonation of C2D2 to yield a higher fraction
6
3
744.
3) Howard, J. B.; Rees, D. C. Chem. ReV. 1996, 96, 2965-2982.
5
80.
(
(
7) Dilworth, M. J. Biochim. Biophys. Acta 1966, 127, 285-294.
8) Lin-Vien, D.; Fateley, W. G.; Davis, L. C. Appl. EnViron. Microbiol.
12
of trans-1,2-C2D2H2. In the present work the stereospecificity
1
989, 55, 354-359.
9) McKenna, C. E.; McKenna, M. C.; Huang, C. W. Proc. Natl. Acad.
of C2D2 reduction catalyzed by other altered MoFe proteins was
(
Sci. U.S.A. 1979, 76, 4773-4777.
(12) Fisher, K.; Dilworth, M. J.; Kim, C. H.; Newton, W. E. Biochemistry
2000, 39, 2970-9.
(13) Scott, D. J.; May, H. D.; Newton, W. E.; Brigle, K. E.; Dean, D. R.
Nature 1990, 343, 188-190.
(
10) Kelly, M. Biochim. Biophys. Acta 1969, 191, 527-540.
(11) Hardy, R. W. F.; Holsten, R. D.; Jackson, E. K.; Burns, R. C. Plant
Physiol. 1968, 43, 1185-1207.
1
0.1021/ja003662x CCC: $20.00 © 2001 American Chemical Society
Published on Web 02/07/2001

Stereospecific C2D2 Reduction
J. Am. Chem. Soc., Vol. 123, No. 9, 2001 1823
Table 1. Acetylene Reduction Properties for Wild-Type and Selected Altered MoFe Proteins
a
1,2-C
H
2 2
D
2
c
C
2
H
6
V
max
b
K c
d
MoFe protein
wild type
cis (%)
trans (%)
production (%)
m
(atm)
(nmol/min/mg)
96
87
86
88
95
67
65
4
13
14
12
5
33
35
0
0
0
0
0
19
21
0.006 ( 0.001
0.008 ( 0.001
0.002 ( 0.0002
0.005 ( 0.0004
0.14 ( 0.01e
0.007 ( 0.001
0.17 ( 0.08
2020 ( 118
1540 ( 17
303 ( 15
Gln
R-96
Leu
R-96
Ala
R-96
1710 ( 51
Ser
e
R-69
1800 ( 50
110 ( 2
Asn
R-195
R-69Ser/R195Asn
270 ( 20
a
Determined using a 20:1 molar ratio of Fe protein to MoFe protein. b Value quoted as a percent of total hydrocarbon production. Zero indicates
the lower level of detection, which is <0.1%. Determined using a 40:1 molar ratio of Fe protein to MoFe protein. d Per milligram of MoFe
protein. Taken from ref 22.
c
e
examined. On the basis of these results a different model is
proposed.
Quantification of Deuterated Ethylenes by Fourier Transform
Infrared (FTIR)1 Spectroscopy. The generation of the gaseous
substrate C
O to a specific quantity of CaC
milligrams of CaC was transferred to a 25 mL vial that was
subsequently capped and evacuated. D O (0.6 mL) was slowly added
to the sample and C evolution was allowed to proceed for 1 h. A 5
2 2
D was accomplished by adding an appropriate volume of
1
2
D
2
2
as previously described. Eighty
Experimental Procedures
2
Nitrogenase Proteins. Wild-type Fe and MoFe proteins were
2
14
expressed in Azotobacter Vinelandii cells and purified to homogeneity
2 2
D
1
5
as previously described. The specific activity for C
H
2 2
reduction
mL aliquot of this gas was then transferred by gastight syringe to an
assay flask under 1.0 atm of argon. Each 120 mL assay vial contained
200 µmol of phosphocreatine, 600 µmol of MOPS buffer, pH 7.0, 100
catalyzed by wild-type nitrogenase was >2000 nmol of product/min/
mg of protein. Strains that produce altered MoFe proteins were
constructed as previously described for other mutant strains of A.
µmol of sodium dithionite, 60 µmol of MgCl , 30 µmol of ATP, 1.85
2
16
Vinelandii. MoFe proteins were purified by an immobilized Zn affinity
mg of creatine phosphokinase, and 12 mg of bovine serum albumin in
a liquid volume of 10 mL. MoFe protein (5 mg) was then added and
the reaction initiated by the addition of wild-type Fe protein. The assays
were incubated with shaking in a water bath at 30 °C with the reaction
being allowed to proceed for 15 min before termination by addition of
2.5 mL of a 0.4 M EDTA, pH 7.5, solution. The assay flask was then
connected to an evacuated 100 mL volume infrared gas cell (10 cm
path length) via a gastight needle and the gaseous contents equilibrated.
The gaseous products were analyzed by FTIR spectroscopy by using a
Mattson Galaxy Series 5000 spectrometer (Madison, WI) in conjunction
with WinFirst software. All spectra were the average of 16 scans with
liquid chromatography method using a step gradient of 250 mM
1
7
imidazole for elution. Protein concentrations were determined by a
1
8
modified biuret method with bovine serum albumin as the standard.
Proteins were judged to be homogeneous on the basis of SDS-PAGE19
with coomassie blue staining. All protein manipulations were conducted
in the absence of O
Liquid and gas transfers were accomplished with gastight syringes.
Reduction Assays. C reduction rates for the wild-type and
2
in septum sealed vials under an argon atmosphere.
C H
2 2
2 2
H
altered MoFe proteins were measured at 30 °C with each assay
containing 0.05 mg of MoFe protein together with a 40-fold molar
excess of wild-type Fe protein. Each assay vial contained 20 µmol of
phosphocreatine, 60 µmol of MOPS buffer, pH 7.0, 10 µmol of sodium
-
1
2 2 2
a 1 cm resolution. The ratio of cis- and trans-1,2-C D H produced
was established by measuring the heights of the absorption maxima in
the infrared spectrum for each of the isomers. The vibrational modes
dithionite, 6 µmol of MgCl
phosphokinase, and 1.2 mg of bovine serum albumin in a total volume
of 1.0 mL. C was added as an overpressure to each assay vial
2
, 3 µmol of ATP, 0.185 mg of creatine
monitored in the infrared spectrum for each ethylenic species were V
7
-
1
-1
H
2
for nondeuterated ethylene at 949 cm , V
8
for C
for trans-1,2-C
Concentrations were determined from
2 3
H D at 943 cm , V
7
2
-
1
-1
containing 1.0 atm of argon, which was then vented to atmospheric
pressure. The assays were initiated by addition of the wild-type Fe
protein and were incubated for 8 min with gentle shaking in a water
bath at 30 °C. All reactions were terminated by the addition of 250 µL
for cis-1,2-C
2
H
2
D
3
2
at 843 cm , V
4
2 2 2
H D at 988 cm ,
-
1 8,20
and V for C HD at 918 cm .
8
2
the peak height of the noted absorption band. The molar absorptivities
for each species were taken as equal, with the exception of that for the
8
of a 0.4 M EDTA, pH 7.5, solution. The production of C
2
H
4
and C
2
H
6
trans-1,2-C H D
2 2 2
isomer, which has been reported to be about half of
was quantified by gas chromatography on a Shimadzu GC-8A gas
the other values.
chromatograph with a flame ionization detector fitted with a 30 cm ×
To assess how varying flux affects the relative formation of cis-
and trans-1,2-C D H catalyzed by wild-type and altered MoFe proteins,
0
.3 cm Porapak N column with nitrogen as the carrier gas.
2
2
2
Kinetic Parameters for Acetylene Reduction. To determine the
the procedure described above was repeated with Fe protein-to-MoFe
protein molar ratios in the range between 20:1 and 1:10. This experiment
was done in two different ways. In one approach, the concentration of
MoFe protein was held constant and the Fe protein concentration was
varied. In the second approach, the total sum of the protein concentration
was held constant while the relative amounts of Fe protein and MoFe
protein varied to achieve the desired molar ratio. The actual Fe protein-
to-MoFe protein ratios are listed in Table 1 and the figure legends.
kinetic parameters for C
described above except the partial pressure of C
2
H
2
reduction, assays were performed as
introduced into a
2 2
H
particular assay vial was varied. Each assay series was independently
run three times with individual data points being measured in triplicate.
The final quoted values for the K and Vmax (Table 1) were calculated
m
by averaging the results obtained from triplicate assays. The kinetic
constants were determined by fitting the rate versus concentration data
to the Michaelis-Menten equation with use of the computer program
Igor Pro (Wavemetrics, Lake Oswego, OR).
C H reduction assays were also performed as described above except
2
2
that either C H or C D was used in a reaction solution of >90% D O.
2
2
2
2
2
Product distribution was determined by FTIR analysis of the gas phase
as described above.
(
14) Seefeldt, L. C.; Morgan, T. V.; Dean, D. R.; Mortenson, L. E. J.
Biol. Chem. 1992, 267, 6680-6688.
15) Burgess, B. K.; Jacobs, D. B.; Stiefel, E. I. Biochim. Biophys. Acta
980, 614, 196-209.
16) Brigle, K. E.; Setterquist, R. A.; Dean, D. R.; Cantwell, J. S.; Weiss,
M. C.; Newton, W. E. Proc. Natl. Acad. Sci. U.S.A. 1987, 84, 7066-7069.
17) Christiansen, J.; Goodwin, P. J.; Lanzilotta, W. N.; Seefeldt, L. C.;
Dean, D. R. Biochemistry 1998, 37, 12611-12623.
18) Chromy, V.; Fischer, J.; Kulhanek, V. Clin. Chem. 1974, 20, 1362-
363.
19) Hathaway, G. M.; Lundak, T. S.; Tahara, S. M.; Traugh, J. A.
Methods Enzymol. 1979, 60, 495-511.
(
1
Results
(
Acetylene Reduction by Altered MoFe Proteins. Substitu-
tion of certain amino acids which provide the first shell of
noncovalent interactions with FeMo-cofactor is known to affect
(
(
1
(
(20) Crawford, B. L.; Lancaster, J. E.; Inskeep, R. G. J. Chem. Phys.
1953, 21, 678-686.

1824 J. Am. Chem. Soc., Vol. 123, No. 9, 2001
Benton et al.
Figure 1. FeMo-cofactor protein environment. Wall-eyed stereoview
9
6
191
195
of the FeMo-cofactor and R-Arg , R-Gln , and R-His . The figure
3
4
was generated from the MoFe protein coordinates by using the
programs MOLSCRIPT and RASTER 3D.
3
5
36
the reactivity of nitrogenase toward various substrates and
2
1
inhibitors. Figure 1 is a stereoview illustrating the FeMo-
cofactor structure and the amino acids examined here. The effect
of substitution of these residues on the stereochemical mech-
anism of proton addition during nitrogenase reduction of C2D2
was examined. Our rationale for these experiments was to test
the hypothesis that the [4Fe-4S] face of FeMo-cofactor ap-
Arg
proached by the R-subunit 96
residue provides the initial
substrate-binding site.2
1-23
It was expected that, if this face
Figure 2. The effect of decreasing electron flux on the production of
2 2 2
trans-1,2-C H D for wild type and altered MoFe proteins. Assay
conditions and ethylene stereoisomer quantification were conducted as
described in the Experimental Procedures. The percentage of trans-
provides the substrate-binding site, substitution of amino acids
approaching this face could alter steric constraints that might
be manifested in a loss of stereospecificity for proton addition
during C2D2 reduction. A list of altered MoFe proteins examined
and the affects of amino acid substitutions on kinetic parameters
and stereospecificity are shown in Table 1. The Km values for
1
2 2 2
,2-C H D relative to the total ethylene produced is plotted against
the MoFe protein to Fe protein ratio. Panel A: MoFe proteins with R-
Ala
Gln
Leu
9
6
(3), R- 96 (b), and R- 96 (O). Panel B: MoFe proteins
Asn Ser Asn
with R-195 /R-69 (4) and R-195 (2). Panel C: Wild-type MoFe
protein (0) and MoFe protein with R-69 (9).
C2H2 reduction for altered MoFe proteins range from values
Ser
Leu
near the wild-type Km of 0.0064 atm to 3-fold lower (R-96
)
Ser
Asn
and 27-fold higher (R-69 /R-195 ) than wild type. The Vmax
values ranged from those nearly the same as the wild type (2020
nmol/min/mg) to values as low as 5% of the wild type (110
nmol/min/mg for the R-195 -substituted protein). The R-
95 -substituted MoFe protein has been shown to contain less
Stereospecificity of Acetylene Reduction Is Flux Depend-
ent. For the nitrogenase system, flux is defined as the overall
rate at which sufficient electrons are accumulated within the
MoFe protein to achieve substrate reduction. Adjustment of
electron flux through nitrogenase is known to alter the pattern
Asn
Asn
1
12
than the full complement of FeMo-cofactor, so a true quantita-
tive value for Vmax is difficult to determine.
,8,10-12
1
of nitrogenase substrate specificity. For example, under a
nitrogen gas atmosphere, as the electron flux decreases, the
As previously reported,7
the wild-type nitrogenase
+
percentage of electrons going to H reduction increases while
2
4
under high electron flux conditions is highly specific for
catalyzing the reductive addition of two protons to C2D2,
yielding ∼96% of cis-1,2-C2H2D2, with the remaining ∼4%
the percentage going to N2 reduction decreases. One way to
regulate electron flux in vitro is by adjusting the ratio of Fe
protein to MoFe protein. At a high molar ratio of Fe protein to
MoFe protein (20:1), the electron flux is high, while at a low
ratio (1:5), the electron flux is low. The effect of varying the
MoFe protein to Fe protein ratio (i.e., electron flux) on the
stereospecificity of proton addition during C2D2 reduction is
shown in Figure 2. For all proteins examined, it was found that,
as the electron flux decreased, the percentage of trans-1,2-
C2H2D2 formed significantly increased. For the wild-type
nitrogenase (panel C), the percentage of trans isomer more than
doubled, increasing from 4% to 11% as the flux was decreased.
For the R-96-substituted proteins, the percentage of trans isomer
Ser
being the trans isomer. The R-69 -substituted MoFe protein
catalyzed formation of a similar product distribution of di-
deuterated ethylene isomers as observed for the wild-type
Leu
Gln
Ala
MoFe protein. In contrast, the R-96 -, R-96 -, R-96 -,
Asn
Ser
Asn
R-195 -, and R-69 /R-195 -substituted MoFe proteins
catalyzed formation of significantly higher percentages of the
trans isomer of dideuterated ethylene, ranging from 12 to 35%.
Asn
The ethylene product distribution for the R-195 -substituted
MoFe protein (33% trans isomer) is similar to the value reported
previously.12 Of the altered MoFe proteins examined in the
Asn
Asn
present study, only MoFe proteins having the R-195 substitu-
tion produced detectable levels of C2H6 (Table 1).
increased from ∼14% to between 27 and 36%. For the R-195
MoFe protein, the percentage increased from 33% to 40%. For
Asn
Ser
the doubly substituted R-195 /R-69 MoFe protein, the
percentage of trans-1,2-C2H2D2 formed increased from 35% to
a value approaching 50% as electron flux was decreased. In
the above experiments, the MoFe protein concentration was held
constant (5 mg in 10 mL) and the Fe protein concentration was
(
21) Christiansen, J.; Chan, J. M.; Seefeldt, L. C.; Dean, D. R. In
Prokaryotic Nitrogen Fixation: A Model System for Analysis of a Biological
Process; Triplett, E., Ed.; Horizon Scientific Press: Wymondham, UK,
2
000; pp 101-112.
22) Christiansen, J.; Cash, V. L.; Seefeldt, L. C.; Dean, D. R. J. Biol.
Chem. 2000, 275, 11459-64.
23) Christiansen, J.; Seefeldt, L. C.; Dean, D. R. J. Biol. Chem. 2000,
75, 36104-36107.
(
(
(24) Thorneley, R. N. F.; Lowe, D. J. In Molybdenum Enzymes; Spiro,
T. G., Ed.; Wiley: New York; 1985; pp 221-284.
2

Stereospecific C2D2 Reduction
J. Am. Chem. Soc., Vol. 123, No. 9, 2001 1825
Scheme 1
substitution of R-191Gln by Lys altered proton addition resulting
in ∼21% of the trans-dideuterated ethylene product compared
to 4% for the wild-type protein. It was observed that an increase
in trans-proton addition was correlated with both a change in
the Km for C2H2 and with the reduction of a small portion of
C2H2 by four electrons to yield C2H6. From this correlation, a
model was proposed linking the stereospecificity of proton
addition with the Km for C2H2 reduction and with the formation
Figure 3. FTIR spectra of the gas phases for wild-type and R-195Asn
2 2 2 2 2 2
reductions of C H or C D in D O with some H O. Assay conditions
and FTIR parameters were as described in the Experimental Procedures.
All reactions were conducted with an Fe protein to MoFe protein molar
ratio of 20:1 at 30 °C. Trace 1: Wild-type MoFe protein was allowed
1
2
of C2H6. In the present study, we have examined a number of
altered MoFe proteins for their stereospecificity of proton
addition to C2H2 as a way to test this model and to ask if such
changes in stereospecificty could be attributed to steric con-
straints imposed by the active site environment. As previously
2 2 2 2
to react with C H in a reaction solution of D O with some H O for 15
min. The spectrum was scaled by a factor of 0.05 for comparison to
Asn
the other spectra. Trace 2: R-195 MoFe protein was allowed to react
2 2 2 2
with C H in a reaction solution containing D O with some H O for
Asn
3
C
0 min. Trace 3: R-195 MoFe protein was allowed to react with
in a reaction solution containing D O with some H O for 15 min.
12
reported, we find that substitution of amino acid residues
D
2 2
2
2
located within close contact to the FeMo-cofactor can result in
perturbations in the stereochemistry of proton addition to C2D2.
However, we do not find that these perturbations correlate with
either the Km for C2H2 reduction or the production of C2H6.
varied to achieve the desired molar ratio. To rule out possible
protein concentration effects, the above experiment was repeated
except that the Fe protein and MoFe protein concentrations were
varied to achieve the desired molar ratio while maintaining the
same total protein concentration (10 mg in 10 mL). Essentially
the same ratio of cis- and trans-1,2-C2H2D2 were obtained by
both experimental approaches.
We also examined whether the Km for C2H2 reduction
changed as a function of electron flux under our experimental
conditions. For these experiments, a series of assay vials were
prepared with a fixed molar ratio of Fe protein and MoFe protein
and the concentration of C2H2 was varied. A fit of the C2H2
reduction rate versus initial substrate concentration to the
Michaelis-Menten equation provided the Km values at a
particular electron flux. It was observed that the Km for C2H2
did not change with changing electron flux. However, when
the above experiment was repeated, with the total protein
concentration held constant, the Km was found to increase 2-fold
with decreasing flux over the range examined.
Arg
For example, substituting R-96 by Leu results in a MoFe
protein with a 3-fold lower Km for C2H2 reduction compared to
the wild-type MoFe protein. This protein catalyzes 14% trans-
ethylene formation during C2D2 reduction compared with 4%
for the wild type. Further, no reduction of C2H2 to C2H6 could
Ala
be detected for this protein. The R-96 MoFe protein has a
Km nearly the same as that measured for wild type, and produces
1
2% cis-dideuterated ethylene. In the opposite direction, the
Ser
Asn
R-69 /R-195
MoFe protein has a 28-fold higher Km for
acetylene and also produces significant cis-ethylene (35%).
Thus, from these results, it must be concluded that the
stereospecificity of proton addition during reduction of C2D2
does not correlate with either the Km for acetylene reduction or
with the formation of C2H6.
These new findings require reformulation of the model
proposed earlier to explain the stereospecificity of proton
addition during acetylene reduction. It was previously proposed
that an enzyme-bound ethylenic intermediate (e.g., 3B in
Scheme 1) led to formation of C2H6 and loss of stereospecificity
Reduction of C2H2 and C2D2 in D2O. To assess the nature
of intermediate states that might occur during C2H2 reduction,
Asn
the product distribution for the wild-type and R-195
-
12
of proton addition to acetylene. A longer residence time of
substituted MoFe protein catalyzed reactions was examined with
C2H2 as the substrate in a reaction solution containing >90%
D2O, with ∼10% H2O. As can be seen in Figure 3 (traces 1
and 2), the principal products detected for both proteins were
cis- and trans-1,2-C2H2D2, with differing percentages of the two
isomers for the two proteins. A small quantity of C2H3D was
also detected, whereas no C2HD3 was observed. As a control
the proposed intermediate at the active site was proposed to be
responsible for the loss in stereospecificity and the increased
formation of C2H6. It was also proposed that an increase in the
population of the ethylenic intermediate was related to the Km
for acetylene. When the Km was very low, it was postulated,
acetylene would displace the intermediate, resulting in less
accumulation of intermediate and less C2H6 formation and
retention of the cis-proton addition stereochemistry. The results
from the present study indicate that the Km cannot be responsible
for the increased population of the intermediate and that the
stereochemistry of proton addition does not correlate with C2H6
formation. Further, in the case of equilibrium displacement, such
as has been indicated for displacement of bound C2H4 from the
to show that formation of C2HD3 could be experimentally
Asn
detected, the R-195
MoFe protein was used to catalyze
reduction of C2D2 in a D2O/H2O solution (trace 3).
Discussion
Fisher et al.12 reported that certain substitutions within the
MoFe protein located near the active site FeMo-cofactor can
significantly change the stereochemistry of proton addition
during C2D2 reduction. For example, it was shown that
25
active site by C2H2, the binding rate for C2H2 is not necessarily
expected to affect the off rate for C2H4. If free C2H2 does affect
the residence time of the proposed ethylenic intermediate, and

1826 J. Am. Chem. Soc., Vol. 123, No. 9, 2001
Benton et al.
Figure 5. Proposed mechanism for nitrogenase reduction of C
scheme shows a model for the reduction of C in H O with the
production of either cis- or trans-1,2-C . Each step intermediate
2 2
D . The
D
2 2
2
2 2 2
H D
is numbered and each reaction step is defined by a rate constant (k)
with a numerical subscript. The numbering interleaves with the
numbering shown in Figure 4.
the variable levels of C2H6 produced by different nitrogenases.
For the wild-type MoFe protein, it is expected that the reaction
step characterized by the rate constant k5 dominates over the
reaction with the rate constant k6, resulting only in C2H4
production. In contrast, certain alterations around FeMo-cofactor
Lys
(
e.g., the R-191 substitution) could result in a decrease in
the k5/k6 ratio resulting in a greater percentage of intermediate
going to C2H6. It has also been observed that C2H4 can be
2
Figure 4. Proposed mechanism for nitrogenase reduction of acetylene
to ethylene and ethane. The scheme depicts a model for the nitrogenase-
catalyzed reduction of acetylene to both ethylene and ethane. Each
intermediate is numbered and each reaction is defined by a rate constant
2
7
reduced to C2H6 by the V-based nitrogenase, indicating the
possibility for equilibrium between released and bound C2H4
such as described by the equilibrium constant K7 (k7/k-7) in
Figure 4. This feature can also account for the lag observed for
C2H6 formation catalyzed by the V-based nitrogenase relative
to C2H4 or H2 formation during C2H2 reduction. Thus, by
following the left-hand reaction pathway (as proposed for the
MoFe protein) no lag in C2H6 production is expected compared
to C2H4 production. In contrast, in the right-hand pathway (as
proposed for the V-based nitrogenase) a lag could be expected
for C2H6 production compared to C2H4 production since a
rebinding step is required for C2H6 formation.
(k) with a numerical subscript.
C2H6 formation and stereospecificity are correlated to that
residence time,12 then C2H6 formation and stereospecificity of
proton addition should be dependent on the concentration of
C2H2, and they are not. The percentage of cis- and trans-1,2-
C2H2D2 was found not to change when the concentration of
C2D2 was changed (data not shown).
We propose a different model based on the available
information and known model chemistry (Figures 4 and 5).
Dilworth et al. put forth two arguments against a possible
2
6
2 4
C H dissociation and association event to account for the lag
Based on the recent spectroscopic results from Lee et al., the
initial binding of C2H2 to FeMo-cofactor is proposed to occur
via an acetylenic bridge between two Fe atoms indicated as
intermediate 1 in Figure 4. Reduction of this bound intermediate
by one electron and one proton results in formation of
in C2H6 formation catalyzed by the V-dependent nitrogenase.28
First, the rate of C2H6 formation from added C2H4 is much lower
than the rate of C2H6 formation from C2H2. Second, the amount
of “free” C2H4 that could accumulate at equilibrium within the
1
2
2 2
reaction vessel via C H reduction in a typical assay would be
intermediate 2 as in the model proposed by Fisher et al. To
explain the reduction of C2H2 by two or four electrons to yield
either C2H4 or C2H6, two divergent reaction pathways from
intermediate 2 are now proposed. For one pathway (Figure 4,
left), intermediate 2 is further reduced by one electron and one
proton to an enzyme-bound ethylenic intermediate such as 4.
If this intermediate remains bound to the active site further
reduction by two more electrons and two protons could occur
to yield C2H6.
far below the Km for reduction of C2H4. Thus, our suggestion
would demand a local concentration effect related to the
proximity of released C2H4 to the active site. This possibility
does not seem unreasonable considering that the active site
FeMo-cofactor is buried within the polypeptide matrix and that
released C2H4 would already be in the correct place and perhaps
in the correct position for rebinding.
We also attempt to explain the observed stereochemistry of
proton addition upon reduction of C2D2 in H2O by the model
shown in Figure 5. This model is an expanded view of steps k2
and k5 presented in Figure 4. Based on recent spectroscopic
The other pathway (Figure 4, right) results in the reduction
of intermediate 2 to give C2H4, which is subsequently released.
The competition between these two pathways can account for
(
25) Lowe, D. J.; Eady, R. R.; Thorneley, R. N. F. Biochem. J. 1978,
73, 277-290.
26) Lee, H. I.; Sorlie, M.; Christiansen, J.; Song, R.; Dean, D. R.; Hales,
B. J.; Hoffman, B. M. J. Am. Chem. Soc. 2000, 122, 5582-5587.
(27) Dilworth, M. J.; Eady, R. R.; Robson, R. L. Nature 1987, 327, 167-
1
8.
(
(28) Dilworth, M. J.; Eady, R. R.; Eldridge, M. E. Biochem. J. 1988,
249, 745-751.

Stereospecific C2D2 Reduction
J. Am. Chem. Soc., Vol. 123, No. 9, 2001 1827
findings,26 intermediates 1A and 2A for bound C2D2 and the
one-electron reduced intermediate, respectively, seem reason-
able. Proceeding from intermediate 2A, we suggest two compet-
ing reaction pathways. Following the downward pathway in
Figure 5, intermediate 2A becomes reduced by one electron,
with one proton being added to the RC on the same face as the
earlier proton addition to the âC, resulting in cis-1,2-C2H2D2.
In the rightward pathway, intermediate 2A would convert to
an intermediate that permits rotation about the C-C bond,
consequently resulting in loss of the stereochemical addition
of the proton to the RC relative to the âC. An intermediate (3B
shown in Scheme 1) was previously proposed12 and we initially
favored this model. We tested the possibility for formation of
this intermediate by examining the products resulting from
reduction of C2H2 in D2O. Under the appropriate conditions
where loss of stereospecificty is observed, accumulation of
intermediate 3B is expected. Thus, removal of a proton or
deuteron from the âC of intermediate 3B should result in
formation of either intermediate 2C or 2D. Based on the
predicted higher energy of breaking a C-D bond when
compared to a C-H bond,29 intermediate 2D is expected to
predominate. The subsequent reduction of intermediate 2D by
one electron and one deuteron would be expected to yield
primarily C2HD3. However, this prediction is not experimentally
observed. For example, in experiments with either the wild-
type or the R-195Asn-substituted MoFe protein for catalysis, no
detectable level of C2HD3 is observed (Figure 3). The lack of
this product clearly suggests that an intermediate like 3B is
unlikely, demanding an alternative intermediate to explain the
loss of C2D2 reduction stereospecficity catalyzed by certain
altered MoFe proteins.
propose an η -vinyl intermediate as indicated by intermediate
2
3 in Figure 5. The occurrence of this intermediate is consistent
with the observed product distribution of C2D2 reduction
catalyzed by altered nitrogenases. It is known that the η f η
1
2
1
f η isomerization (2A f 3 f 2A or B) will result in either
trans- or cis-bound vinyl products (e.g., 2B and 2A),3
0-33
with
the isomer product being controlled by the nature of ligands to
the metal. In the case of nitrogenase, this isomerization could
result in formation of either cis- or trans-1,2-C2H2D2, with the
protein environment around the FeMo-cofactor dictating the
product distribution. Thus, it would be possible to explain the
stereochemistry of proton addition observed for all nitrogenase
proteins based on the protein influence controlling the transition
from intermediate 3 to 2A or 2B without invoking the downward
reaction at the far left. An intermediate such as 3 also explains
why no trideuterated ethylene is observed when C2H2 is reduced
by certain altered nitrogenases when incubated under catalytic
conditions in the presence of D2O.
The effect of electron flux on stereospecificity of C H2
reduction in vitro is consistent with the sequential one-electron
reduction of nitrogenase substrates. Thus, in the case of C D2
reduction, a lower flux should favor a longer residence time of
intermediate 2A (shown in Figure 5) at the active site, thus
giving more opportunity for the formation of intermediate 3. It
seems reasonable to expect that the formation of intermediate
3 is dependent not only on flux but also on steric constraints
within the substrate-binding site. This suggestion is based on
the observation that, for the wild-type protein, even under very
2
2
low flux conditions, only a relatively small proportion of C D2
is reduced to give the trans isomer. In contrast, lowering flux
2
Ser
Asn
for the doubly substituted R-69 /R-195 MoFe protein raises
the trans isomer product from 35% to nearly 50%. It seems
reasonable that the range of trans product produced is a function
of the additive affects of both a lower flux and the alteration of
steric constraints around the FeMo-cofactor environment im-
posed by the protein. However, given that the specific effect of
amino acid substitutions within the FeMo-cofactor environment
on flux is not known, and experimentally difficult to evaluate,
the interplay between these two effects remains an open
question.
Insight concerning the possible nature of the intermediate
comes from well-characterized metal-alkyne complexes de-
30
scribed in the inorganic literature. Among these complexes,
2
metal-bound η -vinyl complexes represent the most relevant to
the present discussion.30 On the basis of these complexes, we
(
29) Sugimoto, N.; Sasaki, M.; Osugi, J. J. Phys. Chem. 1983, 86, 3418-
3
1
423.
(30) Frohnapfel, D. S.; Templeton, J. L. Coord. Chem. ReV. 2000, 207,
99-235.
(31) Davidson, J. L.; Shiralian, M.; Manojlovic-Muir, L.; Muir, K. W.
J. Chem. Soc., Dalton Trans. 1984, 2167-2176.
32) Allen, S. R.; Beevor, R. G.; Green, M.; Norman, N. C.; Orpen, N.
A. G.; Williams, I. D. J. Chem. Soc., Dalton Trans. 1985, 435-450.
33) Davidson, J. L.; Wilson, W. F.; Manojlovic-Muir, L.; Muir, K. W.
(
Acknowledgment. The authors thank Wei Wu for assistance
with the Km and Vmax determinations, Valerie Cash for assistance
with the genetic constructions, and Professor John Hubbard for
helpful discussions regarding mechanisms. This work was
supported by National Institutes of Health Grant R01-GM59087.
(
J. Organomet. Chem. 1983, 254, C6-C10.
(
(
(
34) Kim, J.; Rees, D. C. Nature 1992, 360, 553-560.
35) Kraulis, P. J. J. Appl. Crystallogr. 1991, 24, 946-950.
36) Merritt, E. A.; Murphy, M. E. P. Acta Crystallogr. 1994, D50, 869-
8
73.
JA003662X