ATP- and iron-protein-independent activation of nitrogenase catalysis by light

Article abstract of DOI:10.1021/ja1071866

We report here the light-driven activation of the molybdenum-iron-protein (MoFeP) of nitrogenase for substrate reduction independent of ATP hydrolysis and the iron-protein (FeP), which have been believed to be essential for catalytic turnover. A MoFeP variant labeled on its surface with a Ru-photosensitizer is shown to photocatalytically reduce protons and acetylene, most likely at its active site, FeMoco. The uncoupling of nitrogenase catalysis from ATP hydrolysis should enable the study of redox dynamics within MoFeP and the population of discrete reaction intermediates for structural investigations.

Full text of DOI:10.1021/ja1071866

Published on Web 09/15/2010
ATP- and Iron-Protein-Independent Activation of Nitrogenase Catalysis by
Light
Lauren E. Roth, Joey C. Nguyen, and F. Akif Tezcan*
UniVersity of California, San Diego, Department of Chemistry and Biochemistry, La Jolla, California 92093-0356
Received August 10, 2010; E-mail: tezcan@ucsd.edu
Abstract: We report here the light-driven activation of the
molybdenum-iron-protein (MoFeP) of nitrogenase for substrate
reduction independent of ATP hydrolysis and the iron-protein
(
FeP), which have been believed to be essential for catalytic
turnover. A MoFeP variant labeled on its surface with a Ru-
photosensitizer is shown to photocatalytically reduce protons and
acetylene, most likely at its active site, FeMoco. The uncoupling
of nitrogenase catalysis from ATP hydrolysis should enable the
study of redox dynamics within MoFeP and the population of
discrete reaction intermediates for structural investigations.
The enzyme nitrogenase performs nitrogen fixation under ambient
conditions, in stark contrast to the Haber-Bosch process that
1
,2
requires extreme temperatures and pressures to the same end.
Biological nitrogen fixation comes at a high energetic cost, however,
in that it requires 16 ATP molecules per turnover reaction (2 ATP
molecules per electron):
Figure 1. (a) FeP-MoFeP complex structure obtained in the presence of
AMPPCP (an ATP analog) highlighting key components. (b) IA-RuBP.
(c) Cartoon representations of labeling sites on wt, R-L158C, and R-H196C
MoFeP and corresponding edge-to-edge distances to MoFeP clusters.
N + 8 e- + 8 H+ + 16 ATP f 2NH + H + 16ADP + 16P
2
3
2
i
Despite decades of investigations, it is still not understood in detail
how ATP hydrolysis is coupled to redox catalysis by nitrogenase
and what the mechanistic details of substrate activation by the
catalytic core of nitrogenase (the iron-molybdenum cofactor,
While many electron donors (e.g., dithionite; E° ) -400 to -650
6
IV/III
7
mV; Ti
[citrate], E° e -800 mV ) with comparable or lower
reduction potentials than FeP (-280 mV as isolated, -620 mV as
complexed to MoFeP) can reduce oxidized states of the P-cluster
(P and P ), they cannot support catalysis alone. This has led to
the proposal that MgATP binding to FeP and its hydrolysis within
the FeP-MoFeP complex must mechanically gate ET among the
3
,4
8
FeMoco) are. The second question is intimately linked to the
first: because of the requirement of continuous ATP hydrolysis to
maintain the electron flow for the catalytic activation of FeMoco,
it has been challenging to populate discrete reaction intermediates
bound to the cofactor in sufficient quantities for structural and
spectroscopic interrogation. Here we report for the first time the
light-driven, ATP-independent activation of nitrogenase toward the
catalytic two-electron reduction of alternative nitrogenase substrates,
2
+
1+
9
three redox centers, explaining the exclusivity of FeP as a catalytic
electron donor. Indeed, structures of FeP-MoFeP complexes
obtained at various stages of MgATP hydrolysis revealed large
conformational changes in FeP that enable it to assume different
docking geometries on MoFeP and modulate the distance (thus,
electronic coupling) between the [4Fe:4S] cluster and the P-
+
5
2 2
protons (H ) and acetylene (C H ). Our findings challenge the
long-standing assumption that ATP hydrolysis is obligatory for
nitrogenase catalysis and open up new avenues for studying the
mechanistic intricacies of biological nitrogen fixation.
1
0
cluster. Nevertheless, these structures also showed that there are
no discernible conformational changes within MoFeP that would
readily indicate an ET gate.
Mo-nitrogenase comprises the Fe-protein (FeP) and the
Mo-Fe-protein (MoFeP) (Figure 1a). FeP is an ATPase that
functions as an electron shuttle to MoFeP, where substrate binding
and activation take place at FeMoco (a [7Fe:1Mo:9S:1X] cluster).
FeP can provide only one or two electrons at a time; hence it has
to undergo multiple ATP-dependent association-dissociation cycles
with MoFeP before enough electrons accumulate on FeMoco for
We set out to probe whether it would be possible to mimic the
low-potential, highly coupled state of the FeP [4Fe:4S] cluster
formed within the MgATP-activated FeP-MoFeP complex and
bypass the need for ATP hydrolysis and FeP to initiate nitrogenase
reactivity. Toward this end, we covalently labeled three different
A. Vinelandii MoFeP constructs with a Cys-specific iodoacetamido
3
,4
2+
catalysis. Electron transfer (ET) during catalysis is believed to
proceed from a [4Fe:4S] cluster located in FeP to a buried [8Fe:
2
derivative of [Ru(bpy) (phen)] (IA-RuBP). These Ru-labeled
MoFeP variants (wild type, R-L158C, and R-H196C) are illustrated
in Figure 1c. Wild-type (wt) MoFeP contains several Cys residues.
However, only one (R-Cys45) is appreciably solvent exposed to
be a candidate for functionalization with IA-RuBP. The additional
Cys residues on each of the other two mutants, which also include
R-C45, are in relatively solvent exposed positions, with R-C196 in
7S] cluster (P-cluster) in MoFeP and finally to FeMoco. Association
of the two proteins, the ET step, and the dissociation reaction are
strictly controlled through the ATPase activity of FeP.
MgATP-bound FeP is the only known electron donor that has
been shown to effect the reduction of any substrate by MoFeP.
1
3672 9 J. AM. CHEM. SOC. 2010, 132, 13672–13674
10.1021/ja1071866  2010 American Chemical Society

C O M M U N I C A T I O N S
utilized dithionite, which should allow for diffusion-limited quench-
II
II
ing of the Ru BP excited state (*Ru BP) to generate the reducing
I
14
Ru BP (E° ≈ -1.28 V) species in high yield. Fulfilling a second
role, dithionite is a dioxygen scavenger universally employed to
protect MoFeP clusters from oxidative damage and to maintain the
N
P-cluster in an all-ferrous state (P ).
Ru-wt and Ru-C196 MoFeP show little to no production of
H or C H in the presence of 0.1 atm of C H even after 200 min
2 2 4 2 2
of irradiation (Figure S2). In contrast, irradiation of Ru-C158 leads
to the evolution of both products in equal quantities, with average
velocities of 16 nmol C
over 50 min (Figure 2b,c). Photodriven C
2
H
4
/min and 14 nmol H
2
/min per mg MoFeP
and H production
2
H
4
2
reaches a plateau after 50 min despite the presence of excess
dithionite, yielding a turnover number of ∼110 per active site for
both products. The EPR spectrum of Ru-C158 MoFeP shows no
changes in the characteristic S ) 3/2 feature of FeMoco after
turnover, indicating that the cofactor stays intact (Figure S3). On
the other hand, the absorption features of Ru-BP steadily disappear
during turnover (Figure S4), which may be attributable to Ru-ligand
dissocation and explain the loss of activity.
2 4 2
We next investigated whether C H and H production indeed
stems from our photoreduction scheme and not from an unforeseen
reactive site or species. The elimination of any of the reaction
components (light, Ru-photosensitizer, MoFeP, or dithionite) leads
2 4 2
to a complete abolishment of C H and H evolution (Figure 2b,c).
When free RuBP is included in solution instead of being covalently
linked to R-C158 MoFeP, no activity is observed, indicating that a
surface immobilized Ru-photosensitizer is necessary to ensure
efficient electronic coupling to the MoFeP clusters and electron
injection.
Interestingly, when dithionite is replaced with other typical
sacrificial donors like triethanolamine (TEOA, 200 mM), 2-(N-
morpholino)ethanesulfonic acid (MES, 200 mM), or NADH (40
mM), little photocatalytic activity is observed (Figure S7). In order
to test whether this unique ability of dithionite (λmax ) 314 nm)
for supporting catalysis is due to its photochemistry, we carried
out activity assays and control experiments using monochromatic
455-nm radiation from an LED source that should only excite the
Ru-photosensitizer and not dithionite (Figure S8). These experi-
ments show that (a) photocatalytic activity is maintained at this
wavelength, albeit with slower kinetics, and (b) Ru-C158 MoFeP
is considerably more stable under these conditions with steady
activity for at least 300 min of irradiation, likely due to the
elimination of UV-based damage. While these findings confirm
again that the light-driven reactivity stems from the surface-
immobilized Ru-photosensitizer, they also suggest a possible role
of dithionite beyond acting as a sacrificial donor. Current studies
are underway to probe the potential ability of dithionite to directly
reduce a P-cluster intermediate during photocatalysis along with
RuBP.
1
1
Figure 2. (a) A possible photoreduction scheme for Ru-C158 MoFeP.
b) Acetylene and (c) proton reduction assays and corresponding controls,
(
each of which has one indicated component of the photocatalytic system
missing.
the immediate vicinity of FeMoco and R-C158 in a cleft right above
the P-cluster. Incubation of the variants (2 h) with a 10-fold molar
excess of IA-RuBP leads to quantitative labeling of MoFeP as
determined by inductively coupled optical emission spectroscopy,
with one label per Rꢀ-dimer on wt-MoFeP and two labels per Rꢀ-
dimer on R-C158 and R-C196 variants as planned (Table S1 and
Figure S1).
The distances between the Ru-labels on each variant and the
MoFeP clusters are listed in Figure 1c. R-C45 is too distant from
either cluster to factor in the redox activation of nitrogenase, thus
serving as a control. Ru-labels on R-C158 and R-C196, on the other
hand, should be coupled to either the P-cluster or FeMoco,
respectively, allowing the examination of alternative electron
transfer routes to FeMoco. In order to investigate light-driven
activation of MoFeP (Figure 2a), we pursued the reduction of the
Having established the catalytic competence of the photosensi-
tized Ru-C158 MoFeP system, we sought to determine whether
FeMoco is the site of substrate activation. Carbon monoxide (CO)
has been shown to interact with FeMoco under turnover conditions
and to be a strong inhibitor of all nitrogenase substrates except
alternative substrates, H+ and C
electrons and whose products (H
H
2 2
, which require only two
and C ) are readily detected
by gas chromatography. A typical reaction included 2.7 mg of
+
15-17
2
2
H
4
H .
the substrate C
3a). On the other hand, H
2 2
of the same amount of CO or CO/C H
The inclusion of 0.05 atm of CO alongside 0.1 atm of
abolishes light-driven C production (Figure
is produced at high levels in the presence
, matching the inhibition
2
H
2
2 4
H
RuBP-labeled protein (∼20 nmol of active sites) in a solution of
2
2
00 mM dithionite and 200 mM NaCl at pH 7.75. In deoxygenated
vials, the reaction solutions were irradiated in a 20 °C water bath
with a Xe/Hg lamp using UV- (<300 nm) and IR-cutoff filters under
constant stirring, and the headgas was analyzed for products. As a
profile for ATP-driven nitrogenase catalysis. Similarly, cyanide
-
(CN ) stalls the photolytic C
2
H
2
reduction activity of Ru-C158
MoFeP (Figure 3b), with an apparent inhibition midpoint concen-
1
2,13
-
sacrificial electron donor (D) in a flash-quench scheme,
we
tration of ∼0.25 mM CN (at 0.05 atm C
2 2
H ) that agrees well with
J. AM. CHEM. SOC. 9 VOL. 132, NO. 39, 2010 13673

C O M M U N I C A T I O N S
4
,19
a previously reported value (K
under ATP-driven turnover conditions. These findings strongly
imply that FeMoco is the ultimate destination for the photogenerated
electrons and the site of catalysis.
i
) 0.5 mM) obtained for wt MoFeP
interactions on FeMoco,
a detailed picture of nitrogenase
1
8
catalysis is still not available. We have now uncoupled nitrogenase
catalysis from ATP hydrolysis and protein-protein interactions by
introducing a light-triggered electron delivery system, which should
enable the study of ET dynamics within MoFeP and the population
of discrete redox intermediates for structural investigations.
1
1
Acknowledgment. We are grateful to Prof. Dennis Dean (Va
Tech) for his generous donation of MoFeP mutant strains L158C
and H196C, Prof. Markus Ribbe (UC Irvine) and his group for the
use of their fermentor, Dr. Will Myers and Prof. Dave Britt (UC
Davis) for help with EPR experiments, and Prof. Harry Gray and
Dr. Jay Winkler (Caltech) for helpful discussions. This work was
supported by the NSF (predoctoral fellowship to L.E.R., MCB-
0
643777 to F.A.T.) and ACS (PRF-G 46939-G3).
Supporting Information Available: Protein preparation, charac-
terization, additional assays, and materials/methods. This material is
available free of charge via the Internet at http://pubs.acs.org.
References
(
(
(
1) Burgess, B. K.; Lowe, D. J. Chem. ReV. 1996, 96, 2983–3011.
2) Howard, J. B.; Rees, D. C. Chem. ReV. 1996, 96, 2965–2982.
3) Howard, J. B.; Rees, D. C. Proc. Natl. Acad. Sci. U.S.A. 2006, 103, 17088–
1
7093.
(
4) Hoffman, B. M.; Dean, D. R.; Seefeldt, L. C. Acc. Chem. Res. 2009, 42,
6
09–619.
(5) While light-driven nitrogenase catalysis was previously reported (Druzhinin,
S. Y.; Syrtsova, L. A.; Uzenskaja, A. M.; Likhtenstein, G. I. Biochem. J.
1993, 290, 627–631), this was accomplished under ATP turnover conditions.
(
(
6) Mayhew, S. G. Eur. J. Biochem. 1978, 85, 535–547.
7) Guo, M. L.; Sulc, F.; Ribbe, M. W.; Farmer, P. J.; Burgess, B. K. J. Am.
Chem. Soc. 2002, 124, 12100–12101.
Figure 3. (a) Carbon monoxide and (b) cyanide inhibition of light-driven
substrate reduction by Ru-C158 MoFeP. Experiments were conducted
under the same conditions as those shown in Figure 1 and described in the
(
8) Lanzilotta, W. N.; Seefeldt, L. C. Biochemistry 1997, 36, 12976–12983.
9) Danyal, K.; Mayweather, D.; Dean, D. R.; Seefeldt, L. C.; Hoffman, B. M.
J. Am. Chem. Soc. 2010, 132, 6894–6895.
(
-
text, except that CN inhibition experiments were performed with ∼0.1
(10) Tezcan, F. A.; Kaiser, J. T.; Mustafi, D.; Walton, M. Y.; Howard, J. B.;
Rees, D. C. Science 2005, 309, 1377–1380.
mg Ru-C158 MoFeP.
(
11) Although our hypothetical photocatalytic scheme involves stepwise reduction
of MoFeP clusters and invokes a higher electronation/protonation state of the
Presently, the quantum yield (φ ) catalytically useful electrons/
photons absorbed by Ru BP) of our light-driven system is <1%,
which likely is the reason why Ru-C158 fails to produce significant
P-cluster beyond its all-ferrous P state, it is introduced as a possible model
N
II
based on available data, which do not exclude other ET mechanisms.
Likewise, although the CO and CN- inhibition experiments are highly
suggestive of substrate reduction at FeMoco, they do not definitively exclude
reduction at the P-cluster in our system.
2
levels of ammonia from N . We are currently pursuing the
(
12) Chang, I. J.; Gray, H. B.; Winkler, J. R. J. Am. Chem. Soc. 1991, 113,
optimization of our system for increased yields of electron injection
to potentially enable this 6-e- process. Ultimately, however, our
experiments show that ATP hydrolysis and FeP are not absolutely
essential for substrate reduction by nitrogenase as commonly
assumed.
The study of any major aspect of nitrogenase catalysis has been
hampered due to its dependence on continuous ATP hydrolysis and
transient FeP-MoFeP interactions, which lead to a heterogeneous
mixture of redox and nucleotide-bound states of nitrogenase in
solution. While recent studies have provided glimpses into substrate
7
056–7057.
(13) Gray, H. B.; Winkler, J. R. Chem. Phys. Lett. 2009, 483, 1–9.
(
(
14) Gray, H. B.; Maverick, A. W. Science 1981, 214, 1201–1205.
15) George, S. J.; Ashby, G. A.; Wharton, C. W.; Thorneley, R. N. F. J. Am.
Chem. Soc. 1997, 119, 6450–6451.
(
16) Lee, H. I.; Hales, B. J.; Hoffman, B. M. J. Am. Chem. Soc. 1997, 119,
1
1395–11400.
(17) Maskos, Z.; Fisher, K.; Sorlie, M.; Newton, W. E.; Hales, B. J. J. Biol.
Inorg. Chem. 2005, 10, 394–406.
(
18) Rivera-Ortiz, J. M.; Burris, R. H. J. Bacteriol. 1975, 123, 537–545.
(19) Seefeldt, L. C.; Hoffman, B. M.; Dean, D. R. Annu. ReV. Biochem. 2009,
8, 701–722.
7
JA1071866
1
3674 J. AM. CHEM. SOC. 9 VOL. 132, NO. 39, 2010