ATP-independent formation of hydrocarbons catalyzed by isolated nitrogenase cofactors

Article abstract of DOI:10.1002/anie.201108916

Reduce to produce: Molybdenum- and vanadium-nitrogenase cofactors have been isolated and shown to reduce carbon monoxide and cyanide ions to a mixture of alkanes and alkenes in the presence of a strong reductant, europium(II) diethylenetriaminepentaacetate (see scheme). Various hydrocarbons of up to seven carbon atoms in length are detected as products in these ATP-free reactions.

Full text of DOI:10.1002/anie.201108916

Angewandte
Chemie
DOI: 10.1002/anie.201108916
Biocatalysis
ATP-Independent Formation of Hydrocarbons Catalyzed by Isolated
Nitrogenase Cofactors**
Chi Chung Lee, Yilin Hu,* and Markus W. Ribbe*
Nitrogenases are highly complex and uniquely versatile
metalloenzymes that are capable of reducing a broad
spectrum of substrates, such as N2, CO, and CNÀ ions, under
ambient conditions.[1–4] The molybdenum- and vanadium-
nitrogenases are two homologous members of this enzyme
family, both of which utilize a specific reductase (Fe protein)
to donate electrons to the cofactor site (FeMoco or FeVco) of
a catalytic component (MoFe or VFe protein) during
catalysis. The buried location of the cofactor poses a challenge
for electron transfer in this process and renders it strictly
dependent on the ATP-assisted formation of an electron-
transport chain (within a complex between the reductase and
the catalytic component) that extends all the way from the
[Fe4S4] cluster of the reductase, through the P-cluster, to the
cofactor site of the catalytic component.[5] On the other hand,
both FeMoco and FeVco can be extracted as intact entities
into organic solvents,[6–8] which has spurred interest in seeking
an ATP-independent reaction system in which electrons can
be directly delivered to the isolated cofactors for reduction of
a substrate. In particular, the recent discovery that nitro-
genases can reduce CO to hydrocarbons[3,4] makes it attractive
to explore the capacity of cofactors to directly catalyze the
formation of hydrocarbons from CO, as well as CNÀ ions,
which are isoelectronic with CO.
Such a task can be accomplished by extracting the
cofactors with N-methylformamide (NMF)[6–8] and combining
them with a strong reductant, europium(II) diethylenetria-
minepentaacetate [EuII–DTPA],[9] in an ATP-free buffer
system. The isolated cofactors remain sufficiently stable in
this buffer system, keeping more than 90% integrity within
the first hour (Figure S1 in the Supporting Information) and
thereby permitting the determination of the activity over this
time period. Driven by [EuII–DTPA] (E0’ = À1.14 V at pH 8),
both FeMoco and FeVco reduce CO to CH4, C2H4, C2H6,
C3H6, C3H8, 1-C4H8, n-C4H10, and 1-C5H10, under ambient
conditions (Figure 1; see also Figure S2 in the Supporting
Information). When CNÀ ions are used as a substrate, the
same set of products are generated together with NH3 in both
FeMoco- and FeVco-based reactions. In the reaction cata-
lyzed by FeMoco, n-C5H12, 1-C6H12, n-C6H14, and n-C7H16 are
also detected as products (Figure 1; see also Figure S2 in the
Supporting Information). The product profiles for the reduc-
tion of CNÀ ions and CO are similar, which is consistent with a
[10]
À
previously proposed, common C C coupling pathway.
However, the rates of product formation in the reaction
with CNÀ ions are considerably higher than those in the
reaction with CO, which likely results from a stabilizing effect
of CNÀ ion-binding on the isolated cofactors.[11] GC-MS
analysis further confirms that CO and CNÀ ions are the
carbon sources for the hydrocarbons that are generated in
these reactions, as all of the products display the expected
mass shifts when 12CO and 12CNÀ ions are replaced by 13CO
and 13CNÀ, respectively, in the reaction (Figure 2).
There are interesting discrepancies in how the cofactors
react with the two carbonaceous substrates in the solvent-
extracted/[EuII–DTPA]-driven and protein-bound/ATP-
driven states. Both isolated cofactors are less active than
their protein-bound counterparts for the reduction of CO;
however, the total amounts of hydrocarbons formed by
isolated FeMoco and FeVco are 67.9% and 0.05%, respec-
tively, of the totals produced by the protein-bound FeMoco
and FeVco.[3,4,12] Such a disparate decrease in CO-reducing
efficiency renders FeMoco, which is only 0.1% as active as
FeVco within the protein, comparably active with FeVco in
the isolated state (Figure 1). With regard to CNÀ ions, the
protein-bound cofactors normally reduce this substrate to
CH4 and NH3.[1] This is not the case when CNÀ ions are
reduced by the isolated cofactors, as CH4 is no longer the
major carbonaceous product, and alkenes/alkanes of two to
seven carbon atoms in length are also detected as products in
this reaction (Figure 1).
The differences between the isolated and protein-bound
cofactors in hydrocarbon formation highlight the significant
impact of the protein environment on the reactivity of
nitrogenase cofactors.[13] Nevertheless, the ability of isolated
cofactors to catalyze the ATP-independent formation of long-
chain, liquid-phase hydrocarbons suggests the possibility of
developing electrocatalysts for the production of fuel under
ambient conditions. Understanding how nitrogenases catalyze
the formation of hydrocarbons is crucial for achieving this
goal. These enzymes not only provide a prototype for such an
electrocatalyst, but also serve as a biological blueprint for a
synthetic matrix that immobilizes the catalyst and mimics the
protein machinery for enhanced, ATP-independent electron
transfer.
[*] C. C. Lee, Dr. Y. Hu, Prof. Dr. M. W. Ribbe
Molecular Biology and Biochemistry
University of California, Irvine
2236/2448 McGaugh Hall, Irvine, CA 92697-3900 (USA)
E-mail: yilinh@uci.edu
mribbe@uci.edu
[**] We thank Prof. Dr. D. C. Rees and Dr. Nathan Dalleska of Caltech
(Pasadena) for help with the GC-MS analysis. This work was
supported by Herman Frasch Foundation grant 617-HF07 (M.W.R.)
and NIH grant GM-67626 (M.W.R.).
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.201108916.
Angew. Chem. Int. Ed. 2012, 51, 1947 –1949
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1947

.
Angewandte
Communications
Figure 1. GC-MS analysis of hydrocarbons generated from the reduction of 12CO and 12CNÀ ions by isolated cofactors. Specific activities of hydrocarbon
formation are stated in nmol per mmol of cofactor per hour above the corresponding traces and presented as (meanÆstandard deviation) (N=5).
ND=not determined by GC-based activity analysis. For time-dependent formation of these products, see Figure S2 in the Supporting Information. Other
than hydrocarbons, NH3 is formed at (18270Æ129) and (15128Æ107) nmol per mmol of cofactor per hour in the FeMoco- and FeVco-catalyzed reduction
of CNÀ ions, respectively. Relative to the amounts of nitrogen that appear in NH3, the total amounts of carbon that appear in hydrocarbons are 95% and
85% in the FeMoco- and FeVco-catalyzed reduction of CNÀ ions, respectively.
Figure 2. GC-MS analysis of hydrocarbons generated by isolated cofactors from the reduction of 13CO and 13CNÀ ions.
the purification of these nitrogenase proteins.[14] FeVco and FeMoco
Experimental Section
were extracted into NMF from VFe protein (1.5 g) and MoFe protein
(1.5 g), respectively, by using a previously described method.[6]
Reduction of CNÀ ions and CO by isolated cofactors: A stock
solution of [EuII–DTPA] was prepared by dissolving equal molar
amounts of europium(II) chloride and diethylenetriaminepentaacetic
acid at a final concentration of 200 mm in 1m Tris buffer (pH 8.0). The
reduction reaction with CNÀ ions contained Tris (25 mm, pH 7.8),
[EuII–DTPA] (5 mm),[9] isolated FeMoco or FeVco (0.4 mL), and
Unless otherwise specified, all chemicals were purchased from
Sigma–Aldrich (St. Louis, MO). Natural abundance 12CO (99.5%
purity) was purchased from Airgas (Lakewood, CA). All isotope-
labeled compounds (ꢀ 98% isotopic purity) were purchased from
Cambridge Isotopes (Andover, MA).
Protein purification and cofactor extraction: Azotobacter vine-
landii strains expressing His-tagged VFe and MoFe proteins were
grown as described elsewhere.[14] Published methods were used for
1948
www.angewandte.org
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2012, 51, 1947 –1949

Angewandte
Chemie
NaCN (100 mm) in a total volume of 25 mL. The reduction reaction
with CO was of the same composition, except that CO (100%) was
used instead of NaCN. Both reactions were run at room temperature
and pressure for varying lengths of time. For GC-MS experiments
with isotope labels, natural abundance Na12CN and 12CO were
electron impact (EI) ionization and selected ion monitoring (SIM)
mode.
Received: December 18, 2011
Revised: January 3, 2012
Published online: January 17, 2012
replaced by Na13CN and 13CO, respectively.
Activity analysis of the reduction of CNÀ ions and CO: The
Keywords: carbon monoxide · cofactors · enzymes ·
formation of products CH , C H , C H , C H , C H , 1-C H , n-C H ,
.
4
2
4
2
6
3
6
3
8
4
8
4
10
hydrocarbons · nitrogenases
1-C5H10, n-C5H12, 1-C6H12, n-C6H14, and n-C7H16 was determined by
GC on an activated alumina column (Grace, Deerfield, IL), which
was held at 408C for 2 min, heated to 2008C at 108CminÀ1, and held
at 2008C for another 2 min. The twelve hydrocarbon products were
quantified as described previously[3,4] and their detection thresholds
were, 1.1 (CH4), 1.3(C2H4), 1.3 (C2H6), 1.5 (C3H6), 1.3 (C3H8), 1.4
(1-C4H8), 1.4 (n-C4H10), 3.1 (1-C5H10), 3.7 (n-C5H12), 6.1 (1-C6H12), 5.8
(n-C6H14) and 7.0 (n-C7H16) nmol per mmol of cofactor. NH3 was
[1] B. K. Burgess, D. J. Lowe, Chem. Rev. 1996, 96, 2983.
[2] R. R. Eady, Chem. Rev. 1996, 96, 3013.
[3] C. C. Lee, Y. Hu, M. W. Ribbe, Science 2010, 329, 642.
[4] Y. Hu, C. C. Lee, M. W. Ribbe, Science 2011, 333, 753.
[5] H. Schindelin, C. Kisker, J. L. Schlessman, J. B. Howard, D. C.
Rees, Nature 1997, 387, 370.
[6] A. W. Fay, M. A. Blank, C. C. Lee, Y. Hu, K. O. Hodgson, B.
Hedman, M. W. Ribbe, J. Am. Chem. Soc. 2010, 132, 12612.
[7] A. W. Fay, C. C. Lee, J. A. Wiig, Y. Hu, M. W. Ribbe, Methods
Mol. Biol. 2011, 766, 239.
[8] V. K. Shah, W. J. Brill, Proc. Natl. Acad. Sci. USA 1977, 74, 3249.
[9] K. A. Vincent, G. J. Tilley, N. C. Quammie, I. Streeter, B. K.
Burgess, M. R. Cheesman, F. A. Armstrong, Chem. Commun.
2003, 2590.
[10] E. M. Carnahan, J. D. Protasiewicz, S. J. Lippard, Acc. Chem.
Res. 1993, 26, 90.
[11] C. J. Pickett, K. A. Vincent, S. K. Ibrahim, C. A. Gormal, B. E.
Smith, S. A. Fairhurst, S. P. Best, Chem. Eur. J. 2004, 10, 4770.
[12] These numbers were derived from comparisons between the
activities of protein-bound and solvent-isolated cofactors and
are given in nmol of total carbon in hydrocarbon products per
mmol of cofactor per hour.
[13] Although comparable in activity in the current study, the
isolated FeVco, similar to its protein-bound counterpart, may
be much more active than the isolated FeMoco for the reduction
of CO. The [EuII–DTPA] system used in this work may have a
limited capacity to deliver electrons to the two cofactors, which
renders them similarly inefficient in the isolated state.
[14] C. C. Lee, Y. Hu, M. W. Ribbe, Proc. Natl. Acad. Sci. USA 2009,
106, 9209.
determined by an HPLC fluorescence method.[15]
Determination of cofactor stability: The stability of isolated
FeMoco or FeVco was determined by its ability to reconstitute the
cofactor-deficient DnifB MoFe protein after incubation with the
buffer system that was used for the [EuII–DTPA]-driven reduction
reactions with CO and CNÀ ions. Specifically, the isolated FeMoco or
FeVco was incubated with the buffer system for varying lengths of
time before DnifB MoFe protein was added to the mixture and the
resulting reconstituted MoFe protein was assayed for activity.[7]
GC-MS: The hydrocarbon products were identified by GC-MS
on a Hewlett–Packard 5890 GC and 5972 MSD. The identities of CH4,
C2H4, C2H6, C3H6, C3H8, 1-C4H8, n-C4H10, 1-C5H10, n-C5H12, 1-C6H12,
n-C6H14, and n-C7H16 were confirmed by comparing the masses and
retention times with those of the Scott standard n-alkane and
1-alkene gas mixture (Plumsteadville, PA). A total of 50 mL of gas was
injected into a split/splitless injector operated at 1258C in splitless
mode. A liner (1 mm internal diameter (ID)) was used to optimize the
sensitivity. Gas separation was achieved on a Restek PLOT-QS
capillary column (Bellafonte, PA; 0.320 mm ID ꢀ 30 m length), which
was held at 408C for 1 min, heated to 2208C at 108CminÀ1, and held
at 2208C for another 3 min. The He carrier gas was passed through
the column at 1.0 mLminÀ1. The mass spectrometer was operated in
[15] J. L. Corbin, Appl. Environ. Microbiol. 1984, 47, 1027.
Angew. Chem. Int. Ed. 2012, 51, 1947 –1949
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org 1949