Biosynthesis of the iron-guanylylpyridinol cofactor of [Fe]-hydrogenase in methanogenic archaea as elucidated by stable-isotope labeling

Article abstract of DOI:10.1021/ja211594m

[Fe]-hydrogenase catalyzes the reversible hydride transfer from H ₂ to methenyltetrahydromethanoptherin, which is an intermediate in methane formation from H₂ and CO₂ in methanogenic archaea. The enzyme harbors a unique active site iron-guanylylpyridinol (FeGP) cofactor, in which a low-spin Fe^II is coordinated by a pyridinol-N, an acyl group, two carbon monoxide, and the sulfur of the enzymes cysteine. Here, we studied the biosynthesis of the FeGP cofactor by following the incorporation of ¹³C and ²H from labeled precursors into the cofactor in growing methanogenic archaea and by subsequent NMR, matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF-MS), electrospray ionization Fourier transform ion cyclotron resonance mass spectrometry (ESI-FT-ICR-MS) and IR analysis of the isolated cofactor and reference compounds. The pyridinol moiety of the cofactor was found to be synthesized from three C-1 of acetate, two C-2 of acetate, two C-1 of pyruvate, one carbon from the methyl group of l-methionine, and one carbon directly from CO₂. The metabolic origin of the two CO-ligands was CO₂ rather than C-1 or C-2 of acetate or pyruvate excluding that the two CO are derived from dehydroglycine as has previously been shown for the CO-ligands in [FeFe]-hydrogenases. A formation of CO from CO₂ via direct reduction catalyzed by a nickel-dependent CO dehydrogenase or from formate could also be excluded. When the cells were grown in the presence of ¹³CO, the two CO-ligands and the acyl group became ¹³C-labeled, indicating either that free CO is an intermediate in their synthesis or that free CO can exchange with these iron-bound ligands. Based on these findings, we propose pathways for how the FeGP cofactor might be synthesized.

Full text of DOI:10.1021/ja211594m

Article
pubs.acs.org/JACS
Biosynthesis of the Iron-Guanylylpyridinol Cofactor of
[Fe]-Hydrogenase in Methanogenic Archaea as Elucidated
by Stable-Isotope Labeling
Michael Schick,† Xiulan Xie,‡ Kenichi Ataka,§ Jorg Kahnt,† Uwe Linne,‡ and Seigo Shima*,†,⊥
̈
†Max Planck Institute for Terrestrial Microbiology, Karl-von-Frisch Strasse 10, 35043 Marburg, Germany
‡Department of Chemistry, Philipps-Universitat Marburg, Hans-Meerwein Strasse, 35032 Marburg, Germany
̈
§Department of Physics, Freie-Universitat Berlin, Arnimallee 14, 14195 Berlin, Germany
̈
⊥PRESTO, Japan Science and Technology Agency (JST), Honcho, Kawaguchi, Saitama 332-0012, Japan
S
* Supporting Information
ABSTRACT: [Fe]-hydrogenase catalyzes the reversible hydride
transfer from H2 to methenyltetrahydromethanoptherin, which is an
intermediate in methane formation from H2 and CO2 in
methanogenic archaea. The enzyme harbors a unique active site
iron-guanylylpyridinol (FeGP) cofactor, in which a low-spin FeII is
coordinated by a pyridinol-N, an acyl group, two carbon monoxide,
and the sulfur of the enzyme’s cysteine. Here, we studied the
biosynthesis of the FeGP cofactor by following the incorporation of
2
13C and H from labeled precursors into the cofactor in growing
methanogenic archaea and by subsequent NMR, matrix-assisted
laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF-MS), electrospray ionization Fourier transform ion
cyclotron resonance mass spectrometry (ESI-FT-ICR-MS) and IR analysis of the isolated cofactor and reference compounds.
The pyridinol moiety of the cofactor was found to be synthesized from three C-1 of acetate, two C-2 of acetate, two C-1 of
pyruvate, one carbon from the methyl group of L-methionine, and one carbon directly from CO2. The metabolic origin of the two
CO-ligands was CO2 rather than C-1 or C-2 of acetate or pyruvate excluding that the two CO are derived from dehydroglycine as
has previously been shown for the CO-ligands in [FeFe]-hydrogenases. A formation of CO from CO2 via direct reduction
catalyzed by a nickel-dependent CO dehydrogenase or from formate could also be excluded. When the cells were grown in the
presence of 13CO, the two CO-ligands and the acyl group became 13C-labeled, indicating either that free CO is an intermediate in
their synthesis or that free CO can exchange with these iron-bound ligands. Based on these findings, we propose pathways for
how the FeGP cofactor might be synthesized.
that in the active site of [FeFe]-hydrogenases and [NiFe]-
hydrogenases. The iron differs, however, from that in [FeFe]-
hydrogenases and [NiFe]-hydrogenases in being a mononuclear
rather than a dinuclear metal center and in being complexed
bidentately by a pyridinol involving an acyl carbon−iron bond
that is so far unique in biochemistry.
The FeGP cofactor can be reversibly extracted from [Fe]-
hydrogenase under mildly alkaline conditions in the presence of
mercaptoethanol or under mildly acid conditions in the presence
of acetic acid.10 The extracted cofactor is, however, very labile and
decomposes rapidly even at room temperature and when irra-
diated by UV-A/blue light yielding mainly 2CO, FeII/III, and the
iron-free FeGP cofactor (Figure 1 bottom). The same decomposi-
tion products are found upon matrix-assisted laser desorption/
ionization time-of-flight mass spectrometry (MALDI-TOF-MS)
analysis of the FeGP cofactor. The mass of the entire FeGP cofactor
is found only using electrospray ionization mass spectrometry.10
INTRODUCTION
■
Methane formation from H2 and CO2 is the main energy
yielding process of many methanogenic archaea.1 The pathway
involves many novel coenzymes and unique metal-containing
prosthetic groups such as methanofuran, tetrahydromethanop-
terin (H4MPT), coenzyme F420, coenzyme M, coenzyme B, and
coenzyme F430.2 Recently, to this list the iron-guanylylpyridinol
(FeGP) cofactor was added, which has a unique structure
(Figure 1 top) and a yet unresolved pathway of biosynthesis.
The FeGP cofactor is the prosthetic group of [Fe]-
hydrogenase which catalyzes the reduction of methenyltetrahy-
dromethanopterin (methenyl-H4MPT+) with H2 to form
methylenetetrahydromethanopterin (methylene-H4MPT) and
H+. This reaction is involved in methane formation from H2
and CO2 in many methanogenic archaea.1,3 In the FeGP
cofactor, a low-spin FeII is coordinated bidentately by an acyl-
carbon and an sp2-hybridized nitrogen of the pyridinol moiety
as well as by two cis-CO ligands and a cysteine sulfur atom from
protein (Figure 1 top).4−9 With respect to the CO-ligands,
the iron in the FeGP cofactor of [Fe]-hydrogenase resembles
Received: December 12, 2011
Published: January 18, 2012
© 2012 American Chemical Society
3271
dx.doi.org/10.1021/ja211594m | J. Am. Chem. Soc. 2012, 134, 3271−3280

Journal of the American Chemical Society
Article
moiety of FeGP cofactor binds to the nucleotide monophos-
phate binding site on the N-terminal Rossmann-fold domain of
this enzyme.12 The electron-donating methyl groups of the
pyridinol ring may tune the electronic properties of this ring
system. Furthermore, density functional theory (DFT) cal-
culations indicated that the hydroxyl group of the pyridinol
moiety could be involved in the catalytic reactions.13 In addi-
tion, the strong trans-influencing character of the acyl ligand14
is considered to play an important role for hydrogen activation
because the iron site trans to the acyl ligand is proposed to be a
H2-binding site.5
In this study, we carried out in vivo labeling experiments with
13C- and 2H-labeled precursors to obtain first insights into how
the FeGP cofactor might be synthesized. For these experiments,
Methanothermobacter marburgensis and Methanobrevibacter
smithii were selected. Mt. marburgensis grows autotrophically
on H2/CO2 up to very high cell densities and overproduces
[Fe]-hydrogenase under nickel-limiting conditions up to 7% of
the soluble proteins,15 allowing us to obtain micromole
amounts of labeled cofactor. Despite the autotrophic growth,
the organism does assimilate acetate, pyruvate, and methionine
when added to the growth medium,16 albeit only low amounts,
resulting in partially isotope-labeled cell components. The
partial labeling and the amounts of cofactor obtained were,
however, sufficient for the determination of the labeling pattern
via NMR. Mb. smithii, the second organism used, is not an
autotroph;17 it requires exogenous acetate as a carbon source in
addition to CO2 because it lacks the nickel-dependent
CO dehydrogenase and therefore cannot synthesize acetyl-
coenzyme A (acetyl-S-CoA) from 2CO2. Consequently, in
Mb. smithii, the cell components derived from acetate are in
principle fully enriched with the label of 13C-labeled acetate
added to the medium. However, Mb. smithii grows poorly and
contains relatively low amounts of [Fe]-hydrogenase. There-
fore, the FeGP cofactor labeled in Mb. smithii could only be
analyzed by mass spectrometry rather than by NMR.
RESULTS
■
Mt. marburgensis and Mb. smithii were grown on H2 and CO2 in
media supplemented with 13C- or 2H-labeled putative
precursors in FeGP cofactor biosynthesis. The position of
incorporated 13C and H stable isotopes in the iron-free
2
FeGP cofactor was analyzed by NMR and the number of 13
C
isotopes incorporated was analyzed by mass spectrometry.
NMR analysis was only possible with the iron-free FeGP co-
factor because of line broadening caused by high-spin FeII/III
released upon decomposition of the FeGP cofactor. During
MALDI-TOF-MS analysis a decarboxylated derivative of the
iron-free cofactor was generated whose labeling was also
analyzed.
In methanogens, pyruvate is synthesized from acetyl-S-CoA
via reductive carboxylation. C-2 and C-3 of pyruvate are
therefore derived from C-1 and C-2 of acetate, respectively, and
C-1 of pyruvate is derived from CO2. Therefore, labeling
experiments with 13C from [1-13C]acetate, [2-13C]acetate,
[1-13C]pyruvate, and 13CO2 complement each other.
Acetate, pyruvate, methionine, glycine, formate, CO2, and
CO were found to be incorporated into the FeGP cofactor by
growing Mt. marburgensis or Mb. smithii. The degree of labeling
in Mt. marburgensis was 20% in the case of acetate (5 mM),
<10% in the case of pyruvate (7 mM), 80% in the case of
methionine (5 mM), and 90% in the case of formate (40 mM).
Figure 1. Structures of the FeGP cofactor and its light decomposition
product (iron-free FeGP cofactor). The FeGP cofactor binds to [Fe]-
hydrogenase with the sulfur of a conserved cysteine. The FeGP
cofactor can be reversibly extracted from the protein by denaturation
in the presence of acetic acid or 2-mercaptoethanol. The proposed
structures of the extracted cofactors are shown. The FeGP cofactor
form detected by ESI-MS is generated from the extracted forms
during electrospray ionization.10 The individual carbons of the iron-
free FeGP cofactor were labeled with stable isotopes as shown in
the figure.
Pyridinol/pyridone derivatives in the form of pyridone
polyketides are ubiquitous among alkaloids produced in plants,
fungi, and actinomycetes;11 however, such a highly substituted
pyridinol ring structure as found in the FeGP cofactor is rare.
All of the substituents of the pyridinol ring of the FeGP co-
factor should play crucial roles in its function. Crystal struc-
ture analyses have indicated that the guanosine monophosphate
3272
dx.doi.org/10.1021/ja211594m | J. Am. Chem. Soc. 2012, 134, 3271−3280

Journal of the American Chemical Society
Article
Labeling with glycine (19%) was only possible in Mb. smithii
because Mt. marburgensis is not capable of taking up glycine.
The degree of labeling of the cofactor in Mb. smithii was around
90% in the case of CO2 and 80% in the case of CO. The
enrichment was estimated by NMR or MALDI-TOF-MS of the
iron-free FeGP cofactor.
MS/MS revealed that the pyridinol moiety contained three 13C,
which were almost 100% derived from [1-13C]acetate (Table 2).
Via MS/MS, it was also observed that two 13C were incorporat-
ed into the ribose moiety and one 13C into the purine ring
(summarized in Table 2).
Carbon Atoms Derived from [2-13C]Acetate. 13C NMR
analysis of the iron-free FeGP cofactor labeled with 13C from
[2-13C]acetate in Mt. marburgensis revealed three intense peaks
at 13.2, 38.1, and 67.9 ppm. The three peaks were assigned as
C(i), C(g), and C(5′) from the 13C-chemical shifts (Table 1
and Figure S1, Supporting Information) and the 1H−13C
HSQC spectrum (Figure 2B). The 38.1 ppm C(g) peak was
relatively small and was line broadened. Such line broadening of
the C(g) peak has also been observed in our earlier study.18
The reason appears to be the partial exchange of the hydrogen
Carbon Atoms Derived from [1-13C]Acetate. The 13C
NMR spectrum of iron-free FeGP cofactor labeled with 13C
from [1-13C]acetate in Mt. marburgensis showed six 13C-
enriched peaks, which corresponded to C(b), C(f), and C(e)
of the pyridinol, C(1′) and C(4′) of the ribose, and C(5) of
guanine, based on the chemical shift data reported previously
(Table 1; Figure S1, Supporting Information).18 1H−13C
Table 1. 13C-Chemical Shifts of the Iron-Free FeGP Cofactor
from Mt. marburgensis Cultivated with 13C-Labeled
Substrates in D2O at 300 K
2
atoms at this carbon with H2O. Based on our published 13C
NMR shift data, the 13.2 ppm peak had been assigned as C(j),
which is one of the two methyl groups of the pyridinol ring.18
a
chemical shift of the observed 13C NMR peak
1
Because the earlier assignment did not fit to the H−13C
c
[1-13C]acetate
167.3
[2-13C]acetate
[1-13C]pyruvate
HMBC and HSQC data in this study, the 13C chemical shifts of
position
1
C(j) and C(i) were reassigned as follows. H−13C HSQC
pyridinol
analysis revealed that the carbon with δ = 13.2 ppm connects to
proton at 1.78 ppm, which was assigned as H(i) by the
published data, and it was confirmed as described below. The
1H signal of H(i) shows enhanced 13C-satellites due to partial
b
c
d
e
f
120
163
116.2
140.2
1
13C enrichment on this position (Figure 2B, H projection).
1
g
38.1
13.2
Integral of the H signals revealed an enrichment rate of 18%.
b
1
The assignment of H(i) was confirmed by H−13C HMBC,
i
Ribose
which indicated connectivity of H(i) with C(d), C(e), and C(f)
but not with C(b) and C(c) (Figure S2, Supporting
Information). Consequently, the carbon with δ = 13.2 ppm
was reassigned as C(i). Accordingly, the carbon with δ =
12.4 ppm was reassigned as C(j). It is important to note that
although the 13C-chemical shifts of the two methyl carbons
were reassigned, the chemical structure of the iron-free FeGP
cofactor was not changed.
1′
2′
3′
4′
5′
90.2
86.4
76
73
68.3
Guanine
4
5
154
118.1
The NMR results were confirmed by MALDI-TOF-MS
performed with the FeGP cofactor labeled in Mb. smithii. The
analysis revealed that three 13C from [2-13C]acetate were
incorporated into the iron-free FeGP cofactor and its
decarboxylated derivative (Figure 3C, left) with an enrichment
of about 90% (Figure S4, Supporting Information). MS/MS
analysis indicated that two 13C were incorporated into the
pyridinol moiety and one 13C was incorporated into the ribose
moiety (Table 2).
a
Chemical shifts of the nonlabeled carbons are not shown (See Figure
b
S1, Supporting Information). Revised assignment, see text for details.
c
Chemical shifts without calibration with 4,4-dimethyl-4-silapentane
sodium sulfonate.
heteronuclear multiple-bond correlation (HMBC) indicated
strong correlations between H(j) and C(b), H(i) and C(e), and
H(i) and C(f) (Figure 2A), while 1H−13C heteronuclear single-
quantum correlation (HSQC) showed strong peaks between
H(1′) and C(1′), and H(4′) and C(4′), which corroborated the
assignments of the labeled carbons (spectrum not shown).
The NMR results were confirmed by MALDI-TOF-MS
(positive mode) performed with the FeGP cofactor labeled in
Mb. smithii. The analysis revealed the incorporation of six 13C
from [1-13C]acetate into the iron-free FeGP cofactor and six
13C into decarboxylated cofactor generated during laser ioniza-
tion (Figure 3).10,18 Their corresponding mass peaks were ob-
served at m/z 549 and at m/z 505, respectively (Figure 3B left).
The presence of the multiple isotopic peaks indicated that the
13C enrichment in the FeGP cofactor was about 80% (Figure
S4, Supporting Information). We attributed this to the salvage
biosynthesis of GMP in Mb. smithii, which uses exogenous
guanine from yeast extract added to the medium (2 g/L), as
previously described by Choquet et al.19 Consistently, the
enrichment increased to almost 90% when Mb. smithii was
grown on media supplemented with only 0.05 g/L yeast extract.
Carbon Atoms Derived from [1-13C]Pyruvate. Even
though the intensity of the 13C-enriched NMR peaks of the
cofactor labeled with 13C from [1-13C]pyruvate in Mt.
marburgensis was lower than the intensity of the 13C-enriched
NMR peaks of the cofactor labeled with 13C from [1-13C]-
acetate and [2-13C]acetate, the 13C NMR spectrum clearly
revealed five 13C-enriched peaks at 73, 76, 120, 154, and 163
ppm. These peaks were assigned as C(3′), C(2′), C(c), C(4),
and C(d) based on the chemical shifts of the previous report
(Figure S1C, Supporting Information).18 1H−13C HMBC
indicated the correlation between H(j) and C(c), H(j) and
C(d), H(1′) and C(2′), and H(1′) and C(4). Because of the
lower enrichment in 13C and the lower concentration of the
sample, the background peaks were relatively high. However,
most of the extra peaks can easily be excluded from the data set
because the chemical shifts did not fit to the original data at
all.18 Only the peak observed at 174 ppm was similar to that of
1
C(h) (δ = 177.6 ppm), but H−13C HMBC did not show the
3273
dx.doi.org/10.1021/ja211594m | J. Am. Chem. Soc. 2012, 134, 3271−3280

Journal of the American Chemical Society
Article
2
Figure 2. Two-dimensional NMR analyses of the iron-free FeGP cofactor labeled with 13C and H from labeled compounds. The samples were
dissolved in 2H2O and measured at 300 K. (A) 1H−13C HMBC spectrum of the cofactor labeled with 13C from [1-13C]acetate. (B) 1H−13C HSQC
spectrum of the cofactor labeled with 13C from [2-13C]acetate. Signals arising from contaminants are marked with an asterisk. (C) 1H−13C HMBC
spectrum of the cofactor labeled with 13C from [1-13C]pyruvate. The 1H NMR signals were broadened, most probably by contaminating iron ions.
1
2
(D) H−13C HMBC spectrum of the nonlabeled cofactor (left) and the cofactor labeled with H from L-[methyl-2H]methionine (right).
1
cross peak with H(g) (Figure 2C). Therefore, we concluded
that C(h) was not enriched with [1-13C]pyruvate. Thus,
1H−13C HMBC unequivocally assigned the positions of the
labeled carbons.
at 1.82 ppm (Figure S3, Supporting Information). H−13C
HMBC indicated that the 13C/1H cross peaks at H(j) were not
visible, but the H(i) peak at 1.78 ppm reasonably revealed cross
peaks with C(d), C(e), and C(f), which showed that the methyl
group attached to C(c) was completely deuterated (Figure 2D).
Carbon Atoms from 13CO2. In the MALDI-TOF-MS
spectrum of FeGP cofactor labeled with 13C from 13CO2 in
Mb. smithii, a peak at m/z 553 (+10) (90% enrichment) was
detected (Figure 4A left), which indicated that 10 13C from
13CO2 were incorporated into the iron-free FeGP cofactor
(Figure S4, Supporting Information). In case of the
decarboxylated form, it was nine carbon atoms (m/z 508)
from CO2 indicating that the C(h) carboxyl group is derived
from 13CO2.
The NMR results were confirmed by MALDI-TOF-MS
performed with the FeGP cofactor labeled in Mt. marburgensis.
Mass peaks at m/z 543−548 (+0 to +5) and m/z 499−504 (+0
to +5), respectively, were observed (Figure 3D, left). The
analysis revealed that five 13C from [1-13C]pyruvate were
incorporated into the iron-free FeGP cofactor and into its
decarboxylation derivative with an enrichment of about 50%
(Figure S4, Supporting Information). MS/MS analysis
indicated that two 13C from [1-13C]pyruvate were incorporated
into the pyridinol moiety, two into the ribose moiety, and one
into the purine moiety (Table 2).
Carbon Atoms from[1-13C]Glycine and [13C]Formate.
MALDI-TOF-MS and MS/MS data indicated that [1-13C]-
glycine was incorporated by Mb. smithii exclusively into the
purine part of the iron-free FeGP cofactor, which is in
agreement with common purine biosynthesis (Table 2).19 In
the [13C]formate-labeling experiment with Mt. marburgensis,
one purine carbon was labeled with 13C from [13C]formate.
This result is consistent with the known incorporation of
formate into the C(2) of purine.21 [13C]Formate was not
incorporated into the pyridinol moiety.
Hydrogen Atoms Derived from L-[Methyl-2H3]-
methionine. In a preliminary study, we had found that the
iron-free FeGP cofactor isolated from Mt. marburgensis grown
in the presence of L-[methyl-13C]methionine showed an
increase of m/z = 1, which indicates that one of the methyl
groups on the pyridinol ring originates from L-methionine.20
1
The H NMR spectrum of the methyl-2H-labeled iron-free
FeGP cofactor now revealed the methyl group to be attached to
C(c) as indicated by a strong decrease in the H(j) peak height
3274
dx.doi.org/10.1021/ja211594m | J. Am. Chem. Soc. 2012, 134, 3271−3280

Journal of the American Chemical Society
Article
Figure 3. MALDI-TOF-MS and ESI-FT-ICR-MS mass spectra of (A) nonlabeled and (B−D) 13C-labeled FeGP cofactor from Mb. smithii. The 13
C
substrates used for labeling are indicated. The increase of m/z owing to incorporation of 13C is shown in red. In MALDI-TOF-MS spectra, the iron-
free cofactor and a decarboxylated derivative were detected. (A, right) The typical 54Fe isotopic peak could not be observed due to the low FeGP
cofactor concentration of the sample.10
Table 2. Incorporation of Stable Isotopes from
[1-13C]Acetate, [2-13C]Acetate, 13CO2, L-
[Methyl-2H]methionine, and [1-13C]Glycine into the
FeGP Cofactor of Mb. smithii and from [1-13C]Pyruvate
and [1-13C]Formate into the FeGP Cofactor of
a
Mt. marburgensis
number of stable isotopes in the FeGP cofactor parts
13C-labeled
substrate
CO
acyl
e
b
c
d
total pyridinol ribose guanine ligands
ligand
[1-13C]acetate
[2-13C]acetate
13CO2
6
3
3
2
3
2
1
2
1
12
3
4
2
2
1
1
13CO
[1-13C]pyruvate
5
2
1
2
1
1
[1-13C]formate
1
L-[methyl-2H]
methionine
1
f
[1-13C]glcyine
1
1
f
f
[1-13C]
propionate
[2,3-13C2]
succinate
f
a
Stable isotopes were detected by MALDI-TOF-MS and MS/MS.
b
c
Total carbons of the FeGP cofactor. Carbons of pyridinol other than
d
C(h) that is indicated as acyl ligand. Determined by ESI-FT-ICR-MS.
e
The number of incorporation of 13C into the carboxyl group C(h).
f
Not tested.
Figure 4. MALDI-TOF-MS and ESI-FT-ICR-MS mass spectra of the
FeGP cofactor labeled with 13C from (A) 13CO2 and (B) 13CO. The
increase in m/z owing to the incorporation of 13C is shown in red.
Lack of Labeling of the Iron-Free FeGP Cofactor by
[1-13C]Propionate and [2,3-13C]Succinate. [1-13C]-
Propionate and [2,3-13C]succinate were not incorporated into
the iron-free FeGP cofactor by Mt. marburgensis, although these
acids were assimilated into other cell components,16 for example,
propionate into isoleucine and succinate into F430 as revealed by
mass spectrometry (Figures S5 and S6, Supporting Information).
3275
dx.doi.org/10.1021/ja211594m | J. Am. Chem. Soc. 2012, 134, 3271−3280

Journal of the American Chemical Society
Article
Labeling of the CO- and Acyl Ligands in the FeGP
Cofactor. As indicated above, the complete FeGP cofactor
could not be analyzed by NMR because of line broadening due
to high-spin iron contaminations nor by MALDI-TOF because
of decomposition of the FeGP cofactor by irradiation with the
laser. Therefore, the following analyses were performed by ESI-
FT-ICR-MS and ATR-IR, which allowed detection of the
complete FeGP cofactor.
only one of the two cis CO ligands exchanged by 13CO. The
authentic CO frequency at 2009 cm−1 was not visible in the
spectrum of the 13CO-labeled sample, which indicated that at
least one CO ligand of the individual iron site was fully ex-
changed. Because of the very large amide peak at ∼1600 cm−1,
the acyl−iron ligand peak (∼1700 cm−1) was not visible; there-
fore, we could not determine the labeling of the acyl ligand
by 13CO.
The labeling experiments with [1-13C]acetate, [2-13C]acetate
(Mb. smithii), and [1-13C] pyruvate (Mt. marburgensis) revealed
the FeGP cofactor and the iron-free FeGP cofactor to have the
same mass increases (Figure 3B−D) indicating that the CO
ligands are not biosynthetically derived from these precursors.
When CO2 was the labeled precursor (Mb. smithii) the
complete FeGP cofactor showed a mass increase of 12 (Figure
4A, right), whereas the iron-free FeGP cofactor determined by
MALDI-TOF-MS showed a mass increase of only 10 (Figure
4A, left, m/z = 553). The decarboxylated form showed a mass
increase of nine (Figure 4A, left, m/z = 508). These findings
indicate that the two CO and the acyl group are biosyntheti-
cally derived from CO2. With CO as the 13C-labeled precursor
(Mb. smithii) the complete FeGP cofactor showed a maximal
mass increase of three (Figure 4B, right) (13C enrichment 70%,
see detailed analysis in the legend of Figure S4, Supporting
Information), whereas the iron-free FeGP cofactor determined
by MALDI-TOF-MS showed a maximal mass increase of one
(Figure 4B, left, m/z = 544) (80%). The decarboxylated form
showed no mass increase (Figure 4B, left, m/z = 499). These
findings indicate that the two CO and the acyl group are
derived from CO when added to the cultures. 13C from formate
was not incorporated in one of the iron ligands (Table 2).
Infrared Spectrum of FeGP cofactor Labeled with
13CO. To determine the incorporation of 13CO into the CO
ligands of the FeGP cofactor, we analyzed [Fe]-hydrogenase by
ATR-IR spectroscopy. To obtain large amounts of [Fe]-
hydrogenase, Mt. marburgensis was used for the experiment.
With the nonlabeled enzyme, peaks at 2009 and 1942 cm−1
were observed, which correspond to asymmetric and symmetric
coupling vibrations of two CO bound to the iron in cis
orientation, respectively (Figure 5).7 With the enzyme labeled
DISCUSSION
■
The FeGP cofactor contains 21 carbon atoms, 5 in the guanine
moiety, 5 in the ribose moiety, 9 in the pyridinol moiety, and 2
in the two CO ligands. In Mt. marburgensis growing
autotrophically on H2 and CO2, all 21 C atoms are derived
from CO2. The number of carbon atoms derived exclusively
from CO2 decreased to 12 when the growth medium was
supplemented with acetate. In Mb. smithii, which is dependent
on acetate for growth, always 12 carbon atoms are derived from
CO2 and 9 carbon atoms from acetate: 6 from C-1 of acetate
and 3 from C-2 of acetate. This is the result of the labeling
experiments with [1-13C]acetate, [2-13C]acetate, and 13CO2
reported above. The carbon atoms derived from the labeled
precursors are shown in Figure 1.
In the two methanogens, acetate is assimilated via activation
to acetyl-CoA, which is either condensed to acetoacetyl-CoA,
from which the isoprenoid membrane lipids are synthesized, or
reductively carboxylated to pyruvate, from which almost all
other cell components are generated. Labeling studies with
[1-13C]pyruvate revealed that 5 of the 21 C-atoms in the FeGP-
cofactor are derived from C-1 of pyruvate, one less than from
C-1 of acetate. This finding indicates that either one acetate is
incorporated into the FeGP cofactor without pyruvate being an
intermediate or one pyruvate is incorporated into the cofactor
loosing C-1 on the way.
The observation that six C-atoms are derived from C-1 of
acetate and only three C atoms from C-2 of acetate can be
explained by two findings: (i) Two of the 21 C-atoms of the
FeGP cofactor at the purine moiety are derived from glycine,
which in methanogens is synthesized from C-1 and C-2 of
pyruvate via 3-hydroxypyruvate and serine, C-3 of serine (=C-3
of pyruvate = C-2 of acetate) yielding methylene-H4MPT,
which is reduced to methane;22 (ii) The labeling pattern of the
pyridinol ring of the iron-free FeGP cofactor at C(i), C(e),
C(d), C(c), and C(b) is identical to that of the ribose moiety
indicating the involvement of two pentoses in the biosynthesis
of the FeGP cofactor.23 Pentoses in methanogens are
synthesized from hexulose-6-phosphate, C-1 of the hexulose
(=C-3 of pyruvate = C-2 of acetate) yielding methylene-
H4MPT, which is reduced to methane.24 The hexulose
phosphate is synthesized from pyruvate via reversal of the
Embden−Meyerhof pathway as indicated by the finding that
both in the ribose moiety and the pyridinol moiety two
neighboring carbon atoms are derived from C-1 of pyruvate.
When Mt. marburgensis was grown in the presence of
methionine, the FeGP cofactor contained only 20 carbon atoms
from CO2 and 1 C atom from the methyl group of methionine,
which in methanogens is synthesized from CO2 via methyl-
H4MPT, the methyl group of the latter being transferred to
homocysteine.25 The methyl group of methionine was
incorporated intact into the C(j) position of the pyridinol
moiety as revealed by MS and NMR analysis of the
L-[methyl-2H3]methionine-labeled cofactor indicating that
methyl transfer from methionine is catalyzed by a
Figure 5. Attenuated total reflection infrared spectra of the partially
purified [Fe]-hydrogenase preparation showing the CO ligand region
of the FeGP cofactor. Black line, the nonlabeled FeGP cofactor; red
line, FeGP cofactor labeled with 13C from 13CO.
with 13CO, the two CO frequencies were shifted to 1957 and
1892 cm−1 and broadened. The major peaks appear to
correspond to the CO ligands fully exchanged by 13CO, and
the shoulders at 1988 and ∼1900 cm−1 appear to be those of
3276
dx.doi.org/10.1021/ja211594m | J. Am. Chem. Soc. 2012, 134, 3271−3280

Journal of the American Chemical Society
Article
SAM-dependent methyltransferase rather than by a radical
SAM enzyme.26,27
CO ligands to iron in the FeGP cofactor is different from that
in both [FeFe]- and [NiFe]-hydrogenases.28−30
This leaves six C atoms in the FeGP cofactor not derived
from C-1 and C-2 of acetate, C-1 of pyruvate, or the methyl
group of methionine. Three of the six carbons are attributed to
three carbon atoms in the purine ring, which are synthesized
from CO2 (one via formate).19,21 The other three carbon atoms
are those of the two CO- and the one acyl-ligands to iron in the
cofactor. Labeling experiments with 13CO2 showed that the three
ligands are derived from CO2 and labeling experiments with acetate
and pyruvate that pyruvate is not an intermediate in their bio-
synthesis. The latter finding excludes dehydroglycine as an
intermediate, which was recently shown to be the precursor of the
CO ligands to iron in [FeFe]-hydrogenases.28,29 Dehydroglycine is
generated from p-hydroxyphenylalanine (tyrosine) in a radical
SAM reaction yielding p-cresol as second product, the two C-
atoms in dehydroglycine being derived from C-1 and C-2 of
tyrosine (=C-1 and C-2 of pyruvate). In the case of [NiFe]-
hydrogenase, glycerol has been shown to be an intermediate in
CO-ligand biosynthesis in Ralstonia eutropha when grown
under heterotrophic growth conditions.30 In methanogens,
glycerol is generated from pyruvate, which has been shown not
to be an intermediate in CO-ligands biosynthesis of [Fe]-
hydrogenase. The findings show that the biosynthesis of the
When Mb. smithii was grown in the presence of 13CO, the
three carbon-containing ligands became labeled. However, the
results cannot be interpreted to indicate that CO2 is
incorporated into the three iron ligands via free CO since
Mb. smithii lacks the only enzyme known to catalyze the
reduction of CO2 to CO, namely a nickel-dependent CO
dehydrogenase.17 Incorporation of CO was thus probably via a
salvage pathway although it cannot be excluded that Mb. smithii
contains a noncanonical CO dehydrogenase not yet identified.
Labeling of the acyl carbon by 13CO could occur by scrambling
of the CO- and acyl ligands, as observed in model compounds
irradiated by light.31,32 How the unique acyl iron bond is
formed remains elusive. The formation of the acetyl-nickel
bond from acetyl-CoA in the acetyl-CoA synthase/decarbon-
ylase of acetogens and methanogens could stand as model.33,34
Based on these results, several biosynthetic pathways can be
thought of following the retrosynthesis concept in organic
chemistry described by Eisenreich and Bacher.35 Each of these
pathways has a few unknowns. The pathway shown in Figure 6
assumes that the 2,3-dihydroxy-4-oxo-pentanoate and aspartate
condense to yield the pyridinol moiety most probably with a
pyridoxal phosphate bound β-alanine as intermediate.
Figure 6. Putative biosynthetic pathway of the FeGP cofactor of [Fe]-hydrogenase from hydrogenotrophic methanogenic archaea as deduced from
13C labeling experiments. Mt. marburgensis or the acetate auxotrophic Mb. smithii were grown on H2 and CO2 as energy sources in the presence of
[1-13C]acetate(∗), [2-13C]acetate(°), [1-13C]pyruvate, 13CO2, 13CO, or L-[methyl-2H3]methionine. After harvest and isolation of the FeGP cofactor
the labeling of the cofactor with 13C was determined via mass spectrometry and NMR. The labeling pattern was found to be consistent with a
synthesis of the FeGP cofactor from one acetate, two pyruvate, three CO2, and a methyl-group of methionine. Aspartate, 2,3-dihydroxy-4-oxo-
pentanoate and S-adenosyl-methionine (SAM) are proposed intermediates although alternative pathways can be thought of. The formation of the
pyridinol moiety from aspartate and the pentanoate could proceed via β-alanine as pyridoxal phosphate bound intermediate. The CO-ligands and the
acyl carbon are exclusively derived from CO2 but not via pyruvate C-1. In Mb. smithii, which is devoid of a nickel-dependent carbon mono-
xide dehydrogenase, CO2 reduction to CO by the nickel enzyme is excluded. The unknown reaction(s) are indicated by arrows with a broken line.
The predicted biochemical reactions in analogy to the retrosynthesis concept in organic chemistry are shown by the retrosynthesis arrow.
3277
dx.doi.org/10.1021/ja211594m | J. Am. Chem. Soc. 2012, 134, 3271−3280

Journal of the American Chemical Society
Article
The putative intermediate 2,3-dihydroxy-4-oxo-pentanoate
has a labeling pattern identical to that of the ribose moiety of
the FeGP cofactor. This finding strongly suggested that a C5
compound similar to ribose could be one of the precursors of
pyridinol biosynthesis (Figure 6). Although 2,3-dihydroxy-4-
oxo-pentanoate is not a known compound in methanogenic
archaea, it could be synthesized involving known reactions such
as the L-fuculose-phosphate aldolase reaction in which
methylglyoxal and dihydroxyacetone phosphate are condensed
to 6-deoxy-5-ketoallulose-1-phosphate (Figure 6).36 The
isomer 6-deoxy-5-ketofructose-1-phosphate has been proposed
to be an intermediate in aromatic amino acid biosynthesis in
Methanococcus maripaludis.37,38 One of these C6 compounds
could be converted to the proposed C5 intermediate via a series
of enzyme reactions involving a phosphatase, a mutase, and an
enzyme similar to hexulose-6-phosphate synthase to cleave the
terminal hydroxyl methyl group. The C5 product would have to
be further converted in dehydrogenase reactions to form the
carboxyl group. Alternatively the putative intermediate 2,3-
dihydroxy-4-oxo-pentanoate could be synthesized via aldol
condensation of methylglyoxal with glycolate or glycol
aldehyde. Decarboxylation of aspartate yields the second
putative intermediate β-alanine in the free or pyridoxal
phosphate bound form, which has the same labeling pattern
as the pyridinol ring at N(a), C(f), C(g), and C(h) (Figures 1
and 6). β-Alanine is an intermediate in the biosynthesis of
coenzyme A in bacteria.39 In the genome of archaea, a gene for
aspartate decarboxylase is not found. However, the gene
encoding the enzyme catalyzing pantothenate synthesis from
β-alanine and pantoate is present.40,41
In the biosynthesis of most pyridone-polyketides, the
pyridinol moieties are produced from a polyketide chain and
an amino acid, such as phenylalanine, tyrosine, serine, and
β-alanine.42−44 The amino acid backbones are incorporated
into the pyridone moiety at N(a), C(f), C(e), and C(d)
(following the designation of the atoms of the iron-free FeGP
cofactor). From the proposed biosynthetic pathway of
pyridone-polyketides, the 13C-labeling pattern of their pyridinol
ring, which was reconstructed from the biosynthetic pathway, is
distinct from that of the FeGP cofactor. This shows that the
biosynthetic precursors and reactions of the FeGP cofactor and
pyridone-polyketides are not the same.
Genome analysis of methanogenic archaea has identified
seven genes (hcgA−G) in addition to the [Fe]-hydrogenase
structural gene hmd as a gene cluster that are predicted to be
the FeGP cofactor biosynthesis genes.1 The homology search
revealed that one of the genes in the hcg gene cluster could be
for radical-SAM-dependent enzyme (hcgA) and one for a SAM-
dependent methyltransferase (hcgG). The hcgA gene shows
high sequence similarity to that of biotin synthases and other
radical-SAM-dependent enzymes.1 However, HcgA lacks the
CX3CX2C motif or CX4CX2C motif for [4Fe4S]-cluster that is
characteristic of the N-terminal region for the radical-SAM
protein superfamily and essential for radical formation.26,27
Instead, HcgA universally harbors a unique CX5CX2C motif.45
The protein encoded by the hcgG gene is annotated as a
fibrillarin-like protein based on sequence similarity. Interest-
ingly, the crystal structure of the C-terminal domain of the
fibrillarin homologue MJ0697 from Methanocaldococcus janna-
schii46 shows a similarity to the catalytic domain of SAM-
dependent methyltransferases.47 Sequence analysis of the hcgB
gene product indicates that it may catalyze the ligation of the
GMP and pyridinol parts of the cofactor. The crystal structure
of HcgB from Mt. thermautotrophicum has been determined
(PDB code: 3BRC). The homodimeric protein shares
structural similarities with a human inosine triphosphatase,
which catalyzes cleavage of pyrophosphate from ITP.48 The
labeling pattern of the GMP part of the cofactor follows the
canonical biosynthetic pathway also found in methanogenic
archaea.
CONCLUSION
■
The stable-isotope-labeling experiments have set the frame in
which the biosynthesis of the FeGP cofactor of [Fe]-
hydrogenase in methanogenic archaea has to occur. The
biosynthesis of the pyridinol moiety is definitely distinct from
that of pyridone-polyketides. 2,3-Dihydroxy-4-oxo-pentanoate
and β-alanine are discussed as intermediates to have at least a
hypothesis of how the biosynthesis may occur and of what gene
products could be involved. The C(j) methyl group attached to
the pyridinol ring is derived from the methyl group of
methionine and is most probably introduced via a SAM-
dependent methyltransferase reaction. The CO ligands are
synthesized from CO2, but pyruvate C-1 does not appear to be
involved indicating that the synthesis of the CO ligands in the
FeGP cofactor differs from that in [FeFe]-hydrogenases. For
the synthesis of the unique acyl−iron bond, the mechanism
proposed for the formation of the acetyl−nickel bond in the
catalytic cycle of acetyl-CoA synthase/decarbonylase is
considered.
MATERIALS AND METHODS
■
Stable Isotopes. [1-13C]Acetate (99% 13C), [2-13C]acetate (99%
13C), [1-13C]glycine (99% 13C), [1-13C]propionate (99% 13C),
2
L-[methyl-2H]methionine (99% H), [2,3-13C2]succinate (99% 13C),
and [1-13C]formate (99% 13C) were purchased from Euriso-tope.
13CO2 (99% 13C) and 13CO (99% 13C) were from Sigma-Aldrich.
Growth of Mt. marburgensis and of Mb. smithii.
Mt. marburgensis (DSM 2133) was grown at 65 °C in a mineral salt
medium described by Schonheit et al.49 For cultivation under nickel-
̈
limited conditions, the nickel concentration in the medium was de-
creased: in the 10 L fermenter, the nickel in the medium was omitted,
and in the 500 mL glass fermenter and 2 L glass bottle cultures, the
nickel concentration in the medium was lowered to 0.05 μM.15
For NMR sample preparation, Mt. marburgensis was cultivated in a
10 L fermenter with a continuous gas flow of H2/CO2 (80:20, vol/vol;
2
1.5 L/min). The 13C- or H-labeled compounds were added to the
medium at the following final concentrations: 5 mM [2-13C]acetate,
5 mM [1-13C]acetate, 6.8 mM [1-13C]pyruvate, or 5 mM L-[methyl-2H]-
methionine. For the electrospray ionization Fourier transform ion
cyclotron resonance mass spectrometry (ESI-FT-ICR-MS) sample
preparation, Mt. marburgensis was grown in the presence of 40 mM
[1-13C]pyruvate, 5 mM [2,3-13C2]succinate, or 40 mM [1-13C]formate
in a 500 mL fermenter in 360 mL of medium with a continuous gas
flow of H2/CO2 (80:20, vol/vol; 0.2 L/min); the pH of the medium
was adjusted to 6.0 to facilitate the incorporation of the 13C-
compounds.50,51 For 13C-incorporation from 13CO, Mt. marburgensis
was batch cultivated in 500 mL of medium in 2 L glass bottles sealed
with rubber stoppers under H2/CO2/13CO (80:20:2.5, by vol). The
gas phase was exchanged three times during the daytime.
Mb. smithii (DSM 861) was grown at 37 °C under H2/CO2 (80:20,
vol/vol) in a modified basal medium based on that of Balch et al.,52 in
which sodium formate was omitted and the concentration of yeast
extract was decreased from 2 g/L to 0.05 g/L. For the labeling, the
medium contained 30 mM [1-13C]acetate, 30 mM [2-13C]acetate,
5 mM [1-13C]glycine, 2.5 mM [1-13C]propionate, or 2.5 mM
L-[methyl-2H]methionine. For the 13CO2 or 13CO incorporation
experiments, the standard gas phase was substituted with H2/13CO2
(80:20, vol/vol) or H2/CO2/13CO (80:20:2.5, by vol).
3278
dx.doi.org/10.1021/ja211594m | J. Am. Chem. Soc. 2012, 134, 3271−3280

Journal of the American Chemical Society
Article
Purification of the FeGP Cofactor from Mt. marburgensis
and Generation of the Iron-Free FeGP Cofactor. [Fe]-hydro-
genase was purified from 90 g of Mt. marburgensis cells by ammonium
sulfate precipitation and anion-exchange chromatography as described
previously.53 The FeGP cofactor was extracted from the purified
protein by 60% methanol/1 mM 2-mercaptoethanol/1% ammonia for
15 min at 40 °C under strictly anoxic conditions.54 The cofactor
preparation thus obtained was essentially homogeneous.
Protein and Enzyme Activity Assays. The protein concentration
was determined by the Bradford method using the assay solution from
Bio-Rad and bovine serum albumin as the standard. The enzyme
activity of [Fe]-hydrogenase purified from methanogens and of that
reconstituted from the isolated FeGP cofactor and the apoenzyme
heterologously produced in Escherichia coli were assayed as described
previously.53
To obtain iron-free FeGP cofactor, 10 mL of cofactor preparation
containing ∼0.3 mM FeGP cofactor was placed on ice and exposed to
high-intensity light (Cold Light Sources KL 2500 LC, Schott); the pH
of the solution was then adjusted to 1−2. To completely remove the
released iron ions, 0.6 mM EDTA (final concentration) was added to
the light-exposed solution. The light decomposition product (iron-free
FeGP cofactor) was purified by HPLC on a Synergi 4 m Polar RP 80A
column (250 mm × 4.6 mm) using a linear H2O/CH3OH gradient.
The purified iron-free FeGP cofactor was dried by evaporation at 4 °C
and then dissolved in 200 μL of D2O prior to NMR measurements.
Purification of the FeGP Cofactor from Mb. smithii. [Fe]-
hydrogenase was partially purified from Mb. smithii cells under strictly
anoxic conditions under red light. The cells (2−3 g fresh weight) were
harvested at the end of the exponential growth phase by centrifugation
(16 000 × g for 30 min at 4 °C) and suspended in 5 mL of 50 mM
phosphate buffer, pH 7.0. The cells were disrupted on ice by
ultrasonication (Sonopuls GM200, Ti73 tip, Bandelin) for 15 min with
50% cycle and 60% power. The cell extract exhibited 0.03 U/mg [Fe]-
hydrogenase activity (Figure S7, Supporting Information). After
centrifugation (16 000 × g for 30 min at 4 °C), the soluble fraction
was separated from the membrane fraction by ultracentrifugation (140
000 × g for 30 min at 4 °C). Ammonium sulfate powder was added to
60% saturation to the supernatant on ice, and the precipitated proteins
were removed by centrifugation at 13 000 × g for 15 min at 4 °C. The
[Fe]-hydrogenase fraction was precipitated by adding ammonium
sulfate to 90% saturation to the supernatant. After centrifugation
(16 000 × g for 30 min at 4 °C), the precipitate was dissolved in anoxic
water and desalted using an ultrafilter (30-kDa cutoff). The FeGP
cofactor was extracted either in 60% methanol/1 mM 2-mercapto-
ethanol/1% ammonia for MALDI-TOF-MS or in 50% acetic acid for
ESI-FT-ICR-MS, as described previously.10 The FeGP cofactor was
separated from the unfolded proteins by ultrafiltration (5-kDa cutoff)
before mass spectrometry.
ASSOCIATED CONTENT
■
S
* Supporting Information
Standard 13C NMR spectra of the 13C-labeled iron-free FeGP
cofactor with various labeling patterns, 1H−13C HMBC of iron-
free FeGP cofactor labeled with 13C from [2-13C]acetate, H
1
2
NMR spectrum of iron-free FeGP cofactor labeled with H
from L-[methyl-2H3]methionine, simulation of the isotopic
profiles of the mass spectra, MALDI-TOF-MS of the FeGP
cofactor detected in the cell extract of Mb. smithii cultivated in
the presence of 2.5 mM [1-13C]propionate, MALDI-TOF-MS
spectrum of the FeGP cofactor detected in the cell extract of
Mt. marburgensis cultivated in the presence of 5 mM
[2,3-13C2]succinate, and [Fe]-hydrogenase activity in cell
extract of Mb. smithii. This material is available free of charge
via the Internet at http://pubs.acs.org.
AUTHOR INFORMATION
■
Corresponding Author
shima@mpi-marburg.mpg.de
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
This work was supported by a grant of the Max Planck Society
to R. K. Thauer, by a grant of the PRESTO program, Japan
Science and Technology Agency (JST), to S. Shima and by the
BMBF BioH2 project. We thank W. Buckel (Philipps University
Marburg) for discussions on the reaction mechanism and R. K.
Thauer for many helpful suggestions including the critical
reading of the manuscript. Also the technical assistance of Jan
Bamberger during ESI-MS measurements is acknowledged.
Mass Spectrometry and NMR Spectroscopy. The FeGP
cofactor was analyzed by MALDI-TOF-MS using a 4800 Proteomics
Analyzer (Applied Biosystems/MDS Sciex) with α-cyano-4-hydrox-
ycinnamic acid dissolved in 70% (v/v) acetonitrile and 0.1% (v/v)
trifluoroacetic acid as matrix. The cofactor was analyzed by ESI-FT-
ICR mass spectrometry using an LTQ-FT-Ultra (Thermo Scientific)
equipped with HPLC (Macherey-Nagel 125/2 Nucleodur C18ec
column and water/methanol as solvents). For NMR, about 0.5 mg of
the iron-free cofactors with various labeling patterns were dissolved in
REFERENCES
■
(1) Thauer, R. K.; Kaster, A. K.; Goenrich, M.; Schick, M.; Hiromoto,
T.; Shima, S. Annu. Rev. Biochem. 2010, 79, 507.
(2) DiMarco, A. A.; Bobik, T. A.; Wolfe, R. S. Annu. Rev. Biochem.
1990, 59, 355.
2
200 μL of H2O and filled into Wilmad 3 mm NMR tubes (Rototec
Spintec). Two-dimensional (2D) experiments were performed on a
(3) Zirngibl, C.; van Dongen, W.; Schworer, B.; von Bunau, R.;
̈
̈
1
Bruker AV600 spectrometer with a TXI H−13C/15N probe installed
Richter, M.; Klein, A.; Thauer, R. K. Eur. J. Biochem. 1992, 208, 511.
(4) Hiromoto, T.; Ataka, K.; Pilak, O.; Vogt, S.; Stagni, M. S.; Meyer-
Klaucke, W.; Warkentin, E.; Thauer, R. K.; Shima, S.; Ermler, U. FEBS
Lett. 2009, 583, 585.
(5) Hiromoto, T.; Warkentin, E.; Moll, J.; Ermler, U.; Shima, S.
Angew. Chem., Int. Ed. 2009, 48, 6457.
(6) Korbas, M.; Vogt, S.; Meyer-Klaucke, W.; Bill, E.; Lyon, E. J.;
Thauer, R. K.; Shima, S. J. Biol. Chem. 2006, 281, 30804.
(7) Lyon, E. J.; Shima, S.; Boecher, R.; Thauer, R. K.; Grevels, F. W.;
Bill, E.; Roseboom, W.; Albracht, S. P. J. J. Am. Chem. Soc. 2004, 126,
14239.
(8) Salomone-Stagni, M.; Stellato, F.; Whaley, C. M.; Vogt, S.;
Morante, S.; Shima, S.; Rauchfuss, T. B.; Meyer-Klaucke, W. Dalton
Trans. 2010, 39, 3057.
with z-gradient, while 13C and 31P spectra were recorded on a Bruker
AV500 spectrometer with broadband forward probe BBFO. The one-
dimensional (1D) spectra, 1H, 13C, and 31P, and the 2D spectra,
1H−13C HSQC and HMBC spectra, were recorded with standard
pulse programs at 300 K.55 The 1D spectra were acquired with 65 536
data points, while 2D spectra were collected using 2048 points in the
F2 dimension and 512 increments in the F1 dimension. For the HSQC
1
spectra, 32−64 transients, for the HMBC 64−128 transients, for H
128−1024, and for 13C 24576−92160 scans were used. The relaxation
delay was 3.0 s. Chemical shifts of 1H and 13C spectra were referenced
to the signals of 4,4-dimethyl-4-silapentane sodium sulfonate. The
spectra were processed by Bruker Topspin 3.0.
Infrared Spectroscopy. Attenuated total reflection infrared
(ATR-IR) spectroscopy was performed using a Bruker Vertex 70
FTIR spectrometer equipped with an MCT detector. ATR optical
configuration (DuraSamplIR, SensIR Technologies) was employed
with a silicon prism.56
(9) Shima, S.; Lyon, E. J.; Thauer, R. K.; Mienert, B.; Bill, E. J. Am.
Chem. Soc. 2005, 127, 10430.
(10) Shima, S.; Schick, M.; Ataka, K.; Steinbach, K.; Linne, U. Dalton
Trans. 2012, 41, 767.
3279
dx.doi.org/10.1021/ja211594m | J. Am. Chem. Soc. 2012, 134, 3271−3280

Journal of the American Chemical Society
Article
(11) Henry, T. A. Plant Alkaloids, 4th ed.; The Blakiston Company:
Philadelphia, PA, 1949.
(12) Shima, S.; Pilak, O.; Vogt, S.; Schick, M.; Stagni, M. S.; Meyer-
Klaucke, W.; Warkentin, E.; Thauer, R. K.; Ermler, U. Science 2008,
321, 572.
(13) Yang, X. Z.; Hall, M. B. J. Am. Chem. Soc. 2009, 131, 10901.
(14) Dent, S. P.; Eaborn, C.; Pidcock, A.; Ratcliff, B. J. Organomet.
Chem. 1972, 46, C68.
(45) McGlynn, S. E.; Boyd, E. S.; Shepard, E. M.; Lange, R. K.;
Gerlach, R.; Broderick, J. B.; Peters, J. W. J. Bacteriol. 2010, 192, 595.
(46) Wang, H. M.; Boisvert, D.; Kim, K. K.; Kim, R.; Kim, S. H.
EMBO J. 2000, 19, 317.
(47) Sankpal, U. T.; Rao, D. N. Crit. Rev. Biochem. Mol. Biol. 2002,
37, 167.
(48) Stenmark, P.; Kursula, P.; Flodin, S.; Graslund, S.; Landry, R.;
Nordlund, P.; Schuler, H. J. Biol. Chem. 2007, 282, 3182.
(49) Schonheit, P.; Moll, J.; Thauer, R. K. Arch. Microbiol. 1980, 127, 59.
̈
(15) Afting, C.; Hochheimer, A.; Thauer, R. K. Arch. Microbiol. 1998,
169, 206.
(50) Huster, R.; Thauer, R. K. FEMS Microbiol. Lett. 1983, 19, 207.
(51) Diekert, G.; Gilles, H. H.; Jaenchen, R.; Thauer, R. K. Arch.
Microbiol. 1980, 128, 256.
(16) Kaster, A. K.; Goenrich, M.; Seedorf, H.; Liesegang, H.;
Wollherr, A.; Gottschalk, G.; Thauer, R. K. Archaea 2011, 2011,
No. 973848.
(17) Samuel, B. S.; Hansen, E. E.; Manchester, J. K.; Coutinho, P. M.;
Henrissat, B.; Fulton, R.; Latreille, P.; Kim, K.; Wilson, R. K.; Gordon,
J. I. Proc. Natl. Acad. Sci. U.S.A. 2007, 104, 10643.
(18) Shima, S.; Lyon, E. J.; Sordel-Klippert, M. S.; Kauss, M.; Kahnt,
J.; Thauer, R. K.; Steinbach, K.; Xie, X. L.; Verdier, L.; Griesinger, C.
Angew. Chem., Int. Ed. 2004, 43, 2547.
(52) Balch, W. E.; Fox, G. E.; Magrum, L. J.; Woese, C. R.; Wolfe,
R. S. Microbiol. Rev. 1979, 43, 260.
(53) Shima, S.; Schick, M.; Tamura, H. Methods Enzymol. 2011, 494,
119.
(54) Lyon, E. J.; Shima, S.; Buurman, G.; Chowdhuri, S.; Batschauer,
A.; Steinbach, K.; Thauer, R. K. Eur. J. Biochem. 2004, 271, 195.
(55) Berger, S.; Braun, S. 200 and More NMR experiments, a practical
course; Wiley-VCH: Weinheim, Germany, 2004.
(19) Choquet, C. G.; Richards, J. C.; Patel, G. B.; Sprott, G. D. Arch.
Microbiol. 1994, 161, 471.
(56) Shima, S.; Ataka, K. FEBS Lett. 2011, 585, 353.
(20) Vogt, S. Doctor thesis, Philipps University, 2007.
(21) Kang, Y. N.; Tran, A.; White, R. H.; Ealick, S. E. Biochemistry
2007, 46, 5050.
(22) Angelaccio, S.; Chiaraluce, R.; Consalvi, V.; Buchenau, B.;
Giangiacomo, L.; Bossa, F.; Contestabile, R. J. Biol. Chem. 2003, 278,
41789.
(23) Choquet, C. G.; Richards, J. C.; Patel, G. B.; Sprott, G. D. Arch.
Microbiol. 1994, 161, 481.
(24) Goenrich, M.; Thauer, R. K.; Yurimoto, H.; Kato, N. Arch.
Microbiol. 2005, 184, 41.
(25) Schroder, I.; Thauer, R. K. Eur. J. Biochem. 1999, 263, 789.
̈
(26) Buckel, W.; Thauer, R. K. Angew. Chem., Int. Ed. 2011, 50,
10492.
(27) Zhang, Q.; van der Donk, W. A.; Liu, W. Acc. Chem. Res. 2012,
DOI: 10.1021/ar200202c.
(28) Shepard, E. M.; Duffus, B. R.; George, S. J.; McGlynn, S. E.;
Challand, M. R.; Swanson, K. D.; Roach, P. L.; Cramer, S. P.; Peters, J. W.;
Broderick, J. B. J. Am. Chem. Soc. 2010, 132, 9247.
(29) Swanson, K. D.; Duffus, B. R.; Beard, T. E.; Peters, J. W.;
Broderick, J. B. Eur. J. Inorg. Chem. 2011, 2011, 935.
(30) Burstel, I.; Hummel, P.; Siebert, E.; Wisitruangsakul, N.; Zebger,
I.; Friedrich, B.; Lenz, O. J. Biol. Chem. 2011, 286, 44937.
(31) Chen, D.; Scopelliti, R.; Hu, X. Angew. Chem., Int. Ed. 2010, 49,
7512.
(32) Chen, D. F.; Scopelliti, R.; Hu, X. L. Angew. Chem., Int. Ed. 2011,
50, 5671.
(33) Ragsdale, S. W. J. Inorg. Biochem. 2007, 101, 1657.
(34) Tan, X. S.; Surovtsev, I. V.; Lindahl, P. A. J. Am. Chem. Soc.
2006, 128, 12331.
(35) Bacher, A.; Rieder, C.; Eichinger, D.; Arigoni, D.; Fuchs, G.;
Eisenreich, W. FEMS Microbiol. Rev. 1998, 22, 567.
(36) Grochowski, L. L.; Xu, H. M.; White, R. H. J. Bacteriol. 2006,
188, 2836.
(37) Porat, I.; Sieprawska-Lupa, M.; Teng, Q.; Bohanon, F. J.; White,
R. H.; Whitman, W. B. Mol. Microbiol. 2006, 62, 1117.
(38) White, R. H. Biochemistry 2004, 43, 7618.
(39) Begley, T. P.; Kinsland, C.; Strauss, E. Vitam. Horm. 2001, 61,
157.
(40) Genschel, U. Mol. Biol. Evol. 2004, 21, 1242.
(41) Ronconi, S.; Jonczyk, R.; Genschel, U. FEBS J. 2008, 275, 2754.
(42) Fujita, Y.; Oguri, H.; Oikawa, H. Tetrahedron Lett. 2005, 46,
5885.
(43) Weber, T.; Laiple, K. J.; Pross, E. K.; Textor, A.; Grond, S.;
Welzel, K.; Pelzer, S.; Vente, A.; Wohlleben, W. Chem. Biol. 2008, 15,
175.
(44) Eley, K. L.; Halo, L. M.; Song, Z.; Powles, H.; Cox, R. J.; Bailey,
A. M.; Lazarus, C. M.; Simpson, T. J. ChemBioChem 2007, 8, 289.
3280
dx.doi.org/10.1021/ja211594m | J. Am. Chem. Soc. 2012, 134, 3271−3280