Structural and functional synthetic model of mono-iron hydrogenase featuring an anthracene scaffold

Article abstract of DOI:10.1038/nchem.2707

Mono-iron hydrogenase was the third type of hydrogenase discovered. Its Lewis acidic iron(II) centre promotes the heterolytic cleavage of the H-H bond and this non-redox H₂ activation distinguishes it from the well-studied dinuclear [FeFe] and [NiFe] hydrogenases. Cleavage of the H-H bond is followed by hydride transfer to the enzyme's organic substrate, H₄ MPT⁺, which serves as a CO₂ carrier' in methanogenic pathways. Here we report a scaffold-based synthetic approach by which to model mono-iron hydrogenase using an anthracene framework, which supports a biomimetic fac-C,N,S coordination motif to an iron(II) centre. This arrangement includes the biomimetic and organometallic Fe-C σ bond, which enables bidirectional activity reminiscent of the native enzyme: the complex activates H₂ under mild conditions, and catalyses C-H hydride abstraction plus H₂ generation from a model substrate. Notably, neither H₂ activation nor C-H hydride abstraction was observed in the analogous complex with a pincer-type mer-C,N,S ligation, emphasizing the importance of the fac-C,N,S-iron(II) motif in promoting enzyme-like reactivity.

Full text of DOI:10.1038/nchem.2707

ARTICLES
PUBLISHED ONLINE: 23 JANUARY 2017 | DOI: 10.1038/NCHEM.2707
Structural and functional synthetic model of mono-
iron hydrogenase featuring an anthracene scaffold
Junhyeok Seo, Taylor A. Manes and Michael J. Rose
*
Mono-iron hydrogenase was the third type of hydrogenase discovered. Its Lewis acidic iron(II) centre promotes the
heterolytic cleavage of the H–H bond and this non-redox H2 activation distinguishes it from the well-studied dinuclear
[FeFe] and [NiFe] hydrogenases. Cleavage of the H–H bond is followed by hydride transfer to the enzyme’s organic
substrate, H4MPT+, which serves as a CO2 ‘carrier’ in methanogenic pathways. Here we report a scaffold-based synthetic
approach by which to model mono-iron hydrogenase using an anthracene framework, which supports a biomimetic fac-C,N,
S coordination motif to an iron(II) centre. This arrangement includes the biomimetic and organometallic Fe–C σ bond,
which enables bidirectional activity reminiscent of the native enzyme: the complex activates H2 under mild conditions, and
catalyses C–H hydride abstraction plus H2 generation from a model substrate. Notably, neither H2 activation nor C–H
hydride abstraction was observed in the analogous complex with a pincer-type mer-C,N,S ligation, emphasizing the
importance of the fac-C,N,S-iron(II) motif in promoting enzyme-like reactivity.
he implementation of molecular scaffolds for synthetic mod-
elling of metalloenzymes can achieve structural and functional
mimetics of active sites1–5. For example, as far back as 1996,
a
CO
Fe
N
CO
Fe
N
OH2
H
S
C
– H2O
+ H2O
S
C
Cys176
Cys176
T
O
O
+ H2
– H2
Holm used a tribenzylthiolate scaffold to nucleate an [Fe3S4]
cuboid moiety1. More recently, Agapie used an analogous tri-
substituted benzene scaffold to prepare a photosystem II synthetic
model, Mn3O4Ca cubane2, and a naphthalene dithiolate framework
was used to support diiron centres in [FeFe] hydrogenases models3,
prior to the elucidation of the importance of the azadithiolate
moiety for enzyme function6. In another scaffold-supported
system, Murray and co-workers have assembled tri-iron clusters
by utilizing a sandwich cryptand featuring a benzene framework
with enclosed tris-nacnac chelates4. However, the use of such scaf-
folds to arrange a heteroleptic facial triad of donor sites (including
an organometallic M–C bond) for mononuclear biomimetics has
been less explored, possibly due to a lack of bio-inorganic systems
that inspire such work.
Nature uses dinuclear and mononuclear hydrogenases as
catalysts to utilize and generate dihydrogen (H2). The [FeFe] and
[NiFe] hydrogenases are known for proton reduction and H2
oxidation. In 2009, the X-ray structure of the third type of hydroge-
nase—mono-[Fe] hydrogenase—revealed a unique active site
structure7,8 which has led to an array of synthetic studies9–12. The
active site contains an octahedral Fe(II) ion ligated by three donors
(Cacyl, Npyridone and Scysteine) in facial coordination (Fig. 1a, in blue).
The remaining coordination sites are occupied by two molecules of
carbon monoxide, and one solvent/substrate binding site (Fig. 1a,
maroon). The enzyme catalyses the reversible conversion of methenyl
H4MPT+ to methylene H4MPT through the heterolytic cleavage of
H2 to a hydride and a proton. Notably, H4MPT+ acts as a C1 carrier:
CO
O
CO
OH
O
O
O
–
O
–
P
P
Guanyl
O
Guanyl
O
O
O
b
CO
Fe
I
N
S
HN
N
S
C
CO
Fe
CO
O
Br
HN
CO
C
Present model
Facial chelation
Previous model
Equatorial chelation
O
Figure 1 | Comparison of enzyme active site and model complexes. a,
Active site of mono-iron hydrogenase8, highlighting the facial arrangement
of ligands (blue and maroon), and including a proposed mechanism of H2
activation. b, The target ‘anthracene scaffold’ complex and our previously
reported synthetic model (note: the present model complex uses a
thioether-S donor to model the Cys176-thiolate)13.
C14 in the cofactor is derived from CO2 in the methanogenic models, a combined structural–functional modelling study of the
pathway of CO2 reduction.
enzyme’s activity (with respect to hydride abstraction, H2 activation
Several structural models of the [Fe]-hydrogenase active site have and hydride transfer) has remained a challenge. One recent report
been reported13–24. Hu17–20 and Song21,22 reported biomimetic Fe(II) by Hu and coworkers about H2 activation included a mononuclear
complexes with an acyl group, and Pickett has described a simple iron-acyl model complex; however, a PNP (Et2PCH2NMeCH2PEt2)-
synthetic method to prepare the closely related Fe(II)-carbamoyl type ligand unrelated to the mono-iron hydrogenase active site was
model complexes23,24, in which the methylene unit is replaced with necessary to achieve the H2 reactivity20. In related work, DuBois
an amide linkage. However, despite these families of biomimetic and Bullock reported that an analogous non-biomimetic complex
Department of Chemistry, The University of Texas at Austin, Austin, Texas 78712, USA. e-mail: mrose@cm.utexas.edu
*
NATURE CHEMISTRY | ADVANCE ONLINE PUBLICATION | www.nature.com/naturechemistry
1
© 2017 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.

NATURE CHEMISTRY DOI: 10.1038/NCHEM.2707
ARTICLES
a
(i) [Pd(dba)2], SPhos
KOAc, B2Pin2, dioxane
H2N
(ii) K3PO4, H2O
Br
Cl
N
Cl
Cl
N
NH2
B(OH)2
Na2CO3
XPhos, [Pd(dba)2]
THF/H2O, Δ
S
I
[Fe(CO)4(I)2]
N
S
CO
CO
Fe
CH2Cl2
–30 °C to RT
HN
N
C
S
1
NH2
O
b
c
I1
I
C47
Fe1
N2
N
S
Fe
P1
C46
C1
C
O
O1
S1
Figure 2 | Synthesis and DFT-optimized structure of compound 1, and the structure of 2 determined by X-ray crystallography. a, Synthesis of the
anthracene-based ligand Anth•NH Py•S and its iron carbamoyl derivative 1. b, The DFT calculated, geometry optimized structure of 1 (PW91/6-31G**, except
2
for iodine, 6-311G**). c, Molecular structure (50% thermal ellipsoids) for the PPh3 derivative of 1, namely [(Anth•CNHNS)Fe(CO)2(PPh3)(I)] (2); hydrogen
atoms are omitted and PPh3 is truncated for clarity. Selected bond distances (Å): Fe1–C1, 1.933(5); Fe1–C46, 1.814(5); Fe1–C47, 1.881(7); Fe1–N2, 2.008(4);
Fe1–I1, 2.689(8); Fe1–P1, 2.335(14).
[(CpC6F5)Fe(PNP)]+ also achieved H2 activation25. Thus, the key using
a tandem borylation and Suzuki coupling. Briefly,
components and geometry required for H2 activation in the Hmd the aminopyridine was borylated with B2Pin2 in the presence
active site (no PNP-type ligand) remain unexplored. Considering of [Pd(dba)2]/SPhos and a weak base (KOAc) in refluxing
the facial geometry of the C,N,S donors in the active site, a precise dioxane. Following hydrolysis of the pinacol with aqueous K3PO4,
chelation-based approach to the facial motif has not yet been pursued. 1,8-dichloroanthracene was coupled to the in situ-prepared boronic
The strategy of our research programme is to utilize an anthra- acid, thus affording Anth•NH2Py•Cl (42% yield, 1.0 g). Subsequent
cene ‘scaffold’ to deploy a complex set of donors in three dimensions coupling of Anth•NH2Py•Cl to 3-(methylthio)phenylboronic acid
(that is, unlike planar or pincer-type ligands), thereby restricting the with [Pd(dba)2]/XPhos in THF/H2O at 85 °C afforded the final
chelation to a facial motif. And although other facially chelating ligand Anth•NH2Py•S in 37% yield (0.50 g). For metallation, reaction
platforms certainly exist (Tp, P3, N3tacn, etc.)26–30, we have of Anth•NH2Py•S with [Fe(CO)4(I)2] in CH2Cl2 at −30 °C → 25 °C
developed a new ligand system for the incorporation of complex afforded a dark yellow solution, and precipitation with Et2O
and biomimetic donors for accurate structural modelling of the afforded the target complex [(Anth•CNHNS)Fe(CO)2(I)] (1).
enzyme. Herein, we report (i) the synthesis of an asymmetric,
anthracene-based ligand; (ii) the synthesis and characterization of Structural and spectroscopic characterization. The infrared (IR)
its iron carbonyl complex; and (iii) its functional behaviour regard- spectrum of 1 (Supplementary Fig. 1) exhibits two strong ν(C≡O)
ing hydride abstraction, H2 generation, H2/D2 activation and scram- features at 2,031 and 1,981 cm–1, as well as a sharp ν(C==O) feature at
bling, and the detection of the critical Fe–H/D intermediate.
1,661 cm–1 emanating from the carbamoyl moiety. The shifted
aromatic (pyridine) and anthracene-based 1H NMR resonances
further support the formation and purity of 1. The metallation of the
anthracene-based ligand is also indicated in the anthracene 9,10-H
resonances, which shift upfield upon Fe complexation (8.32, 8.59 ppm
in d8-THF) compared to the free ligand (8.57, 8.74 ppm in d8-THF).
Solutions of 1 are stable for weeks if kept in the dark (CHCl3,
CH2Cl2, THF, DCE (1,2-dichloroethane) and FPh; not DMSO, H2O,
Results and discussion
Syntheses. The key aminopyridine/thioether ligand Anth•NH2Py•S
was obtained through a novel synthetic route (Fig. 2a). Following
conversion of 1,8-dichloroanthraquinone to 1,8-dichloroanthracene
via sequential NaBH4 reductions, the anthracene scaffold was
selectively coupled to one equiv. of 5-bromo-2-aminopyridine
2
NATURE CHEMISTRY | ADVANCE ONLINE PUBLICATION | www.nature.com/naturechemistry
© 2017 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.

NATURE CHEMISTRY DOI: 10.1038/NCHEM.2707
ARTICLES
+
6.12 ppm
[Tl](BArF)
THF
THF
N
F
F
1
S
Im–H
d8-THF
H
CO
CO
Fe
HN
N
N
3
C
F
F
3
O
Im–H
R.T.
THF
O
O
2 h
F
N
F
F
H2
+
N
–
C
F
H
12 h
N
S
O
CO
CO
Fe
HN
C
O
7.0
6.5
6.0
OH
ppm
Figure 4 | Hydride abstraction from imidazolidine by complex 3 monitored
by 1H NMR. Series of 1H NMR spectra for stoichiometric hydride abstraction
from 1,3-bis(2,6-difluorophenyl)-2-(p-tolyl)imidazolidine (Im–H) by complex
3 in d8-THF at 25 °C, where the Im–H methine C–H peak at 6.12 ppm
decreases over the course of the reaction.
BArF
F
F
H
N
S
N
N
CO
CO
Fe
+
HN
F
F
C
Im+
4
Critically, the iodide ligand in 1 (the substrate binding site after
halide abstraction) is in the trans position to the organometallic
Fe–C=O) motif, analogous to the H2O/substrate binding site
observed in the enzyme active site. The synthetic solvato species
[(Anth•CNHNS)Fe(CO)2(THF)](BArF) (3) is formed by a reaction
of 1 with [Tl]BArF (BArF = tetrakis[(3,5-trifluoromethyl)phenyl]
borate) in THF. The solvato species remains diamagnetic and
the ν(CO) values are minimally perturbed compared to 1, owing
to the cis location of the CO ligands relative to the I–/THF
binding site (1H NMR and IR characterization detailed in the
O
Figure 3 | C–H hydride abstraction from imidazolidine and H2 evolution
mediated by complex 3. Generation of the THF-solvato complex 3 by halide
removal from 1 with [Tl](BArF), and its hydride abstraction reactivity from
1,3-bis(2,6-difluorophenyl)-2-(p-tolyl)imidazolidine (Im–H) to afford the
corresponding imidazolium (Im+) and the Fe–H species 4. Intermediate 4
proceeds to form H2 via hydride transfer to a non-coordinating proton
source (2,6-di-tert-butyl-4-methoxyphenol).
MeOH). However, exposure to ambient light causes slow degradation of Supplementary Information).
the complex (t ≈ 3 h) via loss of CO ligands as monitored by solution
½
IR spectroscopy (40 mM of 1 in THF). Crystallization attempts for 1 C–H hydride abstraction reactivity. Next, we turned to the
were not successful; however, evidence for the structural feasibility of reactivity of the complexes (Fig. 3). It is notable that there is no
1 is provided along two fronts: (i) by DFT calculation (Fig. 2b) and precedent for hydride abstraction reactions in biomimetic models
(ii) by the X-ray structure of the corresponding PPh3 derivative of mono-iron hydrogenase. To demonstrate the function of this
(Fig. 2c). The geometry-optimized structure of 1 reveals the roughly synthetic model, we first investigated hydride abstraction activity.
perpendicular orientation of the pyridine and aryl-thioether moieties We utilized the model substrate 1,3-bis(2,6-difluorophenyl)-
with respect to the anthracene scaffold. This structural motif promotes 2-(4-tolyl)imidazolidine (Im–H) and its corresponding hydride-
the binding of the C,N,S donor set exclusively in the fac orientation. abstracted imidazolium congener (Im+), reported by Meyer and
Further evidence for the S-bound state (that is, not THF-bound state) co-workers31. Upon treatment of the solvato complex 3 with ∼1
1
of 1 is provided by DFT calculation (Supplementary Tables 2 and 3) equiv. of Im–H (d8-THF, 25 °C), the H NMR spectrum showed
and ligand competition studies with AsPh3. Briefly, treatment of 1 disappearance of the majority of the methine resonance at
with AsPh3 results in no change in spectroscopic properties, and 6.12 ppm within 2 h; the resonance diminished to completion
unbound AsPh3 is quantitatively recovered (see Supplementary over the course of 12 h (Fig. 4). Quenching the reaction mixture
Information, page 8 for details).
with H2O allowed detection of the imidazolium product, as
X-ray evidence for the core organometallic structure of 1 was determined by electrospray ionization (ESI)–MS (ESI–MS(+) m/z:
obtained via derivatization of 1 with an exogenous L-type ligand, 385.1321). To demonstrate multiple turnovers, we conducted the
which proved fruitful in crystallization. Treatment of 1 with PPh3 analogous hydride abstraction in the presence of 5 equiv. of
resulted in the formation and crystallization of [(Anth•CNHNS)Fe imidazolidine and 5 equiv. of a bulky, non-coordinating proton
(CO)2(PPh3)(I)] (2) (Fig. 2c), which retains the core structural source, namely 2,6-di-tert-butyl-4-methoxyphenol (dBPhOH). Indeed,
features of its precursor 1. Complex 2 exhibits a facial arrangement analysis of the headspace gas (gas chromatography) following the
of the CNP donor set, wherein the PPh3 ligand simply substitutes catalytic reaction revealed the presence of H2 as a concomitant product
for the weaker thioether-S; the two CO ligands and iodide comprise (turnover number = 4, Supplementary Fig. 8). We thus demonstrated
the orthogonal facial triad. For comparison, DFT calculation was also the enzymatic model reaction (H+ + [C–H] → H2 + [C+]) mediated by
performed on 2, and the resulting bond metrics (Supplementary an Fe-based structural model of Hmd.
Table 1) are consistent with the X-ray structure. This lends further
support to our claim regarding the structure of 1. An elaboration H2 activation. We next tested the ability of the anthracene-based
1
regarding the bond metrics, ν(CO) features, and H and 31P NMR model to interact with or activate H2 (or D2). Incubation of the
properties of 1 and 2 may be found in the Supplementary Information. halide starting complex 1 under H2 (or D2) did not change any
NATURE CHEMISTRY | ADVANCE ONLINE PUBLICATION | www.nature.com/naturechemistry
© 2017 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.
3

NATURE CHEMISTRY DOI: 10.1038/NCHEM.2707
ARTICLES
a
b
21 d
D2
THF
20
6
Max intensity
in 2 h
N–D
3
0
5
10
15
20
days
1
8
6
4
2
0
8
6
4
2
0
ppm
ppm
Figure 5 | D2 gas activation by complex 3 evidenced by NH → ND exchange in complex 3. a, 2H NMR spectrum indicating the formation of [(Anth•CNDNS)
Fe(CO)2(THF)](BArF) (3D) from 3 in THF under 1 atm D2 at 25 °C over t ≈ 2 h. b, Timecourse of NH → ND exchange (1, 3, 6, 20 and 21 days) observed
under 1 atm of mixed H2/D2 (1:1) at 25 °C.
a
b
H2
BArF
J(HD): 43 Hz
THF
N
S
CO
CO
Fe
HN
C
3
O
O
4.6
4.5
4.4
4.3
+
D–H
D2
ppm
–
O
c
O
BArF
D2
+
D
dBPhOD
OH
d-THF
Fe–D
D
N
S
CO
CO
Fe
HN
C
5
0
–15
–20
4
O
ppm
Figure 6 | Deuteride (D−) transfer from D2 to H+ and formation of Fe–D species. a, Formation of 4 from 3 in THF under D2 atmosphere, and subsequent
iron-mediated deuteride transfer to a proton source, thereby generating HD. b, 1H NMR spectrum indicating the presence of H2 and HD as scrambling
products from D2/H+ as catalysed by 3; experimental conditions: 3, D2, dBPhOH in d8-THF. c, 2H NMR spectrum indicating the presence of the [Fe–D]
species, 4; experimental conditions: 3D, D2, dBPhOD in THF.
1
2
feature in the H (or H) NMR spectrum. However, incubation of indicated the activation of D2, and formation of the deuterated
the Tl-activated solvato species 3 under D2 (1 atm, 25 °C) led to a species [(Anth•CNDNS)Fe(CO)2(THF)]BArF (3D, 2H NMR
different result. After 2 h, a new resonance at 7.78 ppm was spectrum in Fig. 5a). To obtain tractable kinetics, another
2
observed in the H NMR spectrum of 3 in THF, corresponding to experiment was performed under a mixed H2/D2 atmosphere
the original, broad NH resonance at 7.70 ppm identified in 3. The (1:1 v/v, 1 atm), which slowed the H/D exchange but resulted in
2
analogous experiment performed in d8-THF (for the purpose of the same ND peak in H NMR spectrum (Fig. 5b). As a result,
1
1H NMR analysis) revealed that the corresponding 7.70 ppm H the kinetics of the NH → ND exchange were observable (Fig. 5b,
resonance was abolished under the D2 conditions. These results inset; t = 3.7 d). The solid state infrared spectrum of the resulting
½
4
NATURE CHEMISTRY | ADVANCE ONLINE PUBLICATION | www.nature.com/naturechemistry
© 2017 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.

NATURE CHEMISTRY DOI: 10.1038/NCHEM.2707
3D species exhibited nearly identical carbonyl stretches as 3H, give an Fe–(D2) adduct, which subsequently forms the Fe–D
ARTICLES
indicating no other change in composition of the complex. species. The iron-deuteride then reacts with the bulky dBPhOH to
This set of positive results motivated us to consider the importance generate HD. The cycle can continue: the in situ generated HD
of coordination geometry in the model complexes. In previous reacts with 3 to generate an Fe–H species, which—following reac-
work12, we reported the synthetic model complex [(CNHNS)Fe tion with dBPhOH—finally affords H2. At any time during this
(CO)2(Br)] with the same C,N,S donor set, but arranged instead in cycle, the protonated complex 3H can convert to 3D via D2 or HD
an equatorial fashion (Fig. 1b, right). As a result, the complex exhibits activation.
a solvent/substrate binding site located cis to the organometallic Fe–C
Next, we investigated hydride transfer from H2 to the cationic
(=O) unit, rather than trans in the present case. Intriguingly, under Im+ substrate mediated by the iron complex. Under a variety of
the same experimental conditions for C–H abstraction, the equatorial reducing atmospheres (H2, D2 or mixtures thereof), pressures
C,N,S complex did not exhibit any Im–H hydride abstraction (was (1–7 atm) and with different substrates [Im+, (perfluoro)quinones
activated similarly with [Tl]BArF). Additionally, the analogous H/D and (fluoro)ketones], no hydrogenation of organic substrates was
exchange reactivity was performed with the mer-C,N,S complex observed. This indicates that either (i) the Im+ substrate is not a
[(CNHNS)Fe(CO)2(THF)]BArF, and no H/D exchange was observed strong enough hydride acceptor to abstract the hydride from 4
as determined by 1H or 2H NMR spectroscopy. This specifically high- (as expected from the hydride abstraction from ImH), or that
lights the requirement for designing a stable, fac-C,N,S chelate to (ii) the present complex is more active for hydride abstraction,
enforce the solvent/H2 binding site trans to the organometallic Fe– and cannot self-modulate its Lewis acidity/basicity like the
C(=O) unit. This allows for more accurate modelling of the structure enzyme active site. The latter may be attributable to the lack of a
and function of the enzyme, thus warranting our designed approach pyridone moiety, the thiolate → thioether substitution, or the
with the anthracene scaffold.
lack of an H-bonding network in the second coordination
We next investigated the reverse reaction—‘substrate’ hydrogen- sphere. Unfortunately, stronger hydride acceptors like B(C6F5)3
ation—by using the simplest, analytically viable hydride or deuteride or trityl cation—that is, [CPh3]BF4 – reacted directly with the
acceptor: D+ or H+, respectively for crossover conversions. A reaction complex (likely via the carbamoyl NHC=O•••ZR3 interaction),
mixture of 3 in d8-THF and dBPhOH (10 equiv.) was prepared in a thus precluding related studies.
high-pressure NMR tube inside a glovebox (N2 atmosphere). The
In summary, we have implemented a novel scaffold approach
solution was stored under 7 atm of D2 gas (∼10 equiv.), and after 1 that utilizes an anthracene framework to display the requisite biomi-
day showed both H2 and HD resonances in the 1H NMR spectrum metic coordination sphere of [Fe]-hydrogenase in a chelated, facial
(Fig. 6b). A similar result was observed in the analogous yet isotopi- arrangement. As a result, we have demonstrated bidirectional
cally opposite combination of 3D, dBPhOD, and H2 (Supplementary activity including catalytic C–H hydride abstraction, and H2/D2
Fig. 9) by the corresponding detection of D2 gas in the 2H NMR spec- activation plus hydride (deuteride) transfer to D+/H+. This was
trum. No H/D scrambling was observed in the absence of the achieved in a mononuclear, biomimetic system with a strong resem-
hydride/deuteride acceptor, dBPhOH/D. Indeed, this is quite analo- blance to the Hmd-[Fe] active site—including a solvent/substrate
gousto the enzyme situation, wherein H/D exchange is observed only site trans to the organometallic Fe–C(=O) unit. Importantly, the
in the presence of substrate32.
analogous model complex that exhibits mer-C,N,S ligation and a
Both of the above conversions are predicated on the existence of an solvent/substrate binding site cis to the Fe–C(=O) σ bond (trans
Fe–H (or Fe–D) intermediate, which we attempted to detect next. site occupied by S) did not exhibit any of these reactivity patterns.
Focusing on the Fe–D intermediate, we repeated the above experiment Our continuing efforts to further emulate the Hmd reactivity
under all-deuterated conditions (without the intention for crossover): include incorporation of more biomimetic features of the active
3D, dBPhOD, and D2. Scanning of the extreme upfield region in the 2H site—such as the pyridone moiety and a thiolate donor—to demon-
NMR spectrum revealed a resonance at δ ≈ −18 ppm (Fig. 6c; full spec- strate the hydrogenation of model substrates related to the native
trum, Supplementary Fig. 10), indicating the presence of an observa- enzyme’s activity.
ble, Fe–D species. Interestingly, this same resonance was generated
under different conditions (for example, 3D or 3H; D2 or H2), and Methods
The title complex [(Anth•CNHNS)Fe(CO)2(I)] (1) was synthesized from [Fe
was observed as long as the H+/D+ source remained as dBPhOD
(Supplementary Fig. 11). Indeed, this indicates that the high relative
(CO)4I2] and the Anth•NH2Py•S ligand in dichloromethane. The PPh3-bound
derivative [(Anth•CNHNS)Fe(CO)2(I)(PPh3)] (2) was synthesized by reaction of 1
concentration of phenol (300 mM) versus dissolved H2/D2 (7 atm,
10 equiv., ∼2 mM dissolved gas) determines the final Fe–D
product. Indeed, substitution of dBPhOD for dBPhOH (using 3D
and D2) minimized the Fe–D resonance. However, both the Fe–D
and the Fe–H integration in NMR spectra indicated about 20% con-
version under 7 atm of H2 (Supplementary Fig. 12). We attribute this
low conversion to the small driving force of the forward reaction due
to the low concentration of dissolved H2 (∼2 mM of H2 under 7 atm
versus 30 mM of iron complex 3, and 300 mM of dBPhOH). An
applied pressure of ∼1 atm of H2 (only ∼0.3 mM H2 versus 30
mM complex 3) diminished the conversion to Fe–H to less than
2%. Thus, the conversion yield appears to be limited by the technical
constraints of the thick-walled NMR tube (max pressure, 150 psi).
Additionally, the in situ IR spectrum of the reaction mixture exhib-
ited carbonyl stretch frequencies at 2,035, 1,982 and 1,662 cm–1 con-
firming retention of the core coordination features in complex. Asthe
H/D exchange process is dynamic at ambient temperature, it is not
surprising that two discrete sets of resonances are not observed for
the Fe–THF and Fe–H species (Supplementary Fig. 12).
with PPh3 in dichloromethane. [(Anth•CNHNS)Fe(CO)2(THF)](BArF)] (3) was
prepared via halide abstraction from 1 with [Tl](BArF) in THF and used for the
following reactions in situ. The hydride abstraction from the imidazolidine
substrate was performed by reaction of a solution of 3 in THF with 1,3-bis(2,6-
difluorophenyl)-2-(p-tolyl)imidazolidine (Im–H) in a J. Young (J-Y) NMR tube,
and the resulting solution was analysed by 1H NMR spectroscopy and ESI–MS.
H2 evolution from Im–H and a bulky phenol substrate was carried out using 5
equiv. of Im–H and 5 equiv. of 2,6-di-tert-butyl-4-methoxyphenol in a J-Y NMR
tube, and the headspace gas was analysed by a gas chromatography. For NH →
ND exchange, a THF solution of 3 under 1 atm of D2 gas in a J-Y NMR tube was
monitored at various time points by 2H NMR spectroscopy. The deuteride (D–)
transfer experiments from D2 to H+ were performed by reacting 3 with 10 equiv.
of the bulky phenol under 7 atm of D2 gas in a high pressure NMR tube. The DFT
calculations were performed in Firefly using a PW91/6-31G* (except for iodine,
6-311G*) combination of functional and basis set.
More detailed synthetic procedures, experimental details, spectra, GC–MS
data, X-ray metrics and computational details are given in the
Supplementary Information.
Note: All experiments were performed under N2 atmosphere. Thallium salts are
extremely toxic and the reagents and resulting solid waste (thallium iodide) should
be handled with caution.
Data availability. The data that support the plots within this paper and other
findings of this study are available from the corresponding author upon reasonable
All of the above observations are reasonably explained in the
following way (Fig. 6a) starting from 3H: D2 is activated by 3 to request. X-ray crystallographic data for complex 2 are available from the Cambridge
NATURE CHEMISTRY | ADVANCE ONLINE PUBLICATION | www.nature.com/naturechemistry
© 2017 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.
5

NATURE CHEMISTRY DOI: 10.1038/NCHEM.2707
ARTICLES
Crystallographic Data Centre (CCDC 1512432). These data can be obtained free of
charge (http://www.ccdc.cam.ac.uk/data_request/cif).
19. Chen, D., Scopelliti, R. & Hu, X. Reversible protonation of a thiolate ligand in an
[Fe]-hydrogenase model complex. Angew. Chem. Int. Ed. 51, 1919–1921 (2012).
20. Xu, T. et al. A functional model of [Fe]-hydrogenase. J. Am. Chem. Soc. 138,
3270–3273 (2016).
21. Song, L.-C., Xie, Z.-J., Wang, M.-M., Zhao, G.-Y. & Song, H.-B. Biomimetic
models for the active site of [Fe] hydrogenase featuring an acylmethyl
(hydroxymethyl)pyridine ligand. Inorg. Chem. 51, 7466–7468 (2012).
22. Song, L.-C., Hu, F.-Q., Zhao, G.-Y., Zhang, J.-W. & Zhang, W.-W. Several new
[Fe] hydrogenase model complexes with a single Fe center ligated to an
acylmethyl(hydroxymethyl)pyridine or acylmethyl(hydroxy)pyridine ligand.
Organometallics 33, 6614–6622 (2014).
23. Turrell, P. J., Wright, J. A., Peck, J. N. T., Oganesyan, V. S. & Pickett, C. J. The
third hydrogenase: a ferracyclic carbamoyl with close structural analogy to the
active site of Hmd. Angew. Chem. Int. Ed. 49, 7508–7511 (2010).
24. Turrell, P. J., Hill, A. D., Ibrahim, S. K., Wright, J. A. & Pickett, C. J. Ferracyclic
carbamoyl complexes related to the active site of [Fe]-hydrogenase. Dalton
Trans. 42, 8140–8146 (2013).
25. Liu, T., DuBois, D. L. & Bullock, R. M. An iron complex with pendent amines as a
molecular electrocatalyst for oxidation of hydrogen. Nat. Chem. 5, 228–233 (2013).
26. Parkin, G. Synthetic analogues relevant to the structure and function of zinc
enzymes. Chem. Rev. 104, 699–768 (2004).
27. England, J. et al. A synthetic high-spin oxoiron(IV) complex generation, spectroscopic
characterization, and reactivity. Angew. Chem. Int. Ed. 48, 3622–3626 (2009).
28. Dias, A. R. et al. Li, Ti(III), and Ti(IV) trisamidotriazacyclononane complexes.
syntheses, reactivity, and structures. Inorg. Chem. 42, 2675–2682 (2003).
29. Park, Y. J., Ziller, J. W. & Borovik, A. S. The effects of redox-inactive metal ions
on the activation of dioxygen: isolation and characterization of a
heterobimetallic complex containing a MnIII–(μ-OH)–CaII core. J. Am. Chem.
Soc. 133, 9258–9261 (2011).
30. Marinescu, S. C., Winkler, J. R. & Gray, H. B. Molecular mechanisms of
cobalt-catalyzed hydrogen evolution. Proc. Natl Acad. Sci. USA 109,
15127–15131 (2012).
31. Kalz, K. F., Brinkmeier, A., Dechert, S., Mata, R. A. & Meyer, F. Functional model
for the [Fe] hydrogenase inspired by the frustrated lewis pair concept. J. Am.
Chem. Soc. 136, 16626–16634 (2014).
Received 5 April 2016; accepted 25 November 2016;
published online 23 January 2017
References
1. Zhou, J., Hu, Z., Münck, E. & Holm, R. H. The cuboidal Fe3S4 cluster: synthesis,
stability, and geometric and electronic structures in a non-protein environment.
J. Am. Chem. Soc. 118, 1966–1980 (1996).
2. Kanady, J. S., Tsui, E. Y., Day, M. W. & Agapie, T. A synthetic model of the
Mn3Ca subsite of the oxygen-evolving complex in photosystem II. Science 333,
733–736 (2011).
3. Wright, R. J., Lim, C. & Tilley, T. D. Diiron proton reduction catalysts possessing
electron-rich and electron-poor naphthalene-1,8-dithiolate ligands. Chem.
Eur. J. 15, 8518–8525 (2009).
4. Lee, Y. et al. Dinitrogen activation upon reduction of a triiron(II) complex.
Angew. Chem. Int. Ed. 54, 1499–1503 (2015).
5. Teramoto, Y., Kubo, K., Kume, S. & Mizuta, T. Formation of a hexacarbonyl
diiron complex having a naphthalene-1,8-bis(phenylphosphido) bridge and
the electrochemical behavior of its derivatives. Organometallics 32,
7014–7024 (2013).
6. Rauchfuss, T. B. Diiron azadithiolates as models for the [FeFe]-hydrogenase
active site and paradigm for the role of the second coordination sphere. Acc.
Chem. Res. 48, 2107–2116 (2015).
7. Shima, S. et al. The crystal structure of [Fe]-hydrogenase reveals the geometry of
the active site. Science 321, 572–575 (2008).
8. Hiromoto, T. et al. The crystal structure of C176A mutated [Fe]-hydrogenase
suggests an acyl-iron ligation in the active site iron complex. FEBS Lett. 583,
585–590 (2009).
9. Yang, X. & Hall, M. B. Trigger mechanism for the catalytic hydrogen activation
by monoiron (iron−sulfur cluster-free) hydrogenase. J. Am. Chem. Soc. 130,
14036–14037 (2008).
10. Yang, X. & Hall, M. B. Monoiron hydrogenase catalysis: hydrogen activation
with the formation of a dihydrogen, Fe−Hδ−···Hδ+−O, bond and methenyl-
H4MPT+ triggered hydride transfer. J. Am. Chem. Soc. 131, 10901–10908 (2009).
11. Gubler, J., Finkelmann, A. R. & Reiher, M. Theoretical 57Fe Mössbauer
spectroscopy for structure elucidation of [Fe] hydrogenase active site
intermediates. Inorg. Chem. 52, 14205–14215 (2013).
12. Finkelmann, A. R., Senn, H. M. & Reiher, M. Hydrogen-activation mechanism
of [Fe] hydrogenase revealed by multi-scale modeling. Chem. Sci. 5,
4474–4482 (2014).
13. Durgaprasad, G., Xie, Z.-L. & Rose, M. J. Iron hydride detection and
intramolecular hydride transfer in a synthetic model of mono-iron hydrogenase
with a CNS chelate. Inorg. Chem. 55, 386–389 (2016).
14. Li, B., Liu, T., Popescu, C. V., Bilko, A. & Darensbourg, M. Y. Synthesis and
Mössbauer characterization of octahedral iron(II) carbonyl complexes
FeI2(CO)3L and FeI2(CO)2L2: developing models of the [Fe]-H2ase active site.
Inorg. Chem. 48, 11283–11289 (2009).
15. Royer, A. M., Salomone-Stagni, M., Rauchfuss, T. B. & Meyer-Klaucke, W. Iron
acyl thiolato carbonyls: structural models for the active site of the [Fe]-
hydrogenase (Hmd). J. Am. Chem. Soc. 132, 16997–17003 (2010).
16. Liu, T. et al. Analysis of a pentacoordinate iron dicarbonyl as synthetic
analogue of the Hmd or mono-iron hydrogenase active site. Chem. Eur. J. 16,
3083–3089 (2010).
17. Chen, D., Scopelliti, R. & Hu, X. [Fe]-hydrogenase models featuring
acylmethylpyridinyl ligands. Angew. Chem. Int. Ed. 49, 7512–7515 (2010).
18. Chen, D., Scopelliti, R. & Hu, X. A five-coordinate iron center in the active site of
[Fe]-hydrogenase hints from a model study. Angew. Chem. Int. Ed. 50,
5671–5673 (2011).
32. Lyon, E. J. et al. UV-A/blue-light inactivation of the ‘metal-free’ hydrogenase
(Hmd) from methanogenic archaea. Eur. J. Biochem. 271, 195–204 (2004).
Acknowledgements
This research was supported by the Welch Foundation (F-1822) and the ACS Petroleum
Research Fund (53542-DN13), as well as the UT College of Natural Sciences. We also
thank V. Lynch for assistance with X-ray data analysis, and S. Sorey for assistance
with NMR study.
Author contributions
M.J.R. and J.S. designed the experiments. J.S. synthesized and characterized the model
complexes, and J.S. performed the experiments and analysed the data. M.J.R. and J.S.
carried out the DFT calculations. J.S. and T.A.M. synthesized ligands. M.J.R. and J.S.
co-wrote the paper.
Additional information
Supplementary information and chemical compound information are available in the
online version of the paper. Reprints and permissions information is available online at
www.nature.com/reprints. Correspondence and requests for materials should be addressed
to M.J.R.
Competing financial interests
The authors declare no competing financial interests.
6
NATURE CHEMISTRY | ADVANCE ONLINE PUBLICATION | www.nature.com/naturechemistry
© 2017 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.