Biomimetic peptide-based models of [FeFe]-hydrogenases: Utilization of phosphine-containing peptides

Article abstract of DOI:10.1039/c5dt01796c

Two synthetic strategies for incorporating diiron analogues of [FeFe]-hydrogenases into short peptides via phosphine functional groups are described. First, utilizing the amine side chain of lysine as an anchor, phosphine carboxylic acids can be coupled via amide formation to resin-bound peptides. Second, artificial, phosphine-containing amino acids can be directly incorporated into peptides via solution phase peptide synthesis. The second approach is demonstrated using three amino acids each with a different phosphine substituent (diphenyl, diisopropyl, and diethyl phosphine). In total, five distinct monophosphine-substituted, diiron model complexes were prepared by reaction of the phosphine-peptides with diiron hexacarbonyl precursors, either (μ-pdt)Fe₂(CO)₆ or (μ-bdt)Fe₂(CO)₆ (pdt = propane-1,3-dithiolate, bdt = benzene-1,2-dithiolate). Formation of the complexes was confirmed by UV/Vis, FTIR and ³¹P NMR spectroscopy. Electrocatalysis by these complexes is reported in the presence of acetic acid in mixed aqueous-organic solutions. Addition of water results in enhancement of the catalytic rates.

Full text of DOI:10.1039/c5dt01796c

Dalton
Transactions
View Article Online
View Journal
PAPER
Biomimetic peptide-based models of [FeFe]-
hydrogenases: utilization of phosphine-containing
Cite this: DOI: 10.1039/c5dt01796c
peptides†
Souvik Roy,‡ Thuy-Ai D. Nguyen,§ Lu Gan and Anne K. Jones*
Two synthetic strategies for incorporating diiron analogues of [FeFe]-hydrogenases into short peptides via
phosphine functional groups are described. First, utilizing the amine side chain of lysine as an anchor,
phosphine carboxylic acids can be coupled via amide formation to resin-bound peptides. Second, artifi-
cial, phosphine-containing amino acids can be directly incorporated into peptides via solution phase
peptide synthesis. The second approach is demonstrated using three amino acids each with a different
phosphine substituent (diphenyl, diisopropyl, and diethyl phosphine). In total, five distinct monophos-
phine-substituted, diiron model complexes were prepared by reaction of the phosphine-peptides with
diiron hexacarbonyl precursors, either (μ-pdt)Fe2(CO)6 or (μ-bdt)Fe2(CO)6 (pdt = propane-1,3-dithiolate,
bdt = benzene-1,2-dithiolate). Formation of the complexes was confirmed by UV/Vis, FTIR and 31P NMR
spectroscopy. Electrocatalysis by these complexes is reported in the presence of acetic acid in mixed
aqueous-organic solutions. Addition of water results in enhancement of the catalytic rates.
Received 13th May 2015,
Accepted 23rd July 2015
DOI: 10.1039/c5dt01796c
www.rsc.org/dalton
Introduction
Hydrogenases, the biological catalysts for the reversible
reduction of protons to molecular hydrogen, employ earth-
abundant base metals, either iron or nickel and iron, to carry
out this transformation under mild conditions.1 The utility of
this reaction for developing technologies to produce sustain-
able solar fuels has engendered widespread interest in [FeFe]-
hydrogenases, the most efficient biological system for produ-
cing hydrogen. X-ray crystallographic studies revealed that the
structure of the active site of [FeFe]-hydrogenase, known as the
H-cluster, consists of a [2Fe] subsite covalently connected to a
[4Fe4S] cubane.2 The H-cluster is connected to the protein via
only a single cysteinyl thiolate, which also bridges the [4Fe4S]
cubane and the [2Fe] subsite (Scheme 1A). The [2Fe] subsite
features CO and CN− as terminal diatomic ligands and a bio-
logically unusual non-proteinaceous dithiolate ligand bridges
the two iron centers.
Scheme 1 (A) [FeFe]-hydrogenase active site (H-cluster) and (B) model
representation of peptide-bound diiron complexes described in this
study.
(pdt = propane-1,3-dithiolate) which serves as a convenient
starting point for building a multitude of sophisticated bio-
mimetic diiron analogues.3 Although many of these bio-inspired
organometallic models show moderate electrocatalytic proton
reduction activity, none of them catalyzes hydrogen production
with the same exquisite combination of high turnover fre-
quency and low activation energy as the enzymes.4 Moreover,
the hydrophobic nature of these diiron dithiolate clusters
renders most of them soluble only in organic solvents. They
are therefore unsuitable for applications in fuel cells employ-
ing aqueous solutions, and their activity in such a setting
cannot be evaluated.
From a structural perspective, the [2Fe] subunit is remini-
scent of the well-known organoiron complex [(μ-pdt)Fe2(CO)6]
Department of Chemistry and Biochemistry, Arizona State University, Tempe,
AZ 85287, USA. E-mail: jonesak@asu.edu; Tel: +1-480-965-0356
†Electronic supplementary information (ESI) available. See DOI: 10.1039/
c5dt01796c
‡Current Address: Laboratory of Chemistry and Biology of Metals, CEA-
Grenoble, 17 Rue des Martyrs, 38054 Grenoble Cedex-9, France.
§Current Address: Department of Chemistry and Biochemistry, University of
California Santa Barbara, Santa Barbara, CA 93106, USA.
Recently, the importance of the protein environment for
the catalytic activity of hydrogenases and other metallo-
enzymes has become increasingly clear.5 For example, the
This journal is © The Royal Society of Chemistry 2015
Dalton Trans.

View Article Online
Paper
Dalton Transactions
hydrogenase protein component has been shown to contribute on the hydrophilicity, redox properties and electrocatalytic
to stabilization of the diiron site in a strained “rotated” structure activities of coordinated diiron carbonyl complexes were
that leaves one of the irons with an open coordination site explored in acetonitrile and acetonitrile/water mixtures.
thought to be essential for catalysis.6 Additionally, starting Notably, the complexes display favorable energetics towards
from the apo-enzyme and a model diiron complex with vir- electrocatalytic proton reduction from acetic acid in the pres-
tually no catalytic activity, complete reconstitution of fully ence of water.
functional [FeFe]-hydrogenases has been achieved.5b This
clearly demonstrates the importance of the protein environ-
ment for facilitating efficient, bidirectional catalysis and has
spurred efforts to incorporate bio-inspired catalysts and hydro-
Results and discussion
genase models into supramolecular constructs to investigate
the impact of secondary coordination sphere interactions
on chemical properties and produce synthetic catalysts with
greater activity.7
Synthesis and characterization of the metallopeptides
We present two strategies to construct phosphine-functiona-
lized artificial peptides. The first is analogous to our previous
method for creating dithiol functionalized peptides. As shown
in Scheme 2, using the amine side chain of a lysine as an
anchor, a resin-bound peptide is modified via amide for-
mation with a phosphine-containing, aliphatic carboxylic acid
creating a phosphine-bearing lysine derivative. In the second
approach, several phosphine-containing amino acids have
been directly incorporated via solution-phase peptide syn-
thesis. After synthesis, the phosphine-functionalized peptide
can be reacted with diironhexacarbonyl complexes [(μ-SRS)-
Fe2(CO)6] (R = organic group) in the presence of a decarbony-
lating agent (trimethylamine-N-oxide) resulting in substitution
of one of the CO ligands by the phosphine.
Designed protein and peptide models that bind natural or
artificial metallocofactors are proving to be
a powerful
approach both for tuning the reactivity of embedded metallo-
centers and for probing the mechanisms of the metallo-
enzymes that have inspired them.8 Small artificial peptides
binding diiron hydrogenase models serve as bridges between
natural metalloenzymes and organometallic analogues and
present an opportunity to introduce secondary coordination
sphere interactions into the models. Synthetic methods have
been developed to incorporate thiolate bridged diiron hexacarb-
onyl clusters into suitably designed peptides via both natural
cysteine and artificial dithiol groups.9 However, these hexa-
carbonyl [FeFe]-hydrogenase models are less electron rich and
thus poorer catalysts than pentacarbonyl complexes in which a
terminal CO ligand is replaced by a stronger donor like a phos-
phine. Furthermore, substitution of one of the carbonyls in
these peptide-based models is challenging because low yields
result from the awkward solvent mixtures that are required to
simultaneously solubilize both the peptide and the substitut-
ing ligand.
On-resin modification of peptide. Scheme 2 outlines the
approach for attaching a phosphine to a resin-bound peptide,
Phosphines have been widely employed in models as a sur-
rogate for the CN− ligand found in hydrogenases,10 and con-
struction of phosphine-functionalized peptides represents a
tantalizing route to peptide-based diiron models with better
electrocatalytic properties. Moreover, phosphino-peptides offer
the possibility of synthesizing a wide range of ligands through
variation of the peptide sequence as well as the substituents
on the phosphines. In this report, we describe two general
methods for introducing a phosphine functionality into small
peptides. The first approach is to tether the phosphine to the
side chain amine of a lysine within a resin-bound peptide. The
second method is to incorporate a protected artificial phos-
phine amino acid directly into the peptide during solution
phase peptide synthesis. To demonstrate the efficiency of the
second synthetic approach, three amino acids were prepared
with distinct substituents on the phosphines using the
Scheme 2 Synthetic scheme for on-resin modification of the lysine
residue of an eight amino acid peptide to incorporate a phosphine func-
tional group and subsequent reaction with diiron-hexacarbonyl to
produce the [(μ-pdt){Fe(CO)3{peptide-Fe(CO)2}] complex (1). The pro-
tecting groups on all the amino acid residues except lysine remain intact
method developed by Gilbertson and co-workers: N-Boc-3-(di- until the peptide is cleaved from the resin. The overall yield of 1, starting
from the peptide synthesis, was approximately 5%. The sphere rep-
ethylphosphorothioyl)alanine
(Boc-Epa-OH),
N-Boc-3-(di-
resents the resin bead. Reaction conditions: (a) 2% N2H4, DMF; (b)
Ph2PC2H4COOH, HATU, DIEA, DMF; (c) Ph2P(S)C2H4COOH, HATU, DIEA,
DMF; (d) 95% TFA, 2.5% water, 2.5% TIPS; (e) (i) 95% TFA, 2.5% water,
2.5% TIPS, (ii) Raney™ nickel, aq. CH3CN; (f) (μ-pdt)Fe2(CO)6, Me3NO,
isopropylphosphorothioyl)alanine (Boc-Ipa-OH), and N-Boc-
3-(diphenylphosphorothioyl)alanine (Boc-Ppa-OH).11 Four
diiron-peptide complexes were synthesized using these phos-
phine amino acids, and the influence of the peptide ligands aq. CH3CN.
Dalton Trans.
This journal is © The Royal Society of Chemistry 2015

View Article Online
Dalton Transactions
Paper
and the subsequent incorporation of a diiron cluster. The
peptide utilized, WASKLPSG, is derived from the N-terminal
sequence of nickel-superoxide dismutase from Streptomyces
coelicolor.12 Previously, we employed this sequence to create an
artificial 1,3-dithiol peptide coordinating a diiron-hexacarbo-
nyl cluster.9c To synthesize the phosphine peptide, the desired
eight residue peptide was first synthesized via solid-phase
peptide synthesis using the Fmoc/t-Bu strategy. Lysine with an
orthogonally protected amine group was selectively modified
with diphenylphosphinopropionic acid to generate the iron-
binding site. The side chain amine of tryptophan and the
hydroxyl groups of the serines remain protected during the on-
resin modification steps. Cleavage of the phosphine–peptide
WASK(PPh2)LPSG (1a) from the resin and reaction with
[(μ-pdt)Fe2(CO)6; pdt = propane-1,3-dithiolate] produced the
desired metallopeptide (μ-pdt)[Fe-(CO)3][Fe(CO)2WASK(PPh2)-
LPSG] (1). However, moderate sensitivity of the phosphines to
oxidation reduced the overall yield. To prevent the loss of phos-
phine-peptide during synthesis and purification, an approach
in which the free phosphine was replaced by a sulfur-protected
phosphine synthon {3-(diphenylphosphorothioyl) propanoic
acid} was evaluated (Scheme 2). After peptide cleavage from
the resin, the phosphine sulfide was reduced by Raney™
Scheme 3 Synthetic scheme for preparing diiron-tripeptide complexes
with artificial N-Boc phosphine sulfide amino acids. Reaction conditions:
(a) (i) H-Leu-OMe·HCl, PyBOP, DIEA, CH2Cl2, (ii) Ac-Val-OH, PyBOP,
DIEA, CH2Cl2; (b) Raney™ nickel, CH3CN, heat; (c) (μ-SR1S)Fe2(CO)6,
Me3NO, CH3CN.
nickel to produce the free phosphine-peptide which was (Boc-Ppa-OH). Each was embedded into the tripeptide
immediately used, without further purification, for the reac- sequence, Val-Xpa-Leu (Xpa = unnatural phosphine amino
tion with (μ-pdt)Fe2(CO)6. Unfortunately, the use of sulfur- acid) via solution phase peptide synthesis generating Val-Ppa-
protected phosphine did not significantly improve the overall Leu, Val-Ipa-Leu, and Val-Epa-Leu (Scheme 3). Valine and
yield due to inefficiency of the desulfurization.
leucine were chosen for hydrophobicity, to facilitate large-scale
The metallopeptide was purified via HPLC, and a range of purification of the tripeptides on a normal-phase silica
analytical techniques confirm the presence of the desired column. The peptides’ N- and C-termini were protected as
cluster (Fig. S1–S4†). Most importantly, the IR spectrum of the acetyl and methyl esters, respectively. Treatment of purified tri-
complex contains three characteristic bands at 2044, 1981, peptides with Raney™ nickel generated the free phosphines.
and 1925 cm−1 corresponding to the C–O stretching modes, Notably, 31P NMR revealed that desulfurization also caused
similar to those observed for analogous monosubstituted racemization at the α-carbon of the phosphine amino acid
diiron models.4b,13 Furthermore, the 31P NMR spectrum of the resulting in diasteromeric peptides. The phosphine peptides
complex has a single resonance at 55.4 ppm, confirming were metalated by reaction with (μ-pdt)Fe2(CO)6 or (μ-bdt)-
formation of a unique complex with an Fe–P bond. Finally, Fe2(CO)6 (bdt = benzene-1,2-dithiolate) in CH3CN/CH2Cl2 pro-
ESI-MS provides further confirmation of the presence of an ducing phosphine-substituted, diiron-peptide complexes as
Fe2(CO)5 cluster.
red solids in good yields. Four distinct complexes were pre-
Metallopeptides with unnatural amino acids. The necessity pared. Three complexes contain pdt as bridging ligand with
of including an orthogonally protected amino acid in the variation of the phosphine substituents: [(Val-Epa-Leu)-
peptide sequence and carefully designing selective protection/ {(μ-pdt)Fe2(CO)5}] (2), [(Val-Ipa-Leu)-{(μ-pdt)Fe2(CO)5}] (3), and
deprotection steps limits the range of phosphines that can be [(Val-Ppa-Leu)-{(μ-pdt)Fe2(CO)5}] (4). A fourth complex, [(Val-
employed using the synthetic scheme described above. There- Ppa-Leu)-{(μ-bdt)Fe2(CO)5}] (5), was synthesized using Val-Ppa-
fore, a second synthetic route in which an artificial phosphine Leu as the ligand and bdt bridged diiron hexacarbonyl
amino acid is directly incorporated during solid-phase peptide [(μ-bdt)Fe2(CO)6] as the metallo-precursor.
synthesis was developed. This second method has the added
Despite numerous attempts, X-ray quality crystals were not
benefit of shortening the side chain of the phosphine bearing obtained for the diiron-peptide complexes. Thus, the com-
amino acid. Since fewer rotamers are available, a more pounds were characterized via spectroscopic and electrochemi-
compact metal binding site with decreased solvent accessibil- cal methods. Spectroscopic data for 1–5 are summarized in
ity can be created. To demonstrate the utility of this strategy, Table 1. As shown in Fig. 1, the UV-vis spectra of the com-
using the method developed by Gilbertson et al.,11b,14 three plexes in acetonitrile feature an Fe–S charge transfer band (CT)
phosphine amino acids with different phosphine substituents in the 340–370 nm region. Modifying the phosphine substitu-
were synthesized: N-Boc-3-(diethylphosphorothioyl)alanine ents but keeping the bridging ligand unchanged causes only a
(Boc-Epa-OH),
N-Boc-3-(diisopropylphosphorothioyl)alanine subtle shift in the wavelength of the charge transfer band. On
(Boc-Ipa-OH), and N-Boc-3-(diphenylphosphorothioyl)alanine the other hand, comparing the UV-vis spectra of 4 and 5 shows
This journal is © The Royal Society of Chemistry 2015
Dalton Trans.

View Article Online
Paper
Dalton Transactions
Table 1 Wavelengths of the Fe–S charge transfer band (visible) and IR
stretching frequencies in the ν(CO) region data for complexes 1–5
λmax (nm)
Complex
(ε, M−1 cm−1
)
νCOb (cm−1
)
1a
2
3
4
5
350
2044 (s), 1981 (s), 1958 (sh), 1925 (w)
2040 (s), 1981 (vs), 1966 (sh), 1921 (w)
2043 (s), 1979 (vs), 1960 (sh), 1919 (w)
2046 (s), 1984 (s), 1961 (sh), 1928 (w)
2053 (s), 1994 (s), 1937 (w)
342 (9300)
351 (13 250)
354 (12 050)
365 (25 700)
a Extinction coefficient was not determined. b Intensity of the peaks:
s = strong, vs = very strong, sh = shoulder, w = weak.
Fig. 2 Cyclic voltammograms of (a) 2 (0.88 mM), (b) 3 (0.69 mM), (c) 4
(1.39 mM), and (d) 5 (1 mM) in acetonitrile with 0.1 M [NBu4][PF6] at a
scan rate 0.2
direction.
V . Arrows mark the starting potential and scan
s−1
Fig. 1 UV-vis spectra of 2 (black), 3 (green), 4 (blue), and 5 (red) in
acetonitrile. Spectra were obtained from solutions of approximately
0.1 mM complex.
compound is potentially oxidizable, and this could explain the
presence of several small oxidative waves in the oxidative scan
following reduction. On the other hand, cyclic voltammograms
of 5 show an irreversible oxidation wave at 0.47 V and a par-
tially reversible reduction wave (iopx/ipred = 0.47) at −1.62 V. As
shown in Fig. 2, 5 also undergoes a second oxidation at
ca. −1.29 V. This feature does not appear when the reductive
scan is stopped prior to the reduction event, again suggesting
an EC process in which the reduced species undergoes a
chemical reaction forming a new species. Based on previous
studies, we hypothesize that the chemical step following the
reduction is probably slow exchange of a CO ligand with
solvent to produce a phosphine/MeCN complex.10b,15 The electro-
chemical behavior of 5 is very similar to that of analogous
monosubstituted diiron-bdt-phosphine complexes recently
reported by Pandey et al.16 The reductive and oxidative peak
potentials (Erped and Eopx, respectively) of the complexes corre-
late well with the donating abilities of the phosphines and the
thiolates as deduced from their CO stretching frequencies.
Among the pdt bridged complexes, the diphenylphosphino
complex 4 is the most easily reduced. Changing the bridging
thiolate from pdt to bdt has a more pronounced influence on
the electronic properties of the diiron core as demonstrated by
the 200 mV and 160 mV anodic shift of the FeIFeI/FeIFe0 and
FeIFeII/FeIFeI couples, respectively, in 5 compared to 4.
that replacing pdt by bdt results in significant change in the
position and intensity of the CT transition (Table 1).
FTIR spectra of the complexes show a characteristic three-
band pattern in the C–O stretching region originating from the
monosubstituted phosphine–Fe2(CO)5. The stretching frequen-
cies of the carbonyls are a proxy for the electron density on the
iron centers. As shown in Table 1, replacing aromatic groups
on the phosphine with more electron donating alkyl groups
(isopropyl or ethyl) results in a shift of the CO stretching fre-
quencies by 5–6 cm−1 to lower wavenumbers. The weaker
donor strength of bdt compared to pdt is also reflected in the
higher C–O stretching frequencies observed for 5.
Electrochemical studies in acetonitrile
Cyclic voltammetry demonstrates that the electrochemical pro-
perties of these complexes are similar to those reported for
analogous monosubstituted diiron compounds. Cyclic voltam-
mograms of the pdt complexes, 2, 3 and 4, feature an irrevers-
ible wave corresponding to FeIFeI/FeIFe0 reduction and an
irreversible wave corresponding to FeIFeI/FeIFeII oxidation
(Fig. 2 and Table 2). The irreversibility of the reduction
suggests it is followed by a chemical reaction or, more specifi-
cally, decomposition of the complex. The reductively formed
The ability of complexes 2–5 to electrocatalyze proton
reduction from acetic acid was investigated in acetonitrile.
Dalton Trans.
This journal is © The Royal Society of Chemistry 2015

View Article Online
Dalton Transactions
Paper
Table 2 Reduction potentials for 2–5 in acetonitrile and acetonitrile/water mixed solvents
Epc (FeIFeI/FeIFe0), V
Epa (FeIIFeI/FeIFeI), V
Complex
CH3CN
3 : 1 CH3CN/H2O
3 : 2 CH3CN/H2O
CH3CN
3 : 1 CH3CN/H2O
3 : 2 CH3CN/H2O
2
3
4
5
−1.93
−1.90
−1.82
−1.62
NDa
−1.78
−1.77
−1.51
ND
+0.22
+0.26
+0.31
+0.47
ND
ND
−1.74
−1.72
−1.47
+0.22
+0.29
+0.48
+0.22
+0.32
+0.46
a ND = not determined.
catalytic current is not observed at the FeIFeI/FeIFe0 reduction
potential of the complexes (ca. −1.8 to −1.9 V). Instead, cata-
lytic current is observed at more negative potentials in the
range −2.1 to −2.3 V. Only a small enhancement of the peak
current at the FeIFeI/FeIFe0 reduction potential is observed.
This observation is consistent with the ECEC mechanism pro-
posed in previous electrochemical studies of less electron rich
analogues.17 A likely explanation is that oxidative addition
of a proton to the reduced FeIFe0 species forms the hydride
FeIFeII-H, prompting subsequent displacement of CO by
MeCN.10b The solvent-coordinated species is then reduced at
more negative potential, leading to a negative shift in the
potential at which catalytic current is observed (Fig. S5†). Com-
pared to the pdt complexes, the bdt analogue, 5, behaves
slightly differently in the presence of acid. Addition of acetic
acid renders the reduction peak of 5 completely irreversible
with a small increase in peak current, but catalysis is observed
only at a more negative potential, ca. −2.08 V. A similar ECEC
mechanism is probably operative in which 5 undergoes
reduction followed by protonation to form 5H. The protonated
species 5H is then reduced at −2.08 V generating the electron
rich 5H− which is basic enough to accept another proton. The
catalytic cycle is then completed following reductive elimin-
ation of hydrogen. In light of possible concerns about hetero-
geneous catalysis, we performed a rinse test to confirm that no
catalytically active species is deposited on the electrode. The
result of this test is consistent with homogeneous catalysis
(Fig. S6†).
Current enhancement observed in the presence of acetic
acid is an indicator of the ability of the catalysts, and the ratio
of the catalytic current (icat) to reductive peak current in the
absence of acid (ip) can be used as a marker to compare the
performances of different complexes.4a A higher (icat/ip) value
indicates faster catalysis. To employ this analysis we deter-
mined peak currents (ip) from the irreversible reduction waves
of 2–5 in the absence of acid. Interestingly, at the highest
acetic acid concentration (50 mM) investigated, 5 shows the
largest current enhancement (icat/ip = 23), 4 shows the least
Fig. 3 Cyclic voltammograms of (a) 2 (0.88 mM), (b) 3 (0.69 mM), (c) 4
(1.39 mM), and (d) 5 (1 mM) in acetonitrile with various concentrations of
acetic acid. The acid concentrations used are 5, 10, 20, 30, 40, and
50 mM. Reductive current increases with increasing acid concentration.
Other experimental conditions are as described in Fig. 2.
To ensure that any observed electrocatalytic response was not a
result of proton reduction by the glassy carbon electrode,
control experiments without added catalyst were undertaken
(Fig. S7 and S8†). For the complexes with a pdt bridge, sequen-
tial addition of AcOH resulted in increased reductive current
corresponding to production of hydrogen (Fig. 3). Notably, this
(icat/ip = 6), and 2 and 3 show intermediate increases (icat/ip
=
18.5). However, relative to 4 and 5, electrocatalysis by 2 and 3
occurs at more negative potentials [catalytic peak potential,
Ecpat = −2.30 V (2 and 3), −2.17 V (4 and 5)] due to the more
electron-rich alkyl-phosphine substituents. The overpotentials
This journal is © The Royal Society of Chemistry 2015
Dalton Trans.

View Article Online
Paper
Dalton Transactions
Fig. 4 Cyclic voltammograms of (a) 3, (b) 4, and (c) 5 in 3 : 2 acetonitrile-water (top) and 3 : 1 acetonitrile-water (bottom) at a potential scan rate of
0.2 V s−1. In acetonitrile, 0.1 M [NBu4][PF6] is used as supporting electrolyte. Similarly, 0.1 M KCl is used as the supporting electrolyte in water. The
solvent mixtures include both electrolytes at the concentration formed by mixing. Concentrations of the complexes: (i) 3 : 1 CH3CN/H2O: 3
(0.63 mM), 4 (1.06 mM), and 5 (0.86 mM); (ii) 3 : 2 CH3CN/H2O: 3 (0.64 mM), 4 (1.25 mM), and 5 (0.72 mM).
for proton reduction by the four complexes, defined as the
difference between the theoretical half-wave potential for
reduction of the acid and the experimental half-wave potential
for the catalytic process, were determined using the method
reported by Artero and co-workers.18 At the highest concen-
tration of acetic acid (50 mM), all four complexes require rela-
tively high overpotential for proton reduction (0.75–0.9 V). The
overpotential requirement for the diphenylphosphino com-
plexes (0.75 V for 4 and 5) is slightly lower than that for the
dialkylphosphino analogs (0.85 and 0.90 V for 2 and 3,
respectively).
Electrochemistry in acetonitrile/water mixtures
The improved water solubility of the diiron-peptide complexes
relative to analogous diiron-phosphine complexes previously
reported allowed electrocatalytic experiments in mixed aceto-
nitrile-water solvents. Fig. 4 shows cyclic voltammograms from
complexes 3, 4, and 5 in 3 : 1 and 3 : 2 CH3CN/H2O. The
reduction potentials determined from these experiments are
summarized in Table 2. Since the diethylphosphino complex
(2) and the diisopropylphosphino complex (3) have similar
electrochemical properties in acetonitrile, only 3 was studied
in mixed solvents. The reductive wave associated with each of
the FeIFeI/FeIFe0 couples shifts to less reducing potentials
with increasing water content: shifts of 160, 100 and 150 mV
for 3, 4, and 5, respectively. Similar shifts were observed when
electrochemical measurements of analogous diiron complexes
with hydrophilic phosphine ligands, included to improve
Fig. 5 Cyclic voltammograms of (a) 3 (0.63 mM), (b) 4 (1.06 mM), and
(c) 5 (0.86 mM) in 3 : 1 acetonitrile-water with various concentrations of
water solubility, were performed in mixed acetonitrile/
acetic acid. The acid concentrations used are 5, 10, 20, 30, 40, and
water solvents.15a,19 However, the FeIIFeI/FeIFeI oxidation 50 mM. Other experimental conditions are as described in Fig. 4.
was virtually unaffected by water. This suggests that the
reduction may be coupled to interaction of the reduced species
control experiments without catalyst included clearly demon-
strate that the electrocatalytic currents observed in the pres-
ence of catalyst are considerably higher than those from direct
with solvent.
Fig. 5 and 6 show electrocatalytic reduction of protons from
AcOH by complexes 3–5 in CH3CN/H2O mixtures. Negative
Dalton Trans.
This journal is © The Royal Society of Chemistry 2015

View Article Online
Dalton Transactions
Paper
Fig. 7 Dependence of the (icat/ip) on the amount of water present in
■
◆
▲
the solvent: 3 ( ), 4 ( ), 5 ( ). The TOFs are calculated from data shown
in Fig. 3, 5 and 6.
ence of water, Fig. 7 shows the dependence of the catalytic
(icat/ip) of 3, 4, and 5 in 50 mM acetic acid on the amount of
water present. This shows that addition of water results in sig-
nificant improvement in the catalytic activities of all three
complexes. Similar improvement upon addition of water has
been reported for Dubois-type nickel catalysts.20 At such high
water concentrations, the dielectric constant of the solution is
significantly increased, and the acidity of the solution is also
impacted. Either of these changes could lead to the enhanced
catalytic current. Alternatively, faster proton transfer to the
iron center facilitated by water could be another contributing
factor. Additional experiments using different solvents and
acids will be necessary to distinguish these possibilities.
Finally, it should be noted that in the case of 5, increasing the
amount of water from 25% to 40% led to a slight drop in (icat/ip).
It is possible there is an optimal water concentration for cataly-
sis by this complex, but the reason remains unclear.
Fig. 6 Cyclic voltammograms of (a) 3 (0.64 mM), (b) 4 (1.25 mM), and
(c) 5 (0.72 mM) in 3 : 2 acetonitrile-water with various concentrations of
acetic acid. The acid concentrations used are 5, 10, 20, 30, 40, and
50 mM. Other experimental conditions are as described in Fig. 4.
reduction of protons by the electrode surface (Fig. S7 and S8†).
For complexes 3 and 4, there are notable changes in the inten-
sities and positions of the catalytic waves when mixed aqueous
solvents are used. Unlike the electrocatalysis in acetonitrile,
cyclic voltammograms of 3 and 4 in CH3CN/H2O mixtures
show catalysis at a potential very close to that of FeIFeI/FeIFe0
reduction. This is consistent with the hypothesis above that
the catalytic waves in acetonitrile are associated with reduction
of solvent-coordinated species. The presence of water likely sig-
nificantly shifts the reduction potential of this solvent co-
ordinated species (the second electrochemical step of the
ECEC process) leading to a shift of the observed catalysis to
more positive potentials. Furthermore, in 40% water, increas-
ing the concentration of AcOH from 10 mM to 50 mM resulted
in 19-fold and 21-fold enhancement of peak current for com-
plexes 3 and 4, respectively. In contrast, only 9-fold or 7-fold
current increase was observed in neat acetonitrile. On the
other hand, although the onset of catalysis by 5− also shifts to
less reducing potentials with the addition of water, it is still
significantly more negative than the 5/5− reduction. To quan-
tify the trends of electrocatalytic activities of 3–5 in the pres-
Conclusions
In summary, we have demonstrated two synthetic schemes to
introduce phosphines into peptides and to covalently anchor a
(μ-SRS){Fe(CO)3}{Fe(CO)2(PR2R′)} cluster through the phos-
phine. This approach offers major advantages over other
methods reported for building peptide-based diiron com-
plexes: (1) the phosphine amino acids can be directly incorpor-
ated into designated peptide sequences; (2) electron-rich,
stable diiron complexes can be prepared in only one step with
peptide as the ligand resulting in better yields than syntheses
requiring two peptide modifying steps; and (3) the synthesis of
a diverse array of complexes can be achieved through variation
of the artificial amino acid. Importantly, utilization of peptide
ligands considerably increases the polarity of the complexes,
allowing electrochemical studies in partially aqueous solvents.
The presence of water has a significant influence on the
electrochemical properties of the complexes as demonstrated
by the shift of the reduction potential of the FeIFeI/FeIFe0
couple towards less reducing potentials and increased catalytic
This journal is © The Royal Society of Chemistry 2015
Dalton Trans.

View Article Online
Paper
Dalton Transactions
currents. Thus the synthetic approach described here provides 1H NMR (400 MHz, CDCl3): δ = 7.83 (m, 4H, Ph), 7.5 (m, 6H,
convenient access to monosubstituted diiron-peptide model Ph), 2.73 (m, 4H, P–(CH2)2–COOH); 31P{1H} NMR (161.9 MHz,
complexes and opens the door for the design of more sophisti- CDCl3): δ = 41.9.
cated peptides.
N-Boc-3-(diethylphosphorothioyl)-alanine methyl ester
(Boc-Epa-OMe)
Experimental section
N-Boc phosphine sulfide methyl ester was synthesized from
N-Boc-3-iodo-alanine methyl ester following the method
All reactions were carried out under an atmosphere of dry
previously described.14b Yield 63%. 1H NMR (400 MHz,
nitrogen using standard Schlenk and vacuum line techniques
CDCl3): δ = 5.68 (d, 1H, NH), 4.55–4.64 (m, 1H, N–CH–COO),
3.76 (s, 3H, COOCH3), 2.52–2.64 (m, 2H, N–CH–CH2–P), 2.02
unless otherwise mentioned. Anhydrous solvents were
purchased from Sigma-Aldrich, deuterated solvents from
Cambridge Isotope Laboratories, and Fmoc protected amino
(m, 4H, P–CH2–CH3), 1.42 (s, 9H, NH-Boc), 1.14–1.24 (m, 6H,
P–CH2–CH3); 31P{1H} NMR (161.9 MHz, CDCl3): δ = 52.8. Rf =
acids and peptide coupling reagents from Protein Techno-
0.3 (35% EtOAc/hexane).
logies. The compounds 3-(diphenylphosphino)-propionic
acid,21 N-Boc-3-iodo-alanine methyl ester,22 N-Boc-3-(diphenyl-
N-Boc-3-(diisopropylphosphorothioyl)-alanine methyl ester
(Boc-Ipa-OMe)
phosphorothioyl)-alanine (Boc-Ppa-OH),14b (μ-pdt)Fe2(CO)6,23
24
and (μ-bdt)Fe2(CO)6 were synthesized using literature pro-
This compound was prepared using the same strategy
cedures. The other N-Boc protected phosphine-sulfide amino
acids (Boc-Epa-OH and Boc-Ipa-OH) were synthesized by slight
modification of the literature method. All other starting
1
employed for Boc-Epa-OMe.14b Yield 69%. H NMR (400 MHz,
CDCl3): δ = 5.86 (d, 1H, NH), 4.61–4.52 (m, 1H, N–CH–COO),
3.75 (s, 3H, COOCH3), 2.38–2.46 (m, 2H, N–CH–CH2–P), 2.13
materials were commercially available and used as received.
(m, 2H, P-CHMe2), 1.43 (s, 9H, NH-Boc), 1.15–1.24 (m, 12H,
1H, 13C and 31P NMR spectra were recorded at room temp-
P-CHMe2); 31P{1H} NMR (161.9 MHz, CDCl3): δ = 64.7. Rf = 0.45
erature on a Varian NMR spectrometer (400 or 500 MHz for
(40% EtOAc/hexane).
1H). NMR chemical shifts are quoted in ppm; spectra are refer-
1
enced to tetramethylsilane for H and 13C NMR. Splitting pat-
N-Boc-3-(diethylphosphorothioyl)-alanine (Boc-Epa-OH)
terns are designated as follows: s, singlet; d, doublet; t, triplet;
m, multiplet; br, broad singlet; dd, doublet of doublets; td,
triplet of doublets. FTIR spectra were recorded on a Bruker
Vertex 70 spectrophotometer with MCT detector using a sealed
liquid spectrophotometer cell with CaF2 windows. UV-vis
measurements were performed on a Hewlett-Packard 8453
spectrophotometer using quartz cuvettes with a 1 cm path-
length. MALDI-MS (matrix assisted laser desorption/ionization
The amino acid was synthesized by the hydrolysis of
the methyl ester following the literature procedure.14b Yield
82%. 1H NMR (400 MHz, CDCl3): δ = 9.54 (br, 1H, COOH), 5.87
(d, 1H, NH), 4.58–4.64 (m, 1H, N–CH–COO), 2.51 (m, 1H,
N-CH-CHH′-P), 2.39 (m, 1H, N-CH-CHH′-P), 1.84–1.99 (m, 4H,
P–CH2–CH3), 1.43 (s, 9H, NH-Boc), 1.13–1.22 (m, 6H, P–CH2–
CH3); 13C{1H} NMR (100 MHz, CDCl3): δ = 174.6 (s), 155.6 (s),
80.8 (s), 50 (s), 29.9 (d, JC–P = 45 Hz), 24.72 (d, JC–P = 51 Hz),
mass spectrometry) was performed on a Voyager DE STR using
α-cyano-4-hydroxycinnamic acid as matrix. ESI-MS of metallo-
peptides with molecular weights above 1200 was performed
24.3 (d, JC–P = 51 Hz), 6.26 (d, JC–P = 5 Hz); 31P{1H} NMR
(161.9 MHz, CDCl3): δ = 51.6. Rf = 0.13 (1 : 1 EtOAc/hexane).
using a Thermo Quantum Discovery Max triple-quadrupole
mass spectrometer. Measurements were conducted in positive
N-Boc-3-(diisopropylphosphorothioyl)-alanine (Boc-Ipa-OH)
This compound was also prepared using the procedure
described by Greenfield et al.14b Yield 69%. 1H NMR
(400 MHz, CDCl3): δ = 9.21 (br, 1H, COOH), 5.95 (d, 1H, NH),
4.56–4.62 (m, 1H, N–CH–COO), 2.42 (m, 1H, N–CH–CHH′–P),
2.32 (m, 1H, N–CH–CHH′–P), 2.09–2.21 (m, 2H, P-CHMe2),
1.42 (s, 9H, NH-Boc), 1.15–1.27 (m, 12H, P-CHMe2); 13C{1H}
NMR (100 MHz, CDCl3): δ = 171.1 (s), 155.4 (s), 80.6 (s), 50.4
and negative ionization modes using a methanol/water (50 : 50
by volume) mobile phase at a flow rate of 10 mL min−1. Mass
spectra of lower molecular weight metallopeptides were
recorded on a JEOL LCmate using atmospheric-pressure
chemical ionization (APCI) in positive mode. HPLC purifi-
cation of peptides was performed on a Waters 600E HPLC
system. For analytical HPLC, a 3 × 50 mm C-18 column was
used and for semi-preparative a PrepLC 25 mm module C-18
column.
(s), 28.7 (d, JC–P = 49 Hz), 28.3 (d, JC–P = 49 Hz), 25.5 (d, JC–P
=
48 Hz), 16.1 (m), 15.8 (m); 31P{1H} NMR (161.9 MHz, CDCl3):
δ = 65.1. Rf = 0.2 (1 : 1 EtOAc/hexane).
3-(diphenylphosphorothioyl)propionic acid
Solid phase peptide synthesis
3-(Diphenylphosphino)propionic acid (2.5 g, 9.7 mmol) and
sulfur (0.316 g, 9.8 mmol) were suspended in toluene The peptide WASKLPSG was synthesized on a Protein Techno-
(17.5 mL) and refluxed under nitrogen for two hours. The reac- logies PS3 automated peptide synthesizer using the
tion mixture was cooled to room temperature, and the toluene standard Fmoc/t-Bu protection strategy and HBTU as coupling
removed under reduced pressure to yield pure 3-(diphenylphos- reagent on Fmoc-Gly wang resin (Aapptec 0.54 mmol g−1
,
phorothioyl)propionic acid as a pale yellow solid (2.8 g, 99%). 100–200 mesh) at 0.1 mmol scale. Following synthesis, the
Dalton Trans.
This journal is © The Royal Society of Chemistry 2015

View Article Online
Dalton Transactions
Paper
peptide N-terminus was acetylated by treatment with 1 : 1 (v/v) (m, 1H, N–CH–CO), 4.44 (m, 1H, N–CH–CO), 4.12–4.22 (m, 1H,
acetic anhydride: N-methylmorpholine in DMF for 30 min.
N–CH–CO), 3.69 (s, 3H, COOCH3), 2.1–2.5 (m, 3H, Epa: N–CH–
CH2–P, Val: CH–CHMe2), 2.06 (s, 3H, NH-COCH3), 1.7–1.94 (m,
6H, Epa: P–CH2–CH3, Leu: N–CH–CH2–CHMe2), 1.37 (m, 1H,
On-resin modification of peptide
Resin-bound peptide (0.05 mmol) was treated with 2% hydra- Leu: CH2–CHMe2), 1.11–1.21 (m, 6H, P–CH2–CH3), 0.89–0.95
zine (v/v) in DMF (10 mL) for 1 h under nitrogen at room temp- (m, 12H, Leu: CH2–CHMe2, Val: CH–CHMe2); 31P{1H} NMR
erature. The resin beads were then thoroughly washed with (161.9 MHz, CDCl3): δ = 52.0 (major isomer 92%), 51.9 (minor
DMF. The resin beads were then treated with a solution of diastereomer 8%); Rf = 0.2 (EtOAc); MALDI-TOF MS: m/z =
3-(diphenylphosphino)propionic acid or 3-(diphenylphosphor- 500.31 (M + Na)+. The two diastereomeric tripeptides could not
othioyl)propionic acid (0.15 mmol) together with HATU be separated because of nearly identical polarity.
(1-[bis(dimethylamino)methylene]-1H-1,2,3-triazolo[4,5-b]pyri-
Ac-Val-Ipa-Leu-OMe (3a). 1H NMR (400 MHz, CDCl3): δ =
dinium 3-oxid hexafluorophosphate) (54 mg, 0.14 mmol) and 7.94 (d, 1H, NH), 7.85 (d, 1H, NH), 6.04 (m, 1H, NH), 4.88 (m,
N,N-diisopropylethylamine (DIEA) (30 μL, 0.16 mmol) in 8 mL 1H, N–CH–CO), 4.45 (m, 1H, N–CH–CO), 4.28 (m, 1H, N–CH–
DMF for 45 min. The coupling was repeated a second time CO), 3.72 (s, 3H, COOCH3), 2.11–2.48 (m, 5H, Ipa: βH and
to ensure complete reaction. The resin was washed with DMF P–CHMe2, Val: βH), 2.09 (s, 3H, NH–COCH3), 1.63–1.78 (m, 3H,
(4 × 2 mL) and dichloromethane (5 × 2 mL) followed by clea- Leu: βH and γH), 1.19–1.31 (m, 12H, P-CHMe2), 0.92–1 (m,
vage of the peptide using 95 : 2.5 : 2.5 TFA/water/TIPS (v/v) 12H, Leu: CH2–CHMe2, Val: CH–CHMe2); 31P{1H} NMR
for two hours (TFA = trifluoroacetic acid, TIPS = triisopropyl- (161.9 MHz, CDCl3): δ = 65.0. Rf = 0.22 (EtOAc). MALDI-TOF
silane). After concentrating the TFA solution, the peptide was MS: m/z = 528.36 (M + Na)+.
precipitated with diethyl ether at −20 °C and purified by
Ac-Val-Ppa-Leu-OMe (4a). 1H NMR (400 MHz, CD2Cl2): δ =
reverse phase HPLC using aqueous acetonitrile gradients con- 7.83 (m, 4H, P–Ph), 7.75 (d, 1H, NH), 7.49 (m, 6H, P–Ph), 7.10
taining 0.1% TFA (v/v). General yields of the peptides after (d, 1H, NH), 5.98 (d, 1H, NH), 4.69–4.79 (m, 1H, N–CH–CO),
purification were ca. 20%.
4.21 (m, 1H, N–CH–CO), 4.07 (m, 1H, N–CH–CO), 3.64 (s, 3H,
Peptide WASK{–COC2H4PPh2}LPSG, 1a. 31P{1H} NMR COOCH3), 3.14 (m, 2H, Ppa: βH), 2.09 (m, 1H, Val: βH), 2.06 (s,
(161.9 MHz, CD3CN): δ = −14.4; Rt = 60 min (59% acetonitrile, 3H, NH-COCH3), 1.52–1.54 (m, 3H, Leu: βH and γH), 0.81–0.90
flow rate: 0.5 mL min−1); MALDI-TOF MS: m/z = 1149.8 (m, 12 H, Leu: CH2–CHMe2, Val: CH–CHMe2); 31P{1H} NMR
(M + Na)+.
(161.9 MHz, CD2Cl2): δ = 40.35. Rf = 0.1 (2 : 1 EtOAc/hexane).
NMR MALDI-TOF MS: m/z = 595.8 (M + Na)+.
Peptide
WASK{–COC2H4P(S)Ph2}LPSG. 31P{1H}
(161.9 MHz, CD3CN): δ = 43.08; Rt = 54 min (53.6% aceto-
nitrile, flow rate: 0.5 mL min−1); MALDI-TOF MS: m/z = 1181.6
(M + Na)+.
Desulfurization of WASK{–COC2H4P(S)Ph2}LPSG
Raney™ nickel slurry (800 mg) was transferred to a Schlenk
flask and washed with deaerated methanol until the washings
were clear followed by washing with CH3CN (3 × 5 mL). The
Solution phase peptide synthesis
N-Boc protected phosphino amino acid (1 equiv.), leucine peptide (29 μmol) in 4 : 1 CH3CN/water (v/v, 10 mL) was added
methyl ester hydrochloride (1 equiv.), and PyBOP {(Benzotriazol- to the Raney™ nickel and stirred at room temperature for 2 h.
1-yloxy)tripyrrolidinophosphonium
hexafluorophosphate} Completion of desulfurization was confirmed by MALDI-TOF
(1.5 equiv.) were dissolved in CH2Cl2 under nitrogen and MS. The Raney™ nickel was allowed to settle, and the reaction
cooled to 0 °C. DIEA (3 equiv.) was added, and the reaction mixture was filtered into a Schlenk flask under an argon
mix was allowed to warm to room temperature. After stirring at atmosphere. The Raney™ nickel was rinsed with 5 × 5 mL
room temperature for 40 min, the reaction mixture was CH3CN and filtered into the Schlenk flask. The CH3CN solu-
washed with 1 M HCl, saturated NaHCO3, and brine. The tion was concentrated under reduced pressure and lyophilized
CH2Cl2 solution was dried over MgSO4, filtered, concentrated overnight to obtain phosphine-peptide (1a) as white powder
under reduced pressure, and purified over silica using hexane/ that was immediately used for the cluster incorporation.
ethyl acetate as eluent. The N-Boc methyl ester protected
Synthesis of 1 from 1a
dipeptide was treated with 1 : 1 TFA/CH2Cl2 until complete
removal of Boc was indicated by TLC. The reaction mixture The phosphine-peptide 1a (3 μmol) was dissolved in CH3CN
was then concentrated under reduced pressure, dissolved in (5 mL). In a separate flask, (μ-pdt)Fe2(CO)6 (4 mg) and tri-
toluene and concentrated until white solid was obtained. That methylamine-N-oxide (2 mg) were dissolved in CH3CN (1 mL)
solid was dissolved in CH2Cl2 and concentrated followed by and stirred in the dark for 10 min. This solution was trans-
drying under vacuum for 12 h. The N-acetyl valine (1 equiv.) ferred via cannula to the peptide solution. After stirring for
was then coupled to the TFA salt of the dipeptide using the two hours in the dark, the solvent was removed under reduced
same procedure to the point of the brine washing. The CH2Cl2 pressure, and the product metallopeptide (1) was purified via
solution containing tripeptide was dried, concentrated and reverse phase HPLC. Yield 15%. 31P{1H} NMR (161.9 MHz,
purified over silica using 2% methanol/ethyl acetate as eluent.
Ac-Val-Epa-Leu-OMe (2a). 1H NMR (400 MHz, CDCl3): δ = 1981, 1958, 1925. ESI-MS: (EI+) m/z = 1458.5 (M + H − CO)+;
7.98 (d, 1H, NH), 7.82 (d, 1H, NH), 6.06 (m, 1H, NH), 4.89 (EI−) m/z = 1456.4 (M–H–CO)−, 1373.6 (M–H–4CO)−. Rt
10% D2O/CD3CN): δ = 55.43. IR (KBr, cm−1): ν(CO) = 2044,
=
This journal is © The Royal Society of Chemistry 2015
Dalton Trans.

View Article Online
Paper
Dalton Transactions
76 min (73.4% acetonitrile, flow rate 0.5 mL min−1). The mer 40%); IR (CH2Cl2, cm−1): ν(CO) = 2043, 1979, 1960, 1919;
extinction coefficient of the complex was not determined due Rf = 0.6 (EtOAc); MS (APCI+): m/z = 832.1449 (calculated
to difficulty obtaining an accurate concentration value for solu- 832.1452).
tions. Determination of the concentration of 1a from 280 nm
absorbance of tryptophan is not possible due to overlap of the (400 MHz, CD2Cl2): δ = 7.62–7.76 (m, 4H, P–Ph), 7.41–7.51 (m,
(Val-Ppa-Leu)-{(μ-pdt)Fe2(CO)5} (4). Yield 87%. 1H NMR
280 and 350 nm bands.
6H, P–Ph), 6.56 (d, 1H, NH), 5.77 (d, 1H, NH), 5.52 (d, 1H, NH),
4.74 (m, 1H, N–CH–CO), 4.36 (m, 1H, N–CH–CO), 3.75 (m, 1H,
N–CH–CO), 3.69 (s, 3H, COOCH3), 2.83–3.30 (m, 2H, Ppa:
Desulfurization of tripeptides (2a, 3a, 4a)
To a slurry of Raney™ nickel (washed with deaerated MeOH P–CH2), 1.93 (s, 3H, NH-COCH3), 1.85 (m, 5H, Val: βH, pdt:
and CH3CN) was added a solution of tripeptide (0.25 mmol) in SCH2), 1.50–1.58 (m, 6H, Leu: βH and γH, pdt: CH2–CH2–CH2),
CH3CN (15 mL). The reaction mixture was stirred at 45–55 °C 0.73–0.85 (m, 12H, Leu: CH2–CHMe2, Val: CH–CHMe2); 31P{1H}
under argon until complete reduction of the phosphine NMR (161.9 MHz, CD2Cl2): δ = 55.3 (major isomer 70%), 55.9
sulfide as indicated by either 31P NMR or MALDI. The Raney™ (minor diastereomer, 30%); IR (CH2Cl2, cm−1): ν(CO) = 2046,
nickel was allowed to settle, and the supernatant was filtered 1984, 1961, 1928; Rf = 0.3 (2% MeOH/CH2Cl2); MS (APCI+): m/z
under argon. The filtrate was concentrated under reduced = 900.1159 (calculated 900.1139).
pressure and lyophilized overnight. The phosphine tripeptide
was immediately used for the reaction with diiron hexacarbo- (400 MHz, CD2Cl2): δ = 7.65 (m, 2H, bdt), 7.47–7.53 (m, 4H,
nyl precursor. P–Ph), 7.32–7.4 (m, 2H, P–Ph), 7.24–7.27 (4H, P–Ph), 6.94 (d,
Reduction of 2a. Reaction condition: 2 h at 25 °C, then 1 h 1H, NH), 6.54 (d, 1H, NH), 6.48 (td, 1H, bdt), 6.39 (m, 1H, bdt),
at 44 °C. MALDI-TOF MS: m/z = 446.37 (M + H)+.
4.84 (m, 1H, N–CH–CO), 4.41 (m, 1H, N–CH–CO), 3.73 (s, 3H,
Reduction of 3a. Reaction condition: 4 h at 48 °C. 31P{1H} COOCH3), 3.7 (m, 1H, N–CH–CO), 2.73 (m, 2H, Ppa: PCH2),
NMR (161.9 MHz, CDCl3): δ = −4.1. 1.94 (s, 3H, NH–COCH3), 1.49–1.60 (m, 4H, Leu: βH and γH,
(Val-Ppa-Leu)-{(μ-bdt)Fe2(CO)5} (5). Yield 84%. 1H NMR
Reduction of 3a. Reaction condition: 4 h at 55 °C. 31P{1H} Val: βH), 0.72–0.90 (m, 12H, Leu: CH2–CHMe2, Val: CH–
NMR (161.9 MHz, CDCl3): δ = −23.3.
CHMe2); 31P{1H} NMR (161.9 MHz, CD2Cl2): δ = 51.3; IR
(CH2Cl2, cm−1): ν(CO) = 2053, 1994, 1937; Rf = 0.37 (2% MeOH/
CH2Cl2); MS (APCI+): m/z = 934.0973 (calculated 934.0983).
General procedure for cluster incorporation into tripeptide
A solution of (μ-SRS)Fe2(CO)6 (0.3 mmol) (R = pdt or bdt) in
CH3CN (3 mL) was added dropwise to a CH2Cl2 (2 mL) solu-
Electrochemistry
tion of Me3NO·2H2O (0.3 mmol) at room temperature. After Electrochemical measurements were made using a CHI 1200A
stirring the reaction mixture in the dark for 10 min, phosphine electrochemical analyzer under an atmosphere of argon. A con-
peptide (0.2 mmol) in CH2Cl2 (10 mL) was added dropwise. ventional three-electrode cell was used. The working electrode
After 3 h, the solvent was removed under reduced pressure, was a 3 mm diameter glassy carbon disk polished successively
and the residue was purified by silica gel chromatography with with 1 mm and 0.3 mm deagglomerated alpha alumina and
2% MeOH/CH2Cl2 to produce the peptide-diiron complex as sonicated for 15 min in ultrapure water prior to use. The Ag/
red powder.
Ag+ reference electrode was prepared by immersing a silver
(Val-Epa-Leu)-{(μ-pdt)Fe2(CO)5} (2). Yield 77%. 1H NMR wire anodized with AgCl in an CH3CN solution of 0.1 M [NBu4]-
(400 MHz, CD2Cl2): δ = 6.78 (d, 1H, NH), 6.63 (d, 1H, NH), 5.98 [PF6]. A platinum wire was used as counter electrode. For electro-
(d, 1H, NH), 4.93 (m, 1H, N–CH–CO), 4.47 (m, 1H, N–CH–CO), chemical studies in CH3CN, a 0.1 M solution of [NBu4][PF6]
4.18 (m, 1H, N–CH–CO), 3.69 (s, 3H, NH–COCH3), 2.1–2.56 (m, was used as supporting electrolyte. Mixtures of 0.1 M [NBu4]-
5H, Epa: βH, Val: βH, pdt: S–CHH′), 1.99 (s, 3H, NH–COCH3), [PF6] in CH3CN and 0.1 M KCl in water were used for mixed
1.72–1.88 (m, 6H, Epa: P–CH2–CH3, pdt: S–CHH′) 1.56–1.60 solvent experiments. Argon was sparged through the solution
(m, 5H, Leu: βH and γH, pdt: CH2–CH2–CH2), 1.23–1.26 for 15 min before experiments and anaerobicity was main-
(m, 6H, P–CH2–CH3), 0.90–0.94 (m, 12H, Leu: CH2–CHMe2, tained during the electrochemical measurements by an atmos-
Val: CH–CHMe2); 31P{1H} NMR (161.9 MHz, CD2Cl2): δ = 54.8 phere of argon. All potentials are reported relative to the
(major isomer 78%), 54.9 (minor diastereomer 22%); ferrocene couple (Fc+/Fc). Concentrations of the complexes
IR (CH2Cl2, cm−1): ν(CO) = 2040, 1981, 1966, 1921; Rf = 0.5 were determined spectrophotometrically using the extinction
(EtOAc); MS (APCI+): m/z = 804.1117 (calculated 804.1139).
(Val-Ipa-Leu)-{(μ-pdt)Fe2(CO)5} (3). Yield 80%. 1H NMR
coefficients of the Fe–S charge transfer bands.
The catalytic activities of the complexes were compared
(400 MHz, CD2Cl2): δ = 6.85 (d, 1H, NH), 6.63 (d, 1H, NH), 5.95 using the ratio (icat/ip) instead of turnover frequencies (TOF)
(d, 1H, NH), 4.98 (m, 1H, N–CH–CO), 4.47 (m, 1H, N–CH–CO), because determination of TOF is difficult without additional
4.03–4.18 (m, 1H, N–CH–CO), 3.67 (s, 3H, COOCH3), 1.97–2.24 mechanistic information. Since catalysis occurs at a potential
(m, 12H, Leu: COOCH3, Ipa: βH and PCHMe2, Val: βH, pdt: more negative than that of the FeIFeI/FeIFe0 couple and likely
SCH2), 1.55–1.68 (m, 7H, Leu: βH and γH, pdt: CH2–CH2–CH2), involves solvent-substituted complex, the potential of this
1.26–1.35 (m, 12H, Ipa: P–CHMe2), 0.89–0.98 (m, 12H, Leu: second reduction process in the absence of acid is unknown.
CH2–CHMe2, Val: CH–CHMe2); 31P{1H} NMR (161.9 MHz, Therefore, application of ‘foot-of-the-wave’ analysis is also not
CD2Cl2): δ = 70.2 (major isomer 60%), 69.8 (minor diastereo- appropriate for these voltammograms.
Dalton Trans.
This journal is © The Royal Society of Chemistry 2015

View Article Online
Dalton Transactions
Paper
C. S. Mocny, L. Ruckthong, H. Qayyum and V. L. Pecoraro,
Chem. Rev., 2014, 114, 3495–3578.
Acknowledgements
This research was supported through the Center for Bio-
Inspired Solar Fuel Production, an Energy Frontier Research
Center funded by the U.S. Department of Energy, Office of
Science, Office of Basic Energy Sciences under Award Number
770 DE-SC0001016.
9 (a) A. K. Jones, B. R. Lichtenstein, A. Dutta, G. Gordon and
P. L. Dutton, J. Am. Chem. Soc., 2007, 129, 14844–14845;
(b) U.-P. Apfel, M. Rudolph, C. Apfel, C. Robl,
D. Langenegger, D. Hoyer, B. Jaun, M.-O. Ebert,
T. Alpermann, D. Seebach and W. Weigand, Dalton Trans.,
2010, 39, 3065–3071; (c) S. Roy, S. Shinde, G. A. Hamilton,
H. E. Hartnett and A. K. Jones, Eur. J. Inorg. Chem., 2011,
2011, 1050–1055; (d) A. Roy, C. Madden and G. Ghirlanda,
Chem. Commun., 2012, 48, 9816–9818.
References
1 (a) M. W. W. Adams and E. I. Stiefel, Science, 1998, 282, 10 (a) F. Gloaguen, J. D. Lawrence and T. B. Rauchfuss,
1842–1843; (b) M. Frey, ChemBioChem, 2002, 3, 153–160;
(c) J. C. Fontecilla-Camps, A. Volbeda, C. Cavazza and
Y. Nicolet, Chem. Rev., 2007, 107, 4273–4303.
J. Am. Chem. Soc., 2001, 123, 9476–9477; (b) D. S. Chong,
I. P. Georgakaki, R. Mejia-Rodriguez, J. Samabria-
Chinchilla, M. P. Soriaga and M. Y. Darensbourg, Dalton
Trans., 2003, 4158–4163.
2 (a) J. W. Peters, W. N. Lanzilotta, B. J. Lemon and
L. C. Seefeldt, Science, 1998, 282, 1853–1858; (b) Y. Nicolet, 11 (a) S. R. Gilbertson, G. H. Chen and M. McLoughlin, J. Am.
C. Piras, P. Legrand, C. E. Hatchikian and J. C. Fontecilla-
Camps, Structure, 1999, 7, 13–23.
3 (a) D. J. Evans and C. J. Pickett, Chem. Soc. Rev., 2003, 32,
Chem. Soc., 1994, 116, 4481–4482; (b) A. Agarkov,
S. Greenfield, D. Xie, R. Pawlick, G. Starkey and
S. R. Gilbertson, Biopolymers, 2006, 84, 48–73.
268–275; (b) T. R. Simmons, G. Berggren, M. Bacchi, 12 J. Shearer and L. M. Long, Inorg. Chem., 2006, 45, 2358–
M. Fontecave and V. Artero, Coord. Chem. Rev., 2014,
270–271, 127–150.
4 (a) G. A. N. Felton, C. A. Mebi, B. J. Petro, A. K. Vannucci,
D. H. Evans, R. S. Glass and D. L. Lichtenberger, J. Organo-
2360.
13 L.-C. Song, H.-T. Wang, J.-H. Ge, S.-Z. Mei, J. Gao,
L.-X. Wang, B. Gai, L.-Q. Zhao, J. Yan and Y.-Z. Wang,
Organometallics, 2008, 27, 1409–1416.
met. Chem., 2009, 694, 2681–2699; (b) C. Tard and 14 (a) S. R. Gilbertson, S. E. Collibee and A. Agarkov, J. Am.
C. J. Pickett, Chem. Rev., 2009, 109, 2245–2274.
5 (a) G. Berggren, A. Adamska, C. Lambertz, T. R. Simmons,
J. Esselborn, M. Atta, S. Gambarelli, J. M. Mouesca,
Chem. Soc., 2000, 122, 6522–6523; (b) S. J. Greenfield,
A. Agarkov and S. R. Gilbertson, Org. Lett., 2003, 5, 3069–
3072.
E. Reijerse, W. Lubitz, T. Happe, V. Artero and 15 (a) R. Mejia-Rodriguez, D. Chong, J. H. Reibenspies,
M. Fontecave, Nature, 2013, 499, 66–69; (b) J. Esselborn,
C. Lambertz, A. Adamska-Venkatesh, T. Simmons,
G. Berggren, J. Noth, J. Siebel, A. Hemschemeier, V. Artero,
E. Reijerse, M. Fontecave, W. Lubitz and T. Happe,
M. P. Soriaga and M. Y. Darensbourg, J. Am. Chem. Soc.,
2004, 126, 12004–12014; (b) F. Gloaguen, D. Morvan,
J.-F. Capon, P. Schollhammer and J. Talarmin, J. Electro-
anal. Chem., 2007, 603, 15–20.
Nat. Chem. Biol, 2013, 9, 607–609; (c) W. K. Myers, 16 I. K. Pandey, S. M. Mobin, N. Deibel, B. Sarkar and
T. A. Stich, D. L. M. Suess, J. M. Kuchenreuther, J. R. Swartz
and R. D. Britt, J. Am. Chem. Soc., 2014, 136, 12237–
12240.
6 (a) R. D. Bethel and M. Y. Darensbourg, Nature, 2013, 499,
40–41; (b) S. Roy and A. K. Jones, Nat. Chem. Biol, 2013, 9,
603–605.
7 (a) X. de Hatten, E. Bothe, K. Merz, I. Huc and N. Metzler-
Nolte, Eur. J. Inorg. Chem., 2008, 2008, 4530–4537;
(b) W. J. Shaw, Catal. Rev.: Sci. Eng., 2012, 54, 489–550;
S. Kaur-Ghumaan, Eur. J. Inorg. Chem., 2015, 2015, 2875–
2882.
17 (a) J.-F. Capon, S. Ezzaher, F. Gloaguen, F. Y. Petillon,
P. Schollhammer, J. Talarmin, T. J. Davin, J. E. McGrady
and K. W. Muir, New J. Chem., 2007, 31, 2052–2064;
(b) J.-F. Capon, F. Gloaguen, F. Y. Petillon, P. Schollhammer
and J. Talarmin, Coord. Chem. Rev., 2009, 253, 1476–1494;
(c) P. Surawatanawong, J. W. Tye, M. Y. Darensbourg and
M. B. Hall, Dalton Trans., 2010, 39, 3093–3104.
(c) B. Ginovska-Pangovska, A. Dutta, M. L. Reback, 18 V. Fourmond, P.-A. Jacques, M. Fontecave and V. Artero,
J. C. Linehan and W. J. Shaw, Acc. Chem. Res., 2014, 47, Inorg. Chem., 2010, 49, 10338–10347.
2621–2630; (d) G. Caserta, S. Roy, M. Atta, V. Artero and 19 Y. Na, M. Wang, K. Jin, R. Zhang and L. Sun, J. Organomet.
M. Fontecave, Curr. Opin. Chem. Biol., 2015, 25,
36–47.
Chem., 2006, 691, 5045–5051.
20 (a) A. Jain, S. Lense, J. C. Linehan, S. Raugei, H. Cho,
D. L. DuBois and W. J. Shaw, Inorg. Chem., 2011, 50, 4073–
4085; (b) U. J. Kilgore, J. A. S. Roberts, D. H. Pool,
A. M. Appel, M. P. Stewart, M. R. DuBois, W. G. Dougherty,
W. S. Kassel, R. M. Bullock and D. L. DuBois, J. Am. Chem.
Soc., 2011, 133, 5861–5872; (c) M.-H. Ho, M. O’Hagan,
M. Dupuis, D. L. DuBois, R. M. Bullock, W. J. Shaw and
S. Raugei, Dalton Trans., 2015, 44, 10969–10979;
8 (a) M. Faiella, C. Andreozzi, R. T. M. de Rosales, V. Pavone,
O. Maglio, F. Nastri, W. F. DeGrado and A. Lombardi, Nat.
Chem. Biol, 2009, 5, 882–884; (b) V. Nanda and R. L. Koder,
Nat. Chem., 2010, 2, 15–24; (c) D. W. Watkins,
C. T. Armstrong and J. L. R. Anderson, Curr. Opin. Chem.
Biol., 2014, 19, 90–98; (d) F. Yu, V. M. Cangelosi,
M. L. Zastrow, M. Tegoni, J. S. Plegaria, A. G. Tebo,
This journal is © The Royal Society of Chemistry 2015
Dalton Trans.

View Article Online
Paper
Dalton Transactions
(d) W. A. Hoffert, J. A. S. Roberts, R. Morris Bullock and 22 S. van Zutphen, V. J. Margarit, G. Mora and P. Le Floch,
M. L. Helm, Chem. Commun., 2013, 49, 7767–7769; Tetrahedron Lett., 2007, 48, 2857–2859.
(e) J. Hou, M. Fang, A. J. P. Cardenas, W. J. Shaw, 23 E. J. Lyon, I. P. Georgakaki, J. H. Reibenspies and
M. L. Helm, R. M. Bullock, J. A. S. Roberts and
M. Y. Darensbourg, Angew. Chem., Int. Ed., 1999, 38, 3178–
M. O’Hagan, Energy Environ. Sci., 2014, 7, 4013–4017.
3180.
21 E. N. Tsvetkov, N. A. Bondarenko, I. G. Malakhova and 24 J. Zhao, Z. Wei, X. Zeng and X. Liu, Dalton Trans., 2012, 41,
M. I. Kabachnik, Synthesis, 1986, 198–208. 11125–11133.
Dalton Trans.
This journal is © The Royal Society of Chemistry 2015