Enhanced Photocatalytic Hydrogen Production by Hybrid Streptavidin-Diiron Catalysts

Article abstract of DOI:10.1002/chem.202000204

Hybrid protein–organometallic catalysts are being explored for selective catalysis of a number of reactions, because they utilize the complementary strengths of proteins and of organometallic complex. Herein, we present an artificial hydrogenase, StrepH2, built by incorporating a biotinylated [Fe–Fe] hydrogenase organometallic mimic within streptavidin. This strategy takes advantage of the remarkable strength and specificity of biotin-streptavidin recognition, which drives quantitative incorporation of the biotinylated diironhexacarbonyl center into streptavidin, as confirmed by UV/Vis spectroscopy and X-ray crystallography. FTIR spectra of StrepH2 show characteristic peaks at shift values indicative of interactions between the catalyst and the protein scaffold. StrepH2 catalyzes proton reduction to hydrogen in aqueous media during photo- and electrocatalysis. Under photocatalytic conditions, the protein-embedded catalyst shows enhanced efficiency and prolonged activity compared to the isolated catalyst. Transient absorption spectroscopy data suggest a mechanism for the observed increase in activity underpinned by an observed longer lifetime for the catalytic species Fe^IFe⁰ when incorporated within streptavidin compared to the biotinylated catalyst in solution.

Full text of DOI:10.1002/chem.202000204

Full Paper
doi.org/10.1002/chem.202000204
Chemistry—A European Journal
&
Photochemistry
Enhanced Photocatalytic Hydrogen Production by Hybrid
Streptavidin-Diiron Catalysts
Anindya Roy,[a, b] Michael D. Vaughn,[a] John Tomlin,[a] Garrett J. Booher,[a] Gerdenis Kodis,[a]
Chad R. Simmons,[a] James P. Allen,[a] and Giovanna Ghirlanda*[a]
Abstract: Hybrid protein–organometallic catalysts are being
explored for selective catalysis of a number of reactions, be-
cause they utilize the complementary strengths of proteins
and of organometallic complex. Herein, we present an artifi-
cial hydrogenase, StrepH2, built by incorporating a biotiny-
lated [Fe–Fe] hydrogenase organometallic mimic within
streptavidin. This strategy takes advantage of the remarkable
strength and specificity of biotin-streptavidin recognition,
which drives quantitative incorporation of the biotinylated
diironhexacarbonyl center into streptavidin, as confirmed by
UV/Vis spectroscopy and X-ray crystallography. FTIR spectra
of StrepH2 show characteristic peaks at shift values indica-
tive of interactions between the catalyst and the protein
scaffold. StrepH2 catalyzes proton reduction to hydrogen in
aqueous media during photo- and electrocatalysis. Under
photocatalytic conditions, the protein-embedded catalyst
shows enhanced efficiency and prolonged activity compared
to the isolated catalyst. Transient absorption spectroscopy
data suggest a mechanism for the observed increase in ac-
tivity underpinned by an observed longer lifetime for the
catalytic species FeIFe0 when incorporated within streptavi-
din compared to the biotinylated catalyst in solution.
Introduction
organometallic mimics of [Fe–Fe] hydrogenases built around a
bridging dithiolate ligand have been extensively explored, pro-
viding a rich repertoire of chemical structures. These models
serve as an important testing ground to assess the features
that impart catalytic prowess to the H-cluster.[12–24] While most
are soluble in organic solvent, several examples of complexes
built on a (m-S-(CH2)3-S)[Fe2(CO)6] scaffold and modified with
polar groups to facilitate derivatization and/or attachment
have been reported.[12–16] Compared to the hydrogenases,
these models show moderate catalytic activity, and generally
are inefficient catalysts in photoinduced hydrogen production
in part because they degrade during long irradiation, often re-
sulting in the loss of the CO ligands.[17–25] Simple diiron-dithio-
late complexes have also been encased within biopolymers or
non-biological scaffolds, such as cyclodextrins, surfactants, den-
drimers, or metal-organic frameworks (MOFs), showing that
supramolecular confinement enhances photoinduced hydro-
gen production.[20,26–29] However, these systems cannot provide
the tailored second-sphere interactions that in natural hydro-
genases activate the organometallic center by stabilization of
the catalytically active conformation.[8,30,31]
Hydrogen metabolism is regulated by hydrogenases, a class of
metalloenzymes that catalyze reduction of protons as well as
hydrogen oxidation under mild conditions at near-thermody-
namic potential, using bimetallic centers (either [Fe–Fe] or [Ni–
Fe]) at the active site.[1,2] Of these, the [Fe–Fe] hydrogenases
are the most efficient for hydrogen production.[3] They incorpo-
rate an unusual diiron organometallic complex coordinated by
a azadithiolate bridging ligand, carbon monoxide, and cyanide
ligands, tethered via a single cysteine to a cubane-type [4Fe–
4S] cluster to form the H-cluster.[1] The diiron cluster, inactive in
isolation, is stabilized and activated via second-sphere interac-
tions when incorporated within the protein.[4–6] The apo hydro-
genase also accepts synthetic, non-native cofactors, such as a
propanedithiolate bridging ligand, resulting in negligible resid-
ual activity compared to WT hydrogenase, further indicating
the role of the protein environment.[7–11] Bioinspired structural
[a] Dr. A. Roy, Dr. M. D. Vaughn, Dr. J. Tomlin, G. J. Booher, Dr. G. Kodis,
C. R. Simmons, Prof. J. P. Allen, Prof. G. Ghirlanda
School of Molecular Sciences, Arizona State University
Tempe, AZ 85287-1604 (USA)
In a quest for robust, minimalist artificial hydrogenases, our
group and others have incorporated these organometallic cen-
ters into peptides and proteins using distinct strategies.[32–36]
Promising results in photocatalyzed hydrogen production were
obtained by anchoring diiron hexacarbonyl complexes to a
helical peptide or within Cyt c by exploiting the thiol moiety of
cysteine in CXXC motifs,[32,34] or using an artificial amino acid
to tether the diiron hexacarbonyl complex to a helical com-
plex.[35] Encapsulation of organometallic catalysts within host
proteins yielded mixed results: No improvements were ob-
E-mail: gghirlanda@asu.edu
[b] Dr. A. Roy
Present Address: Molecular Engineering and Sciences
Institute for Protein Design, University of Washington
Seattle, WA 98195-1655 (USA)
Supporting information and the ORCID identification number(s) for the au-
thor(s) of this article can be found under:
https://doi.org/10.1002/chem.202000204. It contains experimental proto-
cols, NMR spectra, electrochemical data, transient absorption data, and
HRMS.
Chem. Eur. J. 2020, 26, 1 – 8
1
ꢀ 2020 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ

Full Paper
doi.org/10.1002/chem.202000204
Chemistry—A European Journal
served when incorporating diiron hexacarbonyl complexes in
nitrobindin.[37] A synthetically convenient strategy to build
hybrid metalloproteins is based on the streptavidin-avidin
complex; since its introduction,[38] this strategy has been suc-
cessfully used to catalyze enantioselective reduction of a varie-
ty of organic substrates, indicating that catalysts linked to
biotin interact with the protein.[39–44] In this system, cobalt cata-
lysts for hydrogen production retained or improved catalytic
activity, especially when probed at high pH.[45,46] Mutations
close to the organometallic complex improved pH dependence
and initial rates/TON as compared to the naked catalyst.[46]
Inspired by these results, we utilized the streptavidin-avidin
system to stably encapsulate a biomimetic diiron hexacarbonyl
complex, reasoning that the biotin scaffold could stabilize its
catalytically active state, in analogy to what observed in natu-
ral [FeFe] hydrogenases.[30]
by a displacement assay using 2-(4’-hydroxyazobenzene)ben-
zoic acid (HABA) (Supporting Information). In solution, strepta-
vidin exists as a tetramer formed by units parallel to each
other in pairs, with the biotin cavity oriented accordingly (two
„up“ and two „down“) in the assembly. We found that the stoi-
chiometry of binding for BiotH2 to streptavidin is approximate-
ly 0.7:1 (expressed as BiotH2 to streptavidin monomer), similar
to values reported for the biotin-streptavidin system,[39–41] and
indicating that introduction of the diiron-hexacarbonyl moiety
at the valeric chain of biotin is well tolerated.
The structural integrity of the diironhexacarbonyl moiety of
3 in StrepH2 was verified by UV/Vis and FTIR spectroscopy. Pu-
rified StrepH2 assembly showed two distinct bands in the UV/
Vis spectrum, centered around 329 nm and 474 nm. These fea-
tures are characteristic of Fe2S2 butterfly complexes and consis-
tent with the reported literature for similar organometallic
complexes (Figure 1, Panel 1).[32,36] Possible changes to the en-
vironment surrounding the organometallic cluster upon incor-
poration into streptavidin were assessed by FTIR spectroscopy,
which reports on the environment of the CO groups via
changes to the C=O stretching modes. To interpret the FTIR
spectra, we used as benchmark the changes observed for the
1,3 propane-dithiolate diiron-hexacarbonyl complex as a func-
tion of solvent polarity.[51,52] In nonpolar solvents, a total of five
stretching frequencies were observed, reflecting a pseudo C2v
symmetry. In polar solvents, three of these low frequency
bands collapse to give rise to one broad absorption band.[51,52]
We compared the thin film FTIR signatures of BiotH2 with
those of StrepH2, and found evidence of increased hydropho-
bic interactions in StrepH2. FTIR spectra of BiotH2 showed two
sharp peaks at 2073 and 2030 cmÀ1, and a broad absorption
band at 1990 cmÀ1, consistent with related organometallic ana-
logues in polar solvents, and were assigned to the coupled
stretching frequency of -CO (Figure 2, Panel 2).[51] In StrepH2,
Building on the extensive literature on this class of com-
pounds, summarized above, we chose a simple organometallic
diiron hexacarbonyl center functionalized with a hydroxyl
moiety, and tethered it to biotin; reconstitution with streptavi-
din generated an artificial hydrogenase, StrepH2. The biotiny-
lated hydrogenase mimic is readily incorporated into streptavi-
din and displayed spectroscopic features typical of intact
diiron hexacarbonyl centers. StrepH2 catalyzes hydrogen pro-
duction under photocatalytic conditions with enhanced effi-
ciency compared to the isolated diiron hexacarbonyl center.
The crystal structure of StrepH2 provides a blueprint for the
design of improved versions by selective mutation of residues
in contact within the active site.
Results and Discussion
Synthesis and characterization of the biotinylated-[Fe–Fe]
hydrogenase mimic
the two sharp peaks were shifted to 2074 and 2034 cmÀ1
,
while the broad absorption feature showed three distinct
shoulders (2001, 1991, 1971 cmÀ1) implying a change in the en-
vironment of the organometallic cluster after binding to strep-
tavidin. The splitting of the low wavenumber absorption band
into three bands indicated a change in polarity around the CO
groups to a more hydrophobic environment upon binding to
the protein scaffold compared to the naked catalyst in solu-
tion.[52]
The biotinylated [Fe–Fe] hydrogenase mimic (referred to as
BiotH2 or 3 herein) was obtained as described in Scheme 1 by
coupling 2 to biotin in the presence of HATU with high yield.
Compound 2 was synthesized following established proce-
dures.[47–49] Briefly, the synthesis of 2 started from 3-mercapto-
2-(mercaptomethyl)propan-1-ol (1),[50] followed by refluxing
with Fe3(CO)12 in toluene to yield the metallic complex.[47,48]
BiotH2 was then incorporated into streptavidin to form the
complex StrepH2. The stoichiometry of binding was assessed
Figure 1. UV/Vis and FTIR spectra. StrepH2 panel 1: UV/Vis spectra of strepta-
vidin (black, dotted) and StrepH2 (black, solid). Panel 2: FTIR spectra of
BiotH2 in the absence (dotted line, bottom trace) and presence (solid, top
trace) of streptavidin.
Scheme 1. Synthesis of BiotH2.
&
&
Chem. Eur. J. 2020, 26, 1 – 8
www.chemeurj.org
2
ꢀ 2020 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Paper
doi.org/10.1002/chem.202000204
Chemistry—A European Journal
Catalytic activity of StrepH2
We used cyclic voltammetry to analyze the electrochemical
properties of BiotH2 in the presence and absence of streptavi-
din; experiments were carried out at 10 mm catalyst concentra-
tion, which ensures complete solubility of the complex (Sup-
plementary Information). Cyclic voltammograms of BiotH2 at
pH 6.0 showed an irreversible reductive peak at À980 mV vs.
SHE, corresponding to FeIFeI/FeIFe0 reduction consistent with
other water soluble [Fe-Fe] hydrogenase mimics.[35,53,54] When
the pH of the solution was lowered to 4.5, an increase in a re-
ductive wave corresponding to proton reduction was observed
(Figure 3, left panel). Controls were carried out at low pH to
confirm the catalytic current was not due to proton reduction
with a bare electrode at negative potential (Supplementary In-
formation). Direct electrochemistry of StrepH2 assembled in
situ by addition of 15 mm of streptavidin to a 10 mm solution of
BiotH2 at pH 6.0 resulted in a significant decrease of the reduc-
tive current, most likely due to the encapsulation of the com-
plex within streptavidin (SI). To address this issue, we per-
formed protein film voltammetry (PFV), to ensure rapid elec-
tron transfer with the electrode.
Figure 2. X-ray structure of StrepH2 showing BiotH2 incorporated in the
streptavidin pocket (left panel). The 2FoÀFc electron density reveals the
binding pocket used to build the BiotH2 bound model (middle panel), and
the arrangement of two adjacent streptavidin monomers with ligand inter-
actions of neighboring substituents are shown (right panel).
Because streptavidin has a pI of approximately 5–6, interac-
tion with the negatively charged electrode surface might have
been more favorable at low pH, resulting in conditions appro-
priate for PFV. Cyclic voltammetry experiments on StrepH2 im-
mobilized on a glassy carbon electrode showed a reductive
current that corresponded to FeIFeI/FeIFe0 reduction at pH 6
(Figure 3, right panel). When the pH was lowered to 4.5, the re-
ductive transition became catalytic, indicating proton reduc-
tion. Controls with streptavidin films (in the absence of BiotH2)
showed no redox activity in the range examined, indicating
that the catalytic process is due to the protein-organometallic
complex on the electrode.
Structure of StrepH2
The X-ray crystal structure of StrepH2 was determined to a res-
olution of 2.05 ꢁ confirming placement of the ligand into the
protein pocket (Figure 2, PDB ID 5VCQ). The structure was
solved by molecular replacement, using PDB 1MK5 as the
search model (see SI for details). The residual Fo-Fc difference
electron density in the binding pocket of was further examined
to assess ligand occupancy. The 2FoÀFc map revealed a well
resolved structure for the portion containing the biotin cofac-
tor, including the valeric chain and the 2-methyl-propanol por-
tion, while the area corresponding to the diiron hexacarbonyl
site was not resolved in the experimental density. Partial occu-
pancy of biotinylated ligands in the streptavidin pocket is ob-
served in all reported X-ray structures, because the protein
monomers arrange so that ligands incorporated within the
protein cavity abut onto one another causing steric clashes
and preventing full occupancy.[41] To gain further insights on
the environment of the complex in the protein, the remaining
portion of the ligand was modeled in Rosetta (see Supporting
Information) using the crystal structure of 1,3-propanedithiol in
complexed with diiron hexacarbonyl as starting point (CSDS
accession file AQIQIB).
We investigated the ability of StrepH2 to catalyze hydrogen
production under photocatalytic conditions by measuring the
catalytic efficiency of BiotH2 and StrepH2. In general, diiron
hexacarbonyl complexes are inefficient catalysts in photoin-
duced hydrogen production because they degrade during
long irradiation, often resulting in the loss of the CO li-
gands.[17–25] However, the streptavidin scaffold could provide
shielding via specific interactions with the catalyst. In this
assay, the catalyst at 5.14 mm concentration in the presence of
An analysis of the binding pocket with the ligand reconsti-
tuted as described shows that the diiron hexacarbonyl com-
plex is accommodated in the streptavidin pocket, and is sur-
rounded by protein loops on three sides, with the fourth side
facing a second streptavidin unit. The diiron hexacarbonyl por-
tion is partly exposed to the solvent, thus facilitating exchange
of protons and electrons. While we did not observe direct con-
tacts with any of the side chains or backbone amide moieties,
the complex is close enough for interaction with several resi-
dues, in particular we note possible interactions with Lys 106
from the neighboring monomer.
Figure 3. Cyclic voltammetry. Left: Cyclic voltammetry traces of BiotH2 at
pH 4.5 (black), pH 6.0 (gray) and StrepH2 at pH 6.0 (light gray) in solution.
Right: Protein film voltammetry of StrepH2 at pH 4.5 (black), pH 6.0 (gray)
and streptavidin at pH 4.5 (light gray).
Chem. Eur. J. 2020, 26, 1 – 8
www.chemeurj.org
3
ꢀ 2020 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ

Full Paper
doi.org/10.1002/chem.202000204
Chemistry—A European Journal
150 mm [Ru(bpy)3]2+ as a photosensitizer and 100 mm ascorbic
acid as a sacrificial electron donor at pH 4.5 was irradiated by
light, allowing for the amount of hydrogen evolved upon irra-
diation to be monitored over time by GC (see the Supporting
Information).
We found that BiotH2 gives rise to a TON of 6.42(Æ0.05).
The naked catalyst was deactivated approximately after
3 hours under continuous irradiation. Under the same condi-
tions, StrepH2 resulted in a TON of 47.63(Æ3.16) over 11 hours.
While the initial rate of production is comparable for the two
species, the net increase in TON appears to derive, in large
part, from a much longer lifetime of the catalyst when embed-
ded in the protein (Figure 4): while the activity of BiotH2 starts
to level off after approximately 100 minutes, StrepH2 continues
to produce hydrogen for at least 9 hours.
Figure 5. Left panel: Evolution-associated-difference spectra (EADS) obtained
after global analysis of the transient absorption data for BiotH2 and StrepH2
in presence of [Ru(Bpy)3]2+ as a photosensitizer and ascorbic acid as a sacrifi-
cial electron donor upon excitation at 480 nm in argon degassed ascorbate
buffer at pH 4.5 and 45% (v/v) DMSO solution. (a) EADS for BiotH2: lifetimes
17 ns (black), 10 ms (green), 51 ms (red) and 486 ms (blue). (b) EADS for
StrepH2 system: lifetimes 50 ns (black), 61 ms (red), 823 ms (green), 2.39 ms
(blue) and nondecaying in 9 ms time delay window component (magenta).
Right panel: Schematic proposed electron transfer mechanism in the photo-
catalytic hydrogen production system.
Previous work incorporating a similar (m-S-(CH2)3-S)[Fe2(CO)6]
complex into nitrobindin showed no improvement in activity
when compared to the isolated synthetic catalyst, and much
slower initial rates, attributed by the authors to decreased ac-
cessibility of the ruthenium photosensitizer to the diiron active
site. The activity for the nitrobindin complex leveled off after
approximately 2 hours.[37] Our results, however, show a marked
increase in activity, and point to an effect due in large part to
protection of the catalyst from inactivation.
in the photoinduced electron transfer reaction as depicted in
Figure 5 (right panel). Transient absorption data for BiotH2
were satisfactorily fitted with 4 lifetimes: 17 ns, 10 ms, 51 ms
and 486 ms. The EADS corresponding to the species with a
17 ns lifetime shows broad induced absorption at 500–800 nm
and ground state bleaching at 450 nm characteristic of the
[Ru(Bpy)3]2+excited state.[27] This state decays via an electron
transfer from the ascorbic acid to form the transient
[Ru(Bpy)3]+ species, which has a characteristic strong induced
absorption at 520 nm.[55] The transient [Ru(Bpy)3]+ (see 51 ms
EADS) consequently decays in 51 ms via secondary electron
transfer and give rise to the FeIFe0 species, with a lifetime of
486 ms and characteristic induced absorption at 570 nm and
700 nm.[55] The minor EADS associated with a lifetime of 10 ms
corresponds to the [Ru(Bpy)3]+ species which are quenched by
some other means and is not involved in the electron transfer
to FeIFeI. There was no transient absorption signal at time
delays longer than 1 ms (see Supporting Information for more
transient absorption data). In contrast, transient absorption
data for the StrepH2 system were satisfactorily fitted with 5
lifetimes: 50 ns, 61 ms, 823 ms, 2.39 ms and a nondecaying com-
ponent. Similarly to the BiotH2 case, the EADS with lifetimes of
50 ns and 61 ms can be attributed to the [Ru(Bpy)3]2+ excited
state and transient [Ru(Bpy)3]+, respectively. The lifetime of the
FeIFe0 species can be fitted with two time constants, one with
2.39 ms and a nondecaying component (see the 2.39 ms EADS
and constant EADS), suggesting complex interaction and multi-
ple decay pathways in the protein environment. The 823 ms
EADS displays characteristic induced absorption at around
500 nm and is most likely associated with the oxidized and/or
reduced streptavidin species: identical transient was observed
in the control experiment with apo streptavidin (see Support-
ing Information).[56–58] Incorporation of BiotH2 into streptavidin
thus resulted in an extended (more than 4 times) lifetime for
the catalytically active form of the catalyst, with FeIFe0 oxida-
tion state, from 486 ms to 2.39 ms. The FeIFe0 is the first inter-
To verify this hypothesis, we interrogated the system by
transient absorption spectroscopy. This technique allows for
the elucidation of the electron transfer mechanism in photoca-
talyzed systems by comparing the behavior of BiotH2 (3) and
StrepH2. Experiments were carried out in conditions mirroring
closely those used for the hydrogen production studies (vide
supra), in the presence of ascorbic acid as a sacrificial electron
donor and [Ru(bpy)3]2+ as a photosensitizer, upon excitation at
480 nm. The evolution-associated-difference spectra (EADS) of
BiotH2 and StrepH2 obtained after global analysis of the data
measured in the 9 ms delay time window are shown in
Figure 5. Each EADS has an associated lifetime and shows
spectral features characteristic of the transient species involved
Figure 4. Visible light-driven hydrogen production by StrepH2 (5.14 mm)
(open circles) and by BiotH2 (5.14 mm) (solid circles) in the presence of
150 mm [Ru(bpy)3]2+ and 100 mm ascorbate buffer at pH 4.5. Samples were
irradiated with visible light (410–770 nm).
&
&
Chem. Eur. J. 2020, 26, 1 – 8
www.chemeurj.org
4
ꢀ 2020 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Paper
doi.org/10.1002/chem.202000204
Chemistry—A European Journal
3.81 (d, 1H), 4.11 (m, 1H), 4.30 (t, 1H), 6.35 (s, 1H), 6.41 ( s, 1H).
13C NMR (DMSO, d [ppm]): 24.82, 28.42, 33.61, 33.92, 41.29, 43.20,
55.78, 59.62, 61.48, 63.43, 66.75, 163.14, 172.96, 208.41. Mass: Cal-
culated (M+) 641.92. Obtained 642.92.
mediate in hydrogen evolution by diiron hexacarbonyl com-
plexes, as summarized in Figure 5 (right panel).
Conclusions
Electrochemistry
Here, we have demonstrated that incorporation of a biotinylat-
ed hydrogenase mimic into streptavidin results in increased ac-
tivity compared to the bare catalyst. Streptavidin-based cata-
lysts that interface an organometallic complex with a protein
scaffold have been used successfully in particular to impart ste-
reospecificity to a reaction.[38,41,42,59–62] This approach, however,
had been partially successful in the catalysis of proton reduc-
tion: incorporation of cobalt catalysts did not enhance activity
at neutral pH, although significant effects were observed at
high pH when using [Ir(bpy)(ppy)2]PF6 as photosensitizer and
triethanolamine as sacrificial electron donor; these conditions
require the use of acetonitrile as co-solvent.[45,46] Herein, we
chose to explore the incorporation of a different type of hydro-
gen evolving complex, a biomimetic diiron hexacarbonyl
active center. Hydrogenase mimics based on the diironhexacar-
bonyl moiety present low activity and are generally unstable.
We found that photocatalytic activity expressed as TON for
StrepH2 is about ꢀ8 times higher than for the biotinylated
complex BiotH2. The increase in TON is largely due to in-
creased lifetime of the catalytically competent intermediate
form of the diiron hexacarbonyl complex, FeIFe0, in the case of
StrepH2 but not of BiotH2, as demonstrated by transient spec-
troscopy. The origins of this increase in lifetime are not entirely
clear. It is possible that interactions of the complex with
nearby residues serve to stabilize the reduced form, and pre-
vent possible side reactions, in analogy with what observed
upon incorporation of diiron hexacarbonyl complexes in
MOFs[26] In addition, the proximity of Lys106 may provide a
proton relay that facilitates generation of molecular hydrogen.
The crystal structure of StrepH2 provides information for fur-
ther optimization of the system by engineering mutations in
proximity to the iron-sulfur complex. Such mutations could in-
crease activity for example by providing a proton relay during
the catalytic cycle, or by stabilizing the relevant redox state of
the metal center.[63–67] Finally, we note that the streptavidin
scaffold provides a robust environment, which is stable upon
immobilization on electrodes. This feature may provide addi-
tional avenues for exploring light-aided electrocatalysis.
All electrochemical studies were performed on a CH-Instruments
model 1242B electrochemical workstation using a SCE reference
electrode, a Pt-mesh counter electrode and a glassy carbon elec-
trode (3 mm diameter) in a Coy anaerobic chamber. Cyclic Voltam-
metry scans were performed in 100 mm acetate, 50 mm NaCl
buffer of desired pH (4.5 and 6.0) at 100 mVsÀ1 scan rate. Working
electrode was cleaned in between different samples and a blank
was recorded to ensure cleanliness of the electrode. For protein
film voltammetry, 1 mL of 100 mm catalyst was placed directly on
the active electrode surface and dried under vacuum. This elec-
trode was then placed into fresh buffer of different pH and CVs
were recorded.
Photoinduced hydrogen production
Photocatalytic hydrogen was monitored by quantifying the
evolved hydrogen gas by gas-chromatography. For these experi-
ments, solutions containing 150 mm Ru(Bpy)32+, 100 mm ascorbate
pH 4.5 and the catalyst at desired concentration were placed in a
custom made gas tight cuvette (0.1 cm pathlength), and degassed
extensively. The cuvette was then irradiated with a white LED (Phi-
lips Luxeon S5000 part number LXS9-PW27-0041), and the light in-
tensity at the sample was maintained at 1100 WmÀ2 for all the ex-
periments (one sun). The amount of hydrogen evolved as a func-
tion of time was monitored by injecting 100 mL of argon, sampling
100 mL of the overhead gas with a gas tight syringe, and analyzing
it on a GC (SRI instruments, Model no. 310C) using a 5 ꢁ molecular
sieve column, a thermal conductivity detector and Ar as carrier
gas. The GC instrument was calibrated using a gas standard (1%
H2 and bulk N2, Figure S2).
Transient absorption measurements
Transient absorption spectra and kinetics on the ps-ms time scale
were obtained using EOS spectrometer (Ultrafast Systems). Excita-
tion at 480 nm was from an optical parametric amplifier (Spectra
Physics) pumped with laser pulses of 100 fs at 800 nm generated
by an amplified, mode-locked titanium sapphire laser system (Mil-
lennia/Tsunami/Spitfire, Spectra Physics) operating at 1 kHz or
100 Hz repetition rate and allowing to utilize the 400 ms and 9 ms
EOS spectrometer kinetics decay time windows, respectively. The
instrument response function (IRF) was ca. 800 ps.
Crystallization of streptavidin-BiotH complex
Experimental Section
All manipulations for crystallography were carried out in a Coy
anaerobic chamber (95% N2; 5% H2). The complex, StrepHyd1, was
crystallized using hanging drop vapor diffusion techniques. The
protein-organometallic complex was obtained by mixing
27 mgmLÀ1 of streptavidin with 3 molar excess of BiotH. The re-
sulting solution was centrifuged for 10 minutes at 10,000 g to
remove any precipitated protein. The composition of the reservoir
solution for successful crystallization was as follows: 100 mm TRIS
pH 8.5; 200 mm MgCl2; 25% PEG4000. Drops of the protein and
ligand solution were mixed with an equal volume of the reservoir
solution before equilibration. Crystal growth was observed within
4–7 days. The yellow colored crystals were subsequently harvested
Synthesis of 3
Compound 3 (BiotH) was synthesized by coupling d-Biotin and 2
in DMF using HATU as coupling auxiliary. 180.56 mg (0.74 mmol) of
biotin, 308 mg (0.74 mmol) of compound 3, and 281.37 mg
(0.74 mmol) of HATU were dissolved in 6 mL of anhydrous DMF
and 185.22 mg of dry DIPEA was added to the mixture. The result-
ing mixture was stirred under an inert atmosphere overnight, and
DMF was evaporated under reduced pressure; the yellowish red
crude oil was purified on a silica gel column using 10% MeOH in
1
DCM (RF: 0.65). H NMR (DMSO, d [ppm]): 1.19–1.67 (m, 9H), 2.28–
2.34 (m, 2H), 2.59 (d, 1H), 2.77 (dd, 2H), 3.09 (m, 1H), 3.62 (m, 2H),
Chem. Eur. J. 2020, 26, 1 – 8
www.chemeurj.org
5
ꢀ 2020 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ

Full Paper
doi.org/10.1002/chem.202000204
Chemistry—A European Journal
[12] J. M. Camara, T. B. Rauchfuss, Nat Chem. 2012, 4, 26–30.
[13] Y. Li, T. B. Rauchfuss, Chem. Rev. 2016, 116, 7043–7077.
[14] U.-P. Apfel, C. R. Kowol, Y. Halpin, F. Kloss, J. Kꢃbel, H. Gorls, J. G. Vos,
B. K. Keppler, E. Morera, G. Lucente, et al., J. Inorg. Biochem. 2009, 103,
1236–1244.
using cryo-loops (Hampton Research), and cryo-cooled by plunging
in a liquid nitrogen bath.
Crystallographic data collection, processing, and structure
solution
[15] U.-P. Apfel, C. R. Kowol, E. Morera, H. Gorls, G. Lucente, B. K. Keppler, W.
Weigand, Eur. J. Inorg. Chem. 2010, 5079–5086.
Crystallographic data collection was performed at the Advanced
Light Source (ALS) at Lawrence Berkeley National Lab beamline
8.2.2. Diffraction data were collected on an ADSC Q315R detector
at l=1.0 ꢁ. 180ꢂ18 oscillation frames were collected and were in-
dexed, refined, integrated, and scaled using the HKL2000 pack-
age.[68] Initial crystallographic phases were determined by molecu-
lar replacement using wild-type core streptavidin (PDB: 1MK5) as
the search model. The structure was built upon a single solution,
and model building and refinement were performed using PHENIX,
CCP4, and COOT.[69–71] Calculations of Rfree for each of the datasets
were based upon 5% of the unique reflections, with a final model
that was in very good agreement with the data with Rwork and Rfree
of 16.5% and 21.8%. Crystallographic and refinement statistics are
given in Table 1, Supporting Information. The atomic coordinates
and structure factors have been deposited in the Protein Data
Bank (PDB) with accession code 5VCQ. Both structures shown in
Figures 1 and 3 were generated with PYMOL.
[16] E. Morera, M. Nalli, F. Pinnen, D. Rossi, G. Lucente, Bioorg. Med. Chem.
Lett. 2000, 10, 1585–1588.
[17] A. M. Kluwer, R. Kapre, F. Hartl, M. Lutz, A. L. Spek, A. M. Brouwer,
P. W. N. M. van Leeuwen, J. N. H. Reek, Proc. Natl. Acad. Sci. USA 2009,
106, 10460–10465.
[18] L.-C. Song, L.-X. Wang, M.-Y. Tang, C.-G. Li, H.-B. Song, Q.-M. Hu, Organo-
metallics 2009, 28, 3834–3841.
[19] D. Streich, Y. Astuti, M. Orlandi, L. Schwartz, R. Lomoth, L. Hammar-
strom, S. Ott, Chem. Eur. J. 2010, 16, 60–63.
[20] F. Wang, W.-G. Wang, X.-J. Wang, H.-Y. Wang, C.-H. Tung, L.-Z. Wu,
Angew. Chem. Int. Ed. 2011, 50, 3193–3197; Angew. Chem. 2011, 123,
3251–3255.
[21] M. Wang, L. Chen, X. Li, L. Sun, Dalton Trans. 2011, 40, 12793.
[22] H.-Y. Wang, G. Si, W.-N. Cao, W.-G. Wang, Z.-J. Li, F. Wang, C.-H. Tung, L.-
Z. Wu, Chem. Commun. 2011, 47, 8406–8408.
[23] P. Zhang, M. Wang, Y. Na, X. Li, Y. Jiang, L. Sun, Dalton Trans. 2010, 39,
1204–1206.
[24] X. Li, M. Wang, D. Zheng, K. Han, J. Dong, Energy Environ. Sci. 2012, 5,
8220-8224.
[25] Y. Na, M. Wang, J. Pan, P. Zhang, B. Akermark, L. Sun, Inorg. Chem. 2008,
47, 2805–2810.
Acknowledgements
[26] S. Pullen, H. Fei, A. Orthaber, S. M. Cohen, S. Ott, J. Am. Chem. Soc.
2013, 135, 16997–17003.
[27] S. R. Soltau, P. D. Dahlberg, J. Niklas, O. G. Poluektov, K. L. Mulfort, L. M.
Utschig, Chem. Sci. 2016, 7, 7068–7078.
[28] R. Goy, L. Bertini, T. Rudolph, S. Lin, M. Schulz, G. Zampella, B. Dietzek,
F. H. Schacher, L. De Gioia, K. Sakai, W. Weigand, Chem. Eur. J. 2017, 23,
334–345.
This work was supported by NSF award CHE-1508301. We
thank Dr. Y. Su at the Molecular mass spectrometry facility,
UCSD for HRMS analysis, and Dr. Brian Cherry for assistance
with NMR data collection.
[29] K. Menzel, U.-P. Apfel, N. Wolter, R. Rꢃger, T. Alpermann, F. Steiniger, D.
Gabel, S. Fçrster, W. Weigand, A. Fahr, J. Liposome Res. 2014, 24, 59–68.
[30] M. Y. Darensbourg, E. J. Lyon, X. Zhao, I. P. Georgakaki, Proc. Natl. Acad.
Sci. USA 2003, 100, 3683–3688.
[31] D. W. Mulder, E. M. Shepard, J. E. Meuser, N. Joshi, P. W. King, M. C. Pose-
witz, J. B. Broderick, J. W. Peters, Structure 2011, 19, 1038–1052.
[32] A. K. Jones, B. R. Lichtenstein, A. Dutta, G. Gordon, P. L. Dutton, J. Am.
Chem. Soc. 2007, 129, 14844–14845.
Conflict of interest
The authors declare no conflict of interest.
Keywords: catalysis · hybrid enzymes · hydrogen production ·
streptavidin-avidin · X-ray crystallography
[33] G. Caserta, S. Roy, M. Atta, V. Artero, M. Fontecave, Curr. Opin. Chem.
Biol. Sect. C 2015, 25, 36–47.
[1] J. C. Fontecilla-Camps, A. Volbeda, C. Cavazza, Y. Nicolet, Chem. Rev.
2007, 107, 4273–4303.
[2] Y. Nicolet, C. Piras, P. Legrand, C. E. Hatchikian, J. C. Fontecilla-Camps,
Structure 1999, 7, 13–23.
[3] M. W. W. Adams, Biochim. Biophys. Acta 1990, 1020, 115–145.
[4] P. Knçrzer, A. Silakov, C. E. Foster, F. A. Armstrong, W. Lubitz, T. Happe, J.
Biol. Chem. 2012, 287, 1489–1499.
[5] M. Y. Darensbourg, R. D. Bethel, Nat. Chem. 2012, 4, 11–13.
[6] G. Berggren, A. Adamska, C. Lambertz, T. R. Simmons, J. Esselborn, M.
Atta, S. Gambarelli, J.-M. Mouesca, E. Reijerse, W. Lubitz, T. Happe, V.
Artero, M. Fontecave, Nature 2013, 499, 66–69.
[34] Y. Sano, A. Onoda, T. Hayashi, Chem. Commun. 2011, 47, 8229–8231.
[35] A. Roy, C. Madden, G. Ghirlanda, Chem. Commun. 2012, 48, 9816–9818.
[36] S. Roy, S. Shinde, G. A. Hamilton, H. E. Hartnett, A. K. Jones, Eur. J. Inorg.
Chem. 2011, 1050–1055.
[37] A. Onoda, Y. Kihara, K. Fukumoto, Y. Sano, T. Hayashi, ACS Catal. 2014, 4,
2645–2648.
[38] M. E. Wilson, G. M. Whitesides, J. Am. Chem. Soc. 1978, 100, 306–307.
[39] M. Skander, N. Humbert, J. Collot, J. Gradinaru, G. Klein, A. Loosli, J.
Sauser, A. Zocchi, F. Gilardoni, T. R. Ward, J. Am. Chem. Soc. 2004, 126,
14411–14418.
[40] G. Klein, N. Humbert, J. Gradinaru, A. Ivanova, F. Gilardoni, U. E. Rusban-
di, T. R. Ward, Angew. Chem. Int. Ed. 2005, 44, 7764–7767; Angew. Chem.
2005, 117, 7942–7945.
[41] M. Dꢃrrenberger, T. Heinisch, Y. M. Wilson, T. Rossel, E. Nogueira, L.
Knçrr, A. Mutschler, K. Kersten, M. J. Zimbron, J. Pierron, et al., Angew.
Chem. Int. Ed. 2011, 50, 3026–3029; Angew. Chem. 2011, 123, 3082–
3085.
[7] J. F. Siebel, A. Adamska-Venkatesh, K. Weber, S. Rumpel, E. Reijerse, W.
Lubitz, Biochemistry 2015, 54, 1474–1483.
[8] J. Esselborn, C. Lambertz, A. Adamska-Venkatesh, T. Simmons, G. Bergg-
ren, J. Noth, J. Siebel, A. Hemschemeier, V. Artero, E. Reijerse, M. Fonte-
cave, W. Lubitz, T. Happe, Nat. Chem. Biol. 2013, 9, 607–609.
[9] L. Kertess, F. Wittkamp, C. Sommer, J. Esselborn, O. Rudiger, E. J. Rei-
jerse, E. Hofmann, W. Lubitz, M. Winkler, T. Happe, U.-P. Apfel, Dalton
Trans. 2017, 46, 16947–16958.
[10] C. Sommer, C. P. Richers, W. Lubitz, T. B. Rauchfuss, E. J. Reijerse, Angew.
Chem. Int. Ed. 2018, 57, 5429–5432; Angew. Chem. 2018, 130, 5527–
5530.
[42] J. M. Zimbron, T. Heinisch, M. Schmid, D. Hamels, E. S. Nogueira, T.
Schirmer, T. R. Ward, J. Am. Chem. Soc. 2013, 135, 5384–5388.
[43] V. Kçhler, Y. M. Wilson, M. Dꢃrrenberger, D. Ghislieri, E. Churakova, T.
Quinto, L. Knçrr, D. Hꢄussinger, F. Hollmann, N. J. Turner, T. R. Ward, Nat.
Chem. 2013, 5, 93–99.
[11] C. Papini, C. Sommer, L. Pecqueur, D. Pramanik, S. Roy, E. J. Reijerse, F.
Wittkamp, V. Artero, W. Lubitz, M. Fontecave, ACS Catal. 2019, 9, 4495–
4501.
[44] S. I. Mann, T. Heinisch, A. C. Weitz, M. P. Hendrich, T. R. Ward, A. S. Boro-
vik, J. Am. Chem. Soc. 2016, 138, 9073–9076.
&
&
Chem. Eur. J. 2020, 26, 1 – 8
www.chemeurj.org
6
ꢀ 2020 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Paper
doi.org/10.1002/chem.202000204
Chemistry—A European Journal
[45] A. Call, C. Casadevall, A. Romero-Rivera, V. Martin-Diaconescu, D. J.
Sommer, S. Osuna, G. Ghirlanda, J. Lloret-Fillol, ACS Catal. 2019, 9,
5837–5846.
[46] S. G. Keller, B. Probst, T. Heinisch, R. Alberto, T. R. Ward, Helv. Chim. Acta
2018, 101, e1800036.
[61] M. Dꢃrrenberger, T. R. Ward, Curr. Opin. Chem. Biol. Sect. C 2014, 19, 99–
106.
[62] V. Kçhler, J. Mao, T. Heinisch, A. Pordea, A. Sardo, Y. M. Wilson, L. Knçrr,
M. Creus, J.-C. Prost, T. Schirmer, T. R. Ward, Angew. Chem. Int. Ed. 2011,
50, 10863–10866; Angew. Chem. 2011, 123, 11055–11058.
[63] F. Schwizer, V. Kçhler, M. Dꢃrrenberger, L. Knçrr, T. R. Ward, ACS Catal.
2013, 3, 1752–1755.
[47] A. Winter, L. Zsolnai, G. Z. Huttner, Naturforsch. B Anorg. Chem. Org.
Chem. 1982, 37, 1430–1436.
[48] U.-P. Apfel, Y. Halpin, H. Gorls, J. G. Vos, B. Schweizer, G. Linti, W. Wei-
gand, Chem. Biodiversity 2007, 4, 2138–2148.
[49] F. Zhuanli Shenqing (2017), CN 106632512 a 20170510., n.d.
[50] M. R. Hicks, A. K. Rullay, R. Pedrido, D. H. Crout, T. J. T. Pinheiro, Synth.
Commun. 2008, 38, 3726–3750.
[64] T. Heinisch, M. Pellizzoni, M. Dꢃrrenberger, C. E. Tinberg, V. Kçhler, J.
Klehr, D. Hꢄussinger, D. Baker, T. R. Ward, J. Am. Chem. Soc. 2015, 137,
10414–10419.
[65] C. Lo, M. R. Ringenberg, D. Gnandt, Y. Wilson, T. R. Ward, Chem.
Commun. 2011, 47, 12065–12067.
[51] J. Brown-McDonald, S. Berg, M. Peralto, C. Works, Inorg. Chim. Acta
2009, 362, 318–324.
[66] K. K.-W. Lo, W.-K. Hui, D. C.-M. Ng, J. Am. Chem. Soc. 2002, 124, 9344–
9345.
[52] M. G. I. Galinato, C. M. Whaley, N. Lehnert, Inorg. Chem. 2010, 49, 3201–
3215.
[53] M. L. Singleton, J. H. Reibenspies, M. Y. Darensbourg, J. Am. Chem. Soc.
2010, 132, 8870–8871.
[54] W.-N. Cao, F. Wang, H.-Y. Wang, B. Chen, K. Feng, C.-H. Tung, L.-Z. Wu,
Chem. Commun. 2012, 48, 8081–8083.
[55] Y. Na, J. Pan, A. M. Wang, L. Sun, Inorg. Chem. 2007, 46, 3813–3815.
[56] M. Sjçdin, S. Styring, B. Akermark, L. Sun, L. Hammarstrom, J. Am. Chem.
Soc. 2000, 122, 3932–3936.
[67] K. K.-W. Lo, W.-K. Hui, Inorg. Chem. 2005, 44, 1992–2002.
[68] Z. Otwinowski, W. Minor, Meth. Enzymol. 1997, 276, 307–326.
[69] P. D. Adams, R. W. Grosse Kunstleve, L.-W. Hung, T. R. Ioerger, A. J.
McCoy, N. W. Moriarty, R. J. Read, J. C. Sacchettini, N. K. Sauter, T. C. Ter-
williger, Acta Crystallogr. Sect. D 2002, 58, 1948–1954.
[70] Collaborative Computational Project, Number 4, Acta Crystallogr. Sect. D
1994, 50, 760–763.
[71] P. Emsley, K. Cowtan, Acta Crystallogr. Sect. D 2004, 60, 2126–2132.
[57] H. Murakami, T. Hohsaka, Y. Ashizuka, M. Sisido, J. Am. Chem. Soc. 1998,
120, 7520–7529.
[58] A. Fꢃrstenberg, O. Kel, J. Gradinaru, T. R. Ward, D. Emery, G. Bollot, J.
Mareda, E. Vauthey, ChemPhysChem 2009, 10, 1517–1532.
[59] T. R. Ward, Acc. Chem. Res. 2011, 44, 47–57.
Manuscript received: January 14, 2020
Revised manuscript received: February 24, 2020
Accepted manuscript online: March 22, 2020
[60] T. Heinisch, T. R. Ward, Acc. Chem. Res. 2016, 49, 1711–1721.
Version of record online: && &&, 0000
Chem. Eur. J. 2020, 26, 1 – 8
www.chemeurj.org
7
ꢀ 2020 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ

Full Paper
doi.org/10.1002/chem.202000204
Chemistry—A European Journal
FULL PAPER
An artificial hydrogenase, StrepH2, was
built by incorporating a biotinylated
[Fe–Fe] hydrogenase organometallic
mimic within streptavidin. Under photo-
catalytic conditions, the protein-embed-
ded catalyst shows enhanced efficiency
and prolonged activity compared to the
isolated catalyst (see figure).
&
Photochemistry
A. Roy, M. D. Vaughn, J. Tomlin,
G. J. Booher, G. Kodis, C. R. Simmons,
J. P. Allen, G. Ghirlanda*
&& – &&
Enhanced Photocatalytic Hydrogen
Production by Hybrid Streptavidin-
Diiron Catalysts
&
&
Chem. Eur. J. 2020, 26, 1 – 8
www.chemeurj.org
8
ꢀ 2020 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!