NANO
LETTERS
Biomimetic Synthesis of a H2 Catalyst
Using a Protein Cage Architecture
2005
Vol. 5, No. 11
2306-2309
Z. Varpness,†,‡ J. W. Peters,†,‡ M. Young,‡,§ and T. Douglas*,†,‡
Department of Chemistry & Biochemistry, Department of Plant Sciences, and Center
for Bio-Inspired Nanomaterials, Montana State UniVersity, Bozeman, Montana 59717
Received September 2, 2005
ABSTRACT
A biomimetic approach has been used to develop an artificial hydrogenase that catalyses the efficient reduction of protons producing hydrogen
gas. Analogous to the unique biological metal clusters found in hydrogenase enzymes, the engineered active sites are small, well-defined Pt
clusters deposited on the interior of a heat shock protein cage architecture with stoichiometries of 150 to 1000 Pt per protein cage. The proton
reduction reaction is driven by visible light through a coupled reaction with Ru(bpy)32+ and methyl viologen as an electron-transfer mediator.
Hydrogen production rates are comparable to those of hydrogenase on a per protein basis and exceed production rates of other reported
Pt-based catalysts. These results demonstrate the utility of a biomimetic approach toward addressing the needs of hydrogen production.
There is significant interest in the use of hydrogen gas as
an alternative fuel. The practicality of the increased use of
hydrogen fuel cell technologies is dependent on the ability
to produce stores of hydrogen gas in an efficient, economically feasible, and environmentally sound manner.1-4 The
majority of hydrogen produced for energy yielding applications is generated by the process of reforming methane or
fossil fuels, and thus a hydrogen energy economy based on
these approaches does little to reduce the dependence on
nonrenewable fossils fuels. Therefore, the development of
catalysts for the production of hydrogen gas efficiently from
a renewable source is of paramount importance.
Biological production of hydrogen is likely to play a key
role in the emerging hydrogen economy.5-7 In particular, a
group of enzymes, hydrogenases (H2ases), have attracted a
great deal of attention because of their ability to efficiently
catalyze the reversible reduction of protons to form H2. While
rates of up to 9000 H2 per enzyme per second have been
observed8,9 from these enzymes, the extreme sensitivity of
these enzymes to oxygen, limited expression, and difficult
isolation have hindered their use as a practical means of
hydrogen production.
Hydrogenase enzymes are produced by a variety of
microorganisms where they function in either hydrogen
oxidation or proton reduction. Hydrogen oxidation is coupled
to the generation of reducing equivalents to drive energy
yielding or biosynthetic processes.8,10-12 During anaerobic
* To
whom
correspondence
should
[email protected].
† Department of Chemistry & Biochemistry.
‡ Center for Bio-Inspired Nanomaterials.
§ Department of Plant Sciences.
10.1021/nl0517619 CCC: $30.25
Published on Web 10/19/2005
be
addressed:
© 2005 American Chemical Society
fermentation, some microorganisms are capable of coupling
the oxidization of sugars to the regeneration of electron
carriers, which are used for proton reduction and the
production of hydrogen gas. Hydrogenases, like enzymes in
general, are assembled to display precise organizational
motifs that use their protein architecture to position and
chemically poise an active site in which pathways for
substrate access and product removal are key “design”
features. Substrate, reductant, and product must have access
to and from the catalytic site. Furthermore, continuous
cycling of the catalyst requires ongoing addition of reactants and removal of products. The catalytic site of hydrogenase enzymes consists of unique biological metal clusters
(Fe or NiFe) with carbon monoxide and cyanide ligands.10,13-16
A major goal of biomimetic chemistry is to couple enzyme
design principles with synthetic capacity to create new
functional materials showing dramatic property enhancements.17,18 Artificial enzyme systems can combine the
advantages of biology with the directed synthesis capacity
of chemistry. Through this approach, hard inorganic materials
(metals) can be interfaced with soft biological materials
(proteins) to create new composite functionalities.
With the high cost and limited supply of platinum, it is
necessary to explore ways of maximizing the catalytic
efficiency of Pt on a per atom basis in order to develop
economically feasible catalysts.2 In a particle based approach
toward developing a Pt catalyst, it is necessary to minimize
the diameter of the particle and thus increase the surface
area (i.e., the number of exposed Pt atoms per particle). A
number of different synthetic approaches have been used to
Figure 1. (A) Space filling representation of the small heat shock
protein (Hsp) cage from Methanococcus jannaschii (pdb: 1shs).
(B) Cut-away view of Hsp showing the interior cavity of the cage.
synthesize platinum nanoparticles with different passivating
layers.19-30 The passivating layer generally interferes with
the exposed Pt atoms and reduces efficiency.19 In this study,
we have employed a protein cage as a synthetic platform
which, unlike a passivating layer, does not coat the entire
surface of the nanoparticles but still isolates the particle in
solution and prevents aggregation.
Protein cage architectures have been used as biotemplates
to create interfaces between proteins and metals.31,32 Protein
cage architectures are self-assembled from a limited number
of protein subunits to create well-defined, container-like
morphologies in which the interior and exterior surfaces can
be chemically distinct.17,32 In addition, molecular access to
the interior can be controlled by pores at the subunit
interfaces.33 We have previously shown that protein cage
architectures can be utilized as size- and shape-constrained
reaction environments for nanomaterials synthesis.17,18,31,32,34
These cages have been shown to stabilize inorganic nanoparticles in defined sizes and crystal forms. In addition,
through genetic and chemical modifications, active sites can
be created at precise locations within the cage architecture
to create dramatically new functionality.35
We have adopted a biomimetic approach to create a
protein-based catalyst that functions as a hydrogenase mimic.
The self-assembled cage-like architecture of the small heat
shock protein (Hsp) from Methanococcus jannaschii has been
used to encapsulate metal clusters with a defined spatial
arrangement.32 Hsp assembles from 24 subunits into a 12
nm cage defining a 6.5 nm interior cavity with pores through
the cage architecture, by which molecules can shuttle
between the inside and outside environment (Figure 1).33
Cage-like architectures have previously been shown to act
as a molecular container for the encapsulation of both organic
and inorganic materials.32,36 By using the interior of the cage
for spatially selective mineralization, the protein architecture
of the Hsp mimics the controlled molecular access to the
active site displayed in H2ase.
In the present work, we describe the synthesis of Pt
nanoparticles encapsulated within the Hsp cage. Briefly
outlined, purified Hsp was incubated with PtCl42- at 65 °C
for 15 min. The protein cage was incubated with either 150,
250, or 1000 Pt per protein cage. Subsequent reduction with
dimethylamine borane complex ((CH3)2NBH3) resulted in the
formation of a brown solution. Characterization of the
reaction product by size exclusion chromatography revealed
retention volumes identical with untreated Hsp and showed
Nano Lett., Vol. 5, No. 11, 2005
Figure 2. Size exclusion chromatography of Hsp: (A) unmineralized Hsp; (B) Hsp mineralized with 250 Pt/Hsp showing coelution
of protein (280 nm) and mineral (350 nm); (C) Hsp mineralized
with 1000 Pt/Hsp showing coelution of protein (280 nm) and
mineral (350 nm).
Figure 3. (A) TEM of Hsp 1000 Pt unstained. The inset shows
electron diffraction of Pt0 from Hsp 1000 Pt. (B) TEM of Hsp 1000
Pt stained with 2% uranyl acetate. (C) Histogram of Pt particle
diameters in Hsp 1000 Pt. Average 2.2 ( 0.7 nm. Scale bars ) 20
nm.
coelution of protein (280 nm) and Pt (350 nm) components
(Figure 2). Dynamic light scattering indicated no change in
the particle diameter after the reaction. Visualization of the
Pt-treated Hsp (Pt-Hsp) by transmission electron microscopy
(TEM) revealed electron dense cores identified as Pt metal
by electron diffraction (Figure 3A inset) and intact 12 nm
protein cages when negatively stained (Figure 3B). In the
stained samples, the Pt particles can clearly be seen above
the background stain and are localized within the cage
structure. For average loadings of 1000 Pt/cage, metal
particles of 2.2 ( 0.7 nm were observed (Figure 3C). At
theoretical loadings of 250 Pt/cage, particles of 1 ( 0.2 nm
were observed (Figure 4), while at loadings of 150 Pt/cage,
no particles could be distinguished due to the limitation of
the electron microscope. Hsp-free control reactions resulted
in the formation of aggregated Pt colloids, which rapidly
precipitated from solution. Control reactions using bovine
serum albumin (BSA), at the same total protein concentration, also resulted in bulk precipitation and only a fraction
of the Pt remained in solution with a wide distribution of
particle sizes (3-120 nm) when observed by TEM (Supporting Information).
The Pt-Hsp protein cage composites are highly active
artificial catalysts able to reduce H+ to form H2 at rates
2307
Figure 6. H2 production from Pt-Hsp in 0.2 mM Ru(bpy)32+,
0.5 mM methyl viologen, 200 mM EDTA, and 500 mM acetate
pH 5.0: blue triangles, 1000 Pt/Hsp 5.1 × 10-10 mol Pt; red squares,
250 Pt/Hsp 8.2 × 10-10 mol Pt.
Figure 4. (A) TEM of Hsp 250 Pt stained with 2% uranyl acetate.
(B) Histogram of Pt particle diameters in Hsp 250 Pt. Scale bar )
20 nm.
Figure 5. Schematic of the light-mediated H2 production from PtHsp. Methyl viologen (MV2+) is used as an electron-transfer
mediator between the Ru(bpy)32+ photocatalyst and the Pt-Hsp
responsible for H2 production.
comparable to the highly efficient hydrogenase enzymes.
Typical of hydrogenase assays, we used reduced methyl
viologen (MV+) as a source of reducing equivalents to drive
the reaction. Unlike most in Vitro hydrogenase assays, which
use dithionite to generate MV+, we have used visible light
and a cocatalyst (Ru(bpy)32+) to generate MV+ through
oxidation of simple organics such as EDTA19,24 (Figure 5).
In this modified assay, the solution was illuminated at 25
°C with a 150 W Xe arc lamp equipped with an IR filter
and a UV cutoff filter (<360 nm). The Pt-Hsp (0.51 µM)
was illuminated in the presence of MV2+ (0.5 mM), Ru(bpy)32+ (0.2 mM), and EDTA (200 mM) at pH 5.0, and the
resulting H2 was quantified by gas chromatography. When
calculated on a per cage basis, the initial rate of H2 formation
was 4.47 × 103 H2/s ((394 H2/s) for a loading factor of
1000 Pt per Hsp and 7.63 × 102 H2/s ((405 H2/s) for a
loading factor of 250 (Figure 6). These rates are comparable
to those reported for hydogenase enzymes (4 × 103 to 9 ×
103 H2/s per protein molecule).8
No H2-producing activity was detected for the lowest Pt
loading (150 Pt/Hsp). This is consistent with previous reports
of size-dependent activity of Pt.37 It is also consistent with
our inability to detect discrete Pt particles by TEM in these
samples, implying that the synthesized Pt particles are below
some threshold limit required for activity.
When the H2 production rates are calculated on a per Pt
basis, they compare very favorably with other reported Pt
nanoparticles. The initial rates for Pt-Hsp with 1000 Pt/
cage are 268 H2/Pt/min, which is significantly better than
2308
reported literature values (20 H2/Pt/min,19 16 H2/Pt/min,25
and 6.5 H2/Pt/min27), where comparisons are possible. In
addition, initial H2 production rates for Pt-Hsp are approximately 20-fold greater than those obtained for the Pt
particles produced in protein-free control reactions. The longterm stability of the coupled photochemical reaction to
produce H2 has not been optimized, and a significant slowing
down of the reaction is observed after the first 20 min (Figure
6). The H2 production decay is mainly due to the degradation
of the photocatalyst (Ru(bpy)32+, Supporting Information),
and the electron mediator (MV2+), which is subject to Ptcatalyzed hydrogenation.25
Unlike the hydrogenase enzymes, the artificial Pt-Hsp
systems are not sensitive to O2 and show no significant
inhibition of H2 production by CO but are poisoned by thiols.
The Pt-Hsp catalyzed reaction was driven by the presence
of the reduced viologen (MV+). The MV+ could be generated
either by the photoreduction described above or by using
the Jones reductor38 (Zn amalgam), which yielded rates for
H2 production approximately 40% slower than the coupled
photoreduction reactions. Also, the Pt-Hsp is able to catalyze
the reverse reaction (H2 f 2H+ + 2e-) as monitored by the
in situ reduction and bleaching of methylene blue.39 Importantly for the utility of this artificial system, the Pt-Hsp
construct is remarkably stable and can be heated to 85 °C
without precipitation of the composite or loss of the catalytic
activity.
We have used a well-defined thermally stable protein cage
architecture to generate an artificial hydrogenase having
many of the features common to those biological catalysts.
We have introduced small metal clusters in a spatially
selective manner to the interior of the cage-like structure of
Hsp that act as active sites for the reduction of H+ to form
H2. The specific activities of these artificial enzymes are
comparable to known hydrogenase enzymes and significantly
better than previously described Pt nanoparticles. The protein
cage architecture of Hsp acts to maintain the integrity of the
small clusters, preventing agglomeration, and controlling
access to these “active sites”. The Pt-Hsp composite is stable
up to 85 °C illustrating the utility of using protein architectures for the design and implementation of functional
nanomaterials.
Nano Lett., Vol. 5, No. 11, 2005
Acknowledgment. This work was funded in part by
grants from the Office of Naval Research (19-00-R0006).
Molecular graphics images were produced using the UCSF
Chimera package from the Resource for Biocomputing,
Visualization, and Informatics at the University of California,
San Francisco (supported by NIH P41 RR-01081).
Supporting Information Available: Text of experimental
details and figures of a TEM photomicrograph of BSA
control with 1000 Pt per protein and UV-vis spectra of Ru(bpy)32+ photodegradation. This material is available free of
charge via the Internet at http://pubs.acs.org.
References
(1) Armor, J. N. Catal. Lett. 2005, 101, 131-135.
(2) Spiegel, R. J. Transp. Res., Part D: Transp. EnViron. 2004, 9, 357371.
(3) Tseng, P.; Lee, J.; Friley, P. Energy 2005, 30, 2703-2720.
(4) Winter, C. J. Int. J. Hydrogen Energy 2005, 30, 681-685.
(5) Varfolomeyev, S. D.; Aliev, T. K.; Efremenko, E. N. Pure Appl.
Chem. 2004, 76, 1781-1798.
(6) Wunschiers, R.; Lindblad, P. Int. J. Hydrogen Energy 2002, 27,
1131-1140.
(7) Chum, H. L.; Overend, R. P. Fuel Process. Technol. 2001, 71, 187195.
(8) Adams, M. W. Biochim. Biophys. Acta 1990, 1020, 115-45.
(9) Cammack, R. Nature 1999, 397, 214-5.
(10) Peters, J. W.; Lanzilatta, W. N.; Lemon, B. J.; Seefeldt, L. C. Science
1998, 282, 1853-1858.
(11) Przybyla, A. E.; Robbins, J.; Menon, N.; Peck, H. D., Jr. FEMS
Microbiol. ReV. 1992, 8, 109-35.
(12) Vignais, P. M.; Billoud, B.; Meyer, J. FEMS Microbiol. ReV. 2001,
25, 455-501.
(13) Happe, R. P.; Roseboom, W.; Pierik, A. J.; Albracht, S. P.; Bagley,
K. A. Nature 1997, 385, 126.
(14) Nicolet, Y.; Lemon, B. J.; Fontecilla-Camps, J. C.; Peters, J. W.
Trends Biochem. Sci. 2000, 25, 138-43.
(15) Pierik, A. J.; Hulstein, M.; Hagen, W. R.; Albracht, S. P. Eur. J.
Biochem. 1998, 258, 572-8.
(16) Volbeda, A.; Charon, M. H.; Piras, C.; Hatchikian, E. C.; Frey, M.;
Fontecilla-Camps, J. C. Nature 1995, 373, 580-7.
Nano Lett., Vol. 5, No. 11, 2005
(17) Douglas, T.; Young, M. Nature 1998, 393, 152-155.
(18) Usselman, R. J.; Klem, M.; Allen, M.; Walter, E. D.; Gilmore, K.;
Douglas, T.; Young, M.; Idzerda, Y.; Singel, D. J. J. Appl. Phys.
2005, 97.
(19) Brugger, P. A.; Cuendet, P.; Gratzel, M. J. Am. Chem. Soc. 1981,
103, 2923-2927.
(20) Chen, C. W.; Tano, D.; Akashi, M. J. Colloid Interface Sci. 2000,
225, 349-358.
(21) Chen, D. H.; Yeh, J. J.; Huang, T. C. J. Colloid Interface Sci. 1999,
215, 159-166.
(22) Eklund, S. E.; Cliffel, D. E. Langmuir 2004, 20, 6012-6018.
(23) Gomez, S.; Erades, L.; Philippot, K.; Chaudret, B.; Colliere, V.;
Balmes, O.; Bovin, J. O. Chem. Commun. 2001, 1474-1475.
(24) Jiang, D. L.; Choi, C. K.; Honda, K.; Li, W. S.; Yuzawa, T.; Aida,
T. J. Am. Chem. Soc. 2004, 126, 12084-12089.
(25) Keller, P.; Moradpour, A. J. Am. Chem. Soc. 1980, 102, 7193-7196.
(26) Narayanan, R.; El-Sayed, M. A. Nano Lett. 2004, 4, 1343-1348.
(27) Song, Y. J.; Yang, Y.; Medforth, C. J.; Pereira, E.; Singh, A. K.;
Xu, H. F.; Jiang, Y. B.; Brinker, C. J.; van Swol, F.; Shelnutt, J. A.
J. Am. Chem. Soc. 2004, 126, 635-645.
(28) Teranishi, T.; Kurita, R.; Miyake, M. J. Inorg. Organomet. Polym.
2000, 10, 145-156.
(29) Tu, W. X.; Lin, H. F. Chem. Mater. 2000, 12, 564-567.
(30) Zhao, S. Y.; Chen, S. H.; Wang, S. Y.; Li, D. G.; Ma, H. Y. Langmuir
2002, 18, 3315-3318.
(31) Allen, M.; Willits, D.; Young, M.; Douglas, T. Inorg. Chem. 2003,
42, 6300-6305.
(32) Flenniken, M. L.; Willits, D. A.; Brumfield, S.; Young, M. J.;
Douglas, T. Nano Lett. 2003, 3, 1573-1576.
(33) Kim, K. K.; Kim, R.; Kim, S. H. Nature 1998, 394, 595-599.
(34) Allen, M.; Willits, D.; Mosolf, J.; Young, M.; Douglas, T. AdV. Mater.
2002, 14, 1562-1565.
(35) Klem, M.; Willits, D.; Solis, D.; Belcher, A.; Young, M.; Douglas,
T. AdV. Funct. Mater. 2005, 15, 1489-1494.
(36) McMillan, R. A.; Paavola, C. D.; Howard, J.; Chan, S. L.; Zaluzec,
N. J.; Trent, J. D. Nat. Mater. 2002, 1, 247-252.
(37) Greenbaum, E. J. Phys. Chem. 1988, 92, 4571-4574.
(38) Harris, D., QuantitatiVe Chemical Analysis. 5th ed.; W.H. Freeman
and Co.: New York, 1999; p 425.
(39) Seefeldt, L. C.; Arp, D. J. Biochemistry 1989, 28, 1588-1596.
NL0517619
2309