Home Search Collections Journals About Contact us My IOPscience

Structure and function of nanoparticle–protein conjugates

This article has been downloaded from IOPscience. Please scroll down to see the full text article.

2008 Biomed. Mater. 3 034001

(http://iopscience.iop.org/1748-605X/3/3/034001)

View the table of contents for this issue, or go to the journal homepage for more

Download details:
IP Address: 140.254.87.101
The article was downloaded on 20/06/2012 at 00:19

Please note that terms and conditions apply.

IOP PUBLISHING BIOMEDICAL MATERIALS
Biomed. Mater. 3 (2008) 034001 (17pp) doi:10.1088/1748-6041/3/3/034001

Structure and function of

nanoparticle–protein conjugates
M-E Aubin-Tam1 and K Hamad-Schifferli1,2
1
Department of Biological Engineering, Massachusetts Institute of Technology, 77 Massachusetts Ave,
Cambridge, MA 02139, USA
2
Department of Mechanical Engineering, Massachusetts Institute of Technology, 77 Massachusetts Ave,
Cambridge, MA 02139, USA
E-mail: [email protected]

Received 31 December 2007

Accepted for publication 12 March 2008
Published 8 August 2008
Online at stacks.iop.org/BMM/3/034001

Abstract
Conjugation of proteins to nanoparticles has numerous applications in sensing, imaging,
delivery, catalysis, therapy and control of protein structure and activity. Therefore,
characterizing the nanoparticle–protein interface is of great importance. A variety of covalent
and non-covalent linking chemistries have been reported for nanoparticle attachment.
Site-specific labeling is desirable in order to control the protein orientation on the nanoparticle,
which is crucial in many applications such as fluorescence resonance energy transfer. We
evaluate methods for successful site-specific attachment. Typically, a specific protein residue
is linked directly to the nanoparticle core or to the ligand. As conjugation often affects the
protein structure and function, techniques to probe structure and activity are assessed. We also
examine how molecular dynamics simulations of conjugates would complete those
experimental techniques in order to provide atomistic details on the effect of nanoparticle
attachment. Characterization studies of nanoparticle–protein complexes show that the
structure and function are influenced by the chemistry of the nanoparticle ligand, the
nanoparticle size, the nanoparticle material, the stoichiometry of the conjugates, the labeling
site on the protein and the nature of the linkage (covalent versus non-covalent).
(Some figures in this article are in colour only in the electronic version)

1. Introduction NP labeling of proteins remains challenging, particularly due

to the interface between the biomolecule and the nanostructure.
With the advent of nanotechnology in biology and Inorganic-biological interfaces have always been problematic
medicine, the need for conjugation of nanoparticles (NPs) to on many length scales [14]. Macroscopically, medical devices
biomolecules has grown. This has resulted in rich emerging and implants must be compatible with living tissue. On
area. In particular, the attachment of NPs to proteins has found the nanometer scale, surface effects become dramatically
applications in imaging [1–3], in catalysis [4], in delivery [5], enhanced and could impair the integrity of biomolecules. A
in the control of protein activity with an external field [6, 7] and central issue is that both the structure and function of the
in understanding a local structure in protein folding [8]. NP– protein must be maintained. Clearly, in order to create bio-
protein conjugates are not limited to biological applications, inorganic hybrid devices that properly function biologically,
but have reached many other fields, such as material science, the interface must be optimized in such a manner to preserve
physics and energy, with the design of devices [9, 10] such the biological function of the biomolecule.
as nanosensors and biofuel cells [11] and of new tools for Site-specific labeling of large and even very complex
assembly [12, 13]. proteins with small dyes has been utilized for some time now,
Most research efforts in this area are focused on and has led to many advances in dynamics and structural
developing novel applications. However, efficient and useful studies as well as in imaging both in vitro and in vivo [15]

1748-6041/08/034001+17$30.00 1 © 2008 IOP Publishing Ltd Printed in the UK

Biomed. Mater. 3 (2008) 034001 M-E Aubin-Tam and K Hamad-Schifferli

(a ) (b )

Figure 1. (a) Structure of 2.3 nm Au NP with hexanethiol ligands from molecular dynamics simulations. The red molecule represents the
least crowded thiolate, from [30]. (b) Interactions of maltose binding protein with a quantum dot. The quantum dot surface sulfur atoms are
in teal and the zinc atoms are in pink. The red shell shows the estimated outer radius of the dihydrolipoic acid ligand. The terminal Lys-370
chain to which a His-tag is cloned is represented in blue, from [29].

(e.g., fluorescence resonance energy transfer (FRET)). Site-

directed mutagenesis allows placement of the dye or spin
labels on specific amino acids on the protein. The labeling
reaction can be done in living cells using enzymes that
react specifically with a unique motif on the protein to label
[16, 17]. It is possible to even crystallize the labeled complex
and see the effect on the structure [18, 19]. This has
resulted in sophisticated understanding of cellular trafficking
and protein behavior and mechanism [15, 20, 21], along with
new applications that exploit site-specific labeling.
NPs have been of interest to the science community
for many years because of their unique material and size-
Figure 2. Parameters that affect the interaction of proteins with NPs.
dependent properties [22]. It is therefore desirable to gain
the same versatility with NPs as for small dyes. Although
tremendous progress has been made, analogous capabilities and co-workers have exploited this rearrangement of ligands
for NP labeling are challenging. NPs are considerably larger to optimize the interaction with a protein of interest [27, 28].
than dyes, and possess a lot of surface area which increase Consequently, bioconjugation strategies must consider the NP
non-specific interactions of amino acid side chains of the as a chemical system as nearly as complex as proteins [29]
protein with the NP ligand or surface. The details of these (figure 1(b)).
interactions are still unclear, because they are often numerous Furthermore, NP conjugation may result in unfolding
and difficult to physically and chemically characterize. As a of the protein. It is difficult to use most of the traditional
result, the ability to control the site of NP attachment has been techniques for assaying the protein structure. It is technically
achieved primarily for Au NPs, which have well-defined and challenging to get a crystal structure of a conjugate, which
controllable chemistry, and only on a limited set of proteins is complicated by NP polydispersity, and NMR requires
[23–25]. large sample amounts. Therefore, the question of how to
A NP is more complex than a simple hard sphere, as characterize the structure of the protein when linked to a NP
it is often approximated. Instead, it has multiple surface- and identify the nature of its interactions with the NP surface
coating ligands that have end groups and side chains that atoms and ligands remains.
can interact with the protein. NP surfaces represent a The purpose of this paper is to study the strategies to link
significant portion of the particle. Below ∼2 nm the surface- NPs to proteins and to characterize the NP–protein structure
to-volume atomic ratio can exceed 50%. NP surfaces are not and activity as a function of parameters such as the NP
uniform but have multiple crystal lattices, vertices and edges properties (material, ligand, size, shape), the labeling site and
(figure 1(a)), each of which can have different chemical the protein coverage (figure 2). We seek to design rules for
affinities for adsorption [26]. This is further complicated by efficient conjugation. Where is the best site on the protein
the fact that NP samples are typically polydisperse, with finite for NP conjugation? How does NP surface chemistry affect
size distributions. Many spectroscopic techniques used to protein structure? What types of interactions dominate the
probe bulk surfaces are not applicable or difficult to implement NP–protein interface? Can the NP be used to locally perturb
on soluble NPs. Furthermore, surface-coating ligands on the the protein structure? Currently the nature of the interface
NP are labile. They can adopt multiple conformations and can between NPs and proteins are not well understood, despite the
migrate to different sites on the NP surface. In fact, Rotello fact that it is a significant hurdle for the use of NPs in medical

2
Biomed. Mater. 3 (2008) 034001 M-E Aubin-Tam and K Hamad-Schifferli

Appropriate conditions for the NP ligands and protein side

chains to become attracted to each other must be determined.
The interaction may be modulated by the pH or charge
(a ) screening via controlling the ionic strength of the medium.
Because it is inherently a non-specific interaction, the protein
may interact with the NP in any of a number of orientations
or have the NP positioned at any number of labeling sites.
However, strategic modification of NP surface chemistry has
(b )
enabled regio-specific interactions with the protein [39]. By
changing the charge on the end of the NP ligand and the ligand
hydrophobicity, a specific face of the protein cytochrome c
can be targeted for adsorption on the NP surface. Hydrogen
exchange experiments are used to experimentally verify the
binding faces. When the NPs are functionalized with L-
(c ) phenylalanine, cytochrome c binds on the face of the protein
near the heme crevice which contains many polar residues. If
the ligand is L-aspartic acid, the NPs bind to the large front
face of cytochrome c which is cationic.
(d )
2.2. Linking to the NP ligand
Figure 3. NP–protein labeling strategies: (a) electrostatic Another general method for NP–protein conjugation is
attachment of protein, (b) covalent attachment to the NP ligand, covalently linking a protein to the NP ligand (figure 3(b)).
(c) attachment of a protein cofactor on NP, and (d) direct linkage of
amino acid on the NP core.
If the NP has multiple ligands that can react with the protein,
this can result in a distribution in the number of proteins on the
NP. The stoichiometry can be influenced by varying the ratio
and biological applications. These questions have begun to be
of the reaction. A population of NP–protein conjugates with
probed for proteins that are non-covalently adsorbed on NPs,
various protein:NP ratios is usually produced.
and studies have found that the interface between the protein
This approach has been greatly advanced by extreme
and NP can have significant biological ramifications [31–33].
control over the surface chemistry of the NPs. For example,
It should be noted that this paper will not cover
pioneering work by Hainfeld et al have been able to isolate
biomineralization or NP crystallization by natural systems
Au NPs with exactly one reactive group by HPLC [40], which
such as caged proteins or viruses [34–37]. It is a fascinating enables limiting the coverage of proteins on the NP surface.
topic in itself, which has led to exciting advancements in The synthesis of these Au clusters for labeling purposes has
materials and biological applications. While it touches contributed to the success of Nanoprobes, Inc. Although this
on many similar issues (protein interactions with inorganic chemistry was originally intended for imaging by electron
surfaces), it is beyond the scope of this paper. microscopy (EM), it has enabled capabilities well beyond EM
imaging such as electron transfer [10].
2. Linking chemistries Other strategies to restrict the possible number of linking
groups on a NP include control over the morphology of the
Strategies to link a protein to the NP have taken four surface ligands. This has been demonstrated in strategic use
main approaches (figure 3): (1) electrostatic adsorption, of ligands that phase separate into domains on the NP surface,
(2) conjugation to the ligand on the NP surface, (3) conjugation creating only two sites on which a reactive ligand may be
to a small cofactor molecule that the protein can recognize and placed [41]. In addition, the synthesis of large ligands that
bind, and (4) direct conjugation to the NP surface. Other coordinate in a particular way to the NP surface has also
strategies are described in the review of Medintz et al [29]. permitted a precise number of linking groups [42].
Issues involved in these labeling strategies include sterics, In addition, many have exploited the binding of Ni-NTA
or whether the protein can ‘get past’ the ligand to the NP or Co-NTA groups to His tags on protein [18, 43–47]. Because
surface or the relevant linking group. A choice of chemistries the His tag can be cloned into a protein sequence, it enables
that result in a specific link (i.e., do not cause extensive site-specific labeling. Typically the His tag is put at the termini
cross-linking) and are stable for the desired purpose are also of the proteins and thus is relatively simple to be cloned in.
necessary considerations. This results in a strong bond, with a dissociation constant of
approximately Kd = 10−13 M at pH 8 [48].
2.1. Electrostatic adsorption Another popular labeling chemistry utilizes the covalent
binding of primary amines with sulfo-NHS esters or R-COOH
The most widely used linkage approach consists of groups via reaction with EDC [49, 50]. NPs labeled with NHS
electrostatic adsorption (figure 3(a)). This is the simplest as it esters can react to form covalent bonds with the primary amine
requires no chemical reaction, and is already used routinely as of lysine on a protein. In addition, NPs coated with maleimide
an electron dense marker in histology [38]. groups can react with the thiol of cysteine on a protein.

3
Biomed. Mater. 3 (2008) 034001 M-E Aubin-Tam and K Hamad-Schifferli

2.3. Linking using specific affinity of protein for a cofactor of protein, which could potentially compromise the stability
of the protein and lead to denaturation. However, because
Alternatively, NP–protein conjugation can be achieved
disulfide bridges tend to be buried in the protein, this is unlikely
by using specific labeling strategies in bioconjugation
if the protein remains folded upon NP attachment. In the case
(figure 3(c)).
of proteins that have free cysteines close to the protein surface
The classic linkage used in biolabeling is biotin–
that are not tied up as dithiols, this is a highly convenient
streptavidin binding. In effect, labeling NPs with biotin
way to conjugate a protein to an Au NP. Similarly, for sulfur
ligands allows linkage to streptavidin, which due to its
containing NPs such as ZnS/CdSe, cysteine can directly form
multivalent nature can be bound to other biotin-labeled species.
a disulfide bridge with the surface S atom [56].
The biotin–streptavidin interaction is strong enough to be
Direct linkages can also be achieved by His tags, which
nearly covalent, with a dissociation constant of on the order
can attach directly to Zn, Ni, Cu, Co, Fe, Mn atoms. Mattoussi
of Kd = 10−14 M [51]. Because there are a wide variety of
et al and other groups have exploited the fact that His6 groups
linkers that can be functionalized with biotin, it is a versatile
can coordinate to Ni NPs [57] and to the Zn on the surface of
way to achieve a specific linkage. Consequently, it has been
CdSe/ZnS or CdS/ZnS quantum dots [6, 47], obviating the
utilized extensively for both biological conjugation as well as
need for the addition of a metal ion cofactor.
inorganic-biological conjugation.
An important concern in direct labeling of the NP core
Finally, an attractive route for binding a NP to a specific
is steric crowding on the NP surface. It may be difficult for
protein utilizes antibodies [24, 50]. Because antibodies can
the targeted protein residue to reach the NP surface if the
specifically bind to a target protein, they are suitable for NP
NP ligand is too densely packed on the surface or too long.
labeling. NP–antibody conjugates are useful for imaging
For example, polyethylene glycol is a popular ligand for NPs
the presence of specific proteins and receptors in cells by
as it prevents non-specific adsorption of protein side chains
fluorescence via quantum dots, or sensing protein binding via
with the NP core [23, 58, 59]. However, its long chain may
FRET, and other applications such as diagnostics. The NP is
hinder access to the conjugation site. Also, if the strength
first linked to the antibody, and the NP–antibody conjugate
of interaction of the ligand with the NP surface is similar or
can then bind to a specific protein. Conjugation of NPs to
higher than the bond-linking proteins to NPs, it may not be
antibodies generally adopts a global labeling strategy, such as
able to displace the ligand from the NP surface and attach the
targeting any of the primary amines distributed over the entire
protein. In this case, a large excess of protein in the reaction
antibody, or reducing the disulfides in the hinge region to have
is necessary.
free thiols that can react with a NP.
Recent advances in aptamer chemistry have been
exploited to use them as binders to proteins [52, 53]. As 3. Site-specific labeling
with antibodies, aptamers can be designed to target-specific
proteins. Because DNA chemistry is much simpler to modify For numerous applications, it is important for the NP–protein
artificially than proteins, this is a versatile linker for NP– conjugate function to label the protein at a particular site. For
protein conjugation. example, NPs are linked to proteins to sense when the protein
binds to its substrate, typically by FRET. For this purpose, it is
crucial that the NP be placed in a particular site of the protein
2.4. Direct reaction with NP surface atoms
so that it is regio specific or even site specific. Also, for
A direct reaction of a chemical group on the protein without NPs used as EM imaging tags to localize a specific structure
the use of a linker [23–25] is usually desired if the particle within a large protein [25], it is advantageous to attach the NP
is used as a biosensor where FRET or electron transfer is in close vicinity of the residues or motifs of interest. Even
used (figure 3(d)). These processes are sensitive to distance if one can control the stoichiometry of the resulting product
changes on the Ångstrom scale, and long floppy linkers can (the NP:protein ratio), and assuming that the protein is fully
adopt multiple conformations, resulting in a variable protein– folded upon labeling, most labeling chemistries do not have
NP distance that can be problematic in sensing. In addition, the ability to single out a single amino acid. Instead, there is
long linkers can result in conjugates that are larger than the NP a probability that the NP can be placed at any number of the
or protein, which can decrease circulation times in the blood available sites (figure 4), such as labeling the primary amines or
or cause problems in cellular uptake. Direct linkage to the carboxylic acids. This results in the protein assuming random
protein in these cases would be much more desirable. orientations on the NP surface. Moreover, if one of these
For Au NPs, this can be achieved by the Au-thiol sites is in a non-ideal position, such as within the binding site,
chemistry where a protein with a cysteine covalently bonds those proteins will not be able to bind to their substrate. If
to an Au NP. The conjugation requires only incubation of the the experimental probe is an ensemble measurement, then one
two species together as the Au–S bond is strongly favored. would not be able to distinguish a 20% decrease in activity
This results in a short, direct link from the protein side chain due to all proteins having a lower activity due to the presence
to the NP surface. Disulfides of self-assembled monolayers of the NP, or due to 20% of them being totally inactive. This
(SAMs) are known to break up in order to interact with Au [54] distribution of labeling sites is also a problem if one wants to
because the Au–S bond strength approximates S–S strength use the NP to obtain quantitative information about the protein
(∼40 kcal mol−1 compared to ∼50 kcal mol−1) [55]. This function, such as via FRET pairs in which the distance of the
suggests that NP labeling can potentially break up dithiols NP to a chromophore matters. This is further complicated

4
Biomed. Mater. 3 (2008) 034001 M-E Aubin-Tam and K Hamad-Schifferli

Table 1. Strategies for labeling a specific residue that involve direct linkage to the NP core.
Labeled residue—NP material Examples of labeled protein Molecular weight Ref.
Cys–Au Cytochrome c 12.6 kDa [23]
Ribonuclease S 13.7 kDa [60]
70 s ribosome 2.5 MDa [25]
Single chain Fv antibody 25 kDa [24]
Chaperonin 60 kDa [56]
Cys–CdSe/ZnS Chaperonin 60 kDa [56]
Human serum albumin 67.0 kDa [8]
α-chymotrypsin 21.6 kDa [61]
His Tag–CdSe/ZnS Maltose binding protein 40.6 kDa [62]
His Tag–Ni Green florescent protein 27 kDa [57]

Table 2. Strategies for labeling a specific residue that involve linkage to the NP ligand.
Labeled residue—NP ligand Examples of labeled protein Molecular weight Ref.
Cys—dimercaptosuccinic acid Cytochrome c 12.6 kDa [63]
His Tag–Ni-NTA Adenovirus serotype 12 knob 60.6 kDa [44]
20S proteasome 750 kDa [45]
Streptopain 42 kDa [64]
Green florescent protein 27 kDa [65]
His Tag–Co-NTA Horseradish peroxidase 44 kDa [43]
Lys–azide Lipase 30 kDa [66]

Figure 4. Protein with several NP binding sites. NP attachment can

inactivate the protein via denaturation or blocking of the active site.

by the fact that the protein has only 20 amino acids, so for
larger proteins (more than approximately 100 residues) the Figure 5. Strategy for site-specific labeling RNase S. S-peptide and
possibility of having a unique amino acid that can link to a NP S-protein spontaneously bind to form RNase S (upper). Site-specific
is difficult. For instance, if labeling is achieved via chemistry labeling of an engineered peptide, S18, allows specific labeling of
conjugating to the primary amine of lysine, and a protein has the complex, adapted from [60].
multiple lysines, a distribution of labeling sites would result,
some of which may be detrimental for the protein function. by comparing to cytochrome c from horse. This protein has
Nevertheless, labeling of NPs to a specific amino acid on nearly an identical sequence but lacks C102. Therefore, if
a protein has been achieved by a variety of methods. Table 1 NPs are incubated under identical conditions with both yeast
lists examples where proteins are linked directly to a NP core, and horse cytochrome c, but link only to the yeast, then it can
and table 2 those which link to a ligand on the NP surface. be inferred that the NP is attached specifically to C102 [23].
An example of a protein in which there is a unique thiol Unfortunately, this approach is not applicable for proteins
that can be labeled is yeast cytochrome c, which has a cysteine which have more than 1 cysteine available for conjugation.
at position 102. This cysteine is close to the surface of the For the case of ribonuclease S (RNase S), the self-
protein and near the C-terminus, and thus has been a viable assembling properties of the protein can be utilized to achieve
site for labeling. Two other cysteines do exist in the protein site-specific labeling (figure 5) [60]. RNase S is the two-part
but these are linked to the heme and thus not available for version of the enzyme ribonuclease A, a 13.7 kDa protein
labeling. Site-specific labeling has been assayed indirectly, that cleaves RNA. The S form results from the incubation

5
Biomed. Mater. 3 (2008) 034001 M-E Aubin-Tam and K Hamad-Schifferli

Figure 6. Functionalization of Au NPs with (1), and linking of lipase using click chemistry, from [66].

of RNase A with subtilisin, which cleaves the protein to columns [68], ion exchange [24] and gel filtration [24, 50].
result in a short peptide (1–19), the S-peptide, and the rest In addition, multiple washes and centrifugation have been
of the enzyme, S-protein. When incubated together, the S- successful in isolating the NP–protein conjugate [66].
peptide and S-protein spontaneously self-assemble to form the If the NP is magnetic, magnetic separation can be used
active enzyme (upper panel). Because the S-peptide does not to purify the NP–protein conjugate away from free protein
contain any cysteines (KETAAAKFERQHDMS), one can be [65]. This strategy has been developed for commercial use,
added to serve as a unique labeling site at the N-terminus and large beads containing multiple magnetic species can be
(CGGKETAAAKFERQHDMS). Therefore, the NP can be used to purify proteins. However, from a labeling perspective,
incubated first with the S-peptide to form a NP-peptide bond. an introduction of a magnet precipitates all the magnetic NPs,
Then, the NP–S–peptide complex can be incubated with the S- including those which are unlabeled. While this is adequate for
protein to form a NP–RNase S complex that is site-specifically applications in which a particular protein needs to be pulled out
labeled (lower panel). of solution away from other proteins, it may not be feasible for
More recently, lipase was genetically engineered so that synthesizing and purifying a NP–protein conjugate, because
it contains only one solvent-accessible lysine. This lysine you would obtain a sample that would have a mixture of both
was then modified by carbodiimide chemistry to result in NPs linked to protein and free NPs. Nevertheless, adjustment
an acetylene group. This acetylene group could then be of reaction stoichiometry may ameliorate this problem.
coupled to Au NPs functionalized with azide ligands via HPLC has also been successful in isolating NP–protein
click chemistry, which is the 1,2,3-triazole formation between conjugates [23, 45]. Au NPs with aminoethanethiol ligands
azides and terminal acetylenes that is catalyzed by copper(I) conjugated to yeast cytochrome c were separable from both
[66] (figure 6). free protein and free NP [23]. Absorption spectra of the elution
Other recent strategies include functionalizing the protein curves confirmed that the isolated species were indeed the
with a highly specific ligation reaction, which introduces a conjugate.
reactive group that links with a NP. This strategy has been Gel electrophoresis has been used for many years
used to label the GTPase Rab6a with an Au NP [67] (figure 7). to separate biomolecules of different lengths and even
Rab6A was modified on its N-terminal with a PEG linker with conformations. It was demonstrated in 2001 by Alivisatos
a free thiol. As Rab6A lacks cysteines, only this exposed thiol et al that it can be used to purify NP-DNA conjugates of desired
will react with the NP. Once labeled, the Rab6A still binds to stoichiometry [69], and soon after for NPs linked to proteins
its substrate (BodipyFL-GTP) such as the wild-type version, [70]. Electrophoresis can be used both as a means to isolate the
which can be detected by fluorescence quenching. conjugate from other species, or as a means to assay the purity
of a sample. It should be noted that the mobility of a sample is
4. Techniques to purify and to assay purification of due to both its charge and size, and a particle size distribution
conjugates and charge would lead to a smearing of bands, which can often
obscure good separation. Electrophoresis utilizes the fact that
Once the NP–protein conjugation is achieved, it must be the conjugate has a mobility that differs from both the free
purified from both the free protein and free NP species. This protein and free NP. This change in mobility is influenced
has been achieved by a variety of approaches, such as spin by both a change in the size or charge. Figure 8 shows an

6
Biomed. Mater. 3 (2008) 034001 M-E Aubin-Tam and K Hamad-Schifferli

Figure 7. N-terminal modification of N-Cys-Rab6A (7) with Acm-protected PEG thioester (4) by a native chemical ligation leading to
PEGylated Rab6A (5). The Acm group is subsequently removed by treatment with Hg2+ to give the free thiol group needed for
immobilization on Au NPs (6). Immobilization reaction of PEG-Rab6A (6) on citrate-stabilized Au NPs (9) and DNA-functionalized Au
NPs (10). Upon reaction of the thiol group attached to the Rab6A protein via a PEG linker, a covalent bond is formed, from [67].

5. Biophysical characterization of NP–protein

conjugates

Once the NP–protein conjugate is synthesized and isolated

from other species, the conjugate must be biophysically
characterized. This requires measuring both the structure and
function of the protein when labeled with the NP.

5.1. Structure of NP–protein conjugates

First, the structure of the protein when it is labeled must
be probed. It is necessary to know if the protein is folded,
Figure 8. Gel electrophoresis to assay conjugation of herceptin with or if denaturation upon labeling has occurred. Because
magnetic NPs, from [50]. the protein structure is related to the function, it is a key
for assaying whether the protein in the conjugate form is
example of how gel electrophoresis can assay the conjugation functional. Unfolding is likely upon conjugation to a NP,
of 9 nm Fe3O4 NPs with the antibody herceptin [50]. Lane 1 as the interaction of the protein’s side chains with the NP
shows the NPs by themselves, which are negatively charged surface or ligands is likely to occur, resulting in non-specific
and thus run toward the positive electrode. Lane 2 shows adsorption and perturbation of the native structure.
the NPs incubated with herceptin. A new band of lower There are numerous techniques that have been applied to
mobility appears, indicative of a NP-herceptin conjugate. Blue study the structure of proteins. Unfortunately, only a few have
staining of this band shows the presence of protein. Both been applied to study the structure of proteins when conjugated
polyacrylamide (PAGE) and agarose gel electrophoresis can to NPs. However, this can also be viewed as an opportunity
be used. However, agarose gels which are horizontal can be to expand the set of techniques to characterize these systems.
run in both directions, so both positive and negative species Challenges include requirements in sample quantity, as yields
can be viewed simultaneously, which is useful for conjugating for purified NP–protein conjugates can be low.
proteins to oppositely charged NPs. In addition, the band The techniques described below have been able to evaluate
containing the conjugate can also be cut from the gel and the effect of the NP size on the structure of non-covalently
the conjugate can be extracted from the agarose by spin attached proteins such as lysozyme [72] and cytochrome c
centrifugation columns, allowing purification of the conjugate [73]. In the case of covalently attached proteins, it has been
from both free NP and free protein [23, 60, 71]. Sample able to study the effect of the NP ligand on the structure of
wells in agarose gels tend to be larger than in PAGE so higher cytochrome c on Au NPs [23]. Structural studies have also
quantities of protein and NP can be loaded into the gel. determined the orientation of cytochrome c on Au or Ag NPs

7
Biomed. Mater. 3 (2008) 034001 M-E Aubin-Tam and K Hamad-Schifferli

(A )

(B )

Figure 10. CD spectra of chymotrypsin (CHT) and CdS-CHT

conjugates, from [61].

certain CD signatures in the UV range. Furthermore, one

Figure 9. Stereo representations of the ribosome recycling factor can deconvolute a spectrum to determine the relative amounts
binding positions on the 70S ribosome (A) and on the dissociated of each motif. Spectra can be compared to a database of
50S ribosomal subunit by cryo-EM (B), from [78].
known protein structures, which are usually determined by
x-ray crystallography and NMR.
[74, 75], and how it is affected by the NP ligand functional The benefits of CD are that it is a standard technique and
groups [39], as will be discussed in greater detail below. has been used to characterize many proteins, so signatures
Measuring the protein structure is not only important of numerous proteins are well known. Furthermore, it can
for creating biologically functioning NP–protein conjugates. be performed in solution which is the standard form of NP–
With such information, one can take this further and study protein conjugates. Therefore, it is a direct probe of the protein
how to engineer the NP ligand so that proteins or peptides structure on the NP. It has been the most widely used technique
can adopt a particular conformation when adsorbed on the NP to assay the protein structure in conjugates, and has been used
surface. If more methods to study the protein structure when successfully for several NP–protein systems: lysozyme and
conjugated to NPs were to be explored, it could expand the peptides on silica [72, 83], α-chymotrypsin on CdS (figure 10)
functional utility of NP–protein conjugates. [61] and on Au [84], albumin on CdTe [70] and cytochrome c
on Au [23, 27].
5.1.1. Electron microscopy. Direct imaging of the protein However, CD spectroscopy of NP–protein conjugates
when conjugated to the NP can be used to assay the is complicated by the fact that NP contribution to optical
structure. Small Au NPs have been used as imaging absorption at short wavelengths can be significant enough to
agents. By reconstructing several images, one can obtain a swamp the signal. This makes difference measurements very
three-dimensional map of the protein. Large proteins have noisy unless averaging over long times is done. Corrections
been successfully imaged by this method down to 6–8 Å for absorption flattening were found to be necessary for large
resolution [45]. This requires imaging of many complexes NPs which have stronger absorption [72]. In addition, it is an
and sophisticated image deconvolution techniques. Labeling ensemble measurement so it cannot distinguish between 50%
of proteins with Au NPs has been used in cryo-EM to image of all the proteins being fully denatured from all of the proteins
many proteins, such as MAP2 and tau on microtubules [76], being half denatured.
F-actin [77], the ribosome (figure 9) [78, 79], the scrapie prion
protein [80], and conformational changes in kinesin [81] and 5.1.3. Optical spectroscopy. Optical spectroscopy can be
myosin [82], among many others. used to probe the protein in special cases where the protein
has a chromophore or an optically absorbing cofactor, such
5.1.2. Circular dichroism spectroscopy. Circular dichroism as a heme for proteins such as cytochrome c and hemoglobin.
(CD) spectroscopy has been traditionally used to measure the In this case, optical absorption can yield information on the
secondary structure of a protein. It has also served as a redox state of protein, which can be used to infer the secondary
useful tool for probing a secondary structure of the protein structure of the protein. For example, for cytochrome c, it is
when it is linked to a NP. The basis of CD spectroscopy is known that as the protein is more unfolded, it is more readily
the difference of absorption of right-handed and left-handed oxidized, and this is reflected by a shift in the Soret band and
circularly polarized light through a sample. This difference changes in Q-band intensity. Therefore, optical absorption
spectrum is related to chirality of the protein. In the UV can yield information on the degree of folding in the protein.
region, it reflects the secondary structure of the protein. Protein However, because these features are also sensitive to redox
motifs such as α-helices, β-sheets and random coils have potential, their changes can also be the result of the presence

8
Biomed. Mater. 3 (2008) 034001 M-E Aubin-Tam and K Hamad-Schifferli

Fluorescence spectroscopy. Another optical technique that

has been successful in determining unfolding is fluorescence
spectroscopy. Aromatic residues (Trp, Tyr, Phe) exhibit
fluorescence in the UV which changes as the protein unfolds.
They are typically buried in a hydrophobic environment in
the interior of the protein, and when the protein unfolds or is
destabilized by the presence of the NP, they become exposed
to the hydrophilic surroundings [8]. If measured as a function
of temperature, it also gives information on the stability of
the protein in conjugate. However, this technique is often
unsuitable, as most NPs absorb strongly in this wavelength
range (350 nm), resulting in strong quenching.
Figure 11. Protein denaturation upon interaction with Au
monolayer capped NPs leads to a close packed assembly. If the
Electrophoretic mobility. Because the mobility of a species
protein is denatured (a) the interparticle spacing is larger than if the in gel electrophoresis is a function of both size and charge
protein is not denatured (b), which can be observed by SAXS, from of the complex, Ferguson analysis can yield quantitative
[84]. information on size and charge of the complex. This requires
running the complex through gels of several concentrations,
of the NP ligand or species free in solution, so they should be and measuring the mobility of the complex. A fit of the
accompanied by structure-only probes such as CD. Again, as mobility versus the gel concentration obtains an effective size
in CD spectroscopy, the NP contribution to the optical spectra of the complex. If the mobility is extrapolated to zero gel
can swamp out the heme spectra, as extinction coefficients percentage, one obtains the free mobility, M0, which is the
mobility of the sample if it were unencumbered by a gel. M0
near heme absorption features can be an order of magnitude
is dependent on the charge of the complex. Therefore, gel
higher. Subtraction of the NP contribution, when possible, can
electrophoresis contains information about both the charge
aid in identifying heme features [23, 63].
and size of a NP–biomolecule conjugate. It has been used
successfully to elucidate the conformation of the DNA on the
5.1.4. SAXS. Small angle x-ray scattering (SAXS) have been NP surface [69, 87, 88]. However, similar success has not yet
used to measure the periodicity of the NPs and thus infer the been achieved for proteins.
inter-particle spacing, which can be correlated to the protein These are only a few approaches. Other techniques
size [84] (figure 11). If the protein denatures, the interparticle include mass spectrometry, which can assay the ratio
spacing decreases and thus changes the SAXS spectra. The of labeling (stoichiometry) in a manner similar to gel
sample form is solid for SAXS, so it can be employed in electrophoresis, but by mass [46, 61]. Single-molecule
samples that are not soluble. spectroscopy has also been used to calculate the number
of proteins per NP [89]. Femtosecond time-resolved
spectroscopic methods studying the photophysical properties
5.1.5. Other possible techniques. As aforementioned, of the photosynthetic bacterial light harvesting the complex on
there are many other techniques that have been crucial in TiO2 NPs can yield information on the structural deformation
studying the protein structure that have not yet been employed of the protein on the NP [90]. Other structural techniques
extensively in NP–protein systems. The following methods include H/D exchange [39] which examines the exposed
have been employed only on one or two systems, at most. protons on the protein. To assay the thermodynamic properties
of complex, calorimetry can obtain the heat capacity CP and
H of unfolding, as well as Tm. This information can be used
Atomic force microscopy. Atomic force microscopy (AFM)
to understand the stability of protein when it is linked to the
can be used to probe the structure of the protein in the complex.
NP. A recent study by Rotello et al has utilized calorimetry
In essence, the structure of protein is inferred from a height [91].
measurement [41]. Phase imaging also allows mapping areas
of adhesion and friction. Imaging of proteins on surfaces
5.2. Activity of protein
has improved due to new fabrication techniques for AFM
tips and imaging methodologies. AFM can now resolve a NP-induced protein structural disturbance is prone to affect
secondary structure in proteins such as individual α-helices its activity, as structure and function are closely linked in
[85, 86]. If the tip is chemically functionalized, it can be used proteins. The assay for activity depends on the protein. It is
to chemically image a sample. AFM offers large benefits as typically measured via reaction with the enzyme substrate, or
it consists of a single molecule imaging technique (does not via electrochemistry for a redox active enzyme.
represent an ensemble average) and the sample requirements
are small. However, because it requires the samples to be on 5.2.1. Enzymatic activity measurements by reaction with a
a substrate, protein interactions with the surface may denature substrate. For an enzyme to be active, its substrate must
the protein, and obscure the effects of the NP. reach the active site, the protein must act on it, and finally

9
Biomed. Mater. 3 (2008) 034001 M-E Aubin-Tam and K Hamad-Schifferli

release the product. All of these steps are going to be A Laviron analysis of the peak-to-peak potential separation
dramatically influenced by the presence of a relatively large at variable scan rates yields the electron transfer rate constant
particle with numerous molecules on its surface. Also, as the [112].
NP–protein net charge usually differs from that of the protein, The protein structural integrity can also be probed
the altered electrostatic interaction can affect the substrate by electrochemistry. The redox potential depends on the
binding or the product release. The enzyme activity could be environment of the redox center, which almost certainly
significantly reduced if, for example, both the substrate and changes when the protein unfolds. Denaturation of the protein
NP are negatively (or positively) charged, or if the product could also make it electroinactive. The model protein horse
remains electrostatically linked to the NP–protein complex. cytochrome c is found to be active when attached on various
Hamad-Schifferli and co-workers found that the cleavage nanostructures. On electrochemically grown Au nanorods
of RNA by RNase S is significantly slower when the enzyme is and nanopyramid, its redox potential is 58 mV and 61 mV
covalently linked to NPs [60]. The decrease in ribonucleolytic versus Ag/AgCl (respectively) [113], which is very close to
activity is rationalized by steric and electrostatic arguments. the 60 mV versus Ag/AgCl reported for well-folded horse
The RNA cleavage site is at the interface between the S-peptide cytochrome c [114]. The same protein on negatively charged
and the S-protein, near the NP surface (figure 5). The access Au NPs gives similar redox potentials (68 mV versus Ag/AgCl
of the negatively charged RNA to the active site is hindered for 20 nm NPs [107] and 72 mV versus Ag/AgCl for 12 nm
by crowding and electrostatic repulsion as the NP-RNase S NPs [109]). No cytochrome c signal could be obtained for
conjugates also have a negative net charge. electrodes with aggregates of NPs, suggesting that the protein
The activity of several other NP–enzyme conjugates has unfolds on those surfaces [109].
been studied. Sastry and co-workers found that 3.5 nm Au
NPs conjugated with fungal protease (F-prot) [92] or with 5.2.3. Antibody activity. NP labeling of antibodies is a
pepsin [93] have a proteolytic activity comparable to that widely used technique for sensing, FRET, general imaging
of the free enzyme when incubated with hemoglobin. F- in applications such as cancer detection. While the structure
prot immobilized on NPs produces only slightly less digest of antibodies is not known, the activity assay is relatively
fragments than the free enzyme. Huang and co-workers straightforward. One needs only to test if the conjugate binds
studied the proteolytic activity of trypsin attached on 13 nm to its target, such as in CdS labeled anti-CD4 monoclonal
Au NPs and found a lower enzymatic activity and a different antibody binding to target cells. This is assayed by
specificity for cleavage [94]. While a decrease in activity fluorescence spectroscopy or FACS [115].
of enzymes on NPs was also observed by several groups
[61, 95–97], we should note that others report unchanged or There are other assays of activity such as those involved
enhanced activity [98–100]. in more complex biological processes. For example, peptide–
gold conjugates used to bind to neurons still exhibit neuronal
5.2.2. Electrochemistry. The activity of redox proteins activity [116].
such as heme proteins, copper-containing proteins and iron–
sulfur proteins are typically measured via electrochemical 6. Molecular dynamics simulations
measurements [101, 102]. Many experiments of a
plain unlabeled protein on electrode surfaces have gained Experimental methods to probe individual NP–protein
information on the protein conformation and orientation on interactions (e.g. IR, Raman and NMR) are limited, as they
the surface [103, 104]. Analogous electrochemical studies for require a high amount of sample and the NPs often contribute
protein on NP surfaces are desirable. to spectral dampening and broadening. Therefore, molecular
The use of NPs in bioelectrochemistry has gained interest dynamics simulations are a natural complement to ensemble
in the past decade [105]. In particular, NPs are found capable spectroscopic techniques such as CD spectroscopy in order
of wiring redox proteins to electrodes. They can act as an to yield insight on the molecular interactions involved in NP-
antenna that provides a conducting path to facilitate electron induced structural disturbance.
transfer between the protein redox center and the electrode Peptides and proteins were simulated on flat surfaces
[10, 106–109]. [117, 118]. For example, yeast cytochrome c was tethered
The electrode/NP/protein interface is generally self- to an alkanethiol SAM by covalent disulfide bonding between
assembled using one of the following strategies. (1) NPs are the thiol of C102 and a thiolated alkanethiol. Simulations
linked to the electrode and then immerse in a solution of redox of hydrophobic and hydrophilic SAMs showed that the
enzymes [10, 107, 109]. Or, (2) the NP–protein conjugates hydrophilic SAM interacts more with the protein [118].
are prepared in solution, purified, and then immobilized on the Simulations of proteins on NP surfaces are more
electrode [10, 110, 111]. Willner and co-workers [10] obtained challenging. Several complex interactions occur between the
the same density of NP–proteins attached on the electrode via NP core, the NP ligands and the protein side chains. The nature
routes 1 and 2 (figure 12). of the interactions between metal surfaces and amino acids is
From the rate of electron transfer, we can estimate the not completely understood. Furthermore, simulating all the
distance between the NP and the protein redox center, and NP atoms with its ligands results in computationally expensive
thus determine the protein orientation [103]. The rate of simulations, especially for large NPs. Molecular dynamic
electron transfer is typically measured by cyclic voltammetry. simulations of NPs with ligands have been done mostly for

10
Biomed. Mater. 3 (2008) 034001 M-E Aubin-Tam and K Hamad-Schifferli

(A )

(a)

(b)

(B )

Figure 12. (A) Assembly of Au-NP-reconstituted GOx electrode by (a) the adsorption of Au-NP-reconstituted GOx to a dithiol monolayer
associated with a Au electrode and (b) the adsorption of Au-NPs functionalized with FAD on the dithiol-modified Au electrode followed by
the reconstitution of apo-GOx on the functional NPs. (B) A STEM image of GOx reconstituted with the Au-FAD hybrid NP. Arrows show
Au clusters, from [10].

small (diameter < ∼2 nm) Au NP with simple alkanethiol the NP ligand, the NP size, the NP material, the stoichiometry
ligands [30, 119, 120] (figure 1(a)). Simulations of SAMs of and the labeling site on the protein (figure 2).
alkanethiols on NPs and flat surfaces showed that the structure
of the SAM depends strongly on the alkanethiol chain length, 7.1. Choice of the NP ligand
the surface curvature and the temperature [120]. In particular,
the SAM is more tilted on NPs compared to flat surfaces. The NP ligand in a NP–protein conjugate can greatly influence
Additionally, the ligand density is found to vary on the NP the protein structure of both non-covalently and covalently
surface. The thiolate chains located near the edges of the NP attached proteins. It is in close contact with the protein, and
can interact with nearby residues, so choosing the right NP
faces are more extended and accessible to the outside [30].
ligand is crucial. In covalent NP–protein conjugates, ligands
Interestingly, this is also where ligand exchange is believed to
with charged species as endgroups are likely to denature the
take place [26].
protein via electrostatic effects [23, 63]. Charged ligands can
also prevent or enhance linking chemistry via electrostatic
7. General labeling strategies repulsion or attraction and dramatically affect non-specific
adsorption. PEG has been shown to be an ideal ligand
Overall, in synthesizing NP–protein conjugates, a general for preventing non-specific adsorption for bulk 2D surfaces,
strategy for creating a biologically functional NP–protein NP and nanorod surfaces [23, 58, 59], thus has been used
conjugate is necessary. Because non-specific binding often extensively. PEG’s floppy chains and charge neutrality prevent
leads to protein unfolding, the general goal of this strategy non-specific adsorption, and thus have been a popular choice
is to determine how to block non-specific adsorption on NP for NP ligands.
and maintain the protein structure and function. The protein For cytochrome c, it has been determined that electrostatic
structure and function are influenced by the chemical nature of interactions between charged amino acids and the NP ligand

11
Biomed. Mater. 3 (2008) 034001 M-E Aubin-Tam and K Hamad-Schifferli

(a ) (b )

nm

Figure 13. Effect of the NP ligand charge of structure of site-specific covalently labeled yeast cytochrome c. (a) CD spectra of cytochrome
c as a function of the NP ligand. (b) Schematic of NP ligand and cytochrome c functionalization (aminoethane thiol (AET),
bis(p-sulfonatopehnyl)phenylphosphine (BPS), hexanethylene glycol (PEG)). Adapted from [23] with a normalization of AET-NP-YCC
signal at the isodichroic point (θ 206).

are the dominant interaction influencing the protein structure aspartates positioned at specific intervals. Upon incubation
(figure 13). Yeast cytochrome c linked to 1.5 nm Au NPs of the peptide with the NPs, the interaction between the two
at the C102 position is strongly denatured when the NP has species induced folding of the peptide into an α-helix as
either positively or negatively charged ligands, whereas it is not measured by CD spectroscopy [28].
denatured at all when linked to NPs with PEG (neutral) ligands.
Spectra in figure 13 show strong deviation from the typical α- 7.2. Choice of the NP size, shape and material
helical structure for yeast cytochrome c linked to positively
charged aminoethanethiol (AET) and negatively charged bis- NP size can also affect the protein structure and activity. For
(p-sulfonatopehnyl) phenylphosphine (BPS) coated NPs. The larger NPs, the effective protein surface area it can access
structure is closer to a random coil, especially for the AET is larger, so it may increase the likelihood of denaturing the
particles. Deconvolution of the spectra reveals that the % α- protein. Also, smaller NPs due to their higher curvature will
helicity is normally 35% for wild-type yeast cytochrome c, have a fewer number of NP ligands that can interact with the
but decreases to 7% for the positively charged NPs and 19% protein side chains. In addition, the surface area on the NP
for the negatively charged NPs. However, for cytochrome accessible to the protein will be lower, and could potentially
c linked to PEG-SH coated Au NPs, the protein structure result in less denaturation. With regards to activity, larger NPs
is fully folded, with a 35% α-helicity. Increasing the salt may sterically prevent substrates from reaching the binding
concentration for the charged ligands decreases the amount site or active site more than for smaller NPs. Furthermore,
of denaturation, indicating that charge screening can reduce increasing the NP size too much may result in NP solubility
these interactions and allow the protein to fold. This confirms problems, as larger NPs are more difficult to keep soluble.
that the interaction between the protein and NP that results in This effect of curvature on protein structure has been
denaturation is largely electrostatic, and that neutral ligands observed for lysozyme on silica particles (figure 14) [72].
are best for maintaining the protein structure in the conjugate. Dordick et al perform a systematic study in which they study
Tabulating the amino acids in the general vicinity of the linking the structure of lysozyme adsorbed onto silica particles that
site, C102, shows that there are several that are positively range in diameter from 4 nm to 100 nm. Both structure and
or negatively charged, which can interact with the charged activity of the lysozyme are measured, by CD spectroscopy and
ligands and result in denaturation. It should be noted that colorimetric activity assays, respectively. The particle size is
these experiments do not rule out the possibility of amino acid found to strongly influence the protein–particle interactions
side chains interacting directly with the Au on the NP surface. and consequently the protein behavior. The lysozyme retains
Many studies of proteins on bulk Au surfaces have found that its structure when on the 4 nm silica particles, and denaturation
amines can coordinate with the Au on the surface, possibly by increases with increasing silica particle size. In addition,
the amine acting as a weak electron donor. However, designing activity is highest on the smallest particles and lowest on the
experiments that would directly probe this is difficult for the largest particles. This is most likely a general effect, which
NP. is probably present in conjugates for proteins on NPs of other
In the event of non-covalently bound proteins, the NP materials.
ligand also influences the protein structure. There are In addition, this surface curvature effect will most likely
examples of how one can exploit the ability of the surface- affect proteins that are covalently linked to the NP. Although
coating ligand of the NP to interact with proteins in a specific this has not yet been systematically studied for this case, it
way, permitting control over the protein structure [27]. For is probable that the effect of the NP size may become more
example, NPs functionalized with positively charged NP pronounced. In general, when a protein is allowed to adsorb
ligands could interact with a peptide of a sequence with onto a NP, it is more likely to retain its structure than if it

12
Biomed. Mater. 3 (2008) 034001 M-E Aubin-Tam and K Hamad-Schifferli

Figure 14. Effect of the NP size on the behavior of adsorbed lysozyme, from [72].

is covalently linked. Cytochrome c bound to Au NPs via These studies show that the size, shape and material of
electrostatic adsorption has been studied by several groups, the NP can greatly influence the structure and activity of
and in comparison to the covalently linked case it is more likely conjugated proteins.
to retain its structure or biological activity when adsorbed on
the NP [23, 109, 111]. However, some cases do exist when 7.3. Coverage (Protein: NP ratio)
adsorption alone does result in denaturation, such as for α-
chymotrypsin on Au NPs [27] and horse cytochrome c on The NP–protein stoichiometry can also influence the protein
CoFe2O4 NPs [63]. It is important to note that the effect of structure and activity. By immobilizing several proteins
the NP size on the protein structure is further complicated by per NP, one can take advantage of multivalency [124] and
the fact that most NP samples have finite size distributions that exploit the cooperative effects. Polyvalency is especially
are in the range of anywhere from 1% to as high as 20%, so useful for recognition and sensing applications, because of
the interaction of the protein with the NP surface and ligands increased affinity for substrate. Considerable enhancement of
is not homogeneous in the sample. Therefore, this can lead to the conjugate efficiency and selectivity is thus achieved [125].
different amounts of denaturation within the same sample. However, higher protein coverages on the NP surface
NP shape has also been found to influence protein activity. could lead to crowding, and thus affect protein folding and
Deng et al study how the electrocatalytical activity of adsorbed also activity. Steric issues such as blocking of binding of the
cytochrome c varies when the shape of the NP substrates is substrate may occur. Single-molecule spectroscopy has been
used to calculate the number of proteins per NP [89], and it has
changed [113]. The Au NPs are synthesized to be either
been determined that the coverage is actually a distribution.
spherical, rod-like, or pyramidal in shape. Electron transfer
In the case of Qdots functionalized with a maltose-binding
is strongly morphology dependent, and enhanced when the
protein, a Poisson distribution was obtained [126]. Higher
protein is immobilized on the nanorod or nanopyramidal
coverage on the NP surface has been shown to improve protein
substrates, but not on the spherical ones.
folding. In the case of linked and adsorbed cytochrome c on
In addition, the NP material can affect the protein structure
CoFe2O4 NPs, it is thought that higher coverage prevents the
and activity in the conjugate. For example, if amino acid side carboxylic acid side chains of residues in the protein from non-
chains have a strong affinity for a particular material, it could specifically adsorbing onto the NP surface [63]. However, the
lead to strong non-specific interactions between the protein opposite effect is observed for lysozyme on silica particles,
and NP and result in protein denaturation. NP materials for and the tendency of lysozyme to self-associate is thought to be
bioconjugates have been mostly of noble metals such as Au, responsible [72]. Apparently, the structural behavior depends
magnetic oxide materials and semiconductors for quantum on the protein–protein interactions of the specific protein.
dots. Quantum dots are typically coated with silica shell
which largely resists protein adsorption. Au is relatively inert,
though primary amines and carboxylic acid groups have been 7.4. Choice of the NP labeling site
known to interact with Au. Magnetic nanoparticles utilized Finally, the choice of the site on the protein can affect the
in biological applications are typically magnetic oxides, such protein structure and activity. As aforementioned, labeling in
as Fe3O4, Fe2O3, CoFe2O4 and MnFe2O4. Carboxylic acid the protein’s active site is undesirable as it can lead to decreased
groups have a strong affinity for the metal ions in the surface activity. In addition, as described in sections 7.1 and 7.2,
of the NP, and can coordinate to them. Observation of COO− interactions between the NP ligand and NP surface is typically
interacting with Fe [121], Co [122] and Ti [123] have been with amino acids in the local vicinity of the labeling site.
observed. For Fe-based magnetic oxide NPs, a change in Therefore, these interactions would be strongly influenced if
structure and decrease in activity were observed for lipase [95] the labeling site on the NP has a different local environment.
and cytochrome c [63], most likely due to such interactions. Also, NP curvature effects on the protein structure are probably
Coordination of COO- with Ti could also be in part responsible due to local interactions with the protein and will most likely
for the decrease in activity of the photosystem light-harvesting vary with the labeling site as well. Therefore, choosing a
complex on TiO2 NPs [90]. specific labeling site on the protein is crucial for creating a

13
Biomed. Mater. 3 (2008) 034001 M-E Aubin-Tam and K Hamad-Schifferli

functional NP–protein conjugate. Despite the importance of Acknowledgments

this effect, studies that systematically vary the position of the
NP labeling sites on a protein and examine the effect of the We would like to thank S Jayasinghe for the invitation to
labeling site on the protein structure were performed only participate in this review issue. This paper is dedicated to the
recently [127]. memory of Dr Jianping Shi (2005).
As a first approximation, strategies for dye labeling of
proteins can be utilized. His6 linkers are often put on the N
or C terminus of a protein, which are often far from the active References
site or binding sites of the protein. Alternatively, they can
be put into floppy loops within the sequence of the protein [1] Kim J et al 2006 Designed fabrication of multifunctional
magnetic gold nanoshells and their application to magnetic
that are not crucial for the stability and folding of the protein resonance imaging and photothermal therapy Angew.
structure. This requires knowledge of the protein’s structure Chem. Int. Ed. 45 7754–8
and also folding pathways. Site-directed mutagenesis studies [2] Hainfield J F and Powell R D 2000 New frontiers in gold
on unlabeled proteins that identify which amino acids are labeling J. Histochem. Cytochem. 48 471–80
[3] Bruchez M, Moronne M, Gin P, Weiss S and Alivisatos A P
necessary for stability can aid in the choice of a labeling site.
1998 Semiconductor nanocrystals as fluorescent biological
Some well-studied proteins have had every single amino acid labels Science 281 2013–6
in the sequence mutated. In addition, protein folding studies [4] Pengo P, Baltzer L, Pasquato L and Scrimin P 2007 Substrate
often identify which structures on the protein fold first or are modulation of the activity of an artificial nanoesterase
necessary for maintaining the entire protein structure. made of peptide-functionalized gold nanoparticles Angew.
Chem. Int. Ed. 46 400–4
Even if potential labeling sites can be identified, it is [5] Neuberger T, Schöpf B, Hofmann H, Hofmann M and
difficult to predict if a particular labeling site will not affect Rechenberg B V 2005 Superparamagnetic nanoparticles
the protein structure and activity when linked to a NP. One must for biomedical applications: possibilities and limitations
take into consideration that the NP is large with a considerable of a new drug delivery system J. Magn. Magn. Mater.
surface area accessible to the protein. If the NP is linked to the 293 483–96
[6] Ipe B, Shukla A, Lu H, Zou B, Rehage H and Niemeyer C
protein terminus, it can still interact with many other amino 2006 Dynamic light-scattering analysis of the electrostatic
acids on the protein. interaction of hexahistidine-tagged cytochrome P450
Finally, once researchers begin to systematically enzyme with semiconductor quantum dots Chem. Phys.
investigate the site dependence of NP labeling on protein, it Chem. 7 1112–8
[7] Pitsillides C, Joe E, Wei X, Anderson R and Lin C 2003
is our hope that a general strategy will emerge. Results from
Selective cell targeting with light-absorbing microparticles
these studies could identify which type of motif to label in and nanoparticles Biophys. J. 84 4023–32
a protein—such as those that are the most or least stable, or [8] Sarkar R, Narayanan S S, Palsson L-O, Dias F, Monkman A
closest to the termini, or that have the least-charged amino and Pal S K 2007 Direct conjugation of semiconductor
acids in the local vicinity. nanocrystals to a globular protein to study protein-folding
intermediates J. Phys. Chem. B 111 12294–8
[9] Niemeyer C 2003 Functional hybrid devices of proteins and
8. Issues and challenges ahead inorganic nanoparticles Angew. Chem. Int. Ed.
42 5796–800
[10] Xiao Y, Patolsky F, Katz E, Hainfeld J F and Willner I 2004
The field of NP labeling of proteins is relatively new and Plugging into enzymes : nanowiring of redox enzymes by
rapidly developing. Because most of the proteins that have a gold nanoparticle Science 299 1877–81
been labeled with NPs so far are small and relatively simple, [11] Willner I, Baron R and Willner B 2007 Integrated
the future of this field will focus on the labeling of larger nanoparticle-biomolecule systems for biosensing and
bioelectronics Biosens. Bioelectron. 22 1841–52
proteins that have much more complex functions. This
[12] Patolsky F, Weizmann Y and Willner I 2004 Actin-based
will undoubtedly be more challenging as engineering unique metallic nanowires as bio-nanotransporters Nat. Mater.
labeling sites may not be possible using all the strategies 3 692–5
above. Hainfeld and co-workers [45] have demonstrated [13] Gao X, Djalali R, Haboosheh A, Samson J, Nuraje N and
that site-specific labeling of a large protein complex can be Matsui H 2005 Peptide nanotubes: simple separation
using size-exclusion columns and use as templates for
achieved, enabling EM imaging that can reconstruct a three- fabricating one-dimensional single chains of an
dimensional representation of the complex. This also shows nanoparticles Adv. Mater. 17 1753–7
that spectroscopic techniques are continually improving and [14] Ratner B D 1995 Surface modification of polymers:
becoming more sophisticated. Finally, the next frontier with chemical, biological and surface analytical challenges
NP labeling is to move to labeling intracellularly and in vivo. Biosens. Bioelectron. 10 797–804
[15] Giepmans B N G, Adams S R, Ellisman M H and Tsien R Y
There are numerous exciting applications being developed for 2006 The fluorescent toolbox for assessing protein
NP in biology, which become all the more powerful when location and function Science 312 217–24
employed in cells, tissue and organisms. The strategies for [16] Griffin B A, Adams S R and Tsien R Y 1998 Specific
labeling will become much more complex as protein folding covalent labeling of recombinant protein molecules inside
is quite different inside the crowded environments in cells, live cells Science 287 269–72
[17] Chen I and Ting A Y 2005 Site-specific labeling of proteins
and techniques for assaying the structure and function will with small molecules in live cells Curr. Opin. Biotechnol.
also face new challenges. 16 35–40

14
Biomed. Mater. 3 (2008) 034001 M-E Aubin-Tam and K Hamad-Schifferli

[18] Hirshberg M, Henrick K, Haire L L, Vasisht N, Brune M, [38] Geoghegan W and Ackerman G 1977 Adsorption of
Corrie J E T and Webb M R 1998 Crystal structure of horseradish peroxidase, ovomucoid and
phosphate binding protein labeled with a coumarin anti-immunoglobulin to colloidal gold for the indirect
fluorophore, a probe for inorganic phosphate Biochemistry detection of concanavalin A, wheat germ agglutinin and
37 10381–5 goat anti-human immunoglobulin G on cell surfaces at the
[19] Honegger A, Spinelli S, Cambillau C and Pluckthun A 2005 electron microscopic level: a new method, theory and
A mutation designed to alter crystal packing permits application J. Histochem. Cytochem. 25 1187–200
structural analysis of a tight-binding fluorescein-scFv [39] Bayraktar H, You C-C, Rotello V M and Knapp M J 2007
complex Protein Sci. 14 2537–49 Facial control of nanoparticle binding to cytochrome c J.
[20] Brandenburg B, Lee L Y, Lakadamyali M, Rust M J, Zhuang Am. Chem. Soc. 129 2732–3
X and Hogle J M 2007 Imaging poliovirus entry in live [40] Hainfeld J F and Powell R D 2000 New frontiers in gold
cells PLoS Biol. 5 1543–55 labeling J. Histochem. Cytochem. 48 471–80
[21] Chamberlain C and Hahn K M 2000 Watching proteins in the [41] Jackson A M, Myerson J W and Stellacci F 2004
wild: fluorescence methods to study protein dynamics in Spontaneous assembly of subnanometre-ordered domains
living cells Traffic 1 755–62 in the ligand shell of monolayer-protected nanoparticles
[22] Daniel M-C and Astruc D 2004 Gold nanoparticles: Nat. Mater. 3 330–6
assembly, supramolecular chemistry, quantum-size-related [42] Pankau W M, Mönninghoff S and Kiedrowski G V 2006
properties, and applications toward biology, catalysis, and Thermostable and monoconjugable gold clusters with a
nanotechnology Chem. Rev. 104 293–346 dodecadentate thioether ligand gripper Angew. Chem. Int.
[23] Aubin-Tam M-E and Hamad-Schifferli K 2005 Gold Ed. 45 1889–91
nanoparticle cytochrome c complexes: the effect of [43] Abad J M, Mertens S F L, Pita M, Fernndez V M and
nanoparticle ligand charge on protein structure Langmuir Schiffrin D J 2005 Functionalization of thioctic acid-
21 12080–4 capped gold nanoparticles for specific immobilization of
[24] Ackerson C J, Jadzinsky P D, Jensen G J and Kornberg R D histidine-tagged proteins J. Am. Chem. Soc.
2006 Rigid, specific, and discrete gold nanoparticle/ 127 5689–94
antibody conjugates J. Am. Chem. Soc. 128 2635 [44] Hainfeld J F, Liu W, Halsey C M R, Freimuth P and Powell
[25] Montesano-Roditis L, Glitz D G, Traut R R and Stewart P L R D 1999 Ni-NTA-gold clusters target His-tagged proteins
2001 Cryo-electron microscopic localization of protein J. Struct. Biol. 127 185–98
L7/L12 within the Escherichia coli 70 S ribosome by [45] Hu M, Qian L, Briñas R P, Lymar E S and Hainfeld J F 2007
difference mapping and nanogold labeling J. Biol. Chem. Assembly of nanoparticle-protein binding complexes:
276 14117–23 from monomers to ordered arrays Angew. Chem. Int. Ed.
[26] Hostetler M J, Templeton A C and Murray R W 1999 46 5111–4
Dynamics of place-exchange reactions on [46] Kim S H, Jeyakumar M and Katzenellenbogen J A 2007
monolayer-protected gold cluster molecules Langmuir Dual-mode fluorophore-doped nickel nitrilotriacetic
15 3782–9 acid-modified silica nanoparticles combine
[27] Verma A and Rotello V M 2005 Surface recognition of histidine-tagged protein purification with site-specific
biomacromolecules using nanoparticle receptors Chem. fluorophore labeling J. Am. Chem. Soc. 129 13254–64
Commun. 3 303–12 [47] Medintz I L, Clapp A R, Mattoussi H, Goldman E R,
[28] Verma A, Nakade H, Simard J M and Rotello V M 2004 Fisher B and Mauro J M 2003 Self-assembled nanoscale
Recognition and stabilization of peptide a-helices using biosensors based on quantum dot FRET donors Nat.
templatable nanoparticle receptors J. Am. Chem. Soc. Mater. 2 630–8
126 10806–7 [48] Schmitt J, Hess H and Stunnenberg H 1993 Affinity
[29] Medintz I L, Uyeda H T, Goldman E R and Mattoussi H purification of histidine-tagged proteins Mol. Biol. Rep.
2005 Quantum dot bioconjugates for imaging, labelling 18 223–30
and sensing Nat. Mater. 4 435–46 [49] Wang S, Mamedova N, Kotov N A, Chen W and Studer J
[30] Rapino S and Zerbetto F 2007 Dynamics of thiolate chains on 2002 Antigen/antibody immunocomplex from CdTe
a gold nanoparticle Small 3 386–8 nanoparticle bioconjugates Nano Lett. 2 817–22
[31] Cedervall T, Lynch I, Lindman S, Thulin T B E, Nilsson H, [50] Huh Y M et al 2005 In vivo magnetic resonance detection of
Dawson K A and Linse S 2007 Understanding the cancer by using multifunctional magnetic nanocrystals
nanoparticle—protein corona using methods to quantify J. Am. Chem. Soc. 127 12387–91
exchange rates and affinities of proteins for nanoparticles [51] Green N M 1990 Avidin and streptavidin Methods Enzymol.
Proc. Natl. Acad. Sci. USA 104 2050–5 184 51–67
[32] Cedervall T, Lynch I, Foy M, Berggård T, Donnelly S C, [52] Bagalkot V, Zhang L, Levy-Nissenbaum E, Jon S,
Cagney G, Linse S and Dawson K A 2007 Detailed Kantoff P W, Langer R and Farokhzad O C 2007 Quantum
identification of plasma proteins adsorbed on copolymer dot-aptamer conjugates for synchronous cancer imaging,
nanoparticles Angew. Chem. Int. Ed. 46 5754–6 therapy, and sensing of drug delivery based on
[33] Lynch I, Dawson K A and Linse S 2006 Detecting cryptic bi-fluorescence resonance energy transfer Nano Lett.
epitopes created by nanoparticles Sci. STKE 327 14 7 3065–70
[34] Uchida M et al 2007 Biological containers: protein cages as [53] Chu T C, Shieh F, Lavery L A, Levy M, Richards-Kortum R,
multifunctional nanoplatforms Adv. Mater. 19 1025–42 Korgel B A and Ellington A D 2006 Labeling tumor cells
[35] Colfen H and Mann S 2003 Higher-order organization by with fluorescent nanocrystal-aptamer bioconjugates
mesoscale self-assembly and transformation of hybrid Biosens. Bioelectron. 21 1859–66
nanostructures Angew. Chem. Int. Ed. 42 2350–65 [54] Ulman A 1996 Formation and structure of self-assembled
[36] Sweeney R Y, Mao C B, Gao X X, Burt J L, Belcher A M, monolayers Chem. Rev. 96 1533
Georgiou G and Iverson B L 2004 Bacterial biosynthesis [55] Whitesides G M and Laibinis P E 1990 Wet chemical
of cadmium sulfide nanocrystals Chem. Biol. 11 1553–9 approaches to the characterization of organic surfaces:
[37] Reiss B D, Mao C B, Solis D J, Ryan K S, Thomson T and self-assembled monolayers, wetting, and the physical-
Belcher A M 2004 Biological routes to metal alloy organic chemistry of the solid–liquid interface Langmuir
ferromagnetic nanostructures Nano Lett. 4 1127–32 6 87–96

15
Biomed. Mater. 3 (2008) 034001 M-E Aubin-Tam and K Hamad-Schifferli

[56] McMillan R A, Paavola C D, Howard J, Chan S L, [72] Vertegel A A, Siegel R W and Dordick J S 2004 Silica
Zaluzec N J and Trent J D 2002 Ordered nanoparticle nanoparticle size influences the structure and enzymatic
arrays formed on engineered chaperonin protein templates activity of adsorbed lysozyme Langmuir 20 6800–7
Nat. Mater. 1 247–52 [73] Jiang X, Jiang J, Jin Y, Wang E and Dong S 2005 Effect of
[57] Patel J D, O’Carra R, Jones J, Woodward J G and colloidal gold size on the conformational changes of
Mumper R J 2007 Preparation and characterization of adsorbed cytochrome c: probing by circular dichroism,
nickel nanoparticles for binding to his-tag proteins and UV-visible, and infrared spectroscopy Biomacromolecules
antigens Pharm. Res. 24 343–52 6 46–53
[58] Veiseh M, Zareie M H and Zhang M 2002 Highly selective [74] Macdonald I D G and Smith W E 1996 Orientation of
protein patterning on gold-silicon substrates for biosensor cytochrome c adsorbed on a citrate-reduced silver colloid
applications Langmuir 18 6671–8 surface Langmuir 12 706–13
[59] Zheng M, Li Z and Huang X 2004 Ethylene glycol monolayer [75] Keating C D, Kovaleski K M and Natan M J 1998
protected nanoparticles: synthesis, characterization, and Protein:colloid conjugates for surface-enhanced Raman
interactions with biological molecules Langmuir scattering: stability and control of protein orientation
20 4226–35 J. Phys. Chem. B 102 9404–13
[60] Aubin M-E, Morales D G and Hamad-Schifferli K 2005 [76] Al-Bassam J, Ozer R S, Safer D, Halpain S and Milligan R A
Labeling ribonuclease S with a 3 nm Au nanoparticle by 2002 MAP2 and tau bind longitudinally along the outer
two-step assembly Nano Lett. 5 519–22 ridges of microtubule protofilaments J. Cell Biol.
[61] Narayanan S S, Sarkar R and Pal S K 2007 Structural and 157 1187–96
functional characterization of enzyme-quantum dot [77] Milligan R A, Whittaker M and Safer D 1990 Molecular
conjugates: covalent attachment of CdS nanocrystal to structure of F-actin and location of surface binding sites
a-chymotrypsin J. Phys. Chem. C 111 11539–43 Nature 348 217–21
[62] Mattoussi H, Mauro J M, Goldman E R, Anderson G P, [78] Barat C, Datta P P, Raj V S, Kaji M R S H, Kaji A and
Sundar V C, Mikulec F V and Bawendi M G 2000 Agrawal R K 2007 Progression of the ribosome recycling
Self-assembly of CdSe-ZnS quantum dot bioconjugates factor through the ribosome dissociates the two ribosomal
using an engineered recombinant protein J. Am. Chem. subunits Mol. Cell 27 250–61
Soc. 122 12142–50 [79] Datta P P, Sharma M R, Qi L, Frank J and Agrawal R K 2005
[63] Aubin-Tam M-E, Zhou H and Hamad-Schifferli K 2008 Interaction of the G0 domain of elongation factor G and
Structure of cytochrome c at the interface with magnetic the C-terminal domain of ribosomal protein L7/L12
CoFe2O4 nanoparticles Soft Matter. 4 554–9 during translocation as revealed by Cryo-EM Mol.
[64] Shieh D-B, Su C-H, Chang F-Y, Wu Y-N, Su W-C, Hwu J R, Cell 20 723–31
Chen J-H and Yeh C-S 2006 Aqueous nickel- [80] Wille H, Michelitsch M D, Guénebaut V, Supattapone S,
nitrilotriacetate modified Fe3O4-NH+3 nanoparticles for Serban A, Cohen F E, Agard D A and Prusiner S B 2002
protein purification and cell targeting Nanotechnology Structural studies of the scrapie prion protein by electron
17 4174–82 crystallography Proc. Natl. Acad. Sci. USA 99 3563–8
[65] Xu C J, Xu K M, Gu H W, Zhong X F, Guo Z H, [81] Rice S et al 1999 A structural change in the kinesin motor
Zheng R K, Zhang X X and Xu B 2004 Nitrilotriacetic protein that drives motility Nature 402 778–84
acid-modified magnetic nanoparticles as a general agent to [82] Wells A L, Lin A W, Chen L-Q, Safer D, Cain S M, Hasson
bind histidine-tagged proteins J. Am. Chem. Soc. T, Carragher B O, Milligan R A and Sweeney H L 1999
126 3392–3 Myosin VI is an actin-based motor that moves backwards
[66] Brennan J L et al 2006 Bionanoconjugation via click Nature 401 505–8
chemistry: the creation of functional hybrids of lipases [83] Lundqvist M, Nygren P, Jonsson B-H and Broo K 2006
and gold nanoparticles Bioconjug. Chem. 17 1373–5 Induction of structure and function in a designed peptide
[67] Becker C F W, Marsac Y, Moser P H J, Goody R S and upon adsorption on a silica nanoparticle Angew. Chem. Int.
Niemeyer C M 2007 Functional immobilization of the Ed. 45 8169–73
small GTPase Rab6A on DNA—gold nanoparticles by [84] Srivastava S, Verma A, Frankamp B L and Rotello V M 2005
using a site-specifically attached poly(ethylene Controlled assembly of protein-nanoparticle composites
glycol)linker and thiol place-exchange reaction Chem. Bio. through protein surface recognition Adv. Mater.
Chem. 8 32–6 17 617–21
[68] Xie H, Tkachenko A G, Glomm W R, Ryan J A, [85] Ionov R, El-Abed A, Goldman M and Perett P 2004
Brennaman M K, Papanikolas J M, Franzen S and Structural organization of alpha-helical peptide antibiotic
Feldheim D L 2003 Critical flocculation concentrations, alamethicin at the air/water interface J. Phys. Chem.
binding isotherms, and ligand exchange properties of B 108 8485–8
peptide-modified gold nanoparticles studied by [86] Sugiyama Y, Inoue Y, Muneyuki E, Haneda H and
UV-visible, fluorescence, and time-correlated single Fujimoto M 2006 AFM and TEM observations of a-helix
photon counting spectroscopies Anal. Chem. 75 5797–805 to b-sheet conformational change occurring on carbon
[69] Zanchet D, Micheel C M, Parak W J, Gerion D and nanotubes J. Electron. Microsc. 55 143–9
Alivisatos A P 2001 Electrophoretic isolation of discrete [87] Parak W J, Pellegrino T, Micheel C M, Gerion D,
Au nanocrystal/DNA conjugates Nano Lett. 1 32–5 Williams S C and Alivisatos A P 2003 Conformation of
[70] Mamedova N N, Kotov N A, Rogach A L and Studer J 2001 oligonucleotides attached to gold nanocrystals probed by
Albumin-CdTe nanoparticle bioconjugates: Preparation, gel electrophoresis Nano Lett. 3 33–6
structure, and interunit energy transfer with antenna effect [88] Park S, Brown K A and Hamad-Schifferli K 2004 Changes in
Nano Lett. 1 281–6 oligonucleotide conformation on nanoparticle surfaces by
[71] Soto C M, Blum A S, Wilson C D, Lazorcik J, Kim M, modification with mercaptohexanol Nano Lett.
Gnade B and Ratna B R 2004 Separation and recovery of 4 1925–9
intact gold-virus complex by agarose electrophoresis and [89] Casanova D, Giaume D, Moreau M, Martin J-L, Gacoin T,
electroelution: application to the purification of cowpea Boilot J-P and Alexandrou A 2007 Counting the number of
mosaic virus and colloidal gold complex Electrophoresis proteins coupled to single nanoparticles J. Am. Chem. Soc.
25 2901–6 129 12592–3

16
Biomed. Mater. 3 (2008) 034001 M-E Aubin-Tam and K Hamad-Schifferli

[90] Chen X-H et al 2005 Protein structural deformation induced [108] Zhang Y, He P and Hu N 2004 Horseradish peroxidase
lifetime shortening of photosynthetic bacteria immobilized in TiO2 nanoparticle films on pyrolytic
light-harvesting complex LH2 excited state Biophys. J. graphite electrodes: direct electrochemistry and
88 4262–73 bioelectrocatalysis Electrochim. Acta 49 1981–8
[91] De M, You C-C, Srivastava S and Rotello V M 2007 [109] Brown K R, Fox A P and Natan M J 1996 Morphology-
Biomimetic interactions of proteins with functionalized dependent electrochemistry of cytochrome c at Au colloid-
nanoparticles: a thermodynamic study J. Am. Chem. Soc. modified SnO2 electrodes J. Am. Chem. Soc. 118 1154–7
129 10747–53 [110] Huang Y, Zhang W, Xiao H and Li G 2005 An
[92] Gole A, Dash C, Soman C, Sainkar S R, Rao M and Sastry M electrochemical investigation of glucose oxidase at a CdS
2001 On the preparation, characterization, and enzymatic nanoparticles modified electrode Biosens. Bioelectron. 21
activity of fungal protease-gold colloid bioconjugates 817–21
Bioconjug. Chem. 12 684–90 [111] Zhao J, O’Daly J P, Henkens R W, Stonehuerner J and
[93] Gole A, Dash C, Ramakrishnan V, Sainkar S R, Crumbliss A L 1996 A xanthine oxidase/colloidal gold
Mandale A B, Rao M and Sastry M 2001 Pepsin-gold enzyme electrode for amperometric biosensor applications
colloid conjugates: preparation, characterization, and Biosens. Bioelectron. 11 493–502
enzymatic activity Langmuir 17 1674–9 [112] Laviron E 1979 General expression of the linear potential
[94] Huang Y-F, Huang C-C and Chang H-T 2003 Exploring the sweep voltammogram in the case of diffusionless
activity and specificity of gold nanoparticle-bound trypsin electrochemical systems J. Electroanal. Chem. 101 19–28
by capillary electrophoresis with laser-induced [113] Liu H, Tian Y and Deng Z 2007 Morphology-dependent
fluorescence detection Langmuir 19 7498–502 electrochemistry and electrocatalytical activity of
[95] Dyal A, Loos K, Noto M, Chang S W, Spagnoli C, cytochrome c Langmuir 23 9487–94
Shafi K, Ulman A, Cowman M and Gross R A 2003 [114] Pascher T, Chesick J P, Winkler J R and Gray H B 1996
Activity of Candida rugosa lipase immobilized on Protein folding triggered by electron transfer Science
gamma-Fe2O3 magnetic nanoparticles J. Am. Chem. Soc. 271 1558–60
125 1684–5 [115] Sondi I, Siiman O, Koester S and Matijević E 2000
[96] You C-C, De M, Han G and Rotello V M 2005 Tunable Preparation of aminodextran-CdS nanoparticle complexes
inhibition and denaturation of chymotrypsin with amino and biologically active antibody-aminodextran-CdS
acid-functionalized gold nanoparticles J. Am. Chem. Soc. nanoparticle conjugates Langmuir 7 3107–18
127 12873–81 [116] von Banchet G S and Schaible H-G 1999 Localization of the
[97] Bjerneld E J, Földes-Papp Z, Käll M and Rigler R 2002 neurokinin 1 receptor on a subset of substance P-positive
Single-molecule surface-enhanced Raman and and isolectin B4-negative dorsal root ganglion neurons of
fluorescence correlation spectroscopy of horseradish the rat Neurosci. Lett. 274 175–8
peroxidase J. Phys. Chem. B 106 1213–8 [117] Friedel M, Baumketner A and Shea J-E 2007 Stability of a
protein tethered to a surface J. Chem. Phys. 126 095101
[98] Kouassi G K, Irudayaraj J and McCarty G 2005 Examination
[118] Nordgren C E, Tobias D J, Klein M L and Blasie J K 2002
of cholesterol oxidase attachment to magnetic
Molecular dynamics simulations of a hydrated protein
nanoparticles J. Nanobiotechnol. 3 1–9
vectorially oriented on polar and nonpolar soft surfaces
[99] Asuri P, Karajanagi S S, Vertegel A A, Dordick J S and
Biophys. J. 83 2906–17
Kane R S 2007 Enhanced stability of enzymes adsorbed
[119] Luedtke W D and Landman U 1998 Structure and
onto nanoparticles J. Nanosci. Nanotechnol. 7 1675–8
thermodynamics of self-assembled monolayers on gold
[100] Phadtare S, Vinod V P, Mukhopadhyay K, Kumar A, nanocrystallites J. Phys. Chem. B 102 6566–72
Rao M, Chaudhari R V and Sastry M 2004 Immobilization [120] Ghorai P K and Glotzer S C 2007 Molecular dynamics
and biocatalytic activity of fungal protease on gold simulation study of self-assembled monolayers of
nanoparticle-loaded zeolite microspheres Biotechnol. alkanethiol surfactants on spherical gold nanoparticles
Bioeng. 85 629–37 J. Phys. Chem. C 111 15857–62
[101] Shleev S, Tkac J, Christenson A, Ruzgas T, Yaropolov A I, [121] Rocchiccioli-Deltcheff C, Franck R, Cabuil V and Massart R
Whittaker J W and Gorton L 2005 Direct electron transfer 1987 Surfacted ferrofluid: Interactions at the surfactant-
between copper-containing proteins and electrodes magnetic iron oxide interface J. Chem. Res. (S) 5 126–7
Biosens. Bioelectron. 20 [122] Wu N, Fu L, Su M, Aslam M, Wong K C and Dravid V P
[102] Wu Y and Hu S 2007 Biosensors based on direct electron 2004 Interaction of fatty acid monolayers with cobalt
transfer in redox proteins Microchim. Acta 159 1–17 nanoparticles Nano Lett. 4 383–6
[103] Chen X X, Ferrigno R, Yang J and Whitesides G A 2002 [123] Roddick-Lanzilotta A D and McQuillan A J 2002 An in situ
Redox properties of cytochrome c adsorbed on infrared spectroscopic study of glutamic acid and of
self-assembled monolayers: a probe for protein aspartic acid adsorbed on TiO2: implications for the
conformation and orientation Langmuir 18 7009–15 biocompatibility of titanium J. Colloid Interface Sci.
[104] Kasmi A E, Wallace J M, Bowden E F, Binet S M and 227 48–54
Linderman R J 1998 Controlling interfacial electron- [124] Mammen M, Choi S-K and Whitesides G M 1998 Polyvalent
transfer kinetics of cytochrome c with mixed interactions in biological systems: implications for design
self-assembled monolayers J. Am. Chem. Soc. and use of multivalent ligands and inhibitors Angew.
120 225–6 Chem. Int. Ed. 37 2754–94
[105] Katz E and Willner I 2004 Integrated nanoparticle- [125] Pasquato L, Pengo P and Scrimin P 2004 Functional gold
biomolecule hybrid systems: synthesis, properties, and nanoparticles for recognition and catalysis J. Mater. Chem.
applications Angew. Chem. Int. Ed. 43 6042–108 14 3481–7
[106] Gan X, Liu T, Zhong J, Liu X and Li G 2004 Effect of silver [126] Pons T, Medintz I L, Wang X, English D S and Mattoussi H
nanoparticles on the electron transfer reactivity and the 2006 Solution-phase single quantum dot fluorescence
catalytic activity of myoglobin Chem. Bio. Chem. resonance energy transfer J. Am. Chem.
5 1686–91 Soc. 128 15324–31
[107] Wang L and Wang E 2004 Direct electron transfer between [127] Aubin-Tam M-E, Hwang W and Hamad-Schifferli K 2008
cytochrome c and a gold nanoparticles modified electrode Site-directed nanoparticle labeling of cytochrome c (in
Electrochem. Commun. 6 49–54 preparation)

17