Metalloprotein-Based MRI Probes

Metalloprotein-based MRI probes
The MIT Faculty has made this article openly available. Please share
how this access benefits you. Your story matters.
Citation Matsumoto, Yuri and Alan Jasanoff. "Metalloprotein-based MRI

probes." FEBS Letters 587, 8 (January 2013): 1021-1029 © 2013
Federation of European Biochemical Societies
As Published http://dx.doi.org/10.1016/j.febslet.2013.01.044
Publisher Wiley
Version Author's final manuscript
Citable link https://hdl.handle.net/1721.1/126108
Terms of Use Creative Commons Attribution-NonCommercial-NoDerivs License
Detailed Terms http://creativecommons.org/licenses/by-nc-nd/4.0/

NIH Public Access
Author Manuscript
FEBS Lett. Author manuscript; available in PMC 2014 April 17.
Published in final edited form as:
NIH-PA Author Manuscript
FEBS Lett. 2013 April 17; 587(8): 1021–1029. doi:10.1016/j.febslet.2013.01.044.
Metalloprotein-based MRI probes

Yuri Matsumoto1 and Alan Jasanoff1,2,3,†
1Department of Biological Engineering, Massachusetts Institute of Technology, 77 Massachusetts
Ave., Rm. 16-561, Cambridge, MA 02139

2Departmentof Brain & Cognitive Sciences, Massachusetts Institute of Technology, 77
Massachusetts Ave., Rm. 16-561, Cambridge, MA 02139
3Departmentof Nuclear Science & Engineering, Massachusetts Institute of Technology, 77
Massachusetts Ave., Rm. 16-561, Cambridge, MA 02139
Abstract
Metalloproteins have long been recognized as key determinants of endogenous contrast in
magnetic resonance imaging (MRI) of biological subjects. More recently, both natural and
engineered metalloproteins have been harnessed as biotechnological tools to probe gene
expression, enzyme activity, and analyte concentrations by MRI. Metalloprotein MRI probes are
paramagnetic and function by analogous mechanisms to conventional gadolinium or iron oxide-
based MRI contrast agents. Compared with synthetic agents, metalloproteins typically offer worse
sensitivity, but the possibilities of using protein engineering and targeted gene expression
approaches in conjunction with metalloprotein contrast agents are powerful and sometimes
definitive strengths. This review summarizes theoretical and practical aspects of metalloprotein-
based contrast agents, and discusses progress in the exploitation of these proteins for molecular
imaging applications.
Keywords
metalloprotein; magnetic resonance imaging; protein engineering; contrast agent; sensor;
molecular imaging
Introduction
Metalloproteins are essential to life. Roughly a third of proteins are associated with metals,
most frequently magnesium, zinc, iron, and manganese, in order from most to least abundant
[1]. Metalloprostheses enable many of these proteins to play important roles in biological
processes, prominently including photosynthesis, respiration, and various oxidation
reduction reactions [2]. A particularly impressive example, the cytochrome b6f from plants
and photosynthetic bacteria, contains four types of heme, a magnesium porphyrin, and an
[2Fe-2S] cluster all in a single supramolecular assembly designed to drive proton gradient
formation using energy from light absorbed by the metal complexes [3]. Both optical and
electronic properties of the metal centers thus contribute to a reaction that ultimately gives
© 2013 Federation of European Biochemical Societies. Published by Elsevier B.V. All rights reserved.
†
address correspondence to: AJ, jasanoff@mit.edu.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our
customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of
the resulting proof before it is published in its final citable form. Please note that during the production process errors may be
discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Matsumoto and Jasanoff Page 2
rise to the nutrients used by most organisms on the planet. A far simpler but also famously
vital metalloprotein is hemoglobin (Hb). Hb contains four polypeptides, each bound to a
heme group. Oxygen binding to the metal centers induces a change in the coordination
geometry, which in turn drives a global conformation change that favors further oxygen
binding [3]. Via this mechanism, the chemistry of metal-ligand interactions governs the
ability of Hb to bind and release oxygen in physiologically appropriate concentration ranges.
Biophysical properties of metalloproteins and protein-metal interactions have led to a

number of biotechnological applications. The strength and specificity of metal chelation by
polypeptides gives rise to the well-known Ni2+ affinity purification method used to isolate
polyhistidine-tagged proteins [4]. Catalytic oxidation of chlorinated alkanes by a diiron
active center in methane monooxygenase from Methylosinus trichosporium OB3b has been
used for bioremediation of contaminated groundwater [5,6]. A redox-active metalloprotein,
azurin, has been used as a biotransistor [7] and component of a protein-based biomemory
device [8]. Metalloproteins are also a basis for monitoring physiology in living animals. The
most famous example is provided again by Hb, which due to the oxygen-dependent spectral
properties of its heme groups, and the relative transparency of tissue to wavelengths
differentially absorbed by oxy- and deoxy-Hb, is the basis for vital signs monitoring in
virtually every hospital in the world [9].
It has been recognized more recently that magnetic properties of metalloproteins also
provide important biotechnological capabilities. Many metalloproteins contain ions with

unpaired electrons, rendering them paramagnetic and detectable or manipulable by magnetic
tools. Accompanying the excitement around optogenetics [10] has been a particular interest
in using metalloproteins such as ferritin (Ft) to provide magnetic “handles” on cell function.
In a recent example, implanted cells overexpressing a Ft derivative were induced by
application of oscillating magnetic fields to secrete insulin in mice [11]. There is also a
suggestion that protein-catalyzed paramagnetic metal accumulation in cells could be used
for magnetic pull-down assays [12,13]. Finally, paramagnetic metalloproteins have been
used as contrast agents for magnetic resonance imaging (MRI). Here, the prospect of finding
MRI-detectable analogs to green fluorescent protein (GFP) has been an important
inspiration; a hope is that suitable proteins could be used as gene reporters and sensors
analogous to the various fluorescent and luminescent proteins that have transformed
research in cell biology over the past two decades.
Although the majority of MRI contrast agents have historically been based on small organic
molecules such as Gd3+ chelates [14,15], metalloprotein MRI contrast agents present a
variety of advantages, each of which may be important in distinct contexts. Thanks to the
advancement of molecular biology techniques, proteins are much easier to synthesize and
modify than organic molecules; numerous protein engineering techniques may be applied to
tune metalloprotein properties for MRI (reviewed in [16]). Due to their larger size, protein-
based contrast agents are retained in the blood pool for a longer time than small molecule
agents, allowing for longer imaging times in some types of experiments [17,18]. In some
cases, metalloproteins can indeed be targeted and expressed using gene delivery methods, a
la GFP [19,20]. Such reporters generate relatively static contrast, but can allow particular
types of cells to be tracked in vivo over time [21–23]. By furthermore sensitizing
metalloprotein contrast agents to analytes [24,25], additional information about
physiologically relevant signals can be obtained. Development of such environmentally-
sensitive agents is underway but largely in its infancy. In this article, we discuss the
characteristics of metalloprotein-based MRI contrast agents and review recent progress in
the development and applications of magnetically active proteins in these new spheres of
investigation.

Theoretical basis of MRI contrast agents

In proton MRI [26], the form of MRI most commonly applied in laboratories and clinics,
populations of nuclear spins arising from hydrogen nuclei primarily in water are perturbed
and monitored to generate images. At thermal equilibrium in a strong magnetic field (B0),
the water proton spins align weakly with the applied field and give rise to a net
“longitudinal” magnetization aligned with B0. This magnetization is unobservable, but can
be detected following application of radiofrequency energy pulses which tilt the
magnetization vector off of the B0 axis and give rise to a nonzero “transverse”
magnetization component. After excitation, the transverse magnetization component decays
away with a time constant T2 (the transverse relaxation time) and the overall magnetization
returns to thermal equilibrium with a time constant T1 (the longitudinal relaxation time). The
shorter T1 is, the more frequently an MRI signal can be repeatedly measured per unit time;
areas of a specimen with short T1 therefore give rise to a larger average MRI signal.
Conversely, the shorter T2 is, the more rapidly the observable component of magnetization
disappears, and the lower the MRI signal becomes.
Paramagnetic species influence contrast in MRI by reducing the T1 and T2 relaxation times
[27,28] (Figure 1). In both cases, accelerated relaxation arises from coupling between the
magnetic dipole of the contrast agent and the nuclear spins of water molecules that interact
with the agent through bonds (“inner sphere” interactions) or through space (“outer sphere”).
Relaxation rates R1 (= 1/T1) and R2 (= 1/T2) are generally linear with contrast agent
concentration. The slopes of these relationships are referred to as the T1 and T2 relaxivities,
r1 and r2, respectively, which measure the strength of the contrast agent and are expressed in
units of mM−1s−1. Greater relaxivity is beneficial to an MRI contrast agent, because the
agent can then be applied at lower doses or to greater effect at any given concentration. Both
r1 and r2 vary strongly with B0 and depend on physical parameters including the electron
spin number (S) or magnetic moment of the contrast agent, the number of coordinated inner
sphere water molecules (q), the time constant for inner sphere water exchange (τM), and the
rotational correlation time of the agent (τR). Inner sphere contributions to relaxivity are
described by the theory of Solomon, Bloembergen, and Morgan [29–31], and apply to
metalloproteins with adjustments to account for slow rotation in macromolecules [32]. Outer
sphere contributions are described by related theories applicable to mononuclear [33] and
particulate [34–36] metal complexes. Determinants of relaxivity are summarized in the
appendix to this article, and are also thoroughly discussed in a number of secondary
references [14,37–39]. Relaxivity determinants are important not only because they explain
how contrast agents may be optimized, but also because they provide potential mechanisms
for designing MRI-detectable sensors.
Most MRI contrast agents or sensors, including paramagnetic proteins, tend to affect T1-
weighted MRI scans more than T2-weighted scans, or vice versa, and are correspondingly
referred to as T1 or T2 agents. T1 agents have an r1/r2 ratio of 1–2 and generally contain one
or a small number of paramagnetic ions. Classical T1 agents are exemplified by complexes
of Gd3+ with small chelators like diethylenetriaminepenaacetic acid (DTPA) [40,41], and
are the most commonly applied agents in clincial MRI; analogous proteins can include
porphyrin prostheses or directly bound metal ions (Figure 1A). At typical field strengths for
clinical or preclinical MRI (> 1 T), most T1 agents have r1 values from 1–10 mM−1s−1. In
biological samples with T1 values typically near 1 s, T1 agents need to be applied at
concentrations near 100 μM to induce substantial contrast effects; 100 μM of a contrast
agent with r1 = 5 mM−1s−1, for instance, would induce an MRI signal increase of ~20%
against a background with a T1 of 1 s under conditions of optimal signal to noise ratio. Inner
sphere r1 contributions are usually particularly important for T1 agents, and T1-based

sensors typically undergo analyte-dependent changes in the parameters that most affect inner
sphere relaxation mechanisms, such as q and τR.
T2 agents have r2/r1 ratio greater than ~10 (a necessary condition because T2 values are
much shorter than T1 values in vivo) and are best exemplified by superparamagnetic
nanoparticles (SPNs) [42–44], contrast agents that incorporate discrete crystalline domains
that exhibit highly cooperative magnetic behavior. A biosynthetic analog to SPNs is ferritin
(Ft) [34], an iron storage protein that accumulates minerals in a 12 nm shell like structure
formed from 24 polypeptide chains (Figure 1B). SPN T2 agents produce high r2/r1 ratio
because they become magnetized by the B0 field and create microscopic magnetic
perturbations experienced by diffusing water molecules in solution. These perturbations
affect T2 relaxation more than T1, particularly for particles with highly magnetic mineral
cores over ~3 nm in diameter and at high B0 strengths (> 1 T), where SPN magnetizations
tend toward an asymptotic “saturation” point [45]. The physics of the interaction between
diffusing protons and SPNs also leads to a strong dependence of r2 on R2/D, where R is the
mineral core radius and D is the solvent self-diffusion constant (see appendix); SPNs with
larger size shorten T2 more effectively, up to a so-called static dephasing limit [46] near ~50
nm for typical SPNs. Most SPN contrast agents incorporate iron oxide; synthetic iron-
containing SPNs usually have r2 values of 50–500 (mM Fe)−1s−1. Given typical background
T2 values near 100 ms in tissue, synthetic SPNs applied at concentrations of 1–10 μM Fe
can produce substantial effects under optimal T2-weighted imaging conditions. For instance,
an agent with r2 = 200 mM−1s−1 could produce a ~20% decrease in MRI signal at a
concentration of 10 μM Fe. Because of the dependence of r2 on particle size for SPN agents,
sensors can be constructed by coupling analyte concentration to the clustering of these
particles [47,48], which results in de facto size changes [49].
Relaxivity of protein contrast agents

Protein contrast agents have distinct advantages and disadvantages compared with
conventional synthetic contrast agents in terms of relaxivity. In most cases, proteins are
handicapped by incorporating metal ions with electronic properties that are suboptimal for
R1 or R2 enhancement [50]. Naturally-occurring paramagnetic proteins tend to contain Cu2+,
Mn2+, Mn3+, Fe2+, or Fe3+ ions, ranging in spin number from 1/2 to 5/2. On the other hand,
Gd3+ ions used in most synthetic small molecule agents provide S = 7/2. Since relaxivity
values are approximately proportional to S(S+1), the maximum relaxivity of a gadolinium
agent is about twice what would be achieved in principle with a transition metal-containing
protein with otherwise equivalent relaxivity parameters. Electronic relaxation times T1e and
T2e have a great influence on inner sphere relaxivities and also tend to be less advantageous
(shorter) for transition metals than for gadolinium. Short electronic relaxation times limit
relaxivity when they are less than the molecular motion timescale τR (~10 ns), as has been
reported for some metalloproteins. A further limitation on the relaxivity of metalloproteins,
compared with small metal complexes, can arise from the relative inaccessibility of protein-
coordinated metal ions to outer sphere water molecules. At field strengths above 1.5 T, outer
sphere interactions account for a sizeable fraction of the r1 of typical small molecule contrast
agents [37,51], a contribution that could be reduced due to steric effects in a macromolecule.
The limitations of metalloprotein contrast agents are partially offset by properties that are
predicted to benefit relaxivity. At moderate magnetic field strengths and for molecules with
sufficiently long electronic relaxation times, inner sphere relaxivity tends to be greater for
molecules with longer τR [52]. Approximate τR values can be estimated by the Stokes-
Einstein relationship [53] and are proportional to molecular size; values of 10 ns or above
are typical of proteins (> 20 kD), whereas small molecules generally have τR less than 1 ns.
Water structure around metalloproteins may also be conducive to greater relaxivity. Many

paramagnetic proteins provide more than one water-accessible coordination site per metal
ion (q > 1), compared with q = 1 for conventional Gd3+-containing contrast agents; this is a
substantial benefit because inner sphere relaxivity scales with q. Although complexes with
higher q bind metal ions less tightly, the potential health risk from complex dissociation is
mitigated by the fact that naturally abundant transition metals are far less toxic than Gd3+ at
low doses. The presence of additional bound water molecules associated with
metalloproteins, including “second sphere” waters near but not directly coordinated to
paramagnetic ions, may confer a further relaxivity gain [54]. A final relaxivity-related
advantage of protein contrast agents over some small molecule agents is their relative
solubility. Most cytosolic or secreted metalloproteins are evolved to remain in solution or
interact with well-defined ligands; this limits the potential for biological environments to
substantially degrade relaxivity or analyte sensing capabilities by adversely affecting water
proton interaction parameters.
The preceding discussion of advantages and disadvantages relates most directly to synthetic
vs. metalloprotein T1 agents, but analogous criteria differentiate synthetic T2 agents (SPNs)
from Ft as well. Synthetic iron oxide SPN contrast agents contain magnetite or maghemite,
magnetic materials with saturation magnetization (MS) values of 92–100 or 60–80 emu/g,
respectively [55], whereas Ft naturally contains a hydrated iron oxide called ferrihydrite,
which has a reported MS of only 0.9–1.2 emu/g [56]. Because the expected T2 effects of iron
oxide cores are proportional to their magnetization, the relaxivity of Ft at saturating B0 is in
principle only about 1% that of synthetic iron oxides for equivalent core sizes. The situation
with ferrihydrite-loaded Ft may be more complex, however, as its r2 appears to increase
linearly with field at least up to 11.7 T, as opposed to reaching a saturating value around 1 T
like most SPNs [57]. Further, it has been shown that Ft r2 is more directly dependent on
stored iron concentration than on core size per se [58], suggesting an important role for inner
sphere relaxivity mechanisms akin to those of conventional paramagnetic, as opposed to
superparamagnetic, contrast agents. Practical advantages of Ft compared with synthetic
SPNs include its regular structure and predictable size, both of which can aid in
characterization or engineering the relaxivity of Ft-based contrast agents.
Natural metalloproteins in MRI

The first substantial investigation of magnetic properties in a metalloprotein centered on
hemoglobin, the oxygen transporting heme protein in blood. Pauling and Coryell reported in
a 1936 paper that hemoglobin is paramagnetic in the absence of ligand but diamagnetic
when bound to oxygen [59]. This discovery eventually led to the development of blood
oxygenation level dependent (BOLD) technique for functional neuroimaging (fMRI) studies
[60], which is by far the most commonly exploited example of metalloprotein-induced
contrast in MRI. In the BOLD effect, T2-related signal changes are produced by variations
in the oxygenation of hemoglobin in blood vessels. Although deoxyhemoglobin can act as a
contrast agent through direct interactions with water molecules, its dominant contribution in
the BOLD effect is to change the overall magnetic susceptibility of erythrocytes and blood
vessels in their entirety, converting these structures into cellular-scale “contrast agents” that
alter MRI signal by outer sphere effects analogous to those of SPN agents [61]. The BOLD
effect is enabled by the large concentration of hemoglobin, ~150 g/L in whole blood [62],
giving rise to ~4 mM Fe, which ensures that even a relatively low percentage of
deoxygenation translates into a formidable deoxyhemoglobin concentration. In BOLD
fMRI, brain activity-induced increases in the blood supply to affected areas induces transient
drops in the concentrations of deoxyhemoglobin, which are detectable as localized MRI
signal increases [63–65]. BOLD imaging has been applied to detect hemodynamic changes
in other tissues as well [66,67], and a BOLD-like effect produced by paramagnetic
deoxymyoglobin in muscle can also be used to monitor aspects of muscular physiology [68].

More recently, interest in exploiting natural metalloproteins for MRI contrast has revolved
largely around the potential for expressed paramagnetic proteins to act as gene reporters. An
early effort to express myoglobin in transgenic mice did not produce substantial MRI
contrast changes [69], but contrast changes have been achieved by overexpressing Ft in cells
and animals. Ft has been reported to bind up to ~4500 Fe atoms per 24-mer [70], providing a
stoichiometric advantage of over a hundred, compared with heme proteins, in terms of sheer
iron accumulation. Although many of the iron atoms in Ft are organized into
“antiferromagnetic” ferrihydrite domains [71], the limited per-iron relaxivity of Ft is still
compensated for by the number of atoms that can be stored. The effect of overexpressing Ft
was first demonstrated in C6 glioma cells transfected with murine heavy chain Ft (one of
two isoforms, heavy and light, expressed in mammals) [72]. A 2005 study then showed for
the first time that ectopic (adenovirus driven) Ft overexpression leads to detectable T2
contrast in rodent brains [73], and later papers reported T2 effects in transgenic mice [74]
and transplanted cells in vivo [21–23].
A requirement for efficacy of the Ft reporter approach is that highly regulated endogenous
iron dynamics must be effectively perturbed by overexpression of the metalloprotein without
toxic effects on cells; this criterion might be satisfied to varying extents in different tissues
or cell types. Indeed, as potential gene reporters, metalloproteins in general are limited by
the kinetics of metal ion transport and processing. Iron uptake by cells, for instance, depends
on the import rate and availability of extracellular iron sources, and can require exposure
times on the order of hours to achieve saturation [75–77]. To address the potentially
restrictive role of iron import machinery, one study proposed co-expressing Ft with the
transferrin receptor, a natural participant in cellular iron transport [78]. A conceptually
related strategy involves using overexpression of paramagnetic metal ion transporters
themselves to induce MRI contrast; contrast changes in rodents have been reported using the
magnetotactic bacterial protein MagA [79] and the mammalian divalent cation transporter
DMT1 [80], but it is not clear which specific metalloproteins might be involved in binding
the metal ions once they have entered cells.
Engineering protein-based contrast agents to detect biological targets

One of the greatest strengths of protein contrast agents with respect to small molecules is in
the area of biomolecular target detection, the objective of “molecular imaging.” The large
surface area and numerous functional groups presented by protein interfaces render these
macromolecules especially suited to binding potential ligands with high affinity and
specificity. Although the potential utility of natural metalloproteins for MRI is far from fully
explored, however, it is unusual to be able to find a naturally occurring paramagnetic
molecule that spontaneously fits the demands of a particular biosensing application. Here
another strength of proteins—their amenability to engineering approaches—comes into play.

Whereas modifying small molecule agents often requires complicated synthetic schemes,
proteins may be engineered using simple DNA-level changes or straightforward post-
translational approaches [16].
The simplest examples of how proteins may be modified to facilitate analyte detection
include the use of bioconjugation strategies to attach protein targeting domains to metal
compounds, creating metal-protein hybrids with desirable binding and relaxivity properties.
This approach was first applied to attach Gd3+ ions to antibodies esterified to DTPA or
similar chelating groups [81–83], in the expectation that the resulting complexes could be
used to home in on antigens in vivo and facilitate their visualization by MRI. A recent
achievement using this type of approach involved the modification of the neuronal tract
tracer cholera toxin subunit B with Gd3+-tetraazacyclododecanetetraacetic acid (Gd-DOTA)
complexes [84]. The resulting molecules were injected into rat brains, where they were

taken up by neurons and transported to distal areas connected to the injection site, exposing
patterns of connectivity in the brain. Protein bioconjugation strategies have also been
extensively applied in combination with SPNs, which provide substantially greater contrast
per metal atom than paramagnetic T1 agents [36]. Individual SPN particles can be attached
to multiple targeting proteins, adding the potential for improved targeting through avidity
effects. SPN-protein conjugates have been produced to target genetically-expressed
reporters [85], markers of apoptosis [86], vascular pathology [87], and cancer cells [88,89].
Analyte sensors have also been actuated by proteins that bring about reversible clustering of
SPNs in the presence of various targets [90,91]. In one instance, a calcium ion sensor was
made by attaching SPNs to two proteins that bind to each other in the presence but not the
absence of Ca2+ [91]. Calcium-dependent SPN aggregation was observed and response
properties could be tuned by further engineering of the protein domains.
Canonical metalloproteins have also been engineered to act as MRI probes. Heme proteins
are particularly attractive bases for MRI sensors because of their high metal affinity and
diverse ligand binding functionality, but their properties typically need to be adjusted. For
instance, Hb is the quintessential metalloprotein-based sensor [60], but its use as an explicit
probe for measuring tissue oxygen pressure (pO2) by MRI is hindered by the fact that its
natural binding midpoint (EC50) is near a pO2 of 8 mmHg, below concentrations normally
found in tissue. Crosslinking Hb with glutaraldehyde improves its stability and shifts to the
EC50 to 38 mmHg [92], allowing it to sense physiological relevant pO2 levels in tissue with
a Δr2 of ~7 mM−1s−1 at 14.1 T [93]. In a far more generalizable example of heme protein
engineering for MRI, a bacterial cytochrome P450 heme domain (BM3h) that normally
binds unsaturated fatty acids was retuned through the process of directed evolution [94] to
bind and sense neural signaling molecules [24] in which binding of the analytes competes
with inner sphere water to bring about a change in q (Figure 2A). Over five rounds of
random mutagenesis and selection based on an optical titration screen, BM3h variants were
obtained with micromolar affinity for dopamine, a neurotransmitter involved in reward-
related signaling in the brain. These sensors had Δr1 of ~1 mM−1s−1 at 4.7 T (Figure 2B), a
modest change, but one that nevertheless permitted detection of stimulus-induced dopamine
release in rats. The same directed evolution strategy could also be applied to generate
sensors for other targets, such as the neurotransmitter serotonin [24].
Ft is an obvious platform for engineering MRI probes because of its similarity to SPNs and
the evidence that endogenous [95–97] or ectopic [73,74] Ft expression affects MRI contrast
in vivo. The most natural way to engineer Ft without disrupting its metal storage
functionality is to “display” targeting moieties on the protein surface. In one example of this
approach, Uchida et al. introduced a tumor targeting peptide, RGD-4C, at the N-terminus of
human heavy chain Ft and confirmed that the engineered protein (RGD4C-Fn) can still
mineralize iron oxide in its cavity [98]. RGD4C-Fn was shown to bind melanoma cells
better than unmodified Ft. The solvent exposed Ft N-terminus can also be modified to create
responsive MRI contrast agents. Following this strategy, Shapiro et al. developed a protein
kinase A (PKA) activity sensor based on Ft, which produces aggregation-dependent T2-
weighted MRI contrast changes analogous to those obtained using synthetic SPNs [99]. The
PKA sensor contains two populations of engineered Ft particles, one fused with the kinase
inducible domain (KID) of the transcription factor CREB, and the other with KIX domain of
the protein CBP. Upon phosphorylation of the KID domain by PKA, KID and KIX domains
interact with each other [100] and induce clustering of KID-Ft with KIX-Ft, increasing the r2
of the sensor (Figure 2C–D). These sensors have not yet been applied in vivo, but may have
the potential to allow genetically targeted functional imaging of cell or tissue types because
of the all-protein nature of the Ft-based sensors and the important role of kinases in
mediating cell signaling processes. Responsive Ft derivatives have also been constructed by
chemically crosslinking Ft to reversible polymerizing domains [101].

Engineering metalloproteins for high relaxivity

The chief limitation of engineered metalloproteins for targeting, sensing, and gene reporting
applications in MRI is their low relaxivity. For this reason, there is great interest in
engineering metalloproteins with higher r1 and r2 than natural proteins. One approach to this
problem is to mutate metalloprotein polypeptide sequences and examine effects on
relaxivity. In one instance of this approach, a library of P450 BM3h domains selected for
neurotransmitter affinity was examined for r1 variation [24]. A range of values from 0.7 to
1.9 mM−1s−1 at 4.7 T was found among mutants that differed primarily in residues near the
ligand binding site proximal to the heme. The shapes of so-called nuclear magnetic
relaxation dispersion (NMRD) curves, which plot relaxation rate as a function of B0 field
strength, combined with X-ray crystallographic analysis, suggested that r1 differences arose
from minor variations among multiple relaxivity determinants, with subtle changes in the
metal-proton distance for inner sphere water (2.6–3.0 Å) having greatest effect on r1 (Figure
3). Unfortunately it was not possible to verify the inferred distance changes at the resolution
of the crystal structures, but it could be possible in the future to apply rational design
principles to bring about similar changes. The NMRD results more generally indicate that
nontrivial enhancement of metalloprotein relaxivity is possible through mutagenesis and
screening-based approaches. A gain in the MRI contrast induced by human Ft expression
has also been achieved by amino-acid level changes. Functionality of mammalian Ft is
normally quite sensitive to the balance between the Ft heavy and light chains. Iordanova et
al. fused H and L chains together and showed that the resulting chimera improved relaxation
rates in U2OS cells were improved by roughly 50% [102]. Although it is unclear whether
any difference in r2 per iron or per Ft was achieved, cells harboring the construct appeared to
contain significantly more iron than cells expressing Ft heavy and light chains separately.
Some of the most severe limits on metalloprotein relaxivity come from the characteristics of
naturally bound metal ions themselves, and in particular from the spin numbers (S ≤ 5/2) of
bound transition metal ions. For this reason, a second strategy for creating metalloproteins
with enhanced relaxivity is to engineer the metal content of proteins, rather than simply
modifying their polypeptide sequences. This strategy was applied to BM3h, which normally
contains a low spin (S = 1/2) Fe3+ ion [103]. By substituting the native ferric heme with a
Mn3+ protoporphyrin complex (S = 2), a gain in relaxivity by a factor of 2.5 was obtained.
In principle, the strategy should have resulted in an eight-fold improvement in r1, all else
being equal, suggesting that further relaxivity enhancements might be possible by using
mutagenesis approaches in parallel to metal substitution. Non-native Mn3+ incorporation
into BM3h was made possible within bacterial cells by coexpressing BM3h with a porphyrin
transporter called ChuA. This facilitates large scale production of metal-substituted protein
in bacteria, as well as possible applications requiring intracellular compartmentalization of
the Mn3+-containing protein variants.
Metal substitution has also been possible with Ft, which can be engineered to contain
mineral species with higher magnetization than the natural ferrihydrite core material. The
most impressive example of this approach has been the creation of “magnetoferritin”, an Ft
complex in which magnetite is mineralized in the protein shell [104–106]. This can be
accomplished by incubating purified apoferritin under controlled pH and oxygenation
conditions in the presence of iron salts, and results in particles with a reported r2 of 78 (mM
Fe)−1s−1 at 1.5 T and 37 °C, comparable to synthetic SPNs and two orders of magnitude
higher than the r2 of ferrihydrite-loaded Ft. Yet higher relaxivity has been reported
following mineralization of gadolinium oxide (r2 = 240 mM−1s−1 at 1.5 T) [107] in Ft
variants, and moderate T1 relaxivity has been reported following mineralization of
manganese oxyhydroxide (r1 = 6 mM−1s−1 at 0.47 T) [108]. This strategy has also been
adapted to convert viral capsids into protein-based MRI contrast agents [109–111]. Finally,

Ft has also been used as a compartment for entrapment of soluble Gd-DOTA-related T1

agents [112]. This approach takes advantage of the protein’s shell structure and availability
of exchangeable protons at the protein surface, but not its mineral nucleation capabilities per
se. Very high r1 values, near 80 mM−1s−1, were reported at 0.47 T. A variant of this strategy
was also used to entrap Mn2+ ions in a Ft variant engineered to contain metal ions better
than native Ft [113]; documented r1 and r2 values were 10 mM−1s−1 and 74 mM−1s−1,
respectively, at 3 T. A disadvantage of most metal substitution approaches for Ft and other
proteins is that they cannot be implemented in cells expressing a metalloprotein, with
examples like the BM3h/ChuA strategy being occasional exceptions. Metal substituted
proteins could nevertheless be useful contrast reagents for exogenous application in MRI
experiments, however.
A different technique for engineering metalloprotein contrast agents is to graft metal binding
functionality onto proteins or polypeptides that do not normally bind metals. One version of
this route was taken with the gadolinium compound MS-325, which was designed to bind
noncovalently to serum albumin in the bloodstream [17]. The resulting complex has higher
relaxivity at B0 values because of the long τR of the albumin protein. Karfeld et al. took a
somewhat different approach by conjugating a large number of gadolinium chelators
covalently via lysine side chains in a family of engineered repetitive polypeptides [114,115].
The authors found that the amino acid sequence of the polypeptide could be modified to
produce substantial changes in T1 relaxivity. A parent sequence with ~22 kD molecular
weight bearing eight Gd3+ per molecule displayed a per gadolinium r1 of 8.8 mM−1s−1 at 1.4
T, but bioconjugates with denser spacing of conjugation sites and more Gd3+ ions per
polypeptide achieved r1 values up to 12.1 mM−1s−1. Relaxivities of up to 14.6 mM−1s−1
could also be obtained by increasing the molecular weight of the complexes up to ~80 kD.
Polypeptide contrast agents that chelate lanthanides without the need for bioconjugation
have also been devised. In one example, an immunoglobulin domain was mutated to
introduce multiple acidic sidechains in a cluster at the protein surface [116]. One variant
formed a Gd3+ complex with a reported r1 = 117 mM−1s−1 at 1.5 T, and with an apparent Kd
< 1 pM and q = 2. Several efforts to construct artificial metalloprotein complexes have made
use of the EF hand motif [117,118], a sequence derived from calcium binding proteins such
as parvalbumin. Caravan et al. inserted an EF hand motif into a DNA-binding helix-loop-
helix motif to create a chimeric DNA-sensing contrast agent [119]. The Gd3+ complex has
reported r1 values of 21 and 42 mM−1s−1 at 1.4 T in the absence and presence of DNA,
respectively. In another study, the inherent promiscuity of metal binding by the EF hand
motif was used to generate a sensor that functions by calcium-dependent displacement of
Mn2+ ions associated with the protein calmodulin [120]. Longitudinal relaxivity values of 11
and 8 mM−1s−1 at 4.7 T were reported in the absence and presence of 1 mM Ca2+,
respectively. To improve the affinity and specificity of EF hand motifs for lanthanide
binding, another group used a luminescence assay to screen libraries of troponin-derived
peptides for Tb3+ binding [121,122]. T1 relaxivities were subsequently measured from Gd3+
complexes of several variants of an optimized peptide sequence [123]. Values of up to 5.9
mM−1s−1 at 14 T were observed, despite an apparent q = 0 for the sequence with the highest
relaxivity. Several of the lanthanide binding tags also retained r1 values of 2.3–5.0 mM−1s−1
when fused to the small protein ubiquitin, and crystallographic analysis of the fusion
proteins reveals that the Gd3+-binding peptide domains are capable of forming q = 1 chelate
complexes similar to Gd-DOTA (Figure 4). Although r1 values of these domains do not
exceed those of conventional MRI contrast agents, the ability to incorporate high affinity
Gd3+-binding moieties into genetically encodable proteins may facilitate application of
protein engineering techniques to further enhance relaxivity or target analytes of interest.

Conclusions
The continued development and exploitation of metalloprotein contrast agents for MRI
poses unique opportunities in the fields of molecular imaging and protein engineering. It is
still uncertain what applications protein-based MRI probes will be most successful for, but
utility for measuring macromolecular biokinetics and sensing a variety of ligands in vivo
seems within reach. Only metalloproteins that do not require chemical modification or
artificial transmetallation are applicable as endogenously expressed reporters, but efforts to
improve relaxivity of these agents may ultimately yield valuable biotechnological tools,
perhaps alongside recently-engineered diamagnetic proteins detectable by chemical
exchange saturation contrast [124]. Further analysis of structure-activity relationships in
metalloprotein contrast agents will be interesting from a basic chemical perspective, and
could also guide design of synthetic contrast agents with enhanced properties. In each of
these contexts, the array of powerful genetic techniques for engineering and expressing
proteins will be a tremendous benefit, and helps make continued research in this area a
potential rewarding investment.
Supplementary Material
Refer to Web version on PubMed Central for supplementary material.
Acknowledgments
The authors were supported by NIH grants DP2-OD002114, R01-DA028299, and R01-NS076462, DARPA grant
W911NF-10-0059, and a grant from the Simons Center for the Social Brain at MIT.
Abbreviations
BBB Blood-Brain Barrier

BOLD Blood Oxygen Level Dependent
CNS Central Nervous System
BM3h Cytochrome P450 BM3 Heme Domain
EC50 Effective Concentration Midpoint
T1e Electronic Relaxation Time
S Electron Spin Number
Ft Ferritin
fMRI Functional Magnetic Resonance Imaging

Gd-DOTA Gadolinium-Tetraazacyclododecanetetraacetic Acid
Gd-DTPA Gadolinium-, Diethylenepentaaminetetrace-tic Acid
GFP Green Fluorescent Protein
Hb Hemoglobin
q Inner Sphere Coordination Number
τM Inner Sphere Water Exchange Time Constant
KID Kinase Inducible Domain
T1 Longitudinal Relaxation Time

R1 Longitudinal Relaxation Rate

r1 Longitudinal Relaxivity
B0 Magnetic Field Strength
MRI Magnetic Resonance Imaging
NMRD Nuclear Magnetic Relaxation Dispersion
pO2 Oxygen Pressure
PKA Protein Kinase A
τR Rotational Correlation Time
SPN Superparamagnetic Nanoparticle
Tf Transferrin
T2 Transverse Relaxation Time
R2 Transverse Relaxation Rate
r2 Transverse Relaxivity
References
1. Waldron KJ, Rutherford JC, Ford D, Robinson NJ. Metalloproteins and metal sensing. Nature.
2009; 460:823–830. [PubMed: 19675642]
2. Lu Y, Yeung N, Sieracki N, Marshall NM. Design of functional metalloproteins. Nature. 2009;
460:855–62. [PubMed: 19675646]
3. Voet, D.; Voet, JG. Biochemistry. J. Wiley & Sons; New York: 2004.
4. Hochuli E, Bannwarth W, Döbeli H, Gentz R, Stüber D. Genetic approach to facilitate purification
of recombinant proteins with a novel metal chelate adsorbent. Nat Biotechnol. 1988; 6:1321–1325.
5. Hazen, TC. Cometabolic Bioremediation. In: Timmis, KN., editor. Handbook of Hydrocarbon and
Lipid Microbiology. Springer Berlin Heidelberg; Berlin, Heidelberg: 2010. p. 2505-2514.
6. Fox BG, Borneman JG, Wackett LP, Lipscomb JD. Haloalkene oxidation by the soluble methane
monooxygenase from Methylosinus trichosporium OB3b: mechanistic and environmental
implications. Biochemistry. 1990; 29:6419–6427. [PubMed: 2207083]
7. Alessandrini A, Salerno M, Frabboni S, Facci P. Single-metalloprotein wet biotransistor. Appl Phys
Lett. 2005; 86:133902–4.
8. Yagati AK, Lee T, Min J, Choi JW. A robust nanoscale biomemory device composed of
recombinant azurin on hexagonally packed Au-nano array. Biosens Bioelectron. 2013; 40:283–90.
[PubMed: 22884649]
9. Kram HB. Noninvasive tissue oxygen monitoring in surgical and critical care medicine. Surg Clin
North Am. 1985; 65:1005–24. [PubMed: 3901341]
10. Boyden ES, Zhang F, Bamberg E, Nagel G, Deisseroth K. Millisecond-timescale, genetically
targeted optical control of neural activity. Nat Neurosci. 2005; 8:1263–8. [PubMed: 16116447]
11. Stanley SA, Gagner JE, Damanpour S, Yoshida M, Dordick JS, Friedman JM. Radio-wave heating
of iron oxide nanoparticles can regulate plasma glucose in mice. Science. 2012; 336:604–608.
[PubMed: 22556257]
12. Nishida K, Silver PA. Induction of biogenic magnetization and redox control by a component of
the target of rapamycin complex 1 signaling pathway. PLoS Biol. 2012; 10:e1001269. [PubMed:
22389629]
13. Kim T, Moore D, Fussenegger M. Genetically programmed superparamagnetic behavior of
mammalian cells. J Biotechnol. 2012; 162:237–245. [PubMed: 23036923]

14. Toth, E.; Helm, L.; Merbach, AE. Relaxivity of gadolinium (III) complexes: theory and
mechanism. In: Merbach, AE.; Toth, E., editors. The chemistry of contrast agents in medical
magnetic resonance imaging. J. Wiley & Sons; New York: 2001. p. 45-120.
15. Werner EJ, Datta A, Jocher CJ, Raymond KN. High-relaxivity MRI contrast agents: where
coordination chemistry meets medical imaging. Angew Chem, Int Ed Engl. 2008; 47:8568–80.
[PubMed: 18825758]
16. Hsieh V, Jasanoff A. Bioengineered probes for molecular magnetic resonance imaging in the
nervous system. ACS Chem Neurosci. 2012; 3:593–602. [PubMed: 22896803]
17. Caravan P, et al. The interaction of MS-325 with human serum albumin and its effect on proton
relaxation rates. J Am Chem Soc. 2002; 124:3152–3162. [PubMed: 11902904]
18. Schmiedl U, Ogan MD, Moseley ME, Brasch RC. Comparison of the contrast-enhancing
properties of albumin-(Gd-DTPA) and Gd-DTPA at 2.0 T: and experimental study in rats. Am J
Roentgenol. 1986; 147:1263–1270. [PubMed: 3535459]
19. Gilad AA, Ziv K, McMahon MT, van Zijl PC, Neeman M, Bulte JW. MRI reporter genes. J Nucl
Med. 2008; 49:1905–8. [PubMed: 18997049]
20. Westmeyer GG, Jasanoff A. Genetically controlled MRI contrast mechanisms and their prospects
in systems neuroscience research. Magn Reson Imaging. 2007; 25:1004–10. [PubMed: 17451901]
21. Vande Velde G, et al. Quantitative evaluation of MRI-based tracking of ferritin-labeled
endogenous neural stem cell progeny in rodent brain. Neuroimage. 2012; 62:367–80. [PubMed:
22677164]
22. Iordanova B, Ahrens ET. In vivo magnetic resonance imaging of ferritin-based reporter visualizes
native neuroblast migration. Neuroimage. 2012; 59:1004–12. [PubMed: 21939774]

23. Naumova AV, Reinecke H, Yarnykh V, Deem J, Yuan C, Murry CE. Ferritin overexpression for
noninvasive magnetic resonance imaging-based tracking of stem cells transplanted into the heart.
Mol Imaging. 2010; 9:201–10. [PubMed: 20643023]
24. Brustad EM, Lelyveld VS, Snow CD, Crook N, Jung ST, Martinez FM, Scholl TJ, Jasanoff A,
Arnold FH. Structure-guided directed evolution of highly selective p450-based magnetic
resonance imaging sensors for dopamine and serotonin. J Mol Biol. 2012; 422:245–62. [PubMed:
22659321]
25. Shapiro MG, Westmeyer Gil G, Romero Philip A, Szablowski JO, Küster B, Shah A, Otey CR,
Langer R, Arnold FH, Jasanoff A. Directed evolution of a magnetic resonance imaging contrast
agent for noninvasive imaging of dopamine. Nat Biotechnol. 2010; 28:264–270. [PubMed:
20190737]
26. Lauterbur P. Image formation by induced local interactions: examples employing nuclear magnetic
resonance. Nature. 1973; 242:190–191.
27. Bloembergen N, Purcell EM, Pound RV. Relaxation effects in nuclear magnetic resonance
absorption. Phys Rev. 1948; 73:679–712.
28. Bloch F, Hansen WW, Packard M. The nuclear induction experiment. Phys Rev. 1946; 70:474–
485.
29. Solomon I, Bloembergen N. Nuclear magnetic interactions in the HF molecule. J Chem Phys.
1956; 25:261–166.
30. Bloembergen I. Proton relaxation times in paramagnetic solutions. J Chem Phys. 1957; 27:572–
573.
31. Bloembergen N, Morgan LO. Proton relaxation times in paramagnetic solutions. Effects of
electron spin relaxation. J Chem Phys. 1961; 34:842–850.
32. Kowalewski, J.; Kruk, D.; Parigi, G. In Advances in inorganic chemistry. Elsevier; 2005. NMR
relaxation in solution of paramagnetic complexes: recent theoretical progress for S≥1; p. 41-104.
33. Freed JH. Dynamic effects of pair correlation functions on spin relaxation by translational
diffusion in liquids. II Finite jumps and independent T1 processes. J Chem Phys. 1978; 68:4034–
4037.
34. Gillis P, Koenig SH. Transverse relaxation of solvent protons induced by magnetized spheres:
application to ferritin, erythrocytes, and magnetite. Magn Reson Med. 1987; 5:323–45. [PubMed:
2824967]

35. Roch A, Muller RN, Gillis P. Theory of proton relaxation Induced by superparamagnetic particles.
J Chem Phys. 1999; 110:5403–5411.
36. Koenig SH, Kellar KE. Theory of 1/T1 and 1/T2 NMRD profiles of solutions of magnetic
nanoparticles. Magn Reson Med. 1995; 34:227–33. [PubMed: 7476082]

37. Caravan P, Farrar CT, Frullano L, Uppal R. Influence of molecular parameters and increasing
magnetic field strength on relaxivity of gadolinium- and manganese-based T1 contrast agents.
Contrast Media Mol Imaging. 2009; 4:89–100. [PubMed: 19177472]
38. Muller, RN.; Roch, A.; Colet, J-M.; Ouakssim, A.; Gillis, P. The chemistry of contrast agents in
medical magnetic resonance imaging. J. Wiley & Sons; New York: 2001. Particulet magnetic
contrast agents.
39. Lauffer RB. Paramagnetic metal complexes as water proton relaxation agents for NMR imaging:
theory and design. Chem Rev. 1987; 87:901–927.
40. Caravan P, Ellison JJ, McMurry TJ, Lauffer RB. Gadolinium(III) chelates as MRI contrast agents:
structure, dynamics, and applications. Chem Rev. 1999; 99:2293–2352. [PubMed: 11749483]
41. Carr DH, Brown J, Bydder GM, Weinmann HJ, Speck U, Thomas DJ, Young IR. Intravenous
chelated gadolinium as a contrast agent in NMR imaging of cerebral tumours. Lancet. 1984;
1:484–6. [PubMed: 6142210]
42. Renshaw PF, Owen CS, McLaughlin AC, Frey TG, Leigh JS Jr. Ferromagnetic contrast agents: a
new approach. Magn Reson Med. 1986; 3:217–25. [PubMed: 3713487]
43. Mendonca Dias MH, Lauterbur PC. Ferromagnetic particles as contrast agents for magnetic
resonance imaging of liver and spleen. Magn Reson Med. 1986; 3:328–30. [PubMed: 3713497]
44. Olsson MBE, Persson BRB, Salford LG, Schröder U. Ferromagnetic particles as contrast agent in
T2 NMR imaging. Magn Reson Imaging. 1986; 4:437–40.
45. Jun YW, Seo JW, Cheon J. Nanoscaling laws of magnetic nanoparticles and their applicabilities in
biomedical sciences. Acc Chem Res. 2008; 41:179–89. [PubMed: 18281944]
46. Gillis P, Moiny F, Brooks RA. On T(2)-shortening by strongly magnetized spheres: a partial
refocusing model. Magn Reson Med. 2002; 47:257–63. [PubMed: 11810668]
47. Josephson L, Perez JM, RW. Magnetic Nanosensors for the Detection of Oligonucleotide
Sequences. Angew Chem, Int Ed. 2001; 40:3204–6.
48. Perez JM, Josephson L, O’Loughlin T, Hogemann D, Weissleder R. Magnetic relaxation switches
capable of sensing molecular interactions. Nat Biotechnol. 2002; 20:816–20. [PubMed: 12134166]
49. Matsumoto Y, Jasanoff A. T2 relaxation induced by clusters of superparamagnetic nanoparticles:
Monte Carlo simulations. Magn Reson Imaging. 2008; 26:994–8. [PubMed: 18479873]
50. Schwert, DD.; Davies, JA.; Richardson, N. Contrast agents I. Springer Berlin Heidelberg; Berlin,
Heidelberg: 2002. Non-gadolinium-based MRI contrast agents; p. 165-199.
51. Bonnet, ClS; Fries, PH.; Crouzy, S.; Delangle, P. Outer-sphere investigation of MRI relaxation
contrast agents. Example of a cyclodecapeptide gadolinium complex with second-sphere water. J
Phys Chem B. 2010; 114:8770–81. [PubMed: 20552972]
52. Toth, E.; Helm, L.; Merbach, AE. Contrast agents I. Springer Berlin Heidelberg; Berlin,
Heidelberg: 2002. Relaxivity of MRI contrast agents; p. 61-101.

53. Einstein A. Über die von der molekularkinetischen theorie der wärme geforderte bewegung von in
ruhenden flüssigkeiten suspendierten teilchen. Ann Phys. 1905; 17:549–60.
54. Diakova G, Goddard Y, Korb JP, Bryant RG. Water-proton-spin-lattice-relaxation dispersion of
paramagnetic protein solutions. J Magn Reson. 2011; 208:195–203. [PubMed: 21134772]
55. Rebodos RL, Vikesland PJ. Effects of oxidation on the magnetization of nanoparticulate magnetite.
Langmuir. 2010; 26:16745–53. [PubMed: 20879747]
56. Mohie-Eldin MEY, Frankel RB, Gunther L. A comparison of the magnetic properties of
polysaccharide iron complex (PIC) and ferritin. J Magn Magn Mater. 1994; 135:65–81.
57. Gossuin Y, Roch A, Muller RN, Gillis P. Relaxation induced by ferritin and ferritin-like magnetic
particles: the role of proton exchange. Magn Reson Med. 2000; 43:237–43. [PubMed: 10680687]
58. Vymazal J, Zak O, Bulte JW, Aisen P, Brooks RA. T1 and T2 of ferritin solutions: effect of
loading factor. Magn Reson Med. 1996; 36:61–5. [PubMed: 8795021]

59. Pauling L, Coryell CD. The magnetic properties and structure of hemoglobin, oxyhemoglobin and
carbonmonoxyhemoglobin. Proc Natl Acad Sci U S A. 1936; 22:210–6. [PubMed: 16577697]
60. Ogawa S, Lee TM, Kay AR, Tank DW. Brain magnetic resonance imaging with contrast
dependent on blood oxygenation. Proc Natl Acad Sci U S A. 1990; 87:9868–72. [PubMed:
2124706]
61. Boxerman JL, Bandettini PA, Kwong KK, Baker JR, Davis TL, Rosen BR, Weisskoff RM. The
intravascular contribution to fMRI signal change: Monte Carlo modeling and diffusion-weighted
studies in vivo. Magn Reson Med. 1995; 34:4–10. [PubMed: 7674897]
62. Kapff, CT.; Jandl, JH. Blood, atlas and sourcebook of hematology. Little, Brown; Boston: 1991.
63. Bandettini PA, Wong EC, Hinks RS, Tikofsky RS, Hyde JS. Time course EPI of human brain
function during task activation. Magn Reson Med. 1992; 25:390–7. [PubMed: 1614324]
64. Ogawa S, Tank DW, Menon R, Ellermann JM, Kim SG, Merkle H, Ugurbil K. Intrinsic signal
changes accompanying sensory stimulation: functional brain mapping with magnetic resonance
imaging. Proc Natl Acad Sci U S A. 1992; 89:5951–5. [PubMed: 1631079]
65. Kwong KK, et al. Dynamic magnetic resonance imaging of human brain activity during primary
sensory stimulation. Proc Natl Acad Sci U S A. 1992; 89:5675–9. [PubMed: 1608978]
66. Prasad PV, Priatna A. Functional imaging of the kidneys with fast MRI techniques. Eur J Radiol.
1999; 29:133–48. [PubMed: 10374661]
67. Carlier PG. Skeletal muscle perfusion and oxygenation assessed by dynamic NMR imaging and
spectroscopy. Adv Exp Med Biol. 2011; 701:341–6. [PubMed: 21445807]
68. Tran TK, Sailasuta N, Hurd R, Jue T. Spatial distribution of deoxymyoglobin in human muscle: an
index of local tissue oxygenation. NMR Biomed. 1999; 12:26–30. [PubMed: 10195326]
69. Shonat RD, Koretsky AP. Expression of myoglobin in the transgenic mouse brain. Adv Exp Med
Biol. 2003; 530:331–45. [PubMed: 14562729]
70. Fischbach FA, Harrison PM, Hoy TG. The structural relationship between ferritin protein and its
mineral core. J Mol Biol. 1969; 39:235–8. [PubMed: 5408658]
71. Harrison PM, Fischbach FA, Hoy TG, Haggis GH. Ferric oxyhydroxide core of ferritin. Nature.
1967; 216:1188–90. [PubMed: 6076063]
72. Cohen B, Dafni H, Meir G, Harmelin A, Neeman M. Ferritin as an endogenous MRI reporter for
noninvasive imaging of gene expression in C6 glioma tumors. Neoplasia (N Y). 2005; 7:109–17.
73. Genove G, DeMarco U, Xu H, Goins WF, Ahrens ET. A new transgene reporter for in vivo
magnetic resonance imaging. Nat Med. 2005; 11:450–454. [PubMed: 15778721]
74. Cohen B, Ziv K, Plaks V, Israely T, Kalchenko V, Harmelin A, Benjamin LE, Neeman M. MRI
detection of transcriptional regulation of gene expression in transgenic mice. Nat Med. 2007;
13:498–503. [PubMed: 17351627]
75. Kaplan J, Jordan I, Sturrock A. Regulation of the transferrin-independent iron transport system in
cultured cells. J Biol Chem. 1991; 266:2997–3004. [PubMed: 1993673]
76. Karin M, Mintz B. Receptor-mediated endocytosis of transferrin in developmentally totipotent
mouse teratocarcinoma stem cells. J Biol Chem. 1981; 256:3245–52. [PubMed: 6259157]
77. Rothman RJ, Serroni A, Farber JL. Cellular pool of transient ferric iron, chelatable by
deferoxamine and distinct from ferritin, that is involved in oxidative cell injury. Mol Pharmacol.
1992; 42:703–10. [PubMed: 1435746]
78. Deans AE, Wadghiri YZ, Bernas LM, Yu X, Rutt BK, Turnbull DH. Cellular MRI contrast via
coexpression of transferrin receptor and ferritin. Magn Reson Med. 2006; 56:51–9. [PubMed:
16724301]
79. Zurkiya O, Chan AWS, Hu X. MagA is sufficient for producing magnetic nanoparticles in
mammalian cells, making it an MRI reporter. Magn Reson Med. 2008; 59:1225–31. [PubMed:
18506784]
80. Bartelle BB, Szulc KU, Suero-Abreu GA, Rodriguez JJ, Turnbull DH. Divalent metal transporter,
DMT1: A novel MRI reporter protein. Magn Reson Med. 2012 [Epub ahead of print].
81. Lauffer RB, Brady TJ. Preparation and water relaxation properties of proteins labeled with
paramagnetic metal chelates. Magn Reson Imaging. 1985; 3:11–6. [PubMed: 3923289]

82. Unger EC, Totty WG, Neufeld DM, Otsuka FL, Murphy WA, Welch MS, Connett JM, Philpott
GW. Magnetic resonance imaging using gadolinium labeled monoclonal antibody. Investig
Radiol. 1985; 20:693–700. [PubMed: 4066240]
83. Shreve P, Aisen AM. Monoclonal antibodies labeled with polymeric paramagnetic ion chelates.
Magn Reson Med. 1986; 3:336–40. [PubMed: 3713499]
84. Wu CWH, et al. Development of a MR-visible compound for tracing neuroanatomical connections
in vivo. Neuron. 2011; 70:229–43. [PubMed: 21521610]
85. Weissleder R, Moore A, Mahmood U, Bhorade R, Benveniste H, Chiocca EA, Basilion JP. In vivo
magnetic resonance imaging of transgene expression. Nat Med. 2000; 6:351–5. [PubMed:
10700241]
86. Zhao M, Beauregard DA, Loizou L, Davletov B, Brindle KM. Non-invasive detection of apoptosis
using magnetic resonance imaging and a targeted contrast agent. Nat Med. 2001; 7:1241–4.
[PubMed: 11689890]
87. Ruehm SG, Corot C, Vogt P, Kolb S, Debatin JF. Magnetic resonance imaging of atherosclerotic
plaque with ultrasmall superparamagnetic particles of iron oxide in hyperlipidemic rabbits.
Circulation. 2001; 103:415–22. [PubMed: 11157694]
88. Cerdan S, Lotscher HR, Kunnecke B, Seelig J. Monoclonal antibody-coated magnetite particles as
contrast agents in magnetic resonance imaging of tumors. Magn Reson Med. 1989; 12:151–63.
[PubMed: 2615625]
89. Tiefenauer LX, Kuhne G, Andres RY. Antibody-magnetite nanoparticles: in vitro characterization
of a potential tumor-specific contrast agent for magnetic resonance imaging. Bioconjug Chem.
1993; 4:347–52. [PubMed: 8274518]

90. Sun EY, Weissleder R, Josephson L. Continuous analyte sensing with magnetic nanoswitches.
Small. 2006; 2:1144–7. [PubMed: 17193579]
91. Atanasijevic T, Shusteff M, Fam P, Jasanoff A. Calcium-sensitive MRI contrast agents based on
superparamagnetic iron oxide nanoparticles and calmodulin. Proc Natl Acad Sci USA. 2006;
103:14707–12. [PubMed: 17003117]
92. Chang, TMS. Blood substitutes: principles, methods, products, and clinical trials. Karger Landes
Systems; Basel; New York: 1997.
93. Sun PZ, Schoening ZB, Jasanoff A. In vivo oxygen detection using exogenous hemoglobin as a
contrast agent in magnetic resonance microscopy. Magn Reson Med. 2003; 49:609–14. [PubMed:
12652529]
94. Bloom JD, Arnold FH. In the light of directed evolution: pathways of adaptive protein evolution.
Proc Natl Acad Sci USA. 2009; 106(Suppl 1):9995–10000. [PubMed: 19528653]
95. Vymazal J, Hajek M, Patronas N, Giedd JN, Bulte JW, Baumgarner C, Tran V, Brooks RA. The
quantitative relation between T1-weighted and T2-weighted MRI of normal gray matter and iron
concentration. J MRI. 1995; 5:554–60.
96. Mani V, Briley-Saebo KC, Hyafil F, Fayad ZA. Feasibility of in vivo identification of endogenous
ferritin with positive contrast MRI in rabbit carotid crush injury using GRASP. Magn Reson Med.
2006; 56:1096–106. [PubMed: 17036302]
97. Yao B, Li TQ, Gelderen P, Shmueli K, de Zwart JA, Duyn JH. Susceptibility contrast in high field
MRI of human brain as a function of tissue iron content. Neuroimage. 2009; 44:1259–66.
[PubMed: 19027861]
98. Uchida M, Flenniken Michelle L, Allen Mark, Willits Deborah A, Crowley Bridgid E, Brumfield
S, Willis AF, Jackiw L, Jutila M, Young MJ, Douglas T. Targeting of cancer cells with
ferrimagnetic ferritin cage nanoparticles. J Am Chem Soc. 2006; 128:16626–33. [PubMed:
17177411]
99. Shapiro MG, Szablowski JO, Langer R, Jasanoff A. Protein nanoparticles engineered to sense
kinase activity in MRI. J Am Chem Soc. 2009; 131:2484–6. [PubMed: 19199639]
100. Shaywitz AJ, Dove SL, Kornhauser JM, Hochschild A, Greenberg ME. Magnitude of the CREB-
dependent transcriptional response is determined by the strength of the interaction between the
kinase-inducible domain of CREB and the KIX domain of CREB-binding protein. Mol Cell Biol.
2000; 20:9409–22. [PubMed: 11094091]

101. Bennett KM, Shapiro EM, Sotak CH, Koretsky AP. Controlled aggregation of ferritin to modulate
MRI relaxivity. Biophys J. 2008; 95:342–51. [PubMed: 18326661]
102. Iordanova B, Robison CS, Ahrens ET. Design and characterization of a chimeric ferritin with
enhanced iron loading and transverse NMR relaxation rate. J Biol Inorg Chem. 2010; 15:957–65.
[PubMed: 20401622]
103. Lelyveld VS, Brustad E, Arnold FH, Jasanoff A. Metal-substituted protein MRI contrast agents
engineered for enhanced relaxivity and ligand sensitivity. J Am Chem Soc. 2011; 133:649–51.
[PubMed: 21171606]
104. Meldrum FC, Heywood BR, Mann S. Magnetoferritin: in vitro synthesis of a novel magnetic
protein. Science. 1992; 257:522–3. [PubMed: 1636086]
105. Bulte JW, et al. Magnetoferritin: characterization of a novel superparamagnetic MR contrast
agent. J MRI. 1994; 4:497–505.
106. Jordan VC, Caplan MR, Bennett KM. Simplified synthesis and relaxometry of magnetoferritin for
magnetic resonance imaging. Magn Reson Med. 2010; 64:1260–6. [PubMed: 20677230]
107. Sánchez P, Valero E, Gálvez N, Domínguez-Vera JM, Marinone M, Poletti G, Corti M,
Lascialfari A. MRI relaxation properties of water-soluble apoferritin-encapsulated gadolinium
oxide-hydroxide nanoparticles. Dalton Trans. 2009:800–4. [PubMed: 19156273]
108. Kálmán FK, Geninatti-Crich S, Aime S. Reduction/dissolution of a beta-MnOOH nanophase in
the ferritin cavity to yield a highly sensitive, biologically compatible magnetic resonance
imaging agent. Angew Chem, Int Ed Engl. 2010; 49:612–5. [PubMed: 20013829]
109. Allen M, Bulte JWM, Liepold L, Basu G, Zywicke HA, Frank JA, Young M, Douglas T.
Paramagnetic viral nanoparticles as potential high-relaxivity magnetic resonance contrast agents.

Magn Reson Med. 2005; 54:807–12. [PubMed: 16155869]
110. Liepold L, Anderson S, Willits D, Oltrogge L, Frank JA, Douglas T, Young M. Viral capsids as
MRI contrast agents. Magn Reson Med. 2007; 58:871–9. [PubMed: 17969126]
111. Qazi S, Liepold LO, Abedin MJ, Johnson B, Prevelige P, Frank JA, Douglas T. P22 viral capsids
as nanocomposite high-relaxivity MRI contrast agents. Mol Pharm. 2012 [Epub ahead of print].
112. Aime S, Frullano L, Geninatti Crich S. Compartmentalization of a gadolinium complex in the
apoferritin cavity: a route to obtain high relaxivity contrast agents for magnetic resonance
imaging. Angew Chem, Int Ed Engl. 2002; 41:1017–9. [PubMed: 12491298]
113. Sana B, Poh CL, Lim S. A manganese-ferritin nanocomposite as an ultrasensitive T2 contrast
agent. Chem Commun (Cambridge, U K). 2012; 48:862–4.
114. Karfeld LS, Bull SR, Davis NE, Meade TJ, Barron AE. Use of a genetically engineered protein
for the design of a multivalent MRI contrast agent. Bioconjugate Chem. 2007; 18:1697–700.
115. Karfeld-Sulzer LS, Waters EA, Davis NE, Meade TJ, Barron AE. Multivalent protein polymer
MRI contrast agents: controlling relaxivity via modulation of amino acid sequence.
Biomacromolecules. 2010; 11:1429–36. [PubMed: 20420441]
116. Yang JJ, et al. Rational design of protein-based MRI contrast agents. J Am Chem Soc. 2008;
130:9260–7. [PubMed: 18576649]
117. Grabarek Z. Structural basis for diversity of the EF-hand calcium-binding proteins. J Mol Biol.
2006; 359:509–25. [PubMed: 16678204]
118. Gifford JL, Walsh MP, Vogel HJ. Structures and metal-ion-binding properties of the Ca2+-
binding helix-loop-helix EF-hand motifs. Biochem J. 2007; 405:199–221. [PubMed: 17590154]
119. Caravan P, Greenwood JM, Welch JT, Franklin SJ. Gadolinium-binding helix-turn-helix peptides:
DNA-dependent MRI contrast agents. Chem Commun (Cambridge, U K). 2003:2574–5.
120. Atanasijevic T, Zhang XA, Lippard SJ, Jasanoff A. MRI sensing based on the displacement of
paramagnetic ions from chelated complexes. Inorg Chem. 2010; 49:2589–91. [PubMed:
20141114]
121. Franz KJ, Nitz M, Imperiali B. Lanthanide-binding tags as versatile protein coexpression probes.
ChemBioChem. 2003; 4:265–71. [PubMed: 12672105]
122. Nitz M, Franz KJ, Maglathlin RL, Imperiali B. A powerful combinatorial screen to identify high-
affinity terbium(III)-binding peptides. ChemBioChem. 2003; 4:272–6. [PubMed: 12672106]

123. Daughtry KD, Martin LJ, Sarraju A, Imperiali B, Allen KN. Tailoring encodable lanthanide-
binding tags as MRI contrast agents. ChemBioChem. 2012; 13:2567–74. [PubMed: 23150430]
124. Gilad AA, et al. Artificial reporter gene providing MRI contrast based on proton exchange. Nat
Biotechnol. 2007; 25:217–9. [PubMed: 17259977]

125. Hempstead PD, Yewdall SJ, Fernie AR, Lawson DM, Artymiuk PJ, Rice DW, Ford GC, Harrison
PM. Comparison of the three-dimensional structures of recombinant human H and horse L
ferritins at high resolution. J Mol Biol. 1997; 268:424–48. [PubMed: 9159481]
126. Li H, Poulos TL. The structure of the cytochrome p450BM-3 haem domain complexed with the
fatty acid substrate, palmitoleic acid. Nat Struct Biol. 1997; 4:140–6. [PubMed: 9033595]
127. Spirlet MR, Rebizant J, Desreux JF, Loncin MF. Crystal and molecular structure of sodium
aqua(1,4,7,10-tetraazacyclododecane-1,4,7,10-tetraacetato)europate(III) tetrahydrate Na+
(EuDOTA.H2O)−.4H2O, and its relevance to NMR studies of the conformational behavior of the
lanthanide complexes formed by the macrocyclic ligand DOTA. Inorg Chem. 1984; 23:359–63.

Figure 1. Mechanisms of MRI contrast enhancement by paramagnetic metalloproteins

(A) Protein-based T1 contrast agents operate largely through an inner sphere relaxation
mechanism dependent on metal-coordinated water molecules in fast exchange with bulk
solvent. In the example shown (top), the heme iron (brown) of an engineered P450 BM3
variant (gray ribbon structure) interacts with an axial inner sphere water ligand (blue ball,
arrowhead) [24]. Bound water protons (not shown) experience dipole coupling with the
metal ion and undergo relaxation. T1 relaxivity of the complex may also be promoted by the
presence of second sphere water molecules (additional blue balls) that participate in weak
dipole coupling with the iron atom and exchange with water in the inner sphere position.
The T1 contrast affects the saturation of MRI signals during repeated application of a pulse
sequence (schematics at bottom; pulses in gray and raw MRI signal in black). In the absence
of the contrast agent, the MRI signal tends to decline due to saturation over time (top trace).
The T1 agent relieves this effect (bottom trace), resulting in a enhanced signal and relative
hyperintensity in images. (B) The best-characterized protein T2 contrast agent is ferritin (Ft,
ribbon structure), a shell-shaped oligomer of 24 subunits (one shown in purple) that contains
a paramagnetic hydrated iron oxide core. The core induces a dipolar field perturbation (field
lines shown) over a length scale comparable to the core diameter, and water molecules (blue
balls) undergo transverse relaxation by diffusing through the dipole field. The amplitude of
an MRI signal obtained using a T2-weighted pulse sequence (schematics at bottom) depends
on the amount of T2 relaxation that has occurred prior to acquisition of the signal with each
repetition of the pulse sequence. A T2 contrast agent like Ft promotes T2 relaxation and
leads to attenuation of the signal (bottom trace) and relative hypointensity in MRI scans. Ft
structural model from reference [125].

Figure 2. Engineered metalloprotein-based MRI sensors

(A) Metalloprotein-based T1 contrast agents can sense analytes via a so-called “q-
modulation” mechanism. In this mechanism, exemplified by BM3h dopamine sensors, inner
sphere water molecules bound to the paramagnetic center in the ligand free structure (left)
are displaced by analyte binding (right). For the BM3h-based sensors, neurotransmitter
binding reduces q of the heme iron atom from one to zero. (B) The ligand dependent change
in q induces a sharp drop in the r1 of BM3h variants, from roughly 1 to 0.2 mM−1s−1 for the
best two dopamine-binding variants, BM3h-B7 and -8C8, identified in reference [25]. The
relaxivity decrease upon dopamine binding also leads to a reduction in the corresponding
T1-weighted MRI signal intensities (inset). (C) Metalloprotein-based T2 sensors can be
constructed by modifying protein domains in Ft to include analyte-sensitive moieties. In the
example of reference [99], a kinase-sensitive Ft-based probe was constructed by genetically
fusing the KID domain of the protein CREB and the KIX domain of the protein CBP to Ft
light chain to make chimeric KID-Ft (blue) and KIX-Ft (magenta) variants. In the presence
of protien kinase A (PKA), KID domains are phosphorylated and tend to bind KIX domains,
leading to clustering of the multivalent KID-Ft and KIX-Ft proteins. (D) Kinase-dependent
Ft clustering leads to a change in per-particle r2 values. Relaxivities measured from
prephosphorylated KID-Ft (pKID) mixed with KIX-Ft or from KID-Ft mixed with KIX-Ft
in the presence of PKA (middle two bars) are approximately twice the r2 values measured
from KID-Ft/KIX-Ft in the absence of phosphorylation (left bar), or in the presence of the
ATP phosphate source but not the kinase (right bar). Corresponding MRI image intensities
are shown in the inset.

Figure 3. Nuclear magnetic relaxation dispersion of BM3h variants

(A) Mutant low spin (S = 1/2) BM3h proteins show differing T1 relaxivities as a function of
magnetic field strength [24]. NMRD curves were fit to a Solomon-Bloembergen model
equation with an iron proton distance r, water exchange time constant τM, and field-
independent electronic relaxation time τS. Best fit values for five protein variants are shown
in the inset, along with corresponding measured relaxivity values (circles) and fitted curves
(solid lines), color coded by mutant. Substantial relaxivity variation among mutants is
apparent, showing that relaxivity changes are accessible by mutagenesis of the

metalloprotein, even without changing the nature of the bound metal complex. (B) Location
of amino acid substitutions in the BM3h variants of panel A. Cα positions of mutated
residues are denoted by blue balls in the protein backbone trace (gray) [126], with the heme
shown in pink and the native fatty acid ligand shown in black. Color-coded dots denote
which residues are mutated with respect to the wild type protein in each of the variants listed
in panel A. Mutations were selected to alter ligand binding near the heme site and most are
clustered in the ligand binding region of the protein.

Figure 4. Structures of lanthanide binding sites

(A) Structure of europium-DOTA, a compound isomorphous to the canonical contrast agent
Gd-DOTA [127]. The structure shows coordination of the lanthanide (magenta ball) by
carboxylate oxygens and amine nitrogens on the chelator, with a single water molecule
(cyan) bound at a Eu–O distance of 2.5 Å. (B) The structure of a polypeptide lanthanide
binding tag fused to ubiquitin, in complex with gadolinium (magenta), is surprising similar
to the Eu-DOTA structure. The lanthanide is again coordinated by a mixture of oxygen and
nitrogen ligands and exhibits a q of 1, with a single inner sphere water molecule bound at a
Gd–O distance of 2.9 Å and a second sphere bound water, which may also contribute to r1,
at a distance of 5.4 Å from the lanthanide.

Metalloprotein-Based MRI Probes

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Metalloprotein-Based MRI Probes

Uploaded by

Copyright:

Available Formats

Metalloprotein-based MRI probes

Citation Matsumoto, Yuri and Alan Jasanoff. "Metalloprotein-based MRI

Version Author's final manuscript

Citable link https://hdl.handle.net/1721.1/126108

Terms of Use Creative Commons Attribution-NonCommercial-NoDerivs License

Detailed Terms http://creativecommons.org/licenses/by-nc-nd/4.0/

FEBS Lett. 2013 April 17; 587(8): 1021–1029. doi:10.1016/j.febslet.2013.01.044.

Metalloprotein-based MRI probes

Ave., Rm. 16-561, Cambridge, MA 02139

Biophysical properties of metalloproteins and protein-metal interactions have led to a

provide important biotechnological capabilities. Many metalloproteins contain ions with

FEBS Lett. Author manuscript; available in PMC 2014 April 17.

Theoretical basis of MRI contrast agents

FEBS Lett. Author manuscript; available in PMC 2014 April 17.

Relaxivity of protein contrast agents

FEBS Lett. Author manuscript; available in PMC 2014 April 17.

Natural metalloproteins in MRI

FEBS Lett. Author manuscript; available in PMC 2014 April 17.

Engineering protein-based contrast agents to detect biological targets

another strength of proteins—their amenability to engineering approaches—comes into play.

FEBS Lett. Author manuscript; available in PMC 2014 April 17.

FEBS Lett. Author manuscript; available in PMC 2014 April 17.

Engineering metalloproteins for high relaxivity

the Mn3+-containing protein variants.

FEBS Lett. Author manuscript; available in PMC 2014 April 17.

Ft has also been used as a compartment for entrapment of soluble Gd-DOTA-related T1

FEBS Lett. Author manuscript; available in PMC 2014 April 17.

BBB Blood-Brain Barrier

fMRI Functional Magnetic Resonance Imaging

FEBS Lett. Author manuscript; available in PMC 2014 April 17.

R1 Longitudinal Relaxation Rate

FEBS Lett. Author manuscript; available in PMC 2014 April 17.

native neuroblast migration. Neuroimage. 2012; 59:1004–12. [PubMed: 21939774]

FEBS Lett. Author manuscript; available in PMC 2014 April 17.

nanoparticles. Magn Reson Med. 1995; 34:227–33. [PubMed: 7476082]

Heidelberg: 2002. Relaxivity of MRI contrast agents; p. 61-101.

FEBS Lett. Author manuscript; available in PMC 2014 April 17.

FEBS Lett. Author manuscript; available in PMC 2014 April 17.

1993; 4:347–52. [PubMed: 8274518]

FEBS Lett. Author manuscript; available in PMC 2014 April 17.

Paramagnetic viral nanoparticles as potential high-relaxivity magnetic resonance contrast agents.

FEBS Lett. Author manuscript; available in PMC 2014 April 17.

Biotechnol. 2007; 25:217–9. [PubMed: 17259977]

FEBS Lett. Author manuscript; available in PMC 2014 April 17.

Figure 1. Mechanisms of MRI contrast enhancement by paramagnetic metalloproteins

FEBS Lett. Author manuscript; available in PMC 2014 April 17.

Figure 2. Engineered metalloprotein-based MRI sensors

FEBS Lett. Author manuscript; available in PMC 2014 April 17.

Figure 3. Nuclear magnetic relaxation dispersion of BM3h variants

apparent, showing that relaxivity changes are accessible by mutagenesis of the

FEBS Lett. Author manuscript; available in PMC 2014 April 17.

Figure 4. Structures of lanthanide binding sites

FEBS Lett. Author manuscript; available in PMC 2014 April 17.

You might also like