The interaction of MS-325 with human serum albumin and its effect on proton relaxation rates

Article abstract of DOI:10.1021/ja017168k

MS-325 is a novel blood pool contrast agent for magnetic resonance imaging currently undergoing clinical trials to assess blockage in arteries. MS-325 functions by binding to human serum albumin (HSA) in plasma. Binding to HSA serves to prolong plasma half-life, retain the agent in the blood pool, and increase the relaxation rate of water protons in plasma. Ultrafiltration studies with a 5 kDa molecular weight cutoff filter show that MS-325 binds to HSA with stepwise stoichiometric affinity constants (mM^-1) of K_a1 = 11.0 ± 2.7, K_a2 = 0.84 ± 0.16, K_a3 = 0.26 ± 0.14, and K_a4 = 0.43 ± 0.24. Under the conditions 0.1 mM MS-325, 4.5% HSA, pH 7.4 (phosphate-buffered saline), and 37°C, 88 ± 2% of MS-325 is bound to albumin. Fluorescent probe displacement studies show that MS-325 can displace dansyl sarcosine and dansyl-L-asparagine from HSA with inhibition constants (K_i) of 85 ± 3 μM and 1500 ± 850 μM, respectively; however, MS-325 is unable to displace warfarin. These results suggest that MS-325 binds primarily to site II on HSA. The relaxivity of MS-325 when bound to HSA is shown to be site dependent. The Eu(III) analogue of MS-325 is shown to contain one inner-sphere water molecule in the presence and in the absence of HSA. The synthesis of an MS-325 analogue, 5, containing no inner-sphere water molecules is described. Compound 5 is used to estimate the contribution to relaxivity from the outer-sphere water molecules surrounding MS-325. The high relaxivity of MS-325 bound to HSA is primarily because of a 60-100-fold increase in the rotational correlation time of the molecule upon binding (τ_R = 10.1 ± 2.6 ns bound vs 115 ps free). Analysis of the nuclear magnetic relaxation dispersion (T₁ and T₂) profiles also suggests a decrease in the electronic relaxation rate (1/T_1e at 20 MHz = 2.0 × 10⁸ s^-1 bound vs 1.1 × 10⁹ s^-1 free) and an increase in the inner-sphere water residency time (τ_m = 170 ± 40 ns bound vs 69 ± 20 ns free).

Full text of DOI:10.1021/ja017168k

Published on Web 02/27/2002
The Interaction of MS-325 with Human Serum Albumin and Its
Effect on Proton Relaxation Rates
Peter Caravan,*,† Normand J. Cloutier,† Matthew T. Greenfield,†
Sarah A. McDermid,† Stephen U. Dunham,† Jeff W. M. Bulte,‡ John C. Amedio, Jr.,†
Richard J. Looby,† Ronald M. Supkowski,§ William DeW. Horrocks, Jr.,§
Thomas J. McMurry,† and Randall B. Lauffer†
Contribution from EPIX Medical, Inc., 71 Rogers Street, Cambridge, Massachusetts 02142-1118,
Laboratory of Diagnostic Radiology Research (CC), National Institutes of Health, BLDG 10,
Room B1N256, 10 Center DriVe, MSC 1074, Bethesda, Maryland 20892, and Department of
Chemistry, The PennsylVania State UniVersity, UniVersity Park, PennsylVania 16802
Received September 26, 2001
Abstract: MS-325 is a novel blood pool contrast agent for magnetic resonance imaging currently undergoing
clinical trials to assess blockage in arteries. MS-325 functions by binding to human serum albumin (HSA)
in plasma. Binding to HSA serves to prolong plasma half-life, retain the agent in the blood pool, and increase
the relaxation rate of water protons in plasma. Ultrafiltration studies with a 5 kDa molecular weight cutoff
filter show that MS-325 binds to HSA with stepwise stoichiometric affinity constants (mM-1) of Ka1 ) 11.0
( 2.7, Ka2 ) 0.84 ( 0.16, Ka3 ) 0.26 ( 0.14, and Ka4 ) 0.43 ( 0.24. Under the conditions 0.1 mM MS-
325, 4.5% HSA, pH 7.4 (phosphate-buffered saline), and 37 °C, 88 ( 2% of MS-325 is bound to albumin.
Fluorescent probe displacement studies show that MS-325 can displace dansyl sarcosine and dansyl-L-
asparagine from HSA with inhibition constants (Ki) of 85 ( 3 µM and 1500 ( 850 µM, respectively; however,
MS-325 is unable to displace warfarin. These results suggest that MS-325 binds primarily to site II on
HSA. The relaxivity of MS-325 when bound to HSA is shown to be site dependent. The Eu(III) analogue
of MS-325 is shown to contain one inner-sphere water molecule in the presence and in the absence of
HSA. The synthesis of an MS-325 analogue, 5, containing no inner-sphere water molecules is described.
Compound 5 is used to estimate the contribution to relaxivity from the outer-sphere water molecules
surrounding MS-325. The high relaxivity of MS-325 bound to HSA is primarily because of a 60-100-fold
increase in the rotational correlation time of the molecule upon binding (τR ) 10.1 ( 2.6 ns bound vs 115
ps free). Analysis of the nuclear magnetic relaxation dispersion (T1 and T2) profiles also suggests a decrease
in the electronic relaxation rate (1/T1e at 20 MHz ) 2.0 × 108 s-1 bound vs 1.1 × 109 s-1 free) and an
increase in the inner-sphere water residency time (τm ) 170 ( 40 ns bound vs 69 ( 20 ns free).
Introduction
coupling of this biophysical trick with retention in a particular
tissue, by virtue of the binding, provided a unique way to
Magnetic resonance imaging (MRI) contrast agents are unique
diagnostic drugs which alter water proton relaxation times of
tissue and increase contrast between normal and abnormal
structures on MRI scans. While early contrast agents were
simple chelates of gadolinium, more recent attempts have
involved targeting the agents to particular binding sites in vivo
and exploiting the unique biophysical properties inherent in the
electron-nuclear interaction.1 The most important of these
features is the dependence of water relaxation on the rotational
correlation time of a metal complex. Theory and early results
with metal ions bound to proteins suggest that an increase in
relaxivity could be obtained by protein-chelate binding.1,2 The
increase contrast. MS-325 is the first representative member of
this class of agents targeted to the vasculature.3,4
Gadolinium(III) complexes function as contrast agents by
shortening the T1 of water protons wherever the complex
localizes. In a T1-weighted MR image, water molecules with
the shortest T1 will give the largest signal. Consequently, areas
of the body containing contrast agent will appear as bright
regions in an MR image.
All Gd(III) complexes used as contrast agents contain a
common structural feature: at least one coordinated water
molecule. The Gd(III) complex acts as a catalyst to relax bulk
water by providing efficient relaxation of the coordinated water
molecule. As this water molecule is in fast exchange with the
* To whom correspondence should be addressed. E-mail: pcaravan@
epixmed.com. Fax: +1 617 250 6127.
† EPIX Medical, Inc.
(2) Lauffer, R. B. Chem. ReV. 1987, 87, 901-927.
(3) Parmelee, D. J.; Walovitch, R. C.; Ouellet, H. S.; Lauffer, R. B. InVest.
Radiol. 1997, 32, 741-747.
(4) Lauffer, R. B.; Parmelee, D. J.; Dunham, S. U.; Ouellet, H. S.; Dolan, R.
P.; Witte, S.; McMurry, T. J.; Walovitch, R. C. Radiology 1998, 207, 529-
538.
‡ National Institutes of Health.
§ The Pennsylvania State University.
(1) Caravan, P.; Ellison, J. J.; McMurry, T. J.; Lauffer, R. B. Chem. ReV. 1999,
99, 2293-2352.
9
3152 VOL. 124, NO. 12, 2002 J. AM. CHEM. SOC.
10.1021/ja017168k CCC: $22.00 © 2002 American Chemical Society

Interaction of MS-325 with Human Serum Albumin
A R T I C L E S
Chart 1
bulk, the net effect is an increase in the relaxation rate of bulk
water molecules, which is dependent upon the gadolinium(III)
concentration, the rate of water exchange, and the rate of
relaxation of the protons on the coordinated water molecule.
This is often referred to as inner-sphere relaxation because it
arises from a water molecule in the first coordination sphere.
There is also a relaxation enhancement provided by the
paramagnetic molecule to water molecules which are diffusing
close to the Gd(III) ion, and to those which occupy the “second
coordination sphere”; these relaxation mechanisms are different
but are often grouped together as outer-sphere relaxation. For
the commercially available extracellular contrast agents, the
observed relaxation enhancement arises from similar outer- and
inner-sphere contributions. The degree of relaxation enhance-
ment provided by the paramagnetic ion, normalized to 1 mM,
is commonly referred to as relaxivity, r1 and r2, where r1 refers
to longitudinal and r2 to transverse relaxivity, respectively (eq
1).
high-relaxivity agents have involved the incorporation of Gd-
(III) chelates into high-molecular-weight structuresseither
natural7-13 or synthetic14-20 polymers or dendrimers.21-25 In
these instances, the relaxivity increase is realized by increasing
the number of Gd(III) ions per molecule and by slowing down
rotation as a result of increased molecular weight. Increases in
relaxation per gadolinium are generally not as great as would
be expected on the basis of molecular weight because of fast
segmental motions within the polymers.1
An alternate to the macromolecular approach is the so-called
receptor-induced magnetization enhancement (RIME) strat-
egy.26,27 The RIME method involves targeting a small Gd(III)
complex to a particular protein. Binding to a macromolecule
causes an increased concentration of the Gd(III) complex in the
area of the receptor molecule, giving selective enhancement of
(5) Lauffer, R. B.; Brady, T. J. Magn. Reson. Imaging 1985, 3, 11-16.
(6) Lauffer, R. B.; Brady, T. J.; Brown, R. D., III; Baglin, C.; Koenig, S. H.
Magn. Reson. Med. 1986, 3, 541-548.
(7) Armitage, F. E.; Richardson, D. E.; Li, K. C. P. Bioconjugate Chem. 1990,
1, 365-374.
(1/Ti)obs ) (1/Ti)diamagnetic + ri[Gd]
i ) 1, 2 (1)
(8) Casali, C.; Janier, M.; Canet, E.; Obadia, J. F.; Benderbous, S.; Corot, C.;
Revel, D. Acad. Radiol. 1998, 5, S214-S218.
(9) Meyer, D.; Schaefer, M.; Bouillot, A.; Beaute, S.; Chambon, C. InVest.
Radiol. 1991, 26, S50-S52.
To effect relaxation of a coordinated water proton, a local
fluctuating magnetic field is required. The gadolinium(III) ion
with its seven unpaired electrons is well suited to this task. The
rate at which this magnetic field fluctuates depends primarily
on the rate at which the complex rotates and the rate of
relaxation of the unpaired f electrons. The faster of the two rates
will dominate the relaxation enhancement process. Gd(III) has
(10) Meyer, D.; Schaefer, M.; Chambon, C.; Beaute, S. InVest. Radiol. 1994,
29, S90-S92.
(11) Rebizak, R.; Schaefer, M.; Dellacherie, E. Bioconjugate Chem. 1997, 8,
605-610.
(12) Rebizak, R.; Schaefer, M.; Dellacherie, E. Bioconjugate Chem. 1998, 9,
94-99.
(13) Siauve, N.; Clement, O.; Cuenod, C.-A.; Benderbous, S.; Frija, G. Magn.
Reson. Imaging 1996, 14, 381-390.
(14) Schuhmann-Giampieri, G.; Schmitt-Willich, H.; Frenzel, T.; Press, W. R.;
Weinmann, H.-J. InVest. Radiol. 1991, 26, 969-974.
(15) Spanoghe, M.; Lanens, D.; Dommisse, R.; Van der Linden, A.; Alder
Weireldt, F. Magn. Reson. Imaging 1992, 10, 913-917.
(16) Kellar, K. E.; Henrichs, P. M.; Hollister, R.; Koenig, S. H.; Eck, J.; Wei,
D. Magn. Reson. Med. 1997, 38, 712-716.
8
a symmetric electronic ground state, S, which results in a
relatively slow electronic relaxation rate, especially compared
to the other lanthanide(III) ions. For clinically approved contrast
agents, it is primarily the rate of rotation which determines the
degree of relaxation enhancement at clinical imaging field
strengths.
(17) Toth, E.; Van Uffelen, I.; Helm, L.; Merbach, A. E.; Ladd, D.; Briley-
Saebo, K.; Kellar, K. E. Magn. Reson. Chem. 1998, 36, S125-S134.
(18) Toth, E.; Helm, L.; Kellar, K. E.; Merbach, A. E. Chem. Eur. J. 1999, 5,
in press.
(19) Aime, S.; Botta, M.; Crich, S. G.; Giovenzana, G.; Palmisano, G.; Sisti,
M. Bioconjugate Chem. 1999, 10, 192-199.
Proton relaxation will be optimum when the fluctuating
magnetic field is in resonance with the proton Larmor frequency.
For clinical imaging fields which range from 0.5 to 1.5 T, this
corresponds to Larmor frequencies of 20-65 MHz. Gd(III)
chelates used clinically have rotational frequencies in the 1-2
GHz range. The current generation of contrast agents has far
from optimal relaxation enhancing properties because of the fast
rate of molecular rotation.
(20) Bogdanov, A. A. J.; Weissleder, R.; Brady, T. J. AdV. Drug DeliVery ReV.
1995, 16, 335-348.
(21) Margerum, L. D.; Campion, B. K.; Koo, M.; Shargill, N.; Lai, J.-J.;
Marumoto, A.; Sontum, P. C. J. Alloys Compd. 1997, 249, 185-190.
(22) Raduchel, B.; Schmitt-Willich, H.; Platzek, J.; Ebert, W.; Frenzel, T.;
Misselwitz, B.; Weinmann, H.-J. Polym. Mater. Sci. Eng. 1998, 79, 516-
517.
(23) Toth, E.; Pubanz, D.; Vauthey, S.; Helm, L.; Merbach, A. E. Chem. Eur.
J. 1996, 2, 1607-1615.
(24) Wiener, E. C.; Brechbiel, M. W.; Gansow, O. A.; Foley, G.; Lauterbur, P.
C. Polym. Mater. Sci. Eng. 1997, 77, 193-194.
It was recognized early on that slowing down molecular
tumbling would increase relaxivity.2,5,6 Recent approaches for
(25) Wu, C.; Brechbiel, M. W.; Kozak, R. W.; Gansow, O. A. Bioorg. Med.
Chem. Lett. 1994, 4, 449-454.
9
J. AM. CHEM. SOC. VOL. 124, NO. 12, 2002 3153

A R T I C L E S
Caravan et al.
the target relative to background. Unlike other imaging modali-
ties such as X-ray and nuclear, binding also increases the signal
observed (relaxation enhancement) because the Gd(III) complex
now tumbles at a rotational rate that is similar to that of the
target molecule. Since the relaxivity increase only occurs upon
binding, there is the additional advantage of an improved target-
to-background ratio.
MS-325 (Chart 1) is the first RIME contrast agent3,4 designed
to image the blood pool by exploiting noncovalent binding to
the plasma protein human serum albumin (HSA). This interac-
tion limits the amount of free drug, which can extravasate from
the blood pool into nonvascular space and provides selective
vascular relaxation rate enhancement. HSA, which constitutes
about 4.5% of plasma (∼0.67 mM), is a large (67 kDa) globular
protein that is known to bind a variety of molecules (drugs,
metabolites, fatty acids) at different sites on the protein.28
Relaxation enhancement of water protons with MS-325 is much
greater in plasma or in HSA solution than in pure water. Strong
binding to albumin also reduces the fraction of free chelate
available for glomerular filtration in the kidneys, thus slowing
the renal excretion rate and contributing to an extended blood
half-life. The result of this is a longer time period for imaging,
which allows the radiologist to perform multiple imaging
experiments and to image under steady-state conditions. MS-
325 is currently undergoing Phase III clinical trials for imaging
arteries and veins.
Figure 1. ꢀ* versus [HSA]/[MS-325] at 37 °C in phosphate-buffered saline
(pH 7.4) for a 0.1 mmolal solution of MS-325 and 0-22.5% w/v HSA at
20 MHz (9) and 64.5 MHz (b).
HSA at 20 and 64.5 MHz (0.47 and 1.50 T, respectively) at 37
°C. It was previously shown that the observed relaxation rate
increases as MS-325 is added to a 4.5% solution of HSA or to
a solution of phosphate-buffered saline.4 The relaxation rate
increase is much larger in the presence of HSA, and the
relaxation rate does not increase linearly with concentration,
indicative of protein binding.29 To demonstrate the extent of
relaxation enhancement, an ꢀ* titration is shown in Figure 1.
Here, the ratio of paramagnetic relaxation rates in the presence
and in the absence of HSA is plotted versus increasing HSA/
MS-325 ratio, eq 2, for a constant molal concentration of MS-
325. There is a 9-fold increase in relaxivity at 20 MHz upon
binding to HSA and a 5-fold increase at 64 MHz.
From the biophysics perspective, increasing the rotational
correlation time unmasked the contributions from other param-
eters such as electronic relaxation and the water residency time.
Although relaxivities have been reported, there have been few
systematic studies to measure each of these parameters under
physiologically relevant conditions.
The purpose of this paper is to identify and quantify the
factors which contribute to the high relaxivity of MS-325 when
it is bound to HSA. Do parameters such as water residency time
and electronic relaxation also change upon binding, and if so,
how do they influence relaxivity? First, the binding of MS-325
to HSA has been investigated by direct (ultrafiltration) and
indirect (relaxivity) methods over a wide range of protein-to-
drug ratios. Establishment of binding constants of MS-325 to
HSA allows the choice of conditions whereby one MS-325
molecule is bound predominantly to a single site on HSA. Under
these conditions, information on the parameters influencing
relaxivity can be gleaned. To this end, variable-temperature and
variable-magnetic-field 1H and 17O T1 and T2 relaxation
measurements have been performed on MS-325 in the presence
and in the absence of HSA. A related compound, 5 (Chart 1),
has been prepared which contains no coordinated water mol-
ecules; this serves as an important outer-sphere control.
Fluorescence lifetime measurements on the Eu(III) analogues
of MS-325 in the presence and in the absence of HSA establish
the inner-sphere hydration number, q, of the complexes.
(1/T1)HobSsA - (1/T1)HdiSaA (1//T1)pHaSrAa
(1/T1)PobBsS - (1//T1)PdiBaS (1/T1)pPaBrSa
ꢀ* )
)
(2)
Albumin Binding. Before the nature of the protein-bound
relaxation enhancement can be understood, the noncovalent
interaction between MS-325 and HSA must be explored.
Protein-complex binding was determined by ultrafiltration
across a membrane with a MW 5000 cutoff. The amounts of
unbound MS-325 and total MS-325 were measured directly by
ICP-MS. The extent of binding of MS-325 to HSA is shown in
Figure 2. The total chelate concentration ranged from 0.05 to
5.3 mM while the protein concentration was fixed at 4.5%.
Under these conditions, the percent bound to HSA ranged from
88 to 37%, respectively. The data are presented in Figure 2 as
a plot of the ratio of bound MS-325 per HSA molecule (nj)
versus the concentration of unbound MS-325. It is clear from
Figure 2 that nj increases steeply as the molar concentration (on
a logarithmic scale) increases. It is apparent that at higher free
MS-325 concentrations, nj would increase well above 3 equiva-
lents bound MS-325 per HSA. A complete binding curve would
show an inflection and then reverse its slope so that ultimately
Results and Discussion
Relaxation Enhancement. Proton relaxation rates were
measured for MS-325 in the presence and in the absence of
(26) Jenkins, B. G.; Armstrong, E.; Lauffer, R. B. Magn. Reson. Med. 1991,
17, 164-178.
(27) Lauffer, R. B. Magn. Reson. Med. 1991, 22, 339.
(28) Peters, T. J. All About Albumin: Biochemistry, Genetics, and Medical
Applications; Academic Press: San Diego, 1996.
(29) Dwek, R. A. Nuclear Magnetic Resonance (N.M.R.) in Biochemistry; Oxford
University Press: Oxford, 1973.
9
3154 J. AM. CHEM. SOC. VOL. 124, NO. 12, 2002

Interaction of MS-325 with Human Serum Albumin
A R T I C L E S
Table 1. Stepwise Association Constants for the Binding of
MS-325 to HSA in PBS at 37 °Ca
binding event (i)
K (mM-1
)
∆G° (kJ‚mol-1
)
ai
i
1
2
3
4
11.0 (2.7)
24.0 (0.6)
17.5 (0.3)
14.3 (1.2)
15.6 (2.1)
0.84 (0.16)
0.26 (0.14)
0.43 (0.24)
a Numbers in parentheses represent 3σ.
limit F f 0, nj/F f Ka1.30 Table 1 also gives the corresponding
free energy changes of binding, ∆Gi. It should be explicitly
stated that the stoichiometric constants in Table 1 are not site
binding constants, ki. The latter cannot be addressed without
additional molecular information.
Sudlow and co-workers identified two sites on HSA which
bound a variety of drugs, and these were denoted site I and site
II.31,32 It has been shown that site I is a large site existing on
subdomain IIA capable of binding a diverse range molecules
including warfarin and salicylate. Within site I, it has been
proposed that there are at least three partially overlapping
regions. Yamasaki et al.33 suggest using warfarin, dansyl-L-
asparagine (DNSA), and n-butyl p-aminobenzoate as fluorescent
probes of these site I regions. Site II, located on subdomain
IIIA, binds diazepam, ibuprofen, and naproxen among other
drugs.28,34 Dansylsarcosine has often been used as a fluorescent
probe of site II.32
Figure 2. Binding isotherm for MS-325 (37 °C, phosphate-buffered saline,
pH 7.4) to 4.5% (w/v) HSA.
a saturation plateau would appear. The value at which nj reaches
a plateau would give the total number of occupied sites on HSA
when it is saturated with MS-325. Under the conditions
employed herein, this plateau cannot be determined.
The effect of incremental additions of MS-325 to solutions
containing equal amounts of a fluorescent probe (either warfarin,
DNSA, or dansylsarcosine) and HSA was studied by monitoring
the change in fluorescence with added MS-325. These probes
only fluoresce when they are bound to HSA, and a decrease in
fluorescence is taken as displacement by MS-325. The data for
these probe displacement studies are summarized in Figure 3.
The graph shows probe displacement for three separate experi-
ments; i.e., only one fluorophore is present during each titration.
The decrease in fluorescence intensity, expressed as a percentage
of initial fluorescence, is plotted as a function of the amount of
MS-325 added. It is clear that MS-325 is most effective at
displacing dansylsarcosine, the site II probe. MS-325 is a weak
displacer of DNSA. Warfarin (5 µM) shows no sign of
displacement from HSA (5 µM) by MS-325 at MS-325
concentrations out to 400 µM. This suggests that MS-325 binds
primarily to site II and has a lower affinity for site I.
The probe displacement data can be fit to obtain an inhibition
equilibrium constant, Ki, which reflects the affinity of MS-325
for a given probe’s binding site (Table 2). The model used
ignores MS-325 binding to sites other than that occupied by
the probe: the effect that this may have on the affinity of HSA
for the probe when a molecule of MS-325 is bound elsewhere,
or the effect it may have on the site affinity for MS-325 when
a molecule of MS-325 is already bound elsewhere, is not
addressed. Despite these caveats, it is interesting to note that
the affinity for the dansylsarcosine site (site II) is similar to the
first stoichiometric equilibrium constant. This suggests that at
Nevertheless, one can evaluate binding constants for the
earliest steps, particularly if a classical stoichiometric model of
the equilibria is used. For the following stepwise equilibria,
MS-325 + HSA h (MS-325)HSA;
[(MS-325)HSA]
K1 )
(3)
(4)
[MS-325][HSA]
MS-325 + (MS-325)HSA h (MS-325)2HSA;
[(MS-325)2HSA]
[MS-325][(MS-325)HSA]
K2 )
MS-325 + (MS-325)2HSA h (MS-325)3HSA;
[(MS-325)3HSA]
K3 )
(5)
(6)
[MS-325][(MS-325)2HSA]
MS-325 + (MS-325)3HSA h (MS-325)4HSA;
[(MS-325)4HSA]
K4 )
[MS-325][(MS-325)3HSA]
the function nj can be expressed in terms of these four association
constants, eq 7, where F represents the concentration of free
(unbound) MS-325. The experimental binding data were fit to
K1F + 2K1K2F2 + 3K1K2K3F3 + 4K1K2K3K4F4
1 + K1F + K1K2F2 + K1K2K3F3 + K1K2K3K4F4
nj )
(7)
(30) Klotz, I. M. Ligand Receptor EnergeticssA Guide for the Perplexed; John
Wiley & Sons: New York, 1987.
(31) Sudlow, G.; Birkett, D. J.; Wade, D. N. Mol. Pharmacol. 1975, 11, 824-
832.
eq 7, and the stoichiometric binding constants K2, K3, and K4
were evaluated (Table 1). In the fitting, K1 was fixed to 11.0
mM-1. This value was chosen based upon 36 data points (Table
S1) collected in the range 0.003 < F < 0.015 mM, since in the
(32) Sudlow, G.; Birkett, D. J.; Wade, D. N. Mol. Pharmacol. 1976, 12, 1052-
1061.
(33) Yamasaki, K.; Maruyama, T.; Kragh-Hansen, U.; Otagiri, M. Biochim.
Biophys. Acta 1996, 1295, 147-157.
(34) He, X. M.; Carter, D. C. Nature 1992, 358, 209-215.
9
J. AM. CHEM. SOC. VOL. 124, NO. 12, 2002 3155

A R T I C L E S
Caravan et al.
Figure 4. Mean bound relaxivity at 20 MHz (b) and 64.5 MHz (9) versus
total MS-325 concentration at 37 °C, in 4.5% (w/v) HSA, phosphate-
buffered saline, pH 7.4 solution.
Figure 3. Summary plot of displacement of fluorescent probes dansylsar-
cosine (b, 5 µM), dansyl-L-asparagine (9, 15 µM), and warfarin (2, 5 µM)
from equimolar HSA by MS-325 at 37 °C, phosphate-buffered saline, pH
7.4.
Table 3. jr1b for MS-325 as [MS-325]free f 0
Table 2. Inhibition Constants for MS-325 and 5 at 37 °C in PBS
for the Displacement of Various Fluorescent Probes
frequency
r¯ (mM-1 s-1
)
1b
20
64.5
50.8
25.0
K
i
complex
dansyl−sarcosine
dansyl-asparagine
warfarin
MS-325
5
85 (3)
62 (10)
1500 (850)
1500 (1060)
no displacement
1260 (530)
bound relaxivity is site dependent and variable. As the concen-
tration of MS-325 is decreased, the mean bound relaxivity
increases. This indicates that the high-affinity binding site is
also a high-relaxivity site for MS-325. Combining these data
with the results of the ꢀ* titration, one can estimate relaxivities
at the high-affinity site for the magnetic fields used in this study.
These are given in Table 3; the value at 20 MHz is similar to
that reported by Muller et al.36
Molecular Factors Affecting Relaxivity. To better under-
stand the mode of action of MS-325, the molecular factors which
determine relaxivity were probed. As outlined in the Introduc-
tion, relaxivity depends on (1) the number of water molecules
coordinated in the inner-sphere, (2) the rate of water exchange,
(3) the electronic relaxation rate of the Gd(III) ion, (4) the
rotational diffusion time of the Gd-H vector, and (5) the so-
called outer-sphere relaxivity wherein the Gd ion relaxes water
molecules not directly coordinated to it.
Hydration Number. Fluorescence lifetime experiments were
performed on the Eu(III) analogue of MS-325 in H2O and D2O
solutions in the presence and in the absence of HSA. It is well
established that the ratio of fluorescence decay constants in both
solvents is proportional to q, the hydration number of the
complex.37,38 Hydration numbers are given in Table 4 and
demonstrate that MS-325 has one coordinated water molecule
either free in PBS buffer or when bound to HSA.
a Numbers in parentheses represent 3σ.
low concentrations of drug, MS-325 will be primarily bound
to this high-affinity site (site II).
There have been two recent reports on the interaction of MS-
325 with HSA.35,36 In both reports an association constant was
estimated from relaxation rate enhancement data assuming a
model where there is only one binding site. The Ka1 measured
directly in this study using classical binding techniques is
intermediate between the values determined indirectly in the
previous two reports.
Site-Dependent Relaxivity. The combination of binding plus
proton relaxation rate enhancement data provides a unique
opportunity to evaluate the dependence of site occupancy on
relaxivity. One can define an average bound relaxivity, jr1b,
which is analogous to nj, eq 8. For the ultrafiltration binding
1
/
- 1/
- [MS - 325]r1f
T1dia
T1obs
jr1b
)
(8)
[MS - 325]b
1
experiment summarized in Figure 2, H relaxation rates were
measured at 20 MHz. Since the concentrations of MS-325 free
and bound to HSA are determined by the ultrafiltration
experiment, and the relaxivity of MS-325 free, r1f, is known,
then the function jr1b can be calculated. Figure 4 shows the mean
bound relaxivity as a function of total MS-325 concentration
in the presence of 4.5% HSA. It is apparent from Figure 4 that
Outer-Sphere Relaxivity. To estimate the extent of outer-
sphere relaxivity, the q ) 0 analogue 5 (Scheme 1) was
synthesized. The synthesis of 5 commenced by coupling
(37) (a) Horrocks, W. DeW., Jr.; Sudnick, D. R. J. Am. Chem. Soc. 1979, 101,
334-340. (b) Supkowski, R. M.; Horrocks, W. DeW., Jr. Inorg. Chem.
1999, 38, 5616-5619.
(38) Beeby, A.; Clarkson, I. M.; Dickins, R. S.; Faulkner, S.; Parker, D.; Royle,
L.; de Sousa, A. S.; Williams, J. A. G.; Woods, M. J. Chem. Soc., Perkin
Trans. 2 1999, 493-504.
(35) Aime, S.; Chiaussa, M.; Digilio, G.; Gianolio, E.; Terreno, E. JBIC, J.
Biol. Inorg. Chem. 1999, 4, 766-774.
(36) Muller, R. N.; Raduchel, B.; Laurent, S.; Platzek, J.; Pierart, C.; Mareski,
P.; Vander Elst, L. Eur. J. Inorg. Chem. 1999, 1949-1955.
9
3156 J. AM. CHEM. SOC. VOL. 124, NO. 12, 2002

Interaction of MS-325 with Human Serum Albumin
A R T I C L E S
Scheme 1
Table 4. Hydration Numbers of the Eu(III) Analogues of MS-325
and 5 in HEPES buffer (pH 7.0) or in 4.5% HSA/HEPES Buffer
Solution at 37 °C
a slightly higher affinity for HSA than MS-325 (95% bound at
0.1 mM 5, 4.5% HSA vs 88% for MS-325 under the same
conditions). It also binds to site II, as evinced from fluorescent
probe displacement studies (Ki values given in Table 2).
Luminescence lifetime measurements established that the Eu-
(III) analogue of 5 has no inner-sphere water molecule in the
presence or in the absence of HSA (Table 4).
lifetime (µs)
system
H O
2
D O
2
q
MS-325/HEPES
MS-325/HSA
5/HEPES
597
654.4
1222
1307
2199
2535
1928
1933
1.02
0.93
0.0
Inner-Sphere Relaxivity. Longitudinal (r1) and transverse
(r2) molal relaxivities were measured for MS-325 and 5 over
5/HSA
-0.06
-1
-1
2
a q ) τH O - τD O - 0.30; τ ) lifetime in ms.37b
1
2
the H Larmor frequency range 4-64 MHz using 0.1 mmolal
Gd in the presence and in the absence of 22.5% HSA. A high
(∼34:1) protein-to-drug ratio was used in order to maximize
the amount of drug bound to HSA and also to maximize the
amount of drug bound in the highest affinity site. Plots of molal
relaxivity for MS-325 and 5 with and without HSA are shown
in Figures 5 and 6, respectively. In the presence of HSA, the
relaxivity of MS-325 increases dramatically: the r1 curve
increasing and then decreasing with field, typical of a slow
rotating molecule, and the r2 curve rising to a maximum and
then reaching a plateau. The relaxivity of 5 also increases upon
binding to HSA; although the effect is not as dramatic as that
of MS-325, there is still a 2-4-fold increase depending on field.
To ascertain the amount of relaxivity conferred to the solvent
by the inner-sphere water molecule on MS-325, an inner-sphere
bound relaxivity was calculated. First, jr1,2b was calculated for
N-BOC-L-serine(Obn)-OSu to triethylenetriamine. The resultant
amide was treated with TFA, then borane-tetrahydrofuran
complex to furnish amine 2, which was alkylated with tert-
butyl bromoacetate, providing the alcohol hexaester 3. A
phosphorylation sequence utilizing phosphorous trichloride and
3, followed by chelation with gadolinium chloride, yielded 5.
Compound 5 is based on the well-studied triethylenetet-
raamine hexaacetate (TTHA) ligand core, which is known to
coordinatively saturate lanthanide(III) ions.39-44 Complex 5 has
(39) Alpoim, M. C.; Urbano, A. M.; Geraldes, C. F. G. C.; Peters, J. A. J. Chem.
Soc., Dalton Trans. 1992, 463-467.
(40) Chang, C. A.; Brittain, H. G.; Telser, J.; Tweedle, M. F. Inorg. Chem.
1990, 29, 4468-4473.
(41) Ruloff, R.; Gelbrich, T.; Sieler, J.; Hoyer, E.; Beyer, L. Z. Naturforsch.,
B: Chem. Sci. 1997, 52, 805-809.
(42) Ruloff, R.; Prokop, P.; Sieler, J.; Hoyer, E.; Beyer, L. Z. Naturforsch., B:
Chem. Sci. 1996, 51, 963-968.
(43) Wang, R.-Y.; Li, J.-R.; Jin, T.-Z.; Xu, G.-X.; Zhou, Z.-Y.; Zhou, X.-G.
Polyhedron 1997, 16, 1361-1364.
(44) Wang, R.-Y.; Li, J.-R.; Jin, T.-Z.; Xu, G.-X.; Zhou, Z.-Y.; Zhou, X.-G.
Polyhedron 1997, 16, 2037-2040.
9
J. AM. CHEM. SOC. VOL. 124, NO. 12, 2002 3157

A R T I C L E S
Caravan et al.
Figure 5. Observed longitudinal (r1obs, circles) and transverse (r2obs, squares)
relaxivity for 0.1 mmolal MS-325 in the presence (filled symbols) and in
the absence (open symbols) of 22.5% (w/v) HSA at 37 °C, phosphate-
buffered saline, pH 7.4. Solid lines are fits to the data as described in the
text.
Figure 6. Observed longitudinal (r1obs, circles) and transverse (r2obs, squares)
relaxivity for 0.1 mmolal 5 in the presence (filled symbols) and in the
absence (open symbols) of 22.5% (w/v) HSA at 37 °C, phosphate-buffered
saline, pH 7.4.
Table 5. Four-, Five-, and Six-Parameter Fits of the NMRD Data
at 37 °C in PBS, 22.5% HSA, 0.1 mmolal MS-325a
both MS-325 and 5 since both the relaxivity free and the
concentration free (from ultrafiltration binding studies) were
known. Since 5 is chemically very similar to MS-325 (it has
the same donor atom set except that a carboxylato oxygen
replaces a water oxygen, and it binds to the same site on HSA),
it was assumed that the relaxivity of 5 bound to HSA reflected
the outer-sphere contribution to relaxivity for MS-325. Thus,
the inner-sphere bound relaxivity, jrI1S,2b, for MS-325 is given by
eq 9.
four-parameter
five-parameter
six-parameter
τ
m (ns)
τv (ps)
207 (200-213)
21 (18-24)
172 (130-213)
20 (17-23)
192 (144-240)
23 (19-27)
∆2 (×1018 s-2
)
8.1 (7.6-8.6)
7.0 (5.9-8.1)
7.1 (5.7-8.4)
τ
R (ns)
13.3 (11.6-15.1) 10.1 (7.4-12.9) 9.4 (6.6-12.1)
3.10 (fixed)
r
Gd-H (Å)
3.19 (3.10-3.29) 3.17 (3.06-3.28)
0 (fixed) 3.1 (1.2-5.1)
A/p (×106 rad/s) 0 (fixed)
a Numbers in parentheses represent 95% confidence range.
improve the fits, and the original four fitted parameters are
largely unchanged.
jrIS ) jrM1,2Sb-325 - jr15,2b
(9)
1,2b
Comparison with Unbound MS-325. To understand which
parameters change upon binding to HSA, variable-temperature
1H and 17O relaxation measurements were carried out on MS-
325 in the absence of HSA. Merbach and co-workers have
established that 17O relaxation rates for Gd complexes are
determined solely by inner-sphere mechanisms;46 this is not true
Once the inner-sphere bound relaxivity has been calculated,
the r1 and r2 curves can be fit to the Solomon-Bloembergen-
Morgan equations (see Supporting Information) which describe
paramagnetic relaxation.45 Both transverse and longitudinal
relaxation rate data were fit simultaneously to a dipolar
relaxation mechanism allowing for chemical exchange. In total,
four parameters were varied: τm, the water residency time; τR,
the rotational correlation time which describes the rotational
diffusion of the Gd-H vector; ∆t2, the square of the trace of
the transient zero-field-splitting (ZFS) tensor, which is a measure
of how much the electronic structure of the Gd ion is transiently
distorted from spherical symmetry; and τv, a correlation time
which describes the modulation of the transient ZFS. The latter
two parameters define electronic relaxation, T1,2e. In the fitting,
the electron spin-nuclear spin distance, rGd-H, was fixed at 3.1
Å. This can also be allowed to vary. Finally, at higher fields a
scalar relaxation mechanism may contribute to 1/T2. This
mechanism involves the Gd-H hyperfine coupling constant AH/
p. Table 5 shows results for the four-, five-, and six-parameter
fits. Introduction of additional parameters did not significantly
1
of H relaxation rates for rapidly tumbling molecules, where
outer-sphere contributions can contribute up to 50% of the
observed relaxation rate.1 T2 and chemical shift measurements
of H217O at high field (7.05 T in this instance) yield information
on the water exchange rate and T1e of the Gd ion. The 1/T2
curve versus temperature was parabolic, with a maximum at
ca. 30 °C (Figure 7). On the low-temperature side of the
maximum, the transverse relaxation rate is dominated by water
exchange, and the slope at the low-temperature extreme gives
the activation enthalpy for exchange. At higher temperatures,
1/T2 is modulated by both electronic relaxation and exchange.
Because only one field strength was used, the electronic
relaxation rate was calculated at each temperature and assumed
to have an exponential temperature dependence given by ∆ET
.
1e
(46) Micskei, K.; Helm, L.; Brucher, E.; Merbach, A. E. Inorg. Chem. 1993,
32, 3844-3850.
(45) Bertini, I.; Luchinat, C. Coord. Chem. ReV. 1996, 150, 1-295.
9
3158 J. AM. CHEM. SOC. VOL. 124, NO. 12, 2002

Interaction of MS-325 with Human Serum Albumin
A R T I C L E S
Table 6. Parameters Obtained from the Simultaneous Fit of T1,
T2, and Chemical Shift Data for H217O in the Presence of MS-325
or Magnevist and Literature Valuesa
this work
ref 36
ref 35
Magnevistb,c
τm37 (ns)
69 (20)
83
5.1
51.3
+51.3 +59.0
3.6 17.0
105
4.0
53.0
140,b 130c
3.1,b 3.3c
55.7,b 51.6c
+66,b +53.0b
3.2b
kex298 (×106 s-1
)
5.8 (0.6)
53.7 (5.6)
+65 (10)
5.0 (2.6)
-7.7 (14.5) 6.4
115 (10)
31.5 (3.0)
-4.46 (1.2) -4.1
0.23 (0.14) NR
∆Hq (kJ mol-1
)
∆Sq (J K-1mol-1
)
)
1/Τ1e37(×107 s-1
∆ET (kJ mol-1
)
10.0 (fixed) -12.5b
1e
τR37 (ps)
117
NR
NR
NR
44c
∆ER (kJ mol-1
)
17.3c
A/p (×106 rad/s)
-3.8 (fixed) -3.8c (fixed)b
Cos
NR
0.18c
a Numbers in parentheses represent 3σ. b This work, c Reference 48. NR
) not reported.
relaxivity at fields >0.5 T is determined by the rotational
correlation time and the electron spin-nuclear spin distance,
and these two parameters are strongly correlated; decreasing
(increasing) rGd-H by 0.1 Å decreases (increases) τR by 25 ps.
A τR of 117 ps was reported for the La analogue of MS-325
2
determined by H NMR at 37 °C.36
Why Binding to HSA Increases Relaxivity. The increase
in relaxivity upon HSA binding is primarily because of a
dramatic reduction in the rotational diffusion rate of the
molecule. The rotational rate is decreased by almost a factor of
100. There are two other favorable interactions. The outer-sphere
relaxivity increases by a factor of 2-4, which may be due to
the presence of a long-lived (>1 ns) water molecule in the
second coordination sphere. The electronic relaxation rate, 1/T1e,
also slows down upon binding. On the other hand, the water
exchange rate is slowed by a factor of 2-3, which limits the
amount of relaxation enhancement observed.
Figure 7. Reduced 17O relaxation rates (1/T1r and 1/T2r) and chemical shifts
(ωr) of H217O in the presence of MS-325 in phosphate-buffered saline, pH
7.4, as function of reciprocal temperature. Solid lines represent the
simultaneous fit to the data as described in the Experimental Section.
The hyperfine coupling constant, Ao/p, between the electron spin
and the 17O nucleus was determined primarily through chemical
shift measurements, although the T1, T2, and chemical shift data
were fit simultaneously. The Gd T1 measurements on H217O
yield the rotational correlation time, provided the gadolinium-
water oxygen distance is known and the nuclear quadrupolar
coupling constant for H217O coordinated to Gd is known. The
distance rGd-O was set to 2.48 Å, which is the value determined
from the X-ray crystal structure of MS-325.47 The nuclear
quadrupolar coupling constant was assumed to be the same as
that determined for [Gd(DTPA)(H2O)]2- by Powell et al.48
The water residency time for MS-325 is shorter than that of
[Gd(DTPA)(H2O)]2- (Table 6).48 This difference is signifi-
cant: there is ca. a 5 K shift to lower temperature for the 1/T2
maximum for MS-325 compared to [Gd(DTPA)(H2O)]2- mea-
sured under identical conditions (Table S2, Supporting Informa-
tion). While the water exchange rate increases upon introduction
of the phosphodiester functionality on the DTPA backbone, there
appears to be no difference in the electronic relaxation rate,
1/T1e, at 7.05 T between MS-325 and [Gd(DTPA)(H2O)]2-. The
water exchange rate determined here for MS-325 is in good
agreement with that published by Muller et al.,36 and the water
exchange rate for [Gd(DTPA)(H2O)]2- is in good agreement
with that reported by Powell et al.48
The rotational correlation time reported here (∼10 ( 3 ns)
is longer than that reported by Muller et al.,36 but the NMRD
profiles are very similar in shape and magnitude. These authors
1
give values of 3.3 and 4.4 ns for τR determined by H NMRD
and 6-7 ns for τR determined from 2H line broadening studies.
The rotational correlation time determined in this study is more
in line with their 2H result than their NMRD analysis. However,
under the conditions employed in their study, a considerable
amount of MS-325 (28%) is unbound, implying that the mean
bound relaxivity (cf. Figure 4) is lower. It is likely that these
authors were interpreting data in a region where the distribution
of MS-325 bound to HSA was spread over several sites. In
addition, these authors may have underestimated the outer-
sphere contribution as compared to what was experimentally
observed for 5. These authors also assume that the water
exchange rate does not change upon binding; this will have the
effect of shortening τR in the fitting procedure. On the other
hand, the higher protein concentration in this study may alter
the microviscosity such that rotation within the whole system
is slowed. Stanisz and Henkelman49 have shown that increasing
amounts of macromolecule can increase the relaxation enhance-
ment of water by [Gd(DTPA)(H2O)]2-, although there is no
specific interaction between the macromolecule and the metal
complex. The difference in microviscosity may slow the
rotational rate by as much as 50% upon increasing the protein
content from 4 to 22.5%.
The rotational correlation time at 37 °C is about 120 ps (Table
1
6). Analysis of the 17O T1 data gives 115 ps, while H T1 and
T2 data give 125 ps using a Gd-H distance of 3.1 Å. Proton
(47) Tyeklar, Z.; Midelfort, K.; Dunham, S.; Lauffer, R.; McMurry, T.;
Hollander, F., unpublished results.
(48) Powell, D. H.; Ni Dhubhghaill, O. M.; Pubanz, D.; Helm, L.; Lebedev, Y.
S.; Schlaepfer, W.; Merbach, A. E. J. Am. Chem. Soc. 1996, 118, 9333-
9346.
(49) Stanisz, G. J.; Henkelman, R. M. Magn. Reson. Med. 2000, 44, 665-667.
9
J. AM. CHEM. SOC. VOL. 124, NO. 12, 2002 3159

A R T I C L E S
Caravan et al.
The increased outer-sphere relaxivity upon binding is likely
because of long-lived (∼1 ns) water molecule(s) in close
proximity (<5 Å) to the Gd(III) ion. This effect has been
observed previously by others on q ) 0 complexes.50,51 The q
) 0 Mn(II) analogue of MS-325 also shows an increased
relaxivity upon HSA binding.52
The slower electronic relaxation rate upon binding is an
unexpected benefit. If the electronic relaxation parameters were
identical to those reported for [Gd(DTPA)]2- 48
, then the r1 peak
it is likely that the water exchange rate slows down upon HSA
binding, this has not been unequivocally shown.
Conclusion
MS-325 functions as a contrast agent by targeting HSA in
blood. The affinity for HSA is moderate: at typical circulating
concentrations of 0.1 mM, 88% of the complex is noncovalently
bound to HSA. Binding to HSA causes a dramatic increase in
the relaxivity of the molecule, a 9-fold increase at 20 MHz. It
was shown that MS-325 binds preferentially to the site II region
of HSA and that this high-affinity site is also a high-relaxivity
site. The relaxivity increase is primarily a result of slow
rotational diffusion of the molecule upon binding to HSA.
maximum in Figure 5 would occur at a higher frequency (40-
45 MHz), and its amplitude would be lower. Apart from
molecular symmetry, there are no predictors of electronic
relaxation in Gd(III) complexes. It is unclear why the electronic
relaxation rate of the Gd(III) ion should change when its
complex is associated with HSA (Gd(III) electronic relaxation
is not modulated by rotation). One possible explanation may
lie in the determination of the electronic parameters. For fast
tumbling [Gd(DTPA)]2-, relaxation at higher fields is deter-
mined primarily by the rotational rate of the molecule; the
electronic parameters ∆t2 and τv are determined from modeling
the low-field relaxivity data to the SBM equations. However,
at low fields the SBM equations may not be valid for this non-
spherically-symmetric molecule. When the rotational rate is
much slower, as is the case here, the electronic relaxation rate
(which decreases with the square of the magnetic field) continues
to influence the relaxivity of the complex at higher fields (where
the SBM equations are valid). In this study, ∆t2 and τv were
determined over the frequency range 4-64 MHz, while for
[Gd(DTPA)]2-, these parameters were determined in the range
0.01-4 MHz. It may be that the electronic relaxation rate does
not change upon HSA binding; rather, the difference in
electronic parameters may be due to the method used to
determine them.
Experimental Section
Materials. Human serum albumin (HSA), product no. A-1653
(Fraction V Powder 96-99% albumin, containing fatty acids), and HSA
absorbance standards, along with the fluorescent probes dansyl L-
asparagine, dansylsarcosine (piperidinium salt), and warfarin (sodium
salt), were purchased from Sigma Chemical Co. (St. Louis, Mo.).
Ultrafiltration units (UFC3LCC00, regenerated cellulose membrane of
5000 Da nominal molecular weight cutoff) were obtained from
Millipore Corp. (Bedford, MA). Other reagents were supplied by
Aldrich Chemical Co., Inc., and were used without further purification.
Solvents (HPLC grade) were purchased from various commercial
suppliers and used as received. Column chromatography was conducted
using silica gel from EM Merck. The gadolinium sodium salt of MS-
325 was prepared by Raylo Chemicals (Alberta, Canada); this was
formulated in-house at a concentration of 243 mM.
Concentrations. All gadolinium concentrations were determined by
ICP-MS on a PE-SCIEX ELAN 5000 system.
Synthesis. Intermediates and compounds were characterized by NMR
with a Varian Unity 300 NMR spectrometer. HPLC purity analysis
(both UV and MS detection) was carried out with a HP-1100 MSD
system using a gradient of 20 mM ammonium formate (pH 6.8) with
5% (9:1 ACN:20 mM ammonium formate) to 95% (9:1 ACN:20 mM
ammonium formate) over 12 min (0.8 mL/min, Vydac C4, 4.6 × 250
mm).
The apparent decrease in the water exchange rate is also
somewhat surprising. This change in τm was also noted by Aime
et al.,35 even though these authors were working under condi-
tions where several equivalents of MS-325 were bound to
albumin. The mechanism of this is unclear. It is known that
[Gd(DTPA)(H2O)]2- exchanges water via a dissociative mech-
anism,48 and as the activation parameters for MS-325 in the
(S)-{1-[2-(2-Aminoethylamino)ethylcarbamoyl]-2-hydroxyethyl}-
carbamic Acid tert-Butyl Ester, 1. A suitable reaction flask, equipped
with an addition funnel, nitrogen inlet, and temperature probe, was
charged with diethylenetriamine (17.6 g, 170.1 mmol, 10 equiv). To
the flask was added a mixture consisting of N-Boc-L-serine (OBzl)-
OSu (6.7 g, 17.1 mmol, 1.0 equiv) and CH2Cl2 (40 mL) over a period
of 40 min while maintaining an internal temperature 0-5 °C. The
mixture was allowed to stir while warming to room temperature and
stirred for an additional 3.0 h. The reaction mixture was diluted with
water (50 mL) and stirred for 20 min. The layers were separated, and
the organic layer was concentrated under vacuum (40-45 °C water
bath, 10-15 mmHg) to provide a colorless foam (5.9 g). The foam
was purified by preparative HPLC chromatography (ACN:water:0.1%
absence of protein are similar to those of [Gd(DTPA)(H2O)]2-
,
it is highly likely that MS-325 also has a dissociative water
exchange mechanism.
The NMRD data were fit to an isotropic model. Strictly
speaking, this is invalid. However, the data fit quite well to
this model; a large degree of anisotropy would noticeably alter
the shape of the dispersion profile. In addition, the X-band EPR
spectrum of vanadyl-labeled MS-325 in HSA is very isotropic,
suggesting that the bound Gd(III) complex here tumbles in an
isotropic fashion.53 One consequence of a small degree of
anisotropy would be to lower the relaxivity. From the present
data, it is impossible to distinguish a small degree of rotational
anisotropy from a lengthening of the water residency time. While
1
TFA, C18 Vydac column) to provide 1 (2.7 g). H NMR (300 MHz,
CDCl3): δ 7.3 (m, 5H), 7.2 (br s, 1H), 5.5 (d, J ) 6.8, 1H), 4.54-4.44
(q, J ) 11.8 Hz, 2H), 4.2 (br s, 1H), 3.88-3.83 (dd, J ) 3.9 and 5.3
Hz, 1H), 3.58-3.53 (q, J ) 6.0 Hz, 1H), 3.35-3.29 (q, J ) 11.5 Hz,
2H), 2.6 (m, 4H), 2.5 (m, 2H), 1.4 (s, 9H), 1.3 (m, 3H). 13C NMR (75
MHz, CDCl3): δ 169.4, 154.6, 136.6, 127.5, 126.9 (2), 126.7 (2), 79.2,
72.4, 69.1, 53.2, 50.9, 47.3, 40.6, 38.3. MS (m/z): M + 1 ) 381.
(R)-2-Amino-3-[2-(2-aminoethylamino)ethylamino]propan-1-ol Tet-
rahydrochloride Salt, 2. A suitable reaction flask was charged with
amide 1 (5.9 g) and 10 mL of trifluoroacetic acid (TFA). The mixture
was stirred for 2.0 h at room temperature. The mixture was concentrated
under reduced pressure to give an oil. The oil was triturated with diethyl
ether. The solid was collected and dried until a constant weight to give
a sticky white foam (5.9 g). The foam was combined with MeOH (50
(50) Aime, S.; Batsanov, A. S.; Botta, M.; Howard, J. A. K.; Parker, D.;
Senanayake, K.; Williams, G. Inorg. Chem. 1994, 33, 4696-4706.
(51) Aime, S.; Botta, M.; Crich, S. G.; Giovenzana, G. B.; Pagliarin, R.; Piccinini,
M.; Sisti, M.; Terreno, E. J. Biol. Inorg. Chem. 1997, 2, 470-479.
(52) Dunham, S. U.; Caravan, P., unpublished results.
(53) Clarkson, R. B.; Caravan, P., unpublished results.
(54) Banci, L.; Bertini, I.; Luchinat, C. Nuclear and Electron Relaxation;
VCH: Weinheim, 1991.
(55) Toth, E.; Helm, L.; Merbach, A. E.; Hedinger, R.; Hegetschweiler, K.;
Janossy, A. Inorg. Chem. 1998, 37, 4104-4113.
9
3160 J. AM. CHEM. SOC. VOL. 124, NO. 12, 2002

Interaction of MS-325 with Human Serum Albumin
A R T I C L E S
mL) and Amberlite IRA-440C (hydroxide form) resin (which was
previously washed with H2O, H2O/MeOH (1:1), and MeOH). The slurry
was stirred for 1.5 h and filtered. The resin was washed with MeOH
(2 × 50 mL), and the combined filtrates were concentrated under
reduced pressure to give an oil. The oil was diluted with THF (50 mL)
and concentrated under reduced pressure (this removed traces of
methanol) to provide an oil (2.3 g).
over a period of 3 min while maintaining an internal temperature of
12-15 °C. The reaction mixture was warmed to room temperature and
stirred for 2.5 h. The layers were separated, and the organic layer was
washed with 10% aqueous sodium thiosulfate (2 × 75 mL).
To the above organic layer was added tetraoctylammonium bromide
(0.39 g, 0.708 mmol, 0.1 equiv), followed by concentrated HCl (11.51
M, 21 mL) over a period of 22 min while maintaining an internal
temperature of 22-35 °C. The mixture was stirred for 16.0 h at room
temperature. The layers were separated, and the organic layer was
discarded.
To the above aqueous layer was added 8 M NaOH until a pH of 6.8
was achieved. The solution was concentrated under reduced pressure
(50-55 °C, vacuum 85 mmHg) until a volume of 80 mL. The solution
was loaded onto C-18 reversed-phase silica gel (50 g, packed wet in
MeOH and then washed with 200 mL of MeOH, 200 mL of MeOH/
H2O (1:1), and finally 200 mL of H2O) and eluted with water. The
first 225 mL collected was discarded, and the next 225 mL collected
was retained.
To the above retained solution was added 6 M HCl until a pH of
1.63 was achieved. The formed slurry was stirred for 1 h and filtered.
The solid was washed with pH 1.67 aqueous solution and dried (48-
50 °C, 4-6 mmHg) to a constant weight (18.0 h) to obtain 4 as an
off-white solid (0.9 g). 1H NMR (300 MHz, D2O): δ 7.4-7.2 (m, 10H),
4.2 (br s, 1H), 3.8 (s, 2H), 3.4-2.8 (m, 13H), 2.75-2.25 (m, 11H),
2.20-2.15 (m, 3H), 1.9 (br d, 2H), 1.7 (m, 2H). 31P (121 MHz, D2O/
phosphoric acid): δ -5.5. MS (m/z): M + 1 ) 839.35.
Tetrasodium {(2-(R)-[(4,4-Diphenylcyclohexyl)phosphonooxy-
methyl]diethylenetetraaminehexaaacetato)(aquo)gadolin-
ium(III)}, 5. Compound 4 (38 mg, 45 µmol) was dissolved in distilled,
deionized water (4 mL), and the pH was adjusted to 6.5 with NaOH.
The exact ligand concentration of the solution was determined by
photometric titration with standardized gadolinium chloride in 0.02 M
xylenol orange (pH 4.9, acetate buffer, monitor at 572 nm). There is a
marked increase in absorbance once the endpoint has been reached.
One equivalent of GdCl3‚6H2O (19.3 mg, 37.5 µmol) was added to
the solution of 4 and the pH adjusted to 6.5 by the addition of NaOH
to give an aqueous solution of the title compound (1.42 mg of Gd per
mL, 9.03 mM). Analytical HPLC showed a single peak (5.9 min with
both UV and MS (negative ion, m/z ) 992 [M - H2O + H]-) detection.
Preparation of 4.5% (w/v) Human Serum Albumin (HSA) and
MS-325/HSA Samples. HSA was dissolved in a buffer (PBS) of 10
mM sodium phosphate and 150 mM sodium chloride (pH 7.4). The
protein concentration was determined by measuring its absorbance at
280 nm and comparing the reading to a standard curve constructed
from the absorbance of HSA standards with 2, 4, 6, 8, and 10 g/dL,
respectively, in PBS buffer. Once the HSA concentration was deter-
mined, enough PBS was added to dilute the protein to 4.5 g/dL (i.e.,
4.5% w/v).
The oil (2.2 g) was transferred to a reaction flask equipped with an
addition funnel, temperature probe, and nitrogen inlet and diluted with
THF (50 mL). To the mixture was added BH3‚THF (1.6 L of a 1 M
solution in THF, 50 mL) over a period of 30 min while maintaining an
internal temperature of 25-35 °C. The mixture was heated to reflux,
stirred for 18 h, and then cooled to 0-5 °C. A solution of 2 M HCl
(50 mL) was added over 30 min while maintaining an internal
temperature of 0-5 °C, followed by addition of concentrated HCl (50
mL) over a period of 20 min while maintaining an internal temperature
of 0-5 °C. The mixture was warmed to room temperature and
concentrated under reduced pressure until 50 mL of solution remained.
The solution was refluxed at 6 h and then concentrated under reduced
pressure until 10 mL of solution remained. The solution was diluted
with EtOH (10 mL) and stirred at room temperature for 1.0 h.
Additional EtOH was added until the solution became cloudy, and the
solution was concentrated under reduced pressure to give an oil. The
oil was triturated with diethyl ether to give a solid. The solid was
collected by filtration and dried until a constant weight to give a sticky
white powder (hydroscopic, 1.8 g). 1H NMR (300 MHz, D2O): δ 3.77-
3.64 (m, 3H), 3.57-3.14 (m, 10 H). 13C NMR (75 MHz, D2O): δ
59.1, 49.2, 47.0, 44.7, 44.0, 43.4, 35.4. MS (m/z): M + 1 ) 177.15.
[[2-(Bis-tert-butoxycarbonylmethylamino)ethyl]-(2-{[2-(bis-tert-
butyoxycarbonylmethylamino)-3-hydroxypropyl]-tert-
butoxycarbonylmethylamino}ethyl)amino]acetic Acid tert-Butyl
Ester, 3. A reaction flask equipped with a nitrogen inlet, addition funnel,
and temperature probe was charged with 2 (1.7 g, 5.3 mmol, 1.0 equiv),
KI (0.88 g, 10.6 mmol, 2.0 equiv), diisopropylethylamine (12.9 mL,
74.2 mmol, 14.0 equiv), and DMF (40 mL). The mixture was warmed
to 45-50 °C, and tert-butyl bromoacetate (5.5 mL, 37.1 mmol, 7.0
equiv) was added while maintaining an internal temperature of 45-50
°C. The reaction mixture was stirred for 18 h and then cooled to room
temperature. Ethyl acetate (80 mL) and water (80 mL) were added,
and the mixture was stirred for 15 min. The layers were separated, and
the organic layer was concentrated under reduced pressure to provide
an oil. The oil was purified by flash column chromatography (30 g of
silica gel, hexane/IPA). The fractions containing only desired product
1
were combined to provide 3 as an oil (0.3 g). H NMR (300 MHz,
CDCl3): δ 1.4 (br s, 54H), 2.5-2.4 (dd, J ) 9.2 and 3.7 Hz, 1H),
2.6-2.9 (m, 10H), 3.2-3.5 (m, 13 H), 3.6 (dd, J ) 4.5 and 6.8 Hz,
1H). MS (m/z): M + 1 ) 861.
(R)-[[2-(Carboxymethyl-{2-dicarboxymethylamino-3-[(4,4-diphe-
nylcyclohexyloxy)hydroxyphosphoryloxy]propyl}amino)ethyl]-2-di-
carboxymethylaminoethyl)amino]acetic Acid, 4. A reaction flask,
equipped with a nitrogen inlet, temperature probe, and addition funnel,
was charged with PCl3 (0.61 mL, 7.08 mmol, 1.0 equiv) and THF (12
mL). A solution consisting of 4,4-diphenylcyclohexanol (1.78 g, 7.08
mmol, 1.0 equiv) in THF (15 mL) was added over a period of 20 min
while maintaining an internal temperature of -6 to 0 °C. The mixture
was stirred for 30 min. A solution consisting of imidazole (2.4 g, 35.4
mmol, 5.0 equiv) in THF (15 mL) was added over a period of 15 min
while maintaining an internal temperature of -6 to 0 °C. The mixture
was stirred for 30 min. A solution consisting of alcohol 3 (5.48 g, 6.4
mmol, 0.9 equiv) in THF (12 mL) was added over a period of 30 min
while maintaining an internal temperature of -6 to 0 °C. The mixture
was stirred for 30 min. Water (12 mL) was added all at once while
maintaining an internal temperature of -6 to 6 °C. The mixture was
stirred for 5 min. Heptane (36 mL), toluene (4 mL), and 5 M HCl (12
mL) were added over 5 min while maintaining an internal temperature
of 6-12 °C. Sodium periodate (1.2 g, 5.7 mmol, 0.8 equiv) was added
Two 4.5% (w/v) HSA solutions were made: one containing MS-
325 at a concentration of 5.3 mM and one without MS-325. The
uncertainty in HSA concentration of these 4.5% (w/v) HSA solutions
was estimated to be (0.045% (w/v). An HSA molecular weight of
66 435 Da was used to convert % (w/v) to a molar concentration.
Ultrafiltration Measurements of Binding of MS-325 to HSA. MS-
325/HSA samples ranging from 0.051 to 5.3 mM MS-325 in 4.5%
(w/v) HSA were made by combining appropriate amounts of 4.5% (w/
v) HSA and 4.5% (w/v) HSA/5.3 mM MS-325.
Aliquots (400 µL) of these samples were placed in 5 kDa ultrafil-
tration units, incubated at 37 °C for 20 min, and then centrifuged at
3500g for 7 min. The filtrates from these ultrafiltration units were used
to determine the free concentration of MS-325 of each of the samples.
Duplicate aliquots were processed for each concentration sample of
MS-325 in 4.5% (w/v) HSA, with the exception of the lowest and
highest concentration samples, which were processed seven times.
Drug concentrations of MS-325/HSA samples and ultrafiltrates were
determined by measuring the Gd concentration using ICP-MS. Average
9
J. AM. CHEM. SOC. VOL. 124, NO. 12, 2002 3161

A R T I C L E S
Caravan et al.
relative uncertainties in the ICP-MS Gd concentrations (with a 95%
confidence level) of 2.8% and 4.3% were found for the original MS-
325/HSA samples and ultrafiltrates, respectively.
of fluorescent probe bound to HSA, [FP]bound. The concentration of
bound probe is also given by eq 11, where Kd is the apparent
app
[FP]bound ) {([HSA]t + [FP]t + Kadpp) -
Determination of Dissociation Constants for Fluorescent Probes.
All fluorescence measurements were performed in a Perkin-Elmer HTS-
7000+ fluorescence plate reader. The fluorescence of 2 µM solutions
of each of the fluorescent probes (in PBS buffer) was measured while
increasing the HSA concentration from 0 to 50 µM. Binding to HSA
caused an enhancement of fluorescence of each of the probes. Under
the wavelengths chosen, there was no fluorescence due to unbound
ligand. The fluorescence background due to HSA was determined by
measuring the fluorescence of 0-50 µM HSA (in PBS buffer), without
the presence of any fluorescent probe; this background was then
subtracted from fluorescence data in the HSA titration into 2 µM
fluorophore. The excitation and emission wavelengths used for the
dansyl molecules were 360 and 465 nm, respectively (obtained using
the standard narrow-band fluorescence filters supplied with the plate
reader). The excitation and emission wavelengths used for warfarin
were 320 and 380 nm, respectively (filters obtained from Corion Corp.,
Franklin, MA). The 320 nm filter was rated for fluorescence, with a
bandwidth of (10 nm. The 380 nm filter was rated for absorbance,
with a bandwidth of (6 nm.
([HSA] + [FP] + K app)2 - 4[HSA]t[FP]t}/2 (11)
x
t
t
d
[MS - 325]free
Kadpp ) Kd 1 +
(12)
(
)
Ki
dissociation constant of the probe in the presence of MS-325 and is
related to Kd by eq 12. Ki represents the site-specific dissociation
constant for MS-325. The [FP]bound measured from fluorescence was
compared to the [FP]bound calculated using eqs 11 and 12. Three
parameters were iteratively varied using the Solver function of Microsoft
Excel 98 (Microsoft, Redmond, WA): Ki, the initial fluorescence
((15%; this represented the scatter in the initial fluorescence and took
into account any changes in solution composition upon adding MS-
325), and the fluorescence in the absence of the probe, total displace-
ment ((15%). The program varied these three parameters to minimize
the sum of the squares of the residuals from measured and calculated
[FP]bound. The experiment was performed five times with dansyl
sarcosine, four times with dansyl-L-asparagine, and twice with warfarin.
Relaxivity. Relaxivities were determined at 20 MHz (0.47 T) and
64.5 MHz (1.5 T) using a Bruker NMS 120 Minispec and a modified
Varian XL-300, respectively. T1 was measured with an inversion
recovery pulse sequence, and all samples were measured at 37 °C. The
T1 and T2 NMRD profiles (37 °C) were recorded on a custom-built
relaxometer using saturation transfer and CPMG pulse sequences,
respectively.
The fluorescence data were fit using eq 10, where F represents the
P
2
F ) [(HSAt] + [FPt] + Kd) -
([HSA ] + [FP ] + K )2 - 4[HSA ][FP ] (10)
x
t
t
d
t
t
17O NMR. H217O chemical shifts and relaxation rates were deter-
mined for a PBS buffer solution in the presence and in the absence of
10 mmolal MS-325 as a function of temperature (0-100 °C) on a
Varian Unity 300 NMR operating at 40.6 MHz. Probe temperatures
were determined from ethylene glycol or methanol chemical shift
calibration curves. T1 was determined by inversion recovery and T2 by
a CPMG pulse sequence.
background-corrected fluorescence, [HSA]t and [FP]t are the total
concentrations of HSA and the fluorescent probe, respectively, Kd is
the dissociation constant of the fluorescent probe bound to HSA, and
P is the fluorescence of the fluorescent probe bound to HSA per unit
concentration. The parameters P and Kd were obtained by fitting the
fluorescence data using eq 10 and the program KaleidaGraph (version
3.08, Synergy Software, Reading, PA).
Determination of Site-Specific Binding Dissociation Constants
for MS-325. A solution containing [fluorescent probe] ∼ [HSA] ∼ Kd
and 400 µM MS-325 was prepared. Aliquots of this solution were 2-fold
diluted serially by addition of a solution which contained the same
concentrations of fluorescent probe and HSA but no MS-325. In all,
this produced eight samples of different concentrations of MS-325
(differing by a factor of 2), but all containing the same concentration
of fluorophore and HSA. The fluorescence of 100 µL aliquots of each
of these samples was measured in quadruplicate in 96 well plates.
Sixteen 100 µL aliquots of the solution containing no MS-325
(maximum fluorescent intensity) were also measured in the same 96
well plate, along with sixteen 100 µL aliquots of HSA in PBS
(minimum fluorescent intensity). These two controls set the dynamic
range for fluorescence change when MS-325 displaces a fluorescent
probe.
The program Scientist (ver. 2.0, Micromath, Salt Lake City, UT)
was used to fit the relaxation rate data and the binding data to the
stoichiometric model.
Acknowledgment. Prof. I. M. Klotz is thanked for an
insightful commentary on the binding studies. Drs. J. Ellison
and A. Kolodziej are kindly acknowledged for useful discus-
sions. W.D.H. acknowledges NSF Grant CHE-9705788 for
partial support. The analytical department at EPIX Medical, Inc.,
is thanked for numerous ICP determinations.
Supporting Information Available: Analysis of 1H and 17
O
relaxation rate data; Table S1 listing total, percent bound and
unbound concentrations of MS-325, and corresponding ratios
of bound MS-325:HSA:free MS-325; Table S2 listing reduced
transverse relaxation rates for MS-325 and [Gd(DTPA)(H2O)]2-
as a function of temperature (PDF). This material is available
free of charge via the Internet at http://pubs.acs.org.
The Ki of MS-325 was determined in the following way. The
fractional drop in fluorescence at a particular MS-325 concentration
was interpreted as a fractional displacement of the bound fluorescent
probe. Since the initial concentration of bound probe was known from
its Kd, the observed fluorescence could be converted to a concentration
JA017168K
9
3162 J. AM. CHEM. SOC. VOL. 124, NO. 12, 2002