Using magnetic levitation to distinguish atomic-level differences in chemical composition of polymers, and to monitor chemical reactions on solid supports

Full text of DOI:10.1021/ja8074727

Published on Web 12/08/2008
Using Magnetic Levitation To Distinguish Atomic-Level Differences in
Chemical Composition of Polymers, and To Monitor Chemical Reactions on
Solid Supports
Katherine A. Mirica, Scott T. Phillips, Sergey S. Shevkoplyas, and George M. Whitesides*
Department of Chemistry and Chemical Biology, HarVard UniVersity, Cambridge, Massachusetts 02138
Received September 19, 2008; E-mail: gwhitesides@gmwgroup.harvard.edu
This communication describes a density-based method that uses
magnetic levitation to monitor chemical reactions on solid supports,
and to distinguish differences in chemical composition of polymers.
Solid-supported chemistry is widely used to prepare peptides,
nucleic acids, libraries of small molecules, and capturing reagents
for affinity purification and protein target identification.1-3 Solid-
phase chemistry has the inconvenient feature that there is no
inexpensive and rapid method for monitoring the progress of
reactions quantitatively on insoluble polymeric supports.4 The
methods currently available can be segregated into two categories:
(i) colorimetric tests for the presence or absence of certain functional
groups5,6 and (ii) instrumental techniques (e.g., infrared and
ultraviolet spectroscopy) that identify functional groups or provide
characteristic spectroscopic information (e.g., mass spectroscopy,
1H NMR using a magic angle spinning probe, and 13C NMR).1,3
Figure 1. Schematic representation of the apparatus used for monitoring
solid-supported reactions. A polymer bead levitates in this device at an
Colorimetric tests are rapid and provide qualitative information
about conversion but are subject to artifacts arising from competing
side reactions (false positives) and incomplete reactions (false
negatives).6 Spectroscopic instruments are substantially more
informative than colorimetric tests, but they are expensive (>$10,000),
usually time-consuming, and inconvenient to use, since they often
must be shared by multiple users. An enabling addition to solid-
supported chemistry, particularly for the development stages of a
solid-supported synthesis, would be a rapid and quantitative method
for following the progress of a reaction that does not require
specialized or expensive equipment. In essence, what is needed is
the procedural equivalent of thin-layer chromatography (TLC) for
solid-phase chemistry.
equilibrium position z0 when the magnetic (Fmag) and gravitational (Fg) forces
acting on the bead balance each other.
In this configuration of magnets, the “levitation height” z0
(m)sthe position along the z-axis at which the magnetic and
gravitational forces balance each other on the centerline of the z-axis
between the magnets for a bead with density Fssis given by eq 2,
where B0 (T) is the magnitude of the magnetic field at the surface
of the magnet in the center of the xy plane, and h (m) is the distance
between the magnets (see Supporting Information for details).
gµoh2
(ꢀs - ꢀm)4B0
Fmgµoh2
h
2
z0 )
Fs +
-
(2)
2
2
[
]
[
]
(ꢀs - ꢀm)4B0
This communication describes an inexpensive, rapid, and straight-
forward benchtop method that can be used to quantify the progress
and kinetics of a reaction on a solid support. The method is based
on the concept of magnetic levitation7 and involves levitating a
sample of beads (taken as an aliquot from a reaction mixture) in a
paramagnetic solution (e.g., GdCl3 dissolved in H2O or N,N-
dimethylformamide (DMF)) between two 5 cm × 5 cm × 2.5 cm
NdFeB magnets oriented in the anti-Helmholtz configuration (Figure
1).8
Polymer beads levitate in the environment depicted in Figure 1
when the gravitational (Fg) and magnetic forces (Fmag) acting on
the beads balance one another (eq 1). In eq 1, Fm is the density of
the paramagnetic medium (kg·m-3), Fs is the density of the
suspended particle (kg·m-3), V is the volume of the particle (m3),
g is the acceleration due to gravity (m ·s-2), ꢀm and ꢀs are the
magnetic susceptibilities (unitless) of the paramagnetic medium and
the suspended particle respectively, µ0 is the magnetic permeability
of free space (T ·m·A-1), and B is the applied magnetic field (T).
We have previously applied magnetic levitation for density-based
separation of Merrifield resin beads that differed in the level of
chlorine-containing functionality per bead and for detection of the
binding of streptavidin to resin-bound biotin.8 Here we show that
(i) covalent modification of polymeric beads sufficiently alters the
density of those beads to produce easily measurable changes in z0
and (ii) changes in density (and z0) correlate with the progress and
kinetics of a chemical reaction on a solid support.
The ability to resolve differences in the chemical composition
of polymeric beads by levitation is evident in Figure 2. This figure
shows the correlation of z0 with Fs for 10 derivatives of 4-benzy-
loxybenzaldehyde polystyrene (diameter ) 35-75 µm, loading level
) 3.5 mmol -CHO/g resin, ∼350 pmol -CHO/bead). We prepared
these derivatives using reductive amination reactions (eq 3):
(ꢀm - ꢀs)
b
We used 10 equiv of amine and NaBH3(CN) in each reaction
(dissolved in 5% CH3COOH-DMF) and agitated the beads for 24 h
b
b
b
b
Fg + Fmag ) (Fs - Fm)Vbg +
V(B· )B ) 0
(1)
µ0
9
17678 J. AM. CHEM. SOC. 2008, 130, 17678–17680
10.1021/ja8074727 CCC: $40.75
2008 American Chemical Society

C O M M U N I C A T I O N S
The method also is sensitive to changes in chemical composition
(and, hence, density) of a polymer during the course of a chemical
reaction (Figure 3). We demonstrate this sensitivity by monitoring
the condensation reaction of 2,5-diiodobenzoic acid to leucine-
derivatized Wang polystyrene (diameter ) 75-150 µm, 1.8 mmol
-NH2/g resin, ∼1 nmol -NH2/bead) at 0 °C using 5 equiv of
O-benzotriazole-N,N,N′,N′-tetramethyl-uronium-hexafluoro-phos-
phate (HBTU) and 5 equiv of diisopropylethyl amine (DIEA) in
DMF (Figure 3). We withdrew aliquots of beads from the reaction
mixture at different times throughout the course of the reaction
and washed the beads immediately to remove excess reagents. We
suspended ∼100 beads from each aliquot in a cuvette containing
650 mM GdCl3 in DMF, placed the cuvette between the NdFeB
magnets, and waited ∼15 min for the beads to reach an equilibrium
levitation height.
Figure 3. Schematic describing the process for monitoring conversion for
a condensation reaction of a solid-supported amine with a carboxylic acid.
Figure 4A demonstrates the ease with which levitation can be
used to monitor solid-phase reactions. The levitation height of the
beads decreases as their density increases until the reaction reaches
completion. The beads formed tight clusters at the beginning and
end of the reaction but displayed increased dispersion when the
reaction approached 50% completion. The polymer beads are
polydisperse in size (they vary in diameter between 75 and 150
µm) but are equal in density. During the course of the reaction,
however, this polydispersity leads to variations in accessibility of
reagents to the interior of the beads (∼99% of the amines are on
the interior of the beads)10 and results in slight differences in
chemical composition (and density) between beads.1,11 Once the
reaction reaches completion and all of the available amines react,
the chemical composition of the beads becomes uniform. Figure
4B correlates z0 and Fs (measured independently using the sink-
float technique) as the reaction progresses for each of the data points
summarized in Figure 4A.
The conversion of starting material to product measured by
magnetic levitation matches the conversion measured by 1H NMR
within the 95% confidence interval. For measurements of conversion
by levitation, we assumed that the value of z0 for the beads reflected
the mole fraction of starting material, and we used z0 for each set
of beads to calculate the concentration of unreacted amine at
different time points during the course of the reaction using eq 4:
Figure 2. (A) Graph of z0 versus Fs for 10 different derivatives of the
4-benzyloxybenzaldehyde polystyrene beads levitated in a 650 mM aqueous
solution of GdCl3 at 23 °C. The linear least-squares fit is described by the
following equation: y ) -221x + 276; R2 ) 0.984. (B) Chemical structures
for each of the data in (A).
to ensure complete conversion. We washed the beads (3 × 3 min
each of DMF, CH2Cl2, CH3OH, and H2O) to remove excess reagent
from the polymer and stained aliquots of beads from each reaction
with 2,4-dinitrophenylhydrazine to confirm complete consumption
of the aldehyde.9
We measured z0 by suspending ∼100 beads from each reaction
in a cuvette containing 650 mM GdCl3 in water and positioning
the cuvette along the central axis between the magnets. The
concentration of GdCl3 was chosen empirically to adjust the density
of the medium approximately to the density of the beads. Initially
the beads were dispersed in solution, but after ∼5 min they began
to coalesce to form a cloud in one position between the magnets.
After ∼15 min the cloud of beads had coalesced to form a tight
cluster; we took the center point of this cluster to be z0, which we
measured using a ruler with mm-scale markings.
Each bead in this experiment contained ∼350 pmol of small
molecule. A difference in chemical composition of a single atom
between the small molecules led to measurable differences in z0
(and Fs) between the corresponding beads (e.g., the differences in
z0 between derivatives b, g, and j shown in Figure 2B reflect
differences of one fluorine atom). A plot of z0 versus Fs for each
set of beads reveals a linear relationship between the density of a
polymer and its equilibrium levitation height (Figure 2A), as
expected from eq 2. We measured Fs for each polymer using
sink-float techniques at various concentrations of CaCl2 in water
(see Supporting Information); these measurements are tedious and
impractical as a method of following chemical reactions.
[-NH2]experimental
)
zexperimental - z100%conversion
[-NH2]0%conversion (4)
(
)
z0%conversion - z100%conversion
1
For H NMR experiments, we cleaved the product and starting
material from the polymer (using trifluoroacetic acid) for each
sample of beads collected from the reaction. We obtained the ratio
1
of product to starting material by integration of H NMR spectra
(we integrated leucine R-hydrogens, which are resolved between
product and starting material by 0.71 ppm in the 1H NMR
spectrum).
9
J. AM. CHEM. SOC. VOL. 130, NO. 52, 2008 17679

C O M M U N I C A T I O N S
yield similar rates for the pseudo-first-order reaction (T1/2 ) 23 (
4 min (NMR) and T1/2 ) 18 ( 2 min (levitation)) (Figure 4C).
We conclude that magnetic levitation provides a sensitive density-
based approach for distinguishing relative differences in chemical
functionality on polymeric beads and for monitoring the progress
and kinetics of solid-phase chemical reactions. This technique has
the following useful characteristics: (i) it is exceedingly simple (the
levitation height can be measured easily by eye and quantified using
a ruler); (ii) it is inexpensive (5 cm × 5 cm × 2.5 cm NdFeB
magnets cost $5 each, and GdCl3 costs $0.39 per gram);12 (iii) it is
rapid (measurements require 15 min); (iv) it requires only a small
amount of sample (a single bead in a capillary tube works as well
as groups of beads in a cuvette); (v) and it does not destroy the
sample. The speed and ease of this method are reminiscent of thin-
layer chromatography (TLC) for solution-phase chemistry. The
method, however, neither provides information about the chemical
composition of the sample nor reveals potential byproducts formed
during the reaction.
We believe this technique will find broad applications as a rapid
benchtop tool for monitoring and analyzing chemical and biochemi-
cal transformations on solid supports; in-depth information about
the exact chemical composition of a polymer bead is best obtained
with more accurate (but more expensive) techniques such as NMR
spectroscopy. Studies focusing on the limitations and sensitivity
of this technique are underway.
Acknowledgment. This work was supported by the NIH
(ES0166665), a predoctoral fellowship from Eli Lilly (K.A.M.),
and postdoctoral fellowships from the Damon Runyon Cancer
Research Foundation (S.T.P.) and NIH (S.T.P.).
Supporting Information Available: Additional experimental in-
formation. This material is available free of charge via the Internet at
http://pubs.acs.org.
References
(1) Handbook of Combinatorial Chemistry; Nicolau, K. C., Hanko, R., Hartwig,
W., Eds.; Wiley-VCH: Weinheim, 2002.
(2) (a) Beaucage, S. L.; Iyer, R. P. Tetrahedron 1992, 48, 2223. (b) Engels,
J. W.; Uhlmann, E. Angew. Chem., Int. Ed. Engl. 1989, 28, 716. (c) Lam,
K. S.; Lebl, M.; Krchnak, V. Chem. ReV. 1997, 97, 411. (d) Meldal, M.
Curr. Opin. Chem. Biol. 2004, 8, 238; Verlander, M. Int. J. Pept. Res.
Ther. 2007 13, 75. (e) Yingyongnarongkul, B. E.; How, S. E.; Diaz-Mochon,
J. J.; Muzerelle, M.; Bradley, M. Comb. Chem. High Throughput Screening
2003, 6, 577.
(3) Dorwald, F. Z. Organic Synthesis on Solid Phase: Supports, Linkers,
Reactions, 2 ed.; Wiley-VCH: 2002.
Figure 4. (A) Photographs of clusters of levitating polymer beads (leucine-
derivatized Wang polystyrene (diameter ) 75-150 µm, 1.8 mmol -NH2/g
resin, ∼1 nmol -NH2/bead) (∼100 beads/cluster) taken at different times
throughout the course of the reaction shown in Figure 3. We levitated the
beads in 650 mM GdCl3 in DMF at 23 °C. (B) Plot showing the correlation
of density of the polymer beads with their levitation height; each data point
corresponds to a sample of beads shown in (A). The error bars represent
the standard deviation from three independent measurements. (C) Pseudo-
first-order kinetics plots showing the rate of consumption of polymer-bound
amine determined by 1H NMR (O) and by levitation (b) from three
independent measurements. The data were fit with linear least-squares lines:
(i) levitation (s): y ) -0.038x + 0.730, R2 ) 0.992; and (ii) 1H NMR
(----): y ) -0.031x + 0.583, R2 ) 0.993.
(4) Cironi, P.; Alvarez, M.; Albericio, F. QSAR Comb. Sci. 2004, 23, 61.
(5) Gaggini, F.; Porcheddu, A.; Reginato, G.; Rodriquez, M.; Taddei, M.
J. Comb. Chem. 2004, 6, 805.
(6) Vazquez, J. V.; Qushair, G.; Albericio, F. Methods Enzymol. 2003, 369,
21.
(7) (a) Beaugnon, E.; Tournier, R. Nature 1991, 349, 470. (b) Catherall, A. T.;
Eaves, L.; King, P. J.; Booth, S. R. Nature 2003, 422, 579. (c) Hirota, N.;
Kurashige, M.; Iwasaka, M.; Ikehata, M.; Uetake, H.; Takayama, T.;
Nakamura, H.; Ikezoe, Y.; Ueno, S.; Kitazawa, K. Physica B 2004, 346,
267. (d) Ikezoe, Y.; Hirota, N.; Nakagawa, J.; Kitazawa, K. Nature 1998,
393, 749. (e) Ikezoe, Y.; Kaihatsu, T.; Sakae, S.; Uetake, H.; Hirota, N.;
Kitazawa, K. Energy ConV. Manag. 2002, 43, 417.
(8) Winkleman, A.; Perez-Castillejos, R.; Gudiksen, K. L.; Phillips, S. T.;
Prentiss, M.; Whitesides, G. M. Anal. Chem. 2007, 79, 6542.
(9) Shannon, S. K.; Barany, G. J. Comb. Chem. 2004, 6, 165.
(10) Li, W. B.; Yan, B. J. Org. Chem. 1998, 63, 4092.
(11) Kress, J.; Rose, A.; Frey, J. G.; Brocklesby, W. S.; Ladlow, M.; Mellor,
G. W.; Bradley, M. Chem.sEur. J. 2001, 7, 3880.
(12) These values are for bulk pricing; individual 5 cm × 5 cm × 2.5 cm NdFeB
magnets are available for $15-20 at www.magnet4less.com, and
GdCl3 · 6H2O costs $2 per gram from Sigma-Aldrich.
Since the reagents for the reaction were used in 5-fold excess
relative to the quantity of polymer-bound -NH2, the reaction
followed pseudo-first-order kinetics. Figure 4C gives the data from
three independent reactions. Both magnetic levitation and NMR
JA8074727
9
17680 J. AM. CHEM. SOC. VOL. 130, NO. 52, 2008