Synthesis, characterization, and utility of thermoresponsive natural/unnatural product macroligands for affinity chromatography

Article abstract of DOI:10.1021/ol062045w

The synthesis and characterization of thermoresponsive, water-soluble poly-N-isopropyl acrylamide (PNIPAM) derived macroligands displaying cyclosporin A (CsA) and dexamethasone (Dex) for use as novel affinity resins are described. Characterization of thes

Full text of DOI:10.1021/ol062045w

ORGANIC
LETTERS
2006
Vol. 8, No. 23
5247-5250
Synthesis, Characterization, and Utility
of Thermoresponsive Natural/Unnatural
Product Macroligands for Affinity
Chromatography
Min Zhou,† Ananthapadmanab Sivaramakrishnan,† Krishan Ponnamperuma,†
Woon-Kai Low,§ Chunmei Li,† Jun O. Liu,§,‡ David E. Bergbreiter,† and
Daniel Romo*,†
Department of Chemistry, Texas A&M UniVersity, P.O. Box 30012,
College Station, Texas 77842-3012, and Department of Pharmacology and Molecular
Sciences and Department of Neuroscience, Johns Hopkins School of Medicine,
Baltimore, Maryland 21205
romo@mail.chem.tamu.edu
Received August 19, 2006
ABSTRACT
The synthesis and characterization of thermoresponsive, water-soluble poly-N-isopropyl acrylamide (PNIPAM) derived macroligands displaying
cyclosporin A (CsA) and dexamethasone (Dex) for use as novel affinity resins are described. Characterization of these soluble macroligands,
including ligand loading and integrity, was determined by 1H NMR spectroscopy. One of the CsA macroligands was used in a protein affinity
experiment to capture known binding proteins of CsA, the cyclophilins, from Jurkat T-cell lysates.
Organisms utilize biopolymers in the form of DNA for
information storage and in the form of proteins for cellular
structure and machinery. Most functional biopolymers re-
spond to external stimuli such as heat and pH in an all-or-
nothing or, at least, a highly nonlinear mode.1 Most common
among such biopolymers are proteins that precipitate or de-
nature on heating. Recently, progress has been made toward
the synthesis of functional polymers that mimic biopolymers
by responding in a desired way to external environmental
stimuli such as temperature, pH, and either electric or mag-
netic fields.2 These highly nonlinear responses of synthetic
polymers occur primarily in water but also in organic sol-
vents3 or polymer blends.4 When this stimulus-responsive
behavior occurs in aqueous solution, these polymers have
potential as useful tools for biotechnological and medicinal
research. Among the synthetic stimulus-responsive polymers,
temperature-responsive polymers have been the most exten-
sively studied.5 This kind of polymer is soluble below its
lower critical solution temperature (LCST) due to a combi-
nation of entropic and enthalpic effects. Above the LCST,
the polymer precipitates out of solution presumably due to
loss of hydration and exposure of hydrophobic chains.6,7
Copolymers of poly-N-isopropyl acrylamide (PNIPAM)
and derived functionalized polymers have been shown to be
thermoresponsive, exhibiting quantitative inverse tempera-
ture-dependent solubility in water.8,9 These and other ther-
moresponsive polymers have enabled the design of smart
(3) Osada, Y.; Gong, J. J. Prog. Polym. Sci. 1993, 18, 187.
(4) Sato, T.; Katayama, K.; Suzuki, T.; Shiomi, T. Polymer 1998, 39,
773.
(5) Galeav, I. Y.; Mattiasson, B. Trends Biotechnol. 1999, 17, 335.
(6) Schild, H. G. Prog. Polym. Sci. 1992, 17, 163. Idziak, I.; Avoce, D.;
Lessard, D.; Gravel, D.; Zhu, X. X. Macromolecules 1999, 32, 1260.
(7) Zhang, Y.; Furyk, S.; Bergbreiter, D. E.; Cremer, P. S. J. Am. Chem.
Soc. 2005, 127, 14505.
† Department of Chemistry, Texas A&M University.
‡ Department of Pharmacology and Molecular Sciences, Johns Hopkins
School of Medicine.
§ Department of Neuroscience, Johns Hopkins School of Medicine.
(1) Galaev, I. Y.; Mattiasson, B. TIBTECH 1999, 17, 335.
(2) Galaev, I. Y.; Gupta, M. N.; Mattiasson, B. CHEMTECH 1996, 26,
19.
10.1021/ol062045w CCC: $33.50
© 2006 American Chemical Society
Published on Web 10/18/2006

chemical catalysts having activities that can be switched off
and on by precipitation and redissolution based on reaction
temperature.8,9 In addition, these polymers are useful in
syntheses using biological catalysts.10 Most pertinent to the
described studies herein are previous uses of PNIPAM-
derived thermoresponsive polymers to purify known proteins.
An imidazole-PNIPAM thermoresponsive polymer loaded
with Cu(II) ions was used for the affinity purification of
proteins from cereals (Figure 1).11 PNIPAM coupled to an
ings, derived from coupling of a representative hydrophobic
natural product derivative, dexamethasone (Dex), and a repre-
sentative hydrophilic natural product, cyclosporin A (CsA),
to a copolymer of N-isopropyl acrylamide and N-(acroyloxy)-
succinimide (PNIPAM-NASI). In addition, a pull-down
experiment with the CsA macroligand demonstrates the
utility of these soluble macroligands in affinity experiments.
The synthesis of dexamethasone macroligands 4a-d
commenced with the known R-hydroxy acid derivative 217
that was coupled with mono-N-Boc-1,9-diaminoundecane
providing amide 3a (Scheme 1). Following Boc deprotection
Scheme 1. Synthesis of Dexamethasone Macroligands
Figure 1. PNIPAM-derived macroligands previously employed for
protein purification.
iminobiotin affinity ligand was used to isolate a lysozyme
from fetal calf sera.12 In connection with our interest in
coupling natural products to their cellular receptors,13 we
considered the attachment of natural products to PNIPAM
polymers to facilitate characterization of the immobilized
natural product.14
Affinity chromatography is based on the principle of
specific protein-ligand interactions and is a common tool
for purification of proteins, antibodies, and receptor-ligand
complexes.15 However, traditional affinity chromatography
techniques are plagued by nonspecific binding of proteins
due to physical inclusion as a result of the solid state of the
matrix and nonspecific interactions with the polymer back-
bone. In addition, it is typically not possible to assess the
integrity of the natural product ligand following coupling to
the insoluble matrix. In efforts to minimize these limitations,
we considered the use of the soluble polymer, PNIPAM, for
attachment of natural product ligands for isolation of their
cellular receptors. There are several possible advantages of
using PNIPAM. The parent polymer is soluble in aqueous
solution at 4 °C and has an LCST of 32 °C. The LCST can
be subtly increased or decreased by using more hydrophilic
or hydrophobic comonomers, respectively.16 This tactic could
be used for more sensitive proteins that are unable to
withstand higher LCSTs. Herein, we describe the synthesis
and properties of two macroligands, with various ligand load-
under standard conditions, the trifluoroacetate amine salt 3b
was coupled to the PNIPAM-NASI copolymer, available
by radical copolymerization.18 The remaining unreacted
N-acryloyloxy succinimide acid esters were capped by
addition of excess isopropyl amine to provide the macroli-
gands 4a-d with various ligand loadings, accomplished by
varying the stoichiometry of 3b (0.01, 0.04, 0.029, and
0.0075 equiv of 3b relative to the final N-iPr monomer,
respectively) added during the coupling step (Figure 2a). The
1
ratio of ligand to backbone monomer was verified by H
NMR analysis (Figure 2b).
The data in Figure 2 provide convincing evidence that
covalent attachment of dexamethasone occurred. First, the
water solubility of the polymer changes significantly.
Incorporation of larger amounts of the hydrophobic dexam-
ethasone lowered the water solubility of the product copoly-
mer as expected. If the dexamethasone had not been
incorporated as shown in Scheme 1, the NASI active esters
(8) Bergbreiter, D. E.; Zhang, L.; Mariagnanam, V. M. J. Am. Chem.
Soc. 1993, 115, 9295.
(9) Bergbreiter, D. E.; Caraway, J. W. J. Am. Chem. Soc. 1996, 118,
6092.
(10) Huang, X.; Witte, K. L.; Bergbreiter, D. E.; Wong, C.-H. AdV. Synth.
Catal. 2001, 343, 675.
(11) Kumar, A.; Galaev, I. Y.; Mattiasson, B. Bioseparation 1998, 7,
185.
(12) Garret-Flaudy, F.; Freitag, R. Biotechnol. Bioeng. 2000, 71, 223.
(13) Liu, J. O.; Low, W.-K.; Dang, Y.; Schneider-Poetsch, T.; Zanggao,
S.; Choi, N. S.; Romo, D.; Merrick, W. Mol. Cell 2005, 20, 709.
(14) (a) Affinity Chromatography: A Practical Approach; Dean, P. D.
G., Johnson, W. S., Middle, F. A., Eds.; IRL Press: Oxford, 1985. (b)
Scopes, R. K. Protein Purification: Principles and Practice, 3rd ed.;
Springer-Verlag: New York, 1994.
(16) (a) Taylor, L. D.; Cerankowski, L. D. J. Polym. Sci., Polym. Chem.
Ed. 1975, 13, 2551. (b) Mao, H.; Li, C.; Zhang, Y.; Furyk, S.; Cremer, P.
S.; Bergbreiter, D. E. Macromolecules 2004, 37, 1031.
(17) Lustenberger, P.; Formstecher, P.; Dautrevaux, M. J. Steroid
Biochem. Mol. Biol. 1981, 14, 697.
(18) Bergbreiter, D. E.; Hughes, R.; Besinaiz, J.; Li, C.; Osburn, P. L.
J. Am. Chem. Soc. 2003, 125, 8244.
(15) Cuatrecasas, P. J. Biol. Chem. 1970, 245, 3059.
5248
Org. Lett., Vol. 8, No. 23, 2006

Figure 3. Comparison of the width at half-height in 1H NMR (500
MHz, DMSO-d6) of Hb in (a) macroligand 4c (w1/2 ) 4.55 Hz) vs
(b) a 1:53 mixture of dexamethasone amine 3b and parent PNIPAM
(w1/2 ) 4.12 Hz).
As a representative macroligand bearing a hydrophilic
natural product, we synthesized microligands loaded with
varying amounts of CsA (Scheme 2). Using the strategy
Scheme 2. Synthesis of Cyclosporin A Macroligands
Figure 2. Water solubility data and 1H NMR (500 MHz, DMSO-
d6) spectra of dexamethasone polymer macroligands 4a-d. (a)
Structure and water solubility of dexamethasone macroligands
synthesized. (b) Expansion of 1H NMR (δ 3.5-6.5) spectra showing
the relative ratio of ligand to polymer based on relative integration
of Ha and Hb.
would have been converted to isopropyl acrylamides and the
water solubility of the polymer would have remained
unchanged. These experiments also show that acceptable
solubility of a macroligand can be achieved with 0.5-2.0
mol % of loaded polymer and also suggest that for
hydrophobic natural products such as dexamethasone a
loading of 1:30-50 may be ideal for precipitation of the
macroligand around 25 °C. In addition, it should be noted
that there is of course a limit for measurement of the ligand/
polymer ratio by 1H NMR integration at low ligand loadings
(see Figures 2b, 4a). Second, the amount of macroligand
incorporated as measured by the change in the integral ratio
for protons Ha and Hb in the polymer products 4a-d is
consistent with the equivalents of 3b used in the synthesis
of the macroligands. Finally, as was noted previously for
other PNIPAM derivatives with groups tethered by modest-
sized alkyl chains, there is a modest change in the 1H NMR
spectra for the immobilized amine. Figure 3 illustrates this,
reported by Diver, cross metathesis enabled coupling of the
olefin in CsA to olefin 720 to give modified CsA derivative
8 as a mixture of E/Z isomers.21 Simultaneous removal of
the Cbz group and reduction of the alkene under Lindlar
conditions gave the CsA derivative 9 bearing an amino linker
for subsequent attachment to the polymer. Varying amounts
of amine 9 were reacted with the PNIPAM-NASI copoly-
mer in the presence of triethylamine. To lower the LCST of
the resulting polymer due to the hydrophilic nature of CsA,
the unreacted NASI groups were capped with tert-butyl
amine, to give various CsA macroligands 10a-e. The degree
1
showing that the H NMR spectrum of the macroligand 4c
has an Hb signal that is 0.43 Hz broader relative to the same
proton in an admixture of dexamethasone salt 3b and
PNIPAM (Figure 3). Other procedures for analysis of the
macroligand were less successful. Previously, analysis by
thin-layer chromatography (TLC) could be used to demon-
strate covalent attachment of dyes to PNIPAM.19 However,
in the case of macroligands 4a-d, this analysis was frustrated
by streaking of the macroligand on the TLC plate.
1
of loading was again determined by H NMR spectroscopy
by comparing intensities of the N-CH3 of cyclosporin and
(19) Gonzalez, S. O.; Furyk, S.; Li, C.; Tichy, S. E.; Bergbreiter, D. E.;
Simanek, E. E. J. Polym. Sci., Part A: Polym. Chem. 2004, 42, 6309.
(20) This linker was synthesized in three steps from ω-undecenyl alcohol
in 44% yield. See Supporting Information for details.
Org. Lett., Vol. 8, No. 23, 2006
5249

Figure 5. Pull-down experiments to isolate cyclophilins from
Jurkat T-cell lysates using (a) CsA macroligand 10b and (b) CsA-
sepharose resin. Protein samples were mixed with equal volumes
of SDS-PAGE sample buffer, loaded, and run on 12% SDS-PAGE
gels which were silver stained. (a) Lane 1: denatured proteins
released from CsA-PNIPAM showing cyclophilin bands (18-20
kDa). Lane 2: competition experiment; same as lane 1 but in the
presence of 100 µM sanglifehrin A. (b) Lane 1: denatured proteins
released from CsA-sepharase showing cyclophilin bands (18-20
kDa). Lane 2: competition experiment; same as lane 1 but in the
presence of 100 µM sanglifehrin A.
1
Figure 4. Expansion of H NMR spectra (δ 3.0-4.4) showing
the relative ratio of ligand to polymer based on relative integration
of Ha of PNIPAM and N-CH3 of CsA in macroligands 10a-e.
b
Ha of PNIPAM in the polymer (Figure 4). In preparation
for affinity chromatography experiments, the solubility of
macroligand 10b in pH 7.4 lysis buffer was determined to
be ∼22 mg/mL and the LCST was ∼10 °C.
Finally, we tested the ability of the CsA macroligand to
retain and capture cyclophilins from Jurkat T-cell lysates.
In our initial studies, we considered the possibility that
binding of proteins to the macroligand may alter the solubility
properties of the macroligand-protein(s) complex. We also
considered the loading required to detect binding proteins
and settled initially on high loading of CsA to ensure
detection. A comparison was run simultaneously with a
CsA-sepharose resin that has been used previously to pull
down cyclophilins.21 CsA-PNIPAM 10b was incubated with
Jurkat T-cell lysates with or without sanglifehrin A (as a
positive control to compete with cyclophilin binding) at 4
°C in pH 7.4 lysis buffer (20 mM Tris‚HCl, 0.1 M KCl, and
0.2% Triton X-100) for 1 h. The CsA-PNIPAM was
precipitated by incubation at 32 °C (to ensure complete
precipitation) for 5 min. After recovering the polymer by
centrifugation, the pellets were washed by redissolving in
lysis buffer at 4 °C, and the precipitation cycle was repeated
once. Bound proteins in the pellets were solubilized in
denaturing buffer (pH 7.5, 20 mM Tris‚HCl, 2 mM DTT,
and 8 M urea), followed by heating with gel loading buffer
containing sodium dodecyl sulfate in a boiling water bath
for 5 min, and analyzed by sodium dodecyl sulfate poly-
acrylamide gel electrophoresis (SDS-PAGE) (Figure 5).
These polymers clearly retained the 18 kDa cyclophilin and
other less-abundant cyclophilins and did so to the exclusion
of other nonspecific binding proteins. However, we found
that the amount of protein captured is less than what might
be anticipated on the basis of initial protein concentration
in the cell lysates and on comparison with CsA-sepharose
under similar conditions (Figure 5). This may be due to the
precipitation/redissolution cycles which may also remove
some of the lower-abundance cyclophilins that may bind
initially but may be incompatible with the hydrophobic
interior of the precipitated polymers. These results suggest
avenues to explore for improving pull-down capacity.
In summary, we have synthesized thermoresponsive mac-
roligands from dexamethasone and cyclosporin A by cou-
pling to PNIPAM polymers. Characterization including lig-
and loading and integrity of the ligand was determined by
1
solution-phase H NMR spectroscopy, which highlights a
very useful aspect of these affinity resins. Successful pull-
down experiments with the CsA macroligand indicate that
nonspecific binding can indeed be minimized with these
soluble affinity resins; however, the amount of protein cap-
tured is less compared to that with the use of sepharose resin
with an equivalent CsA loading. These proof of principle
studies lay the groundwork for potential application of these
polymers for isolation of unknown natural product cellular
receptors and have provided general guidelines for optimal
ligand loading for both hydrophobic and hydrophilic ligands.
Acknowledgment. We thank the NIH (GM 52964-06,
D.R.; CA10021, J.O.L.), NSF (CHE-0446107, D.E.B.),
CIHR Fellowship (W.K.L), and Pfizer (D.R.) for support of
these investigations. We thank NSF (CHE-0077917) for
funds to purchase NMR instrumentation, Prof. Steve Diver
(SUNY Stony Brook) for a sample of sangliferin A, Linda
G. Cortina for technical assistance, and Neal Audeneart with
graphics assistance.
Supporting Information Available: General experimen-
tal procedures and characterization data for compounds 3a,b
and 7-9. Representative 1H NMR spectra for macroligands
4a-d and 10a-e and procedure for the affinity experiment.
This material is available free of charge via the Internet at
http://pubs.acs.org.
(21) Pan, F.; Liu, J. O.; Smulik, J. A.; Diver, S. T. Org. Lett. 2002, 4,
2051.
OL062045W
5250
Org. Lett., Vol. 8, No. 23, 2006