Development of chemically stable solid phases for the target isolation with reduced nonspecific binding proteins

Article abstract of DOI:10.1016/j.bmcl.2005.09.011

Poly(methacrylate) matrices for affinity resins were designed and synthesized based on our previous results that nonspecific protein absorption on affinity resins strongly depended on their hydrophobic property. The novel affinity resins bearing FK506 (6a, 6b) captured specific binding protein, FKBP12, with a small amount of nonspecific binding proteins. The amount of nonspecific binding proteins on 6a-6b was much reduced compared to that on commercially available poly(methacrylate) resins, Toyopearl (8), and was almost the same as that on one of the most popular resins, Affigel (9). Interestingly, 6a and 6b could isolate FKBP52 as a specific binding protein as well, although 8 and 9 could not.

Full text of DOI:10.1016/j.bmcl.2005.09.011

Bioorganic & Medicinal Chemistry Letters 16 (2006) 447–450
Development of chemically stable solid phases for the target
isolation with reduced nonspecific binding proteins
Teruki Takahashi,a Takaaki Shiyama,a Ken Hosoyab and Akito Tanakaa,*
aChemistry Department, Reverse Proteomics Research Institute Co., Ltd., 2-6-7 Kazusa-Kamatari, Chiba 292-0818, Japan
bDepartment of Polymer Science and Engineering, Kyoto Institute of Technology, Matsugasaki, Sakyo-ku, Kyoto 606-8585, Japan
Received 28 June 2005; revised 27 August 2005; accepted 2 September 2005
Available online 14 November 2005
Abstract—Poly(methacrylate) matrices for affinity resins were designed and synthesized based on our previous results that nonspe-
cific protein absorption on affinity resins strongly depended on their hydrophobic property. The novel affinity resins bearing FK506
(6a, 6b) captured specific binding protein, FKBP12, with a small amount of nonspecific binding proteins. The amount of nonspecific
binding proteins on 6a–6b was much reduced compared to that on commercially available poly(methacrylate) resins, ToyopearlTM
(8), and was almost the same as that on one of the most popular resins, AffigelTM (9). Interestingly, 6a and 6b could isolate FKBP52
as a specific binding protein as well, although 8 and 9 could not.
Ó 2005 Elsevier Ltd. All rights reserved.
Affinity resins bearing bioactive compounds such as nat-
ural products, drugs, and toxins play an important role
in the discovery of novel drug targets and the elucida-
tion of drug mechanisms.1,2 The successful isolation of
target proteins by affinity chromatography depends
on the synthesis of polymeric resins that can bind to
the cellular target with maximum selectivity and efficien-
cy. Nonspecific binding of cellular proteins to affinity
matrices is therefore a significant limitation to this
approach.3,4
these methacrylate polymers bearing bioactive com-
pounds often show high levels of nonspecific binding
protein with the target protein.3 Therefore, development
of novel solid phases that are chemically stable as
poly(methacrylate) derivatives and are hydrophilic en-
ough for elimination of the nonspecific absorption as
agarose ones is an important goal.
The phrase Ônonspecific protein bindingÕ is usually used
to represent proteins that bind to affinity resins based
on physical adsorption.7,8 In a previous paper,3 we
reported a linear relationship between hydrophobicity
of ligands and the amount of nonspecific binding pro-
teins, and effectiveness of introduction of hydrophilic
spacers for reduction of those nonspecific binding pro-
teins. However, it was difficult to eliminate such nonspe-
cific binding proteins completely by introduction of
them. We thought that it was difficult entirely to cover
the hydrophobic surface by the introduction of hydro-
philic spacers because those spacers were not always
spatially close to the hydrophobic surfaces. Thus, we
next attempted to introduce those hydrophilic spacers
into monomers before polymerization reactions for
improvement in reducing the nonspecific binding pro-
teins because it could be effective to construct more
hydrophilic surfaces on the solid phases comparing with
introduction of hydrophilic spacers after polymerization
reactions. We herein report the design and synthesis of
novel methacrylate derivative bearing hydrophilic spac-
ers and its polymers for the target isolation studies, and
There are now some commercially available solid mate-
rials for preparation of affinity resins. AffigelTM, agarose
derivatives,5 is one of the most popular matrices for this
purpose. In our previous study, it was shown that the
amount of nonspecific protein absorptions to AffigelTM
was very small, which could be due to its high hydro-
philic property.3 However, AffigelTM is not suitable solid
phases for organic synthesis, because it is easily dena-
tured under organic synthesis conditions.3 Thus, chemi-
cal approaches using this resin are often limited. In
contrast, ToyopearlTM, a poly(methacrylate) derivative,6
is stable under most synthetic conditions, which allow
the synthesis of more effective affinity resins. However,
Keywords: Nonspecific binding proteins; Affinity chromatography;
Poly(methacrylate); FK506; FKBP12; FKBP52.
*
Corresponding author. Tel.: +81 438 52 3990; fax: +81 438 52
3986; e-mail: tanaka-a@reprori.jp
0960-894X/$ - see front matter Ó 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2005.09.011

448
T. Takahashi et al. / Bioorg. Med. Chem. Lett. 16 (2006) 447–450
their effectiveness in reduction of nonspecific binding
proteins maintaining the specific interactions.
on an X-ray structure of FK506 and FKBP12 that
exhibited a hydroxyl group at the 32 position was not in-
volved in the interaction.3 FK506-affinity resins using
the commercially available matrices, ToyopearlTM (8)
and AffigelTM (9), were also prepared.
We designed methacrylate monomers bearing poly(eth-
yleneglycol) (3a) and D-glucamine (3b) moieties because
introductions of those moieties to ToyopearlTM as hydro-
philic spacers resulted in much reduction of the nonspe-
cific binding proteins maintaining binding with the
specific binding protein in our previous studies.3,4 These
novel monomers were synthesized as shown in Scheme
1.9 Replacement of a hydroxyl group of hexaethylene
glycol (1a) with an amino group was carried out via a
phthalimide derivative, followed by introduction of a
tert-butoxycarbonyl (Boc) group on primary amine to
give 2a. A Boc group was introduced on amino group
of D-glucamine (1b) which afforded 2b. 2a and 2b were
coupled with methacryloyl chloride in the presence of
N,N,N0, N0-tetramethyl-1,3-diaminopropane to afford
the methacrylate monomers 3a and 3b, respectively.
These monomers were polymerized by a solution poly-
merization method with 2% glycerol dimethacrylate, a
well-known crosslinker reagent, in dioxane using azobis-
isopropyronitrile (AIBN) as an initiator, which gave N-
Boc protected solid phases (4a–4b). AIBN was pur-
chased from Wako Pure Chemical and purified through
re-crystallization technique. After grinding by a mortar
grinder, Boc groups were removed by treatment with tri-
fluoroacetic acid (TFA) to give desired solid phases 5a
and 5b. A FK506 derivative with a linker moiety (7)
was immobilized on 5a and 5b to give 6a and 6b, respec-
tively. FK506, an immunosuppressive drug, was specifi-
cally bound to FKBP12, a FK506 binding protein
12 kDa, with a Kd of 0.4 nM,9 and was often used as
standard compound for assessment of affinity chroma-
tography.3,4,10 The design of 7 was performed based
To assess capacity of the FK506-affinity resins (6a, 6b, 8,
and 9) to capture the specific binding protein, FKBP12,
specifically, these resins were mixed with lysate obtained
from rat brain, and proteins bound on them were com-
paratively analyzed.11 This lysate was known to include
FKBP12 with other miscellaneous proteins.3,4 After
extensive washes with lysate buffer, binding proteins
were completely eluted by SDS sample buffer solution
and analyzed by SDS–polyacrylamide gel electrophore-
sis (Fig. 1). As shown in Figure 1, all FK506 affinity res-
ins successfully captured FKBP12. However, a large
amount of nonspecific binding proteins such as tubulin
(Fig. 1A), actin (panel B), and glyceraldehyde-3-phos-
phate dehydrogenase (panel C) were found on FK506-
Toyopearl resins (8). On the other hand, these nonspe-
cific binding proteins were little observed on 6a and
6b, while these resins consisted of the same poly(methac-
rylate) polymers as well as 8. The amount of nonspecific
binding proteins was almost the same as that on the
hydrophilic polymer, AffigelTM (9). These results support-
ed the speculations that the amounts of nonspecific
binding proteins strongly depends on hydrophobicity
of affinity resins. Specificity of binding proteins to
FK506-affinity resins was confirmed by the competition
method.11 Interestingly, another specific binding protein
(D, Fig. 1A) was observed on 6a and 6b. Analysis by
LC–MS/MS ions search method using an ESI ion trap
mass spectrometry (Thermoelectron, LTQ) after in-gel
digestion exhibited that this band was another FK506
R1
R1=A (1a)
HO
CH2OH
a, b, c
A=
B=
-(CH2CH2O)5-CH2-
OHOH
e
O R1CH2NHBoc
R1
R1=A (2a)
HO
CH2NHBoc
O
R1=A (3a)
=B (3b)
HO R1CH2NH2
R1=B (1b)
d
OHOH
=B (2b)
h
R1
f
g
R1
R1
O
O
CH2NH-CO-FK506
O
O
CH2NH2
O
O
CH2NHBoc
R1=A (6a)
=B (6b)
R1=A (5a)
=B (5b)
R1=A (4a)
=B (4b)
32
R2
MeO
O
R2
h
Me
NHCO-FK506
NH2
Me
N
O
H
FK506:
7:
solid phase=ToyopearlTM (8)
=AffigelTM (9)
O OH
O
HOOC-(CH2)6CO-
Me
Me
O
O
Me
OH
O
OMe
OMe
Scheme 1. Synthesis of affinity resins bearing FK506.13 Reagents and conditions: (a) phthalimide, triphenylphosphine, and diisopropyl
azodicarboxylate/THF; (b) NH2NH2/MeOH; (c) Boc2O, K2CO3/MeOH; (d) Boc2O/MeOH; (e) methacryloyl chloride, N,N,N0,N0-tetramethyl-1,3-
diaminopropane/acetonitrile; (f) AIBN, 2% glycerol dimethacrylate/dioxane, 100 °C; (g) TFA/CH2Cl2/dioxane/H2O = 50:39:10:1; (h) 7, EDC HCl,
HOBt/NMP.

T. Takahashi et al. / Bioorg. Med. Chem. Lett. 16 (2006) 447–450
9 6b 6a
449
A
8
(+)
(-)
(+)
(-)
(+)
(-)
(+)
(-)
competition
116
97.4
D (FKBP52)
66.3
55.4
A
B
C
36.5
31
FKBP12
21.5
14.4
6.0
CBB stained
B
8
9
6b
6a
FKBP52
competition
(-)
(+)
(-)
(+)
(-)
(+)
(-)
(+)
55.4
silver stained
8
6b
C
(+)
competition
(-)
(+)
(-)
anti-FKBP52 antibodies
Figure 1. Binding proteins on affinity matrices bearing FK506 (6a, 6b, 8, and 9). (A) Each resin (10 lL, 1 lmol FK506 on resins) was mixed with
1 mL of lysate obtained from rat brain at 4 °C for 1 h.12 After separation of the resins by centrifugation and extensive washes with lysate buffer
(25 mM Tris–HCl, pH 7.4, 0.25 M sucrose, 0.3 mM N,N-diethylthiocarbamate, 2 mM CaCl2, and 2 mM MgCl2), binding proteins were completely
eluted by SDS sample buffer solution and analyzed by SDS–polyacrylamide gel electrophoresis (stained by Coomassie brilliant blue (CBB)). FKBP12
and FKBP52 (D) are known as specific binding proteins to FK506. Other proteins such as tubulin (A), actin (B), and glyceraldehyde-3-phosphate
dehydrogenase (C) are known to be nonspecific binding proteins. Identification of proteins was performed by MS/MS ions search method using an
ESI ion trap mass spectrometry (Thermoelectron, LTQ) after in-gel digestion, respectively. A commercially available apparatus (Dainippon Seiki
Co., Ltd., cat. code: 1D-SDS) was used for cutting off each strip including the desired protein from CBB or silver-stained SDS-gels. One micromole of
FK506 was added for the competition experiment. (B) FKBP52 was successfully isolated by 6a and 6b, not by 8 and 9. The same SDS gel to the above
was stained by silver after decolorization of CBB stain. (C) Western blot studies using anti-FKBP52 antibodies on 8 and 6a.14
specific binding protein, FK506-binding protein 52
(FKBP52, MW = 51.7 kDa).12 FKBP52 could not be
detected on AffigelTM bearing the same ligand (9). The
reason for this difference between our resins (6a–6b)
and agarose ones (9) is not clear now. FKBP52 was
not detected on ToyopearlTM resins bearing FK506 (8)
(Fig. 1B) because of the existence of large amounts of
nonspecific binding proteins such as tubulin while
FKBP52 was detected on it by Western blot analysis
on FKBP52 (Fig. 1C). These results demonstrated the
importance of reduction of the nonspecific binding pro-
teins for identifications target finding.
resins, isolation studies using other ligands are in process
and those results are going to be reported. Polymeriza-
tion of other matrices using the monomers (3a–3b) by
other methods is in progress because 6a and 6b were
powdered by a mortar grinder after a solution polymer-
ization and had heterogeneous distributions of particles
(Fig. S1). Results on these novel solid phases for target
identification will be reported in the future.
Acknowledgments
We thank to Dr. Masayuki Haramura and Mr. Mikio
Takeuchi (Reverse Proteomics Research Institute Co.,
Ltd.) for their critical reading of the manuscript. The
financial support from NEDO (The New Energy and
Industrial Technology Development Organization) is
gratefully acknowledged.
Since 6a and 6b are poly(methacrylate) polymers, they
are thought to be stable in several synthetic conditions
in contrast to AffigelTM and other sugar polymers. For
example, the poly(methacrylate) polymers (4–6) were
stable under the synthetic conditions studied in this
work. Among them, the stability of 4b, 5b, and 6b was
notable because these polymers had a sugar-like struc-
ture that was similar to that of AffigelTM.
Supplementary data
Supplementary data associated with this article can be
found, in the online version, at doi:10.1016/j.bmcl.2005.
09.011.
These results demonstrated the effectiveness of our novel
solid phases for affinity resins. To assess generality of the

450
T. Takahashi et al. / Bioorg. Med. Chem. Lett. 16 (2006) 447–450
N,N,N0,N0-tetramethyl-1,3-diaminopropane
(0.99 mL,
References and notes
5.9 mmol), and acetonitrile (50 mL), and was stirred at rt
for 2 h. The reaction was poured into a mixture of a
saturated aqueous solution of sodium bicarbonate and
CHCl3. The organic layer was washed with brine, dried
over Na2SO4, and filtered. The filtrate was evaporated in
vacuo and was purified by column chromatography (silica
gel 50 g, eluted with 50% ethyl acetate (EA)/n-hexane, 67%
1. Harding, M. W.; Galat, A.; Uehling, D. E.; Schreiber, S.
L. Nature 1989, 341, 758.
2. Taunton, J.; Hassig, C. A.; Schreiber, S. L. Science 1996,
272, 408.
3. Tamura, T.; Terada, T.; Tanaka, A. Bioconjugate Chem.
2003, 14, 1222.
4. Shiyama, T.; Furuya, M.; Yamazaki, A.; Terada, T.;
Tanaka, A. Bioorg. Med. Chem. 2004, 12, 2831.
5. http://www.bio-rad.com/.;
6. http://www.tosoh.com/EnglishHomePage/tchome.htm.;
7. Hofstee, B. H. J. Adv. Exp. Med. Biol. 1974, 42, 43.
8. Hofstee, B. H. J. Biochem. Biophys. Res. Commun. 1973,
53, 1137.
9. Siekierka, J. J.; Hung, S. H.; Poe, M.; Lin, C. S.; Sigal, N.
H. Nature 1989, 341, 755.
10. Shimizu, N.; Sugimoto, K.; Tang, J.; Nishi, T.; Sato, I.;
Hiramoto, M.; Aizawa, S.; Hatakeyama, M.; Ohba, R.;
Hatori, H.; Yoshikawa, T.; Suzuki, F.; Oomori, A.;
Tanaka, H.; Kawaguchi, H.; Watanabe, H.; Handa, H.
Nat. Biotechnol. 2000, 18, 877.
11. The preparation of lysate, binding experimental for
capturing the binding proteins, and analysis of proteins
were carried out in similar ways to that described in Ref. 3.
12. Peattie, D. A.; Harding, M. W.; Fleming, M. A.; DeCen-
zo, M. T.; Lippke, J. A.; Livingston, D. J.; Benasutti, M.
Proc. Natl. Acad. Sci. USA 1992, 89, 10974.
13. A 40% toluene solution of diisopropyldiazocarboxylate
(93.5 mL, 0.19 mmol) was added to a mixture of hexaeth-
ylene glycol (1a, 50 g, 0.18 mmol), phthalimide (29.1 g,
0.2 mmol), triphenylphosphine (52.4 g, 0.2 mmol), and
tetrahydrofuran (THF, 500 mL) over 15 min at 0 °C,
and was stirred at room temperature (rt) for 15 h. After
evaporation in vacuo, the resulting residue was dissolved
with a mixture of chloroform (CHCl3) and water. The
separated organic layer was dried over Na2SO4. After
filtration, the filtrate was evaporated in vacuo to give
crude mono-phthalimide derivative, which was used for
the next reaction without further purification. This crude
mono-phthalimide derivative was mixed with hydrazine
(17 mL, 0.36 mmol) and methanol (MeOH, 300 mL), and
was heated at reflux for 3 h. After evaporation in vacuo,
the resulting residue was dissolved by chloroform. After
removal of insoluble matters by filtration, the filtrate was
dried over Na2SO4. After filtration, the filtrate was
evaporated in vacuo. The residue was applied to the
column chromatography (silica gel, 200 g), and was eluted
with CHCl3 and 50% MeOH/CHCl3. Fractions including
the objective were collected and evaporated in vacuo, to
give a mono-amino derivative (21.4 g) A mixture of the
mono-amino derivative (20.7 g, 73.5 mmol), di-tert-butyl
dicarbonate (16.1 g, 73.5 mmol), potassium carbonate
(20.3 g, 147 mmol), and MeOH (50 mL) was stirred at rt.
After 15 h, the reaction was poured into a mixture of a
saturated aqueous sodium bicarbonate and CHCl3. The
organic layer was washed with saturated sodium chloride
(300 mL) and dried over Na2SO4. After filtration, the
filtrate was evaporated in vacuo to afford 2a (28 g, 42%).
1H NMR (CDCl3) d : 1.44 (9H, s, tBu), 3.32 (2H, m, –
NHCH2CH2), 3.53–3.74 (22H, m, OCH2CH2O–), 5.22
(1H, br s, NH). ESI-MS (m/z): 382.2 (382.24 Calcd for
C17H36NO8, M++H). Methacryloyl chroride (0.91 mL,
9.4 mmol) was added to a mixture of 2a (1.5 g, 3.9 mmol),
1
EA/n-hexane, and EA) to give 3a (0.9 g, 51%). H NMR
(CDCl3) d: 1.44 (9H, s, tBu), 1.95 (3H, dd, J = 1.3, 6.6 Hz,
CH2@C(CH3)CO–), 3.31 (2H, m,–NHCH2CH2), 3.53–
3.78 (20H, m, OCH2CH2O–), 4.30 (2H, m, –COOCH2-
CH2O–), 5.05 (1H, br s, NH), 5.57 (1H, m,
CH2@C(CH3)CO–), 6.13 (1H, m, CH2@C(CH3)CO–).
Anal. Calcd for C21H39NO9/0. 4H2O: C, 55.2; N, 3.07;
O, 32.9. Found: C, 55.2; N, 3.03; O, 33.1. 3b were prepared
by a similar manner to that of 3a, 1H NMR (DMSO-d6) d:
1.38 (9H, s, tBu), 1.89 (3H, s, CH2@C(CH3)CO–), 2.94
(1H, m, –NHCH2–), 3.13 (1H, m, –NHCH2–), 3.45 (1H,
m, 4-CH), 3.59 (2H, m, 2-CH and 3-CH), 3.75 (1H, m, 5-
CH), 4.04 (5H, dd, J = 6.6, 11.3Hz, –COOCH2CH), 4.26–
4.30 (2H, m, –COOCH2CH and 2- or 3-OH), 4.53 (1H, d,
J = 6.3 Hz, 4-OH), 4.73 (1H, d, J = 4.7 Hz, 2-or 3-OH),
4.90 (1H, d, J = 6.1 Hz, 5-OH), 5.66 (1H, m,
CH2@C(CH3)CO–), 6.08 (1H, s, CH2@C(CH3)CO–),
6.52 (1H, t, J = 5.4Hz, NH). Anal. Calcd for
C15H27NO8: C, 51.6; H, 7.79; N, 4.01. Found: C, 51.4;
H, 7.85; N, 3.90. A mixture of 3a (399 mg, 0.89 mmol),
glycerol dimethacrylate (4.3 lL, 0.018 mmol), azobisi-
sobutyronitrile (1.46 mg, 0.0089 mmol), and dioxane
(0.259 mL) was heated at 80 °C for 15 h. The resulting
polymer was ground by a mortar grinder and washed
successively with MeOH, 2-methyl-N-pyrrolidone (NMP)
to give 4a (210 mg, 53%). 4a (170 mg) was treated with a
mixture of TFA, CH2Cl2, dioxane, and water (10 mL,
50:39:10:1) at rt for 15 h, and then washed successively
with dioxane, NMP, water, MeOH, and ether, to give 5a
(125 mg). The amount of amino group of 5a was deter-
mined by the Ninhydrin test and was 2.27 lmol/mg. A
mixture of 5a (11 mg, 25 lmol), 7 (2.4 mg, 2.5 lmol), 1-
hydroxybenzotriazole (HOBt, 0.7 mg, 5 lmol), 1-ethyl-3-
(3-dimethylaminopropyl)carbodiimide
hydrochloride
(EDC HCl, 1.0 mg, 5 lmol), and NMP (1.5 mL) was
stirred at rt for 15 h. After collection by filtration, the solid
phase was washed by NMP and was treated with a mixture
of water, NMP, and acetyl anhydride (10 mL, 1:7:2) at rt
for 1 h. The resulting resin was washed with NMP and
acetonitrile to give 6a (80 lL). 6b were prepared by a
similar manner.
14. Western blot analysis on FKBP52 was carried out the
following way. The proteins were subjected to SDS–
PAGE followed by electroblotting onto PVDF membrane
using the Invitrogen XCell IITM blot module. After
blocking with Blocking oneTM (Nacalai Tesque Inc., cat.
03953-95) for 30 min at rt, the membrane was incubated
with anti-FKBP52 IgG (Santa Cruz Biotechnology, Inc.,
cat. sc-1803) for 1 h at rt. Following washing, the
membrane was incubated with HRP-conjugated anti-goat
antibodies (Santa Cruz Biotechnology, Inc., cat. sc-2033)
for 1 h at rt and washed again. The membrane was soaked
for 5 min in the detection reagent ECL plus (Amersham
Biosciences Corp. cat. RPN2132). The resulting light was
detected on Hyperfilm ECL (Amersham Biosciences Corp.
cat. RPN1674K).