Phosphotyrosine binding by ammonium- and guanidinium-modified cyclodextrins

Full text of DOI:10.1021/jo971979i

J . Org. Chem. 1998, 63, 1737-1739
1737
P h osp h otyr osin e Bin d in g by Am m on iu m -
a n d Gu a n id in iu m -Mod ified Cyclod extr in s
Edward S. Cotner and Paul J . Smith*
Department of Chemistry and Biochemistry, University of
Maryland, Baltimore County, Baltimore, Maryland 21250
Received October 28, 1997
In tr od u ction
An important challenge in host-guest chemistry is the
preparation of host molecules that are capable of affecting
important biochemical processes in cellular settings. This
requires the identification of biological targets (or guests)
for which small synthetically accessible complexing agents
can be rationally designed; moreover, there should be a
reasonable expectation that complexation will bring
about some desirable physiological outcome. One such
target is phosphotyrosine. The phosphorylation of ty-
rosine residues on the cytosolic surfaces of receptor
tyrosine kinases represents the quintessential early step
in growth factor-induced mitogenesis.1 Sites of modifica-
tion are recognized and bound by specific secondary
signaling proteins; these as well as subsequent protein-
protein binding events provide a means for transmission
of the signal from one component in the pathway to the
next, ultimately leading to the activation of transcription
factors that initiate cell division.2
F igu r e 1. Structure of a guanidinium-substituted â-CD
derivative and a schematic representation of an inclusion
complex with phosphotyrosine on a protein surface (showing
only one of the unmodified primary hydroxyl groups for
clarity).
Sch em e 1a
In a number of human cancers, including breast and
ovarian cancers, there is a strong correlation between
abnormally high levels of tyrosine-phosphorylated pro-
teins and tumor cell growth, which appears to result from
the overexpression or constitutive activity of cellular
tyrosine kinases.3 One approach to inhibiting inap-
propriate mitogenic signaling is to inhibit the binding of
secondary cytosolic proteins to tyrosine-phosphorylated
receptors. We envision that this can be accomplished
using small molecules that form strong specific complexes
with phosphotyrosine residues on protein surfaces. Ac-
cordingly, we have initiated a program to identify host
molecules with high affinity and selectivity for phospho-
tyrosine, employing cyclodextrins (CDs) as starting ar-
chitectures. The parent compounds possess reasonably
good water solubility and their ability to form inclusion
complexes with aromatic guests in aqueous solution is
well documented.4 In addition, a variety of approaches
to regiospecific chemical modification have been identi-
fied, allowing for the introduction of desired functionality
to enhance binding affinity and selectivity.5 Initially, we
have chosen to examine derivatives of â-CD with pendant
ammonium or guanidinium groups on the primary rim
a
Reagents: (a) TsCl, pyridine (30%); (b) NaN3, KI, DMF (80%);
(c) H2, Pd/C, H2O (90%); (d) 5, (i-Pr)2NEt, DMF (33%).
of the cavity, with the expectation that these groups will
take part in favorable electrostatic interactions with the
phosphate group of phosphotyrosine within a complex
(Figure 1).6-8
Resu lts a n d Discu ssion
Guanidinium-containing â-CD derivative 4 has been
prepared via known amino derivative 39 using the
guanylating agent 1H-pyrazolecarboxamidine hydrochlo-
ride (5)10 (Scheme 1). A bis-guanidinium derivative of
â-CD was also synthesized (Scheme 2). Here, the seven
small circles represent the seven glucose units in â-CD
(6) Phosphate binding by ammonium and guanidinium com-
pounds: (a) Springs, B.; Haake, P. Bioorg. Chem. 1977, 6, 181-190.
(b) Dietrich, B.; Fyles, D. L.; Fyles, T. M.; Lehn, J .-M. Helv. Chim.
Acta 1979, 62, 2763-2787.
(7) Biological recognition of phosphotyrosine: (a) Waksman, G.;
Shoelson, S. E.; Pant, N.; Cowburn, D.; Kuriyan, J . Cell 1993, 72, 779-
790. (b) Shoelson, S. E.; Eck, M. J .; Nolte, R. T.; Dhe-Paganon, S.; Tru¨b,
T. Cell 1996, 85, 695-705. (c) Stuckey, J . A.; Schubert, H. L.; Fauman,
E. B.; Zhang, Z.-Y.; Dixon, J . E.; Saper, M. A. Nature 1994, 370, 571-
575.
(8) Binding of phosphates and amino acids by ammonium cyclo-
dextrins: (a) Boger, J .; Knowles, J . R. J . Am. Chem. Soc. 1979, 101,
7631-7633. (b) Eliseev, A. V.; Schneider, H.-J . J . Am. Chem. Soc. 1994,
116, 6081-6088. (c) Bonomo, R. P.; Pedotti, S.; Vecchio, G.; Rizaqrelli,
E. Inorg. Chem. 1996, 35, 6873-6877.
(1) Fantl, W. J .; J ohnson, D. E.; Williams, L. T. Annu. Rev. Biochem.
1993, 62, 453-481.
(2) Koch, C. A.; Anderson, D.; Moran, M. F.; Ellis, C.; Pawson, T.
Science 1991, 252, 668-674.
(3) (a) Luttrell, D. K.; Lee, A.; Lansing, T. J .; Crosby, R. M.; J ung,
K. D.; Willard, D.; Luther, M.; Rodriguez, M.; Berman, J .; Gilmer, T.
M. Proc. Natl. Acad. Sci. U.S.A. 1994, 91, 83-87. (b) Mishra, S.;
Hamburger, A. W. Int. J . Cancer 1994, 58, 538-542. (c) Clark, J . W.;
Santos-Moore, A.; Stevenson, L. E.; Frackelton, A. R., J r. Int. J . Cancer
1996, 65, 186-191.
(4) Clarke, R. J .; Coates, J . H.; Lincoln, S. F. Adv. Carbohydr. Chem.
Biochem. 1988, 46, 205-248.
(5) Croft, A. P.; Bartsch, R. A. Tetrahedron 1983, 39, 1417-1474.
(9) (a) Melton, L. D.; Slessor, K. N. Carbohydr. Res. 1971, 18, 29-
37. (b) Petter, R. C.; Salek, J . S.; Sikorski, C. T.; Kumaravel, G.; Lin,
F.-T. J . Am. Chem. Soc. 1990, 112, 3860-3868.
(10) Bernatowicz, M. S.; Wu, Y.; Matsueda, G. R. J . Org. Chem.
1992, 57, 2497-2502.
S0022-3263(97)01979-8 CCC: $15.00 © 1998 American Chemical Society
Published on Web 02/14/1998

1738 J . Org. Chem., Vol. 63, No. 5, 1998
Notes
Ta ble 1. Associa tion Con sta n ts (M-1) for â-CD a n d
Sch em e 2a
Der iva tivesa ,b
host
m-cresol purple
tyrosine
phosphotyrosine
â-CD
1130
1300
1590
850
n.b.d.
40 ( 10
n.b.d.
30 ( 3
n.b.d.
n.b.d.
30 ( 10
n.b.d.
350 ( 8
310 ( 8
3
4
8
9
1410
a
Conditions: 25.0 °C, 100 mM sodium phosphate, pH 7.00.
n.b.d. indicates that no binding was detected.
b
a
Reagents: (a) NaN3, KI, DMF (50%, crude); (b) H2, Pd/C, H2O
F igu r e 2. Fluorescence emission spectrum of phosphotyrosine
in the absence (solid line) and presence (broken line) of 8: 205
µM phosphotyrosine, 2.9 mM 8, 25.0 °C, 100 mM sodium
phosphate, pH 7.00. Excitation wavelength 274.5 nm.
(90%); (c) 5, (i-Pr)2NEt, H2O (74%).
and X is a functional group covalently bonded to the
primary (6) carbon of the A and D sugars in the
cyclodextrin ring. The required starting material 6 was
prepared by the reaction of â-CD with 4,4′-biphenyldi-
sulfonyl chloride in pyridine as reported by Tabushi and
co-workers.11 This method also produces a minor amount
of the A,C isomer (ca. 10% based on isolated bis-sulfonate
esters). After reaction with sodium azide and reduction,
the regioisomeric amines were separated on carboxy-
methyl cellulose ion-exchange resin using a linear gradi-
ent of aqueous ammonium bicarbonate; pure 8 was then
converted to 9 with 5.12 Compounds 4, 8, and 9 were
isolated as bicarbonate salts after purification and were
converted to chloride salts prior to use in binding experi-
ments; pure 3 was isolated as the free base.
Competitive binding experiments were carried out
using â-CD and derivatives with tyrosine and phospho-
tyrosine, using the indicator dye m-cresol purple. Ini-
tially, 1:1 association constants for the cyclodextrins and
the indicator were determined from titration data using
an iterative linear least-squares approach.13 Subsequent
competition experiments involved repeating the 1:1
experiments in the presence of a fixed concentration of
amino acid, and the resulting data were analyzed as
described by Pendergast and Connors.14 Results are
summarized in Table 1.
significantly enhanced binding of phosphotyrosine rela-
tive to 3 and 4. Importantly, both 8 and 9 also show a
significant preference for the phosphorylated amino acid
(-∆∆G g 1.4 kcal/mol), indicating that they would ex-
hibit selectivity for phosphorylated vs nonphosphorylated
tyrosine residues on a protein surface. Our results
suggest that this preference results from an increase in
favorable electrostatic interactions between the pendant
cationic groups and the phosphate of phosphotyrosine,
as no similar increase is seen for tyrosine binding. In
the case of 8, the energetic benefit of the second am-
monium group (1.4 kcal/mol) is similar to that observed
for the formation of a salt bridge in a wide variety of
host-guest systems (1.2 ( 0.2 kcal/mol).16
To investigate whether binding of phosphotyrosine
involves inclusion within the CD cavity, preliminary
fluorescence experiments were carried out with 8 and 9
(Figure 2). Compound 8 produces a 74% increase in the
quantum yield of phosphotyrosine (extrapolated to 100%
bound), which is similar in magnitude to that previously
observed for phenol with unmodified â-CD (71%) and is
consistent with the formation of an inclusion complex.17,18
A much smaller increase was seen with 9 (12% extrapo-
No binding was detected between â-CD and either
tyrosine or phosphotyrosine.15 The addition of a single
ammonium group provides a host (3) that binds both
amino acids with similar (low) affinity while monoguani-
dinium derivative 4 shows no detectable binding with
either amino acid. The addition of a second positively
charged group provides hosts (8 and 9) that display
(15) Binding constants for tyrosine and unmodified â-CD of 33.3 M-1
and 48 ( 5 M-1 have previously been reported: Matsuyama, K.; El-
Gizawy, S.; Perrin, J . H. Drug Devel. Ind. Pharm. 1987, 13, 2687-
2691. Bekos, E. J .; Gardella, J . A., J r.; Bright, F. V. J . Inclusion
Phenom. Mol. Recognit. Chem. 1996, 26, 185-195. Under our assay
conditions, we were unable to detect binding in replicate experiments
(Ka < 30 M-1 based on our experience that association constants of 30
M-1 can be reliably reproduced in this system).
(16) (a) Schneider, H.-J .; Schiestel, T.; Zimmerman, P. J . Am. Chem.
Soc. 1992, 114, 7698-7703. (b) Schneider, H.-J .; Blatter, T.; Palm, B.;
Pfingstag, U.; Ru¨diger, V.; Theis, I. J . Am. Chem. Soc. 1992, 114, 7704-
7708.
(17) Hoshino, M.; Imamura, M.; Ikehara, K.; Hama, Y. J . Phys.
Chem. 1981, 85, 1820-1823.
(18) Ratios of integrated fluorescence intensities of uncorrected
emmission spectra were used; absorbances of solutions with and
without added cyclodextrin were matched at the excitation wavelength.
(11) Tabushi, I.; Yamamura, K.; Nabeshima, T. J . Am. Chem. Soc.
1984, 106, 5267-5270.
(12) The synthesis of diamine 8 has recently been reported: ref 8c.
(13) Cowart, M. D.; Sucholeiki, I.; Bukownik, R. R.; Wilcox, C. S. J .
Am. Chem. Soc. 1988, 110, 6204-6210.
(14) Pendergast, D. D.; Connors, K. A. J . Pharm. Sci. 1984, 73,
1779-1783.

Notes
J . Org. Chem., Vol. 63, No. 5, 1998 1739
R f ) 0.04; 1H NMR (D2O, 500 MHz) δ 5.13-5.03 (m, 7 H), 4.04-
3.78 (m, 26 H), 3.70-3.50 (m, 16 H); 13C NMR (D2O, 125 MHz)
δ 158.61, 102.77, 102.64, 102.35, 83.20, 82.16, 81.96, 81.90, 73.86,
73.84, 73.79, 73.73, 73.53, 72.82, 72.68, 72.63, 72.56, 71.31, 61.34,
61.13, 60.92, 42.98. Anal. Calcd for C43H84ClN3O39‚5H2O: C,
39.65; H, 6.50; Cl, 2.72; N, 3.23. Found: C, 39.48; H, 6.40; Cl,
3.07; N, 3.31.
lated to 100% bound). Increases in fluorescence intensity
resulting from CD binding are thought to result primarily
from decreased collisional quenching of the excited state
by solvent molecules.17 The reduced enhancement with
9 relative to that with 8 then suggests that phosphoty-
rosine is more solvent exposed in the former complex,
although the fact that binding solutions contain 100 mM
phosphate argues against purely electrostatically driven
association.
The bis-ammonium and -guanidinium cyclodextrins
described represent promising lead compounds in the
development of phosphotyrosine-specific binding com-
pounds. However, hosts exhibiting dissociation constants
in the low micromolar to nanomolar range will be
required to effectively compete for sites on tyrosine-
phosphorylated receptors in vivo. More conformationally
restricted analogues of 8 and 9 are currently being
synthesized with the expectation that these will display
higher affinities for phosphotyrosine. Experiments with
phosphotyrosine-containing peptides and proteins are
also presently underway.19
6A,6D-Did eoxy-6A,6D-d igu a n id in ocycloh ep ta a m ylose (9).
8 (304 mg, 0.268 mmol), 5 (162 mg, 1.11 mmol), and N,N-
diisopropylethylamine (187 µL, 1.07 mmol) were stirred in 2 mL
of water for 17 h at room temperature and then frozen and
lyophilized. The resulting solid was dissolved in 3 mL of water,
additional 5 (163 mg, 1.11 mmol) and N,N-diisopropylethylamine
(93 µL, 0.54 mmol) were added, and the reaction was stirred at
room temperature for 17 h. The solution was frozen and
lyophilized, and the resulting solid was stirred with 40 mL of
ether, collected by filtration, and washed with 60 mL of ether
to provide crude 9 as a fine white powder (707 mg). This was
dissolved in 250 mL of 50 mM NH4HCO3 and applied to a
(carboxymethyl)ellulose column that was eluted with a linear
gradient of aqueous NH4HCO3 (50-400 mM) followed by 400
mM NH4HCO3. Fractions containing 9 were combined, concen-
trated under reduced pressure, and repeatedly lyophilized from
deionized water to provide the bicarbonate salt of 9 as a fluffy
-
white solid (264 mg, 74% yield). Data for 9 (HCO3 salt): mp
210 °C dec; R f ) 0.0; FAB-MS m/z calcd for C44H78N6O33 (M -
H+) 1217.8, measured 1218.4; 1H NMR (D2O, 500 MHz) δ 5.12-
5.04 (m, 7 H), 4.03-3.78 (m, 24 H), 3.69-3.50 (m, 18 H); 13C
NMR (D2O, 125 MHz) δ 164.12, 158.69, 102.79, 102.66, 102.42,
83.28, 82.28, 82.25, 82.04, 82.02, 73.85, 73.82, 73.75, 73.54, 72.90,
72.85, 72.69, 72.61, 71.41, 61.39, 61.20, 60.97, 42.99. Anal. Calcd
for C46H80N6O39‚5H2O: C, 38.12; H, 6.40; N, 5.80. Found: C,
37.93; H, 6.11; N, 5.71.
Exp er im en ta l Section
Gen er a l Meth od s. TLCs were run on precoated EM Science
silica gel 60 plates and developed in 7:7:5:4 EtOAc/2-propanol/
NH4OH/water. Whatman Express-Ion Exchanger C was used
for purification of 3, 4, 8, and 9. 1H chemical shifts are reported
downfield from external sodium 3-(trimethylsilyl)propionate-d4
in D2O; acetone in D2O was used as an external standard for
13C NMR spectra (δ ) 31.07 ppm20). â-CD was obtained from
Cerestar USA and used without further purification; elemen-
tal analysis indicated that the sample used in binding experi-
ments contained 9.5 waters per CD molecule (Anal. Calcd for
C42H70O35‚9.5H2O: C, 38.62; H, 6.87. Found: C, 38.54; H, 6.58).
m-Cresol purple and tyrosine were obtained from Aldrich and
used without further purification. Phosphotyrosine was ob-
tained from Sigma and was converted to its ammonium salt by
lyophilization from aqueous ammonium bicarbonate followed by
repeated lyophilization from deionized water; elemental analysis
was consistent with the bis-ammonium salt containing 1.5
waters (Anal. Calcd for C9H18N3O6P‚1.5H2O: C, 33.52; H, 6.57;
N, 13.04. Found: C, 33.48; H, 6.58; N, 12.77).
Mon o-6-deoxy-6-gu an idin ocycloh eptaam ylose (4). Crude
3 (909 mg, 0.382 mmol) was suspended in 1.5 mL of dry DMF
and 5 (235 mg, 1.60 mmol) and N,N -diisopropylethylamine (224
µL, 1.6 mmol) were added. The reaction was stirred for 22 h
under N2, transferred to a 40 mL beaker, and 20 mL of ether
was added dropwise. The resulting suspension was stirred for
2 h, and the precipitate was collected by filtration, washed with
40 mL of ether, and dried in vacuo, providing crude 4 as a fine
white powder (1.15 g). A portion of the product (810 mg) was
dissolved in 250 mL of water and applied to a (carboxymethyl)-
cellulose column that was eluted with a linear gradient of
aqueous NH4HCO3 (10-200 mM). Fractions containing 4 were
combined, concentrated under reduced pressure, and repeatedly
lyophilized from deionized water to provide the bicarbonate salt
of 4 as a fluffy white solid (308 mg, 33% yield). Data for 4
(HCO3- salt): R f ) 0.04; FAB-MS m/z calcd for C43H75N3O34 (M+)
1177.4, measured 1177.0. Anal. Calcd for C44H75N3O37‚4H2O:
C, 40.34; H, 6.39; N, 3.21. Found: C, 40.33; H, 6.49; N, 3.57.
The bicarbonate salt of 4 was dissolved in water, loaded on a
column of Dowex 1 × 2-100 (chloride form), and eluted with
water. Fractions containing 4 were combined, concentrated
under reduced pressure, and lyophilized, providing the chloride
salt of 4 as a clear solid. Data for 4 (Cl- salt): mp 230 °C dec;
The bicarbonate salt of 9 was dissolved in water, loaded on a
column of Dowex 1 × 2-100 (chloride form), and eluted with
water. Fractions containing 9 were combined, concentrated under
reduced pressure, and lyophilized, providing the chloride salt
of 9 as a clear solid. Data for 9 (Cl- salt): 1H NMR (D2O, 500
MHz) δ 5.13-5.03 (m, 7 H), 4.04-3.78 (m, 24 H), 3.70-3.50 (m,
18 H). Anal. Calcd for C44H88Cl2N6O38‚5H2O: C, 38.29; H, 6.43;
Cl, 5.14; N, 6.09. Found: C, 38.17; H, 6.37; Cl, 5.22; N, 6.03.
Bin d in g Exp er im en ts. A solution of m-cresol purple (ca.
60 µM) and the desired amino acid (2-3 mM) was titrated with
a solution of the cyclodextrin (ca. 8 mM) that also contained
m-cresol purple and the amino acid at the same concentrations
present in the cuvette. Absorbance at 432 nm was used for
quantitation. Duplicate experiments were performed, and un-
certainties are based on the error in the slopes of the resulting
linear plots using least-squares analysis (95% confidence inter-
val). In all titrations, a distinct isosbestic point was observed,
consistent with 1:1 binding between the dye and the cyclodex-
trins examined. The pH’s of solutions were checked before and
after experiments and were found to change less than 0.01
during the course of an experiment; pH changes of this magni-
tude were found to produce changes in dye absorbance on the
order of 0.002, which is within the instrumental error for
individual measurements. Dye absorbance was found to be
unaffected by the presence of tyrosine or phosphotyrosine at the
concentrations used in competitive experiments.21
Ack n ow led gm en t. We thank Professor Craig Wil-
cox (University of Pittsburgh) for making the Hostest
program available to us for analysis of 1:1 binding data,
Professor Todd Houston (Virginia Commonwealth Uni-
versity) for discussions regarding guanylating agents,
and Professor Lisa Kelly and Dr. Qui-Wei Xu of this
Department for discussions regarding fluorescence and
assistance in obtaining 13C NMR spectra, respectively.
â-Cyclodextrin was a gift from Cerestar USA (Ham-
mond, IN). This work was funded by the University of
Maryland, Baltimore County.
(19) Previously determined association constants for phenylalanine
and (uncharged) phenylalanineamide with â-CD suggest that the
charged ammonium and carboxylate groups of phosphotyrosine may
significantly impede binding: Rekharsky, M. V.; Schwarz, F. P.;
Tewari, Y. B.; Goldberg, R. N. J . Phys. Chem. 1994, 98, 10282-10288.
(20) Abeygunawardana, C.; Bush, C. A. Biochemistry 1991, 30,
8568-8577.
J O971979I
(21) Gray, J . E.; MacLean, S. A.; Reinsborough, V. C. Aust. J . Chem.
1995, 48, 551-556.