Tuning of Binding Selectivity: Metal Control of Organic Guest Binding and Allosteric Perturbation of Fluorescent Metal Sensor

Article abstract of DOI:10.1021/jo991246t

Ligand 1, bearing two ethylenediamine groups, was designed to form a Hydrophobic cavity upon binding to metals. The shape of its nonpolar cavity depends on the metal: a reversal in binding preference for naphthalene or biphenyl groups is found when the metal is changed from zinc to copper, with a selectivity change of 260-fold. In the presence of dansylated amino acids, the new ligand constitutes a fluorescent sensor for zinc ion. Variations are seen in affinity for dansylamino acid with minor structural changes, and organic selectivity changes with complexes of variant metals. These findings suggest that sensor tuning of affinity and selectivity for metal is possible by choice of simple organic guest and for organic guest by choice of metal.

Full text of DOI:10.1021/jo991246t

8922
J . Org. Chem. 1999, 64, 8922-8928
Tu n in g of Bin d in g Selectivity: Meta l Con tr ol of Or ga n ic Gu est
Bin d in g a n d Alloster ic P er tu r ba tion of F lu or escen t Meta l Sen sor
Fen Wang and Alan W. Schwabacher*
Department of Chemistry, University of WisconsinsMilwaukee, Milwaukee, Wisconsin 53211
Received August 5, 1999
Ligand 1, bearing two ethylenediamine groups, was designed to form a hydrophobic cavity upon
binding to metals. The shape of its nonpolar cavity depends on the metal: a reversal in binding
preference for naphthalene or biphenyl groups is found when the metal is changed from zinc to
copper, with a selectivity change of 260-fold. In the presence of dansylated amino acids, the new
ligand constitutes a fluorescent sensor for zinc ion. Variations are seen in affinity for dansylamino
acid with minor structural changes, and organic selectivity changes with complexes of variant
metals. These findings suggest that sensor tuning of affinity and selectivity for metal is possible
by choice of simple organic guest and for organic guest by choice of metal.
In tr od u ction
geometry restricts the conformation, varying the geom-
etry and cavity shape of a macrocyclic receptor. Schneider
and DeShayes have studied metal-organized cyclophane
receptors that enhance the fluorescence intensity of a
fluorophore.12
Sensors for determination of metals and organic ana-
lytes are of great technical significance.1,2 Biosensors3
offer exquisite selectivities because of their subtle posi-
tioning of functional groups. One of the important areas
in contemporary sensor development is the design and
synthesis of functional molecules as fluorescent receptors
for the spectrometric determination of metal ions with
high sensitivity and selectivity.4,5 Imperiali has recently
demonstrated6 that hybrid peptide structures are advan-
tageous: the high metal affinity of simple organic ligands
can be tuned by incorporation into peptide structures.
Synthesis of self-assembled structures has received
considerable attention in the field of molecular recogni-
tion,7 due to the efficiency provided for construction of
well-defined large structures.8,9 The ability of a metal ion
to orient, as well as to gather organic fragments around
its coordination sphere, provides a versatile way to build
specific recognition sites.7a,10,11 The metal coordination
Metal tuning in such a structure would be of particular
interest. We decided to investigate the ability of a metal
to tune the selectivity for organic ligand binding and its
converse, tuning of metal binding by an organic guest.
We report here that flexible bis diamine 1 forms a
complex with both zinc and copper (Figure 1) and in doing
so closes the macrocycle. Furthermore, the distinct
coordination geometries of these metals perturb the
shape of the nonpolar cavity. Our studies were under-
taken to investigate the effect of the metal identity on
the organic binding selectivity and organic guest on metal
affinity. Here we confirm that the identity of the metal
can have a profound effect on the selectivity of the derived
receptor and also demonstrate a new approach to the
tuning of metal selectivity in sensors.13
* To whom correspondence should be addressed. E-mail:
awschwab@uwm.edu.
(1) Chemosensors of Ion and Molecular Recognition; Desvergne, J .-
P., Czarnik, A. W., Ed.; Kluwer Academic Press: Dordrecht, The
Netherlands, 1997; Vol. 492.
(2) de Silva, A. P.; Gunaratne, H. Q. N.; Gunnlaugsson, T.; Huxley,
A. J . M.; McCoy, C. P.; Rademacher, J . T.; Rice, T. E. Chem. Rev. 1997,
97, 1515-1566.
(3) Advances in Biosensors; Turner, A. P. F., Yevdokimov, Y. M.,
Ed.; J AI Press: Greenwich, CT, 1995; Vol. 3.
(4) Dujols, V.; Ford, F.; Czarnik, A. W. J . Am. Chem. Soc. 1997, 119,
7386-7387. Corradini, R.; Dossena, A.; Galaverna, G.; Marchelli, R.;
Panagia, A.; Sartor, G. J . Org. Chem. 1997, 62, 6283-6289. Mortellaro,
M. A.; Nocera, D. G. J . Am. Chem. Soc. 1996, 118, 7414-7415. Czarnik,
A. W. Chem. Biol. 1995, 2, 423-428.
(5) Czarnik, A. W. Acc. Chem. Res. 1994, 27, 302-308. Valeur, B.;
Bardez, E. Chem. Brit. 1995, 216-220.
(6) Torrado, A.; Walkup, G. K.; Imperiali, B. J . Am. Chem. Soc. 1998,
120, 609-610. Walkup, G. K.; Imperiali, B. J . Am. Chem. Soc. 1997,
119, 3443-3450. Cheng, R. P.; Fisher, S. L.; Imperiali, B. J . Am. Chem.
Soc. 1996, 118, 11349-11356.
Resu lts a n d Discu ssion
Syn th esis of Liga n d 1. We have designed and studied
several diarylphosphinic acid derivatives because of their
favorable properties: the aryl groups provide a hydro-
phobic concave surface for the cavity, and the phosphi-
nate anion provides water solubility and a hydrophilic
(10) Fujita, M.; Aoyagi, M.; Ibukuro, F.; Ogura, K.; Yamaguchi, K.
J . Am. Chem. Soc. 1998, 120, 611-612. Rivera, J . M.; Martin, R.;
Rebek, J ., J r. Science 1998, 279, 1021-1023. Kuroda, Y.; Kawashima,
A.; Hayashi, Y.; Ogoshi, H. J . Am. Chem. Soc. 1997, 119, 4929-4933.
Olenyuk, B.; Whiteford, J . A.; Stang, P. J . J . Am. Chem. Soc. 1996,
118, 8221-8230. Cole, K. L.; Farran, M. A.; Deshayes, K. Tetrahedron
Lett. 1992, 33, 599-602. Maverick, A. W.; Ivie, M. L.; Waggenspack,
J . H.; Fronczek, F. R. Inorg. Chem. 1990, 29, 2403-2409.
(11) Fox, O. D.; Dulley, N. K.; Harrison, R. G. J . Am. Chem. Soc.
1998, 120, 7111-7112.
(12) Baldes, R.; Schneider, H. J . Angew. Chem., Int. Ed. Engl. 1995,
34, 321-323. Schneider, H. J .; Ruf, D. Angew. Chem., Int. Ed. Engl.
1990, 29, 1159-1160. Cole, K. L.; Farran, M. A.; Deshayes, K.
Tetrahedron Lett. 1992, 33, 599-602.
(13) Part of this work has been communicated: Wang, F.; Schwa-
bacher, A. W. Tetrahedron Lett. 1999, 40, 7641-7644.
(7) (a) Linton, B.; Hamilton, A. D. Chem. Rev. 1997, 97, 1669-1680.
(b) Conn, M. M.; Rebek, J ., J r. Chem. Rev. 1997, 97, 1647-1668.
(8) Zeng, F.; Zimmerman, S. C. Chem. Rev. 1997, 97, 1681-1712.
Seto, C. T.; Whitesides, G. M. J . Am. Chem. Soc. 1993, 115, 905-916.
Lindsey, J . S. New J . Chem. 1991, 15, 153-180.
(9) Wyler, R.; de Mendoza, J .; Rebek, J ., J r. Angew. Chem., Int. Ed.
Engl. 1993, 32, 1699. Zimmerman, S. C.; Zeng, F.; Reichert, D. E. C.;
Kolotuchin, S. V. Science 1996, 271, 1095-1098.
10.1021/jo991246t CCC: $18.00 © 1999 American Chemical Society
Published on Web 11/26/1999

Tuning of Binding Selectivity
J . Org. Chem., Vol. 64, No. 24, 1999 8923
F igu r e 1.
Sch em e 1
convex surface.14 A Co2+-assembled bis amino acid mac-
rocycle formed a cavity of ideal shape for indole and
naphthalene derivatives.14a,b A related structure with
ethylenediamine groups was found to ligate Zn2+ to form
a complex that showed substantial discrimination upon
binding to naphthalene guests.14c The metals play both
structural and functional roles, as substrates bind best
that can simultaneously fill the cavity and interact
directly with the metal. Variation of the metal causes a
perturbation of the binding abilities,14 but relatively few
metals cause the formation of good binding sites. With
the expectation that the requirements for a single metal
would be less stringent than those placed upon metals
that must both assemble and organize a receptor, we now
report a related molecule that binds a single metal ion
to form a hydrophobic binding site. Dibromoxylene was
chosen as an appropriate sized spacer because of its ready
reactivity. Simply by changing the spacer, aniline, or
iodoaryl group, a family of related structures could be
prepared similarly.
lyzed coupling15 of phosphinate 2 with iodide 3 leads to
a diastereomeric mixture of diarylphosphinates 4. Alky-
lation with dibromide 5 in the presence of AgOTf gave
6, a protected form of 1. Compound 6, formed initially
as a triflate salt, was isolated as the tetrafluoroborate
salt by chromatography on silica gel with NaBF4. Depro-
tection with acid provides the desired ligand.
The syntheses of dansylamino16 acids were carried out
in the usual way with dansyl chloride in the presence of
excess amino acid under basic conditions, followed by
acidification.
Com p lex F or m a tion . Association of ligand 1 with
metals is readily discerned: changes in NMR and CD
spectra are consistent with 1:1 complexes (Figure 2) with
Zn2+ and Cu2+, as well as a 2:1 complex in the presence
of excess Cu2+. The shape of the hydrophobic cavity, as
well as the relative orientation of the nearby metal
center, should be distinctly different for these two
complexes since Cu2+ should lead to square planar
coordination, while Zn2+ is probably octahedral.
Met a l E ffect s on Or ga n ic Bin d in g Select ivit y.
NMR titrations of diamagnetic zinc complex 1‚Zn with
(1-naphthoxy)acetate 7 and with 2-phenylphenoxyacetate
8 provided strong evidence for a two-state binding event.
In each titration, resonances of host metal complex were
fit by multidimensional nonlinear least squares to a
single Kd. Titration of paramagnetic copper complex 1‚
The desired ligand 1 was prepared as shown in Scheme
1. N,N-Dimethylaniline is converted to tert-butyl arylphos-
phinate 2 by heating in a sealed tube with PCl3 and
pyridine, followed by quenching with tert-butyl alcohol.
Iodide 3 was prepared in enantiomerically pure form from
phenylalanine as we have reported.14c Palladium-cata-
(14) (a) Schwabacher, A. W.; Lee, J .; Lei, H. J . Am. Chem. Soc. 1992,
114, 7597-7598. (b) Lee, J ., Schwabacher, A. W. J . Am. Chem. Soc.
1994, 116, 8382-8383. (c) Schwabacher, A. W.; Stefanescu, A. D.;
Rehman, A. J . Org. Chem. 1999, 64, 1784-1788.
(15) Lei, H.; Stoakes, M.; Schwabacher, A. W. Synthesis 1992, 1255-
1260. Schwabacher, A. W.; Stefanescu, A. D. Tetrahedron Lett. 1996,
37, 425-428.
(16) Dansyl ) N-(5-dimethylaminonaphthalene-1-sulfonyl).

8924 J . Org. Chem., Vol. 64, No. 24, 1999
Wang and Schwabacher
Sch em e 2a
a
Schematic diagram showing how octahedral Cu leads to cavity
where methyl interferes with naphthalene, but not biphenyl guest.
F igu r e 2. NMR titration of 4 × 10-4 M 1 at 23 °C with Zn2+
in 10-2 M pH 9 borate. Linear changes until saturation
indicate high Zn affinity and a 1:1 stoichiometry.
Ta ble 1.
F igu r e 3. Fluorescence spectra of 2.0 × 10-4 M dansyl glycine
at 23 °C in 10-2 M pH 9 borate in the presence and absence of
2.0 × 10-4 M 1 and 2.0 × 10-4 M ZnSO4: (A) Dans-Gly + 1 +
Zn; (B) Dans-Gly + 1; (C) Dans-Gly + Zn2+ (spectrum of Dans-
Gly alone is identical); (D) 1 + Zn2+ (1 alone is identical).
guest
1. Zn Kda (M)
1‚Cu Kdb (M)
(7.25 ( 1.42) × 10-5
7, 2-(1-naphthoxy)-
acetate
(1.69 ( 0.11) × 10-4
8, 2-(2-phenylphen- (6.00 ( 0.43) × 10-3 e(9.73 ( 2.50) × 10-6
c
oxy)acetate
9, 1-naphthyl
phosphate
(5.04 ( 0.18) × 10-4
(1.66 ( 0.54) × 10-4
between the CD spectra of 1‚Zn and 1‚Cu. 1‚Zn shows
little if any CD, while 1‚Cu exhibits a distinct negative
extremum at 234 nm. These data confirm a significant
difference between the Zn and Cu complex conformations,
but do not allow detailed structure assignment.
a
b
Measured by NMR titration. Measured by CD titration.
c [Host] ) 10-5 M, so this value is an upper limit. Error bars reflect
only random error.
Cu was monitored by changes in circular dichroism (CD)
and fit in the same manner.
We believe that the metal centers interact with the
anionic portions of the guest molecules.14 While both
naphthalene and biphenyl guests bear the same anionic
groups, the polar groups will be presented differently to
the metal upon binding of the nonpolar aromatic group
into the cavity. It is the cooperativity between the two
types of binding interactions that leads to the observed
selectivity.
As shown in Table 1, a modest 2.3-fold stronger binding
of naphthalene 7 by copper complex 1‚Cu compared to
zinc complex 1‚Zn becomes a g600-fold stronger binding
by the Cu complex on replacement of the aromatic portion
of the guest with biphenyl 8. Phosphate 9 rather than
carboxylate slightly decreases the affinity, which high-
lights the nonpolar component to the binding. This
corresponds to a reversal in aromatic group preference
for these two metal complexes. Approximately a 35-fold
preference for the naphthalene 7 over the biphenyl 8 is
found for complex 1‚Zn; in contrast, 1‚Cu prefers the
biphenyl substrate by at least a factor of 7.4. Metal causes
a very substantial nonpolar shape selectivity with a
change of 260-fold.
Or ga n ic Su bstr a te Affects Meta l Affin ity. Com-
plexation of dansyl glycine and 1‚Zn causes a substantial
increase in fluorescence (Figure 3), because of the well-
documented sensitivity of dansylamide fluorescence to
environmental polarity. The increase in dansyl amino
acid fluorescence intensity was dependent on the pres-
ence of both 1 and Zn2+ and reversed upon the addition
of a strong chelating agent such as ethylenediamine
tetraacetic acid (EDTA) or iminodiacetate.
These results can be rationalized using models. Simple
CPK model-building exercises lead to the prediction that
bis diamine 1 appears capable of forming an unstrained
box-like Zn2+ complex 1‚Zn with a cavity that readily
accommodates naphthalenes. In contrast, square planar
coordination by 1 provides a somewhat strained and
lengthwise macrocyclic Cu2+complex 1‚Cu that should
better accommodate biphenyl than naphthalene in its
cavity. A favorable conformation places an inward-
pointing methyl group toward the center of one side
(Scheme 2). This conformation would be compatible with
biphenyl binding, as the central twist prevents steric
interference. However the methyl group could interfere
with naphthalene binding, so a different conformation
would be anticipated. That different conformations of Zn
and Cu species are present is indicated by the difference
This fluorescence change allowed determination (Fig-
ure 4) of the affinity of the Zn complex of 1 for a series of
dansyl amino acids, shown in Table 2. The fluorescence
intensities at 10 distinct wavelengths were simulta-
neously fit by multidimensional nonlinear least squares
to obtain a dissociation constant.
The R-carboxylate of the dansyl amino acid guest most
likely interacts with metal on binding, since lengthening
its distance from the naphthyl group by two carbons
decreases binding 39-fold (compare 10a -c). Further
evidence that the carboxylate interacts with the metal
on binding is found in the λmax of the emission: dansylg-
lycine emission shifts from 555 nm to 520 nm, equivalent
to a solvent change from H2O to MeOH. In contrast,

Tuning of Binding Selectivity
J . Org. Chem., Vol. 64, No. 24, 1999 8925
F igu r e 5.
F igu r e 4. Titration of 2.0 × 10-5 M dansyl glycine with 1. Zn
at 23 °C in pH 9 borate at 340 nm excitation. Emission
wavelengths (nm) from top to bottom at right: 510, 500, 520,
490, 530, 480, 540, 470, 550, 560, 570. All curves are theoreti-
cally fit to a single Kd.
Ta ble 2.
guest
dissoc constant (M)
10a
10b
10c
10d
10e
10f
10g
10h
10i
dansylglycine
dansyl-â-alanine
(9.70 ( 0.71) × 10-5
(1.27 ( 0.08) × 10-3
(3.77 ( 0.33) × 10-3
(5.77 ( 0.51) × 10-5
(1.03 ( 0.07) × 10-4
(1.09 ( 0.06) × 10-4
(2.48 ( 0.10) × 10-4
(8.05 ( 0.38) × 10-4
(1.45 ( 0.15) × 10-4
(1.10 ( 0.11) × 10-4
(2.60 ( 0.18) × 10-4
(1.15 ( 0.88) × 10-4
(2.05 ( 0.14) × 10-5
(2.14 ( 0.12) × 10-5
dansyl-4-aminobutyric acid
dansyl-aminomalonic acid
dansyl-L-aspartic acid
dansyl-D-aspartic acid
dansyl-L-glutamic acid
dansyl-D-glutamic acid
dansyliminodiacetic acid
dansyl-L-alanine
F igu r e 6. Fluorescence of 10-5 M dansyl glycine (9) or dansyl
â-alanine (b) and 5 × 10-4 M 1 in 100 mM pH 9 borate at 23
°C as Zn2+ is added. Background fluorescence of each dansyl
amino acid has been subtracted: 199 arbitrary units for dansyl
glycine and 553 units for dansyl â-alanine.
10j
10k
10l
constitute a very narrow class of substrates; nonetheless,
a 180-fold range of affinities is observed. Interestingly,
11 and 12 (Figure 5), the sulfonyl imidazole isomers of
10s and 10t, behave totally differently than the other
dansyl derivatives: no changes in fluorescence intensity
are observed upon the addition of 1‚Zn up to 5 × 10-4 M,
indicating a dissociation constant well above 10-3 M. As
the fluorescence of these derivatives is also sensitive to
solvent polarity, this structural change appears to have
had a profound effect on binding, consistent with a quite
specific geometric preference of the host for binding both
aromatic and carboxylate groups.
dansyl-D-alanine
dansyl-L-serine
10m
10n
10o
10p
10q
10r
10s
10t
dansyl-o-phospho-D,L-serine
dansyl-o-phospho-L-serine
dansyl aminomethylphosphonate (2.07 ( 0.11) × 10-4
dansyl l-ala-l-ala
dansyl L-proline
(5.67 ( 0.70) × 10-4
(5.17 ( 0.90) × 10-4
(1.81 ( 0.53) × 10-3
(6.30 ( 0.46) × 10-5
(8.65 ( 0.68) × 10-5
dansyl D-proline
dansyl l-his methyl ester
dansyl L-histidine
dansyl-â-alanine emission shifts upon binding only to 540
nm, indicating a more polar environment, and suggesting
a different orientation of the naphthalene in the cavity.
Presumably this conformation allows metal ligation. A
second carboxylate provides limited assistance: substrate
10d is bound only 2 times better than 10e, and 4-fold
stronger than 10g which is consistent with the conclusion
that the shorter the tether between carboxylate and
naphthalene, the better the binding, as long as an
appropriate conformation can be attained. Charge is not
the dominant factor: although dansyl-O-phosphoserine
(10m and 10n ), bearing three negative charges, has a
10-fold higher affinity than dansylaminomethylphos-
phonic acid (10o) with two negative charges, substrate
10o is bound two times less strongly than dansyl L-serine
(10l), which has only one negative charge. Moreover,
there is no significant different between the ester and
acid form of dansyl-L-histidine (10s and 10t). Therefore,
it is not the charge, but a specific structural feature,
presumably metal ligation, that is important. This must
be geometrically compatible with the hydrophobic com-
ponent of the binding.
Sen sin g Meta l Ion s by Or ga n ic Su bstr a tes. These
complexes that act as hosts for organic substances can
be used as sensors for metals.6,7 A mixture of 1 and 10a
constitutes a fluorescent sensor, as it becomes substan-
tially more fluorescent as Zn2+ is added and a complex
is formed. The significance of such behavior is that a
metal sensor composed of 1 and a dansyl amino acid can
be tuned by choice of readily available dansyl amino acid.
This is analogous to the tuning of a protein active site
by second-sphere ligand interactions that restrict con-
formation.6 However, our approach is operationally much
simpler. One could incorporate the fluorophore into a
large molecule and sense the metal ion upon binding to
the molecule. Perturbation of metal binding affinity and
selectivity would require structural changes to a large
molecule. In contrast, we simply change the small
substrates that are much more easily accessible.
Figure 6 shows the fluorescence of a mixture of 1 and
two different dansyl amino acids as they are titrated with
Zn2+. Each of these mixtures acts as a fluorescent sensor
for zinc, but dansylglycine causes a much more rapid
increase in fluorescence than does dansyl-â-alanine, even
though each is present at the same concentration.
Eventually, each enhances fluorescence to a similar
extent when all the dansyl groups are bound. In this way,
If both aromatic group and carboxylate interact with
host, enantioselective binding is reasonable. The enan-
tioselectivity of host 1‚Zn has been tested, and a small
2.4-fold preference for (S)-dansylalanine is observed
(10j and 10k ). The dansyl amino acid guest molecules

8926 J . Org. Chem., Vol. 64, No. 24, 1999
Wang and Schwabacher
Exp er im en ta l Section
the different amino acids lead to sensors whose affinity
can be tuned. The photophysics remains constant, as the
chromophore is not substantially perturbed. It is likely
that the metal selectivity can also be tuned in the same
way as the affinity, since we have demonstrated13 above
that complexes of different metals bind with substantially
different selectivity for organic molecular shape. Since
these are equilibria, the variation of organic ligand will
cause a similar variation in metal binding selectivity.
Such variations may be particularly useful in arrays of
sensors, where a pattern of changes can lead to very
specific information.17
Gen er a l P r oced u r es. Melting points were determined
without correction. 1H NMR spectra were recorded at 300
MHz, 13C NMR at 75 MHz. All chemical shifts are reported
downfield as δ values (ppm) from TMS in organic solvents or
DSS in D2O. 13C NMR spectra are 1H decoupled; multiplets
are 31P coupled. 31P NMR are reported vs external 85% H3-
PO4. TLC was performed on 0.25 mm Merck silica gel 60-F
and flash chromatography on 230-400 mesh Merck Kieselgel
60 as described.18 Substituted naphthalenes and biphenyl used
as binding substrates were purchased or prepared by literature
procedures. CH3CN was freshly distilled from P4O10; CH2Cl2,
tert-butyl alcohol, and triethylamine were freshly distilled from
CaH2. Other chemicals were used as obtained from commercial
suppliers unless otherwise specified.
ter t-Bu tyl (4-N,N-Dim eth yla m in o)p h en ylp h osp h in a te
(2). Freshly distilled N,N-dimethylaniline (1.0 mL, 1.89 mmol)
under N2 in a screw cap tube was cooled to -5 °C, and PCl3
(2.0 mL, 23 mmol) was added, followed by freshly distilled
pyridine (1.8 mL, 22 mmol), and the tube was capped. The
mixture was heated at 115 °C behind a shield for 4 h and
cooled to room temperature, and the excess reagents were
removed under reduced pressure. The solid was redissolved
in 2 mL of CH2Cl2, the solution under N2 was cooled to -10
°C, and tert-butyl alcohol (4.4 mL, 46 mmol) was added
dropwise. After 1 h at room temperature, the excess reagents
were removed by rotary evaporation, and the crude product
was redissolved in CH2Cl2 (100 mL), washed with satd
NaHCO3 (3 × 30 mL) and satd NaCl, and purified by flash
chromatography (silica, 20:1 CH2Cl2/MeOH) to yield 1.60 g
(85%) of tert-butyl 4-(dimethylamino)phenylphosphinate as a
white solid: mp 77.8-79.0 °C; 1H NMR (CDCl3) δ 7.70 (d, J )
547.3 Hz, 1H), 7.59 (dd, J ) 13.0, 8.9 Hz, 2H), 6.73 (dd, J )
8.9, 2.7 Hz, 2H), 3.02 (s, 6H), 1.55 (s, 9H); 13C NMR (CDCl3) δ
152.86 (d, J ) 2.7 Hz), 132.30 (d, J ) 13.0 Hz), 116.70 (d, J )
149.7 Hz), 111.36 (d, J ) 14.6 Hz), 82.01 (d, J ) 8.1 Hz), 40.07,
30.46 (d, J ) 4.8 Hz); 31P NMR (CDCl3) δ 16.55; HRMS calcd
for C12H20NO2P 241.1232, found 241.1253. Anal. Calcd for
In this example, the sensitivity difference is due
directly to a difference in binding of dansylamino acid to
the same host complex. Some of the benefits of tunable
sensors derive from an actual difference in the metal
affinity for sensor. For instance, if the intracellular zinc
concentration of a living cell were to be measured, a
sensor species must provide a signal at the concentration
range of interest, but too strong binding would be
deleterious. This is because a sufficiently strongly binding
species would compete for bound zinc, and not simply
signal the presence of the free ion. As the total intrac-
ellular zinc concentration is several orders of magnitude
higher than the concentration of free zinc, unsatisfactory
behavior would result.
Variation of detection sensitivity, involving the forma-
tion of the receptor complex described here, would not
appear to be as useful as a tunable sensor where the
metal affinity varies. Note, however, that there are two
equilibria involved: first the binding of metal to 1 to form
a host complex, and then the binding of dansyl amino
acid to this host with fluorescence enhancement. If the
metal binding equilibrium is driven by binding to the
signal-causing organic substrate, an effective variation
in metal affinity will be achieved.
C
12H20NO2P: C, 59.74; H, 8.36; N, 5.81. Found: C, 59.99; H,
8.40; N, 5.80.
ter t-Bu tyl 4-(N,N-Dim eth yla m in o)p h en yl-4-((2S)-2,3-
d i(t er t -b u t o x y c a r b o n y la m in o )p r o p y l)p h e n y l)p h o s -
p h in a te (4). A suspension of iodoarene 3 (0.45 g, 0.94
mmol), phosphinate 2 (0.217 g, 0.90 mmol), PPh3 (0.012 g,
0.046 mmol), and Pd2(dibenzylideneacetone)3 (0.020 g, 0.022
mmol) in CH3CN (8 mL) and Et3N (1.3 mL, 9.37 mmol) was
sealed in a screw-capped tube under N2 and heated in a 90 °C
oil bath for 36 h. Solvent removal by rotary evaporation and
flash chromatography (silica, 50:3 CH2Cl2/2-propanol) gave
The specific sensors described here are not likely to
be appropriate for intracellular studies because of the
high pH optimum and difficulties associated with a three-
component sensor. Endogenous aromatic anionic species
can compete with fluorophore for binding, and the ratio
of sensor components could vary. Both difficulties could
be addressed to some extent by linkage of 1 and fluoro-
phore with a flexible tether. Information readily obtained
by evaluation of separated molecules could lead to an
appropriate choice of groups.
1
0.278 g (53%) 4 as a white solid: mp ) 96.5-98 °C; H NMR
(CDCl3) δ 7.66 (dd, J ) 12.0, 7.8 Hz, 2H), 7.610 ( 0.0035 (dd,
J ) 11.4, 9.0 Hz 2H, diastereomers resolved), 7.21 (dd, J )
8.0, 3.0 Hz, 2H), 6.67 (dd, J ) 8.8, 2.5 Hz, 2H), 5.13 (br m,
2H), 3.85 (br m, 1H), 3.15 (br m, 2H), 2.98 (s, 6H), 2.83 (br m,
1H), 2.74 (dd, J ) 13.5, 6.8 Hz, 1H), 1.48 (s, 9H), 1.46 (s, 9H),
1.37 (s, 9H); 13C NMR (CDCl3) δ 156.71, 155.87, 152.18 (d, J
) 2.5 Hz), 141.01, 133.90 (d, J ) 146.9 Hz), 132.85 (d, J )
11.5 Hz), 131.21 (d, J ) 10.3 Hz), 129.10 (d, J ) 13.4 Hz),
119.51 (d, J ) 148.7 Hz), 111.15 (d, J ) 13.8 Hz), 82.58 (d, J
) 8.3 Hz), 79.36, 79.20, 52.46, 43.67, 39.99, 38.86, 30.95 (d, J
) 4.0 Hz), 28.35; 31P NMR (CDCl3) δ 28.23; HRMS calcd for
Con clu sion
We have presented an effective synthesis of the new
ligand 1 that forms a metal complex of 1:1 stoichiometry.
The hydrophobic cavity shape upon binding to metal
depends on the metal with the selectivity as hoped: the
octahedral complex 1‚Zn prefers naphthalene, while the
square planar 1‚Cu better binds biphenyl. Complexation
of the metal ion with ligand 1, and subsequent dansyl
amino acid inclusion in the cavity were demonstrated by
NMR, CD, and fluorescence spectra. A substantial change
is seen in the hydrophobic binding selectivity with change
of metal, and a variation of affinity with position of
ancillary binding groups. The convergent synthesis of 1
allows ready structural variation.
C
C
31H48N3O6P 589.3281, found 589.3282. Anal. Calcd for
31H48N3O3P: C, 63.14; H, 8.20; N, 7.13. Found: C, 63.16; H,
7.89; N, 7.16.
Bis-p h osp h in a te (6). A mixture of amine 4 (0.632 g, 1.07
mmol), R,R′-dibromo-p-xylene (0.113 g, 0.430 mmol), and
AgOTf (0.276 g, 1.07 mmol) was evacuated and flushed three
times with N2 before CH3CN (15 mL) and 2,6-di-tert-butylpy-
ridine (500 µL, 2.23 mmol) were added by syringe. An ash-
colored precipitate formed upon heating in an oil bath at 38
°C. After 12 h, the solvent was removed by rotary evaporation,
(17) Dickinson, T. A.; Michael, K. L.; Kauer, J . S.; Walt, D. R. Anal.
Chem. 1999, 71, 2192-2198.
(18) Still, W. C.; Kahn, M.; Mitra, A. J . Org. Chem. 1978, 43, 2923-
2925.

Tuning of Binding Selectivity
J . Org. Chem., Vol. 64, No. 24, 1999 8927
and crude 6 was applied to a silica gel column, rinsed with
10:3 CH2Cl2/MeOH, and eluted as the tetrafluoroborate with
10:3 MeOH/0.33 M aqueous NaBF4. The eluent was removed
by rotary evaporation, and the solid was partitioned between
H2O and CH2Cl2 and extracted with CH2Cl2 several times.
Organic aliquots were combined, and the solvent was removed,
affording 0.412 g (83%) 6 as a white solid mp > 200 °C dec;
1H NMR (CDCl3) δ 7.90 (br dd, 4H), 7.86 (br dd, 4H), 7.62 (br
dd, 4H), 7.30 (br dd, 4H), 6.97 (s, 4H), 5.29 (br m, 4H), 4.87 (s,
4H), 3.87 (br m, 2H), 3.48 (s, 12H), 2.99 (br m, 4H), 2.80 (br
m, 4H), 1.45 (s, 18H), 1.40 (s, 18H), 1.33 (s, 18H); 13C NMR
(CDCl3) δ 155.83, 154.93, 146.11, 141.96, 136.98 (d, J ) 136.2
Hz), 132.40, 132.03, 130.48 (d, J ) 9.7 Hz), 129.87 (d, J ) 142.4
Hz), 128.73 (d, J ) 7.5 Hz), 128.68, 120.43 (d, J ) 12.5 Hz),
84.01 (d, J ) 8.2 Hz), 78.37, 78.18, 71.17, 51.66, 42.76, 38.00,
29.84 (d, J ) 3.9 Hz), 27.37, 27.33; 31P NMR (CDCl3) δ 26.02;
FAB-MS calcd for [C70H104N6O12P2BF4]+ 1370, found 1370.
Anal. Calcd for C70H104N6O12P2B2F8‚2CH2Cl2: C, 53.15; H, 6.69;
N, 5.17. Found: C, 53.21; H, 6.63; N, 5.56.
1.0 Hz, 1H), 7.60 (dd, J ) 8.1, 7.8 Hz, 1H), 7.51 (dd, J ) 8.4,
7.6 Hz, 1H), 7.20 (d, J ) 7.5 Hz, 1H), 5.88 (d, J ) 8.3 Hz, 1H),
4.65 (d, J ) 8.3 Hz, 1H), 3.96 (m, 4H), 2.88 (s, 6H), 1.05 (t, J
) 7.1 Hz, 6H); 13C NMR (CDCl3) δ 165.39, 134.29, 130.93,
129.84, 129.63, 129.30, 128.58, 128.52, 123.17, 119.06, 115.47,
62.70, 58.92, 45.46, 13.67.
Da n syla m in om a lon ic a cid (10d ): 1H NMR (DMSO) δ
9.16 (d, J ) 9.3 Hz, 1H), 8.75 (d, J ) 8.5 Hz, 1H), 8.58 (d, J )
8.3 Hz, 1H), 8.23 (d, J ) 7.2 Hz, 1H), 7.70 (dd, J ) 16.2, 7.6
Hz, 2H), 4.49 (d, J ) 9.1 Hz, 1H), 3.66 (d, J ) 6.0 Hz, 1H),
3.05 (s, 6H).
Da n sylim in od ia cetic a cid (10i): 1H NMR (D2O) δ 8.47
(d, J ) 8.7 Hz, 1H), 8.42 (d, J ) 7.3 Hz, 1H), 8.32 (d, J ) 8.7
Hz, 1H), 7.68 (dd, J ) 8.0, 7.8 Hz, 2H), 7.43 (d, J ) 7.2 Hz,
1H), 4.06 (s, 4H), 2.90 (s, 6H); 13C NMR (D2O) δ 176.13, 150.19,
135.23, 129.98, 129.89, 129.79, 129.20, 128.68, 124.46, 120.68,
116.37, 51.06, 45.37.
Da n syl O-p h osp h o-D,L-ser in e tr is(tr ieth yla m in e) sa lt
1
(10m ): H NMR (D2O) δ 8.60 (dd, J ) 4.8, 4.7 Hz, 1H), 8.41
(d, J ) 8.6 Hz, 1H), 8.32 (d, J ) 7.3 Hz, 1H), 7.76 (m, 3H),
4.03 (m, 1H), 3.92 (m, 2H), 3.23 (s, 6H), 3.20 (q, J ) 7.4 Hz,
18H), 1.28 (t, J ) 7.3 Hz, 27H); 13C NMR (D2O) δ 174.57,
143.35, 135.30, 130.19, 129.18, 128.52, 127.67, 127.11, 126.04,
124.27, 118.37, 66.46 (d, J ) 4.7 Hz), 58.79 (d, J ) 8.8 Hz),
46.97, 46.36, 8.57; 31P NMR (D2O) δ 1.15; FAB-MS calcd
for [C15H20N2O8PS‚2Et3N]+ 621, [C15H20N2O8PS‚Et3N]+ 520,
[C15H20N2O8PS]+ 419, found 621, 520, 419.
Bis-d ia m in e (1). A solution of tetrafluoroborate 6 (0.200
g, 0.137 mmol) in CH2Cl2 (50 mL) and CF3CO2H (0.3 mL, 3.89
mmol) was stirred at room temperature for 0.5 h, and excess
reagents were removed under reduced pressure. After three
cycles of dissolution in H2O and rotary evaporation, the residue
was precipitated from 1 mL of H2O by addition of 30 mL of
MeOH. The white precipitate was isolated by centrifugation
and dried in vacuo to yield 0.154 g of 1 (96%): mp > 200 °C
dec; 1H NMR (D2O) δ 7.79 (dd, J ) 10.9, 8.9 Hz, 4H), 7.67 (dd,
J ) 11.7, 8.0 Hz, 4H), 7.54 (dd, J ) 7.4, 1.6 Hz, 4H), 7.43 (dd,
J ) 8.0, 2.4 Hz, 4H), 6.89 (s, 4H), 4.91 (s, 4H), 3.97 (ddd, J )
10.4, 10.4, 6.4 Hz, 2H), 3.57 (s, 12H), 3.42 (dd, J ) 13.7, 5.7
Hz, 2H), 3.36 (d, J ) 13.0, 4.7 Hz, 2H), 3.23 (dd, J ) 14.3, 6.2
Hz, 2H), 3.06 (dd, J ) 14.3, 8.6 Hz, 2H); 13C NMR (D2O) δ
146.14 (d, J ) 3.1 Hz), 138.75 (d, J ) 2.8 Hz), 137.31 (d, J )
140.0 Hz), 133.03 (d, J ) 138.2 Hz), 133.27 (d, J ) 11.1 Hz),
132.91, 132.05 (d, J ) 10.7 Hz), 130.15 (d, J ) 13.4 Hz), 129.85,
121.95 (d, J ) 13.1 Hz), 73.25, 53.35, 50.86, 41.04, 36.46; 31P
NMR (D2O) δ 22.82; FAB-MS calcd for [C42H56N6O4P2BF4]+
857, found 857.
Gen er a l P r oced u r e for Syn th eses of Da n syl Am in o
Acid s (10a -t, 11, a n d 12). Compounds 10a ,19 10b,20 10c,21
10e,22 10f,22 10g,23 10h ,23 10j,20 10k ,20 10l,24 10q,25 10r ,25 10s,20
10t,20 11,26 and 1226 were prepared as described. A 0.65 mmol
portion of amino acid was dissolved in 2.5 mL of solvent along
with an equivalent amount of base, 0.13 mmol of dansyl
chloride in 2.5 mL of CH3CN was added dropwise, and the
mixture was stirred at rt until TLC indicated dansyl chloride
reacted completely. For dansyl diethylaminomalonate (solvent,
CH2Cl2; base, pyridine), 10i, 10p (solvent, H2O; base, Na2CO3),
the solution was acidified to pH 4-5 and extracted with
EtOAc. After removal of EtOAc, the crude product was purified
by flash chromatography. 10d was obtained by hydrolysis of
its ethyl ester. For 10m -o (solvent, H2O; base, Et3N), the
solution was extracted with CH2Cl2 several times to remove
excess dansyl chloride. The unreacted amino acid was pre-
cipitaed from the H2O layer by adding CH3CN and removed
by centrifuge. The precipitation was repeated until unreacted
amino acid was removed completely. The solvent was evapo-
rated and the product was dried under vacuum.
Da n syl O-p h osp h o-L-ser in e t r is(t r iet h yla m in e) sa lt
1
(10n ): H NMR (D2O) δ 8.56 (d, J ) 8.3 Hz, 1H), 8.41 (d, J )
8.6 Hz, 1H), 8.32 (d, J ) 7.3 Hz, 1H), 7.75 (m, 3H), 4.04 (m,
1H), 3.95 (m, 2H), 3.19 (q, J ) 7.2 Hz, 18H), 3.16 (s, 6H), 1.28
(t, J ) 7.3 Hz, 27H); 13C NMR (D2O) δ 174.63, 144.08, 135.24,
130.13, 129.20, 128.57, 127.93, 127.33, 125.88, 123.84, 118.17,
66.49 (d, J ) 4.45 Hz), 58.82, 46.97, 46.25, 8.57; 31P NMR (D2O)
δ
1.14; FAB-MS calcd for [C15H20N2O8PS‚2Et3N]+ 621,
[C15H20N2O8PS‚Et3N]+ 520, [C15H20N2O8PS]+ 419, found 621,
520, 419.
Da n syl a m in om eth ylp h osp h on ic bis(tr ieth yla m in e)
sa lt (10o): 1H NMR (D2O) δ 8.44 (d, J ) 8.7 Hz, 1H), 8.35 (d,
J ) 8.8 Hz, 1H), 8.27 (d, J ) 7.3 Hz, 1H), 7.67 (dd, J ) 8.3,
8.3 Hz, 2H), 7.34 (d, J ) 7.7 Hz, 1H), 3.17 (q, J ) 7.3 Hz, 12H),
2.99 (s, 1H), 2.94 (s, 1H), 2.79 (s, 6H), 1.27 (t, J ) 7.3 Hz, 18H);
13C NMR (D2O) δ 150.98, 133.57, 130.25, 129.81, 129.32,
129.20, 129.09, 124.37, 119.49, 116.30, 46.85, 45.18, 41.27(d,
J ) 140.07 Hz), 8.52; 31P NMR (D2O) δ 14.73; FAB-MS calcd
for [C13H18N2O5PS‚Et3N]+ 446, [C13H18N2O5PS]+ 345, found
446, 345.
Da n syl L-Ala -L-Ala :. 1H NMR (CDCl3) δ 8.50 (br s, 1H),
8.49 (d, J ) 8.5 Hz, 1H), 8.31 (d, J ) 8.6 Hz, 1H), 8.21 (d, J )
7.1 Hz, 1H), 7.48 (ddd, J ) 12.0, 11.9, 8.0 Hz, 2H), 7.17 (dd, J
) 7.7, 7.6 Hz, 2H), 6.53 (d, J ) 7.6 Hz, 1H), 4.17 (dd, J ) 6.9,
6.9 Hz, 1H), 3.88 (dd, J ) 7.4, 7.4 Hz, 1H), 2.86 (s, 6H), 1.10
(s, 3H), 1.08 (s, 3H); 13C NMR (CDCl3) δ 175.35, 172.10, 150.76,
134.50, 130.46, 129.92, 129.43, 129.40, 128.58, 123.58, 119.39,
116.50, 115.71, 52.55, 48.37, 45.49, 19.00, 17.37.
Titr a tion P r oced u r e. 1H NMR titrations were performed
at 20.0 ( 0.5 °C in 0.1 M borate buffer in D2O, prepared from
anhydrous Na2B4O7. In each titration, the host concentration
was kept constant by adding a solution of host and guest to a
solution of host at the same concentration. All distinguishable
resonances of host were fit to yield a single dissociation
constant, using multidimensional nonlinear least squares
using Scientist (MicroMath) version 2.01. Error limits are 95%
confidence (s plane).
Da n syl d ieth yla m in om a lon a te: 1H NMR (CDCl3) δ 8.56
(d, J ) 8.5 Hz, 1H), 8.30 (d, J ) 8.6 Hz, 1H), 8.24 (dd, J ) 7.4,
(19) Malamas, M. S.; Sestanj, K.; Millen, J . J . Med. Chem. Chim.
Ther. 1991, 26, 369-374.
(20) Froelich, P. M.; Murphy, L. D. Anal. Chem. 1977, 49, 1606-
1608.
(21) Hartkoorn, A.; Lunden, R. J . Med. Chem. 1974, 17, 649-651.
(22) Nakamura, M.; Ikeda, T.; Nakamura, A.; Ikeda, H.; Ueno, A.;
Toda, F. Chem. lett. 1995, 5, 343-344.
(23) Corradini, R.; Sartor, G.; Marchelli, R.; Dossena, A.; Spisni, A.
J . Chem. Soc., Perkin Trans. 2 1992, 11, 1979-1983.
(24) Armstrong, D. W.; Tang, Y.; Chen, S.; Zhuo, Y.; Bagwill, C.;
Chen, J . Anal. Chem. 1994, 66, 1473-1484.
(25) Reitsma, B. H.; Yeung, E. S. Anal. Chem. 1987, 59, 1059-1061.
(26) Mostovnikov, V. A.; Nechaev, C. V.; Solonenko, M. G. J . Appl.
Spectrosc. 1990, 53, 712-717.
Resonances were fit using the following equation:
∆δ
δi ) δ0i
+
maxi(S - S2 - 4H G )
x
0
0
H0
where Kd ) dissociation constant, H0 ) total host added, G0 )
total guest added, and S ) H0 + G0 + KD, all in M.
Fluorescence titrations were performed at 23 °C in 0.1 M
borate. Guests at 10-5-10-4 M (always [guest] < Kd/5) were
titrated with a solution of host, and the fluorescence intensities

8928 J . Org. Chem., Vol. 64, No. 24, 1999
Wang and Schwabacher
at 10 distinct wavelengths between 470 and 570 nm were fit
to yield a Kd, as described for NMR titrations.
Ack n ow led gm en t. This work was supported by
NIH grant GM47736-02. We acknowledge Kamal P. B.
Herath for preliminary studies in a related synthesis.27
CD titrations were performed at 23 °C in 0.1 M borate. As
in NMR titrations, hosts (always [host] < Kd/5) were titrated
with a solution of guest and host, and ellipticities at 6-10
wavelengths (320-330 nm for naphthalene guests, 240-260
nm for biphenyl) were simultaneously fit to yield a Kd, as
described above.
Su p p or tin g In for m a tion Ava ila ble: Copies of 1H and 13C
1
NMR spectra for 1 and H NMR spectra for 1‚Zn. This material
is available free of charge via the Internet at http://pubs.acs.org.
(27) Herath, K. P. B. Synthesis of Arylphosphinic Acid Derivatives
as Building Blocks for Binding Sites; MS Thesis, Iowa State University.
Ames, 1994.
J O991246T