Fluorophore-linked zinc(II)dipicolylamine coordination complexes as sensors for phosphatidylserine-containing membranes

Article abstract of DOI:10.1016/j.tet.2004.08.052

A series of Zn²⁺-2,2′-dipicolylamine (Zn²⁺-DPA) coordination complexes with an attached NBD fluorophore are synthesized and evaluated as fluorescent sensors. The sensors do not respond to vesicles composed of zwitterionic phosphatidy

Full text of DOI:10.1016/j.tet.2004.08.052

Tetrahedron 60 (2004) 11307–11315
Fluorophore-linked zinc(II)dipicolylamine coordination
complexes as sensors for
phosphatidylserine-containing membranes
*
C. Lakshmi, Roger G. Hanshaw and Bradley D. Smith
Department of Chemistry and Biochemistry and The Walther Center for Cancer Research, University of Notre Dame, 251 Nieuwland Hall,
Notre Dame, IN 46556-5670, USA
Received 19 February 2004; revised 8 July 2004; accepted 19 August 2004
Available online 15 September 2004
Abstract—A series of Zn2C–2,20-dipicolylamine (Zn2C–DPA) coordination complexes with an attached NBD fluorophore are synthesized
and evaluated as fluorescent sensors. The sensors do not respond to vesicles composed of zwitterionic phosphatidylcholine, but the NBD
fluorescence emission is enhanced in the presence of anionic vesicles. A sensor with two Zn2C–DPA units and a hydrophilic
tris(ethyleneoxy) linker produced a larger emission enhancement than an analogue with a butyl linker, and titration with 1:1 POPC:POPS
vesicles lead to an apparent phospholipid association constant of 5.3!104 MK1. The sensor can detect the presence of vesicles containing as
little as 5% phosphatidylserine. The sensing effect apparently requires a membrane surface because the sensors do not respond to a
phosphatidylserine derivative that is monodispersed in aqueous solution.
q 2004 Elsevier Ltd. All rights reserved.
1. Introduction
detected. Monitoring the appearance of phosphatidylserine
(PS, Fig. 1) on a cell surface is a more suitable assay for the
early stages of apoptosis.9–11 PS is maintained almost
exclusively in the inner monolayer of the plasma membrane
of healthy animal cells, but a scrambling effect is initiated
early in apoptosis which results in externalization of PS.
New and improved assays for the in vitro and in vivo
detection of apoptosis (programmed cell death) would be a
useful contribution to research in cell biology and clinical
medicine. Current methods used in detecting apoptosis
focus on changes in cytosolic biochemical events, degra-
dation of cell nuclear material, and changes in the cell
plasma membrane structure.1–4 Each assay has its own
inherent advantages and shortcomings. For example,
monitoring of cytosolic caspase activity and detection of
nucleic acid fragmentation are classic assays for apop-
tosis.5–8 However, these processes often occur too late in the
apoptosis process to allow other important events to be
Keywords: Molecular recognition; Phospholipid binding; Phosphate
recognition; Chemosensor; Apoptosis; Phospholipid asymmetry;
Phosphatidylserine.
Figure 1. Head group structure of common phospholipids.
Abbreviations: DHPC, 1,2-dihexanoyl-sn-glycero-3-phosphocholine;
DHPS, 1,2-dihexanoyl-sn-glycero-3-[phospho-L-serine]; NBD, 7-nitro-
1,3-benz-2-oxa-1,3-diaza-4-yl; PC, phosphatidylcholine; POPA, 1-palmi-
toyl-2-oleoyl-sn-glycero-3-phosphate; POPC, 1-palmitoyl-2-oleoyl-sn-gly-
cero-3-phosphocholine; POPG, 1-palmitoyl-2-oleoyl-sn-glycero-3-
phosphoglycerol; POPS, 1-palmitoyl-2-oleoyl-sn-glycero-3-[phospho-L-
serine]; PS, phosphatidylserine; TEO, tris(ethyleneoxy); TES, N-tris[hy-
droxymethyl]methyl-2-aminoethanesulfonic acid.
Detection of externalized PS is currently achieved using
annexin V, a member of a class of Ca2C-dependent
phospholipid binding proteins.12,13 Annexin V exhibits
high affinity for PS, and binding of fluorescent labeled
annexin V (e.g., annexin-FITC) to the cell surface is
currently used in microscopic and flow cytometric assays
for the detection of apoptosis. However, use of annexin V
is not without limitations. PS binding by annexin V is a
Ca2C-dependent process, which raises the concern of
* Corresponding author. Tel.: C1-574-631-8632; fax: C1-574-631-6652;
e-mail: smith.115@nd.edu
0040–4020/$ - see front matter q 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tet.2004.08.052

11308
C. Lakshmi et al. / Tetrahedron 60 (2004) 11307–11315
activation of non-specific Ca2C-dependent phospholipid
scramblases within the membrane bilayer when Ca2C-rich
annexin binding buffers are used. Further, the protein is
susceptible to degradation and exhibits slow binding
kinetics.14,15 Low molecular weight synthetic mimics of
annexin V that bind rapidly to PS-enriched membrane
surfaces in a Ca2C-independent manner would be an
attractive alternative.
systematically alter the three molecular components shown
in Figure 3: the PS binding group, the linker, and the
fluorophore. Herein, we report the design and synthesis of
the first examples of this series, namely, coordination
complexes 1–4 (Fig. 4). The design retains the Zn2C–DPA
affinity units, attached in a meta orientation to a phenyl
ring,18 which is in turn linked to a 7-nitrobenz-2-oxa-1,3-
diaza-4-yl (NBD) fluorophore. The NBD fluorophore is
excited at 470 nm and exhibits a broad emission centered at
530 nm, which limits its use in microscopy and flow
cytometry applications; nevertheless, its environmental
sensitivity makes it useful for the mechanistic experiments
performed here.
An X-ray crystal structure of an annexin–glycero-
phosphoserine complex shows that the phosphoserine head
group is coordinated to one of the four canonical binding
sites of the protein through two bridging Ca2C ions.16 This
picture suggested to us that synthetic metal complexes with
appropriate charge, geometry, and spatial orientation may
bind to the head group of anionic phospholipids preferen-
tially over zwitterionic phospholipids. In other words,
rationally designed coordination complexes may be useful
as mimics of annexin V. Several phosphate chemosensors
with Zn2C-coordinated binding sites have been described in
the literature.17–19 For example, Hamachi and co-workers
recently described an anthracene-based sensor with two zinc
dipicolylamine (Zn2C–DPA) units (Fig. 2).20 The sensor
binds to the dianionic phosphotyrosine groups in phos-
phorylated peptides with association constants of up to
Figure 3. Modular design of PS sensor.
107 MK1
.
We subsequently demonstrated that this sensor, which we
call PSS-380, can be used to detect the presence of anionic
phospholipids, specifically PS, on the surface of vesicles
and cells.21 While the fluorescence emission of PSS-380
was unchanged upon treatment with vesicles composed
entirely of zwitterionic phosphatidylcholine (PC), a 10-fold
increase in fluorescence emission was observed upon
addition of anionic vesicles composed of 1:1 POPC:POPS
(Fig. 1). PSS-380 was successfully used to detect apoptosis
in Jurkat cells treated with camptothecin (an apoptosis
inducing agent) via fluorescence microscopy and flow
cytometry. Although PSS-380 is an attractive alternative
to annexin V-FITC, its wavelength of excitation (380 nm) is
in the UV range, which is not compatible with the lasers in
most flow cytometers. Replacing the anthracene unit with
fluorophores that absorb at longer wavelengths would
alleviate this drawback.
In an effort to develop second generation sensors, we have
chosen to pursue a modular design which allows us to
Figure 4. Coordination complexes 1–4 as potential PS sensors.
We examined two types of linkers, butyl (relatively
hydrophobic) and tris(ethyleneoxy) (TEO, relatively hydro-
philic), anticipating that upon binding, the nature of the
linker may influence the interaction between the fluorophore
and the bilayer membrane. We illustrate how sensor
Figure 2. Structure of PSS-380.

C. Lakshmi et al. / Tetrahedron 60 (2004) 11307–11315
11309
response to PS-enriched membrane surfaces is altered by
changes in the number of Zn2C–DPA binding units, as well
as the structure of the linker.
2. Results and discussion
2.1. Synthesis
The synthesis of 1 was achieved in eight steps starting from
commercially available 5-hydroxyisophthalic acid
(Scheme 1). The butyl linker was attached to 3,5-
bis(hydroxymethyl)phenol,22,23 using N-(4-bromo-
butyl)phthalimide and K2CO3 in acetonitrile, and the
product converted in high yield to the corresponding
dibromide 7 using CBr4 and Ph3P. The zinc binding groups
were introduced in the next step by treating 7 with DPA in
the presence of K2CO3 in DMF. The phthalimide protecting
group was removed using hydrazine hydrate and the
resulting amine was treated with NBD-Cl in the presence
of K2CO3 in dry THF to afford 9 in 70% yield. The
complexation of 9 with Zn(NO3)2 in methanol/water
mixture was quantitative and the complex was used for
the binding studies without further purification. Compound
2, with a TEO linker was synthesized using a similar
synthetic strategy (Scheme 2). In addition, the syntheses of
control compounds 3 and 4, each with one Zn2C–DPA unit
are depicted in Schemes 3 and 4.
Scheme 2. (a){N-Boc-{2-[2-(2-bromoethoxy)-ethoxy]-ethyl}monoamine,
K2CO3, acetonitrile, reflux (89%); (b) CBr4, Ph3P, CH2Cl2 (70%); (c) 2,20-
dipicolylamine, K2CO3, DMF (87%); (d) TFA, CH2Cl2 (85%); (e) NBD-Cl,
K2CO3, THF (70%); (f) zinc nitrate, MeOH/H2O (100%).
2.2. Evaluation of vesicle sensing properties
The ideal candidate for apoptosis detection via externalized
PS should have a strong affinity for anionic PS embedded in
a membrane that is primarily composed of zwitterionic
phospholipids. The sensor binding, however, must not
disrupt the membrane structure. The effect of 1 on
membrane integrity was measured in two ways. The first
was vesicle leakage, where 1 was found to be incapable of
inducing bilayer permeabilization. Specifically, addition of
1 at concentrations up to 10 mM to vesicles composed of
either 100% POPC or 1:1 POPC:POPS failed to induce
carboxyfluorescein leakage. Furthermore, 1 and 2 do not
penetrate into vesicles. This was proved by adding sensors 1
or 2 at concentrations up to 2.0 mol% to a dispersion of
vesicles (25 mM total phospholipid) consisting either of
100% POPC or 1:1 POPC:POPS. Subsequent addition of the
membrane-impermeable reducing agent sodium dithionite
at 2 h after sensor addition resulted in greater than 95%
quenching of fluorescence. The near total quenching of
NBD fluorescence by sodium dithionite indicates that 1 and
2 do not cross the vesicle membrane and become protected
from chemical reduction. The ability of 1 and 2 to bind to
PS-enriched membranes (see below) and not disrupt the
membrane structure establishes the utility of the meta-aryl
bis (Zn2C–DPA) unit as a PS affinity group in modular
designs of PS-sensors.
Scheme 1. (a) EtOH, concd H2SO4, reflux, (96%); (b) LiAlH4, THF, 0 8C to
rt, (96%); (c) N-(4-bromobutyl)phthalimide, K2CO3, acetonitrile, reflux,
(90%); (d) CBr4, Ph3P, CH2Cl2, (89%); (e) 2,20-dipicolylamine, K2CO3,
DMF, (91%); (f) hydrazine hydrate, CH2Cl2/EtOH, reflux, (90%); (g)
NBD-Cl, K2CO3, THF (70%); (h) zinc nitrate, MeOH/H2O, (100%).

11310
C. Lakshmi et al. / Tetrahedron 60 (2004) 11307–11315
Scheme 3. (a) N-(4-Bromobutyl)phthalimide, K2CO3, acetonitrile, reflux
(67%); (b) CBr4, Ph3P, CH2Cl2 (85%); (c) 2,20-dipicolylamine, K2CO3,
DMF (93%); (d) hydrazine hydrate, EtOH/CH2Cl2, reflux (90%); (e) NBD-
Cl, K2CO3, THF (72%); (f) zinc nitrate, MeOH/H2O (100%).
Scheme 4. (a){N-Boc-{2-[2-(2-bromoethoxy)-ethoxy]-ethyl}monoamine,
K2CO3, acetonitrile, reflux (89%); (b) CBr4, Ph3P, CH2Cl2 (65%); (c) 2,20-
dipicolylamine, K2CO3, DMF (87%); (d) TFA, CH2Cl2 (85%); (e) NBD-Cl,
K2CO3, THF (76%); (f) zinc nitrate, MeOH/H2O (100%).
The NBD fluorophore is known to exhibit an enhancement
in fluorescence intensity upon transfer from a polar
environment to an apolar environment.24,25 Thus, the
fluorescence emission from sensors 1 and 2 was expected
to increase upon binding to the surface of a bilayer
membrane. Indeed, titration of both sensors with anionic
vesicles produced moderate to large fluorescence enhance-
ments. While such an enhancement is not a necessary
feature for employment in apoptosis detection, the observed
NBD fluorescence enhancement upon membrane binding
provides insight into the interactions between the sensor and
the membrane surface. The titration experiments involved
addition of unilamellar vesicles composed of 1:1 POPC:-
POPS, 1:1 POPC:POPA, 1:1 POPC:POPG, or 100% POPC
to a 1 mM solution of 1 or 2. The resulting isotherms (Fig. 5)
were fitted to a 1:1 binding model, which allowed
calculation of apparent phospholipid binding constants.†
With sensors 1 and 2, the order of binding affinities to
vesicles was 1:1 POPC:POPSO1:1 POPC:POPGw1:1
POPC:POPA[100% POPC (Table 1). The sensors were
not expected to bind the zwitterionic POPC head group, and
this titration was performed as a control to determine
specificity of the sensors for anionic PS. Although 1 binds to
anionic vesicles with slightly higher binding constants than
2, the more important sensing property is the difference in
emission intensity for anionic vesicles versus 100% POPC.
In this case, the response with sensor 2 is much more
selective. For example, when the total phospholipid
concentration is 60 mM, the emission of butyl-linked sensor
1 with 1:1 POPC:POPS vesicles is about two times that
observed with 100% POPC vesicles, whereas, the intensity
ratio with the TEO-linked sensor 2 is about five.
Sensors 1 and 2 only respond to membrane-bound
phospholipids, as evidenced by the lack of response to the
short acyl chain phospholipids, dihexanoylphosphatidyl-
choline (DHPC) and dihexanoylphosphatidylserine
(DHPS), which exist as monomeric dispersions in aqueous
media. For example, addition of DHPC (100 mM) or a 1:1
mixture of DHPC:DHPS to a 1 mM solution of 2 results in
no detectable increase in NBD fluorescence intensity (data
not shown).
The effect of only one Zn2C–DPA unit in the sensor
structure was evaluated using control compounds 3 and 4.
Vesicles composed of 100% POPC or 1:1 POPC:POPS were
added to a 1 mM solution of 3 or 4. As in the case above, the
emission of butyl-linked sensor 3 with 1:1 POPC:POPS
vesicles (60 mM phospholipid) is about two times that
observed with 100% POPC vesicles, whereas the intensity
† At this point it is not known if the sensors induce lateral phospholipid
aggregation; however, control experiments showed that the meta-aryl bis
(Zn2C–DPA) unit does not promote phospholipid transmembrane flip
flop. The binding constants in Tables 1 and 2 are considered apparent
because they do not consider Zn2C dissociation from the DPA units in 1
and 2 in the absence of vesicles.

C. Lakshmi et al. / Tetrahedron 60 (2004) 11307–11315
11311
Figure 6. Fluorescence intensity of a 1 mM solution of either 3 or 4 upon
addition of phospholipid vesicles consisting of 100% POPC (% for 3, ,
for 4), or 1:1 POPC-POPS (B for 3, ! for 4) in 5 mM TES buffer, 145 mM
NaCl, pH 7.4 at 25 8C. For comparison, the fluorescence intensity for 2
upon addition of vesicles composed of 1:1 POPC:POPS is represented by
(C). Fluorescence intensity of 2 upon addition of vesicles composed of 1:1
POPC:POPS in the presence of 1 mM Zn(NO3)2 is denoted by (6). All
experiments were reproduced at least three times.
significant fraction of 1 and 2 may exist as the correspond-
ing mononuclear Zn2C complexes. To test if Zn2C
dissociation induces a measurable effect on sensor perform-
ance, the titration of sensor 2 with 1:1 POPC:POPS vesicles
was repeated in the presence of 1 mM Zn(NO3)2. The
resulting titration isotherm (Fig. 6) shows a 20% increase in
fluorescence enhancement, but no measurable change in
binding constant. This is consistent with an equilibrium
picture that converts a less responsive mononuclear Zn2C
form of sensor 2 into a more responsive binuclear Zn2C
form.
Figure 5. Fluorescence intensity upon addition of vesicles to a 1 mM
solution of 1 or 2 in 5 mM TES buffer, 145 mM NaCl, pH 7.4 at 25 8C. (A)
Addition of 1:1 POPC:POPS (B), 1:1 POPC:POPG (,),1:1 POPC:POPA
(!), or 100% POPC (%) to 1. (B) Addition of 1:1 POPC:POPS (B), 1:1
POPC:POPG (,),1:1 POPC:POPA (!), or 100% POPC (%) to 2.
The PS sensitivity of sensor 2 was evaluated by conducting
additional titration experiments with POPC vesicles
enriched with various amounts of POPS. Specifically, a
solution of 2 (1 mM) was titrated with vesicles composed of
POPC and 0–50% POPS (Fig. 7). The emission response of
2 increases with increasing fraction of PS, until about 20%
PS, after which there is a plateau. This trend is also reflected
by the apparent binding constants listed in Table 2. The data
in Figure 7 shows that sensor 2 can readily detect the
presence of bilayer membranes containing 5% PS. It is
worth noting that this is approximately the fraction of PS
that is externalized during the early-to-intermediate stages
of cell apoptosis.27
ratio with TEO-linked sensor 4 is about five. Thus, it is clear
that the hydrophilic TEO-linker is more useful for detecting
PS-containing membranes over PC-only membranes. The
emission of sensor 2 with two Zn2C–DPA units is
significantly higher than that obtained with 4, indicating
that the second Zn2C–DPA unit increases the polar
interactions that produce a stronger response to PS-
containing membranes.
The Hamachi group has measured association constants for
the two Zn2C in PSS-380 and found K1w7!106 MK1 and
K2w7!104 MK1 26
It is reasonable to think that sensors
.
1–4 have similar Zn2C association constants which means
that at 1 mM, the sensor concentration used in this study, a
In conclusion, we have shown that coordination complexes
with Zn2C–DPA units do not respond to anionic
Table 1. Sensor/phospholipid binding constants (Ka)
Sensor
Ka!104 (MK1 a
1:1 POPC:POPS
)
100% POPC
1:1 POPC:POPG
1:1 POPC:POPA
1
2
3
4
!1
!1
23.3G1.7
5.3G2.0
11.5G5.5
7.7G3.3
14.2G3.8
2.0G0.5
—
11G6.0
2.0G0.3
—
!1b
!1b
—
—
a
Values are average of at least three independent measurements.
Values are average of two independent measurements.
b

11312
C. Lakshmi et al. / Tetrahedron 60 (2004) 11307–11315
3. Experimental
3.1. Synthesis
3.1.1. Synthesis of compound 6. 3,5-Bis-hydroxymethyl
phenol (1.06 g, 6.88 mmol), N-(4-bromobutyl)phthalimide
(2.13 g, 7.57 mmol) and K2CO3 (4.75 g, 34.4 mmol) were
mixed in acetonitrile and refluxed for 24 h. The reaction
mixture was cooled to room temperature and the solvent
removed. The residue was redissolved in CH2Cl2 and the
insoluble residue filtered off. The filtrate was washed with
water, dried over MgSO4 and the solvent removed under
vacuum. The pure product was obtained as a semi-solid after
purification on a silica column using EtOAc. Yield 90%; 1H
NMR (300 MHz, CDCl3): d 1.82–1.88 (br, 4H), 3.73–3.78
(m, 2H), 3.99–4.03 (m, 2H), 4.63 (s, 4H), 6.81 (s, 2H), 6.90
(s, 1H), 7.70–7.73 (m, 2H), 7.83–7.85 (m, 2H); 13C NMR
(75 MHz, CDCl3): d 25.3, 26.6, 37.7, 64.7, 67.3, 112.0,
117.5, 123.2, 132.1, 133.9, 143.0, 159.2, 168.5.
Figure 7. Fluorescence intensity of 2 (1 mM) upon addition of POPC
vesicles enriched with 0% (%), 5% (B), 10% (,), 20% (!), or 50% (C)
POPS in 5 mM TES buffer, 145 mM NaCl, pH 7.4 at 25 8C.
Table 2. Sensor 2/phospholipid binding constants (Ka) for vesicles with varying PS content
95:5 POPC:POPS
90:10 POPC:POPS
80:20 POPC:POPS
50:50 POPC:POPS
Ka!104 (MK1 a
)
1.0G0.3
2.2G0.5
5.7G0.5
5.3G2.0
a
Values are averages of at least three independent measurements.
3.1.2. Synthesis of compound 7. To a solution of
compound 6 (3.05 g, 8.62 mmol) and CBr4 (6.3 g,
18.96 mmol) in dry CH2Cl2 at 0 8C, was slowly added
Ph3P (4.85 g, 18.52 mmol) in dry CH2Cl2. The reaction
mixture was stirred at room temperature overnight. The
reaction mixture was concentrated under vacuum and the
pure product obtained as a white solid after silica gel
chromatography (EtOAc/hexanes (1:1). Yield 89%; 1H
NMR (300 MHz, CDCl3): d 1.86–1.89 (m, 4H), 3.78 (m,
2H), 4.00 (m, 2H), 4.42 (s, 4H), 6.83 (m, 2H), 6.98 (br, 1H),
7.71–7.74 (m, 2H), 7.84–7.87 (m, 2H); 13C NMR (75 MHz,
CDCl3): d 25.4, 26.6, 33.0, 37.7, 67.4, 115.3, 121.9, 123.3,
132.2, 134.1, 139.7, 159.4, 168.5.
3.1.3. Synthesis of compound 8. Compound 7 (1 g,
2.08 mmol), and K2CO3 (1.44 g, 10.4 mmol) were mixed in
dry DMF. 2,20-Dipicolylamine (1.04 g, 5.21 mmol) was
added and the reaction mixture stirred overnight. The DMF
was removed and the residue dissolved in CHCl3. The
insoluble residue was filtered off. The filtrate was washed with
water, brine and dried over MgSO4. Solvent was removed and
the residue chromatographed on a neutral alumina column
using CHCl3 as eluent to afford the pure product. Yield 91%;
1H NMR (300 MHz, CDCl3): d 1.85 (br, 4H), 3.62 (s, 4H),
3.74–3.79 (m, 10H), 3.97 (m 2H), 6.84 (s, 2H), 7.04 (s, 1H),
7.08–7.13 (m, 4H), 7.57–7.64 (m, 8H), 7.67–7.70 (m, 2H),
7.81–7.84 (m, 2H), 8.48–8.50 (m, 4H); 13C NMR (75 MHz,
CDCl3): d 25.6, 26.9, 37.8, 58.7, 60.2, 67.2, 113.6, 121.6,
122.1, 122.9, 123.3, 132.2, 134.1, 136.6, 140.8, 149.1, 159.2,
159.9, 168.6; FAB MS m/z 718 [MCHC].
Figure 8. Sensing of PS-containing membranes. PS is represented with
filled head group, PC is unfilled head group.
phospholipids (such as PS) when they are monodispersed in
aqueous solution, but there is a significant fluorescence
enhancement when the anionic phospholipids are part of a
vesicle membrane (Fig. 8). The magnitude of the enhance-
ment depends on the structure of the linker between the
Zn2C–DPA affinity units and the NBD fluorophore. The
more hydrophilic TEO linker produces a sensor with
significantly enhanced emission response to anionic PS-
enriched membranes compared to zwitterionic PC-only
membranes. The results reported here show that the NBD
fluorophore can be used as a probe for rapid testing of new
sensor designs, and for investigating fundamental questions
in molecular recognition at the membrane surface. Future
studies will determine if analogous sensors with more
practically useful fluorophores can detect PS externalization
and cell apoptosis.
3.1.4. Synthesis of compound 9.
3.1.4.1. Generation of free amine from compound 8.
Compound 8 (0.550 g, 0.77 mmol) was dissolved in a
mixture of CH2Cl2 and EtOH (25:75). Hydrazine hydrate

C. Lakshmi et al. / Tetrahedron 60 (2004) 11307–11315
11313
(0.074 g, 2.31 mmol) was added and refluxed for 1 h. The
solvent was removed and the residue dissolved in CH2Cl2.
The insoluble residue was removed by filtration. The filtrate
was collected, solvent removed under high vacuum to afford
the free amine as a semi-solid. The free amine was directly
used in Section 3.1.4.2 without further purification. Yield
(300 MHz, CDCl3): d 1.42 (s, 9H), 3.32 (m, 2H), 3.54 (m, 2H),
3.63–3.66 (m, 6H), 3.70–3.73 (m, 2H), 3.80 (s, 8H), 3.86
(m, 2H), 4.14 (m, 2H), 5.05 (br, 1H), 6.89 (s, 2H), 7.07 (s, 1H),
7.10–7.15 (m, 4H), 7.56–7.65 (m, 8H), 8.49–8.52 (m, 4H);
13C NMR (75 MHz, CDCl3): d 28.6, 40.6, 58.8, 60.3, 67.4,
70.0, 70.5, 70.9, 113.8, 121.9, 122.2, 122.9, 136.6, 140.8,
149.2, 159.1, 160.0; FAB MS m/z 748 [MCHC].
1
(90%); H NMR (300 MHz, DMSO-d6): d 1.60–1.70 (m,
2H), 1.76–1.84 (m, 2H), 2.78–2.83 (m, 2H), 3.28 (br, 2H),
3.63 (s, 4H), 3.79 (s, 8H), 3.93–3.97 (m, 2H), 6.23 (s, 2H),
7.05 (s, 1H), 7.10–7.14 (m, 4H), 7.56–7.64 (m, 8H),
8.48–8.50 (m, 4H).
3.1.9. Synthesis of compound 13.
3.1.9.1. Generation of free amine from 12. Compound
12 (1.073 g, 1.43 mmol) was dissolved in 6.9 mL of 50%
TFA in dry CH2Cl2 at 0 8C under N2 atmosphere. The
reaction mixture was stirred at room temperature for 2 h.
After removing TFA, the residue was taken in CH2Cl2 and
neutralized with satd Na2CO3. The aqueous layer was
extracted with CH2Cl2 (3!20 mL) and the combined
organic layer dried over MgSO4. The solvent was removed
under vacuum to yield the free amine (85%), which was
directly used for the next step.
3.1.4.2. Preparation of compound 9. Amine (0.75 g,
1.28 mmol) and K2CO3 (0.88 g, 6.4 mmol) was stirred in
dry THF under Ar atmosphere. NBD-Cl (0.28 g,
1.41 mmol) dissolved in dry THF was slowly added and
the reaction mixture stirred in dark for 36 h. The reaction
mixture was filtered to remove the K2CO3 and concentrated
under vacuum. The dark green semi-solid obtained was
dissolved in CH2Cl2, washed with water, brine and dried
over MgSO4. The crude product was purified on an alumina
column using CHCl3 as eluent. Yield 70%; 1H NMR
(300 MHz, CD3OD): d 1.86 (br, 4H), 3.52 (s, 6H), 3.66
(s, 8H), 3.98 (m, 2H), 6.20 (d, JZ9 Hz, 1H), 6.73 (s, 2H),
6.92 (s, 1H), 7.13–7.18 (m, 4H), 7.54–7.68 (m, 8H),
8.30–8.33 (m, 5H); 13C NMR (75 MHz, CDCl3): d 25.6,
26.7, 43.9, 58.6, 60.1, 67.3, 98.7, 113.7, 113.9, 122.2, 122.3,
123.0, 136.6, 140.7, 140.9, 144.2, 144.5, 149.2, 158.8,
159.9, 168.3; FAB MS m/z 751 [MCHC].
3.1.9.2. Synthesis of compound 13. Synthesized accord-
ing to procedure given in Section 3.1.4.2. Yield 70%; H
1
NMR (300 MHz, CDCl3): d 3.64 (br, 6H), 3.74–3.78 (m,
12H), 3.85–3.92 (m, 4H), 4.16–4.19 (m, 2H), 6.11 (d, JZ
8.7 Hz, 1H), 6.85 (s, 2H), 7.07 (s, 1H), 7.11–7.20 (m, 5H),
7.55–7.64 (m, 8H), 8.39 (d, JZ8.7 Hz, 1H), 8.49–8.51 (m,
4H); 13C NMR (75 MHz, CDCl3): d 43.9, 58.7, 60.2, 67.5,
68.5, 70.1, 70.8, 71.1, 98.9, 113.8, 122.0, 122.2, 122.9,
128.7, 130.7, 136.5, 136.6, 140.7, 144.1, 144.4, 149.2,
158.9, 159.8; FAB MS m/z 811 [MCHC].
3.1.5. Synthesis of compound 1. A methanolic solution of
compound 9 (0.187 g, 0.25 mmol) and an aqueous solution
of zinc nitrate (0.1521 g, 0.512 mmol) were mixed and
stirred for 0.5 h. The solvent was removed and the residue
lyophilized to afford the complex 1 in quantitative yield.
3.1.10. Preparation of compound 2. Synthesized accord-
ing the procedure given in Section 3.1.5.
3.1.11. Synthesis of compound 14. 3-Hydroxy benzyl-
alcohol
(1 mmol),
N-(4-bromobutyl)phthalimide
3.1.6. Synthesis of compound 10. 3,5-Bis-hydroxymethyl
phenol (1.35 g, 8.74 mmol), N-Boc-{2-[2-(2-bromoethoxy)-
ethoxy]-ethyl}monoamine (3.0 g, 9.61 mmol) and K2CO3
(6.63 g, 48.05 mmol) were mixed in acetonitrile and
refluxed for 24 h. The reaction mixture was then cooled to
room temperature and the solvent removed. The residue was
dissolved in CH2Cl2 and the insoluble residue filtered off.
The filtrate was washed with water and dried over MgSO4.
The pure product was obtained as a semi-solid after
purification on a silica column using EtOAc. Yield 89%;
1H NMR (300 MHz, CDCl3): d 1.44 (s, 9H), 3.05–3.30 (br,
2H), 3.30 (br, 2H), 3.45–3.50 (m, 2H), 3.55–3.70 (m, 4H),
3.78 (m, 2H), 4.05 (m, 2H), 4.53 (s, 4H), 5.20 (br, 1H), 6.73
(s, 2H), 6.83 (s, 1H); 13C NMR (75 MHz, CDCl3): d 28.6,
40.8, 60.6, 65.1, 67.6, 70.0, 70.5, 71.0, 79.6, 112.3, 117.9,
143.1, 156.3, 159.4.
(1.01 mmol) and K2CO3 (5 mmol) were mixed in aceto-
nitrile and refluxed for 12 h. The work up and purification
were done according to procedure given in Section 3.1.1.
1
Yield 67%; H NMR (300 MHz, CDCl3): d 1.96–2.05 (m,
4H), 2.39 (br, 1H), 3.91 (m, 2H), 4.15 (t, JZ5.7 Hz, 2H),
4.80 (s, 2H), 6.93–6.96 (m, 1H), 7.05–7.07 (m, 2H), 7.36–
7.41 (m, 1H), 7.85–7.89 (m, 2H), 7.96–8.01 (m, 2H); 13C
NMR (75 MHz, CDCl3): d 25.4, 26.7, 37.7, 65.2, 67.2,
113.0, 113.8, 119.2, 123.3, 129.6, 132.1, 134.1, 142.8,
159.2, 168.6.
3.1.12. Synthesis of compound 15. Compound 14
(1 mmol), CBr4 (1.1 mmol) were dissolved in dry CH2Cl2
and Ph3P (1.05 mmol) dissolved in dry CH2Cl2 slowly
added over a period of 15 min. The reaction mixture was
stirred at room temperature for 12 h. The solvent was
removed and the residue chromatographed on a silica
column using gradient elution with EtOAc/Hex (1:1). Yield
3.1.7. Synthesis of compound 11. Synthesized according to
procedure given in Section 3.1.2. Yield 70%; 1H NMR
(300 MHz, CDCl3): d 1.44 (s, 9H), 3.34 (m, 2H), 3.56 (m,
2H), 3.64–3.67 (m, 2H), 3.71–3.74 (m, 2H), 3.85–3.88 (m,
2H), 4.15–4.18 (m, 2H), 4.43 (s, 4H), 5.00 (br, 1H), 6.90 (s,
2H), 7.01 (s, 1H); 13C NMR (75 MHz, CDCl3): d 28.6, 33.0,
40.0, 67.6, 69.7, 70.5, 71.0, 115.6, 122.6, 140.0, 159.0.
1
85%; H NMR (300 MHz, CDCl3): d 1.83–1.92 (m, 4H),
3.78 (m, 2H), 4.00 (m, 2H), 4.45 (s, 2H), 6.79–6.83 (m, 1H),
6.90 (m, 1H), 6.94–6.96 (m, 1H), 7.20–7.25 (m, 1H),
7.70–7.75 (m, 2H), 7.84–7.87 (m, 2H); 13C NMR (75 MHz,
CDCl3): d 25.6, 26.8, 33.8, 37.9, 67.4, 114.9, 115.3, 121.5,
123.5, 130.0, 132.4, 134.2, 139.5, 159.4, 168.7.
3.1.8. Synthesis of compound 12. Synthesized according to
3.1.13. Synthesis of compound 16. Compound15(1 mmol),
K2CO3 (5 mmol) and 2,20-dipicolylamine (1.2 mmol) were
procedure given in Section 3.1.3. Yield 87%; 1H NMR

11314
C. Lakshmi et al. / Tetrahedron 60 (2004) 11307–11315
mixed in dry DMF under N2 atmosphere and the reaction
mixture stirred for 5 h. The work up and purification were
carried out according procedure given in Section 3.1.3. Yield
93%; 1H NMR (300 MHz, CDCl3): d 1.86–1.91 (m, 4H), 3.67
(s, 2H), 3.75–3.79 (m, 2H), 3.83 (s, 4H), 3.97–4.01 (m, 2H),
6.74–6.77 (m, 1H), 6.97–7.00 (m, 2H), 7.12–7.23 (m, 3H),
7.59–7.74 (m, 6H), 7.83–7.87 (m, 2H), 8.51–8.53 (m, 2H);
13C NMR (75 MHz, CDCl3): d 25.6, 26.9, 37.9, 58.6, 60.0,
67.3, 113.3, 115.3, 121.4, 122.3, 123.2, 123.4, 129.5, 132.3,
134.2, 136.8, 149.1, 159.3, 168.7.
(300 MHz, DMSO-d6): d 3.59–3.72 (m, 16H), 4.01–4.04
(m, 2H), 6.44 (d, JZ9 Hz, 1H), 6.75–6.78 (m, 1H), 6.94–
6.97 (m, 2H), 7.18–7.26 (m, 3H), 7.56 (d, JZ7.8 Hz, 2H),
7.75–7.80 (m, 2H), 8.45–8.49 (m, 3H), 9.44 (br, 1H); 13C
NMR (75 MHz, DMSO-d6): d 43.4, 57.3, 59.0, 66.9, 68.0,
69.0, 69.9, 99.5, 112.8, 114.6, 120.8, 122.2, 122.4, 129.3,
136.6, 137.8, 140.3, 144.4, 145.3, 148.8, 158.4, 159.1; FAB
MS m/z 600 [MCHC].
3.1.20. Synthesis of compound 4. Synthesized according to
procedure given in Section 3.1.15.
3.1.14. Synthesis of compound 17. Synthesized according
to procedure given in Section 3.1.4. Yield 72%; H NMR
1
3.2. Vesicle preparations
(300 MHz, DMSO-d6): d 1.83 (br, 4H), 3.54 (br, 2H), 3.60
(s, 2H), 3.70 (s, 4H), 4.01 (m, 2H), 6.42 (d, JZ9 Hz, 1H),
6.68–6.81 (m, 1H), 6.95–6.97 (m, 2H), 7.19–7.26 (m, 3H),
7.56 (d, JZ7.5 Hz, 2H), 7.75–7.80 (m, 2H), 8.47–8.49 (m,
3H), 9.58 (m, 1H); 13C NMR (75 MHz, DMSO-d6): d 24.4,
26.1, 43.0, 57.3, 59.0, 66.9, 99.1, 113.0, 114.7, 120.7, 122.2,
122.5, 129.3, 136.6, 137.9, 140.1, 144.2, 144.4, 145.2, 148.8
158.6, 158.9; FAB MS m/z 540 [MCHC].
All lipids were purchased from Avanti Polar Lipids, Inc.
(Alabaster, AL) and stored as 10 mg/mL stock solutions in
CHCl3 at K20 8C. Lipids were added in the appropriate
ratios to a 10 mL round bottom flask and solvent removed
by rotary evaporation. Residual solvent was removed under
oil pump vacuum for R1 h. Dried lipid was then rehydrated
in TES buffer (5 mM TES, 145 mM NaCl, pH 7.4). A glass
pyrex ring was added to the flasks to ensure complete
removal of lipid from the flask walls, and then flasks were
vortexed vigorously. The resulting lipid dispersion was then
extruded approximately 30 times through a 19 mm diameter
polycarbonate membrane with 200 nm pore diameter. All
vesicles were prepared at room temperature and used the
same day.
3.1.15. Preparation of Zn complex 3. A methanolic
solution of compound 17 (1 mmol) and an aqueous solution
of zinc nitrate (1.025 mmol) were mixed and stirred for
0.5 h. The solvent was removed and the residue lyophilized
to afford the complex 3 in quantitative yield.
3.1.16. Synthesis of compound 18. 3-Hydroxy benzyl-
alcohol (1 mmol), N-Boc-{2-[2-(2-bromoethoxy)-ethoxy]-
ethyl}monoamine (1.01 mmol) and K2CO3 (5 mmol) were
mixed in acetonitrile and refluxed for 12 h. The work up and
purification were done according to procedure given in
3.3. Fluorescence spectroscopy
All fluorescence spectroscopy was performed on a Perkin
Elmer LS50B fluorimeter with FT WinLab software using
standard 1!1!5 cm3 cuvettes. NBD fluorescence was
measured using excitation and emission wavelengths of 470
and 530 nm, respectively. A 515 nm cutoff filter was used in
all NBD acquisitions. Carboxyfluoresecin fluorescence was
measured with monochromater excitation and emission
wavelengths of 495 and 520 nm, respectively, with an open
filter. All measurements were performed at 25 8C without
degassing of samples.
1
Section 3.1.6. Yield 89%; H NMR (300 MHz, CDCl3): d
1.46 (s, 9H), 2.01 (br, 1H) 3.32 (m, 2H), 3.56 3.58 (m, 2H),
3.61–3.76 (m, 4H), 3.87–3.90 (m, 2H), 4.12–4.20 (m, 2H),
4.65–4.69 (m, 2H), 5.09 (br, 1H), 6.85–6.88 (m, 1H), 6.95
(m, 2H), 7.26–7.31 (m, 1H); 13C NMR (75 MHz, CDCl3): d
28.6, 40.8, 65.4, 67.6, 70.0, 70.5, 71.0, 79.6, 113.3, 114.1,
119.6, 129.8, 143.0, 156.3, 159.3.
3.1.17. Synthesis of compound 19. Synthesized according
to procedure given in Section 3.1.12. Yield 65%; H NMR
1
3.4. Carboxyfluorescein leakage
(300 MHz, CDCl3): d 1.43 (s, 9H), 3.20 (br, 2H), 3.55 (m,
2H), 3.63–3.66 (m, 2H), 3.70–3.74 (m, 2H), 3.85–3.88 (m,
2H), 4.13–4.17 (m, 2H), 4.46 (s, 2H), 5.01 (br, 1H), 6.84–
6.88 (m, 1H), 6.95–7.00 (m, 2H), 7.23–7.27 (m, 1H); 13C
NMR (75 MHz, CDCl3): d 28.7, 33.7, 40.7, 67.7, 70.0, 70.6,
71.0, 115.1, 115.5, 121.8, 130.1, 139.4, 156.2, 159.2.
Vesicles of the appropriate composition were prepared as
described, with the following exceptions. Lipids were
rehydrated after drying with TES buffer (5 mM TES,
100 mM NaCl, pH 7.4) containing 50 mM (5–6)carboxy-
fluorescein. Following extrusion, vesicles were separated
from excess carboxyfluorescein using a Sephadex G-45
column. The fluorescence intensity of a 3 mL sample of
50 mM vesicles was monitored for 400 s. At time 100 s, an
appropriate volume of a 1 mM solution of 1 in TES buffer
was added to the cuvette, followed by Triton X-100
detergent (0.5%) at time 300 s. Zero and 100% leakage
was determined from the initial fluorescence and the
fluorescence after addition of detergent, respectively.
3.1.18. Synthesis of compound 20. Synthesized according
to procedure given in Section 3.1.13. Yield 87%; H NMR
1
(300 MHz, DMSO-d6): d 1.35 (s, 9H), 3.03–3.09 (m, 2H),
3.32–3.40 (m, 2H), 3.50–3.53 (m, 2H), 3.56–3.60 (m, 4H),
3.71–3.75 (m, 6H), 4.05–4.08 (m, 2H), 6.75–6.83 (m, 2H),
6.97–7.00 (m, 2H), 7.21–7.28 (m, 3H), 7.57–7.60 (m, 2H),
7.76–7.82 (m, 2H), 8.48–8.50 (m, 2H); 13C NMR (75 MHz,
DMSO-d6): d 28.2, 39.7, 57.3, 59.1, 66.9, 68.9, 69.2, 69.5,
69.9, 112.9, 114.6, 120.8, 122.2, 122.4, 129.3, 136.6, 140.4,
148.8, 155.6, 158.5, 159.1.
3.5. Vesicle titrations
Stock solutions of compounds 1, 2, 3 or 4 were diluted in
TES buffer to a final concentration of 1 mM in a 5 mL
cuvette. With stirring, aliquots of 10 mM phospholipid
3.1.19. Synthesis of compound 21. Synthesized according
to procedure given in Section 3.1.9. Yield 76%; H NMR
1

C. Lakshmi et al. / Tetrahedron 60 (2004) 11307–11315
11315
7. Laxman, B.; Hall, D. E.; Bhojani, S. B.; Hamstra, D. A.;
Chenevert, T. L.; Ross, B. D.; Rehemtulla, A. Proc. Natl.
Acad. Sci. U.S.A. 2002, 99, 16551–16555.
vesicles of the appropriate composition were sequentially
added to the solution to give the desired phospholipid
concentration over the range 0–100 mM. After each
addition, the fluorescence intensity was measured after a
twenty-second incubation. Curves of fluorescence intensity
(l530) versus available phospholipid concentration (taken as
60% of the total phospholipid concentration) were gener-
ated and fitted to a 1:1 binding model.28 An iterative curve-
fitting method yielded the apparent binding constants listed
in Tables 1 and 2, all of which represent the average of at
least three independent measurements.
`
¨
8. Biven, K.; Erdal, H.; Hagg, M.; Ueno, T.; Zhou, R.; Lynch, M.;
Rowley, B.; Wood, J.; Zhang, C.; Toi, M.; Shoshan,
M. C.; Linder, S. Apoptosis 2003, 8, 263–268.
9. Koopman, G.; Kuijten, R. M. J.; Keehnen, R. M. J.; Pals,
S. T.; van Oers, M. H. J. Blood 1994, 84, 1415–1420.
10. Vermes, I.; Haanen, C.; Steffens-Nakken, H.;
Reutelingsperger, C. J. Immunol. Methods 1995, 184, 39–51.
11. van Engeland, M.; Nieland, L. J. W.; Ramaekers, F. C. S.;
Schutte, B.; Reutelingsperger, C. P. M. Cytometry 1998, 31,
1–9.
3.6. Sensor permeation
12. Liemann, S.; Huber, R. CLMS, Cell. Mol. Life Sci. 1997, 53,
516–521.
To a 45 mL solution of 25 mM phospholipid vesicles
composed either of 100% POPC or 1:1 POPC-POPS,
compound 1 or 2 was added to a final concentration of either
0.5 or 2.0 mol%. A 3.0 mL aliquot of the resulting solution
was removed at predetermined intervals over a two-hour
period and placed in a 5 mL cuvette. Fluorescence intensity
of each aliquot was monitored over a 200 s interval. At 50 s,
sodium dithionite was added to the cuvette to a final
concentration of 55 mM. At 150 s, 0.2% octaethylene glycol
monododecyl ether was added to disrupt vesicle membranes
and expose any internalized NBD-labeled compounds to the
aqueous environment.
13. Gerke, V.; Moss, S. E. Physiol. Rev. 2002, 82, 331–371.
14. Kamp, D.; Sieberg, T.; Haest, C. W. M. Biochemistry 2001,
40, 9438–9446.
15. Zweifach, A. Biochem. J. 2000, 349, 255–260.
16. Sairjo, M. A.; Concha, N. M.; Dedman, J. R.; Seaton,
B. A. Nat. Struct. Biol. 1995, 2, 968–974.
17. Czarnik, A. W. Acc. Chem. Res. 1994, 27, 302–308.
18. Matsuda, S.; Ishikubo, A.; Kuzuya, A.; Yashiro, M.;
Komiyama, M. Angew. Chem., Int. Ed. 1998, 37, 3284–3286.
19. Lee, D. H.; Im, J. H.; Son, S. U.; Chung, Y. K.; Hong, J. I.
J. Am. Chem. Soc. 2003, 125, 7752–7753.
20. Ojida, A.; Mito-oka, Y.; Inoue, M.; Hamachi, I. J. Am. Chem.
Soc. 2002, 124, 6256–6258.
Acknowledgements
21. Koulov, A. V.; Stucker, K. A.; Lakshmi, C.; Smith, B. D. Cell
Death Differ. 2003, 10, 1357–1359.
This work was supported by the National Institutes of
Health, the Department of Defense, and the Phillip Morris
External Research Program.
22. Ashton, R. A.; Anderson, D. W.; Brown, C. L.; Shipway,
A. N.; Stoddart, J. F.; Tolley, M. S. Chem. Eur. J. 1998,
4, 781–795.
23. Delphine, F.; Gutierrez, N. M.; del Pilar, M.; Jean-Francois,
G.; Michaeal, L.; Cornie, S.; Patrick, M.; Jean-Louis, G.;
Benoit, H.; Daniel, G.; Jean-Francois, N. Helv. Chim. Acta
2002, 85, 288–319.
References and notes
24. Lancet, D.; Pecht, I. Biochemistry 1977, 16, 5150–5157.
25. Lin, Su.; Struve, W. S. Photochem. Photobiol. 1991, 54,
361–365.
1. Sgonc, R.; Gruber, J. Exp. Gerentol. 1998, 33, 525–533 and
references cited therein.
26. Ojida, A.; Mito-oka, Y.; Sada, K.; Hamachi, I. J. Am. Chem.
Soc. 2004, 126, 2454–2463.
2. Gorczyca, W. Endocr-Relat. Cancer 1999, 6, 17–19.
3. Brauer, M. Prog. Neuro-Psychoph. 2003, 27, 323–331.
4. Waterhouse, N. J.; Trapani, J. A. Cell Death Differ. 2003, 10,
853–855.
27. Borisenko, G.; Matsura, T.; Liu, S.; Tyurin, V. A.; Jainfei, J.;
Serinkan, F. B.; Kagan, V. E. Arch. Biochem. Biophys. 2003,
413, 41–52.
5. Stadelmann, C.; Lassmann, H. Cell Tissue Res. 2000, 301,
19–31.
6. Lecoeur, H. Exp. Cell Res. 2002, 277, 1–14.
28. Xie, H.; Wu, S. J. Chem. Soc., Perkin Trans. 2 1999,
2751–2754.