Facilitated phosphatidylserine flip-flop across vesicle and cell membranes using urea-derived synthetic translocases

Article abstract of DOI:10.1039/b314006g

Tris(2-aminoethyl)amine derivatives with appended urea and sulfonamide groups are shown to facilitate the translocation of fluorescent phospholipid probes and endogenous phosphatidylserine across vesicle and erythrocyte cell membranes. The synthetic translocases appear to operate by binding to the phospholipid head groups and forming lipophilic supramolecular complexes which diffuse through the non-polar interior of the bilayer membrane.

Full text of DOI:10.1039/b314006g

View Article Online / Journal Homepage / Table of Contents for this issue
Facilitated phosphatidylserine flip-flop across vesicle and cell
membranes using urea-derived synthetic translocases
Yoshihiro Sasaki,a,b Rameshwer Shuklaa and Bradley D. Smith*a
a Department of Chemistry and Biochemistry, University of Notre Dame,
251 Nieuwland Science Hall, Notre Dame, IN 46556-5670, USA. E-mail: smith.115@nd.edu
b Graduate School of Materials Science, Nara Institute of Science and Technology,
8916-5 Takayama, Ikoma, Nara 630-0101, Japan
Received 5th November 2003, Accepted 27th November 2003
First published as an Advance Article on the web 10th December 2003
Tris(2-aminoethyl)amine derivatives with appended urea and sulfonamide groups are shown to facilitate the
translocation of fluorescent phospholipid probes and endogenous phosphatidylserine across vesicle and erythrocyte
cell membranes. The synthetic translocases appear to operate by binding to the phospholipid head groups and
forming lipophilic supramolecular complexes which diffuse through the non-polar interior of the bilayer membrane.
Introduction
Phospholipids are asymmetrically distributed across biological
membranes; the aminophospholipids, such as phosphatidyl-
serine (PS), are enriched on the cytoplasmic surface of animal
plasma membranes, while choline-based lipids are enriched on
the outer surface.1,2 The phospholipid asymmetry is vital
not only for exocytosis and intracellular fusion processes, but
also for membrane–protein interactions and signal trans-
duction pathways.3–6 In model vesicle membranes, spontaneous
membrane translocation (or flip-flop, see Scheme 1) of
phospholipids is a very slow process with a typical half-life of
hours to days.7 In cell membranes, the asymmetric distribution
is generated and maintained by a family of membrane-bound
transport proteins called phospholipid translocases. The
best known translocase is the aminophospholipid flippase
which pumps phosphatidylserine to the membrane inner
monolayer.8,9
Scheme 2
multitopic synthetic receptor with appropriate charge and
lipophilicity.15 While compound 1 has no PS-scramblase activ-
ity, we recently screened a library of tren-derived receptors and
uncovered some active derivatives with appended urea groups.16
This unexpected observation has motivated us to prepare some
simple tren-urea derivatives and evaluate them as synthetic
phospholipid translocases and more specifically as PS-scram-
Scheme 1 Phospholipid translocation or flip-flop.
blases. Here, we report that tren derivatives 2 and 3 (Scheme 2)
with appended urea and sulfonamide groups are remarkably
effective at promoting the translocation of phosphatidylcholine
and phosphatidylserine across vesicle and cell membranes. We
also characterize the anion and proton binding properties of
Our research group is attempting to design synthetic trans-
locases for use as pharmaceuticals or as chemical tools for bio-
logical membrane research.10–16 For instance, we reported that
tris(2-aminoethyl)amine (tren) sulfonamide 1 (Scheme 2) facili-
these synthetic translocases, and propose a translocation
tates phosphatidylcholine translocation across synthetic vesicle
mechanism that is consistent with the data.
and erythrocyte membranes. Mechanistic studies indicate that 1
forms a supramolecular complex with the phosphatidylcholine
Results
head group, which diminishes head group hydrophilicity
and promotes diffusion through the non-polar interior of the
bilayer membrane. More recently we have focused on synthetic
translocases that facilitate phosphatidylserine flip-flop. In
particular, our aim is to develop PS-scramblases that raise the
concentration of phosphatidylserine on the cell surface, which
in turn is known to induce various biological processes such as
blood coagulation and cell clearance. Our previous work has
shown that the facilitated translocation of phospholipids with
multiply charged head groups, like phosphatidylserine, is a dif-
ficult supramolecular challenge that requires a complementary,
Synthesis
Tren sulfonamide 1 was prepared using a previously reported
method.10–12 Mono-urea derivatives 2 and 3 were synthesized in
three steps (Scheme 3). Treatment of tris(2-aminoethyl)amine
with di-tert-butyl dicarbonate gave the desired mono-Boc-
protected 4 in 69% yield. Reaction of 4 with toluenesulfonyl
chloride and subsequent deprotection with TFA gave bisulfon-
amide 5, which was reacted with the appropriate aryl-
isocyanate to give 2 or 3.
O r g . B i o m o l . C h e m . , 2 0 0 4 , 2, 2 1 4 – 2 1 9
T h i s j o u r n a l i s © T h e R o y a l S o c i e t y o f C h e m i s t r y 2 0 0 4
214

View Article Online
amount as NBD-PS (Table 1). Concentration studies showed
that the rate of translocation increases as the concentration of 2
or 3 is raised. Control experiments verified that compounds 1–3
do not induce leakage of entrapped carboxyfluorescein from
the vesicles.
In addition to enhanced phosphatidylserine translocation
across vesicle membranes, mono-urea derivatives 2 and 3 can
facilitate the inward-translocation of phosphatidylserine across
erythrocyte cell membranes. The inward-translocation assay
was initiated by treating a population of erythrocytes with a
small aliquot of PS-NBD (1% of total phospholipid). The
probe molecules that did not insert into the outer monolayer of
the erythrocyte membrane were removed by washing. Sub-
sequent translocation of the probe to the inner monolayer of
the erythrocyte membrane was measured by removing a small
sample of cells at periodic intervals and using the above
mentioned NBD/dithionite assay. When the translocase candi-
date was added as an ethanolic solution to human erythrocytes,
less than 2% of hemolysis19 occurred over a 6 h time period.
Therefore, it appears that these compounds do not compromise
membrane integrity. To ensure that dithionite did not enter
the erythrocytes and quench the probe on the inside, it was
necessary to inhibit the band 3 anion exchanger protein by
pretreating the cells with 4,4Ј-diisothiocyanatostilbene-2,2Ј-
disulfonate.20,21 In addition, the cells were pretreated with
the aminophospholipid flippase inhibitor N-ethylmaleimide
(NEM) which prevents endogenous, enzyme-mediated, trans-
location of phosphatidylserine to the membrane inner mono-
layer.8 The data in Fig. 2 clearly show that urea-derived 2 and 3
can significantly facilitate inward translocation of NBD-PS
probe as compared to sulfonamide 1, although the activity of 3
in erythrocytes is lower than in vesicles.
Scheme 3
Flip into vesicles and cells
Phospholipid translocation was monitored via the well
established NBD/dithionite quenching assay which uses fluor-
escent phospholipid probes containing a 7-nitrobenz-2-oxa-
1,3-diazol-4-yl (NBD) group in one of its acyl chains (Scheme
2).10,17,18 Exo-labeled vesicles were prepared by addition of a
small aliquot of NBD-lipid (2 mol% of total phospholipid) in
ethanol to a dispersion of vesicles. The NBD-lipid readily
inserts into the outer monolayer of fluid-phase vesicles. Upon
treatment with sodium dithionite (Na2S2O4), the NBD fluor-
escence is quenched due to reduction of the NBD nitro group.
Bilayer membranes are effectively impermeable to dithionite;
therefore, only NBD-phospholipid located in the outer leaflet is
chemically quenched. The inward translocation experiment is
initiated by adding a small aliquot of synthetic translocase can-
didate to the exo labeled-vesicles. At any given time, the fraction
of probe located in the outer monolayer can be determined
from the drop in fluorescence intensity when a portion of the
vesicles is subjected to dithionite quenching.
Inward translocation of NBD-PS induced by addition of
12.5 µM of one of the synthetic translocase candidates 1, 2 or 3
was measured using vesicles composed of POPC/cholesterol
(7 : 3, 25 µM total lipid concentration). The background rate for
inward translocation of NBD-PS is very slow, with a trans-
location half-life of much greater than 3 h. Addition of sulfon-
amide 1 produces no enhancement of inward translocation,
whereas the presence of mono-urea derivatives 2 or 3 produces
a significant rate increase with translocation half-lives of 12 min
and 60 min respectively (Fig. 1). Also, compounds 2 and 3
facilitated NBD-PC translocation by approximately the same
Fig. 2 Inward translocation of NBD-PS (1 mol%) across erythrocyte
membranes in the presence of 40 µM 2 (᭿), 3 (᭹), or no addition (ꢀ).
The cells were pretreated with the aminophospholipid flippase inhibitor
N-ethylmaleimide (NEM, 10 mM) which prevents endogenous
translocation of phosphatidylserine.
Phosphatidylserine scrambling in blood cells
Flow cytometry experiments show that compound 3 can
scramble the distribution of endogenous phosphatidylserine
across erythrocyte membranes. The appearance of phospha-
tidylserine on the cell surface following treatment with
compound 3 (10 µM, 3 h, 1.5 × 107 cells cmϪ3) was detected
using the fluorescein-labeled, PS-binding protein, annexin V
(annexin V-FITC). The distribution of fluorescence intensity
from a representative flow cytometry experiment is shown
in Fig. 3. Results obtained with normal erythrocytes are
represented in the top half of Fig. 3(a,b), while the cells in the
bottom half of Fig. 3 (c,d) were pretreated with the amino-
phospholipid flippase inhibitor (NEM). As expected, hardly
Fig. 1 Inward translocation of NBD-PS (2 mol%) across POPC/
cholesterol (7 : 3) vesicles (25 µM) in the presence of 12.5 µM 1 (ꢀ),
2 (᭿), or 3 (᭹) at pH 7.4, 25 ЊC.
O r g . B i o m o l . C h e m . , 2 0 0 4 , 2, 2 1 4 – 2 1 9
215

View Article Online
Table 1 Half-lives for inward translocation of NBD-lipid facilitated by translocase candidates and phospholipid association constants
Lipid
Compound
Translocation half-lifea/min
K/104 MϪ1
NBD-PS
1
2
3
1
2
3
>180
60
12
>180
60
14
—
0.12 ( 0.02)b
1.9 ( 0.2)b
—
NBD-PC
—
2.2 ( 0.1)c
a Inward translocation of 2 mol% NBD-lipid across POPC/Cholesterol (7 : 3) vesicles (25 µM) in the presence of 12.5 µM synthetic translocase
candidates at pH 7.4, 25 ЊC. b Association with POPS in 99 : 1 CHCl3/CH3OH at 25 ЊC. c Association with POPC in 99 : 1 CHCl3/CH3OH at 25 ЊC.
Fig. 4 Change in 1H NMR chemical shift (δ) for NCH2 signal in 1 (ꢀ)
and 2 (᭹) in 9 : 1 CD3OD/D2O.
field shifts of aryl urea NH protons, alkyl urea NH protons,
Fig. 3 Flow cytometry analysis of annexin V-FITC binding to normal
and sulfonamide NH protons, suggesting that both the urea
(a,b) or NEM treated (c,d, 10 mM) erythrocytes. Before exposure to
and sulfonamide moieties are involved in the anion binding as
annexin V-FITC, the erythrocytes in (b) and (d) were incubated at 37 ЊC
shown in Scheme 4. Titration isotherms were generated by
monitoring aryl urea NH signals (Fig. 5), and the association
constants were determined by fitting the curves to a 1 : 1 bind-
ing model using iterative computer methods. In the case of
for 3 h with 10 µM of 3.
any normal cells (which sequester most of the phospha-
tidylserine in the inner monolayer of the membrane) were
stained by annexin V-FITC (Fig. 3a), but 34% of the cells were
stained after treatment with compound 3 (Fig. 3b). If the
endogenous aminophospholipid flippase is inhibited with
NEM, then the PS-externalization effect induced by compound
3 is approximately double (Fig. 3d). The lack of hemoglobin
leakage (< 3% leakage over 3 h) indicates that the compounds
do not induce nonselective membrane transport.
Ϫ
CH3CO2 the 1 : 1 binding stoichiometry was confirmed by a
pKa measurements
The translocase candidates are only moderately soluble in pure
water. Therefore, the pKa values of their protonated forms were
determined by 1H NMR titration method in 9 : 1 CD3OD/D2O
(Fig. 4), and found to be 4.01 ( 0.04) for sulfonamide 1 and
4.91 ( 0.01) for mono-urea derivative 2. These values are in
line with previous measurements on related systems,12 and sug-
gest that the translocases exist predominantly as their free bases
in the bulk solution of the phospholipid translocation assay
(pH = 7.4).
Scheme 4 Supramolecular complex of 3 with anion.
Anion and phospholipid binding
The translocation data suggest that tren-urea derivatives 2 and
3 are able to form hydrogen-bonded complexes with the phos-
pholipid head groups. To gain structural insight into the com-
plexes, a series of NMR titration studies were conducted. The
anion binding ability of the most active translocase 3 was
1
investigated by H NMR spectroscopy in DMSO-d6, a highly
polar solvent that minimizes ambiguities due to inter- and
intramolecular aggregation. Compound 3 was titrated with
tetrabutylammonium salts of H2PO4Ϫ, CH3CO2Ϫ, and ClϪ. In
all cases the addition of anionic guests resulted in large down
Fig. 5 Change in 1H NMR chemical shift of the aryl urea NH signal
for translocase 3 (∆δ) upon addition of tetrabutylammonium salts of
CH3CO2Ϫ (᭹), H2PO4Ϫ, (᭿) and ClϪ (ꢀ) in DMSO-d6 and at 25 ЊC.
O r g . B i o m o l . C h e m . , 2 0 0 4 , 2, 2 1 4 – 2 1 9
216

View Article Online
Job plot (Fig. 6). The observed association constants are
to complexes of composition [3 ϩ DHPSϪ ϩ 2Hϩ] (m/z 1074.4),
Ϫ
Ϫ
CH3CO2 (4500 MϪ1), H2PO4 (900 MϪ1), and ClϪ (32 MϪ1).
[3 ϩ DHPSϪ ϩ Hϩ ϩ Naϩ] (m/z 1096.4), and [3 ϩ DHPSϪ
ϩ
Ϫ
The relatively strong affinity for CH3CO2 is attributed to its
2Naϩ] (m/z 1118.4). Thus, mass spectrometry shows that 3 can
bind to the anionic head group of DHPS, and that this binary
complex can also associate with Hϩ or Naϩ ions.
shape and relatively high basicity.
Discussion
We have previously demonstrated that tren-sulfonamide 1 can
facilitate the membrane flip-flop of phosphatidylcholine. Com-
pound 1 associates with the phosphate diester residue in the
zwitterionic phosphatidylcholine head group, which diminishes
its hydrophilicity and promotes diffusion across the lipophilic
interior of the membrane.2 Compound 1 is unable to accelerate
the flip-flop of anionic phosphatidylserine, presumably because
it does not provide an appropriate level of lipophilicity and
charge balance.
Tren-urea derivatives 2 and 3 have strong affinities for anions
and it is not surprising that they can facilitate phospha-
tidylcholine flip-flop (Table 1). Less easily rationalized are their
impressive abilities to promote phosphatidylserine flip-flop.
Compound 3 with its more acidic urea NH residues, has a
stronger affinity for phospholipid head groups and is a more
efficient translocase. Not only can compound 3 translocate
NBD-PS into vesicle and cells, it can promote phosphatidyl-
serine flop, or translocation from inside to outside. In other
words, it can act as a synthetic PS-scramblase. The evidence
presented in this paper combined with the results of our pre-
vious studies suggest that compounds 2 and 3 operate by form-
ing a supramolecular complex with the phosphatidylserine head
group, and that the complex diffuses across the bilayer mem-
brane. Charge balance demands that facilitated translocation
of anionic phosphatidylserine, by 2 or 3, must also include a
cationic residue in the kinetically active complex. The mass
spectrometry data suggests that this is likely to be a Hϩ or Naϩ.
We envision that compounds 2 and 3 promote phosphatidyl-
serine scrambling across the membranes of red blood cells in
the following way. After addition to the sample, the neutral
form of 2 or 3 partitions into the cell membrane and diffuses to
the inner leaflet either by itself or hydrogen bonded to a phos-
pholipid like phosphatidylcholine. It can return to the outer
leaflet as a supramolecular complex with anionic phosphatidyl-
serine and a balancing cation like Hϩ or Naϩ.
Fig. 6 Job plot for translocase 3 and tetrabutylammonium acetate by
1H NMR in DMSO-d6 at 25 ЊC. [translocase] ϩ [acetate] = 5 mM.
Compounds 2 and 3 were also evaluated for their abilities to
associate with phospholipid head groups. Attempts to measure
the binding of phospholipids by 1H NMR in weakly polar
solvents were hampered by self-aggregation effects, however,
the changes in chemical shifts were consistent with hydrogen-
bonding between the phospholipid head groups and the urea
and sulfonamide NH groups. The self-aggregation problems
were overcome by conducting UV titration studies in 99 : 1
CHCl3/CH3OH at 25 ЊC. The typical titration isotherm shown
in Fig. 7 was generated by monitoring the change in UV
absorption of the N-aryl urea groups in 3 upon addition of
POPS, and the association constant was determined by fitting
the curve to a 1 : 1 binding model using an iterative computer
method. Compounds 2 and 3 bind POPS with association con-
stants of (0.12 0.02) × 104 MϪ1 and (1.9 0.2) × 104 MϪ1
,
respectively. The stronger binding exhibited by 3 is attributed to
the increased acidity of its urea NH residues. The data listed in
Table 1 shows that an increase in phospholipid association con-
stant correlates with a decrease in NBD-lipid translocation
half-life.
Conclusion
We have shown that tren-urea derivatives 2 and 3 can greatly
facilitate the inward and outward translocation of phos-
pholipids, including phosphatidylserine and its fluorescent
derivative, NBD-PS, across vesicle and erythrocyte membranes.
Furthermore, these synthetic translocases can function as
PS-scramblases; that is, they can work against the inward
pumping action of the endogenous aminophospholipid flippase
and increase the amount of phosphatidylserine on the cell sur-
face. The structural simplicity of these compounds makes them
attractive candidates for further development as either chemical
tools for biological membrane research, or perhaps eventually
as chemotherapuetic agents. From a molecular design perspec-
tive, it is becoming increasingly clear that the construction of a
successful synthetic translocase for phosphatidylserine should
include a relatively large hydrogen-bonding pocket that can effi-
ciently desolvate the multiply-charged phosphatidylserine head
group.15,16
Fig. 7 Change in absorbance (∆abs) upon titration of 3 (15 µM) with
POPS in 99 : 1 CHCl3/CH3OH at 25 ЊC and its curve fitting using a 1 : 1
binding model.
Additional structural information was obtained by using
electrospray ionization mass spectrometry to analyze an equi-
molar mixture of 3 and DHPS, a phosphatidylserine derivative
with short C6 acyl chains (structure shown in Scheme 2). In the
negative ion mode, the base peak corresponds to a 1 : 1 complex
of 3 with DHPS ([3 ϩ DHPSϪ], m/z 1072.4). In the positive ion
mode, three signals are detected, whose m/z ratios correspond
Experimental
General
Compounds 1,10 4 and 512 were prepared as reported previously.
All lipids were purchased from Avanti Polar Lipids, Inc (USA).
O r g . B i o m o l . C h e m . , 2 0 0 4 , 2, 2 1 4 – 2 1 9
217

View Article Online
Synthesis of urea, 2
ethanolic solution of NBD-lipid (final NBD-lipid concen-
tration was 0.125 µM) to a 45 cm3 solution of POPC/cholesterol
(7 : 3) vesicles (final total lipid concentration was 25 µM) at
room temperature. To each vesicle solution, a synthetic scram-
blase candidate was added from a ethanolic stock solution
(solvent alone does not induce translocation). Over the time
course of 3 h, a series of 3 cm3 aliquots were removed and
assayed for extent of translocation. The 200 s assay consisted of
a dithionite injection (60 mM, 0.18 cm3 of a 1 M solution) at
t = 50 s and a Triton X-100 injection (0.5%, 0.075 cm3 of a 20%
solution) at t = 150 s. Data points were collected every second.
The percentage of exo NBD-lipid located in the vesicle outer
monolayer was calculated according to the following equation,
(%exo NBD-lipid) = [(Fi Ϫ Ff)/Fi] × 100, where Fi and Ff are the
intensities just prior to the addition of dithionite and Triton
X-100, respectively. All % exo NBD-lipid values contain 5%
error. In some cases, the translocation curves appeared to have
biexponential character, but because of the uncertainty in the
data a double exponential analysis was not attempted. Instead,
the reported translocation half-lives simply indicate the time
taken to reach 80% exo NBD-lipid, which is halfway toward an
equilibrium value of 60% exo NBD-lipid.
A solution of 4-methylphenylisocyanate (0.020 g, 0.16 mmol) in
dichloromethane (10 cm3) was added dropwise to a stirred solu-
tion of 5 ( 0.075 g, 0.16 mmol) in dichloromethane (30 cm3) at
ice bath temperature. The mixture was allowed to stir overnight
at room temperature. Removal of solvents and purification by
chromatography on silica (EtOAc : hexane, 8 : 2) gave a color-
less oil (0.055 g, 59%) δH(300 MHz; CDCl3; Me4Si) 2.35 (6H, s,
ArCH3), 2.43 (6H, m, NCH2CH2), 2.98 (4H, m, NCH2CH2-
SO2Ar), 3.21 (2H, m, NCH2CH2NHCO), 5.98 (1H, s, br,
CH2NHCONHAr), 6.20 (2H, s, br, CH2NHSO2Ar), 6.98 (1H,
t, J 6.0, ArH ), 7.21 (6H, m, ArH), 7.42 (2H, d, J 12.0, ArH ),
7.49 (1H, s, br, CH2NHCONHAr) and 7.76 (2H, d, J 11.9,
ArH ); δC (75 MHz; CDCl3; Me4Si) 21.6, 37.6, 41.3, 54.1, 119.4,
122.6, 127.2, 129.0, 130.0, 136.7, 139.7, 143.7 and 156.5;
HRMS m/z (FAB) 574.2189 (Mϩ. C27H35N5O5S2 requires
574.2189).
Synthesis of urea, 3
The above procedure was used to prepare 3 as a colorless oil
(0.058 g, 43%) δH(300 MHz; CDCl3; Me4Si) 2.36 (6H, s,
ArCH3), 2.54 (6H, m, NCH2CH2), 3.01 (4H, m, NCH2CH2-
SO2Ar), 3.26 (2H, m, NCH2CH2NHCO), 6.04 (2H, s, br,
CH2NHSO2Ar), 6.16 (1H, s, br, CH2NHCONHAr), 7.23 (4H,
d, ArH, J12.0), 7.55 (4H, d, ArH, J 12.0), 7.73 (2H, d, ArH,
J 9.0), 8.01 (1H, s, CH2NHCONHAr) and 8.05 (2H, d, ArH,
J 9.0); δC (75 MHz; CDCl3; Me4Si) 22.1, 38.7, 42.0, 54.4, 118.4,
125.7, 127.7, 130.6, 144.6, 147.0 and 155.8; HRMS m/z (FAB)
619.1997 (Mϩ. C27H35N6O7S2 requires 619.2009).
Carboxyfluorescein leakage assay
Excitation was set at 495 nm, and fluorescence emission was
measured at 520 nm using an open filter. Vesicles encapsu-
lating carboxyfluorescein were freeze-thawed 10 times prior to
extrusion. The fluorescence of a 3 cm3 sample of vesicles
(25 µM) in TES buffer was monitored. After 50 s, an aliquot of
translocase (40 µM) was added. Detergent (0.5% Triton X-100)
was added at t = 250 s. Total assay time was 300 s. A constant
fluorescent intensity up until the point of detergent addition
indicated no vesicle leakage.
pKa determination
Translocase candidates were dissolved in 9 : 1 CD3OD/D2O
(1 mM, 0.75 cm3), and 4 molar equivalents of DCl were added
to fully protonate the tertiary amine. Chemical shifts of the
NCH2and NCH2CH2 protons were monitored upon the addi-
tion of small aliquots of NaOD stock solution (0.1 M). A pD
reading was also recorded after each addition. The changes in
chemical shift were plotted against the pD values, and the pKa
values were extracted using a curve fitting procedure.12
Hemolysis assay19
Absorbance was monitored from 600 to 300 nm. A series of
1 cm3 samples was prepared: 0% hemolysis (only buffer), 100%
hemolysis (only H2O), 1 (40 µM in buffer), 2 (40 µM in buffer),
and 3 (40 µM in buffer). To each, 100 µL of packed red blood
cells was added (10% hematocrit). Samples were incubated at
37 ЊC for varying time periods before removal of 100 µL
aliquots for hemolysis analysis. The aliquots were centrifuged
for 10 min at 1000 rpm, and the supernatant was diluted to
1 cm3 with buffer before recording the absorbance of the hemo-
globin in solution (A at 414 nm). The percent hemolysis was
calculated according to the following equation, %hemolysis =
100 Ϫ [(A100% Ϫ A)/(A100% Ϫ A0%) × 100].
Vesicle preparation
A film of the vesicle lipids was dried under vacuum for at least
1 h. Hydration was performed at room temperature with an
appropriate amount of TES buffer (5 mM TES, 100 mM NaCl,
pH 7.4). Multilamellar vesicles were generated using a Vortex
mixer; use of a glass bead ensured complete lipid removal from
the flask wall. The multilamellar vesicles were extruded to form
large unilamellar vesicles with a hand-held Basic LiposoFast
device purchased from Avestin, Inc., Ottawa, Canada. The
vesicles were extruded 29 times through a 19 mm polycarbonate
Nucleopore filter with 100 nm diameter pores.
1H NMR titration
Small aliquots of guest stock solution (0.375 M) were added to
an NMR tube containing a solution of translocase in DMSO-
d6 (5 mM, 0.75 cm3), followed by the acquisition of a 1H NMR
spectrum. Concentrations and equivalents were adjusted to give
the optimum change in Weber p values (0.2–0.8).22 Titration
isotherms were generated for the NH protons of sulfonamide or
urea residues. Fitting the data to a 1 : 1 binding model using an
iterative curve-fitting method yielded association constants and
maximum change in chemical shift.23
Erythrocyte preparation
Fresh human blood was treated with EDTA solution
(dipotassium salts) and erythrocytes were isolated by centri-
fugation at 7500 rpm for 5 min, followed by washing three times
with 4 volumes of ice cold solution of 138 mM NaCl, 5 mM
KCl, 6.1 mM Na2HPO4, 1.4 mM NaH2PO4, 1 mM MgSO4
at pH 7.4. The cells were used the same day. All incubations
were carried out in the above mentioned buffer at 37 ЊC, 20%
hematocrit in capped plastic tubes unless otherwise stated.
Job plot
Job’s method was used to determine the stoichiometry of the
complex formed between 3 and tetrabutylammonium (TBA)
acetate.23 Stock solutions (5 mM) of 3 (host, H) and TBA
acetate (guest, G) were prepared in DMSO-d6. In separate
NMR tubes, 10 samples of varying 3 : TBA acetate ratios were
prepared. A 1H NMR spectrum was acquired for each sample,
and the concentration of the complex ([HG]eq) was determined
according to the following equation: [HG]eq = [H]0 × [(δobs Ϫ δ0)/
Inward NBD-lipid translocation assay
The inward translocation assay using phospholipids with
7-nitrobenz-2-oxa-1,3-diazol-4-yl (NBD) labels was adapted
from the original paper by McIntyre and Sleight.17,18 Exo-
labeled vesicles were generated upon addition of a concentrated
O r g . B i o m o l . C h e m . , 2 0 0 4 , 2, 2 1 4 – 2 1 9
218

View Article Online
(δmax Ϫ δ0)], where [H]0 is the pre-equilibrium host concen-
tration, δobs is the observed chemical shift, δ0 is the chemical
shift of the free host, and δmax is the chemical shift of the com-
plex (δmax was obtained from curve-fitting methods used to
derive the association constant). The conventional Job’s plot
([HG]eq vs. host mole fraction) was then prepared.
Acknowledgements
We thank Dr J. Middleton Boon (Duke University) and
Professor Jun-ichi Kikuchi (Nara Institute of Science and
Technology) for advice and useful discussions. This work was
supported by the US National Institutes of Health, the US
Department of Defense, and the Ministry of Education,
Culture, Sports, Science and Technology of Japan.
UV titration
UV titration experiments were conducted by adding POPC or
POPS to solutions of compounds 2 or 3 in 99 : 1 CHCl3 :
CH3OH at 295 K. Specifically, small aliquots of phospholipid
stock solution (0.315 mM phospholipid in 15 µM scramblase
solution) were added to a cuvette containing a solution of
scramblase (15 µM, 1 cm3), followed by the acquisition of a UV
spectrum. The total volume of added phospholipid was 0.4 cm3
or 6 molar equivalents. Titration isotherms were generated
from the changes in absorbance, and the data were fitted to a
1 : 1 binding model.24 An iterative curve-fitting method yielded
association constant and maximum change in absorbance. The
association constants listed in the text are the average of three
independent measurements.
References
1 J. M. Boon and B. D. Smith, Curr. Opin. Chem. Biol., 2002, 6, 749–
756.
2 J. M. Boon and B. D. Smith, Med. Res. Rev., 2002, 22, 251–281.
3 N. Kato, M. Nakanishi and N. Hirashima, Biochemistry, 2002, 41,
8068–8074.
4 S. Manno, Y. Takakuwa and N. Mohandas, Proc. Natl. Acad. Sci. U.
S. A., 2002, 99, 1943–1948.
5 A. J. Verkleij and J. A. Post, J. Membr. Biol., 2000, 178, 1–10.
6 R. F. Zwaal and A. J. Shroit, Blood, 1997, 89, 1121–1132.
7 R. D. Kornberg and H. M. McConnell, Biochemistry, 1971, 10,
1111–1120.
8 D. L. Daleke and J. V. Lyles, Biochim. Biophys. Acta, 2000, 1486,
108–127.
9 D. L. Daleke, J. Lipid Res., 2003, 44, 233–242.
10 J. M. Boon and B. D. Smith, J. Am. Chem. Soc., 1999, 121, 11924–
11925.
Flow cytometry
11 B. D. Smith and T. N. Lambert, Chem. Commun., 2003, 2261–2268.
12 J. M. Boon, T. N. Lambert, B. D. Smith, A. M. Beatty, V. Ugrinova
and S. N. Brown, J. Org. Chem., 2002, 67, 2168–2174.
13 J. M. Boon, R. Shukla, B. D. Smith, G. Licini and P. Scrimin, Chem.
Commun., 2002, 260–261.
Erythrocytes were isolated by centrifugation at 7900 rpm for
5 min, washed three times with 4 volumes of ice cold 137 mM
NaCl, 2.7 mM KCl, 10.1 mM Na2HPO4, 1.8 mM KH2PO4, pH
7.4 (PBS), and suspended in PBS at a density of 1.5 × 108 cells
cmϪ3. A portion of the cells was pretreated with N-ethylmale-
imide (NEM, 10 mM in PBS) for 30 min at room temperature.
After 2 PBS washings, the NEM pretreated cells were sus-
pended in PBS at a density of 1.5 × 108 cells cmϪ3. A 10 mM
translocase solution was prepared by dilution of a DMSO or
ethanol stock solution into PBS, and a 0.05 cm3 aliquot of
either normal or NEM-pretreated cells was added to 0.45 cm3
of this translocase solution (1.5 × 107 cells cmϪ3). The resulting
solutions were incubated at 37 ЊC for 3 h. To prepare the
samples for flow cytometry analysis,25 the above solutions were
resuspended in 0.5 cm3 of 10 mM HEPES, 140 mM NaCl,
2.5 mM CaCl2, pH 7.4 (binding buffer). In a 5 cm3 culture tube,
0.015 cm3 annexin V-FITC (BD Biosciences/Pharmingen) was
added to 100 µL of the erythrocyte solution (1.5 × 106 cells),
and the mixture was incubated at 37 ЊC for 15 min. An addi-
tional 0.6 cm3 of binding buffer was added before measuring
fluorescence intensity and size with a Beckman Coulter Epics
XL flow cytometer. Control experiments indicate that ethanol
or DMSO alone has no effect on the amount of bound annexin
V-FITC.
14 T. N. Lambert and B. D. Smith, Coord. Chem. Rev., 2003, 240, 129–
141.
15 J. M. Boon, T. N. Lambert, A. L. Sisson, A. P. Davis and B. D.
Smith, J. Am. Chem. Soc., 2003, 125, 8195–8201.
16 R. Shukla, Y. Sasaki, V. Krchnak and B. D. Smith, manuscript in
preparation.
17 J. C. McIntyre and R. G. Sleight, Biochemistry, 1991, 30, 11819–
11827.
18 R. A. Moss and S. Bhattacharya, J. Am. Chem. Soc., 1995, 117,
8688–8689.
19 S. Castano, I. Cornut, K. Buttner, J. L. Dasseux and J. Dufourcq,
Biochim. Biophys. Acta, 1999, 1416, 161–175.
20 T. Pomorski, A. Herrmann, A. Zachowski, P. F. Devaux and
P. Muller, Mol. Membr. Biol., 1994, 11, 39–44.
21 M. V. Serra, D. Kamp and C. W. M. Heast, Biochim. Biophys. Acta,
1996, 1282, 263–273.
22 C. S. Wilcox, in Frontiers in Supramolecular Organic Chemistry and
Photochemistry, ed. H. J. Schneider and H. Durr, VCH, Weinham,
1991, pp. 123–143.
23 M. P. Hughes and B. D. Smith, J. Org. Chem., 1997, 62, 4492–4499.
24 H. Xie, S. Yi and S. Wu, J. Chem. Soc., Perkin Trans. 2, 1999, 2751–
2754.
25 I. Vermes and C. Heanen, Immunol. Methods, 1995, 184, 39–51.
O r g . B i o m o l . C h e m . , 2 0 0 4 , 2, 2 1 4 – 2 1 9
219