Space-filling effects in membrane disruption by cationic amphiphiles

Article abstract of DOI:10.1016/S0968-0896(00)00326-6

We studied the hemolytic activity towards bovine erythrocytes of novel synthetic steroid-polyamine conjugate consisting of a rigid hydrophobic steroid unit, and a flexible hydrophilic polyamine unit connected by a linker. The steroid structure, polyamine chain length, and the presence of a hydrophobic substituent on the steroid, all influenced the activity. Analysis of the time dependence of hemolysis suggested that these structurally related cationic amphiphiles have different mechanisms of membrane perturbation. Copyright

Full text of DOI:10.1016/S0968-0896(00)00326-6

Bioorganic & Medicinal Chemistry 9 (2001) 1013±1024
Space-Filling E€ects in Membrane Disruption by Cationic
Amphiphiles
Tomoko Fujiwara, Naohide Hirashima, Seiji Hasegawa, Mamoru Nakanishi and
Tomohiko Ohwada*
Faculty of Pharmaceutical Sciences, Nagoya City University, Tanabe-dori, Mizuho-ku, Nagoya 467-8603, Japan
Received 6 October 2000; accepted 21 November 2000
AbstractÐWe studied the hemolytic activity towards bovine erythrocytes of novel synthetic steroid±polyamine conjugates consist-
ing of a rigid hydrophobic steroid unit, and a ¯exible hydrophilic polyamine unit connected by a linker. The steroid structure,
polyamine chain length, and the presence of a hydrophobic substituent on the steroid, all in¯uenced the activity. Analysis of the
time dependence of hemolysis suggested that these structurally related cationic amphiphiles have di€erent mechanisms of mem-
brane perturbation. # 2001 Elsevier Science Ltd. All rights reserved.
Introduction
their membrane disruption activity, using bovine ery-
throcytes. These steroid±polyamine conjugates (A and
B, Chart 2) consist of a hydrophobic, structurally rigid
steroid, a ¯exible hydrophilic polyamine, the nitrogen
atoms of which can be protonated under physiological
conditions, and a linker which connects the hydro-
phobic and hydrophilic units. The steroid moieties used
were cholestane and lithocholic acid, of which the for-
mer has a trans-decalin structure and the latter, a cis-
decalin structure. In the case of lithocholic acid-poly-
amine conjugates (B, Chart 2), a hydrophobic sub-
stituent (R2) was also introduced. The polyamine chain
length was varied. Thus, in the present series of syn-
thetic cationic amphiphiles, the space-®lling characters
of the hydrophobic and hydrophilic units are di€erent
(Chart 2). We found that the membrane-perturbing
e€ect of the steroid±polyamine conjugates is in¯uenced
by both polyamine chain length and steroid structure,
and we postulate that di€erent mechanisms of action
Molecular interactions of amphiphilic molecules with
lipid membranes are ubiquitous, being involved in a
wide range of biological events. An understanding of
bilayer interactions with natural and synthetic com-
pounds would be bene®cial in many areas, for example,
in developing novel antimicrobial agents which can
speci®cally perturb bacterial cell membranes.1,2 Many
mechanisms of bilayer interactions have been proposed
so far, including channel formation,3 direct entry of
compounds into lipid membranes to form holes4 and
others. For example, polyene macrolide antibiotics,
such as amphotericin B, are proposed to form ion
channels in biomembranes.2 On the other hand, natu-
rally occurring saponins such as digitonin are surface-
active, inducing hemolysis and having other biological
activities.3,5 It has been proposed that digitonin and
amphotericin B enter lipid membranes and form com-
plexes with sterols such as cholesterol and ergosterol.2,4
Recently a naturally occurring sterol±polyamine con-
jugate, squalamine, was isolated from stomach tissue of
shark and its antimicrobial and hemolytic properties
were evaluated (Chart 1).6 Squalamine includes a long
and rigid hydrophobic sterol unit, a hydrophilic poly-
amine, and a zwitterionic (sulfate anion) head group.7
In the present work, we synthesized novel polyamine±
steroid conjugates, and evaluated structural e€ects on
*Corresponding author. Fax: +81-52-836-3407; e-mail: ohwada@
phar.nagoya-cu.ac.jp
Chart 1. Structure of squalamine.
0968-0896/01/$ - see front matter # 2001 Elsevier Science Ltd. All rights reserved.
PII: S0968-0896(00)00326-6

1014
T. Fujiwara et al. / Bioorg. Med. Chem. 9 (2001) 1013±1024
are involved in the hemolytic activities of these structu-
rally related compounds. These structural features rele-
vant to hemolytic activities are apparently di€erent
from those of gene transfection activities of these com-
pounds as cationic liposomal vehicles.8
Results
Synthesis of cationic amphiphiles
Polyamines were synthesized in a modi®ed manner of
Nakanishi et al.9 Michael addition of the terminal
Chart 2. Cationic Amphiphiles.
Scheme 1. Preparation of polyamines.

T. Fujiwara et al. / Bioorg. Med. Chem. 9 (2001) 1013±1024
1015
amine of diamine 15 to acrylonitrile, followed by N-Boc
protection, reduction of the nitrile with lithium alumi-
num hydride led to extension of a propylamine unit
(Scheme 1). Addition of a linker moiety to cholestane 24
was carried out by Michael reaction of acrylonitrile with
the 3b-hydroxyl group of 24, acid-catalyzed alcoholysis
of the nitrile 25 to give the ester (26) and alkaline
hydrolysis of the ester 26 to the acid 27 (Scheme 2).10
Amide coupling of the cholestane acid 27 and an N-Boc
(t-butoxycarbonyl) protected polyamine (14, 17, 20 and
23) was carried out with the conventional N-hydro-
xysuccinimide (NHS)-DCC method. The polyamine
chain length was varied. Deprotection of the Boc group
was carried out in tri¯uoroacetic acid (TFA). The
cationic amphiphiles (1±4) were obtained as TFA salts
(Scheme 3). 3-a-Acetyl lithocholic acid 30 was readily
obtained by acetylation of lithocholic acid 29. Amide
coupling of 30 with the N-Boc-protected polyamine (14
or 18) was performed with the NHS-DCC chemistry,
followed by deprotection of the BOC group in TFA to
give 8 and 9 (Scheme 4). Allylation of 3a-hydroxy group
of the methyl ester (33) of lithocholic acid with allyl
bromide in the presence of Hunich base (N,N-diisopro-
pylethylamine) was performed to give 34. Catalytic
hydrogenation over Pd/C and alkaline hydrolysis gave
O-propyl lithocholic acid 35 (Scheme 5). Amide cou-
pling of the acid 36 with the N-protected polyamine (14
or 18) gave 37 and 38, respectively followed by depro-
tection of the Boc group in TFA to give 10 and 11
(Scheme 6).
Hemolytic activities of polyamine±cholestane conjugates
towards bovine erythrocytes
We prepared four derivatives of cholestane±polyamine
conjugates (1±4) and the corresponding N-Boc-pro-
tected counterparts (5±7 and 28) (Chart 2). The steroid
moiety is cholestane and the cationic hydrophilic moiety
(R1) is a polyamine. The polyamine moiety of 1 was
diaminobutane, that of 2 was spermidine, and that of 3
Scheme 2. Preparation of cholestane bearing a linker.
Scheme 3. Preparation of cationic amphiphiles based on cholestane.

1016
T. Fujiwara et al. / Bioorg. Med. Chem. 9 (2001) 1013±1024
was a non-natural isomer of spermine. The linker bond
of these cholestane±polyamine conjugates is an amide
group. In order to evaluate the membrane-disrupting
ability of these cholestane±polyamine conjugates, we
measured their dose-dependent hemolytic activity
towards bovine erythrocytes at 37 ꢀC at 60min after the
addition of compounds (Fig. 1). The conjugates 2 and 3
showed dose-dependent hemolytic activity while the
short polyamine analogue 1 showed no signi®cant
activity at concentrations below 100 mM. The e€ective
concentrations (EC50) of 2 and 3 that induce 50%
hemolysis under the present experimental conditions
are 15.8 and 20.1 mM, respectively. Basicity of the
amino nitrogen atoms of the polyamine chain is crucial
for the hemolytic activity, because the N-Boc-protected
polyamine analogues 5±7 (Chart 2) showed no sig-
ni®cant hemolytic activity (Fig. 1). A polyamine by
itself may not induce hemolysis because spermidine
showed no hemolytic activity at concentrations below
100 mM (Fig. 2). The importance of the two amine
Scheme 4. Cationic amphiphiles derived from 3-acetyl lithocholic acid.
Scheme 5. Alkylation of 3-hydroxy group of lithocholic acid.

T. Fujiwara et al. / Bioorg. Med. Chem. 9 (2001) 1013±1024
1017
Scheme 6. Cationic amphiphiles derived from 3-propyl lithocholic acid.
nitrogen atoms of 2 for hemolytic activity is suggested
by the observation that the carbon analogue 4, in which
the secondary amine of 2 is replaced with a methylene
group, has little hemolytic activity (Fig. 2).
jugates 8 and 10 showed no signi®cant hemolytic activ-
ity, while 9 and 11 showed activity (Fig. 3), the latter
being as active as 2. The e€ective concentrations (EC50)
of 8, 9 and 11 are 216.8, 32.3 and 8.6 mM, respectively.
The polyamine chain lengths of 9 and 11 are the same,
but the hemolytic activity of 11 is stronger than that of
9. Thus, the substituent R2 (Chart 2) is another factor
which modi®es the hemolytic activity, that is, the non-
polar hydrophobic n-propyl group is more e€ective than
the polar acetyl group.
Membrane perturbation arising from polyamine±
lithocolic acid conjugates.E€ect of steroid structure on
hemolytic activity
We prepared six derivatives of the lithocholic acid±
polyamine conjugates (8±13) (Chart 2). The polyamine
moiety was diaminobutane (in 8 and 10) or spermidine
(in 9 and 11), both of which are surface-active in the
case of the cholestane±polyamine conjugates. The
hydroxyl group at C-3 of lithocholic acid (substituent
R2, Chart 2) was substituted with an acetyl group (8 and
9), or a propyl group (10 and 11). Hemolytic activities
of these cationic amphiphiles were measured in the same
manner as those of the cholestane±polyamine con-
jugates, i.e. 60min incubation after the addition of the
compounds at 37 ꢀC (Fig. 3). The short-polyamine con-
Time-dependence of membrane perturbation
In order to shed light on the kinetics of hemolysis, the
magnitude of hemolysis was recorded over 120min after
the addition of compounds. The results at various con-
centrations (1, 25 and 100 mM) of compounds 2, 9
and 11 are shown in Figure 4A±C. Compounds 2, 9
and 11 all have the same polyamine chain, while the
hydrophobic units and the substituents (R2) are di€er-
ent, i.e. the steroid unit is cholestane in 2 and lithocholic
Figure 1. Dose-dependent hemolysis induced by cholestane±poly-
amine conjugates. Incubation at 37 ꢀC for 60min. Each value is the
mean Æ S.E. (n=3).
Figure 2. Structural requirements for hemolytic activity of cholestane±
polyamine conjugates. Incubation at 37 ꢀC for 60min. Each value is
the mean Æ S.E. (n=3).

1018
T. Fujiwara et al. / Bioorg. Med. Chem. 9 (2001) 1013±1024
acid in 9 and 11, while the hydrophobic substituent (R2)
is acetyl (9) or n-propyl (11). At the concentration of
1 mM, these compounds showed practically no hemo-
lysis (Fig. 4A). Addition of 11 to the cells at the ®nal
concentration of 25 mM induced 100% hemolysis within
20min (Fig. 4B). In contrast, compounds 2 and 9 at the
concentration of 25 mM induced a plateau level of 60%
hemolysis after about 60min (Fig. 4B). The times
required to reach the half-maximum level of hemolysis
(t50) with 25 mM of 2, 9 and 11 were 9.2, 27.8, and
3.7 min, respectively. Even at the concentration of
100 mM, hemolysis induced by 9 was slow, and reached
a plateau at 80% after 120 min. Compounds 2 and 11
both induced complete hemolysis within 20min at the
concentration of 100 mM (Fig. 4C). At 100 mM, the t50
values of 2, 9 and 11 were 2.3, 21.2, and 3.7 min,
respectively. The hemolysis curves of 2 and 11 are
very similar. In the case of 2, the t50 value depends
inversely on the concentration of the reagent, i.e., as
the concentration of the reagent was increased four-
fold (25 to 100 mM), the t50 value decreased to one-
fourth (9.2 to 2.3 min). On the other hand, the t50 value
Figure 3. Hemolytic activity of lithocholic acid±polyamine conjugates.
incubation at 37 ꢀC for 60min. Each value is the mean Æ S.E. (n=3).
Figure 4. Time-dependency of hemolytic activity of steroid±spermidine conjugates. Incubation at 37 ꢀC. Each value is the mean Æ S.E. (n=3).

T. Fujiwara et al. / Bioorg. Med. Chem. 9 (2001) 1013±1024
1019
of 11 seems to be insensitive to the concentration of the
reagent. Di€erences in the rate of partition of the com-
pounds into the plasma membrane of the erythrocytes
may account for the di€erence of kinetics.
against both Gram-positive and Gram-negative bacteria
at lower concentrations than those at which hemolytic
activity is observed, the present study may provide a
new approach for the development of novel, potent,
broad-spectrum antibiotics.6,12
Discussion
Experimental
Although the hemolytic activities of some steroid-sapo-
nins and triterpenoid-saponins have been well stud-
ied,4,5 structural e€ects remain controversial, i.e., it is
unclear whether di€erences of hemolytic activity and
time-dependency are attributable to the hydrophilic
sugar moiety or the hydrophobic aglycon structure.
Intuitively, a hydrophilic moiety should be inert for
membrane disruption11 while the hydrophobic moiety
of a cationic amphiphile should be crucial for mem-
brane disruption, because anchoring of the hydrophobic
moiety to the cell membrane would be the initial step of
hemolysis. However, we found that the polyamine chain
length has a striking e€ect on hemolytic activity: the
analogue bearing a short polyamine chain (1) showed
no hemolytic activity whereas the analogues (2 and 3)
with longer polyamine chains are surface-active. Fur-
thermore, the time-dependent hemolytic behavior at the
concentration of 25 mM highlighted the di€erence in
activities of the lithocholic acid±spermidine conjugate 11
and the cholestane±spermidine conjugate 2 (Fig. 4B):
i.e., the order of hemolytic activity is 11>2. Thus,
space-®lling of the hydrophobic steroid unit (shape and
orientation) also a€ects the hemolytic activities. In the
time-dependency experiments at the concentration of
25 mM (Fig. 4B), hemolysis due to the lithocholic acid±
polyamine conjugate 11 was very rapid being completed
in 20min, whereas the lithocholic acid±polyamine deri-
vative 9 induced slow hemolysis, which reached a pla-
teau. The only di€erence in structural units between 9
and 11 is the substituent R2 (Chart 2). In addition,
lithocholic acid±polyamine derivatives (12 and 13,
Chart 2) bearing a hydroxyl group (R2=H) showed
signi®cantly reduced hemolytic activities (Fig. 3), the
ED50 value of 13 being 150.3 mM, much larger than
those of 9 and 11. Thus a small hydrophobic appendant
(i.e. the R2 substituent) can substantially in¯uence the
interaction with the bilayer.
General methods
All the melting points were measured with a Yanaco
Micro Melting Point Apparatus (MP-500D) and are
uncorrected. Proton (400 MHz) NMR spectra were
measured on a JEOL Caliber-GX400 NMR spectro-
meter with TMS as an internal reference in CDCl3 as
the solvent, except otherwise speci®ed. Chemical shifts
are shown in ppm. Coupling constants are given in
hertz. High-resolution mass spectra (HRMS, EI) and
FAB mass spectra (FABMS) were recorded on a JEOL
JMS-SX 102A instrument. High-performance liquid
chromatography (HPLC) was run on a Shimadzu
LC-10A system and a Shimadzu SPD-M10A system
equipped with a UV±vis photo diode array detector on
ODS gel (RP-18 GP, Mighty sil, Kanto Chemicals,
Japan) packing (20mm Â25 cm) with the speci®ed elu-
ent. The eluents were 0.1% (v/v) TFA±99.9% H2O
solution and 0.08% (v/v) TFA±99.92% CH3CN. Col-
umn chromatography was performed on silica gel [silica
gel 60(63±20 mm, Merck], and ¯ash column chroma-
tography was carried out on silica gel [silica gel 60(40±
63 mm), Merck]. The combustion analyses were carried
out in the microanalytical laboratory of this faculty.
Preparation of the steroid±polyamines conjugates
4-Aza-8-amino-octacarbonitrile (16). To a solution of
1,4-diaminobutane (15) [1.32 g (15.0mmol)] in 0.35 mL
of methanol, acrylonitrile [1.25 g (1.54 equiv)] was
added at 0 ꢀC (in an ice-water bath) over 2 min.9 The
mixture was stirred at ambient temperature for 17 h.
The organic solvent was evaporated to give the residue
which was column-chromatographed (iPrNH2:MeOH:
CHCl3=1:5:15) to give the nitrile 16 (1.21 g (57% yield)
1
as colorless liquid. H NMR: 2.930(2H, t, J=6.60Hz),
2.711 (2H, t, J=6.78 Hz), 2.654 (2H, t, J=6.97 Hz),
2.529 (2H, t, J=6.60Hz), 1.542 (3H, m), 1.509 (4H, m).
MS (EI): (M+) 141.
Conclusion
We prepared cationic amphiliphiles based on the cho-
lestane and lithocholic acid structures. These present
novel steroid±polyamine conjugates can serve as useful
probes to study space-®lling e€ects, i.e., the e€ects of
shape and orientation, of hydrophilic and hydrophobic
structural units upon interactions with a bilayer.
N4,N8-di-Boc-4-Aza-8-amino-octacarbonitrile (17). To a
solution of 16 [282.0mg (2.0mmol)] in 2.5 mL of
CH2Cl2, a solution of (Boc)2O (di t-butyl dicarbonate)
[894.8 mg (4.1 mmol)] in 2.5 mL of CH2Cl2 was added at
0 ꢀC (in an ice-water bath).10 The whole was stirred at
ambient temperature for 4 h. The whole was poured into
30mL of water, and was extracted with ethyl acetate
(20mL Â2). The organic layer was washed with brine,
and was dried over magnesium sulfate. The organic
solvent was evaporated to give the residue which was
¯ash-chromatographed (CHCl3, and then CHCl3:
MeOH=99:1) to give 17 [88.7 mg (86% yield)] as color-
The hemolytic activity of a structurally similar com-
pound, squalamine (Chart 1), was measured after
10min incubation with erythrocytes and no signi®cant
hemolysis was observed at 25 mM even though some
compounds tested in this study showed distinct hemo-
lytic activities at this concentration (Fig. 4B).6 Con-
sidering that squalamine exerts antibiotic activity
1
less liquid. H NMR: 4.585 (1H, br s), 3.461 (2H, t,
J=6.60Hz), 3.280(2H, t, J=7.24 Hz), 2.564 (2H, br s),

1020
T. Fujiwara et al. / Bioorg. Med. Chem. 9 (2001) 1013±1024
1.598±1.441 (6H, m), 1.468 (9H, s), 1.441 (9H, s). MS
(EI): 341 (M+); HRMS (EI): calcd for C17H31N3O4,
341.2316. Found: 341.2317.
3ꢀ-O-Cyanoethyl cholestane (25). To a solution of
7.77 g (1.0mmol) of a-cholestane 24 in 30mL of
CH2Cl2, a mixture of 6.0mL of aqueous KOH (w/w
40%) and 4.0 mL of acrylonitril was added at ambient
temperature, followed by the addition of 18-crown-6
(520.0 mg). The whole was stirred at ambient tempera-
ture for 18 h. The whole was diluted with 100 mL of
H2O, and was acidi®ed with aqueous 1 N HCl solution.
The mixture was extracted with CHCl3 (50mL Â2), and
the organic layer was washed with brine, and was dried
over magnesium sulfate. The organic solvent was eva-
porated to give 10.17 g of the crude 25. Recrystallization
from methanol gave 7.53 g (85% yield) of pure nitrile 25
N4,N8-Di-Boc-4-aza-octane-1,8-diamine (18). To a sus-
pension of LiAlH4 (1.26 g) in 150mL of dry ether, a
solution of 17 [3.16 g (9.27 mmol)] in 50mL of dry ether
was added at 0 ꢀC (in an ice-water bath) with stirring
over 30min. After 1 h, aqueous 1 N NaOH solution was
added, and the inorganic salts were ®ltered o€. The
obtained organic layer was washed with brine, and was
dried over magnesium sulfate. The organic solvent was
evaporated to give the amine 18 [2.86 g (90% yield)] as
colorless liquid. H NMR: 4.570(1H, br s), 3.252 (2H,
as colorless plates. Mp 102.5±104.5 ꢀC. H NMR: 3.684
1
1
br s), 3.139 (4H, m), 2.683 (2H, t, J=6.70Hz), 1.638
(2H, m), 1.542±1.242 (6H, m), 1.453 (9H, s), 1.440(9H,
s). HRMS (EI, M+): calcd for C17H35N3O4, 345.2627.
Found: 345.2622.
(2H, d,d, J=1.65, 6.42 Hz), 3.289 (1H, m), 2.569 (2H, t,
J=6.51 Hz), 1.962 (1H, m), 0.897 (3H, d, J=6.60Hz),
0.864 (3H, d, J=6.60Hz), 0.860 (3H, d, J=6.60Hz)
0.795 (3H, s), 0.646 (3H, s), 1.829±0.577 (total 31H, m).
Anal. calcd for C30H51NO: C, 81.57; H, 11.64; N, 3.17.
Found: C, 81.41; H, 11.91; N, 3.23.
N8,N12 -di-Boc-4,8-diaza-12-amino-dodecacarbonitrile
(19). The nitrile 19 was prepared from 18 in a similar
1
manner of 16. Yield: 58%. Colorless liquid. H NMR:
Ethyl cholestane-3-O-propylate (26). A solution of nitrile
25 [110.3 mg (0.25 mmol)] in 95% EtOH (1 mL) in the
presence of 0.5 mL of concentrated H2SO4, was re¯uxed
for 5 h. The whole was diluted with 50mL of
water, followed by extraction with CHCl3 (30mL Â2).
The organic layer was washed with 10% aqueous
NaHCO3 (20mL Â2) and brine. The organic layer was
dried over magnesium sulfate. The organic solvent was
evaporated to give the residue which was ¯ash-chroma-
tographed (hexane:AcOEt=19:1 ) to give 63.4 mg (52%
yield) of the ester 26. Mp 57.5±59.5 ꢀC (recrystallized
4.561 (1H, br s), 3.247 (1H, br s), 3.137 (4H, m), 2.912
(2H, t, J=6.69 Hz), 2.618 (2H, t, J=6.60Hz), 2.510(2H,
t, J=6.69 Hz), 1.692 (2H, quintet, J=6.87 Hz), 1.596±
1.440(6H, m), 1.452 (9H, s), 1.440(9H, s). MS(EI): 398
(M+). HRMS (EI, M+): calcd for C20H38N4O4:
398.2893. Found: 398.2908.
N4,N8,N12-tri-Boc-4,8-diaza-12-amino-dodecacarbonitrile
(20). The tri-Boc protected nitrile 20 was prepared from
19 in a similar manner to 17. Yield: 94%. Colorless
liquid. 1H NMR (CDCl3): 4.635 (1H, br s), 3.478 (4H, t,
J=6.60Hz), 3.265 (2H, t, J=7.42 Hz), 3.172 (6H, m),
2.635 (2H, br s), 1.763 (2H, quintet, J=6.87 Hz), 1.573±
1.440(6H, m), 1.471 (9H, s), 1.452 (9H, s), 1.439 (9H, s).
MS(EI): 498 (M+). HRMS (EI, M+): calcd for
C25H46N4O6: 498.3417. Found: 498.3416.
1
from ethanol, colorless plates). H NMR: 4.147 (2H, q,
J=7.15 Hz), 3.730(2H, dd, J=2.02, 6.60 Hz), 3.232
(1H, m), 2.547 (2H, t, J=6.51 Hz), 1.956 (1H, m), 1.256
(3H, t, J=7.15 Hz), 0.896 (3H, d, J=6.60Hz), 0.863
(3H, d, J=6.60Hz), 0.859 (3H, d, J=6.60Hz), 0.781
(3H, s), 0.641 (3H, s), 1.826±0.572 (total 31H, m). Anal.
calcd for C32H56O3: C, 78.63; H, 11.55; N, 0.00. Found:
C, 78.48; H, 11.69; N, 0.00.
N4,N8,N12-tri-Boc-4,8-diaza-12-amino-dodecane-1,12-
diamine (21). The amine 21 was prepared in a similar
1
manner of 18. Yield: 68%. H NMR: 4.650(1H, br s),
Cholestane-3-O-propionic acid (27). To a solution of 26
[500.2 mg (0.976 mmol)] and 18-crown-6 (50.8 mg) in
3.0mL, 1.5 mL of 40% (w/w) aqueous NaOH solution
was added. The whole was stirred at ambient tempera-
ture for 17 h, and then was poured into 30mL of water.
The whole was acidi®ed with aqueous 1 N HCl solution,
and was extracted with CHCl3 (30mL Â2). The organic
layer was washed with water (20mL), and was dried
over magnesium sulfate. The organic solvent was eva-
porated to give crude 27 (470.5 mg), followed by
recrystallization from methanol to give 27 (421.4 mg,
3.279 (2H, br s), 3.153 (6H, br s), 2.729 (2H, t,
J=7.24 Hz), 1.534±1.438 (6H, m), 1.455 (9H, s), 1.449
(9H, s), 1.438 (9H, s). MS(EI): 502 (M+); HRMS (EI,
M+): calcd for C25H50N4O6: 502.3730. Found: 502.3731.
N8-Boc-octane-1,8-diamine (23). To a solution of 1,8-
diaminooctane (22) [5.77 g (40.0 mmol)] in 60 mL of 1,4-
dioxane, a solution of (Boc)2O 1.09 g (5 mmol) in 30 mL
of 1,4-dioxane was added at 25 ꢀC over 2 h. The whole
was stirred at 25 ꢀC for 24 h. The solvent was evapo-
rated, and to the residue 100 mL of water was added.
Undissolved materials were separated by ®ltration with
suction, and the ®ltrate was extracted with carbon tet-
rachloride (30mL Â3). The organic layer was washed
with water (20mL Â4), and was dried over sodium sul-
fate. The organic solvent was evaporated to give 23
[756.4 mg (63% yield)] as colorless solid. 1H NMR:
4.499 (1H, br s), 3.014 (2H, m), 2.673 (2H, t,
J=6.98 Hz), 1.443 (9H, br s), 1.299 (8H, br s). MS (EI):
224 (M+). HRMS (EI, M+): calcd for C13H28N2O2:
224.2152. Found: 224.2152.
90% yield) as colorless plates. Mp 139.5±141.0 ꢀC. H
1
NMR: 3.754 (2H, dd, J=1.65, 6.23 Hz), 3.310(1H, m),
2.625 (2H, t, J=6.14 Hz), 1.962 (1H, m), 0.897 (3H, d,
J=6.42 Hz), 0.864 (3H, d, J=6.60Hz), 0.856 (3H, d,
J=6.60 Hz), 0.795 (3H, s), 0.646 (3H, s), 1.876±0.580
(total 31H, m). Anal. calcd for C30H52O3: C, 78.21; H,
11.38; N, 0.00. Found: C, 78.03; H, 11.54; N, 0.00.
N-Boc-Diamine±cholestane conjugate (5). To a stirred
mixture of 27 [172.5 mg (0.375 mmol)], N-Boc-1,4-di-
aminobutane (14) (70.5 mg (1.0 equiv)) and NHS (N-

T. Fujiwara et al. / Bioorg. Med. Chem. 9 (2001) 1013±1024
1021
hydroxysuccinimide) [48.3 mg (1.1 equiv)] in 2.0mL of
dry CH2Cl2, DCC (N,N-dicyclohexylcarbodiimide)
[77.1 mg (1.0equiv)] was added at ambient temperature,
and the whole was stirred for 20h. After removal of the
precipitate of DCurea by ®ltration with suction, and the
®ltrate was washed with saturated aqueous NaHCO3,
water, and brine. The organic phase was dried over
magnesium sulfate. The residue after evaporation of the
organic solvent was ¯ash-chromatographed (hexane:
AcOEt=2:1 and then 1:3) to give 5 (202.0 mg (85%
yield)). Mp 96.0±98.0 ꢀC (recrystallized from CH2Cl2/n-
J=5.87 Hz), 1.455 (9H, s), 1.447 (9H, s), 1.438 (9H, s),
0.896 (3H, d, J=6.60Hz), 0.863 (3H, d, J=6.60Hz),
0.859 (3H, d, J=6.60 Hz), 0.785 (3H, s), 0.643 (3H, s),
1.974±0.586 (total 39H, m). FABMS: 946 (MH+).
Tetramine±cholestane conjugate (3). The amine con-
jugate 3 was prepared from 7 in a similar manner of 1
(60% yield). Pale yellow oil. The crude product was
puri®ed by HPLC [retention time: 6.48 min (ODS ana-
lytical column: 0.1% TFA±H2O: 0.08% TFA±
CH3CN=10:90)]. 1H NMR: 6.885 (1H, br s), 3.704
(2H, t, J=5.04 Hz), 3.326 (1H, m), 3.257 (1H, m), 2.863
(2H, t, J=6.23 Hz), 2.769 (6H, m), 2.676 (2H, t,
J=6.32 Hz), 2.434 (2H, t, J=5.96 Hz), 0.897 (3H, d,
J=6.60Hz), 0.864 (3H, d, J=6.60Hz), 0.859 (3H, d,
J=6.42 Hz), 0.789 (3H, s), 0.645 (3H, s), 1.978±0.592
(total 31H, m). FABMS: 646 (MH+).
1
hexane). H NMR: 6.520(1H, br s), 4.756 (1H, br s),
3.690(2H, m), 3.257 (3H, m), 2.428 (2H, t, J=5.68 Hz),
1.440(9H, s), 0.898 (3H, d, J=6.60Hz), 0.864 (3H, d,
J=6.60Hz), 0.860 (3H, d, J=6.60Hz), 0.795 (3H, s),
0.646 (3H, s), 2.045±0.582 (total 37H, m). FABMS
(MH+): 631. HRMS (FAB+, MH+): calcd for
C39H71N4O2: 631.5414. Found: 631.5435. Anal. calcd
for C39H70N2O4: C, 74.24; H, 11.18; N, 4.44. Found: C,
74.31; H, 11.06; N, 4.22.
N-Boc-diamine±cholestane conjugate (28). The amide 28
was prepared from 27 and 23 in a similar manner of 5
(82% yield). Mp 80.3±81.9 ꢀC (recrystallized from
CH2Cl2/n-hexane, colorless cubes). 1H NMR: 6.468
(1H, m), 4.498 (1H, br s), 3.693 (2H, t, J=6.23 Hz),
3.254 (1H, m), 3.229 (2H, q, J=5.68 Hz), 3.091 (2H, m),
2.429 (2H, t, J=5.50Hz), 1.443 (9H, s), 0.897 (3H, d,
J=6.42 Hz), 0.859 (3H, d, J=6.42 Hz), 0.795 (3H, s),
0.647 (3H, s), 1.949±0.620 (total 43H, m). FABMS: 687
(MH+). HRMS (FAB+, MH+): calcd for C43H79N2O4:
687.6030. Found: 687.6019. Anal. calcd for
C43H78N2O4+1/4 H2O: C, 74.67; H, 11.44; N, 4.05.
Found: C, 74.64; H, 11.40; N, 4.15.
Diamine±cholestane conjugate (1). Under Ar atmosphere
a solution of 5 [90.0 mg (0.143 mmol)] in 2.0 mL of TFA
was stirred at ambient temperature for 30min. The acid
was evaporated to give the residue which was ¯ash-
chromatographed (CHCl3:MeOH=9:1 and then CHCl3:
MeOH:iPrNH2=30:5:1), followed by freeze-drying in
vacuum to give pale yellow amorphous 1 (129.9 mg).
The product was puri®ed with HPLC (retention time
20.25 min (ODS analytical column. Eluent: 0.1% TFA±
H2O:0.08% TFA±CH3CN=10:90) 1H NMR: 3.746
(2H, t, J=5.59 Hz), 3.325 (3H, m), 3.163 (2H, m), 2.555
(2H, t, J=5.32 Hz), 0.897 (3H, d, J=6.60Hz), 0.864
(3H, d, J=6.60Hz), 0.859 (3H, d, J=6.60Hz), 0.789
(3H, s), 0.647 (3H, s), 2.045±0.582 (total 37H, m).
FABMS (MH+): 531. HRMS (FAB+, MH+): calcd for
C34H63N2O2: 531.4889. Found: 531.4905.
Diamine±cholestane conjugate (4). The conjugate 4 was
prepared from 28 (86% yield). Further puri®cation was
performed on HPLC (retention time: 26.50min (ODC
analytical column: 0.1% TFA±99.9%H2O:0.08% aqu-
1
eous TFA±CH3CN=10:90)). H NMR: 7.542 (2H, br
s), 7.113 (1H, br s), 3.627 (3H, br m), 3.192 (2H, br s),
2.881 (2H, br s), 2.448 (2H, br s), 0.896 (3H, d,
J=5.68 Hz), 0.860 (3H, d, J=6.23 Hz), 0.787 (3H, s),
0.645 (3H, s), 1.976±0.645 (total 46H, m). FABMS: 587
(MH+); HRMS (FAB+, MH+): calcd for C39H71N2O2:
587.5515. Found: 587.5510.
N,N0-DiBoc triamine±cholestane conjugate (6). A similar
coupling of 27 and 18 was carried out in a similar man-
ner of 1 (43% yield). Colorless liquid. H NMR: 4.559
(1H, br s), 3.712 (2H, t, J=5.50 Hz), 3.300±3.000 (9H,
m), 2.437 (2H, t, J=5.89 Hz), 1.452 (9H, s), 1.440(9H,
s), 0.896 (3H, d, J=6.60Hz), 0.863 (3H, d, J=6.60Hz),
0.859 (3H, d, J=6.60 Hz), 0.785 (3H, s), 0.643 (3H, s),
1.975±0.607 (total 37H, m). FABMS: 789 (MH+)
1
3-O-acetyl lithocholic acid (30). A solution of lithocholic
acid (29) [1.00 g (2.66 mmol)] and 5.0 mL of acetic
anhydride in 5.0mL of pyridine was stirred at ambient
temperature for 72 h. The whole was poured into
100 mL of water, followed by addition of saturated
aqueous NaHCO3. The whole was extracted with
CHCl3 (30mL Â3). The organic layer was washed with
1N aqueous HCl and with water, and was dried over
magnesium sulfate. The organic solvent was evaporated
to give 1.07 g (97% yield). Mp 50.5±52.5 ꢀC (recrys-
Triamine±cholestane conjugate (2). Deprotonation of
the Boc group of 6 was carried out in a similar manner
of 1. Yield: 73%. The product was puri®ed with HPLC
[retention time: 8.21 min (ODC analytical column: 0.1%
TFA±H2O:0.08% TFA±CH3CN=10:90)]. 1H NMR:
3.712 (2H, m), 3.480(2H, m), 3.257 (1H, m), 2.857±
2.802 (6H, m), 2.455 (2H, t, J=5.87 Hz), 0.897 (3H, d,
J=6.42 Hz), 0.864 (3H, d, J=6.60Hz), 0.859 (3H, d,
J=6.60 Hz), 0.783 (3H, s), 0.645 (3H, s), 1.975±0.607
(total 37H, m). FABMS: 588 (M+), 589 (MH+).
1
tallized from ethanol; colorless plates); H NMR: 4.720
(1H, m), 2.031 (3H, s), 0.927 (3H, s), 0.926 (3H, d,
J=5.50 Hz), 0.649 (3H, s), 2.401±0.996 (total 32 H, m).
Anal. calcd for C25H42O4+1/4 C2H5OH: C, 73.81; H,
10.12; N, 0.00. Found: C, 73.65; H, 10.40; N, 0.00.
N,N0,N00-Tri-Boc tetramine±cholestane conjugate (7). The
N-Boc protected amide 7 was prepared from 27 and 21
in a similar manner of 5 (59% yield). Colorless liquid.
1H NMR: 6.998 (1H, br s), 4.632 (1H, br s), 3.711 (2H,
t, J=5.87 Hz), 3.258±3.135 (13H, m), 2.435 (2H, t,
N-Boc-Diamine-3-acetyl-lithocholic acid conjugate (31).
To a stirred mixture of 30 (203.0 mg [0.486 mmol)], 14
(94.0mg (1.0 equiv)) and NHS [64.5 mg (1.1 equiv)] in

1022
T. Fujiwara et al. / Bioorg. Med. Chem. 9 (2001) 1013±1024
3.0mL of CH 2Cl2, DCC 102.5 mg (1.0 equiv) was
added, and the whole was stirred at ambient tempera-
ture for 24 h. After removal of DCurea by ®ltration with
suction, the ®ltrate was washed with saturated aqueous
sodium bicarbonate, water and brine, and was dried
over sodium sulfate. The residue obtained after eva-
poration was ¯ash-chromatographed (n-hexane:
AcOEt=2:1, and then 3:2) to give 31 (253.6 mg (89%
yield)). Mp 40.5±42.4 ꢀC (recrystallized from CH2Cl2/
was dried over magnesium sulfate. The organic solvent
was evaporated to give the crude 33 (4.08 g (98%)). Mp
129.0±130.4 ꢀC (recrystallized from AcOEt, colorless
1
needles). H NMR: 3.664 (3H, s), 3.624 (1H, m), 2.354
(1H, m), 2.327 (1H, m), 0.917 (3H, s), 0.907 (3H, d,
J=4.03 Hz), 0.641 (3H, s), 1.974±0.970 (total 26 H, m).
FABMS: 390(M +). Anal. calcd for C25H42O3; C, 76.87;
H, 10.84; N, 0.00. Found: C, 76.77; H, 10.92; N, 0.00.
1
n-hexane). H NMR: 5.653 (1H, br s), 4.720(1H, m),
Methyl 3-O-allyl-lithocholate (34). A solution of methyl
lithocholate [(33) 4.68 g (12.0mmol)] and N,N-diiso-
propylethylamine 4.66 g (3.0equiv) in 40mL of dry
DMF was heated to re¯ux. After re¯ux for 30min, allyl
bromide (5.0mL) was added to this solution over 5 min.
The whole was re¯uxed for 9 h. The whole was poured
into 30mL of H 2O, and was acidi®ed with 1 N aqueous
HCl. The mixture was extracted with Et2O (30mL Â3)
and the organic layer was washed with brine, and was
dried over magnesium sulfate. Evaporation of the sol-
vent gave the residue which was ¯ash-chromatographed
(n-hexane:AcOEt=30:1) to give 34 (3.72 g (72% yield)
as pale yellow solid. Mp 76.0±77.9 ꢀC (recrystallized
4.598 (1H, br s), 3.495 (1H, d, J=4.77 Hz), 3.267 (2H,
q, J=5.87 Hz), 3.137 (2H, d, J=6.05 Hz), 2.233 (1H,
m), 2.047 (1H, m), 2.033 (3H, s), 1.455 (9H, s), 0.925
(3H, s), 0.917 (3H, d, J=6.05 Hz), 0.639 (3H, s), 1.977±
0.957 (total 35 H, m). FABMS: 589 (MH+). Anal. calcd
for C35H60N2O5+3/4 CH2Cl2: C, 65.80; H, 9.50; N,
4.29. Found: C, 65.97; H, 9.49; N, 4.01.
Diamine-3-acetyl-lithocholic acid conjugate (8). Under
Ar atmosphere, a solution of 31 [72.2 mg (0.123 mmol)]
in 2.0mL of TFA was stirred at ambient temperature
for 3.5 h. The acid was evaporated to give the residue
which was ¯ash-chromatographed (CHCl3:MeOH=9:1,
and then 8:2), followed by freeze-drying in vacuum to
give 8 [60.8 mg (82% yield)] as colorless liquid. Further
puri®cation was performed on HPLC (retention time
10.36 min (analytical column: 0.1% TFA±H2O:0.08%
TFA±CH3CN=40:60)). 1H NMR (CD3OD): 4.854 (1H,
m), 3.194 (2H, t, J=6.87 Hz), 2.936 (2H, t, J=7.42 Hz),
2.213 (1H, m), 2.107 (1H, m), 1.997 (3H, s), 0.963 (3H,
s), 0.961 (3H, d, J=6.42 Hz), 0.688 (3H, s), 2.073±0.953
(total 34 H, m). FABMS: 490(MH +).
1
from methanol, colorless plates). H NMR: 5.936 (1H,
m), 5.275 (1H, dd, J=1.23, 17.41 Hz), 5.155 (1H, d,
J=10.26 Hz), 4.019 (2H, d, J=5.50Hz), 3.664 (3H, s),
3.306 (1H, m), 2.349 (1H, m), 2.214 (1H, m), 0.910 (3H,
s), 0.903 (3H, d, J=5.50 Hz), 0.632 (3H, s), 1.956±0.957
(total 26H, m). FABMS: 429 (MÀ1+). Anal. calcd for
C28H46O3: C, 78.09; H, 10.77; N, 0.00. Found: C, 77.88;
H, 10.72; N, 0.00.
Methyl 3-O-propyl-lithocholate (35). The allyl group of
34 was hydrogenated over 5% Pd-C in ethyl acetate at
ambient temperature for 8 h. Yield: 93%. Mp 60.3±
62.1 ꢀC (recrystallized from MeOH, colorless plates). 1H
NMR: 3.663 (1H, m), 3.411 (2H, t,d, J=2.02, 6.97 Hz),
3.224 (1H, m), 2.353 (1H, m), 2.213 (1H, m), 1.579 (2H,
q, J=7.33 Hz), 0.918 (3H, t, J=7.33 Hz), 0.909 (3H, s),
0.903 (3H, d, J=5.32 Hz), 0.632 (3H, s), 1.953±0.895
(total 26 H, m). FABMS: 431 (MÀ1+). Anal. calcd for
C28H48O3: C, 77.72; H, 11.18; N, 0.00. Found: C, 77.64;
H, 11.31; N, 0.00.
N,N0-diBoc-triamine-3-acetyl-lithocholic acid conjugate
(32). The N-Boc protected conjugate 32 was prepared
from 29 and 18 in a similar manner of 31. Yield: 89%.
1
Colorless liquid. H NMR: 6.724 (1H, brs), 4.179 (1H,
m), 4.548 (1H, brs), 3.277 (2H, br s), 3.204 (2H, br),
3.140 (4H, m), 2.248 (1H, m), 2.063 (1H, m), 2.030 (3H,
s), 1.463 (9H, s), 1.441 (3H, s), 0.925 (3H, s), 0.924 (3H,
d, J=6.23 Hz), 0.639 (3H, s), 1.980±1.027 (total 32H,
m). FABMS: 746 (MH+). HRMS (FAB+, MH+):
calcd for C43H76N3O7: 746.5729. Found: 746.5706.
3-O-Propyl-lithocholic acid (36). A mixture of 35 (1.50g
(3.47 mmol)), 18-crown-6 (700.0 mg) and 1 N aqueous
KOH solution in 30mL of dioxane was heated at 40 C
Triamine-3-acetyl-lithocholic acid conjugate (9). The
polyamine conjugate 9 was prepared from 32 through
deprotection of the Boc group in TFA. Yield: 58%
yield. Colorless liquid. Further puri®cation was per-
formed on HPLC [retention time 4.89 min (ODS analy-
tical column: 0.1% TFA±H2O: 0.08% TFA±
ꢀ
for 19 h. The whole was diluted with 50mL of water
(50mL), acidi®ed with 1 N aqueous HCl solution, and
extracted with diethyl ether (30mL Â3). The organic
layer was washed with water, and with brine, and was
dried over sodium sulfate. Evaporation of the organic
solvent gave the solid 36 (1.41 g, 97% yield). Mp 126.1±
128.0 ꢀC (recrystallized from methanol, colorless plates).
1H NMR: 3.415 (2H, t,d, J=2.02, 6.97 Hz), 3.229 (1H,
m), 1.581 (2H, q, J=7.15 Hz), 0.918 (3H, t, J=7.33 Hz),
0.918 (3H, d, J=6.42 Hz), 0.910 (3H s), 0.637 (3H, s),
1.958±0.884 (total 27 H, m). Anal. calcd for C27H46O3:
C, 77.46; H, 11.07; N, 0.00. Found: C, 77.19; H, 10.96;
N, 0.00.
1
CH3CN=40:60)]. H NMR (CD3OD): 4.670(1H, m),
3.282 (2H, t, J=6.42 Hz), 2.995 (6H, m), 2.286 (1H, m),
2.251 (1H, m), 1.998 (3H, s), 0.966 (3H, d, J=6.42 Hz),
0.963 (3H, s), 0.690 (3H, s), 2.034±0.958 (total 36 H, m).
FABMS: 546 (MH+). HRMS (FAB+, MH+): calcd for
C33H60N3O3: 546.4634. Found: 546.4611.
Methyl lithocholate (33). A mixture of lithocholic acid
[(29) 4.01 g (10.7 mmol)] and 4.05 mL of concd sulfuric
acid in 30mL of methanol was stirred at ambient tem-
perature for 1 h. After evaporation of methanol, the
residue was dissolved in 60mL of CHCl 3, and the
organic layer was washed with water (40mL Â2), and
N-Boc-Diamine-3-O-propyl lithocholic acid conjugate
(37). A mixture of 36 [209.0 mg (0.50 mmol)], and 14
(94.0mg (1.0 equiv)), NHS [64.5 mg (1.1 equiv)] and

T. Fujiwara et al. / Bioorg. Med. Chem. 9 (2001) 1013±1024
1023
DCC 102.8 mg (1.0 equiv) in methylene chloride was
stirred at ambient temperature for 18 h. After addition
of another portion of DCC (30.2 mg), the whole was
stirred for another 42 h. After removal of DCurea by
®ltration with suction, the organic layer was washed
with saturated aqueous sodium bicarbonate, water and
brine, and was dried over sodium sulfate. The residue
obtained after evaporation was ¯ash-chromatographed
(n-hexane:AcOEt=3:2, and then 2:3) to give 37
deprotection of the Boc group in TFA gave 13. Mp
93.5±97.1 ꢀC. 1H NMR(CDCl3+D2O): 3.647±3.575
(1H, m), 3.337 (2H, t, J=6.23 Hz), 2.717 (4H, t,
J=6.23 Hz), 2.625 (2H, t, J=6.23 Hz), 2.252±2.175 (1H,
m), 2.087±2.010 (1H, m), 1.971±0.916 (total 38H, m),
0.640 (3H, s). HRMS (ESI+, MH+): calcd for
C31H58N3O2: 504.4529. Found: 504.4510.
Biological materials
1
(282.1 mg (96%)) as colorless liquid. H NMR: 5.657
(1H, br s), 4.599 (1H, br s), 3.412 (2H, t,d, J=1.65,
6.97 Hz), 3.265 (2H, d, J=5.87 Hz), 3.241 (1H, m),
3.135 (2H, br d, J=5.31 Hz), 2.219 (1H, m), 2.043 (1H,
m), 1.580(2H, quartet, J=7.15 Hz), 1.444 (9H, s), 0.918
(2H, t, J=7.33 Hz), 0.913 (3H, d, J=4.22 Hz), 0.908
(3H, s), 0.629 (3H, s), 1.953±0.908 (total 32H, m).
FABMS: 589 (MH+). HRMS (FAB+, MH+): calcd for
C36H65N2O4: 589.4944. Found: 589.4957.
Triton X-100 was purchased from Sigma (St. Louis,
MO, USA). Spermidine was purchased from Aldrich
(Milwaukee, WI, USA), and was used without further
puri®cation. Bovine erythrocytes were purchased from
Nippon Bio-Supply Center (Japan).
Hemolytic activity towards bovine erythrocytes
Bovine erythrocytes were centrifuged at 800Âg
(2200 rpm) for 10 min and the plasma and bu€y coat
were discarded. To the precipitated cells was added
phosphate-bu€ered saline (PBS; 137 mM NaCl, 2.68 mM
Diamine-3-O-propyl lithocholic acid conjugate (10).
Under Ar atmosphere, a solution of 37 [100.1 mg
(0.170 mmol)] in 2.0 mL of TFA was stirred at ambient
temperature for 1 h. The acid was evaporated to give the
residue which was ¯ash-chromatographed (CHCl3:
MeOH=9:1), followed by freeze-drying in vacuum to
give 10 [90.2 mg (88%)] as colorless liquid. Further
puri®cation was performed with HPLC (Retention
Time: 8.95 min (ODS analytical column: 0.1% TFA±
H2O: 0.08% TFA±CH3CN=40:60). 1H NMR: 7.680
(2H, br s), 6.890(1H, brs), 3.477 (2H, t, J=6.96 Hz),
3.38 (1H, m), 3.266 (2H, br s), 3.042 (2H, br s), 2.311
(1H, m), 2.111 (1H, m), 1.789 (2H, q, J=7.14 Hz), 0.913
(3H, t, J=7.42 Hz), 0.913 (3H, s), 0.889 (3H, d,
J=4.22 Hz), 0.622 (3H, s), 1.945±0.884 (total 30H, m).
FABMS: 489 (MH+).
.
KCl, 8.1 mM Na2HPO4 12H2O, 1.47 mM KH2PO4;
pH:7.4), followed by centrifugation to collect the cells.
This process was repeated twice. The cells were sus-
pended in PBS to obtain 1Â108 cells/mL of erythrocyte
suspension. To 1-mL aliquots of the above suspension,
1 mL of 0.1, 1,10, 50 or 100 mM solution of the com-
pound in dimethylsulfoxide (DMSO; the ®nal con-
centration of DMSO was 0.1% v/v) was added, and the
whole was mixed gently by using a vortex mixer, then
ꢀ
the whole was incubated for 60min at 37 C. The
supernatant, after centrifugation at 800Âg (2200 rpm)
for 5 min, was collected. The absorbance (Ix) at 576 nm
of an aliquot of the supernatant (750 mL) was measured
with a spectrophotometer U-2000 (Hitachi; Tokyo,
Japan). The absorbance of the supernatant of the cell
suspension treated with vehicle only (0.1% (v/v)
DMSO) was used as a control value (Icontrol). The
intensity of absorption corresponding to 100% hemo-
lysis (I100) was determined by adding 50 mL of a 10% (v/
v) aqueous solution of Triton X-100 to a 1-mL aliquot
of the cell suspension in each hemolysis experiment. The
percentage hemolysis was then obtained according to
eq 1:
N,N0-diBoc-Triamine-3-O-propyl lithocholic acid conju-
gate (38). The N-Boc protected amine conjugate 38 was
prepared from 36 and 18 in a similar manner of 37.
Yield: 66%. Colorless liquid. 1H NMR: 6.718 (1H, brs),
4.543 (1H, brs), 3.411 (2H, t, d, J=6.78, 2.20Hz),
3.215±3.125 (9H, m), 1.464 (9H, s), 1.411 (9H, s), 0.918
(3H, t, J=7.33 Hz), 0.922 (3H, d, J=6.78 Hz), 0.908
(3H, s), 0.629 (3H, s), 1.975±0.900 (total 36H, m).
FABMS: 746 (MH+).
Triamine-3-O-propyl lithocholic acid conjugate (11).
Deprotection of the Boc group of 38 was carried out in
a similar manner of 37 to give 11. Yield: 84%. Colorless
liquid. Further puri®cation was performed with HPLC
[retention time 14.37 min (ODS analytical column: 0.1%
TFA±H2O: 0.08% TFA±CH3CN=40:60)]. 1H NMR:
6.736 (1H, t, J=5.02 Hz), 3.411 (2H,td, J=2.20,
6.90Hz), 3.355 (2H, q, J=6.42 Hz), 3.225 (1H, m),
2.848 (4H, m), 2.800 (2H, t, J=6.05 Hz), 2.232 (1H, m),
2.082 (1H, m), 1.580 (2H, q, J=6.70Hz), 0.918 (3H,t,
J=7.33 Hz), 0.913 (3H, d, J=4.03 Hz), 0.908 (3H, s),
0.628 (3H, s), 1.955±0.908 (total 35 H, m). FABMS: 546
(MH+). HRMS (FAB+, MH+): calcd for C34H64N3O2:
546.4998. Found: 546.4976.
hemolysis % ˆ 100 ꢀIx À Icontrol†= ꢀI100 À Icontrol
†
ꢀ1†
No signi®cant hemolysis was observed when the cells
were incubated with 0.1% (v/v) DMSO. All absorption
measurements were carried out at ambient temperature.
Non-linear curve ®tting of concentration (x)Àpercent
hemolysis (y) curves (Figs 1±3) was carried out with the
assumption of the relation y=100/(1+EC50/x) wherein
EC50 is an e€ective concentration that induces 50%
hemolysis under the present experimental conditions.
Time-dependent hemolysis curves (Fig. 4 B,C) were also
®tted non-linearly with the assumption of the relation
% hemolysis (y)=ymax(1ÀeÀkt) wherein t represents
time (min). Fitting was carried out with PRISM
(GraphPad, CA, U.S.A).
Triamine±lithocholic acid conjugate (13). Alkaline
hydrolysis of the acetyl group of 32, followed by

1024
T. Fujiwara et al. / Bioorg. Med. Chem. 9 (2001) 1013±1024
5. Hu, M.; Konoki, K.; Tachibana, K. Biochim. Biophys. Acta
1996, 1299, 252.
Acknowledgements
6. Moore, K. S.; Wehrli, S.; Roder, H.; Rogers, M.; Forrest,
J. J. N.; McCrimmon, D.; Zaslo€, M. Proc. Natl. Acad. Sci.
U.S.A. 1993, 90, 1354.
7. Merritt, M.; Lanier, M.; Deng, G.; Regen, S. L. J. Am.
Chem. Soc. 1998, 120, 8494.
The authors thank Professor Kentaro Yamaguchi,
Chemical Analysis Center, Chiba University for the
FAB+ high-resolution mass spectra measurement.
8. Fujiwara, T.; Hasegawa, S.; Hirashima, N.; Nakanishi, M.;
Ohwada, T. Biochim. Biophys. Acta 2000, 1468, 396.
9. Hunag, D.; Jiamg, H.; Nakanishi, K.; Usherwood, P. N. R.
Tetrahedron 1997, 53, 12391.
References
10. Sliedregt, L. A. J. M.; Rensen, P. C. N.; Rump, E. T.; van
Santbrink, P. J.; Bijsterbosch, M. K.; Valentijn, A. R. P. M.;
van der Marel, G. A.; van Boom, J. H.; van Berkel, T. J. C.;
Biessen, E. A. L. J. Med. Chem. 1999, 42, 609.
11. Azaz, E.; Segal, R.; Milo-Goldzweig, I. Biochim. Biophys.
Acta 1981, 646, 444.
1. Li, C.; Peters, A. S.; Meredith, E. L.; Allman, G. W.;
Savage, P. B. J. Am. Chem. Soc. 1998, 120, 2961.
2. Bolard, J. Biochim. Biophys. Acta 1986, 864, 257.
3. Yamashita, K.; Janout, V.; Bernard, M.; Armstrong, D.;
Regen, S. L. J. Am. Chem. Soc. 1995, 117, 6249.
4. Nishikawa, M.; Nojima, S.; Akiyama, T.; Sankawa, U.;
Inoue, K. J. Biochem. 1984, 96, 1231.
12. Kikuchi, K.; Bernard, E. M.; Sadownik, A.; Regen, S. L.;
Armstrong, D. Antimicrob. Agents Chemother. 1997, 41, 1433.