Transport of the highly charged myo-Inositol Hexakisphosphate molecule across the red blood cell membrane: A phase transfer and biological study

Article abstract of DOI:10.1016/S0968-0896(02)00162-1

To address the problem of delivering highly charged small molecules, such as phytic acid (InsP₆ or IHP), across biological membranes, we investigated an approach based on a non-covalent interaction between transport molecule(s) and IHP. Thus, w

Full text of DOI:10.1016/S0968-0896(02)00162-1

Bioorganic & Medicinal Chemistry 10 (2002) 2825–2834
Transport of the Highly Charged myo-Inositol
Hexakisphosphate Molecule across the Red Blood Cell
Membrane: A Phase Transfer and Biological Study
Stephane P. Vincent,a Jean-Marie Lehn,a,* Jaime Lazarteb and Claude Nicolaub
a
´
´
Laboratoire de Chimie Supramoleculaire, ISIS-Universite Louis Pasteur CNRS, 4 rue Blaise Pascal, 67000 Strasbourg, France
bCenter for Blood Research Laboratories, Harvard Medical School, 1256 Soldiers Field Road, Boston, MA 02135, USA
Received 4February 2002; accepted 14May 2002
Abstract—To address the problem of delivering highly charged small molecules, such as phytic acid (InsP6 or IHP), across biolo-
gical membranes, we investigated an approach based on a non-covalent interaction between transport molecule(s) and IHP. Thus,
we synthesized a collection of compounds containing IHP ionically bound to lipophilic (but non-lipidic) ammonium or poly-
ammonium cations. First, we assessed the ability of these water-soluble salts to cross a biological membrane by measuring the
partition coefficients between human serum and 1-octanol. In view of the ability of IHP to act as potent effector for oxygen release,
the O2–hemoglobin dissociation curves were then measured for the most efficient salts on whole blood. From both the biological
and the physical properties of IHP-ammoniums salts we determined that cycloalkylamines (or poly-amines) were the best transport
molecules, especially cycloheptyl- and cyclooctylamine. Indeed, the octanol/serum partition coefficient of IHP undecacycloocty-
lammonium salt, is superior to 1, which is very favorable for potential uptake into the red blood cell membrane. A qualitative
correlation was found between the partitioning experiments and the biological evaluations performed on whole blood. # 2002
Elsevier Science Ltd. All rights reserved.
Phytic acid (myo-Inositol HexakisPhosphate, InsP6 or
IHP) is the most abundant form of phosphate in
plants.1 This organic polyphosphate also has the prop-
erty to tightly bind to human hemoglobin.2 Hemoglobin
is a tetrameric protein that cooperatively fixes four O2
molecules and whose affinity for molecular oxygen is
allosterically regulated by 2,3-diphosphoglycerate
(DPG). By binding to hemoglobin, polyphosphates such
as DPG or phytic acid, trigger a decrease of the O2/
hemoglobin affinity and the subsequent release of oxy-
gen. Moreover, IHP interacts with hemoglobin 1000
times more tightly than DPG.3 This interesting property
makes IHP a good pharmaceutical candidate in the case
of diseases characterized by a limited oxygen flow to
organs or tissues (generally called ischemia, or ischemic
insult). Heart attacks and stroke are the most widely
recognized examples of the damage resulting from
ischemia. Furthermore, if brain and nerve cells are
deprived of oxygen for too long, irreversible brain
damage or cell death may occur. In a general manner,
IHP loaded erythrocytes would help to alleviate the
ischemic problems as well as diseases or damages
requiring repeated transfusions (such as anemia, sickled
cell disease, surgery. . .).
However, in physiological conditions, IHP bears at least
seven charges, making it very difficult to be transported
across cell membranes. Although IHP delivery has
already been achieved using a liposome delivery sys-
tem4,5 and electroporation techniques,6 a more classical
transport approach would be of much interest, invol-
ving in particular a IHP derivative that would be solu-
ble in both aqueous and low polarity media.
*Corresponding author. Fax: +33-3-9024-1117; e-mail: lehn@chimie.
u-strasbg.fr
0968-0896/02/$ - see front matter # 2002 Elsevier Science Ltd. All rights reserved.
PII: S0968-0896(02)00162-1

2826
S. P. Vincent et al. / Bioorg. Med. Chem. 10 (2002) 2825–2834
As emphasized in phase transfer catalysis, it has long
been recognized that organic and inorganic anions can
be efficiently solubilized in organic media when asso-
ciated to tri- or tetra-alkyl ammonium counter-cations.
IHP is a highly charged polyphosphate insoluble in
many organic solvents, but is also capable of multiple
ionic bonds with organic ammonium cations. Since pKa
values of all acidic protons of IHP have been measured
either by NMR,7 or by potentiometric methods,8 it has
been established that IHP bears 7 or 8 charges at
physiological pH. It also means that IHP can be asso-
ciated to at least seven lipophilic cations in physi-
ological conditions. Thus, depending on the associated
ammonium ions, the lipophilicity can theoretically be
shifted to a cell-membrane compatible IHP complex,
the ideal case being a poly-ammonium IHP derivative
that would be soluble in both water and low polarity
media. As a result of the very active research in the
development of artificial vectors for gene transfer, a
large number of cationic lipids have been described in
the literature, and extensively reviewed.9 In most of the
cases these chemical vectors are lipids functionalised
with amines,9 poly-amines,9 (poly-) guanidiniums,10,11
and, in rare cases, phosphonium cations.12 These catio-
nic lipids are designed for gene delivery, and the
mechanism by which they transport oligonucleotides
across biological membranes may be very different for
the delivery of a small poly-anionic molecule such as
phytic acid. Moreover, to avoid the technical problems
associated with drug delivery mediated by liposomes or
vesicles, we decided, as a first approach, to prepare
water-soluble lipophilyzed IHP derivatives. Thus, a col-
lection of IHP/ammoniums salts has first been prepared
from commercially available non-lipidic amines in order
to assess the structural parameters allowing both the
transport (by increasing the lipophilicity) and the water
solubility of the salts (for the improvement of the bioa-
vailability). Both the biological and physical properties
of each salt has been evaluated by the measurement of
O2 dissociation curves, performed on whole blood, and
by the determination of 1-octanol/serum partition coef-
ficients, respectively.
Partition Coefficients of IHP-ammoniums Derivatives
Partition coefficients relate to the distribution of a
solute between two immiscible liquid phases and are
defined as the ratios of concentrations (or molar frac-
tion) of the distributed solute. These data have been
used to predict or rationalize numerous drug properties
such as quantitative structure–activity relationship,13
lipophilicity14 and pharmacokinetic characteristics.15
1-Octanol has been selected to mimic biological mem-
branes, and it has been estimated than 1-octanol/water
(Kow) partition coefficients of more than 18,000 sub-
stances are now available in the literature.15
½IHPꢀoctanol
Kos
¼
½IHPꢀserum
Using 31P NMR, we have determined the partition
coefficients of IHP salts between a 1-octanol phase and
two different aqueous phases: water and human serum.
In Figure 1 are represented the structures of the differ-
ent ammonium ions preassociated to IHP and the cor-
responding partition coefficients Kow and Kos. These
coefficients are measured after equilibration, at a con-
centration of 30 mM, close to the typical concentration
employed for the biological evaluations (the standard
concentration used for O2–dissociation curve measure-
ments was 22 mM).
Choice of the Aqueous Phases
Because of the large number of comparable data
described in the literature we first measured 1-octanol/
water partition coefficients Kow. To be closer to physi-
ological conditions, we determined the partition coeffi-
cients Kos between 1-octanol and human serum.
Interestingly, we observed that Kos values are system-
atically lower (by a factor of 2–5) than Kow values.
Therefore, we only measured Kos’s when Kow values
were significant. To illustrate this observation we may
compare the partitions of two IHP derivatives, 3 and 4,
in different aqueous systems: water, artificial serum (for
its composition, see the experimental part) without and
with albumin, and human serum. Table 1 summarizes
these results.
IHP-ammonium Salts Preparation
IHP ammonium salts can be efficiently prepared from
IHP, dodecasodium form and the corresponding free
amine. IHP was first protonated using a cation-exchange
resin and then mixed with an ethanolic solution of the
desired amine. After evaporation the number of ammo-
This simple series of results emphasizes the problems
that can be encountered when experiments are trans-
posed from water to real physiological conditions.
Indeed, the proteins present in serum, such as albumin,
1
nium per IHP molecules was determined by H NMR.
Table 1. Partition coefficients (at 18 mM) of IHP salts 3 and 4 with different aqueous phases
Compds
Kow
(water)
Koa
(artificial serum without albumin)
Koas
(artificial serum with albumin)
Kos
(human serum)
3
IHP, 9
cycloheptylammonium
4
IHP, 11
cyclooctylammonium
0.411
8.98
0.088
2.16
0.041
1.81
0.031
1.85

S. P. Vincent et al. / Bioorg. Med. Chem. 10 (2002) 2825–2834
2827
Figure 1. Partition coefficients of IHP ammonium salts (at 30 mM); Kow=octanol/water partition coefficient, Kos=octanol/serum partition coeffi-
cient; the figures indicate the number of ion units per IHP molecule in the salt used in these experiments. Compound numbers are given in brackets,
counter-cation numbers are written in italics.
Analysis of the Results
as well as the cations that can be exchanged after equi-
libration, may play critical roles during a drug delivery
process. For instance, albumin has been shown to par-
ticipate in cholesterol transport in blood, thus demon-
strating its propensity to bind lipophilic compounds.16
The partition coefficients depicted in Table 1 show that
albumin only slightly decreases the ability of a given
organic ammonium to transport a polyphosphate into
an apolar phase. This conclusion may be correlated to
the results of a comparative study between octanol/
water an octanol/buffer partition coefficients, correlated
to diffusion across brain microvessel endothelium
(Fig. 1).17
The analysis of this first series of partition coefficients
already leads to interesting conclusions: cyclic aliphatic
amines seem to display the best characteristics with
regard to both lipophilicity and water solubility. Indeed,
we observed a very significant increase in the Kow values
from cyclopentyl- (<10ꢀ3) to cyclooctyl-ammoniums
(9.98). Furthermore, the IHP cyclooctylammonium salt
is still reasonably soluble in water (the aqueous solubi-
lity limit is between 25 and 30 mM), whereas the corre-
sponding n-octylammonium IHP salt presents
a
solubility limit below 1 mM (thus rendering impossible
its Kow measurement as well as its biological evalua-
tion). Interestingly, hydrophobic amino-acids, even
esterified, do not possess satisfactory transport proper-
ties into the octanol phase, and will not be considered
suitable for further studies aiming to IHP delivery into
This first series of experiments also showed that artifi-
cial serum, that can be easily prepared from commer-
cially available chemicals and BSA is a very good model
of human serum.

2828
S. P. Vincent et al. / Bioorg. Med. Chem. 10 (2002) 2825–2834
red blood cells. Thus, from this first library we selected
cycloalkyl groups for a further design of carrier
molecules.
Kos Variation as a Function of Cyclooctylammonium
Concentration
The results (Fig. 3) show two important characteristics
of the IHP transfer into an apolar phase: at a constant
IHP concentration (22 mM), 8 equivalents of cyclo-
octylammoniums are required to reach a Kos value equal
to 1, corresponding to an identical distribution between
human serum and octanol. Secondly, cyclooctylammo-
nium ions, initially present in their hydrochloride form
in the organic phase, are able to extract IHP, exclusively
present as a sodium salt in human serum at the begin-
ning of the experiment as a result of chloride/polyphos-
phate exchange between 1-octanol and human serum.
From a quantitative point of view, it is remarkable that
a hexaphosphate could reach Kos values around or
above
1
(compound
3
Kos=0.84, compound
4
Kos=1.85) indicating that more than 40% of the start-
ing polyanion diffused from human serum to the octa-
nolic phase after equilibration. As a comparison, several
studies described buffer/1-octanol partition coefficients
of uncharged antiviral nucleosides or nucleotides to
predict their ability to diffuse across cell membranes.
For instance, the Kow value of 30-azido-30-deoxy-
thymidine (AZT) was determined to be 1.26 (1-octanol/
100 mM sodium phosphate, pH 7.0) while the Kow of
thymidine in the same conditions was 0.064.18 More
recently, the partitioning properties of uncharged phos-
phonic acid esters of 20,30-dideoxythymidine gave Kow
values between 0.23 and 2.06.19
To improve the polyphosphate transfer efficiency we
reasoned that multiple interactions between a polyanion
and a single (or two) transport molecule(s) will increase
the binding strength between the two partners, thus
preventing or limiting the exchange, under physiological
conditions, with other cations present in high con-
centration in serum. For this reason, lipophilic poly-
amines seem to be the best candidates for transporting
polyphosphates, such as IHP or DPG, across non-polar
biological membranes. The 1-octanol/water partition
measurements showed that amines bearing cycloalkyl
groups display the best transport properties. Thus, we
optimized two general synthetic procedures in order to
obtain a new library of polyamines bearing cycloalkyl
groups: a reductive amination procedure leading to
acyclic tri- or tetra-amines, and a coupled acylation/
borane-reduction procedure for the preparation of
macrocyclic polyamines.
However, new questions arise from this first study con-
cerning the importance of the number of lipophilic
amines associated to IHP and the variation of IHP dis-
tribution in apolar phases as a function of the polyanion
concentration. To answer them, we measured Kow and
Kos variations for the two best IHP salts, 3 (9 cyclo-
heptylammoniums) and 4 (11 cyclooctylammoniums) as
a function of concentration. The resulting curves are
depicted in Figure 2.
Synthesis of Lipophilized Polyamines
A first procedure used for the derivatization of primary
amines with different cyclic ketones (Scheme 1), proved
efficient, giving good to excellent yields, and versatile
enough to allow reactions with hindered ketones such as
2-decalone. The two central tetra-amine cores employed
are TREN [tris-(2-aminoethy)-amine] and BAP [N,N0-
bis-(3-aminopropyl)-piperazine]. BAP is a linear tetra-
amine presenting two primary amines susceptible to be
alkylated, whereas TREN is a branched molecule pre-
senting three possible N-alkylation sites through reductive
Figure 2. Kow dependence on IHP starting concentration. (top) Com-
pound 4 (IHP, 11 cyclooctyl-ammonium) was dissolved in 1-octanol at
various concentrations and agitated with an equal volume of water or
human serum. (bottom) Compound 3 (IHP, 9 cycloheptyl-ammonium)
was dissolved in octanol at various concentrations and agitated with
an equal volume of water.
Figure 3. IHP uptake in 1-octanol by cyclooctylammonium ions. IHP,
dodecasodium form, was dissolved in human serum ([IHP]=22 mM,
pH=7.4). Each sample was agitated 2 days with an equal volume of
octanolic solution of cyclooctylamine,HCl at different concentrations.
Partition coefficients Kos were measured by 31P NMR.

S. P. Vincent et al. / Bioorg. Med. Chem. 10 (2002) 2825–2834
2829
Scheme 1. Poly-(cycloalkyl-amines) obtained by reductive amination.
amination. Seven new polyamines were thus obtained
from these two central cores, bearing either cyclohexyl
groups (25, 29 and 30), cycloheptyl group (27),
cyclooctyl- (28 and 32), or 2-decalinyl groups (26 and
31).
were performed only with water-soluble salts. Poly-
amines bearing aromatic or decalinyl groups generally
lead to insoluble salts. With its three lipophilized
amines, the TREN core structure (42) gave rise to the
most favorable uptake into 1-octanol, with both water
and human serum. At a concentration of 30 mM, the
Kow value for compound 42 is 2.71 (meaning that 73%
of starting IHP was found in 1-octanol after equilibra-
tion) while the Kos value is 0.64(39% of the total
amount of IHP stays in the apolar phase), which is an
excellent distribution for a hexaphosphate under nearly
physiological conditions. To compare with the partition
experiments of the first library (IHP/monoammonium
salts), it should be emphasized that, in the case of com-
pound 42, only two transfer molecules are required to
efficiently solubilize IHP into 1-octanol whereas at least
eight cycloalkylamines were necessary to reach similar
Kos values.
We then prepared a lipophilic polyamine derived from
Cyclam, a well studied and commercially available 14-
membered ring tetra-aza macrocycle.
This macrocycle was chosen in order to take advantage
of the strengthened interaction of polyammonium ions,
especially of macrocyclic type, with polyanions such as
polyphosphates.20,21 One of the best procedures to
obtain tetra-N-alkyl-cyclam derivatives consists in a
per-acylation of the four secondary amines of cyclam,
followed by reduction of the resulting tetra-amide using
an excess of borane-THF complex (Scheme 2).22 Taking
into account the lack of solubility of IHP n-octyl-
ammonium salts in water, we chose to introduce shorter
chains, n-hexyl lipophilic groups, for compound 34.
Biological Evaluation and Comparison with the
Physico-chemical Data
The IHP transport into red blood cells can be indirectly
determined by measuring the partial pressure of oxygen
of whole blood. Indeed, by binding to intracellular
hemoglobin, IHP triggers an oxygen release from the
erythrocytes. This release can be measured by standard
hemoglobin–O2 dissociation curves: P50 values refer to
the partial pressure of molecular oxygen for which half
of the hemoglobin molecules are bound to O2. A shift to
the right of the dissociation curve indicates a loss of
affinity of hemoglobin for O2 due to the interaction with
phytic acid (when compared to a control experiment
performed without IHP). The extent of IHP delivery
into erythrocytes may be considered as proportional to
the P50 shift. A 30 mM mother solution of a given IHP
salt was first prepared, the pH was adjusted to 7.2 and
the osmolality measured. This solution was then mixed
to the proper volume of whole blood (final IHP con-
centration: 22 mM) and the dissociation curve was
measured after homogenization. To properly assess
Partition Coefficient of IHP Associated with Polyamines
A second library of IHP salts of di-, tri- and tetra-
cycloalkylammonium ions was prepared using the stan-
dard procedure. The structures are listed in Figure 4.
Compounds 35, 36, 37, 38, 44 and 45 have been
obtained from commercially available diamines.
The partition coefficients Kow and Kos were measured by
31P NMR. Once again, the partitioning experiments
Scheme 2. Synthesis of a lipophilic macrocyclic polyamine.

2830
S. P. Vincent et al. / Bioorg. Med. Chem. 10 (2002) 2825–2834
Figure 4. Partition coefficients of IHP/poly-cycloalkylammoniums salts (at 30 mM); Kow=octanol/water partition coefficient, Kos=octanol/serum
partition coefficient; the figures indicate the number of ion units per IHP molecule in the salt used in these experiments. Counter-cation numbers are
written in italics.
which amines or polyamines could trigger an IHP
transport, we only compared data corresponding to
experiments displaying similar osmolality values, in the
range of 200–240 mOsM.
50% shift of the P50 value (compared to the
control experiment);
ꢁ amino-esters did not display transfer properties;
ꢁ tributyl-ammonium salts and isoleucine-tBu ester
provoked a fast hemolysis of red blood cells.
This lysis might be due either to detergent prop-
erties of the salts or to the blockage of potassium
channels.
From the biological evaluation described above (see
Table 2), several conclusions can be drawn:
ꢁ the best results were obtained with two IHP
cycloalkylammonium salts 2 (undeca cyclohexyl-
ammonium) and 3 (nona cycloheptyl-ammo-
nium). At 220 mOsM, these IHP salts triggered a
It is interesting to note that there is a good correlation
between the biophysical study (1-octanol/serum parti-
tion coefficients) and the biological evaluation (P50 shift
Table 2. Representative data from P50 shift measurements performed on whole blood under low osmolality conditions
a
Compound
P50 (Torr)
P50 controlb (Torr)
Shift (%)
Osmolalityc (mOsM)
1 (11 cyclopentylammonium)
2 (11 cyclohexylammonium)
3 (9 cycloheptylammonium)
IHP, 6 tributylammonium
8 (8 iLeu-CO2tBu)
38
26.8
23
Cell lysis
Cell lysis
32.5
25
38.5
24.5
23
41.5
25
39
42
31.5
—
3
57
37
—
—
6
14220
0
10
2424
2
173
220
224
—
17 (12 quinuclidinium)
34.5
28.5
38.5
27
28.5
42.5
25
240
22 (6 indanammonium)
5 (7, norbornylammonium)
11 (7 decahydroquinolinium)
9 (10 Tyr-CO2Et)
18 (6 N,N-dimethylammonium)
43 (3 tetraamine 32)
200
221
7
225
221
0
aMeasured on whole blood with [IHP]=22 mM, pH=7.2 at 37 ꢂC.
bWithout IHP derivative.
cMeasured after pH adjustment.

S. P. Vincent et al. / Bioorg. Med. Chem. 10 (2002) 2825–2834
2831
measurements), since both techniques lead to the con-
clusion that cycloalkylamines display the highest effi-
ciency for both partitioning of IHP into an octanolic
phase and transport of IHP across the erythrocyte
membrane.
NaCl, pH=7.45). Effector characteristics prior to incu-
bation were: c=30 mM, pH=7.1–7.4(at 37 ꢂC), osmo-
larity=170–340 mOsM. 75 mL of whole blood was
incubated 2 min at 37 ꢂC with 300 mL of a 30 mM
solution of IHP derivative. After incubation with/
without effector, the blood cells were washed four
times with HBS-BSA (20 mM HEPES, 130 mM NaCl,
pH=7.42 at 37 ꢂC) by centrifugal pelleting to remove
exogenous effector and evaluate hemolysis. 20 mL were
used for measurement of the hemoglobin–O2 dissocia-
tion curve at 37 ꢂC. Blood oxygen dissociation of
samples were determined using a Hemox Analyzer
Model B: the control sample contained 2.5 mL buffer
HBS (20 mM HEPES, 130 mM NaCl, 20 mL BSA per
5 mL buffer, pH=7.2–7.4at 37 ꢂC) and 25 mL of whole
blood; the effector sample contained 2.5 mL HBS and
20 mL of pelleted blood cells incubated with the IHP
derivative. The P50 values were calculated from the
dissociation curves compared to the same day control
sample.
Conclusion
The octanol/water and octanol/serum partition studies
described in this report provide significant information
regarding both polyanion uptake into apolar phases and
design of lipophilic polyamines as polyphosphate trans-
port molecules. Lipophilic ammonium ions bearing
cycloalkyl residues are of particular interest and
increasing their ring size leads to a remarkable
improvement of IHP distribution in an octanol phase.
At the same time, the water solubility of IHP associated
to cycloalkylamines is preserved compared to linear n-
alkylamines, which is a valuable feature for the devel-
opment of a drug delivery strategy. The biological tests
performed on the two IHP-ammonium cations libraries
are globally in good, qualitative, agreement with the
physico-chemical results obtained from the phase dis-
tribution studies. Further improvements in transfer effi-
ciencies may yield improved vectors for the delivery of
effectors such as IHP into red blood cells.
Partition coefficients
IHP derivatives were dissolved in 1 mL of aqueous
phase at a concentration of 30 mM in an eppendorf.
1 mL of 1-octanol was then added and each sample was
shaken at 36 rpm during 12 h. The equilibrated biphasic
solutions were centrifuged for 2 h at 3000 rpm. 350 mL of
the octanolic phase was first taken off with a syringe and
transferred into a NMR tube. 350 mL of the aqueous
phase was then taken off. 31P NMR spectra were
performed on a Bruker AC-300 spectrometer. An
external standard (triphenylphosphine oxide, 60 mM
in DMSO-d6) was placed inside the NMR tubes to
allow both the locking process of the spectrometer and
an accurate integration of IHP peaks. Partition coeffi-
cients were determined as the ratio of IHP integrations,
relative to the external standard, in the octanol and
aqueous phases. The detection limits of this technique
do not allow partition coefficient measurements below
Experimental
Materials and procedures
All chemicals (starting amines or diamines, IHP,
cycloalkanones. . .) were purchased from Sigma, Aldrich
or Fluka and were used without further purification.
The resin Dowex 50WX8–200 was purchased from
Aldrich and washed with distilled water before the first
1
use. H, 13C and 31P NMR spectra were recorded with
Bruker AC-200 or AC-300 spectrometers. Mass spectra
were determined by the Service commun de spectro-
metrie de masse at the Institut de Chimie, Universite
Louis Pasteur. Human serum was purchased at the
Centre de Transfusion Sanguine de Strasbourg, stored
at ꢀ18 ꢂC and used such as received. Artificial serum
was based on blood composition: [Na+]=140 mM,
[Mg2+]=1 mM, [K+]=5 mM, [Ca2+]=1 mM, [Clꢀ]=
106 mM, [PO34ꢀ]=1 mM, [SO42ꢀ]=0.5 mM, [CO23ꢀ]=
30 mM, [Brꢀ]=0.5 mM, bovine serum albumin (fraction
10ꢀ3
.
Preparation of the IHP polyammoniums derivatives
A 100 mM IHP dodecasodium salt solution was applied
on a cation exchange column (Dowex 50WX8–200, H+
form) and eluted with distilled water. The fractions
containing the perprotonated IHP were collected and
poured into an ethanolic solution of the desired amine.
The solution was then concentrated to dryness in vacuo.
The product was then dissolved in ethanol or in a 1/1
toluene/EtOH mixture, concentrated and dried in
vacuo. The IHP-ammonium salts were characterized by
1H and 31P NMR. The final salt composition was
V, >98%) 30 g Lꢀ1, pH=7.41. Blood oxygen dissocia-
.
tion of samples were determined using a Hemox Analy-
zer Model B (TCS Medical Products Company, New
Hope, PA, USA).
1
determined by H NMR.
Blood preparation
Whole blood was collected from one subject. The blood
was stored in a Vacutainer and stored at 4–8 ꢂC.
Tris-(N-cyclohexyl-2-aminoethyl)-amine (25). To a solu-
tion of TREN [tris-(2-aminoethyl)-amine, 2.24g,
15.3 mmol], cyclohexanone [15 mL, 156.4mmol,
10.2 equiv], in 60 mL methanol and 2.7 mL acetic acid,
was added, at 0 ꢂC, sodium cyanoborohydride (3 g,
79.6 mmol, 9.4equiv) in three portions. The solution
was stirred 16 h at room temperature, then fresh sodium
Bioassay conditions
Effector stocks were prepared at 100–120 mM using
water or Bis–Tris buffer (20 mM Bis–Tris, 140 mM

2832
S. P. Vincent et al. / Bioorg. Med. Chem. 10 (2002) 2825–2834
cyanoborohydride (2 g, 52.9 mmol, 6.3 equiv) was
added. After 20 h, the reaction was quenched with
20 mL NaOH, 1 N. The mixture was stirred 30 min,
diluted with 30 mL H2O and washed three times with
70 mL CH2Cl2. The combined organic phases were
concentrated, dissolved in 20 mL ethanol, and 10 mL of
HCl 12 N were added. After concentration, the resulting
white solid was abundantly rinsed with Et2O and
recrystallized in a 95/5 EtOH/H2O mixture. Yield: 92%.
White solid, C24H48N4Cl4, Mr=538.7. 1H NMR
(200MHz, D2O) d 3.15 (m, 8H), 2.89 (m, 6H), 2.1–1.0 (m,
30H). 13C NMR (50 MHz, D2O) d 57.80, 49.09, 41.19,
28.95, 24.54, 24.06. MS (elec. ionization): 100%=280.1
[M-(N-cyclohex.-methyl-amine)], 50%=268.1 [M-(N-
cyclohex.-ethyl-amine)], 10%=393.3 (M+1).
2.7 equiv) in three portions. The solution was stirred 3
days at room temperature (every day, 0.3 g sodium cya-
noborohydride was added). The reaction was quenched
with 20 mL NaOH, 1 N. The mixture was stirred 30 min,
diluted with 30 mL H2O and washed three times with
70 mL CH2Cl2. The combined organic phases were
concentrated, dissolved in 20 mL ethanol, and 10 mL of
HCl 12 N were added. After concentration, the resulting
white solid was abundantly rinsed with Et2O. Yield :
93%. White solid, C30H64N4Cl4, Mr=622.1H NMR
(200 MHz, CDCl3/CD3OD 1/1) d 3.39 (m, 3H), 3.27 (bt,
J=4.9 Hz, 6H), 2.97 (bt, J=4.6 Hz, 6H), 2.10 (m, 6H),
1.85 (m, 12H), 1.70–1.40 (m, 24H). 13C NMR (50 MHz,
CDCl3) ꢀ 60.61, 51.97, 40.94, 29.00, 26.28, 25.47,
23.67.MS (FAB, positive mode): M+1=477.5. El.
analysis: calcd (%) C 57.87, H 10.36, N 9.00; meas. C
58.14, H 10.55, N 9.17.
Tris-(N-(2-decalinyl)-2-aminoethyl)-amine (26). To
a
solution of TREN [tris-(2-aminoethyl)-amine, 0.7 mL,
4.7 mmol], 2-decalone (mixture of isomers, 2.17 g,
14.3 mmol, 3.2 equiv), in 15 mL methanol and 1 mL
acetic acid, was added, at 0 ꢂC, sodium cyanoborohy-
dride (1.1 g, 17.5 mmol, 3.7 equiv) in three portions. The
solution was stirred overnight at room temperature,
then 0.3 g sodium cyanoborohydride was added. After
6 h, the reaction was quenched with 10 mL NaOH, 1 N.
The mixture was stirred 30 min, diluted with 30 mL H2O
and washed three times with 70 mL CH2Cl2. The com-
bined organic phases were concentrated, dissolved in
20 mL ethanol, and 10 mL of HCl 12 N were added.
After concentration, the resulting white solid was
abundantly rinsed with Et2O. Yield: 88%. White solid,
Bis-(N-cyclohexyl-3-aminopropyl)-amine (29). To
a
solution of dipropylene triamine (2.1 g, 16.0 mmol),
cyclohexanone (5.1 mL, 48.9 mmol, 3.1 equiv), sodium
cyanoborohydride (2.9 g, 46.1 mmol, 2.9 equiv) in 40 mL
methanol, was added, at 0 ꢂC, 2.7 mL acetic acid. The
solution was stirred 24h at room temperature then
quenched with 30 mL saturated sodium carbonate. The
mixture was stirred 30 min, diluted with 30 mL H2O and
washed three times with 70 mL CH2Cl2. The combined
organic phases were concentrated, dissolved in 20 mL
ethanol and 10 mL of HCl 12 N were added. After con-
centration, the resulting white solid was abundantly
rinsed with Et2O and recrystallized in a 95/5 EtOH/H2O
mixture. Yield: 68%. White solid, C18H40N3Cl3,
1
C36H70N4Cl4, Mr=700.6. H NMR (200 MHz, CDCl3)
1
d 9.5 and 9.4 (2 bs, 4H), 4.16 (m, 5H), 3.78 (m, 6H), 3.01
(m, 4H), 1.90–1.40 (m, 48H). MS (FAB, positive mode):
M+1=555.4.
Mr=405.0. H NMR (200 MHz, D2O) d 3.15 (m, 10H),
2.21 (m, 8H), 1.79 (m, 4H), 1.63 (bd, 2H), 1.40–1.00 (m,
12H). 13C NMR (50 MHz, D2O) d 57.61, 44.82, 41.33,
28.94, 24.57, 23.98, 22.99. MS (EI): 100%=112.1 (N-
cyclohexyl-methyl-iminium); 30%=296.2 (M+1); El.
analysis: calcd (%) C 53.40, H 9.96, N 10.38; meas. C
53.67, H 10.16, N 10.15.
Tris-(N-cycloheptyl-2-aminoethyl)-amine (27). To
a
solution of TREN [tris-(2-aminoethyl)-amine, 2.52 g,
17.2 mmol], cycloheptanone (7.0 mL, 59.3 mmol,
3.5 equiv), in 60 mL methanol and 3 mL acetic acid, was
added, at 0 ꢂC, sodium cyanoborohydride (3.5 g,
55.7 mmol, 3.3 equiv) in three portions. The solution
was stirred 16 h at room temperature, then 1 g sodium
cyanoborohydride was added. After 20 h, the reaction
was quenched with 20 mL NaOH, 1 N. The mixture was
stirred 30 min, diluted with 30 mL H2O and washed
three times with 70 mL CH2Cl2. The combined organic
phases were concentrated, dissolved in 20 mL ethanol,
and 10 mL of HCl 12 N were added. After concentra-
tion, the resulting white solid was abundantly rinsed
with Et2O and recrystallized in methanol. Yield: 84%.
White solid, C27H58N4Cl4, Mr=580. 1H NMR
(200 MHz, D2O) d 3.23 (h, J=4.6 Hz, 3H), 3.14 (t,
J=6.1 Hz, 6H), 2.89 (t, J=6.1 Hz, 6H), 1.97 (m, 6H),
1.70–1.30 (m, 30H). 13C NMR (50 MHz, D2O) d 62.06,
51.07, 43.16, 32.51, 29.16, 25.38. MS (FAB, positive
mode): M+1=435.4.
N,N0-di-[3-Cyclohexylamino-propyl]-piperazine (30). To
a solution of bis-(3-aminopropyl)-piperazine (3.80 g,
19.0 mmol), cyclohexanone (4.0 g, 40.8 mmol, 2.2 equiv),
in 40 mL methanol and 2.0 mL acetic acid, was added,
at 0 ꢂC, sodium cyanoborohydride (2.0 g, 31.8 mmol,
1.7 eq.) in three portions. The solution was stirred over-
night at room temperature. Sodium cyanoborohydride
(1.0 g, 15.9 mmol, 0.85 equiv) was added, and the reac-
tion mixture was stirred 6 h at room temperature. The
reaction was quenched with 10 mL H2O and 40 mL satd
sodium carbonate. The mixture was stirred 30 min,
diluted with 30 mL H2O and washed three times with
70 mL CH2Cl2. The combined organic phases were
concentrated, dissolved in 20 mL ethanol, and 10 mL of
HCl 12 N were added. After concentration, the resulting
white solid was abundantly rinsed with Et2O and
recrystallized in a 9/1 EtOH/H2O mixture. Yield: 82%.
White solid, C22H48N4Cl4, Mr=510. 1H NMR
(200 MHz, D2O) d 3.71 (s, 8H), 3.37 (t, J=7.9 Hz, 4H),
3.15 (t, J=7.6 Hz, 6H), 2.15 (m, 8H), 1.78 (m, 4H), 1.60
(bd, 2H), 1.40–1.15 (m, 10H). 13C NMR (50 MHz, D2O)
d 59.63, 55.66, 50.99, 43.13, 30.85, 26.51, 25.93, 23.06.
MS (FAB, positive mode): M+1=465.3.
Tris-(N-cyclooctyl-2-aminoethyl)-amine (28). To a solu-
tion of TREN [tris-(2-aminoethyl)-amine, 2.35 g,
16.7 mmol], cyclooctanone (9.3 g, 73.7 mmol, 4.4 equiv),
in 30 mL methanol and 2 mL acetic acid, was added, at
0 ꢂC, sodium cyanoborohydride (2.8 g, 44.6 mmol,

S. P. Vincent et al. / Bioorg. Med. Chem. 10 (2002) 2825–2834
2833
N,N0-di- [3- (N- (2-Decalinyl)- amino)- propyl] piperazine
(31). To a solution of bis-(3-aminopropyl)-piperazine
(0.80 mL, 4.0 mmol), 2-decalone (mixture of isomers,
1.23 g, 8.1 mmol, 2.1 equiv), in 20 mL methanol and
1 mL acetic acid, was added, at 0 ꢂC, sodium cyanobor-
ohydride (1.9 g, 30.2 mmol, 7.7 equiv) in three portions.
The solution was stirred 60 h at room temperature. The
reaction was quenched with 10 mL NaOH, 1 N. The
mixture was stirred 30 min, diluted with 30 mL H2O and
washed three times with 70 mL CH2Cl2. The combined
organic phases were concentrated, dissolved in 20 mL
ethanol, and 10 mL of HCl 12 N were added. After
concentration, the resulting white solid was abundantly
rinsed with Et2O and recrystallized in a 9/1 EtOH/H2O
mixture. Yield: 71%. White solid, C30H60N4Cl4,
Mr=618. 1H NMR (200 MHz, D2O) d 3.27 (s, 8H), 3.38 (t,
J=7.9 Hz, 4H), 3.14 (m, 6H), 2.17 (m, 4H), 1.90–1.15 (m,
32H). 13C NMR (50MHz, D2O) d (the spectrum displays
two sets of peaks corresponding to two different isomers of
the decaline moiety, the major isomer is labelled M, the
minor one m) 60.47(M), 56.29(m), 55.70(M), 50.96(M),
43.27(m), 43.06(M), 37.06(m), 36.58(M), 36.50(m),
36.14(M), 33.13(M), 32.58(m), 31.66(M), 30.85(m),
30.08(M), 28.31(M), 27.91(m), 26.81(M), 25.45(M),
23.06(M), 22.47(m), 22.21(M). MS (FAB, positive
mode): M+1=473.4. El. analysis: calcd (%) C 58.14, H
9.97, N 8.91; meas. C 58.25, H 9.78, N 9.06.
as a white solid. 1H NMR (200 MHz, CDCl3) d 3.46 (m,
16H), 2.37 (t, J=7.3 Hz, 4H), 2.28 (t, J=7.9 Hz, 4H),
1.88 (m, 4H), 1.64 (m, 8H), 1.30 (m, 16H), 0.87 (m,
12H). 13C NMR (50 MHz, CDCl3) d 174.56, 173.83,
48.06, 47.55, 45.71, 33.36, 32.88, 31.66, 28.91, 25.27,
24.94, 22.55, 13.99. MS (FAB, positive mode)
M+1=593. El. analysis: calcd (%) C 68.88, H 10.88, N
9.45; meas. C 68.79, H 11.14, N 9.26.
N,N0,N00,N000-Tetrahexyl-cyclam (34). The general proce-
dure of amide reduction22 was applied from the cyclam
derivative 33. The tetraamine was isolated as a colorless
oil. C34H72N4, Mr=536.4. 1H NMR (200 MHz, CDCl3)
d 2.53 (s, 8H), 2.47 (t, J=7.0 Hz, 4H), 2.37 (dd,
J=7.6 Hz J=7.0 Hz, 4H), 1.57 (m, 4H), 1.45–1.15 (m,
32H), 0.88 (t, J=6.4MHz, 12H). 13C NMR (50 MHz,
CDCl3) d 55.83, 51.64, 50.72, 31.83, 27.42, 27.27, 22.64,
14.00. MS (FAB, positive mode) M+1=537.5 ; El.
analysis: calcd (%) C 76.05, H 13.52, N 10.43; meas. C
73.31, H 13.72, N 10.20.
Acknowledgements
The authors are grateful to GMP Oxycell for financial
support.
N,N0-di-[3-Cyclooctylamino-propyl]-piperazine (32). To a
solution of bis-(3-aminopropyl)-piperazine (1.33 g,
6.6 mmol), cyclooctanone (4.6 g, 36.4 mmol, 5.5 eq.), in
30 mL methanol and 1.0 mL acetic acid, was added, at
0 ꢂC, sodium cyanoborohydride (2.3 g, 36.6 mmol,
5.5 equiv) in three portions. The solution was stirred
12 h at 50 ꢂC. Sodium cyanoborohydride (1.0 g,
15.9 mmol) was added, and the reaction mixture was
stirred 12 h at 50 ꢂC. The reaction was cooled down at
room temperature and quenched with 10 mL H2O and
10 mL NaOH, 1 N. The mixture was stirred 30 min,
diluted with 30 mL H2O and washed three times with
70 mL CH2Cl2. The combined organic phases were
concentrated, dissolved in 20 mL ethanol, and 10 mL of
HCl 12 N were added. After concentration, the resulting
white solid was abundantly rinsed with Et2O. Yield:
96%. White solid, C26H56N4Cl4, Mr=566. 1H NMR
(200 MHz, D2O) d 3.72 (s, 8H), 3.38 (dd, J=7.9 Hz
J=11.1 Hz, 6H), 3.15 (t, J=7.9 Hz, 4H), 2.19 (m, 4H),
1.99 (m, 4H), 1.80–1.30 (m, 24H). 13C NMR (50 MHz,
D2O) d 61.47, 55.66, 50.99, 43.61, 30.74, 27.87, 27.14,
25.19, 23.13. MS (FAB, positive mode): M+1=421.5.
El. analysis: calcd (%) C 55.12, H 9.96, N 9.89; meas. C
55.36, H 10.23, N 9.93.
References and Notes
1. Skoglund, E.; Carlsson, N.-G.; Sandberg, A.-S. J. Agric.
Food Chem. 1998, 46, 1877.
2. Benesch, R.; Benesch, R. E.; Yu, C. Y. Proc. Nat. Acad.
Sci. USA 1968, 59, 526.
3. Benesch, R. E.; Edalji, R.; Benesch, R. Biochemistry 1977,
16, 2594.
4. Teisseire, B.; Ropars, C.; Vallez, M. O.; Herigault, R. A.;
Nicolau, C. J. Appl. Phys. 1985, 58, 1810.
5. Teisseire, B.; Ropars, C.; Villereal, M.-C.; Nicolau, C.
Proc. Nat. Acad. Sci. U.S.A. 1987, 84, 6894.
6. Bruggeman, U.; Roux, E. C.; Hannig, J.; Nicolau, C.
Transfusion 1995, 35, 478.
7. Isbrandt, L. R.; Oertel, R. P. J. Am. Chem. Soc. 1980, 102,
3144.
8. Bieth, H.; Spiess, B. J. Chem. Soc., Faraday Trans. 1986, 1,
25 6701.
9. Miller, A. D. Angew. Chem., Int. Ed. 1998, 37, 1768.
10. Vigneron, J.-P.; Oudrhiri, N.; Fauquet, M.; Vergely, L.;
Bradley, J.-C.; Basseville, M.; Lehn, P.; Lehn, J.-M. Proc. Nat.
Acad. Sci. U.S.A. 1996, 96, 9682.
11. Oudrhiri, N.; Vigneron, J.-P.; Peuchmaur, M.; Leclerc, T.;
Lehn, J.-M.; Lehn, P. Proc. Nat. Acad. Sci. U.S.A. 1997, 94, 1651.
12. Guenin, E.; Herve, A.-C.; Floch, V.; Loisel, S.; Yaouanc,
J.-J.; Clement, J.-C.; Ferec, C.; des Abbayes, H. Angew.
Chem., Int. Ed. 2000, 39, 629.
13. Hansch, C.; Leo, A. Exploring QSAR; Fundamentals and
Applications in Chemistry and Biology; American Chemical
Society: Washington, DC, 1995.
14. Balzarini, J.; Cools, M.; Clercq, E. D. Bioch. Biophys. Res.
Comm. 1986, 26, 6701.
15. Sangster, J. Octanol/Water Partition Coefficients: Funda-
mentals and Physical Chemistry; John Wiley & Sons: Chich-
ester, 1997.
N,N0,N00,N000-Tetrahexanoyl-cyclam
(33).
Hexanoyl
chloride (0.8 mL, 5.82 mmol, 5.2 equiv) was added
dropwise, at 0 ꢂC, to a solution of cyclam (224mg,
1.12 mmol) and 1.0 mL Et3N in 15 mL anhydrous
CH2Cl2. This solution was stirred 24h at room tem-
perature, washed twice with satd NaHCO3, dried over
MgSO4, concentrated under vacuum, and chromato-
graphed on silica gel (elution conditions: AcOEt/hexane
1/1 until all the unreacted acyl chloride was totally
eluted, then pure AcOEt). The tetra-amide was isolated
16. Zhao, Y.; Marcel, Y. L. Biochemistry 1996, 35, 7174.
17. Rim, S.; Audus, K. L.; Borchardt, R. T. Int. J. Pharm.
1986, 3, 79.

2834
S. P. Vincent et al. / Bioorg. Med. Chem. 10 (2002) 2825–2834
18. Zimmerman, T. P.; Mahony, W.; Prus, K. L. J. Biol.
Chem. 1987, 262, 5748.
19. Meier, C. Angew. Chem., Int. Ed. 1993, 32, 1704.
20. Dietrich, B.; Hosseini, M. W.; Lehn, J. M.; Sessions, R. B.
J. Am. Chem. Soc. 1981, 103, 1282.
21. Hosseini, M. W.; Lehn, J. M.; Maggiora, L.; Mertes,
K. B.; Mertes, M. P. J. Am. Chem. Soc. 1987, 109, 537.
22. Alcock, N. W.; Benniston, A. C.; Grant, S. J.; Omar,
H. A. A.; Moore, P. Chem. Commun. 1991, 1573.