Substrate specificity of the bovine serum amine oxidase and in situ characterisation of aminoaldehydes by NMR spectroscopy

Article abstract of DOI:10.1016/j.bmc.2005.03.020

The oxidation of spermidine or homospermidine with bovine serum amine oxidase (BSAO) was monitored in situ, using proton nuclear magnetic resonance spectroscopy in water with 10% D₂O. NMR assignments were performed by spin decoupling and COSY spectra or by comparison with data from synthetic aminoaldehydes. The results represent the first in situ characterisation of the highly reactive aminoaldehydes and showed oxidation at the N¹ amino group of spermidine and homospermidine. Comparison of homospermidine with a variety of substrates revealed that among straight chain di- and polyamines both an aminopropyl group and two primary amino groups separated by seven (norspermidine) or eight (spermidine) carbon atoms were required for optimal substrate ability. However, highest activity was seen with the substrate N-(4-aminobutyl)hexahydropyrimidine, showing that the substrate channel of BSAO has a dual substrate preference, with moderately bulky substituents at the distal end of a diamine contributing equally well as an alkyl amino group. Cytotoxic investigations of a variety of substrates for BSAO, confirmed previous results, that cytotoxicity is primarily linked to polyamines encompassing the aminopropyl moiety. No acrolein was observed at any time during the oxidation showing that it reacts very fast with available amino groups forming a variety of derivatives.

Full text of DOI:10.1016/j.bmc.2005.03.020

Bioorganic & MedicinalChemistry 13 (2005) 3783–3796
Substrate specificity of the bovine serum amine oxidase and
in situ characterisation of aminoaldehydes by NMR spectroscopy
Gunnar Houen,a,* Casper Struve,b Roar Søndergaard,b Tina Friis,a Uffe Anthoni,b
Per H. Nielsen,b Carsten Christophersen,b Bent O. Petersenc and Jens Ø. Duusc
aDepartment of Research and Development, Statens Serum Institute, Artillerivej 5, DK-2300 Copenhagen, Denmark
bMarine Chemistry Section, Department of Chemistry, University of Copenhagen, The H. C. Ørsted Institute,
Universitetsparken 5, DK-2100 Copenhagen, Denmark
cCarlsbergLaboratory, Gamle CarlsbergVej 10, DK-2500 Valby, Denmark
Received 9 July 2004; revised 7 March 2005; accepted 11 March 2005
Available online 6 April 2005
Abstract—The oxidation of spermidine or homospermidine with bovine serum amine oxidase (BSAO) was monitored in situ, using
proton nuclear magnetic resonance spectroscopy in water with 10% D2O. NMR assignments were performed by spin decoupling and
COSY spectra or by comparison with data from synthetic aminoaldehydes. The results represent the first in situ characterisation of
the highly reactive aminoaldehydes and showed oxidation at the N1 amino group of spermidine and homospermidine.
Comparison of homospermidine with a variety of substrates revealed that among straight chain di- and polyamines both an
aminopropylgroup and two primary amino groups separated by seven (norspermidine) or eight (spermidine) carbon atoms were
required for optimalsubstrate ability. However, highest activity was seen with the substrate N-(4-aminobutyl)hexahydropyrimidine,
showing that the substrate channel of BSAO has a dual substrate preference, with moderately bulky substituents at the distal end of
a diamine contributing equally well as an alkyl amino group. Cytotoxic investigations of a variety of substrates for BSAO, con-
firmed previous results, that cytotoxicity is primarily linked to polyamines encompassing the aminopropyl moiety. No acrolein
was observed at any time during the oxidation showing that it reacts very fast with available amino groups forming a variety of
derivatives.
ꢀ 2005 Elsevier Ltd. All rights reserved.
1. Introduction
Abbreviations: CAO, copper-containing amine oxidase; BSAO, bovine
serum amine oxidase; HRP, horseradish peroxidase; p-NPP, p-nitro-
The intracellular polyamine level is intimately linked to
cell proliferation and apoptosis1 and is regulated by
multiple pathways including acetylation and oxidative
phenylphosphate substrate tabelts; bFGF, recombinant human basic
fibroblast growth factor; VEGF, recombinant human vascular endo-
degradation by polyamine oxidase (PAO).2,3 Extracellu-
thelial growth factor; HUVEC, human umbilical vein endothelial cells;
larly, the oxidative deamination of polyamines and other
PET, polyethylene terephthalate; DSS, 2,2-dimethyl-2-silapentane-5-
amines is catalysed by copper-containing amine oxidases
(CAOs).4–8 The bovine serum amine oxidase (BSAO) is a
sulfonate; EBM-2, endothelial cell basal medium-2; FBS, foetal bovine
serum; hbFGF, human basic fibroblast growth factor; hVEGF, human
particularly interesting enzyme, which is easily purified
and gives rise to highly cytotoxic products by the oxida-
tion of the naturalpoylamines spermidine and sperm-
ine.9–24 BSAO is present in foetalcafl serum (FCS),
and cellular studies carried out in the presence of FCS
must take into account the possible oxidation of polyam-
ines by BSAO. The importance of being able to identify
the oxidation products has given rise to severalstudies of
substrate selectivity and regiospecificity. Previous inves-
tigations were based on studies of secondary transforma-
tions followed by product analysis, using TLC, HPLC,
MS and NMR spectroscopy.25–27 Tabor et al., using
vascular endothelial growth factor; hEGF, human epidermal growth
factor; R3-IGF-1, long R insulin-like growth factor-1; GA-1000, gen-
tamicin sulfate; PBS, phosphate-buffered saline; VLC, vacuum liquid
chromatography; ECAO, Escherichia coli amine oxidase; PSAO, pea
seedling amine oxidase; AGAO, Arthrobacter globiformis amine oxi-
dase; HPAO, Hansenula polymorpha amine oxidase; PAO, polyamine
oxidase; DQF-COSY,
spectroscopy.
double-quantum
filtered
correlation
Keywords: Bovine serum amine oxidase; Spermidine; Homospermi-
dine; N-(4-Aminobutyl)hexahydropyrimidine; NMR spectroscopy;
Cytotoxicity; Acrolein.
*
Corresponding author. Tel.: +45 3268 3276; fax: +45 3268 3149;
e-mail: gh@ssi.dk
0968-0896/$ - see front matter ꢀ 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmc.2005.03.020

3784
G. Houen et al. / Bioorg. Med. Chem. 13 (2005) 3783–3796
Scheme 1. The BSAO oxidation of spermidine and homospermidine.
Table 1. Substrate activities towards BSAO
TLC in combination with NaBH4 reduction of alde-
hydes, concluded that spermidine and spermine were oxi-
dised at primary amino groups25 (Scheme 1). Houen et
al., using NMR spectroscopy, HPLC and MS, concluded
that oxidation of these two polyamines occurred at sec-
ondary amino groups.26 In both of these studies, b-elim-
ination of acrolein was a major complicating problem.
Consequently, Lee and Sayre27 used homospermidine
as a substrate incapable of b-elimination (Scheme 1) in
combination with NaBH4 reduction and acetylation fol-
lowed by NMR spectroscopy of the reduced and acety-
lated aminoaldehydes and concluded that oxidation
had occurred at the primary amino group. The stereo-
chemistry of the oxidation of dopamine was elucidated
Substrate
Relative activitya
Polyamines
Norspermidine
100
86
53
49
8
Spermidine
Spermine
Homospermidine
1,1,9,9-Tetradeuterio-homospermidine
N-(2-Aminoethyl)-putrescine
7
Straight chain mono- and diamines
Octylamine
46
38
30
23
13
13
9
Octane-1,8-diamine
Decane-1,10-diamine
Dodecane-1,12-diamine
Putrescine
1
by deuterium labelling and reduction prior to H NMR
analysis28 and the structure of reaction intermediates,
for example, from benzylamine has been probed using
UV/vis and Resonance Raman spectroscopy.29 We have
carried out a comparative analysis of homospermidine
and other amines with respect to substrate specificity
and cytotoxicity, and report an improved method for
monitoring and identifying the oxidation products of
polyamines in situ using 1H NMR spectroscopy in water
containing 10% D2O.
Propane-1,3-diamine
Heptane-1,7-diamine
Hexane-1,6-diamine
Cadaverine
8
7
Moderately bulky amines
N-(4-Aminobutyl)hexahydropyrimidine
N-(3-Aminopropyl)hexahydropyrimidine
Benzylamine
110
61
30
19
11
7
N-Aminobutylpyrrolidine
Decane-2,9-diamine
Histamine
a Relative substrate activities of selected amines (0.1 mM) with BSAO
(0.01 mg/mL), in 20 mM bis–tris pH 7.0.
2. Results
2.1. The substrate specificity of BSAO
In order to examine the substrate specificity of BSAO,
and in particular to compare homospermidine with sper-
midine and related polyamines, a number of amines
were selected for comparison. Compounds were either
commercially available or synthesised, and the relative
substrate activities were determined using a peroxi-
dase-coupled chemiluminescence assay. The results ap-
pear in Table 1. The data first of all demonstrate that
the presence of long chains and hydrophobic character
of the amines favour an accelerated rate of enzymatic
oxidation. The diamines (H2N(CH2)nNH2) in particular
demonstrate this trend as illustrated graphically in Fig-
ure 1, that is, a long chain is required for maximum
activity (n = 8 to 10) while the shorter diamines (n = 3
to 7) show diminished substrate activity. In the case of
decane-2,9-diamine a lower activity is observed com-
pared to those of both decane-1,10-diamine and oc-
tane-1,8-diamine. Secondly, substrates containing the
aminopropyl moiety are always superior to those lack-
ing this group but with comparable chain length. Thus,
norspermidine containing two aminopropylgroups re-
acts faster than both homospermidine (two aminobutyl
groups) and N-(2-aminoethyl)-putrescine (carrying one
Figure 1. Relative substrate activities of straight-chain diamines at a
concentration of 0.1 mM.
aminobutyland one aminoethylgroup). The presence
of moderately bulky substituents in the non-oxidised
end of the amines does not appear to have a major
influence on the substrate activities, provided that the
substituents are removed from the amino group. In
particular the high activity of N-(4-aminobutyl)-hexahy-
dropyrimidine renders it among the best substrates for
BSAO.

G. Houen et al. / Bioorg. Med. Chem. 13 (2005) 3783–3796
3785
2.2. Enzymatic reaction with spermidine
around 3 ppm arising from the CH2 groups attached
to the protonated amino groups, a quintet at 2 ppm
due to the centralCH 2 group of the trimethylene moiety
and two superimposed quintets around 1.65 ppm attrib-
uted to the corresponding CH2CH2 groups of the
tetramethylene grouping. Addition of three portions of
enzyme resulted in almost complete conversion to the
aminoaldehyde after 24 h (30% after 2 h, 40% after
5 h) as apparent from the replacement of the quintet
at 2 ppm with a new signalat 1.86 ppm in the olwer
spectrum (Fig. 2A). The spectrum is identicalwith the
spectrum previously obtained in D2O26 with the sole
exception that the signalat 1.86 ppm was absent in
D2O, corresponding to a CH2 group, which is easily
deuterated. Irradiation at 3 ppm verified that this signal
is due to a CH2 group attached to a protonated amino
group, and since the remaining doublet showed a cou-
pling constant of J(H,H) = 4.8 Hz it must be situated
a to a hydrated aldehyde group. The DQF-COSY spec-
tra (Fig. 2C) further served to establish the presence of
the sequence N+–CH2CH2CH(OH)2. The deuteration
with formation of N+–CH2CD2CH(OH)2 must be due
to an equilibrium with free aldehyde, however, in
amounts too small to be observed in the NMR spectra.
Nevertheless, the NMR spectra confirm that oxidation
does in fact occur at the primary amino group. When
The oxidation of spermidine by purified BSAO in H2O
with 10% D2O was monitored by NMR spectroscopy
using a jump-return sequence or water presaturation.
Using 100% D2O as a solvent, it was previously reported
that spermidine is converted to putrescine.26 However,
the intermediate aldehyde could not be identified with
certainty from the NMR pattern due to deuteration of
the CH2 group a to the aldehyde group. Based upon
other evidence, an oxidation sequence involving the sec-
ondary amino group seemed probable, and the NMR
spectra could, in fact, be interpreted to support this con-
tention. Lee and Sayre27 later obtained results for homo-
spermidine supporting a regiospecific oxidation at the
primary amino group. Therefore, we decided to reinves-
tigate the enzymatic oxidation of spermidine and homo-
spermidine by NMR spectroscopy. The concentration of
the aqueous solution was chosen to be 1 mM in spermi-
dine. This is small enough to ascertain the natural func-
tion of the enzyme but simultaneously allows a
satisfactory signal-to-noise ratio. The enzymatic conver-
sion was followed either at 400 or 800 MHz, and the sig-
1
nals assigned from 1D H and 2D DQF-COSY spectra
at 800 MHz.
1
Before the enzyme is added to the NMR tube, spermi-
taking the signal-to-noise ratio of the H NMR spectra
F
dine,
NH3+CH2ACH2BCH2CCH2DNH2+CH2ECH2
-
into account, it can be seen, that over 90% of the oxida-
tion takes place at the aminopropyl moiety of
spermidine.
CH2GNH3+, displayed the following NMR spectrum:
dH 3.00 and 2.98 (2H, t, J(H,H) = 7.8 Hz, CH2E) and
CH2G, 2.96 (2H, t, J(H,H) = 7.5 Hz, CH2D), 2.93 (2H,
t, J(H,H) = 7.2 Hz, CH2A), 1.99 (2H, quintet,
J(H,H) = 7.8 Hz, CH2F), 1.68 (2H, quintet, CH2C),
1.65 (2H, quintet, CH2B).
2.3. The pH equilibrium of 1-(40-aminobutyl)-pyrrolinium
dihydrochloride investigated by H NMR
1
A totalof 19 proton NMR spectra (400 MHz) were re-
corded at the following pH: 1.14, 1.43, 1.35, 2.75, 3.78,
6.56, 7.22, 7.75, 7.95, 8.40, 8.86, 9.05, 9.50, 10.16,
10.52, 11.49, 11.51, 12.25 and 12.70 (bold-faced values
indicate spectra recorded using 3.2 mg instead of
6.8 mg substrate per sample). The spectra proved inde-
pendent of concentration. Scheme 2 shows the structure
of aminoaldehyde 1 in different ranges of pH based on
assignment of the spectra. The entropically disfavoured
cyclisation to a 10-membered ring, and the possibility of
imines arising from bimolecular reactions have been
omitted in Scheme 2.
Hydrated 3-(x-aminobutylamino)-propionaldehyde, or 3-
(x-aminobutylamino)-propane-1,1-dioldihydrochloride,
NH3 CH2ACH2 CH2 CH2DNH2 CH2 CH2 CHG(OH)2
as present in almost pure state after addition of three
portions of enzyme exhibited the following NMR
signals: 3.05 (2H, t, J(H,H) = 7.3 Hz, CH2E), 2.98 (2H,
t, J(H,H) = 7.5 Hz, CH2D), 2.93 (2H, t, J(H,H) = 7.2
Hz, CH2A), 1.88 (2H, dt, J(H,H) = 4.8 Hz and 6.6 Hz,
CH2F), 1.67 (2H, quintet, CH2C), 1.65 (2H, quintet,
CH2B).
+
B
C
+
E
F
The signalarising from –CH G(OH)2 was not observed
because it is assumed to be almost coincident with the
water signal. However, the presence of the –CHO group
could be excluded from the absence of signals in the rel-
evant region and the observation of a coupling constant,
J(H,H), for the CH2F group of 4.8 Hz, which is well be-
yond the upper limit (3 Hz) for couplings to the –CHO
group. It was confirmed by spin decoupling that the sig-
At acidic pH (1.14–6.56) the NMR spectra displayed all
the characteristic signals of 1a, which had previously
been assigned (1H, 13C NMR, COSY, HETCOR).
At pH 7.22–7.95 the eight individualpeaks arising from
1a broaden and loses the coupling information although
they still retain the overall appearance of spectra re-
corded at more acidic pHÕs.
nal at 1.88 ppm collapsed to
J(H,H) = 4.8 Hz on irradiation at the triplet at
3.05 ppm.
a
doublet with
In the region from pH 7.95 to 10.5 it proved difficult to
obtain a clear-cut assignment of the spectrum due to
lack of characteristic signals and broad overlapping sig-
nals. Deprotonation of 1a yields very little 1b evidenced
by the lack of an imine signal around 7.7 ppm, although
in the range pH 7.22–7.95 a broad singlet at 6.9 ppm can
be observed. This suggests that upon deprotonation 1b
is immediately transformed to one or more of the species
The spectra are shown in Figure 2A. The upper spec-
trum was recorded of the spermidine solution prior to
addition of enzyme and is characterised by three multi-
plet patterns. DQF-COSY spectra (Fig. 2B) demon-
strated the presence of four partly overlapping triplets

3786
G. Houen et al. / Bioorg. Med. Chem. 13 (2005) 3783–3796
Figure 2. Oxidation of spermidine. A: Proton NMR spectra (400 MHz) of: Spermidine, pH 7.2 (top) and enzymatically oxidised spermidine, pH 7.2
(bottom). B: Proton NMR spectrum and DQF-COSY (800 MHz) of spermidine. C: Proton NMR spectrum and DQF-COSY (800 MHz) of partly
oxidised spermidine.
1c, 1d, 1e and 1f. The aldehyde 1e is not observed at any
pH, as the aldehyde proton around d 9.7 ppm is never
observed.
Spectra recorded in the region of pH 9.05–9.50 reveala
proton appearing at d 5.08 ppm from the hydrated alde-
hyde 1f. A mixture of hemiaminal 1d and hydrated
Scheme 2. The structures of oxidised homospermidine at different pH.

G. Houen et al. / Bioorg. Med. Chem. 13 (2005) 3783–3796
3787
aldehyde 1f is found at pH 9.05 indicated by a broad sig-
nalat 5.24 ppm. When pH is changed from 9.05 to 9.50
the signalat 5.24 ppm sharpens and moves to 5.08 ppm
(characteristic of the hydrated aldehyde), which indi-
cates that the equilibrium between 1d and 1f is shifted
towards 1f. The remaining peaks around pH 7.95–10.5
may be due to protonated 1c, provided that the charac-
teristic triplet from the aminal proton is disguised by the
broad HDO signalat 4.8 ppm.
spectra of synthetic 1-(40-aminobutyl)-pyrrolinium dihy-
drochloride at pH 7.2 and at pH 1. An expansion of the
relevant spectra is shown in Figure 3. The upper spectra
were recorded of the synthetically prepared aminoalde-
hyde (1-(40-aminobutyl)-pyrrolinium dihydrochloride)
at neutralpH and acidic pH, whiel the corresponding
lower spectra were obtained from enzymatic oxidation
of homospermidine. A primary kinetic isotope effect,
kH/kD ꢀ 6.2, was found for homospermidine deuterated
a to the two primary amino groups using the lumines-
cence assay (Table 1). These results demonstrate that
the oxidation of homospermidine is regioselective and
occurs principally at the primary amino groups (Scheme
1).
In the basic region from pH 11.49 to 12.70, the charac-
teristic triplet at 3.99 ppm due to the N–CH–N proton
clearly demonstrates the presence of bicyclic 1c.
Although the rest of the signals are broad, due to the
different conformations for the two fused five- and se-
ven-membered rings of 1c, they correlate with a previous
assignment of 1,6-diazabicyclo[5.3.0]decane 1c based on
1H, 13C NMR, DEPT, COSY and HETCOR spectra.
2.5. Cytotoxic investigations
The cytotoxic effect (IC50 values) for some of the amines
used in Table 1 was determined using human umbilical
vein endothelial cells (HUVEC) (Table 2). It may be
seen from the table that an aminopropyl moiety confers
cytotoxicity. Thus, even though homospermidine is a
relatively good substrate, it does not exert any cytotoxic
effect. It may thus be concluded that H2O2 and NH3 do
not contribute appreciably to cytotoxicity, and hence
that the aminoaldehyde is the major cytotoxic agent.
Similar results were obtained with HeLa and EaHY
926 cells (results not shown). Decane-1,10-diamine,
dodecane-1,12-diamine and octylamine also exhibited
cytotoxic effects. These, however, were not influenced
by aminoguanidine, showing that cytotoxicity was not
due to enzymatic oxidation, but rather due to an inher-
ent cytotoxicity of these compounds.
2.4. Enzymatic reaction with homospermidine
Before addition of BSAO, the only observable peaks in
the 1H NMR spectrum (400 MHz) arose from homospe-
rmidine NH2+(CH2ACH2BCH2BCH2ANH3+)2. dH 2.94–
3.02 (8H, m, CH2A), 1.66–1.76 (8H, m, CH2B). After
addition of 25 lL BSAO solution, the reaction subsided
after 7–24 h yielding 10% turnover and no further oxida-
tion could be detected over the course of 48 h. If the oxi-
dation was carried out at 37 ꢁC, the signal-to-noise ratio
of the resulting spectrum was much lower suggesting
either inactivation of the enzyme or decomposition of
the products. Using 50 lL BSAO resulted in 15% turn-
over after 24 h. Adding 50 lL BSAO every 7 h (four
times in total), and leaving the sample for a total of
48 h, resulted in a maximum of 40% turnover. Since this
is sufficient to characterise the oxidation product ade-
quately, no further attempts were made to oxidise the
remaining homospermidine. The aminoaldehyde could
be unambiguously identified by comparison with the
The prerequisite of an aminopropylgroup for cytotoxi-
city suggested that it was mediated by acrolein. However,
acrolein was not observed in the NMR spectra of the
reaction mixture from oxidation of spermidine, not even
when oxidised spermidine was spontaneously converted
Figure 3. Oxidation of homospermidine. Proton NMR spectra (400 MHz) of: (a) synthetically prepared Ôhomospermidine aldehydeÕ (1-(40-
aminobutyl)-pyrrolinium dihydrochloride), pH 7.2 and 1.0 and (b) enzymatically oxidised homospermidine, pH 7.2 and 1.0.

3788
G. Houen et al. / Bioorg. Med. Chem. 13 (2005) 3783–3796
Table 2. Cytotoxic effect towards human umbilical vein endothelial
cells
channelel ading from the outer surface to the active site.
33–42
Comparisons of crystalstructures
show a partial
IC50 (lM)a
AG effectb
hindrance of the channelin Escherichia coli amine oxi-
dase (ECAO) and a narrower channelin pea seedilng
amine oxidase (PSAO) compared to the more accessible
channels of Arthrobacter globiformis amine oxidase
(AGAO), Hansenula polymorpha amine oxidase
(HPAO) and BSAO, implying that the internal structure
of the channelis different in each protein. Furthermore,
steric and electrostatic differences are observed near the
channel entrance and along the channel enabling sub-
strate differentiation.
Substrate
Spermine
Spermidine
3
14
>1000
>1000
>1000
ND
ND
ND
ND
ND
ND
ND
ND
ND
NE
Norspermidine
Homospermidine
26
>1000
>1000
>1000
>1000
>1000
>1000
>1000
>1000
>1000
380
N-(2-Aminoethyl)-putrescine
Histamine
Propane-1,3-diamine
Butane-1,4-diamine (Putrescine)
Pentane-1,5-diamine (Cadaverine)
Hexane-1,6-diamine
Heptane-1,7-diamine
Octane-1,8-diamine
Decane-1,10-diamine
Dodecane-1,12-diamine
Benzylamine
Our results on the substrate specificity for BSAO (Table
1) are consistent with previous results26,43–49 and con-
firm the existence of a substrate channel with a lipophilic
interior and an attachment point for a secondary amino
group with a spacing of 3–4 CH2 groups to the primary
amino group. Due to the lipophilic interior, an increased
chain length confers increased substrate activity. Both
straight-chain diamines and polyamines lacking the
aminopropylmoiety are poor substrates when the spac-
ing between the primary amino groups is less than eight
atoms. This could indicate an attachment point fitting
the aminooctylmoiety, but the fact that octyalmine is
a better substrate than octane-1,8-diamine shows that
the lipophilic properties of the amine play a key role
in the differentiation between good and poor substrates.
The presence of moderately bulky substituents does not
seem to have a major influence on the substrate activi-
ties, suggesting a flexible substrate channel able to
accommodate a wide substrate variety. In particular,
the high activity of BSAO towards N-(4-aminobut-
yl)hexahydropyrimidine renders this substrate the best
in this investigation. The presence of a secondary amino
group containing an aminopropylmoiety aslo enhances
substrate activity due to the attachment point for the po-
sitive charge. For example, when comparing the sub-
strate activities of norspermidine and (N-(2-amino-
ethyl)-putrescine), the former is by far the better
substrate. Although both have a spacing of seven atoms
between the two primary amino groups, only norspermi-
dine contains the aminopropylgroup. These results indi-
cate the presence of an attachment point fitting the
aminopropylmoiety near the active site. The fact that
the hydrophilic propane-1,3-diamine is a poor substrate,
even though it contains the aminopropylmoiety, can be
explained from the general tendency in lipophilicity. The
lack of a secondary amino group in the d position also
leads to reduced substrate activity (e.g., heptane-1,7-di-
amine vs norspermidine and octane-1,8-diamine vs sper-
midine). In conclusion, among straight chain amines
both an aminopropylgroup and two primary amino
groups separated by seven or eight carbon or nitrogen
atoms were required for optimal substrate ability, while
for the lipophilic bulky hexahydropyrimidines a tetra-
methylene spacing between the primary and tertiary
amino groups was superior to a trimethylene spacing.
Di Paolo et al.43 studied the kinetics of BSAO oxidation
of spermine, spermidine, N8-Ac-spermidine, 1,8-diami-
nooctane, nonylamine, butylamine, 1,3-diaminopropane
and benzylamine, and concluded that the enzyme had
a dualspecificity reflecting a substrate channelwith a
50
>1000
250
NE
ND
NE
Octylamine
a Cytotoxic effect (IC50) of selected amines added to human umbilical
vein endothelial cells (HUVEC) in the presence of bovine serum
amine oxidase.
b AG, aminoguanidine; ND, not determined; NE, no effect.
to putrescine. This result presumably reflects that acro-
lein reacts promptly with amino groups present in the
reaction mixture, that is, NH3 and protein lysine side
chains. To test this hypothesis we mixed acrolein
(1 mM) with the same amount of enzyme as used for
the enzymatic oxidation reactions. Controlspectra re-
corded for 1 mM acrolein showed the signals character-
istic for acrolein, but no signals were observed when
acrolein had been mixed with enzyme (result not
shown). This showed that all acrolein was transformed
to products of a concentration too low to be observed
in the NMR spectra.
3. Discussion and conclusion
3.1. The substrate specificity of BSAO
Copper-containing amine oxidases (CAOs) have been
the subject of extensive biochemical, spectroscopic and
kinetic studies for many years.8,30–32 Crystalstructure
determinations of bacterial, yeast and plant CAOs have
revealed these as mushroom-shaped homodimers with a
close similarity between the active site structures of the
enzymes.33–41 The active site is buried inside the enzymes
and the substrate access is provided by a channelleading
from the outer surface. Negatively charged groups pri-
marily surround the entrance to the channel, providing
excellent conditions for trapping protonated amines.
The channel becomes increasingly lipophilic as the ac-
tive site is approached, favouring the unprotonated
amines necessary for oxidation.
CAOs generally oxidise a wide variety of amines, but the
relative rates and stereochemistries vary significantly
from enzyme to enzyme.8,30–32 These differences in sub-
strate selectivity are not consistent with the close resem-
blance of the active sites, and Wilce et al.38 suggest that
the differences in catalytic activities to be caused by the

G. Houen et al. / Bioorg. Med. Chem. 13 (2005) 3783–3796
3789
lipophilic interior and attachment points for positive
charges corresponding to the geometry of the natural
polyamines and this conclusion is supported by the X-
tures from BSAO oxidation of spermidine and spermine,
has an extreme cytotoxicity55,56,64–67 and has been
shown to induce gene products promoting cell sur-
vival.68 The fast reaction of acrolein with NH3 and other
amines explains the appearance of 3-aminopropanal as
previously reported26 and explains why no acrolein is
observed since reaction with protein side chains leads
to a heterogeneous mixture. The extreme cytotoxicity
of acrolein also indicates why intracellular oxidation
of polyamines takes place by acetylation and oxidation
by polyamine oxidase (PAO). Apparently, it is impera-
tive to prevent b-elimination and generation of acrolein.
As an illustration of this, it has been shown that micro-
injection of CAOs causes cell death by a process resem-
bling apoptosis.69,70 Extracellularly, polyamines are
present only at low concentrations. Therefore, CAOs
presumably do not oxidise polyamines to any apprecia-
ble extent except in cases of cell death by apoptosis or
necrosis, where the intracellular polyamines are released.
42
ray crystalstructure of the enzyme.
Apparently, the
geometry of the substrate channel does allow bulky
amines to reach the active site (e.g., benzylamine) and
also allows some secondary amines to react with the ac-
tive site TPQ.50,51
3.2. Enzymatic reaction with spermidine and homospe-
rmidine monitored by NMR spectroscopy
The NMR method for identifying the major products of
enzymatic oxidation is a time consuming but rewarding
method. It is currently the only method of following the
oxidation process with detection of unstable intermedi-
ates. The results show that NMR spectroscopy in 10%
D2O allows the monitoring of the major products hour
by hour, at physiological concentrations, using the
jump-return sequence or water presaturation. Analysing
the spectra using COSY and spin decoupling identified
the products arising from enzymatic oxidation of sper-
midine. Moreover, indication may appear in the spectra
for the formation and structure of minor products not
apparent after secondary transformations. The prob-
lems encountered using the above method, however,
are considerable. In order to characterise the spectra
of the aminoaldehydes adequately it may be necessary
to prepare authentic samples. Since the spectra are re-
corded of extremely low concentrations in aqueous solu-
tion, they often lack information, for example, coupling
constants. Signals close to the water signal are not ob-
servable, due to the inherent nature of the experiment.52
It has proved important to use H2O containing 10%
D2O in order not to deuterate exchangeable protons.
For identification purposes it has furthermore proved
usefulto compare the product from enzymatic oxidation
at a different pH even though the oxidation with the en-
zyme occurs at physiological pH. Thus, for homosper-
midine a much more informative spectrum is obtained
by adjusting the pH of the solution to the acidic range
and comparison with a synthetically prepared reference
compound proved necessary.
4. Experimental
4.1. General
All reagents and solvents were obtained from commer-
cialsuppielrs and used without further purification.
Hydroxylammonium chloride (99%) was from Riedel-
de Hae¨n. 4-Chlorobutyraldehyde diethylacetal is 98%
pure from ABCR, Germany. n-Octylamine, 1 M Bor-
ane–tetrahydrofurane complex solution in THF,
ꢁ10 M Borane–methylsulfide complex solution and
LiAlD4 (98% Atom D) were obtained from Aldrich.
Putrescine, spermidine, spermine, norspermidine, benz-
ylamine, histamine, pentane-1,5-diamine, hexane-1,6-
diamine, heptane-1,7-diamine, octane-1,8-diamine,
dodecane-1,12-diamine, bis–tris, horseradish peroxidase
(HRP), alkaline phosphatase-conjugated goat anti-
mouse IgG, diethanolamine, p-nitrophenylphosphate
substrate tablets (p-NPP), a-Me-mannopyranoside,
amino guanidine and Triton X-100 were purchased
from Sigma (St. Louis, MO, USA). Decane-1,10-di-
amine was from Fluka (Buchs, Germany). Propane-
1,3-diamine, MgCl2, 37% HCl, NaCl, Na2HPO4,
NaH2PO4 and Tween 20 were from Merck (Darmstadt,
Germany). EAH Sepharose, Con A Sepharose, Q
Sepharose Fast Flow and ECL reagents were from
Amersham Pharmacia (Buckinghamshire, UK). Mouse
anti-human ds DNA, TTN buffer (0.05 M tris buffer,
pH 7.5, 0.3 M NaCl, 1% v/v Tween 20) and purified
BSAO were from Statens Serum Institute (Copenha-
gen, Denmark). Recombinant human basic fibroblast
growth factor (bFGF) and recombinant human vascu-
lar endothelial growth factor (VEGF) were from R &
D Systems (Minneapolis, MN, USA). Human umbili-
cal vein endothelial cells (HUVEC) and the HUVEC
media-kit EGM-2 Bullet kit were obtained from Clo-
netics, BioWhittaker (Walkersville, MD, USA). Poly-
ethylene terephthalate (PET) plates were from Wallac
(Turku, Finland).
3.3. Cytotoxicity of substrates for BSAO
An aminopropylgroup was a prerequisite for cytotoxic-
ity. Thus, although homospermidine was almost as good
a substrate as spermine only the latter exhibited cytotoxi-
city. These results are in agreement with previous
results obtained by others9–27,53 and show that acrolein,
and not the aminoaldehydes, is the main cytotoxic prod-
uct. The lack of cytotoxicity of putrescine and homospe-
rmidine could in principle be due to cyclisation of the
corresponding aminoaldehydes. The oxidation products
of putrescine and homospermidine can both cyclise to
five-membered rings at physiological pH,54 whereas
spermidine oxidised at N1 does not cyclise due to the
fact that the resulting four-membered ring would be
too strained. The relative contribution of the various
reaction products (NH3, H2O2, aminoaldehydes and
acrolein) has been the subject of several studies.55–65
Acrolein, which has been isolated from reaction mix-
The 1H NMR spectra were recorded at 25 ꢁC on a
Varian XL-400 FT NMR spectrometer operating at

3790
G. Houen et al. / Bioorg. Med. Chem. 13 (2005) 3783–3796
400 MHz for H and 100 MHz for 13C , respectively, or
on a Varian UNITY INOVA 800 spectrometer operat-
ing at 799.6 MHz for 1H. The DQF-COSY spectra were
obtained using the standard Varian pulse program tndq-
cosy. 400 MHz NMR experiments and synthesis of
novelcompounds were conducted at the University of
Copenhagen. 800 MHz NMR spectra were obtained at
the Danish Instrument Center for NMR Spectroscopy
of Biological Macromolecules. The spectra are reported
in d units (ppm) relative to tetramethylsilane while the
coupling constants, J(H,H), are given in Hz. The largest
solvent peak is used as reference, unless otherwise sta-
ted. Mass spectra were recorded on a JEOL JMS-HX/
HX 110A Tandem Mass spectrometer. Samples for
GC–MS were injected into a HP 5890 Gas chromato-
graph connected to a HP 5972 mass spectrometer. IR
spectra were obtained on a Perkin–Elmer 1760X FT-
IR spectrophotometer from KBr pellets or from a film.
Elemental analyses were obtained on a Flash EA 1112
series elemental analyser. Purification of BSAO, lumi-
nescence and cytotoxic assays were performed at Statens
Serum Institut. A Wallac 1450 Microbeta Trilux (Turku,
Finland) was used as luminescence counter in the chemi-
luminescence assays.
for the acidic range demonstrating that the equilibrium
is reversible.
1
4.3. Enzymatic reaction with spermidine investigated by
1H NMR spectroscopy
The oxidation of a terminal–CH 2NH2 group in spermi-
dine with formation of the corresponding aminoalde-
hyde was followed by 1H NMR spectroscopy
(400 MHz). In order to mimic the physiological condi-
tions as closely as possible very dilute solutions were
used. In a NMR tube, an aqueous solution (54 lL) of
spermidine trihydrochloride (10 mM, 0.078 mg) was
mixed with a 10 mM sodium phosphate buffer (486 lL,
pH 7.2), and D2O was added (60 lL) to give a totalof
600 lL solution containing 10% D2O. After recording
the spectrum BSAO (75 lL, 1.65 mg/mL, dialysed
against a 5 mM phosphate buffer at pH 8.0 to remove
bis–tris) was added. The oxidation proceeded rapidly
initially but then subsided (30% conversion after 2 h,
40% after 5 h). Addition of two additionalportions of
enzyme with an interval of 2 h resulted in virtually com-
plete oxidation of spermidine as evidenced by the lack of
spermidine signals.
4.2. The pH equilibrium of 1-(40-aminobutyl)-pyrrolinium
dihydrochloride investigated by H NMR spectroscopy
(Scheme 2)
The oxidation was also monitored on a Varian Unity In-
ova 800 instrument using identicalconditions except
that spectra were recorded at 799.6 MHz at 25 ꢁC in
5 mm tubes.
1
The samples (0.6 mL) were prepared in NMR tubes
(4 mm) using the following procedure, which ensures
uniformity. A stock solution was prepared by dissolv-
ing 1-(40-aminobutyl)-pyrrolinium dihydrochloride
(113.5 mg, 0.532 mmol) in D2O (9.7 mL) and adding
DSS (2,2-dimethyl-2-silapentane-5-sulfonate, 0.5 mM)
in D2O (50 mM, 0.1 mL). One-half of the sample
(4.9 mL) is transferred to a beaker fitted with the pH
electrode and DCl is added (20%, 0.1 mL) to give a
total volume of 5.0 mL. Next, the pH (close to 1) is
measured accurately, 0.6 mL is placed in a NMR tube,
The 1H NMR spectra indicated almost complete
hydration of the initially formed aminoaldehyde in
the dilute aqueous solution allowing for further
NMR spectroscopic investigations necessary to analyse
the spectra (spin decoupling, COSY). However, the
small amount of free aminoaldehyde present in the
equilibrium after 24 h gives rise to almost complete
conversion to putrescine by elimination of acrolein.
Signals which could be attributed to acrolein or its
dimerisation or hydration products were not observed
at any time although the oxidation was followed clo-
sely by recording an NMR spectrum for each hour.
From the observation of a large number of small peaks
scattered around in the finalspectrum we concul de that
acrolein is converted to a large number of products
each with intensity too small to allow identification.
Since addition of 1 equiv of acrolein to a solution of
enzyme did not change the activity, reaction of acrolein
with the active site can be excluded. Therefore acrolein
probably is removed by reaction with, for example, free
amino groups of the enzyme.
1
and the H NMR spectrum was recorded (400 MHz).
D2O (0.1 mL) is added to the remaining stock solution
to a totalvoul me of 5.0 mL. Transferring 0.6 mL of the
stock solution to the beaker leaves the concentration
unhanged and restores the volume to 5.0 mL. The pH
can now be adjusted to a higher value with 12 M
NaOD using capillary tubes without significantly
changing the volume, and thus the concentration.
1
Again, 0.6 mL is placed in a NMR tube, and the H
NMR spectrum was recorded. By repeating this proce-
1
dure, H NMR spectra covering the pH range 1–12 can
1
be recorded. H NMR spectra of the samples were re-
corded within 3 h after preparation, using 40 scans per
spectrum. DSS is used as internalstandard and the
largest peak arising from DSS is referenced at
0.00 ppm. DSS is preferred to the readily available sol-
vent peak HDO because of the narrow line width and
1H NMR spectra were recorded in H2O containing 10%
D2O with chemicalshifts being referenced to the sovl ent
peak (d = 4.7). The very strong water signalwas sup-
pressed using the jump-return procedure. In the jump-
return sequence two 90ꢁ pulses are applied directly on
the water signalwith a dealy time of 200 ls between
the pulses. Alternatively a water presaturation for 1.5 s
was used for the data at 800 MHz. The assignments of
all spectra were confirmed by COSY spectra performed
in the phase-sensitive mode combined with spin decou-
pling in order to resolve overlapping signals.
71
chemicalshift insensitivity to pH in the range 2–11.
pH values are calculated by adding 0.40 units72 to
the pH meter reading, obtained from a standard pH-
meter calibrated to read pH in aqueous solutions.
One of the alkaline samples was titrated to an acidic
pH. The spectrum proved identicalto that recorded

G. Houen et al. / Bioorg. Med. Chem. 13 (2005) 3783–3796
3791
4.4. Enzymatic reaction with homospermidine investigated
by H NMR spectroscopy
tions were chromatographed on a Q Sepharose column,
which was eluted with a stepwise gradient of NaCl
(0.1 M, 0.2 M, 0.3 M, 0.4 M, 0.5 M, 1 M) in 5 mM
Na2HPO4 (pH 8.0). Enzyme activity was determined in
fractions as above and fractions containing enzyme were
pooled and dialysed against 5 mM Na2HPO4 (pH 8.0).
A280 was measured to be 1.65 and the finalspecific activ-
ity using 1 mM benzylamine as substrate was found to
be 0.96 lmolmin mg ꢂ1 (20 ꢁC, 20 mM bis–tris, pH 7.0).
1
In contrast to spermidine, the reaction of BSAO with
homospermidine gives rise to an aminoaldehyde unable
to undergo b-elimination (retro-Michael addition). The
reaction was followed by 1H NMR using the same
jump-return sequence described previously. In a 4 mm
NMR tube, an aqueous solution of 10 mM homosper-
midine trihydrochloride (75 lL) was mixed with
10 mM sodium phosphate buffer (pH 7.2, 415 lL) and
D2O (60 lL). Finally, purified BSAO (50 lL, 1.65 mg/
mL) was added to yield a total volume of 600 lL.
Furthermore, BSAO was added in portions of 50 lL
at intervals of 7 h, amounting to a final volume
of 750 lL. After 48 h, the reaction subsided (as moni-
tored via 1H NMR) resulting in approximately 40% con-
version. Since substantialamounts of homospermidine
were still present, direct assignment of the resulting spec-
trum proved difficult. However, the aminoaldehyde
could be unambiguously identified by comparison with
the spectra of synthetic 1-(40-aminobutyl)-pyrrolinium
dihydrochloride both at pH 7.2 and by acidification to
pH = 1.
4.7. Cell culture
4.7.1. Human umbilical vein endothelial cell propagation.
Human umbilical vein endothelial cells (HUVEC) were
maintained in culture at 37 ꢁC, 5% CO2, in an EGM-2
Bullet kit composed of endothelial cell basal medium-2
(EBM-2 medium) supplemented with ascorbic acid, foe-
talbovine serum (FBS), hydrocortisone, human basic
fibroblast growth factor (hbFGF), human vascular
endothelial growth factor (hVEGF), human epidermal
growth factor (hEGF), long R insulin-like growth
factor-1 (R3-IGF-1), gentamicin sulfate (GA-1000)
and heparin as described by the manufacturer
(BioWhittaker).
4.5. Enzyme activity assays
When HUVEC were used for an experiment, the cells
were seeded in 96-microwell polystyrene plates from
Nunc (Kamstrup, Denmark) in a reduced medium
called TFSM + 10%. This is composed of 500 mL
EBM-2 medium supplemented with 0.55 mL ascorbic
acid, 2.2% FBS, 0.55 mL gentamicin sulfate, 0.55 mL
heparin (all from the HUVEC medium kit), 1.1 ng/mL
hbFGF, and 11 ng/mL hVEGF.
The amine was dissolved in 20 mM bis–tris (pH 7.0) to a
concentration of 10 mM. HRP was dissolved in 20 mM
bis–tris (10 mg/mL) and BSAO was diluted to 0.1 mg/
mL in 20 mM bis–tris. HRP was diluted in ECL-2 solu-
tion to furnish a finalconcentration of 10 lg/mL and
amines were added untilthe desired concentration was
reached (0.1 mM). Reactions were then started by the
addition of 10 lL BSAO (finalconcentration 0.01 mg/
mL) to the substrate- and HRP-containing ECL-2 solu-
tion (90 lL). All measurements were carried out in 96-
well PET plates, which were read immediately in the
luminescence reader.
4.7.2. Cytotoxicity assays. In each well of a 96-microwell
plate, 1 · 103 HUVEC were seeded in 90 lL
TFSM + 10% with or without the presence of
0.11 mM amino guanidine (TFSM + 10% and amino
guanidine had been incubated over night at 4 ꢁC).
Immediately after seeding the cells each well was added
10 lL vehicle or diamine/polyamine resulting in a final
concentration ranging from 1 to 1000 lM. After three
days of culture at 37 ꢁC and 5% CO2 the medium was
removed from the 96-microwell plate, and the cells were
washed with phosphate-buffered saline (PBS) (8 mM
Na2HPO4, 2 mM NaH2PO4, 2H2O, 0.15 M NaCl,
pH 7.3) for 1 min. All washings in this assay were car-
ried out with 200 lL/well at room temperature on an
orbitalshaker. Incubations were performed with
100 lL/well. The cells were fixed with ice-cold 96% eth-
anol for 2 h at room temperature. This was followed by
three washings with TTN buffer (50 mM tris, pH 7.5, 1%
Tween 20, 0.3 M NaCl) and permeabilisation for 2 h at
room temperature with 1% Triton X-100. The wells were
washed three times with TTN buffer and blocked for 1 h
with 200 lL TTN buffer. Cells were then incubated
either overnight at 4 ꢁC or for 1 h at room temperature
with 100 lL mouse anti-human ds DNA diluted 1:2000
in TTN buffer. Subsequently, the cells were washed three
times with TTN buffer, followed by 1 h of incubation
with alkaline phosphatase-conjugated goat anti-mouse
IgG diluted 1:1000 in TTN buffer. After being washed
three times with TTN buffer the amount of detected
4.6. Purification of bovine serum amine oxidase
BSAO was purified from bovine serum essentially as de-
scribed previously.73 Bovine serum (915 mL) was stirred
overnight at 5 ꢁC with ammonium sulfate (209 g/L). The
precipitate was collected by centrifugation (20 min,
16.300 g) and the supernatant stirred overnight at 5 ꢁC
with ammonium sulphate (129 g/L). The precipitate
was collected by centrifugation as above, dissolved in
10 mM Na2HPO4 (170 mL, pH 8.0) and dialysed three
times against 5 mM Na2HPO4 (pH 8.0). The dialysed
solution was clarified by centrifugation and chromato-
graphed on an EAH column (100 mL) eluted with step-
wise increasing concentrations of Na2HPO4 (pH 8.0,
0.1 M, 0.2 M, 0.3 M, 0.4 M). The enzyme-containing
fraction was located by activity assays as described
above using spermine as substrate. Pooled fractions
were chromatographed on a 100 mL con A Sepharose
column, which was equilibrated and washed with
5 mM Na2HPO4 (pH 8.0). Enzyme was eluted by
0.3 M a-Me-mannopyranoside. Enzyme-containing
fractions were located as above, pooled and dialysed
against 5 mM Na2HPO4 (pH 8.0). The dialysed frac-

3792
G. Houen et al. / Bioorg. Med. Chem. 13 (2005) 3783–3796
antigen was quantitated using 1 mg/mL p-nitro phenyl
phosphate in alkaline substrate buffer (1 M diethanol-
amine, 0.5 mM MgCl2, pH 9.8). The absorbance was read
on a Versamax plate reader at 405 nm with background
subtraction at 690 nm. After being washed three times
with 200 lL Milli-Q water the cells were inspected in
an Olympus CK2 microscope.
16.27%); dH (DMSO-d6) 1.35 (4H, m, CH2), 1.77 (6H,
s, CH3), 3.0 (4H, m, CH2NH), 7.81 (2H, s, NH); dC
(DMSO-d6) 22.6, 26.7, 38.2, 168.9 (CO); m/z (FAB)
173 ([M+H]+).
4.8.4. 1,11-Diacetylhomospermidine. Essentially as previ-
ously reported,75 1-acetamido-4-aminobutane (5.5 g,
0.042 mol) was refluxed in toluene with dehydrated Ra-
ney-Ni (3 g) for 6.5 h to give the pure product in quan-
titative yield. Mp 74–75 ꢁC (Et2O–MeCN 95:5); (Found:
C, 58.95; H, 10.39; N, 17.02. C12H25N3O2 requires
C, 59.23; H, 10.36; N, 17.27%); mmax (KBr)/cmꢂ1
3424vs br (NH), 1635vs (CO), 1567s (NH bend); dH
(D2O) 1.47 (8H, s), 1.92 (6H, s, CH3CO), 2.60 (4H, s),
3.12 (4H, s); dC (D2O) 22.5, 26.6, 38.8, 38.9, 48.7,
170.6 (CO); m/z (FAB, glycerol) 244 ([M+H]+).
4.8. Synthesis
4.8.1. N-Benzyl-5-aza-nonanedinitrile. The dinitrile was
prepared essentially by the method previously de-
scribed.74 However, substantialamounts (20–30%,
GC–MS) of N-benzyl-2-pyrrolidone were formed by
cyclisation and hydrolysis of the initially formed N-benz-
yl-4-aminobutanenitrile. This by-product was removed
by taking the filtered reaction mixture to dryness, dis-
solving it in hydrochloric acid, and washing repeatedly
with ether. The acidic mixture was then made basic with
the addition of sodium carbonate and extracted with
ether (3 · 100 mL). Evaporation of the solvent left the
crude dinitrile still containing minor impurities (TLC).
The crude product was purified by vacuum liquid chro-
matography (VLC) using heptane–ethylacetate as the
eluent to give the pure dinitrile (50%) as a colourless
oil(Found: C, 74.51; H, 8.00; N, 17.35. C 15H19N3 re-
quires C, 74.65; H, 7.94; N, 17.41%); mmax (film)/cmꢂ1
2245s (CN), 1602m, 1564w and 1494s; dH (CDCl3)
1.76 (8H, q, J(H,H) = 6.8 Hz, CH2), 2.32 (4H, t,
J(H,H) = 7.0 Hz, CH2), 2.52 (4H, t, J(H,H) = 6.8 Hz,
CH2), 3.51 (2H, s, CH2Ph), 7.21–7.34 (5H, m, Ph); dC
(CDCl3) 14.3, 22.8, 51.8, 58.0 (CH2Ph), 119.3 (CN),
126.8, 128.0, 128.3, 138.2; m/z (EI) 241 (M+, 1%), 187
(33), 91 (100). Catalytic hydrogenation at atmospheric
pressure and room temperature with PtO2 in glacial ace-
tic acid has been reported72 to furnish homospermidine
triacetate. However, complete conversion of both the
benzyland nitreil groups proved impossibel without
simultaneous formation of a considerable amount of
polymeric material. This method was thus abandoned
for the synthesis of larger quantities of homospermidine
in favour of the following approach. However, the puri-
fied N-benzyl-5-aza-nonanedinitrile was still a conve-
nient precursor for preparation of the deuterium
labelled homospermidine.
4.8.5. N-(4-Aminobutyl)-pyrrolidine. This compound was
prepared using the two-step procedure previously re-
ported.27 The crude slightly yellow oil was purified by
distillation under reduced pressure (120–140 ꢁC/ 5–10
mbar) to yield a colourless oil. Spectroscopic data (1H,
13C NMR, MS) proved identicalto the previousyl
reported.27
4.8.6. N-(4-Aminobutyl)hexahydropyrimidine. This com-
pound was prepared from spermidine essentially as de-
scribed76 except that formaldehyde was used in 10%
excess to ensure complete conversion of spermidine. Dis-
tillation of the crude product (66–68 ꢁC, 0.1 mbar) left a
colourless oil (47%). However, after allowing it to stand
for some time, colourless crystals (mp 42–44 ꢁC) were se-
cured. The product was very hygroscopic and analysed
as C8H19N3 with varying amounts of water. mmax
(KBr)/cmꢂ1 3316s, 3268m, 3215s and 3200s (mNH),
1614s (dNH2); Other spectroscopic data (1H, 13C
NMR, MS) proved identicalto the previouysl
reported.76
4.8.7. N-(3-Aminopropyl)hexahydropyrimidine. This
compound was prepared from norspermidine using the
same procedure as described for N-(4-aminobutyl)hexa-
hydropyrimidine. The crude product was distilled in va-
cuo (59 ꢁC/0.2 mbar) to yield a colourless oil (38%),
which crystallised near room temperature, mp 33–
34 ꢁC. The product is extremely hygroscopic and ana-
lysed as C7H17N3 with varying amounts of water. It rap-
idly liquefies when exposed to air but recrystallises when
dried in vacuo over NaOH. mmax (KBr)/cmꢂ1 3200–
3400s (mNH), 1596s (dNH2); dH (CDCl3) 1.61 (2H, quin-
tet, J(H,H) = 5.5 Hz, CH2), 1.63 (2H, quintet,
J(H,H) = 7.1 Hz, CH2), 2.27 (3H, s, NH and NH2),
2.31 (2H, t, J(H,H) = 7.2 Hz, NCH2), 2.57 (2H, br s,
NCH2), 2.77 (2H, t, J(H,H) = 6.8 Hz, NCH2), 2.81
(2H, t, J(H,H) = 5.1 Hz, NCH2), 3.38 (2H, s, NCH2N);
dC (CDCl3), 26.6, 29.9, 40.2, 44.6, 52.6, 52.8, 69.5; m/z
(FAB) 144 (100%, [M+H]+), 142 (64, [M+HꢂH2]+).
4.8.2. 1-Acetamido-4-aminobutane. Following the re-
ported procedure,75 a mixture of 1,4-butanediamine
(23.1 g, 0.262 mol) and ethylacetate (7.4 g, 0.084 mol)
was sealed in a tube and heated to 100 ꢁC for 24 h. Ku-
gelrohr distillation afforded the pure compound (7.2 g,
65%), dH (CDCl3) 1.25–1.40 (6H, m, CH2CH2 and
NH2), 1.79 (3H, s, CH3CO), 2.54 (2H, t, J(H,H) =
6.6 Hz, CH2NH2), 3.03 (2H, q, J(H,H) = 6.7 Hz,
CH2NH), 6.96 (1H, s, NH); dC (CDCl3) 22.6, 26.5,
30.4, 38.9, 41.3, 169.9 (CO); m/z (EI) 130 (M+, 3%),
101 (90), 73 (88), 43 (100).
4.8.3. 1,4-Diacetamidobutane. The remanent from the
distillation of 1-acetamido-4-aminobutane crystallised
to give almost pure diacetylated product (2.0 g, 28%).
Mp 137–138.5 ꢁC (EtOAc); (Found: C, 55.49; H, 9.76;
N, 16.27. C8H16N2O2 requires C, 55.79; H, 9.36; N,
4.8.8. N,N0-Di-(methyl-d3)-1,8-octanediamine dihydro-
chloride. This compound proved readily available from
1,8-octane-diamine using LiAlD4 for reduction using
the two-step procedure described by Skinner et al.77
for the parent compound. Recrystallisation from EtOH

G. Houen et al. / Bioorg. Med. Chem. 13 (2005) 3783–3796
3793
furnished white crystals, mp 240–241 ꢁC; (Found: C,
47.64; H, 10.57; N, 11.12. C10H20D6Cl2N2 requires
C, 47.80; ÔHÕ, 10.43; N, 11.15%); mmax (KBr)/cmꢂ1
2582m, 2363m (mCD3), 1601m (dNH2); dH (DMSO-d6)
1.28 (8 H, m, CH2), 1.59 (4H, quintet, J(H,H) = 7.0 Hz,
CH2), 2.81 (4H, t, J(H,H) = 8.0 Hz, NCH2), 8.92 (4H, s,
NH2); dC (DMSO-d6) 25.0, 25.2, 27.6, 31.5 (N-CD3),
48.8 (N-CH2); m/z (FAB) 179 (100%, [M+H]+), 215
(52, [M+H+HCl]+).
vent was evaporated to yield the product as a colourless
oil(0.78 g, 28%), which was used without further purifi-
cation in the subsequent step. m/z (FAB) 254.27
([M+H]+); m/z (EI) 193 (3%), 179 (6), 162 (16), 91 (100).
4.8.13. N,N-Bis(4-amino-4,4-dideuteriobutyl)amine trihy-
drochloride. N,N-Bis(4-amino-4,4-dideuteriobutyl)benz-
ylamine (0.60 g, 2.37 mmol) dissolved in ethanol
(30 mL) was hydrogenated for 22 h at atmospheric pres-
sure using palladium hydroxide on carbon (0.36 g, 20%
pH) as the catalyst. The catalyst was removed by filtra-
tion, and the clear ethanolic solution was treated with
gaseous HCl to give the trihydrochloride as white col-
ourless crystals. The trihydrochloride was collected on
a sintered glass funnel, washed repeatedly with dry
ether, and dried in an exsiccator over KOH. The pure
trihydrochloride of the deuterated triamine (0.32 g,
50%) was obtained as fine white crystals, mp 279–
281 ꢁC (dec. on slow melting); (Found: C, 35.15; ÔHÕ,
9.05; N, 15.28; Cl, 38.73. C8H20D4Cl3N3 requires C,
35.24; ÔHÕ, 8.87; N, 15.41; Cl, 39.00%); mmax (KBr)/
cmꢂ1 3426, 2948 (CH2), 2020 (CD2), 1763, 1688 and
1489 (CH2); dH (D2O) 1.73 (8 H, m, CH2CH2), 3.05 (4
H, t, CH2N); dC (D2O) 22.6, 23.6, 38.2(t), 46.8; m/z
(FAB, glycerol + 5% TFA) 164.17 (100%, [M(d4)+
H]+), 163 (7, [M(d3H)+H]+), 162 (3, [M(d2H2)+H]+),
161 (<1, [M(d1H3)+H]+), 160 (<1, [M(H4)+H]+). Ap-
prox. 96% deuterium incorporation is found (FAB).
4.8.9. Homospermidine. In contrast to previous sugges-
tions,75 the diacetylcompound had to be heated under
reflux with 6 M HCl for 9 h to ensure complete hydroly-
sis. After evaporation of the solvent the crude hydro-
chloride (4.1 g) was thoroughly mixed with finely
powdered KOH (3 g) and distilled in a Kugelrohr appa-
ratus to give pure homospermidine as a colourless oil.
dH (DMSO-d6) 1.3–1.4 (8H, m, CH2CH2), 2.42 (4H, t,
J(H,H) = 6.9 Hz, CH2N), 2.48 (4 H, t, J(H,H) = 6.6 Hz,
CH2N); dC (DMSO-d6) 27.3, 31.5, 41.9, 49.7; m/z (FAB,
glycerol) 160 ([M+H]+).
4.8.10. Homospermidine dihydrochloride. An ethereal
solution of homospermidine was treated with gaseous
HClto give the product as white crystasl. Mp 230–
233 ꢁC (Et2O); (Found: C, 41.16; H, 10.08; N, 17.95;
Cl, 30.09. C8H23Cl2N3 requires C, 41.38; H, 9.98; N,
18.10; Cl, 30.54%); mmax (KBr)/cmꢂ1 2800–3000vs, br
(NH), 1631s, 1597s and 1557s (NH3 and NH2); dH
(D2O) 1.52 (8H, m, CH2CH2), 2.74 (4H, t,
J(H,H) = 7.3 Hz, CH2N), 2.77 (4H, t, J(H,H) = 7.1 Hz,
CH2N); dC (D2O) 23.9, 25.6, 39.2, 47.2; m/z (FAB, glyc-
erol) 160 ([M+H]+).
4.8.14. 5,10-Diazadecanal diethyl acetal. A solution of 4-
chlorobutyraldehyde diethyl acetal (5.0 g, 0.028 mol)
was added to 1,4-diaminobutane (25 mL, 0.25 mol),
which was stirred at 85 ꢁC overnight. After cooling to
rt, the mixture was slowly added to a stirred cold solu-
tion of 2 M NaOH (50 mL, 0.1 mol) and extracted with
ether (3 · 25 mL). From the combined organic extracts
the solvent is evaporated. The remanent is redissolved
in cold 2 M NaOH (50 mL) and extracted with ether
(3 · 25 mL). The combined organic extracts are dried
over anhydrous K2CO3 followed by evaporation of the
solvent. The crude oil is distilled in vacuo (115 ꢁC/
0.5 mbar) to yield the aminoacetal (1.36 g, 21%) as a
hygroscopic colourless oil dH (CDCl3) 1.02 (3H, br s,
NH2+NH), 1.05 (6H, t, J(H,H) = 7.0 Hz, CH3), 1.28–
1.53 (8H, m), 2.46 (2H, t, J(H,H) = 7.0 Hz), 2.47 (2H,
t, J(H,H) = 7.1 Hz), 2.55 (2H, t, J(H,H) = 7.0 Hz),
3.34 and 3.49 (4H, dq, J(H,H) = 9.5 Hz and 7.1 Hz,
CH2), 4.34 (1H, t, J(H,H) = 5.5 Hz, H-1); dC (CDCl3)
15.0 (CH3), 25.0, 27.2, 31.1, 31.3, 41.8 (C-9), 49.4 and
49.5 (C-6 and C-4), 60.6 (CH2), 102.4 (C-1); m/z
4.8.11. Homospermidine trihydrochloride. When an eth-
anolic solution of homospermidine is treated with gas-
eous HCl, a white precipitate of homospermidine
trihydrochloride is formed. Mp 278–280 ꢁC (EtOH,
Lit.78 283–285 ꢁC, dec. on slow melting); (Found: C,
35.74; H, 9.13; N, 15.48; Cl, 39.60. C8H24Cl3N3 requires
C, 35.77; H, 9.00; N, 15.64; Cl, 39.59%); mmax (KBr)/
cmꢂ1 2700–3100vs, br (NH), 1610s, 1598s, and 1516s
+
(NH3 and NH2+); dH (D2O) 1.73 (8H, m, CH2CH2),
3.01 (4H, t, J(H,H) = 7.0 Hz, CH2N), 3.07 (4H, t,
J(H,H) = 7.6 Hz, CH2N); dC (D2O) 22.7, 23.9, 38.8,
46.8; m/z (FAB, glycerol) 160 ([M+H]+).
4.8.12. N,N-Bis(4-amino-4,4-dideuteriobutyl)benzylamine.
Following the reported procedure,74 a solution of AlCl3
(3.2 g, 24.0 mmol) in dry ether (75 mL) was added to a
suspension of LiAlD4 (1.03 g, 24.6 mmol) in dry ether
(225 mL). The mixture was stirred under a nitrogen
atmosphere for 15 min, followed by dropwise addition
of N-Benzyl-5-aza-nonanedinitrile (2.69 g, 11.2 mmol)
dissolved in dry ether (50 mL). Stirring was continued
for 24 h at room temperature. After cooling to 0 ꢁC
the reaction mixture was quenched by dropwise addition
of H2O (10 mL) followed by 1 M HCl (150 mL). Stirring
was continued until a clear slightly yellow solution was
obtained. Slow addition of solid NaOH (6 g) gave a pre-
cipitate, which was separated from the mother liquor by
filtration and washed repeatedly with ether. The com-
bined ether extracts were dried over Na2SO4 and the sol-
(FAB, 3-nitrobenzylacloho)l 233.12 ([M+H]
+); m/z
(EI) 203 (22%, MꢂC2H5), 141 (11), 84 (100), 72 (70),
70 (64), 43 (47), 30 (77).
4.8.15. 1-(40-Aminobutyl)-pyrrolinium dihydrochloride.
5,10-Diazadecanaldiethylacetal(0.625 g, 2.93 mmo)l
was hydrolysed in water (20 mL) with 2 M HCl added
(3.25 mL, 6.5 mmol) for 25 min at rt. Lyophilisation
for two days furnished hygroscopic white amorphous
crystals, which were used without further purification.
dH (D2O) 1.74 (2H, m, H-30), 1.93 (2H, m, H-20), 2.35
(2H, quintet, J(H,H) = 8.1 Hz, H-4), 3.06 (2H, br t,
J(H,H) = 7.7 Hz, H-40), 3.22 (2H, br t, J(H,H) =

3794
G. Houen et al. / Bioorg. Med. Chem. 13 (2005) 3783–3796
7.0 Hz, H-3), 3.98 (2H, br t, J(H,H) = 7.5 Hz, H-10),
4.21 (2H, tq, J(H,H) = 8.1 Hz and 1.8 Hz, H-5), 8.73
(1H, m, J(H,H) = 1.5 Hz, H-2); dC (D2O) 19.1 (C-4),
23.2 and 23.5 (C-20 and C-30), 35.6 (C-3), 38.6 (C-40),
53.0 (C-10), 58.7 (C-5), 181.3 (C-2); m/z (FAB, 3-nitro-
4.8.19. N-(Boc-glycyl)-N0-Boc-putrescine. Boc-putrescine
(5.65 g, 30 mmol) dissolved in dichloro-methane
(300 mL) was added slowly to a stirred mixture of
Boc-glycine (5.35 g, 30 mmol) and dicyclohexylcarbodi-
imide (6.80 g, 33 mmol) cooled to 0 ꢁC. The stirred mix-
ture was kept at 0 ꢁC for 1 h and then at ambient
temperature for 24 h, whereupon the precipitated di-
cyclohexylurea was filtered off and washed with dichloro-
methane. Washing with saturated NaHCO3 (30 mL),
cold 2 M NaHSO4 (30 mL) and water (2 · 30 mL) en-
sured removalof unreacted Boc-glycine and Boc-putres-
cine. The organic phase was dried over MgSO4 and the
solvent evaporated. Minor impurities of dicyclohexylu-
rea were removed by dissolving the residue in ether fol-
lowed by filtration and evaporation. Recrystallisation
from ether afforded the desired compound as colourless
crystals (8.16 g, 79%), mp 79–80 ꢁC. (Found: C, 55.67;
H, 9.21; N, 12.08. C16H31N3O5 requires C, 55.63; H,
9.05; N, 12.16%); mmax (KBr)/cmꢂ1 3349vs and 3267vs
(mNH), 3053m, 2978s, 2940m, 2871w, 1718vs (mCO),
1695vs (mCO), 1657vs (mCO), 1555 (dNH), 1381s,
1364m, 1294vs, 1251vs, 1181vs, 1134s; dH (CDCl3)
1.37 (18H, s, CH3), 1.44 (4H, m, 2 · CH2), 3.04 (2H, t,
J(H,H) = 6.1 Hz, CH2–N), 3.25 (2H, q, J(H,H) =
5.3 Hz, CH2–N), 3.71 (2H, s, N–CH2–CON), 4.72 (1H,
s, NH), 5.40 (1H, s, NH), 6.57 (1H, s, NH); dC (CDCl3)
26.4, 27.3, 28.2, 39.0, 40.0, 44.2, 79.1, 80.0, 156.0, 169.4;
m/z (FAB) 346 ([M+H]+), 691 ([(2 · M)+H]+).
benzylacloho)l 141.11 ([C
1,6-diazabicyclo[5.3.0]decane in gas phase.
8H16N2+H]+). Analyses as
4.8.16. 1,6-Diazabicyclo[5.3.0]decane. 2 M HCl(5.92
mL, 11.84 mmol) was added to a solution of 5,10-di-
azadecanaldiethylacetal(1.25 g, 5.38 mmo)l in water
(25 mL). The solution was stirred for 25 min at rt, fol-
lowed by addition of 1 M K2CO3 (25 mL). Extraction
with CH2Cl2 (3 · 25 mL), followed by drying over
K2CO3, and evaporation of the solvent afforded the
bicyclic aminal (0.47 g, 62%) as a colourless oil. dH
(CDCl3, 0 ꢁC) 1.40–1.50 (1H, m), 1.60–1.87 (6H, m),
1.94–2.30 (1H, vbr s, NH), 2.12–2.22 (1H, m), 2.44–
2.54 (2H, m), 2.73–2.81 (1H, m), 2.88–3.02 (3H, m),
3.61 (1H, t, J(H,H) = 7.3 Hz, H-1); dC (75 MHz,
CDCl3) 22.0, 26.7, 30.4, 33.9, 47.3, 53.7, 54.6, 78.9; dH
(D2O, 2 ꢁC) 1.10 and 1.83 (2H, br s, H-2), 1.28 and
1.40 (6H, br s, H-3, H-7, H-8), 2.25 and 2.42 (4H, br
s, H-4, H-6), 2.25 and 2.59 (2H, br s, H-9), 3.37 (1H,
br t, J(H,H) = 6.4 Hz, H-1); dC (D2O, 2 ꢁC) 21.5 and
25.4 and 28.8 (C-3, C-7, C-8), 32.4 (C-2), 46.5 (C-9),
51.9 and 53.4 (C-4, C-6) 77.4 (C-1); dC (75 MHz,
CD3OD) 23.4, 27.7, 31.0, 34.5, 48.7, 54.3, 55.7, 80.1;
m/z (EI) 140 (M+, 65%), 139 (59, M–H), 83 (100).
4.8.20. N-(2-Aminoacetyl)-putrescine dihydrochloride. N-
(Boc-glycyl)-N0-Boc-putrescine (4.14 g, 12 mmol) was
dissolved in dichloromethane (40 mL), cooled to 0 ꢁC
and treated with TFA (9.18 mL, 0.12 mol). The reaction
mixture was stirred at room temperature for 30 min and
then evaporated to dryness leaving a slightly pink oil. In
order to remove excess TFA, toluene (50 mL) was added
and subsequently evaporated at reduced pressure. This
procedure was repeated three times. The resulting oil
was dissolved in HCl-saturated methanol followed by
slow addition of HCl-saturated diethyl ether causing
precipitation of the hydrochloride to occur. The white
precipitate was filtered off, washed with ether and dried
at reduced pressure over NaOH pellets. Recrystallisa-
tion from methanol gave white needle-like crystals
(1.91 g, 73%), mp 183–184 ꢁC. (Found C, 33.08; H,
7.89; N, 19.06. C6H17Cl2N3O requires C, 33.04; H,
7.86; N, 19.26%); mmax (KBr)/cmꢂ1 3500–2200vs
(mNH3+), 3252vs (mNH), 1969m, 1679vs (mCO); dH
(DMSO-d6) 1.46 (2H, m, CH2), 1.59 (2H, m, CH2),
2.74 (2H, m, CH2), 3.10 (2H, m, CH2), 3.51 (2H, s,
N–CH2–CO), 8.21 (3H, s, NH3+), 8.29 (3H, s, NH3+),
8.70 (1H, t, J(H,H) = 5.7 Hz, NH); dC (DMSO-d6)
24.3, 25.8, 38.0, 38.3, 40.1, 165.8; m/z (FAB) 146
([M+H]+).
4.8.17. Boc-glycine. Boc-glycine was prepared from gly-
cine and di-tert-butyldicarbonate essentially by the pro-
cedure described,79 scaled to 1:10. Recrystallisation
yielded pure white crystals (89%), mp 87–88 ꢁC. The
presence of both the E- and Z-isomers of the amide is
observed (1H and 13C NMR). (Found: C, 47.91;
H, 7.55; N, 8.08. C7H13NO4 requires C, 47.99; H,
7.48; N, 8.00%); mmax (KBr)/cmꢂ1 3500–2400s (mCOOH),
3407m, 3343s (dNH), 1749vs (mCOOH), 1668vs
(mCONH), 1541vs (dNH), 1411s, 1369m, 1215vs,
1198vs, 1058m, 959m, 860m; dH (CDCl3) 1.43 (9H, s,
CH3), 3.86 and 3.93 (2H, s, CH2), 5.23 and 6.65 (1H,
s, NH), 10.05 (1H, s, COOH); dC (CDCl3) 28.1, 42.1
and 43.2, 80.3 and 81.7, 155.9 and 157.1, 173.8 and
174.4; m/z (EI) 160 (2%, MꢂCH3), 130 (11, MꢂCOOH),
120 (60), 116 (4, MꢂCH2COOH), 102 (5,
MꢂOC(CH3)3), 76 (21), 59 (90, Mꢂ(CH3)3COCONH),
57 (100, MꢂO2CNCH2COOH), 41 (40), 29 (24).
4.8.18. Boc-putrescine. Using the previously described
procedure80 Boc-putrescine was prepared from putres-
cine and di-tert-butyldicarbonate (83%). The colourless
oilwas shown to be pure by GC–MS but too hygroscopic
for elemental analysis. mmax (KBr)/cmꢂ1 3363vs, 2934vs,
2175w, 1688vs, 1525vs, 1365s, 1173vs, 1042m, 993m,
868m; dH (CDCl3) 1.3–1.5 (13H, m, CH3+2 · CH2),
2.07 (2H, s, NH2), 2.64 (2H, t, J(H,H) = 6.6 Hz, N–
CH2), 3.03 (2H, t, J(H,H) = 6.6 Hz, N–CH2), 4.88 (1H,
s, NH); dC (CDCl3) 27.3, 28.2, 30.1, 40.2, 41.3, 78.7,
155.9; m/z (EI) 188 (0.1%, M+), 143 (2), 132 (28), 115
(22, MꢂOC(CH3)3), 114 (19), 103 (19), 98 (17), 74 (33),
70 (64), 57 (79, (CH3)3C), 59 (58), 41 (45), 30 (100,
CH2@NH2).
4.8.21. N-(2-Aminoethyl)-putrescine trihydrochloride. A
suspension of N-(2-aminoacetyl)-putrescine dihydro-
chloride (1.45 g, 6.6 mmol) in dry THF (135 mL) was
added slowly to a stirred solution of borane–THF com-
plex (1.0 M, 75 mL, 75 mmol) under nitrogen. The reac-
tion mixture was refluxed for 48 h and cooled to room
temperature, whereupon it was quenched by slow addi-
tion of methanol (180 mL) followed by reflux overnight.

G. Houen et al. / Bioorg. Med. Chem. 13 (2005) 3783–3796
3795
After filtration and evaporation of the solvents the crude
products were dissolved in methanol (100 mL), cooled
to 0 ꢁC and HCl-saturated methanol (150 mL) was
added. Evaporation of methanoland excess hydrogen
chloride left a mixture of white crystals and yellowish
oil. It was found that the starting materials had only
partially reacted (1H NMR), and the crude product re-
quired recrystallisation from methanol in three turns
to provide pure N-(2-aminoethyl)-putrescine trihydro-
chloride as white crystals (0.51 g, 32%), mp 217–
218 ꢁC. (Found C, 30.02; H, 8.34; N, 17.14. C6H20Cl3N3
requires C, 29.95; H, 8.38; N, 17.46%); mmax (KBr)/cmꢂ1
tion of 2.3 M KOH (60 mL, 140 mmol) and refluxed
for an additional12 h. The organic sovlents were re-
moved by evaporation and the remaining water phase
was extracted with ether (4 · 40 mL). The combined ex-
tracts were dried over MgSO4 followed by evaporation
of the solvent. The crude product was mixed with finely
powdered KOH and distilled using Kugelrohr to obtain
the pure product as a colourless oil (1.07 g, 58%). The
compound was too hygroscopic for elemental analysis.
mmax (KBr)/cmꢂ1 3325s and 3260s (mNH2), 2955s,
2925s, 2854s, 1592s (dNH2), 1461s, 1373s, 1152m,
1075m (mCN), 819m br, 724m; dH (DMSO-d6) 2.71
(2H, sextet, J(H,H) = 6.3 Hz, CH), 1.59 (4H, br s,
NH2), 1.17–1.27 (12H, m, CH2), 0.93 (6H, d,
J(H,H) = 6.3 Hz, CH3); dC (DMSO-d6) 46.3, 39.9,
29.4, 25.9, 24,1; m/z (EI) 171 (<1%, MꢂH), 157 (3,
MꢂCH3), 145 (2, MꢂHCN), 129 (6, C8H19N), 112 (4),
70 (9), 56 (12), 44 (100, C2H6N), 30 (6, CH2@NH2).
+
3500–2400vs (mNH3 and mNH2+), 2049m, 1865w,
1600s, 1521s, 1466s, 1396m, 1345m, 1028s, 879m,
815m; dH (D2O) 1.75 (4 H, m, 2 · CH2), 3.01 (2H, t,
J(H,H) = 7.0 Hz, CH2–N), 3.15 (2H, t, J(H,H) =
7.0 Hz, CH2–N), 3.39 (4H, m, 2 · CH2–N); dC (D2O)
22.5, 23.6, 35.2, 38.6, 43.9, 47.2; m/z (FAB) 132
([M+H]+).
4.8.22. Decane-2,9-dione dioxime. Pure decane-2,9-dione
was prepared in two steps using the procedure
previously described81 and used as starting material.
Decane-2,9-dione (5.45 g, 32 mmol) and hydroxylam-
monium-chloride (7.02 g, 0.1 mol) were dissolved in
60% ethanol(50 mL) and cooeld to 0 ꢁC. Potassium
carbonate (13.8 g, 0.1 mol) was added to the stirred
mixture that was allowed to return to rt, followed by
reflux for 3 h. Addition of water (100 mL) followed by
cooling to 0 ꢁC led to precipitation of the crude product,
which was filtered off and washed with water.
The remaining product was extracted with ether
(3 · 50 mL). The ether phase was dried over MgSO4, fol-
lowed by evaporation of the solvent. The extracted and
the precipitated crude products were combined and
recrystallised from water–ethanol (60:40) to yield pure
white crystals (3.95 g), mp 129–133 ꢁC. Additionalpure
product (1.35 g, mp 124–131 ꢁC) could be secured by
evaporation of the mother liquor and recrystallisation
of the remanent. The difference in melting points of
the first and second crop was ascribed to the presence
of more than one isomer of the dioxime. (totalyiedl,
5.30 g, 83%). (Found: C, 59.78; H, 10.10; N, 13.96.
C10H20N2O2 requires C, 59.97; H, 10.07; N, 13.99%);
mmax (KBr)/cmꢂ1 3234s br (mOH), 2926vs, 2883s,
2852s, 2824m (mCH), 1679m (mC@N), 1462s (dCH2),
1411s (dOH), 1368s (dCH3), 951vs (mN–O), 729s
(qCH2); dH (DMSO-d6) 1.24 (4H, m, CH2), 1.42
(4H, m, CH2), 2.08 (4H, t, J(H,H) = 7.2 Hz, CH2–
CNOH), 10.18 (2H, s, OH); dC (DMSO-d6) 13.0, 25.8,
28.4, 35.0, 155.1; m/z (EI) 200 (1%, M+), 183 (14, Mꢂ
OH), 168 (39), 167 (46), 150 (9), 128 (100,
MꢂC3H6NO), 100 (10), 86 (42), 73 (95, MꢂC7H13NO),
55 (36), 42 (50), 41 (51).
Acknowledgements
The spectra at 800 MHz were obtained on the Varian
Unity Inova spectrometer of the Danish Instrument
Center for NMR Spectroscopy of Biological
Macromolecules.
References and notes
1. Thomas, T.; Thomas, T. J. Cell Mol. Life. Sci. 2001, 58,
244–258.
2. Seiler, N. Progr. Brain Res. 1995, 106, 333–344.
3. Seiler, N. Neurochem. Res. 2000, 25, 471–490.
4. Buffoni, F.; Ignesti, G. Mol. Gen. Met. 2000, 71, 559–564.
5. Lyles, G. A. Int. J. Biochem. Cell Biol. 1996, 28, 259–274.
6. Sessa, A.; Perin, A. Agents Actions 1994, 43, 69–77.
7. Mondovi, B.; Riccio, P.; Agostinelli, E. Adv. Exp. Med.
Biol. 1988, 250, 147–161.
8. Houen, G. APMIS 1999, 107(Suppl. 96), 1–46.
9. Hirsch, J. G. J. Exp. Med. 1953, 97, 327–343.
10. Alarcon, R. A.; Foley, G. E.; Modest, E. J. Arch. Biochem.
Biophys. 1961, 94, 540–541.
11. Israel, M.; Rosenfield, J. S.; Modest, E. J. J. Med. Chem.
1964, 7, 710–716.
12. Bachrach, U.; Abzug, S.; Bekierkunst, A. Biochim.
Biophys. Acta 1967, 134, 174–181.
13. Higgins, M. L.; Tillman, M. C.; Rupp, J. P.; Leach, F. R.
J. Cell Physiol. 1969, 74, 149–154.
14. Katsuta, H.; Takaoka, T.; Nose, K.; Nagai, Y. Jpn. J.
Exp. Med. 1975, 45, 345–354.
15. Allen, J. C.; Smith, C. J.; Curry, M. C.; Gaugas, J. M.
Nature 1977, 267, 623–625.
16. Byrd, W. J.; Jacobs, D. M.; Amoss, M. S. Nature 1977,
267, 621–623.
17. Gahl, W. A.; Pitot, H. C. Chem. -Biol. Interact. 1978, 22,
91–98.
18. Gaugas, J. M.; Dewey, D. L. Br. J. Cancer 1978, 39, 548–
557.
4.8.23. Decane-2,9-diamine. Decane-2,9-dione dioxime
(2.00 g, 10 mmol) dissolved in THF (80 mL) was slowly
added to an ice-cold solution of ꢁ10 M boran–dimethyl-
sulfide complex (12 mL, 120 mmol) under nitrogen. The
mixture was heated to rt and refluxed for 24 h, where-
upon it was cooled to 0 ꢁC and quenched with ethanol
(120 mL). The THF–ethanolsoul tion was then refluxed
for 6 h, followed by slow addition of an aqueous solu-
19. Allen, J. C.; Smith, C. J.; Hussain, J. I.; Thomas, J. M.;
Gaugas, J. M. Eur. J. Immunol. 1979, 102, 153–158.
20. Gaugas, J. M. In Polyamines in Biomedical Research;
Gaugas, J. M., Ed.; Wiley and Sons: New York, 1980; pp
343–362.
21. Swanson, T. L.; Gibbs, G. E. In Vitro 1980, 16, 761–
766.

3796
G. Houen et al. / Bioorg. Med. Chem. 13 (2005) 3783–3796
22. Gilad, G. M.; Gilad, V. H. Int. J. Devl. Neurosci. 1986, 4,
195–208.
50. Lee, Y.; Huang, H.; Sayre, L. M. J. Am. Chem. Soc. 1996,
118, 7241–7242.
23. Nishida, Y.; Miyamoto, T. Toxicol. Lett. 1986, 31, 203–
209.
24. Morgan, D. M. Biochem. J. 1987, 242, 347–352.
25. Tabor, C. W.; Tabor, H.; Bachrach, U. J. Biol. Chem.
1964, 239, 2194–2203.
26. Houen, G.; Bock, K.; Jensen, A. J. Acta Chem. Scand.
1994, 48, 52–60.
27. Lee, Y.; Sayre, L. M. J. Biol. Chem. 1998, 273, 19490–
19494.
28. Coleman, A. A.; Hindsgaul, A. A.; Palcic, M. M. J. Biol.
Chem. 1989, 264, 19500–19505.
29. Nakamura, N.; Moe¨nne-Loccoz, P.; Tanizawa, K.; Mure,
M.; Suzuki, S.; Klinman, J. P.; Sanders-Loehr, J. Bio-
chemistry 1997, 36, 11479–11486.
51. Lee, Y.; Ling, K. Q.; Lu, X.; Silverman, R. B.; Shepard, E.
M.; Dooley, D. M.; Sayre, L. M. J. Am. Chem. Soc. 2002,
124, 12135–12143.
52. Cavanagh, J.; Fairbrother, W.; Palmer, A. G., III;
Skelton, N. J. Protein NMR Spectroscopy; Academic:
San Diego, CA, 1996.
53. Israel, M.; Zoll, E. C.; Muhammad, N.; Modest, E. J. J.
Med. Chem. 1973, 16, 1–5.
54. Struve, C.; Christophersen, C. Heterocycles 2003, 60(8),
1907–1914.
55. Alarcon, R. A. Arch. Biochem. Biophys. 1964, 106, 240–
242.
56. Alarcon, R. A. Arch. Biochem. Biophys. 1970, 137, 365–
372.
30. Klinman, J. P.; Mu, D. Annu. Rev. Biochem. 1994, 63,
299–344.
57. Hussain, J. I.; Smith, C. J.; Allen, J. C. Cell Tissue Kinet
1983, 16, 583–591.
31. Mondovi, B. Structure and Function of Amine Oxidases;
CRC: Boca Raton, USA, 1985.
58. Hegre, O. D.; Marshall, S.; Hickey, G. E. In Vitro 1984,
20, 198–204.
32. Kapeller-Adler, R. Amine Oxidases and Methods for Their
Study; Wiley Interscience: New York, USA, 1970.
33. Parsons, M. R.; Convery, M. A.; Wilmot, C. M.; Yadav,
K. D. S.; Blakeley, V.; Corner, A. S.; Phillips, S. E. V.;
McPherson, M. J.; Knowles, P. F. Structure 1995, 3, 1171–
1184.
34. Wilmot, C. M.; Murray, J. M.; Alton, G.; Parsons, M. R.;
Convery, M. A.; Blakeley, M. A.; Corner, A. S.; Palcic, M.
M.; Knowles, P. F.; McPherson, M. J.; Phillips, S. E.
Biochemistry 1997, 36, 1608–1620.
35. Murray, J. M.; Saysell, C. G.; Wilmot, C. M.; Tambyra-
jah, W. S.; Jaeger, J.; Knowles, P. F.; Phillips, S. E. V.;
Mcpherson, M. J. Biochemistry 1999, 38, 8217–8227.
36. Murray, J. M.; Kurtis, C. R.; Tambyrajah, W.; Saysell, C.
G.; Wilmot, C. M.; Parsons, M. R.; Phillips, S. E. V.;
Knowles, P. F.; Mcpherson, M. J. Biochemistry 2001, 40,
12808–12818.
59. Henle, K. J.; Moss, A. J.; Nagle, W. A. Cancer Res. 1986,
46, 175–182.
60. Averill-Bates, D. A.; Agostinelli, E.; Przybytkowski, E.;
Mondovi, B. Biochem. Cell Biol. 1994, 72, 36–42.
61. Agostinelli, E.; Przybytkowski, E.; Mondovi, B.; Averill-
Bates, D. A. Biochem. Pharmacol. 1994, 48, 1181–1186.
62. Agostinelli, E.; Przybytkowski, E.; Averill-Bates, D. A.
Free Rad. Biol. Med. 1996, 20, 649–656.
63. Averill-Bates, D. A.; Agostinelli, E.; Przybytkowski, E.;
Mateescu, M. A.; Mondovi, B. Arch. Biochem. Biophys.
1993, 300, 75–79.
64. Bonneau, M.-J.; Poulin, R. Exp. Cell Res. 2000, 259, 23–
34.
65. Shamin, S.; Sakata, K.; Kashiwagi, K.; Ueda, S.; Iwasaki,
S.; Shirahata, A.; Igarashi, K. Biochem. Biophys. Res.
Commun. 2001, 282, 228–235.
66. Izard, C.; Libermann, C. Mutat. Res. 1978, 47, 115–138.
67. Sakata, K.; Kashiwagi, K.; Sharmin, S.; Ueda, S.;
Igarashi, K. Biochem. Soc. Trans. 2003, 31, 371–374.
68. Kwak, M.-K.; Kensler, T. W.; Casero, R. A., Jr. Biochem.
Biophys. Res. Commun. 2003, 305, 662–670.
69. Bachrach, U.; Ash, I.; Abu-Elheiga, L.; Hershkovitz, M.;
Loyter, A. J. Cell Physiol. 1987, 131, 92–98.
70. Bachrach, U.; Ash, I.; Rahamim, E. Tissue Cell 1987, 19,
39–50.
37. Matsuzaki, R.; Tanizawa, K. Biochemistry 1998, 37,
13947–13957.
38. Wilce, M. C. J.; Dooley, D. M.; Freeman, H. C.; Guss, J.
M.; Matsunami, H.; McIntire, W. S.; Ruggiero, C. E.;
Tanizawa, K.; Yamaguchi, H. Biochemistry 1997, 36,
16116–16133.
39. Li, R. B.; Klinman, J. P.; Mathews, F. S. Structure 1998,
6, 293–307.
40. Chen, Z.; Schwartz, B.; Williams, N. K.; Li, R.; Klinman,
J. P.; Mathews, F. S. Biochemistry 2000, 39, 9709–
9717.
41. Kumar, V.; Dooley, D. M.; Freeman, H. C.; Guss, J. M.;
Harvey, I.; McGuirl, M. A.; Wilce, M. C.; Zubak, V. M.
Structure 1996, 4, 943–955.
71. Wishart, D. S.; Sykes, B. D. Methods Enzymol. 1994, 239,
363–392.
72. Glasoe, P. K.; Long, F. A. J. Phys. Chem. 1960, 64, 188–
190.
73. Houen, G.; Jørgensen, J.; Leonardsen, L.; Larsson, L. I.
Acta Chem. Scand. 1993, 47, 902–909.
42. Lunelli, M.; Di Paolo, M. L.; Biadene, M.; Calderone, V.;
Battistutta, R.; Scarpa, M.; Rigo, A.; Zanotti, G. J. Mol.
Biol. 2005, 346, 991–1004.
43. Di Paolo, M. L.; Stevanato, R.; Corazza, A.; Vianello, F.;
Lunelli, L.; Scarpa, M.; Rigo, A. Biochem. J. 2003, 371,
549–556.
74. Bergeron, J.; Burton, P. S.; McGovern, K. A.; Kline, S. J.
Synthesis 1981, 9, 732–733.
75. Lee, Y. B.; Park, M. H.; Folk, J. E. J. Med. Chem. 1995,
38, 3053–3061.
76. Siddiqui, A. O.; Merson-Davies, L.; Cullis, P. M. J. Chem.
Soc., Perkin Trans. 1 1999, 22, 3243–3252.
77. Skinner, W. A.; Cory, M.; Schellenberger, T. E.; DeGraw,
J. I. J. Med. Chem. 1966, 9, 520–522.
44. Israel, M.; Modest, E. J. J. Med. Chem. 1971, 14, 1042–
1047.
45. De Matteis, G.; Agostinelli, E.; Mondovi, B.; Morpurgo,
L. JBIC 1999, 4, 348–353.
78. Kuttan, R.; Radhakrishnan, A. N.; Spande, T.; Witkop,
B. Biochemistry 1971, 10, 361–365.
46. Werle, E.; Roever, F. Biochem Z. 1954, 325, 550–554.
47. Yamada, H.; Yasunobu, K. T. J. Biol. Chem. 1962, 237,
1511–1516.
48. Gahl, W. A.; Pitot, H. C. Biochem. J. 1982, 202, 603–611.
49. Mondovi, B.; Turini, P.; Befani, O.; Sabatini, S. Methods
Enzymol. 1983, 94, 314–318.
79. Keller, O.; Keller, W. E.; Van Lock, G. Org. Synth., Coll.
1990, VII, 70–71.
80. Krapcho, A. P.; Kuell, C. S. Synth. Commun. 1990, 20,
2559–2564.
81. Taherirastgar, F.; Brandsma, L. Chem. Ber./Recueil 1997,
130, 45–48.