New efficient substrates for semicarbazide-sensitive amine oxidase/VAP-1 enzyme: Analysis by SARs and computational docking

Article abstract of DOI:10.1021/jm051076e

Structure activity relationships for semicarbazide-sensitive amine oxidase/vascular adhesion protein-1 (SSAO/VAP-1) were studied using a library of arylalkylamine substrates, with the aim of contributing to the discovery of more efficient SSAO substrates. Experimental data were contrasted with computational docking studies, thereby allowing us to examine the mechanism and substrate-binding affinity of SSAO and thus contribute to the discovery of more efficient SSAO substrates and provide a structural basis for their interactions. We also built a model of the mouse SSAO structure, which provides several structural rationales for interspecies differences in SSAO substrate selectivity and reveals new trends in SSAO substrate recognition. In this context, we identified novel efficient substrates for human SSAO that can be used as a lead for the discovery of antidiabetic agents.

Full text of DOI:10.1021/jm051076e

J. Med. Chem. 2006, 49, 6197-6208
6197
New Efficient Substrates for Semicarbazide-Sensitive Amine Oxidase/VAP-1 Enzyme: Analysis
by SARs and Computational Docking
Francesc Yraola,†,‡ Silvia Garc´ıa-Vicente,†,§ Juan Ferna´ndez-Recio,§ Fernando Albericio,§,| Antonio Zorzano,§,
Luc Marti,*,§ and Miriam Royo*,‡
Combinatorial Chemistry Unit and Institute for Research in Biomedicine, Barcelona Science Park, Josep Samitier 1, E-08028-Barcelona,
Spain, Departament de Qu´ımica Orga`nica, Facultat de Qu´ımica, UniVersitat de Barcelona, E-08028-Barcelona, Spain, and Departament de
Bioqu´ımica i Biologia Molecular, Facultat de Biologia, UniVersitat de Barcelona, E-08028-Barcelona, Spain
ReceiVed October 24, 2005
Structure activity relationships for semicarbazide-sensitive amine oxidase/vascular adhesion protein-1 (SSAO/
VAP-1) were studied using a library of arylalkylamine substrates, with the aim of contributing to the discovery
of more efficient SSAO substrates. Experimental data were contrasted with computational docking studies,
thereby allowing us to examine the mechanism and substrate-binding affinity of SSAO and thus contribute
to the discovery of more efficient SSAO substrates and provide a structural basis for their interactions. We
also built a model of the mouse SSAO structure, which provides several structural rationales for interspecies
differences in SSAO substrate selectivity and reveals new trends in SSAO substrate recognition. In this
context, we identified novel efficient substrates for human SSAO that can be used as a lead for the discovery
of antidiabetic agents.
Introduction
aminoacetone, whose product of oxidation, formaldehyde, has
been strongly implicated in cardiovascular pathologies associated
with diabetes.12
Semicarbazide-sensitive amine oxidase/vascular adhesion
protein-1 (SSAO/VAP-1) is a bifunctional protein with copper-
containing amine oxidase activity (EC 1.4.3.6) that converts
primary amines to aldehydes, with the concomitant production
of hydrogen peroxide and ammonia (Figure 1).1,2
On the basis of these observations, our research focuses on
the identification of potent SSAO/VAP-1 substrates based on
arylalkylamines. Benzylamine or arylalkylamine derivatives are
good candidates, as shown by previous modeling studies of the
catalytic domain of SSAO/VAP-1. We previously found that
the aromatic residues Tyr 384, Phe 389, and Tyr 394 define a
pocket of stable size (7 Å width), which may participate in the
binding of apolar substrates.13 After defining the pharmacophoric
moiety of SSAO substrates based on benzylamine/arylalkyl-
amine derivatives, we designed a strategy to evaluate other
substrate requirements that enhance substrate-binding affinity.
SSAO/VAP-1 is highly expressed in adipocytes where it is
localized mainly in plasma membrane in an insulin-independent
manner.3,4 Substrates of SSAO/VAP-1 exert a variety of insulin-
like effects in human, rat, and mouse adipose cells.4-7 Thus,
substrates such as benzylamine or tyramine stimulate glucose
transport in isolated human adipocytes.7 Furthermore, in isolated
rat adipocytes, the combination of substrates of SSAO with low
ineffective vanadate concentrations produces a potent stimulation
of glucose transport, which is abolished by semicarbazide and
catalase.4,5 This combination also induces insulin-sensitive
glucose transporter isoform 4 (GLUT4) recruitment to the cell
surface,4,5 lipogenesis8 and an inhibition of lipolysis.7 These
observations indicated that the SSAO-dependent generation of
hydrogen peroxide may be responsible for these effects via a
chemical interaction with vanadate, which can form peroxo-
vanadate, a powerful insulin-mimetic agent.9
More recently, in vivo studies have also demonstrated the
antidiabetic properties of the combination of benzylamine or
other arylalkylamines with vanadium in experimental models
of type 1 and type 2 diabetes.9-11 As a beneficial effect of the
administration of exogenous SSAO substrates, beyond their
antidiabetic properties, we could expect a decrease in the
oxidation of potential endogenous substrates, methylamine and
Here we studied SSAO/VAP-1 structure activity relationships
(SARs) and identified several molecular features using benzyl-
amine as a hit. Our study provides key data on the differences
between human and mouse SSAO substrate selectivity and
contributes to the identification of new highly potent substrates.
In addition, we used computational docking and homology-
based modeling to analyze the results and provide a structural
basis for the interaction between the arylalkylamine substrates
and SSAO.
Results and Discussion
Combinatorial Chemistry. A combinatorial approach was
applied to define the structural ligand requirements for new
SSAO substrates. To explore the substitution of the benzylamine
aromatic ring, compounds with electron-withdrawing and
electron-donating groups were evaluated as ligands of the phenyl
residues near the catalytic site of the enzyme. Thus, both
electronic and steric effects were studied. In this context, 48
compounds were tested, combining commercial and synthesized
ones (Tables 1-4 in Supporting Information). After testing the
first set, we synthesized a second generation of compounds using
* Corresponding authors. Phone: +34934037122 (M.R.); +34934034700
(L.M.). Fax: +34934037126 (M.R.); +34934034717 (L.M.). E-mail:
mroyo@pcb.ub.es (M.R.); lmarti@pcb.ub.es (L.M.).
† These authors contributed equally to the study.
‡ Combinatorial Chemistry Unit.
§ Institute for Research in Biomedicine.
| Departament de Qu´ımica Orga`nica.
Departament de Bioqu´ımica i Biologia Molecular.
10.1021/jm051076e CCC: $33.50 © 2006 American Chemical Society
Published on Web 09/22/2006

6198 Journal of Medicinal Chemistry, 2006, Vol. 49, No. 21
Yraola et al.
Figure 1. SSAO turnover is schematically represented. (1) Benzylamine condensation to topaquinone cofactor of the enzyme (TPQ); (2)
tautemerization of Schiff base intermediate (rate-limiting step); (3), (4), and (5) hydrolysis followed H2O2 and ammonia production and aldehyde
byproduct release.
Scheme 1a
compound 12 as a starting point. This second generation, based
on the modification of the amino group of m-xylylenediamine,
was also synthesized by a solid-phase approach (Table 3 in
Supporting Information). The members of this new family were
designed and synthesized, exploring size, hydrophobicity, and
stability by new hydrogen-binding interactions in this extended
position. The function of this extended group was to fill the
a Reagents and conditions: (i) NH2OH, 2 h at room temperature; (ii) H2
(4 bar)/ PtO2‚1H2O, 24 h at room temperature or LiBH4 in THF for 8 h;
(iii) HCl/dioxane.
extra apolar cavity (defined by Leu 469 and Phe 389 residues),
found in previous studies, to enhance binding capacity to the
SSAO active site.13,17 After screening, the compounds that
showed substrate properties were selected, resynthesized, and
purified to confirm the biological data and establish SARs. The
selected compounds were classified in two structural families:
(a) arylalkylamines, especially with para substitution on the
aromatic ring, and (b) m-xylene derivatives.
ment with HCl in dioxane using ethyl acetate or diethyl ether
as precipitating solvents (Scheme 1).14
A second generation of compounds derived from compound
12 were synthesized using the commercial 3-(Boc-amino-
methyl)-benzylamine hydrochloride as starting material. Next,
several standard protocols were applied to obtain the desired
compounds in the reactions represented in Scheme 2. To obtain
compound 12, 3-(Boc-aminomethyl)-benzylamine hydrochloride
was acylated quantitatively with acetic anhydride and base. In
the case of N-(3-aminomethyl-benzyl)-acylamide derivatives
(compounds 31, 32, 34, 38, 39, and 40), 3-(Boc-aminomethyl)-
benzylamine hydrochloride was acylated with the corresponding
carboxylic acids and as coupling mixture reagents HOBt,
DIPCDI, and DIEA. The Boc-protected compounds were
yielded after purification. The allylcarbamate derivative 33 was
obtained by reaction of 3-(Boc-aminomethyl)-benzylamine
hydrochloride with allyl chloroformate, giving quantitative yields
of the corresponding Boc-protected compound 33. Compound
35 was obtained first by acylating 3-(Boc-aminomethyl)-
benzylamine with FmocPhg and as coupling reagents HOBt and
DIPCDI. The Fmoc protecting group was then removed, and
the final compound was achieved by acylation with acetic
anhydride. The sulfonamide derivative 36 was achieved by
reacting 3-(Boc-aminomethyl)-benzylamine hydrochloride with
tosyl chloride, yielding the Boc-protected derivative 36.15 The
hydrochloric salts were obtained quantitatively by treatment with
HCl in dioxane of the Boc derivatives using ethyl acetate or
diethyl ether as solvent.
Synthesis of Selected Compounds. The arylalkylamines
selected were prepared from commercially available aldehydes
by reductive amination. First, the hydroxylamine was condensed
with aldehyde in methanol at reflux for 30 min. In all the cases,
the reaction quantitatively yielded the corresponding aldoximes.
The aldoxime was then catalytically reduced with PtO2‚1H2O
and H2 (4 bar) in acetic acid at room temperature. The
conversion is directly proportioned to the nature of the substitu-
tion. In this regard, the conversions of (4-ethylphenyl)methyl-
amine (5) and (3-fluoro-4-methylphenyl)methylamine (16) were
60 and 53%, respectively. In contrast, the conversion of (4-
butylphenyl)methylamine (11) was 80%. In the case of strong
p-EWG in the aryl moiety, such as (4-fluorophenyl)methylamine
(3) and (4-(trifluoromethyl)phenyl)methylamine (8), the reduc-
tion by PtO2‚1H2O was not successful. Other catalysts were
tested, such as Pd/C or Pd(OH)2, with similar unsuccessful
results. This could be explained by strong electropositive density
of the carbon between the Ph rings of the corresponding
aldoximes. In these cases, the reduction was carried out using
hydride reagent LiBH4 under Ar refluxed in dry THF. Under
these conditions, compounds 3 and 8 were obtained with a
moderated yield but with good purity (over 99% in both cases).
The corresponding hydrochloride salts were obtained by treat-

Substrates for Amine Oxidase/VAP-1 Enzyme
Journal of Medicinal Chemistry, 2006, Vol. 49, No. 21 6199
Scheme 2a
a Reagents and conditions: (i) 3-(Boc-aminomethyl)-benzylamine hydrochoride/RCOOH/HOBt/DIPCDI/DIEA in DMF, 12 h at room temperature; (ii)
HCl/dioxane, 2 h at room temperature; (iii) ClSO2p-Ts/K2CO3 in acetone/water, 4 h at room temperature; (iv) allyl chloroformate/DIEA in DCM for 4 h at
room temperature; (v) piperidine/DMF, 4 h at room temperature; (vi) Ac2O/DIEA in DCM, 4 h at room temperature.
Table 1. Human and Mouse SSAO Activity in the Presence of
Arylalkylamine Derivatives Synthesized as SSAO Substratesa
SSAO activity
(% rel. to benzylamine oxidation)
entry
compound
benzylamine
4-fluorobenzylamine
4-ethylbenzylamine
4-trifluoromethyl-
benzylamine
human
mouse
1
3
5
8
100 ( 7
134 ( 11
14 ( 1
4 ( 1
100 ( 15
127 ( 4
26 ( 3
35 ( 3
11
16
4-butylbenzylamine
3-fluoro-4-methyl-
benzylamine
5 ( 1
88 ( 2
24 ( 3
133 ( 2
41
42
44
46
phenylethylamine
3-phenylpropylamine
4-phenylbutylamine
2-(4-fluorophenyl)-
ethylamine
21 ( 2
44 ( 4
166 ( 11
35 ( 3
19 ( 2
40 ( 5
71 ( 4
18 ( 2
a Human and mouse SSAO activity was determined by detecting the
production of hydrogen peroxide in the presence of distinct compounds
present at 1 mM for human and 100 µM for mouse activity measurements,
as indicated in Materials and Methods. Values are means of percentages
relative to benzylamine-related hydrogen production.
Figure 2. Comparison of mouse and human SSAO active sites. Active
site residues in human SSAO are shown in cyan. Residues near the
human SSAO active site that are nonconserved in mouse SSAO are in
white. The equivalent residues in mouse SSAO are in green. Binding
cavities are shown as dots (human in yellow; mouse in red), which
were computed with Pocket-Finder (www.bioinformatics.leeds.ac.uk/
pocketfinder/). Figure prepared with ICM-Browser from Molsoft
(www.molsoft.com).
because these same residues are conserved in mouse SSAO/
VAP-1, their nature cannot explain the difference in substrate
selectivity observed between the two species. To perform a more
detailed comparison of mouse and human SSAO active sites at
the structural level, a model of mouse SSAO was built based
on the structure of the homologous human SSAO protein (PDB
code 1us1). Superimposition of mouse and human SSAO active
sites showed that the conformation of the catalytic residues is
well-conserved, as expected (Figure 2). However, the active site
in mouse SSAO model was slightly larger than that in human
SSAO X-ray structure. To evaluate whether this was artifactual
(because we were comparing a homology-based model and an
X-ray structure) and also to check the dynamic behavior of the
active site, we performed molecular dynamics simulations on
both structures, as described in the Experimental Section. We
extracted snapshots from the simulations every 10 ps and
computed the solvent-excluded volume of the active site in these
conformations using Pocket, a program that identifies active
Exploration of Human and Mouse Substrate Selectivity
by Docking. Slight changes of the substituents on the phenyl
group of benzylamines produced a considerable decrease in the
catalytic activity of the two enzymes. As expected, more
substrates for mouse than for human SSAO were found (Tables
1-4 in Supporting Information). These results are in agreement
with previous observations that human SSAO/VAP-1 is more
restrictive in substrate selectivity than the mouse enzyme.16 This
observation could be attributed to differences in the distribution
of the mouse and human catalytic site residues. As previously
proposed by Wilce et al., amine oxidase substrate recognition
may be influenced by the amino acid composition of the channel
near the junction of domains D3 and D4 of the protein, which
form the entrance of the substrate from the protein surface to
the active site.17 Because of their critical position within the
catalytic site, Tyr 384 or Leu 468 and Leu 469 residues may
affect the substrate selectivity of the enzyme.9,17-19 However,
sites of proteins on the basis of the Alpha Shape theory.20
A
few outlier conformations for which Pocket artificially detected
the active site as “fused” with other larger cavities, or

6200 Journal of Medicinal Chemistry, 2006, Vol. 49, No. 21
Yraola et al.
Table 2. Human and Mouse SSAO Activity in the Presence of Arylalkylamine Derivatives Synthesized as SSAO Substratesa
a Human and mouse SSAO activity was determined by detecting the production of hydrogen peroxide in the presence of distinct compounds present at
1 mM for human and 100 µM for mouse activity measurements, as indicated in Materials and Methods. Values are means of percentages relative to benzylamine-
related hydrogen production.
“fragmented” in small cavities, were not considered for calcula-
tions. The average volumes of the SSAO active sites during
simulations were similar in mouse (V ) 190.6 Å3 ( 13.4 SD)
and in human (V ) 185.0 Å3 ( 37.9 SD). However, the larger
deviation in volume values for human SSAO indicate that its
active site is more flexible than that of mouse. There were also
interesting differences in the residues near the active site. For
instance, Phe 173 at the entrance of the active site in human
SSAO is Asp 173 in mouse, which leaves more space at the
entrance of the active site and introduces an attraction point for
positively charged substrates. Another residue, Leu 177 in
human SSAO, is Gln 177 in mouse, whose side-chain confor-
mation makes the entrance of the active site wider. All the
considerations described above might explain the differences
in substrate affinity between human and mouse SSAO. Another
potentially interesting difference is that human His 242 is
mutated to Pro 242 in mouse, which creates a small cavity
adjacent to the catalytic site. Interestingly, this proline residue
is also conserved in bovine SSAO (Pro 241). The role of this
cavity is unknown. Some insights will be provided by docking
simulations, as discussed at the end of the section. Finally, in
the amine oxidase family of enzymes, intersubunit interactions
are formed by two long hairpins. One of these, the â-ribbon
arm I, which stretches along the surface of the molecule to the
channel entrance, may also be involved in substrate specificity
because the residues at the end of the arm differ in each
protein.17 For instance, Leu 447 in this loop in human SSAO is
Phe 447 in mouse. This loop differs between human and bovine
SSAO in only two residues (Leu 447, Tyr 448 in human vs
Phe 446, Leu 447 in bovine). Despite this small difference in

Substrates for Amine Oxidase/VAP-1 Enzyme
Journal of Medicinal Chemistry, 2006, Vol. 49, No. 21 6201
Figure 3. Qualitative representation of % mouse (A) and human SSAO activity (B). Values extracted from Tables 1 and 2.
sequence, the overall loop conformation is distorted. In our
model of mouse SSAO, the loop conformation was inherited
mostly from the human structure; nevertheless, there was
significant conformational change with respect to the human
template (Figure 2). Conformational variability of this loop
during molecular dynamics was very similar in human
and in mouse SSAO, ranging from 0.2 to 1.3 Å RMSD in human
and from 0.2 to 1.5 Å RMSD in mouse (only CR atoms
considered). However, given the sensitivity of this loop to
the residue sequence, we cannot rule out that further con-
formational deviation in mouse with respect to human SSAO
affects the active site entrance. To our knowledge, this is the
first study to provide a rational explanation of the basis of
structural differences to support the interspecies differences
in substrate selectivity observed between mouse and human
SSAO.

6202 Journal of Medicinal Chemistry, 2006, Vol. 49, No. 21
Table 3. Kinetic Parameters for Mouse SSAO/VAP-1 Substratesa
Yraola et al.
a
Vmax
kcat
(s-1
kcat/Km
entry
compound
K
m (µM)
(nmol/min/mg protein)
)
(s-1 M-1
)
1
3
benzylamine
4-fluorobenzylamine
15.4 ( 2.2
15.7 ( 0.4
14.8 ( 1.7
29.8 ( 3.9
11.5 ( 1.2
12.8 ( 1.6
34.8 ( 8.2
27.0 ( 5.1
11.3 ( 0.8
88.8 ( 0.8
3.5 ( 0.4
4.4 ( 0.7
4.0 ( 0.6
2.5 ( 0.3
1.8 ( 0.2
2.1 ( 0.2
1.2 ( 0.1
2.6 ( 0.5
3.6( 0.2
1.5 ( 0.2
0.18
0.23
0.21
0.13
0.09
0.11
0.06
0.13
0.19
0.08
11 763
14 553
14 064
4428
8099
8747
1773
4957
16 664
883
16
32
33
34
41
42
44
46
3-fluoro- 4-methylbenzylamine
N-(3-(aminomethyl)benzyl)-2-iodoacetamide
N-allyl 3-(aminomethyl)benzylcarbamate
N-(3-(aminomethyl)benzyl)benzamide
phenylethylamine
3-phenylpropylamine
4-phenylbutylamine
2-(4-fluorophenyl)-ethylamine
a Kinetic parameters of human SSAO activity in the presence of the substrates indicated were determined as explained in Materials and Methods. Values
are the means of four independent measurements.
Table 4. Kinetic Parameters and Docking Energy Values for Human SSAO/VAP-1 Substratesa
Vmax
kcat
k
cat/Km
docking energy
(kcal/mol)
entry
compound
K
m (µM)
(nmol/min/mg protein)
(s-1
)
(s-1 M-1
)
1
3
16
32
benzylamine
4-fluorobenzylamine
3-fluoro-4-methylbenzylamine
N-(3-(aminomethyl)benzyl)-
2-iodoacetamide
224 ( 54
205 ( 18
405 ( 34
525 ( 36
315 ( 17
490 ( 45
398 ( 44
263 ( 17
0.45
0.71
0.57
0.38
2023
3439
1413
720
-68.8
-67.6
-65.3
-46.4
33
N-allyl 3-(aminomethyl)-
benzylcarbamate
569 ( 32
114 ( 9
0.16
289
-30.5
34
41
42
44
46
N-(3-(aminomethyl)benzyl)benzamide
phenylethylamine
3-phenylpropylamine
4-phenylbutylamine
2-(4-fluorophenyl)ethylamine
630 ( 55
3725 ( 1007
1672 ( 192
280 ( 40
112 ( 5
0.16
0.41
0.49
0.82
0.43
256
111
294
2938
228
-26.6
-72.7
-63.8
-69.1
-72.2
288 ( 39
342 ( 42
571 ( 20
300 ( 40
1895 ( 347
a Kinetic parameters of mouse SSAO activity in the presence of the substrates indicated were determined as described in Materials and Methods. Values
are means of four independent measurements.
Qualitative Structure Activity Relationships. All resyn-
thesized compounds were screened again as potential substrates
of mouse and human SSAO. As described above, these selected
compounds can be classified in two structural families: (a)
phenylalkylamines (Table 1) and (b) derivatives of m-xilene
diamines (Table 2).
The phenylalkylamine active compounds present diverse
substitutions in the aryl ring, especially in the para position,
which appears to be crucial for the SSAO activity (Table 1).
The percentage of mouse SSAO activity of the p-substituted
benzyl compounds decreased with the presence of bulky para-
substitution groups (Figure 3A). A similar effect was observed
for human SSAO (Figure 3B), where compound p-F (3) showed
remarkably higher activity than benzylamine (1).
To minimize the possible substitution aryl ring effects in the
interaction with the SSAO active site, distinct compounds with
diverse C chain lengths between the amino group and the phenyl
group were explored. With this purpose, commercial arylalkyl-
amines were tested (Table 1, compounds 41, 42, and 44), among
these, phenylethylamine, and already known SSAO substrate.27
For these compounds, mouse SSAO activity increased as the
chain length between the amine function and the phenyl ring
was enlarged. In this regard, these compounds showed the
following decreasing relation: C n ) 4 (44) > C n ) 3 (42) >
C n ) 2 (41) (Figure 3A). When these compounds were tested
with human SSAO (Figure 3B), similar effects were observed
and compound C n ) 4 (44) showed higher activity than
benzylamine (1).
The biological activity of compounds with modified amino
group of m-xylylenediamine (Table 2) showed that the increase
in volume was well-tolerated by the catalytic site of mouse
SSAO (Figure 3A). For the human enzyme, the addition of
hydrophobic groups also induced the recovery of significant
activity (compounds 32-34, 39, and 40). The affinity prefer-
ences of the hydrophobic group are consistent with the previous
structural studies of the active site, which defined an extra cavity
of highly hydrophobic character.13,17
QSAR Based on Kinetic Studies. The kinetic parameters
for the selected novel mouse and human SSAO substrates were
determined (Tables 3 and 4). The kcat of human and mouse
VAP-1 for benzylamine was relatively low (0.18 and 0.45 s-1
,
respectively). As previously reported, this low catalytic turnover
may support the implication of VAP-1/SSAO in lymphocyte
adhesion.21 High affinity substrates for mouse and human SSAO
were found.
Interestingly, compound 44, with a chain length of four C,
showed a 2-fold Vmax and, similar to fluorinated benzylamine
compounds 3 and 16, a better catalytic efficiency (kcat/Km) than
benzylamine in both species. In the case of 4-phenybutylamine
(44), the aromatic ring may be at a sufficient distance to prevent
π-stacking interactions between the phenyl ring of the aryl-
alkylamine and the SSAO active site.
We observed a significant trend between the kcat for the mouse
and human SSAO (Figure 4). A first interpretation of this
observation is that, beyond differences in substrate selectivity,
mouse and human SSAO have a similar capacity to catalyze
amines. Up to certain point, a good substrate for mouse SSAO
has high probability of also being a good substrate for human
SSAO. Similarly, a bad substrate for mouse SSAO will probably
be a bad substrate for the human enzyme as well (Figure 4).
This is relevant for the design of future screenings. However,
as shown by the intercept of the fitted line, a bad substrate for
mouse SSAO may not be a substrate for human SSAO22 (Tables
1-2).22 It is interesting to note that (1) SSAO activity is sensitive
to meta substitutions, (2) compounds 41 and 46 phenylethyl-
amine compounds are equally poor substrates for mouse and
human SSAO, and (3) the presence of a fluor atom in the para
position and the length of the aliphatic chain greatly influence

Substrates for Amine Oxidase/VAP-1 Enzyme
Journal of Medicinal Chemistry, 2006, Vol. 49, No. 21 6203
Figure 4. Correlation of mouse SSAO/VAP-1 kcat versus human kcat
SSAO/VAP-1 values extracted from Tables 3 and 4, respectively.
the capacity of the compound to be oxidized by both human
and mouse SSAO. Among the compounds, we identified 3, 16,
and 44 as the best substrates in both species.
Qualitative SARs and Docking. To provide a structural basis
to our observations, we performed computational docking
studies using the X-ray structure of human SSAO, which is
available in the PDB with code 1us1. Compounds 1, 3, 16, 32-
34, 41, 42, 44, and 46 were computationally docked onto the
coordinates of human SSAO. Given that Leu469 partially
occluded the active site, we modeled a distinct rotamer of the
leucine side-chain, which resulted in a slightly more open active
site. The conformation of TPQ in the human structure (1us1)
was incompatible with catalysis. Therefore, it was modeled to
the active conformation (so-called “off-Cu”) using the coordi-
nates of SSAO from Escherichia coli of code 1spu as template.23
The overall docking results did not depend heavily on the TPQ
conformation, but the ammonium group of the ligand com-
pounds was consistently closer to the TPQ moiety when the
latter was in the “off-Cu” conformation. Figure 5 shows the
lowest-energy docking conformations obtained for selected
compounds. As expected, the docking orientation obtained for
benzylamine (1) fitted the active site perfectly (Figure 5A), with
the ammonium group close to the catalytic oxygen atom of TPQ.
Regarding the docking results of 4-phenylbutylamine (44), the
lowest-energy docking conformation superimposed nicely onto
the benzylamine (1) docking structure (Figure 5B). The am-
monium group of 44 was located in the same spatial coordinates
as that in benzylamine, and the four Caliphatic chain super-
imposed perfectly onto the aromatic ring of benzylamine. Thus,
the aromatic ring of 44 was closer to the active site entrance,
without major atomic clashes with surrounding protein atoms.
As expected, this interaction mode proposed by the docking
calculations indicates that the aliphatic chain has favorable
interactions with the active site, while the phenyl ring prevents
excessive π-stacking with this site, perhaps favoring the exit
of the aldehyde byproduct. Several of the docking orientations
of benzylamine (1) were found in the adjacent cavity around
His 242; this was not observed for 44. We can hypothesize that
bulkier compounds such as 44 are not trapped inside the human
SSAO active site as easily as benzylamine (1), which would
explain why the Vmax of 44 is larger than that of 1 (Table 4). In
contrast, the larger active site (including the mouth entrance
and this adjacent cavity of unknown role) in mouse SSAO
(Figure 2) might equally trap small-sized and bulkier compounds
(Table 3). This hypothesis is therefore consistent with the
experimental data. Compounds 3, 16, 41, 42, and 46 had
similarly good docking energy values (Table 4), and their
Figure 5. Computational docking results. (A) Best docking solution
(lowest binding energy) calculated by CMIP for compound 1 (stick
representation, color cyan), surrounded by human SSAO active site
residues (represented in CPK). (B). Best docking solutions for
compounds 1 (cyan), 41 (pink), 42 (blue), and 44 (yellow) on the active
site of human SSAO. Figure prepared with ICM-Browser from Molsoft
(www.molsoft.com).
lowest-energy docking solutions were found within the active
site in a similar orientation as in 1 and 44. Interestingly, the
docking structure of phenylethylamine (41) deviated slightly
from compounds 1 and 44, indicating that it does not fit as
perfectly into the active site. This observation is consistent with
the lower affinity found for 41 and for its fluoroderivative
(compound 46). Regarding 42, the atoms of the phenyl ring
were located between the corresponding atoms of compounds
41 and 44, but slightly closer to the latter. The distance between
the center of atomic coordinates of the phenyl ring in compounds
41 and 44 is 3.21 Å, whereas the equivalent distance between
42 and 41 is 2.03 Å, and the one between 42 and 44 is 1.21 Å.
This may explain why the activity found experimentally is below
that of benzylamine (1). In contrast, compounds 33 and 34,
which showed poor experimental kcat values with human SSAO,
did not yield any docking orientation within the active site where
the ammonium group can be at a reasonable distance from TPQ.
In addition, the binding energy of 33 and 34 docking solutions,
located at the entrance of the active site but not within it, were
much worse than those of the “good” substrates. Interestingly,
compound 32, which has an intermediate experimental Kcat

6204 Journal of Medicinal Chemistry, 2006, Vol. 49, No. 21
Yraola et al.
Figure 6. Stimulation of glucose transport in isolated mouse adipocytes. Mouse adipocytes were incubated for 45 min without (bas) or with 100
nM insulin, amines at the indicated concentration (1, benzylamine, 44, 4-phenyl-butylamine) alone or in combination with 1 mM sodium orthovanadate
(Van) and/or 1 mM semicarbazide (SCZ), as indicated. Subsequently, 2-deoxy-D-glucose uptake was measured over a 10-min period. Results are
mean ( SEM of two experiments performed in triplicate.
value, had a few docking orientations within the active site,
with a binding energy between that of the good substrates and
the poor ones. This observation correlates with the experimental
Compound 44 is, to our knowledge, the best substrate
described to date for human SSAO. Interestingly, this compound
is structurally related to the side chain of lysine, an endogenous
amino acid that interacts with the human catalytic site of SSAO/
VAP-1.32 We also show that compound 44 strongly stimulated
glucose transport in isolated mouse adipocytes.
k
cat values of these compounds. This is remarkable considering
that enzymatic activity can be affected by determinants other
than binding affinity of the substrates, and that our model is
limited by the use of a rigid-body approach for the active-site
side-chains. In fact, compounds 1, 41, and 46 had slightly larger
computed binding energy values than the ones expected from
the energy/kcat relation for the similar compounds 3, 42, and
44. Indeed, the side-chain of Leu469 was very close to the
aromatic ring of these compounds, so the conformation of this
residue might affect compounds 1, 41, and 46 in a different
way than to the others. Although this is out of the scope of this
work, we are planning a systematic study in which protein
flexibility will be analyzed.
Activation of Glucose Transport in Isolated Mouse Adi-
pocytes. The incubation of isolated rat or mouse adipocytes in
the presence of SSAO substrates, such as benzylamine and
vanadate, stimulates glucose transport.4,5 Here we analyzed
whether one of our best novel SSAO substrates (44) exerted
such an insulin-mimetic effect. The incubation of isolated mouse
adipocytes with 0.1 or 1 mM of compound 44 plus vanadate
produced a strong stimulation of glucose transport (3.3-fold and
3.4-fold, respectively), similar to that obtained with insulin or
benzylamine plus vanadate (Figure 6). As for benzylamine, this
effect was completely inhibited by semicarbazide. The incuba-
tion of 1 mM of compound 44 without vanadate did not
significantly affect glucose transport (1.1-fold).
Experimental Section
Materials and Methods. DIPCDI was obtained from Fluka
Chemika (Buchs, Switzerland) and HOBt from Albatross Chem,
Inc. (Montreal, Canada). 3-(Boc-aminomethyl)-benzylamine hy-
drochoride was purchased from Neosystem (Strasbourg, France).
Semicarbazide hydrochloride, benzylamine hydrochloride, hydrogen
peroxide, sodium orthovanadate horseradish peroxidase, and other
chemicals were purchased from Sigma Aldrich (St. Louis, MO,
U.S.A.). Purified human VAP-1 was a kind gift from BioTie
Therapeutics (Turku, Finland). Amplex red reagent (10-acetyl-3,7-
dihydroxyphenoxazine) was from Molecular Probes (Eugene, OR,
U.S.A.). 2-Deoxy-D-[2,6-3H]glucose (53 Ci/mmol) was purchased
from Amersham Pharmacia Biotech (Buckinghamshire, U.K.).
Purified porcine insulin was a kind gift from Eli Lilly (Indianapolis,
IN). Collagenase type I was obtained from Worthington (Lakewood,
NJ). Other chemicals were obtained from Aldrich (Milwaukee, WI.,
U.S.A.), with the highest purity grade available. All commercial
1
reagents and solvents were used as received. The H/13C NMR
spectra were recorded on a Varian Mercury 400-MHz spectrometer.
HPLC analysis was performed using an Alliance 2795 Waters
Chromatography system with a reverse-phase column C18 X-Terra
3.5 µm 2.1 × 50 mm column with Waters 996 PDA detector.
HPLC-MS spectra were recorded on a Waters Alliance HT 2795
system with dual λ absorbance detector 2487 and Micromass ZQ
mass spectrometer. The purity of the arylalkylamines synthesized
was determined by HPLC using two gradients, see Supporting
Information for HPLC purity data table and spectroscopy data (1H/
13C NMR) of selected compounds resynthesized.
Synthesis of Aldoxime Derivatives. The corresponding aldehyde
(300 mg) was dissolved in methanol (5 mL). After that, a solution
of hydroxylamine in water at 50% (4 mmol) was added, and the
mixture was refluxed for a period of 30 min. The solvent was then
evaporated, and the resulting solid was dissolved in ethyl acetate
(100 mL) and washed with HCl 1 N (2 × 100 mL), saturated
NaHCO3 (2 × 100 mL) solution, and brine (2 × 100 mL). The
organic phase was dried with MgSO4 and concentrated at reduced
pressure to give the corresponding aldoxime. The compounds were
analyzed by HPLC-MS.
Conclusions
Here we derived SARs for SSAO /VAP-1 with 20 aryalkyl-
amine analogues selected from a library of 48 compounds. The
chemical nature of the substitution in the aromatic ring and the
length of the alkyl chain modulates their activity versus SSAO/
VAP-1. This finding introduces a new concept in SSAO
substrate design that could be useful in subsequent studies.
More importantly, three novel SSAO substrates, namely,
4-fluorobenzylamine (3), 3-fluoro-4-methyl-benzylamine (16),
and 4-phenylbutylamine (44), were identified. These compounds
presented high affinity for mouse and human SSAO, and their
Vmax values were even higher for the latter. Results from docking
Synthesis of the Arylalkylamine Derivatives. Method A. The
corresponding aldoxime (300 mg) and PtO2‚1H2O (0.022 mmol)
were dissolved in acetic acid (10 mL). The mixture was pressurized
with H2 to 4 bar and reacted for 8 h at room temperature. The
calculations were consistent with these experimental values.
Lowest-energy binding modes fitted nicely within the active
site and had the ammonium group near the TPQ residue.

Substrates for Amine Oxidase/VAP-1 Enzyme
Journal of Medicinal Chemistry, 2006, Vol. 49, No. 21 6205
crude was then filtered through Celite, the acetic acid was removed
at low pressure, and the crude product was dissolved with HCl 1
N (300 mL). The aqueous phase was washed with ethyl acetate (3
× 100 mL), and the organic layer was discarded. The aqueous phase
was then basified with a solution of NaOH 1 M and washed with
ethyl acetate (3 × 100 mL). The organic phase was dried with
MgSO4 and concentrated at reduced pressure to give the corre-
sponding amine. Afterward all the compounds were treated with 2
mL of HCl/dioxane 4.0 M in 10 mL of ethyl acetate for 2 h at
room temperature. The final hydrochloride salts were filtered and
dried.
Method B. The corresponding aldoxime (300 mg) was dissolved
with 5 mL of LiBH4 4.0 M in dry THF and refluxed for 8 h. Then
100 mL of ethyl acetate and 100 mL of HCl 6 N were added to the
crude product. The aqueous phase was washed with ethyl acetate
(3 × 100 mL), and the organic layer was discarded. The aqueous
phase was then basified with a solution of NaOH 1 M and was
washed with ethyl acetate (3 × 100 mL). The organic phase was
dried with MgSO4 and concentrated at reduced pressure to give
the corresponding amine. Afterward, all the compounds were treated
with 2 mL of HCl/dioxane 4.0 M in 10 mL of ethyl acetate for 2
h at room temperature. The final hydrochloride salts were filtered
and dried.
(m, 2H). 13C NMR (CD3OD, 400 MHz): 42.9, 43.0, 127.6, 128.0,
128.17, 129.2, 133.5, 139.7, 172.4. ES+ (m/z): calcd for C10H15N2O
(M + H)+, 126.1179; found, 179.1176.
General Method for the Synthesis of N-(3-Aminomethyl-
benzyl)-acylamide Derivatives. 3-(Boc-aminomethyl)-benzylamine
hydrochloride (1 equiv) was dissolved in DMF, and HOBt (1 equiv),
the corresponding carboxylic acid (1 equiv), and DIEA (2 equiv)
were then added. Next, DIPCDI (1 equiv) was added, and the
mixture was stirred for 12 h at room temperature. DMF was then
removed at low pressure. The Boc derivative crude product was
dissolved in 300 mL of ethyl acetate and washed with 5% NaHCO3
(3 × 100 mL). The organic fractions were washed with brine and
dried with MgSO4, and the solvent was removed at low pressure.
All the compounds were purified by silica chromatography and
eluted with a gradient of ethyl acetate/hexane. Afterward, all the
compounds were treated with 2 mL of HCl/dioxane 4.0 M in 10
mL of ethyl acetate for 2 h at room temperature. The final
hydrocloride salts were filtered and dried.
N-(3-(Aminomethyl)benzyl)propionamide Hydrochloride Salt
(31). A white solid was obtained (yield ) 86%) with 100% purity.
1H NMR (CD3OD, 400 MHz): 1.14 (t, 2H, J ) 7.6 Hz), 2.27 (q,
3H, J ) 7.6 Hz), 4.10 (s, 2H), 4.38 (s, 2H), 7.352 (m, 2H), 8.43 (s,
NH), 7.38 (m, 2H). 13C NMR (CD3OD, 400 MHz): 9.2, 28.9, 42.5,
43.0, 127.5, 127.9, 128.0, 129.2, 133.4, 140.2, 175.8. ES+ (m/z):
calcd for C11H17N2O (M + H)+, 193.1335; found, 192.1340.
N-(3-(Aminomethyl)benzyl)-2-iodoacetamide Hydrochloric
Salt (32). A white solid was obtained (yield ) 81%) with 99/97%
(4-Fluorophenyl)methylamine Hydrochloride Salt (3). A white
solid with 99% purity was obtained following method B (yield )
1
40%). H NMR (CD3OD, 400 MHz): 4.06 (s, 2H), 7.06 (dd, 2H,
J ) 6.1 Hz, J ) 11.4 Hz), 7.31 (dd, 2H, J ) 5.4 Hz, J ) 8.2 Hz).
13C NMR (CD3OD, 400 MHz): 40.7, 115.5, 130.4, 132.7, 161.0,
163.4. ES+ (m/z): calcd for C7H9F1N1 (M + H)+, 126.0714; found,
126.0715.
(4-Ethylphenyl)methylamine Hydrochloride Salt (5). A white
solid with 100% purity was obtained following method A (yield )
60%). Compound 5 was recrystallized with i-PrOH/MeOH (2:1)
at 75 °C. 1H NMR (CD3OD, 400 MHz): 1.22 (t, 3H, J ) 7.6 Hz),
2.66 (q, 2H, J ) 7.6 Hz), 7.28 (d, 2H, J ) 8.1 Hz), 7.36 (d, 2H,
J ) 8.1 Hz). 13C NMR (CD3OD, 400 MHz): 14.8, 28.3, 42.9, 128.4,
128.8, 130.4, 145.6. ES+ (m/z): calcd for C9H14N (M + H)+,
136.1121; found, 136.1123.
1
purity. H NMR (CD3OD, 400 MHz): 4.10 (s, 2H), 4.44 (s, 2H),
7.39 (m, 4H). 13C NMR (CD3OD, 400 MHz): 41.9, 42.9, 43.0,
127.6, 128.0, 128.1, 129.3, 133.4, 139.6, 168.1. ES+ (m/z): calcd
for C10H15N2O (M + H)+, 126.0714; found, 126.0714.
N-(3-(Aminomethyl)benzyl)benzamide Hydrochloride Salt
(34). A white solid was obtained (yield ) 85%) with 99% purity.1H
NMR (CD3OD, 400 MHz): 4.10 (s, 2H), 4.60 (s, 2H), 7.36 (m,
1H), 7.42 (m, 2H), 7.43 (m, 3H), 7.54 (m, 1H), 7.85 (m, 2H). 13C
NMR (CD3OD, 400 MHz): 43.0, 43.1, 127.1, 127.5, 128.0, 128.1,
128.4, 129.2, 131.6, 133.4, 134.1, 140.4, 169.0. ES+ (m/z): calcd
for C10H15N2O (M + H)+, 241.1335; found, 241.1335.
(4-(Trifluoromethyl)phenyl)methaylamine Hydrochloride Salt
(8). A white solid with 95% purity was obtained following method
B (yield ) 45%). 1H NMR (CD3OD, 400 MHz): 4.22 (s, 2H), 7.76
(d, 2H, J ) 8.1 Hz), 7.67 (d, 2H, J ) 8.2 Hz). 13C NMR (CD3OD,
400 MHz): 42.5, 125.7, 125.8, 125.8, 125.9, 129.5, 137.6. ES+
(m/z): calcd for C8H9F3N (M + H)+, 176.0682; found, 176.0680.
(4-Butylphenyl)methylamine Hydrochloride Salt (11). A white
solid with 100% purity was obtained following method A (yield )
N-(3-(Aminomethyl)benzyl)-4-hydroxybenzamide Hydrochlo-
ride Salt (38). A white solid was obtained (yield ) 84%) with
1
98% purity. H NMR (CD3OD, 400 MHz): 4.09 (s, 2H), 4.56 (s,
2H), 6.83 (d, 2H, J ) 8.7 Hz), 7.33 (m, 1H), 7.41 (m, 2H), 7.43 (s,
H), 7.75 (d, 2CH, J ) 8.7 Hz). 13C NMR (CD3OD, 400 MHz):
42.9, 43.1, 114.9, 124.9, 127.4, 127.9, 128.1, 129.1, 129.2, 133.3,
140.6, 161.0, 168.8. ES+ (m/z): calcd for C15H17N2O2 (M + H)+,
257.1285; found, 257.1280.
1
79%). H NMR (CD3OD, 400 MHz): 0.93 (t, 3H, J ) 7.4 Hz),
N-(3-(Aminomethyl)benzyl)-4-bromobenzamide Hydrochlo-
ride Salt (39). A white solid was obtained (yield ) 80%) with
1.35 (m, 2H), 1.59 (m, 2H), 2.64 (t, 2H, J ) 7.7 Hz), 4.06 (z, 2H),
7.26 (d, 2H, J ) 7.9 Hz), 7.36 (d, 2H, J ) 8.0 Hz). 13C NMR
(CD3OD, 400 MHz): 13.0, 22.0, 33.6, 35.0, 42.9, 128.7, 129.0,
130.4, 144.1. ES+ (m/z): calcd for C11H18N (M + H)+, 164.1434;
found, 164.1433.
1
100% purity. H NMR (CD3OD, 400 MHz): 4.103 (s, 2H), 4.58
(s, 2H), 7.35 (m, 1H), 7.414 (m, 2H), 7.453 (m, 1H), 7.78 (d, J )
8.68 Hz, 2H), 7.64 (d, J ) 8.68 Hz, 2H). 13C NMR (CD3OD, 400
MHz): 43.1, 43.2, 126.0, 127.5, 128.0, 128.2, 129.0, 129.2, 131.6,
133.2, 133.4, 140.2, 167.8. ES+ (m/z): calcd for C15H16BrN2O (M
+ H)+, 319.0441; found, 319.0441.
(3-Fluoro-4-methylphenyl)methylamine Hydrochloride Salt
(16). A white solid with 98% purity was obtained following method
A (yield ) 53%). Compound 16 was recrystallized with i-PrOH/
N-(3-(Aminomethyl)benzyl)-4-(aminomethyl)benzamide (40).
1
1
MetOH (85:5) at 75 °C. H NMR (CD3OD, 400 MHz): 2.28 (s,
A white solid was obtained (yield ) 82%) with 100% purity. H
3H), 4.06 (s, 2H), 7.09 (t, 1H, J ) 9.0 Hz), 7.30 (ddd, 1H, J ) 2.3
Hz, J ) 4.7 Hz, J ) 7.3 Hz), 7.35 (d, 1H, J ) 7.2 Hz). 13C NMR
(CD3OD, 400 MHz): 13.1, 13.2, 42.4, 115.2, 128.3, 132.4, 160.5,
163.0. ES+ (m/z): calcd for C7H9F1N1 (M + H)+, 140.0870; found,
140.0871.
NMR (CD3OD, 400 MHz): 4.11 (s, 2H), 4.19 (s, 2H), 4.60 (s,
2H), 7.36 (m, 1H), 7.42 (m, 2H), 7.47 (m, 1H), 7.96 (d, J ) 8.39
Hz, 1H), 7.58 (d, J ) 8.27 Hz, 1H). 13C NMR (CD3OD, 400 MHz):
42.6, 43.0, 43.1, 127.6, 128.0, 128.1, 128.9, 129.2, 133.4, 134.8,
136.8, 140.2, 168.0. ES+ (m/z): calcd for C16H21N3O (M + H)+,
270.1601; found, 270.1596.
N-(3-(Aminomethyl)benzyl)Phg(Ac) Hydrochloride Salt (35).
A mixture of of 3-(Boc-aminomethyl)-benzylamine hydrochloride
(300 mg, 1 equiv) was dissolved in DMF (100 mL) and HOBt (149
mg, 1 equiv). Fmoc-phenylglycine carboxylic acid (410 mg, 1
equiv) and DIPCDI (162 µL, 1 equiv) were then added. DIEA (204
µL, 1.1 equiv) was then added, and the mixture was stirred for 12
h at room temperature. After that, the DMF was removed at low
pressure. The crude of N-(3-(aminomethyl)benzyl)Phg(Fmoc) was
dissolved in 300 mL of ethyl acetate and washed with 5% NaHCO3
N-(3-(Aminomethyl)benzyl)acetamida Hydrochloride Salt
(12). 3-(Boc-aminomethyl)-benzylamine hydrochloride (300 mg,
1 equiv) was acetylated with Ac2O/DIEA (1:2) in DCM for 4 h at
room temperature. The crude product was then dried at low pressure
and purified by silica chromatography and eluted with ethyl acetate/
hexane (8:2). Afterward, 12 was treated with 2 mL of HCl/dioxane
4.0 M in 10 mL of ethyl acetate for 2 h at room temperature. The
final hydrochloride salt was filtered and dried. A white solid with
1
99% purity was obtained (yield ) 82%). H NMR (CD3OD, 400
MHz): 2.04 (s, 3H), 4.10 (s, 2H), 4.40 (s, 2H), 7.35 (m, 2H), 7.38

6206 Journal of Medicinal Chemistry, 2006, Vol. 49, No. 21
Yraola et al.
(3 × 100 mL). The organic layers were pooled, washed with brine
(3 × 100 mL), and dried with MgSO4, and the solvent was removed
at low pressure. N-(3-(Aminomethyl)benzyl)Phg(Fmoc) was dis-
solved in DCM (10 mL). Piperidine 20% in DMF (2 mL) was then
added, and the mixture was stirred for 4 h at room temperature.
The crude product was then washed with diethyl ether (3 × 100
mL) to remove the dibenzofulvene. N-(3-(Aminomethyl)benzyl)-
PhgNH2 was dissolved in DCM (100 mL) and acetylated with
Ac2O/DIEA (1:2) for 4 h at room temperature to yield the final
product. The crude product was purified by silica chromatography
with AcOEt/hexane (8:2, v/v). Afterward, the compound was treated
with 2 mL of HCl/dioxane 4.0 M in 10 mL of ethyl acetate for 2
h at room temperature. The final hydrochloride salt was filtered
and dried. A white solid was obtained (yield ) 70%) with 100%
-80 °C until use. Protein concentrations were measured by the
Bradford method with γ-globulin as standard.
Fluorimetric Detection of SSAO-Mediated H2O2 Formation.
The SSAO activity of human recombinant VAP-1 (0.1 µg protein/
assay) and mouse adipose tissue membranes (1 µg protein/assay)
was measured using Amplex Red reagent, a highly sensitive and
stable probe for H2O2.24 The reaction was performed in 200 µL of
0.2 mol/L phosphate buffer at pH 7.4 for 50 min at 37 °C in black
nonphosphorescent microplates (Nunc). Catalytic reaction was
initiated by the addition of the amines indicated as putative SSAO
substrates and a H2O2-detecting mixture containing horseradish
peroxidase and Amplex Red, as previously described.15 Fluores-
cence intensity was measured (excitation, 545 nm; emission, 590
nm; Bio-Tek fluorescence plate reader), and H2O2 concentration
was calculated from calibration curves generated by serial dilutions
of standard H2O2. Fluorescence readings were performed every 5
min. SSAO-independent H2O2 production was measured by pre-
incubating adipocyte membranes or human SSAO protein with
250 µM semicarbazide for 20 min to totally inhibit SSAO activity,
and fluorescence values were subtracted from the total amount of
H2O2 formed. The kinetic parameters (Km and Vmax) were calculated
using appropriate nonlinear curve-fitting formula following the
Michaelis-Menten equation and using GraphPad Prism 4.0 soft-
ware.
Glucose Transport Measurements in Isolated Mouse Adipo-
cytes. Adipocytes were isolated from internal fat pads of healthy
male Swiss mice (30-35 g) by digestion in KRBHA contain-
ing 0.66 mg colagenase/mL.4,5 After a preincubation period of
45 min at 37 °C, each vial, which contained 200 µL of cell
suspension in KRBHA and the drugs being tested, received an
isotopic dilution of 2-deoxy-D-[3H]glucose (2-DG), giving a final
concentration of 0.1 mM, equivalent to approximately 1 300 000
dpm/vial. 2-DG transport assays were performed as reported.4,5 The
values of basal transport represented 0.33 ( 0.03 nmol 2-DG/mg
lipid/10 min.
Molecular Modeling and Docking. Homology modeling of the
biological dimer of mouse SSAO was performed on the basis of
the crystallographic structure of homologous human SSAO (83.8%
sequence identity), whose coordinates are deposited in the Protein
Data Bank (PDB) with code 1us1.25-27 Given the high sequence
identity between mouse and human SSAO, no other template was
used. We followed the standard protocol in Modeler.15 Fifteen side
chains were automatically placed by the default procedure, followed
by a standard minimization in Modeler, with no additional restraints
applied to the active site or to any other part of the structure. The
five lowest-scoring conformations were automatically selected
according to the Modeler objective function. From these five
conformations, as our final model, we selected the one that yielded
the best values of PROCHECK test.28 This model was subjected
to molecular dynamics simulations using NAMD with ff99
parameterization.29,30 Molecular dynamics of the mouse SSAO
model built was performed with a protocol comprising an initial
10 000-step conjugate gradient minimization in full explicit solvent,
followed by 100 ps of equilibration (gradually reducing positional
restraints on protein atoms, with no restraints on explicit solvent
atoms) and finally 200 ps of free molecular dynamics in full explicit
solvent at constant temperature (T ) 300 K). Protein was placed
in a truncated octahedron box filled by water molecules, and all
atoms (protein and solvent) were free to move during the molecular
dynamics step. The complete protocol took around 50 h in 16 CPUs
(Intel Xeon 2.80 MHz). The model was stable throughout both the
equilibration and the molecular dynamics phases, in which atomic
coordinates remained mostly within 1.6 Å of RMSD from the initial
coordinates. The same protocol was applied to the X-ray structure
of human SSAO, yielding similarly stable dynamics. Selected
molecules for docking (1, 34, 41, 42, 44) were minimized in the
MM3 force field using TINKER.31,32 Conformational sampling of
these molecules was performed by molecular dynamics in the MM3
force field, as implemented in TINKER, a molecular modeling
package previously developed in our laboratory.33 For each of these
conformations, CMIP performed an exhaustive rotational and
1
purity. H NMR (CD3OD, 400 MHz): 2.0 (s, 3H), 4.06 (s, 2H),
4.41 (ddd, 2H, J ) 5.9 Hz, J ) 15.5 Hz, J ) 56.6 Hz), 5.36 (s, H),
7.25 (m, 1H), 7.34 (m, 6H), 7.44 (m, 2H), 8.78 (s bb, NH). 13C
NMR (CD3OD, 400 MHz): 19.6, 21.2, 42.4, 43.1, 58.6, 60.3, 127.3,
127.4, 127.7, 127.8, 128.3, 128.6, 129.0, 133.3, 136.9, 139.9, 171.9,
172.1. ES+ (m/z): calcd for C10H15N2O (M + H)+, 312.1707;
found, 312.1703.
N-(3-(Aminomethyl)benzyl)allylcarbamate Hydrocloride Salt
(33). 3-(Boc-Aminomethyl)-benzylamine hydrochloride (300 mg,
1 equiv) was dissolved in DCM. DIEA (932 µL,1 equiv) and allyl
chloroformate (88 µL,1 equiv) were then added, and the reaction
mixture was stirred for 4 h at room temperature. Afterward, the
solvent was evaporated, and the crude product was dissolved in
ethyl acetate (300 mL), washed with H2O (2 × 100 mL) and brine
(2 × 100 mL), and dried with MgSO4. The product was dried at
low pressure and analyzed by HPLC-MS. The compound was then
treated with 2 mL of HCl/dioxane 4.0 M in 10 mL of ethyl acetate
for 2 h at room temperature. The final hydrochloride salt was filtered
and dried. A white solid was obtained (yield ) 92%) with 99%
1
purity. H NMR (CD3OD, 400 MHz): 4.10 (s, 2H), 4.31 (s, 2H),
4.54 (d, 2H, J ) 5.4 Hz), 5.24 (dd, 2H, J ) 13.9 Hz, J ) 47.6
Hz), 5.93 (ddd, 1H, J ) 5.4 Hz, J ) 10.6 Hz, J ) 22.1 Hz), 7.38
(m, 4H), 7.60 (s, NH). 13C NMR (CD3OD, 400 MHz): 43.1, 43.9,
65.3, 116.3, 127.5, 127.7, 127.8, 129.2, 133.2, 133.4, 140.7, 157.6.
ES+ (m/z): calcd for C12H17N2O2 (M + H)+, 221.1285; found,
221.1281.
(3-((Tosylamino)methyl)phenyl)methanamonium Hydrochlo-
ride Salt (36). An amount of p-toluensulfonyl chloride (208 mg,
1.0 equiv) and 3-(Boc-aminomethyl)-benzylamine hydrochloride
(300 mg, 1.0 equiv) were dissolved in 20 mL ice-cooled acetone
and stirred for 10 min. Afterward, K2CO3 (379 mg, 2.5 equiv)
dissolved in water (10 mL) was added, and the mixture was stirred
for 4 h at room temperature. The solvent was then evaporated, and
the crude product was dissolved in ethyl acetate (300 mL). The
organic layer was washed with water (3 × 100 mL), and brine (3
× 100 mL), and dried with MgSO4, and the solvent was then
removed. The product was dried at low pressure and analyzed by
HPLC-MS. Afterward, the Boc derivative was treated with 2 mL
of HCl/dioxane 4.0 M in 10 mL of ethyl acetate for 2 h at room
temperature. The final hydrochloride salt was filtered and dried. A
pale yellow solid was obtained (yield ) 86%) with 99% purity. 1H
NMR (CD3OD, 400 MHz): 2.41 (s, 3H), 4.05 (s, 2H), 4.06 (s,
2H), 5.48 (s, SO2NH), 7.28 (m, 1H), 7.33 (m, 3H), 7.36 (d, 2H, J
) 8.4 Hz), 7.73 (d, 2H, J ) 8.2 Hz). 13C NMR (CD3OD, 400 MHz):
20.3, 43.0, 46.3, 126.9, 127.8, 128.2, 128.4, 129.1, 129.5, 133.3,
137.7, 138.9, 143.5. ES+ (m/z): calcd for C15H19N2O2S (M + H)+,
291.1162; found, 291.1161.
Preparation of Mouse Adipose Tissue Membranes. Internal
adipose tissue was obtained from Swiss mice weighing between
20 and 25 g. The tissue was cut and homogenized in HES buffer
(25 mmol/L HEPES, 2 mmol/l EDTA, 255 mmol/L sucrose) with
antiproteases (1 µmol/L pepstatin, 1 µmol/L leupeptin, 0.14 trypsin
inhibitor units per mL aproptinin and 1 mmol/L PMSF). Lysates
were then centrifuged at 5000 × g at 4 °C for 15 min to remove
the fat cake and nonhomogenized material, and supernatants were
collected and centrifuged at 200 000 × g for 2 h at 4 °C. Pelleted
membranes were resuspended in 30 mmol/ L HEPES and stored at

Substrates for Amine Oxidase/VAP-1 Enzyme
Journal of Medicinal Chemistry, 2006, Vol. 49, No. 21 6207
(6) Morin, M.; Visentin, V.; Calisse, D.; Marti, L.; Zorzano, A.; Testar,
X.; Valet, P.; Fischer, Y.; Carpe´ne´, C. Tyramine stimulates glucose
uptake in insulin-sensitive tissues in vitro and in vivo via its oxi-
dation by amine oxidases. J. Pharmacol. Exp. Ther. 2002, 303, 1238-
1247.
(7) Morin, N.; Lizcano, J. M.; Fontana, E.; Marti, L., Smih, F.; Rouet,
P.; Prevot, D.; Zorzano, A.; Unzeta, M.; Carpe´ne´, C. Semicarbazide-
sensitive amine oxidase substrates stimulate glucose transport and
inhibit lipolysis in human adipocytes. J Pharmacol. Exp. Ther. 2001,
297, 563-572.
(8) Fontana, E.; Boucher, J.; Marti, L.; Lizcano, J. M.; Testar, X.,
Zorzano, A.; Carpe´ne´, C. Amine oxidase substrates mimic several
of the insulin effects on adipocyte differentiation in 3T3 F442A cells.
Biochem J. 2001, 356, 769-777.
(9) Abella, A.; Marti, L.; Camps, M.; Claret, M.; Ferna´ndez-Alvarez,
J.; Gomis, R.; Guma`, A.; Viguerie, N.; Carpe´ne´, C.; Palac´ın, M.;
Testar, X.; Zorzano, A. Semicarbazide-sensitive amine oxidase/
vascular adhesion protein-1 activity exerts an antidiabetic action in
Goto-Kakizaki rats. Diabetes 2003, 52, 1004-1013.
(10) Marti, L.; Abella, A.; Carpe´ne´, C.; Palac´ın, M.; Testar, X.; Zorzano,
A. Combined treatment with benzylamine and low dosages of
vanadate enhances glucose tolerance and reduces hyperglycemia in
streptozotocin-induced diabetic rats. Diabetes 2003, 50, 2061-2068.
(11) Zorzano, A.; Abella, A.; Marti, L.; Carpe´ne´, C.; Palac´ın, M.; Testar,
X. Semicarbazide-sensitive amine oxidase activity exerts insulin-like
effects on glucose metabolism and insulin-signaling pathways in
adipose cells. Biochim. Biophys. Acta 2003, 1647, 3-9.
(12) Yu, P. H.; Zuo, D. M. Oxidative deamination of methylamine by
semicarbazide-sensitive amine oxidase leads to cytotoxic damage in
endothelial cells. Possible consequences for diabetes. Diabetes 1993,
42, 594-603.
(13) Marti, L.; Abella, A.; De la Cruz, X.; Garc´ıa-Vicente, S.; Unzeta,
M.; Carpe´ne´, C.; Palac´ın, M.; Testar, X.; Orozco, M.; Zorzano, A.
Exploring the binding mode of semicarbazide-sensitive amine oxi-
dase/VAP-1: identification of novel substrates with insulin-like
activity. J. Med. Chem. 2004, 47, 4865-4874.
(14) Bair, K. W.; Tuttle, R. L.; Knick, V. C.; Cory, M.; McKee, D.D.
[(1-Pyrenylmethyl)amino]alcohols, a new class of antitumor DNA
intercalators. Discovery and initial amine side chain structure-activity
studies. J. Med. Chem. 1990, 33, 2385-2393.
(15) Scozzafava, A.; Supuran, C. T. Carbonic anhydrase and matrix
metalloproteinase inhibitors: sulfonylated amino acid hydroxamates
with MMP inhibitory properties act as efficient inhibitors of CA
isozymes I, II, and IV, and N-hydroxysulfonamides inhibit both these
zinc enzymes. J. Med. Chem., 2000, 43, 3677-3687.
(16) Lyles, G. A. Substrate-specificity of mammalian tissue-bound semi-
carbazide-sensitive amine oxidase. Prog. Brain Res. 1995, 106, 293-
303.
(17) 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. Crystal structures of the copper-containing amine
oxidase from Arthrobacter globiformis in the holo and apo forms:
implications for the biogenesis of topaquinone. Biochemistry 1997,
36, 16116-16133.
(18) Salminen, T. A.; Smith, D. J.; Jalkanen, S.; Johnson, M. S. Structural
model of the catalytic domain of an enzyme with cell adhesion
activity: human vascular adhesion protein-1 (HVAP-1) D4 domain
is an amine oxidase. Protein Eng. 1998, 11, 1195-1204.
(19) Lunelli, M.; Di Paolo, M. L.; Biadene, M.; Calderone, V.; Battistutta,
R.; Scarpa, M.; Rigo, A.; Zanotti, G. Crystal structure of amine
oxidase from bovine serum. J. Mol. Biol. 2005, 346, 991-1004.
(20) Liang, J.; Edelsbrunner, H.; Woodward, C. Anatomy of protein
pockets and cavities: measurement of binding site geometry and
implications for ligand design. Protein Sci. 1998, 7, 1884-1897.
(21) Salmi, M.; Yegutkin, G. G.; Lehvonen, R.; Koskinen, K.; Salminen,
T.; Jalkanen, S. A cell surface amine oxidase directly controls
lymphocyte migration. Immunity 2001, 14, 265-276.
translational search in a series of steps for which the binding energy
is computed using binding potentials precalculated on a 3D grid
(0.5 Å grid step) within a box of size of 30 × 28 × 24 Å3 covering
the full SSAO active site. These binding energy potentials accounted
for the Lennard-Jones van der Waals interactions and the Poisson-
Boltzmann electrostatics, with ff99 parameterization.20 In addition
to the one in the X-ray structure (PDB code 1us1), several
conformational states of SSAO were considered, for which a
number of models were generated as follows. Two conformations
of the TPQ residue were considered: on-copper inactive conforma-
tion (the one in the X-ray structure of human SSAO) and off-copper
active conformation (with coordinates taken from bovine SSAO
with modified residue mimicking benzylamine covalently attached
to the TPQ). To roughly mimic the flexibility of the binding pocket,
an alternative side-chain rotamer was considered for the Leu469
residue, in addition to that in the X-ray structure. The TPQ residue
was parameterized using LEAP from Amber, with atomic charges
calculated by Gaussian 03 (www.Gaussian.com).34 Binding energy
potentials were calculated by CMIP on a 3D grid around the SSAO
active site residues. Given that the protein coordinates were fixed
during each docking simulation, the van der Waals radii of protein
and substrate atoms were scaled by a factor of 0.8, to avoid strong
clashes derived from the suboptimal geometries of the docked
orientations.
Acknowledgment. This study was supported by research
grants from the Ministerio de Ciencia y Tecnolog´ıa (SAF2002-
02125, SGR01-118, BQU2002-02047, CTQ2005-0315, and
BQU2003-00089), Grant SGR01-118 from the Generalitat de
Catalunya, the Fundacio´ Marato´ de TV3 (300720), and from
the Instituto de Salud Carlos III RCMN (C03/08), RGDM (G03/
212), and RGTO (G03/028), the Generalitat de Catalunya (Grup
Consolidat and Centre de Refere`ncia en Biotecnologia), and the
Barcelona Science Park. J.F.-R. is supported by the Ramon y
Cajal program (Ministerio de Ciencia y Tecnolog´ıa, Spain).
We thank Molsoft for the use of ICM-Browser and Modesto
Orozco and Josep Lluis Gelpi for useful comments and the use
of CMIP. NAMD was developed by the Theoretical and
Computational Biophysics Group in the Beckman Institute for
Advanced Science and Technology at the University of Illinois
at Urbana-Champaign. We thank Tanya Yates for her editorial
support.
Supporting Information Available: Combinatorial chemistry,
preliminary biological HPLC data, purity data table, and spectros-
copy data (1H/13C NMR) of selected compounds. This material is
available free of charge via the Internet at http://pubs.acs.org
References
(1) Murray, M. J.; Saysell, C. G.; Wilmot, C. M.; Tambyrajah, W. S.;
Jaeger, J.; Knowles, P. F.; Pillips, S.; McPherson, M. J. The active
site base controls cofactor reactivity in Escherichia coli amine
oxidase: X-ray crystallographic studies with mutational variants.
Biochemistry 1999, 38, 8217-8227.
(2) Jalkannen, S.; Salmi, M. Cell surface monoamine oxidases: enzymes
in search of a function. EMBO J. 2001, 20, 3893-3901.
(3) Morris, N. J.; Ducret, A.; Aebersold, R.; Ross, S. A.; Keller, S. R.;
Lienhard, G. E. Membrane amine oxidase cloning and identification
as a major protein in the adipocyte plasma membrane. J. Biol. Chem.
1997, 272, 9388-9392.
(4) Enrique-Taranco´n, G.; Marti, L.; Morin, M.; Lizcano, J., M.; Unzeta,
M.; Sevilla, L.; Camps, M.; Palac´ın, M.; Testar, X.; Carpe´ne´, C.;
Zorzano, A. Role of semicarbazide-sensitive amine oxidase on
glucose transport and GLUT4 recruitment to the cell surface in
adipose cells. J. Biol. Chem. 1998, 273, 8025-8032.
(5) Marti, L.; Morin, M.; Enrique-Taranco´n, G.; Prevot, D.; Lafontan,
M.; Testar, X.; Zorzano, A.; Carpe´ne´, C. Tyramine and vanadate
synergistically stimulate glucose transport in rat adipocytes by amine
oxidase-dependent generation of hydrogen peroxide. J. Pharmacol.
Exp. Ther. 1998, 285, 342-349.
(22) Lyles, G. A. Substrate-specificity of mammalian tissue-bound semi-
carbazide-sensitive amine oxidase. Prog. Brain Res. 1995, 106, 293-
303.
(23) Wilmot, C. M.; Murray, J. M.; Alton, G.; Parsons, M. R.; Convery,
M. A.; Blakeley, V.; Corner, A. S.; Palcic, M. M.; Knowles, P. F.;
McPherson, M. J.; Phillips, S. E. V. Catalytic mechanism of the
quinoenzyme amine oxidase from Escherichia coli: exploring the
reductive half-reaction. Biochemistry 1997, 36, 1608-1620.
(24) Zhou, M.; Panchuk-Voloshina, N. A one-step fluorometric method
for the continuous measurement of monoamine oxidase activity. Anal.
Biochem. 1997, 253, 169-174.
(25) Sali, A.; Blundell, T. L. Comparative protein modeling by satisfaction
of spatial restraints. J. Mol. Biol. 1993, 234, 779-815.

6208 Journal of Medicinal Chemistry, 2006, Vol. 49, No. 21
Yraola et al.
(26) Bernstein, F. C.; Koetzle, T. F.; Williams, G. J.; Meyer, E. F., Jr.;
Brice, M. D.; Rodgers, J. R.; Kennard, O.; Shimanouchi, T.; Tasumi,
M. The protein data bank: a computer-based archival file for
macromolecular structures. J. Mol. Biol. 1977, 112, 535-542.
(27) Airenne, T. T.; Nymalm, Y.; Kidron, H.; Smith, D. J.; Pihlavisto,
M.; Salmi, M.; Jalkanen, S.a; Johnson, M. S.; Salminen, T. A. Crystal
structure of the human vascular adhesion protein-1: Unique structural
features with functional implications. Protein Sci. 2005, 14, 1964-
1974.
(28) Laskowski, R. A.; MacArthur, M. W.; Moss, D. S.; Thornton, J. M.
PROCHECK: a program to check the stereochemical quality of
protein structures. J. Appl. Crystallogr. 1993, 26, 283.
(29) Phillips, J. C.; Braun, R.; Wang, W.; Gumbart, J.; Tajkhorshid, E.;
Villa, E.; Chipot, C.; Skeel, R. D.; Kale, L.; Schulten, K. Scalable
molecular dynamics with NAMD. J. Comput. Chem. 2005, 26, 1781-
1802.
(31) Allinger, N. L.; Yuh, Y. H.; Lii, J. H. Molecular mechanics. The
MM3 force field for hydrocarbons. 1. J. Am. Chem. Soc. 1989, 111,
8551-8566.
(32) Ponder, J. W.; Richards, F. M. An efficient Newton-like method for
molecular mechanics energy minimization of large molecules. J.
Comput. Chem. 1987, 8, 1016-1024.
(33) Gelpi, J. L.; Kalko, S. G.; Barril, X.; Cirera, J.; de La Cruz, X.;
Luque, F. J.; Orozco, M. Classical molecular interaction potentials:
improved setup procedure in molecular dynamics simulations of
proteins. Proteins 2001, 45, 428-437.
(34) Pearlman, D. A.; Case, D. A.; Caldwell, J. W.; Ross, W. S.;
Cheatham, T. E., III; DeBolt, S.; Ferguson, D.; Seibel, G.e; Kollman,
P. “AMBER”, a package of computer programs for applying
molecular mechanics, normal-mode analysis, molecular dynamics
and free energy calculations to stimulate the structural and
energetic properties of molecules. Comput. Phys. Commun. 1995,
91, 1-42.
(30) Wang, J.; Cieplak, P.; Kollman, P. A. How well does a restrained
electrostatic potential (RESP) model perform in calculating confor-
mational energies of organic and biological molecules? J. Comput.
Chem. 2000, 21, 1049-1074.
JM051076E