Synthesis, in vitro activity, and three-dimensional quantitative Structure-activity relationship of novel hydrazine inhibitors of human vascular adhesion protein-1

Article abstract of DOI:10.1021/jm100337z

Vascular adhesion protein-1 (VAP-1) belongs to the semicarbazide-sensitive amine oxidases (SSAOs) that convert amines into aldehydes. SSAOs are distinct from the mammalian monoamine oxidases (MAOs), but their substrate specificities are partly overlapping. VAP-1 has been proposed as a target for anti-inflammatory drug therapy because of its role in leukocyte adhesion to endothelium. Here, we describe the synthesis and in vitro activities of novel series of VAP-1 selective inhibitors. In addition, the molecular dynamics simulations performed for VAP-1 reveal that the movements of Met211, Ser496, and especially Leu469 can enlarge the ligand-binding pocket, allowing larger ligands than those seen in the crystal structures to bind. Combining the data from molecular dynamics simulations, docking, and in vitro measurements, the three-dimensional quantitative Structure-activity relationship (3D QSAR) models for VAP-1 (q²_LOO: 0.636; r²: 0.828) and MAOs (q²_LOO: 0.749, r²: 0.840) were built and employed in the development of selective VAP-1 inhibitors.

Full text of DOI:10.1021/jm100337z

J. Med. Chem. 2010, 53, 6301–6315 6301
DOI: 10.1021/jm100337z
Synthesis, in Vitro Activity, and Three-Dimensional Quantitative Structure-Activity Relationship of
Novel Hydrazine Inhibitors of Human Vascular Adhesion Protein-1
†
‡
§
§
†
§
Elisa M. Nurminen, Marjo Pihlavisto, Laszlo Lazar, Zsolt Szakonyi, Ulla Pentikainen, Ferenc Fulop, and
ꢀ
ꢀ
ꢀ ꢀ
€
€ €
,†
€
Olli T. Pentikainen*
†Department of Biological and Environmental Science & Nanoscience Center, P.O. Box 35, FI-40014 University of Jyva€skyla€, Finland,
‡BioTie Therapies Corporation, Tykisto€katu 6, FI-20520 Turku, Finland, and §Institute of Pharmaceutical Chemistry, University of Szeged,
H-6720 Eotvos 6, Szeged, Hungary
Received February 17, 2010
Vascular adhesion protein-1 (VAP-1) belongs to the semicarbazide-sensitive amine oxidases (SSAOs)
that convert amines into aldehydes. SSAOs are distinct from the mammalian monoamine oxidases
(MAOs), but their substrate specificities are partly overlapping. VAP-1 has been proposed as a target for
anti-inflammatory drug therapy because of its role in leukocyte adhesion to endothelium. Here, we
describe the synthesis and in vitro activities of novel series of VAP-1 selective inhibitors. In addition, the
molecular dynamics simulations performed for VAP-1 reveal that the movements of Met211, Ser496,
and especially Leu469 can enlarge the ligand-binding pocket, allowing larger ligands than those seen in
the crystal structures to bind. Combining the data from molecular dynamics simulations, docking, and
in vitro measurements, the three-dimensional quantitative structure-activity relationship (3D QSAR)
models for VAP-1 (q2LOO: 0.636; r2: 0.828) and MAOs (q2LOO: 0.749, r2: 0.840) were built and employed
in the development of selective VAP-1 inhibitors.
Introduction
VAP-1 has adhesive properties, in addition to its catalytic
function. It is a type II transmembrane glycoprotein with a
large glycosylated extracellular domain and a short N-terminal
intracellular domain of four amino acids, which have a role in
the adhesion of lymphocytes to endothelial cells and lympho-
cyte recirculation in an L-selectin-independent manner.8,9 The
crystal structures determined for human VAP-110,11 show that
its structure is homodimeric. The active site is buried deep
within the protein, including the TPQ cofactor, three conserved
histidines which are coordinated to the Cu2þ required for TPQ
biogenesis, and the catalytic base (Asp386).10,11 The accessi-
bility of the active site is restricted by Leu469, which blocks the
entrance to the ligand binding site.11
Changes in the expression levels or enzymatic activity of
VAP-1 are related to numerous inflammation-associated dis-
eases such as arthritis,12,13 type 1 and type 2 diabetes,14-16
obesity,17 and congestive heart failure.18 In addition to the
harmful effects of changes in expression levels of VAP-1, some
reaction products, such as methylglyoxal, which is able to
form advanced-glycation end-products, can be cytotoxic.19
The number of VAP-1-related diseases and the populations
they affect make them an interesting target for drug discovery.
Several small molecule VAP-1 inhibitors have already been
identified, such as phenylallylhydrazines,20 thiocarbamoyl
derivatives,21 carboxamides,22 and sulfonamides (Figure 1).23
In addition to small molecule inhibitors, VAPþ-1 function can
be altered by using (1) peptides with NH3 functionality
(lysine side chain),24-26 (2) function blocking antibodies
(that inhibit leukocyte trafficking),27 and (3) small interfering
RNAs(thatinhibitbothadhesionand enzymatic functions).28
The major challenge in the design of VAP-1 inhibitors is to
gain enzyme subtype selectivity against other amine oxidases.
Amine oxidases (AOs) are a group of enzymes that catalyze
amine metabolism; they convert amine compounds into cor-
responding aldehydes with concomitant hydrogen peroxide
and ammonia production. All reaction products are known
to be cytotoxic at high concentrations and may play an impor-
tant role in the pathogenesis of many types of vasculopathies.1,2
AOs can be divided into two families based on their cofactors,
namely, (1) posttranslationally modified amino acids [either
topaquinone (TPQa)3 or lysine tyrosylquinone4 or (2) flavin.
Semicarbazide-sensitive amine oxidases (SSAOs) such as vascular
adhesion protein-1 (VAP-1) (EC 1.4.3.21) and diamine oxidase
(EC 1.4.3.22) use TPQ as a cofactor, whereas lysyl oxidase (EC
1.4.3.13) uses lysine tyrosylquinone. Flavin-containing AOs
are, for example, monoamine oxidases (MAO) A and B (EC
1.4.3.4), and polyamine oxidases (EC 1.5.3.11.). The natural
substrate for VAP-1 is not known, but primary amines are
known to be accepted as substrates (e.g., benzylamine and
methylamine).5 MAOs’ natural substrates include, for example,
neurotransmitters such as dopamine, epinephrine, and nore-
pinephrine.6 Both VAP-1 and MAOs can degrade serotonin
and tyramine.7
*To whom correspondence should be addressed. Address: Department
of Biological and Environmental Science & Nanoscience Center, Com-
putational Bioscience Laboratory, P.O. Box 35, FI-40014 University of
€
€
Jyvaskyla, Finland. Phone: þ358-14-260-4186. Fax: þ358-14-260-2221.
E-mail: olli.t.pentikainen@jyu.fi.
a Abbreviations: 3D QSAR, three dimensional quantitative struc-
ture-activity relationship; CoMSIA, comparative molecular similarity
indices analysis; LOO, leave one out; MAO, monoamine oxidase; MD,
molecular dynamics; PLS, partial least square; SEE, standard error of
estimate; SSAO, semicarbazide-sensitive amine oxidase; TPQ, topaqui-
none; VAP-1, vascular adhesion protein-1.
r
2010 American Chemical Society
Published on Web 08/06/2010
pubs.acs.org/jmc

6302 Journal of Medicinal Chemistry, 2010, Vol. 53, No. 17
Nurminen et al.
their hydrogen maleate salt. For the methoxy-substituted
derivatives, this separation procedure could only result in the
main products (5b,c), and all of our efforts to isolate the
minor isomers (6b,c) failed (Scheme 2).
The synthesis of 1,2-hydrazino alcohols 9a-j and 10a-d
(Table 1) was also accomplished by sequential nitrosation
and reduction reactions starting from the corresponding
amino alcohol 7. C-Unsubstituted 1,2-amino alcohols (7:
R3 =R4 =H) were obtained by the ring-opening reactions
of the corresponding styrene oxides with amines.38,39
Ephedrine-homologue amino alcohols (7: R3 = Et, Pr, R4 =
H; or R3 = R4 = Me) were also prepared by using the
methods described in the literature.40-42 The N-nitroso
intermediates 8 were also converted to their O-methyl deri-
vatives (11),43,44 the LiAlH4 reduction of which resulted in
hydrazino ethers 12a-j (Scheme 3).
In Vitro Activity. In vitro potency was determined for the
synthesized inhibitors 3a-3s, 9a-9j, 12a-12j, and 10a-10d
(Table 1). The results indicate that most of the compounds
are specific inhibitors of human VAP-1 activity. The IC50 for
VAP-1 inhibition varies from 20 nM to 2.24 μM, which is
110-fold variation in inhibition (Table 1). The major diffe-
rences were in the selectivity of inhibitors against the total
MAO potency, with selectivity ranging from 6- to 200-fold
(Table 1).
Flexibility of the Ligand-Binding Pocket. All synthesized
compounds are larger (Table 1) than the 2-hydrazinopyr-
idine seen in the only crystal structure of VAP-1 with a bound
inhibitor.10 The biological data (Table 1) also show that
notably larger compounds also bind strongly to VAP-1, for
example, compounds 3b, 3f, 3o, and 5a-c (Table 1). For
future improvement of VAP-1 inhibitors, it is crucial to
understand how the conformation of the ligand-binding
pocket changes when larger inhibitor molecules bind. The
flexibility of the ligand-binding pocket was hence examined
using MD simulations with a bound ligand. In particular, we
wanted to investigate the flexibility of Leu469, as Leu469
seems to act as a gate residue into the substrate binding site.11
From the synthesized molecules, compounds 5a-c (Table 1)
have a hydrogen atom and a hydroxyl group at positions R1
and R2, respectively, binds more strongly to VAP-1 than
those having a methyl group or larger at position R1 (9a-9j,
Table 1) or methyl and methoxy groups at positions R1
and R2, respectively (12a-j, Table 1). As the size of the
aryl group is not critical for binding to VAP-1 (Table 1,
compounds 5a-c), the molecule 5a was chosen for MD
simulations, as it has the simplest structure and shows
strong binding and good VAP-1 selectivity over total MAO
(Table 1).
In the crystal structure of VAP-1 with the bound inhibitor
2-hydrazinopyridine, Leu469 is in a conformation where its
side chain blocks the ligand-binding cavity (Figure 2A,B).
Interestingly, during the course of the MD simulations of
VAP-1 with the bound ligand 5a (Table 1), which is con-
siderably larger than 2-hydrazinopyridine as seen in the
crystal structure, the side chain of Leu469 rotates and un-
blocks the ligand-binding cavity, providing ligands with free
access from the solvent to the ligand-binding pocket
(Figure 2C,D). This new conformation of Leu469 is stabi-
lized by hydrophobic interactions with Phe239 (Figure 2D).
Another remarkable conformational change observed in
MD simulations is that the side chains of Met211 and
Ser496 rotate, and additional space becomes available for
ligand binding (Figure 2B,D); however, none of inhibitors in
Figure 1. Structures of some known small molecule VAP-1 inhibi-
tors. Phenylallylhydrazines a-d, thiocarbamoyl derivatives a-c,
sulfonamide, and carboxamide.
Only a few molecules such as phenylallylhydrazines20 have so
far been reported to exhibit VAP-1 selectivity over MAOs.
Available crystal structuresfor VAP-1 withboundinhibitor,10
and for other amine oxidases such as MAO A29,30 and MAO
B,31-34 however, enable the rational structure-based inhibitor
design. The crystal structures of VAP-110 and MAO B34 with
bound hydrazine inhibitors reveal that the reaction mechan-
isms for these two enzymes are different: in MAO B the
hydrazine moiety of the ligand is removed in the course of
the reaction, whereas in VAP-1 it remains in the covalently
bound product.
Here, we show the ability of synthesized novel 46 hydrazine
molecules to inhibit VAP-1 and total MAO activities (both
MAO A and B). Among these compounds, there are several
molecules that are highly selective for VAP-1 over total MAO.
In addition, we have investigated the flexibility of the ligand-
binding pocket of VAP-1 using molecular dynamics (MD)
simulations. The resultsreveal that the flexibility of the VAP-1
ligand-binding pocket enables the binding of much larger
ligands than have thus far been shown. On the basis of the
biological data and information obtained from MD simula-
tion, we have built three-dimensional quantitative structure-
activity relationship (3D QSAR) models for VAP-1 and
total MAO. Both 3D QSAR models are based on the same
ligand superposition, allowing a rapid prediction not only of
inhibitor potency but also of inhibitor selectivity between
VAP-1 and total MAO. This is advantageous for VAP-1
ligand discovery, since only one docking simulation is needed.
To further validate the quality of these 3D QSAR models,
novel VAP-1 inhibitors were developed by employing the
models.
Results and Discussion
Synthesis. The N1,N1-disubstituted hydrazines 3a-s(Table 1)
were prepared starting from the corresponding secondary
amines 1a-s by a conventional two-step procedure,35,36 includ-
ing N-nitrosation and subsequent LiAlH4 reductions of the
N-nitroso intermediates 2a-s (Scheme 1).
Unsubstituted hydrazino alcohols 5 and 6 were synthe-
sized by ring-opening reactions of the corresponding styrene
oxides 4a-c with hydrazine hydrate, resulting in a mixture of
the regioisomeric products 5 and 6 in a ratio of ca. 2:1
(Scheme 2).37 For the unsubstituted compounds 5a and 6a,
both isomers could be isolated by fractional crystallization of

Article
Journal of Medicinal Chemistry, 2010, Vol. 53, No. 17 6303
Table 1. Structures of the Synthesized Hydrazine Molecules and Commercially Available Phenelzine, and Their Measured Biological Potency,
Predicted pIC50 Values for Both VAP-1 and MAO as Well As the Selectivity of VAP-1 over MAO
VAP1
pIC50 ( pSEMa
pred. pIC50
MAO
pIC50 ( pSEMa
pred. pIC50
selectivity
IC50
c
b
b
d
molecule
phenelzineg
R
R1
R2
R3
R4
pred. IC50
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
7.70 ( 0.22
7.38
6.15 ( 0.06
6.11
40
20
3a
3b
3c
3d
3e
3f
H
CH3
CH3
CH3
CH3
CH3
CH3
CH3
CH3
CH3
CH3
CH3
CH3
CH3
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
6.96 ( 0.04
6.70
5.27 ( 0.06
4.58
50
130
40
2-F
4-F
2-Cl
3-Cl
4-Cl
7.04 ( 0.05
7.10
5.41 ( 0.07
5.74
23
21
6.96 ( 0.12
7.13
5.64 ( 0.11
5.70
27
22
6.77 ( 0.08
6.94
5.43 ( 0.14
5.51
27
28
6.92 ( 0.11
6.92
5.48 ( 0.08
5.54
24
50
7.15 ( 0.06
7.09
5.50 ( 0.09
5.74
22
17
3g
3h
3i
2-OCH3
6.46 ( 0.05
6.48
5.22 ( 0.12
5.11
23
20
3-OCH3
7.05 ( 0.05
6.95
5.70 ( 0.22
5.34
40
40
4-OCH3
7.22 ( 0.07
7.15
5.68 ( 0.17
5.36
60
11
3j
2,5-(OCH3)2
3,4-(OCH3)2
2,3,4-(OCH3)3
3,4,5-(OCH3)3
H
6.43 ( 0.06
6.23
5.41 ( 0.07
5.19
11
18
3k
3l
6.46 ( 0.06
6.70
5.20 ( 0.07
5.22
30
33
6.48 ( 0.07
6.37
4.96 ( 0.16
4.92
28
72
3m
3n
3o
3p
3q
3r
3s
5a
5b
5c
9a
9b
9c
9d
9e
9f
6.49 ( 0.05
6.30
4.64 ( 0.09
5.08
17
10
CH2CH3
6.74 ( 0.02
6.68
5.74 ( 0.04
5.48
16
40
4-OCH3
CH2CH3
7.10 ( 0.11
6.87
5.46 ( 0.16
5.55
20
6
3,4,5-(OCH3)3
H
CH2CH3
5.65 ( 0.09
5.78
4.89 ( 0.07
5.17
4
CH2CH(CH3)2
CH2CH(CH3)2
6.36 ( 0.03
6.65
5.47 ( 0.02
5.62
8
11
4-OCH3
6.59 ( 0.05
6.67
5.57 ( 0.31
5.55
10
13
6.62 ( 0.14
6.59
5.41 ( 0.16
5.59
16
10
H
H
OH
OH
OH
OH
OH
OH
OH
OH
OH
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
7.19 ( 0.01
7.18
5.05 ( 0.05
5.27
140
80
3-OCH3
4-OCH3
H
H
7.28 ( 0.08
7.28
5.04 ( 0.02
5.17
200
130
200
120
190
140
85
H
7.54 ( 0.15
7.28
5.22 ( 0.00
5.19
CH3
CH3
CH3
CH3
CH3
CH3
6.68 ( 0.04
6.72
4.40 ( 0.01
4.57
2-Cl
6.57 ( 0.08
6.69
4.64 ( 0.02
4.63
120
68
4-Cl
6.51 ( 0.04
6.64
4.68 ( 0.02
4.61
110
160
150
120
180
120
140
3-OCH3
4-OCH3
4-F
6.55 ( 0.05
6.69
4.36 ( 0.02
4.51
6.51 ( 0.03
6.74
4.44 ( 0.05
4.49
6.49 ( 0.07
6.72
4.42 ( 0.01
4.59

6304 Journal of Medicinal Chemistry, 2010, Vol. 53, No. 17
Table 1. Continued
Nurminen et al.
VAP1
pIC50 ( pSEMa
pred. pIC50
MAO
pIC50 ( pSEMa
pred. pIC50
selectivity
IC50
c
b
b
d
molecule
9g
R
R1
R2
OH
R3
R4
pred. IC50
2,3,4-(OCH3)3
CH3
CH3
CH3
H
H
H
H
H
6.66 ( 0.04
6.58
4.35 ( 0.02
4.52
200
120
140
240
53
9h
3,4-[O(CH2)2O]
OH
H
H
H
H
H
H
CH3
H
H
H
H
H
H
H
H
H
H
6.70 ( 0.07
6.70
4.55 ( 0.03
4.32
9i
2-naphthyle
OH
6.37 ( 0.02
6.31
4.64 ( 0.04
4.65
46
31
9j
H
CH2CH(CH3)2
CH3
OH
6.46 ( 0.04
6.39
4.96 ( 0.08
4.91
30
10af
10bf
10cf
10d
12a
12b
12c
12d
12e
12f
12g
12h
12i
H
OH
CH3
6.46 ( 0.14
6.29
4.41 ( 0.02
4.50
110
62
H
CH3
OH
CH2CH3
6.59 ( 0.10
6.35
4.39 ( 0.03
4.19
160
150
110
60
H
CH3
OH
(CH2)2CH3
6.59 ( 0.07
6.47
4.55 ( 0.02
4.70
H
CH3
OH
CH3
H
6.25 ( 0.05
6.70
4.26 ( 0.02
4.65
100
110
47
3-F
CH3
OCH3
OCH3
OCH3
OCH3
OCH3
OCH3
OCH3
OCH3
OCH3
OCH3
6.36 ( 0.02
6.58
4.69 ( 0.02
4.56
110
53
4-F
CH3
H
6.24 ( 0.04
6.21
4.51 ( 0.01
4.35
72
41
3-Cl
CH3
H
6.39 ( 0.04
6.39
4.77 ( 0.10
4.41
95
44
4-Cl
CH3
H
6.37 ( 0.04
6.16
4.72 ( 0.04
4.26
80
3-CH3
3-OCH3
4-OCH3
2,3,4-(OCH3)3
1-naphthyle
2-naphthyle
CH3
H
6.49 ( 0.04
6.35
4.47 ( 0.08
4.65
110
50
CH3
H
6.39 ( 0.01
6.45
4.36 ( 0.01
4.63
110
70
CH3
H
6.38 ( 0.03
6.35
4.36 ( 0.01
4.61
110
60
CH3
H
6.20 ( 0.03
6.29
4.15 ( 0.01
4.68
110
40
CH3
H
6.40 ( 0.03
6.34
4.57 ( 0.16
4.58
68
58
12j
CH3
H
6.64 ( 0.04
6.62
4.88 ( 0.07
4.67
57
89
a Experimental pIC50 data in M concentration. The pIC50-errors (for 2-4 similar experiments) are given as a relative error (SEM/IC50)*0.434 that
results significant progressive error with large IC50-errors. b Predicted pIC50 values (final 3D QSAR models; see Table 2). c Selectivity of molecules from
IC50-data MAOover VAP-1[IC50 (MAO)]/[IC50 (VAP-1)]. d Predicted selectivity calculated from the pIC50-values 10[predicted pIC50 (VAP-1)-predicted pIC50 (MAO)]
.
e 1-Naphtyl or 2-naphtyl group instead of phenyl group, does not refer to R. f Relative configuration 1R*,2S*; 1R*,2S*; 1R*,2S*. g Commercially
available.
Scheme 1a
Scheme 3a
a Reagents: (a) NaNO2, AcOH, H2O; (b) LiAlH4, THF (for the
meaning of R, R1, and n see Table 1).
Scheme 2a
a Reagents: (a) NaNO2, AcOH, H2O; (b) LiAlH4, THF; (c) NaH,
MeI, THF (for the meaning of R, R1, R3, and R4 see Table 2).
a Reagents: (a) H2NNH2 H2O, EtOH.
3
this study seem to use this additionally available space.
However, the site could be very important for VAP-1
selectivity over total MAOs and should be taken into account
when new VAP-1 selective molecules are designed. None of the

Article
Journal of Medicinal Chemistry, 2010, Vol. 53, No. 17 6305
Figure 2. Flexibility of VAP-1 ligand binding site (A) Leu469 acts as a gate residue that blocks the entry into substrate binding site; (B) in the
crystal structure of VAP-1 with bound 2-hydrazinopyridine amino acids Met211 and Ser496 form a wall to the substrate binding site that
separates the substrate binding site and much larger side pocket. (C) Ligands can enter into substrate binding site, when the gate residue Leu469
takes another conformation. (D) During the MD simulation of VAP-1 with bound 5a the gate residue Leu469 is held “open” and Met211 and
Ser496 rotate and reveal a direct entrance from substrate binding site into side pocket (marked with X). In (A) and (C) the entrance into
substrate binding site is pointed with a yellow arrow. In (D) the starting structure (light green carbon atoms) and the key conformational
changes (pink carbon atoms) are shown.
Table 2. Statistics of 3D QSAR Models
0
b
pred
c
yy
d
bs
model (protein)
q2
r2
SEEa
no. of components
q2
q**2c
cSDEPc
dq2 /dr2
q2
LOO
submodel 1 (VAP-1)
submodel 2 (VAP-1)
submodel 3 (VAP-1)
submodel 4 (VAP-1)
final model (VAP-1)
submodel 1 (MAO)
submodel 2 (MAO)
submodel 3 (MAO)
submodel 4 (MAO)
final model (MAO)
0.602
0.522
0.623
0.601
0.636
0.676
0.717
0.776
0.673
0.749
0.804
0.855
0.881
0.861
0.828
0.838
0.843
0.866
0.824
0.840
0.150
0.162
0.141
0.157
0.163
0.208
0.205
0.200
0.223
0.209
3
4
4
4
4
3
3
2
3
3
0.63
0.66
0.50
0.58
0.5457
0.5661
0.2664
0.3612
0.3373
0.9379
0.868
0.880
0.79
0.63
0.53
0.88
a Standard error of estimate. b q2pred value acquired from the prediction of the molecules in test set (Table S1, column Set). c Model validation with
progressive scrambling-method. d Model validation with bootstrapping-method.
other amino acids of the VAP-1 ligand-binding pocket undergo
any noteworthy conformational changes. Taken together, these
simulations clearly reveal that ligands having considerably
larger substituents at position R1 than hydrogen, as seen in 5a
(Table 1), could also bind to VAP-1. This information is
valuable for designing more selective and better inhibitors for
VAP-1.
3D QSAR Model Building. To fully understand the rela-
tionship between the ligand structure and its associated
biological activity, we used the 3D QSAR technique to build
statistical models that could explain and predict the binding
affinities for synthesized hydrazines. The superpositioning of
the ligands for 3D QSAR was obtained by using molecular
docking into the crystal structure of VAP-1,11 where the
conformation of Leu469 was changed to resemble that seen
in MD simulations. The representative conformation for
each ligand was selected based on the fitness function
implemented in the docking software. The molecules were
divided into four subsets, out of which three were used to
build the model and one was used as a test set (see Supporting
Information: Table S1, column: set), in order to validate the
built models for a full ligand set (four submodels, Table 2).
These validation sets resulted in a final model that predicted
the potency of test set molecules with reasonable accuracy
(Figure 3; Table 2: q2pred). All 47 molecules were also
predicted using final models (Table 2). All 47 molecules fit
into the final 3D QSAR models; that is, no outliers were
detected (Figure 3). Accordingly, this model should yield
predictions with even higher reliability than those of the
submodels. The final comparative molecular similarity in-
dices analysis (CoMSIA) models were built by using leave
one out (LOO) cross-validation (Table 2). The QSAR model
for total MAO was constructed using the ligand superposi-
tion obtained from the docking simulations to VAP-1.
Similarly as for VAP-1, four submodels and a final model
were also built for MAO (Table 2). Accordingly, single-
ligand superpositioning can be used to obtain meaningful
predictions for both VAP-1 and MAOs. For the VAP-1 3D
QSAR model, the correlation coefficients q2LOO, r2 0.636,
and 0.828, respectively, were obtained by means of a steric
CoMSIA field (Table 2). Four partial least-squares (PLS)
components were used and the standard error of estimate
(SEE) value was 0.163 (Table 2). For the MAO 3D QSAR
model, the best correlation was found with the hydrophobic

6306 Journal of Medicinal Chemistry, 2010, Vol. 53, No. 17
Nurminen et al.
covalent mechanism of inhibition, IC50-values include asso-
ciation, reactivity, and dissociation components. In the case
of MAO, the hydrazine molecules react with flavin adenine
dinucleotide34 most likely via the mechanism where dissocia-
tion of the N(R1)NH2 moiety leads to the same covalently
bound products if the remaining substitution patterns are
identical (e.g., in the case of phenelzine, 3a, 3n, and 3q).
However, in case of VAP-1 the hydrazine moiety remains in
the covalently bound product.10 These different components
of binding affinity should also affect the QSAR models.
Accordingly, reaction ability especially influences the mole-
cular fields near the hydrazine-moiety in both VAP-1 and
MAO. For example, there are unfavorable fields in VAP-1
(steric: Figure 4A, yellow) and total MAO (hydrophobic:
Figure 4B, red) 3D QSAR models adjacent to the hydrazine
moiety, which implies that substituents larger than hydrogen
atoms at position R1 would be disadvantageous for both
VAP-1 and MAO inhibition (Table 1). This can also be seen
from the experimental data when, for example, phenelzine,
3a, 3n, and 3q, or 5a, 9a, and 9j are compared (Table 1).
However, if the size of the R1-groups exceeds that of the
methyl-group, the selectivity toward VAP-1 against MAO is
decreased dramatically (e.g., compare compounds 3a and 3n,
3q, and 9a and 9j: Table 1).
The hydroxyl group at position R2 decreases both VAP-1
and MAO potency (compare e.g., phenelzine-5a, 3a-9a,
3d-9b, 3f-9c, 3h-9d, 3i-9e) except in compounds where
the aryl group has a fluoro group at the para-position
(9f-12b), or there is 2-naphtyl in place of the benzene ring
(9i-12j). Also, when the aryl group is trisubstituated
(2,3,4-(OCH3)3, 3l, 9g), the addition of the hydroxyl group
decreases MAO potency but increases slightly VAP-1 po-
tency. In those molecules, where the addition of the hydroxyl
group at position R2 decreases both VAP-1 and MAO
potency, the influence is more prominent with MAO and,
consequently, the VAP-1 selectivity over MAO is clearly
improved. Actually, the molecules containing a hydroxyl-
group at position R2 are the most selective inhibitors in the
molecule set (Table 1). Modification of the R2 hydroxyl-
group to a methoxy-group does not have any clear influence
on either VAP-1 or MAO potency (Table 1).
Figure 3. Correlation between the experimental and predictions of
submodels -log IC50. Blue squares represent VAP-1 correlation,
red spots MAO correlation.
field (q2LOO: 0.749; r2: 0.840), where three PLS components
was used, and the SEE was 0.209 (Table 2). With neither of
the enzymes, other CoMSIA fields did not have as good
correlations with the experimental data. For example, the
hydrophobic field for VAP-1 and the steric field for total
MAO also correlated (see Supporting Information Table
S2), but especially q2
values were lower than the steric
LOO
field for VAP-1 and the hydrophobic field for total MAO,
respectively (Table 2). Combining the steric and electrostatic
field, in turn, correlated for both VAP-1 and MAO, but with
an outlier (see Supporting Information in Table S2 and
Figure S1).
Numerically both of the developed 3D QSAR models
indicate that the IC50-values for similar molecules can be
predicted with reasonable accuracy, and they should be able
to separate the active inhibitors from inactive molecules. For
example, the predictions made using submodels (Figure 3)
show that the predicted potencies are clearly within one log-
unit of those experimentally measured. In practice, for both
VAP-1 and total MAO, the range of inhibitor potencies is
fairly narrow to build a quantitative QSAR model: approxi-
mately two logarithmic units. Thus, the inhibitor potency
predictions cannot be reliably made outside this potency
range. Furthermore, because of the redundancy in the data
set, further validation of the models was done by employing
progressive scrambling45 and bootstrapping methods.46,47
Addition of substituents to positions R3 and R4 (compare
9a and 10a-d) has only a modest effect on MAO inhibitory
potency but slightly decreases VAP-1 potency, which clearly
decreases VAP-1 selectivity over MAO, especially when both
R3 and R4 positions are substituted (10d: Table 1).
The second favorable steric field (Figure 4A,C, green II)
shows that the addition of R-substituents to the para-
position (R-group, Table 1) of the phenyl ring improves the
binding to VAP-1. This is logical, as this area is toward the
VAP-1 surface. This is consistent with the inhibition data as
well. For example, with the addition of the methoxy-group at
the para-position of compounds 3a and 5a, resulting in
compounds 3i and 5c, respectively, the inhibition of VAP-1
is slightly increased. However, if the R1-substituent is the
methyl-group instead of the hydrogen atom, the VAP-1
inhibition ability is decreased when a methoxy-group is
added to the para-position of the phenyl ring (compare 9a
and 9e, Table 1). The influence of the para-substitutions to
selectivity is complex; when R2 is the hydroxyl-group rather
than the hydrogen atom, the VAP-1 selectivity over MAO is
increased (3a-3i, 5a-5c).
These data show that the models are resilient to chance
(Table 2: dq2 /dr2yy-values are lower than one), as well as
being predictive (Table 2).
0
Interpretation of Molecular Fields Using in Vitro Data. The
advantage of using enzyme structure-based inhibitor overlay
in 3D QSAR rather than superposition based on ligand
structures alone is that the molecular fields can be visualized
within the inhibitor binding site. This allows for the evalua-
tion of the reasonableness of the developed 3D QSAR model
by analyzing the possible interaction sites between the in-
hibitors and the amino acids at the binding cavity. As the
VAP-1 data correlates with the steric CoMSIA field, the
areas of favorable steric fields have space that could be filled
by the ligand’s substituents. According to the proposed
In addition to the favorable steric field near the phenyl
group, two sterically unfavorable fields are also found in the

Article
Journal of Medicinal Chemistry, 2010, Vol. 53, No. 17 6307
Figure 4. CoMSIA fields and the reference molecule (phenelzine, Table 1). (A) CoMSIA fields for VAP-1. Sterically favorable areas are green
and unfavorable areas yellow and they are shown with contribution levels of 85% and 25%, respectively. (B) CoMSIA fields for MAO.
Hydrophobically favorable areas are blue and unfavorable areas are red, and they are shown with contribution levels of 80% and 15%,
respectively. In (C), the parallel stereoview of the same spatial distribution of the CoMSIA fields in the ligand binding pocket.
VAP-1 3D QSAR model (Figure 4A,C: small yellow areas).
The clearest effect is seen with the compound 9i, where the
phenyl group is replaced by 2-naphtyl and the VAP-1
inhibition potency is clearly decreased compared to that of
the parent compound 9a.
model building (Table 1). Specifically, we wanted to chal-
lenge our 3D QSAR models with molecules whose properties
vary significantly from those compounds that were used in
the model development. On the basis of predicted inhibitor
potency and selectivity toward VAP-1 over MAO, three
molecules were synthesized (Table 3: 13, 15, 16). Also, two
molecules with less selectivity were synthesized, since they
have novel substituents (Table 2: 14a, 14b). Compounds
13-16 were synthesized similarly to the analogous deriva-
tives 3, 9, and 12. In general, the developed molecules were
predicted to be slightly better inhibitors of VAP-1 than they
actually are, while the MAO potency was predicted quite
accurately (Table 3). First, we predicted a compound that
has a shorter linker between the hydrazine moiety and the
phenyl ring, that is, only one CH2-group in between (Table 3:
13). This was very challenging for the 3D QSAR models, as
they do not contain any molecules with such a short link.
This compound 13 was predicted to be highly active
and selective (pIC50-values for VAP-1: 7.70; MAO: 5.32,
Table 3); however, the measured IC50 values (pIC50-values
for VAP-1: 6.57; MAO: 5.74, Table 3) show that the models
cannot reliably predict this type of compound. Second, as the
VAP-1 3D QSAR model was built by using a steric field that
cannot discriminate between polar and hydrophobic groups,
we challenged the VAP-1 3D QSAR model by developing
molecules 14a and 14b, which have a methyl group at
position 2, instead of the hydroxyl group seen in their
counterpart compounds 5a and 9a, respectively. Logically,
the predicted pIC50 values for these two new compounds
should be more highly similar than for their counterparts
(14a: 7.41; 5a: 7.19; 14b: 6.67; 9a: 6.68, Tables 1 and 3). It is
noteworthy that the measured pIC50-values are also quite
similar to those predicted (14a: 7.23; 14b: 6.33 μM; Table 3).
When both the hydroxyl and methyl group were introduced
into position 2 of the ligand (15), both 3D QSAR models
Overall, the fields from the 3D QSAR model of MAO
(Figure 4B) are obviously more difficult to rationalize, as the
model is based on ligand superposition obtained from a
docking experiment to the VAP-1 structure. However, this
ligand-based alignment delineates the effect of field variables
on potency and leads to a descriptive and predictive QSAR
model. In particular, the location of MAO fields around the
ligand can be used to analyze where hydrophobic groups
would be beneficial for MAO but not necessarily for VAP-1.
Thus, these MAO-favorable fields can be used to guide the
synthesis of VAP-1 inhibitors in a way that favorable sites for
MAO inhibition would be avoided, especially if they do not
simultaneously contribute toward VAP-1 binding. Compar-
ison of the favored and unfavorable fields of MAO and VAP-
1 shows an interesting area where the hydrophobically
favorable field of MAO and the sterically unfavorable field
of VAP-1 overlap, in the area between the ligand and Phe389
(Figure 4). This MAO-favoring region can also be seen by
comparing the selectivity of compound-series 9 and 12
(Table 1), where the methylation of hydroxyl-group (R2-
substitution) decreases the binding into VAP-1 (except in the
compound pair 9i and 12j, where the aryl group is 2-naphtyl),
while the MAO potency is generally not affected. Thus, the
selectivity is lowered 2-3 times (Table 1). Accordingly,
addition of a group at this position should be avoided, since
it reduces VAP-1 selectivity.
Prediction, Synthesis, and in Vitro Potency of Novel Li-
gands. To further validate the built 3D QSAR models, they
were used to predict the IC50 values for hydrazine molecules
that have novel features, in contrast to those used in the

6308 Journal of Medicinal Chemistry, 2010, Vol. 53, No. 17
Nurminen et al.
Table 3. Structures of Five Novel Hydrazines and Their Predicted and Measured pIC50 Values for the Inhibition of VAP-1 and MAO
a pIC50 data for VAP-1 in M concentration, mean ( pSEM of 2-4 similar experiments [pSEM = (SEM/IC50)*0.434]. b pIC50 data for MAO in M
concentration, mean ( pSEM of 2-4 similar experiments.
slightly overestimated the binding (VAP-1:0.72 log-units;
MAO: 0.67 log-units: Table 3). This is quite a small error in
the context that none of the original compounds contained
two substituents in position 2. What is notable for future
ligand discovery is that compound 15 uses both sterically
favored regions in VAP-1 3D QSAR models (fields shown
in Figure 4A), which also resulted in high selectivity of this
compound (2.5 log-units, i.e., >300-fold selectivity to
VAP-1 over MAO, Table 3). Finally, we introduced a
clearly larger substituent into ligand position 2 (16;
Table 3). Again, the 3D QSAR models should not be able
to predict this compound accurately, as none of the com-
pounds used in the 3D QSAR model building had such a
large substituent in position 2. However, the predictions
were surprisingly good, and also this compound shows
selectivity toward VAP-1 against MAO (1.8 log-units:
Table 3).
The models obtained in this study are excellent tools to
guide future inhibitor discovery toward more potent and
selective ligands for VAP-1. The statistics (Table 2) suggest
that when using 3D QSAR models to predict potencies for
unmeasured molecules, these models can sufficiently pre-
dict IC50 values. The molecules that have functional
groups at the positions that have not been varied within
this ligand series are beyond the limits of these models and
cannot therefore be predicted reliably. However, as shown
for five molecules that do not fall into same ligand set, the
predictions can be quite accurate (Table 3). In addition to
direct usage of these 3D QSAR models in the prediction of
new molecules, the developed 3D QSAR model can in-
dicate which properties are favorable for the design of
ligands, no matter whether they are similar to these
covalently binding hydrazines, or designed to bind rever-
sibly. The binding mode of possible reversible ligands is
not yet known, so the descriptor fields near TPQ should be
ignored, since the orientation of TPQ could be either
oxidized (inactive, bound to the copper ion11) or reduced
(active,10 also used in this study).
The model presented here is the first published 3D QSAR
model for VAP-1 inhibitors, although structure-activity
relationship studies for SSAO substrates43,44 and 3D QSAR
study for MAO A inhibitors32 have been published previously.

Article
Journal of Medicinal Chemistry, 2010, Vol. 53, No. 17 6309
Conclusions
3.06 (s, 3H, NCH3), 3.11-3.18 (m, 2H, ArCH2), 3.47-
3.53 (m, 2H, NCH2), 7.15-7.25 (m, 2H, C6H4), 7.34-7.41 (m,
2H, C6H4). Anal. Calcd for C9H14ClFN2: C, 52.81; H, 6.89; N,
13.69. Found: C, 52.54; H, 6.61; N, 13.48.
1-[2-(4-Fluorophenyl)ethyl]-1-methylhydrazine Hydrogen Maleate
(3c Hydrogen Maleate). White crystalline solid, yield 1.96 g (69%);
mp 82-84 °C (EtOH-Et2O); 1H NMR (400 MHz, D2O) δ 3.04-
3.12 (m, 5H, NCH3, ArCH2), 3.47-3.54 (m, 2H, NCH2), 6.28 (s,
2H, CH=CH), 7.09-7.17 (m, 2H, C6H4), 7.29-7.37 (m, 2H,
C6H4). Anal. Calcd for C13H17FN2O4: C, 54.92; H, 6.03; N, 9.85.
Found: C, 54.66; H, 5.83; N, 9.72.
1-[2-(2-Chlorophenyl)ethyl]-1-methylhydrazine Hydrochloride
(3d Hydrochloride). Beige crystalline solid, yield 1.62 g (73%);
mp 97-99 °C (EtOH-Et2O); 1H NMR (400 MHz, D2O) δ 3.08
(s, 3H, NCH3), 3.20-3.26 (m, 2H, ArCH2), 3.45-3.52 (m, 2H,
NCH2), 7.31-7.44 (m, 3H, C6H4), 7.47-7.53 (m, 1H, C6H4).
Anal. Calcd for C9H14Cl2N2: C, 48.88; H, 6.38; N, 12.67. Found:
C, 48.60; H, 6.29; N, 12.55.
1-[2-(3-Chlorophenyl)ethyl]-1-methylhydrazine Hydrochloride
(3e Hydrochloride). Yellowish white crystalline solid, yield 1.65 g
(75%); mp 73-76 °C (EtOAc-Et2O); 1H NMR (400 MHz,
D2O) δ 3.02-3.13 (m, 5H, ArCH2, NCH3), 3.46-3.54 (m, 2H,
NCH2), 7.25-7.31 (m, 1H, C6H4), 7.34-7.43 (m, 3H, C6H4).
Anal. Calcd for C9H14Cl2N2: C, 48.88; H, 6.38; N, 12.67. Found:
C, 49.02; H, 6.25; N, 12.51.
A series of hydrazine compounds has been synthesized and
shown to selectively inhibit VAP-1 activity over MAO. The
built 3D QSAR model for VAP-1 is based on a rather large set
of biologically tested molecules (47 molecules) and can hence
be considered to be rather accurate for the prediction of the
properties required for selective and effective VAP-1 inhibi-
tors. Combination of the data from the 3D QSAR model with
the results obtained from MD simulations allows for the
design of even better inhibitors with different scaffolds.
However, when reversible inhibitors are developed, it must
bekept in mind thatthe TPQ can be in bothinactiveandactive
conformations, and accordingly the 3D QSAR fields near the
TPQ should be used with care.
Materials and Methods
Materials. Benzylamine, tyramine, 2,4-dichlorophenol, 4-amino-
antipyrine, and horseradish peroxidase were purchased from Sigma
(St. Louis, MO). Cell culture reagents were supplied by Gibco.
Other chemicals were of analytical or reagent grade, and were
purchased from commercial suppliers.
Methods. Synthesis. General Methods. Melting points were
measured on a Kofler hot-plate microscope apparatus and are
uncorrected. For routine thin-layer chromatography (TLC),
Silica gel 60 F254 plates (Merck, Germany) were used. Elemental
analyses were performed with a Perkin-Elmer 2400 CHNS
elemental analyzer. 1H NMR spectra were recorded in CDCl3,
in DMSO-d6 or in D2O solutions on a Bruker DRX 400 or DRX
500 spectrometer by using completely dissolved samples. These
spectra indicated the >95% purities of the prepared com-
pounds. Chemical shifts are given in δ (ppm) relative to TMS
(CDCl3, DMSO-d6) or to TSP (D2O) as internal standards.
General Procedure for the Preparation of N-Nitroso Amines 2
and Amino Alcohols 8. A solution of NaNO2 (1.38 g, 20 mmol) in
H2O (5 mL) was added dropwise to a suspension/emulsion/
solution of the corresponding amine (1a-s) or amino alcohol (7)
(10 mmol) in H2O (20 mL) with vigorous stirring on an ice-cold
bath, and then AcOH (0.90 g, 15 mmol) was added dropwise.
The mixture was stirred at room temperature for 5-10 h and
then was extracted with EtOAc (4 ꢀ 30 mL). The combined
organic phases were dried (Na2SO4) and evaporated under
reduced pressure to give the N-nitroso derivative which was
used in the next steps without further purification.
General Procedure for the Preparation of Hydrazines 3a-s, 13,
and 14b and Hydrazino Alcohols 9a-j and 10a-d and 15. A
solution of the corresponding N-nitroso compound (2 or 8,
10 mmol) in THF (10 mL) was added dropwise to a strirred and
ice-cooled suspension of LiAlH4 (0.76 g, 20 mmol) in THF
(25 mL), and the mixture was stirred at ambient temperature for
3 h. The excess of LiAlH4 was decomposed with a mixture of
H2O (1.5 mL) and THF (20 mL), and the resulting precipitate
was filtered off and washed with EtOAc (3 ꢀ 25 mL). The
combined filtrate and washings were dried (Na2SO4) and
evaporated under reduced pressure. The oily residue was
treated with an equivalent amount of maleic or fumaric acid,
or with an excess of 22% ethanolic HCl in a mixture of EtOH
and Et2O to give crystalline salts which was filtered off and
recrystallized.
1-[2-(4-Chlorophenyl)ethyl]-1-methylhydrazine Hydrochloride
(3f Hydrochloride). Pale yellow crystalline solid, yield 1.72 g
(78%); mp 96-98 °C (EtOH-Et2O); 1H NMR (400 MHz, D2O)
δ 3.02-3.12 (m, 5H, ArCH2, NCH3), 3.46-3.53 (m, 2H,
NCH2), 7.32 (d, 2H, J = 8.3 Hz, C6H4), 7.42 (d, 2H, J = 8.3
Hz, C6H4). Anal. Calcd for C9H14Cl2N2: C, 48.88; H, 6.38; N,
12.67. Found: C, 48.69; H, 6.46; N, 12.63.
1-[2-(2-Methoxyphenyl)ethyl]-1-methylhydrazine Hydrogen
Maleate (3g Hydrogen Maleate). Beige crystalline solid, yield
1.68 g (57%); mp 61-63 °C (EtOH-Et2O); 1H NMR (400
MHz, DMSO-d6) δ 2.50 (s, 3H, NCH3) 2.78-2.90 (m, 2H,
ArCH2), 2.99-3.12 (m, 2H, NCH2), 3.80 (s, 3H, OCH3), 6.03
(s, 2H, CH=CH), 6.87-7.03 (m, 2H, C6H4), 7.15-7.30 (m,
2H, C6H4). Anal. Calcd for C14H20N2O5: C, 56.75; H, 6.80; N,
9.45. Found: C, 56.62; H, 6.97; N, 9.41.
1-[2-(3-Methoxyphenyl)ethyl]-1-methylhydrazine Hydrogen
Maleate (3h Hydrogen Maleate). Beige crystalline solid, yield
1.84 g (62%); mp 88-90 °C (EtOH-Et2O); 1H NMR (400
MHz, D2O) δ 3.03-3.14 (m, 5H, NCH3, ArCH2), 3.48-3.57
(m, 2H, NCH2), 3.84 (s, 3H, OCH3), 6.28 (s, 2H, CH=CH),
6.93-7.02 (m, 3H, C6H4), 7.33-7.40 (m, 1H, C6H4). Anal.
Calcd for C14H20N2O5: C, 56.75; H, 6.80; N, 9.45. Found: C,
56.55; H, 6.69; N, 9.54.
1-[2-(4-Methoxyphenyl)ethyl]-1-methylhydrazine Hydrogen
Maleate (3i Hydrogen Maleate). White crystalline substance,
yield 1.90 g (64%); mp 113-115 °C (EtOH-Et2O); 1H NMR
(400 MHz, D2O) δ 3.00-3.08 (m, 5H, ArCH2, NCH3),
3.45-3.52 (m, 2H, NCH2), 3.82 (s, 3H, OCH3), 6.27 (s, 2H,
CH=CH), 6.99 (d, 2H, J = 8.6 Hz, C6H4), 7.28 (d, 2H, J =
8.6 Hz, C6H4). Anal. Calcd for C14H20N2O5: C, 56.75; H,
6.80; N, 9.45. Found: C, 56.49; H, 6.52; N, 9.30.
1-[2-(2,5-Dimethoxyphenyl)ethyl]-1-methylhydrazine Hydrogen
Maleate (3j Hydrogen Maleate). Pinkish white crystalline sub-
stance, yield1.60 g(49%);mp84-87 °C(EtOH-Et2O); 1HNMR
(400 MHz, D2O) δ 3.02-3.09 (m, 5H, NCH3, ArCH2), 3.43-3.49
(m, 2H, NCH2), 3.81 (s, 3H, OCH3), 3.85 (s, 3H, OCH3), 6.27 (s,
2H, CH=CH), 6.91-6.98 (m, 2H, C6H3), 7.04 (d, 1H, J = 8.9 Hz,
C6H3). Anal. Calcd for C15H22N2O6: C, 55.21; H, 6.79; N, 8.58.
Found: C, 55.03; H, 6.58; N, 8.41.
1-[2-(3,4-Dimethoxyphenyl)ethyl]-1-methylhydrazine Hydro-
chloride (3k Hydrochloride). White crystalline substance,
yield 1.95 g (79%); mp 127-130 °C (MeOH-Et2O); 1H
NMR (400 MHz, D2O) δ 3.02-3.08 (m, 5H, ArCH2, NCH3),
3.46-3.52 (m, 2H, NCH2), 3.86 (s, 3H, OCH3), 3.88 (s, 3H,
OCH3), 6.92-6.97 (m, 1H, C6H3), 6.99-7.07 (m, 2H, C6H3).
1-Methyl-1-(2-phenylethyl)hydrazine Hydrochloride (3a Hy-
drochloride). White crystalline solid, yield 1.18 g (63%); mp
93-94 °C (EtOH-Et2O); 1H NMR (400 MHz, D2O) δ 3.06 (s,
3H, NCH3), 3.07-3.14 (m, 2H, ArCH2), 3.48-3.55 (m, 2H,
NCH2), 7.31-7.46 (m, 5H, C6H5). Anal. Calcd for C9H15ClN2:
C, 57.90; H, 8.10; N, 15.04. Found: C, 57.63; H, 7.98; N, 14.88.
1-[2-(2-Fluorophenyl)ethyl]-1-methylhydrazine Hydrochloride
(3b Hydrochloride). White crystalline solid, yield 1.66 g (81%);
1
mp 138-140 °C (EtOH-Et2O); H NMR (400 MHz, D2O) δ

6310 Journal of Medicinal Chemistry, 2010, Vol. 53, No. 17
Nurminen et al.
Anal. Calcd for C11H19ClN2O2: C, 53.55; H, 7.76; N, 11.35.
Found: C, 53.31; H, 7.87; N, 11.18.
(dd, 1H, J = 3.4, 13.3 Hz, NCH2), 3.47 (dd, 1H, J = 9.9, 13.3
Hz, NCH2), 5.18 (dd, 1H, J = 3.4, 9.9 Hz, OCH), 6.30 (s, 2H,
CH=CH), 7.39-7.51 (m, 5H, C6H5). Anal. Calcd for C13H18-
N2O5: C, 55.31; H, 6.43; N, 9.92. Found: C, 55.26; H, 6.70; N,
9.87.
1-[2-(2,3,4-Trimethoxyphenyl)ethyl]-1-methylhydrazine Hydro-
chloride (3l Hydrochloride). White crystalline solid, yield 1.80 g
1
(65%); mp 129-131 °C (EtOH-Et2O); H NMR (400 MHz,
D2O) δ 3.01-3.09 (m, 5H, ArCH2, NCH3), 3.39-3.46 (m, 2H,
NCH2), 3.89 (s, 6H, 2 ꢀ OCH3), 3.93 (s, 3H, OCH3), 6.91 (d, 1H,
J = 8.6 Hz, C6H2), 7.08 (d, 1H, J = 8.6 Hz, C6H2). Anal. Calcd
for C12H21ClN2O3: C, 52.08; H, 7.65; N, 10.12. Found: C, 51.87;
H, 7.42; N, 9.94.
1-(2-Chlorophenyl)-2-(1-methylhydrazino)ethanol Hydrogen
Fumarate (9b Hydrogen Fumarate). White crystalline solid, yield
2.06 g (65%); mp 127-128 °C (MeOH-Et2O); 1H NMR (400
MHz, D2O) δ 3.07 (s, 3H, NCH3), 3.32 (dd, 1H, J = 9.9, 13.4
Hz, NCH2), 3.43 (dd, 1H, J = 2.8, 13.3 Hz, NCH2), 5.56 (dd,
1H, J = 2.6, 9.8 Hz, OCH), 6.62 (s, 2H, CH=CH), 7.32-7.45
(m, 3H, C6H4), 7.59-7.63 (m, 1H, C6H4). Anal. Calcd for
C13H17ClN2O5: C, 49.30; H, 5.41; N, 8.84. Found: C, 49.05;
H, 5.15; N, 8.65.
1-[2-(3,4,5-Trimethoxyphenyl)ethyl]-1-methylhydrazine Hydro-
gen Maleate (3m Hydrogen Maleate). White crystalline solid, yield
2.56 g (72%); mp 76-78 °C (EtOH-Et2O); 1H NMR (400 MHz,
D2O) δ 3.01-3.12 (m, 5H, ArCH2, NCH3), 3.46-3.55 (m, 2H,
NCH2), 3.77 (s, 3H, OCH3), 3.88 (s, 6H, 2 ꢀ OCH3), 6.27 (s, 2H,
CH=CH), 6.69 (s, 2H, C6H2). Anal. Calcd for C16H24N2O7: C,
53.92; H, 6.79; N, 7.86. Found: C, 53.99; H, 7.86; N, 7.73.
1-Ethyl-1-(2-phenylethyl)hydrazine Hydrogen Maleate (3n
Hydrogen Maleate). White crystalline solid, yield 1.70 g
(61%); mp 75-77 °C (EtOH-Et2O); 1H NMR (400 MHz,
D2O) δ 1.33 (t, 3H, J = 7.3 Hz, CH3), 3.08-3.16 (m, 2H,
ArCH2), 3.37 (q, 2H, J = 7.3 Hz, CH2CH3), 3.47-3.57 (m,
2H, NCH2), 6.29 (s, 2H, CH=CH), 7.32-7.45 (m, 5H, C6H5).
Anal. Calcd for C14H20N2O4: C, 59.99; H, 7.19; N, 9.99.
Found: C, 59.76; H, 6.93; N, 9.85.
1-Ethyl-1-[2-(4-methoxyphenyl)ethyl]hydrazine Hydrochloride
(3o Hydrochloride). Yellowish white crystalline solid, yield 1.50
g (65%); mp 132-134 °C (MeOH-Et2O); 1H NMR (400 MHz,
D2O) δ 1.33 (t, 3H, J = 7.3 Hz, CH2CH3), 3.03-3.11 (m, 2H,
ArCH2), 3.35 (q, 2H, J = 7.3 Hz, CH2CH3), 3.43-3.51 (m, 2H,
NCH2), 3.84 (s, 3H, OCH3), 7.02 (d, 2H, J = 8.8 Hz, C6H4), 7.31
(d, 2H, J = 8.8 Hz, C6H4). Anal. Calcd for C11H19ClN2O: C,
57.26; H, 8.30; N, 12.14. Found: C, 57.49; H, 8.16; N, 11.96.
1-Ethyl-1-[2-(3,4,5-trimethoxyphenyl)ethyl]hydrazine Hydro-
chloride (3p Hydrochloride). White crystalline substance, yield
2.02 g (69%); mp 189-192 °C (MeOH-Et2O); 1H NMR (400
MHz, D2O) δ 1.35 (t, 3H, J = 7.3 Hz, CH2CH3), 3.06-3.13 (m,
2H, ArCH2), 3.37 (q, 2H, J = 7.3 Hz, CH2CH3), 3.48-3.56 (m,
2H, NCH2), 3.79 (s, 3H, OCH3), 3.89 (s, 6H, 2 ꢀ OCH3), 6.73 (s,
2H, C6H2). Anal. Calcd for C13H23ClN2O3: C, 53.70; H, 7.97; N,
9.63. Found: C, 53.84; H, 8.08; N, 9.51.
1-(4-Chlorophenyl)-2-(1-methylhydrazino)ethanol (9c). White
crystalline solid, yield 1.44 g (72%); mp 86-88 °C (Et2O-
1
n-hexane); H NMR (400 MHz, CDCl3) δ 2.48 (dd, 1H, J =
9.2, 12.8 Hz, NCH2), 2.56 (s, 3H, NCH3), 2.64 (dd, 1H, J = 2.3,
12.7 Hz, NCH2), 4.98 (dd, 1H, J = 2.3, 9.3 Hz, OCH), 7.28-
7.35 (m, 4H, C6H4). Anal. Calcd for C9H13ClN2O: C, 53.87; H,
6.53; N, 13.96. Found: C, 53.59; H, 6.37; N, 13.75.
1-(3-Methoxyphenyl)-2-(1-methylhydrazino)ethanol Hydrogen
Fumarate (9d Hydrogen Fumarate). White crystalline solid, yield
2.16 g (69%); mp 88-90 °C (EtOH-Et2O); 1H NMR (400 MHz,
D2O) δ 3.06 (s, 3H, NCH3), 3.36 (dd, 1H, J = 9.9, 13.4 Hz,
NCH2), 3.47 (dd, 1H, J = 10.1, 13.3 Hz, NCH2), 3.84 (s, 3H,
OCH3), 5.13 (dd, 1H, J = 3.2, 10.0 Hz, OCH), 6.66 (s, 2H,
CH=CH), 7.03 (d, 2H, J = 8.4 Hz, C6H4), 7.41 (d, 2H, J = 8.8
Hz, C6H4). Anal. Calcd for C14H20N2O6: C, 53.84; H, 6.45; N,
8.97. Found: C, 53.57; H, 6.25; N, 8.73.
1-(4-Methoxyphenyl)-2-(1-methylhydrazino)ethanol Hydrogen
Fumarate (9e Hydrogen Fumarate). White crystalline solid, yield
1
2.06 g (66%); mp 142-144 °C (EtOH-Et2O); H NMR (400
MHz, D2O) δ 3.09 (s, 3H, NCH3), 3.39 (dd, 1H, J = 3.4, 13.5 Hz,
NCH2), 3.50 (dd, 1H, J = 10.1, 13.4 Hz, NCH2), 3.86 (s, 3H,
OCH3), 5.16 (dd, 1H, J = 3.3, 10.1 Hz, OCH), 6.68 (s, 2H,
CH=CH), 7.06 (d, 2H, J = 8.8 Hz, C6H4), 7.41 (d, 2H, J = 8.8
Hz, C6H4). Anal. Calcd for C14H20N2O6: C, 53.84; H, 6.45; N,
8.97. Found: C, 53.59; H, 6.21; N, 8.70.
1-(4-Fluorophenyl)-2-(1-methylhydrazino)ethanol Hydrogen
Maleate (9f Hydrogen Maleate). White crystalline solid, yield
1.89 g (63%); mp 101-104 °C (MeOH-Et2O); 1H NMR (400
MHz, D2O) δ 3.05 (s, 3H, NCH3), 3.36 (dd, 1H, J = 3.4, 13.4
Hz, NCH2), 3.44 (dd, 1H, J = 10.1, 13.4 Hz, NCH2), 5.16 (dd,
1H, J = 3.4, 9.9 Hz, OCH), 6.28 (s, 2H, CH=CH), 7.16 (dd, 2H,
J = 8.8, 9.0 Hz, C6H4), 7.44 (dd, 2H, J = 5.5, 8.5 Hz, C6H4).
Anal. Calcd for C13H17FN2O5: C, 52.00; H, 5.71; N, 9.33.
Found: C, 52.24; H, 5.97; N, 9.58.
2-(10-Methylhydrazino)-1-(2,3,4-trimethoxyphenyl)ethanol
Hydrogen Fumarate (9g Hydrogen Fumarate). White crystal-
line solid, yield 2.42 g (65%); mp 139-141 °C (MeOH-Et2O); 1H
NMR (500 MHz, D2O) δ 3.06 (s, 3H, NCH3), 3.26-3.45 (m, 2H,
NCH2), 3.86 (s, 3H, OCH3), 3.88 (s, 3H, OCH3), 3.92 (s, 3H,
OCH3), 5.36 (d, 1H, J = 8.4 Hz, OCH), 6.70 (s, 2H, CH=CH),
6.95 (d, 1H, J = 8.7 Hz, C6H2), 7.22 (d, 1H, J = 8.7, C6H2). Anal.
Calcd for C16H24N2O8: C, 51.61; H, 6.50; N, 7.52. Found: C,
51.46; H, 6.43; N, 7.40.
1-(2,3-Dihydrobenzo[b][1,4]dioxin-6-yl)-2-(1-methylhydrazino)-
ethanol Hydrogen Maleate (9h Hydrogen Maleate). White crystal-
line solid, yield 2.48 g (73%); mp 127-129 °C (MeOH-Et2O); 1H
NMR (400 MHz, CDCl3) δ 3.05 (s, 3H, NCH3), 3.30-3.45 (m,
2H, NCH2), 4.30 (s, 4H, OCH2CH2O), 5.07 (dd, 1H, J = 2.9, 9.8
Hz, OCH), 6.29 (s, 2H, CH=CH), 6.91-6.98 (m, 3H, C6H3).
Anal. Calcd for C15H20N2O7: C, 52.94; H, 5.92; N, 8.23. Found:
C, 52.66; H, 5.63; N, 8.00.
2-(1-Methylhydrazino)-1-(2-naphthyl)ethanol Hydrogen Maleate
(9i Hydrogen Maleate). Yield 2.26 g (68%); mp 143-145 °C
(MeOH-Et2O); 1H NMR (400 MHz, CDCl3) δ 3.11 (s, 3H,
NCH3), 3.35-3.59 (m, 2H, NCH2), 5.35 (dd, 1H, J = 2.9, 9.9
Hz, OCH), 6.19 (s, 2H, CH=CH), 7.53-7.62 (m, 3H, C10H7),
1-Isobutyl-1-(2-phenylethyl)hydrazine Hydrogen Maleate (3q
Hydrogen Maleate). White crystalline solid, yield 1.72 g (56%);
mp 96-97 °C (EtOH-Et2O); 1H NMR (400 MHz, D2O) δ 0.99
(d, 6H, J = 6.7 Hz, CH(CH3)2), 2.10-2.22 (m, 1H CH(CH3)2),
3.08-3.17 (m, 4H, ArCH2, NCH2), 3.45-3.54 (m, 2H, NCH2),
6.29 (s, 2H, CH=CH), 7.31-7.45 (m, 5H, C6H5). Anal. Calcd
for C16H24N2O4: C, 62.32; H, 7.84; N, 9.08. Found: C, 62.05; H,
7.58; N, 8.86.
1-Isobutyl-1-[2-(4-methoxyphenyl)ethyl]hydrazine Hydrogen
Maleate (3r Hydrogen Maleate). Yellowish white crystalline
1
solid, yield 2.15 g (64%); mp 109-111 °C (MeOH-Et2O); H
NMR (400 MHz, D2O) δ 0.99 (d, 6H, J = 6.6 Hz, CH(CH3)2),
2.09-2.22 (m, 1H CH(CH3)2), 3.02-3.16 (m, 4H, ArCH2,
NCH2), 3.41-3.51 (m, 2H, NCH2), 3.82 (s, 3H, OCH3), 6.27
(s, 2H, CH=CH), 6.99 (d, 2H, J = 8.6 Hz, C6H4), 7.28 (d, 2H,
J = 8.6 Hz, C6H4). Anal. Calcd for C17H26N2O5: C, 60.34; H,
7.74; N, 8.28. Found: C, 60.09; H, 7.63; N, 8.16.
1-Methyl-1-(3-phenylpropyl)hydrazine Hydrogen Maleate (3s
Hydrogen Maleate). White crystalline solid, yield 2.10 g (75%);
mp 71-72 °C (EtOH-Et2O); 1H NMR (400 MHz, D2O) δ
2.03-2.14 (m, 2H, CH2CH2CH2), 2.75 (t, 2H, J = 7.5 Hz,
CH2CH2NCH3), 3.02 (s, 3H, NCH3), 3.23-3.29 (m, 2H, CH2-
CH2C6H4), 6.31 (s, 2H, CH=CH), 7.29-7.45 (m, 5H, C6H5).
Anal. Calcd for C14H20N2O4: C, 59.99; H, 7.19; N, 9.99. Found:
C, 60.21; H, 7.35; N, 10.15.
2-(1-Methylhydrazino)-1-phenylethanol Hydrogen Maleate
Hemiethanolate (9a Hydrogen Maleate Hemiethanolate). White
crystalline substance, yield 1.52 g (54%); mp 63-66 °C (EtOH-
1
Et2O); H NMR (400 MHz, D2O) δ 3.06 (s, 3H, NCH3), 3.40

Article
Journal of Medicinal Chemistry, 2010, Vol. 53, No. 17 6311
7.91-7.99 (m, 4H, C10H7). Anal. Calcd for C17H20N2O5: C, 61.44;
H, 6.07; N, 8.43. Found: C, 61.71; H, 5.90; N, 8.23.
mixture was kept at 60 °C for 10 min. The solvent was evapo-
rated off and the oily residue was dissolved in EtOH and treated
with an equivalent amount of maleic acid. The separated
crystals of the corresponding 5a-c maleate were filtered off
and recrystallized from EtOH. In case of compounds a, the
filtrate was treated with Et2O to give crystalline 6a maleate
which was filtered off and recrystallized twice from a mixture of
EtOH and Et2O.
2-Hydrazino-1-phenylethanol Hydrogen Maleate (5a Hydro-
gen Maleate). Beige crystalline solid, yield 3.22 g (48%); mp
120-122 °C (EtOH-Et2O); 1H NMR (D2O) δ 3.38 (d, 2H, J =
7.9 Hz, NCH2), 5.07 (t, 1H, J = 7.0 Hz, OCH), 6.29 (s, 2H,
CH=CH), 7.39-7.52 (m, 5H, C6H5). Anal. Calcd for C12H16-
N2O5: C, 53.73; H, 6.01; N, 10.44. Found: C, 53.50; H, 5.83; N,
10.28.
2-(1-Isobutylhydrazino)-1-phenylethanol Hydrogen Fumarate
(9j Hydrogen Fumarate). White crystalline solid, yield 1.91 g
(59%); mp 90-92 °C (EtOH-Et2O); 1H NMR (400 MHz,
CDCl3) δ 3.11 (d, 6H, J = 6.5 Hz, CH(CH3)2), 1.98-2.13
(m, 1H, CH(CH3)2), 2.98-3.92 (m, 2H, NCH2), 3.30-3.48
(m, 2H, NCH2) 5.24 (dd, 1H, J = 3.8, 9.1 Hz, OCH), 6.69
(s, 2H, CH=CH), 7.42-7.53 (m, 5H, C6H5). Anal. Calcd for
C16H24N2O5: C, 59.24; H, 7.46; N, 8.64. Found: C, 59.32; H,
7.67; N, 8.85.
(1R*,2S*)-2-(1-Methylhydrazino)-1-phenylpropan-1-ol Hydro-
chloride (10a Hydrochloride). White crystalline solid, yield 1.63 g
1
(75%); mp 145-146 °C (EtOH-Et2O); H NMR (400 MHz,
D2O) δ 1.22 (d, 3H, J = 7.0 Hz, CHCH3), 2.99 (s, 3H, NCH3),
3.66-3.73 (m, 1H, CHCH3), 5.40-5.47 (m, 1H, CHOH), 7.41-
7.54 (m, 5H, C6H5). Anal. Calcd for C10H17ClN2O: C, 55.42; H,
7.91; N, 12.93. Found: C, 55.17; H, 7.65; N, 12.66.
2-Hydrazino-1-(3-methoxyphenyl)ethanol Hydrogen Maleate
(5b Hydrogen Maleate). Pale yellow crystalline solid, yield 3.80 g
1
(51%); mp 130-133 °C (EtOAc-MeOH); H NMR (D2O) δ
3.38 (d, 2H, J = 6.5 Hz, NCH2), 3.87 (s, 3H, OCH3), 5.07 (t, 1H,
J = 6.5 Hz, OCH), 6.29 (s, 2H, CH=CH), 6.99-7.10 (m, 3H,
C6H4), 7.42 (t, 1H, J = 7.9 Hz, C6H4). Anal. Calcd for C13H18-
N2O6: C, 52.34; H, 6.08; N, 9.39. Found: C, 52.52; H, 6.21; N,
9.65.
2-Hydrazino-1-(4-methoxyphenyl)ethanol Hydrogen Maleate
(5c Hydrogen Maleate). Pale yellow crystalline solid, yield 3.51 g
(47%); mp 117-120 °C (EtOH-Et2O); 1H NMR (D2O) δ
3.32-3.45 (m, 2H, NCH2), 3.87 (s, 3H, OCH3), 5.04 (dd, 1H,
J = 4.4, 8.8 Hz, OCH), 6.32 (s, 2H, CH=CH), δ 7.06 (d, 2H,
J = 8.8 Hz, C6H4), 7.41 (d, 1H, J = 8.8 Hz, C6H4). Anal. Calcd
for C13H18N2O6: C, 52.34; H, 6.08; N, 9.39. Found: C, 52.50; H,
6.17; N, 9.69.
(1R*,2S*)-2-(1-Methylhydrazino)-1-phenylbutan-1-ol Hydro-
gen Fumarate (10b Hydrogen Fumarate). White crystalline solid,
yield 1.90 g (61%); mp 141-143 °C (EtOH-Et2O); H NMR
(400 MHz, DMSO-d6) δ 0.63 (t, 3H, J = 7.5 Hz, CH2CH3),
1.26-1.38 (m, 1H, CH2), 1.49-1.62 (m, 1H, CH2), 2.58-2.64 (m
4H, NCH3, ArCH), 5.15 (d, 1H, J = 2.1 Hz, NCH), 6.55 (s, 2H,
CH=CH), 7.17-7.38 (m, 5H, C6H5). Anal. Calcd for C15H22-
N2O5: C, 58.05; H, 7.15; N, 9.03. Found: C, 57.89; H, 7.02; N,
8.86.
1
(1R*,2S*)-2-(1-Methylhydrazino)-1-phenylpentan-1-ol Hydrogen
Fumarate (10c Hydrogen Fumarate). White crystalline solid, yield
1
2.23 g (69%); mp 132-134 °C (EtOH-Et2O); H NMR (400
MHz, DMSO-d6) 0.64(t, 3H, J=7.3Hz, CH2CH3), 0.79-0.93 (m,
1H, CH2), 1.07-1.31 (m, 2H, CH2), 1.46-1.57 (m, 1H, CH2), 2.60
(s, 3H, NCH3), 2.66-2.71 (m, 1H, ArCH), 5.15 (d, 1H, J = 2.1 Hz,
NCH), 6.57 (s, 2H, CH=CH), 7.16-7.37 (m, 5H, C6H5). Anal.
Calcd for C16H24N2O5: C, 59.24; H, 7.46; N, 8.64. Found: C, 59.03;
H, 7.28; N, 8.45.
2-Hydrazino-2-phenylethanol Hydrogen Maleate (6a Hydro-
gen Maleate). White crystalline solid, yield 0.80 g (12%); mp
1
107-110 °C (EtOH-Et2O); H NMR (D2O) δ 3.91 (2H, m,
OCH2), 4.32 (1H, m, NCH), 6.31 (2H, s, CH=CH), 7.47 (5H, m,
C6H5). Anal. Calcd for C12H16N2O5: C, 53.73; H, 6.01; N, 10.44.
Found: C, 53.48; H, 5.78; N, 10.57.
2,2-Dimethyl-2-(10-methylhydrazino)-1-phenylethanol Hydro-
gen Fumarate (10d Hydrogen Fumarate). White crystalline solid,
yield 2.10 g (68%); mp 184-186 °C (MeOH-Et2O); 1H NMR
(400 MHz, DMSO-d6) 1.04 (s, 6H, 2 ꢀ CCH3), 2.20 (s, 3H,
NCH3), 6.59 (s, 2H, CH=CH), 7.22-7.43 (m, 5H, C6H5). Anal.
Calcd for C15H22N2O5: C, 58.05; H, 7.15; N, 9.03. Found: C,
57.92; H, 6.94; N, 8.95.
General Procedure for the Preparation Hydrazino Alcohol
O-Methyl Ethers 12a-j and O-Benzyl Ether 16. 55% Sodium
hydride suspension (1.40 g, 32 mmol) was washed with n-hexane
and suspended in THF (35 mL). A solution of the corresponding
N-nitroso amino alcohol (8, 10 mmol) in THF (60 mL) was
degassed with N2 flushing and added dropwise to the NaH
suspension with stirring and continuous N2 flushing at 0 °C over
a period of 1 h. Stirring was continued at 0 °C for 2 h, and then
a solution of MeI (2.41 g, 17 mmol) or benzyl bromide (2.05 g,
12 mmol) in THF (20 mL) was added dropwise to the stirred
suspension at 0 °C. The mixture was stirred at 0 °C for 30 min
and then was allowed to warm to room temperature. The excess
of NaH was decomposed by addition of MeOH. The solution
evaporated to dryness, and the residue was dissolved in H2O
(40 mL) and extracted with EtOAc (3 ꢀ 40 mL). The combined
EtOAc extracts were washed with H2O (40 mL) and then dried
(Na2SO4) and evaporated under reduced pressure to give the
O-methyl-substituted product (11a-j) as a yellow oil.
The crude O-methyl-N-nitroso amino alcohol (11a-j) was
dissolved in THF (20 mL) and was added dropwise to a strirred
and ice-cooled suspension of LiAlH4 (1.26 g, 33.2 mmol) in THF
(60 mL). The mixture was stirred at 0 °C for 3 h, and then
allowed to warm to room temperature. The excess of LiAlH4
was decomposed with a mixture of H2O (2.5 mL) and THF (20
mL), and the resulting precipitate was filtered off and washed
with EtOAc (2 ꢀ 75 mL). The combined filtrates were dried
(Na2SO4) and evaporated under reduced pressure. The oily
residue was treated with an equivalent amount of maleic or
fumaric acid in a mixture of EtOH and Et2O to give crystalline
hydrogen maleate or hydrogen fumarate salt which was filtered
off and recrystallized.
1-Benzyl-1-methylhydrazine Hydrogen Maleate (13 Hydrogen
Maleate). White crystalline solid, yield 1.18 g (47%); mp 88-90
1
°C (MeOH-Et2O); H NMR (400 MHz, D2O) δ 2.99 (s, 3H,
CH3), 4.38 (s, 2H, CH2), 6.29 (s, 2H, CHdCH), 7.46-7.58 (m,
5H, C6H5). Anal. Calcd for C12H16N2O4: C, 57.13; H, 6.39; N,
11.10. Found: C, 57.01; H, 6.20; N, 10.98.
1-Methyl-1-(2-Phenylpropyl)hydrazine Hydrogen Fumarate
(14b Hydrogen Fumarate). White crystalline solid, yield 1.88 g
1
(67%); mp 143-146 °C (MeOH-Et2O); H NMR (400 MHz,
D2O) δ 1.31 (d, 3H, J = 7.0 Hz, CCH3), 2,96 (s, 3H, NCH3),
3.25-3.37 (m, 1H, CH), 3.44 (dd, 1H, J = 12.9, 5.9 Hz, CH2),
3.55 (dd, 1H, J = 12.9, 9.9 Hz, CH2), 6.67 (s, 2H, CH=CH),
7.34-7.49 (m, 5H, C6H5). Anal. Calcd for C14H20N2O4: C,
59.99; H, 7.19; N, 9.99. Found: C, 60.14; H, 7,01; N, 9.82.
2-(4-Methoxyphenyl)-1-(10-methylhydrazino)-2-propanol Hydro-
gen Maleate (15 Hydrogen Maleate). White crystalline solid, yield
1
2.02 g (62%); mp 127-129 °C (MeOH-Et2O); H NMR (400
MHz, D2O) δ 1.62 (s, 3H, CCH3), 2.77 (s, 3H, NCH3), 3.26-3.48
(m, 2H, CH2), 3.84 (s, 3H, OCH3), 6.30 (s, 2H, CH=CH), 7.05 (d,
2H, J=8.8Hz, C6H4), 7.47 (d, 2H, J=8.8Hz, C6H4). Anal. Calcd
for C15H22N2O6: C, 55.21; H, 6.79; N, 8.58. Found: C, 55.03; H,
6.68; N, 8.50.
General Procedure for the Preparation of Hydrazino Alcohols
5a-c and 6a. To a stirred solution of hydrazine hydrate (12.5 g,
0.25 mol) in EtOH (10 mL), the corresponding styrene oxide
(4a-c, 25 mmol) was added dropwise. The exothermic reaction
brought the mixture to reflux. After addition was complete, the
1-[2-(3-Fluorophenyl)-2-methoxyethyl]-1-methylhydrazine
Hydrogen Fumarate (12a Hydrogen Fumarate). White crystalline

6312 Journal of Medicinal Chemistry, 2010, Vol. 53, No. 17
Nurminen et al.
solid, yield 1.67 g (53%); mp 108-110 °C(EtOH-Et2O); 1HNMR
(400 MHz, D2O) δ 3.21 (s, 3H, NCH3), 3.47 (s, 3H, OCH3), 3.52
(dd, 1H, J = 3.4, 13.8 Hz, NCH2), 3.43 (dd, 1H, J = 10.1, 13.7 Hz,
NCH2), 4.96 (dd, 1H, J = 3.2, 10.3 Hz, OCH), 6.84 (s, 2H,
CH=CH), 7.30-7.45 (m, 3H, C6H4), 7.61-7.69 (m, 1H, C6H4).
Anal. Calcd for C14H19FN2O5: C, 53.50; H, 6.09; N, 8.91. Found:
C, 53.24; H, 5.95; N, 8.70.
1-[2-(4-Fluorophenyl)-2-methoxyethyl]-1-methylhydrazine
Hydrogen Fumarate (12b Hydrogen Fumarate). White crystalline
solid, yield 1.70 g (54%); mp 117-119 °C(EtOH-Et2O); 1HNMR
(500 MHz, D2O) δ3.02 (s, 3H, NCH3), 3.24 (s, 3H, OCH3), 3.30 (d,
1H, J = 13.8 Hz, NCH2), 3.44 (dd, 1H, J = 10.1, 13.4 Hz, NCH2),
4.75 (d, 1H, J = 9.8 Hz, OCH), 6.65 (s, 2H, CH=CH), 7.18 (dd,
2H, J = 7.4, 9.0 Hz C6H4), 7.41 (dd, 2H, J = 5.8, 7.8 Hz, C6H4).
Anal. Calcd for C14H19FN2O5: C, 53.50; H, 6.09; N, 8.91. Found:
C, 53.27; H, 5.91; N, 8.75.
1-(2-(3-Chlorophenyl)-2-methoxyethyl)-1-methylhydrazine
Hydrogen Fumarate (12c Hydrogen Fumarate). White crystalline
solid, yield 1.62 g (49%); mp 116-118 °C(EtOH-Et2O); 1HNMR
(400 MHz, D2O) δ 3.21 (s, 3H, NCH3), 3.47 (s, 3H, OCH3), 3.51
(dd, 1H, J = 3.1, 14.1 Hz, NCH2), 3.63 (dd, 1H, J = 10.1, 13.7 Hz,
NCH2), 4.95 (dd, 1H, J = 3.1, 10.1 Hz, OCH), 6.85 (s, 2H,
CH=CH), 7.51-7.55 (m, 1H, C6H4), 7.62-7.67 (m, 3H, C6H4).
Anal. Calcd for C14H19ClN2O5: C, 50.84; H, 5.79; N, 8.47. Found:
C, 51.00; H, 5.99; N, 8.70.
1-[2-(4-Chlorophenyl)-2-methoxyethyl]-1-methylhydrazine
Hydrogen Maleate (12d Hydrogen Maleate). Yield 1.95 g (59%);
mp 98-100 °C (EtOH-Et2O); 1H NMR (500 MHz, D2O) δ 2.98
(s, 3H, NCH3), 3.20-3.30 (m, 5H, OCH3, NCH2), 3.35-3.43 (m,
1H, NCH2), 4.72 (dd, 1H, J = 2.5, 10.1 Hz, OCH), 6.25 (s, 2H,
CH=CH), 7.34(d, 2H, J=8.3Hz, C6H4), 7.43 (d, 2H, J=8.3Hz,
C6H4). Anal. Calcd for C14H19ClN2O5: C, 50.84; H, 5.79; N, 8.47.
Found: C, 50.59; H, 5.55; N, 8.19.
J = 8.6 Hz, C6H2). Anal. Calcd for C17H26N2O8: C, 52.84; H, 6.78;
N, 7.25. Found: C, 52.60; H, 6.63; N, 7.19.
1-[2-Methoxy-2-(1-naphthyl)ethyl]-1-methylhydrazine Hydro-
gen Fumarate (12i Hydrogen Fumarate). Yield 2.01 g (58%); mp
147-149 °C (MeOH-Et2O); 1H NMR (400 MHz, D2O) δ 3.11
(s, 3H, NCH3), 3.37 (s, 3H, OCH3), 3.48 (dd, 1H, J = 3.0, 13.8
Hz, NCH2), 3.63 (dd, 1H, J = 10.4, 13.8 Hz, NCH2), 5.61 (dd,
1H, J = 3.4, 10.8 Hz, OCH), 6.67 (s, 2H, CH=CH), 7.60-7.72
(m, 4H, C10H7), 8.01 (d, 1H, J = 8.1 Hz, C10H7), 8.05 (d, 1H,
J = 8.1 Hz, C10H7), 8.22 (d, 1H, J = 8.1 Hz, C10H7). Anal.
Calcd for C18H22N2O5: C, 62.42; H, 6.40; N, 8.09. Found: C,
62.70; H, 6.67; N, 8.38.
1-[2-Methoxy-2-(2-naphthyl)ethyl]-1-methylhydrazine Hydro-
gen Maleate (12j Hydrogen Maleate). Yield 1.94 g (56%); mp
119-120 °C (MeOH-Et2O); 1H NMR (400 MHz, D2O) δ 3.09
(s, 3H, NCH3), 3.33 (s, 3H, OCH3), 3.41 (dd, 1H, J = 2.8, 13.6
Hz, NCH2), 3.63 (dd, 1H, J = 10.5, 13.6 Hz, NCH2), 4.95 (dd,
1H, J = 2.6, 10.4 Hz, OCH), 6.25 (s, 2H, CH=CH), 7.54 (dd,
1H, J = 1.7, 8.8 Hz, C10H7), 7.60-7.65 (m, 2H, C10H7),
7.95-8.04 (m, 4H, C10H7). Anal. Calcd for C18H22N2O5: C,
62.42; H, 6.40; N, 8.09. Found: C, 62.17; H, 6.15; N, 8.38.
1-[2-Benzyloxy-2-(4-methoxyphenyl)ethyl]-1-methylhydrazine
Hydrogen Fumarate (16 Hydrogen Fumarate). Beige crystalline
1
solid, yield 1.52 g (38%); mp 118-120 °C (EtOH-Et2O); H
NMR (400 MHz, D2O) δ 2.97 (s, 3H, NCH3), 3.31 (dd, 1H, J =
13.4, 3.0 Hz, NCH), 3.54 (dd, 1H, J = 13.4, 10.4 Hz, NCH), 3.86
(s, 3H, OCH3), 4.37 (d, 1H, J = 11.3 Hz, OCH2), 4.49 (d, 1H,
J = 11.3 Hz, OCH2), 4.87 (dd, 1H, J = 10.4, 3.0 Hz, OCH), 6.66
(s, 2H, CH=CH), 7.07 (d, 2H, J = 8.6 Hz, C6H4), 7.32-7.46 (m,
7H, C6H4, C6H5). Anal. Calcd for C21H26N2O6: C, 62.67; H,
6.51; N, 6.96. Found: C, 62.49; H, 6.71; N, 7.11.
1-(2-Phenylpropyl)hydrazine Hydrogen Fumarate (14a Hydro-
gen Fumarate). 2-Phenylpropyl bromide (1.99 g, 10 mmol) was
added dropwise to a stirred refluxing solution of hydrazine
hydrate (6.00 g, 0.12 mol) in EtOH (10 mL). The solution was
refluxed for 22 h and evaporated in vacuo. Et2O (25 mL) and
K2CO3 (5 g) was added to the residue and the mixture was stirred
for 15 min at ambient temperature. It was then filtered and
evaporated. The oily residue was treated with an equivalent
amount of fumaric acid in a mixture of EtOH and Et2O to give a
crystalline hydrogen fumarate salt. The crystalline product was
filtered off and recrystallized.
White crystalline solid; yield 1.24 g (47%); mp 126-129 °C
(EtOH-Et2O); 1H NMR (400 MHz, D2O) δ 1.32 (d, 3H, J = 6.8
Hz, CH3), 3.10-3.23 (m, 1H, CH), 3.30-3.42 (m, 2H, CH2),
6.67 (s, 2H, CH=CH), 7.28-7.47 (m, 5H, C6H5). Anal. Calcd
for C13H18N2O4: C, 58.63; H, 6.81; N, 10.52. Found: C, 58.75;
H, 6.62; N, 10.37.
Expression of Recombinant VAP-1. Recombinant VAP-1
protein was obtained from Chinese hamster ovary (CHO) cells
stably transfected with a full-length human VAP-1 cDNA.
Recombinant human VAP-1 SSAO expressed in CHO cells
was used as a source of VAP-1 SSAO for activity measurements.
Native CHO cells have negligible SSAO activity. These cells and
their culture have previously been described.8 A cell lysate was
prepared by suspending approximately 3.6 ꢀ 108 cells in 25 mL
lysis buffer (150 mM NaCl, 10 mM Tris-Base pH 7.2, 1.5 mM
MgCl2, 1% NP40) and incubating at 4 °C overnight on a
rotating table. The lysate was clarified by centrifugation at
18000g for 5 min at room temperature and the supernatant
was used directly in the assay.
Mitochondrial MAO Extract. Rat total monoamine oxidase
enzyme (MAO, mixture of MAO A and MAO B) was prepared
from rat liver tissues by rinsing the liver sample several times in
KCl-EDTA-solution to remove all blood. Then the liver sample
was homogenized in ice-cold potassium phosphate buffer
(0.1 M, pH 7.4) with an Ultra-Turax homogenizer (setting
11 000 rpm, 4 ꢀ 10 s). After centrifugation at 500g for 10 min
at 4 °C the supernatant was carefully withdrawn and was
centrifuged at 12300g for 15 min at 4 °C. The supernatant was
1-[2-Methoxy-1-(3-tolyl)ethyl]-1-methylhydrazine Hydrogen
Fumarate (12e Hydrogen Fumarate). White crystalline solid,
1
yield 1.74 g (56%); mp 116-119 °C (EtOH-Et2O); H NMR
(400 MHz, D2O) δ 2.37 (s, 3H, (CH3)C6H4), 3.05 (s, 3H, NCH3),
3.28 (s, 3H, OCH3), 3.32 (dd, 1H, J = 2.8,14.0 Hz, NCH2),
3.45 (dd, 1H, J = 10.4, 13.4 Hz, NCH2), 4.74 (dd, 1H, J = 3.0,
10.3 Hz, OCH), 6.67 (s, 2H, CH=CH), 7.21-7.32 (m,
3H, C6H4), 7.37-7.42 (m, 1H, C6H4). Anal. Calcd for C15H22-
N2O5: C, 58.05; H, 7.15; N, 9.03. Found: C, 57.77; H, 6.93;
N, 8.77.
1-[2-Methoxy-2-(3-methoxyphenyl)ethyl]-1-methylhydrazine
Hydrogen Fumarate (12f Hydrogen Fumarate). Yield 1.99 g
(61%); mp 140-143 °C (MeOH-Et2O); 1H NMR (400 MHz,
D2O) δ 3.03 (s, 3H, NCH3), 3.27 (s, 3H, OCH3), 3.31 (dd, 1H,
J = 3.5, 13.8 Hz, NCH2), 3.45 (dd, 1H, J = 10.3, 13.6 Hz,
NCH2), 3.84 (s, 3H, (CH3O)C6H4), 4.74 (dd, 1H, J = 3.2, 10.4
Hz, OCH), 6.65 (s, 2H, CH=CH), 6.99-7.05 (m, 3H, C6H4),
7.41 (t, 1H, J = 8.1 Hz, C6H4). Anal. Calcd for C15H22N2O6: C,
55.21; H, 6.80; N, 8.58. Found: C, 55.48; H, 7.05; N, 8.95.
1-[2-Methoxy-1-(4-methoxyphenyl)ethyl]-1-methylhydrazine
Hydrogen Fumarate (12g Hydrogen Fumarate). Pale yellow
crystalline solid, yield 1.96 g (60%); mp 104-107 °C (EtOH-
Et2O); 1H NMR (400 MHz, D2O) δ 3.07 (s, 3H, NCH3), 3.27 (s,
3H, OCH3), 3.33 (dd, 1H, J = 3.3, 13.8 Hz, NCH2), 3.51 (dd,
1H, J = 10.7, 13.8 Hz, NCH2), 3.87 (s, 3H, (CH3O)C6H4), 4.74
(dd, 1H, J = 3.1, 10.5 Hz, OCH), 6.68 (s, 2H, CH=CH), 7.08
(d, 2H, J = 8.8 Hz, C6H4), 7.39 (d, 2H, J = 8.8 Hz, C6H4).
Anal. Calcd for C15H22N2O6: C, 55.21; H, 6.80; N, 8.58.
Found: C, 55.47; H, 7.01; N, 8.85.
1-[2-Methoxy-2-(2,3,4-trimethoxyphenyl)ethyl]-1-methylhydrazine
Hydrogen Fumarate (12h Hydrogen Fumarate). Beige crystalline
1
solid, yield 1.97 g (51%); mp 122-124 °C (MeOH-Et2O); H
NMR (500 MHz, CDCl3) δ2.80 (s, 3H, NCH3), 3.44 (s, 3H, OCH3),
3.76-3.85 (m, 1H, NCH), 3.88 (s, 3H, OCH3), 3.91 (s, 3H, OCH3),
3.94 (s, 3H, OCH3), 4.07-4.15 (m, 1H, NCH), 4.70-4.80 (br s,
overlapped with the peak of the residual H2O, 1H, OCH), 6.61
(s, 2H, CH=CH), 6.96 (d, 1H, J = 8.6 Hz, C6H2), 7.22 (d, 1H,

Article
Journal of Medicinal Chemistry, 2010, Vol. 53, No. 17 6313
discharged and sedimented mitochondria were resuspended in
fresh phosphate buffer and centrifuged as previously. The
mitochondria were suspended in phosphate buffer and homo-
genized with an Ultra-Turax homogenizer (setting 11 000 rpm,
2 ꢀ 10 s). Mitochondrial preparate was aliquoted and stored
at -70 °C.
[(HID(520, 522)Nδ: -0.76; HIE(684)Nε -0.93]. Finally, hydro-
gens, neutralizing counterions, and TIP3P water molecules
˚
extending 13 A around the protein complex (rectangular box)
were added with LEAP.52
Molecular Dynamics Simulations. All energy minimizations
and molecular dynamics simulations were performed with
NAMD 2.653 using ff03 force field parameters54 for the protein
and gaff99 parameters55 for the ligands (TPQ þ compound 5a,
Table 1). The whole MD simulation procedure comprised of two
minimization runs and two MD simulation runs. (1) The water
molecules, counterions, and amino acid side chains were mini-
mized with the conjugate gradient algorithm (3000 steps) as rest
of the complex was constrained by restraining CR-atoms into
In Vitro Inhibition of VAP-1 SSAO Activity. VAP-1 SSAO
activity was measured using the coupled colorimetric method
essentially as described for monoamine oxidase and related
enzymes.48 The VAP-1 SSAO assay was performed in 96 well
microtiter plates as follows. To each well was added a prede-
termined amount of inhibitor if required. The amount of inhi-
bitor varied in each assay but was generally at a final concentra-
tion of between 1 nM and 50 μM. Controls lacked inhibitor. The
inhibitor was in a total volume of 20 μL in water. The assay was
performed in a final volume of 200 μL consisting of 0.2 M
potassium phosphate buffer (pH 7.6) and freshly made chromo-
genic solution containing 1 mM 2,4-dichlorophenol, 500 μM
4-aminoantipyrine and 4 U/mL horseradish peroxidase and an
amount of CHO cell lysate containing VAP-1 SSAO that caused
a change of 0.6 A490 per h. This was within the linear response
range of the assay. The plates were incubated for 30 min at 37 °C
and the background absorbance measured at 490 nm using a
Wallac Victor II multilabel counter. To initiate the enzyme
reaction 10 mM benzylamine (final concentration = 1 mM)
was added and the plate incubated for 1 h at 37 °C. The increase
in absorbance, reflecting VAP-1 SSAO activity, was measured
at 490 nm. Inhibition was presented as percent inhibition
compared to control after correcting for background absor-
bance and IC50 values calculated using GraphPad Prism.
Comparison of VAP-1 SSAO Activity versus Rat Total MAO
Activity. Total MAO activity was measured in a similar way as
for VAP-1 SSAO except that 2,4-dichlorophenol was replaced
by 1 mM vanillic acid. To each well was added a predetermined
amount of inhibitor if required. The amount of inhibitor varied
in each assay but was generally at a final concentration of
between 10 nM and 800 μM. Controls lacked inhibitor. The
inhibitor was in a total volume of 20 μL in water. The assay was
performed in a final volume of 300 μL consisting of 0.2 M
potassium phosphate buffer (pH 7.6) and of freshly made
chromogenic solution (as above) and total MAO preparation.
The plates were incubated for 30 min at 37 °C and the back-
ground absorbance measured at 490 nm using a Wallac Victor II
multilabel counter. To initiate the enzyme reaction 5 mM
tyramine (final concentration 0.5 mM) was added and the plate
incubated for 1 h at 37 °C. The increase in absorbance, reflecting
total MAO activity, was measured at 490 nm. Inhibition was
presented as percent inhibition compared to control after cor-
recting for background absorbance and IC50 values calculated
using GraphPad Prism. Clorgyline and pargyline (inhibitors of
MAO A and MAO B, respectively) at 0.5 μM were added to
some wells as positive controls for MAO inhibition.
their initial positions with the harmonic force of 5 kcal mol-1
˚
A
-2. (2) The whole system was minimized without constraints
(3000 steps) to make sure complete equilibration; (3) the first
MD simulation (360 ps), where the CR-atoms were restrained as
in step (1), was performed in the constant pressure, and (4) the
2.4 ns production simulation was performed without con-
straints. Since large conformation changes are not likely on
inhibitor binding, the 2.4 ns were considered to be long enough
simulation time.
The simulated complex was held at constant temperature
(300 K) with Langevin dynamics for all non-hydrogen atoms,
using a Langevin damping coefficient of 5 ps-1. A constant
ꢀ
pressure of 1 atm was upheld by a Nose-Hoover Langevin
piston56 with an oscillation time scale of 200 fs and a damping
time scale of 100 fs. An integration time step of 2 fs was used
under a multiple time stepping scheme.57 The bonded and short-
range interactions were calculated every time step and long-
range electrostatic interactions every third step. A cutoff value
˚
of 12 A was used for the van der Waals and short-range
electrostatic interactions. A switching function was enforced for the
van der Waals forces to smoothen the cutoff. The simulations were
conducted under the periodic boundary conditions with the full-
system, and the long-range electrostatics were counted with the
particle-mesh Ewald method.58 The bonds involving hydrogen
atoms were restrained by the SHAKE algorithm.59
Starting Structures for 3D QSAR. The crystal structure of
VAP-1 in complex with 2-hydrazinopyridine10 was retrieved
from the Protein Data Bank.60 The conformation of the side
chain of Leu469 was changed to resemble that obtained from
MD simulations (see Results and Discussion) using amino acid
side-chain rotamer library61 incorporated into BODIL model-
ing environment.62 In addition, the water molecules and the
atoms of bound inhibitor molecule 2-hydrazinopyridine (C1,
C2, C3, C4, C5, and N3 in the crystal structure) were removed.
Hydrogen atoms were added using SYBYL.49
Construction of Ligands for 3D QSAR. Ligands were sketched
in SYBYL49 and energy was minimized by using MMFF94s63
force field with MMFF94 charges and conjugate gradient
method until the energy gradient was less than 0.05 kcal/mol.
All possible enantiomers were built.
Starting Structures for Molecular Dynamics Simulations. A
3D structure of VAP-1 with TPQ in an active conformation
Docking Simulations. Bioactive conformations for studied
ligands were predicted with GOLD 3.1.1.64,65 The ligand bind-
ing pocket was defined using the CβH from Leu468 as a central
10
(PDB ID 2C11, resolution of 2.90 A) without a bound
˚
2-hydrazinopyridine ligand was used in the MD simulations.
All studied inhibitor molecules bind covalently to TPQ (called
thereafter as TPQ-ligand complex) and they were parametrized
as following: (1) TPQ-ligand complex was build with SYBYL
7.3,49 (2) optimized with GAUSSIAN0350 at the HF/6-31þG*
level, (3) the electrostatic potentials of the ligand were calculated
with GAUSSIAN03 (HF/6-31þG*) for the optimized TPQ-
ligand complex, and (4) the RESP methodology51 was used to
create the atom-centered point charges from the electrostatic
potentials. The charges of chemically comparable atoms in the
ligand molecules were set to identical values.
˚
atom with radius of 20 A. The ligands were docked covalently to
nitrogen atom N1 of PAQ1729 (PAQ is used in PDB file (PDB-
code: 2C1110) to assign TPQ and the 2-hydrazinopyridine
covalently bonded to it). Each ligand was docked 10 times;
however, the docking simulation was allowed to terminate in
cases where the root-mean-squared deviations (rmsd) of three
˚
best conformations were within 1.5 A.
3D QSAR Analysis. The superpositioning of ligands for 3D
QSAR was obtained by using the first ranked conformation for
each ligand from docking by using the fitness function of
GOLD. The CoMSIA method in SYBYL49 was used for model
building to explore physicochemical properties that contribute
to the correlation. The CoMSIA descriptors for steric and
hydrophobic fields were calculated in SYBYL49 using default
To maintain the coordination of three histidines (HID520,
HID522, and HIE684) to the Cu2þ ion, the negative charges of
the coordinating N-atoms were increased in order to keep the
Cu2þ ion in the coordination site throughout the MD simulation

6314 Journal of Medicinal Chemistry, 2010, Vol. 53, No. 17
Nurminen et al.
settings, and were used for model building. The correlation of
these fields with the -log IC50 values was calculated by using
the partial least-squares (PLS) method.66 Column filtering of
0.5 kcal/mol was used for leave-one-out (LOO) cross-valida-
tions. The built models were validated as follows: (1) the
47 molecules were sorted according to their experimental data,
(2) molecules were divided into four sets that where equal in size
(one containing 11 and three containing 12 molecules) and
diversity in VAP-1 potency, (3) three out of four sets were used
to build a QSAR model, and the potencies for the molecules in
the fourth set were predicted. Finally, all 47 molecules were used
to build a 3D QSAR model for both VAP-1 and MAO. Thus,
in total, 10 QSAR models were made: five for VAP-1 (one
that contained all 47 molecules and four containing 3/4 of
molecules), and similarly five for MAO.
adipocytes and search for novel functions. Inflammopharmacology
2003, 11, 119–126.
(6) Houslay, M. D.; Tipton, K. F. A kinetic evaluation of monoamine
oxidase activity in rat liver mitochondrial outer membranes.
Biochem. J. 1974, 139, 645–652.
(7) Ochiai, Y.; Itoh, K.; Sakurai, E.; Adachi, M.; Tanaka, Y. Substrate
selectivity of monoamine oxidase A, monoamine oxidase B, dia-
mine oxidase, and semicarbazide-sensitive amine oxidase in COS-1
expression systems. Biol. Pharm. Bull. 2006, 29, 2362–2366.
(8) Smith, D. J.; Salmi, M.; Bono, P.; Hellman, J.; Leu, T.; Jalkanen, S.
Cloning of vascular adhesion protein 1 reveals a novel multifunc-
tional adhesion molecule. J. Exp. Med. 1998, 188, 17–27.
(9) Jalkanen, S.; Salmi, M. Cell surface monoamine oxidases: enzymes
in search of a function. EMBO J. 2001, 20, 3893–3901.
(10) Jakobsson, E.; Nilsson, J.; Ogg, D.; Kleywegt, G. J. Structure of
human semicarbazide-sensitive amine oxidase/vascular adhesion
protein-1. Acta Crystallogr. D Biol. Crystallogr. 2005, 61, 1550–
1562.
The validation of the VAP-1 and MAO models was done
using progressive scrambling45 and bootstrapping.46,47 In pro-
gressive scrambling the maximum number of bins was set to 10
and the minimum number of bins was set to 2. The default value
of 0.85 was used for critical point, and a total of 100 scramblings
were used for both VAP-1 and MAO models. Progressive
scrambling was performed five times for both models and
average values were calculated for parameters. The parameters
acquired from the progressive scrambling are (1) q**2 which
indicates the predictivity of the model after potential effects of
redundancy have been removed, that is, the expected value of q2
at the specified critical point, (2) cSDEP which is the estimated
crossvalidated standard error at the specified critical point for
(11) Airenne, T. T.; Nymalm, Y.; Kidron, H.; Smith, D. J.; Pihlavisto,
M.; Salmi, M.; Jalkanen, S.; 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.
(12) Marttila-Ichihara, F.; Smith, D. J.; Stolen, C.; Yegutkin, G. G.;
Elima, K.; Mercier, N.; Kiviranta, R.; Pihlavisto, M.; Alaranta, S.;
€
€
€ €
Pentikainen, U.; Pentikainen, O.; Fulop, F.; Jalkanen, S.; Salmi,
M. Vascular amine oxidases are needed for leukocyte extravasation
into inflamed joints in vivo. Arthritis Rheum. 2006, 54, 2852–2862.
(13) Akin, E.; Aversa, J.; Steere, A. C. Expression of adhesion molecules
in synovia of patients with treatment-resistant lyme arthritis.
Infect. Immun. 2001, 69, 1774–1780.
(14) Boomsma, F.; Derkx, F. H.; van den Meiracker, A. H.; Man in ’t
Veld, A. J.; Schalekamp, M. A. Plasma semicarbazide-sensitive
amine oxidase activity is elevated in diabetes mellitus and correlates
with glycosylated haemoglobin. Clin. Sci. (London) 1995, 88, 675–
679.
(15) Boomsma, F.; Bhaggoe, U. M.; van der Houwen, A. M.; van den
Meiracker, A. H. Plasma semicarbazide-sensitive amine oxidase in
human (patho)physiology. Biochim. Biophys. Acta 2003, 1647,
48–54.
(16) Garpenstrand, H.; Ekblom, J.; Backlund, L. B.; Oreland, L.;
Rosenqvist, U. Elevated plasma semicarbazide-sensitive amine
oxidase (SSAO) activity in Type 2 diabetes mellitus complicated
by retinopathy. Diabet. Med. 1999, 16, 514–521.
r2 (the correlation of the 0scrambled responses with the un-
0
yy
perturbed data), and (3) dq2 /dr2yy which is the slope of q2 (the
crossvalidated correlation coefficient) evaluated at the specified
critical point with respect to the correlation of the original
dependent variables versus the perturbed dependent variables.
The value of the slope should be under 1 as the values over
1 indicate the overfitting of the model. For bootstrapping
validation 100 runs were performed.
(17) Weiss, H. G.; Klocker, J.; Labeck, B.; Nehoda, H.; Aigner, F.;
Klingler, A.; Ebenbichler, C.; Foger, B.; Lechleitner, M.; Patsch,
J. R.; Schwelberger, H. G. Plasma amine oxidase: a postulated
cardiovascular risk factor in nondiabetic obese patients. Metabo-
lism 2003, 52, 688–692.
Acknowledgment. This study was supported by the Sigrid
ꢀ
Juselius Foundation (OTP), Academy of Finland (UP Grant
Number 121393), and CSC-Finnish IT Center for Science
(OTP: jyy2516).
(18) Boomsma, F.; van Veldhuisen, D. J.; de Kam, P. J.; Man in’t Veld,
A. J.; Mosterd, A.; Lie, K. I.; Schalekamp, M. A. Plasma semi-
carbazide-sensitive amine oxidase is elevated in patients with
congestive heart failure. Cardiovasc. Res. 1997, 33, 387–391.
(19) Mathys, K. C.; Ponnampalam, S. N.; Padival, S.; Nagaraj, R. H.
Semicarbazide-sensitive amine oxidase in aortic smooth muscle
cells mediates synthesis of a methylglyoxal-AGE: implications for
vascular complications in diabetes. Biochem. Biophys. Res. Com-
mun. 2002, 297, 863–869.
(20) Wang, E. Y.; Gao, H.; Salter-Cid, L.; Zhang, J.; Huang, L.; Podar,
E. M.; Miller, A.; Zhao, J.; O’rourke, A.; Linnik, M. D. Design,
synthesis, and biological evaluation of semicarbazide-sensitive
amine oxidase (SSAO) inhibitors with anti-inflammatory activity.
J. Med. Chem. 2006, 49, 2166–2173.
(21) Yabanoglu, S.; Ucar, G.; Gokhan, N.; Salgin, U.; Yesilada, A.;
Bilgin, A. A. Interaction of rat lung SSAO with the novel 1-N-
substituted thiocarbamoyl-3-substituted phenyl-5-(2-pyrolyl)-2-
pyrazoline derivatives. J. Neural Transm. 2007, 114, 769–773.
(22) Olarte, A. Z.; Mian, A.; Clauzel, L. M.; Exposito, M. R.; Font,
F. Y.; Palomera, F. A. WO 2006013209, 2006.
(23) Clauzel, L. M.; Font, F. Y.; Vicente, S. G.; Exposito, M. R.;
Palomera, F. A.; Olarte, A. Z. US 2007066646 A1 20070322, 2007.
(24) Olivieri, A.; Tipton, K.; O’Sullivan, J. L-Lysine as a recognition
molecule for the VAP-1 function of SSAO. J. Neural Transm. 2007,
114, 747–749.
Supporting Information Available: Table S1: The melting
points of synthesized hydrazine molecules and the number of
the submodel where a compound belongs in QSAR model
building. Table S2: Statistics of different COMSIA 3D QSAR
models. Figure S1: Visualization of additional COMSIA-
models. This information is available free of charge via the
Internet at http://pubs.acs.org/.
References
(1) Yu, P. H.; Lu, L. X.; Fan, H.; Kazachkov, M.; Jiang, Z. J.;
Jalkanen, S.; Stolen, C. Involvement of semicarbazide-sensitive
amine oxidase-mediated deamination in lipopolysaccharide-
induced pulmonary inflammation. Am. J. Pathol. 2006, 168, 718–
726.
(2) Yu, P. H.; Deng, Y. L. Endogenous formaldehyde as a potential
factor of vulnerability of atherosclerosis: involvement of semicar-
bazide-sensitive amine oxidase-mediated methylamine turnover.
Atherosclerosis 1998, 140, 357–363.
(3) Janes, S. M.; Mu, D.; Wemmer, D.; Smith, A. J.; Kaur, S.; Maltby,
D.; Burlingame, A. L.; Klinman, J. P. A new redox cofactor in
eukaryotic enzymes: 6-hydroxydopa at the active site of bovine
serum amine oxidase. Science 1990, 248, 981–987.
(4) Wang, S. X.; Mure, M.; Medzihradszky, K. F.; Burlingame, A. L.;
Brown, D. E.; Dooley, D. M.; Smith, A. J.; Kagan, H. M.; Klin-
man, J. P. A crosslinked cofactor in lysyl oxidase: redox function
for amino acid side chains. Science 1996, 273, 1078–1084.
(5) Carpene, C.; Visentin, V.; Morin, N.; Prevot, D.; Smih, F.; Rouet,
P.; Jayat, D.; Fontana, E.; Lizcano, J. M. Characterization of
semicarbazide-sensitive amine oxidase in human subcutaneous
(25) Yegutkin, G. G.; Salminen, T.; Koskinen, K.; Kurtis, C.; McPher-
son, M. J.; Jalkanen, S.; Salmi, M. A peptide inhibitor of vascular
adhesion protein-1 (VAP-1) blocks leukocyte-endothelium inter-
actions under shear stress. Eur. J. Immunol. 2004, 34, 2276–2285.
(26) Wang, X.; Pietrangeli, P.; Mateescu, M. A.; Mondovi, B. Extended
substrate specificity of serum amine oxidase: possible involvement
in protein posttranslational modification. Biochem. Biophys. Res.
Commun. 1996, 223, 91–97.

Article
Journal of Medicinal Chemistry, 2010, Vol. 53, No. 17 6315
(27) Kirton, C. M.; Laukkanen, M. L.; Nieminen, A.; Merinen, M.;
Stolen, C. M.; Armour, K.; Smith, D. J.; Salmi, M.; Jalkanen, S.;
Clark, M. R. Function-blocking antibodies to human vascular
adhesion protein-1: a potential anti-inflammatory therapy. Eur. J.
Immunol. 2005, 35, 3119–3130.
(28) Jalkanen, S.; Jalkanen, M.; Salmi, M. WO/2006/134203, 2006.
(29) De Colibus, L.; Li, M.; Binda, C.; Lustig, A.; Edmondson, D. E.;
Mattevi, A. Three-dimensional structure of human monoamine
oxidase A (MAO A): relation to the structures of rat MAO A and
human MAO B. Proc. Natl. Acad. Sci. U. S. A. 2005, 102, 12684–
12689.
(49) SYBYL 7.3, Tripos International: St. Louis, MI.
(50) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Rob,
M. A.; Cheeseman, J. R.; Montgomery, J. A.; Jr., Vreven, T.;
Kudin, K. N.; Burant, J. C.; Millam, J. M.; Iyengar, S. S.; Tomasi,
J.; Barone, V.; Mennucci, B.; Cossi, M.; Scalmani, G.; Rega, N.;
Petersson, G. A.; Nakatsuji, H.; Hada, M.; Ehara, M.; Toyota, K.;
Fukuda, R.; Hasegawa, J.; Ishida, M.; Nakajima, T.; Honda, Y.;
Kitao, O.; Nakai, H.; Klene, M.; Li, X.; Knox, J. E.; Hratchian,
H. P.; Cross, J. B.; Bakken, V.; Adamo, C.; Jaramillo, J.;
Gomperts, R.; Stratmann, R. E.; Yazyev, O.; Austin, A. J.; Cammi,
R.; Pomelli, C.; Ochterski, J. W.; Ayala, P. Y.; Morokuma, K.;
Voth, G. A.; Salvador, P.; Dannenberg, J. J.; Zakrzewski, V. G.;
Dapprich, S.; Daniels, A. D.; Strain, M. C.; Farkas, O.; Malick,
D. K.; Rabuck, A. D.; Raghavachari, K.; Foresman, J. B.; Ortiz,
J. V.; Cui, Q.; Baboul, A. G.; Clifford, S.; Cioslowski, J.; Stefanov,
B. B.; Liu, G.; Liashenko, A.; Piskorz, P.; Komaromi, I.; Martin,
R. L.; Fox, D. J.; Keith, T.; M. A. Al-Laham, Peng, C. Y.;
Nanayakkara, A.; Challacombe, M.; Gill, P. M. W.; Johnson, B.;
Chen, W.; Wong, M. W.; Gonzalez, C.; Pople, J. A. Gaussian 03;
Gaussian, Inc.: Wallingford, CT, 2003.
(51) Bayly, C. I.; Cieplak, P.; Cornell, W. D.; Kollman, P. A. A well-
behaved electrostatic potential based method using charge re-
straints for deriving atomic charges: The RESP model. J. Phys.
Chem. 1993, 97, 10269.
(52) Wang, J.; Wang, W.; Kollman, P. A.; Case, D. A. Automatic atom
type and bond type perception in molecular mechanical calcula-
tions. J. Mol. Graph. Model. 2006, 25, 247–260.
(53) 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.
(54) Duan, Y.; Wu, C.; Chowdhury, S.; Lee, M. C.; Xiong, G.; Zhang,
W.; Yang, R.; Cieplak, P.; Luo, R.; Lee, T.; Caldwell, J.; Wang, J.;
Kollman, P. A point-charge force field for molecular mechanics
simulations of proteins based on condensed-phase quantum mecha-
nical calculations. J. Comput. Chem. 2003, 24, 1999–2012.
(55) Wang, J.; Wolf, R. M.; Caldwell, J. W.; Kollman, P. A.; Case, D. A.
Development and testing of a general amber force field. J. Comput.
Chem. 2004, 25, 1157–1174.
(30) Son, S. Y.; Ma, J.; Kondou, Y.; Yoshimura, M.; Yamashita, E.;
˚
Tsukihara, T. Structure of human monoamine oxidase A at 2.2-A
resolution: the control of opening the entry for substrates/inhibi-
tors. Proc. Natl. Acad. Sci. U. S. A. 2008, 105, 5739–5744.
(31) Binda, C.; Hubalek, F.; Li, M.; Herzig, Y.; Sterling, J.; Edmond-
son, D. E.; Mattevi, A. Crystal structures of monoamine oxidase B
in complex with four inhibitors of the N-propargylaminoindan
class. J. Med. Chem. 2004, 47, 1767–1774.
(32) Binda, C.; Newton-Vinson, P.; Hubalek, F.; Edmondson, D. E.;
Mattevi, A. Structure of human monoamine oxidase B, a drug
target for the treatment of neurological disorders. Nat. Struct. Biol.
2002, 9, 22–26.
(33) Binda, C.; Wang, J.; Pisani, L.; Caccia, C.; Carotti, A.; Salvati, P.;
Edmondson, D. E.; Mattevi, A. Structures of human mono-
amine oxidase B complexes with selective noncovalent inhibitors:
safinamide and coumarin analogs. J. Med. Chem. 2007, 50, 5848–
5852.
(34) Binda, C.; Wang, J.; Li, M.; Hubalek, F.; Mattevi, A.; Edmondson,
D. E. Structural and mechanistic studies of arylalkylhydrazine
inhibition of human monoamine oxidases A and B. Biochemistry
2008, 47, 5616–5625.
(35) Ragnarsson, U. Synthetic methodology for alkyl substituted hy-
drazines. Chem. Soc. Rev. 2001, 30, 205–213.
€ €
(36) Zalan, Z.; Lazar, L.; Fulop, F. Chemistry of hydrazinoalcohols and
their heterocyclic derivatives. Part 1. Synthesis of hydrazinoalco-
hols. Curr. Org. Chem 2005, 9, 357–376.
(37) Kim, M.; White, J. D. Olefins from thermal decomposition of
N-sulfoximino-2-oxazolidones. A novel synthesis of bicyclo-
[3.3.1]non-1-ene. J. Am. Chem. Soc. 1977, 99, 1172–1180.
(38) Roth, H. J.; Brandau, A. On the aminolysis of styroloxide,
preparation of phenylaminoethanols. Part IV. Aminolysis of ep-
oxide. Arch. Pharm. 1959, 292/64, 761–777.
(39) Coote, S. J.; Davies, S. G.; Middlemiss, D.; Naylor, A. Enantios-
pecific synthesis of (þ)-(R)-6,7-dimethoxy-2-methyl-4-phenyl-
1,2,3,4-tetrahydroisoquinoline from (þ)-(S)-2-methylamino-1-
phenylethanol. J. Chem. Soc., Perkin Trans. 1 1989, 2223-2228.
(40) Rosenmund, K. W.; Karg, E.; Marcus, F. K. Preparation of beta-
arylalkylamines. Chem. Ber. 1942, 1850–1859.
(56) Feller, S. E.; Zhang, Y.; Pastor, R. W.; Brooks, B. R. Constant
pressure molecular dynamics simulation: The Langevin piston
method. J. Chem. Phys. 1995, 103, 4613.
ꢀ
(57) Schlick, T.; Skeel, R. D.; Brunger, A. T.; Kale, L. V.; Board, J. A. J.;
Hermans, J.; Schulten, K. Algorithmic challenges in computational
molecular biophysics. J. Comput. Phys. 1999, 151, 9–48.
(58) Darden, T.; York, D.; Pedersen, L. Particle mesh Ewald: An W
log(N) method for Ewald sums in large systems. J. Chem. Phys.
1993, 98, 10089–10092.
(59) Ryckaert, J.; Ciccotti, G.; Berendsen, H. J. Numerical integration
of the Cartesian equations of motion of a system with constraints:
molecular dynamics of n-alkanes. J. Coput. Phys 1977, 23, 327.
(60) Berman, H. M.; Westbrook, J.; Feng, Z.; Gilliland, G.; Bhat, T. N.;
Weissig, H.; Shindyalov, I. N.; Bourne, P. E. The Protein Data
Bank. Nucleic Acids Res. 2000, 28, 235–242.
(61) Lovell, S. C.; Word, J. M.; Richardson, J. S.; Richardson, D. C.
The penultimate rotamer library. Proteins 2000, 40, 389–408.
(62) Lehtonen, J. V.; Still, D. J.; Rantanen, V. V.; Ekholm, J.;
(41) Schulz, H. Preparation of cardiac- and circulatory-active com-
pounds. Pharmazie 1967, 22, 19–22.
€
(42) Lovgren, K.; Nilsson, J. L. Adrenergic receptor agonists. Confor-
mation of erythro and threo forms of compounds related to
ephedrine and isoetharine. Acta Pharm. Suec. 1977, 14, 30–36.
(43) Enders, D.; Eichenauer, H. Asymmetric syntheses via metalated
chiral hydrazones. Enantioselective alkylation of cyclic ketones
and aldehydes. Chem. Ber. 1979, 112, 2933–2960.
€
€
Bjorklund, D.; Iftikhar, Z.; Huhtala, M.; Repo, S.; Jussila, A.;
(44) Porath, B.; Rademacher, P.; Boese, R.; Blaser, D. Hydrogen
€
€
bonding of 2-tetrazenes, 2. Synthesis and structural studies of
hydroxyalkyl-substituted 2-tetrazenes. Z. Naturforsch. 2002, 57,
365–376.
Jaakkola, J.; Pentikainen, O.; Nyronen, T.; Salminen, T.; Gyllen-
berg, M.; Johnson, M. S. BODIL: a molecular modeling environ-
ment for structure-function analysis and drug design. J. Comput.-
Aided Mol. Des. 2004, 18, 401–419.
(45) Clark, R. D.; Fox, P. C. Statistical variation in progressive
scrambling. J. Comput.-Aided Mol. Des. 2004, 18, 563–576.
(46) Eriksson, L.; Jaworska, J.; Worth, A. P.; Cronin, M. T.; McDo-
well, R. M.; Gramatica, P. Methods for reliability and uncertainty
assessment and for applicability evaluations of classification- and
regression-based QSARs. Environ. Health Perspect. 2003, 111,
1361–1375.
(63) Halgren, T. A. Merck molecular force field. I. Basis, form, scope,
parameterization, and performance of MMFF94. Comput. Chem.
1996, 17, 490–519.
(64) Jones, G.; Willett, P.; Glen, R. C. A genetic algorithm for flexible
molecular overlay and pharmacophore elucidation. J. Comput.-
Aided Mol. Des. 1995, 9, 532–549.
(47) Henderson, A. R. The bootstrap: a technique for data-driven
statistics. Using computer-intensive analyses to explore experi-
mental data. Clin. Chim. Acta 2005, 359, 1–26.
(48) Holt, A.; Smith, D. J.; Cendron, L.; Zanotti, G.; Rigo, A.; Di
Paolo, M. L. Multiple binding sites for substrates and modulators
of semicarbazide-sensitive amine oxidases: kinetic consequences.
Mol. Pharmacol. 2008, 73, 525-538.
(65) Jones, G.; Willett, P.; Glen, R. C.; Leach, A. R.; Taylor, R.
Development and validation of a genetic algorithm for flexible
docking. J. Mol. Biol. 1997, 267, 727–748.
(66) Cramer, R. D., III; Bunce, J. D.; Patterson, D. E.; Frank, I. E.
Crossvalidation, bootstrapping, and partial least squares com-
pared with multiple regression in conventional QSAR studies.
Quant. Struct.-Activity Relationships 1988, 7, 18–25.