Thio- and aminocaffeine analogues as inhibitors of human monoamine oxidase

Article abstract of DOI:10.1016/j.bmc.2011.10.036

In a recent study it was shown that 8-benzyloxycaffeine analogues act as potent reversible inhibitors of human monoamine oxidase (MAO) A and B. Although the benzyloxy side chain appears to be particularly favorable for enhancing the MAO inhibition potency

Full text of DOI:10.1016/j.bmc.2011.10.036

Bioorganic & Medicinal Chemistry 19 (2011) 7507–7518
Contents lists available at SciVerse ScienceDirect
Bioorganic & Medicinal Chemistry
journal homepage: www.elsevier.com/locate/bmc
Thio- and aminocaffeine analogues as inhibitors of human monoamine oxidase
Hermanus P. Booysen a, Christina Moraal a, Gisella Terre’Blanche a, Anél Petzer b, Jacobus J. Bergh a,
Jacobus P. Petzer a,
⇑
a Pharmaceutical Chemistry, School of Pharmacy, North-West University, Private Bag X6001, Potchefstroom 2520, South Africa
b Unit for Drug Research and Development, School of Pharmacy, North-West University, Private Bag X6001, Potchefstroom 2520, South Africa
a r t i c l e i n f o
a b s t r a c t
Article history:
In a recent study it was shown that 8-benzyloxycaffeine analogues act as potent reversible inhibitors of
human monoamine oxidase (MAO) A and B. Although the benzyloxy side chain appears to be particularly
favorable for enhancing the MAO inhibition potency of caffeine, a variety of other C8 oxy substituents of
caffeine also lead to potent MAO inhibition. In an attempt to discover additional C8 substituents of caf-
feine that lead to potent MAO inhibition and to explore the importance of the ether oxygen for the MAO
inhibition properties of C8 oxy-substituted caffeines, a series of 8-sulfanyl- and 8-aminocaffeine
analogues were synthesized and their human MAO-A and -B inhibition potencies were compared to those
of the 8-oxycaffeines. The results document that the sulfanylcaffeine analogues are reversible competi-
tive MAO-B inhibitors with potencies comparable to those of the oxycaffeines. The most potent inhibitor,
Received 26 August 2011
Revised 4 October 2011
Accepted 13 October 2011
Available online 20 October 2011
Keywords:
Monoamine oxidase
Reversible inhibition
Caffeine
Sulfanylcaffeine
Thiocaffeine
8-{[(4-bromophenyl)methyl]sulfanyl}caffeine, exhibited an IC50 value of 0.167 lM towards MAO-B.
While the sulfanylcaffeine analogues also exhibit affinities for MAO-A, they display in general a high
degree of MAO-B selectivity. The aminocaffeine analogues, in contrast, proved to be weak MAO inhibitors
with a number of analogues exhibiting no binding to the MAO-A and -B isozymes. The results of this
study are discussed with reference to possible binding orientations of selected caffeine analogues within
the active site cavities of MAO-A and -B. MAO-B selective sulfanylcaffeine derived inhibitors may act as
lead compounds for the design of antiparkinsonian therapies.
Aminocaffeine
Ó 2011 Elsevier Ltd. All rights reserved.
1. Introduction
drugs in the treatment of several disorders.5 For example, MAO-A
inhibitors block the central oxidation of serotonin by MAO-A and
The monoamine oxidases (MAO) A and B are mitochondrial
bound flavin adenine dinucleotide (FAD) enzymes which catalyze
the
are used as antidepressants. MAO-B inhibitors reduce the MAO-B
catalyzed oxidative metabolism of dopamine in the brain and are
used in the treatment of Parkinson’s disease. Of importance is the
observation that MAO-B activity and density increase in most brain
regions including the basal ganglia with age while MAO-A activity
remains unchanged.6,7 In the aged parkinsonian brain MAO-B is
therefore thought to be the principal MAO isozyme responsible
for dopamine catabolism. MAO-B inhibitors may conserve dopa-
mine in the basal ganglia and offer a symptomatic benefit in the
treatment of Parkinson’s disease.8–10 MAO-B inhibitors are
frequently combined with levodopa therapy since inhibitors of this
enzyme have been shown to enhance the elevation of dopamine lev-
els derived from levodopa.11 MAO-B inhibitors may permit a reduc-
tion of the dose of levodopa required for a therapeutic effect and
therefore the occurrence of levodopa associated side effects.12
MAO may also play an important role in the neurodegenerative pro-
cesses associated with Parkinson’s disease. The oxidation of dopa-
mine by MAO stoichiometrically yields potentially toxic metabolic
by-products.13 For each mole of dopamine oxidized by MAO, one
mole of hydrogen peroxide (which may lead to oxidative damage)
and dopaldehyde (which may react with exocyclic amino groups
a
-carbon oxidation of a variety of aminyl substrates.1 Human
MAO-A and -B consist of 529 and 520 amino acids, respectively,
and the FAD is covalently bound to a cysteinyl residue in both
enzymes (Cys-406 and Cys-397 in MAO-A and -B, respectively).
While MAO-A and -B are products of separate genes they share
approximately 70% amino acid sequence identity.2 The X-ray crys-
tallographic structures of MAO-A and -B indicate that the amino
acid residues comprising the active sites and their relative geome-
tries are similar with only 6 of the 16 active site amino acid residues
differing between the two enzymes.3,4 In spite of these similarities,
MAO-A and -B have different substrate and inhibitor specificities.
Most notably, MAO-A metabolizes the neurotransmitters, serotonin
and norepinephrine, as well as the dietary amine, tyramine. MAO-B
is well known to metabolize extraneous amines such as benzyl-
amine and phenylethylamine. Dopamine is considered to be a
substrate for both isozymes.5
Since MAO-A and -B are both involved in the degradation of neu-
rotransmitter amines, inhibitors of these enzymes are employed as
of nucleosides and N-terminal and lysine
e-amino groups of pro-
⇑
Corresponding author. Tel.: +27 18 2992206; fax: +27 18 2994243.
E-mail address: jacques.petzer@nwu.ac.za (J.P. Petzer).
teins) are formed.13 Inhibitors of MAO reduce the MAO-catalyzed
0968-0896/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmc.2011.10.036

7508
H. P. Booysen et al. / Bioorg. Med. Chem. 19 (2011) 7507–7518
Table 2
metabolism of DA and as a result reduce the formation of these toxic
by-products. MAO inhibitors are therefore considered as a potential
treatment strategy to slow the progression of Parkinson’s disease
since they may exert neuroprotective effects in the brain.13
The IC50 values for the inhibition of recombinant human MAO-A and -B by
compounds 4a–h
O
Based on the therapeutic value of MAO inhibitors the current
study aims to discover new reversible inhibitors of the MAO en-
zymes, particularly the B isozyme. For this purpose caffeine (1)
serve as lead compound (Fig. 1). Although caffeine is a weak
MAO-B inhibitor (Ki = 3.6 mM), substitution at the C8 position with
a variety of substituents has been shown to enhance the MAO-B
inhibition potency of caffeine to a large degree.14 In previous stud-
ies it was shown that substitution at C8 of caffeine with alkyloxy
N
N
R
N
O
N
SIb
a
IC50
(
l
M)
R
MAO-A
MAO-B
c
c
4a
4b
4c
4d
4e
4f
–NH–C6H5
—
—
—
—
–
–
2.6
–
3.2
–
c
c
–NH–CH2–C6H5
–NH–(CH2)2–C6H5
–NH–(CH2)3–C6H5
–NH–(CH2)4–C6H5
45.2 9.19
17.6 2.48
c
c
—
—
O
N
O
N
30.9 6.74
9.60 0.759
19.4 5.07
24.4 18.0
c
N
N
N
N
–NH–(CH2)2–(pyridin-2-yl)
–NH–(CH2)2–(3-ClC6H4)
–NH–C5H9
—
N
N
4g
4h
5.78 0.411
0.2
–
O
c
c
—
—
R
O
O
a
All values are expressed as the mean SD of triplicate determinations.
The selectivity index is the selectivity for the MAO-B isoform and is given as the
1
2
b
ratio of IC50(MAO-A)/IC50(MAO-B).
Figure 1. The structures of caffeine (1) and oxycaffeine analogues (2).
c
No inhibition observed at a maximum concentration of 100 lM of the test
inhibitor.
O
O
N
N
N
N
N
Table 3
H
N
N
N
N
The IC50 values for the inhibition of recombinant human MAO-A and -B by
compounds 5a–b
S
R
O
O
N
O
O
R
3
4
O
N
N
CH3
N
R
N
N
N
O
N
N
R
a
IC50
(l
M)
SIb
5
R
MAO-A
MAO-B
Figure 2. The structures of sulfanylcaffeine analogues (3), aminocaffeine analogues
(4) and aminomethylcaffeine analogues (5).
5a
5b
–(NCH3)–(CH2)2–C6H5
–(NCH3)–(CH2)4–C6H5
107 9.85
37.7 6.40
16.8 6.83
2.97 0.536
6.4
12.7
a
All values are expressed as the mean SD of triplicate determinations.
The selectivity index is the selectivity for the MAO-B isoform and is given as the
b
Table 1
ratio of IC50(MAO-A)/IC50(MAO-B).
The IC50 values for the inhibition of recombinant human MAO-A and -B by
compounds 3a–l
substituents (2) yielded particularly potent MAO-B inhibitors with
O
a
number of compounds exhibiting IC50 values in the nM
N
range.15,16 Interestingly these oxycaffeines are also MAO-A inhibi-
tors, a property that may be attributed to the relatively large
degree of rotational freedom of the C8 side chain at the carbon–
oxygen ether bond. It has been suggested that structures with a
relatively larger degree of conformational freedom may be better
suited for binding to MAO-A than relatively rigid structures.15
Based on these promising results the present study investigates
the possibility that alkylsulfanyl and alkylamino substituents at
C8 of caffeine may similarly enhance the MAO-A and -B inhibition
potency of caffeine. For this purpose, a series of 12 aryl- and alkyl-
sulfanylcaffeine analogues (3a–l) and 10 aryl- and alkylaminocaf-
feine analogues (4a–h, 5a–b) were synthesized and evaluated as
potential inhibitors of recombinant human MAO-A and -B (Fig. 2
and Tables 1–3).
N
R
N
O
N
a
IC50
(
l
M)
SIb
R
MAO-A
MAO-B
3a
3b
3c
3d
3e
3f
3g
3h
3i
–S–C6H5
–S–CH2–C6H5
–S–(CH2)2–C6H5
–S–(CH2)2–O–C6H5
–S–CH2–(4–Cl–C6H4)
–S–CH2–(4–Br–C6H4)
–S–CH2–(4–F–C6H4)
–S–CH2–(4–CH3O–C6H4)
–S–(CH2)2–CH(CH3)2
–S–C6H11
56.4 12.9
8.22 1.13
20.5 4.49
15.5 2.17
2.77 0.570
2.62 0.104
4.80 0.584
33.2 3.41
1.86 0.034
0.223 0.010
0.332 0.033
0.192 0.025
0.167 0.020
0.348 0.036
1.7
4.4
91.9
46.7
14.4
15.7
13.8
—
5.8
1.9
0.4
1.0
c
c
—
—
15.2 4.09
24.4 8.76
9.40 0.572
3.60 0.291
2.62 0.546
13.1 3.49
20.9 3.11
3.60 1.10
3j
3k
3l
–S–C5H9
–S–2-Naphthalenyl
2. Results
a
All values are expressed as the mean SD of triplicate determinations.
The selectivity index is the selectivity for the MAO-B isoform and is given as the
2.1. Chemistry
b
ratio of IC50(MAO-A)/IC50(MAO-B).
c
The aryl- and alkylsulfanylcaffeine analogues (3a–l) were
synthesized by reacting 8-chlorocaffeine (6) with an appropriate
No inhibition observed at a maximum concentration of 100 lM of the test
inhibitor.

H. P. Booysen et al. / Bioorg. Med. Chem. 19 (2011) 7507–7518
7509
nonselective. 8-(Phenylsulfanyl)caffeine (3a) which displays slight
selectivity for MAO-B, was found to be a relatively weak inhibitor
of both MAO-A and -B. Extension of the C8 side chain by one meth-
O
N
O
N
N
N
N
N
N
N
a
Cl
R
SH
+
S
ylene unit to yield the benzylsulfanyl homolog 3b (IC50 = 1.86
enhances the MAO-B inhibition potency 17-fold compared to 3a
(IC50 = 33.2 M). A further increase in the length of the C8 side
lM)
O
7
R
O
6
3
l
chain yields even more potent MAO-B inhibitors. For example,
compounds 3c and 3d, the phenylethylsulfanyl and pheno-
xyethylsulfanyl homologs exhibited IC50 values of 0.223 and
Scheme 1. Synthetic pathway to sulfanylcaffeine analogues (3). Reagents and
conditions: (a) NaOH, H2O/ethanol, reflux.
0.332
of 3a (IC50 = 56.4
(IC50 = 8.22 M) also results in improved MAO-A inhibition, further
increasing the length of the C8 side chain does not result in a fur-
ther enhancement of inhibition activity. For example, 3c
l
M, respectively. While the extension of the C8 side chain
l
M) by one methylene unit to yield 3b
O
N
O
N
l
N
N
N
N
H
N
N
N
a,b
Cl
R
NH2
+
(IC50 = 20.5
lM) and 3d (IC50 = 15.5
lM) are weaker MAO-A inhib-
8
O
O
R
itors than the benzylsulfanyl homolog 3b (IC50 = 8.22
l
M).
6
4
Interestingly, halogen substitution of the phenyl ring of the C8
side chain is also associated with an increase in MAO-B inhibition
potency. The benzylsulfanyl substituted caffeine homologs
O
N
CH3
N
containing chlorine (3e; IC50 = 0.192
IC50 = 0.167 M) and fluorine (3g; IC50 = 0.348
phenyl ring were found to be 5- to 11-fold more potent than the
corresponding unsubstituted homolog 3b (IC50 = 1.86 M). Meth-
lM), bromine (3f;
N
c
N
l
lM) on the benzyl
N
O
R
l
5
oxy substitution of the phenyl ring of 8-(benzylsulfanyl)caffeine,
in contrast, is associated with a loss of both MAO-A and -B inhibi-
tion activity. Halogen substitution of 8-(benzylsulfanyl)caffeine
(3b) also enhances MAO-A inhibition potency, although by a lesser
degree compared to MAO-B. The homologs containing chlorine
Scheme 2. Synthetic pathway to aminocaffeine analogues (4 and 5). Reagents and
conditions: (a) reflux; (b) acetic acid; (c) KOH, DMSO, CH3I.
thiol reagent (7) in the presence of NaOH and employing a mixture
of ethanol and water as reaction solvent (Scheme 1).17 The aryl-
and alkylaminocaffeine analogues (4a–h) were similarly prepared
by reaction of an appropriate amine reagent (8) with 8-chlorocaf-
feine but with the exception that the addition of base and
additional solvent were not required (Scheme 2).18 Methylation
of 4c and 4e at the C8 amine to yield aminomethyl caffeine ana-
logues 5a and 5b, respectively, were carried out in DMSO using
CH3I as alkylating agent and KOH as base. The structures of all
the target compounds were verified by 1H NMR, 13C NMR and mass
spectrometry. The purities of the compounds were estimated by
HPLC analysis.
(3e; IC50 = 2.77
(3g; IC50 = 4.80
l
M), bromine (3f; IC50 = 2.62
lM) and fluorine
l
M) on the benzyl phenyl ring are 1.7- to 3-fold
more potent than the corresponding unsubstituted homolog 3b
(IC50 = 8.22 M).
l
The results document that compound 3i, the sulfanylcaffeine
analogue containing a branched alkyl side chain at C8, is also a
relatively good MAO-B inhibitor with an IC50 value of 2.62 lM. In
fact, 3i is approximately 12-fold more potent as an MAO-B inhibitor
than was 8-(phenylsulfanyl)caffeine (3a) which contains a C8 phen-
ylsulfanyl side chain. This result demonstrates that a ring system in
the C8 side chain is not an absolute requirement for MAO-B inhibi-
tion by sulfanylcaffeine analogues. In accordance with this view,
compound 3i also proved to be more potent as a MAO-B inhibitor
than the sulfanylcaffeine analogues containing cyclohexyl (3j),
cyclopentyl (3k) and napthalenyl (3l) C8 side chains. Interestingly
the sulfanylcaffeine analogues containing cyclohexyl (3j), cyclopen-
tyl (3k) and napthalenyl (3l) C8 side chains were more potent inhib-
itors of both MAO-A and -B than the phenyl substituted analogue 3a.
This result suggests that, with the appropriate structural modifica-
tion, these moieties may be suitable for the future design of sulfanyl-
caffeine derived MAO inhibitors. An example of such a structural
modification would be the extension of the length of the C8 side
chain. Based on the observations that the naphthalenyl substituted
sulfanylcaffeine analogue 3l is 15- and 9-fold more potent as a
MAO-A and -B inhibitor, respectively, than the phenyl substituted
sulfanylcaffeine analogue 3a and that 3l is a nonselective inhibitor,
the naphthalenyl moiety may be particularly suited for the design
of sulfanylcaffeine derived mixed MAO-A/B inhibitors.
2.2. Inhibition of MAO-A and -B
The MAO inhibition potencies of the sulfanylcaffeine (3a–l) and
aminocaffeine analogues (4a–h, 5a–b) were examined by employ-
ing the recombinant human MAO-A and -B enzymes as enzyme
sources.19 The mixed MAO-A/B substrate, kynuramine, was used
as substrate for the inhibition studies of both MAO-A and -B.
Kynuramine displays similar Km values towards the two enzymes
with values of 16.1 and 22.7 l
M for MAO-A and -B, respectively.15
The MAO-catalyzed oxidation of kynuramine yields 4-hydroxy-
quinoline, a fluorescent compound which is readily measured in
basic solutions at excitation and emission wavelengths of 310
and 400 nm, respectively. Neither the substrate nor the test inhib-
itors fluoresce under these conditions, or quench the fluorescence
of 4-hydroxyquinoline. The inhibition potencies of the sulfanylcaf-
feine and aminocaffeine analogues are expressed as the IC50 values
which were determined from sigmoidal dose–response curves con-
structed in triplicate from six different inhibitor concentrations
spanning at least 3 orders of magnitude.
2.2.2. MAO inhibition by aminocaffeine analogues (4a–h, 5a–b)
The MAO inhibition potencies of the aminocaffeine analogues
4a–h are presented in Tables 2 and 3. The data show that these
compounds were relatively weak inhibitors of both MAO-A and -
2.2.1. MAO inhibition by sulfanylcaffeine analogues (3a–l)
The MAO inhibition potencies of the sulfanylcaffeine analogues
(3a–l) are presented in Table 1. As shown by the selectivity index
(SI) values, the sulfanylcaffeine analogues are selective inhibitors
of MAO-B. The only exceptions are 3k which displays slight
selectivity for the MAO-A isozyme and 3l which is essentially
B with IC50 values ranging from 5.78–45.2 to 9.60–24.4 lM for
the inhibition of MAO-A and -B, respectively. In fact, several of
the compounds exhibited no binding and the MAO-A and -B
isozymes. Even homologs containing extended C8 side chains such
as 4d (–NH–(CH2)3–C6H5) and 4e (–NH–(CH2)4–C6H5) did not

7510
H. P. Booysen et al. / Bioorg. Med. Chem. 19 (2011) 7507–7518
exhibit potent MAO inhibition. Similarly, homolog 4g, which con-
tains a halogen on the phenyl ring of the C8 side chain, was found
to be a relatively weak MAO inhibitor. These results demonstrate
that, in contrast to the sulfanylcaffeine analogues, extension of
the length of the C8 side chain and halogen substitution do not lead
to potent MAO-B inhibition by the aminocaffeine analogues. Com-
pared to the sulfanylcaffeine analogues, the aminocaffeine ana-
logues are therefore weak MAO-B inhibitors. For example,
MAO-A
25
20
15
10
5
sulfanylcaffeine analogue 3c (IC50 = 0.223
tent as an MAO-B inhibitor than its corresponding aminocaffeine
homolog 4c (IC50 = 17.6 M).
lM) is 78-fold more po-
l
To investigate the possibility of enhancing the MAO inhibition
potencies of the aminocaffeines, selected analogues, 4c and 4e,
were methylated at the C8 amine to yield compounds 5a and 5b,
respectively. While methylation improved the MAO-B inhibition
potency of 4e by approximately 3-fold, MAO-A inhibition activity
was slightly reduced. Methylation of 4c did not result in a signifi-
cant improvement of MAO-B inhibition potency and led to a reduc-
tion in MAO-A inhibition. Compared to the sulfanylcaffeine
analogues, the aminocaffeine analogues are also weak MAO-B
inhibitors. For example, sulfanylcaffeine analogue 3c (IC50 = 0.2
0
No Inhibitor
0
15
30
60
Incubation time (min)
MAO-B
8
6
4
2
0
23
homolog 5a (IC50 = 16.8
MAO-B inhibitor among the aminocaffeine analogues was the C8
methylated derivative 5b with an IC50 value of 2.97 M. It can
lM) is 75-fold more potent than its aminomethylcaffeine
lM). It is noteworthy that the most potent
l
therefore be concluded that while methylation of the C8 amine
of aminocaffeine analogues may result in enhanced MAO-B inhibi-
tion potency, the resulting compounds remain relatively weak
inhibitors compared to the sulfanylcaffeine analogues.
2.3. Reversibility of MAO-A and -B inhibition
Based on the nature of the interactions with the MAO enzymes,
inhibitors may be classified as reversible or irreversible. Irrevers-
ible inhibitors normally form covalent interactions with the
enzymes while reversible inhibitors bind via intermolecular inter-
actions. While irreversible inhibitors of MAO have been clinically
used for many years, this mode of inhibition may be associated
with certain shortcomings.20 These include a slow and variable
recovery of enzyme activity following withdrawal of the irrevers-
ible inhibitor.21 The turnover rate for the biosynthesis of MAO-B
in the human brain may be as much as 40 days.22 In contrast,
enzyme activity is regained relatively quickly following with-
drawal of a reversible inhibitor, once the inhibitor is cleared from
the tissues. Based on these observations, the reversibility of
MAO-A and -B inhibition by the sulfanylcaffeine and aminocaffeine
analogues was examined. For this purpose the time-dependency of
the inhibition of MAO-A and -B, respectively, by sulfanylcaffeine
analogue 3f was measured. The time-dependency of the inhibition
of MAO-A and -B by aminocaffiene analogues 4g and 5b, respec-
tively, was also examined. While irreversible inhibitors would lead
to a time-dependent reduction of enzyme activity, the degree of
enzyme inhibition in the presence of a reversible MAO inhibitor
remains unchanged irrespective of the time for which the inhibitor
is incubated with the enzyme.23
No Inhibitor
0
15
30
60
Incubation time (min)
Figure 3. Time-dependent inhibition of recombinant human MAO-A and -B by 3f.
The enzymes were preincubated for various periods of time (0–60 min) with 3f at
concentrations of 5.22 and 0.32
l
M for MAO-A and -B, respectively. The concen-
trations of the enzyme substrate, kynuramine, were 45 and 30
l
M for the studies
with MAO-A and MAO-B, respectively, and the enzyme concentrations were
0.015 mg/mL. The catalytic rates are expressed as nmoles 4-hydroxyquinoline
formed/min/mg protein.
the MAO catalytic rates. These results suggest that the test caffe-
ines are not time-dependent inhibitors of MAO-A and -B and inter-
act reversibly, at least for the time period (0–60 min) and at the
inhibitor concentrations (2 Â IC50) evaluated.
This study also examined the possibility that 3f may act as a
competitive inhibitor of human MAO-A and -B. For this purpose,
sets of Lineweaver–Burk plots were constructed for the inhibition
of these enzymes by 3f. The initial catalytic rates of MAO-A or -B
were measured in the absence and presence of three different con-
centrations of 3f. These measurements were carried out using four
different concentrations of the substrate, kynuramine (15–90 lM).
The test inhibitors, at concentrations of approximately 2-fold
their measured IC50 values for the inhibition of the respective
MAO enzymes, were preincubated with recombinant human
MAO-A or -B for time periods of 0, 15, 30 and 60 min. Following
these preincubations the residual MAO catalytic activities were
measured after the addition of the substrate, kynuramine. The
results of these reversibility studies are presented in Figures 3
and 4. The graphs show that when 3f and 4g are preincubated with
MAO-A and 3f and 5b are preincubated with MAO-B there are no
time-dependent reductions of MAO-A and -B catalytic activities.
Even after a period of 60 min the test compounds do not reduce
The Lineweaver–Burk plots obtained from these experiments are
shown in Figure 5. The graphs show that the Lineweaver–Burk
plots constructed for the inhibition of MAO-A and -B are linear
and intersect at the y-axis. This indicates that the inhibition of
the MAO enzymes by 3f is competitive. This result is further sup-
port that 3f is a reversible MAO inhibitor.
2.4. Molecular modeling
The results of the MAO inhibition studies shows that among the
sulfanylcaffeine analogues evaluated here, several compounds act

H. P. Booysen et al. / Bioorg. Med. Chem. 19 (2011) 7507–7518
7511
MAO-A
MAO-A
0.16
10
5
0.12
0.08
0.04
0.00
0
No Inhibitor
0
15
30
60
-0.02
0.00
0.02
1/[S]
0.04
0.06
Incubation time (min)
MAO-B
MAO-B
0.5
2.0
1.5
1.0
0.5
0.0
0.4
0.3
0.2
0.1
0.0
No Inhibitor
0
15
30
60
-0.02
0.00
0.02
1/[S]
0.04
0.06
Incubation time (min)
Figure 4. Time-dependent inhibition of recombinant human MAO-A and -B by 4g
and 5b, respectively. The enzymes were preincubated for various periods of time
(0–60 min) with 4g (MAO-A) and 5b (MAO-B) at concentrations of 11.56 and
Figure 5. Lineweaver–Burk plots of the recombinant human MAO-A and -B
catalyzed oxidation of kynuramine in the absence (filled squares) and presence of
various concentrations of 3f. For the studies with MAO-A the concentrations of 3f
5.94
lM, respectively. The concentrations of the enzyme substrate, kynuramine,
were: 1.31
the studies with MAO-B the concentrations of 3f were: 0.04
0.08 M (filled circles), 0.16 M (open circles). The rates (V) are expressed as nmol
lM (open squares), 2.61
lM (filled circles), 5.22
lM (open circles). For
were 45 and 30
lM for the studies with MAO-A and MAO-B, respectively, and the
lM (open squares),
enzyme concentrations were 0.0075 mg/mL. The catalytic rates are expressed as
nmoles 4-hydroxyquinoline formed/min/mg protein.
l
l
product formed/min/mg protein.
as potent reversible inhibitors of MAO-B with IC50 values in the nM
range. Interestingly, extension of the length of the C8 side chain
leads to enhanced MAO-B inhibition potency. While the sulfanyl-
caffeine analogues are also MAO-A inhibitors, they display, for
the most part, selectivity for the MAO-B isozyme. In contrast to
the sulfanylcaffeine analogues, the aminocaffeine analogues were
found to be weak MAO-B inhibitors with many analogues exhibit-
ing no binding to either MAO-A or -B. To provide additional insight,
the predicted binding modes of selected analogues (3a–c and 4c) in
the active site cavities of MAO-A and -B were examined using
molecular docking.
The docking studies were carried out using the LigandFit appli-
cation of the Discovery Studio modeling software (Accelrys)
according to a previously reported protocol.23 As enzyme models,
the three-dimensional structures of human MAO-A cocrystallized
with harmine (PDB entry: 2Z5X)3 and human MAO-B cocrystal-
lized with safinamide (PDB entry: 2V5Z)4 were selected. The en-
zyme models were prepared by calculating the protonation states
of the ionizable residues and adding the hydrogen atoms accord-
ingly. After the valences of the FAD cofactor and cocrystallized li-
gands were corrected, hydrogen atoms were added and the
models were subjected to an energy minimization cascade while
the protein backbone was constrained. For the purpose of the dock-
ing, only the crystal waters, which are reported to be conserved
and non-displaceable, were retained (see Section 4).
3,4
The best ranked docking solution for the binding of the selected
analogues (3a–c and 4c) to MAO-B shows one prevailing orienta-
tion for all of the inhibitors. As shown by the binding orientation
of 3c, the caffeine ring binds within the substrate cavity of MAO-
B, in close proximity to the FAD cofactor (Fig. 6). This places the
carbonyl oxygen at C2 of the caffeine ring 3.4 Å from the flavin
N5 and the carbonyl oxygen at C6 within hydrogen bond distance
to the phenolic hydrogen of Tyr-435. The caffeine ring also forms a
potential
p–p interaction with the aromatic ring of Tyr-398. The
region defined by the flavin isoalloxazine ring, Tyr-398 and Tyr-
435 is the only polar space of the MAO-B active site and is also
the site where amine catalysis occurs.24 In the MAO-B model se-
lected for these studies, the side chain of Ile-199 is rotated into
an alternative conformation to allow for the fusion of the substrate
and entrance cavities.25 This rotation of the Ile-199 side chain from
the active site cavity is essential for relatively large inhibitors, such
as safinamide and C8 substituted caffeine derivatives, to be able to
bind to MAO-B. As a result the phenylethyl C8 side chain of 3c is

7512
H. P. Booysen et al. / Bioorg. Med. Chem. 19 (2011) 7507–7518
orientations of 3a and 3b within the MAO-B active site, 3b was
found to be a 17-fold more potent inhibitor. Since 3b protrudes
only slightly deeper (by ꢀ1.2 Å) into the entrance cavity compared
to 3a, this result suggest that a relatively small enhancement of the
space occupied by an inhibitor in the entrance cavity leads to a
large increase in binding affinity. Another possible explanation
may be that the larger degree of conformational freedom afforded
by the longer C8 side chain of 3b may facilitate improved interac-
tion with the entrance cavity. The view that interaction with the
entrance cavity is essential for high affinity inhibitor binding is
supported by the observation that caffeine is a weak MAO-B inhib-
itor.14 Lacking a C8 side chain, caffeine is expected to bind only
within the substrate cavity and is unable to interact with the en-
trance cavity. The lower MAO-B inhibition potencies of 3a and 3b
compared to compound 3c may thus be explained by weaker inter-
action with the entrance cavity.
Interestingly the aminocaffeine analogue 4c adopts a similar
binding mode to that described above for 3c and as a result forms
similar interactions with the MAO-B active site (Fig. 8). The only
additional interaction that may occur between 4c and MAO-B is
a potential hydrogen bond between the C8 amine and the phenolic
group of Tyr-326. The results of the MAO inhibition studies how-
ever, document that the aminocaffeine analogues are weak MAO-
B inhibitors with 4c being 78-fold weaker than the corresponding
sulfanylcaffeine homolog 3c. The docking studies therefore suggest
that differing binding orientations cannot account for the apparent
loss of MAO-B inhibition activity of the aminocaffeine analogues.
The predicted binding orientation of 3c within the active site
of MAO-A is similar to the binding orientation observed in
MAO-B, with the caffeine ring bound in close proximity to the
FAD cofactor and the C8 side chain extending towards the en-
trance of the active site (Fig. 9). Interestingly, the caffeine ring
is rotated by ꢀ180 °C compared to the binding orientation
adopted in the MAO-B active site. This dissimilarity in binding
orientations in the MAO-A and -B active sites has also been
observed in docking studies with 8-benzyloxycaffeine ana-
Figure 6. The predicted binding orientation of 3c (orange) in the MAO-B active site.
allowed to extend into the hydrophobic entrance cavity where it
may be stabilized via Van der Waals interactions. As expected,
the relatively shorter phenyl (3a) and benzyl (3b) C8 side chains
of sulfanylcaffeine homologs 3a and 3b do not extend as deep into
the MAO-B entrance cavity as the side chain of 3c, and may there-
fore undergo interactions with the entrance cavity to a lesser ex-
tent compared to 3c (Fig. 7). Despite the similar binding
Figure 7. The predicted binding orientations of 3a (orange) and 3b (magenta) in the
MAO-B active site.
Figure 8. The predicted binding orientation of 4c (green) in the MAO-B active site.

H. P. Booysen et al. / Bioorg. Med. Chem. 19 (2011) 7507–7518
7513
Figure 9. The predicted binding orientation of 3c (magenta) in the MAO-A active
site.
Figure 10. The predicted binding orientations of 3c within the active sites of MAO-
A
(green) and MAO-B (cyan) with the caffeine moieties of the respective
orientations overlaid.
logues.15 As a result of the flipped orientation of the caffeine ring,
the C2 carbonyl oxygen is within hydrogen bond distance of the
phenolic hydrogen of Tyr-444 and two active site waters. Also,
the caffeine ring of 3c binds more distant from Tyr-407 (ꢀ4.3 Å)
in MAO-A than from the corresponding residue, Tyr-398
(ꢀ3.6 Å), in MAO-B. For this reason, a
p–p interaction similar to
that observed between the caffeine ring and Tyr-398 in MAO-B,
is not observed between 3c and the MAO-A active site. This
may represent a possible reason for the finding that sulfanylcaf-
feine analogues are in general more potent MAO-B inhibitors than
MAO-A inhibitors. Also noteworthy is the observation that, in the
MAO-A active site, the C8 side chain of 3c is bent at the CH2–S
thioether bond from the plane of the caffeine ring while in the
MAO-B active site, the side chain of 3c exhibits a modest devia-
tion from the plane of the caffeine ring (Fig. 10). The differing
binding orientations adopted by C8 substituted caffeine deriva-
tives in MAO-A and -B may, for the most part, be attributed to
steric hindrance caused by the aromatic moieties of Phe-208 in
MAO-A and Tyr-326 in MAO-B. As illustrated in Figure 11A, the
binding orientation and position of 3c in the MAO-B active site
cannot be reproduced in MAO-A because this would result in
structural overlap with the phenyl ring of Phe-208. In MAO-B,
Figure 11. Illustrations of the overlaid active sites of human MAO-A and -B. Panel
A: The predicted binding orientation of 3c as docked within the active site of MAO-B
is shown in the MAO-A active site. The active site residues of MAO-A are displayed
in gray with Phe-208 in magenta while residue Ile-199 in MAO-B is displayed in
green. Panel B: The predicted binding orientation of 3c as docked within the active
site of MAO-A is shown in the MAO-B active site. The active site residues of MAO-B
are displayed in gray with Tyr-326 in magenta while residue Ile-325 in MAO-A is
displayed in green.

7514
H. P. Booysen et al. / Bioorg. Med. Chem. 19 (2011) 7507–7518
the amino acid residue that occupies the same position as Phe-
208 in MAO-A is Ile-199. In MAO-B the side chain of Ile-199
may rotate out of the active site cavity to allow for the observed
binding pose of 3c.25 Similarly, the binding orientation and posi-
tion of 3c in the MAO-A active site cannot be reproduced in MAO-
B because of structural overlap with of Tyr-326 (Fig. 11B). In
MAO-A, the amino acid residue that occupies the analogous posi-
tion as Tyr-398 in MAO-B is Ile-335. The relatively smaller side
chain of Ile-335 compared to the aromatic ring of Tyr-398, does
not sterically prevent the observed binding orientation of 3c in
the MAO-A active site.3
entrance cavity where they are stabilized by Van der Waals interac-
tions. Since extension of the C8 chain length results in enhanced
MAO-B inhibition potency, it may be concluded that longer C8 side
chains form more productive interactions with the MAO-B entrance
cavity, which thus leads to more potent enzyme inhibition. Halogen
substitution on the phenyl ring of the C8 side chain also leads to a
significant enhancement of MAO-B inhibition. This result may be ex-
plained by the possibility that halogen substitution may further im-
prove Van der Waals and dipole interactions between the MAO-B
entrance cavity and the C8 side chain. While it is not clear why meth-
oxy substitution of 8-(benzylsulfanyl)caffeine leads to a loss of both
MAO-A and -B inhibition potency, this result is in accordance with
the findings of a previous study, which showed that MAO-B inhibi-
tion potencies of a series of benzyloxycaffeine analogues correlate
with the electronegativity of substituents on the phenyl ring of the
C8 side chain and that electron-withdrawing groups enhance
MAO-B inhibition potency.15 Interestingly, this study shows that
C8 side chains that do not contain phenyl rings are also suitable
for MAO inhibition. Examples of sulfanylcaffeine analogues contain-
ing such side chains are the 3-methylbutyl (3i), cyclohexyl (3j),
cyclopentyl (3k) and napthalenyl (3l) substituted homologs.
One of the most significant findings of this study is that the
aminocaffeine analogues are weak MAO inhibitors with most
homologs displaying no inhibition. The predicted binding orienta-
tion and interactions of aminocaffeine analogue 4c in the MAO-B
active site is similar to the orientation of sulfanylcaffeine analogue
3c. In fact 4c displays an additional hydrogen bond interaction
with Tyr-326. In spite of these predictions, 4c is approximately
78-fold weaker as a MAO-B inhibitor compared to 3c. Even meth-
ylation of the C8 amines to yield tertiary amines does not produce
inhibitors with similar potencies to those of the sulfanylcaffeine
analogues. While the reasons for this behavior is not clear, differing
ionization states of the sulfanylcaffeine and aminocaffeine ana-
logues do not explain the difference in binding affinities to the
MAO enzymes, since both the sulfanylcaffeine and aminocaffeine
analogues are expected to be uncharged in the buffer used for
the inhibition studies (pH 7.4). Also, it is unlikely that aminocaffe-
ines are excluded from entering the access channel leading to the
active site cavity since aminyl substrates are thought to be depro-
tonated prior to entering the MAO active sites.24
3. Discussion
Based on previous reports that oxycaffeine analogues are MAO
inhibitors,15,16 the present study investigated the possibility that
C8 substituted sulfanylcaffeine and aminocaffeine analogues may
also act as inhibitors of human MAO-A and -B. The results demon-
strated that several of the sulfanylcaffeine analogues act as potent
MAO-B inhibitors and that the inhibition is reversible. For example,
the bromine substituted sulfanylcaffeine analogue 3f was the most
potent MAO-B inhibitor with an IC50 value of 0.167 lM. The rela-
tively high MAO-B inhibition potencies of the sulfanylcaffeine ana-
logues may be evaluated by comparison of the IC50 value of 3f
(IC50 = 0.167
binds to MAO-B with an IC50 value of 0.08
l
M) with the reversible inhibitor safinamide, which
l
M.4 While the sulfanyl-
caffeine analogues are also MAO-A inhibitors, they are for the most
part selective for the MAO-B isoform. Modeling studies predict that
the sulfanylcaffeine analogues adopt dissimilar binding modes in
the MAO-A and -B active site cavities, respectively. Compared to
the predicted orientation in MAO-B, the caffeine ring is flipped
by approximately 180° in the MAO-A active site which results in
differing interactions of the caffeine ring with the polar regions
of the MAO-A and -B substrate cavities. The alternative binding ori-
entation of the caffeine ring in MAO-A may be less optimal for the
formation of stabilizing polar interactions compared to the binding
orientation adopted in MAO-B and may explain, at least in part, the
lower binding affinities of the sulfanylcaffeine analogues to MAO-
A.15 Interestingly, modeling studies suggest that the C8 side chain
of the sulfanylcaffeine analogue 3c is bent to a high degree from
the plane of the caffeine ring at the CH2–S thioether bond while
in the MAO-B active site, the C8 side chain displays only a modest
deviation from the plane of the caffeine moiety. The ability of the
C8 side chain to adopt a bent orientation may be an important
requirement for the inhibition of MAO-A. Rigid C8 substituted caf-
feine analogues such as (E)-8-(3-chlorostyryl)caffeine (CSC)
(Fig. 12) are not MAO-A inhibitors while displaying high affinity
binding to MAO-B.26 The observation that CSC does not bind to
MAO-A may be explained by its low degree of flexibility and inabil-
ity to adopt a bent orientation similar to that observed for 3c in the
MAO-A active site.
In conclusion, the sulfanylcaffeine analogues exhibit similar
MAO-B inhibition potencies to those of the previously reported
oxycaffeine analogues with various homologs from both series
exhibiting IC50 values in the nM range.15,16 The attachment of sub-
stituents at C8 of caffeine via a thioether linkage therefore en-
hances MAO-B inhibition activity to a similar extent compared to
attachment via an oxyether. In contrast, C8 substituted amin-
ocaffeines are not suitable for MAO inhibition. Based on the potent
MAO-B inhibition properties of the sulfanylcaffeine analogues,
they may be considered as lead compounds for the development
of reversible MAO-B inhibitors.
The notion that the C8 side chains of the caffeine analogues are
important structural features for MAO-A and -B inhibition is sup-
ported by the observation that caffeine is a weak MAO inhibitor.14
Modeling shows that, in the MAO-B active site, the C8 side chains
of the sulfanylcaffeine analogues may extend into the hydrophobic
4. Experimental section
4.1. Chemicals and instrumentation
Unless otherwise noted, all starting materials were obtained
from Sigma–Aldrich and were used without purification. Proton
(1H) and carbon (13C) NMR spectra were recorded on a Bruker
Avance III 600 spectrometer at frequencies of 600 and 150 MHz,
respectively. All NMR measurements were conducted in CDCl3
and DMSO-d6 and the chemical shifts are reported in parts per mil-
lion (d) downfield from the signal of tetramethylsilane added to the
deuterated solvent. Spin multiplicities are given as s (singlet),
d (doublet), dd (doublet of doublets), t (triplet), q (quartet), qn
O
Cl
N
N
N
O
N
CSC
Figure 12. The structure of (E)-8-(3-chlorostyryl)caffeine (CSC).

H. P. Booysen et al. / Bioorg. Med. Chem. 19 (2011) 7507–7518
7515
(quintet), sept (septet) or m (multiplet). High resolution mass
spectra (HRMS) were obtained on a Waters Synapt G2 instrument
in electrospray ionization (ESI) mode. The HRMS spectrum of 4a
was recorded on a DFS high resolution magnetic sector mass spec-
trometer (Thermo Electron Corporation) in atmospheric pressure
chemical ionization (APCI) mode. Melting points (mp) were mea-
sured with a Stuart SMP10 melting point apparatus and are uncor-
rected. The purity of the synthesized compounds were determined
via HPLC analyses which were conducted with an Agilent 1100
HPLC system equipped with a quaternary gradient pump and an
Agilent 1100 series diode array detector (see Supplementary data).
HPLC grade acetonitrile (Merck) and Milli-Q water (Millipore) were
used for the chromatography. For fluorescence spectrophotometry,
a Varian Cary Eclipse fluorescence spectrophotometer was em-
ployed. Microsomes from insect cells containing recombinant hu-
man MAO-A and -B (5 mg/mL) and kynuramineÁ2HBr were
obtained from Sigma–Aldrich.
C
16H19N4O3S (MH+), 347.1176, found 347.1173; Purity (HPLC):
95%.
4.2.5. 8-{[(4-Chlorophenyl)methyl]sulfanyl}caffeine (3e)
The title compound was prepared from 4-chlorobenzyl mercap-
tan in a yield of 85.6%: mp 169 °C (ethanol). 1H NMR (Bruker
Avance III 600, CDCl3) d 3.36 (s, 3H), 3.56 (s, 3H), 3.73 (s, 3H),
4.40 (s, 2H), 7.26 (m, 4H); 13C NMR (Bruker Avance III 600, CDCl3)
d 27.9, 29.7, 32.2, 36.4, 108.8, 128.8, 130.3, 133.8, 135.2, 148.3,
149.7, 151.5, 154.6; ESI-HRMS m/z: calcd for
C15H16ClN4O2S
(MH+), 351.0682, found 351.0679; Purity (HPLC): 97%.
4.2.6. 8-{[(4-Bromophenyl)methyl]sulfanyl}caffeine (3f)
The title compound was prepared from 4-bromobenzyl mercap-
tan in a yield of 82.0%: mp 166 °C (ethanol). 1H NMR (Bruker
Avance III 600, CDCl3) d 3.35 (s, 3H), 3.55 (s, 3H), 3.72 (s, 3H),
4.38 (s, 2H), 7.21 (d, 2H, J = 8.3 Hz), 7.40 (d, 2H, J = 8.3 Hz); 13C
NMR (Bruker Avance III 600, CDCl3) d 27.8, 29.7, 32.2, 36.4,
108.7, 121.8, 130.6, 131.8, 135.8, 148.3, 149.6, 151.4, 154.5; ESI-
HRMS m/z: calcd for C15H16BrN4O2S (MH+), 395.0177, found
395.0178; Purity (HPLC): 98%.
4.2. Synthesis of C8-substituted thiocaffeine analogues (3a–l)
A solution of NaOH (4 mmol) in 3.5 mL water and 7 mL ethanol
was cooled in an ice bath and the appropriate thiol (4 mmol) was
added. The reaction mixture was stirred and 8-chlorocaffeine
(4 mmol) was added in a single portion to yield a suspension.
The reaction was heated under reflux for 60 min and then cooled
on ice. The white precipitate was collected by filtration and
washed with 30 mL ethanol. The product was recrystallized from
30 mL ethanol at room temperature and the crystals were washed
with 30 mL ethanol.17
4.2.7. 8-{[(4-Fluorophenyl)methyl]sulfanyl}caffeine (3g)
The title compound was prepared from 4-fluorobenzyl mercap-
tan in a yield of 71.6%: mp 175 °C (ethanol). 1H NMR (Bruker
Avance III 600, CDCl3) d 3.35 (s, 3H), 3.55 (s, 3H), 3.72 (s, 3H),
4.40 (s, 2H), 6.96 (t, 2H, J = 8.3 Hz), 7.30 (q, 2H, J = 5.3 Hz); 13C
NMR (Bruker Avance III 600, CDCl3) d 27.8, 29.7, 32.1, 36.4,
108.7, 115.6 (d), 130.6 (d), 132.4 (d), 148.3, 149.8, 151.4, 154.5,
161.4, 163.1; ESI-HRMS m/z: calcd for
C
15H16FN4O2S (MH+),
4.2.1. 8-(Phenylsulfanyl)caffeine (3a)
335.0976, found 335.0972; Purity (HPLC): 95%.
The title compound was prepared from thiophenol in a yield of
65.1%: mp 149 °C (ethanol). 1H NMR (Bruker Avance III 600, CDCl3)
d 3.37 (s, 3H), 3.54 (s, 3H), 3.90 (s, 3H), 7.32 (m, 5H); 13C NMR (Bru-
ker Avance III 600, CDCl3) d 28.0, 29.9, 33.1, 109.5, 128.2, 129.6,
130.5, 130.9, 146.4, 148.0, 151.4, 154.9; ESI-HRMS m/z: calcd for
4.2.8. 8-{[(4-Methoxyphenyl)methyl]sulfanyl}caffeine (3h)
The title compound was prepared from 4-methoxybenzyl mer-
captan in a yield of 90.5%: mp 159 °C (ethanol). 1H NMR (Bruker
Avance III 600, CDCl3) d 3.36 (s, 3H), 3.58 (s, 3H), 3.71 (s, 3H),
3.76 (s, 3H), 4.39 (s, 2H), 6.80 (d, 2H, J = 8.7 Hz), 7.24 (d, 2H,
J = 8.7 Hz); 13C NMR (Bruker Avance III 600, CDCl3) d 27.8, 29.7,
32.1, 37.0, 55.3, 108.6, 114.1, 128.4, 130.2, 148.4, 150.2, 151.5,
C
98%.
14H15N4O2S (MH+), 303.0916, found 303.0912; Purity (HPLC):
4.2.2. 8-(Benzylsulfanyl)caffeine (3b)
154.6, 159.2; ESI-HRMS m/z: calcd for C
16H19N4O3S (MH+),
The title compound was prepared from benzyl mercaptan in a
yield of 59%: mp 149 °C (ethanol). 1H NMR (Bruker Avance III
600, CDCl3) d 3.35 (s, 3H), 3.57 (s, 3H), 3.69 (s, 3H), 4.42 (s, 2H),
7.27 (m, 3H), 7.31 (m, 2H); 13C NMR (Bruker Avance III 600, CDCl3)
d 27.8, 29.7, 32.2, 37.4, 108.7, 127.9, 128.7, 128.9, 136.6, 148.3,
347.1176, found 347.1171; Purity (HPLC): 94%.
4.2.9. 8-[(3-Methylbutyl)sulfanyl]caffeine (3i)
The title compound was prepared from 3-methyl-1-butanethiol
in a yield of 35.3%: mp 79 °C (ethanol). 1H NMR (Bruker Avance III
600, CDCl3) d 0.91 (d, 6H, J = 6.8 Hz), 1.59 (q, 2H, J = 7.9 Hz), 1.69
(sept, 1H, J = 6.8 Hz), 3.23 (t, 2H, J = 7.5 Hz), 3.34 (s, 3H), 3.51 (s,
3H), 3.79 (s, 3H); 13C NMR (Bruker Avance III 600, CDCl3) d 22.1,
27.4, 27.8, 29.6, 30.8, 32.1, 38.5, 108.4, 148.4, 151.3, 151.5, 154.5;
ESI-HRMS m/z: calcd for C13H21N4O2S (MH+), 297.1385, found
297.1382; Purity (HPLC): 97%.
150.0, 151.5, 154.6; ESI-HRMS m/z: calcd for
C15H17O2N4S,
317.1072 (MH+), found 317.1073; Purity (HPLC): 99%.
4.2.3. 8-[(2-Phenylethyl)sulfanyl]caffeine (3c)
The title compound was prepared from phenylethyl mercaptan
in a yield of 22.9%: mp 95 °C (ethanol). 1H NMR (Bruker Avance III
600, CDCl3) d 3.04 (t, 2H, J = 7.9 Hz), 3.36 (s, 3H), 3.49 (t, 2H, J
= 7.9 Hz), 3.55 (s, 3H), 3.78 (s, 3H), 7.21 (m, 3H), 7.29 (m, 2H);
13C NMR (Bruker Avance III 600, CDCl3) d 27.8, 29.7, 32.1, 33.8,
36.0, 108.5, 126.7, 128.5, 139.3, 148.5, 150.9, 151.5, 154.5; ESI-
4.2.10. 8-(Cyclohexylsulfanyl)caffeine (3j)
The title compound was prepared from cyclohexanethiol in a
yield of 37.2%: mp 133 °C (ethanol). 1H NMR (Bruker Avance III
600, CDCl3) d 1.28 (m, 1H), 1.38 (m, 2H), 1.48 (m, 2H), 1.58 (m,
1H), 1.74 (m, 2H), 2.03 (m, 2H), 3.34 (s, 3H), 3.52 (s, 3H), 3.71
(m, 1H), 3.82 (s, 3H); 13C NMR (Bruker Avance III 600, CDCl3) d
25.4, 25.8, 27.8, 29.7, 32.3, 33.4, 47.2, 108.4, 148.4, 150.3, 151.5,
154.6; ESI-HRMS m/z: calcd for C14H21N4O2S (MH+), 309.1385,
found 309.1385; Purity (HPLC): 99%.
HRMS m/z: calcd for
C
16H19N4O2S, 331.1229 (MH+), found
331.1229; Purity (HPLC): 99%.
4.2.4. 8-[(2-Phenoxyethyl)sulfanyl]caffeine (3d)
The title compound was prepared from 2-phenoxyethanethiol
in a yield of 43.9%: mp 114 °C (ethanol). 1H NMR (Bruker Avance
III 600, CDCl3) d 3.37 (s, 3H), 3.53 (s, 3H), 3.63 (t, 2H, J = 6.4 Hz),
3.83 (s, 3H), 4.30 (t, 2H, J = 6.4 Hz), 6.91 (d, 2H, J = 8.3 Hz), 6.94
(t, 1H, J = 7.2 Hz), 7.25 (m, 2H); 13C NMR (Bruker Avance III 600,
CDCl3) d 27.8, 29.7, 31.5, 32.2, 66.3, 108.7, 114.5, 121.3, 129.5,
148.4, 150.3, 151.5, 154.5, 158.1; ESI-HRMS m/z: calcd for
4.2.11. 8-(Cyclopentylsulfanyl)caffeine (3k)
The title compound was prepared from cyclopentanethiol in a
yield of 60.9%: mp 135 °C (ethanol). 1H NMR (Bruker Avance III
600, CDCl3) d 1.63 (m, 4H), 1.76 (m, 2H), 2.15 (m, 2H), 3.34 (s,

7516
H. P. Booysen et al. / Bioorg. Med. Chem. 19 (2011) 7507–7518
3H), 3.51 (s, 3H), 3.80 (s, 3H), 3.99 (qn, 1H); 13C NMR (Bruker
Avance III 600, CDCl3) d 24.6, 27.8, 29.7, 32.2, 33.8, 46.4, 108.2,
148.5, 151.3, 151.5, 154.6; ESI-HRMS m/z: calcd for C13H19N4O2S
(MH+), 295.1229, found 295.1233; Purity (HPLC): 95%.
III 600, DMSO-d6) d 27.1, 29.2, 29.7, 30.8, 32.3, 42.0, 101.8, 125.7,
128.2, 128.3, 141.7, 148.3, 150.9, 152.8, 154.1; ESI-HRMS m/z:
calcd for C17H22N5O2 (MH+), 328.1773, found 328.1774; Purity
(HPLC): 99%.
4.2.12. 8-(Naphthalen-2-ylsulfanyl)caffeine (3l)
4.3.5. 8-[(4-Phenylbutyl)amino]caffeine (4e)
The title compound was prepared from 2-naphthalenethiol in a
yield of 87.7%: mp 175 °C (ethanol). 1H NMR (Bruker Avance III
600, CDCl3) d 3.37 (s, 3H), 3.53 (s, 3H), 3.91 (s, 3H), 7.35 (dd, 1H,
J = 1.9, 8.3 Hz), 7.48 (m, 2H), 7.73 (m, 1H), 7.78 (m, 2H), 7.84 (s,
1H); 13C NMR (Bruker Avance III 600, CDCl3) d 27.9, 29.8, 33.1,
109.5, 126.9, 127.0, 127.4, 127.5, 127.8, 127.9, 129.4, 129.7,
132.6, 133.6, 146.4, 148.0, 151.4, 154.9; ESI-HRMS m/z: calcd for
The title compound was prepared from 4-phenylbutylamine
and 8-chlorocaffeine in a yield of 63.0%: mp 179–180 °C (ethanol).
1H NMR (Bruker Avance III 600, DMSO-d6) d 1.59 (m, 4H), 2.60 (t,
2H, J = 7.2 Hz), 3.13 (s, 3H), 3.30 (s, 3H), 3.32 (m, 2H), 3.51 (s,
3H), 6.94 (t, 1H, J = 5.6 Hz), 7.14 (t, 1H, J = 7.2 Hz), 7.18 (d, 2H,
J = 7.2 Hz), 7.25 (t, 2H, J = 7.2 Hz); 13C NMR (Bruker Avance III
600, DMSO-d6) d 27.1, 28.2, 28.8, 29.2, 29.7, 34.8, 42.2, 101.7,
125.6, 128.2, 128.3, 142.1, 148.3, 150.9, 152.8, 154.1; ESI-HRMS
m/z: calcd for C18H24N5O2 (MH+), 342.1930, found 342.1929; Purity
(HPLC): 99%.
C
94%.
18H17N4O2S (MH+), 353.1072, found 353.1074; Purity (HPLC):
4.3. Synthesis of C8-substituted aminocaffeine analogues (4a–
h)
4.3.6. 8-{[2-(Pyridin-2-yl)ethyl]amino}caffeine (4f)
The title compound was prepared from 2-(2-pyridyl)ethylamine
and 8-chlorocaffeine in a yield of 20.4%: mp 196–197 °C (ethanol).
1H NMR (Bruker Avance III 600, DMSO-d6) d 3.03 (t, 2H, J = 7.2 Hz),
3.13 (s, 3H), 3.31 (s, 3H), 3.50 (s, 3H), 3.65 (q, 2H, 6.7 Hz), 7.10 (t,
1H, J = 5.6 Hz), 7.20 (t, 1H, J = 5.6 Hz), 7.26 (d, 1H, J = 7.5 Hz), 7.69
(t, 1H, J = 7.5 Hz), 8.49 (d, 1H, J = 4.1 Hz); 13C NMR (Bruker Avance
III 600, DMSO-d6) d 27.1, 29.2, 29.7, 37.5, 42.4, 101.8, 121.5, 123.2,
136.4, 148.3, 149.0, 150.9, 152.8, 153.9, 159.1; ESI-HRMS m/z:
calcd for C15H19N6O2 (MH+), 315.1569, found 315.1570; Purity
(HPLC): 98%.
A mixture of 8-chlorocaffeine (2 mmol) and the appropriate
amine (10 mmol) was heated under reflux (175–180 °C) for 3 h.
The reaction was cooled to room temperature and treated with
50 mL acetic acid (5%). The resulting suspension was stirred for
15 min at room temperature and the precipitate was collected by
filtration. The product was dried at 60 °C and recrystallized twice
from ethanol (30 mL) at 0 °C.18
4.3.1. 8-(Phenylamino)caffeine (4a)
The title compound was prepared from aniline and 8-chlorocaf-
feine in a yield of 24.2%: mp 164–265 °C (ethanol). 1H NMR (Bruker
Avance III 600, DMSO-d6) d 3.15 (s, 3H), 3.35 (s, 3H), 3.74 (s, 3H),
6.96 (t, 1H, J = 7.5 Hz), 7.29 (t, 2H, J = 7.5 Hz), 7.67 (d, 2H,
J = 8.3 Hz), 9.07 (s, 1H); 13H NMR (Bruker Avance III 600, DMSO-
d6) d 27.3, 29.4, 30.5, 102.0, 118.1, 121.7, 128.7, 140.0, 147.2,
149.3, 150.9, 153.3; APCI-HRMS m/z: calcd for C14H15N5O2 (M+),
285.1226, found 285.1230; Purity (HPLC): 98%.
4.3.7. 8-{[2-(3-Chlorophenyl)ethyl]amino}caffeine (4g)
The title compound was prepared from 2-(3-chlorophenyl)eth-
anamine and 8-chlorocaffeine in a yield of 48.5%: mp 111–113 °C
(ethanol). 1H NMR (Bruker Avance III 600, DMSO-d6) d 2.88 (t,
2H, J = 7.2 Hz), 3.13 (s, 3H), 3.32 (s, 3H), 3.50 (s, 3H), 3.52 (m,
2H), 7.10 (t, 1H, J = 5.6 Hz), 7.18 (d, 1H, J = 7.5 Hz), 7.24 (d, 1H,
J = 8.3 Hz), 7.29 (d, 1H, J = 7.5 Hz), 7.31 (s, 1H); 13C NMR (Bruker
Avance III 600, DMSO-d6) d 27.1, 29.2, 29.7, 34.8, 43.7, 101.8,
126.1, 127.5, 128.6, 130.1, 132.9, 142.0, 148.3, 150.9, 152.8,
153.8; ESI-HRMS m/z: calcd for C16H19N5O2Cl (MH+), 348.1227,
found 348.1225; Purity (HPLC): 98%.
4.3.2. 8-(Benzylamino)caffeine (4b)
The title compound was prepared from benzylamine and 8-
chlorocaffeine in a yield of 74.0%: mp 230 °C (ethanol). 1H NMR
(Bruker Avance III 600, DMSO-d6) d 3.14 (s, 3H), 3.35 (s, 3H), 3.58
(s, 3H), 4.53 (d, 2H, J = 5.6 Hz), 7.23 (t, 1H, J = 7.5 Hz), 7.32 (t, 2H,
J = 7.5 Hz), 7.36 (d, 2H, J = 7.5 Hz), 7.56 (t, 1H, J = 5.6 Hz); 13H
NMR (Bruker Avance III 600, DMSO-d6) d 27.1, 29.2, 29.8, 45.7,
102.0, 126.9, 127.4, 128.3, 139.6, 148.2, 150.9, 152.9, 154.0; ESI-
4.3.8. 8-(Cyclopentylamino)caffeine (4h)
The title compound was prepared from cyclopentylamine and
8-chlorocaffeine in a yield of 38.6%: mp 217–218 °C (ethanol). 1H
NMR (Bruker Avance III 600, DMSO-d6) d 1.52 (m, 4H), 1.68 (m,
2H), 1,93 (m, 2H), 3.13 (s, 3H), 3.31 (s, 3H), 3.53 (s, 3H), 4.10 (m,
1H), 6.76 (d, 1H, J = 7.2 Hz); 13H NMR (Bruker Avance III 600,
DMSO-d6) d 23.4, 27.1, 29.2, 29.8, 32.4, 54.2, 101.7, 148.3, 150.9,
HRMS m/z: calcd for
C
15H18N5O2 (MH+), 300.1460, found
300.1459; Purity (HPLC): 99%.
4.3.3. 8-[(2-Phenylethyl)amino]caffeine (4c)
152.8, 153.8; ESI-HRMS m/z: calcd for
278.1617, found 278.1612; Purity (HPLC): 96%.
C
13H20N5O2 (MH+),
The title compound was prepared from 2-phenylethylamine
and 8-chlorocaffeine in a yield of 68.3%: mp 221 °C (ethanol). 1H
NMR (Bruker Avance III 600, DMSO-d6) d 2.88 (t, 2H, J = 7.2 Hz),
3.14 (s, 3H), 3.32 (s, 3H), 3.49 (m, 2H), 3.51 (s, 3H), 7.11 (t, 1H,
J = 5.3 Hz), 7.19 (t, 1H, J = 7.2 Hz), 7.22 (d, 2H, J = 7.5 Hz), 7.29 (t,
2H, J = 7.5 Hz); 13H NMR (Bruker Avance III 600, DMSO-d6) d 27.1,
29.2, 29.7, 35.3, 44.1, 101.8, 126.1, 128.3, 128.7, 139.4, 148.3,
150.9, 152.9, 153.9; ESI-HRMS m/z: calcd for C16H20N5O2 (MH+),
314.1617, found 314.1621; Purity (HPLC): 99%.
4.4. Methylation of the C8-substituted aminocaffeine analogues
(5a–b)
Potassium hydroxide (0.05 g) was powderized and suspended
in 5 mL DMSO. The resulting mixture was stirred for 30 min at
room temperature and the aminocaffeine analogue (3 mmol), dis-
solved in DMSO (5 mL), was added. The reaction was heated to
40 °C (in order for the aminocaffeine analogue to remain in solu-
tion) and iodomethane (0.8 mmol) was added. Stirring of the reac-
tion was continued and another portion of iodomethane
(0.8 mmol) was added every 20 min until silica gel TLC (petroleum
ether/ethyl actetate 30:70) indicated completion of the reaction.
The pH of the reaction was also continually measured, and when
acidic (pH paper), another portion of potassium hydroxide
(0.05 g) was added. Upon completion, the reaction was cooled to
4.3.4. 8-[(3-Phenylpropyl)amino]caffeine (4d)
The title compound was prepared from 3-phenylpropylamine
and 8-chlorocaffeine in a yield of 76.8%: mp 204–205 °C (ethanol).
1H NMR (Bruker Avance III 600, DMSO-d6) d 1.88 (qn, 2H, 7.5 Hz),
2.64 (t, 2H, J = 7.5 Hz), 3.13 (s, 3H), 3.30 (s, 3H), 3. 32 (m, 2H),
3.52 (s, 3H), 6.98 (t, 1H, J = 5.3 Hz), 7.16 (t, 1H, J = 7.2 Hz), 7.22
(d, 2H, J = 7.5 Hz), 7.26 (t, 2H, J = 7.5 Hz); 13H NMR (Bruker Avance

H. P. Booysen et al. / Bioorg. Med. Chem. 19 (2011) 7507–7518
7517
room temperature and water (250 mL) was added. The resulting
solution was incubated for several days at 4 °C and the formed
crystals were collected by filtration.
sium phosphate buffer (100 mM, pH 7.4, made isotonic with KCl)
with MAO-A and -B (0.015–0.03 mg/mL) for periods of 0, 15, 30,
60 min at 37 °C. The concentrations of the inhibitors used were
2-fold their measured IC50 values for the inhibition of the respec-
4.4.1. 8-[Methyl(2-phenylethyl)amino]caffeine (5a)
tive MAO enzymes and were: 5.22
the studies with MAO-A, and 0.32
l
l
M (3f) and 11.56 M (4g) for
M (3f) and 5.94 lM (5b) for
l
The title compound was prepared from 8-[(2-phenyl-
ethyl)amino]caffeine (4c) and iodomethane in a yield of 57.7%:
mp 103 °C (ethanol). 1H NMR (Bruker Avance III 600, DMSO-d6) d
2.88 (t, 2H, J = 7.5 Hz), 2.98 (s, 3H), 3.16 (s, 3H), 3.33 (s, 3H), 3.47
(t, 2H, J = 7.5 Hz), 3.59 (s, 3H), 7.18 (m, 1H), 7.25 (m, 4H); 13C
NMR (Bruker Avance III 600, DMSO-d6) d 27.3, 29.3, 32.5, 33.0,
38.6, 54.5, 103.8, 126.1, 128.3, 128.8, 139.0, 147.1, 150.9, 153.5,
156.6; ESI-HRMS m/z: calcd for C17H22N5O2 (MH+), 328.1774,
found 328.1734; Purity (HPLC): 99%.
the studies with MAO-B. To compensate for a potential time-
dependent loss of enzyme activity, the enzyme preincubations
were firstly incubated at 37 °C and the inhibitors were subse-
quently added at different time points. These time points were se-
lected as to ensure that all enzyme preparations were preheated at
37 °C for exactly 60 min, irrespective of the time period (0–60 min)
for which the enzyme preparations were preincubated in the pres-
ence of the test inhibitor. These reactions were diluted 2-fold by
the addition of kynuramine at concentrations of 45 and 30
for the incubations with MAO-A and -B, respectively, and the
resulting reactions (500 L final volume) were incubated at 37 °C
for a further 15 min. The final enzyme concentration in these reac-
tions was 0.0075–0.015 mg/mL and the concentrations of the se-
lected inhibitors were approximately equal to their IC50 values
for the inhibition of the respective isozymes. The reactions were
lM
4.4.2. 8-[Methyl(4-phenylbutyl)amino]caffeine (5b)
The title compound was prepared from 8-[(4-phenylbu-
tyl)amino]caffeine (4e) and iodomethane in a yield of 49.9%: mp
114 °C (ethanol). 1H NMR (Bruker Avance III 600, DMSO-d6) d
1.57 (m, 4H), 2.57 (t, 2H, J = 7.15 Hz), 2.90 (s, 3H), 3.16 (s, 3H),
3.26 (t, 2H, J = 7.15 Hz), 3.32 (s, 3H), 3.65 (s, 3H), 7.15 (m, 3H),
7.24 (t, 2H, J = 7.91 Hz); 13C NMR (Bruker Avance III 600, DMSO-
d6) d 26.3, 27.3, 27.9, 29.3, 32.6, 34.7, 38.4, 52.5, 103.8, 125.7,
128.2, 128.2, 142.0, 147.1, 150.9, 153.5, 156.9; ESI-HRMS m/z:
calcd for C19H26N5O2 (MH+), 356.2087, found 356.2088; Purity
(HPLC): 98%.
l
terminated with the addition of 400 lL NaOH (2 N) and 1000 lL
distilled water and the rates of MAO catalyzed generation of 4-
hydroxyquinoline were measured and calculated as described
above. All measurements were carried out in triplicate and are ex-
pressed as mean SD.23
4.5. IC50 determinations for the inhibition of human MAO
4.7. Construction of Lineweaver–Burk plots
Microsomal preparations form insect cells containing recombi-
nant human MAO-A and -B (5 mg/mL) served as enzyme sources
and all enzymatic reactions were conducted in potassium phos-
phate buffer (100 mM, pH 7.4, made isotonic with KCl) to a final
A set consisting of four Lineweaver–Burk plots were constructed
for the inhibition of MAO-A and -B by the selected inhibitor 3f. One
plot was constructed in the absence of inhibitor while three plots
were constructed in the presence of three different concentrations
of 3f each. These concentrations were 1.31–5.22 and 0.04–
volume of 500
substrate, kynuramine, at concentrations of 45
the incubations with MAO-A and -B, respectively, various concen-
trations of the test inhibitor (0–100 M) and the MAO enzymes
l
L.19 The reactions contained the MAO-A/B mixed
lM and 30
lM for
0.16
lM for the inhibition studies with MAO-A and -B, respectively.
Four different kynuramine concentrations (15–90
l
M) were em-
l
ployed for each plot and the concentrations of recombinant human
MAO-A and -B used were 0.015 mg/mL. The initial MAO catalytic
rates were measured as described above. Linear regression analysis
was performed using GraphPad Prism.23
(0.0075 mg/mL). The enzyme activities employed for the IC50 value
determinations were 24–28 nmoles 4-hydroxyquinoline formed/
min/mg protein for MAO-A and 6–8 nmoles 4-hydroxyquinoline
formed/min/mg protein for MAO-B. Stock solutions of the test
inhibitors were prepared in DMSO and added to the reactions to
yield a final concentration of 4% (v/v) DMSO. The enzyme reactions
were incubated at 37 °C for 20 min and then terminated with the
4.8. Molecular modeling studies
The modeling studies were carried out with the Windows based
Discovery Studio 1.7 molecular modeling software (Accelrys).23,27
For this purpose the crystallographic structures of MAO-A co-crys-
tallized with harmine (PDB code: 2Z5X)3 and MAO-B co-crystal-
lized with safinamide (PDB code: 2V5Z)4 were obtained from the
Brookhaven Protein Data Bank (www.rcsb.org/pdb). The proton-
ation states of the ionizable amino acids residues were calculated
at pH 7.4 and hydrogen atoms were added to the receptor models.
The valences of the FAD cofactors (oxidized state) and co-crystal-
lized ligands were corrected and hydrogen atoms were added
according to the appropriate protonation states at pH 7.4. The
structures were typed automatically with the Momany and Rone
CHARMm forcefield, the backbone of the protein was constrained
and the structures were subjected to a three step energy minimiza-
tion. The first step was a steepest descent minimization which was
followed by conjugate gradient minimization. For both protocols
the termination criteria was set to a maximum of 2500 steps or a
minimum value of 0.1 for the root mean square of the energy gra-
dient. The final step was an adopted basis Newton–Rapheson min-
imization and the termination criteria was set to a maximum of
5000 steps or a minimum value for the root mean square of the en-
ergy gradient of 0.01. For these minimization steps the implicit
generalized Born solvation model with simple switching was em-
addition of 400 lL NaOH (2 N) and 1000 lL distilled water. After
centrifugation at 16,000g for 10 min, the fluorescence of the MAO
generated 4-hydroxyquinoline in the supernatant fractions were
measured (kex = 310 nm, kem = 400 nm). To determine the concen-
trations of 4-hydroxyquinoline, a linear calibration curve was con-
structed from solutions of 4-hydroxyquinoline (0.047–1.50
potassium phosphate buffer. The calibration standards were pre-
pared to a volume of 500 L and contained 4% DMSO, 400 L NaOH
(2 N) and 1000 L distilled water. The initial rate of MAO catalysis
lM) in
l
l
l
was plotted versus the logarithm of the inhibitor concentration to
obtain a sigmoidal dose–response curve. Each curve was con-
structed from 6 different inhibitor concentrations spanning at least
3 orders of magnitude. These data were fitted to the one site com-
petition model incorporated into the GraphPad Prism software and
the IC50 values were determined in triplicate and are expressed as
mean standard deviation (SD).
4.6. Time-dependent inhibition studies
The reversibility of MAO inhibition was examined by determin-
ing the time-dependence of inhibition of three selected inhibitors,
3f, 4g and 5b. The selected inhibitors were preincubated in potas-

7518
H. P. Booysen et al. / Bioorg. Med. Chem. 19 (2011) 7507–7518
ployed with the dielectric constant set to 4. For both the MAO-A
and -B models, the crystal water molecules were removed with
the exception of three active site waters in each model. The
X-ray crystallographic structures of MAO-B shows that three active
site water molecules (HOH 1155, 1170 and 1351; A-chain) are con-
served, all located in the vicinity of the FAD cofactor.4 In the MAO-
A model, the crystal waters HOH 710, 718 and 739 which occupies
the analogous positions in the MAO-A active site compared to
those cited above for MAO-B, were retained. The co-crystallized li-
gands and the backbone constraints were subsequently removed
from the models and the binding sites were identified by a flood-
filling algorithm. The structures of 3a–c and 4c were constructed
within Discovery Studio, and their hydrogen atoms were added
according to the appropriate protonation states at pH 7.4. The
geometries of the ligands were briefly optimized in Discovery Stu-
dio using a fast Dreiding-like forcefield (1000 iterations) and the
atom potential types and partial charges were assigned with the
Momany and Rone CHARMm forcefield. Docking of the ligands
was carried out with the LigandFit application of Discovery Studio
and the docking solutions were refined using the Smart Minimizer
algorithm. The parameters for the docking runs were set to their
default values, ten possible binding solutions were computed for
each docked ligand and the best-ranked binding conformation of
each ligand was determined according to the DockScore values.
The illustrations were prepared in PyMOL.28 It is interesting to note
that among the 10 best ranked binding orientations, there were
orientations which exhibited a reversed binding mode with the
caffeine ring directed towards the entrance of the MAO-A and -B
active sites while the C8 sulfanyl and amino side chains project to-
wards the FAD cofactor. Based on the low DockScore values these
orientations were however deemed unlikely.
Supplementary data
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.bmc.2011.10.036.
References and notes
1. Binda, C.; Newton-Vinson, P.; Hubálek, F.; Edmondson, D. E.; Mattevi, A. Nat.
Struct. Biol. 2002, 9, 22.
2. Shih, J. C.; Chen, K.; Ridd, M. J. Annu. Rev. Neurosci. 1999, 22, 197.
3. Son, S.-Y.; Ma, J.; Kondou, Y.; Yoshimura, M.; Yamashita, E.; Tsukihara, T. Proc.
Natl. Acad. Sci. U.S.A. 2008, 105, 5739.
4. Binda, C.; Wang, J.; Pisani, L.; Caccia, C.; Carotti, A.; Salvati, P.; Edmondson, D.
E.; Mattevi, A. J. Med. Chem. 2007, 50, 5848.
5. Youdim, M. B. H.; Edmondson, D.; Tipton, K. F. Nat. Rev. Neurosci. 2006, 7, 295.
6. Nicotra, A.; Pierucci, F.; Parvez, H.; Senatori, O. Neurotoxicology 2004, 25, 155.
7. Fowler, J. S.; Volkow, N. D.; Wang, G. J.; Logan, J.; Pappas, N.; Shea, C.;
MacGregor, R. Neurobiol. Aging 1997, 18, 431.
8. Youdim, M. B. H.; Collins, G. G. S.; Sandler, M.; Bevan-Jones, A. B.; Pare, C. M.;
Nicholson, W. J. Nature 1972, 236, 225.
9. Collins, G. G. S.; Sandler, M.; Williams, E. D.; Youdim, M. B. H. Nature 1970, 225,
817.
10. Di Monte, D. A.; DeLanney, L. E.; Irwin, I.; Royland, J. E.; Chan, P.; Jakowec, M.
W.; Langston, J. W. Brain. Res. 1996, 738, 53.
11. Finberg, J. P.; Wang, J.; Bankiewich, K.; Harvey-White, J.; Kopin, I. J.; Goldstein,
D. S. J. Neural Transm. Suppl. 1998, 52, 279.
12. Fernandez, H. H.; Chen, J. J. Pharmacotherapy 2007, 27, 174S.
13. Youdim, M. B. H.; Bakhle, Y. S. Br. J. Pharmacol. 2006, 147, S287.
14. Van der Walt, E. M.; Milczek, E. M.; Malan, S. F.; Edmondson, D. E.; Castagnoli,
N., Jr.; Bergh, J. J.; Petzer, J. P. Bioorg. Med. Chem. Lett. 2009, 19, 2509.
15. Strydom, B.; Malan, S. F.; Castagnoli, N.; Bergh, J. J.; Petzer, J. P. Bioorg. Med.
Chem. 2010, 18, 1018.
16. Strydom, B.; Bergh, J. J.; Petzer, J. P. Eur. J. Med. Chem. 2011, 46, 3474.
17. Long, L. M. J. Am. Chem. Soc. 1947, 69, 2939.
18. Cramer, L. Chem. Ber. 1894, 27, 3089.
19. Novaroli, L.; Reist, M.; Favre, E.; Carotti, A.; Catto, M.; Carrupt, P. A. Bioorg. Med.
Chem. 2005, 13, 6212.
20. Prins, L. H. A.; Petzer, J. P.; Malan, S. F. Eur. J. Med. Chem. 2010, 45, 4458.
21. Tipton, K. F.; Boyce, S.; O’Sullivan, J.; Davey, G. P.; Healy, J. Curr. Med. Chem.
2004, 11, 1965.
22. Fowler, J. S.; Volkow, N. D.; Logan, J.; Wang, G.; MacGregor, R. R.; Schlyer, D.;
Wolf, A. P.; Pappas, N.; Alexoff, D.; Shea, C.; Dorflinger, E.; Kruchowy, L.; Yoo,
K.; Fazzini, E.; Patlak, C. Synapse 1994, 18, 86.
23. Manley-King, C. I.; Bergh, J. J.; Petzer, J. P. Bioorg. Med. Chem. 2011, 19, 261.
24. Binda, C.; Mattevi, A.; Edmondson, D. E. J. Biol. Chem. 2002, 277, 23973.
25. Hubálek, F.; Binda, C.; Khalil, A.; Li, M.; Mattevi, A.; Castagnoli, N., Jr.;
Edmondson, D. E. J. Biol. Chem. 2005, 280, 15761.
26. Vlok, N.; Malan, S. F.; Castagnoli, N., Jr.; Bergh, J. J.; Petzer, J. P. Bioorg. Med.
Chem. 2006, 14, 3512.
27. Accelrys Discovery Studio 1.7, Accelrys Software Inc., San Diego, CA, USA. 2006,
http://www.accelrys.com.
Acknowledgments
The NMR spectra were recorded by André Joubert of the SASOL
Centre for Chemistry, North-West University while the MS spec-
tra were recorded by the Mass Spectrometry Service, University
of the Witwatersrand and the Mass Spectrometry Unit, Stel-
lenbosch University. This work was supported by grants from
the National Research Foundation and the Medical Research
Council, South Africa.
28. DeLano, W. L. The PyMOL Molecular Graphics System. DeLano Scientific, San
Carlos, USA, 2002.