Orientation of oxazolidinones in the active site of monoamine oxidase

Article abstract of DOI:10.1016/j.bcp.2005.05.001

Oxazolidinone inhibitors of monoamine oxidase (MAO) and oxazolidinone antibacterials are two distinct classes of drug, often with linear structures and overlapping activities for some derivatives. By synthesizing novel dimerised derivatives with identical

Full text of DOI:10.1016/j.bcp.2005.05.001

Biochemical Pharmacology 70 (2005) 407–416
www.elsevier.com/locate/biochempharm
Orientation of oxazolidinones in the active site of monoamine oxidase
Tadeusz Z.E. Jones a, Paul Fleming b, Charles J. Eyermann b,
Michael B. Gravestock b, Rona R. Ramsay a,
*
a Centre for Biomolecular Sciences, University of St. Andrews, North Haugh, St. Andrews, Fife KY16 9ST, UK
b Infection Discovery, AstraZeneca R&D Boston, 35 Gatehouse Drive, Waltham, MA 02451, USA
Received 11 March 2005; accepted 4 May 2005
Abstract
Oxazolidinone inhibitors of monoamine oxidase (MAO) and oxazolidinone antibacterials are two distinct classes of drug, often with
linear structures and overlapping activities for some derivatives. By synthesizing novel dimerised derivatives with identical substitution of
the two C-5 side chains, we have obtained experimental evidence for the orientation of oxazolidinones in the active site of MAO A. Two
types of spectral changes, either increasing the absorbance at 510 nm or decreasing it at 495 nm depending on the group nearest to the
flavin cofactor, were seen on ligand binding to MAO A. Side chain derivatives with amine substituents are very poor substrates so that it
was possible to examine the spectral change due to binding of a substrate before reduction of the flavin occurred. Binding of these amino
derivative substrates to MAO A induced a spectral change characterized by a strong decrease in absorbance at 495 nm. These substrates
reduced the enzyme fully without any trace of a semiquinone intermediate. Only oxazolidinone inhibitors with a bromo-imidazole
substituent increased the yield of semiquinone intermediate obtained during chemical reduction. In accord with the experimental data,
results of docking experiments showed that binding of the oxazolidinone ring in the aromatic cage close to the flavin was favored and that
the nitrogen of the derivatives that were substrates was within van der Waals distance of N-5 of the flavin.
# 2005 Elsevier Inc. All rights reserved.
Keywords: Oxazolidinone; Monoamine oxidase; Difference spectra; Flexible docking; Kinetics
1. Introduction
expected to bind close to the flavin. The structure of MAO
A is now available [4], so modelling of inhibitors into the
active site was done to reveal the interactions responsible
for the change in spectral and redox properties.
Monoamine oxidase A (MAO A; EC 1.4.3.4) is the key
enzyme in the nervous system for catabolism of the
catecholamine neurotransmitters and also in the intestine
for scavenging ingested biogenic amines to prevent them
reaching the circulation. MAO A catalyses the oxidation of
a wide range of amines to the imine with concomitant
reduction of the cysteinyl-FAD cofactor that is then reox-
idised by oxygen producing H2O2.
The proximity of the ligands to the flavin also influences
the redox properties of the FAD in MAO A. In the absence
of ligands, chemical reduction of MAO generates a spec-
trum characterized by a peak at 412 nm, typical of the
anionic semiquinone form of the flavin (or possibly of a
tyrosyl radical [5]), as an intermediate before full reduc-
tion. Inhibitors, such as D-amphetamine, increase the
amount of intermediate seen and prevent full reduction
[2,6,7]. In the presence of substrates, no intermediate is
ever seen and reduction proceeds directly to the fully
reduced form. Thermodynamic and kinetic studies have
shown that substrates increase the redox potential of the
flavin [8] and that substrate bound to the reduced enzyme
accelerates the rate of reoxidation [9]. In contrast, inhibi-
tors do not change the redox potential and they inhibit
reoxidation. Since the substrate b-phenylethylamine dif-
fers from the inhibitor, D-amphetamine, by only a methyl
The absorbance spectrum of the FAD in MAO A alters
when ligands are bound [1–3]. The alteration is sensitive to
minor changes in the ligand structure, so provides a tool for
examining the interaction between inhibitors and the active
site of MAO A. In this paper, we use this tool to determine
differences in the interactions of oxazolidinone derivatives
containing three different side chain substituents in the area
Abbreviation: MAO A, monoamine oxidase A
* Corresponding author. Tel.: +44 1334 463411; fax: +44 1334 463400.
E-mail address: rrr@st-and.ac.uk (R.R. Ramsay).
0006-2952/$ – see front matter # 2005 Elsevier Inc. All rights reserved.
doi:10.1016/j.bcp.2005.05.001

408
T.Z.E. Jones et al. / Biochemical Pharmacology 70 (2005) 407–416
group a to the amino group, this indicates that interactions
close to the N-5 of the FAD may be responsible for the
differences. Here, we look at the influence of acetamide,
hydroxyl, and amine substituents on the yield of the
semiquinone in MAO A.
used in this investigation, shown in Fig. 1, were available
from a separate study designed to assess the possibility of
accessing adjacent antibacterial ribosomal binding sites, so
their binding characteristics were compared with mono-
meric compounds known to interact with the active site of
MAO A. In the event, they showed modest activity against
the Gram-positive bacterial species Staphlococcus aureus
and Streptococcus pneumoniae only with minimum inhi-
bitory concentrations in the range of 4–8 mg/L.
Oxazolidinone derivatives are well-known inhibitors of
both MAO A and MAO B [10–15]. They also include a new
class of antibacterial agents that act by preventing protein
synthesis by binding to the bacterial ribosome prior to
initiation of synthesis [16]. The structure–activity relation-
ships of the two activities are similar but differentiation can
be made [14], so that this new class of antibacterials can be
developed devoid of risk from the pressor response to
tyramine or amines in the diet that can occur with irre-
versible MAO A inhibition [17,18]. Most oxazolidinones
inhibit MAO A reversibly and competitively, although
time-dependent inhibition of MAO B and irreversible
inhibition have also been observed for some derivatives
[19]. Oxazolidinones are linear heterocyclic molecules
which, in principle, could bind in either orientation in
the active site of MAO A. The dimerised derivatives 10–12
Structure–function studies on MAO demonstrated that
good binding is obtained with ligands that are generally
lipophilic with an aromatic ring and a nitrogen (or oxygen
or sulfur) close to the ring. Crystal structures of MAO A [4]
and MAO B [20] have now provided a clear picture of the
active site as a long, narrow hydrophobic cavity penetrat-
ing deep into the molecule to the flavin where the substrate
is aligned by two tyrosines (Y407 and Y444 in MAO A) so
that the amine is close to the N-5 of the flavin for catalysis.
In this study, we provide experimental evidence for the
orientation of oxazolidinones in this active site of MAO A.
In addition, to seek an explanation for the ligand effects on
Fig. 1. Structures of oxazolidinone derivatives.

T.Z.E. Jones et al. / Biochemical Pharmacology 70 (2005) 407–416
409
the flavin cofactor, seen as spectral changes on ligand
binding and changes in the redox properties, we have
docked the oxazolidinone derivatives in a model of the
MAO A binding site based on its known structure.
dichloride to afford N,N0-((1,5-dioxopentane-1,5-diyl)-
bis{piperazine-4,1-diyl(3-fluoro-4,1-phenylene)[(5S)-2-
oxo-1,3-oxazolidine-3,5-diyl]methylene})diacetamide,
compound 10. Compound 11 was prepared (Chart 2) by a
multi-step sequence beginning with the addition of homo-
piperazine to two equivalents of 3,4-difluoronitrobenzene
to afford the dinitro compound. Reduction of the dinitro
compound to the dianiline was accomplished using Pd/C in
the presence of ammonium formate. The dianiline com-
pound was then converted to its dibenzyloxycarbamate
which was then treated with LiHMDS and R-(ꢀ)-glycidyl
butyrate to afford (5R,50R)-3,30-[1,4-diazepane-1,4-diyl-
bis(3-fluoro-4,1-phenylene)]bis[5-(hydroxymethyl)-1,3-
oxazolidin-2-one], compound 11. Compound 12 was pre-
pared in an analogous manner to 10 but with amine 3 as the
reactant and succinyl rather than glutaryl dichloride as
acylating agent. All other chemicals were purchased from
Sigma Chemical Company. All final compounds were
isolated by standard procedures, purified by reverse phase
chromatography and characterized by high resolution mass
spectrum and 300 MHz NMR, which showed purities
>98%.
2. Materials and methods
2.1. Materials
Monoamine oxidase A (human liver form) was purified
after expression in S. cerevisiae [21]. The enzyme was
stored at ꢀ20 8C in a solution of 50 mM potassium phos-
phate, pH 7.2, 0.8% n-octyl-b-D-glucopyranoside, 1.5 mM
dithiothreitol, and 0.5 mM D-amphetamine, to which was
then added an equal volume of glycerol. Before use,
dithiothreitol, D-amphetamine, and glycerol were removed
by gel filtration in a spin column of G-50 Sephadex
equilibrated with 50 mM potassium phosphate, pH 7.2,
containing 0.05% Brij.
The 4-(N-morpholino)-aryl, 4-(40-bromo-N-imidazoyl)-
aryl, and 4-(N0-2-(pyrazino)-N-piperazinyl)-aryl oxazoli-
dinone derivatives were synthesized by known methods
[22,23]. The novel dimer compound 10 (Chart 1) was
synthesized by linking two equivalents of the known
N-{[(5S)-3-(3-fluoro-4-piperazin-1-ylphenyl)-2-oxo-1,3-
oxazolidin-5-yl]methyl}acetamide [22] with glutaryl
2.2. MAO A assay and spectral titrations
Initial rates of oxidation were measured at 30 8C in
50 mM potassium phosphate, pH 7.2, containing 0.05%
Chart 1.
Chart 2.

410
T.Z.E. Jones et al. / Biochemical Pharmacology 70 (2005) 407–416
Triton X-100. Ki values were determined using a substrate
range of 0.1–0.9 mM kynuramine [3] or 0.03–1 mM
1-methyl-4-(1-methylpyrrol-2-yl)-1,2,3,6-tetrahydropyr-
idine [24] at six different inhibitor concentrations over a
10-fold range. The formation of the product was followed
phenyl, 4-(40-bromo-N-imidazoyl)-phenyl, and 4-(N0-2-
(pyrazino)-N-piperazinyl)-phenyl (Fig. 1). The N-aryl sub-
stituents differ in electronic and spacial attributes. The
morpholino group is non-aromatic so is not planar and
lacks the p electron system of the 40-bromo-N-imidazoyl
and N0-2-(pyrazino)-N-piperazinyl groups. Since an aro-
matic group favors binding for all MAO ligands, the
aromatic, nitrogen-containing groups may favor location
of that part of the ligand between the two tyrosines (Y407
and Y444 in MAO A) shown by crystal structures to orient
ligands by stacking [20]. Novel compounds 10–12 were
studied to allow assessment of the orientation of the
oxazolidinone molecules in the long channel that forms
the active site of MAO A. These are essentially two
oxazolidinone molecules linked by aryl substituents
‘‘head-to-head’’ or by C-5 side chains ‘‘tail-to-tail’’, so
that both ends of the molecule are identical. The effects of
these compounds on MAO A allow definition of the
influence of the acetamidomethyl, hydroxymethyl, and
N-morpholino groups when bound near the flavin.
spectrophotometrically at 314 nm (e = 12,300 Mꢀ1 cmꢀ1
)
or 343 nm (e = 24,000 Mꢀ1 cmꢀ1), respectively, and the
kinetic constants determined using the Shimadzu software.
Data are presented in a Lineweaver–Burke plot to illustrate
the unchanged Vmax
.
Spectra were recorded in a Shimadzu UV-2101PC spec-
trophotometer in an anaerobic cuvette containing MAO A
at 10–15 mM in 50 mM potassium phosphate, pH 7.2,
containing 0.05% Brij. Inhibitor-induced spectral changes
are known to be fast [1], so spectra were recorded as soon
as the absorbance had stabilized, approximately 5 min after
mixing and repositioning the cuvette. Aliquots of dithio-
nite were then added to the anaerobic enzyme-inhibitor
mixture to reduce the flavin. Spectra were recorded 10 min
after the additions to allow redox equilibration.
The oxazolidinone derivatives used here are all rever-
sible competitive inhibitors of MAO A as shown for line-
zolid 1 in Fig. 2. The observed inhibition increased slightly
with time but was maximal within 2 min. The Ki values for
the derivatives were determined from the apparent Km
values at varying inhibitor concentrations.
2.3. Molecular modelling
Docking studies of oxazolidinones to the homology
model of MAO A were performed using the QXP/FLO
software [25]. Residues in the substrate cavity were allowed
full conformational flexibility in the docking studies.
The Ki values for the compounds vary over a more than
500-fold range (Table 1). In the morpholino series, the
hydroxyl derivative 2 is the most effective inhibitor at
1.2 mM, with the amino derivative 3 100-fold less effective.
A similar pattern is seen in the bromo-imidazoyl series
where the hydroxyl derivative 5 has a Ki of 0.34 mM and
in the pyrazino-piperazinyl series where the Ki for 8 is
0.16 mM. The dimer inhibitors are at least as effective as
the monomeric versions, with some improvement for 10
compared to 1 which might be due to the increased interac-
tions of the longer moleculewith the large cavity of MAOA.
3. Results
The lead compound for this work was linezolid 1, the
first oxazolidinone to be approved for use as an antibacter-
ial agent. The chiral C-5 side chain on the oxazolidinone
ring was varied as acetamidomethyl, hydroxymethyl, or
aminomethyl for each of three aryl substituents on the
oxazolidinone ring nitrogen, namely 4-(N-morpholino)-
Table 1
Ki values for the inhibition of MAO A by oxazolidinones and the influence of their binding on the spectrum of MAO A and stabilization of the semiquinone
during chemical reduction
Series
Compound
R
Ki (mM)
Spectrum (nm)
Reduction
Morpholino
1
2
3
Acetamide
OH
20
1.2
" 510
# 495
# 495
No SQ
No SQ
NH2
116
Substrate
Br-imidazole
4
5
6
Acetamide
OH
5
Hybrid
# 495
–
Some SQ
Some SQ
Substrate
0.34
23
NH2
Pyrazino-piperazinyl
7
8
9
Acetamide
OH
1.6
" 510
# 495
–
No SQ
0.16
5.7
No SQ
NH2
Substrate
Dimer acetamide
Dimer OH
10
11
12
0.5a
1a
90a
" 510
# 495
–
Some SQ
–
–
Dimer morpholino
(–) Not determined; SQ, semiquinone.
a
IC50 at 0.15 mM kynuramine. Trial experiments indicated that the dimers give competitive inhibition, so Ki values can be calculated from the relationship
IC50 = Ki(1 + S/Km). For the conditions of this table, this means that the Ki values would be: for 10, 0.25 mM; 11, 0.5 mM; 12, 45 mM.

T.Z.E. Jones et al. / Biochemical Pharmacology 70 (2005) 407–416
411
Fig. 2. Linezolid inhibits MAO A competitively: Lineweaver–Burke plot for the inhibition of MAO A by 1. MAO A activity was measured spectro-
photometrically with the substrate 4-(1-methyl-2-pyrryl)-1 methyl-1,2,3,6-tetrahydropyridine (31–1000 mM). The inhibitor 1 concentrations were 0 (^), 1.56
(*), 3.2 (*), 6.25 (&), 12.5 (&), 25 (~), and 50 mM (~). The inset shows the secondary plot of the slopes.
However, considering the Ki values for 1, 10, and 12 in
relation to structure, another explanation is possible. The
dimer with acetamide at both ends 10 has an IC50 of 0.5 mM
but the dimer with morpholino at both ends 12 has an IC50 of
90 mM. If themonomer compound 1 with morpholino at one
end and acetamide at the other end can bind in either
orientation, the IC50 obtained would depend on the propor-
tion of enzyme occupied by each mode.
These spectra feature a peak at 510 nm in the difference
spectrum. This 510 nm peak is also observed with D-amphe-
tamine [7]. Otherfeaturesofthe difference spectrafor1, 4, 7,
and 10 include close to zero change at about 480 nm and a
double trough between 410 and 450 nm. The difference
spectrum for 4 (Fig. 3, left column, second row) is a little
different, in that it has a smaller peak at 510 nm and is
dragged down below 500 nm. This would be consistent with
a small proportion of binding in another orientation that
induced the pattern with a 495 nm minimum.
The amino derivatives 3, 6, and 9 are poor inhibitors and
are also poor substrates that reduce MAO A at relatively
slow rates when incubated at 30 8C under anaerobic con-
ditions. The amino derivative in the bromo-imidazole
series, 6, reduces the enzyme within 10 min but the amino
derivative in the morpholino series, 3, requires 4 h to
reduce it (see below).
In general, previous studies showed that inhibitors
increased the yield of the semiquinone form of MAO A
during reduction with dithionite. Both D-amphetamine,
which induces a 510 nm increase in absorbance when
bound to MAO A, and pirlindole, binding of which results
in a decrease in absorbance at 500 nm [7], increased the
yield of semiquinone produced by chemical reduction of
MAO A, and moreover, prevented full reduction. However,
for the reduction of the MAO A–oxazolidinone complexes
by dithionite, the 412 nm peak typical of the semiquinone
was seen for only a few compounds (Fig. 4). Although
compound 1 gave the same difference spectrum on binding
as for D-amphetamine, it did not increase the semiquinone
peak at 412 nm (Fig. 4, top left) and full reduction fol-
lowed. As expected, the amino derivative substrates desta-
bilized the semiquinone so that none was seen in the course
of the substrate reductions by 3 and 6 (Fig. 4, right). For the
remaining inhibitors, different yields of semiquinone are
seen with most stabilization by the bromo-imidazole series
(4 and 5), but in all cases full reduction is possible. For the
dimeric inhibitor 10, the concentration was not high
The spectral changes that occur on ligand binding to
MAO A are shown in Fig. 3. The slow rate of reduction of
MAO A by the amino derivative 3 allowed investigation of
the spectral change for substrate binding. The difference
spectrum for binding of 3 (Fig. 3, top right), taken before
reduction has progressed significantly, shows a minimum
at 495 nm similar to that observed for most of the inhibitors
previously investigated [2,3]. The hydroxyl derivatives 2,
5, 8, and 11 all produce a difference spectrum with a
minimum at 495 nm (Fig. 3, middle column), the same as
for the amino derivative. Other features of the difference
spectrum are another minimum at 465 nm and a set of
triple peaks between 380 and 460 nm where the absor-
bance is increased compared to the unliganded enzyme.
The spectral changes for derivatives bearing the acet-
amide group are distinctly different (Fig. 3, left column).

412
T.Z.E. Jones et al. / Biochemical Pharmacology 70 (2005) 407–416
Fig. 3. Ligand-induced spectral changes in MAO A. Difference spectra are shown for the addition of inhibitors as follows: (1) 40, 320, and 620 mM 1 to 16 mM
MAO A; (2) 16 and 160 mM 2 to 14 mM MAO A; (3) 4, 6, 60, and 200 mM 3 to 15 mM MAO A; (4) 20, 60, 240, 320, and 600 mM 4 to 25 mM MAO A; (5) 4 and
40 mM 5 to 12 mM MAO A; (7) 4, 12, and 20 mM 7 to 10 mM MAO A; (8) 0.5, 1, 1.5, and 2.5 mM 8 to 11 mM MAO A; (10) 8, 12, and 36 mM 10 to 9 mM MAO
A; (11) 8, 24, 32, and 56 mM 11 to 11 mM MAO A; (12) 20 mM 12 to 11 mM MAO A.
enough to ensure full saturation due to solubility limita-
tions, so full reduction could occur via free enzyme.
Fig. 5 shows the most favorable binding mode for 1 and
2 identified by QXP. The models of the inhibitors bound to
MAO A indicate that for most oxazolidinones, the C-5
substituent on the oxazolidinone ring is furthest into the
enzyme cavity, as shown in Fig. 5 for 1 (Fig. 5a) and 2
(Fig. 5b). Each is docked in the active site close to the

T.Z.E. Jones et al. / Biochemical Pharmacology 70 (2005) 407–416
413
Fig. 4. Spectral changes during reduction of the MAO A–oxazolidinone complexes. Spectra for addition of dithionite to anaerobic MAO A in the presence of
inhibitors are shown, except for the substrates 3 and 6 where no reductant was added. Starting from the top left: 16 mM MAO Awith 3.6 mM (1), 14 mM MAO A
with 0.66 mM (2), 15 mM MAO A with 0.2 mM (3) where spectra were recorded over 4 h, 25 mM MAO A with 2 mM (4), 12 mM MAO A with 0.56 mM (5),
17 mM MAO A with 0.1 mM (6) where spectra were recorded at 10 and 20 min, 9 mM MAO A with 0.2 mM (7), 11 mM MAO A with 80 mM (8), and 13 mM
MAO Awith 0.284 mM (10). (Note that the bottom right spectral picture has been placed for compact presentation. The bottom right panel shows the spectra for
compound 10 which has two acetamide end groups and so it should be compared with the other spectra in the left column of this figure.)
flavin (pink) and between tyrosines 407 and 444 (green).
Hydrogen bonding interactions for these inhibitors with the
active site residues are indicated by dashed yellow lines.
Linezolid 1 forms two hydrogen bonds to tyrosine 197 and
to glutamine 215. Compound 4 gave a similar picture to 1
with the acetamide group between the tyrosines 407 and
444 as shown in Fig. 5a. However, with 4, an alternative
possible binding mode was also identified (Fig. 5c). In this
alternative binding mode, the bromo-imidazole ring is
furthest into the cavity and occupies a position close to
the flavin similar to the oxazolidinone ring in 2.
difference in the interactions with the substrate-orienting
tyrosines.
Two energetically similar binding modes were also
found for the substrate 3 as shown in Fig. 6. One
(Fig. 6b) is similar to the result obtained for the hydroxyl
derivative 2 (Fig. 5b) with the same three hydrogen bonds
to asparagine 181, tyrosine 197, and the backbone carbonyl
of glycine 443. In the other mode (Fig. 6a), the amine is
within van der Waals distance of the N-5 of the flavin. This
position is the one likely to allow the observed slow
reduction of the flavin.
The models of binding for 1 and 2 show significantly
different interactions with the two orienting tyrosines 407
and 444. With 2 the oxazolidinone ring stacks with the
tyrosines as expected for a substrate [26] whereas with 1
the acetamide lies between the tyrosines providing no p-
stacking interactions. It is possible that the spectral differ-
ences between the acetamide and OH series seen in Fig. 3
could be due to altered effects on the flavin arising from the
4. Discussion
This study has probed the orientation of the oxazolidi-
none analogues in the active site of MAO A. Comparison of
the Ki values (Table 1) for the dimeric compounds, 10 and
12, with the monomer 1 that has an acetamide at one end

414
T.Z.E. Jones et al. / Biochemical Pharmacology 70 (2005) 407–416
Fig. 5. Modelled positions of oxazolidinone derivatives with respect to the FAD in the active site of MAO A for (a) compound 1, (b) compound 2, and (c)
compound 4. The oxazolidinones are docked in the active site close to the flavin (pink) and between tyrosines 407 and 444 (green). Dotted yellow lines indicate
hydrogen bonds. Models of inhibitor-liganded MAO A indicate that with most oxazolidinones the substituent on the five-membered oxazolidinone ring is
furthest in to the enzyme cavity. However, with 4, an alternative binding mode is seen. The bromo-imidazole ring is able to occupy a position close to the flavin
similar to the oxazolidinone ring in 2.
and morpholino at the other suggested that 1 binds to MAO
A preferably in the orientation that places the acetamide
near the flavin. The spectral change induced by binding of
10 (acetamide–acetamide) to MAO A shows a difference
spectrum with the 510 nm increase and very little decrease
between 440 and 500 nm. However, the compound 12
(morpholino–morpholino) gave a clear decrease at
495 nm and no 510 nm peak. The spectra for 1 (morpho-
lino–acetamide) suggest that the acetamide binding close
to the flavin is the favored mode, consistent with the Ki
values of 10 and 11 and with the optimum docking shown
in Fig. 5a. For the bromo-imidazoyl-acetamide compound
4, the spectra (Fig. 3) are clearly a hybrid of the two types,
suggesting that binding with either end entering the active
site cavity first is possible for this compound. The model-
ling study found two binding modes for 4, one with
acetamide near the flavin (as shown for 1 in Fig. 5) and
one with the bromo-imidazole group inwards (Fig. 5c),
consistent with these experimental results.
must take place for substrates. The data here are consistent
with what is known about binding from the crystal struc-
ture of MAO B, showing that two tyrosine residues align
the molecule for optimum orientation towards the N-5 of
the flavin [20]. Ab initio molecular orbital calculations for
other oxazolidinone derivatives have shown that the elec-
tronic properties of the oxazolidinone ring are good for
stacking with aromatic systems [27].
The docking of energy-minimised models of the oxa-
zolidinone compounds to the structure of MAO A [4] was
performed to seek an explanation for the very clear spectral
difference between an acetamide group near the flavin (left
column of Fig. 3) or an amino (right), or hydroxyl (middle
column of Fig. 3) group near it. The modelled positions are
shown in Fig. 5. For 2 and 4, stacking of the oxazolidinone
and imidazole rings between the tyrosines is consistent
with what has been seen in the crystal structures of MAO B
[26]. For 1, the acetamide group shifts the oxazolidinone
ring away from the tyrosines and the acetamide nitrogen is
seen there instead (Fig. 5a). The amino derivative 3 in
Fig. 6 can either take a position like 2 with ring overlap
The end of the molecule entering first should bind
closest to the N-5 of the flavin where the electron transfer

T.Z.E. Jones et al. / Biochemical Pharmacology 70 (2005) 407–416
415
ing up in Fig. 5a but down in Fig. 5b. The hydroxyl group of
2 forms a hydrogen bond to a backbone residue pulling the
ligand well into the hydrophobic binding pocket so that the
middle ring is under Q215 in 2. In contrast, with 1 it is the
oxazolidinone ring carbonyl that H-bonds to Q215. The
better inhibitor (2, with the OH) also has three hydrogen
bonds whereas only two bonds are formed with 1. Both of
these factors could contribute to the higher affinity
observed for the hydroxymethyl compounds. The apparent
lower affinity for the aminomethyl derivatives compared to
the hydroxymethyl compounds that model similarly into
the active site is simple: the amino group pKa is likely to be
around 9–10 so only a small proportion will be in the
neutral form that binds to MAO.
The models, difference spectra, and the semiquinone
stabilization all indicate that bromo-imidazole can bind
towards the flavin but there is no such indication for the
pyrazinyl piperazino compounds 7 and 8. The pyrazinyl
piperazino group is somewhat similar to the bromo-imi-
dazole but clearly differs from the morpholino in having
the extra heteroaromatic ring. The preferred binding of
such lipophilic aromatic residues over cycloaliphatic and
more polar rings is well documented in the early literature
on MAO A and MAO B inhibitor structure–activity rela-
tionships. Such interactions may be based on p-stacking
interactions with aromatic amino acids (see references
quoted in [14]) in an area of the cleft that influences A/
B specificity [4,28].
Fig. 6. (a and b) Alternative modelled positions of 3 with respect to the FAD
in the active site of MAO A. The distance between N-5 of the flavin and the
amine is indicated.
5. Conclusions
with the tyrosines or one where the amine is in the correct
position for electron transfer to the flavin.
Modelling studies reveal alternate binding modes for
oxazolidinones in the active site of MAO A. Experimental
support for aryl group binding near the flavin comes from
the substrate behaviour of the amino derivatives and
favored binding of dimers with aryl groups at both ends
(compounds 10 and 11). When only the bulky morpholino
group is available as in the dimer 12, the molecule is a poor
inhibitor. Placement of the bromo-imidazole group near
the flavin was found as an alternate binding mode and its
presence there was supported by the mixed spectral pat-
terns obtained for 4. In general, efficacy of inhibition of
MAO A was favored by binding of the oxazolidinone ring
in the aromatic cage close to the flavin, reinforced by
hydrogen bonding.
The models do not provide any obvious explanation for
the influence of inhibitors on the redox properties of the
flavin. The inhibitors used in previous studies, such as D-
amphetamine and pirlindole, prevented full reduction of
the flavin so that the semiquinone state indicated by the 385
and 412 nm peaks in the difference spectra for dithionite
titrations of MAO A was maximised [7]. Dithionite titra-
tions of MAO A in the presence of the oxazolidinone
inhibitors (Fig. 4) either resemble that seen with free
enzyme with an estimated yield of about 35% of the
maximum semiquinone [2] or they resemble substrate
titrations (Fig. 4, right) which show no semiquinone at
all. This could be interpreted as these molecules having a
relatively non-specific fit in the active site, blocking sub-
strate access but not staying sufficiently close to the flavin
to prevent its full reduction. It is interesting that another
oxazolidinone, befloxatone, an antidepressant devoid of
antibacterial activity, also gave a substrate-like reduction
pattern with no trace of the 412 nm peak [2].
Acknowledgment
This work was supported by AstraZeneca.
Taken together the theoretical and experimental results
give insight into the improved potency of the hydroxy-
methyl derivatives over the acetamides. Fig. 5 shows that
the compounds 1 and 2 take up different binding poses in
the active site, with the oxazolidinone ring carbonyl point-
References
[1] Kim H, Sablin SO, Ramsay RR. Inhibition of monoamine oxidase A by
beta-carboline derivatives. Arch Biochem Biophys 1997;337:137–42.

416
T.Z.E. Jones et al. / Biochemical Pharmacology 70 (2005) 407–416
inhibitor. 1. Biochemical profile. J Pharmacol Exp Ther 1996;277:
[2] Ramsay RR, Hunter DJB. Inhibitors alter the spectrum and redox
properties of monoamine oxidase A. BBA Proteins Proteomics
253–64.
2002;1601:178–84.
[16] Hutchinson DK. Oxazolidinone antibacterial agents: a critical review.
Curr Top Med Chem 2003;3:1021–42.
[3] Hynson RMG, Wouters J, Ramsay RR. Monoamine oxidase A inhi-
bitory potency and flavin perturbation are influenced by different
aspects of pirlindole inhibitor structure. Biochem Pharmacol
2003;65:1867–74.
[17] Antal EJ, Hendershot PE, Batts DH, Sheu WP, Hopkins NK, Donald-
son KM. Linezolid, a novel oxazolidinone antibiotic: assessment of
monoamine oxidase inhibition using presser response to oral tyramine.
J Clin Pharmacol 2001;41:552–62.
[4] Ma JC, Yoshimura M, Yamashita E, Nakagawa A, Ito A, Tsukihara T.
Structure of rat monoamine oxidase A and its specific recognitions for
substrates and inhibitors. J Mol Biol 2004;338:103–14.
[5] Rigby SEJ, Hynson RMG, Ramsay RR, Munro AW, Scrutton NS. A
stable tyrosyl radical in monoamine oxidase A. J Biol Chem
2005;280:4627–31.
[18] Cantarini MV, Painter CJ, Gilmore EM, Bolger C, Watkins CL,
Hughes AM. Effect of oral linezolid on the pressor response to
intravenous tyramine. Br J Clin Pharmacol 2004;58:470–5.
[19] Ding CZ, Silverman RB. Transformation of heterocyclic reversible
monoamine oxidase-b inactivators into irreversible inactivators by N-
methylation. J Med Chem 1993;36:3606–10.
[6] Sablin SO, Ramsay RR. Monoamine oxidase contains a redox-active
disulfide. J Biol Chem 1998;273:14074–6.
[20] Binda C, Li M, Hubalek F, Restelli N, Edmondson DE, Mattevi A.
Insights into the mode of inhibition of human mitochondrial mono-
amine oxidase B from high-resolution crystal structures. Proc Natl
Acad Sci USA 2003;100:9750–5.
[7] Hynson RMG, Kelly SM, Price NC, Ramsay RR. Conformational
changes in monoamine oxidase A in response to ligand binding or
reduction. Biochim Biophys Acta Gen Subj 2004;1672:60–6.
[8] Sablin SO, Ramsay RR. Substrates but not inhibitors alter the redox
potentials of monoamine oxidases. Antioxid Redox Signal 2001;3:
723–9.
[21] Tan AK, Weyler W, Salach JI, Singer TP. Differences in substrate
specificities of monoamine oxidase-a from human liver and placenta.
Biochem Biophys Res Commun 1991;181:1084–8.
[9] Tan AK, Ramsay RR. Substrate-specific enhancement of the oxida-
tive half-reaction of monoamine-oxidase. Biochemistry 1993;32:
2137–43.
[22] Brickner SJH, Douglas K, Barbachyn MR, Manninen PR, Ulanowicz
DA, Garmon SA, et al. Synthesis and antibacterial activity of U-
100592 and U-100766, two oxazolidinone antibacterial agents for the
potential treatment of multidrug-resistant Gram-positive bacterial
infections. J Med Chem 1996;39:673–9.
[10] Mazouz F, Lebreton L, Milcent R, Burstein C. 5-Aryl-1,3,4-oxadiazol-
2(3h)-one derivatives and sulfur analogs as new selective and compe-
titive monoamine-oxidase type-B inhibitors. Eur J Med Chem 1990;25:
659–71.
[23] Betts MJ, Zeneca patent WO 9731917, vol. WO 9731917, 1997.
[24] Bissel P, Bigley MC, Castagnoli K, Castagnoli N. Synthesis and
biological evaluation of MAO-A selective 1,4-disubstituted-1,2,3,6-
tetrahydropyridinyl substrates. Bioorg Med Chem 2002;10:3031–41.
[25] McMartin C, Bohacek RS. QXP: powerful, rapid computer algorithms
[11] Kneubuhler S, Thull U, Altomare C, Carta V, Gaillard P, Carrupt PA, et
al. Inhibition of monoamine oxidase-B by 5h-indeno 1,2-C pyrida-
zines—biological-activities, quantitative structure–activity-relation-
ships (Qsars) and 3d-Qsars. J Med Chem 1995;38:3874–83.
[12] Lebreton L, Curet O, Gueddari S, Mazouz F, Bernard S, Burstein C,
et al. Selective and potent monoamine-oxidase type-B inhibitors-2-
substituted 5-aryltetrazole derivatives. J Med Chem 1995;38:4786–
92.
for structure-based drug design.
J Comput Aided Mol Des
1997;11:333–44.
[26] Binda C, Newton-Vinson P, Hubalek F, Edmondson DE, Mattevi A.
Structure of human monoamine oxidase B, a drug target for the
treatment of neurological disorders. Nat Struct Biol 2002;9:22–6.
[27] Moureau F, Wouters J, Vercauteren DP, Collin S, Evrard G, Durant F,
et al. A reversible monoamine-oxidase inhibitor, toloxatone—spectro-
photometric and molecular-orbital studies of the interaction with
flavin adenine-dinucleotide (FAD). Eur J Med Chem 1994;29:269–77.
[28] Reyes-Parada M, Fierro A, Iturriaga-Vasquez P, Cassles BK. Mono-
amine oxidase inhibition in the light of new structural data. Curr
Enzyme Inhib 2005;1:85–95.
[13] Ding CZ, Silverman RB. 4-(Aminomethyl)-1-aryl-2-pyrrolidinones, a
new class of monoamine oxidase-b inactivators. J Enzyme Inhib
1992;6:223–31.
[14] Ramsay RR, Gravestock MB. Monoamine oxidases: to inhibit or not to
inhibit. Mini Rev Med Chem 2003;3:129–36.
[15] Curet O, Damoiseau G, Aubin N, Sontag N, Rovei V, Jarreau FX.
Befloxatone, a new reversible and selective monoamine oxidase-A