Y.-H. Tao et al. / Bioorg. Med. Chem. Lett. 19 (2009) 731–734
733
Cl
formed between GABA-T and 9 is stable and the inactivation is
an irreversible process.
O-
O-
Two published mechanism including a Michael addition and an
enamine mechanism cannot elucidate the present inactivation. We
previously speculated that inactivation of GABA-T by 8 accords
with a Michael addition mechanism (Scheme 4).14 Compound 9
is not a Michael acceptor, so the inactivation cannot be described
by a Michael addition mechanism. On the other hand, 5-fluoro-4-
oxopentanoic acid (16)17 inactivates only the PMP form of GABA-
T by the mechanism shown in Scheme 5. Compound 16 initially
was described as inactivating GABA-T by Michael addition mecha-
nism; however, this work was published prior to the report on the
enamine mechanism. In fact, this mechanism starts as Schiff base
formation with PMP until it reaches the same intermediate 19 pro-
duced by the reaction of 4-amino-5-fluoropentanoic acid (5) with
the PLP form of GABA-T. Consequently, 16 undoubtedly inactivates
the enzyme by an enamine mechanism rather than a Michael addi-
tion mechanism.18 These findings suggest that the inactivation via
an enamine mechanism requires introduction of a halogen atom to
O
NH2
8
NH2
O
9
Pyr
Pyr
X
X
Cl
a
H
H
H+N
O-
O-
B:
N+H
Pyr
13
21
Pyr
b
X
X
B
H
O-
O-
N+H
Pyr
Scheme 4. Potential mechanism of inactivation of GABA-T by 8 and 9.
HN
Pyr
15
14
the
group; in other words, the b-position of the amino group for 5 is
equal to the -position of the carbonyl group for 16. By careful
a-position of the carbonyl group or the b-position of the amino
a
comparison between 9 and 16, it can be found that the halogen
atom is introduced to the b-position of carbonyl group for 9, differ-
ing from a-position for 16. Therefore, an enamine mechanism can-
not account for the inactivation by 9.
Two mechanisms were proposed to rationalize the inactivation
by 9 (Scheme 4). An active-site nucleophile may react directly with
9 via an SN2 mechanism, accounting for the inactivation. When the
electropositive carbon atom adequately approaches to an active-
site nucleophile of GABA-T, an SN2 reaction occurs (Scheme 4).
Silverman et al. have reported that an SN2 pathway is one of the
inactivation mechanisms of 22,19 suggesting that an SN2 reaction
of an active-site nucleophile for GABA-T with halide may take
Inactivation by 9 was prevented by addition of a-ketoglutarate
to the incubation (Fig. 2), indicating that the inactivation occurs at
the active site and 9 is bound to the pyridoxamine 50-phosphate
(PMP) form of the coenzyme to produce a Schiff base. Most of pub-
lished GABA-T inactivators are GABA analogues and thus are bound
to the pyridoxal 50-phosphate (PLP) form of the coenzyme. Like 7
and 8, however, 9 is a a-ketoglutarate analogue.
Because of the surprising inhibition of GABA-T by 9, the inhibi-
tion type was studied further via three methods. As shown in
Figure 3A, incubation of GABA-T with 9 for 30 min followed by
removal of excess inactivator using gel filtration did not result in
recovery of enzyme activity. No activity was also restored when
enzyme that had been inactivated by 9 was dialyzed at 4 °C (data
not shown). Moreover, GABA-T was exposed to 9 for 40 min fol-
place. Alternatively, proton abstraction from the a-carbon and sub-
sequent chloride elimination may result in transient formation of a
Michael acceptor (13) which then irreversibly reacts with an
active-site nucleophile. While participation of a Michael acceptor
in the inactivation process has not been demonstrated, there is
some evidence of other enzymes to support a b-elimination-Mi-
chael addition mechanism, such as inactivation of HMG-CoA syn-
thase by b-chloropropionyl CoA.20 Thus, the above two
hypotheses are realistic possibilities. The distinction between an
SN2 mechanism and b-elimination-Michael addition mechanism
must await the results of future investigation. In contrast to 9, 7
is a reversible inhibitor and 8 inactivates via a Michael addition
pathway,14 suggesting that modifications in the inactivator struc-
tures can lead to major difference in inhibitory mechanisms.
Regardless of the identity of the actual inactivating agent, there
is no doubt that our previous and present results may also give a
clue to the design of a novel class of GABA-T inhibitors inactivating
only the PMP form of the enzyme. GABA analogues could poten-
tially inhibit PLP-dependent glutamate decarboxylase (GAD) and
thus inhibit synthesis of GABA. Gabaculine, for example, an irre-
versible inactivator of GABA-T,21 inactivates GAD potently.22
Therefore, analogues of succinic semialdehyde (2) are useful as
selective inactivators of GABA-T. So far, however, only two ana-
logues of succinic semialdehyde, 3,5-dioxocyclohexane-carboxylic
acid23 and 16,17 were found to inactivate its PMP form.
lowed by addition of increasing concentrations of
a-ketoglutarate.
Even at 30 mM -ketoglutarate, no increase in activity was
a
observed (Fig. 3B). These observations indicated that the adduct
F
F
B
-
H3+N
Pyr
CO2
-
O
CO2
F
H
H+N
5
H
N
16
NH2
-
CO2
17
Pyr
Pyr
H2O
F
F
B
H
N
-
H
CO2
H+N
Pyr
CO2
-
19
18
Pyr
-
H2+N
Pyr
CO2
H
H+N
Pyr
CO2
-
N
NH2
Introduction of phenol group may increase lipophilicity of drug
candidates, which is extremely important to central nervous sys-
tem drugs. Phenol moiety is very useful to design GABA-T inhibi-
tors that could be more effective at crossing the blood–brain
barrier. Moreover, design of conformationally rigid analogues of
vigabatrin has received significant attention in recent years,
because vigabatrin has been found to lead to irreversible visual
H2+N
-
CO2
N+H
H
N
-
H2N
CO2
Pyr
20
Pyr
Scheme 5. Mechanism of inactivation of GABA-T by 5 and 16.