 |
Printable
PDF Version
Identification of amino
acid residues involved
in the binding of
Huperzine A to
cholinesterases
ASHIMA SAXENA,’ NAIFENG
QIAN,2 ILDIKO M.
KOVACH,2 A.P.
KOZIKOWSKI,3
Y.P. PANG: DANIEL C.
VELLOM? ZORAN RADIC;,’
DANIEL QUINN,4,6
PALMER TAYLOR,4 AND
BHUPENDRA P. DOCTOR’
Division of
Biochemistry, Walter
Reed Army Institute of
Research, Washington,
D.C. 20307
Department of Chemistry,
The Catholic University
of America, Washington,
D.C. 20064
The Mayo Clinic,
Jacksonville, Florida
32224
University of California
San Diego, La Jolla,
California 92093
Visiting Fogarty Fellow,
Institute of Medical
Research, University of
Zagreb, Croatia
(RECEIVE January 31,
1994; ACCEPTED Jun2e6 ,
1994)
Abstract
Huperzine A, a potential
agent for therapy in
Alzheimer’s disease and
for prophylaxis of
organophosphate
toxicity, has recently
been characterized as a
reversible inhibitor of
cholinesterases. To
examine the specificity
of this novel compound
in more detail, we have
examined the interaction
of the 2 stereoisomers
of Huperzine A with
cholinesterases and
site-specific mutants
that detail the
involvement of specific
amino acid residues.
Inhibition of fetal
bovine serum
acetylcholinesterase by
(-)-Huperzine A was
35-fold more potent than
(+)-Huperzine A, with K,
values of 6.2 nM and 210
nM, respectively. In
addition, (-)-Huperzine
A was 88-fold more
potent in inhibiting
Torpedo
acetylcholinesterase
than (+)-Huperzine A,
with K, values of 0.25
pM and 22 pM,
respectively. Far larger
K, values that did not
differb etween the 2
stereoisomers were
observed with horse and
human serum
butyrylcholinesterases.
Mammalian
acetylcholinesterase,
Torpedo
acetylcholinesterase,
and mammalian
butyrylcholinesterase
can be distinguishebdy
the amino acid Tyr,
Phoe,r Ala in the 330
position, respectively.
Studies with mouse
acetylcholinesterase
mutants, Ty3r3 7(330)
Phe and Tyr3 37(330) Ala
yielded a difference in
reactivity that closely
mimicked the native
enzymes. In contrast,
mutation of the
conserved Glu 199
residue to Gln in
Torpedo
acetylcholinesterase
produced only a 3-fold
increase in K, value for
the binding of Huperzine
A. Molecular mechanics
energy minimization of
the complexes formebde
tween each of the 2
stereoisomers of
Huperzine A and fetal
bovine seruma
cetylcholinesterase,
Torpedo
acetylcholinesterase, or
human
butyrylcholinesterase
also revealed that (-)-Huperzine
A gave a better fit than
(+)-Huperzine A and
implicated Tyr3 37(330)
in the stereoselectivity
of Huperzine A.
Keywords:
cholinesterases;
Huperzine A; inhibitor;
molecular modeling;
site-directed
mutagenesis Huperzine A,
an alkaloid constituent
oHf uperzia serrata has
been a longstanding
agent in Chinese herbal
medicine for the
treatment of dementia in
the elderly population.
It hasb een synthesized
(Kozikowski et al.,
1991) and characterized
asa potent and selective
inhibitor of
acetylcholinesterase (EC
3.1.1.7) (McKinney et ai.,
1991). The potentially
superior inhibition
characteristics of
Huperzine A, as compared
to other cholinesterase
inhibitors, have been
attributed to the very
slow rate of
dissociation (to.s = 35
min) of AChE-Huperzine A
complex in solution
(Ashani et al., 1992).
Also, the interaction of
Huperzine A with AChE
appearsr eversible and
does notr esult in any
detectable chemical
modification of the
inhibitor (Ashani et
at., 1992). These anti-AChE
properties of Huperzine
A, which are not shared
by other commonly used
anti-AChE drugs such as
physostigmine andt
etrahydroaminoacridine,
may prove to be useful
pharmacological
characteristics, not
onlyf or the treatment
of Alzheimer’s disease
and other nervous
system-related
dementias, but also for
prophylaxis against
organophosphate
toxicity. Compared to
AChE, Huperzine A has
been reported to be
1,000-fold less potent
as an inhibitor of
butyrylcholinesterase
(EC 3.1.1.8). The
difference in the
activity of Huperzine A
toward AChE and BChE
hasb een proposed to be
due to differences in
the aromatic amino acid
residues lining the
pocket of the catalytic
region of the enzyme
molecule (Ashani et al.,
1992). The naturally
occurring (-)-Huperzine
A was a mixed
competitive inhibitor of
AChE (Wang et al1.,9
86). Inhibition studies
with the 2 stereoisomers
of Huperzine A revealed
that (-)-Huperzine A
inhibited crude
preparations of rat
cortical AChE 38-fold
more potently than (+)-Huperzine
A (McKinney
et al., 1991). In the
present study, thiss
tereoselectivity of
Huperzine A was examined
in more detaiwl ith 3
distinct families of
ChEs with known sequence
differences. The basis
for particular residues
conferring selectivity
was then confirmed by
using site-specific
mutants of the
implicated residuien a
single template enzyme.
These findings enabluesd
t o model the
ChEHuperzine A
interactions and propose
an orientation fothr ese
bound stereoisomers in
the complex.
Results
Inhibition of ChEs by
the 2 stereoisomers of
Huperzine A Plots of
percent residuaalc
tivity versus time in
the presenceo f (+)-Huperzine
A, (-)-Huperzine A, and
(+)-HuperzineA for
purified preparations of
FBS AChE and Torpedo
AChE are shown in Figure
1. The results confirm
the previous observation
that (-)-Huperzine A was
the active stereoisomer
(Wang et al., 1986;
McKinney et al., 1991).
There was no inhibition
of either horse or human
serum BChE at
concentrations of up1 0t
op M of the 2
stereoisomers of
Huperzine A (data not
shown). The K, values
for the 2 stereoisomers
of Huperzine A show that
(-)-Huperzine A
inhibited FBS AChE
35-fold more potently
than (+)-Huperzine A
(Table 1). On the other
hand, (-)- Huperzine A
was 88-foldm ore potent
than (+)-HuperzinAe in
inhibiting Torpedo AChE.
The preparation of (+)-Huperzine
A had approximately 1%
contamination of (-)-Huperzine
A, which precluded us
from observing more than
2 orders of magnitude
difference between the
K, values for the 2
stereoisomers. Therefore
the stereoselectivity
reflected by the K,
values for the 2
stereoisomers of
Huperzine A is minimal.
Due to thhei gh
association
and dissociation rates
for the interactiono f
BChE and Huperzine A, K,
values for horse and
human serum BChE for the
2 stereoisomers of
Huperzine Aw ere
determined by the
analyses of steady-state
kinetic data (Table 1).
Unlike AChE, no
significant differences
in the KI values were
observed for the 2
stereoisomers of
Huperzine A with horse
and human serum BChEs,
suggesting a lack of
stereoselectivityo f
this compound for BChE.
Differencesin the
reactivity of (-)-Huperzine
A toward FBS AChE,
Torpedo AChE, and horse
and human serum BChEs
showed that Tyr
337(330)' in mammalian
AChE may be an important
amino acid residue in
the binding of (-)-
Huperzine A to ChEs.
This prediction was
tested by conducting
inhibition studies of
(-)-Huperzine A with
recombinant mouse AChE
mutants, where T3y3r7
(330) in the wild-type
enzyme was modified to
either Phe ians Torpedo
AChE or Ala as in BChE.
Plots of percent
residuacl tivity vs.
time in the presence of
(-)-Huperzine A for
recombinant mouse AChE
and the Tyr 337(330) Phe
and Tyr 337(330) Ala
mutants are shown in
Figure 2. These results
show that (-)-Huperzine
A was a more potent
inhibitor of wild-type
mouse AChE as compared
to the Tyr 337(330) Phe
and Tyr 337(330) Ala
mutants, implicating the
amino acid residue Tyr
337(330) in the binding
of Huperzine A to ChEs.
As shown in Table2 ,
mutation of Tyr 337(330)
in a mouse template to
Phoer Ala yielded a
difference in reactivity
that closely mimicked
the native enzymes
(Table 1). Eliminating
the charge deep in the
active-center gorge by
mutation of Glu 199 in
Torpedo AChE to Gln
resulted in only a
3-fold increase in KI
value of (-)-Huperzine A
for this enzyme.
Energy-minimized
structures of Huperzine
A bound to ChEs
Because AChE fromT
orpedo californica is
the only ChEw hose
structure has been
determined
experimentally at atomic
resolution, docking of
each of the2
stereoisomers of
Huperzine Ain to the
active site and
subsequent molecular
mechanics energy
minimization of the
formed conjugatew ere
carried out inT orpedo
AChE. The procedure used
for Torpedo AChE was
then extended to both
FBS AChE and human BChE.
Panel 82 in Figure 3
shows the interaction of
(-)-Huperzine A with
various amino acidre
sidues in the active
site of Torpedo AChE.
There are 3
energetically favorable
regions for the
interaction of Huperzine
A with the enzyme
molecule: (1) the
hy-drophobic core,
represented in green,
includes the ethylidene
group and the bridge
double bondo f (-)-Huperzine
A, interacting with Trp
84, Phe 330, Tyr 334,
and Tyr 442 of the
enzyme molecule; (2) the
oxyanion region,
represented by default
color, includes the 3
partial hydrogen bonds
formed between the
carbonyl oxygen of
Huperzine A and the
amide backbone hydrogens
of Gly 118, Gly 119, Ala
201, the electrostatic
interaction
between the
electrophilic carbonyl
carbon of Huperzine A,
and the y-oxygen of Ser
200 of the enzyme
molecule; and (3) the
electrostatic region,
represented in red,
indicates thep ossible
interaction of the
primary amine group of
Huperzine A with the
carboxylate of Glu 199,
which is at a distance
of 6-8 A. Panel A2 in
Figure 3 shows the
complex of (-)-HuperzinAe
with FBS AChE. Like the
Torpedo AChE-(-)-Huperzine
A
complex, this complexa
lso has favorable
interactionisn the
hydrophobic core
involving the ethylidene
group and the bridge
double bondo f (-)-Huperzine
A, with Trp 86(84), Tyr
337(330), Tyr 341(334),
and Tyr 449(442) of the
enzyme. In the oxyanion
hole region and
electrostatic region,
the interactions between
the inhibitor and the
enzyme molecule are as
favorable as in Torpedo
AChE. Panel C2 in Figure
3 shows (-)-Huperzine A
complexed with human
BChE. According to the
molecular mechanics
calculations, there are
numerous soft van der
Waals interactions
between Huperzine A and
the activesi te residues
of AChE. With
both stereoisomers, the
total number of these
interactions is smaller
for BChE as compared to
AChE. The predominant
interactions in the
hydrophobic region are
with Trp 86(84), which
is within 3.4-4.0 A from
a number of atoms in the
inhibitor in either
orientation. There arfee
wer contacts between Trp
86(84) of BChE and the 2
stereoisomers of
Huperzine A. In BChE,
Ala 330 provides a
weaker interaction with
both stereoisomers of
Huperzine A as compared
to Ph3e3 0 in Torpedo
AChE and Tyr 337 in
FBSAChE, all of which
are within optimal
(3.4-4.0A) distance from
some of the atoms of
Huperzine A. The
complexes of (+)-Huperzine
A with FBS AChE, Torpedo
AChE, and human BChE are
shoiwn np anels AI, B1,
and C1, respectively.
The stereoselectivity of
Huperzine A for FBS AChE
(35-fold) and Torpedo
AChE (88-fold) can be
correlated to
differences in
interactions in the
oxyanion hole region and
weakelectrostatic forces
between Glu 199 and the
primary amine group of
Huperzine A. Although
the former of these
interactions are also
soft van deWr aals
forces, there arew eak
hydrogen bonds between
the carbonyl group of
Huperzine A and amide
backbone ofG ly 1 18 and
Gly 119, and theh
ydroxyl group of Ser
200, especially for FBS
AChE, with the (-)-Huperzine
A. The active site
serineh as the
strongestte ndency to
hydrogen bond, which is
not surprising because
our model does not
engage this serine in a
covalent bond. The
electrostatic
interactions in this
model are between Glu
199 and the primary
amine of Huperzine A,
but they are maximum at
-2.0 kcal/mol or less
because the distances
between the nitrogen and
one of thcea rboxyl
oxygens are between 6
and 8 A (FBS AChE-(-)-Huperzine
A, 7.12 A; FBS AChE-(+)-Huperzine
A, 7.5 A ; Torpedo AChE-
(-)-Huperzine A, 6.15 A
; Torpedo AChE-(+)-Huperzine
A, 7.02 A; BChE-(-)-Huperzine
A, 7.06 A ; and BChE-(+)-
Huperzine A, 8.25 A).
It is difficult to
assess the quantitative
difference is in binding
energies that are the
suma ogfr eat numbero f
small terms and are not
isolated from the
interactions with the
surroundings. The
35-fold
stereoselectivity
observed with FBS AChE,
for example, would
translate into a
difference of 2.1
kcal/mol in binding
energy.
The weak hydrogen bonds
observed for the (-)-Huperzine
A complexes as well as
more favorable
interactions in the
hydrophobic and
electrostatic regions
can support this
difference.
Discussion
In the 3-dimensional
structure of Torpedo
AChE, the active site is
in a gorge lined with
the side chains of 14
aromatic amino acid
residues (Sussman et
al., 1991). Based on
sequence alignments of
ChEs, these residues are
fully conserved in all
known vertebrate AChEs
(Gentry & Doctor, 1991).
On the other hand, 6 of
14 aromatic amino acid
residues at positions
70, 121,279,288, 290,
and 330 in Torpedo AChE
are replaced by
aliphatic amino acid
residues in BChE. These
structural differences
between AChE and BChE
and recent
investigations employing
molecular modeling,
site-directed
mutagenesis, and studies
of the catalytic and
inhibitory properties of
these mutants lend
strong support to the
involvement of these
aromatic amino acid
residues in the binding
and selectivity of
inhibitors to ChEs. The
role of Trp 86(84) in
the orientation and
stabilization of the
quaternary ammonium
group of the substrate
has been demonstrated by
chemical labeling
studies (Weise et al.,
1990; Kreienkamp et al.,
19911, by
crystallographic data (Sussman
et al., 1991, 1992) and
site-directed
mutagenesis studies for
Trp 86(84) in human AChE
(Ordentlich et al.,
1993). The 2 Phe
residues at positions
295(288) and 297(290)
define the dimensions of
the acyl pocket of
mammalian AChEs,
markedly reducing BTC
hydrolysis and enhancing
ATC hydrolysis by
forming a clamp around
the methyl moiety ofth e
acetoxy group, and
restricting its degrees
of freedom (Vellom et
al., 1993). Replacement
of either of these
residues with
nonaromatic amino acid
residues, as in BChEs,
results in an increased
hydrolysis of BTC by the
mutant enzyme and its
increased sensitivity to
the bulky BChE-specific
organophosphate
inhibitor iso-OMPA (Harel
et al., 1992; Ordentlich
et al., 1993; Vellom et
al., 1993). This
observation is also
supported by a naturally
occurring mutation in
Drosophila AChE that
contains a single Phe at
position 368(290).
Replacement of Phe by
Tyr confers enzyme
resistance to certain
organophosphates
(Fournier et al., 1992).
The Trp residue at
position 279 in Torpedo
AChE is located near the
lip of the gorge and has
been designated as part
of the “peripheral’’
anionic site. Mutation
of this amino acid
residue to a nonaromatic
amino acid residue as in
BChE results in a loss
of sensitivity of the
mutant enzyme to
“peripheral” anionic
site ligands like
propidium (Harel et al.,
1992; Shafferman et al.,
1992; RadiC et al.,
1993). The 2 neighboring
Tyr residues conserved
in AChEs at position
72(70) and 124(121) also
contribute
to the stabilization of
“peripheral” site
inhibitor complexes (RadiC
et al., 1993). Another
aromatic amino acid
residue that is present
near the choline binding
site is Tyr 337(330) (Sussman
et al., 1991). This
residue appears to
stabilize the binding of
ligands such as
edrophonium, acridines,
and 1 end of
bisquaternary compounds
such as BW284C51 and
decamethonium
(Shafferman et al.,
1992; Ordentlich et al.,
1993; Radii et al.,
1993). This residue
destabilizes the binding
of phenothiazines
such as ethopropazine,
which contains a bulky
exocyclic substitution.
Structure-activity
relationships show that
this is a consequence of
steric hindrance between
the
diethylamino-2-isopropyl
moiety with the aromatic
side chain of Tyr
337(330) in the
mammalian enzyme (Radii
et al., 1993). The
results presented here,
using stereoisomers of
Huperzine A also show
the involvement of
certain aromatic amino
acids lining the gorge
in the stabilization of
AChE-Huperzine A
complexes. The passage
of Huperzine A through
the gorge to reach the
active site of AChE was
tested in the Torpedo
AChE model and found to
be completely feasible.
The inhibition of AChE
by Huperzine A is
probably because the
lactam region has a
resemblance to the
staggered conformation
of acetylcholine (Fig.
4). However, unlike
esters, amides are not
good substrates for
AChEs. The 6-membered
lactam ring is a
thermodynamically
favored internal amide
that may resist
hydrolysis because the
catalytic residues for
general acid/base
catalysis like histidine
are not at the proper
distance and
orientation. The lactam
ring may open and close
transiently in a
reversible manner.
The guiding principle in
our modeling of the
Huperzine A
stereoisomers into the
active site pockets of
ChEs was to orient the
carbonyl group of the
lactam ring in the
oxyanion hole. The
interaction energies
arising from the
hydrogen bonding of C=O
or P=O have been the
most significant
stabilizing forces in
the active site region
of serine hydrolase
enzymes (Qian & Kovach,
1993). The optimal
orientation for each of
the 2 stereoisomers of
Huperzine A in the
active-site pockets of
FBS AChE, Torpedo AChE,
and human BChE was then
determined. Our studies
implicate Trp 86(84) and
Tyr 337(330) as the
critical amino acid
residues that interact
with Huperzine Aa nd
ares upported by
experimental data
showing the
stereoselectivity of
AChE for Huperzine A and
differences in the
reactivity of (-)-Huperzine
A toward AChE and BChE.
An analysis of the
differences in free
energy of binding for
the 2 stereoisomers of
Huperzine A to ChEs
(Table 3) reveals
2 interesting facets of
the ChE-Huperzine A
interaction. First, the
contributions of the
functional moieties on
the Tyr337(330) side
chain appear additive
rather than cooperative,
as the effect of
Tyr-to-Ala substitutiona
ppears tob e a linear
summation of the Phe-to-Tyr
and Ala-to-Phe
substitutions. Second,
the contribution of the
hydroxyl group is of
comparable magnitude for
the interaction of both
stereoisomers of
Huperzine A with ChEs,
whereas the aromatic
interactions contributed
by the benzene ring
prevail in the binding
of (-)-Huperzine A.
Differences in the
reactivity of Huperzine
A toward FBS AChE,
Torpedo AChE, and horse
serum BChE implicate the
involvement of Tyr
337(330) in the binding
of (-)-Huperzine A to
AChEs. The lone pair of
electrons on 0 in the OH
of Tyr 337 are H-bond
acceptors from the NH of
Trp 86(84). This
increases the electron
density in the indole
side chain of Trp 86(84)
and draws the 2 residues
closer to provide a
continuous ?r electron
envelope around (-)-Huperzine
A. Trp 86(84) provides
the strongest overlap
with the bridge methyl
and primary amine
segments of the frame,
which pulls the ethylene
bridge of the (-)-Huperzine
A molecule in the
vicinity of the aromatic
ring of Tyr 337. This
interaction is
responsible for the low
KI value of 6 nM for the
mammalian enzyme. In
Torpedo AChE, the
corresponding Tyr
residue has been
replaced by Phe,
resulting in the loss of
this stabilizing
hydrogen bond and a
higher KI value of 0.25
pM for this enzyme. In
human serum BChE, the
Tyr residue has been
replaced by Ala, which
results in a loss of
aromatic interactions
and a further increase
in KI value to 76 pM. In
contrast, both Trp8
6(84) and Tyr 337(330)
are near the NH3+ group
of (+)-Huperzine A,
which-appears to be a
less favorable
arrangementa s indicated
by the higher KI values
of 0.21 pM, 22 pM, and
36 pM for FBS AChE,
Torpedo AChE, and human
serum BChE,
respectively. Inhibition
studies of (-)-HuperzinAe
with mouse AChE mutants,
where Tyr 337 (KI = 8.45
nM) was modified to
either Phe, as in
Torpedo AChE (KI= 0.273
pM), or Ala, as inB ChE
(K, = 8.2 pM), lend
further support to our
models. The
corresponding KI values
for (+)-Huperzine A were
0.36 pM for mouse wild-typeA
ChE, 21 pM for Tyr 337
Phe mutant AChE, and
31.3 pM for Tyr 337 Ala
mutant AChE. The results
of our modeling studies
also suggest that Glu
199 in Torpedo AChE may
be involved inth e
binding of (-)-Huperzine
A to the enzyme. Our
experimental data using
Torpedo Glu 199 Gln
mutant AChE suggest that
an energetically
favorable electrostatic
region generatedb y the
interaction of the
primary amineg roup of
Huperzine A with the
carboxylate of Glu 199,
the carbonyl oxygen of
Gly4 41, and the
hydroxyl group of Tyr
130 contribute minimally
to the stabilization of
AChE-Huperzine A
complex. This is
consistent with the fact
that the distance
between the primary
amine group of
HuperziAne a nd Glu 199
is greater than 5 A in
all cases. In this
regard, it is relevant
to mention that a model
for the orientation of
(-)-Huperzine A into the
active site gorge of
Torpedo AChE using
systematic docking
studiheas s been
proposed
(Pang & Kozikowski,
1993). These modeling
studies implicate the
involvement of Phe 330,
Tyr 121, and Phe 290 in
the binding of Huperzine
A to AChEs and attribute
the substitution of
these aromatic amino
acid residues by
nonaromatic residues to
the increase in thKe I
value of HuperzineA for
BChE. Although the
orientation of HuperzAin
ien the active spioteck
et of AChE proposed in
our modeling studies,
using molecular
mechanics, is different
from that proposed by
systematic docking
studies, both models
implicate the
involvement of Trp 84
and Phe 330 in the
binding of HuperziAne t
o AChEs. It is likely
that Huperzine A may
also interact with other
regions in the active
site of the AChE
molecule; we have not
explored this
possibility eitherb y
molecular modeling
studieos r by testing
the inhibition of
site-specific AChE
mutants of the other 5
aromatic amino acids
located in the gorge and
replaced by nonaromatic
residues in mammalian
BChE. The differences
iKn I values for the
inhibition of FBS AChE,
Torpedo AChE, and
mammalian BChE by
Huperzine A and increase
inK I values caused byt
he mutation of Tyr 337
in mouse AChE to Phe and
Ala are essenDownloaded
tially the same. This
suggests that the other
5 aromatic amino acids
may not affect the
binding of Huperzine A
to ChEs.
Materials and methods
Materials
ATC, BTC, and DTNB were
obtained from Sigma
Chemical Co., St. Louis,
Missouri. The 2
stereoisomers of
Huperzine A were
isolated from synthetic
(f)-Huperzine A as
described (McKinney et
al., 1991).
Electrophoretically pure
AChE from FBS was
purified as described
(De La Hoz et al.,
1986), and AChE from 7:
calijornica was provided
by Professor I. Silman
(Weizman Institute,
Rehovot, Israel). BChE
from horse serum was
purified by affinity
chromatography using the
procedure similar to the
oned escribed for FBS
AChE (De La Hoz et al.,
1986). One milligram of
pure native AChE or BChE
contained approximately
14 and 11 nmol of active
sites, respectively.
Expression of Torpedo
and mouse AChEs
Wild-type and mutant
mouse AChE cDNAs were
inserted into CMV-based
expression vectors (Andersson
et al., 1989; pRC/CMV,
Invitrogen). The
presence of the neomycin
resistance gene in the
CMV vectors enabled the
selection of stable
transfectants (Radii et
al., 1993; Vellom et
al., 1993). Wild-type
mouse AChE and the
mutant AChEs were
expressed from stable
transfectants in HEK-293
cells. The cell culture
media were concentrated
to 1-2% of the original
volume by
ultrafiltration as
described (Vellom et
al., 1993). The cloning,
mutagenesis, and
expression of Torpedo
AChE and its mutant
forms in a baculovirus
Spodoptera system and
subsequent purification
of expressed enzymes
were all carried out as
described (Radii et al.,
1992).
Measurement of ChEs
activity and inhibition
AChE and BChE activities
were measured in 50 mM
sodiumphosphate, pH 8.0,
at 22 "C as described (Ellman
et al., 1961) using ATC
and BTC as substrates,
respectively. Inhibition
of various ChEs was
carried out by diluting
an appropriatev olume of
stock solutions (2-5 mM)
of each of the 2
stereoisomers of
Huperzine A (in 20%
acetonitrile) into the
enzyme solutions (15-20
units/mL in 50 mM sodium
phosphate, pH 8.0,
containing
0.01% BSA) and measuring
residual enzyme activity
at various times. When
the inhibition of
various mutant AChEs by
(-)- Huperzine A was
determined,
approximately 1 unit/mL
of enzyme activity was
used.
Determination of kinetic
and inhibition
parameters The
interaction of each of
the 2 stereoisomers of
Huperzine A with various
ChEs can be described by
the scheme shown in
Figure 5A. K, values for
FBS AChE, Torpedo AChE,
and wildtype mouse AChE
were determined by
equilibrating 0.2 units/mL
of AChE with various
concentrations of the
inhibitor and measuring
residual enzyme activity
after equilibrium was
reached. The data
obtained were analyzed
according to the
following equation,
where [HUP-A] is the
initial Huperzine A
concentration: Due to
the high association and
dissociation rates for
the interaction between
Huperzine A and BChE (Ashani
et al., 1992), the KI
values for the
inhibition of horse and
human serum BChE and
mouse Tyr 337(330) Phe
and Tyr 337(330) Ala
mutant AChEs with either
stereoisomer of
Huperzine A were
determined by analysis
of kinetic data
described according to
the scheme shown in
Figure 5B, where KI and
aKI reflect the
interaction of Huperzine
A with the free enzyme
and the enzymesubstrate
complexes, respectively.
K, is the competitive
inhibition constant.
Data for this analysis
were obtained by
measuring inhibition of
enzyme activity over a
substrate concentration
range of 0.025-0.4 mM
and a series of
Huperzine A
concentrations of 0-66
pM depending on the
enzyme. Plots of
reciprocal velocities
versus reciprocal
substrate concentrations
at a series of Huperzine
A concentrationsy ielded
a family of slopes.
Replots of the slopes
versus Huperzine A
concentrations were used
for the determination of
the KI values of these
enzymes.
Modeling and energy
minimization by
rnolecufar mechanics
Molecular modeling and
all calculations were
carried out on a Silicon
Graphics Personal Iris
W-4D35TG workstation.
Preparation of ChE
structures for molecular
mechanics computation
was done as follows: the
X-ray
diffraction-determined
coordinates of AChE from
T. californica (Sussman
et al., 1991) were
obtained from the
Brookhaven Protein Data
Bank, and a missing
5-amino acid segment
(485-489) of the AChE
structurew as
reconstructed. The 4 and
$ dihedral angles of the
main chain of the
segment were adopted
from the results of a
statistical search of
the database in
molecular modeling
package GEMM (V7.8) (B.K.
Lee, National Institutes
of Health, Bethesda,
Maryland), and the side
chains were optimized
with molecular mechanics
program YETI (V5.3) (Vedani,
1988). Missing atoms
from the side chains of
27 amino acid residues
on the protein surface
were also inserted and
their positions
optimized. YETI was used
to generate the hydrogen
positions at heteroatoms.
Optimization in YETI was
carried out in an
internal/Cartesian
coordinate coordinate
space with a
conjugate-gradient
minimizer. All bond
lengths, bond angles,
and the positions of
main chain atoms were
kept constant during the
calculations. The YETI
force field consisted of
9.5/10.0 A for
electrostatic
interactions, 6.5/7.0 A
for van der Waals
interactions, and
4.5/5.0 A for hydrogen
bonding interactions.
Convergence criteria
were set to 0.025 kcal
mol" deg" for torsional
RMS first derivative, to
0.050 kcal mol" deg" for
rotational FMS first
derivative, and to 0.750
kcal mol" A" for
translational RMS first
derivative. The energy
convergence criterion
was k0.05 kcal mol". The
atomic structures of FBS
AChE (B.P. Doctor & J.L.
Sussman, unpubl.
results) and human serum
BChE (Hare1 et al.,
1992) were reconstructed
models based on the
structure of AChE from
T. californica.
X-ray fractional
coordinates for
Huperzine A (Geig et
al., 1991) were
converted to Cartesian
coordinates. For
molecular mechanics
calculation, the partial
atomic charges for
Huperzine A were
calculated with MNDO, as
implemented in MOPAC
(V6.3) (Dewar et al.,
1985). Each of the 2
stereoisomers of
Huperzine A was
interactively docked
into the active site of
T. californica AChE with
the GEMM 7.3 package,
and the formed
conjugates were
individually subjected
to energy minimization
by molecular mechanics.
Based on the energy
calculated for all the
conjugates, the
orientation with the
lowest energy was
selected, and fine
manual adjustments for
the selected orientation
were systematically
performed. After each
fine adjustment, energy
minimization was carried
out, and the interaction
energy between the
inhibitor and the enzyme
was calculated and
compared to obtain the
best orientation for the
(-)-Huperzine A molecule
in the active site of
Torpedo AChE. The
procedure used for
Torpedo AChE was then
extended to both FBS
AChE and human
serum BChE.
Printable
PDF Version
|
