TOXICOLOGICAL SCIENCES 90(2), 460–469 (2006)
doi:10.1093/toxsci/kfj094
Advance Access publication January 10, 2006
Concentration-Dependent Kinetics of Acetylcholinesterase
Inhibition by the Organophosphate Paraoxon
Clint A. Rosenfeld* and Lester G. Sultatos†,1
*Drug Metabolism and Pharmacokinetics, Schering-Plough Research Institute, Lafayette, New Jersey 07843; and †Department of Pharmacology and
Physiology, University of Medicine and Dentistry of New Jersey, New Jersey Medical School, Newark, New Jersey 07103
Received November 7, 2005; accepted January 3, 2006
For decades the interaction of the anticholinesterase organophosphorus compounds with acetylcholinesterase has been characterized as a straightforward phosphylation of the active site
serine (Ser-203) which can be described kinetically by the inhibitory rate constant ki. However, more recently certain kinetic
complexities in the inhibition of acetylcholinesterase by organophosphates such as paraoxon (O,O-diethyl O-(p-nitrophenyl)
phosphate) and chlorpyrifos oxon (O,O-diethyl O-(3,5,6-trichloro2-pyridyl) phosphate) have raised questions regarding the adequacy of the kinetic scheme on which ki is based. The present
article documents conditions in which the inhibitory capacity of
paraoxon towards human recombinant acetylcholinesterase appears to change as a function of oxon concentration (as evidenced
by a changing ki), with the inhibitory capacity of individual oxon
molecules increasing at lower oxon concentrations. Optimization
of a computer model based on an Ordered Uni Bi kinetic
mechanism for phosphylation of acetylcholinesterse determined
k1 to be 0.5 nM1h1, and k1 to be 169.5 h1. These values were
used in a comparison of the Ordered Uni Bi model versus a ki
model in order to assess the capacity of ki to describe accurately
the inhibition of acetylcholinesterase by paraoxon. Interestingly,
the ki model was accurate only at equilibrium (or near equilibrium), and when the inhibitor concentration was well below its Kd
(pseudo first order conditions). Comparisons of the Ordered Uni
Bi and ki models demonstrate the changing ki as a function of
inhibitor concentrations is not an artifact resulting from inappropriate inhibitor concentrations.
Key Words: acetylcholinesterase; organophosphate; paraoxon;
kinetics.
Organophosphorus insecticides exert their acute mammalian
toxicity through inhibition of the critical enzyme acetylcholinesterase (EC 3.1.1.7) (Mileson et al., 1998). The inhibition
occurs as a result of the phosphylation of Ser-203 (the numbers
used in this report refer to amino acid positions in human
acetylcholinesterase). Much is known regarding the molecular
1
To whom correspondence should be addressed. Fax: (973) 972-4554.
E-mail: [email protected].
events that lead to the phosphylation of Ser-203 by certain
organophosphorus compounds. Ordentlich et al. (1996) have
reported that the acyl pocket of acetylcholinesterase’s active
gorge (Phe-295 and Phe-297) participates in the positioning of
an organophosphorus molecule for the nucleophilic attack by
the catalytic Ser-203 (Fig. 1). This same group (Ordentlich
et al., 1998) have suggested that the oxyanion hole subsite
(Gly-121, Gly-122, and Ala-204) might polarize the P¼O bond
during formation of the Michaelis complex, thereby activating
the phosphorus and promoting the nucleophilic attack by
Ser-203 (Fig. 1). This nucleophilic attack is made possible by
a proton transfer from Ser-203 to a nitrogen in the imidazole
ring of His-447, which in turn transfers a proton from the
second nitrogen of the imidazole ring to the carboxyl group of
Glu-334 (Soreq and Seidman, 2001) (Fig. 1). Additionally, the
peripheral anionic site of acetylcholinesterase, located at the lip
of the active site gorge (Fig. 1), has been shown to play an important role in the stereoselectivity towards enantiomers of the
chemical warfare agent VX (O-ethyl S-[2-(dissopropylamino)
ethyl] methylphosphonothioate) (Ordentlich et al., 2004).
Occupation of the peripheral anionic site by certain ligands
has also been shown to alter rates of substrate catalysis (Taylor
and Radić, 1994; Masson et al., 2004).
The interactions between organophosphorus compounds and
acetylcholinesterase have been summarized kinetically by an
Ordered Uni Bi reaction scheme (Aldridge and Reiner, 1972;
Kardos and Sultatos, 2000):
Scheme ð1Þ
where AB represents the free inhibitor, E is free enzyme, E–AB
represents inhibitor bound reversibly to enzyme (Michaelis
complex), EA is the phosphylated enzyme, B is the leaving
group, and EAge represents aged enzyme. Water, which dephosphylates EA is not included because it is always present in
excess. The reactivation and aging pathways are often excluded
Ó The Author 2006. Published by Oxford University Press on behalf of the Society of Toxicology. All rights reserved.
For Permissions, please email: [email protected]
KINETICS OF ACETYLCHOLINESTERASE INHIBITION
461
where,
FIG. 1. Depiction of the active site gorge of acetylcholinesterase with the
entry of the organophosphate paraoxon. The acyl binding pocket might help
form the Michaelis complex, while the oxyanion hole probably renders the
phosphorus more susceptible to nucleophilic attach from Ser-203 by polarizing
the P¼O bond (Ordentlich et al., 1998). The catalytic triad consists of Ser-203,
His-447, and Glu-334. Binding of certain ligands to the peripheral anionic site
can lead to enzyme inhibition or activation (Masson et al., 2004).
from this kinetic scheme because they are often negligible for
in vitro studies.
Main (1964) simplified the kinetic description in Scheme 1
by deriving the inhibitory rate constant ki (Scheme 2). This
bimolecular rate constant describes both the formation of
a Michaelis complex, as well as the phosphylation of Ser-203
by a specific inhibitor, as shown in the following scheme:
Scheme ð2Þ
ki ¼ k2 =Kd
Equation ð1Þ
Kd ¼ k1 =kþ1
Equation ð2Þ
In the above scheme and equations, the parameters have the
same meaning as in Scheme 1. The reactivation constant (k3)
and aging constant (k4) were excluded. The larger the ki, the
greater is the capacity of an organophosphorus compound to
inhibit acetylcholinesterase. While Schemes 1 and 2 have
served for many years as the basis for our understanding of the
kinetic interactions of organophosphorus compounds with acetylcholinesterase, more recent evidence suggests that it is incomplete. Kardos and Sultatos (2000), as well as Kousba et al.
(2004) have documented that the kis for inhibition of mouse
and rat brain acetylcholinesterase by certain organophosphates
appear to change over a wide range of inhibitor concentrations,
suggesting that at low organophosphate concentrations, individual inhibitor molecules have a greater capacity to inhibit
acetylcholinesterase than individual organophosphate molecules at higher inhibitor concentrations (Fig. 2). If viewed as an
enzymatic reaction where the organophosphate is hydrolyzed
by acetylcholinesterase, this phenomenon would be considered
substrate inhibition. However, the stability of the phosphylated
intermediate precludes the use of this terminology. Kardos and
Sultatos (2000) and Kousba et al. (2004) speculated that certain
organophosphates might bind reversibly to a site on acetylcholinesterase separate from the catalytic triad, and that
this binding decreases, through steric hindrance or allosteric
FIG. 2. Diagram of a plausible explanation for a changing ki with paraoxon concentration. The shaded crescents represent monomeric human recombinant
acetylcholinesterase active sites, while small circles represent paraoxon molecules. The ability of paraoxon to phosphylate the active site (ki) is greatest when no
other molecules of paraoxon are bound elsewhere to the enzyme, and is the lowest when paraoxon is bound to a putative secondary site on the enzyme. This
diagram depicts a phenomenon reported for mouse acetylcholinesterase by Kardos and Sultatos (2000), and for rat acetylcholinesterase by Kousba et al. (2004).
462
ROSENFELD AND SULTATOS
modulation, the capacity of other organophosphate molecules
to inhibit enzyme activity (Fig. 2). The current study extends
the observations of Kardos and Sultatos (2000) and Kousba
et al. (2004) to human recombinant acetylcholinesterase, and
investigates the appropriateness of Schemes 1 and 2 for the
kinetic description of the inhibition of acetylcholinesterase by
paraoxon.
MATERIALS AND METHODS
Chemicals. Paraoxon (O,O-diethyl O-(p-nitrophenyl) phosphate) was
purchased from Chem Services (West Chester, PA), and was further purified
by dissolution in trichloroethylene, followed by washes with 2% sodium
carbonate. HPLC analyses of the washed product revealed a single peak of
paraoxon (Kousba and Sultatos, 2002). Acridine, iodomethane, methanol,
ether, and Dowex (1wx2, 100 mesh) were purchased from Sigma Chemcial
Company (St. Louis, MO). Human recombinant acetylcholinesterase was also
purchased from Sigma Chemical Company and was routinely stored at 70°C.
Previous investigators have established that recombinant acetylcholinesterase
is very unstable in solution unless in the presence of certain ‘‘stabilizers’’ such
as bovine serum albumin (Estrada-Mondaca and Fournier, 1998; Shafferman
et al., 1992). In the present study, bovine serum albumin at a concentration of
1 mg/ml was required in order to stabilize the enzyme in solution for up to 24 h
(data not shown). Equilibrium dialysis at 37°C as described by Kousba and
Sultatos (2002) found no evidence of binding or hydrolysis of paraoxon by
bovine serum albumin under the conditions of the incubations (data not shown).
Additionally, bovine serum albumin had no effect on human recombinant
acetylcholinesterase activity (data not shown).
Synthesis of N-methylacridinium. The fluorophore N-methylacridinium
was synthesized and purified as described by Mooser and Sigman (1974).
Briefly, 3 g of acridine were slowly dissolved in 15 ml acetone in a 200 ml
round-bottom flask. To that solution 10 ml of iodomethane were added
and stirred for 18 h. The solution was poured through filter paper using
excess acetone to remove the entire product form the flask. The filtrate
(consisting of orange crystals) was re-dissolved in 30 ml methanol and the Nmethylacridinium (iodide salt) was re-crystallized (red crystals) by the addition
of anhydrous ether (about ten volumes of ether were required). The methanol
and ether were poured off and the crystals were dried in the hood. The red
crystals were re-dissolved in about 5 ml of water to give a fluorescent green
color. This solution was passed through a Dowex chloride exchange column
(converting the iodide salt into the more stable chloride salt). The chloride salt
was recovered and the water was evaporated under a gentle stream of air. Nmethylacridinium was scanned in the UV-Vis range on a Shimadzu MPS-2000
spectrophotometer (Shimadzu MPS 2000 UV-Vis spectrophotometer, Shimadzu Scientific Instruments, Inc., Columbia, MD), and fluorescence scans
were performed on a Cary Eclipse Fluorescence Spectrophotometer (Varian,
Inc., Palo Alto, CA), with an excitation wavelength of 360 nm and an emission
wavelength of 490 nm (Mooser and Sigman, 1974). The UV-Vis and
fluorescence scans were identical to previously published spectra (Fig. 3)
(Chan et al., 1974; Eastman et al., 1995; Happ et al., 1970; Mooser and Sigman,
1974; Rosenberry et al., 1996).
Determination of active site concentration. The active site concentration in several enzyme solutions was determined by titration with Nmethylacridinium as described by Mooser and Sigman (1974). With this
approach the fluorescence of N-methylacridinium is reduced when it binds
stoichiometrically to the active site, allowing active site quantification.
Binding data were analyzed by fitting to the following equation:
B ¼ ðBmax *LÞ=ðKdm þ LÞ
Equation ð3Þ
Where B is the concentration of bound N-methylacridinium at a given
concentration of free N-methylacridinium, L; Kdm is the dissociation constant;
FIG. 3. Scans of N-methylacridinium. The top panel shows a UV-Vis scan
from 300–600 nm of N-methylacridinium at a concentration of 11 lg/ml in
buffer. The bottom panel contains the spectrofluorometric scan of the same
compound (at a concentration of 618 nM). The spectrofluorometer was set to an
excitation wavelength of 360 nm and emission wavelengths from 400–600 nm.
The maximum fluorescence was found to be 490 nm.
and Bmax is the maximum possible concentration of bound N-methylacridinium,
which is equivalent to the active site concentration. Regression analyses were
performed with Sigmaplot 3 or Sigmaplot 8 (SPSS Science Inc., Chicago, IL).
Once the active site concentration was known, an Ellman assay was performed
in order to link the rate of acetylthiocholine hydrolysis with a known enzyme
active site concentration.
Determinations of ki . A bottle of human recombinant acetylcholinesterase
(0.1 mg protein) was dissolved in 10 ml of 100 mM sodium phosphate buffer
(pH 7.4) containing 1 mg/ml bovine serum albumin (this buffer was used
throughout). Aliquots of 100 ll were placed into vials and frozen at 70°C
until the day of use. For each ki determination a vial was thawed and 14.9 ml
buffer were added to give a final volume of 15 ml. A concentrated stock
solution of paraoxon was prepared in ethanol and stored at 70°C. Paraoxon
solutions to be used for incubations were made from this stock with dilutions
made with buffer such that the ethanol concentration within any incubation
never exceeded 1%. Incubations consisted of a volume of paraoxon solution in
buffer at a given concentration with an equal volume of stock enzyme solution
that when mixed together, gave the desired final paraoxon concentration.
Incubations were carried out in an orbital shaker bath at the indicated
temperatures and time periods. At the end of an incubation, 30 ll were placed
into a well on a 96 well plate (Nunc-Immuno Module, Morris Plains, NJ), and
345 ll of buffer with acetylthiocholine (0.44 mM) were added. This dilution
along with the acetylthiocholine terminated the further inhibition of enzyme by
paraoxon since, in certain samples, addition of paraoxon after, rather than
463
KINETICS OF ACETYLCHOLINESTERASE INHIBITION
before, the addition of the buffer did not result in enzyme inhibition (data not
shown). Next 25 ll of 5,5-dithio-bis (2-nitrobenzoic acid) were added to give
a final concentration of 0.1 mM in a final volume of 400 ll. The changes in
absorbance at 405 nm were monitored at 5-min intervals for up to 20 min on
a microplate reader (BIO-TEK Instruments, Inc., Winoski, VT) (Rosenfeld
et al., 2001). The increasing absorbances were fitted to a straight line by
Sigmaplot 3 or Sigmaplot 8, and the slopes were used as a measurement of
uninhibited acetylcholinesterase activities. From these activities the enzyme
active site concentrations were calculated.
The ki’s were determined by optimization of the following equations
(Kardos and Sultatos, 2000) to the empirical inhibition data with the software
ACSL (Advanced Continuous Simulation Language, Aegis, Huntsville, AL).
These equations are based on Scheme 2, with the addition of aging and
reactivation pathways.
d½EA=dt ¼ ki *½E*½AB k3 *½EA k4 *½EA
Equation ð4Þ
d½EAge =dt ¼ k4 *½EA
Equation ð5Þ
½E ¼ ½ET ½EA ½EAge Equation ð6Þ
½AB ¼ ½ABT ½B
Equation ð7Þ
ET represents the initial active site concentration; and ABT represents the
initial paraoxon concentration (Kardos and Sultatos, 2000). All other parameters used in Equations 4–7 have the same meaning as those in Schemes 1
and 2. The rate constants for reactivation (k3) and aging (k4) were taken from
Masson et al. (2000), and were 0.005379 h1 and 0.001575 h1, respectively.
Determination of the Kd and k2 for paraoxon. Co-incubation studies
where paraoxon and p-nitrophenyl acetate were incubated concurrently with
acetylcholinesterase were performed in a manner similar to those described by
Hart and O’Brien (1973) and Barak et al. (1995). Incubation volumes were
1.6 ml, and were performed in a cuvette in a Shimadzu MPS 2000 UV-Vis
spectrophotometer. Paraoxon and p-nitrophenyl acetate in buffer were initially
added to both sample and reference cuvettes to give final inhibitor concentrations of 31 nM–1000 nM, and a 1 mM substrate concentration. The reaction
was initiated by addition of enzyme (for a final active site concentration of
0.4 pM) to the sample cuvette, while buffer only was added to the reference
cuvette. Product formation was continuously monitored by measuring the
increase in optical density at 402 nm. These data were analyzed by the method
developed by Hart and O’Brien (1973), as described by Barak et al. (1995).
Determination of k1 and k1 for paraoxon. The association (k1) and
dissociation (k1) rate constants for the reversible binding of paraoxon to
acetylcholinesterase were determined by optimizing the following equations
(referred to as the Ordered Uni Bi model, and based on Scheme 1) to the
inhibition profiles used for ki determinations for paraoxon concentrations of
25 nM, 50 nM, and 100 nM, at 24°.
d½E AB=dt ¼ k1 *½AB*½E k1 *½E AB k2 *½E AB
Equation ð8Þ
d½EA=dt ¼ k2 *½E AB k3 *½EA k4 *½EA
Equation ð9Þ
d½EAge =dt ¼ k4 *½EA
Equation ð10Þ
½E ¼ ½ET g ½E AB ½EA ½EAge
Equation ð11Þ
½AB ¼ ½ABT ½E AB ½B
Equation ð12Þ
In Equations 11 and 12 ET represents initial enzyme active site concentration
while ABT represents initial paraoxon concentration. All other parameters in
Equations 8–11 have the same meaning as in Scheme 1.
Modeling studies. All modeling studies were carried out on laptap or
desktop computers using ACSL 11.8 (Advanced Continuous Simulation
Language, Aegis, Hunstville, AL). The sensitivity of a model to an optimized
parameter was routinely assessed following optimization in order to assure the
optimized value yielded the best fit.
RESULTS
In order to minimize the number of variables determined by
the optimization routine in ACSL, acetylcholinesterase active
site concentrations in stock solutions were determined by
titration with the fluorescent compound N-methylacridinium
(Mooser and Sigman, 1974). Analyses of binding data yielded
the enzyme active site concentration (Mooser and Sigman, 1974),
as well as the dissociation constant for N-methylacridinium
(Table 1). Assessment of acetylcholinesterase activity in appropriate dilutions of the stock enzyme solutions indicated that 5 pM
active sites resulted in an increase of 0.088 optical density units
per minute in the Ellman assay, performed under the conditions
described in Materials and Methods.
Incubations of 4–6 pM human recombinant acetylcholinesterase active sites with various concentrations of paraoxon at
37° and 24° produced inhibition profiles that could be fitted
with the ki computer model to determine the ki at each
paraoxon concentration. Figure 4 presents two examples of
these analyses.
A comparison of kis generated over a wide range of inhibitor
concentrations revealed a changing ki as a function of paraoxon
concentration (Fig. 5). The fitted parameters in Figure 5 indicated that the largest ki (the y intercept) at 37° was 29.14 h1,
while the smallest ki (when paraoxon was 100 nM) was
0.18 h1. Similarly at 24°, the largest and smallest ki’s were
0.61 h1 and 0.15 h1, respectively. These data, therefore,
confirm the observations of Kardos and Sultatos (2000) with
dilute mouse brain homogenate where the ki of paraoxon
appeared to vary over a wide range of oxon concentrations.
Interestingly, at 37°, the difference between the highest
estimated ki (29.14 nM1h1) and the lowest estimated ki
(0.18 nM1h1) was considerably greater than that same
comparison at 24° (0.61 nM1h1 vs 0.15 nM1h1) (Fig. 5).
Furthermore, the determination of the ki by the method of Main
(1964) at 24° and 37° yielded a ki nearly equal to the plateau
value generated by the method of Kardos and Sultatos (2000)
at the corresponding temperature (Fig. 5).
TABLE 1
Parameters Descriptive of the Binding of N-methylacridinium
to Stock Solutions of Three Different Batches of Human
Recombinant Acetylcholinesterase
Enzme sample
Bmaxa
Kdma
Batch 1
Batch 2
Batch 3
267.5
307.8
403.5
391.8
427.7
428.6
a
nM.
464
ROSENFELD AND SULTATOS
FIG. 4. Inhibition data for human recombinant acetylcholinesterase by
78 pM (top panel) and 25 nM (lower panel) paraoxon at 24°C. In both panels
the open circles represent empirical data expressed as uninhibited active site
concentration over time, where each circle depicts a single observation. The
solid lines represent the optimized computer fit of the ki model for ki
determination (Kardos and Sultatos, 2000). The ki in the upper panel is
0.410 nM1h1, while the ki of the bottom panel is 0.111 nM1h1.
Incubation of 0.4 pM acetylcholinesterase with variable
concentrations of paraoxon and 1 mM p-nitrophenyl acetate
(Barak et al., 1995; Hart and O’Brien, 1973) yielded a Kd and
k2 for paraoxon of 339 nM and 58.1 h1, respectively (Fig. 6).
The value for k2 was directly substituted into the Ordered
Uni Bi model (Equations 8–12). The dissociation constant Kd
established the relationship between the binding k1 and k1, but
could not provide a single set of values for these binding
parameters. However, optimization of the Ordered Uni Bi
model (including the restrictions on k1 and k1 imposed by Kd)
to inhibition profiles of acetylcholinesterase in the presence of
25 nM, 50 nM, and 100 nM paraoxon determined k1 and k1 to
be 0.5 nM1h1 and 169.5 h1 respectively. The upper panel of
Figure 7 shows the final fit of the Ordered Uni Bi model to the
25 nM paraoxon inhibition profile. These paraoxon concentrations were chosen because they yielded essentially constant
estimates for ki (Fig. 5), and because the lowest paraoxon
FIG. 5. Values of ki as a function of paraoxon concentration using human
recombinant acetylcholinesterase at 37° (top panel) and 24° (bottom panel).
Incubations contained human recombinant acetylcholinesterase (active site
concentrations ranged from 4–6 pM) in 100 mM sodium phosphate buffer (pH
¼ 7.4), and the indicated paraoxon concentrations. Each open circle represents
a ki optimized from a data set containing at least 8 data points. The solid lines in
both panels represent nonlinear regression analyses of the data within each
graph to the following equation for a rectangular hyperbola: y ¼ a/(x þ b) þ c.
At 37°, a ¼ 0.0618; b ¼ 0.0021, and c ¼ 0.1790. At 24°, a ¼ 0.0467, b ¼
0.1101, and c ¼ 0.1334. The kis determined by the method of Main (1964),
using paraoxon concentrations of 10–750 nM, are represented by the dashed
lines, and are 0.125 nM1h1 and 0.180 nM1h1 at 24° and 37°, respectively.
concentration used for determination of the Kd and k2 was
31 nM (Fig. 6).
The construction of the Ordered Uni Bi model where values
for all parameters were known allowed a systematic evaluation
of the accuracy of the ki model and Scheme 2, since the ki
scheme was originally derived from the Ordered Uni Bi scheme
(Scheme 1). If the ki scheme and model accurately describe the
rate of inhibition of acetylcholinesterase by paraoxon, this rate
must equal that of the Ordered Uni Bi scheme and model.
When these two rates are assumed to be equal, the following
equations can be written:
d½EAki =dt ¼ d½EAOUB Equation ð13Þ
where EAki represents phosphylated enzyme from the ki
model, and EAOUB represents phosphylated enzyme from the
KINETICS OF ACETYLCHOLINESTERASE INHIBITION
465
FIG. 6. Double reciprocal plot used to calculate Kd and k2 for paraoxon at
24°. Slope refers to the slope of the velocity curve for the breakdown of
p-nitrophenyl acetate in the presence of 31 nM–1000 nM paraoxon (Hart
and O’Brien, 1973). The term a equals [S]/(Km þ [S]), where [S] is the
p-nitrophenyl concentration, and Km is 4.39 mM (Hart and O’Brien, 1973).
Linear regression analysis on the data gave a k2 of 51.8 h1 (the reciprocal of
the intercept of the fitted line), and a Kd of 339.2 nM (the slope of the fitted line
multiplied by k2).
Ordered Uni Bi model. Substituting Equations 4 and 9 into
Equation 13 gives:
ki *½E*½AB k3 *½EA k4 *½EA
¼ k2 *½E AB k3 *½EA k4 *½EA
Equation ð14Þ
Equation 14 rearranges to:
ki *½E*½AB ¼ k2 *½E AB
Equation ð15Þ
Substituting Equation 1 for ki gives:
Kd *½E AB ¼ ½E*½AB
Equation ð16Þ
which upon rearrangement gives:
Kd ¼ ½E*½AB=½E AB
Equation ð17Þ
Equation 17 can be recognized as descriptive of the relationship between free and reversibly bound molecules at equilibrium. Therefore, Equation 14 is simply a different form of
Equation 17, and therefore, can only be valid at equilibrium,
with respect to the reversible binding of paraoxon to acetylcholinesterase (formation of the Michaelis complex). Consequently, the ki scheme can only be accurately applied to the
inhibition of acetylcholinesterase by paraoxon (or any other
inhibitor), if reversible binding of inhibitor achieves equilibrium. This requirement for establishment of equilibrium was
not discussed by Main (1964). Simulations of the Ordered Uni
Bi model indicated that equilibrium was nearly achieved with
paraoxon and acetylcholinesterase, and that both the ki model
and the Ordered Uni Bi models yielded nearly identical outputs
at paraoxon concentrations well below the Kd (Fig. 8).
Interestingly, however, at concentrations of paraoxon approaching or exceeding its Kd, the ki model output differed
substantially from that of the Ordered Uni Bi model, thereby
FIG. 7. Simulation of the inhibition of acetylcholinesterase by the Ordered
Uni Bi kinetic model. The paraoxon concentration in the upper panel was
25 nM, and the concentration in the lower panel was 78 pM. The upper panel
represents the final fit of the Ordered Uni Bi model (solid line) to the empirical
data (open circles) to give a k1 of 0.5 nM1h1 and a k-1 of 169.5 h1. The lower
panel shows the inability of the Ordered Uni Bi model developed in the upper
panel to simulate accurately the inhibition of acetylcholinesterase by 78 pM
paraoxon. The open circles represent the empirical data while the solid line
represents the model output.
revealing a second limitation to the ki scheme. This necessity
for inhibitor concentrations well below the inhibitor Kd is an
indication of the requirement for psuedo first order conditions
in order for ki to be meaningful. The requirement for psuedo
first order conditions was discussed by Main (1964), although
he nevertheless used inhibitor concentrations larger than
their Kds.
And finally, the Ordered Uni Bi kinetic model which was
based on a Kd and k2 determined at relatively high paraoxon
concentrations (31 nM–1000 nM), did not accurately simulate
the empirically determined inhibition profiles for low paraoxon
concentrations, essentially confirming the results of the ki
model (Fig. 7, lower panel). The Ordered Uni Bi model could
accurately simulate the inhibition profiles when changes were
made in k1, k-1, or k2, but not k3 (Fig. 9).
466
ROSENFELD AND SULTATOS
FIG. 8. Comparisons of the output of the Ordered Uni Bi and ki model outputs over a wide range of paraoxon concentrations. In each panel the dashed and
solid lines represent the simulated inhibition of acetylcholinesterase by the indicated paraoxon concentration by the ki and Ordered Uni Bi models, respectively.
The Kd’ represents the calculated Kd during the Ordered Uni Bi model simulation determined from Equation 17. The experimentally determined Kd was found to be
339 nM (Fig. 6).
DISCUSSION
A changing ki as a function of paraoxon concentration is
inconsistent with the kinetic mechanism described by both the
Ordered Uni Bi and ki schemes (Schemes 1 and 2). Therefore
a consideration of the possible limitations in the use of ki is
appropriate when interpreting these results (Fig. 5). Two
important limitations were identified. The first is the requirement for equilibrium (or near-equilibrium) with regard to
reversible binding of inhibitor to acetylcholinesterase (formation of the Michaelis complex). The lack of achievement of
equilibrium with paraoxon accounted, at least in part, for the
non-identical outputs from the Ordered Uni Bi and ki models at
paraoxon concentrations lower than 100 nM. However, these
differences were likely not great enough to be detected
experimentally, and can therefore be considered negligible
(Fig. 8). Paraoxon concentrations of 200 nM and greater
revealed the second limitation in the use of ki. These high
concentrations resulted in more substantial differences between
the two models (Fig. 8), even though equilibrium conditions
were more closely approximated as the inhibitor concentration
increased (Fig. 8). This is reflective of the need for pseudo-first
order conditions, since the rate of reversible binding of
paraoxon to acetylcholinesterase slows as unbound enzyme
KINETICS OF ACETYLCHOLINESTERASE INHIBITION
467
FIG. 9. Fitting of the 78 pM paraoxon inhibition profile with the Ordered
Uni Bi model. The open circles represent the empirical data. The solid line is
output from the model after adjusting k1 from 0.5 nM1h1 to 1.75 nM1h1.
The dotted line is output after changing k-1 from 169.5 h1 to 12.5 h1. The
short dashed line is model output with k2 adjusted from 51.8 h1 to 500 h1.
The long dashed line shows the model output with k3 set to zero.
molecules become fewer with increasing paraoxon concentrations. The ki does not account for this binding phenomenon,
and will therefore overestimate the degree of inhibition
observed with high inhibitor concentrations. The need for
equilibrium conditions and pseudo-first order conditions are
limitations in the use of ki that have largely been ignored for
many decades, perhaps leading to many erroneous estimations
of this constant. In the case of paraoxon, the use of ki to
quantify its inhibitory capacity seems appropriate below
inhibitor concentrations of 100 nM (Fig. 8). Moreover, the
lower the paraoxon concentration, the more accurate was the
ki model (Fig. 8), dispelling the notion that the inhibitor
concentration must be many times higher than that of
acetylcholinesterase in order for ki to be meaningful.
Since the ki and Ordered Uni Bi models simulated essentially
the same inhibition profiles for paraoxon concentrations of
100 nM and below (Fig. 8), the changing ki as a function of
paraoxon concentration observed in Figure 5 did not result
from an inappropriate application of ki. This is further
supported by the observation that the Ordered Uni Bi model,
which utilized rate constants determined at paraoxon concentrations of 31 nM–1000 nM, also underestimated the capacity
of much lower paraoxon concentrations to inhibit acetylcholinesterase (Fig. 9).
While the presence of two kinetically distinct forms of
acetylcholinesterase could account for a changing ki as
a function of paraoxon concentration (Fig. 5), reports suggest
that such distinct enzymatic forms do not exist. As described by
Stojan et al. (1998), all known posttranslational modifications
to acetylcholinesterase do not appear to change the kinetic
behavior (Estrada-Mondaca and Fournier, 1998; Taylor and
Radić, 1994). One report (Bolger and Taylor, 1979) documented binding of certain bisquaternary ammonium ligands to
FIG. 10. Activation and inhibition paradigms for peripheral and active
gorge events to acetylcholinesterase. The peripheral anionic site is depicted as
PAS, whereas paraoxon and N-methylacridinium are depicted as PO and NMA,
respectively. See text for details and source references.
acetylcholinesterase that suggested the existence of two
conformational states of the enzyme. However these two states
were distinguished only by bisquaternary ammonium ligands,
and not by active and peripheral site ligands. Furthermore,
Bolger and Taylor (1979) established that bisquaternary
ammonium ligands bound to only one of the two possible
conformational states—a phenomenon that cannot account for
the data presented in Figure 5.
The results of the current study support the suggestion of
Kardos and Sultatos (2000) that paraoxon binds reversibly to
a site distinct from the active site, and that occupation of this
site by paraoxon reduces the capacity of other paraoxon
molecules to inhibit the active site, either by allosteric
modification or steric hindrance. Since the Ordered Uni Bi
model could accurately simulate inhibition profiles at lower
paraoxon concentrations by adjusting either k1, k-1, or k2
(Fig. 9), binding of paraoxon to a secondary site might reduce
the rate of formation of the Michaelis complex and/or reduce
the rate of phosphorylation of Ser-203. While reactivation of
phosphorylated enzyme could also be affected by occupation
of a secondary site, such a change could not account for the
concentration-dependent inhibition kinetics documented in
Figure 5 (Fig. 9).
468
ROSENFELD AND SULTATOS
Ligand-induced alterations in acetylcholinesterase activity
have been documented previously, and various mechanisms
have been proposed to account for these observations, which
include both increased and decreased activities (Fig. 10). It has
been known for many years that binding of certain ligands to
the peripheral anionic site reduces substrate hydrolysis through
steric blockade and/or conformation changes in Trp-86 and
Tyr-133 residues within the active site gorge (Barak et al.,
1995; Bourne et al., 2003; Taylor and Radić, 1994). Such
ligands include propidium, thioflavin t, acetylthiocholine, and
acetylcholine (DeFerrari et al., 2001; Taylor and Radić, 1994).
The drug D-tubocurarine also usually inhibits acetylcholinesterase through binding to the peripheral anionic site of this
enzyme (Golicnik et al., 2001). However, in the case of the
Drosphila W359L mutant, D-tubocurarine increased hydrolysis
of acetylthiocholine as a result of its binding to the peripheral
anionic site (Golicnik et al., 2001, 2002). Since in this case,
enzyme inhibition from binding of acetylchothiocholine to the
peripheral binding site was greater than enzyme inhibition from
binding of D-tubocurarine to the peripheral anionic site, the presence of D-tubocurarine appeared to activate enzyme activity.
Although the apparent activation of the W359L mutant by Dtubocurarine resulted from a ‘‘disinhibition,’’ occupation of the
peripheral anionic site of wild type Drosophila acetylcholinesterase by this same compound increased the binding affinity of
the active site for methanesulfonylfluoride, thereby increasing
the methanesulfonylation of the active site serine (Golicnik
et al., 2002). Similarly, binding of the detergent Triton X-100 to
the peripheral anionic site of Drosophila acetylcholinesterase
resulted in activation of phosphorylation and carbamylation of
the active site by certain organophosphates and carbamates,
respectively, while inhibiting these same reactions with other
organophosphates and carbamates (Marcel et al., 2000).
It must also be noted that binding to the peripheral anionic
site is not requisite for activation of acetylcholinesterase. The
chemical N-methylacridinium binds reversibly to the active
site, and accelerates the hydrolysis of certain neutral acetic
acid esters such as methyl and ethyl acetate, while competitively inhibiting the hydrolysis of acetylcholine (Barnett and
Rosenberry, 1977). Similarly, small cationic ligands such as
tetramethyl- and tetraethyl-ammonium have been shown to
bind in the active site and enhance the rate of carbamylation
and sulfonylation by carbamyl and sulfonyl fluorides (Kitz and
Wilson, 1963; Metzger and Wilson, 1963).
The reports summarized above have established that ligand
binding to the peripheral anionic site, or other sites on
acetylcholinesterase, can result, in some cases, in inhibition
of enzyme activity. In other cases, such binding can lead to
apparent activation (from ‘‘disinhibition’’), and still in other
cases true activation of enzyme. The specific action of the
ligand on enzyme activity appears to be dependent on the
ligand itself, where it binds on the molecule, and the chemicals
interacting at the active site. Consequently, it seems plausible,
as first suggested by Kardos and Sultatos (2000), that paraoxon
could bind to a secondary site on acetylcholinesterase, thereby
reducing the capacity of other paraoxon molecules to phosphylate the Ser-203. Regardless of mechanism, the concentration dependent inhibition presented in Figure 5 demonstrates
the inability of Schemes 1 and 2 to fully describe the inhibition
of acetylcholinesterase by paraoxon. The documentation of
concentration dependent inhibition of rat acetylcholinesterase
with chlorpyrifos oxon (Kousba et al., 2004) indicates that this
phenomenon is not unique to paraoxon.
While the potential binding of certain organophosphorus
anticholinesterases to a secondary binding site is of interest for
understanding the fundamental interactions of these toxic
compounds with acetylcholinesterase, a more important consideration is how such binding might influence our understanding of the health risks posed to the public by these insecticides.
Current risk assessments of organophosphorus insecticides
are built on the assumption that the inhibitory capacity (ki) of
molecules of a given oxon at low concentrations is the same as
that of the same oxon molecules at high concentrations.
Evidence presented here (Fig. 5), as well as elsewhere (Kardos
and Sultatos, 2000; Kousba et al., 2004), suggests that the
inhibitory capacity of certain oxons increases as the concentration decreases, thereby raising concern that current risk
assessments might underestimate the risk posed to public health
by low level exposure to certain organophosphorus insecticides.
And finally, further evidence in support of the interaction
of certain organophosphorus compounds with a secondary site
on acetylcholinesterase was provided by Howard et al. (2005),
who showed that the organophosphorus insecticide chlorpyrifos disrupted the morphogenic function of acetylcholinesterase in the absence of inhibition of the catalytic site.
Acetylcholinesterase has been shown to mediate neuronal
morphogenesis independently of its catalytic site (Bigbee
et al., 1999; Johnson and More, 2000; Sharma et al., 2001;
Soreq and Seidman, 2001; Sternfeld et al., 1998). Consequently, binding of certain organophosphorus compounds to
a secondary site(s) distinct from the catalytic site could not
only reduce the capacity of additional organophosphorus
molecules to inhibit the active site (Fig. 5), but might also
have adverse developmental consequences unrelated to phosphylation of Ser-203.
ACKNOWLEDGMENT
This study was supported in part by Grant ES012648 from NIEHS.
REFERENCES
Aldridge, W. N., and Reiner, E. (1972). Enzyme inhibitors as substrates:
Interactions of esterases with esters of organophosphorus and carbamic
acids. In Frontiers of Biology, Vol. 26 (A. Neuberger and E. L. Tatum, Eds.),
p. 41. North-Holland, New York.
Barak, D., Ordentlich, A., Bromberg, A., Kronman, C., Marcus, D., Lazor, A.,
Ariel, N., Velan, B., and Shafferman, A. (1995). Allosteric modulation of
KINETICS OF ACETYLCHOLINESTERASE INHIBITION
469
acetylcholinesterase activity by peripheral ligands involves a conformational
transition of the anionic subsite. Biochemistry 34, 15444–15452.
Main, A. R. (1964). Affinity and phosphorylation constants for the inhibition of
esterases by organophosphates. Science 144, 992–993.
Barnett, P., and Rosenberry, T. L. (1977). Catalysis by acetylcholinesterase.
Acceleration of the hydrolysis of neutral acetic acid esters by certain
aromatic cations. J. Biol. Chem. 252, 7200–7206.
Marcel, V. S., Estrada-Mondaca, S., Magne, F., Stojan, J., Klaebe, A., and
Fournier, D. (2000). Exploration of the Drosophila acetylcholinesterase
substrate activation site using a reversible inhibitor (Triton X-100) and
mutated enzymes. J. Biol. Chem. 275, 11603–11609.
Bigbee, J. W., Sharma, K. V., Gupta, J. J., and Dupree, J. L. (1999).
Morphogenic role for acetylcholinesterase in axonal outgrowth during neural
development. Environ. Health Persp. 107(Suppl. 1), 81–87.
Bolger, M. P., and Taylor, P. (1979). Kinetics of association between
bisquaternary ammonium ligands and acetylcholinesterase. Evidence for
two conformational states of the enzyme from stopped-flow measurements
of fluorescence. Biochemistry 18, 3622–3629.
Bourne, Y., Taylor, P., Radić, Z., and Marchot, P. (2003). Structural insights into
ligand interactions at the acetylcholinesterase peripheral anionic site. EMBO
22, 1–12.
Chan, L. M., Himel, C. M., and Main, A. R. (1974). Active-site directed
fluorescent probes in the kinetics and spectroscopy of purified horse serum
cholinesterase. Biochemistry 13, 86–90.
DeFerrari, G. V., Mallender, W. D., Inestrosa, N. C., and Rosenberry, T. L.
(2001). Thioflavin T is a fluorescent probe of the acetylcholinesterase
peripheral site that reveals conformational interactions between the peripheral and acylation sites. J. Biol. Chem. 276, 23282–23287.
Eastman, J., Wilson, E. J., Cervenansky, C., and Rosenberry, T. L. (1995).
Fasciculin 2 binds to the peripheral site on acetylcholinesterase and inhibits
substrate hydrolysis by slowing a step involving proton transfer during
enzyme acylation. J. Biol. Chem. 270, 19694–19701.
Estrada-Mondaca, S., and Fournier, D. (1998). Stabilization of recombinant
Drosophila acetylcholinesterase. Protein Expr. Purif. 12, 166–172.
Golicnik, M. D., Fournier, D., and Stojan, J. (2001). Interaction of Drosophila
acetylcholinesterase with D-tubocurarine: An explanation of the activation
by an inhibitor. Biochemistry 40, 1214–1219.
Golicnik, M. D., Fournier, D., and Stojan, J. (2002). Acceleration of Drosophila
melangaster acetylcholinesterase methanesulfonylation: peripheral ligand Dtubocurarine enhances the affinity for small methanesulfonylfluoride. Chem.
Biol. Interact. 139, 145–157.
Happ, J. W., Janzen, E. G., and Rudy, B. C. (1970). Electron spin resonance
studies of radical formation in nucleophilic addition reactions. III. Kinetic and
proton magnetic resonance investigations of the interaction of methoxide anions
with 1-methyoxy-2,4,5-trinitronaphthalene. J. Org. Chem. 35, 3382–3390.
Hart, G. J., and O’Brien, R. D. (1973). Recording spectrophotometric method
for determination of dissociation and phosphorylation constants for the
inhibition of acetylcholinesterase by organophosphates in the presence of
substrate. Biochemistry 12, 2940–2945.
Howard, A. S., Bucelli, R., Jett, D. A., Bruun, D., Yang, D., and Lein, P. J.
(2005). Chlorpyrifos exerts opposing effects on axonal and dendritic growth
in primary neuronal cultures. Toxicol. Appl. Pharmacol. 207, 112–124.
Johnson, G., and Moore, S. W. (2000). Cholinesterases modulate cell adhesion
in human neuroblastoma cells in vitro. Int. J. Devl. Neurosci. 18, 781–790.
Kardos, S. A., and Sultatos, L. G. (2000). Interactions of the organophosphates
paraoxon and methyl paraoxon with mouse brain acetylcholinesterase.
Toxicol. Sci. 58, 118–126.
Kitz, R., and Wilson, I. B. (1963). Acceleration of the rate of reaction of
methanesulfonyl fluoride and acetylcholinesterase by substituted ammonium
ions. J. Biol. Chem. 238, 757–748.
Kousba, A. A., and Sultatos, L. G. (2002). Continuous systems modeling of
equilibrium dialysis for determinations of tissue partitioning of parathion and
paraoxon. Toxicol. Lett. 133, 153–159.
Kousba, A. A., Sultatos, L. G., Poet, T. S., and Timchalk, C. (2004).
Comparison of chlorpyrifos-oxon and paraoxon acetylcholinesterase inhibition dynamics: Potential role of a peripheral binding site. Toxicol. Sci. 80,
239–248.
Masson, H. J., Sains, C., Stevenson, A. J., and Rawbone, R. (2000). Rates of
spontaneous reactivation and aging of acetylcholinesterase in human
erythrocytes after inhibition by organophosphorus pesticides. Hum. Exp.
Toxicol. 19, 511–516.
Masson, P., Goldstein, B. N., Debouzy, J.-C., Froment, M.-T., Lockridge, O.,
and Schopfer, L. M. (2004). Damped oscillatory hysteretic behaviour of
butyrylcholinesterase with benzoylcholine as substrate. Eur. J. Biochem. 271,
220–234.
Metzger, H. P., and Wilson, I. B. (1963). Acceleration of the rate of reaction of
dimethylcarbamyl fluoride and acetylcholinesterase by substituted ammonium ions. J. Biol. Chem. 238, 3432–3435.
Mileson, B. E., Chambers, J. E., Chen, W. L., Dettburn, W., Ehrich, M.,
Eldefrawi, A. T., Gaylor, D. W., Hamernik, K., Hodgson, E., Karczmar, A. G.,
et al. (1998). Common mechanism of toxicity: A case study of organophosphorus pesticides. Toxicol. Sci. 41, 8–20.
Mooser, G., and Sigman, D. S. (1974). Ligand binding properties of acetylcholinesterase determined with fluorescent probes. Biochemistry 13, 2299–2307.
Ordentlich, A., Barak, D., Kronman, C., Ariel, N., Segall, Y., Velan, B., and
Shafferman, A. (1996). The architecture of human acetylcholinesterase
active center probed by interactions with selected organophosphate inhibitors. J. Biol. Chem. 271, 11953–11962.
Ordentlich, A., Barak, D., Kronman, C., Ariel, N., Segall, Y., Velan, B., and
Shafferman, A. (1998). Functional characteristics of the oxyanion hole in
human acetylcholinesterase. J. Biol. Chem. 278, 19509–195127.
Ordentlich, A., Barak, D., Sod-Moriah, G., Kaplan, D., Mizrahi, D., Segall, Y.,
Kronman, C., Karton, Y., Lazar, A., Marcus, D., et al. (2004). Stereoselectivity toward VX is determined by interactions with residues of the acyl
pocket as well as of the peripheral anionic site of AChE. Biochemistry 43,
11255–11265.
Rosenberry, T. L., Rabl, C. R., and Neumann, E. (1996). Binding of the
neurotoxin fasciculin 2 to the acetylcholinesterase peripheral site
drastically reduces the association and dissociation rate constants for
N-methylacridinium binding to the active site. Biochemistry 35, 685–690.
Rosenfeld, C., Kousba, A., and Sultatos, L. G. (2001). Interactions of rat brain
acetylcholinesterase with the detergent Triton X-100 and the organophosphate paraoxon. Toxicol. Sci. 63, 208–213.
Shafferman, A., Kronman, C., Flashner, Y., Leitner, M., Grosfeld, H.,
Ordentlich, A., Gozes, Y., Cohen, S., Ariel, N., Barak, D., et al. (1992).
Mutagenesis of human acetylcholinesterase. Identification of residues
involved in catalytic activity and in polypeptide folding. J. Biol. Chem.
267, 17640–17648.
Sharma, K. V., Koenigsberger, C., Brimijoin, S., and Bigbee, J. W. (2001).
Direct evidence for an adhesive function in the noncholinergic role of
acetylcholinesterase in neurite outgrowth. J. Neurosci. Res. 63, 165–175.
Soreq, H., and Seidman, S. (2001). Acetylcholinesterase – new roles for an old
actor. Nat. Rev. Neuro. 2, 294–301.
Sternfeld, M., Ming, G., Song, H., Sela, K., Timberg, R., Poo, M., and Soreq, H.
(1998). Acetylcholinesterase enhances neurite growth and synapse development through alternative contributions of its hydrolytic capacity, core
protein, and variable C termini. J. Neurosci. 18, 1240–1249.
Stojan, J., Marcel, V., Estrada-Mondaca, S., Klaebe, A., Masson, P., and
Fournier, D. (1998). A putative kinetic model for substrate metabolisation
by Drosophila acetylcholinesterase. FEBS Lett. 440, 85–88.
Taylor, R., and Radić, Z. (1994). The cholinesterases: From genes to proteins.
Annu. Rev. Pharmacol. Toxicol. 34, 281–320.