![Inhibition of 2 nM [ 3 H]CGP 39653 binding by 1 μM and 100 μM inhibitors](https://www.researchgate.net/profile/Stephen-Lantz/publication/303935436/figure/fig2/AS:372568302342146@1465838613597/nhibition-of-2-nM-3-HCGP-39653-binding-by-1-mM-and-100-mM-inhibitors_Q320.jpg)
Content uploaded by Stephen R Lantz
Author content
All content in this area was uploaded by Stephen R Lantz on Jun 13, 2016
Content may be subject to copyright.
Atypical Organophosphorus Toxicology of the Herbicides Glufosinate and Ethephon
by
Stephen R. Lantz
A dissertation submitted in partial satisfaction of the
requirements for the degree of
Doctor of Philosophy
in
Molecular Toxicology
in the
Graduate Division
of the
University of California, Berkeley
Committee in charge:
Professor John E. Casida, Chair
Professor Chris D. Vulpe
Professor Diana Bautista
Fall 2013
All rights reserved
INFORMATION TO ALL USERS
The quality of this reproduction is dependent upon the quality of the copy submitted.
In the unlikely event that the author did not send a complete manuscript
and there are missing pages, these will be noted. Also, if material had to be removed,
a note will indicate the deletion.
Microform Edition © ProQuest LLC.
All rights reserved. This work is protected against
unauthorized copying under Title 17, United States Code
ProQuest LLC.
789 East Eisenhower Parkway
P.O. Box 1346
Ann Arbor, MI 48106 - 1346
UMI 3686378
Published by ProQuest LLC (2015). Copyright in the Dissertation held by the Author.
UMI Number: 3686378
1
Abstract
Atypical Organophosphorus Toxicology of the Herbicides Glufosinate and Ethephon
by
Stephen R. Lantz
Doctor of Philosophy in Molecular Toxicology
University of California, Berkeley
Professor John E. Casida, Chair
Organophosphorus (OP) compounds have found tremendous use as industrial
chemicals, pharmaceuticals and chemical warfare agents. The intended toxicity of warfare
agents and the unwanted toxicity of industrial and agrochemicals results primarily from
inhibition of cholinesterase enzymes, predominantly acetylcholinesterase (AChE).
Functionalization of the OP class has led to development of compounds with selective action at
various primary protein targets and the identification of numerous secondary and tertiary off
target effects. Two OP compounds, glufosinate (GLF) and ethephon, have significant
agricultural use as an herbicide and plant growth regulator respectively. They are also
characterized by atypical mammalian OP toxicity. This dissertation explores newly
characterized aspects of GLF and ethephon mammalian toxicology.
GLF can cause convulsions in mammals without AChE inhibition. The structural
similarity of GLF to glutamate and the known excitatory mechanisms of glutamate signaling
implicate the glutamatergic system as a target for GLF neurotoxicity. Known glutamate targets
which could reasonably be expected to be involved in a convulsive phenotype after modulation
were assayed for effect by GLF and related compounds. There appears to be no effect of GLF or
the primary metabolite N-acetyl-GLF (N-Ac-GLF) on the urea cycle regulator enzyme CPS-I.
However, GLF and N-Ac-GLF can interact directly with the N-methyl-D-aspartate (NMDA)
subtype glutamate receptor and the high affinity glutamate reuptake transporter GLT-I.
Potency at the NMDAR is in the range of 10-10,000 μM and potency at GLT-I is >1,000 μM. The
concentrations required to affect transport through GLT-I are likely not to be attained and
therefore not relevant to the neurotoxic mode of action. However, toxicokinetic data from
reports of intentional human poisonings suggest that the concentration of GLF in the brain
could reach high enough levels after acute exposure to account for the convulsive and memory
loss effects. Furthermore, the newly characterized action of N-Ac-GLF at the NMDAR suggests
that both the parent compound and metabolite could contribute to neurotoxicity.
Direct radioligand binding experiments are one of the most unbiased approaches to
target site identification. However, the preparation of high specific activity radioligands for
compounds of interest is specialized and costly. Progress toward the preparation of [
3
H-Me]
GLF is reported with successful synthesis of unlabeled GLF prepared with methyl addition and
deprotection as the last synthetic step. Although the synthetic procedure was not scalable for
2
economical radiosynthesis, the method provides a platform for further optimization and
eventual radioligand synthesis.
Ethephon is well characterized to readily degrade to ethylene, which acts as a regulator
of plant growth. Butyrylcholinesterase (BChE) is selectively inhibited after exposure to
ethephon. Ethephon itself does not inhibit BChE, but is degraded to a BChE inhibitor in alkaline
solutions. In the present study,
31
P NMR monitoring of ethephon degradation over time
demonstrates that a transitory intermediate is formed during the degradation process. The
chemical shift of the intermediate is as expected for 2-oxo-2-hydroxy-1,2-oxaphosphetane, the
proposed BChE inhibitor. The degradation of ethephon proceeds through competitive
reactions, forming either phosphate and ethylene directly or the BChE inhibitor. This
mechanism is supported by kinetic modeling which is used to generate rate constants and T
1/2
values that are in agreement with reported ethephon degradation rates.
OP chemicals are known to interact with a wide variety of biochemical targets.
Functionalization has allowed for the preparation of OP chemicals that have surprising species
and target site selectivity. However, complete target site specificity is never achieved as most
OP chemicals have some secondary or tertiary effects at alternate sites. Furthermore, OP
herbicides are not specific for plants. The main goal of toxicological profiling is therefore to
identify the type and threshold for activity at all targets. The mechanistic insight obtained
through these studies allows for better toxicological evaluation of GLF and ethephon.
i
Dedications
To my family and friends near and far who bestow endless wisdom in humility,
compassion, perseverance, dedication and love.
and
To nature, which never ceases to awe and inspire.
ii
Table of Contents
Abstract ........................................................................................................................................... 1
Dedications ....................................................................................................................................... i
Table of Contents ............................................................................................................................. ii
List of Figures ................................................................................................................................... v
List of Tables ................................................................................................................................... vi
Symbols and Abbreviations ........................................................................................................... vii
Acknowledgements ......................................................................................................................... ix
Chapter I: Glufosinate Neurotoxicity Mediated by NMDAR Activation ......................................... 1
Introduction ........................................................................................................................ 1
Glufosinate Use ....................................................................................................... 1
Herbicidal Mode of Action ...................................................................................... 1
GLA Herbicide Tolerance......................................................................................... 1
Human Poisonings .................................................................................................. 2
Mammalian Neurotoxicity Targets Implicated from Chemical Structure
and Poisonings .................................................................................................. 2
Radioligand Binding and Bioassay in Target Site Analysis ...................................... 4
Materials and Methods ....................................................................................................... 4
Chemicals ................................................................................................................ 4
Buffers ..................................................................................................................... 5
Crude Forebrain Synaptic Plasma Membrane Preparation .................................... 5
[
3
H]CGP 39653 Binding Inhibition ........................................................................... 5
NMDAR Electrophysiology ...................................................................................... 6
HEK293-GLT-I Cell Culture ....................................................................................... 6
GLT-I Mediated [
3
H]-GLU Uptake ............................................................................ 6
Culture of Rat Primary Cortical Neurons ................................................................ 7
MEA Recordings ...................................................................................................... 7
Viability of Rat Primary Cortical Neurons Treated with GLF and Related
Compounds ....................................................................................................... 8
Desalted Mouse Liver Homogenate ....................................................................... 8
CPS-I Activity ........................................................................................................... 8
Data Analysis ........................................................................................................... 9
Results ................................................................................................................................. 9
iii
Effect on Urea Cycle Regulation (CPS-I Activity) ..................................................... 9
Effects on NMDAR (Binding) ................................................................................... 9
Effect on NR1/NR2A subtype NMDAR (Electrophysiology) .................................... 9
Effects on Glutamate Reuptake in the Brain (GLT-I Assay) .................................. 15
Effects on Primary Cultured Rat Cortical Neurons (MEA Recordings) .................. 15
Cell Viability and pH in Primary Cortical Neuron Cultures.................................... 23
Discussion.......................................................................................................................... 23
Neurotoxicity Mechanistic Considerations ........................................................... 23
Urea Cycle Effects ................................................................................................. 23
Glutamatergic Signaling Effects ............................................................................ 26
Neuronal Signaling Effects .................................................................................... 26
Correlation of in Vitro Results to in Vivo Toxicity ................................................. 28
Potential for Exposure .......................................................................................... 29
Summary ............................................................................................................... 30
Chapter II: Progress in Glufosinate Radiosynthesis ...................................................................... 31
Introduction ...................................................................................................................... 31
GLF Radioligand Utility .......................................................................................... 31
GLF Radiosynthetic Approaches ........................................................................... 31
Materials and Methods ..................................................................................................... 31
Chemicals .............................................................................................................. 31
NMR Spectroscopy for Structural Determination ................................................ 31
Synthetic Procedures ............................................................................................ 31
Results ............................................................................................................................... 34
GLF Synthesis Amenable to Radiolabel Incorporation ......................................... 34
Discussion.......................................................................................................................... 35
Chapter III: Characterization of the Transient Oxaphosphetane BChE Inhibitor Formed
from Spontaneously-Activated Ethephon .............................................................................. 36
Introduction ...................................................................................................................... 36
Ethephon Inhibition of BChE ................................................................................. 36
Mechanism of Ethephon Degradation .................................................................. 36
Materials and Methods ..................................................................................................... 36
Chemicals and Buffers .......................................................................................... 36
Buffer Comparison ................................................................................................ 38
iv
31
P NMR Kinetic Time Course ................................................................................ 38
Derivation of Integrated Rate Equations .............................................................. 38
Integrated Rate Equations Fitting and Determination of Parameter
Confidence Intervals ....................................................................................... 41
Determination of T
1/2
Values for all Species ......................................................... 42
Determination of Total % Formation for B-E ........................................................ 42
Ethephon Standard Spectrum............................................................................... 42
Preparation of 2-Hydroxyethylphosphonic acid ................................................... 42
Monomethyl 2-Hydroxyethylphosphonic acid ..................................................... 42
2-Oxo-2-hydroxy-1,2-oxaphosphetane ................................................................. 42
Results ............................................................................................................................... 42
Buffer System for
31
P NMR Time Course Experiments ......................................... 42
Characterization of Ethephon Degradation .......................................................... 44
Kinetic Modeling of
31
P NMR Data ........................................................................ 47
Discussion.......................................................................................................................... 47
Goodness of Fit ..................................................................................................... 47
Ethephon Degradation Rate ................................................................................. 47
Degradation Products Other than Phosphate ...................................................... 51
BChE Inhibitor ....................................................................................................... 51
Multiple Degradation Pathways ........................................................................... 51
Summary ............................................................................................................... 51
Conclusions ................................................................................................................................... 53
References .................................................................................................................................... 54
v
List of Figures
Figure 1: Chemical structures of GLF and related compounds. ..................................................... 2
Figure 2: GLF neurotoxicity targets implicated by structural similarity to GLU and
convulsive phenotype. .............................................................................................................. 3
Figure 3: CPS-I activity in the presence and absence of D,L-GLF or N-Ac-D,L-GLF ....................... 10
Figure 4: Inhibition of 2 nM [
3
H]CGP 39653 binding by L-GLU and D,L-GLF................................. 11
Figure 5: Inhibition of 2 nM [
3
H]CGP 39653 binding by 1 μM and 100 μM inhibitors ................. 12
Figure 6: TEVC current traces for activity of D,L-GLF on NR1/NR2A transfected Xenopus
oocytes .................................................................................................................................... 13
Figure 7: Partial concentration response to D,L-GLF in NR1/NR2A subtype NMDAR .................. 14
Figure 8: TEVC current traces of D,L-GLF activity in untransfected (control) oocytes
compared to NR1/NR2A transfected oocytes ........................................................................ 16
Figure 9: Inhibition of [
3
H]GLU uptake by WAY213613 ................................................................ 17
Figure 10: Inhibition of [
3
H]GLU uptake by D,L-GLF (top) or N-Ac-D,L-GLF (bottom) .................. 18
Figure 11: LDH release in GLT-I expressing HEK-293 cells ............................................................ 19
Figure 12: Cell counting of GLT-I expressing HEK-293 cells treated with D,L-GLF........................ 20
Figure 13: Concentration response for D,L-GLF and related compounds in MEA assay .............. 21
Figure 14: Effect of 1 μM MK801 pretreatment on NMDA and D,L-GLF concentration
responses in MEA .................................................................................................................... 22
Figure 15: Change in MEA active electrodes after application of GLF and related
compounds ............................................................................................................................. 24
Figure 16: Cell viability of primary rat cortical neurons treated with MEA test compounds ....... 25
Figure 17: Attempted [
3
H-2,3]GLF preparation method .............................................................. 32
Figure 18: Attempted [
3
H-Me]GLF preparation method .............................................................. 33
Figure 19: Possible degradation reactions of ethephon in aqueous solution at pH > 6 .............. 37
Figure 20: Comparison of BChE inhibition by ethephon pre-incubated in carbonate or
phosphate buffers for 0 or 20 h .............................................................................................. 43
Figure 21:
31
P NMR time course spectra for 200 mM ethephon degradation in 2 M pH
7.4 potassium carbonate buffer ............................................................................................. 45
Figure 22:
1
H NMR spectra of ethephon degradation in pH 7.4 potassium carbonate
buffer at 1, 21 and 96 h .......................................................................................................... 46
Figure 23:
1
H coupled
31
P NMR spectra of ethephon degradation in pH 7.4 potassium
carbonate buffer at 21 and 96 h ............................................................................................. 48
Figure 24: Graphical fitting results of concentration-time integrated rate equations ................ 49
Figure 25: Comparison of BChE inhibitor concentration-time data from BChE trapping
experiments and
31
P NMR monitoring ................................................................................... 52
vi
List of Tables
Table 1: Comparison of NMDAR Binding and MEA Activity Data ................................................. 27
Table 2: Fitting Results of
31
P NMR Data to Concentration-Time Rate Equations ....................... 50
Table 3: Rate Constants and T
1/2
Values Derived from Monte Carlo Simulation ......................... 50
vii
Symbols and Abbreviations
Ac acetyl
AcCoA acetyl-coenzyme A
ACh acetylcholine
AChE acetylcholinesterase
AMPAR α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid receptor
ATP adenosine triphosphate
bar bialaphos resistance gene
BChE butyrylcholinesterase
BOS barium oocyte solution
CPS-I carbamyl phosphate synthase I
CV cardiovascular
DMEM Dulbecco's Modified Eagle Medium
EAAT excitatory amino acid transporter
EPA US Environmental Protection Agency
GABA γ–aminobutyric acid
GAD glutamic acid decarboxylase
GDH glutamic acid dehydrogenase
GI gastrointestinal
GLA glufosinate ammonium
GLF glufosinate
GLU glutamic acid or glutamate
GluR glutamate receptor
Gly glycine
GM genetically modified
GS glutamine synthetase
HBS HEPES buffered saline
HU hydroxyurea
iGluR ionotropic glutamate receptor
KR kainic acid receptor
LHB liver homogenization buffer
MEA microelectrode array
mGluR metabotropic glutamate receptor
MHz megahertz
MSO methionine sulfoximine
N-Ac-GLF N-acetyl-glufosinate
N-Ac-GLU N-acetyl-glutamate
NAGS N-acetyl-L-glutamate synthase
NMDA N-methyl-D-aspartate
NMDAR N-methyl-D-aspartate receptor
NMR nuclear magnetic resonance
NT neurotransmitter
OP organophosphorus
viii
PAT phosphinothricin acetyltransferase enzyme
pat phosphinothricin acetyltransferase gene
RMSE root-mean-square error
TEVC two electrode voltage clamp
ix
Acknowledgements
General
The essence of the projects described herein derives from the extensive experience with
agricultural chemicals of my graduate advisor, Professor John Casida. His intuition was
essential to the development of this work. For this I am very grateful. Additional dissertation
committee members Professor Chris Vulpe and Professor Diana Bautista offered much
encouragement and suggestion during the progression of my dissertation research. Their
demeanor and varied knowledge of the scientific disciplines involved was a welcomed addition
to committee discussions. The assistance and enthusiasm of two undergraduate students Sean
D. Kodani and Ellen F. Key at UC Berkeley made many of the experiments possible. Brandon
Gaytan, a graduate student in the laboratory of Professor Chris Vulpe at UC Berkeley was
trusting enough of my experience with radioactive materials to seek collaboration on a side
project pertaining to dieldrin effects on leucine bioavailability in yeast, leading to publication.
In addition, Brandon’s weekend camaraderie in the city or the country has left me with many
fond memories of travel, adventure, discovery and humor. Dr. Chris Canlas in the NMR
laboratory at UC Berkeley was always helpful with any questions I had regarding the
instruments in his superbly maintained facility. Fellow graduate students in the Department of
Nutritional Sciences and Toxicology, in particular Dr. Charles Krois, Vanessa De La Rosa, Dr.
Tami Swenson, Kristin Obrochta, Katie Hall, Dr. Holly Nicastro, Dr. Nora Gray and Dr. Max Ruby
were supportive and informative at various stages of my graduate experience. Last and
certainly not least, my parents, brothers, sister, nephew and girlfriend provide essential moral
support and encouragement in all of my endeavors.
Funding
My graduate studies were funded by STAR Fellowship Assistance Agreement number FP917139
awarded by the U.S. Environmental Protection Agency (EPA). This dissertation has not been
formally reviewed by EPA and the views expressed are solely my own. EPA does not endorse
any products or commercial services mentioned in this article.
Glufosinate
Dr. Timothy Shafer, his laboratory members Dr. Cina Mack, Diana Hall, and Jenna Strickland,
and their colleague Dr. Kathleen Wallace all at the Integrated Systems Toxicology Division,
National Health and Environmental Effects Research Laboratory of the EPA made the
microelectrode array experiments possible through their hospitality and expertise at Research
Triangle Park. I am eternally grateful for their collaboration. Dr. Ryan Ahrant in the laboratory
of Professor Ehud Isacoff at UC Berkeley provided generous assistance with electrophysiology
of NMDAR expressed in Xenopous oocytes. Without his time and efforts, these experiments
simply would not have been possible. Dr. Ann Fischer and her technicians in the Cell and Tissue
Culture Facility at UC Berkeley were flawless in cell culture maintenance and plate preparation
for the GLT-I experiments. Visiting with Ann for cell pickup was always an enjoyable
experience. Ellen F. Key performed numerous replicates for the urea cycle inhibition and
glutamate uptake assays. Dr. Marcus Rattray at the University of Reading, UK generously
provided HEK-293 cells stably expressing the high affinity glutamate transporter GLT-I.
Professor Wolfgang Maison at Universität Hamburg, Germany generously provided the
protected vinyl glycine starting material for the attempted synthesis of [
3
H-Me]-GLF. Professor
x
Jean Luc Montchamp at Texas Christian University gladly offered quick and prudent advice
regarding phosphinate synthesis. Vanessa De La Rosa and Mani Tagmount in the laboratory of
Professor Chris Vulpe at UC Berkeley were highly informative and supportive regarding
methods for cytotoxicity testing and sterile technique. Dr. Charles Krois in the laboratory of
Professor Joseph Napoli at UC Berkeley was very kind to provide isolated liver mitochondria for
urea cycle assays and much advice in all areas of molecular biology and analytic chemistry. In
addition, Chuck’s spots at the gym and immense knowledge of Schwarzenegger films added
laughter and variety to my experience at UC Berkeley.
Ethephon
Professor Oksana Lockridge and her laboratory members Dr. Lawrence Shopfer and Dr. Mariya
Liyasova undertook the ethephon project with our laboratory. The initial publication from their
collaboration set the groundwork for the ethephon project described in Chapter III. Sean D.
Kodani contributed to much of the experimental work with ethephon and BChE.
1
Chapter I: Glufosinate Neurotoxicity Mediated by NMDAR Activation
Introduction
Glufosinate Use
(RS)-2-Amino-4-(hydroxy(methyl)phosphonoyl)butanoic acid (phosphinothricin or glufosinate
(GLF)) is a phosphorus containing amino acid that is naturally occurring as a component of the
bacteria-derived bactericidal and fungicidal tripeptides bialaphos
1
and phosalacine.
2,3
The
ammonium salt of GLF (GLA) has widespread use as an herbicide in formulations referred to as
Liberty®, Rely® and Finale®. The use of GLA herbicide has increased dramatically over the
past decade, primarily to fill the void created as a result of weed resistance to the major
herbicide glyphosate (Roundup™).
4
As of 2010 GLA was the 19
th
ranking agrochemical in terms
of sales with a worldwide market share of about $400 million.
GLA herbicide is a favored replacement for glyphosate for several reasons. The two
herbicides are cheap to manufacture, and have similar physical and chemical properties (water
soluble, zwitterionic organophosphorus salts). Both are used in broad spectrum applications to
control a wide variety of weeds. Genetically modified (GM) varieties of major crops with GLF
tolerance are available (see ‘GLA Herbicide Tolerance’ below). Normal application of GLA
herbicide as a directed spray to soil has no detrimental effect to fruit and nut trees. Lastly,
while glyphosate herbicide is a competitive inhibitor of 5-enolpyruvylshikimate-3-phosphate
synthase, an essential enzyme for aromatic amino acid synthesis in bacteria, fungi, and plants,
GLA herbicide is a competitive inhibitor of glutamine synthetase (GS), an essential enzyme in
ammonia detoxification in bacteria, fungi, and plants.
1
Thus glyphosate and GLA have different
herbicidal mechanisms and no cross resistance.
Herbicidal Mode of Action
GLF produces plant, bacterial and fungal toxicity through inhibition of GS. GS catalyzes the
reaction L-glutamate + ammonia L-glutamine + H
2
O. Inhibition of GS in susceptible plants,
fungi and bacteria results in ammonia buildup, altered amino acid homeostasis and ultimately
death.
5–7
Support for GLF herbicidal action at GS comes from measurements of GS inhibitory
potency,
1,8–10
X-ray crystal structure determination of GLF bound to Salmonella typhimurium GS
at the substrate binding pocket in the active site,
11
and GLA herbicide resistance in Italian
ryegrass resulting from a point mutation at the GS active site.
12
GLA Herbicide Tolerance
Two genes which confer GLF tolerance in bacteria have been identified. The bar gene from
Streptomyces hygroscopicus
13
and the pat gene from Streptomyces viridochromogenes
14
both
encode an acetyltransferase enzyme which catalyzes the addition of an acetyl (Ac) group to the
amino moiety of GLF.
15,16
The resulting N-Ac-GLF metabolite is herbicidally inactive.
15
Crops
engineered for GLF tolerance incorporate the bacterial bar or pat gene into the plant genome,
allowing the GM crop to detoxify the herbicide. As of 2011, GM varieties expressing the bar or
pat gene are authorized for canola, cotton, maize, soybean, rice, and sugar beet.
17
The
abundance of tolerant crop varieties contributes to the extensive use of GLA herbicides in
agriculture. The use of these GM crops also significantly changes the available residues and
requires more in depth study on the potential toxicity of the N-Ac-GLF metabolite.
2
Human Poisonings
Increased GLA herbicide use has been accompanied by higher incidence of accidental and
intentional human poisoning cases and as a result there is substantial information on the acute
human health effects of GLF. Three types of acute toxicity result from exposure to GLA
formulations in mammals. Immediate gastrointestinal (GI) symptoms include sore throat, oral
ulcers, nausea, vomiting, GI upset, abdominal pain, and diarrhea.
18
Cardiovascular (CV) effects
include altered vascular resistance and cardiac output.
18,19
GI and CV effects result primarily
from the anionic surfactant sodium polyoxyethylene alkyl ether sulfate used in the
formulation.
18–21
Moderate to severe neurotoxicity is defined by convulsions and memory
loss.
18,21–25
Severity of poisoning is correlated with dose, age and absence of concomitant
ethanol consumption.
18
Hyperammonemia in the blood has been reported after several severe
GLA poisonings but it is unclear whether elevated ammonia levels precede or follow
neurological symptoms.
18,25–27
The traditional OP target acetylcholinesterase (AChE) is not
inhibited in cases of severe neurotoxicity.
18,28
Magnetic resonance imaging of the brain after
the development of neurotoxicity indicates that the hippocampal region, thought to be highly
important in the processes of memory and learning, is disrupted selectively.
22
Mammalian Neurotoxicity Targets Implicated from Chemical Structure and Poisonings
GLF is the methyl phosphinate analog of glutamate (GLU) and the activity of GLF at GS results
from the high degree of structural similarity between GLF and the GS substrate GLU (Figure 1).
Candidate targets for GLF neurotoxicity are those GLU utilizing proteins which can be
modulated to produce the reported neurotoxic effects. In the mammalian body, L-GLU is an
essential component of proteins, but it is also the most prominent excitatory neurotransmitter
and an important precursor for N-Ac-L-GLU, an allosteric modulator of the urea cycle. The
convulsive and amnesic effects of GLF could result from effects on glutamatergic signaling or
the urea cycle. The implicated targets and pathways for development of convulsions are
outlined in Figure 2 which was compiled from extensive literature review for GLU function in
mammals.
Figure 1: Chemical structures of GLF and related compounds.
3
Figure
2
: GLF neurotoxicity targets
implicated by structural similarity to GLU and convulsive phenotype.
Legend:
(
+
) and (
-
) symbols indicate agonist and antagonist action at the indicated targets respectively.
↑ and ↓ represent
increase and decrease respectively. NT = neurotransmitter.
4
Direct activation of a GLU receptor (GluR), inhibition of GLU reuptake through an
excitatory amino acid transporter (EAAT) at the synaptic cleft or inhibition of GLU conversion to
glutamine by GS or to γ-aminobutyric acid (GABA) by GLU decarboxylase (GAD) could result in
overexcitation of neurons (and convulsions) leading to neuronal death (and amnesia). There
are numerous types and subtypes of GLU receptors which are broadly classified as ionotropic
GluRs (iGluRs) including the N-methyl-D-aspartate (NMDA) receptor (NMDAR), α-amino-3-
hydroxy-5-methyl-4-isoxazolepropionic acid receptor (AMPAR), and kainic acid receptor (KR) or
metabotropic GluRs (mGluRs).
29,30
Each iGluR type is further divided into subtypes composed
of varying tetrameric subunits.
29
The mGluRs are organized into groups based on sequence
homology and further distinguished as individual subtypes.
29
Although there are several types
and splice variants of EAATs,
31
EAAT2 is responsible for over 90% of GLU reuptake in the
brain.
32
Inhibition of the urea cycle by GLF would lead to elevated blood ammonia which can be
preferentially absorbed by the brain and cause convulsions.
33–35
Carbamyl phosphate synthase
I (CPS-I), the rate limiting enzyme in the urea cycle which converts ammonium, bicarbonate,
and 2 ATP to carbamyl phosphate requires N-Ac-L-GLU as an allosteric activator. N-Ac-L-GLU is
synthesized in liver mitochondria by N-Ac-L-GLU synthase (NAGS) from GLU and Ac-coenzyme A
(AcCoA). Inhibition of NAGS or CPS-I by GLF would down regulate the urea cycle and result in
ammonia accumulation in the body. For reasons that are unclear, the brain preferentially
accumulates excess ammonia from the blood. In the brain, ammonium changes the
electrochemical gradient of neuronal cells and leads to increased neuronal excitation and
convulsions.
There is some data in the literature on the activity of GLF at several of these structure
implied targets. GLF is an agonist at mGluR4 (EC
50
= 1.1 mM), mGluR6 (EC
50
> 0.5 mM), mGluR7
(EC
50
> 0.5 mM), and mGluR8 (EC
50
= 412 mM).
36
GLF inhibits GS isolated from spinach with a k
i
of 6.1 µM,
10
Escherichia coli with a K
1S
of 18 μm
37
and sheep brain with a K
i
of 28 µM.
38
GLF
inhibits purified rat brain GAD (k
i
= 2.2 mM; GLU K
m
= 1.3 mM).
39
GLF or the metabolite N-Ac-
GLF have not been tested on NMDAR, EAATs, NAGS or CPS-I.
Radioligand Binding and Bioassay in Target Site Analysis
The goal of this research was to elucidate the mechanism(s) of GLF neurotoxicity. The overall
approach utilized radioligand binding and various bioassays of previously untested targets
implicated by structure-activity. The commercially available radioligand [
3
H]CGP 39653 was
used to assess the affinity of GLF for the NMDAR. The activity of GLF at the NMDAR was tested
on the NR1/NR2A subtype NMDAR expressed in Xenopus oocytes. The inhibitory effect of GLF
on the high affinity GLU reuptake transporter GLT-I (the mouse homolog of human EAAT2) was
measured. The effect of GLF on signaling in primary rat cortical neurons was tested in multiwell
microelectrode array (MEA). The effect of GLF on CPS-I and the urea cycle was assessed.
Materials and Methods
Chemicals
[
3
H]CGP 39653 (NET1050, 47.3 Ci/mmol) and [
3
H]L-glutamic acid (NET490, 48.1 Ci/mmol) were
from Perkin NEN. D,L-GLA (93.7%) was from PESTANAL. Choline chloride (98%) was from TCI
America. N-Ac-D,L-GLF (98%) was from Crescent Chemical. D,L-GLF (98%), N-Ac-L-GLU (98%),
5
NMDA (98%), and WAY213613 (99%) were from Santa Cruz Biotech. CGP37849 (98%), L-GLU
(99%), monosodium L-GLU (99%), D,L-methionine-D,L-sulfoximine (MSO) (99%), and all other
chemicals were from Sigma Aldrich and of the highest purity available.
Buffers
Forebrain membrane homogenization medium contained 320 mM sucrose, pH 7.6. Binding
buffer contained 50 mM Tris HCl and 2.5 mM CaCl
2
, pH 7.4. For electrophysiology, barium
oocyte solution (BOS) contained 96 mM NaCl, 2 mM KCl, 5 mM HEPES, 2.8 mM BaCl
2
, pH 7.5
and ND96 recording solution contained 96mM NaCl, 2mM KCl, 1 mM MgCl
2
, 1 mM CaCl
2
and 5
mM HEPES pH adjusted to 7.5 with NaOH. For GLT-I assays, HEPES-buffered saline (HBS)
contained 5 mM Tris base, pH 7.4, 10 mM HEPES, 140 mM NaCl, 2.5mM KCl, 1.2 mM CaCl
2
, 1.2
mM MgCl
2
, 1.2 mM K
2
HPO
4
, and 10 mM glucose. Sodium free HBS (Na-free-HBS) was made
using equimolar replacement of NaCl with choline chloride. For primary neuron culture,
cortical media contained Dulbecco's Modified Eagle Medium (DMEM) with GlutaMax
TM
(Gibco
Cat#. 10569-010) supplemented with 10% heat inactivated horse serum (Gibco Cat# 26050-
088), 10 mM HEPES (Gibco Cat# 15630-080), 100 U/ml penicillin, and 0.1 mg/ml streptomycin
(added as 100x penicillin-streptomycin; [Gibco Cat# 15140-122]). NB/B27 media contained
neurobasal-A medium (Gibco Cat# 10888), supplemented with B-27 (Gibco Cat# 17504-004),
GlutaMax
TM
(Gibco Cat# 35050-061) and Pen-Strep (Lonza Cat# 17-602E), pH to 7.4. Cortical
buffer contained 137 mM NaCl, 5 mM KCl, 170 μM Na
2
HPO
4
, 205 μM KH
2
PO
4
, 5 mM glucose, 59
mM sucrose, 100 U/ml penicillin, and 0.1 mg/ml streptomycin, pH 7.4. Cytosine arabinoside
working solution (0.05 mM) was in Krebs ringer buffer (9 mM Tris, 10 mM glucose, 127 mM
NaCl, 5.5 mM KCl, 9 mM Na
2
HPO
4
, and 4 mM MgSO
4
pH 7.4). Liver homogenization buffer
(LHB) contained 50 mM triethanolamine HCl, 5 mM ATP (disodium salt), 15 mM magnesium
acetate (tetrahydrate), and 1 mM dithiothreitol pH 8.0.
Crude Forebrain Synaptic Plasma Membrane Preparation
Crude synaptic plasma membranes were prepared according to reported procedures
40–42
using
forebrains dissected from Pelfreeze untrimmed frozen whole mouse brains. All steps were at
4°C unless noted otherwise. Tissue was homogenized in 15 equivalents (wet tissue weight)
sucrose medium using Thomas 5098 dounce homogenizer with Teflon pestle (600 rpm, 12
strokes). Homogenate was centrifuged at 1,000 g for 10 min. Resulting supernatant was
transferred to new tubes and centrifuged at 17,000 g for 20 min. Pellet was resuspended in 40
equivalents deionized water, lysed with a polytron, incubated at 35°C for 35 min and
centrifuged at 48,000 g for 15 min. Pellets were washed twice by resuspension in 40
equivalents deionized water and centrifugation at 48,000 g for 15 min. Final pellets were
suspended in 3 equivalents deionized water, homogenized with a polytron and stored as
aliquots at -80°C. Immediately before binding experiments, aliquots were thawed and
centrifuged at 48,000 g. Resulting pellets were rinsed twice with 40 equivalents deionized
water and the final pellet was dissolved in 75 equivalents binding buffer and homogenized with
polytron.
[
3
H]CGP 39653 Binding Inhibition
Displacement binding of [
3
H]CGP 39653 was performed according to reported procedures.
42
Assay tubes in final 1 ml volume contained 2 nM [
3
H]CGP 39653, 25-40 μg protein (as
determined by Bradford
43
using bovine serum albumin standard) and varying concentrations of
inhibitor in duplicate. Each experiment included duplicate filter blanks (without protein),
6
control binding tubes (without inhibitor), and nonspecific binding samples (1 mM L-GLU as the
inhibitor). Assays were initiated by the addition of protein and allowed to equilibrate at 22°C
for 30 min. Assays were stopped by addition of 4.5 ml ice cold binding buffer and filtration
through Whatman GF-B glass microfiber filters using a Brandel model M-24 cell harvester.
Filters were rinsed twice with 4.5 ml ice cold binding buffer, dried briefly, transferred to 20 ml
plastic scintillation vials and 4 ml High Flash Point counting cocktail was added. After at least 18
h, samples were counted three times using a Beckman LS 6000 IC liquid scintillation counter
with auto DPM. Typical experiments yielded a maximum of 650 DPM, with 150 DPM filter blank
and 70-90% specific binding (SB, total binding – control binding) after filter blank subtraction.
Each data point was normalized for protein and then converted to % of control binding. Data
were represented as mean % control binding ± standard error for three biological replicates.
Data from three experiments were independently fit to the Hill function using nonlinear
regression analysis in Excel
44
and IC
50
values were determined from the resulting equations.
Unpaired one tailed Student’s T-tests were used to calculate p-values.
NMDAR Electrophysiology
Xenopous laevis oocytes were isolated and co-transfected with 0.5-1 ng of NR1 and NR2A RNA
constructs to express the NR1/NR2A heterotetramer as reported.
45,46
The two electrode
voltage-clamp (TEVC) experiments utilized the typical protocol.
47
The TEVC instrument
consisted of Dagan oocyte CA-1B amplifier, Axon Digidata 1440A digitizer and pCLAMP
software. Glass electrodes filled with 3 M KCl were pulled to a resistance of 0.2-1.0 MΩ.
Reference electrodes were connected via 1 M NaCl agar bridges with a platinum wire. Test
chemicals were dissolved directly in recording solution at 100 μM (NMDA), 10 μM (glycine, Gly)
and 100-10,000 μM (D,L-GLF) concentration. Recording solution was BOS (for transfected
oocytes) or ND96 (for untransfected oocytes). The BOS recording solution was Ca
2+
and Mg
2+
free to reduce the effect of endogenous Ca
2+
-dependent Cl
-
channels and voltage dependent
Mg
2+
block at the NMDAR respectively. Oocyte responses to perfused test solutions were
recorded with a clamped voltage of -80 mV. The clamped voltage was stepped in +10 mV
intervals for determination of reversal potential.
HEK293-GLT-I Cell Culture
HEK293 cells stably expressing the MAST-KREK isoform of mouse GLT-I
32
were generously
provided by Dr. Marcus Rattray, University of Reading UK. Cells were maintained in 95% O
2
/5%
CO
2
at 37°C and grown in DMEM supplemented with 10% fetal bovine serum, 10 units/ml
penicillin, 10 μg/ml streptomycin, and 400 μg/ml G-418 (Geneticin) as reported.
32
Cells were
seeded at 2 x 10
5
cells/well on poly-D-lysine coated (10 μg/ml) 24 well culture plates 24 h prior
to GLU uptake and cell viability assays.
GLT-I Mediated [
3
H]-GLU Uptake
Inhibition of [
3
H]GLU uptake through GLT-I by GLF and N-Ac-GLF was measured under
conditions that produce linear uptake as reported with minor modifications.
32
HEK293-GLT-I
cells in 24 well plates (see above) were washed twice (1 ml/ well/ wash) with Na-free HBS. HBS
(0.5 ml) containing GLF (0.1-1000 μM), N-Ac-GLF (0.1-1000 μM) or the high potency inhibitor
WAY 213613 (0.15 μM) or no inhibitor was added to triplicate wells. Each experiment included
triplicate controls for maximum uptake (no inhibitor) and triplicate controls for Na-independent
GLU uptake (Na-free-HBS). [
3
H]-L-GLU (48.1 Ci/ mmol, 1 mCi/ml, 28 μl) was added to 93 μl
water or 4.95 mM aqueous GLU (pH 7.4) and 5 μl of resulting solution was transferred to each
7
well resulting in 50 nM or 50 μM final GLU concentrations respectively. Plates were incubated
for 9 min at 25°C. Reactions were stopped by aspiration and rinsing two times with ice cold Na-
free HBS (1 ml/ well/ wash). Cells were visualized under the microscope to verify that most
cells were intact. Cells were then lysed with 0.1 M NaOH (0.5 ml/well). Cell lysates were
transferred to 20 ml disposable plastic scintillation vials and 10 ml Safety Solve scintillation
cocktail added. After at least 18 h, the amount of
3
H per vial was quantified with a Beckman LS
6000IC liquid scintillation counter using three measurements per sample and auto DPM
calculation. GLT-I specific uptake was calculated for each experiment by subtracting Na-
independent uptake from each well and data was expressed as % maximum uptake.
Culture of Rat Primary Cortical Neurons
Cultured rat primary cortical neurons were provided by Dr. Kathleen Wallace at the EPA and
were prepared according to the following unpublished protocol. Freshly isolated Newborn
Long Evans rat pup brains were put in a 60 mm dish containing 3-5 ml cortical buffer. Under a
dissecting scope the cortex was separated from other brain regions, striatum, hippocampus,
optic chasm, and meninges were removed in order. Cortex was transferred back to a 60 mm
dish of ice cold cortical buffer and remaining tissue was discarded. All but 0.2 ml cortical buffer
was removed by aspiration in a sterile hood and cortex was chopped into 1-2 mm
2
pieces with
scissors. Cortex pieces were transferred to a 25 cm
2
flask and cells were dissociated with 1 ml
trypsin (1 g/ 400 ml cortical buffer pH 7.6) per brain, trituration, and shaking in water bath (30
rpm, 37°C, 4.75 min). DNAse I (2 ml of 16 mg/ 100 ml in cortical buffer) was added and cells
were shaken again in water bath (30 rpm, 37°C, 4.75 min). Cortical media (30 ml, 37°C) was
added, cell suspension was transferred to 50 ml conical centrifuge tube, and centrifuged at 300
g for 5 min at 25°C. Supernatant was aspirated and pellet was resuspended in 1 ml DNAse (see
above) followed by gentle trituration with 10 ml cortical media after 1 min and centrifuged at
300 g for 5 min at 25°C. Supernatant was removed and pellet resuspended as before and
suspension was filtered through Nitex (100 μm pore, pre-wetted with 5 ml cortical media) into
a Nitex beaker and washed with 5 ml cortical media. Nitex was discarded and cell suspension
was transferred to a sterile vial. An aliquot of cells was counted (trypan blue, hemocytometer)
and 85% viability was used as the cutoff for continuing. Typical procedure yielded ~3 x 10
7
cells/ cortex. Cell suspension was diluted to 3 x 10
6
cells/ ml with cortical media. Cells were
plated at a density of 1.5 x 10
5
cells/ well on 48 well MEA plates (Axion: M768-KAP) with
laminin (1 nl/ well) coated electrodes at 37°C or at a density of 3.5 x 10
4
cells/ well on 96 well
sterile culture plates coated with poly-L-lysine (10 μg/ well, 3 h prior). Cultures were incubated
at 37
o
C under 5% CO
2
for 2 h to allow for attachment and media was then replaced with
NB/B27 (0.5 ml). After three to five days, cytosine arabinoside was added (5 μM final
concentration) and media was changed completely after 2 days to remove cytosine
arabinoside. Cultures were grown at 37
o
C under 5% CO
2
and media was changed every 7 days.
All steps were performed under sterile conditions and all buffers were sterilized through 20 μm
filters before use.
MEA Recordings
MEA recordings were made essentially as reported.
48
MEA instrumentation consisted of a
Maestro 768-channel amplifier, Middle-man data acquisition interface and Axion Integrated
Software all products of Axion Biosystems operated by a basic personal computer.
48
MEA
plates (48 well, Axion: M768-KAP) contained sixteen 40-50 μm individually nano-textured gold
8
microelectrodes/ well. Baseline spontaneous activity was recorded for 1 h at 13 or 15 cell
divisions. Wells were then treated (10 µl directly into media) with test chemicals (50x in water)
or vehicle control (at least 3/ plate). To assess the effect of MK801 and D,L-GLF, appropriate
wells were pre-treated with 1 μM MK801 for 10 min prior to treatment with D,L-GLF. Activity
was then recorded for 1 h in the presence of the chemicals or controls. An in house program in
the Shafer laboratory was used to calculate the weighted mean firing rate. The program utilizes
a background threshold of eight times the root mean square of the noise to characterize a spike
and a threshold of 5 spikes/ min to establish an active electrode. Only wells with at least 10
active electrodes in the baseline were used in the analysis. Spike rates of active electrodes post
treatment were expressed as a percentage of the spike rate for active electrodes in the baseline
of the same well (i.e. each well served as its own control). Data were also analyzed with
respect to changes in the total number of active electrodes/ well.
Viability of Rat Primary Cortical Neurons Treated with GLF and Related Compounds
Neurons plated on 96 well plates (see above) were treated with the MEA test compounds using
the same dosing as the MEA experiments. Cell viability was assessed using the Promega Cell
Titer-Blue
®
Cell Viability Assay for 1 h according to the manufacturer’s protocol. Each replicate
experiment included a positive (0.05% v/v triton X-100, triplicate) control, vehicle (water)
control and untreated (Neurosbasal A medium) control. Reagent blanks were subtracted from
all values and results were expressed as % viability observed in untreated wells.
Desalted Mouse Liver Homogenate
Desalted mouse whole liver homogenate was isolated as reported with minor modifications.
49
Pelfreeze frozen mouse livers were thawed, weighed, minced, and homogenized using a motor-
drive glass homogenizer with teflon pestle in 20 equivalents (w/v) LHB. Suspension was
centrifuged for 15 min at 37,000 g to remove particulate cell debris. Supernatant was
transferred to center of PD-10 desalting columns (GE Healthcare, spin prepped with LHB per
manufacturer’s protocol, 2 ml sample/ column). Columns were placed into 50 ml Corning
collecting tubes using spin inserts (GE Healthcare). Protein was eluted at 1,000 g for 2 min and
protein content of each sample was assessed by Bradford assay.
43
Aliquots of desalted
homogenate were used immediately or stored at -80°C until use.
CPS-I Activity
The activity of CPS-I was measured using an adaptation of the reported method
49
for reading in
96 well plate format. Glass screw cap reaction tubes (2.5 ml) contained 20 μl desalted
mitochondrial lysate (see above), and 180 μl LHB containing 1 umol N-Ac-L-GLU and 10 umol
NH
4
HC0
3
with and without inhibitor. Negative controls lacked N-Ac-L-GLU. Reactions were
initiated by addition of protein and shaken for 10 min at 37°C. Carbamyl phosphate was
converted to hydroxyurea (HU) by the addition of 2.0 M hydroxylamine and incubation for 10
min at 95°C. HU was quantitated by addition of 800 μl of chromogenic reagent ( equal parts 85
mg antipyrene/ 10 ml 40% v/v H
2
SO
4
in water and 62.5 mg diacetyl monoxime in 10 ml 5% (v/v)
acetic acid in water mixed immediately before use) to the capped reaction tubes and heating
for 20 min at 95°C, cooling to room temperature, transfer of 200 μl from each tube to 96 well
plate in triplicate and measuring the absorbance of the yellow chromophore at 458 nm using
the plate reader. The concentration of the HU was determined from a carbamyl phosphate
standard curve constructed in each plate.
9
Data Analysis
Statistical analyses and graphical representation of results were performed using Microsoft
Excel 2007 or 2010.
50
Unless noted otherwise, results were calculated and expressed as the
mean ± standard error of the mean.
Results
Effect on Urea Cycle Regulation (CPS-I Activity)
GLF (1, 10 mM) and N-Ac-GLF (1 mM) had no inhibitory effect on the activity of the urea cycle
enzyme CPS-I (Figure 3). Attempts to measure NAGS activity by pre-incubating the desalted
mouse liver homogenate with the NAGS cofactors (L-arginine, Ac-Co-A and L-GLU) and then
measuring CPS-I activity as above were unsuccessful. This procedure produced excessive color
(data not shown) likely due to the formation of other ureido compounds from the additional
cofactors rendering the results un-interpretable. No further attempts to measure inhibition of
NAGS activity were made although a procedure to measure NAGS activity is reported that
utilizes heavy isotope labeling and mass spectrometry.
33
Effects on NMDAR (Binding)
The inhibition of [
3
H]CGP 39653 binding to mouse forebrain crude synaptic membranes by L-
GLU (Figure 4) agrees with reported binding to rat brain
42
with an IC
50
value of 0.40 ± 0.13 μM.
D,L-GLF inhibited [
3
H]CGP 39653 binding at concentrations greater than 100 μM (Figure 4) with
an IC
50
value of 668 ± 236 μM. Therefore, D,L-GLF is about 1,000 times less potent than L-GLU
in displacing [
3
H]CGP 39653.
In agreement with binding to the rat brain,
42,51,52
NMDA at 100 µM reduced [
3
H]CGP
39653 specific binding to 13 ± 5% while 1 µM NMDA did not affect binding (Figure 5). The GS
inhibitor MSO (1 and 100 µM) did not decrease [
3
H]CGP 39653 binding. Surprisingly, 100 μM N-
Ac-D,L-GLF was slightly better at displacing [
3
H]CGP 39653 binding than both 100 μM D,L-GLF
(p=0.046) and 100 μM N-Ac-L-GLU (p=0.02) (Figure 5). [
3
H]CGP 39653 binding was not
significantly inhibited by 1 µM N-Ac-D,L-GLF or N-Ac-L-GLU (Figure 5).
Effect on NR1/NR2A subtype NMDAR (Electrophysiology)
GLF exhibited both an agonist and antagonist effect in NR1/NR2A heterotetramer transfected
Xenopus oocytes. D,L-GLF (10 mM) in the presence of the NMDAR coagonist Gly (10 µM)
induced a current in transfected oocytes with a magnitude equal to 23% of the current induced
by NMDA (100 µM) + Gly (10 µM) (Figure 6). This current had a reversal potential near 0 mV
(data not shown). D,L-GLF (10 mM) inhibited the positive cation current induced by the
NMDAR agonist NMDA (100 µM) + Gly (10 µM) (Figure 6). D,L-GLF concentrations of 3.3, 10
and 30 all induced progressively increasing currents in NR1/NR2A transfected oocytes (Figure
7). The concentration required to induce half the current of 100 μM NMDA + 10 μM Gly (EC
50
)
was about 22 mM (Figure 7).
10
Figure 3: CPS-I activity in the presence and absence of D,L-GLF or N-Ac-D,L-GLF
Legend: Hydroxyurea (HU) concentration is proportional to carbamyl phosphate concentration
0
2
4
6
8
10
12
14
0 1 10
CPS-I activity, nmol HU/ min / mg protein
[inhibitor], mM
GLF
NAcGLF
11
Figure 4: Inhibition of 2 nM [
3
H]CGP 39653 binding by L-GLU and D,L-GLF
Legend: Lines represent the fitted Hill functions for aggregate data. Numbers overlaid on the
graph are IC
50
values.
0
50
100
-10 -9 -8 -7 -6 -5 -4 -3 -2 -1
% of control binding
log[inhibitor], M
L-GLU
D,L-GLF
668 ± 236
µ
M
0.40 ± 0.13
µ
M
12
Figure 5: Inhibition of 2 nM [
3
H]CGP 39653 binding by 1 μM and 100 μM inhibitors
0
50
100
% of control binding
inhibitor
100 µM
1 µM
13
Figure 6: TEVC current traces for activity of D,L-GLF on NR1/NR2A transfected Xenopus
oocytes
Legend: Agonists at the indicated concentrations were co-perfused with 10 μM Gly.
14
Figure 7: Partial concentration response to D,L-GLF in NR1/NR2A subtype NMDAR
Legend: Values are normalized to the max current induced by 100 μM NMDA + 10 μM Gly in
NR1/NR2A transfected oocytes.
0
0.2
0.4
0.6
0.8
1
1 10 100
normalized max current
[D,L-GLF], mM
EC
50
= 20 mM
15
D,L-GLF activates a nonselective cation channel in control (untransfected) Xenopus
oocytes. D,L-GLF (10 mM) in the presence of the noncompetitive NMDAR channel blocker
antagonist MK801 (10 μM) induced a current in un-transfected control oocytes with a
magnitude of about 60% of the max current induced in NR1/NR2A transfected oocytes by the
same concentration of D,L-GLF (Figure 8). NMDA (100 μM) + Gly (10 μM) induced no current in
control oocytes (Figure 8).
Effects on Glutamate Reuptake in the Brain (GLT-I Assay)
The mouse high affinity sodium-dependent GLU reuptake transporter GLT-I was assayed for
inhibition by D,L-GLF and N-Ac-D,L-GLF. The positive control and known high affinity GLT-I
inhibitor WAY213613 (140 nM, reported IC
45
)
32
differentially inhibited 50 nM and 50 μM GLU
uptake by 61 ± 6 % and 81 ± 3 % respectively (p=0.028) (Figure 9). Concentrations of D,L-GLF
greater than 1 mM were needed to decrease GLU uptake through GLT-I (Figure 10). There was
no difference in D,L-GLF inhibition of GLT-I activity in the presence of 50 nM or 50 µM GLU
(Figure 10). While preincubating GLT-I expressing cells for 1 h with D,L-GLF did not affect the
potency of D,L-GLF in the uptake assays, 24 h preincubation resulted in increased inhibition
(Figure 10). N-Ac-D,L-GLF (up to 1 mM) did not inhibit GLU uptake through GLT-I (Figure 10).
Twenty four h exposure of GLT-I expressing HEK-293 cells to D,L-GLF (up to 10 mM) did not
induce cytotoxicity or change cell number as measured by LDH-release (Figure 11) and trypan
blue staining and counting (Figure 12) respectively.
Effects on Primary Cultured Rat Cortical Neurons (MEA Recordings)
D,L-GLF and N-Ac-D,L-GLF and the structurally or functionally related compounds MSO, NMDA,
N-Ac-GLU and CGP37849 were used to treat primary rat cortical neurons cultured on 48 well
MEA plates. In a preliminary analysis of the MEA data, the weighted mean firing rate was
calculated in the baseline and post treatment. When plotted as a function of concentration,
the weighted mean firing rate data yield concentration-response curves (Figure 13) with distinct
profiles.
In water treated control wells, MEA activity post-treatment did not change (92 ± 6 %,
n=56) relative to baseline. NMDA produced a biphasic response in the MEA system (Figure 13)
in agreement with published results.
53
NMDA (1, 3, and 10 µM) produced progressively
increasing spike rates with a maximum of 375 ± 95 % of baseline at 10 µM (Figure 13). Higher
concentrations of NMDA resulted in submaximal firing (30 µM = 180 ± 96 %) or completely
abolished activity (100, 300, and 1,000 µM). D,L-GLF and N-Ac-D,L-GLF were active in the MEA
system and showed a biphasic profile similar to NMDA (Figure 13). D,L-GLF (30, 100, 300, and
1,000 µM) and N-Ac-D,L-GLF (300 and 1,000 µM) increased the firing rate with maximum values
of 410 ± 139 % (1,000 µM) and 187 ± 29 % (300 µM) respectively. Higher concentrations of D,L-
GLF or N-Ac-D,L-GLF produced progressively decreasing activation (3,000 µM) or completely
abolished activity (10,000 µM). MSO produced slightly increased activity at 1,000 and 3,000 µM
and slightly decreased activity at 10,000 µM (Figure 13). N-Ac-L-GLU did not increase activity at
any test concentrations but decreased activity at the highest concentration (10,000 µM). The
competitive NMDAR antagonist CGP37849 (0.3 – 100 µM) decreased activity in the MEA system
(Figure 13). The minimum activity measured in the MEA system in the presence of CGP37849
was 14% and the inhibition curve fluctuated at 14-35% from 1 - 100 µM. The NMDAR channel
blocker antagonist MK801 (1 µM) decreased activity in the MEA system (Figure 14). While
16
Figure 8: TEVC current traces of D,L-GLF activity in untransfected (control) oocytes compared
to NR1/NR2A transfected oocytes
17
Figure 9: Inhibition of [
3
H]GLU uptake by WAY213613
0
10
20
30
40
50
60
70
80
90
140 nM WAY
213613
[3H]GLU uptake, % control
50 nM GLU
50 uM GLU
p=0.028
18
Figure 10: Inhibition of [
3
H]GLU uptake by D,L-GLF (top) or N-Ac-D,L-GLF (bottom)
0
20
40
60
80
100
120
140
0.001
0.01
0.1
1
10
100
1000
10000
[3H]GLU uptake, % control
[D,L-GLF], uM
50 nM GLU 50 uM GLU
50 nM GLU/ 1 h preinc 50 nM GLU/ 24 h preinc
0
20
40
60
80
100
120
140
0.1
1
10
100
1000
10000
[3H]GLU uptake, % control
[n-Ac-D,L-GLF], uM
50 nM GLU
19
Figure 11: LDH release in GLT-I expressing HEK-293 cells
0
20
40
60
80
100
10 mM
GLF
1 mM
GLF
0.1 mM
GLF
1 mM RIL Media total lysis
LDH release, % control
treatment
60 min 180 min 22 hr
20
Figure 12: Cell counting of GLT-I expressing HEK-293 cells treated with D,L-GLF
0E+0
1E+5
2E+5
3E+5
4E+5
5E+5
1000010001001010
cells/well
[D,L-GLF], μM
total
live
dead
21
Figure 13: Concentration response for D,L-GLF and related compounds in MEA assay
22
Figure 14: Effect of 1 μM MK801 pretreatment on NMDA and D,L-GLF concentration
responses in MEA
23
MK801 pretreatment eliminated the NMDA induced increase in MEA activity (Figure 14),
MK801 pretreatment reduced, but did not eliminate the D,L-GLF induced increase in MEA
activity (Figure 14). The number of replicates for MK801 pretreatment experiments was only
two, so the results should be interpreted with caution.
The % of baseline active electrodes was indicative of agonist and antagonist potency
(Figure 15). For the agonists NMDA, D,L-GLF and N-Ac-D,L-GLF the % of active electrodes
begins to decrease at the concentration of maximal firing rate increase (compare Figure 13 and
Figure 15). Although MSO increased the firing rate, it did not reduce the number of active
electrodes (compare Figure 13 and Figure 15). N-Ac-L-GLU drastically reduced the number of
active electrodes at the highest (10 mM) concentration (Figure 15), similar to the effect on
firing rate (Figure 13). CGP37849 reduced the number of active electrodes at concentrations
greater than 0.1 µM (Figure 15).
Cell Viability and pH in Primary Cortical Neuron Cultures
The decrease in MEA activity measured at the high concentrations tested could result from
networks shutting down or cells dying. The alamar blue assay
54,55
was used to test cell viability
of primary cultured rat cortical neurons for up to 1 h of treatment (Figure 16). Intermediate
concentrations of NMDA (30 - 1,000 µM) resulted in a decrease in cell viability (Figure 16). The
highest concentration of both N-Ac-D,L-GLF and N-Ac-L-GLU (10,000 µM) dramatically
decreased cell viability (Figure 16). All other test concentrations showed no change in cell
viability after 1 h of treatment (Figure 16).
Activity in the MEA system and viability of primary cultured neurons are sensitive to
changes in pH. In consideration of the high test concentrations, the pH of the solutions in each
well at the end of the MEA experiments was measured with pH paper. N-Ac-D,L-GLF and N-Ac-
L-GLU treatment wells had decreased in pH to 6-6.5. This pH change likely accounts for some of
the decreased cell viability and MEA activity noted for these treatments above. The pH of all
other treatment wells held at 7.4.
Discussion
Neurotoxicity Mechanistic Considerations
The primary phenotype that results from exposure to acutely toxic levels of GLF is convulsions
which may occur with elevated serum ammonia levels.
18,25
Although the herbicidal mode of
action of GLF is ammonia accumulation and amino acid dysregulation resulting from inhibition
of GS, this mechanism is not thought to be significant to mammalian toxicity.
28,56
The structural
similarity between GLF and the endogenous amino acid GLU and the possibility that these
targets could be modulated to yield a convulsive phenotype was used to generate a list of
potential GLF targets (Figure 2). There are two main GLU utilizing systems which might be
affected to yield a convulsive phenotype, the NAGS/CPS-I regulated urea cycle in the liver and
glutamatergic signaling in the brain.
Urea Cycle Effects
Inhibition
33,57
or mutation
58
of the urea cycle regulating enzymes NAGS or CPS-I is known to
produce hyperammonemia
35,59,60
and possibly convulsions as the brain preferentially absorbs
excess ammonia from the blood. D,L-GLF and N-Ac-D,L-GLF up to 10 and 1 mM respectively do
24
Figure 15: Change in MEA active electrodes after application of GLF and related compounds
25
Figure 16: Cell viability of primary rat cortical neurons treated with MEA test compounds
26
not effect CPS-I activity. While NAGS activity could not be measured using a NAGS/CPS-I
coupled procedure, it is unlikely that inhibition of NAGS could produce hyperammonemia in
patients on the reported time scale
25
as the throughput for the NAGS enzyme is low.
61
Considering the variation in reports and the lack of effect on CPS-I, hyperammonemia is likely
explained not by a toxic effect at an ammonia metabolizing enzyme, but as a secondary effect
of convulsions
62,63
and/or as a direct result of absorption of the ammonium ion contained in
GLA formulations.
Glutamatergic Signaling Effects
Excitotoxicity, the over-excitation of neurons that results from excessive activation of GluRs,
can result from an abnormal buildup of endogenous GLU at the synaptic cleft, or direct
activation of GluR by exogenous agonists. The human high affinity GLU reuptake transporter
EAAT2 can be potently blocked by methyl derivatives of GLU
64
making this transporter a viable
candidate for inhibition by GLF. However, very high and probably unattainable concentrations
of D,L-GLF (> 1 mM) are needed to affect transport through the mouse homolog of EAAT2, GLT-
I, and N-Ac-D,L-GLF does not affect GLT-I transporter activity (up to 1 mM). Therefore, as GLT-I
is responsible for >90% of GLU reuptake in the brain
32
it is unlikely that inhibition of EAAT2
(GLT-I) produces the primary GLF toxic insult. However, given the variation in activity between
EAAT1,2,3,4,
31
we cannot rule out the possibility that another subtype is modulated by
biologically relevant concentrations of GLF or N-Ac-GLF to produce a synapse specific or local
inhibitory effect in the brain.
A convulsive GLF phenotype mediated through GluR could result from direct action at
any of the receptor types (see ‘Mammalian Neurotoxicity Targets Implicated from Chemical
Structure and Poisonings’). NMDAR antagonists but not AMPAR/KR antagonists could delay the
onset of GLA induced convulsions in mice
65
and reverse GLA induced changes in striatal
dopamine release,
66
implicating the NMDAR as the target for GLF neurotoxicity. D,L-GLF and N-
Ac-D,L-GLF inhibited binding of the preferred NMDAR radioligand [
3
H]CGP39653 to mouse
forebrain membranes, suggesting direct action of GLF at the NMDAR. Furthermore, D,L-GLF
was able to agonize and antagonize NR1/NR2A subtype NMDAR expressed in Xenopus oocytes
although quantitation of the extent of activation is complicated by D,L-GLF activation of an
endogenous channel. We were limited to experiments with the NR1/NR2A subtype, but given
the pharmacological variation of the NMDAR subtypes, higher activity could be possible in
another subunit combination, making further study worthwhile.
Neuronal Signaling Effects
MEA recordings of cultured rat primary cortical neurons were used to assess the effects of GLF
and related compounds on neuronal signaling. Spike rate, the simplest analysis of the MEA
data, produced dose response curves with potencies that correspond well with binding of the
compounds in the series to the NMDAR. The binding affinity to the NMDAR (IC
50
) corresponds
to the concentration for maximal increase in MEA activity for the agonists NMDA, D,L-GLF and
N-Ac-D,L-GLF (Table 1). A biphasic activation-deactivation MEA profile is indicative of
excitotoxicity.
53
The high affinity (K
i
= 0.2 μM)
67
NMDAR antagonist CGP37849 only reduced
MEA activity further indicating that the biphasic excitotoxicity profile is specific to agonist
compounds. Taken together these data suggest that GLF and also N-Ac-GLF can produce
excitotoxicity through direct action at the NMDAR.
27
Table 1: Comparison of NMDAR Binding and MEA Activity Data
chemical
NMDAR
binding IC
50
,
μM
MEA activity
(mean firing rate)
conc. for
max increase
max increase
,
% action
NMDA 3.6
a
10 μM
320 agonist
N-Ac-D,L-GLF
100
b
300 μM
187 agonist
D,L-GLF 660 1,000 μM
324 agonist
N-Ac-L-GLU >100 none
103 none
MSO >1,000 3,000 μM
168 agonist
CGP37849 0.22
c
none
none antagonist
a
IC
50
reported previously
51
b
100 μM inhibited [
3
H]CGP39653 binding by 44%
c
literature K
i
value determined by inhibition of [
3
H]-CPP binding
67
28
N-Ac-L-GLU did not alter MEA activity up to 3 mM and although it did decrease activity
at 10 mM, the pH effect at this concentration likely accounts for the change in activity. The
inactivity of N-Ac-L-GLU in MEA is noteworthy because while the N-Ac-GLF used in this study
was racemic, only the L isomer of N-Ac-GLF is produced by the PAT enzyme in bacteria and GM
crops. Furthermore, the NMDAR is fantastically stereoselective for D over L isomers.
68
This
suggests that the activity of racemic N-Ac-GLF may come from the less biologically relevant D
isomer. The GLF used was also racemic, but as the herbicide is applied as a racemic mixture,
both isomers are relevant toxicologically. Further study is needed to assess the
stereospecificity of GLF and N-Ac-GLF neuroactivity.
GLF and MSO inhibit Escherichia coli (K
i
= 1.1 µM, K
i
= 2.0 µM respectively)
69
and sheep
brain (K
i
= 28 µM, K
i
= 210 µM respectively)
38
GS. MSO (1, 100 µM) did not inhibit
[
3
H]CGP39653 specific binding to mouse forebrain membranes. MSO (1 mM) inhibited
[
3
H]CGP39653 specific binding to rat cortical membranes by only 25%.
70
MSO (1, 3 mM)
produced slight activation in the MEA system with reduction of activity at higher (10 mM)
concentration. The activity profile indicates that MSO is excitotoxic, but MSO did not decrease
the number of active electrodes up to 10 mM. Excitotoxicity of MSO has been demonstrated
previously to result from increased GLU release and is not correlated with affinity for the
NMDAR.
70
The excitotoxic effect of NMDA and GLF is reduced by pretreatment in the MEA system
with the noncompetitive NMDAR channel blocker antagonist MK801. While MK801 (1 µM)
completely eliminates excitation by NMDA, D,L-GLF is still able to increase spike rate relative to
MK801 treatment alone. Although only 2 replicates have been completed, this experiment
suggests that D,L-GLF can increase activity in the MEA system through an NMDAR independent
mechanism. It is possible that the receptor activated by D,L-GLF in the presence of MK801 in
the MEA system is homologous to the nonselective cation channel activated by D,L-GLF in
untransfected Xenopous oocytes in the presence of MK801. Nevertheless, the reduction but
not elimination of D,L-GLF excitotoxicity by MK801 is consistent with the observations that
intraperitoneal pre-administration of MK801 increases the time to onset of, but does not
protect against GLF induced convulsions in mice.
65
Correlation of in Vitro Results to in Vivo Toxicity
Ammonia concentration and GS activity in the rat brain are not sufficiently altered after GLA
exposure to indicate GS inhibition as a primary mechanism of GLF neurotoxicity.
56
Astrocytes in
the brain utilize GS to detoxify ammonia and cycle GLU, but the urea cycle in the liver provides
the primary mechanism for ammonia detoxification in animals. Several studies have suggested
that the primary toxic insult from GLF exposure might occur outside of the brain as transport of
GLF across the blood-brain barrier would be difficult.
23,71
Although an indirect correlation
between nitric oxide production and GLF neurotoxicity has been reported,
71
no clear primary
mechanism originating outside of the brain has been identified. The human case reports of
hyperammonemia following GLA exposure,
25
prompted us to test for GLF inhibition of the N-Ac-
L-GLU utilizing CPS-I urea cycle regulating enzyme. GLF and N-Ac-GLF did not inhibit CPS-I
activity. Hyperammonemia has been suggested as a biomarker of neurotoxic GLF exposure.
25–27
Given the lack of activity at CPS-I and low inhibition of brain GS after in vivo exposure, it is likely
that hyperammonemia results from convulsions or direct absorption of the ammonium ion in
29
GLA formulations. If elevated ammonia results from convulsions (the more likely explanation
following oral exposure), it is of little utility as a therapeutic indicator for toxicity.
In one human poisoning case (267 mg/kg D,L-GLF exposure), the concentrations
respectively of D- and L-GLF one h after exposure were 1,050 μM and 1,070 μM in plasma and
after 27 h were 7.95 μM and 1.93 μM in plasma and 2.6 μM and 0.66 μM in cerebrospinal
fluid.
72
From the [
3
H]CGP 39653 binding inhibition data presented here, D,L-GLF concentrations
in excess of 100 μM are needed to affect the NMDAR. Therefore, D,L-GLF is only likely to
produce convulsions through direct activation of the NMDAR if it can concentrate at the
synaptic cleft. GLF distribution and localization in the brain could be aided by transport systems
which allow GLU and glutamine to cross the blood brain barrier readily.
73
Interestingly,
elevated serum ammonia levels have been demonstrated to alter blood-brain barrier
permeability and increase solute transport,
74
indicating a potential for increased transport of
GLF or N-Ac-GLF in the hyperammonemic state. Additionally, EAAT1 and EAAT3 (but not
EAAT2) GLU transport activity is increased in the presence of elevated ammonium,
75
indicating
further that the changes seen after GLF exposure could affect susceptibility to GLF. Ammonia
alone can induce changes to the brain through mechanisms that involve the NMDAR.
76
The
ability of biological systems to adapt to GLF is highlighted by the development of resistance in
Chlamydomonas reinhardtii through alteration in transport ability.
77
The specificity for
alteration in the hippocampal region of the brain after acute exposure to GLF is likely influenced
by the prevalence of high affinity NMDAR in this brain region.
78
The GLF metabolite N-Ac-GLF is largely considered to be safe. This study indicates that
N-Ac-GLF has a slightly higher affinity than GLF to the NMDAR and can significantly increase
neuronal firing in the brain suggesting that at high enough doses, both compounds could
contribute to the convulsive effect mediated through the NMDAR. GM crops expressing the
PAT enzyme generate N-Ac-GLF as the primary metabolite and micro flora in the gut are known
to produce N-Ac-GLF. In lieu of these two points, further study of N-Ac-GLF neurotoxicity is
warranted.
Although human GLF poisoning cases are primarily from acute exposure, the possibility
exists that GLF can produce chronic neurotoxic effects
79
or detrimentally impact the developing
brain.
80,81
NMDAR antagonism is associated with schizophrenia.
82,83
Excitotoxicity is of
considerable research interest as it has been linked to various neurodegenerative disorders
84,85
including Alzheimer’s disease,
86,87
multiple sclerosis,
88
autism,
89
amyotrophic lateral sclerosis,
90
and stroke.
91–93
Chronic exposure to GLF is reported to induce structural changes in the
NMDAR rich hippocampal region of the mouse brain.
79
A similar abnormality is documented
from acute human poisoning.
22
The importance of the NMDAR for learning and memory
suggests that significant exposure to GLF or N-Ac-GLF could disrupt the mechanisms for
information storage within the brain. This is supported by retrograde amnesia which follows
acute toxicity in humans
22
and spatial memory impairment measured in mice after chronic
exposure to as low as 2.5 mg/kg GLF.
79
The implications of small structural changes in the brain
are not well understood but may be indicators of some of the disorders listed above.
Potential for Exposure
The acute dose of GLF required to produce convulsions in mice and people (> 100 mg/kg) and
the lowest adverse effect dose for subacute exposure (2.5 mg/kg/day) are fairly high. The
proper use of GLF as a directed spray on the ground is reported to produce little to no drift
30
residues
94
but may result in significant dermal exposure to applicators
95
if personal protective
equipment is not worn. Despite the rise in GLA herbicide use, GLF residues have not been
detected in market basket surveys,
96,97
indicating that potential for adverse effects resulting
from exposure to residues on food is unlikely. Therefore, the main concern for GLF
neurotoxicity is after accidental and intentional exposures which are most likely to occur in the
workplace where GLF or GLA formulations are made or used in high amounts.
Summary
GLF produces neurotoxicity in mice and humans characterized by convulsions without AChE
inhibition, amnesia and structural changes to the hippocampal region of the brain. Despite
reports of hyperammonemia after high dose acute human exposure, brain GS and the urea
cycle regulator CPS-I in the liver do not have sufficiently altered activity by GLF to indicate these
proteins as primary targets for neurotoxicity. Furthermore, GLF and the metabolite N-Ac-GLF
GLF do not affect GLU reuptake through the prominent high affinity mouse brain GLU
transporter GLT-I. GLF and N-Ac-GLF can bind to the NMDAR (implicated previously as a
glutamatergic receptor involved in the convulsive phenotype) and produce an excitotoxic
profile in cultured primary rat cortical neurons. The NMDAR is expressed most highly in the
hippocampal region of the brain. These observations suggest that with high enough exposure
levels, GLF and N-Ac-GLF can both produce excitotoxicity through direct NMDAR activation in
the hippocampal region of the brain.
31
Chapter II: Progress in Glufosinate Radiosynthesis
Introduction
GLF Radioligand Utility
GLF is a major herbicide with mammalian neurotoxic effects (see Chapter I). Numerous
molecular targets for GLF neurotoxicity have been implicated by structure-activity relationships
and bioassay (see Chapter I). Previous approaches for identifying GLF targets relied on known
mechanisms of neuronal systems to identify and prioritize protein targets for testing. An
alternate approach to target screening utilizes radioligand binding to purified cellular and
subcellular fractions and identification of purified high affinity protein fractions. An essential
requirement of radioligand binding is the preparation of the high specific activity radioligand.
GLF Radiosynthetic Approaches
Numerous synthetic procedures are reported for the preparation of racemic, L-, or D-GLF.
98–108
However, these procedures are not readily adaptable to radiolabel incorporation. In a typical
radiolabelling procedure, a precursor to the radioligand of interest is synthesized. This
precursor is usually structurally similar to the ultimate radioligand but either incorporates an
alkene or is dealkylated allowing for one step radioisotope incorporation by tritiation (with
3
H
2
/Pd or similar) or incorporation of a small
3
H labeled alkyl respectively. Two procedures
suitable for the preparation of a high specific activity GLF radioligand were devised. The first
procedure (Figure 17), aimed at generating a protected GLF alkene, is a modification of the L-
amino-4-phosphono butyric acid (L-AP4) radiosynthesis.
109
The second procedure (Figure 18),
aimed at generating protected desmethyl GLF is a reordering of a reported method.
108
Materials and Methods
Chemicals
D,L-GLF (98%) was from Santa Cruz Biotech. Protected des-methyl L-GLF was a generous gift
from Prof. Wolfgang Maison, Universität Hamburg. All other chemicals were from Sigma
Aldrich and of the highest purity available.
NMR Spectroscopy for Structural Determination
Routine
1
H and
31
P NMR spectra for chemical structure determination of synthetic products
were recorded on an Avance 400 MHz spectrometer with Topspin 2.1 acquisition software and
analyzed using MestReNova 8.1 software.
Synthetic Procedures
Protected des-methyl L-GLF:
Methyl (2S)-2-(N-benzyloxycarbonyl)amino-4-[(hydroxy)- phosphinyl]butanoate was prepared
similarly to reported procedures.
36,108,110
A mixture of hypophosphorus acid (H
3
PO
2
, 47 mg,
1.14 mmol, 80% aqueous) was added to N-benzyloxycarbonyl-L-α-vinylGly methyl ester (Z-L-α-
vinylGlyOMe, 50 mg, 0.2 mmol) in a 5 ml round bottom flask. Methanol (1 ml) was added
followed by triethylborane (TEB, 0.2 ml, 1 M solution in n-hexane). Flask was covered with
parafilm and stirred open to air with a hole punched in the parafilm. After 18 h, TEB (0.05 ml, 1
M solution in hexane) was added to drive the reaction toward completion. After 4 h, the
32
Figure 17: Attempted [
3
H-2,3]GLF preparation method
Legend: 1: Oxalyl chloride, N
2
, CHCl
3
, 70% yield, 90% purity. 2. Ethyl magnesium bromide +
ethynyl TMS in ether + chloridate, N
2
, 0°C, 50% yield, 96% purity. 3. Ethyl
ethynyl(methyl)phosphinate + diethy acetamidomalonate + NaH in benzene, 0% yield. 4.
Unattempted due to lack of suitable precursor.
33
Figure 18: Attempted [
3
H-Me]GLF preparation method
Legend: 1: H
3
PO
2
, Z-L-α-vinylGlyOMe, methanol, triethylborane, 70-80% yield, 50-90% purity.
2: Protected desmethyl GLF, N
2
, Bis(trimethylsilyl)acetamide, CH
3
I, chloroform, 90% conversion,
50-80% purity. 3: 6N HCl, 100°C, 84% conversion, 73% purity.
34
methanol and volatile organic byproducts were removed in vacuo The crude residue was taken
up into 8 ml 5% aqueous KHSO
4
and extracted three times with 14 ml ethyl acetate. The
combined organic layers were dried over Na
2
SO
4
, gravity filtered through Whatman No 2 paper,
and concentrated in vacuo to yield 53.6 mg of sufficiently pure (70-90% purity by
31
P{
1
H} and
1
H
NMR; 70-80% crude yield) protected des-methyl L-GLF.
1
H NMR (400 MHz, CD
3
OD): δ 2.05 (m,
4H), 3.76 (s, 3H), 4.11 (m, 1H), 5.14 (s, 2H), 7.06 (d, JPH) 544 Hz, 1H), 7.38 (m, 5H).
31
P{
1
H} NMR
(400 MHz, CDCl
3
): δ 32.2 (s).
Protected L-GLF:
Protected des-methyl GLF (5 mg, .0158 mmol, 1 eq) was dissolved in dry chloroform (250 μl)
and added to a 5 ml round bottom flask which was then sealed with a Suba-Seal
®
red rubber
septa. The flask was flushed under vacuum and charged with nitrogen three times leaving the
flask under slight positive pressure. Bis(trimethylsilyl)acetamide (BSA; 38.73 μl, 0.158 mmol, 10
eq) was added using a dry syringe. After 10 min CH
3
I (2.94 μl, 0.0474 mmol, 3eq) was added
using a dry syringe. Chloroform (1 ml) was added through dry syringe and the reaction was
stirred at room temperature. After 20 h, the reaction was stirred open to air. After 1 h 2 N HCl
(5 ml) was added to quench any remaining silylating agent and the mixture was stirred
vigorously for 2 h. The layers were separated and the upper aqueous layer was extracted three
times with ethyl acetate (15 ml total) until the upper organic layer was colorless. Combined
organic layers were dried over Na
2
SO
4
and concentrated in vacuo to yield the crude product
(yellow residue, 5.6 mg, 70-80% purity by
31
P NMR, ).
31
P{
1
H} NMR (400 MHz, CDCl
3
): δ 58.1 (s).
L-GLF:
Protected-L-GLF (5.6 mg) was dissolved in 6 N HCl (1 ml) in a 2 ml vial with rubber lined screw
cap and placed in heat block at 100
o
C for 18 h. The solvent was removed in vacuo yielding the
crude product (4.2 mg, colorless oil). L-GLF was the major component (
1
H NMR) with 72%
phosphorus purity.
1
H NMR (400 MHz, D
2
O): δ 1.45 (d, JPCH 3H) 12 Hz, δ 1.83 (m, 2H), δ 2.17
(m, 2H), 4.08 (m, 1H).
31
P{
1
H} NMR (400 MHz, D
2
O): δ 50.4 (s). The 28% impurity (
31
P{
1
H} NMR
[400 MHz, D
2
O]: δ 24.2 [s]) is likely L-amino-4-phosphonobutyric acid (L-AP4).
Results
GLF Synthesis Amenable to Radiolabel Incorporation
A synthetic approach appropriate for the preparation of radiolabeled GLF in the alkyl chain
based on the reported radiolabeling procedure for L-AP4
109
was attempted (Figure 17). The
diethyl methyl phosphonate starting material was converted to ethyl
methylphosphonochloridate (
31
P NMR: δ 40.2) in good yield (70%) with excellent purity (>90%).
Ethyl ethynyl(methyl)phosphinate (
31
P NMR: δ 18.0) was then prepared in moderate yield (50%)
and excellent purity (>95%). After numerous attempts changing solvent, reaction time,
catalyst, and temperature (0-100°C), no conversion to ethyl
ethenyldiethylacetamidomalonate(methyl)phosphinate was detectable by
31
P NMR or
1
H NMR.
In an alternate approach, the procedure for the preparation of L-GLF reported by the
Montchamp laboratory
108,110
was modified. Montchamp’s method utilized an asymmetric
phosphinate synthesis adding the methyl group followed by the protected amino acid under
fairly mild conditions. In a reversal of this order (Figure 18) we prepared protected des-methyl
L-GLF from protected vinyl Gly and 80% hypophosphorus acid with 70-80% conversion and 50-
35
90% purity. Several attempts to prepare protected-L-GLF via Sila-Arbuzov coupling
conditions
110,111
(triethyamine, bromotrimethylsilane and CH
3
I) varying temperature and ratio
of reactants were unsuccessful. Monitoring the reaction by
31
P NMR it was apparent that the
trimethylsilylated P(III) intermediate was not being formed. We achieved success after using
bis(trimethylsilyl)acetamide to form the P(III) intermediate (58 ppm) followed by alkylation with
5 equivalents CH
3
I. The reaction yield (90%) and product purity (80%) at 5 mg starting material
scale were excellent. The reaction could be scaled to 1 mg starting material with loss of yield
(50-70%) and purity (~50%). Attempts to reduce the equivalents of CH
3
I needed or reduce the
scale further to 0.5 mg were unsuccessful. Protected-L-GLF could be readily converted to L-GLF
by refluxing in 6N HCl for 18 h and purification through Dowex resin column. Thus, progress
gives the intermediate, but CH
3
I ratio required was prohibitively high for use of [
3
H
3
]CH
3
I in
radiosynthesis.
Discussion
One of the most direct ways of identifying a specific target(s) for toxicity, and thereby
identify mechanisms, is direct binding assay using the high specific activity radiolabelled
chemical of interest.
112–114
There is no such GLF radioligand available. We developed a more
appropriate procedure for radiosynthesis of L-GLF which includes radiolabel incorporation as
the penultimate step. However, the requirement of at least three equivalents CH
3
I at the
labeling step still precluded the economical preparation of [
3
H-Me]L-GLF. The synthesis might
be made cost-effective by screening of additional catalysts which could increase reaction
efficiency for the second alkylation. Furthermore, combination of our two attempted
procedures using Grubb’s or similar catalyst in olefin metathesis
115–118
to join the alkynyl-
methyl-phosphinate (Figure 18) and the alkynyl Gly (Figure 17) might yield the L-isomer of the
alkene targeted in our first synthetic approach (Figure 17) as the precursor to [2,3-
3
H]L-GLF.
The limitations of this procedure would be the possibility of forming additional homocoupling
products. Despite significant effort and advancement in the chemistry related to GLF
preparation, an economical synthesis of a high affinity [
3
H]-GLF radioligand was not achieved.
36
Chapter III: Characterization of the Transient Oxaphosphetane BChE Inhibitor
Formed from Spontaneously-Activated Ethephon
Note: This chapter has been accepted for publication in Chemical Research in Toxicology as a
rapid report with the same title.
Introduction
Ethephon Inhibition of BChE
Ethephon (2-chloroethylphosphonic acid, A) is widely used as an ethylene-generating plant
growth regulator with an intriguing toxicological observation that it inhibits
butyrylcholinesterase (BChE) in vitro and in vivo in rats and humans.
119,120
Even more
surprising, ethephon is not direct-acting, but undergoes spontaneous activation in aqueous
alkaline solutions to the activated BChE inhibitor.
119
The ethephon-inhibited BChE is covalently
derivatized with a transfer of the phosphorous from ethephon.
120
The active site Ser198
contains a mass adduct of 108 Da thought to be the 2-hydroxyethylphosphonate (C) derivative
which forms from ring opening of 2-oxo-2-hydroxy-1,2-oxaphosphetane (the proposed
activated inhibitor, B).
119
Degradation of B in aqueous media in the absence of BChE by
electrophilic attack of the phosphorous by water or phosphate would produce the hydrolysis
product C or the phospholysis product 2-hydroxyethylphosphonyl phosphate (E). The reported
3-5 mol % trapping of B by BChE
119
suggests that the formation of B likely competes with the
direct formation of phosphate (D) and ethylene from A. All reactions are expected to be
irreversible.
Mechanism of Ethephon Degradation
The possibilities for degradation of A in aqueous solution with pH > 6 are depicted in Figure 19.
The presence of phosphorous in each chemical species permits
31
P NMR monitoring of the
concentrations of each product formed from ethephon degradation. Fitting of concentration-
time rate equations to kinetic data can be used to model proposed reaction mechanisms and
facilitates the calculation of individual reaction rates from experimental data for complex
mechanisms.
121
We used 600 MHz
31
P NMR to identify a transient species and two products in
addition to phosphate from the degradation of ethephon in aqueous solution. The chemical
shifts of the individual species correlate with the expected shifts for the structures proposed in
Figure 19. The rates for the degradation of A to B, C, D, and E were determined by
simultaneous nonlinear least squares fitting of the experimental data to the integrated rate
equations for each species in the proposed mechanism. The time course for formation and
degradation of the transient species correlates with BChE inhibitory potency of similar
incubated aqueous ethephon solutions.
Materials and Methods
Chemicals and Buffers
Ethephon (97%) was from ChemService. D
2
O was from Cambridge Isotope Labs. All other
components were from Sigma Aldrich. Potassium carbonate buffer (2 M) for
31
P NMR time
37
Figure 19: Possible degradation reactions of ethephon in aqueous solution at pH > 6
38
course was prepared in 90% H
2
O/ 10% D
2
O at pH 7.4 and for
1
H NMR spectra, 99% D
2
O was
used.
Buffer Comparison
Ethephon (10 mM) was dissolved in 100 mM potassium carbonate or potassium phosphate
buffers. Aliquots were diluted to 0.11 mM and assayed for BChE inhibitory potency
immediately after preparation and after 20.25 h. BChE activity was measured as reported
previously.
119
31
P NMR Kinetic Time Course
A 1 ml solution of ethephon (200 mM in 2 M potassium carbonate buffer) was added to an
appropriate NMR tube. The sample was kept capped at 25°C and protected from light for the
duration of the experiment. Degradation reactions were monitored by 600 MHz
31
P NMR on a
Bruker Avance 600 console with Bruker 14 T magnet. Quantitative measurements were made
using 32 scans and a 10 s pulse delay. It was determined that these conditions optimized the
spectral resolution disrupted by the Nuclear Overhauser effect and the production of gaseous
ethylene and CO
2
in solution. Spectra were acquired with Topspin 2.1 and processed
(automatic Bernstein polynomial baseline correction, followed by manual phase correction,
Savitsky-Golay and Whittaker smoothing, manual integration of peak areas, and normalization
of total integrated peak area in each spectra to 100%) using MestreNova 7.1.1. The % peak
area was considered a measurement of relative concentration. The magnitudes of spectra in
Figure 21 were normalized based on the area of the constant peak at 22.11 ppm. Chemical
shifts are reported relative to the designated phosphate peak.
Derivation of Integrated Rate Equations
The degradation of ethephon in aqueous buffers to produce the species in Figure 19 could
proceed through any of five mechanisms:
where A is ethephon, B is the activated BChE inhibitor, C is the hydrolysis product of the BChE
inhibitor, D is phosphate and ethylene, and E is the product of B with phosphate. Mechanism 1
consists of degradation of ethephon A to phosphate and ethylene D. Mechanism 2 consists of
degradation of A to D in competition with degradation to the transient inhibitor B which further
degrades to the hydrolysis product C. Mechanism 3 incorporates the possibility that B could
degrade to D directly by an intramolecular Wittig reaction. Mechanism 4 further incorporates
the reaction of B with D to generate a di-phosphate. Mechanism 5 is the special case of
Mechanism 4 where no B is converted to D. All of the steps are expected to be irreversible
based on the proposed structures.
Mechanism 4 is the most complicated. Mechanisms 1-3 and 5 are special cases of
Mechanism 4 where the omitted rate constants are equal to zero. Thus, a derived set of rate
39
equations for Mechanism 4 permits fitting of all of the five mechanisms to experimental data.
The differential rate equations for Mechanism 4 are as follows:
ௗ
ௗ௧
= −݇
ଵ
ሾܣሿ−݇
ଷ
ሾܣሿ= −(݇
ଵ
+݇
ଷ
)ሾܣሿ (1)
ௗ
ௗ௧
= ݇
ଵ
ሾܣሿ−݇
ଶ
ሾܤሿ−݇
ସ
ሾܤሿ−݇
ହ
ሾܤሿሾܦሿ (2)
ௗ
ௗ௧
= ݇
ଶ
ሾܤሿ (3)
ௗ
ௗ௧
= ݇
ଷ
ሾܣሿ+݇
ସ
ሾܤሿ−݇
ହ
ሾܤሿሾܦሿ (4)
ௗா
ௗ௧
= ݇
ହ
ሾܤሿሾܦሿ (5)
The differential equation
ௗ
ௗ௧
has the form of a simple first order differential rate equation with
ௗ
ௗ௧
= ݇
ሾܣሿ where ݇
= −(݇
ଵ
+݇
ଷ
) :
and therefore the integrated rate equation for the concentration of A from time 0 to t is:
ሾܣሿ = ሾܣሿ
݁
ି(
ೌ
)௧
(6)
The
31
P NMR data indicate that at every observed time point in the degradation, the phosphate
concentration is at least ten times the concentration of any other species which changes in
concentration. Therefore, the second order reaction of B with D can be considered pseudo first
order where:
݇
ହ
ሾܤሿሾܦሿ≈ k
ହ
’ሾܤሿ (7)
Substituting equation (7) into equations (2), (4) and (5) and combining the individual rate
constants into overall rate constants for each chemical species results in the simplified
differential rate equations:
ௗ
ௗ௧
= ݇
ଵ
ሾܣሿ−(݇
ଶ
+݇
ସ
+k
ହ
’)ሾܤሿ (8)
ௗ
ௗ௧
= ݇
ଷ
ሾܣሿ+(݇
ସ
−k
ହ
’)ሾܤሿ (9)
ௗா
ௗ௧
= k
ହ
’ሾܤሿ (10)
For ݇
= ݇
ଶ
+݇
ସ
+k
ହ
’ and݇
= ݇
ଵ
, the simplified general form of
ௗ
ௗ௧
is:
ௗ
ௗ௧
= ݇
ሾܣሿ−݇
ሾܤሿ (11)
Substituting equation (6) into equation (11) yields:
ௗ
ௗ௧
= ݇
ሾܣሿ
݁
ି(
ೌ
)௧
−݇
ሾܤሿ (12)
40
which rearranges to
ௗ
ௗ௧
+݇
ሾܤሿ= ݇
ሾܣሿ
݁
ି(
ೌ
)௧
(13)
Assuming that [B] has the form:
ሾܤሿ
௧
= ݕ(ݐ)݁
ି
௧
(14)
differentiating
ௗሾሿ
ௗ௧
= −ݕ(ݐ)(݇
݁
ି
௧
)+݁
ି
௧
ቀ
ௗ௬
ௗ௧
ቁ (15)
substituting equation (14) into equation (15) and rearranging
ௗሾሿ
ௗ௧
+݇
ሾܤሿ
௧
= ݁
ି
௧
ቀ
ௗ௬
ௗ௧
ቁ (16)
equating the right side of equation (16) with the right side of equation (13)
݁
ି
௧
ቀ
ௗ௬
ௗ௧
ቁ = ݇
ሾܣሿ
݁
ି(
ೌ
)௧
(17)
rearranging
ቀ
ௗ௬
ௗ௧
ቁ = ݇
ሾܣሿ
݁
ି(
ೌ
)௧
(݁
ା
௧
)= ݇
ሾܣሿ
݁
(
ି
ೌ
)௧
(18)
and integrating
ݕ(ݐ)=
್
ሾሿ
బ
ି
ೌ
݁
(
ି
ೌ
)௧
+ܥ (19)
At t=0, y=0 so: ܥ = −
್
ሾሿ
బ
ି
ೌ
and therefore:
ݕ(ݐ)=
್
ሾሿ
బ
ି
ೌ
൫݁
(
ି
ೌ
)௧
−1൯ (20)
Substituting (14) into (20) and rearranging
ሾܤሿ =
್
ሾሿ
బ
ି
ೌ
൫݁
(
ି
ೌ
)௧
−1൯݁
ି
௧
(21)
or more simply
ሾܤሿ =
್
ሾሿ
బ
ି
ೌ
(݁
ି
ೌ
௧
−݁
ି
௧
) (22)
Substituting (22) into (3) yields:
ௗ
ௗ௧
= ݇
ଶ
್
ሾሿ
బ
ି
ೌ
(݁
ି
ೌ
௧
−݁
ି
௧
) (23)
which integrates to:
ሾܥሿ=
మ
್
ሾሿ
బ
ି
ೌ
ቀ
షೖ
−
షೖೌ
ೌ
−
ଵ
+
ଵ
ೌ
ቁ (24)
41
Substituting (22) and (6) into (9) yields:
ௗ
ௗ௧
= ݇
ଷ
ሾܣሿ
݁
ି
ೌ
௧
+(݇
ସ
−݇
ହ
′)
್
ሾሿ
బ
ି
ೌ
(݁
ି
ೌ
௧
−݁
ି
௧
) (25)
which integrates to:
ሾܦሿ=
య
ሾሿ
బ
ೌ
(1−݁
ି
ೌ
௧
)+
(
ర
ି
ఱ
ᇱ)
್
ሾሿ
బ
ି
ೌ
ቀ
షೖ
−
షೖೌ
ೌ
−
ଵ
+
ଵ
ೌ
ቁ (26)
Substituting (22) into (10) yields:
ௗா
ௗ௧
= ݇
ହ
′
್
ሾሿ
బ
ି
ೌ
(݁
ି
ೌ
௧
−݁
ି
௧
) (27)
which integrates to:
ሾܧሿ=
ఱ
ᇱ
್
ሾሿ
బ
ି
ೌ
ቀ
షೖ
−
షೖೌ
ೌ
−
ଵ
+
ଵ
ೌ
ቁ (28)
Therefore the integrated rate equations for Mechanism 4 expressed in terms of the individual
rate constants are:
ሾܣሿ = ሾܣሿ
݁
ି(
భ
ା
య
)௧
(29)
ሾܤሿ=
భ
ሾሿ
బ
మ
ା
ర
ା
ఱ
ᇱି
భ
ି
య
൫݁
ି(
భ
ା
య
)௧
−݁
ି(
మ
ା
ర
ା
ఱ
ᇱ)௧
൯ (30)
ሾܥሿ=
మ
భ
ሾሿ
బ
మ
ା
ర
ା
ఱ
ᇱି
భ
ି
య
ቀ
ష(ೖమశೖరశೖఱᇲ)
మ
ା
ర
ା
ఱ
ᇱ
−
ష(ೖభశೖయ)
భ
ା
య
−
ଵ
మ
ା
ర
ା
ఱ
ᇱ
+
ଵ
భ
ା
య
ቁ (31)
ሾܦሿ=
య
ሾሿ
బ
భ
ା
య
൫1− ݁
ି(
భ
ା
య
)௧
൯+
(
ర
ି
ఱ
ᇱ)
భ
ሾሿ
బ
మ
ା
ర
ା
ఱ
ᇱି
భ
ି
య
൬
ష(ೖమశೖరశೖఱᇲ)
మ
ା
ర
ା
ఱ
ᇱ
−
ష(ೖభశೖయ)
భ
ା
య
−
ଵ
మ
ା
ర
ା
ఱ
ᇱ
+
ଵ
భ
ା
య
൰
(32)
ሾܧሿ=
ఱ
ᇱ
భ
ሾሿ
బ
మ
ା
ర
ା
ఱ
ᇱି
భ
ି
య
ቀ
ష(ೖమశೖరశೖఱᇲ)
మ
ା
ర
ା
ఱ
ᇱ
−
ష(ೖభశೖయ)
భ
ା
య
−
ଵ
మ
ା
ర
ା
ఱ
ᇱ
+
ଵ
భ
ା
య
ቁ (33)
Integrated Rate Equations Fitting and Determination of Parameter Confidence Intervals
Peak integrations were manually entered into Microsoft Excel
50
and analyzed by nonlinear
regression using the Solver feature as reported.
44
Equations 29-33 were fit to the assigned
peak data sets simultaneously by maximizing the sum of the R
2
values and varying the
parameters k
1
, k
2
, k
3
, k
4
, and k
5
’ with initial values of 0.00085, 0.048, 0.029, 0.004 and 0.007
respectively with an initial concentration of ethephon ([A]
0
) set to 97.5. The goodness of fit was
assessed based on the closeness of the R
2
value to 1. The reported rate constants were
determined by Monte Carlo simulation.
122
Briefly, the individual RMSE for each chemical
species was used to generate 40 virtual data sets per chemical species. Initial parameter values
for k
1
-k
5
’ of ± 25% of the fitted value were generated using the RANDBETWEEN function in
Excel. The reported average values are the results obtained after maximizing the sum of R
2
for
the virtual data sets to fitted values by modifying k
1
-k
5
’ for each set of virtual data by
implementing Solver using 100 iterations, automatic scaling and ‘assume non-negative.’ The
corresponding first order or pseudo first order T
1/2
values were calculated from T
1/2
= ln(2)/k for
each value of k. The reported confidence intervals are 95%.
42
Determination of T
1/2
Values for all Species
The first order T
1/2
value for the degradation of A was calculated using the equation T
1/2
=
ln(2)/(݇
ଵ
+݇
ଷ
). The T
1/2
values for formation of C, D and E were calculated by solving the
corresponding integrated rate equation using the calculated k values for the time at half
maximal concentration. These overall T
1/2
values are given in Table 2.
Determination of Total % Formation for B-E
The maximum concentrations of C (C
max
), D (D
max
), and E (E
max
) were determined by solving the
appropriate rate equation at t=10,000 h. The total % formation of C and E are equivalent to the
maximum value. The total formation of D (D
total
) was calculated as the sum of D
max
and E
max
.
The maximum concentration of D formed from A (D
maxA
) was calculated by solving the first term
of equation (32) at time t=10,000 h. The maximum formation of D from B (D
maxB
) was
calculated by subtracting D
maxA
from D
total
. The total % formation of B (B
tot
) was calculated as
the sum of C
max
, D
maxB
, and E
max
. The total % formation for each species is given in Table 2.
Ethephon Standard Spectrum
2-chloroethylphosphonic acid (ethephon):
1
H NMR (600 MHz, D
2
O) δ 3.68 (dt, J = 13.5, 7.6 Hz,
2H), 1.96 (dt, J = 18.0, 7.6 Hz, 2H).
31
P{
1
H} NMR (600 MHz, potassium carbonate buffer in
H
2
O:D
2
O 90:10, pH=7.4) δ 16.65 ppm.
31
P NMR (600 MHz, 99% D
2
O, pH = 7.4) δ 16.26 (tt, J =
18.0, 13.5 Hz).
Preparation of 2-Hydroxyethylphosphonic acid
Dimethyl 2-hydroxyethylphosphonic acid was refluxed in 6N HCl. Conversion was monitored by
31
P NMR and reached 75% after 120 h.
1
H NMR (400 MHz, D
2
O, pH<4) δ 3.48 (dt, J = 11.8, 7.5
Hz, 3H), 1.71 (dt, J = 18.2, 7.6 Hz, 3H).
31
P{
1
H} NMR (400 MHz, D
2
O, pH<4) δ 25.3 (s).
Monomethyl 2-Hydroxyethylphosphonic acid
was visualized by 31P NMR as an intermediate in the hydrolysis of dimethyl 2-
hydroxyethylphosphonic acid to 2-hydroxyethylphosphonic acid.
31
P{
1
H} NMR (400 MHz, D2O,
pH<4) δ 26.9 (s).
2-Oxo-2-hydroxy-1,2-oxaphosphetane
was visualized in the degradation time course spectra of ethephon.
1
H NMR (400 MHz,
potassium carbonate buffer in 99% D
2
O, pH = 7.4) δ 3.97 (ddt, J = 8.0, 8.0, 7.6 Hz, 2H), 2.85 (dt, J
= 18.0, 7.6 Hz, 2H).
31
P{
1
H} NMR (600 MHz, potassium carbonate buffer in H
2
O:D
2
O 90:10, pH =
7.4) δ 18.1 ppm.
31
P NMR (600 MHz, 99% D
2
O, pH = 7.4) δ 18.1 (tt, J = 18.0, 8.0, 8.0 Hz). The J-
couplings reported here should be taken with caution as many of the peaks were barely greater
than the spectral noise.
Results
Buffer System for
31
P NMR Time Course Experiments
Previous studies of ethephon activation as a BChE inhibitor utilized 10x phosphate buffer to
maintain pH throughout the experiments. This is not an ideal buffering system for
31
P NMR
measurements as the large phosphorous peak could mask other peaks of interest. We tested
inhibition of BChE by ethephon pre-incubated in 10x phosphate and carbonate buffers in order
to establish another buffering system for
31
P NMR experiments. Pre-incubation of high
concentration ethephon in phosphate and carbonate buffers have nearly the same time course
of activation to a BChE inhibitor (Figure 20). This establishes that higher concentrations of
43
Figure 20: Comparison of BChE inhibition by ethephon pre-incubated in carbonate or
phosphate buffers for 0 or 20 h
Legend: Potassium carbonate buffer (0.1 M, pH 9) results are depicted as
■
and potassium
phosphate buffer (0.1 M, pH 9) results are depicted as
♦
.
44
ethephon than previously reported in 10x carbonate buffer can be converted to the activated
BChE inhibitor in a similar manner to lower concentrations of ethephon in phosphate buffer.
Characterization of Ethephon Degradation
Conversion of ethephon to the activated BChE inhibitor is greatest at pH 7-9.
119
For time course
monitoring by
31
P NMR, pH 7.4 was chosen in order to permit formation of the activated
inhibitor at an optimal rate for measurement as a < 1% phosphorous component. The time
course spectra (Figure 21) show degradation of ethephon (16.65 ppm, s) and formation of
phosphate (2.49 ppm, s). Minor impurities not changing in concentration are evident at 10.94
ppm (s), 22.11 ppm (s) and 24.18 ppm (s). One peak (28.11 ppm (s)) increases and then
decreases in concentration and three peaks (-6.29 (d), 13.51 (d) and 18.76 (s) ppm) increase in
concentration over the time course (Figure 21).
31
P NMR chemical shifts of phosphorous acids
are sensitive to pH and standards are often reported in highly acidic solutions complicating
peak identification in our spectra. However, a change from pH < 4 to pH > 7 induces an ~8 ppm
upfield shift of phosphonate peaks in
31
P NMR as illustrated by ethephon in acidic D
2
O at 25.19
ppm
123
and in pH 7.4 carbonate buffer at 16.65 ppm. C has a
31
P NMR chemical shift of 25.30
ppm in acidic D
2
O. The singlet peak at 18.76 ppm is thus in the expected region for C at pH 7.4
and is assigned as such. The doublets at -6.29 and 13.51 ppm have J couplings of 24 Hz and
integrate nearly equally at all time points. The spectra are
1
H decoupled and peak splitting
results from P-P spin spin coupling only. J=24 Hz matches two bond P-P coupling of a
diphosphate.
124
Two doublets indicate unequal phosphorous atoms corresponding to a
phosphate (-6.29 ppm) and a phosphonate (13.51 ppm). We therefore assign these peaks as
the phosphorous atoms in E.
The transient
31
P NMR peak at 28.11 ppm (designated B) is of particular interest. The
atom connectivity to phosphorous in B is similar to monoalkyl-C. We have observed
monomethyl-2-hydroxyethylphosphonic acid (26.9 ppm) as an intermediate in the hydrolysis of
dimethyl 2-hydroxyethylphosphonic acid (35.87 ppm) to C (25.3 ppm) in refluxing HCl. Thus
oxaphosphetane B is expected to have a chemical shift downfield of the 18.76 ppm C signal.
Although numerous mechanistic
31
P NMR studies have proposed an oxaphosphetane
intermediate in the Wittig reaction, to our knowledge, only one of these studies
125
examined
the phosphonate modification of the Wittig reaction where the oxaphosphetane intermediates
would be most similar to B.
31
P NMR chemical shifts of 35 and 32 ppm were reported for these
dialkyl oxaphosphetanes.
125
Thus, a chemical shift at 28.11 ppm, slightly upfield of similar
dialkyl oxaphosphetanes, is as expected for 2-oxo-2-hydroxy-1,2-oxaphosphetane (B).
The
1
H NMR spectra of ethephon incubated in pH 7.4 potassium carbonate buffer at 1 -
96 h are shown in Figure 22. The
1
H NMR spectra of B is expected to be shifted downfield of
the
1
H NMR spectra of C. The transient doublet of triplets at 2.85 ppm corresponds to the alkyl
alpha hydrogens of B and is de-shielded compared to the beta hydrogens of C (1.71 ppm) and A
(1.96 ppm) consistent with the formation of an oxaphosphetane ring. The H-C-C-H coupling
constants are nearly the same for B (7.6 Hz) and C (7.6 Hz) while the H-C-P coupling constant is
slightly smaller for B (18.0 Hz) than C (18.2 Hz). The peaks for the alkyl beta hydrogens near 3.5
ppm overlap too much to allow definitive characterization, but it appears that the peaks slits
into a ddt with two bond splitting to the alpha hydrogens and two sets of three bond splitting
to the phosphorus both directions through the oxaphosphetane ring.
45
Figure 21:
31
P NMR time course spectra for 200 mM ethephon degradation in 2 M pH 7.4
potassium carbonate buffer
Legend: Letters correspond to the chemical structures in Figure 19.
46
Figure 22:
1
H NMR spectra of ethephon degradation in pH 7.4 potassium carbonate buffer at
1, 21 and 96 h
Legend: Insets are expansions of the indicated spectral regions. Letters correspond to the
chemical structures in Figure 19.
47
In
1
H coupled
31
P NMR the 28.11 ppm peaks in the
21 h
1
H coupled
31
P NMR spectra
support the J-coupling calculations (see values in methods) although the magnitude of the
peaks is barely above the background requiring caution in interpretation (Figure 23). Also in
agreement with the assignments from
1
H uncoupled NMR are the
1
H splitting of the C and E
phosphonate peaks but not the E phosphate peak (Figure 23).
Kinetic Modeling of
31
P NMR Data
The reactions in Figure 19 were mathematically modeled using irreversible first order and
pseudo first order kinetics for k
1
≠ k
2
≠ k
3
≠ k
4
≠ k
5
’ and B
0
= C
0
= D
0
= E
0
= 0 and the integrated
concentration-time rate equations for A-E were derived (see Methods). Nonlinear regression
analysis in Excel was used to fit these equations to the experimental peak integration data using
a simultaneous fitting extension of the reported method.
44
Rate equations 29-33 were fit to
the peaks with corresponding peak letter assignments in Figure 19. The results of the fitting are
shown in Table 2 and the graphical fits of the rate equations to the experimental data is
depicted in Figure 24. The calculated rate constants and confidence intervals from Monte Carlo
simulation as well as the corresponding first order T
1/2
values are given in Table 3.
Discussion
Goodness of Fit
The proposed degradation reactions of ethephon can be modeled by irreversible first order and
pseudo first order kinetic equations and fit simultaneously to the assigned experimental data.
The calculated rate constants (Table 3) are representative of the relative rates of conversion
between each chemical species under the experimental conditions. The overall fit of the
31
P
NMR data to the integrated rate equations is good (average R
2
= 0.887). The R
2
values for C
(0.771) and E (0.783) are lower than those for A (0.985), B (0.917) and D (0.979). This likely
results in part from error due to the pseudo-first order approximation (݇
ହ
ሾܦሿ≈ k
ହ
’) used in
deriving the integrated rate equations. While the concentration of D is always at least 10 times
B, it varies from 0-93 mol % and thus k
5
’ is an approximation of the average value over time. At
0 h the relative conversion of B to C or E is expected to shift toward formation of C and as the
concentration of D increases over time, the conversion of B to E would increase. Furthermore,
in phosphate buffer, a higher percentage conversion to E is expected.
Ethephon Degradation Rate
Ethephon degrades to ethylene and a BChE inhibitor important in plants and mammals
respectively. The T
1/2
degradation rate of A (32.7 h) and formation rate of D (33.0 h) are in
agreement with reported rates (11-58 h) summarized previously.
119
Rates derived from
measurements of ethephon degradation,
123
phosphate formation, or ethylene generation
126,127
are all reported as a measure of the rate of ethephon hydrolysis. This study supports that
approximation as the measured overall hydrolysis rate of A does not differ from the formation
rate of D. The measured rate for formation of D from A (37 h) is ten times faster than the rate
for formation of B from A (297 h) indicating that the overall rate of ethephon hydrolysis is
governed by the rate of phosphate formation (k
3
). The rate of ethylene generation would also
approximately equal the measured rate of phosphate generation.
48
Figure 23:
1
H coupled
31
P NMR spectra of ethephon degradation in pH 7.4 potassium
carbonate buffer at 21 and 96 h
Legend: Insets are expansions of the indicated spectral regions. Letters correspond to the
chemical structures in Figure 19.
49
Figure 24: Graphical fitting results of concentration-time integrated rate equations
Legend: Points represent concentrations calculated as area under the curve of spectra in Figure
21. Lines represents fitting of equations 29-33 to experimental data. Letters correspond to
chemical structures in Figure 19.
1
10
100
0 30 60 90 120 150 180
31P NMR area, %
time, h
0.01
0.1
1
10
0 30 60 90 120 150 180
31P NMR area, %
time, h
D
A
C
E
B
50
Table 2: Fitting Results of
31
P NMR Data to Concentration-Time
Rate Equations
equation R
2
95%
CI
T
1/2
, h type
max, %
[A] 0.985
9.63
32.7 degradation
97.5
[B] 0.917
0.19
32.9 formation
10.7
3.02 degradation
[C] 0.769
1.02
45.6 formation
3.19
[D] 0.979
10.7
33.0 formation
93.3
[E] 0.784
0.80
37.1 formation
1.49
Table 3: Rate Constants and T
1/2
Values Derived from Monte Carlo Simulation
rate
constant
k, h
-
1
T
1/2
, h
average SE LCI UCI
average SE LCI UCI
k1
0.00233 0.000112 0.00133 0.00401
297 14.2 173 520
k2
0.0584 0.000565 0.0522 0.0646
12 0.115
10.7
13.3
k3
0.0189 0.000207 0.0167 0.0210
37 0.402
33.0
41.5
k4
0.139 0.0109 0.0399 0.282
5 0.390
2.46
17.4
k5'
0.0317 0.000413 0.0274 0.0359
22 0.285
19.3
25.3
51
Degradation Products Other than Phosphate
Hydroxyethylphosphonate (C) has been reported previously as a degradation product of A in
buffered solutions, formed in 1% and 8% yield at pH 7.4 and 13.8 respectively,
123
suggesting
either differential formation or degradation of B at varying pH. Up to 10% C forms from
photolysis or soil degradation of A and up to 3.7% is formed from anaerobic degradation,
128
likely by direct SN
2
reaction at the β-carbon rather than via the intermediacy of B. To the best
of our knowledge E is previously unreported as a degradation product of A.
BChE Inhibitor
The BChE inhibitor forms spontaneously at neutral to alkaline pH and is short lived. The
concentration-time curves based on enzyme inhibition
119
or
31
P NMR analysis (this study) are
almost identical (Figure 25) indicating that the two methods measure the same compound (the
transient BChE inhibitor, B). The results of our kinetic modeling suggest that in pH 7.4
carbonate buffer in the absence of BChE, the total conversion to B is 10.7 mol % and that B
degrades to D (6 mol %, k
4
=0.139 h
-1
), C (3.2 mol %, k
2
=0.0584 h
-1
), and E (1.5 mol %, k
5
’=0.0317
h
-1
). In the presence of BChE, 3-5 mol % of A was trapped as a BChE adduct over 168 h.
119
The
difference in values for % conversion to B likely results from a 30-50% efficiency for the BChE
trap with the remainder of B degraded to C, D, and E under the conditions of the trapping
experiments. Although the formation and degradation of B depends on conditions, the high
agreement (Figure 25) between the concentration-time curves generated from BChE
trapping
119
and direct
31
P NMR monitoring of ethephon degradation (this study) suggests that
the efficiency of the BChE trap is likely limited by competitive degradation of B.
Multiple Degradation Pathways
There are two separate mechanisms for ethephon degradation; direct production of ethylene
(and phosphate) and formation of the BChE inhibitor. In both forks of the degradation
pathway, the first step is halide dissociation (Figure 19) which is in agreement with the
observed trend of higher BChE inhibitory potency for bromo and iodo analogs of A
120
as well as
analogs with electron withdrawing β-substituents.
129
Degradation of A under different
conditions might alter the percent formation of the BChE inhibitor B and the degradation
products C, D and E. In biological systems additional derivatives of many nucleophilic species
could form, but B reacts fairly specifically with BChE compared to numerous other esterases.
130
Thus, inhibited BChE is a useful monitor of ethephon exposure.
Summary
Fitted curves and rates are reported for the degradation reactions of ethephon in pH 7.4
potassium carbonate buffer. The overall conversion and rate for formation and degradation of
the inhibitor matches the rates measured by BChE trapping reported previously.
31
P NMR
chemical shifts and kinetic modeling of the proposed reaction mechanisms supports the
proposed 2-oxo-2-hydroxy-1,2-oxaphosphetane structure of the active BChE inhibitor formed
from ethephon degradation.
52
Figure 25: Comparison of BChE inhibitor concentration-time data from BChE trapping
experiments and
31
P NMR monitoring
Legend: BChE trapping results are reported previously.
119
53
Conclusions
OP pesticides have a long history of use, wide structural variation, and often surprisingly
selective efficacy and toxicity. This dissertation research addressed questions regarding the
mechanisms of toxicity for GLF and ethephon, two OP pesticides with atypical mammalian OP
toxicity. GLF, which produces convulsions and amnesia in mammals without AChE inhibition,
was further characterized as an excitotoxicity inducing NMDAR agonist. Ethephon, which
results selectively in BChE inhibition was further described as a pro-inhibitor, acting only after
spontaneous activation to 2-oxo-2-hydroxy-1,2-oxaphosphetane. These mechanisms are
important for the appropriate risk characterization of GLF and ethephon as significant
agricultural chemicals.
54
References
(1) Bayer, E.; Gugel, K. H.; Hägele, K.; Hagenmaier, H.; Jessipow, S.; König, W. A.; Zähner, H.
Stoffwechselprodukte von mikroorganismen. 98. mitteilung. phosphinothricin und
phosphinothricyl-alanyl-alanin. Helvetica Chimica Acta 1972, 55, 224–239.
(2) Omura, S.; Hinotozawa, K.; Imamura, N.; Murata, M. The structure of phosalacine, a
new herbicidal antibiotic containing phosphinothricin. J. Antibiot. 1984, 37, 939–940.
(3) Omura, S.; Murata, M.; Hanaki, H.; Hinotozawa, K.; Oiwa, R.; Tanaka, H. Phosalacine, a
new herbicidal antibiotic containing phosphinothricin. Fermentation, isolation,
biological activity and mechanism of action. J. Antibiot. 1984, 37, 829–835.
(4) California Department of Pesticide Regulation (CPDR) 2010 Summary of pesticide use
report data 2011.
(5) Hoerlein, G. Glufosinate (phosphinothricin), a natural amino acid with unexpected
herbicidal properties. Rev Environ Contam Toxicol 1994, 138, 73–145.
(6) Wendler, C.; Barniske, M.; Wild, A. Effect of phosphinothricin (glufosinate) on
photosynthesis and photorespiration of C3 and C4 plants. Photosynth Res 1990, 24, 55–
61.
(7) Wendler, C.; Putzer, A.; Wild, A. Effect of glufosinate (phosphinothricin) and inhibitors
of photorespiration on photosynthesis and ribulose-1,5-bisphosphate carboxylase
activity. Journal of Plant Physiology 1992, 139, 666–671.
(8) Abell, L. M.; Villafranca, J. J. Investigation of the mechanism of phosphinothricin
inactivation of Escherichia coli glutamine synthetase using rapid quench kinetic
technique. Biochemistry 1991, 30, 6135–6141.
(9) Woolfolk, C. A.; Shapiro, B.; Stadtman, E. R. Regulation of glutamine synthetase : I.
Purification and properties of glutamine synthetase from Escherichia coli. Archives of
Biochemistry and Biophysics 1966, 116, 177–192.
(10) Ericson, M. C. Purification and properties of glutamine synthetase from spinach leaves.
Plant Physiol. 1985, 79, 923–927.
(11) Gill, H. S.; Eisenberg, D. The crystal structure of phosphinothricin in the active site of
glutamine synthetase illuminates the mechanism of enzymatic inhibition. Biochemistry
2001, 40, 1903–1912.
(12) Avila-Garcia, W. V.; Sanchez-Olguin, E.; Hulting, A. G.; Mallory-Smith, C. Target-site
mutation associated with glufosinate resistance in Italian ryegrass (Lolium perenne L.
ssp. multiflorum). Pest Management Science 2012.
(13) Block, M. D.; Botterman, J.; Vandewiele, M.; Dockx, J.; Thoen, C.; Gossele, V.; Movva, N.
R.; Thompson, C.; Montagu, M. V.; Leemans, J. Engineering herbicide resistance in
plants by expression of a detoxifying enzyme. EMBO J 1987, 6, 2513–2518.
(14) Wohlleben, W.; Arnold, W.; Broer, I.; Hillemann, D.; Strauch, E.; Pühler, A. Nucleotide
sequence of the phosphinothricin N-acetyltransferase gene from Streptomyces
viridochromogenes Tü494 and its expression in Nicotiana tabacum. Gene 1988, 70, 25–
37.
(15) Wehrmann, A.; Van Vliet, A.; Opsomer, C.; Botterman, J.; Schulz, A. The similarities of
bar and pat gene products make them equally applicable for plant engineers. Nat.
Biotechnol. 1996, 14, 1274–1278.
55
(16) Thompson, C. J.; Movva, N. R.; Tizard, R.; Crameri, R.; Davies, J. E.; Lauwereys, M.;
Botterman, J. Characterization of the herbicide-resistance gene bar from Streptomyces
hygroscopicus. EMBO J 1987, 6, 2519–2523.
(17) Kleter, G. A.; Unsworth, J. B.; Harris, C. A. The impact of altered herbicide residues in
transgenic herbicide-resistant crops on standard setting for herbicide residues. Pest
Management Science 2011, 67, 1193–1210.
(18) Mao, Y.-C.; Hung, D.-Z.; Wu, M.-L.; Tsai, W.-J.; Wang, L.-M.; Ger, J.; Deng, J.-F.; Yang, C.-
C. Acute human Glufosinate-containing herbicide poisoning. Clinical Toxicology 2012, 1–
7.
(19) Koyama, K.; Koyama, K.; Goto, K. Cardiovascular effects of a herbicide containing
glufosinate and a surfactant: In vitro and in vivo analyses in rats. Toxicology and Applied
Pharmacology 1997, 145, 409–414.
(20) Ebert, E.; Leist, K.-H.; Mayer, D. Summary of safety evaluation toxicity studies of
glufosinate ammonium. Food and Chemical Toxicology 1990, 28, 339–349.
(21) Koyama, K.; Andou, Y.; Saruki, K.; Matsuo, H. Delayed and severe toxicities of a
herbicide containing glufosinate and a surfactant. Vet Hum Toxicol 1994, 36, 17–18.
(22) Park, H. Y.; Lee, P. H.; Shin, D. H.; Kim, G. W. Anterograde amnesia with hippocampal
lesions following glufosinate intoxication. Neurology 2006, 67, 914–915.
(23) Watanabe, T.; Sano, T. Neurological effects of glufosinate poisoning with a brief review.
Hum Exp Toxicol 1998, 17, 35–39.
(24) Ohtake, T.; Yasuda, H.; Takahashi, H.; Goto, T.; Suzuki, K.; Yonemura, K.; Hishida, A.
Decreased plasma and cerebrospinal fluid glutamine concentrations in a patient with
bialaphos poisoning. Hum Exp Toxicol 2001, 20, 429–434.
(25) Mao, Y.-C.; Wang, J.-D.; Hung, D.-Z.; Deng, J.-F.; Yang, C.-C. Hyperammonemia following
glufosinate-containing herbicide poisoning: A potential marker of severe neurotoxicity.
Clin Toxicol (Phila) 2011, 49, 48–52.
(26) Kyong, Y. Y.; Choi, K. H.; Oh, Y. M.; Lee, K. U. “Hyperammonemia following glufosinate-
containing herbicide poisoning: A potential marker of severe neurotoxicity” by Yan-
Chido Mao et al., Clin Toxicol (Phila) 2011; 49:48–52. Clinical Toxicology 2011, 49, 510–
512.
(27) Mao, Y.-C.; Yang, C.-C. Response to “Hyperammonemia following glufosinate-containing
herbicide poisoning: A potential marker of severe neurotoxicity” by Yan-Chido Mao et
al., Clin Toxicol (Phila) 2011; 49:48–52. Clinical Toxicology 2011, 49, 513–513.
(28) Hack, R.; Ebert, E.; Ehling, G.; Leist, K.-H. Glufosinate ammonium - Some aspects of its
mode of action in mammals. Food and Chemical Toxicology 1994, 32, 461–470.
(29) Neugebauer, V. Glutamate receptor ligands. Handb Exp Pharmacol 2007, 217–249.
(30) Masu, M.; Nakajima, Y.; Moriyoshi, K.; Ishii, T.; Akazawa, C.; Nakanashi, S. Molecular
characterization of NMDA and metabotropic glutamate receptors. Annals of the New
York Academy of Sciences 1993, 707, 153–164.
(31) Bunch, L.; Erichsen, M. N.; Jensen, A. A. Excitatory amino acid transporters as potential
drug targets. Expert Opin. Ther. Targets 2009, 13, 719–731.
(32) Peacey, E.; Miller, C. C. J.; Dunlop, J.; Rattray, M. The four major N- and C-terminal splice
variants of the excitatory amino acid transporter GLT-1 form cell surface homomeric
and heteromeric assemblies. Mol Pharmacol 2009, 75, 1062–1073.
56
(33) Aires, C. C. P.; van Cruchten, A.; IJlst, L.; de Almeida, I. T.; Duran, M.; Wanders, R. J. A.;
Silva, M. F. B. New insights on the mechanisms of valproate-induced hyperammonemia:
Inhibition of hepatic N-acetylglutamate synthase activity by valproyl-CoA. Journal of
Hepatology 2011, 55, 426–434.
(34) Daniotti, M.; la Marca, G.; Fiorini, P.; Filippi, L. New developments in the treatment of
hyperammonemia: emerging use of carglumic acid. Int J Gen Med 2011, 4, 21–28.
(35) Häberle, J. Role of carglumic acid in the treatment of acute hyperammonemia due to N-
acetylglutamate synthase deficiency. Ther Clin Risk Manag 2011, 7, 327–332.
(36) Selvam, C.; Goudet, C.; Oueslati, N.; Pin, J.-P.; Acher, F. C. l-(+)-2-Amino-4-
thiophosphonobutyric acid (l-thioAP4), a new potent agonist of group III metabotropic
glutamate receptors: Increased distal acidity affords enhanced potency. J. Med. Chem.
2007, 50, 4656–4664.
(37) Colanduoni, J. A.; Villafranca, J. J. Inhibition of Escherichia coli glutamine synthetase by
phosphinothricin. Bioorganic Chemistry 1986, 14, 163–169.
(38) Logusch, E. W.; Walker, D. M.; McDonald, J. F.; Franz, J. E. Substrate variability as a
factor in enzyme inhibitor design: Inhibition of ovine brain glutamine synthetase by α-
and γ-substituted phosphinothricins. Biochemistry 1989, 28, 3043–3051.
(39) Lacoste, A. M.; Mansour, S.; Cassaigne, A.; Neuzil, E. Effect of phosphonic analogues of
glutamic acid on glutamate decarboxylase. Cellular and Molecular Life Sciences 1985,
41, 643–644.
(40) Van Amsterdam, F. T. M.; Giberti, A.; Mugnaini, M.; Ratti, E. 3-[(±)-2-Carboxypiperazin-
4-yl]propyl-1-phosphonic acid recognizes two N-methyl-D-aspartate binding sites in rat
cerebral cortex membranes. Journal of Neurochemistry 1992, 59, 1850–1855.
(41) Olverman, H. J.; Jones, A. W.; Watkins, J. C. L-Glutamate has higher affinity than other
amino acids for [
3
H]-D-AP5 binding sites in rat brain membranes. Nature 1984, 307,
460–462.
(42) Mugnaini, M.; Giberti, A.; Ratti, E.; van Amsterdam, F. T. Allosteric modulation of
[
3
H]CGP 39653 binding by glycine in rat brain. J. Neurochem. 1993, 61, 1492–1497.
(43) Bradford, M. M. A rapid and sensitive method for the quantitation of microgram
quantities of protein utilizing the principle of protein-dye binding. Anal. Biochem. 1976,
72, 248–254.
(44) Brown, A. M. A step-by-step guide to non-linear regression analysis of experimental
data using a Microsoft Excel spreadsheet. Comput Methods Programs Biomed 2001, 65,
191–200.
(45) Ulbrich, M. H.; Isacoff, E. Y. Rules of engagement for NMDA receptor subunits. Proc.
Natl. Acad. Sci. U.S.A. 2008, 105, 14163–14168.
(46) Jones, M. L.; Liao, G.-Y.; Malecki, R.; Li, M.; Salazar, N. M.; Leonard, J. P. PI 3-kinase and
PKCζ mediate insulin-induced potentiation of NMDA receptor currents in Xenopus
oocytes. Brain Research 2012, 1432, 7–14.
(47) Dascal, N. Voltage clamp recordings from Xenopus oocytes. In Current Protocols in
Neuroscience; John Wiley & Sons, Inc., 2001.
(48) McConnell, E. R.; McClain, M. A.; Ross, J.; Lefew, W. R.; Shafer, T. J. Evaluation of multi-
well microelectrode arrays for neurotoxicity screening using a chemical training set.
Neurotoxicology 2012, 33, 1048–1057.
57
(49) Duane L., P. A rapid colorimetric assay for carbamyl phosphate synthetase I. Journal of
Biochemical and Biophysical Methods 1980, 3, 31–37.
(50) Excel 2007; Microsoft Corporation: Redmond, WA, USA.
(51) Sills, M. A.; Fagg, G.; Pozza, M.; Angst, C.; Brundish, D. E.; Hurt, S. D.; Jay Wilusz, E.;
Williams, M. [
3
H]CGP 39653: a new N-methyl-D-aspartate antagonist radioligand with
low nanomolar affinity in rat brain. European Journal of Pharmacology 1991, 192, 19–
24.
(52) Reynolds, I. J. [
3
H]CGP 39653 binding to the agonist site of the N-methyl-D-aspartate
receptor is modulated by Mg
2+
and polyamines independently of the arcaine-sensitive
polyamine site. J. Neurochem. 1994, 62, 54–62.
(53) Frega, M.; Pasquale, V.; Tedesco, M.; Marcoli, M.; Contestabile, A.; Nanni, M.; Bonzano,
L.; Maura, G.; Chiappalone, M. Cortical cultures coupled to Micro-Electrode Arrays: A
novel approach to perform in vitro excitotoxicity testing. Neurotoxicology and
Teratology 2012, 34, 116–127.
(54) Pesch, K. L.; Simmert, U. Combined assays for lactose and galactose by enzymatic
reactions. Milchw. Forsch. 1929, 8, 551.
(55) Anoopkumar-Dukie, S.; Carey, J. B.; Conere, T.; O’sullivan, E.; van Pelt, F. N.; Allshire, A.
Resazurin assay of radiation response in cultured cells. Br J Radiol 2005, 78, 945–947.
(56) Dewhurst, I. Pesticide residues in food, toxicological evaluations, glufosinate ammonium
(addendum); International Programme on Chemical Safety (IPCS), 1999.
(57) Nissim, I.; Horyn, O.; Nissim, I.; Daikhin, Y.; Caldovic, L.; Barcelona, B.; Cervera, J.;
Tuchman, M.; Yudkoff, M. Down-regulation of hepatic urea synthesis by oxypurines:
xanthine and uric acid inhibit N-acetylglutamate synthase. J. Biol. Chem. 2011, 286,
22055–22068.
(58) Caldovic, L.; Ah Mew, N.; Shi, D.; Morizono, H.; Yudkoff, M.; Tuchman, M. N-
acetylglutamate synthase: structure, function and defects. Mol. Genet. Metab. 2010,
100 Suppl 1, S13–19.
(59) Nordenström, A.; Halldin, M.; Hallberg, B.; Alm, J. A trial with N-carbamylglutamate may
not detect all patients with NAGS deficiency and neonatal onset. J. Inherit. Metab. Dis
2007, 30, 400.
(60) Bachmann, C.; Krähenbühl, S.; Colombo, J. P.; Schubiger, G.; Jaggi, K. H.; Tönz, O. N-
acetylglutamate synthetase deficiency: a disorder of ammonia detoxication. N. Engl. J.
Med. 1981, 304, 543.
(61) Caldovic, L.; Lopez, G. Y.; Haskins, N.; Panglao, M.; Shi, D.; Morizono, H.; Tuchman, M.
Biochemical properties of recombinant human and mouse N-acetylglutamate synthase.
Mol. Genet. Metab. 2006, 87, 226–232.
(62) Liu, K.-T.; Su, C.-S. Postictal transient hyperammonemia. Am J Emerg Med 2008, 26,
388.e1–2.
(63) Liu, K.-T.; Yang, S.-C.; Yeh, I.-J.; Lin, T.-J.; Lee, C.-W. Transient hyperammonemia
associated with postictal state in generalized convulsion. Kaohsiung J. Med. Sci. 2011,
27, 453–456.
(64) Vandenberg, R. J.; Mitrovic, A. D.; Chebib, M.; Balcar, V. J.; Johnston, G. A. R. Contrasting
modes of action of methylglutamate derivatives on the excitatory amino acid
transporters, EAAT1 and EAAT2. Mol Pharmacol 1997, 51, 809–815.
58
(65) Matsumura, N.; Takeuchi, C.; Hishikawa, K.; Fujii, T.; Nakaki, T. Glufosinate ammonium
induces convulsion through N-methyl-D-aspartate receptors in mice. Neuroscience
Letters 2001, 304, 123–125.
(66) Ferreira Faro, L. R.; Ferreira Nunes, B. V.; Alfonso, M.; Moraes Ferreira, V. M.; Durán, R.
Role of glutamate receptors and nitric oxide on the effects of glufosinate ammonium, an
organophosphate pesticide, on in vivo dopamine release in rat striatum. Toxicology.
(67) Fagg, G. E.; Olpe, H.-R.; Pozza, M. F.; Baud, J.; Steinmann, M.; Schmutz, M.; Portet, C.;
Baumann, P.; Thedinga, K.; Bittiger, H.; Allgeier, H.; Heckendorn, R.; Angst, C.; Brundish,
D.; Dingwall, J. G. CGP 37849 and CGP 39551: novel and potent competitive N-methyl-D-
aspartate receptor antagonists with oral activity. Br J Pharmacol 1990, 99, 791–797.
(68) Watkins, J. C.; Jane, D. E. The glutamate story. Br J Pharmacol 2006, 147, S100–S108.
(69) Logusch, E. W.; Walker, D. M.; McDonald, J. F.; Franz, J. E.; Villafranca, J. J.; DiIanni, C. L.;
Colanduoni, J. A.; Li, B.; Schineller, J. B. Inhibition of Escherichia coli glutamine
synthetase by alpha- and gamma-substituted phosphinothricins. Biochemistry 1990, 29,
366–372.
(70) Shaw, C. A.; Bains, J. S.; Pasqualotto, B. A.; Curry, K. Methionine sulfoximine shows
excitotoxic actions in rat cortical slices. Can. J. Physiol. Pharmacol. 1999, 77, 871–877.
(71) Lapouble, E.; Montécot, C.; Sevestre, A.; Pichon, J. Phosphinothricin induces epileptic
activity via nitric oxide production through NMDA receptor activation in adult mice.
Brain Research 2002, 957, 46–52.
(72) Hori, Y.; Tanaka, T.; Fujisawa, M.; Shimada, K. Toxicokinetics of DL-glufosinate
enantiomer in human BASTA poisoning. Biol. Pharm. Bull 2003, 26, 540–543.
(73) Lee, W. J.; Hawkins, R. A.; Viña, J. R.; Peterson, D. R. Glutamine transport by the blood-
brain barrier: a possible mechanism for nitrogen removal. Am. J. Physiol. 1998, 274,
C1101–1107.
(74) Ziylan, Y. Z.; Uzüm, G.; Bernard, G.; Diler, A. S.; Bourre, J. M. Changes in the permeability
of the blood-brain barrier in acute hyperammonemia. Effect of dexamethasone. Mol.
Chem. Neuropathol. 1993, 20, 203–218.
(75) Søgaard, R.; Novak, I.; MacAulay, N. Elevated ammonium levels: differential acute
effects on three glutamate transporter isoforms. Am J Physiol Cell Physiol 2012, 302,
C880–C891.
(76) Izumi, Y.; Svrakic, N.; O’Dell, K.; Zorumski, C. F. Ammonia inhibits long-term potentiation
via neurosteroid synthesis in hippocampal pyramidal neurons. Neuroscience 2013, 233,
166–173.
(77) Franco, A. R.; Lopez-Siles, F. J.; Cardenas, J. Resistance to phosphinothricin (glufosinate)
and its utilization as a nitrogen source by Chlamydomonas reinhardtii. Appl. Environ.
Microbiol. 1996, 62, 3834–3839.
(78) Mugnaini, M.; van Amsterdam, F. T.; Ratti, E.; Trist, D. G.; Bowery, N. G. Regionally
different N-methyl-D-aspartate receptors distinguished by ligand binding and
quantitative autoradiography of [3H]-CGP 39653 in rat brain. Br J Pharmacol 1996, 119,
819–828.
(79) Calas, A.-G.; Richard, O.; Même, S.; Beloeil, J.-C.; Doan, B.-T.; Gefflaut, T.; Même, W.;
Crusio, W. E.; Pichon, J.; Montécot, C. Chronic exposure to glufosinate-ammonium
59
induces spatial memory impairments, hippocampal MRI modifications and glutamine
synthetase activation in mice. Neurotoxicology 2008, 29, 740–747.
(80) Watanabe, T. Apoptosis induced by glufosinate ammonium in the neuroepithelium of
developing mouse embryos in culture. Neurosci. Lett. 1997, 222, 17–20.
(81) Fujii, T.; Ohata, T.; Horinaka, M. Alterations in the response to kainic acid in rats
exposed to glufosinate ammonium, a herbicide, during infantile period. Proceedings of
the Japan Academy. Ser. B: Physical and Biological Sciences 1996, 72, 7–10.
(82) Moghaddam, B.; Javitt, D. From revolution to evolution: the glutamate hypothesis of
schizophrenia and its implication for treatment. Neuropsychopharmacology 2012, 37,
4–15.
(83) Kantrowitz, J. T.; Javitt, D. C. N-methyl-d-aspartate (NMDA) receptor dysfunction or
dysregulation: the final common pathway on the road to schizophrenia? Brain Res. Bull.
2010, 83, 108–121.
(84) Doble, A. The role of excitotoxicity in neurodegenerative disease: implications for
therapy. Pharmacol. Ther. 1999, 81, 163–221.
(85) Mehta, A.; Prabhakar, M.; Kumar, P.; Deshmukh, R.; Sharma, P. L. Excitotoxicity: bridge
to various triggers in neurodegenerative disorders. Eur. J. Pharmacol. 2013, 698, 6–18.
(86) Hynd, M. R.; Scott, H. L.; Dodd, P. R. Glutamate-mediated excitotoxicity and
neurodegeneration in Alzheimer’s disease. Neurochem. Int. 2004, 45, 583–595.
(87) Ong, W.-Y.; Tanaka, K.; Dawe, G. S.; Ittner, L. M.; Farooqui, A. A. Slow excitotoxicity in
Alzheimer’s disease. J. Alzheimers Dis. 2013, 35, 643–668.
(88) Kostic, M.; Zivkovic, N.; Stojanovic, I. Multiple sclerosis and glutamate excitotoxicity. Rev
Neurosci 2013, 24, 71–88.
(89) Essa, M. M.; Braidy, N.; Vijayan, K. R.; Subash, S.; Guillemin, G. J. Excitotoxicity in the
pathogenesis of autism. Neurotox Res 2013, 23, 393–400.
(90) Guo, H.; Lai, L.; Butchbach, M. E. R.; Stockinger, M. P.; Shan, X.; Bishop, G. A.; Lin, C. G.
Increased expression of the glial glutamate transporter EAAT2 modulates excitotoxicity
and delays the onset but not the outcome of ALS in mice. Hum. Mol. Genet. 2003, 12,
2519–2532.
(91) Wang, Y.; Denisova, J. V.; Kang, K. S.; Fontes, J. D.; Zhu, B. T.; Belousov, A. B. Neuronal
gap junctions are required for NMDA receptor-mediated excitotoxicity: implications in
ischemic stroke. J. Neurophysiol. 2010, 104, 3551–3556.
(92) Hansen, A. J. The importance of glutamate receptors in brain ischemia. Adv. Exp. Med.
Biol. 1995, 363, 123–131.
(93) Choi, D. Antagonizing excitotoxicity: a therapeutic strategy for stroke? Mt. Sinai J. Med.
1998, 65, 133–138.
(94) Druart, C.; Millet, M.; Scheifler, R.; Delhomme, O.; Raeppel, C.; de Vaufleury, A. Snails as
indicators of pesticide drift, deposit, transfer and effects in the vineyard. Sci. Total
Environ. 2011, 409, 4280–4288.
(95) Delhomme, O.; Raeppel, C.; Teigné, D.; Briand, O.; Millet, M. Analytical method for
assessing potential dermal exposure to pesticides of a non-agricultural occupationally
exposed population. Anal Bioanal Chem 2011, 399, 1325–1334.
(96) United States Department of Agriculture (USDA) Pesticide Data Program - Annual
Summary, Calendar Year 2011 2013.
60
(97) Nougadère, A.; Reninger, J.-C.; Volatier, J.-L.; Leblanc, J.-C. Chronic dietary risk
characterization for pesticide residues: A ranking and scoring method integrating
agricultural uses and food contamination data. Food and Chemical Toxicology 2011, 49,
1484–1510.
(98) Han, F. Method for synthesizing glufosinate. CN102372739 (A), March 14, 2012.
(99) Hideaki, T.; Atsushi, N.; Katsuzou, T.; Nobuyuki, O.; Takeshi, E. Manufacture of L-2-
amino-4-(hydroxymethylphosphinyl)butyric acid with transaminase-producing
microorganisms. 06-245780, September 6, 1994.
(100) Li, J. Isomerization reaction. CN102336783 (A), February 1, 2012.
(101) Li, C.; Li, J.; Niu, Y. Preparation method of glufosinate. CN101747367 (A), June 23, 2010.
(102) Logusch, E. . Facile synthesis of D,L-phosphinothricin from methyl 4-bromo-2-
phthalimidobutyrate. Tetrahedron Letters 1986, 27, 5935–5938.
(103) Logusch, E. W. 2-Amino-4-Phosphonyl-butanoate derivates and preparation of 2-amino-
4-phosphinyl butanoic acids. EP0265412 (A2), April 27, 1988.
(104) Minowa, N.; Hirayama, M.; Fukatsu, S. Asymmetric synthesis of (+)-phosphinothricin
and related compounds by the Michael addition of glycine Schiff bases to vinyl
compounds. Bulletin of the Chemical Society of Japan 1987, 60, 1761–1766.
(105) Minowa, N.; Mitomi, M. Manufacturing method of phosphorus-containing alpha-amino
acid, and phosphorus-containing nitro derivative as production intermediate thereof.
JP2009023942 (A), February 5, 2009.
(106) Suzuki, A.; Inouye, S.; Tsuruoka, T.; Takematsu, T.; Mizutani, K.; Tachibana, K.; Sekizawa,
Y. Process for the production of D,L-2-amino-4-methylsulphosphino-butyric acid.
WO7900405 (A1), July 12, 1979.
(107) Wasielewski, C.; Antczak, K. Aminophosphonic acids; XVI
1
. A new and facile synthesis of
phosphinothricine and 2-amino-4-phosphonobutanoic acid. Synthesis 1981, 1981, 540–
541.
(108) Abrunhosa-Thomas, I.; Sellers, C. E.; Montchamp, J.-L. Alkylation of H-phosphinate
esters under basic conditions. J Org Chem 2007, 72, 2851–2856.
(109) Monaghan, D. T.; McMills, M. C.; Chamberlin, A. R.; Cotman, C. W. Synthesis of [
3
H]2-
amino-4-phosphonobutyric acid and characterization of its binding to rat brain
membranes: a selective ligand for the chloride/calcium-dependent class of L-glutamate
binding sites. Brain Res 1983, 278, 137–144.
(110) Bravo-Altamirano, K.; Coudray, L.; Deal, E. L.; Montchamp, J.-L. Strategies for the
asymmetric synthesis of H-phosphinate esters. Org. Biomol. Chem. 2010, 8, 5541–5551.
(111) Thottathil, J. K.; Przybyla, C. A.; Moniot, J. L. Mild Arbuzov reactions of phosphonous
acids. Tetrahedron Letters 1984, 25, 4737–4740.
(112) Bylund, D. B.; Toews, M. L. Radioligand binding methods: practical guide and tips. Am J
Physiol Lung Cell Mol Physiol 1993, 265, L421–429.
(113) Maguire, J. J.; Kuc, R. E.; Davenport, A. P. Radioligand binding assays and their analysis.
Methods Mol. Biol. 2012, 897, 31–77.
(114) Schütz, W. The pharmacological basis of receptor binding. Wien. Klin. Wochenschr.
1991, 103, 438–442.
61
(115) Grubbs, R. H.; Trnka, T. M. Ruthenium-catalyzed olefin metathesis. In Ruthenium in
Organic Synthesis; Murahashi, essor S.-I., Ed.; Wiley-VCH Verlag GmbH & Co. KGaA,
2005; pp. 153–177.
(116) Nguyen, S. T.; Johnson, L. K.; Grubbs, R. H.; Ziller, J. W. Ring-opening metathesis
polymerization (ROMP) of norbornene by a Group VIII carbene complex in protic media.
J. Am. Chem. Soc. 1992, 114, 3974–3975.
(117) Schwab, P.; France, M. B.; Ziller, J. W.; Grubbs, R. H. A series of well-defined metathesis
catalysts–synthesis of [RuCl
2
(=CHR′)(PR
3
)
2
] and its reactions. Angewandte Chemie
International Edition in English 1995, 34, 2039–2041.
(118) Schwab, P.; Grubbs, R. H.; Ziller, J. W. Synthesis and applications of RuCl
2
(CHR‘)(PR3)
2
:
the influence of the alkylidene moiety on metathesis activity. J. Am. Chem. Soc. 1996,
118, 100–110.
(119) Liyasova, M. S.; Schopfer, L. M.; Kodani, S.; Lantz, S. R.; Casida, J. E.; Lockridge, O. Newly
observed spontaneous activation of ethephon as a butyrylcholinesterase inhibitor.
Chem. Res. Toxicol. 2013, 26, 422–431.
(120) Haux, J. E.; Quistad, G. B.; Casida, J. E. Phosphobutyrylcholinesterase: phosphorylation
of the esteratic site of butyrylcholinesterase by ethephon [(2-chloroethyl)phosphonic
acid] dianion. Chem. Res. Toxicol. 2000, 13, 646–651.
(121) Mucientes, A. E.; de la Peña, M. A. Kinetic analysis of parallel-consecutive first-order
reactions with a reversible step: concentration–time integrals method. J. Chem. Educ.
2009, 86, 390–392.
(122) Lambert, R. J. W.; Mytilinaios, I.; Maitland, L.; Brown, A. M. Monte Carlo simulation of
parameter confidence intervals for non-linear regression analysis of biological data
using Microsoft Excel. Computer Methods and Programs in Biomedicine 2012, 107, 155–
163.
(123) Segall, Y.; Grendell, R. L.; Toia, R. F.; Casida, J. E. Composition of technical ethephon [(2-
chloroethyl)phosphonic acid] and some analogs relative to their reactivity and biological
activity. J. Agric. Food Chem. 1991, 39, 380–385.
(124) Verkade, J. G.; Quin, L. D. Phosphorus-31 NMR spectroscopy in stereochemical analysis
(organic compounds and metal complexes); Methods in Stereochemical Analysis; VCH
Publishers, Inc: Deerfield Beach, Florida, 1987.
(125) Breur, E.; Zbaida, S.; Segall, E. The observation of an oxaphosphetane intermediate in
the phosphonate modification of the wittig oelfin synthesis. Tetrahedron Letters 1979,
20, 2203–2204.
(126) Biddle, E.; Kerfoot, D. G. S.; Kho, Y. H.; Russell, K. E. Kinetic studies of the thermal
decomposition of 2-chloroethylphosphonic Acid in aqueous solution. Plant Physiol.
1976, 58, 700–702.
(127) Domir, S. C.; Foy, C. L. A study of ethylene and CO
2
evolution from ethephon in tobacco.
Pesticide Biochemistry and Physiology 1978, 9, 1–8.
(128) Bureau-REACH-RIVM Proposal for harmonized classification and labeling; substance
ethephon, Annex VI Report 2011.
(129) Zhang, N.; Casida, J. E. Novel irreversible butyrylcholinesterase inhibitors: 2-chloro-1-
(substituted-phenyl)ethylphosphonic acids. Bioorg. Med. Chem. 2002, 10, 1281–1290.
62
(130) Haux, J. E.; Lockridge, O.; Casida, J. E. Specificity of ethephon as a butyrylcholinesterase
inhibitor and phosphorylating agent. Chem. Res. Toxicol. 2002, 15, 1527–1533.
