Highly Selective, Reversible Inhibitor Identified by Comparative Chemoproteomics Modulates Diacylglycerol Lipase Activity in Neurons

Article abstract of DOI:10.1021/jacs.5b04883

Diacylglycerol lipase (DAGL)-α and -β are enzymes responsible for the biosynthesis of the endocannabinoid 2-arachidonoylglycerol (2-AG). Selective and reversible inhibitors are required to study the function of DAGLs in neuronal cells in an acute and temporal fashion, but they are currently lacking. Here, we describe the identification of a highly selective DAGL inhibitor using structure-guided and a chemoproteomics strategy to characterize the selectivity of the inhibitor in complex proteomes. Key to the success of this approach is the use of comparative and competitive activity-based proteome profiling (ABPP), in which broad-spectrum and tailor-made activity-based probes are combined to report on the inhibition of a protein family in its native environment. Competitive ABPP with broad-spectrum fluorophosphonate-based probes and specific β-lactone-based probes led to the discovery of α-ketoheterocycle LEI105 as a potent, highly selective, and reversible dual DAGL-α/DAGL-β inhibitor. LEI105 did not affect other enzymes involved in endocannabinoid metabolism including abhydrolase domain-containing protein 6, abhydrolase domain-containing protein 12, monoacylglycerol lipase, and fatty acid amide hydrolase and did not display affinity for the cannabinoid CB₁ receptor. Targeted lipidomics revealed that LEI105 concentration-dependently reduced 2-AG levels, but not anandamide levels, in Neuro2A cells. We show that cannabinoid CB₁-receptor-mediated short-term synaptic plasticity in a mouse hippocampal slice model can be reduced by LEI105. Thus, we have developed a highly selective DAGL inhibitor and provide new pharmacological evidence to support the hypothesis that "on demand biosynthesis" of 2-AG is responsible for retrograde signaling.

Full text of DOI:10.1021/jacs.5b04883

Article
pubs.acs.org/JACS
Highly Selective, Reversible Inhibitor Identified by Comparative
Chemoproteomics Modulates Diacylglycerol Lipase Activity in
Neurons
Marc P. Baggelaar,† Pascal J. P. Chameau,‡ Vasudev Kantae,§ Jessica Hummel,† Ku-Lung Hsu,∥,#
Freek Janssen,† Tom van der Wel,† Marjolein Soethoudt,† Hui Deng,† Hans den Dulk,† Marco Allara,⊥
̀
Bogdan I. Florea,† Vincenzo Di Marzo,⊥ Wytse J. Wadman,‡ Chris G. Kruse,‡ Herman S. Overkleeft,†
Thomas Hankemeier,§ Taco R. Werkman,‡ Benjamin F. Cravatt,∥ and Mario van der Stelt*,†
†Department of Bioorganic Synthesis, Leiden Institute of Chemistry, Leiden University, Leiden 2300 RA, The Netherlands
‡Center for Neuroscience, Swammerdam Institute for Life Sciences, University of Amsterdam, Amsterdam 1000 GG, The
Netherlands
§Division of Analytical Biosciences, Leiden Academic Centre for Drug Research, Leiden University, Leiden 2300 RA, The
Netherlands
∥Department of Chemical Physiology, The Scripps Research Institute, La Jolla, California 92037, United States
⊥Endocannabinoid Research Group, Institute of Biomolecular Chemistry, Pozzuoli 80078, Italy
S
* Supporting Information
ABSTRACT: Diacylglycerol lipase (DAGL)-α and -β are
enzymes responsible for the biosynthesis of the endocannabi-
noid 2-arachidonoylglycerol (2-AG). Selective and reversible
inhibitors are required to study the function of DAGLs in
neuronal cells in an acute and temporal fashion, but they are
currently lacking. Here, we describe the identification of a
highly selective DAGL inhibitor using structure-guided and a
chemoproteomics strategy to characterize the selectivity of the
inhibitor in complex proteomes. Key to the success of this
approach is the use of comparative and competitive activity-based proteome profiling (ABPP), in which broad-spectrum and
tailor-made activity-based probes are combined to report on the inhibition of a protein family in its native environment.
Competitive ABPP with broad-spectrum fluorophosphonate-based probes and specific β-lactone-based probes led to the
discovery of α-ketoheterocycle LEI105 as a potent, highly selective, and reversible dual DAGL-α/DAGL-β inhibitor. LEI105 did
not affect other enzymes involved in endocannabinoid metabolism including abhydrolase domain-containing protein 6,
abhydrolase domain-containing protein 12, monoacylglycerol lipase, and fatty acid amide hydrolase and did not display affinity
for the cannabinoid CB1 receptor. Targeted lipidomics revealed that LEI105 concentration-dependently reduced 2-AG levels, but
not anandamide levels, in Neuro2A cells. We show that cannabinoid CB1-receptor-mediated short-term synaptic plasticity in a
mouse hippocampal slice model can be reduced by LEI105. Thus, we have developed a highly selective DAGL inhibitor and
provide new pharmacological evidence to support the hypothesis that “on demand biosynthesis” of 2-AG is responsible for
retrograde signaling.
receptor dependent (patho)physiological effects.6,7 Selective
inhibition of the formation of anandamide and 2-AG would be
instrumental to determine which endocannabinoid is respon-
sible for specific CB1-mediated physiological effects. However,
pathway-selective inhibitors for 2-AG and anandamide biosyn-
thesis are currently lacking.
INTRODUCTION
■
Endocannabinoids are endogenous signaling lipids that activate
the cannabinoid CB1 and CB2 receptor. They play an essential
role in human health and disease, regulating processes such as
immunomodulation, energy balance and neurotransmission.1
There are two main endocannabinoids: anandamide and 2-
arachidonoylglycerol (2-AG).2−4 Both endocannabinoids are
2-AG is mainly formed by the action of two diacylglycerol
lipases (DAGL-α and DAGL-β).8 DAGLs are intracellular,
often found together, but their levels vary between species,
tissue type, developmental stage, and pathological condition.5
multidomain integral membrane proteins. The DAGLs share
extensive homology, but differ in size: ∼120 and ∼70 kDa for
Although selective inhibitors of their metabolic pathways have
provided information about the biological function of the
endocannabinoids, it is still unclear to a large extent which
Received: May 11, 2015
endocannabinoid is responsible for specific cannabinoid CB1
© XXXX American Chemical Society
A
DOI: 10.1021/jacs.5b04883
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
Figure 1. Structure-guided modeling and biochemical characterization of LEI105. (a) Structures of α-ketoheterocycle based DAGL inhibitors
LEI104 and LEI105. (b) LEI104 and LEI105 in a homology model of DAGL-α. (c) Representative fluorescent ABPP gel showing dose dependent
inhibition of MB064 (250 nM) labeling of endogenous DAGL-α labeling by LEI105 in the mouse brain membrane proteome. (* = DAGL-α
breakdown product) (d) Dose response curve of DAGL-α inhibition as determined with competitive ABPP (pIC50 7.5 0.07 (IC50 = 32 nM); n =
3). (e) Dose response curve of DAGL-α inhibition by LEI104 (pIC50 6.3 0.1 (IC50 = 501 nM); n = 4) and LEI105 (pIC50 7.9 0.08 nM (IC50
=
13 nM); n = 4) as determined with a glycerol based natural substrate assay. (f) ABPP using MB064 (1 μM) with different hDAGL-β constructs and
anti-FLAG Western blot of the same gel. (g) Competitive ABPP in the mouse spleen membrane proteome using MB064 (1.0 μM) in competition
with LEI105 (10 μM), LEI105 can block labeling of endogenously expressed DAGL-β in the mouse spleen membrane proteome. (H) Schematic
representation of the size exclusion chromatography (SEC) experiment that shows reversibility of LEI105 in recombinant DAGL-α (n = 3, full
fluorescent gel and Western blot are given in the Supporting Information, SI) Statistical analysis: 2-way ANOVA (*** = p < 0.001; ** = p < 0.01 vs
vehicle).
DAGL-α and DAGL-β, respectively.8,9 DAGLs belong to the
class of serine hydrolases that employ the typical Ser-His-Asp
catalytic triad to hydrolyze the ester bond of acyl chains from
arachidonate-containing diacylglycerols in a sn-1 specific
manner. Studies with DAGL knockout mice have shown that
DAGL-α controls to a large extent the formation of 2-AG in the
central nervous system, whereas DAGL-β appears to partake in
2-AG production in the periphery during inflammation.5,10
Importantly, also basal anandamide levels were reduced in
DAGL-α knockout mice. Selective inhibitors for DAGLs, which
can be used in an acute and temporal fashion and do not
modulate anandamide levels, would, therefore, constitute an
important counterpart of the DAGL knockout mice, and allow
the examination of acute versus congenital inhibition.
Although, several classes of DAGL inhibitors have been
described in literature,8,11−15 these inhibitors are based on the
natural substrates and/or have reactive chemical warheads, and
are not selective over other serine hydrolases that modulate
endocannabinoid signaling (e.g., abhydrolase domain-contain-
ing protein 6 and 12 (ABHD6 and ABHD12), monoacylglycer-
ol lipase (MAGL), or fatty acid amide hydrolase (FAAH)),
therefore highly selective DAGL inhibitors are warranted to
study the cellular role of the biosynthetic enzymes.
entity (fluorophore, biotin, bioorthogonal tag) to label the
active site of the enzyme or enzyme family at hand. ABPP is
unique in its ability to rapidly identify inhibitor activity and
selectivity within large enzyme families in complex proteome
samples. The prototypical ABP for serine hydrolases is based
on a fluorophosphonate (FP)-warhead.18 However, this probe
does not recognize DAGL-α, and the signals for DAGL-β,
MAGL, and ABHD6 cannot be unequivocally established
because not all gel bands can be clearly resolved in the brain
proteome.19 Previously, we have reported on the design,
synthesis, and characterization of a specific β-lactone-containing
probe (MB064), which was tailor-made for the detection of
DAGL-α activity.20,21 Using this probe, we identified 1-
(oxazolo[4,5-b]pyridin-2-yl)-6-phenylhexan-1-one (LEI104,
Figure 1), which belongs to the class of α-ketoheterocycles,
as the first reversible inhibitor for DAGL-α. LEI104 was,
however, weakly active in a cellular assay and was not selective
over FAAH, the enzyme responsible for the metabolism of the
other endocannabinoid anandamide (AEA).22,23 Moreover,
activity of LEI104 on DAGL-β was not studied, but in view
of the high homology between DAGL-α and DAGL-β, it is
likely that there is cross-reactivity. In order to apply α-
ketoheterocycles as chemical tools to study 2-AG signaling, it is
important to increase their cellular activity, to have selectivity
over FAAH and to assess their activity on endogenous DAGL-
β.
Activity-based protein profiling (ABPP) has emerged as a
powerful technique for discovering selective enzyme inhibitors
acting in their native physiological context.16 ABPP hinges on
the use of activity-based probes (ABPs) to report on enzyme
activity in cells, tissue or animals.17 An ABP normally consists
of a covalent, irreversible enzyme inhibitor featuring a reporter
Here, we report a structure-guided approach to optimize
LEI104 employing a homology-model of DAGL-α. In addition,
we discovered that our tailor-made β-lactone probe MB064
B
DOI: 10.1021/jacs.5b04883
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
Figure 2. Selectivity of LEI105. (a) Competitive ABPP in the mouse brain membrane proteome with LEI104 (10 μM), LEI105 (10 μM),
OMDM188 (1 μM), and THL (1 μM) using TAMRA-FP (500 nM). (b) Same experiment as in (a) with THL based ABP MB064 (250 nM)
showing excellent selectivity of LEI105 at 10 μM compared to THL and OMDM188. (c) One possible LEI105-like conformer in a previously
reported cocrystal structure of FAAH, showing a steric clash between the toluyl group and FAAH. (d) Activity of LEI104 and LEI105 against FAAH
in a concentration response experiment in the mouse brain membrane proteome using TAMRA-FP (500 nM, full gels are given in the SI). (e)
Chemoproteomic competition between LEI104 (10 μM) and FP-biotin (5 μm) for FAAH, MAGL, and ABHD6, showing that the chemoproteomic
settings using FP-biotin are compatible with reversible α-ketoheterocycles. (f) MB108 and FP-biotin based chemoproteomic analysis of serine
hydrolase activities in the mouse brain membrane proteome treated with LEI105 (10 μM). DAGL-α labeling is reduced to over 70%. (g)
Chemoproteomic control experiment shows equal isotopic labeling and detection of peptides from MB108 and FP-biotin labeled serine hydrolases.
Enzymes known to be directly involved in endocannabinoid biosynthesis and metabolism are highlighted in red. (n = 3−4 independent
chemoproteomics experiments; Error bars represent s.e.m. of medium over light ratios of quantified peptides (minimum of two unique peptides
per enzyme). Only unique proteins are given for MB108, all detected proteins including proteins overlapping between the probes are given in the SI.
See SI data set 1 for complete proteomic data.
could also label DAGL-β in cells and tissues. Using these tools,
we characterized LEI105 as a cellular active, dual DAGL-α/β
inhibitor. Comparative chemoproteomics revealed that LEI105
is selective over ABHD6, ABHD12, MAGL, and FAAH.
Furthermore, targeted lipidomics revealed that LEI105 is able
to concentration-dependently reduce 2-AG levels in neuronal
cells without affecting AEA levels. We showed that cannabinoid
CB1-receptor-dependent short-term synaptic plasticity in a
hippocampal slice model can be reduced by the selective
DAGL-inhibitor LEI105. In summary, comparative and
competitive chemoproteomics were applied to characterize
the most selective DAGL inhibitor to date, that can be used to
study DAGL function in an acute and temporal manner in a
neuronal context.
the binding pose of LEI104 in DAGL-α using a molecular
dynamics simulation in our previously generated homology
model.21 From this analysis, we identified an additional
hydrophobic pocket close to the catalytic site that did not
appear to be occupied by LEI104 (Figure 1B). Introduction of
a phenyl substituent at the 6-position of the oxazolopyridine
would allow us to probe this pocket with the aim of increasing
potency and/or selectivity. We synthesized compound
(LEI105) (see SI for synthesis) and tested its activity on
human DAGL-α in a colorimetric assay using para-nitro-
phenylbutyrate as a surrogate substrate.24 LEI105 proved to be
a potent inhibitor with a pIC50 of 8.5 0.06 (n = 4), thus some
10-fold more potent than LEI104 (7.4
0.05; n = 4).21
Reduction of the α-keto group to the corresponding alcohol
(compound 8 (SI)) led to a ∼150 fold drop of activity against
DAGL-α, thereby indicating that the α-carbonyl in LEI105
reacts with the active site serine hydroxyl to form a covalent,
though reversible, enzyme−inhibitor hemiketal adduct. To
confirm that LEI105 was also able to block conversion of the
natural substrate 1-stearoyl-2-arachidonoyl-sn-glycerol of
DAGL-α to the endocannabinoid 2-AG, we employed our
recently developed real-time, fluorescence-based assay.25 In this
RESULTS AND DISCUSSION
■
Structure-Guided Modeling to Identify LEI105 as
DAGL-α Inhibitor. Previously, we have identified the α-
ketoheterocycle, 1-(oxazolo[4,5-b]pyridin-2-yl)-6-phenylhexan-
1-one (LEI104, Figure 1A), as an inhibitor for DAGL-α. To
improve the potency of LEI104 (Figure 1A) we used a
structure-guided modeling approach. We investigated in detail
C
DOI: 10.1021/jacs.5b04883
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
Figure 3. Cellular activity of LEI105. (a) Representative gel of concentration-dependent inhibition of endogenous DAGL-β in vitro in the Neuro2A
proteome as determined with ABP MB064. (b) Dose response curve of DAGL-β inhibition as determined with competitive ABP MB064 ( SEM, n
= 3). (c) In situ treatment of Neuro2A cells (1 h, 37 °C) dose dependently decreased basal 2-AG levels while keeping anandamide (AEA) levels (d)
constant (mean SEM; n = 4). Statistical analysis: 2-way ANOVA (*** = p < 0.001; ** = p < 0.01; and * = p < 0.05 vs vehicle).
assay, LEI105 inhibited recombinant human DAGL-α with a
pIC50 of 7.9 0.08 (n = 4) (Figure 1E), which is a 40-fold
increase compared to LEI104 (pIC50 6.3 0.1 (n = 4)). The
inhibitory activity of LEI105 was confirmed in a radiometric
assay using 1-[14C]oleoyl-2-arachidonoyl-sn-glycerol (1.0 mCi
mmol−1, 20 μm) as substrate (pIC50 of 6.6; n = 2).8
human DAGL-β using a biochemical assay with para-nitro-
phenylbutyrate as a surrogate substrate (pIC50 of 8.1 0.07, n
= 4; SI). Next, we used the MB064 ABP to profile endogenous
DAGL-β activity in 12 tissues from wild-type and DAGL-β
knockout mice. Spleen tissue was found to display the highest
DAGL-β activity. This activity could be inhibited by LEI105
(Figure 1G). Of note, we could not detect DAGL-β activity
with MB064 in the brain.
Determining Endogenous DAGL Activity Using
MB064 as ABP. To test the activity of LEI105 on
endogenously expressed DAGL-α in mouse membrane
proteome, we used our previously reported ABPP method
with MB064.21 Eleven tissues from wild-type and DAGL-α
knockout mice were screened to obtain a tissue-wide profile of
endogenous DAGL-α activity (SI). DAGL-α activity was found
to be highest in the brain (which is in line with our previous
reported results on a smaller set of tissues).21 LEI105 prevented
DAGL-α labeling in the mouse brain membrane proteome by
MB064 with a pIC50 of 7.5 0.07 (n = 3) (Figure 1C, D).
In view of the high homology between DAGL-α and DAGL-
β, we assessed the activity of LEI105 also on native DAGL-β.
To this end, we tested whether our ABP MB064 was also able
to label DAGL-β. We incubated MB064 with membranes from
mock and hDAGL-β transfected HEK293T cells. MB064
labeled a protein at the expected molecular weight of DAGL-β,
which was not present in the control membranes, or in S443A-
hDAGL-β transfected cells, in which the catalytic serine is
replaced by alanine using site-directed mutagenesis (Figure 1F).
Thus, our tailor-made probe can also detect DAGL-β. Labeling
of hDAGL-β was inhibited by LEI105 with a pIC50 of 7.4
0.07 (n = 3) (SI). We confirmed the activity of LEI105 on
To determine the mode of action (reversible versus
irreversible) of LEI105, we performed an experiment in
which human DAGL-α membranes were preincubated at the
IC80 concentration of LEI105 or the irreversible DAGL
inhibitor KT-109.15 The protein was separated from small
molecule inhibitors via size exclusion column chromatography
and remaining enzyme activity was visualized with MB064. No
recovery of DAGL-α activity was found for KT-109, whereas
MB064 labeled DAGL-α exposed to LEI105 with similar
intensity to DMSO treated DAGL-α (Figure 1H). This
indicates that LEI105 is a reversible inhibitor of DAGL-α.
Determining Proteome-Wide Selectivity of LEI105
Using Comparative and Competitive Chemoproteo-
mics. To investigate the selectivity of LEI105 in the mouse
brain proteome, we have used comparative and competitive
ABPP with a broad-spectrum FP-based probe (TAMRA-FP)
and the tailored DAGL-α ABP MB064. As a reference, we first
tested the widely used β-lactone-based DAGL-inhibitors THL
and OMDM-188, two highly potent nonselective covalent and
irreversible serine hydrolase inhibitors.8,14 In our experimental
setup, both compounds blocked labeling of at least 4 serine
D
DOI: 10.1021/jacs.5b04883
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
Figure 4. DSI is reduced in hippocampal slices treated with 10 μM LEI015. (a) Example of a typical whole-cell voltage-clamp recording of evoked
inhibitory postsynaptic currents (IPSCs) in a CA1 pyramidal neuron under control conditions. IPSCs are evoked every 5 s, but for clarity, IPSCs
shown here are representative traces recorded every 15 s. DSI is induced with a 5-s duration depolarization from −80 to 0 mV and is observed as a
marked and brief reduction of IPSC amplitude following the depolarization. (b) Averaged IPSC amplitude recorded every 5 s in CA1 neurons from
vehicle-treated slices (DMSO) and from slices preincubated at least 30 min and in the continuous presence of LEI 105 (10 μM). The dark lines
represent the single exponential fit of the recovery phase following the depolarizing step. (c) Initial DSI amplitude under control conditions
(DMSO) and in the presence of LEI105 (10 μM). (* p < 0.05 vs vehicle).
hydrolases at concentrations as low as 1 μM, including ABHD6
and ABHD12, enzymes involved in 2-AG metabolism in the
brain (Figure 2A, B).
quantitative chemoproteomics protocol, which was previously
applied to determine the selectivity profile of the irreversible
inhibitor KT-109, using MB108, a biotinylated version of
MB064 and the FP-probe (see SI for synthesis and character-
ization). This methodology allows for a more accurate
quantification avoiding band overlap (as observed with gel-
based assay) and enables screening over a broader range of
specified serine hydrolases. We first validated equal isotopic
labeling and detection of light and medium peptides from
proteins targeted by both ABPs (Figure 2G). Next, we set out
to investigate the selectivity of LEI105 in this chemoproteomic
assay. LEI105 (10 μM, 30 min) did not reduce labeling of the
detected serine hydrolases by more than 20%, except for
DAGL-α (Figure 2F). In addition, we tested LEI105 in
biochemical assays using recombinant human ABHD6 and
MAGL and in a cannabinoid CB1 receptor radioligand
displacement assay. We did not find any significant interaction
(pKi < 5) with these proteins of the endocannabinoid system
(SI). Together, these results indicate that LEI105 is a highly
selective DAGL-inhibitor.
In our previous studies, we identified fatty acid amide
hydrolase (FAAH) as a major off-target for LEI104. In stark
contrast to LEI104, which inhibited FAAH labeling with a
pIC50 of 7.8 0.04 (n = 3) (Figure 2D, SI), LEI105 did not
block FAAH labeling up to a concentration of 10 μM (n = 3)
(Figure 2A, D). Of note, LEI105 did not inhibit labeling of any
other band either, and we conclude that LEI105 is highly
selective at least within the panel of serine hydrolases labeled by
TAMRA-FP and MB064. To explain this remarkable selectivity
of LEI105 over FAAH, we looked in detail into a previously
published cocrystal structure of FAAH with OL-135 (pdb:
2WJ1 and 2WJ2).26 In silico modification of the heterocyclic
part of OL-135, in the presence of the protein, to a LEI105 like
conformer revealed a steric clash of the toluoyl moiety of
LEI105 with the substrate channel of FAAH, thereby possibly
explaining the high selectivity of LEI105 over FAAH compared
to its analog LEI104 (Figure 2C).
Since, the human serine hydrolase family contains approx-
imately 200 family members,27 we wanted to examine the
selectivity of LEI105 in the brain proteome in a broader and
more detailed manner. To this end, we adapted a semi-
LEI105 Reduces 2-AG Levels in Neuro2A Cells. To test
the cellular activity of LEI105, we used Neuro2A cells, which is
a mouse neuroblastoma cell line known to express both DAGL-
α and DAGL-β.28 First, we confirmed the presence of the
E
DOI: 10.1021/jacs.5b04883
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
mRNA transcripts of both DAGLs. Interestingly, we found that
DAGL-β mRNA levels are ∼128-fold higher, compared to
DAGL-α, as determined by qPCR (SI). In line, DAGL-α
protein activity was below the detection limit of our ABPP
assay, whereas we detected a fluorescent band at the expected
molecular weight of DAGL-β using MB064 (Figure 3A; SI for
validation of DAGL-β labeling). LEI105 selectively reduced
DAGL-β labeling in a concentration dependent manner with a
pIC50 of 7.3 0.07 (Figure 3A, B). Of note, no inhibition of
ABHD6 or FAAH was observed (SI). Next, in a targeted
lipidomics experiment, we determined the effect of LEI105 on
the cellular levels of 2-AG and anandamide. A concentration-
dependent reduction of 2-AG was found in Neuro2A cells,
whereas anandamide levels were unaffected (Figure 3C, D). To
test the activity of LEI105 in a human cell line, we used human
PC3 cells. LEI105 reduced 2-AG levels, but no change in
substrate levels of DAGL was detected, which may suggest that
DAG species are rapidly converted into other membrane
constituents, such as phosphatidic acid by DAG-kinases (SI).
Interestingly, arachidonic acid levels were reduced by LEI105
treatment of PC3 cells (SI). This may indicate that downstream
metabolic pathways of 2-AG signaling are also affected and that
the biosynthesis of 2-AG is the rate-limiting step. This is in line
with a previously reported reduction of arachidonic acid levels
in DAGL-α knockout mice and by KT-109 in mouse liver.15,5
LEI105 Reduces Synaptic Transmission in Hippo-
campus. To demonstrate a physiological effect of the
inhibition of 2-AG biosynthesis by LEI105, we focused our
attention on the hippocampal slice preparation in which
depolarization-induced suppression of inhibition (DSI) is
thought to be mediated via 2-AG-induced activation of
presynaptic cannabinoid CB1 receptors.29,30 Experimental
evidence, which supports a role of DAGL-α in DSI, has been
obtained through the use of genetically engineered mice that
constitutively lack DAGL-α. DSI was absent in these DAGL-α
knockout mice, but not in their DAGL-β knockout counter
parts.5,10 Of note, anandamide levels were also reduced in these
DAGL-α knockout mice. Acute DAGL inhibition studies with
chemical tools would, therefore, provide an additional line of
evidence to verify the role of DAGL-α in DSI. Pharmacological
intervention studies with the two highly potent nonselective
DAGL inhibitors, THL and OMDM-188, however, have
challenged the unequivocal role of the DAGLs in DSI. Multiple
groups reported a suppression of DSI by THL and/or OMDM-
188,31−36 whereas other groups did not find any inhibitory
effect of THL and/or OMDM-188.37,38 The discrepancy
between the genetic model and the pharmacological studies
led to the alternative hypothesis that 2-AG is released from
preformed lipid stores instead of the “on demand” production
of 2-AG.39,40
is clear that DSI is not completely blocked in the presence of
LEI105, which suggests that in addition to on demand 2-AG
synthesis, the recruitment of 2-AG pools may contribute to
DSI.40 Alternatively, other endocannabinoids (e.g., ananda-
mide) may be involved. Of note, at 10 μM LEI105 was still
3
selective in our experimental set up and did not displace H−
CP55940 from the CB1 receptor (SI), a result that excludes
direct antagonism of the CB1 receptor by LEI105. Thus, our
data support the hypothesis of a major, but not exclusive, role
of “on demand” production of 2-AG, which is responsible for
the cannabinoid CB1 receptor-mediated synaptic plasticity.33
In conclusion, we have applied comparative and competitive
chemproteomics using a tailor-made ABP (MB064) in
combination with a broad-spectrum probe (TAMRA-FP) to
identify and characterize the reversible DAGL inhibitor LEI105.
This inhibitor was found to be an excellent tool to study the
enzymatic role of DAGL-α in a neuronal setting and provides a
critical counterpart to currently available DAGL knockout mice.
This was exemplified in Neuro2A cells (where LEI105 reduced
2-AG levels, but not AEA levels) and in intact hippocampal
brain slices (where a clear suppression of DSI was observed in
the presence of LEI105). Using this combination of activity-
based probes in a gel-based and chemoproteomic setting many
lipases involved in the regulation of endocannabinoid levels
(e.g., DAGL-α and β, ABHD6, ABHD12, MAGL, and FAAH)
in the brain are targeted with high affinity. This opens the door
to broadly apply this methodology to identify and characterize
not only covalent irreversible inhibitors which can potentially
lead to toxicity and immunogenicity of covalent inhibitor−
protein complexes and selectivity issues using reactive covalent
warheads, but also reversible inhibitors that target DAGL and
other enzymes involved in 2-AG biosynthesis and metabolism.
These reversible inhibitors have the additional intrinsic
advantage to enable a better control of partial inhibition of
their target enzyme. This is especially important with respect to
endocannabinoid signaling in the CNS, since complete
blockade of cannabinoid CB1 receptor signaling in the CNS
may lead to severe side effects.41 Thus, this feature can play an
important factor in the development of clinical candidates to
provide therapeutic solutions for diseases, such as obesity,
related metabolic disorders, and neuroinflammation, in which
excessive 2-AG signaling or its metabolites play an important
factor.
ASSOCIATED CONTENT
■
S
* Supporting Information
Experimental procedures, supporting figures, compound
characterization, and an appendix containing detailed proteo-
mic data. The Supporting Information is available free of charge
on the ACS Publications website at DOI: 10.1021/
jacs.5b04883.
Here, we set out to investigate cannabinoid CB1 receptor-
dependent DSI in CA1 pyramidal neurons in hippocampal
slices with our selective DAGL inhibitor. Preincubation and
continuous application of 10 μM LEI105 had no effect on the
amplitude of the stimulus-evoked inhibitory postsynaptic
currents (IPSC) (271 22 pA) compared to control (276
32 pA). In control conditions, a DSI response could be evoked
which decayed back with an exponential time course (τ = 34 s)
to baseline levels. DSI was quantified as the reduction in IPSC
amplitude averaged over the first 15 s after its induction. DSI
evoked in slices treated with LEI105 was smaller than DSI
induced in control slices (18 7%, n = 15 and 36 4%, n = 18,
respectively; p < 0.05, Student’s t test) (Figure 4). However, it
AUTHOR INFORMATION
■
Corresponding Author
*m.van.der.stelt@chem.leidenuniv.nl
Present Address
#Hsu Laboratories, Department of Chemistry, University of
Virginia, McCormick Road, P.O. Box 400319, Charlottesville,
Virginia 22904−4319, United States.
Notes
The authors declare no competing financial interest.
F
DOI: 10.1021/jacs.5b04883
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
(23) Boger, D. L.; Sato, H.; Lerner, A. E.; Hedrick, M. P.; Fecik, R.
A.; Miyauchi, H.; Wilkie, G. D.; Austin, B. J.; Patricelli, M. P.; Cravatt,
B. F. Proc. Natl. Acad. Sci. U. S. A. 2000, 97, 5044.
(24) Pedicord, D. L.; Flynn, M. J.; Fanslau, C.; Miranda, M.;
Hunihan, L.; Robertson, B. J.; Pearce, B. C.; Yu, X. C.; Westphal, R. S.;
Blat, Y. Biochem. Biophys. Res. Commun. 2011, 411, 809.
(25) van der Wel, T.; Janssen, F. J.; Baggelaar, M. P.; Deng, H.; den
Dulk, H.; Overkleeft, H. S.; van der Stelt, M. J. Lipid Res. 2015, 56,
927.
(26) Mileni, M.; Garfunkle, J.; DeMartino, J. K.; Cravatt, B. F.; Boger,
D. L.; Stevens, R. C. J. Am. Chem. Soc. 2009, 131, 10497.
(27) Bachovchin, D. A.; Ji, T.; Li, W.; Simon, G. M.; Blankman, J. L.;
Adibekian, A.; Hoover, H.; Niessen, S.; Cravatt, B. F. Proc. Natl. Acad.
Sci. U. S. A. 2010, 107, 20941.
(28) Jung, K. M.; Astarita, G.; Thongkham, D.; Piomelli, D. Mol.
Pharmacol. 2011, 80, 60.
(29) Ohno-Shosaku, T.; Maejima, T.; Kano, M. Neuron 2001, 29,
729.
(30) Wilson, R. I.; Nicoll, R. A. Nature 2001, 410, 588.
(31) Hashimotodani, Y.; Ohno-Shosaku, T.; Kano, M. J. Neurosci.:
Off. J. Soc. Neurosci. 2007, 27, 1211.
(32) Hashimotodani, Y.; Ohno-Shosaku, T.; Maejima, T.; Fukami,
K.; Kano, M. Neuropharmacology 2008, 54, 58.
(33) Hashimotodani, Y.; Ohno-Shosaku, T.; Tanimura, A.; Kita, Y.;
Sano, Y.; Shimizu, T.; Di Marzo, V.; Kano, M. J. Physiol. 2013, 591,
4765.
(34) Edwards, D. A.; Zhang, L. H.; Alger, B. E. Proc. Natl. Acad. Sci.
U. S. A. 2008, 105, 8142.
(35) Zhang, L. H.; Wang, M. N.; Bisogno, T.; Di Marzo, V.; Alger, B.
E. PLoS One 2011, 6.
(36) Szabo, B.; Urbanski, M. J.; Bisogno, T.; Di Marzo, V.;
Mendiguren, A.; Baer, W. U.; Freiman, I. J. Physiol.-London 2006, 577,
263.
ACKNOWLEDGMENTS
■
We thank the Dutch Research Council-Chemical Sciences
(ECHO-Grant: 711.014.009; MvdS), Leiden University,
Faculty of Science (“Profiling Programme: Endocannabinoids”;
MvdS, T.H., V.K.), and the Chinese Scholarship Council (DH)
for financial support.
REFERENCES
■
(1) Katona, I.; Freund, T. F. Ann. Rev. Neurosci. 2012, 35, 529.
(2) Devane, W. A.; Hanus, L.; Breuer, A.; Pertwee, R. G.; Stevenson,
L. A.; Griffin, G.; Gibson, D.; Mandelbaum, A.; Etinger, A.;
Mechoulam, R. Science 1992, 258, 1946.
(3) Sugiura, T.; Kondo, S.; Sukagawa, A.; Nakane, S.; Shinoda, A.;
Itoh, K.; Yamashita, A.; Waku, K. Biochem. Biophys. Res. Commun.
1995, 215, 89.
(4) Mechoulam, R.; Ben-Shabat, S.; Hanus, L.; Ligumsky, M.;
Kaminski, N. E.; Schatz, A. R.; Gopher, A.; Almog, S.; Martin, B. R.;
Compton, D. R.; Pertwee, R. G.; Griffin, G.; Bayewitch, M.; Barg, J.;
Vogel, Z. Biochem. Pharmacol. 1995, 50, 83.
(5) Gao, Y.; Vasilyev, D. V.; Goncalves, M. B.; Howell, F. V.; Hobbs,
C.; Reisenberg, M.; Shen, R.; Zhang, M. Y.; Strassle, B. W.; Lu, P.;
Mark, L.; Piesla, M. J.; Deng, K.; Kouranova, E. V.; Ring, R. H.;
Whiteside, G. T.; Bates, B.; Walsh, F. S.; Williams, G.; Pangalos, M. N.;
Samad, T. A.; Doherty, P. J. Neurosci.: Off. J. Soc. Neurosci. 2010, 30,
2017.
(6) Di Marzo, V. Nat. Rev. Drug Discovery 2008, 7, 438.
(7) Blankman, J. L.; Cravatt, B. F. Pharmacol. Rev. 2013, 65, 849.
(8) Bisogno, T.; Howell, F.; Williams, G.; Minassi, A.; Cascio, M. G.;
Ligresti, A.; Matias, I.; Schiano-Moriello, A.; Paul, P.; Williams, E. J.;
Gangadharan, U.; Hobbs, C.; Di Marzo, V.; Doherty, P. J. Cell Biol.
2003, 163, 463.
(37) Edwards, D. A.; Kim, J.; Alger, B. E. J. Neurophysiol. 2006, 95,
(9) Reisenberg, M.; Singh, P. K.; Williams, G.; Doherty, P. Philos.
Trans. R. Soc. London. Ser. B, Biol. Sci. 2012, 367, 3264.
(10) Tanimura, A.; Yamazaki, M.; Hashimotodani, Y.; Uchigashima,
M.; Kawata, S.; Abe, M.; Kita, Y.; Hashimoto, K.; Shimizu, T.;
Watanabe, M.; Sakimura, K.; Kano, M. Neuron 2010, 65, 320.
(11) Appiah, K. K.; Blat, Y.; Robertson, B. J.; Pearce, B. C.; Pedicord,
D. L.; Gentles, R. G.; Yu, X. C.; Mseeh, F.; Nguyen, N.; Swaffield, J.
C.; Harden, D. G.; Westphal, R. S.; Banks, M. N.; O’Connell, J. C. J.
Biomol. Screening 2014, 19, 595.
67.
(38) Min, R.; Testa-Silva, G.; Heistek, T. S.; Canto, C. B.; Lodder, J.
C.; Bisogno, T.; Di Marzo, V.; Brussaard, A. B.; Burnashev, N.;
Mansvelder, H. D. J. Neurosci. 2010, 30, 2710.
(39) Min, R.; Di Marzo, V.; Mansvelder, H. D. Neuroscientist: Rev. J.
Bring. Neurobiol., Neurol., Psych. 2010, 16, 608.
(40) Alger, B. E.; Kim, J. Trends Neurosci. 2011, 34, 304.
(41) Kunos, G.; Tam, J. Br. J. Pharmacol. 2011, 163, 1423.
(12) Janssen, F. J.; Deng, H.; Baggelaar, M. P.; Allara, M.; van der
Wel, T.; den Dulk, H.; Ligresti, A.; van Esbroeck, A. C. M.; McGuire,
R.; Di Marzo, V.; Overkleeft, H. S.; van der Stelt, M. J. Med. Chem.
2014, 57, 6610.
(13) Bisogno, T.; Mahadevan, A.; Coccurello, R.; Chang, J. W.;
Allara, M.; Chen, Y.; Giacovazzo, G.; Lichtman, A.; Cravatt, B.; Moles,
A.; Di Marzo, V. Br. J. Pharmacol. 2013, 169, 784.
(14) Ortar, G.; Bisogno, T.; Ligresti, A.; Morera, E.; Nalli, M.; Di
Marzo, V. J. Med. Chem. 2008, 51, 6970.
(15) Hsu, K. L.; Tsuboi, K.; Adibekian, A.; Pugh, H.; Masuda, K.;
Cravatt, B. F. Nat. Chem. Biol. 2012, 8, 999.
(16) Niphakis, M. J.; Cravatt, B. F. Ann. Rev. Biochem. 2014, 83, 341.
(17) Heal, W. P.; Dang, T. H.; Tate, E. W. Chem. Soc. Rev. 2011, 40,
246.
(18) Liu, Y.; Patricelli, M. P.; Cravatt, B. F. Proc. Natl. Acad. Sci. U. S.
A. 1999, 96, 14694.
(19) Hoover, H. S.; Blankman, J. L.; Niessen, S.; Cravatt, B. F. Bioorg.
Med. Chem. Lett. 2008, 18, 5838.
(20) Yang, P. Y.; Liu, K.; Ngai, M. H.; Lear, M. J.; Wenk, M. R.; Yao,
S. Q. J. Am. Chem. Soc. 2010, 132, 656.
(21) Baggelaar, M. P.; Janssen, F. J.; van Esbroeck, A. C.; den Dulk,
H.; Allara, M.; Hoogendoorn, S.; McGuire, R.; Florea, B. I.;
Meeuwenoord, N.; van den Elst, H.; van der Marel, G. A.; Brouwer,
J.; Di Marzo, V.; Overkleeft, H. S.; van der Stelt, M. Angew. Chem., Int.
Ed. Engl. 2013, 52, 12081.
(22) Fowler, C. J. Fundam. Clin. Pharmacol. 2006, 20, 549.
G
DOI: 10.1021/jacs.5b04883
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX